advances in immunology volume 25

A D V A N C E S I N

I m m u n o l o g y

V O L U M E 25

CONTRIBUTORS T O THIS VOLUME

CLYDE F. BARKER R. E. BILLINGHAM

L. CHESS

DAVID L. GASSER MICHAEL POTTER S, F. SCHLOSSMAN

ANNE-MARIE SCHMITT-VERHULST GENE M. SHEARER

ADVANCES IN

Immunology E D I T E D B Y

HENRY G. KUNKEL FRANK J. DIXON The Rockefeller Univenity

New rod, New Yo& Scrippr Clinic ond Research Foundation

Lo Jollo, California

V O L U M E 25

1 9 7 7

ACADEMIC PRESS New York S a n Francisco London

A Subsidiary of l i a rcour t Brace Jovanovich. Publishers

COPYRIGHT 01977, BY ACADEMIC PRESS, INC.

NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR ALL RIGHTS RESERVED.

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 I

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 61 -1 7057

ISBN 0-12-022425-9

PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS

LIST OF CONTRIBUTORS . . . . . . . . . . . vii PREFACE . . . . . . . . . . . . ix

Immunologically Privileged Sites CLYDE F. BARKER AND R. E. BILLINCHAM

I. Introduction . . . . . . . . . . 1 11. The Anterior Chamber of the Eye . . . . . . . . 3

111. The Cornea . . . . . . . . . . . 8 IV. The Eye Lens . . . . . . . . . . . 11 V. The Brain . . . . . . . . . . . . 11

VI. The Hamster’s Cheek Pouch . . . . . . . . 15 VII. Subcutaneous Tissue . . . . . . . . . . 22

VIII. The Matrix of the Hair Follicle . . . . . . . . 23 IX. The Bone Marrow Space . . . . . . . . . 24 X. The Testicle . . . . . . . . . . . 25

XI. The Prostate . . . . . . . . . . . 27 XII. TheLiver . . . . . . . . . . . . 29

XIII. The Uterus . . . . . . . . . . . . 32 XIV. Artificial Privileged Sites . . . . . . . . . 34

References . . . . . . . . . . . . 49

Major Histocompatibility Complex Restricted Cell-Mediated Immunity GENE M. SHEARER AND ANNE-MARIE SCHMITT-VERHULST

I. Introduction . . . . . . . . . . . 55

Functions . . . . . . . . . . . . 57 111. Fine Specificity ofCytotoxic Effector Cells . . . . . . 64 IV. Immune Response Genes for H-2-Restricted Cytotoxicity. . . . 78 V. Conclusions and Speculation . . . . . . . . . 83

References . . . . . . . . . . . . 87

11. Major Histocompatibility Complex Restriction for Distinct T-Lymphocyte

Current Status of Rat lmmunogeneticr DAVID L. GASSER

I. Introduction . . . . . . . . . . . 93 11. The MajorHistocompatibilityComplex(MHC) . . . . . 98

111. Lymphocyte Alloantigens . . . . . . . . . 107 IV. Other Blood Group and Histocompatibility Polymorphisms . . . 112

VI. Genetics of the Immune Response . . . . . . . 117 VII. Immunoglobulin Genetics . . . . . . . . . 125

V. Evidence for Selection at Histocompatibility Loci . . . . . 115

V

vi CONTENTS

VIII. The Current Linkage Map References . . .

Antigen-Binding Myeloma Proteins of Mice MICHAEL POTTER

I. Introduction . . . . . . . . . . 11. Structures of BALB/c Mouse V Regions . . . . .

111. Groups of Myeloma Proteins That Bind the Same Haptens . IV. Concluding Remarks . . . . . . . .

References . . . . . . . . . .

Human lymphocyte Subpopulations L. CHESS AND s. F. SCHLOSSMAN

I. Introduction . . . . . . . . . . 11. Classical Cell Surface Determinants on Human Lymphocytes .

. .

. .

. .

. .

. .

. .

. .

. .

. . 111. Antigens Distinguishing Human Thymocytes (HTL) and Peripheral Blood

T-cell Subclasses (THJ . . . . . . . . . . IV. Human B-Cell Specific Antigens . . . . . . . . V. Purification of Lymphocyte Subclasses . . . . . . .

VI. The Functional Analysis of Isolated Human Lymphocyte Subpopulations . References . . . . . . . . . . . .

SUBJECT INDEX . . . . . . . . .

131 134

14 1 145 162 20 1 205

213 2 14

2 16 220 222 226 238

243 CONTENTS OF PREVIOUS VOLUMES . . . . . . . . 247

LIST OF CONTRIBUTORS

Numbers in parentheses indicate the pages on which the aiithors' contributions begin.

CLYDE F. BARKER, Department of Surgery, University of Pennsylva- nia School of Medicine, Philadelphia, Pennsylvania ( 1 )

R. E . BILLINGHAM, Department of Cell Biology, The University of Texas Health Science Center at Dallas, Texas (1)

L. CHESS, Dioision of Tumor Immunology, Sidney Farber Cutzcer In- stitute, Hartxird Medical School, Boston, Massachusetts (213)

DAVID L. GASSER, Department of Human Genetics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania (93)

MICHAEL POTTER, Laborutoq of Cell Biology, National Cancer Insti- tute, Bethesda, Maryland (141)

S. F. SCHLOSSMAN, Dioision of Tumor Immunology, Sidney Farber Cci ncer Institute, Haroa rd Medica 1 School, Boston, Mussachusetts (213)

ANNE-MARIE S CHMITT-VEFWULST, Immunology Branch, Nu tional Cancer Institute, Bethesda, Muntland (55)

GENE M. SHEARER, Immunology Branch, National Cancer Institute, Bethesda, Maryland (55)

vii

This Page Intentionally Left Blank

PREFACE

The prediction that the seventies would be the decade of the lymphocyte clearly has been fulfilled. The science of this enigmatic cell no longer can be termed the new immunology, of interest only to a select few, but clearly has begun to permeate wide and diverse branches of biology. Detailed analysis of functional immunology at the cellular level has brought new insight into basic mechanisms of immunity. The application of this knowledge in specialized areas is apparent in much of the content of Volume 25.

The first article, b y Barker and Billingham, on immunologically privileged sites, is especially useful because it brings together a diverse literature from a variety of specialized journals. Many intriguing questions are raised b y the study Sf these privileged sites that are of obvious significance to the ordinary problems of transplantation. The uterus and the protection of the fetus during pregnancy continues to be one of the most challenging problems of immunology and this article is of considerable aid in placing it in proper perspective.

The paper by Shearer and Schmitt-Verhulst on histocompatibility restrictions in cell-mediated immunity is especially timely. One of the most surprising and significant developments stemming from the study of lymphocytes has been the elucidation of the relationship of most T-cell-mediated reactions to the histocompatibility system. This probably has been most thoroughly studied with respect to T- cell-mediated cytotoxicity and the authors have played a major role in this work. Similar conclusions have been reached in the three major systems analyzed: virally infected cells, chemically modified cells, and weak transplantation antigens. Controversy has developed as to whether one receptor or two are involved in the recognition of specific antigens and their associated histocompatibility types. It is an intriguing question which is well discussed in this review.

The paper by Gasser is a very complete review of immunogenetics in the rat. The primary aspect covered in special detail concerns the major histocompatibility antigens and their relation to immune response genes. I t is in this area that the author himself has made important contributions. Immunoglobulin genetics is also a major topic and it is of special utility to have it accompany the histocompatibility section. Such additional topics as blood group immunogenetics are also well covered.

The article by Potter on the antigen-binding myeloma proteins of mice is an extremely thorough presentation of this important branch of

ix

X PREFACE

immunology. Just as the human myeloma proteins played such a significant part in the elucidation of antibody structure, the extensive work now going on with mouse myeloma proteins is providing key answers in V-region genetics. The hapten-binding proteins have been of special utility for these studies as well as for such others as X-ray crystallography for three-dimensional structure. Dr. Potter played an essential role in these developments, in considerable part due to his generous provision of these proteins to other workers.

The last article is by Chess and Schlossman on lymphocyte subpopulations in the human system. This topic is actually quite different from that developed in the mouse, largely because the usual markers obtained by interstrain immunization cannot be obtained similarly in the human. However, other systems which are well discussed in this review are available, as, for example, the T-cell characteristic of sheep cell rosette formation, which would be useful if similarly available in other species. Also included are a number of separation procedures for specific lymphocyte subpopulations, an area in which this laboratory has had wide experience.

H. G. KUNKEL F. J. DIXON

Immunologically Privileged Sites

CLYDE F. BARKER AND R. E. BILLINGHAM

Deparfmennt of Surgery, University of Pennsykania School of Medcim, Philaddphta, Pennsylwnio, and Deportment of Cell Biology, The Univenity of Texas Hwlth Science Center ot Dollas, Texos

I. Introduction ................................................................................. ... 11. The Anterior Chamber of the Eye ....................................................................

111. The Cornea ......................................................................................................... IV. The Eye Lens ......................... V. The Brain .............................................................................................................

VI. T h e Hamster’s Cheek Pouch ......................................... VII. Subcutaneous Tissue ..........................................................................................

VIII. The Matrix of the Hair Follicle ......................................................................... IX. The Bone Marrow Space ................................................................. X. T h e Testicle

XI. T h e Prostate ........................................................................................................

XIII. The Uterus .......................................................................................................... XIV. Artificial Privileged Sites .............

A. Alymphatic Skin Flaps ................................................................................ B. Traumatized Panniculus Camosus Muscle ................................................ C. Muscle ................_. ....... , ................................ D. Skin Islands .................................................................................................. E . Corneal Diffusion Chambers .......................................................................

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

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

. . . . . . . . . ers ........................ ................ .. ...........................

References ..... .. .......... .. .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 3 8

11 11 15 22 23 24 25 27 29 32 34 34 41 42 43 45 46 47 49

I. Introductior.

Genetically alien grafts of a wide variety of both normal and malignant tissues have repeatedly been transplanted to anatomically unnat- ural sites in the body-i.e., transplanted heterotopically-for many different though frequently interrelated purposes. These include: ( 1) determination on an empirical basis of the sitefs) most conducive to the growth and/or prolonged survival of a particular alien tissue in a normal host; (2) confirmation of endocrine function, or demonstration of graft responsiveness to hormones; (3) maintenance of a graft so that it can be visualized and its fate followed directly (for example, after transplantation to the anterior chamber of the eye) or, indirectly, by transillumination (after transplantation to the hamster’s cheek pouch) without recourse to surgery; (4) ease of recovery; (5 ) studies on tissue

1

2 CLYDE F. BARKER AND R. E. BILLINGHAM

interreactions at the morphologic level; and (6) appraisal of the significance of some of the local anatomical and physiological vari- ables for the healing-in of grafts and the elicitation and expression of transplantation immunity, by taking advantage of regional anatomical peculiarities, such as the absence of a lymphatic drainage system in the brain and of blood vessels in the anterior chamber, the cornea, and the lens of the eye.

The pertinent literature, spanning almost a century, is widely scat- tered and confusing because the experiments described were often inadequately controlled and conducted on ill-defined, heterogeneous stocks of animals by investigators who were usually unaware of the basic principles of transplantation immunology. However, the longevity undoubtedly enjoyed by alien grafts in some of the sites employed, as compared with that of similar grafts in other sites, taken in conjunction with the long-recognized and enigmatic success of a high proportion of both experimental and therapeutic orthotopic corneal allografts (Harris and Rathbun, 1972), have given rise to the concept that some of these sites may be “immunologically privileged” or favored-i.e., grafts transplanted to them are in some way partially or fully exempted from the normal rigors imposed by their histoincompatible status.

Contributing to the increased attention that has recently been focused on privileged sites are the following considerations: (1) Privileged sites can be created artificially. (2) It is recognized that better understanding of the modus operandi of privileged sites may lead to improvement in the results obtainable with therapeutic allografts. (3) There is a current search for a host site on which to test the pancreatic islet grafts that can now be prepared for treating diabetics without the need for immunosuppressive drugs (Barker, 1975). (4) Evidence exists that allografts sustained in some privileged sites may weaken on an immunologically specific basis the host’s capacity to harm the alien cells concerned-producing a tolerant or “enhanced” status (see Billingham and Silvers, 1964). (5) Naturally occumng (as well as artificially created) privileged sites afford important “experiments of nature” pertinent to critical evaluation of the theory of immunologic surveillance against neoplastic disease (Bumet, 1970; Schwartz, 1975).

This article presents a critical account of the status of the known or suspected privileged sites in the body and evaluates their significance from both the immunologic and therapeutic viewpoints.

For completeness’ sake, it may be stated, more or less empirically, that a few tissues can survive allotransplantation under conditions in which grafts of nearly all other tissues of similar genetic makeup

IMMUNOLOGICALLY PRIVILEGED SITES 3

would suffer prompt rejection, i.e., there are also immunologically privileged tissues, most notably trophoblast and its malignant derivative, choriocarcinoma, and cartilage (see Beer and Billingham, 1976; Heyner, 1973).

Nude mice deserve a mention here because their basic congenital athymic status renders them “immunologically privileged” hosts that sustain on an indefinite basis both allografts and xenografts from a wide spectrum of vertebrate donors (Manning et al., 1973; Rygaard, 1973).

II. The Anterior Chamber of the Eye

Use of the anterior chamber as a graft site was pioneered by Van Dooremaal(l873) and by Zahn (1884), who observed short-term survival of human malignant tumor tissue and a higher degree of survival with fetal cartilage from both allogeneic and xenogeneic donors in the anterior chambers of rabbits’ eyes. Subsequently, Hegner (1913) reported the short-term growth before regression of mouse tumor tissue grafts in the anterior chambers of rats, mice, guinea pigs, and rabbits, although he had little success with human tumor material in rats’ eyes. By contrast, Smirnova (1937) and Greene and various co-workers (see, e.g., Greene, 1952, 1957; Greene and Arnold, 1945; Greene and Mur- phy, 1945), on the basis of very extensive studies, reported the growth and long-term survival of a variety of human tumors that had acquired the capacity to invade and metastasize in the anterior chamber of rats, guinea pigs, and rabbits. In Greene’s experience, once xenogeneic tumors had become established in the anterior chamber, it was often possible to maintain them by serial transplantation within the eyes of other members ofthe initial host species, and sometimes they could be successfully transferred to the testis. However, neither benign nor malignant tumors at an early stage of their development survived heterotransplantation to the eye. Human melanomas, the slowest growing of the tumors studied, when transplanted to rabbits’ eyes, sometimes persisted apparently unaltered for several months before growth occurred. On the basis of these and other findings, Greene maintained that heterotransplantability could furnish the basis of a biologic test of malignancy. In his hands, unlike normal adult tissue, embryonic tissue and neoplastic brain tissue, which does not metastasize, also survived both xenogeneic and allogeneic transplantation.

However, it is important to note that Greene’s interesting findings on the xenotransplantation and allotransplantation of malignant and embryonic tissues have not been reproducible in the hands of many,


indeed the majority of, other investigators (Morris et al., 1950). In extensive studies on allografts of various normal tissues from fetal and adult donors transplanted to the anterior chambers of mice, Browning (1949) found that, after an initial phase of growth, regression in the fourth week was the invariable fate of the grafts. A possible complica- tion of his experimental design was the use of both eyes in each host. Furthermore, no experiments were conducted to determine the fate of similar grafts in other sites in the body. Dameron (1950, 1951) was much more successful with a variety of fetal endocrine tissue allografts in the eyes of guinea pigs and rats, especially in hosts previously rendered totally deficient in the endocrine tissue concerned. His- tologic evidence of maturation of the endocrine tissue after transplantation was accompanied by functional evidence of its survival. Indeed, one could cite many investigators who have used the anterior chamber with a reasonable degree of success to sustain, for a variety of experimental purposes, endocrine, gonadal, and other tissues from immature and adult allogeneic donors. Markee’s (1932) observation that endometrial tissue allografts in the anterior chambers of guinea pigs, rabbits, and monkeys rapidly acquired a blood supply and underwent estrous cycles for long periods is a familiar classic of reproductive endocrinology. Working with outbred guinea pigs, Woodruff and Woodruff (1950) found that 78% of thyroid tissue allografts in the anterior chambers of thyroidectomized hosts quickly became vascularized, increased in size and survived permanenfly, evok- ing little or no inflammatory reaction. By contrast, only 11% of thyroid allografts transplanted subcutaneously in similar hosts were successful. Of particular interest were the authors’ observations that (1) intraocular allografts gradually lost their initial susceptibility to specific sensitivity elicited in hosts by a subsequent subcutaneous thyroid tissue allograft from the original donor; and (2) when long-established intraocular grafts were recovered and transplanted to a subcutaneous site in the same host, they became vascularized and survived indefinitely in a high proportion of instances. These findings indicated that some kind of adaptation must have taken place, either in the grafts themselves or in their hosts-the Woodruffs favored the latter possibility.

The present authors have been unable to show that, in hamsters and guinea pigs, skin allografts sustained by the cheek pouch milieu or b y the alymphatic skin pedicle flap, respectively, either: (a) weaken the host’s capacity to respond to subsequent orthotopic skin allografts from the original donor strain, or (b) undergo some kind of antigenic attenuation, possibly as a consequence of the loss of passenger leuko-


cytes (see Billingham, 1971; Talmage et al., 1975). However, Warden et al . (1973) have confirmed and extended the observations of the Woodruffs in a study that entailed transplantation of DA strain rat thyroid tissue allografts to Ag-B locus incompatible, thyroidectomized FI strain hosts. They recovered long-established intraocular grafts and compared their survival after subcutaneous implantation into the original hosts and into normal rats syngeneic with the original hosts. A functional criterion of allograft survival-serum thyroxine levels in the thyroidectomized hosts-was used. Only the grafts in the first group survived, indicating that adaptation must have taken place at the level of the host, rather than the graft. These authors suggested that active immunologic enhancement (see Brent and Kilshaw, 1976) of the host was responsible for weakening its reactivity to the subcutaneous allograft. Consistent with these findings is a report that thyroidectomized and parathyroidectomized hamsters bearing thyroid and parathyroid allografts, respectively, of 50-60 days’ standing in their anterior chambers displayed weakened reactivity when tested with orthotopic skin grafts from the same alien donor strain (Weiner, 1965). Evidence will be presented below (see pages 7 and 8) that lends strong support to the concept that the long exemption from rejection that may be enjoyed by intraocular allografts depends upon some kind of induced suppression of the host’s capacity to mount a cellular immune response.

Medawar and Russell (1958) demonstrated that a significant proportion of adrenalectomized mice can subsist for at least several weeks upon allografts of adrenal cortical tissue in the anterior chamber. The fate of skin allografts in the anterior chamber has been studied by several investigators, but, as with other types of grafts in this site, the results are enigmatic because of inconsistency. All investigators are in agreement that skin grafts, like most other tissue grafts in this site, become revascularized within a day or two, usually from the iris. In 1948, in a study of the role of blood and lymph vessels in transplantation immunity, Medawar (1948) reported that skin allografts transplanted to the anterior chambers of specifically immunized rabbits were destroyed if, and only if, they were revascularized. Brown- ing (1949) observed that skin allografts in the anterior chambers of mice were rejected within 30 days, whereas in guinea pigs, according to Connelly (1961), skin allografts grew successfully in a high proportion of subjects, there being no difference in histologic appearance between autografts and allografts of 65 days’ standing. Despite the fact that animals bearing anterior chamber grafts rejected orthotopic skin allografts from the original donor in an accelerated manner, the in-


traocular grafts responsible for the sensitization continued to survive. In the anterior chambers of rabbits’ eyes, skin allografts consistently survived for long periods of time in Raju and Grogan’s (1969) experience, whereas Franklin and Prendergast (1970) found that rejection was always complete by postoperative day 10 as a consequence of a typical allograft reaction. Telling observations substantiating the immunologic basis of this rejection were the indefinite survival of intraocular skin autografts and the prolongation of survival of intraocular skin allografts in rabbits previously exposed to 500 r whole-body ir- radiation.

Recently, Vessella et al. (1974) and Kaplan and various associates (Kaplan and Stevens, 1975; Kaplan et al., 1975a,b) presented the findings of critical systematic analyses of the transplantation immunology of the anterior chamber of the eye, using inbred strains of rats. Their findings go some way toward explaining the highly variable results obtained by other investigators. Although the expectation of survival of intraocular skin allografts significantly exceeded that of orthotopic controls, the immunogenetic disparity between donor and host was an important variable-Ag-B locus compatible grafts living longer than Ag-B locus incompatible grafts. Graft size, or dosage, was another important variable, smaller grafts surviving longer than larger ones. Thyroid tissue allografts enjoyed less protection than skin in the anterior chamber, especially in euthyroid hosts, and various findings sustained the authors’ conclusion that the high degree of susceptibility of thyroid tissue to ischemic necrosis appeared to magnify its immunogenicity. The capacity of a thyroid tissue allograft in one eye to curtail the survival of a concomitantly transplanted skin allograft from the same donor in the opposite eye was indicative of the ability of anterior chamber thyroid grafts to elicit systemic immunity.

Additional evidence that, despite the privilege which the anterior chamber can extend to allografts, this site is certainly not lacking in an immunologically significant afferent connection with the animal’s immune response machinery is afforded by the following observations: (1) Three weeks’ residence, in a viable condition, of Lewis (LE) strain skin in the eyes of Fischer (FI) strain rat hosts sensitizes them in respect of orthotopic Lewis (LE) strain test skin grafts. (2) FI hosts of intraocular DA strain skin develop significant hemagglutinin titers as early as 21 days after transplantation (Kaplan and Stevens, 1975).

Kaplan et al. (1975a) have shown that, when parental strain lymph node cells are inoculated into the anterior chambers of genetically tolerant F1 hybrid rats, graft-versus-host (GVH) reactions develop that are expressed as an anterior uveitis. These reactions resemble GVH


reactions incitable locally in other sites, but with one important exception. Animals that recover from systemic GVH disease usually become refractory to subsequent rechallenge with lymphoid cells from the original donor. This also applies to the local GVH reactivity that underlies the popliteal lymph node assay in the rat (see Grebe and Streilein, 1976). However, (FI x DA)FI hybrid rat hosts that had developed primary intraocular GVH reactions as a consequence of inoculation with 10 x lofi parental strain lymph node cells were able to develop GVH reactions in their local popliteal lymph nodes following subsequent rechallenge with parental strain node cells in their hind foot-pads. Likewise, animals that had given a primary GVH reaction in one eye responded by fulminant GVH reactions when rechal- lenged in either the same or the other eye. The authors tentatively ascribed this disparity in refractoriness, following popliteal and anterior chamber GVH reactions, respectively, to the absence of a lymphatic drainage in the anterior chamber which forces antigen, or antigen-reactive cells introduced into it, to leave exclusively via the intravenous route.

Local GVH reactions were also used by Kaplan and Streilein (1974) further to define the pathway by which antigenic material or lymphocytes can escape from the anterior chamber. Viable suspensions of lymph node cells from FI rats sensitized to DA rat tissue antigens were injected into either the subconjunctival space or the anterior chamber of genetically tolerant (FI x DA)Fl hybrid hosts. Hypertrophy of the cervical nodes and splenomegaly were incited by the former, but not the latter, inocula, substantiating evidence from dye injection studies that the anterior chamber has no lymphatic drainage. Nevertheless, slit-lamp microscopy strongly indicated that intraocular lymphocytic cellular inocula disappear within a few days.

Evidence that these cells enter the host’s blood circulation and can profoundly influence its central machinery of immunologic response was provided by observations that (1) anti-DA strain hemagglutinins appeared within 4 days of inoculating the anterior chambers of normal FI strain rats with (FI x DA)Fl hybrid lymphoid cells; (2) (FI x DA)F, hybrid test skin grafts enjoyed a few days’ prolongation of survival on FI rats that had received an intraocular or an intravenous inoculation of hybrid lymphoid cells 10 days beforehand, whereas similar skin grafts placed on subconjunctivally inoculated hosts underwent summary rejection; and (3) DA rats that had been injected intraocularly with FI strain node cells developed high titers of anti-FI lymphocytotoxic antibodies in addition to hemagglutinins. Further- more, test skin allografts on these animals were rejected in an immune


manner, in contrast to the prolongation of skin graft survival seen in the animals which had received an intraocular injection of F1 hybrid node cells. This, according to Kaplan and Streilein (1974), indicated that the recipient’s immunologic response to the alien lymphocytic inoculum in its anterior chamber depended upon the immunologic reactivity of these cells vis-his the host.

Two well-established key facts about the spleen-first, that, by vir- tue of its size and blood flow, it receives and processes most of the antigenic material administered to an animal by the intravenous route; and, second, that it is the principal source of “enhancing” antibodies-were taken into consideration by Kaplan and Streilein (1974; see also Streilein et al., 1975a) when they postulated that the essential quality of the anterior chamber, and possibly some other immunologically privileged sites apparently devoid of lymphatic drainage, is their ability to allow antigen direct access to the blood stream, bypassing peripheral nodes altogether. The resultant intensive exposure of hosts to antigen via their spleens may then favor the develop- mc nt of unresponsiveness (tolerance and/or enhancement) rather than sensitivity, as a consequence of the synthesis of tissue-protecting enhancing antibodies, the generation of suppressor T lymphocytes (Asherson and Zembala, 1976), or the selective trapping of antigen- reactive cytotoxic lymphocytes within the spleen (Streilein and Read, 1976). Their finding that inoculation of splenectomized FI rats, via the anterior chamber or intravenously, with (FI x DA)F1 hybrid lymphocytes not only failed to prolong the survival of subsequent test skin allografts from the hybrid donors-indeed, it tended to curtail their survival as compared with controls-sustains this interesting concept.

The route by which cells introduced into the anterior chamber gain access to the host’s blood stream has yet to be defined. The obvious possibilities are via the blood vessels supplying the ciliary body and/or the canal of Schlemm. Whether open-ended, epithelial-lined canali- culi run from the anterior chamber into this canal is still equivocal (see Kaplan et aZ., 1975b).

111. The Cornea

The relatively high degree of success that has long been known to attend the use of penetrating corneal allografts to achieve the repair of corneal lesions in the eyes of nonimmunosuppressed patients and the even greater success rate of similar grafts in the normal eyes of experimental animals have long been recognized as apparent exceptions to


the “laws of transplantation,” posing the questions whether corneal tissue is effectively nonantigenic, and whether the cornea as a graft site has unique properties.

That corneal tissue is effectively antigenic has repeatedly been demonstrated. For example, when transplanted heterotopically to vascularized beds, such as subcutaneous pockets or full-thickness skin defects in rabbits, corneal allografts become vascularized and both elicit and succumb to transplantation immunity just as do skin grafts (see Billingham and Boswell, 1953). Furthermore, allografts of corneal epithelium growing on extensive vascular beds prepared by removal of the full thickness of the skin are also rejected like grafts of pure epidermis (Billingham and Boswell, 1953; Khodadoust and Silver- stein, 1966). Khodadoust and Silverstein (1969) have developed an in- genious method for transplanting allogeneic corneal epithelium, stroma, or Descemet’s membrane plus endothelium to the host’s cornea. When vascularization of the recipient bed was procured by positioning the graft eccentrically near the limbus, or by delayed removal of the sutures, each type of graft sensitized the host and underwent rejection. Finally, it has been shown that successful, recently transplanted penetrating corneal allografts in rabbits are vulnerable to transplantation immunity generated b y transplantation of donor skin grafts to the host, but this susceptibility on the part of corneal allografts is usually lost with time (Maumenee, 1951).

The special privilege that orthotopic corneal allografts appear to enjoy cannot be ascribed to surreptitious replacement of alien donor cells by equivalent cells of host origin. Experiments of appropriate design, making use of the sex chromosome marker, karyotype analysis, tritiated thymidine and other labeling techniques, have established unequivocally that in a successful penetrating cornea allograft, there is long-term survival of epithelial cells, keratocytes, and endothelial cells (see Harris and Rathbun, 1972). In corneal grafts that have been stored, the epithelium frequently does become totally detached, in which case it is promptly replaced by centripetal migration of epithelium of host origin, but this does not prejudice the success of the graft. However, the presence and continued viability of the original endothelium appear to be mandatory both for the initial and for the continued success of penetrating corneal allografts because of the great physiologic dependence of the entire cornea on the integrity of this layer. One of its functions is to act as a barrier to the imbibition of fluid from the aqueous humor, as well as a metabolic pump that de- hydrates the stroma. Lamellar, i.e., partial thickness, allografts appear


to enjoy a greater resistance to rejection than penetrating grafts because they are not required to furnish the highly vulnerable endothelial layer, at least Descemet’s membrane and the host endothelium being left intact in the graft bed.

The Zmmunologic Privilege of the Cornea. Corneal tissue is clearly not endowed with any special immunologic privilege, and an explanation for the persistence of donor cells in successful, long-term orthotopic allografts must therefore be sought at the level of the recipient site. Clinical observations and experimental findings have focused attention upon the normal avascularity of the cornea as the most likely basis of its apparently privileged status as a graft site (Polack, 1973). Large grafts or eccentrically placed grafts, whose margins approach the normally vascularized limbal region, are less likely to suc- ceed than smaller grafts, and opacification of a penetrating allograft as a consequence of host sensitization is much more likely when the recipient cornea is vascularized than when it is avascular. Frequently, a typical allograft reaction, leading to the opacification of a graft that has previously been clear for a period of weeks or even years, com- mences near a tuft of vessels or a capillary loop that has grown into or near the graft margin and has affected its endothelium initially. Local steroid therapy can frequently avert this process and reverse endothelial damage (Maumenee, 1973). Rejection of a graft, starting from a single point on its periphery near a capillary, may proceed across the entire graft without being followed by ingrowth of the vessel presumed to be the source of the host cells responsible for mediating the destructive process (Khodadoust and Silverstein, 1969; Silvers tein and Khodadoust, 1973). Local immunologic injury to the endothelium and impairment of its normal physiologic function would result in edema and inflammation-conditions that would amplify the destructive process.

Whether avascularity of the host cornea acts primarily to prevent sensitization of the host, or to prevent rejection once sensitization has been evoked, or both, has yet to be finally resolved. It has been established that skin allografts implanted in shallow, horizontal pockets cut in the corneal stroma survive even in specifically sensitized hosts, SO

long as they remain avascular, and penetrating corneal allografts in experimental animals usually survive specific sensitization of the host if they are avascular, but consistently succumb if they become vascularized (see Billingham and Boswell, 1953). It also appears that neither clear lamellar nor penetrating corneal grafts which have been sustained in continuously avascular beds sensitize their hosts.

The afferent pathway of the immunologic reflex that leads to the rejection of a corneal allograft remains unidentified. The principal


possibilities are that antigenic material from the graft (1) reaches immunologically competent cells via lymphatics known to be present in vascularized corneas and to represent extensions of the abundant lymphatic system of the conjunctiva, which drains into regional lymph nodes of the neck (Collin, 1966, 1970); (2) passes through small veins in the host cornea, ultimately reaching the spleen; or (3) enters the aqueous humor from the endothelium of the graft, gaining access to the venous system and ultimately the spleen via the canal of Schlemm. In diseased eyes in which the corneal and limbal tissues are already infiltrated with lymphocytes, the immune response may be generated locally, i.e., sensitization occurs peripherally (Jones, 1973).

Although the usual efferent pathway of the immune response to a corneal allograft is via the blood vessels in the cornea, the interesting possibility has been raised that, in certain instances, sensitized lymphocytes may be delivered in part by vessels in the anterior uveal tract, crossing the aqueous humor to reach and interact with the endothelium of the graft (see Jones, 1973).

Finally, although it is generally agreed that corneal graft rejection is normally a cell-mediated process, the possible contribution of a humoral response cannot be dismissed entirely since, occasionally, hitherto clear grafts have been observed to undergo opacification in a completely avascular host cornea.

IV. The Eye Lens

Small transplants of benign and malignant human tissues (Morns et al., 1950) and murine tumor cells (Franks, 1957) sometimes survive and grow for many weeks in the avascular lens of a guinea pig’s eye. Success appears to be fitful and dependent upon the lenticular capsule preventing vascularization and cellular infiltration, the transplants being sustained by diffusion-like grafts in cell-impermeable Millipore chambers (see Billingham and Silvers, 1971; see Section XIV,G, on p. 47).

V. The Brain

Ebeling’s report in 1914 that, in mice, allografts of a particular carcinoma were more successful in the brain than in the subcutaneous space initiated studies on this organ as a graft site. Roffo’s (1917) findings with a sarcoma in rats confirmed this observation, and other investigators, notably Shirai (1921) and Murphy and Sturm (1923), corroborated and extended it to xenografts. They demonstrated the successful growth of transplantable mouse sarcomas in the brains of

12 CLYDE F. BARKER AND R. E. BILLINCHAM

rats, guinea pigs, and pigeons, in contradistinction to the failure of such grafts after subcutaneous or intramuscular transplantation. These investigators pointed out that survival of the intracerebral tumors depended upon their not impinging upon the ependymal lining of the ventricle, since this resulted in engorgement of the choroid plexus with lymphocytes, which then proceeded to invade and destroy the tumor. They also carried out an experiment designed to determine whether induced resistance to transplantable adenocarcinoma allografts in mice extended to the brain. The hosts were an undefined “stock” in which about 80% of the subcutaneous grafts and 90% of the intracerebral grafts were successful. Panels of mice were immunized against the tumor antigens by inoculation of defibrinated murine blood. Ten days later these animals were challenged with the tumor, both in the brain and subcutaneously. Whereas 78% ofthem proved to be resistant to their subcutaneous grafts, only in about 10% did the intracerebral grafts fail.

These various findings, taken in conjunction with the brain’s known lack of conventional lymphatic vessels (see Yoffey and Courtice, 1970), have been used by many investigators to relate the apparently privileged status of this organ as a graft site to an interruption in the afferent pathway of the immune response, in the belief that transplantation antigens, passenger leukocytes, and/or peripherally sensitized host lymphocytes cannot easily leave the graft site. How- ever, it is important to recall that blood vessels penetrating the pia mater are surrounded by perivascular spaces that open freely at the brain surface into the subarachnoid space. Fluid from the capillaries, seeping through the tissue, although unable to drain away in lymphatics as in other tissues, may have an effective drainage pathway to the cerebrospinal fluid by passing between the ependymal cells into the ventricles, since these cells are not joined by occluding junctions (see Peters et al., 1976). AIthough a small amount of cerebrospinal fluid does gain access to lymphatic or venous vessels, most of it enters the large intracranial venous sinuses via the arachnoid villi (Bloom and Fawcett, 1975).

Several of Murphy and Sturm’s (1923) observations indicate that the immunologic protection afforded to transplants by the brain is either incomplete or at least delicately balanced: (1) Tumor grafts that did not impinge upon the ventricles, although not subject to cellular reactivity at their margins, were nevertheless surrounded by vessels having col- lars of mononuclear cells, and small vessels in or near the alien tumors were frequently crowded with, and sometimes even blocked by, cells of the lymphoid series. (2) When a small piece of autologous (i.e., host)


spleen was transplanted to a rat’s brain together with the mouse sarcoma graft, the latter rarely grew. However, if the accompanying splenic tissue was of allogeneic origin, it had no inhibitory effect. (3) The success of intracerebral allografts or xenografts was achieved only with “transplantable” tumors (i.e., those previously shown to be easily passageable, and which, presumably, had been maintained in this manner in noninbred, allogeneic hosts). (4) “Spontaneous” mouse tumors failed to grow in the brains of both rats and mice.

Subsequently, Greene and various co-workers (see Greene, 1957) reported the successful transplantability of many different human tumors which had reached the stage of metastasizing, as well as cerebral tumors, in the brains of mice, rabbits, and guinea pigs.

So far as intracerebral tumor allografts are concerned, one of the few analytical studies was that performed by Scheinberg et al. (1964) in mice using a C57BU6 ependymoblastoma and various inbred strains as hosts. Whereas the success rate of intracerebral allografts ranged from 10 to 35% in different strains, all subcutaneous grafts failed. Destruction of the intracerebral grafts was observed as a gradual hyalinization, beginning centrally, with a negligible inflammatory response on the host’s part. Prior sensitization of the hosts by means of skin allografts resulted in the consistent destruction of the intracere- bra1 grafts in association with a marked lymphocytic infiltration. Also pertinent is these authors’ observation that, whereas syngeneic tumor grafts consistently grew in the brains of normal mice, with negligible lymphocytic response, similar grafts in the brains of mice that had been preimmunized by inoculation of glioma-adjuvant mixtures incited massive lymphocytic infiltration and the animals succumbed.

The mammalian brain has also been shown to afford a remarkably hospitable site for allografts and xenografts of a wide variety of tissues and organs from embryonic donors, and such grafts continue to grow and differentiate in this milieu (see Willis, 1935, 1962; Albrink and Greene, 1953; Greene, 1957). In the experience of some investigators, endocrine tissues, too, frequently survive allotransplantation to the brain with retention of function, although again some reports furnish seemingly unequivocal evidence that the privilege extended to grafts by the brain is incomplete and apparently fitful. For example, Siebert (1928) reported that, in guinea pigs, intracerebral thyroid allografts outlived similar grafts transplanted subcutaneously, yet the former, although provoking little or no lymphocytic reaction, were nevertheless inferior histologically to autografts of similar temporal standing. In Woodruffs ( 1960) experience, thyroid allografts that survived in guinea pigs’ brains usually proved to be avascular. Pertinent here is an


earlier finding of Menvin and Hill (1954) that, in mice, small allografts of thyroid or Harderian gland tissue from newborn donors transplanted subcutaneously in such a manner that they failed to become vascularized also survived indefinitely. Such grafts did not sensitize the hosts, although retaining susceptibility to subsequent active immunization.

Using both functional and morphologic criteria of survival, Lance ( 1967) failed to obtain successful intracranial thyroid allografts in thyroidectomized beagle dogs. Evidence of the persistence of viable parenchyma for as long as 8 weeks postoperatively was obtained in a few animals, together with suggestive evidence that these grafts might have functioned evanescently, although they appeared to be under vigorous cellular attack at the time of their removal. It was also found that exposure to intracerebral thyroid allografts led to a state of systemic sensitivity in the host, challenging the notion of the existence of an afferent block in the immunologic reflex. However, at variance with Lance’s findings are previous reports by Athias and Guimarais (1933) of the survival for weeks or months of intracerebral ovarian allografts in castrated male guinea pigs, and by Pomerat et al. (1944) of the long-term survival (based on both functional and histologic criteria) of neonatal adrenal allografts in adult rats’ brains.

Nevertheless, Greene (1957), who claimed such remarkable succes- ses with xenografts of malignant and embryonic tissues in the brains of various species, stated forthrightly that, in his experience, all normal adult tissues failed to survive in either the brains or eyes of alien species hosts. He attributed the inability of many investigators to obtain successful intracerebral xenografts of malignant tissues to vagaries of technique.

Few studies have been made of the fate of intracerebral skin allografts in normal hosts. That skin allografts transplanted to the brains of specifically presensitized rabbits undergo accelerated rejection was established by Medawar in 1948 in a classic but widely misread paper. His observation that the breakdown of such grafts lagged slightly behind that of concomitant control orthotopic skin grafts hinted that an extant state of immunity may be less effectively expressed in the brain than in other tissues. Recently, Raju and Grogan (1977), based on a study of the fate of allografts of Brown Norway (BN) rat skin in the brains of LE rats, concluded that the brain is more effective as a privileged site than the anterior chamber. Since host immunization was demonstrable 20 days after implantation, as evidenced by hemagglutinin formation, yet the inciting intracerebral allografts were capable of surviving for up to 100 days, these authors felt that “blocking” antibodies or suppressor T cells might play an important role.


Clearly, there is a need for additional critical studies, using inbred strains as hosts, better to define the limitations of the brain as a privileged site for alien grafts as well as its modus operandi. It is still impossible to account for the remarkable results obtained with tumor and fetal tissue xenografts by some (but certainly not all) investigators and to reconcile them with the poor results obtained in many instances with allografts, especially from adult donors. Obviously, apart from the magnitude of its histoincompatibility, and its passenger cell content, other properties of the graft must be important, including the possible release of immunosuppressant regulatory agents by malignant tissue grafts (see Udassin et al., 1975).

It is certainly tempting to associate the status of the brain as a graft site with the operation of the ill-defined protective mechanism(s) that normally prevent animals from mounting effective immune responses against the potential autoantigenic components of this organ.

Apart from its lack of a defined lymphatic drainage, two of the properties of cerebral tissue that may underlie its relative hospitality to alien grafts, sometimes even in the face of a state of specific sensitization in the host, are described below. (1) One property is the unique- ness of the cerebral circulation characterized by the blood-brain barrier, which has long been known to prevent the escape of blood- borne dyes into all but a few areas of the brain. We now know that this is due to the presence of occluding junctions between the plasma membranes of adjacent endothelial cells and the paucity of transport vesicles in these cells in brain capillaries (see Bloom and Fawcett, 1975; Peters et al., 1976). It is conceivable either that this anatomic barrier, and/or a layer formed by the end processes of neuroglial cells on the capillary walls, restricts diapedesis and so helps restrain lymphocytic infiltration and also the development of inflammatory responses in the brain, which would surely prejudice the integrity of functional connections between neurons. (2) The other property is that brain tissue may contain chemical ingredients that discourage even normal leukocytic traffic through its parenchyma, substances that might favor allografts by reducing the likelihood of peripheral sensitization.

VI. The Hamster’s Cheek Pouch

Perhaps the most unequivocally effective and most intensively studied, naturally occurring, privileged site is provided by the tissue comprising the wall of the hamster’s cheek pouch (HCP). The large diverticula of the cheek cavities of this species, which are used for food storage, extend caudad beneath the skin of the shoulders and are


lined by a continuation of the mucous membrane of the buccal cavity. Loosely packed areolar tissue knits the outer walls of the pouches to their surrounding structures, allowing them to be everted or pulled inside-out in anesthetized hamsters, facilitating implantation and subsequent observation directly or by transillumination of small grafts beneath the mucosa. It has been recognized for many years that this is a favorable transplant site for normal and malignant tissues (Lutzet al., 1950; Lemon et aZ., 1952; Handler and Shepro, 1968). The submucosa of the HCP is highly vascular and, as early as 24 hours after implantation, proliferating vascular buds of host venous origin, which can be visualized by means of implanted transparent chambers (Greenblatt, 1972), reach the transplanted tissue. Blood flow is initially sluggish, but by the fourth day arteriolization is complete. Once established, allografts usually survive for a surprisingly long time, providing prima facie evidence that the HCP is a privileged site. Even xenografts may escape rejection, though this usually requires “conditioning” of the host with cortisone or other immunosuppressive agents to achieve a high percentage of long-term survivals (Toolan, 1954; Olansky et al., 1959).

Proof of the privileged status of the HCP was established by the report of Billingham et aZ. (1960) that intrapouch skin allografts long outlived similar grafts transplanted orthotopically. A state of systemic sensitivity, however, was found to preclude continued allograft survival on an immunologically specific basis. Skin grafts from the donor strain transplanted elsewhere on recipients currently supporting allogeneic tissue growth in the cheek pouch brought about rejection of the favored graft concomitantly with their own demise. It is aIso possible to bring about destruction of established intrapouch skin allografts by active immunization with intracutaneously injected donor pouch skin epidermal cells or leukocytes or by adoptive transfer of immunity by intraperitoneal injection of host strain lymphoid cells from specifically presensitized donors (Barker and Billingham, 1971). Thus, the hospitality afforded alien tissue in the HCP seems to depend on an interruption of the afferent arc of the immunologic reflex.

The demonstration of Billingham and Silvers (1962) that the HCP “skin” itself can be transplanted in the manner of a free skin graft has allowed further analysis of the modus operandi of the HCP as a privileged site. These workers excised HCPs and transplanted them to full thickness beds prepared on the lateral thoracic walls of syngeneic or autologous hosts. The HCP grafts rapidly became vascularized and, when well established, provided transplant sites that were more readily available than in situ cheek pouches for inspection in nonanes-


thetized animals. Billingham and Silvers then prepared shallow beds at the centers of transplanted HCP isografts to which they transplanted full-thickness ear skin allografts 10-12 mm in diameter. The majority of skin allografts so transplanted across a major histocompatibility bamer (CB + MHA) outlived similar grafts transplanted orthotopically, often surviving upward of 100 days. Variability in survival time was not related to size of the graft, and rejection-if it ensued-was always rapid, suggesting that it resulted from a suddenly provoked, acute response rather than from continued reactivity of low intensity. That the relocated HCP retains the capacity to protect alien skin grafts of allogeneic or xenogeneic origin has been confinned repeatedly (Barker and Billingham, 1971; Kreider et al., 1971).

A likely explanation of the hospitality afforded by in situ or transplanted HCP to grafts of foreign origin is that lymphatic drainage is absent or reduced. If this is so, the effectiveness of lymphatic quarantine would depend on continued prevention of continuity between lymphatic vessels intrinsic to inlaid allografts and lymphatic- rich host skin at the perimeter of the HCP. Barker and Billingham (1971) carried out studies to determine the necessity for preserving a complete, uninterrupted annulus of pouch skin between ear skin allografts and host skin. They found that placement of CB strain skin grafts eccentrically in ectopic pouch skin, so that a small portion of the allograft’s perimeter impinged on surrounding MHA strain host trunk skin, resulted in rejection within 11 days. Furthermore, healthy, centrally located CB skin allografts even of long standing retained their vulnerability to purposeful establishment of lymphatic continuity with their hosts. This continuity was promoted by excising an ellipse of tissue including the edge of the established inlaid allogeneic ear skin and the portion of the CP comprising the annulus on one aspect (see Fig. 1). Donor ear skin and incised host body skin were then sewn together and allowed to heal. This maneuver was invariably followed by rejection. The location of the site of approximation of the inlaid skin allograft with host skin exerted a significant influence on its subsequent longevity. In dorsally located unions, graft survival was longer than with cephalad unions, a finding consistent with the fact that the major lymphatics course principally in the cephalocaudad direction on the side of the hamster’s trunk. The necessity of isolating intrapouch skin allografts from the host tissue beneath them was also investigated. In MHA hosts bearing established HCP isografts, central beds were also cut down through the cheek pouch connective tissue and underlying host panniculus carnosus, exposing the deep fascia of underlying muscle of the chest wall. CB strain skin allografts inlaid into such beds


A~~i~~iPf(~o~~ NO. tests of Survival wr rkin homograft8 timer of central (doyrl Ea~erimant types of excision

Heod - -Ear skin graft

' 4 "Pouch skin graft

2 6 x 100

FIG. 1. Summary of experiments illustrating the dependence of skin allografts inlaid in heterotopic hamster cheek pouch on isolation from contact with host skin.

were protected just as well as those placed in partial thickness beds prepared in pouch isografts, as long as the allograft did not impinge on host body skin at the perimeter of the HCP, a finding consistent with the relatively poor lymphatic supply of muscle (Godart, 1968).

Direct investigations of the lymphatic drainage of the HCP have also indicated this to be absent or greatly reduced. Barker and Billing- ham (1971) found that superficial injections of the colloidal dye, Pat- ent Blue V, into (1) normal intact HCPs; (2) long-established HCP isografts on the chest wall; or (3) established skin isografts inlaid centrally in established cheek pouches, consistently failed to reveal any lymphatic drainage. However, if CB ear skin allografts were inlaid in shallow beds prepared in established MHA body skin isografts, instead of HCP isografts, injections of dye superficially into the inlaid allografts readily demonstrated lymphatic drainage. Furthermore, skin allografts were rejected in this site. Despite the report of Lindenmann and Strauli (1968) that lymphatic vessels can be recognized histologically within HCP tissue, other workers have been unable to demonstrate lymphatic routes from the HCP by means of vital dye or carbon particle injections (Shepro et al., 1963; Witte et al., 1965, 1968). Whether the lymphatic drainage is truly absent or merely greatly reduced remains controversial, since Goldenberg (1970) and Golden-


berg and Steinborn (1970) observed ferritin in the regional nodes by histochemical staining, but not until 24 hours after the material had been injected into the HCP. The amount of ferritin that reached the nodes was much smaller than if the iron compound had been depo- sited in the mucous membrane of the lips, and the time necessary for accumulation of measurable amounts of iron was about 10 times longer than in the case of extrapouch injections.

In view of the probable importance of passenger leukocytes and peripheral sensitization in the rejection process, the ability of lymphocytes of both donor and host origin to reach regional nodes from the HCP may be more critical than whether particulate matter can do so. Barker and Billingham (1971) studied this by introducing suspensions of lymph node cells into transplanted or in situ HCPs. It is known that lymphoid cells from a parental strain donor injected into tissues such as skin, kidney, or uterus of a major histocompatibility locus incompatible F1 hybrid offspring (which is unable to reject them) not only incite local GVH reactions, i.e., normal lymphocyte transfer (NLT) reactions (Ramseier and Billingham, 1966), but by traveling to regional nodes also cause GVH reactions there, reflected in hypertrophy of these organs (Billingham, 1968; Levine, 1968). Lymph node cell inocula from donors presensitized against the host cause more pronounced nodal hypertrophy than do those from normal donors. Lymph node suspensions from MHA anti-CB donors were, therefore, used for injections into (CB x MHA)FI hybrid hamsters. Doses of 20 X 10' MHA anti-CB lymphocytes injected into normal trunk skin or into well-established body skin isografts in (CB x MHA)Fl animals incited both immune lymphocyte transfer (ILT) reactions (Ramseier and Billingham, 1966) at the skin injection site and (by day 7 after inoculation) significant degrees of hypertrophy of the draining, ipsilateral axillary and bra- chial lymph nodes. However, identical cellular inocula in established F, hybrid cheek pouch skin isografts on hybrid hamsters, although inciting local ILT reactions, failed to cause lymph nodal enlargement. The lymphatic drainage of intact HCP was also investigated b y inoculating submucosally into the intact pouches of (CB x MHA)F, hybrid hosts, 20 x lo6 MHA anti-CB cells on the right and 20 x 10' syngeneic (i.e., F1 hybrid) cells on the left. ILT reactions developed at the injection sites of the MHA cells, but there were no significant differences between the weights of right and left cervical nodes, which were excised and weighed 7 days later. These findings are in agreement with those of Chadwick and Blamey (1968), who also presented evidence that the intact HCP prevents lymphoid cell inocula from gaining access to regional lymph nodes.


Another interesting property of HCPs is its partially privileged status. The finding by Billingham and Silvers (1962) that free allografts of HCP skin to the chest wall often enjoyed prolonged survival as compared to orthotopic allografts of trunk or ear skin, is in all probability related closely to the property which endows the HCP with its capacity as a privileged site. Barker and Billingham (1971) later found that HCP allografts were routinely rejected when transplanted across the same histocompatibility barriers, a result they were unable to reconcile completely with the earlier results of Billingham and Silvers in animals derived from the same colony. It remained clear, however, that HCP allografts were favored, since a threshold dose of rabbit antihamster lymphocyte serum (ALS) given only at the time of HCP allograft transplantation and 2 days later (which failed to prolong significantly the survival of body skin allografts) allowed 94% of HCP allografts to remain in good condition for at least 100 days. This unique attribute of the HCP is probably dependent on the presence of the slimy areolar tissue which lies beneath the tough compact connective tissue “dermis” of the pouch skin, since pouch skin epidermis is known to be alloantigenic (Ramseier and Billingham, 1966). That the subdermal areolar tissue constitutes a barrier between the epidermis of the pouch and host tissues, probably impeding lymphatic intercon- nections of graft and host, also follows from the above-outlined experiments with inlaid skin allografts. Billingham and Silvers (1962) noted that sheets of areolar connective tissue dissected from cheek pouches and interposed between skin allografts and their beds conferred significant protection against rejection, even if repeated freezing and thawing had been carried out to devitalize the HCP areolar tissue prior to transplantation. This observation further reinforced the thesis that the HCP areolar tissue plays an important role in the curious immunologic properties of the pouch.

In addition to the importance of the HCP as a means of studying mechanisms of rejection in general, it continues to be of great interest and possibly practical importance as a site of tumor transplantation and storage. Early experiments in which tumors of xenogeneic origin were transplanted to this site focused on the establishment of a few perma- nent tumor lines of human or animal origin by serial passage (Toolan, 1954). More recently, attention has been given to growing individual human tumors removed in the operating room and immediately implanted in HCPs. Possibly an explanation of the interesting findings of Williams et al. (1971) that, of 72 human tumors, 55% of intrapouch uterine cervical carcinomata grew, but only 25-30% of uterine fundal or colorectal cancers and few breast tumors sur-


vived in this milieu would have important biological significance. Of more immediate clinical usefulness is the possibility currently being explored by D. M. Goldenberg (personal communication, 1974) that the growth or regression of human tumors transplanted to the HCPs in response to various treatments of the animals may afford an assay system for the response of residual tumor in the original human host to different therapeutic regimens, i.e., antineoplastic drugs, radiation, hormones, immunotherapy, or combinations of these (see also Hertz, 1967). Also under evaluation is the possibility of using HCP- based human tumors for development of new diagnostic methods, such as detection of tumor-associated antigens (Goldenberg et al., 1975).

The cheek pouch (and other privileged site models) may also be useful in shedding light on the current controversy regarding the importance of tumor surveillance. Prehn’s (1974) contention that the abnormal incidence of malignancies, notably lymphomas, in renal transplant recipients is less likely to be due to effects of immunosuppression on tumor surveillance than to immunostimulation from the allografts is difficult to reconcile with the high incidence of tumors in patients with immunodeficiency diseases and the predilection of lymphosarcomas in immunosuppressed patients for the brain (which, as a privileged site, ought to have impaired surveillance). Although the cheek pouch and cornea are both privileged sites containing populations of actively proliferating epithelial cells, which might be expected to undergo undetected malignant mutation, they are rarely sites of spontaneous tumor formation. Several investigators (Salley, 1954; Silberman and Shklar, 1963) have noted that the application of chemical carcinogens to cheek pouch tissue results in a high incidence of tumor formation, although Hamner (1966) failed to confirm this for autotransplants of pouch skin. Ziegler et al. (1975) found that carcinogens had a profound influence on cheek pouch tissue and on skin grafts transplanted to cheek pouch skin beds on the thoracic wall. The car- cinogenic stimulus was provided either by implantation of a l-cm in diameter paper disk impregnated with 5% 3-methylcholanthrene or b y a weekly injection of 0.4 mg of 3-methylcholanthrene in mineral oil. In the experimental group, the treatment site was the subcutaneous tissue of cheek pouch autografis, and in the control group identical disks were implanted in the subcutaneous tissue of the chest wall. Fibrosar- comas were found in 9 of 20 (45%) of the experimental group after a mean dose of 5.4 mg of 3-methylcholanthrene had been given over 28-161 days. By contrast, only 1 of 10 control animals developed a fibrosarcoma (after 135 days). The carcinogen was also applied


topically either to alymphatic tissue (autotransplanted cheek pouch or ear skin inlaid in cheek pouch previously transplanted to the chest wall) or to tissue with lymphatic circulation (ear skin autotransplanted to the chest wall). Again, the incidence of oncogenesis was much higher in the alymphatic sites (84.6% VS 10%). The significantly greater incidence of tumors induced in the immunologically privileged cheek pouch site as compared with intact skin strongly implies that a break in the afferent limb of the immune mechanism lessens the effectiveness of surveillance.

VII. Subcutaneous Tissue

In a review of endocrine organ transplantation, Krohn (1965) pointed out that endocrine grafts are usually transplanted to fatty, subcutaneous sites, where vascularization is less rapid, ischemia more complete, and lymphatic drainage less complete than the site afforded orthotopic skin grafts. He even raised the possibility that the subcutaneous site might be semiprivileged in an immunologic sense.

Particularly pertinent to this notion is the work of Parrott (1960) on ovarian tissue transplantation in rats of a closed but noninbred colony. This investigator found that intracolony ovarian allografts survived if transplanted to subcutaneous sites, but were rejected in a perfectly typical manner if transplanted orthotopically within the ovarian capsule. Likewise, established subcutaneous ovarian allografts were destroyed if recovered from their subcutaneous site and retransplanted orthotopically on the same animal. On recovery, established subcutaneous grafts were found to be surrounded by a relatively avascular connective tissue capsule, whereas the orthotopic grafts had an abundant blood supply. Consequently, the status of the subcutaneous grafts may have been comparable to that of grafts sequestered in cell- impermeable Millipore chambers. Also pertinent is the fact that the ovary has much more abundant lymphatic drainage than the subcutaneous milieu (Yoffey and Courtice, 1970).

Krohn (1965) also cited some of his own preliminary experiments which indicated that pieces of allogeneic mouse skin, similar in weight to ovarian grafts, may survive for long periods when transplanted into the fatty subcutaneous connective tissue of the chest wall.

Based on experiments making use of the relatively weak Y histocompatibility antigen in C57BL mice, Blair and Moretti (1967) concluded that there is at least a quantitative difference in the immunologic response evoked in females by syngeneic male tissue


implanted either subcutaneously in one case or into the mammary fat pads in the other. Further, they suggested that the mammary fat pad can be considered as protected or privileged, in that the immune response evoked by tissue transplanted into it is limited. These investigators studied the influence of implanting small, 1 mm3, grafts of mammary gland tissue or skin from adult C57BL males subcutaneously or into gland-free sites in the mammary fat pads of 3-week-old syngeneic females, on the fate of challenge grafts of syngeneic male mammary tissue grafts placed in the contralateral fat pad 7 weeks later. After they had been in residence for 8 weeks, these challenge grafts were examined histologically to evaluate the extent of their outgrowth. The results showed very clearly that prior exposure to the Y antigen, via syngeneic mammary tissue or skin inserted into adipose tissue, did not prevent the growth of subsequent male mammary tissue test grafts in the fat pads. Indeed, it actually increased the incidence of successful male mammary grafts as compared with the untreated controls, leading to the suggestion that an enhancing effect was involved. By contrast, pretreatment of the host mice by presentation of the initial graft of male mammary gland or skin via the subcutaneous route curtailed the growth of the challenge grafts as compared with the controls.

Clearly, adipose tissue as a graft site merits further study. Although it has a rich blood supply, there is a paucity of information about its lymphatic drainage status and a basis for suspecting that, in some parts of the body, this tissue may be poorly endowed with lymphatics (Renold and Cahill, 1965; J . M. Yoffey, personal communication).

VIII. The Matrix of the Hair Follicle

With the aid of melanocytes as “markers,” evidence has been obtained that, in the guinea pig, the matrices of the hair follicles behave as partially sequestered sites for cellular allografts. When autologous or syngeneic sheets of black ear skin epidermis, or viable suspensions of cells prepared therefrom, are transplanted to shallow, split- thickness beds prepared in white skin areas of spotted black and white guinea pigs, the grafts heal in rapidly and the defect is resurfaced within a week or less. The regenerated superficial epidermis is produced by commingling of cells of graft origin with native epithelial cells that migrate upward and outward from the ends of the transected pilosebaceous units in the graft bed (Billingham and Silvers, 1963, 1970). The fate of the grafted cell population is reflected by the activity of its included melanocytes. Initially, the regenerated epithelium is a

24 CLYDE F. BARKER AND R. E . BILLINGHAM

dirty gray, but, as the melanocytes settle down and resume functional activity, it begins to darken-usually from about day 10 onward. By 16-20 days postoperatively it is usually intensely black. From about day 14 onward, fine tips of regenerating hair shafts pierce the surface. Most of these are white, but a variable number-about 50-100-on a grafted area of 2 cm2 have become blackened as a consequence of the incorporation of melanocytes (of graft origin) into their matrices. These secondarily blackened hair follicles continue to generate black hairs indefinitely.

When allografts of pure black epidermis were similarly transplanted between spotted guinea pigs of strains Nos. 2 and 13, repigmentation of the superficial epidermis never proceeded very far: “bleaching-out” overtook the darkening superficial epidermis, accompanied by erythema and mononuclear cell infiltration of the underlying dermis by about day 13 (Barker and Billingham, 1972). These changes are indicative of the survival end point of the alien cells, including melanocytes, in the superficial epidermis. Despite this, an appreciable number of black hairs subsequently emerged from the grafted area and retained their pigmentation for 30 to >250 days. Al- though in a few animals synthesis of pigment by the surviving alien melanocytes in the follicles terminated abruptly, as evidenced by hairs that were white proximally and black distally, in most animals the number of black hairs and the intensity of their pigmentation decreased gradually. In addition, the diameters of the hair shafts decreased as they became less pigmented. These findings suggest that immunogenetically alien melanocytes (and presumably keratinocytes, too) that become established in the hair bulbs are significantly less vulnerable to transplantation immunity than similar cells in the superficial epidermis. Probable factors contributing to the protection afforded the melanocytes are the avascular status of the epidermis and the interposition of the acellular, two-layered vitreous or glassy membrane between the outer root sheath and the connective tissue sheath of the follicle (Bloom and Fawcett, 1975; Montagna, 1962).

IX. The Bone Marrow Space

In view of the absence of evidence that lymphatic vessels occur in bone or bone marrow (see Yoffey and Courtice, 1970), it is surprising that so little attention has been paid to the marrow space as a graft site. The work of Iwao et al. (1935), and more recently of Fonkalsrud (1968), indicates that autografts of skin, thyroid, and pituitary glands survive in the marrow spaces of dogs and rabbits, although the grafts

IMMUNOLOGICALLY PRIVILEGED SITES 2 5

become less well vascularized and their epithelia proliferate to a lesser extent than after implantation into muscle. As far as they go, the findings hint that allografts may fare better in the marrow space than in muscle.

X . The Testicle

Along with the brain and anterior chamber of the eye, the testicle has long enjoyed a reputation as a peculiarly hospitable site for allografts. Greene (1940) successfully used the testicle in mice, rats, and hamsters as a transplant site for Brown-Pearce tumors of rabbit origin. However, he observed that only 50-100% of recipients grew tumors compared with 100% when the same tumor was transplanted to the anterior chamber of the eye. After periods varying from several days to more than a month, the intratesticular tumor grafts regressed, and thereafter resistance to reinoculation of this tumor was noted. Lympha- tic extension and metastasis did not occur. Aron et al. (1957) also reported successful allotransplantation of pituitary, thyroid, and kidney tissue to the guinea pig testicle.

More recent confirmation of the effectiveness of the testicle as a privileged site comes from Dib-Kuri et al. (1975), who found that in parathyroidectomized outbred rats parathyroid allografts uniformly sustained normal serum calcium levels for longer than 3 months if placed in the testicle. By contrast, recipients of intramuscular allogeneic parathyroid grafts failed to achieve even transient normocalcemia, presumably because of very early rejection. These workers also claimed that even parathyroid xenografts of guinea pig and rabbit origin remained viable and functional for upward of 25 days if transplanted to the rat testicle.

In an attempt to define accurately the role histocompatibility plays in intratesticular transplantation, Naji and Barker (1976) studied the fate of parathyroid allografts in parathyroidectomized inbred rats. In the testicles of Fischer (FI) hosts the survival time, as evidenced by normocalcemia, of both Ag-B locus compatible Lewis (LE) and incompatible ACI strain parathyroids exceeded by many weeks that observed for allogeneic parathyroids inserted in subdermal pockets (12-16 days). However, histocompatibility was a significant factor, since only 20% of the ACI grafts survived for 150 days while 62% of the recipients of the more compatible LE parathyroid glands remained normoglycemic at this time. Human parathyroid xenografts appeared to be rejected so rapidly that normocalcemia was never seen following intratesticular transplantation in rats.


That the testicle is a frequently effective but incompletely privileged site for other tissue allografts is also borne out by the findings of Whitmore and Gittes (1975) that LE skin transplanted to testes of FI rats and biopsied after 26 days showed histologic evidence of reiection in 9 of 29 animals.

The variability in graft survival times observed even when the genetic relationship of donor and host is constant is difficult to reconcile with the most frequently cited explanation for the protection afforded allografts by the testes. According to this, the lymphatic drainage of the testicle is deficient because of an unusually long lymphatic pathway between this organ and the nearest draining lymph node at the level of the kidney. This explanation is also challenged by Barker and Billingham’s (1973a) finding that skin allografts transplanted orthotopically to beds prepared near the tips of adult rats’ tails were rejected with normal promptitude, despite an afferent lymphatic pathway to the nearest draining node of 19 cm. A more plausible theory, and one that could at least partially account for inconsistencies, would be failure in some individual animals of the testicular lymphatic effluent to transverse any lymph node. This situation was noted by Engeset (1959) in 19 of 65 rats, in which lymphatics leaving the testicle, as visualized by lymphangiography, passed directly into the thoracic duct without traversing a node. In contrast, McCullough (1975) and Tihey (1971), who also studied the lymphatic drainage of the testes in rats, both found that afferent lymphatics uniformly led to at least one subdiaphragmatic lymph node. McCullough (197.5) notes, however, that in dogs and humans these subdiaphragmatic nodes are sometimes b v o a s d

Further studies appear to be in order regarding the immunologically privileged status of the testicle and whether the lymphatic circulation of the organ is responsible.

Physiologic studies have indicated the existence of a blood-testis barrier around the seminiferous tubules, capable of excluding from the lumens of these tubules many substances normally present in blood and lymph, including antibodies, and at the same time, preventing the escape of autoantigenic material associated with the maturing components of the germinal epithelium, i.e., testicular antigens. Ultrastruc- tural studies (Dym and Fawcett, 1970; Neaves, 1977) have revealed that this barrier has two components: (1) the contractile myoid cells united by occluding junctions that surround the seminiferous tubules; (2) a continuous barrier formed by the contiguous Sertoli cells with their tight junctions. These Sertoli cell junctions constitute an adlumi- nal compartment in the seminiferous tubules that “quarantines” the


spermatocytes and their differentiation products. However, it seems unlikely that this anatomically defined barrier could play any significant role in mitigating host reactivity to intratesticular allografts.

XI. The Prostate

For many years, pathologists have been impressed by the unusually high incidence of prostatic carcinoma, a lesion found in 41% of all human males dying in their seventh decade and in 57% of those in their eighth (Scott et uZ., 1969). Fortunately, these tumors, although histologically malignant and fully capable of metastasizing with fatal outcome, generally behave in a benign fashion, apparently remaining confined to the prostate for many years until their incidental discovery at autopsy (Franks, 1956).

Another histological curiosity of the prostate is its lack of lymphatics. An exhaustive search by Rodin et ul. (1967) failed to reveal lymphatic vessels, and Gittes and McCullough (1974) could find none of the iodinated emulsified oil they injected into the prostates of rats or humans in nearby lymph nodes, although the same technique outlined lymphatic drainage of the testicle. McCullough (1975) also failed to demonstrate lymphatic drainage from the prostate radiographically, using the most sensitive modern technique with hyperemulsified radiopaque medium.

Smith (1966) was unable to demonstrate lymphatics in canine and human prostates by injections of Patent Blue V dye. He suggested that, although lymph nodal involvement is common in patients with prostatic carcinoma, this probably does not occur until the tumor has spread beyond the fascia1 envelope. Early extension of prostatic tumors is often by way of the perineural spaces, which do not appear to be lymphatic vessels, although they have sometimes been misinter- preted as such.

Gittes and McCullough (1974) made the interesting suggestion that a relationship exists between the lymphatic status of the prostate and the high incidence of tumors. If immunologic surveillance for tumors is dependent on intact lymphatic circulation to regional nodes, as is rejection of skin allografts (Barker and Billingham, 1968), an organ without lymphatics might be expected to lack efficient surveillance, resulting in a high incidence of de novo tumors springing from mutant cells within itself. A similar breakdown in surveillance has been proposed as an explanation of the predilection of tumors in immunosuppressed kidney transplant recipients for the brain (Penn, 1972), another organ lacking conventional lymphatic drainage (see Section V).


In the brain, these tumors are predominantly lymphosarcomas, which are of mesenchymal rather than epithelial origin, despite the rarity of this type of tumor in the central nervous system of nonimmunosuppressed individuals. Another factor of possible significance in transplant patients, thought by some (Prehn and Lapp&, 1971; Schwartz, 1974) to be of more importance than immunosuppression, is that of immunostimulation by the continuous presence of the allograft itself. Because prostatic carcinoma is rarely found in patients as young as most transplant recipients, it may be of little significance that no increase in incidence of prostatic tumors has been reported in this group.

Common to both prostatic carcinoma and the brain tumors in immunosuppressed patients is the possibility that blocking antibody could be a factor encouraging tumor growth. Since tumor antigen in both instances would be presented to the host predominantly by the intravenous route (the lymphatic one being absent), the formation of enhancing antibody rather than the stimulation of cellular immunity would not be a surprising outcome.

The hypothesis set forth with regard to deficient immunologic surveillance for the prostate would be more plausible if experimental evidence were forthcoming that the gland is also an immunologically privileged site for transplantation of allogeneic tissue. In investigating this possibility, Gittes and his co-workers have encountered conflicting evidence. With regard to growth of tumor allografts, Gittes and McCullough (1974) found the prosta:e to be an unusually hospitable transplant site. Cells of methylcholanthrene-induced squamous tumors of LE rat origin rapidly grew to palpable size when suspensions of them were injected into the prostates of Sprague-Dawley rats, although no growth was apparent if the injections were made intramuscularly. However, tumor allografts growing in the prostate were eventually rejected (after 3 to 4 weeks), and subsequent inocula of the tumor cells failed to grow.

In another experiment, Whitmore and Gittes (1975) transplanted nonmalignant tissue (tiny 2 x 2 mm skin allografts) from LE donors to prostate, muscle, anterior chamber, testes, or orthotopic sites in FI rats. The grafts were removed 26 days later and examined histologically. The intraprostatic grafts (25% of which showed no evidence of rejectiod fared better than those transplanted to muscle but not as well as those in the testes or the anterior chamber. These investigators also studied delayed hypersensitivity responses to intradermal injections of sheep red blood cells in rats which had previously been sensitized by inoculation of sheep red cells into the prostate, anterior


chamber, testis, skin, or muscle. Animals which had been immunized by prostatic or anterior chamber inocula proved less reactive than those exposed to the antigen via testicular, intradermal, or intramuscular injections. Although the experiments with skin allografts were interpreted by the authors as disproving the possibility that the prostate might be a privileged site, further experimental scrutiny of the organ is clearly indicated before final judgment can be pronounced on its possible privileged status. Since it is a relatively “silent” organ, it might prove to be a site for endocrine allografts superior to the brain or anterior chamber of the eye. As compared with other candidates for clinical transplant sites, it would be safer and more accessible than the brain and, although less accessible than the anterior chamber or testicle, more acceptable than the eye on the basis of risk and than the testicle on emotional grounds.

XII. The liver

Several intriguing and perhaps unrelated phenomena suggest that the liver possesses peculiar immunologic characteristics relevant to the subject matter of this review. Chase pointed out as early as 1946 that haptens such as dinitrochlorobenzene (DNCB) or picryl chloride, administered to guinea pigs by mouth and presumably absorbed into the hepatic portal vein, could inhibit the development of delayed hypersensitivity rather than elicit it, as would be expected from topical application of these agents to the skin. More recently, Cantor and Dumont (1967) found that feeding DNCB to dogs suppressed their ability to form circulating antibody to this hapten when it was subsequently injected subcutaneously. If access of the hapten to the liver was bypassed by carrying out a portacaval shunt, prior to feeding DNCB, the ability to form antibody remained intact.

Indirect evidence that this phenomenon may also occur in humans is the observation (Triger et al., 1972) that patients with hepatic disease, in whom spontaneous portosystemic shunts may be formed as a result of the liver damage, frequently show elevated antibody titers to a variety of bacterial and other antigens encountered via the gastrointestinal tract. A similar increase in antibody titer to Escherichia coli has been reported after portacaval shunting (Triger, 1976).

That these observations were not the result of denaturation of the various ingested antigens by gastric acid or digestive enzymes is suggested by Battisto and Miller’s (1962) demonstration that the injection of hapten-protein conjugates directly into the portal vein induced immunologic paralysis, whereas the same antigens presented by other


routes were highly immunogenic. Additional evidence that the liver plays a major role in the humoral response to antigens is the demonstration by Triger et al. (1973) that particulate antigen injected into the portal vein of rats results in a lower level of antibody response than if injected into the inferior vena cava.

Several lines of evidence suggest that the liver could play a regulatory role in the response to allografts as well as in delayed hypersensitivity and humoral immunity. Mandel et al. (1965), noting reports of Manson et al. (1963; see also Manson and Palm, 1968) and Al-Askari et al. (1964) that cell-free extracts prepared from allogeneic livers, unlike extracts of lymphoid and other tissues, were incapable of sensitizing mice to subsequent skin grafts, investigated the possibility that the extraction process released from hepatocytes a material (perhaps an enzyme) which could destroy the sensitizing capacity of transplantation antigens. They found that alloantigenic material extracted from splenic tissue, if incubated with homogenized liver, lost its ability to sensitize mice against subsequent skin allografts of the same alien genetic origin. The possibility that the liver as an intact organ might possess the same capacity for inactivation or destruction of tissue transplantation antigens was investigated by Barker and Corriere (1967), who transplanted dog kidney allografts in such a way that the venous effluent was directed to the liver via a portal vein anastamosis. No prolongation of primary allograft survival was found but, using the same technique, Fukuda et al. (1969) noted abrogation of the usual accelerated rejection of second kidney allografts from the same canine donors.

Sakai (1970), utilizing microsurgical techniques in the rat, found that kidney allografts, the veins of which were anastornosed to the recipient portal system, survived significantly longer than those with venous drainage into the inferior vena cava (9-19 days for portal vein vs. 9-11 days for vena caval transplants). A similar finding was subsequently reported for rat heart allografts transplanted across an Ag-B locus barrier by Boeckx et al. (1975), who noted a brief prolongation (10.9 days) for portal vein hearts compared with 7.0 days when the vena cava was used.

In the case of whole-organ allografts transplanted by the above method, the failure to demonstrate more impressive retardation of rejection could be accounted for b y escape of antigenic material, including cells, from the allograft into the host by other than the venous route. Lymphatic communications from transplanted kidneys to host regional lymphatics are known to be established within a few days, and these lymphatic channels in themselves may constitute an effi-


cient route for sensitization (Pederson and Morris, 1970). To investigate posbible properties of the liver as a privileged site for allografts transplanted in such a way that antigen was unlikely to reach the host by other routes, C. F. Barker and R. E. Billingham (unpublished data, 1970) performed skin grafts in rats, using a bed cut in the highly vascular surface of the organ. These grafts healed in promptly and, although isografts flourished indefinitely, allografts were rejected with normal promptitude.

Since skin is highly exacting in its immunologic requirements for success, transplantation of endocrine organs to the liver might be expected to meet with greater success, because they may constitute partially privileged tissue. Barker et al. (1975) found that allogeneic islets of Langerhans, isolated from the pancreas by collagenase digestion and transplanted to the liver by embolization via portal vein injections, sometimes survived longer than those implanted into the peritoneal cavity. In ACI strain rats with chemically induced diabetes, portal vein injections of Ag-B locus compatible DA strain islets were followed by normoglycemic intervals of up to 80 days [the median survival time (MST) was 30.5 days]. Intraperitoneally administered islets had an MST of only 11.5 days. However, use of the portal vein of the liver did not prolong the survival of implanted Ag-B incompatible islet tissue, which was rejected very rapidly (MST <7 days) no matter what the transplant site.

Pfeffermann et al. (1976) also reported prolonged survival of intrahepatic allografts of endocrine tissue (parathyroid) under circumstances of minor histoincompatibility (FI + LE rats). When allografts were implanted through incisions several millimeters deep in the parenchyma of the liver, mean survival was 35.4 days as compared with 15.7 days if transplanted to intramuscular sites. However, when donor and host differed at the Ag-B locus, intrahepatic allogeneic parathyroids survived no longer than intramuscular ones.

Despite the evidence that the intrahepatic site provides some protection for allografts, the mechanism responsible for this is unclear. Certainly rich vascularity, both blood and lymphatic, distinguishes the liver from classic privileged sites, such as the anterior chamber of the eye and the hamster cheek pouch. Differences in antigen processing, or hepatic enzyme capable of altering antigen, are possible expla- nations for the anomalous survival of intrahepatic allografts. An interesting question is whether the above-described phenomena are related in any way to the unique properties attributable to the liver when this organ is itself used as an allograft. In man, liver allografts rarely fail from rejection (Roddy et al., 1976) and are remarkably resis-


tant to hyperacute rejection (Starzl et al., 1974). In the pig, liver allografts are also (Calne et al., 1969) unlikely to be rejected even in nonimmunosuppressed subjects. In fact, established porcine liver allografts have an immunosuppressive effect with regard to subsequent allografts (from the same donor) of kidney, heart, or skin (Calne et al., 1969; Millard et al., 1971).

XIII. The Uterus

Pregnancy represents a natural, complex, intimate, allograft-host parabiotic relationship that challenges the transplantation im- munologist to explain the unqualified success, or survival to term, of fetuses qua allografts and their resistance to virtually all attempts to prejudice their survival by applying the principles of transplantation immunology. It is now firmly established that the fetus’ exemption from rejection is largely, if not exclusively, dependent upon unique immunologic properties of the trophoblast-primarily upon the ineffective expression or nonexpression of transplantation antigens by the latter (Goodfellow et al., 1976; Faulk et al., 1977). For many years, however, the notion has been entertained that special immunologic dispensations might apply to the uterus, at least as a site for implantation of immunogenetically alien conceptuses (see Beer and Billing- ham, 1971, 19761.

In 1962, Schlesinger tested the possible privileged status of the uterus by surgically implanting small tumor grafts into the uterine horns of rats and mice. Syngeneic tumors grew progressively, whereas allogeneic tumors survived for only a relatively short time, in normal hosts and underwent rapid rejection in the uteri of specifically sensitized hosts, irrespective of the endocrinologic status of the latter, i.e., whether they were pregnant, pseudopregnant, or nonpregnant. Unfor- tunately, since the design of this experiment did not exclude the possibility that the test allografts might have invaded host tissues beyond the physiologic limits of the uterine endometrium, the findings did not deny a privileged status to the uterus. Subsequently, Poppa et al. (1964) showed that allografts of a normal, noninvasive tissue, parathyroid, transplanted to the uterus of pseudopregnant and nonpseudopregnant, parathyroidectomized rats were consistently rejected within 20 days. This observation suggested that transplantation immunity can be elicited as well as expressed within the uterus. How- ever, it did not refute the hypothesis that at least local implantation sites in the gravid herus might have a privileged status, since the parathyroid grafb studied by Poppa et al. (1964) failed to evoke decidual reactions.


Before embarking upon a reevaluation of the rat’s uterus as a site for allografts, Beer and Billingham (1970) first determined the conditions under which it would accept “free” grafts of genetically compatible skin or monodisperse epidermal cells, as “placebo” conceptuses, on its undamaged endometrial surface. They found that small grafts, inserted through longitudinal incisions near the uterotubal junction, consistently implanted on the untraumatized endometrial surface, became well vascularized within a few days and survived indefinitely, provided that the endometrium was in the proliferative phase and that the recipient had not ovulated. Contrary to expectation, no decidual response was elicited in the uterus. In the absence of an estrogen surge at the time of transplantation, only rarely did either skin grafts or epidermal cell suspensions become established of their own accord on the endometrium. When skin allografts from Ag-B locus compatible LE strain or Ag-B locus incompatible DA strain donors were placed in the uteri of estrogen-primed virgin FI strain female hosts, they healed in and succumbed to typical allograft reactions, surviving no longer than they would have if transplanted orthotopically on control hosts. However, when the skin allografts were placed in the uteri of females which were either in the preimplantation stage of pregnancy by males of their own strain, or had been rendered pseudopregnant by prior cervicomechanical stimulation, typical decidual responses developed beneath them, and their survival was prolonged by about 4-5 days (Beer and Billingham, 1974). The observation that neither pregnancy nor pseudopregnancy prolonged the survival of skin allografts transplanted orthotopically dismissed the possibility that the extension of survival enjoyed by the intrauterine grafts was due to the action of hormones released into the circulation, some of which might have immunosuppressive properties (see Beer and Billingham, 1976). Skin allografts healed in poorly and were rejected in a summary manner when placed in the uteri of specifically presensitized, pregnant hosts.

These findings sustain the conclusion that decidual tissue subjacent to the alien intrauterine grafts in nonimmune hosts impairs the afferent pathway of the immunologic reflex. It may act by partial blockage of the afferent lymphatic vessels which, according to McLean and Scothorne (1970), are located at the endometriaVmyometria1 junction. This could prevent antigenic material from reaching the proximal seat of immunologic response in the host-the regional para-aortic lymph nodes-or it could interrupt the passage of peripherally sensitized ma- ternal lymphocytes to these organs. Another possibility is that it facili- tates presentation of fetal antigenic material by the intravenous route, which is known to favor the development of enhancing or blocking antibodies (see Section 11).


Simmons and Russell’s (1966) demonstration in mice of the vulnerability of allogeneic blastocysts placed in decidual-deficient ectopic sites, such as beneath the renal capsule, as compared to their relative invulnerability when transplanted to uteri of specifically presensitized mice, also suggests that decidual tissue affords the early conceptus some degree of immunologic protection.

In a careful study of the intrauterine lymphatics in the rabbit by retrograde dye injection, McLean and Scothorne (1970) found that, whereas these vessels are abundant beneath the uterine peritoneum, between the muscle layers, and at the endometrial/myometrial junction, they are not demonstrable in the endometrium. Subsequently, these workers carried out a comparative study of the fates of skin autografts and allografts implanted at various anatomical levels in the uterine wall, subcutaneously and orthotopically in rabbits. Suggestive evidence of very modest prolongation of survival and delayed infiltration by mononuclear cells of the allografts in the endometrial and myometrial sites led them to conclude that in this species the endometrium is a privileged site (McLean and Scothorne, 1972).

XIV. Artificial Privileged Sites

A. ALYMPHATIC SI<IN FLAPS

Although the absence of lymphatic circulation from several classic privileged sites (e.g., hamster’s cheek pouch, brain) has been cited as a likely explanation of the protection afforded allogeneic implants, evidence for this has been indirect until recently. Landsteiner and Chase (1939) and Frey and Wenk (1957) demonstrated that topical application of chemical allergens elicited delayed hypersensitivity in guinea pigs only if the areas of skin painted had intact lymphatic drainage. The common assumption that lymphatic channels draining skin or tumor allograft beds carry antigenic material to regional lymph nodes was based only on cytological and adoptive transfer studies indicating that these nodes were the primary seat of the immune response (Bil- lingham et al., 1954; Mitchison, 1954). Despite this, excision of the regional lymph nodes near a skin allograft was found to prolong its survival by only a few days in mice (Billingham et al., 1954), dogs (Chrdtein e t al., 1967), guinea pigs (Barker and Billingham, 1968), and man (Stark et al., 1960).

Lambert et al. (1965) reported that dividing the lymphatic vessels in rabbits’ ears proximal to the site of a skin allograft delayed rejection


for 4 days-about the time required for constitution of alternative lymphatic pathways. Since none of these experimental approaches resulted in the loss of lymphatic circulation for more than a few days, the results did not discredit the hypothesis that a permanently lymphatic-free area would constitute an immunologically privileged site. Direct evidence supporting this hypothesis was finally provided by the experiments now to be described (Barker and Billingham, 1968).

In guinea pigs, a successful privileged site (Fig. 2) was created artificially by surgically separating from the flank a 3 cm in diameter disk of skin and its underlying panniculus carnosus muscle. Carefully preserved in the dissection was a single neurovascular bundle which served as an “umbilical cord” to nourish the otherwise completely detached skin flap. Patent Blue V dye injected into such flaps outlined a rich lymphatic network within their confines, but no lymphatic vessels were found accompanying the umbilical cord. The operative defect in flank skin was sutured around the vascular pedicle and the flap

FIG. 2. A freshly prepared isolated skin flap with its neurovascular umbilical cord, which contains no lymphatics. After closure of the wound, the plastic capsule housing the flap is glued to the underlying skin. Inlaid skin allografts enjoy prolonged sulvival in this artificial privileged site.


was placed in a plastic capsule, with a hole to admit the umbilical cord. The protective capsule was then glued to the underlying skin. Shallow beds were cut in the flap skin, and 1 cm in diameter grafts of thin ear skin were fitted in place. These isolated skin flaps usually remained viable and healthy, as evidenced, for example, by continued hair growth, for 20-30 days (occasionally for 100 days) before their ultimate and usually sudden demise from ischemic necrosis, which was caused by twisting or other trauma to the umbilical cord. Intraflap skin allografts exchanged between guinea pigs of syngeneic strains Nos. 2 and 13 healed in promptly and in 22 of25 instances survived for as long as the flaps remained viable (20-100 days, mean 24 days). The survival time of control orthotopic skin allografts exchanged between these strains (which differ at 7 or 8 histocompatibility loci including the MHC) was 8-10 days. Subsequent skin grafting tests demonstrated that intraflap allografts did not sensitize their hosts. However, they were promptly rejected if recipients had been specifically presensitized by orthotopic skin allografts. Furthermore, active or adoptive immunization of hosts bearing established, healthy intraflap allografts brought about their prompt destruction, leaving no doubt that the efferent arc of the immunological reflex was intact.

Several studies designed to determine the consequences of maintaining a channel for lymphatic continuity from an otherwise isolated skin flap provided further evidence that a privileged site does not exist unless lymphatics are totally and permanently ablated. If the standard dissection was carried out, but a narrow bridge of skin, in which lymphatics could be demonstrated by dye injection, was left intact connecting flap to host, an inlaid skin allograft was rejected. Another method of maintaining lymphatic drainage was to dissect the flap from skin near the axilla where Keller’s (1937) studies of guinea pig anatomy indicated that a large lymphatic vessel should accompany the blood vessels used to form the “umbilical cord.” Dye injections confirmed the presence of this lymphatic, which led directly to the axillary lymph nodes. Flaps in this location afforded no protection to inlaid skin allografts unless the lymphatic vessel was ligated. Finally, it was noted by means of dye injections that lymphatic continuity could be restored to flank-based flaps of long standing if they were removed from their protective capsules, trimmed at the margins and undersides to remove epidermis and then sutured into freshly excised skin defects, with the umbilical cords being preserved. Skin allografts in healthy, isolated skin flaps of 22-25 days’ standing were rejected in 14-23 additional days when lymphatic continuity was reestablished in their sustaining flaps by this method.


Use of the isolated flap model has also provided evidence regarding the minimum time required for presence of a skin allograft to sensitize its host fully. A conventional orthotopic skin allograft from a strain No. 2 donor was placed on the right side of a strain No. 13 guinea pig, where it would be expected to elicit sensitivity if left in place. Con- comitantly, on the contralateral side, another strain No. 2 graft was placed in an isolated flap to serve as an indicator. The orthotopic graft was completely excised after 2-6 days. No intraflap indicator graft was rejected unless its accompanying orthotopic graft was left in place a full 3 days. Fifty percent of the animals rejected indicator grafts when orthotopic grafts were excised after 4 days, and 100% if the excisions were after 5 days.

Futrell and Myers (1972a) recently utilized the artificial privileged site represented by the skin-flap model to demonstrate the essential role of regional lymphatics in early tumor recognition, using both allogeneic tumors and syngeneic ones with tumor-specific antigens. Ear- lier experiments with techniques of lymph node excision had provided conflicting evidence on the importance of lymphatics to tumor growth, probably because these studies (Crile, 1965; Bard et al., 1969; Fisher and Fisher, 1971) were subject to the shortcomings outlined above for allogeneic skin, i.e., lymphatic ablation was incomplete or transient. Futrell et uZ. (1972) found that 6 million cells from a methylcholanthrene-induced sarcoma (characterized by a tumor- specific antigen) inoculated into syngeneic guinea pig hosts resulted in small tumors that were soon completely rejected if the injection was made into intact skin. However, similar inoculation into alymphatic skin flaps led to progressive tumor growth, which eventually caused the deaths of the animals.

Unlike animals bearing intraflap skin allografts studied by Barker and Billingham (1968), guinea pigs harboring enlarging intraflap tumors became sensitized within 20 days as shown by their failure to accept a subsequent intraflap inoculum of the same tumor. The authors suggested that embolization of antigenic tumor fragments may have occurred via the venous route, although evidence of distant metasta- ses was not found. Another explanation is that sensitization occurred peripherally” as a consequence of circulation of immunocompetent

cells through the vasculature of the large tumors, which certainly must provide more “contact time” (see Strober and Gowans, 1965) for circulating host lymphocytes to be primed by alloantigens than would the smaller mass of tissue comprising a skin allograft. This study is important and somewhat disturbing, since lymphatic ablation is a common accompaniment of clinical therapy for malignancy. It remains

“


unclear, however, whether lymphatic ablation is ever actually detri- mental in clinically detectable tumors since, in Futrell and Myers’ model, once a critical tumor size had developed, host sensitization was not prevented by lymphatic interruption, nor was immune destruction of the tumor still possible.

Working with rats, Tilney and Gowans (1971) also studied the isolated alymphatic skin flap as a privileged site for skin allografts, repeating many of the studies carried out by Barker and Billingham in the guinea pig. They confirmed the prolonged survival of intraflap allografts, but did find evidence of eventual host sensitization as demonstrated by accelerated rejection of subsequently transplanted orthotopic test skin allografts. They also noted that intraflap skin allografts began undergoing contracture after about 2 weeks. Although this contracture was progressive and interpreted as a manifestation of rejection, the epidermal necrosis usually characterizing the end point of survival of orthotopic skin allografts did not occur and the microscopic appearance of intraflap allografts was not strikingly different from that of intraflap isografts, both types displaying mononuclear cell infiltration.

Tilney and Ford (1974) recently presented evidence of a proliferative response in the thymus-independent areas of the spleen (as determined by [3H]thymidine uptake) in the absence of proliferative responses in the regional nodes or other T-cell areas in rats bearing intraflap allografts. This raises the interesting possibility that antibody, rather than cell-mediated immunity, might have been the mediator of the slowly progressive damage of intraflap allografts seen by Tilney and Gowans.

Barker and Billingham (1968) occasionally (in 3/25 animals) observed signs of rejection in long-standing intraflap skin allografts in guinea pigs, and speculated that lymphatic regeneration might have been responsible, although this could not be demonstrated by dye injections. Tilney and Gowans also sought, unsuccessfully, for evidence of lymphatic regeneration by radioactive counting of regional nodes after injection of I3lI-labeled albumin into long-standing flaps. Their interpretation of the chronic rejection seen was that sensitization of their rats took place via the blood, probably by the process of peripheral sensitization. Since they used trunk skin grafts, which are much thicker than ear skin and might be expected to provide greater “contact time” for host lymphocytes circulating through the graft vasculature than the very thin ear skin allografts used by Barker and Billingham, peripheral sensitization might be a more likely outcome in their experiments. However, Silvers (1977) recently noted that ear


skin allografts fare better than trunk skin when transplanted across weak histoincompatibilities in areas of intact lymphatic drainage. Another possible source of the variation observed in these experiments is the greater number of “passenger” cells present in the thicker grafts of trunk skin than in ear skin.

Studies of these alymphatic skin flaps, while providing definitive information regarding the critical importance of the lymphatic drainage of skin allografts, also intensify interest in the differences between host responses to skin and vascularized whole organ allografts. If kidney allografts are performed in such a way that the lymph from the transplanted organ is diverted from reaching the host, prompt rejection is not prevented in dogs (Hume and Egdahl, 1955), sheep4Peder- sen and Morris, 1970), or man (Lavenderet d., 1968). A likely possibility is that the mechanism of peripheral sensitization, which may play no more than a minor role in the instance of skin allografts, operates effectively when the large mass of a kidney with its extensive vasculature is available for “contact time” with circulating antigen-sensitive host lymphocytes. Despite the minimal contact time possible within a small skin graft, however, it is generally conceded that successful immunosuppression is much harder to achieve for skin than for kidney allografts and that skin allografts are unusually potent stimuli. of last- ing immunity despite their small mass. One possible explanation of this would be a unique property of skin tissue such as the existence of a skin-specific antigen (see Lance et al., 1971). Another is that the mode of grafting skin could provoke an especially vigorous rejection; for example, transplanting skin to a bed rich in lymphatics allows easy access to regional nodes (with their high T-lymphocyte population) of particulate antigen of graft otigin, passenger leukocytes, or host macrophages and antigen-primed lymphocytes. Additionally, it has been suggested that the several days of ischemia suffered by a free skin graft prior to neovascularization could render it more immunogenic by facilitating the release of antigenic material and that tissue degradation products could act as immunologic adjuvahts. Furthermore, a skin graft already suffering areas of ischemic necrosis and nonspecific inflammation might be more vulnerable to the additional onslaught of rejection and thus succumb more promptly. If skin were transplanted in the manner of a whole organ, an ischemic interval being avoided by anastomoses of the nourishing blood vessels, it might be possible to delineate the reasons for differences in rejection of skin and whole organs. Such attempts have been plagued with technical difficulties of the microsurgery, but have been successful in several instances.

Cho et al. (1972) and Salyer and Kyger (1973) found that avoidance


of the usually obligatory ischemic interval for free skin allografts, by carrying out immediate surgical revascularization of allogeneic skin pedicles, did not prolong their survival in rats across either weak or strong histocompatibility barriers. However, in these experiments the revascularized pedicles were sutured into hll-thickness skin defects prepared on the new hosts so that lymphatic continuity was reestablished promptly. Since the lymphatic drainage factor seems to be of overriding importance in rejection of skin allografts, prompt destruction of the allogeneic pedicles was rather expected. More surprising is the report of Wustrack et al. (1975) and Wustrack and Lucas (1974) that immediately revascularized skin allografts, in which an alymphatic state was maintained by isolating them from the host with Silastic sheeting, were also promptly rejected by Ag-B incompatible rat hosts. These workers confirmed, however, that alymphatic skin flaps prepared in the host’s own skin (in manner of Barker and Billing- ham) were privileged sites for inlaid skin allografts. A possible explanation of these apparently paradoxical findings is that the full thickness allogeneic skin pedicles, which were transplanted by vascular anastamoses along with associated subcutaneous fat and blood vessels, constituted a much larger mass of tissue than small free skin grafts inlaid in host-origin alymphatic flaps. The larger allografts, similar in size to a rat kidney, would transfer more passenger leukocytes and provide more opportunity for “contact time” with the antigen-sensitive host lymphocytes responsible for peripheral sensitization. Another pertinent finding of Wustrack et al. was that, if enhancing antisera were administered, the immediately vascularized but alymphatic allogeneic skin pedicles did enjoy prolonged survival. However, when lymphatic continuity with the host was allowed, enhancement of the flap survival could not be obtained.

A possible role of immunologic enhancement in the prolonged survival of skin allografts transplanted to alymphatic flaps has also been suggested since antigen can leave such grafts only by the venous route, known to be efficient for production of enhancing antibody. Previous studies, including those of Barker and Billingham (1968), failed to substantiate this possibility, which was tested simply by applying orthotopic skin allografts to guinea pigs bearing intraflap allografts of long standing. Survival of the test allografts was not enhanced and their prompt rejection was accompanied by the destruction of the intraflap allografts. This test system may be overly exacting since skin allografts are notoriously difficult to enhance. Merriam and Tilney (1976) harvested serum from L E rats bearing healthy intraflap (LE X BN)FI skin allografts transplanted 8 and 12 days earlier. This serum was injected into LE strain recipients of heterotopic accessory h)eart allografts, known to be more susceptible to enhancement than


skin allografts. The heart allografts kept beating for 13-15 days as compared with 7 days in untreated recipients. Although enhancement is unlikely to be the complete explanation of the survival of intraflap allografts, it would not be surprising to find that small amounts of antigen released from allografts exclusively into the venous system stimulate production of blocking antibody (see Billingham and Sil- vers, 1964; Kaplan and Streilein, 1974).

A full explanation of the isolated flap as a privileged site awaits further exploration of possible roles of enhancement, suppressor cells, trapping of passenger leukocytes, and primed circulating host lymphocytes. The critical importance of the lymphatic circulation in the reaction to skin allografts is, however, firmly established.

B. TRAUMATIZED PANNICULUS CARNOSUS MUSCLE

In rodents, the usual bed for skin grafts is the carefully exposed and undamaged fascia1 surface of the panniculus carnosus muscle, on which a network of lymphatic vessels can be seen. Allografts transplanted to this site are rejected with vigor. The substance of the muscle itself is known to be deficient in lymphatic vessels (Godart, 1968; Landsteiner and Chase, 1939), and because of this, Barker and Billingham (1972) investigated the possibility that muscle might be a privileged site for skin allografts if they were transplanted in such a way that their impingement on lymphatics was prevented. It was reasoned that contact with lymphatics would have to be avoided at the perimeter of grafts as well as at the interface of the graft’s dermal surface with host tissue, since full-thickness allografts placed in open-fit fashion at the center of beds of intact panniculus carnosus muscle are rejected promptly (Medawar, 1944). In Medawar’s experiments, carried out in rabbits, the lymphatics coursing on the surface of the muscle that formed the graft bed were intact since in this species a natural cleavage plane exists beneath the dermis, and the fascia remains undamaged when the skin is removed.

In both the guinea pig and the rat, however, Barker and Billingham (1972, 1973b) found that if the limits of the prospective bed were defined by sharp dissection with a scalpel and the skin was grasped with a hemostat and avulsed from the underlying tissue, the surface of the panniculus carnosus muscle was inevitably denuded of its superficial fascia and associated lymphatic vessels, since in these species the skin is tightly knit to the epimysium of this underlying muscle. The survival times of skin allografts placed at the center of large, open, lymphatic-poor beds prepared by the above technique were about double the survival times of control grafts fitted carefully into their beds, the latter technique promoting rapid establishment of 1ymDhatic

42 CLYDE F. BARKER AND R. E . BILLINGHAM

continuity with the host. If grafts were placed on large beds but located eccentrically so that they contacted host skin along one margin, rejection followed promptly. The stress of the large open wounds and the resultant release of corticosteroids appeared not to be responsible for increased longevity of the alien tissue, since preparation of similar wounds on the one side of an animal did not prolong survival of fitted allografts on the contralateral side. “Open-fit” grafts were rejected by presensitized animals in an accelerated manner, indicating an intact efferent limb of the immunologic reflex.

C. MUSCLE

Although muscle is not generally categorized as a privileged site, several experiments tend to substantiate the relative effectiveness of this tissue in protecting implanted allografts. Despite the difficulties of evaluating survival of skin which is completely buried in intramuscular transplant sites, Whitmore and Gittes (1975) obtained histologic evidence that rejection of skin allografts was delayed by implanting them in muscle. Functional endocrine grafts, monitored by chemical assays, have been more convenient in evaluating the intramuscular transplant site. Russell and Gittes (1959) were intrigued with their finding that some parathyroid allografts implanted intramuscularly in previously parathyroidectomized rats functioned as long as 140 days as determined by normocalcemia, although others survived only 9-14 days before rejection, precipitating recurrent hypocalcemia. Since these workers used outbred animals, differences in histocompatibility may have accounted for the variation in graft survival times. However, Raaf et uZ. (1974), in repeating Russell and Gittes’ experiments with inbred rats, also noted two distinct categories of rejection of intramuscular parathyroid allografts which they classified as either “fast” or

slow.” Since inconsistencies in histoincompatibility were excluded as a cause of variation in graft survival, the authors invoked as an explanation the vagaries of the local anatomical situation of the transplant site in individual animals-particularly with regard to the state of neovascularization (either blood or lymphatic). N d i and Barker (1976) also investigated the intramuscular site for parathyroid allo- transplants using both Ag-B compatible and incompatible donor- recipient rat strain combinations. As in the previous studies, which had been done under circumstances of either undefined or major locus mismatched histocompatibility, “fast” and “slow” rejections were again noted. Rejection in less than 20 days occurred in 33% of strongly incompatible grafts (ACI + FI), but 17% survived more than 100 days. Nine percent of Ag-B locus compatible allografts (LE --* FI) were re-

“


jected within 20 days, but 73% survived longer than 100 days. Since rejection of intramuscular parathyroid allografts of more than 100 days’ standing could always be procured promptly by grafting the recipients with donor strain skin, prolonged survival appeared to be an oversight in the afferent limb of the immunologic reflex. Since the paucity of lymphatics in the muscle seemed a likely explanation for this, an effort was made to devise, as a control, an implant site approximating other characteristics of the intramuscular one, but differing to the extent of allowing opportunity for prompt lymphatic revascularization. Parathyroid allografts were therefore placed in subder- ma1 pockets so that they rested immediately on the surface of the intact, and lymphatic-rich, fascia of the panniculus carnosus muscle. In this position, they consistently underwent prompt rejection (MST: 12.5 days for ACI + FI; 16.5 days for LE + FI). Other lymphatic-rich areas such as the renal and hepatic subcapsular spaces also failed to protect parathyroid allografts against prompt rejection, a finding that tends partially to discredit the validity of previous speculation that this tissue lacks full antigenicity (Russell and Gittes, 1959).

D. SI<IN ISLANDS

Although there is agreement that isolated flaps of skin are effective artificial privileged sites, their construction requires surgical finesse and, despite meticulous attention to protection of the nourishing “umbilical cord,” traumatic occlusion of this tenuous blood supply usually shortens the possible observation period (Barker and Billingham, 1968; Tilney and Gowans, 1971). Barker and Billingham (1972) described a less exacting method of achieving isolation from the lymphatic circulation which also proved effective in converting a portion of skin to a privileged site. In these experiments, large rectangular beds of “raw” panniculus muscle were prepared on the flanks of guinea pigs or rats. The epimysium and its contained lymphatics were stripped off the muscle by the technique described above. However, at the center of the large open bed, a 1.5 x 1.5 cm island of skin was allowed to remain undisturbed and attached to the underlying muscle. A similar model in which the entire panniculus carnosus muscle was excised, leaving an island of skin, was noted by Landsteiner and Chase (1939) to prevent egress of injected vital dyes and to be an inefficient site for induction of generalized hypersensitivity to poison ivy in guinea pigs. Barker and Billingham (1973b) found that Patent Blue V dye injected up to 18 days after construction of the islands failed to enter lymphatics or appear in regional lymph nodes. How- ever, beyond 18 days, alternative lymphatic pathways were formed in


most cases as demonstrated by dye injection-indicating that skin islands were likely to be only a transient privileged site. This proved to be the case, since in strain No. 13 guinea pigs the MST of strain No. 2 skin grafts inlaid into skin islands was 18.2 days as compared with only 11.7 days for grafts transplanted to areas of intact skin. In Ag-B locus incompatible rats, prolongation of allograft survival was less impressive, but if Ag-B compatibility existed the MST was 19.8 days and 4 of 15 grafts survived longer than 50 days. Stress, enhancement, and efferent blockage were ruled out by appropriate controls.

Ziegler et al. (1972, 1973) studied skin islands as a site for tumor transplantation. Cell suspensions (15-35 x lofi cells) prepared from a LE rat mammary tumor were injected intradermally into either intact flank skin or skin islands in LE, FI, or BN rat hosts. I n syngeneic animals, the tumor grew rapidly in flank skin to a maximal mean size (MMS) of 92.0 cm3, killing the hosts in a mean of 46.5 days. Tumor allografts were rejected rapidly if placed in intact flank skin of either Ag-B locus compatible FI hosts (MMS of 0.3 cm3 and mean tumor regression time, MTRT, of 7.3 days) or Ag-B locus incompatible BN rats (MMS of 0.5 cm3 and MTRT of 9.9 days). Tumor allografts in alymphatic skin islands fared much better (MMS of 13.9 cm3 and MTRT of 42.1 days in FI rats and MMS of 14.6 cm3 and MTRT of 22.3 days in BN hosts). The tumor allografts eventually disappeared in all cases, but never before wound contracture had restored skin, and thus lymphatic continuity, to the surrounding flank skin. That the efferent limb of the immune response remained intact was evidenced by the accelerated rejection of tumors implanted in skin islands in rats presensitized by donor strain skin or tumor allografts.

The growth pattern of syngeneic tumors expressing a tumor-specific antigen was also found to be influenced significantly by the lymphatic status of the transplant site. Viable cells (0.05-0.5 x lo6) from a LE rat methylcholanthrene-induced fibrosarcoma were injected into intact skin or skin islands of LE hosts. Palpable tumors were noted earlier (10.4 VS 16.6 days), grew more rapidly (38.2 cm3 VS 16.6 cm3 at 25 days) and killed their recipients sooner (38.0 VS 47.4 days) if the cells were injected into alymphatic skin instead of into intact skin.

The skin island was also used by Ziegler et al. (1973) to study the action of neuraminidase (Vibrio cholerae neuraminidase, VCN) on a LE methylcholanthrene-induced tumor in syngeneic hosts. When tumor cells were incubated with VCN (200 U/ml, lo6 cells at 37°C for 1 hour) prior to their injection into intact skin, they grew much more slowly than untreated cells (palpable tumor onset 19.3 days, 4.77 cm3 at 25 days and host death at 51.8 days). VCN-treated cells injected into


skin islands, however, had a similar growth pattern to that of untreated cells (onset 11.3 days, 23.1 cm:’ at 25 days and death at 42.4 days). This provides evidence that the retarded growth of VCN-treated cells is on an immunologic basis. In the instance of some other control models used to study effects of VCN, such as heat inactivation of VCN or immunosuppression of recipient animals, the possibility of a cytotoxic effect of VCN was not completely dispelled. Confirmation of the importance of the afferent side of the immunologic reflex to the heightened response to VCN-treated cells was provided by observation of a test inoculum of untreated tumor cells in intact skin of one side and simultaneous injection of VCN-treated cells on the other side of a host’s trunk, either into the intact skin or a skin island. Inhibition of growth of the untreated tumor was observed only when the previous injection of VCN-treated cells was made into skin having intact lymphatic drainage.

Skin islands have also been used with success as implant sites for endocrine allografts (Naji and Barker, 1976). In rats, the MST of parathyroids transplanted to skin islands across a major histocompatibility barrier (ACI + FI) was 22.0 days, but if the transplant site was subdermal, onto the surface of lymphatic-rich muscle fascia in an area of intact skin, rejection took place in 12.5 days. When Ag-B compatibility prevailed (LE + FI), only one of 10 animals rejected an intra- island parathyroid graft in less than 50 days and the MST was >lo0 days, despite prompt rejection (16.5 days) of allografts in the control site, which was identical except for its possession of intact lymphatics.

E. CORNEAL DIFFUSION CHAMBERS

Sturgis and Castallanos (1962) devised an interesting technique that enabled ovarian allografts to survive for greatly extended periods as a consequence of residence in what may be considered as artificial privileged sites. These “sites” were chambers fashioned like little purses out of allogeneic corneas and implanted between the rectus muscle and the peritoneum in the ventral abdominal wall. This novel approach was motivated by the author’s belief that the corneal chamber would function transiently as a cell-impermeable diffusion chamber, protecting the enclosed alien ovarian tissue, it was hoped, at least until the latter had undergone some form of antigenic adaptation to the host. The eventual disintegration of the chambers would, so they reasoned, “liberate” the enclosed grafts, allowing them to be revascularized.

With this artifice, endocrinologically functional ovarian tissue allografts survived for a year or more in ovariectomized rabbits and


monkeys, and also in humans, in contrast to the short survival times of similar grafts implanted in similar locations without the benefits of the protective corneal chambers. This finding was subsequently confirmed by Shaffer and Hulka (1969), working with ovariectomized rabbits. These investigators also obtained suggestive evidence that some kind of adaptation did indeed take place at the level of the graft; for example, they observed that all vascularized grafts continued to survive in animals that rejected subsequent challenge allografts of skin or ovary from the original donor. They raised the interesting possibility that steroid synthesis on the part of the “luteinized” cells deep in the ovarian stroma might have been responsible for interfering lo- c d y with the host’s ability to mount an effective immune response. No information was provided concerning the status of lymphatic drainage in either the initially established corneal chambers or in the eventually liberated ovarian tissue allografts.

F. SCARS It is well known that scars of various types may be the sites of tumor

formation. Operative scars are frequent sites of tumor recurrency, presumably on the basis of seeding the wound with viable malignant zells. The occurrence of tumors at the site of skin incisions made years earlier for benign conditions or in fistulous tracts is more difficult to explain. Tumorigenesis could possibly be related nonspecifically to injury per se, since a wound or scar subject to repeated trauma seems particularly likely to be the site where an epidermoid carcinoma develops (see Peacock and Van Winkle, 1976). Epithelial migration, a part of normal wound healing which sometimes occurs aIong suture tracts leaving epithelial sinuses or cysts buried in or near a scar, has some of the histological characteristics of neoplasia, but is not necessarily related to tumor formation. Burn scars are even more likely than those produced mechanically to be the sites of development of malignant tumors. Shortwave radiation (X-ray) is particularly dangerous, but longwave radiation (thermal) may also lead to neoplasia by the same mechanism, production of mutant cells.

Another factor that could be important in the development of scar cancer is obliteration of the lymphatics in a cicatrix, with the possible impairment of normal immunologic surveillance mechanisms. Futrell and Myers (1972b) studied experimentally the possibility that bum scars could be privileged sites for tumor growth. In guinea pigs they found that a methylcholanthrene-induced liposarcoma, characterized by tumor-specific antigens, would grow when inoculated into standard experimental burn scars, while a similar dose (5 x 10.’) of tumor cells was always rejected if injected into normal skin.


Another type of scar in which tumors are known to occur is the pulmonary scar resulting from infarction or from old inflammatory processes (see Bal6 et al., 1956; Freant et al., 1974). It has been suggested (Carroll, 1962) that the blockage of lymphatic drainage may play a role in the development of these tumors. If so, the mechanisms could be either abnormal accumulation of inhaled carcinogens in the scarred area, or possibly impaired surveillance for mutant neoplastic cells. That the specific type of injury is probably unimportant etiologi- cally is suggested by the finding of Montgomery (1944) that small incisions in the lungs of experimental animals occasionally heal with frankly malignant transformation.

G. MILLIPORE DIFFUSION CHAMBERS

As early as 1932, Rezzesi described a method of isolating transplanted tissue by placing it in a chamber constructed of semipermeable material. Bisceglie ( 1933) reported survival for 12 days of mouse tumor cells contained in a collodion bag which was placed in guinea pigs. Modem investigation of the diffusion chamber as a method of preventing allograft rejection was pioneered by Algire et al. (1954). As used by these workers and others since (see Amos, 1961), the diffusion chamber is a flat capsule sealed on two faces by a semipermeable membrane. The membrane, which is of a plastic material usually having a pore size of 0.1-0.8 pm, is thought to serve several functions: (1) to confine transplanted tissue in the chamber so that it can be recovered for examination; (2) to prevent escape of antigenic particles or cells so that active immunization is avoided; (3) to exclude antigen-sensitive cells of host origin, thus preventing peripheral sensitization; (4) to exclude immune cells from the host if a state of immunity already exists; and (5 ) to admit nutrient-containing fluids to support the grafted tissue.

Algire’s finding (1954) of the prolonged survival of allografts transplanted in Millipore chambers, which were presumed to permit passage of humoral but not cellular elements even in preimmunized hosts, was taken to constitute strong evidence against a significant role of antibody in the rejection process. Subsequently, however, it was found that antibody and complement sometimes penetrate such chambers with difficulty, but that if this does occur allogeneic tissues may be destroyed (Algire et al., 1957). Thus, it seems that the modus operandi of this artificial privileged site may depend partially on aEe- rent and partially on efferent blockade of the immunologic reflex.

In the case of xenogeneic tissue, Algire et a2. (1954) found that Mil- lipore chambers failed to protect lung or epidermoid tissue of mouse origin from very prompt destruction in either presensitized or normal


rats, probably indicating that humoral factors are especially important in rejection of xenografts. More recently, Reemtsma’s (1970) efforts were also frustrated when piscine endocrine xenografts transplanted in diffusion chambers into rats were destroyed within a few days.

The small mass of normally functioning tissue that is necessary for reversal of endocrine deficiency states has been reported to allow their successful treatment by transplantation of tissue in diffusion chambers at least for short intervals. Strautz (1970) reported that islets of Langerhans, isolated from the pancreas of normal mouse donors and transplanted in Millipore chambers to genetically obese and diabetic (ob/ob) recipients, resulted in reversal of hyperglycemia and hyperin- sulinemia for several weeks. Gates et al. (1972) transplanted normal islets from albino mice to Millipore chambers in diabetic New Zea- land Obese mice with similar reversal of the diabetic parameters to normal. After 10 weeks, the chambers were removed from the recipients, which then returned to the diabetic state. The islets, when recovered, were found to be fully viable as evidenced by their in uitro secretion of insulin when stimulated by exposure to high concentrations of glucose.

A difficulty in indefinitely extending the survival of tissue transplanted to conventional Millipore chambers is that fibrin and adhesions from host tissues tend to cover the membranes in time, clogging the pores (Nettesheim and Makinodan, 1967), thus interfering with diffusion of nutrients across the membrane. A new type of chamber may obviate this problem by allowing continuous circulation of rapidly moving blood past the membrane. Knazek et al. (1972) have reported culturing cells on the outer surface of artificial capillaries while culture medium circulates inside the lumen of the capillary. The material constituting the “capillary wall” of the hollow fiber capillaries consists of a thin retentive skin surrounded by an outer mac- roporous spongy layer on which cells may be seeded. The membrane is semipermeable, with pores large enough to admit glucose or insulin but not cells or y-globulin. Bundles of 100 such fibers have been sealed into a glass jacket and used by Chick et al. (1975a,b) for growing beta cells from rat fetal pancreas for as long as several weeks. Increasing or decreasing the concentration of glucose in the media circulating through the lumen of the fibers brought about similar changes in the quantity of insulin being released from the capillary unit.

Tze et al. (1976) recently implanted similar units into the aortas of diabetic rats, so that blood circulated through the lumen of the hollow fibers which had been provided with isolated islets growing on the


niacroporous outside layer of the membranes. The authors found that the blood glucose levels of their diabetic animals returned toward normal and that serum insulin significantly increased during an observation period of up to 13 hours. Problems with clotting of blood or bleeding from excessive heparinization appear to be major difficulties still to be solved before it can be determined whether this type of artificial privileged site can prevent rejection of allogeneic tissue.

ACKNOWLEDGMENTS The expenses of some of the experimental work cited and the preparation of this

article were defrayed in part by Grants AI-10678 and A1-11051 from the US. Public Health Service.

It is a pleasure to acknowledge o u r indebtedness to Ms. Mary Owens and Ms. Roxanne Holt for preparation of the manuscript and to Dr. Alan E. Beer for helpful criticism and advice.

REFERENCES AI-Askari, S., Dunionde, D. C., Lawrence, H. S., and Thomas, L. (1964).Ann. N.Y. Acad.

Albrink, W. S., and Greene, H. S. N. (1953). Cancer Res. 13, 64-68. Algire, C . H., Weaver, J. M., and Prehn, R. T. (1954).J. Natl. Cancer Inst. 15,493-518. Algire, G. H., Weaver, J . M., and Prehn, R. T. (1957).Ann. N.Y. Acud. Sci. 64,1009-1013. Amos, D. B. (1961). I n “Transplantation of Tissues and Cells” (R. E. Billingham and

W. K. Silvers, eds.), pp. 69-80. Wistar Inst. Press, Philadelphia, Pennsylvania. Arm, M., Maresceaux, J.. and Petrovic, A. (1957). Colloq. Int. C . N . R . S . 78, 2533 . Asherson, C. L., and Zembala, M. (1976). Br. Med. Bull. 32, 158-164. Athias, M., and Cuimarais, A. (1933). C. R. Seances Soc. B i d . Ses F i l . 113,733. Bal6, J . , Jtiltasz, E., and Temes, J. (1956). Cancer 9,918-922. Bard, D. S., Hammond, W. C., and Pilch, Y. H. (1969). Cancer Res. 29, 1379-1384. Barker, C. F. (1975). Diabetes 24, 766-775. Barker, C. F., and Billingham, R. E. (1968).J. E x p . Med. 128, 197-221. Barker, C . F., and Billingham, R. E. (1971).J. E x p . Med. 133,620-639. Barker, C. F., and Billingham, R. E. (1972). Ann. Surg. 176,597-604. Barker, C . F., and Billingham, R. E. (1973a). Ciba Found. Symp. 15 (N.S.), 80-104. Barker, C . F., and Billingham, R. E. (1973b). J . E x p . Med. 138, 289-299. Barker, C. F., and Corriere, J . N., Jr. (1967). Ann. Surg. 165, 279-282. Barker, C. F., Reckard, C . R., Ziegler, M. M., and Naji, A. (1975). Diabetes 24,418. Battisto, J . R., and Miller, J. (1962). Proc. Soc. E x p . Biol. Med. 1 1 1 , 111-115. Beer, A. E., and Billingham, R. E. (197O).J. E x p . Med. 132,721-736. Beer, A. E., and Billingham, R. E. (1971).Ado. Zmmunol. 14, 1-84. Beer, A. E., and Billingham, R. E . (1974).J. Reprod. Fertil., S u p p l . 21, 59-88. Beer, A. E., and Billingham, R. E. (1976). “The Immrinohiology of Mammalian Repro-

Billingham, R. E. (1968). Haroey Lect. 62,21-78. Billingham, R. E. (1971). Cell. Immunol. 2, 1-12. Billingham, R. E., and Boswell, T. (1953). Proc. R. Soc. London, Ser. B 141,392-406. Billingham, R. E., and Silvers, W. K. (1962). Transplantation, Ciba Found. Symp. 1961,

Sci. 120, 261-269.

duction.” Prentice-Hall, Englewood Cliffs, New Jersey.

pp. 90-108.


Billingham, R. E., and Silvers, W. K. (1963). Ann. N.Y. Acad. Sci. 100,348-360. Billingham, R. E., and Silvers, W. K. (1964). Proc. R. Soc. London, Ser. B 161,168-190. Billingham, R. E., and Silvers, W. K. (1970)./. E x p . Med. 131, 101-117. Billingham, R. E., and Silvers, W. K. (1971). “The Immunobiology of Transplantation.”

Billingham, R. E., Brent, L., and Medawar, P. B. (1954). Proc. R. Soc. London, Ser. B 143,

Billingham, R. E., Ferrigan, L. W., and Silvers, W. K. (1960). Science 132, 1488. Bisceglie, V. (1933). 2. Krebsforsch. 40, 141. Blair, P. B., and Moretti, R . L. (1967). Transplantation 5, 542-544. Bloom, W., and Fawcett, D. W. (1975). “A Textbook of Histology.” Saunders, Philadel-

Boeckx, W., Sobis, H., Lacquet, A., Gruwez, J., and Vandeputte, M. (1975). Transplanta-

Brent, L., and Kilshaw, P. J. (1976). Transplant. Proc. 8, 209-215. Browning, H. C. (1949). Cancer 2,646-672. Bumet, M . (1970). “Immunological Surveillance.” Pergamon, Oxford. Calne, R. Y., White, H. J. O., Binns, R. M., Herbertson, B. M., Millard, P. R., Pena, J.,

Cantor, H. M., and Dumont, A. E. (1967). Nature (London) 215,774-745. Carroll, R. (1962). J . Pathol. Bacteriol. 83, 293-297. Chadwick, D. E., and Blarney, R. W. (1968). Transplantation 6, 544-548. Chase, M. W. (1946). Proc. Soc. E x p . B i d . Med. 61, 257-259. Chick, W. L., Like, A. A., and Lauris, V. (1975a). Science 187,847-849. Chick, W. L., Like, A. A., Lauris, V., Galletti, P. M., Richardson, P. D., Panol, G., Mix,

T. W., and Colton, C. K. (1975b). Trans. Am. Soc. ArtiJ Intern. Organs 21,U. Cho, S. I., Marcus, F. S., and Kountz, S. L. (1972). Transplantation 13, 486-492. Chrktein, P. B., Behar, R. J., Kohn, Z., Moldovanu, G., Miller, D. G., and Lawrence, W.,

Collin, H. B. (1966). Znoest. Ophthalmol. 5, 337-354. Collin, H. B. (1970). Znoest. Ophthalmol. 9, 146-155. Connelly, D. M. (1961). Plast. Reconstr. Surg. 28, 1-11. Crile, G., Jr. (1965). Surg., Gynecol. Obstet. 120, 975-982. Dameron, J. T. (1950). Proc. SOC. E x p . Biol. Med. 73,343-345. Dameron, J. T. (1951). Surgery 30, 787-799. Dib-Kuri, A., Revilla, A., and Chavez-Peon, F. (1975). Transplant. Proc. 7, 7<53. Dym, M., and Fawcett, D. W. (1970). Biol. Reprod. 3,308-326. Ebeling, E. (1914). Z. Krebsforsch. 14, 151. Engeset, A. (1959). /. Anat. 93,96-100. Faulk, W. P., Sanderson, A. R., and Temple, A. (1977). Transplant. Proc. 9, 1379-1384. Fisher, B., and Fisher, E. R. (1971). Cancer 27, 1001-1004. Fonkalsrud, E. W. (1968). Surg., Gynecol. Obstet. 127, 71-75. Franklin, R. M., and Prendergast, R. A. (1970).J. Immunol. 104, 463-469. Franks, L. M. (1956). Lancet 2, 1037. Franks, L. M. (1957). Transplant. Bull. 4 , 3 0 3 1 . Freant, L. J . , Joseph, W. L., and Adkins, P. C. (1974).Ann. Thorac. Surg. 17,531-537. Frey, J. R., and Wenk, P. (1957). Int. Arch. Allergy Appl. Immunol. 11, 81-100. Fukuda, A., Hanaoka, T., Solowey, A. C., and Rapaport, F. T. (1969). Transplant. Proc. 1,

Futrell, J. W., and Myers, G. H., Jr. (1972a). Transplantation 13, 551557. Futrell, J. W., and Myers, G. H., Jr. (1972b). Surg. Forum 23, 129-131.

Prentice-Hall, Englewood Cliffs, New Jersey.

58-80.

phia, Pennsylvania.

tion 19, 145-149.

Salaman, J. R., Samuel, J . R., and Davis, D. R. (1969). Transplant. Proc. 1,321324.

Jr. (1967). Anat. Rec. 159, 5-15.

602-604.


Futrell, J. W., Albright, N. L., and Myers, G. H., Jr. (1972).]. Surg. Res. 12, 62-69. Gates, R. J., Hunt, M. I., Smith, R., and Lazarus, N. R. (1972). Luncet 2, 567-570. Gittes, R. F., and McCullough, D. L. (1974).J. Urol. 112,241-244. Godart, S. (1968). Lymphology 1, 80-87. Goldenberg, D. M. (1970). Experientia 26,907-908. Goldenberg, D. M., and Steinborn, W. (1970). Proc. SOC. E x p . Biol. Med. 135,724-726. Goldenberg, D. M., Pegram, C. A., and Valquez, J . J . (1975).J. Zmmunol. 114, 1008. Goodfellow, P. N., Barnstable, C. J., Bodmer, W. F., Snary, D., and Crumpton, M. J.

Grebe, S. C., and Streilein, J. W. (1976). Adu. fmmunol. 22, 120-222. Greenblatt, M. (1972). Prog. E x p . Tumor Res. 16,380-395. Greene, H. S. N. (194O).J . E x p . Med. 71,305-324. Greene, H. S. N. (1952). Cancer 5, 24-44. Greene, H. S. N. (1957). Ann. N.Y. Acad. Sci. 69,818-829. Greene, H. S. N., and Arnold, H. (1945). J . Neurosurg. 2, 315-331. Greene, H. S. N., and Murphy, E. D. (1945). Cancer Res. 5,269-282. Handler, A. H., and Shepro, D. (1968). In “The Golden Hamster; its Biology and Use in

Medical Research” (R. A. H o b a n , P. F. Robinson, and H. Magalhaes, eds.), pp. 195-202. Iowa State Univ. Press, Iowa City.

(1976). Transp lan ta t ion 22,595-603.

Hamner, J. E., 111. (1966). Oral Surg., Oral Med. Oral Pathol. 22, 114-119. Harris, J. E., and Rathbun, W. B. (1972). In “Transplantation” (J. S. Najarian and R. L.

Hegner, C. A. (1913). Miiench. Med. Wochenschr. 60,2722. Hertz, R. (1967). Choriocarcinomo, Trans. Conf., I965 UICC Monogr. Ser., Vol. 3, pp.

Heyner, S. (1973). Surg., Cynecol. Obstet. 136,298-305. Hume, D. M., and Egdahl, R. H. (1955). Surgery 38, 194-214. Iwao, T., Kurihara, M., and Ochiai, A. (1935). Trans. SOC. Pathol. Jpn. 25,610. Jones, B. (1973). Ciba Found. Symp. 15 (NS), 349-354. Kaplan, H. J., and Stevens, T. R. (1975). Transplantation 19, 302-308. Kaplan, H. J., and Streilein, J . W. (1974). Nature (London) 251, 553554. Kaplan, H. J., Stevens, T. R., and Streilein, J. W. (1975a).J. Immunol. 115, 800-804. Kaplan, H. J., Streilein, J. W., and Stevens, T. R. (1975b).J. Immunol. 115, 805-810. Keller, G. (1937). Z. Znfektionskr, Parasit. Krankh. Hyg. Haustiere 52, 250. Khodadoust, A. A,, and Silverstein, A. M. (1966). Surg. Ophthalmol. 11,435443. Khodadoust, A. A., and Silverstein, A. M. (1969). Inuest. Ophthalmol. 8, 180-195. Knazek, R. A., Gullino, P. M., Kohler, P. O., and Dedrick, R. L. (1972). Science 178,

Kreider, J . W., Haft, H. M., and Roode, P. B. (1971).J. Invest. D e m t o l . 57, 66-71. Krohn, P. L. (1965). Br. Med. Bull. 21, 157-161. Lambert, P. B., Frank, H. A., Bellman, S., and Farnsworth, D. (1965). Transplantation 3,

Lance, E. M. (1967). Surg., Cynecol. Obstet. 125,529-539. Lance, E. M., Boyse, E. A,, Cooper, S., and Campbell, E. A. (1971). Transplant. Proc. 3,

Landsteiner, K., and Chase, M. W. (1939).J. E x p . Med. 69,767-784. Lavender, A. R., Forland, M., Rams, J. J., Thompson, 1. S., Russe, H. P., and Spargo, B. H.

Lemon, H. M., Lutz, B. R., Pope, R., Parsons, L., Handler, A. H., and Patt, D. I. (1952).

Levine, S. (1968). Transplantation 6, 799-802.

Simmons, eds.), pp. 613-626. Lea & Febiger, Philadelphia, Pennsylvania.

2 6 3 1.

65-67.

62-73.

864-868.

(1968). J . Am. Med. Assoc. 203,265-271.

Science 115,461-465.

s2 CLYDE F. BARKER AND R. E. BILLINGHAM

Lindenmann, R., and Striuli, P. (1968). Transplantation 6, 557-561. Lutz, B. R., Fulton, G. P., Patt, D. I., and Handler, A. H. (1950). Cancer Res. 10, 231-

McLean, J. M., and Scothorne, R. J . (1970). J . Anat. 107, 39-48. McLenn, J. M., and Scothorne, R. J , (1972).J. Anut. 112, 423-432. McCullough, D. L. (1975). Invest. Urol. 13,211-219. Mandel, M. A., Monaco, A. P., and Russell, P. S . (1965)./. Zmmunol. 95, 673-682. Manning, D. D., Reed, N. D., and Shaffer, C. F. (1973)./. E x p . Med. 138, 488-494. Manson, L. A., and Palm, J . (1968). Transplantation 6, 667-669. Manson, L. A, , Foschi, G. V., and Palm, J. (1963).J. Cell. Comp. Physiol. 61, 109-119. Markee, J. E. (1932). Am. J . Physiol. 100, 32-39. Maumenee, A. E. (1951). Am. J. O ~ ~ h t h u l 3 4 , 142-152. Maumenee, A. E. (1973). Ciba Found. Symp. 15, (N.S.), 5-15. Medawar, P. B. (1944).J. Anat. 78, 176-206. Medawar, P. B. (1948). Br. /. E x p . Pathol. 29,58-69. Medawar, P. B., and Russell, P. S. (1958). Immunology 1, 1-12. Merriam, J. C., and Tilney, N. L. (1976). Surg. Forum 27,336. Menvin, R. M., and Hill, E. L. (19Fj4).J. Natl. Cancer Znst. 14, 819-830. Millard, P. R., Herbertson, B. M., and Calne, R. Y. (1971). Transplant. Proc. 3,505-508. Mitchison, N. A. (1954). Proc. R. SOC. London, Ser. B 142, 72-87. Montagna, W. (1962). “The Struchire and Function of Skin,” 2nd ed. Academic Press,

Montgomery, C. L. (1944). B r . / . Surg. 31,292-299. Morris, D. S., McDonald, J. R., and Mann, F. C. (1950). Cancer Res. 10, 36-48. Murphy, J. B., and Sturm, E. (1923). J . E x p . Med. 38, 183-197. Naji, A., and Barker, C. F. (1976).J. Surg. Res. 20, 261-268. Neaves, W. B. (1977). In “The Testis” (A. D. Johnson and W. K. Comes, eds.), Vol. 4,

pp. 125-162. Academic Press, New York. Olansky, S., Tully, H. T., and Knisely, W. H. (1959).J. Inuest. Dennatol. 32, 117-125. Parrott, D. M. V. (1960). Immunology 3,244-2.53. Peacock, E. E., Jr., and Van Winkle, W., Jr. (1976). “Wound Repair.” Saunders,

Pedersen, N. C., and Morris, B. (1970). J . E x p . Med. 131, 936-969. Penn, I. (1972).J. Am. Med. Assoc. 221, 1412. Peters, A., Palay, S., and Webster, H. de F. (1976). “The Fine Structure of the Nervous

Pfeffermann, R., Sakai, A., Auda, S., and Kountz, S. L. (1976). Surgery 79, 182-187. Polack, F. M. (1973). Ciba Found. Syrnp. 15 (N.S.), 127-139. Pomerat, C. M., Breckenridge, C. G., and Cordon, L. (1944). E n d o c r i n o l o g y 34,60-68. Poppa, G., Simmons, R. L., David, D. S., and Russell, P. S. (1964). Transplantation 2,

Prehn, R. T. (1974). Clin. Immunobiol. 2, 191-204. Prehn, R. T., and Lapp&, M. A. (1971). Transplant. Reu. 7,2644. Raaf, J. H., Farr, H. S., Laird Myers, W. P., and Good, R. A. (1974). Am. /. Surg. 128,

Raju, S., and Grogan, J. B. (1969). Transplantation 7,475-483. Raju, S., and Grogan, J. (1977). Transplant. Pmc. 9, 1187-1191. Ramseier, H., and Billingham, R. E. (1966).J. E x p . Med. 123,629-656. Reemtsma, K. (1970). Transplant. Proc. 2, 513. Renold, A. E., and Cahill, G. F., eds. (1965). “Handbook of Physiology,” Sect. 5. Am.

232.

New York.

Philadelphia, Pennsylvania.

System.” Saunders, Philadelphia, Pennsylvania.

496-502.

478-483.

Physiol. SOC., Washington, D.C.


Rezzesi, F. D. (1932). Arch. E x p . Zellforsch. Besunders Gewebezuecht. 13, 258. Roddy, H., Putnam, C. W., and Fennell, R . H., Jr. (1976). Trunspluntation 22,625-f30. Rodin, A. E., Larson, D. L., and Roberts, D. K. (1967). Cancer 20, 1772-1779. Roffo, A. H. (1917). Reu. Unio. Buenos Aires 35, 249. Russell, P. S., and Gittes, R. F. (1959)./. E r p . Med. 109, 571-588. Rygaard, J. (1973). “Thymus and Self-Immunobiology of the Mouse Mutant Nude.”

Sakai, A. (1970). Trunspluntation 9, 333-334. Salley, J. J . (1954)./. Dent. Res. 33, 253-262. Salyer, K. E., and Kyger, E. R., 111. (1973). Plast. Reconstr. Surg. 51, 672-681. Scheinberg, L. C., Edelman, F. L., and Levy, W. A. (1964).Arch. Neurol. (Chicago) 11,

Schlesinger, M. (1962).J. Nutl. Cancer Znst. 28,927-946. Schwartz, R. S. (1974). Transplant. Proc. 6 ,4547. Schwartz, R. S. (1975). N . Eng l . J . Med. 293, 181-184. Scott, R., Jr., Mutchnik, D. L., Laskowski, T. Z., and Schmalhorst, W. R. (1969)./. Urol.

Shaffer, C. F., and Hulka, J. F. (1969). Am. /. Obstet. Cynecol. 103, 78-85. Shepro, D., Kula, N., and Halkett, J. A. E. (1963)./. Exp. Med. 117, 749-754. Shirai, Y. (1921)./pn. Med. Wwld 1, 14. Siebert, W. J. (1928). Proc. SOC. E x p . Biol. Med. 26, 236-237. Silberman, S., and Shklar, G. (1963). Oral Surg., Oral Med. 0 7 U l P Q t h O l . 16,1344-1355. Silvers, W. K., Murphy, G., and Poole, T. W. (1977). lmmunogenetics 4, 85-100. Silverstein, A. M., and Khodadoust, A. A. (1973).Cibo Found. Symp. 15 (N.S.), 105-120. Simmons, R. L., and Russell, P. S. (1966). Ann. N.Y. Acad. Sci. 129,3545. Smirnova, E. (1937). Bull. Biol. Med. E r p . URSS 4,6. Smith, M. J. V. (1966). lnuest. Urol. 3, 4 3 9 4 4 . Stark, R. B., Dwyer, E. M., and De Forest, M. (1960).Ann N.Y.Acad. Sci. 87,140-148. Starzl, T. E., Ishikawa, C. W., Putnam, C. W., Porter, K. A., Picache, R., Husberg, B. S.,

Strautz, R. L. (1970). Diabetologia 6, 306-312. Streilein, J . W., and Read, C. B. (1976). /. Theor. Biol. 61, 363376. Streilein, J. W., Grebe, S. C., Kaplan, H. J., and Streilein, J. S. (197Sa). Transplant. Proc.

Streilein. J . W., Stone, M. J., and Duncan, W. R. (1975b).]. Zmmunol. 114, 255-260. Strober, S., and Gowans, J. L. (196.5). J . Exp. Med. 122,347-360. Sturgis, S. H., and Castellanos, H. (1962). Ann. Surg. 156, 367-376. Talmage, D. W., Dart, G., Radovich, J., and Lafferty, K. L. (1975).Science 191,385388. Tihey, N. L. (1971).]. Anut. 109,369. Tilney, N. L., and Ford, W. L. (1974). Cel l Zrnmunol. 14, 146-150. Tilney, N. L., and Gowans, J. L. (1971)./. Exp. Med. 133,951-962. Toolan, H. W. (1954). Cancer Res. 14, &O-666. Triger, D. R. (1976). Gastroenterology 71, 162-164. Triger, D. R., Alp, M. H., and Wright, R. (1972). Lancet 1, 60-62. Triger, D. R., Cynamon, M. H., and Wright, R. (1973). Zmmunology 25,941-950. Tze, W. J., Wong, F. C., Chen, L. M., and O’Young, S. (1976). Nature (London) 264,

Udassin, R., Schlesinger, M., and Ben-Hur, N. (1975). Transplant. Proc. 7, 763-769. Van Dooremaal, J. C. (1873). Albrecht von Graefes Arch. Ophthalmol. 19, 359. Vessella, R. L. Jr., Raju, S., and Grogan, J. B. (1974). Fed. Proc., Fed. Am. SOC. E x p . B i d .

Wiley (Interscience), New York.

248-264.

101,602-607.

Halgrimson, C. G., and Schroter, G. (1974). Transplant. Proc. 6, 129-139.

7,349-354.

466-467.

33,783.


Warden, G. D., Reemtsma, K., and Steinmuller, D. (1973). Transplant. Proc. 5,635-639. Weiner, L. J. (1965). Surg. Fwum 16, 240-241. Whitmore, W. F., 111, and Gittes, R. F. (1975). Surg. Fwum 26, 88-340. Williams, D. E., Evans, D. M. D., and Blarney, R. W. (1971). Br.1. Cancer 25,533-537. Willis, R. A. (1935). Proc. R. S O C . London, Ser. B 117,400-412. Willis, R. A. (1962). “The Borderland of Embryology and Pathology.” Butterworth.

Witte, S., Goldenberg, D. M., and Schricker, K. T. (1965). Klin. Wochenschr. 43, 1182. Witte, S., Goldenberg, D. M., and Schricker, K. T. (1968). 2. Cesanite Exp. Med. 148,

Woodruff, M. F. A. (1960). “The Transplantation of Tissues and Organs.” Thomas,

Woodruff, M. F. A., and Woodruff, H. G. (1950). Philos. Trans. R . Soc. London, Ser. B

Wustrack, K. O., and Lucas, Z. J . (1974). Surg. Forum 25,279-282. Wustrack, K. O., Cmber, R. P., and Lucas, Z. J. (1975). Transplantation 19, 156-165. Yoffey, J. M., and Courtice, F. C. (1970). “Lymphatics, Lymph and the Lymphomyeloid

Zahn, F. W. (1884). Virchow’s Arch. Pathol. Anat. Physiol. 95,369. Ziegler, M. M., Miller, E. E., and Barker, C. F. (1972). Surg. Forum 23, 127-129. Ziegler, M. M., Miller, E. E., and Barker, C. F. (1973). Surg. Forum 24,290-293. Ziegler, M. M., Lopez, V., and Barker, C. F. (1975).]. Surg. Res. 18, 201-207.

London.

72-80.

Springfield, Illinois.

234,559-582.

Complex.” Academic Press, New York.

Major Histocompatibility Complex Restricted Cel I-Mediated Immunity

GENE M. SHEARER AND ANNE-MARIE SCHMITT-VERHULST

Immunology Bmnch, National Cancer Institute, Bofhesdo, Maryland

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

T-Lymphocyte Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Major Histocompatibility Complex Restriction for Distinct

A. T-Cell-Mediated Cytotoxicity B. Delayed Hypersensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Proliferation in Mixed Lymphocyte Reactions . . . . . . . . . . . .

111. Fine Specificity of Cytotoxic Effector Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. H-2-Restricted Specificity . . . . . . . . . . . C. Infecting or Modifying Agent Specificity . . . . . . . . . . . . .

IV. Immune Response Genes for H-2-Restricted Cytotoxicity .

55

57 57 63 64 64 65 66 73 78 83 87

I . Introduction

The major histocompatibility complex (MHC) has been known to be involved in some aspects of immune recognition for more than two decades. The MHC was found to code for cell surface antigens involved in allograft rejection (Gorer et al., 1948; S h r e f i r and David, 1975). This genetic region has also been shown to control immune response potential in a number of animal species (Benacerraf and McDevitt, 1972; McDevitt and Bodmer, 1974). In the mouse, the MHC, known as the H-2 complex, has been divided into four major regions: K, I , S, and D (Shreffler and David, 1975). The K and D regions determine the strong serologically detectable transplantation antigens of the mouse, which are important for tissue graft rejection, the generation of anti-H-2 antibodies, and the target cell antigens for thymus-derived (T-) cell-mediated lympholysis (CML) (Gorer et al., 1948; Alter et al., 1973; Abbasi et al., 1973; Schendel et al., 1973; Nabholz et al., 1974; Brondz et al., 1975; Bevan, 1975a). Recent results obtained using mutant strains of mice in which graft-versus-host and CML reactions can be generated between the mutants and the wild types in the absence of detectable serological differences raise the possibility that the targets of CML may be distinct from the cell surface products detected ssrologically, although both products are coded within the K and D regions of H-2 (Berke and Amos, 1973;

55

56 GENE M. SHEARER AND ANNE-MARIE SCHMITT-VERHULST

Widmer et al., 1973; Forman and Klein, 1975; Nabholz et al., 1975). Cytotoxic effector cells can be generated without inducing strong proliferation by culturing mixtures of cells differing only at the K or D region or both, whereas cell mixtures differing only at the Z region induce strong proliferative responses (MLR) in the absence or in the presence of weaker cytotoxic responses (Widmer et al., 1973; Plate, 1974; Wagner et al., 1975; Hodes et al., 1975; Klein et al., 1977). The S region, which maps between the I and D regions, determines the level of a serum a-globulin known as Ss as well as the sex-linked allotypic variant designated SZp (Shreffler and Passmore, 1971). Products of this region appear to be functionally relevant in the complement system (Lachman et al., 1975; Hansen et aZ., 1975). The I region has been divided into at least five subregions, based on mapping studies using recombinant mouse strains for immune response (Zr) gene functional studies (Lieberman et al., 1972; Melchers et al., 1973; Lozner et al., 1974; Benacerraf and Katz, 1975) as well as for the detection of l a antigens on lymphocyte subpopulations (David et al., 1973; Hauptfeld et al., 1973; Sachs and Cone, 1973; Sachs et al., 1975), and suppressor cell functions (Murphy et al., 1976; Tada et al., 1976). These I subregions are currently designated Z-A, I - B , Z-J, Z-E, and Z-C (Shreffler et al., 1977).

One of the more recent functions associated with the murine MHC has been the finding that efficient CML reactions can be generated in uiuo and in uitro to virally infected or chemically ,modified syngeneic cells only when the target cells are H-2 matched with the infected or modified stimulating or responding cells or both (Zinkernagel and Doherty, 1974a; Doherty and Zinkernagel, 1975; Gardner et al., 1975; Koszinowski and Thomssen, 1975; Shearer, 1974; Shearer et al., 1975a; Forman, 1975; Rehn et al., 1976a). Similar requirements for H-2 homology have been demonstrated for sensitization and lysis of cells expressing weak transplantation antigens (Bevan, 1975b; Gordon et al., 1975), and have been suggested for tumor-associated antigens (Germain et al., 1975; Schrader et al., 19i'S). Products of the K and D regions appear to be involved in the recognition for these self- associated CML reactions, irrespective of which system was used to demonstrate the phenomenon. The recognition of antigenic determinants in association with self MHC products need not be limited to the cytotoxic examples cited above, and may be a more general aspect of T-cell recognition. This could involve products of different MHC regions, depending on the functional class of the T-lymphocyte involved (Rosenthal and Shevach, 1973; Erb and Feldmann, 1975; Miller et al., 1975; Schmitt-Verhulst et al., 1977). In this report the role of the MHC products in the various MHC-restricted systems will be reviewed, and

MAJOR HISTOCOMPATIBILITY COMPLEX 57

the experimental evidence obtained in these models will b e considered as it relates to the nature (Binz and Wigzell, 1975a,b); Eichman and Rajewsky, 1975) and specificity of the T-cell receptor(s) (Janeway et al , , 1976; Rehn et al., 1976b).

it. Major Histocompatibility Complex Restriction for Distinct T-Lymphocyte Functions

Within the syngeneic models, MHC restriction has been demonstrated for a number of T-cell functions. These include: T- cell-mediated cytotoxicity (Zinkernagel and Doherty, 1974a; Shearer, 1974; Doherty and Zinkernagel, 1975; Gardner et al., 1975; Kos- zinowski and Thomssen, 1975; Shearer et al., 1975a; Forman, 1975; Rehn et at., 1976a); memory for such cytotoxic responses (Schmitt- Verhulst et al., 1977); delayed hypersensitivity (Miller et al., 1975, 1976); mixed lymphocyte reactions (Schmitt-Verhulst et al., 1977; Thomas and Shevach, 1977) as well as cellular interactions involving helper cells (Katz et al., 1973, 1975; Sprent and von Boehmer, 1976); macrophages (Rosenthal and Shevach, 1973; Erb and Feldmann, 1975); and suppressor cells (Rich and Rich, 1976), as well as suppression factors (Rich and Rich, 1975; Tada et al., 1976; Moorehead, 1977). In this section a review will be presented of T-cell-mediated cytotoxic responses, delayed hypersensitivity, and mixed lymphocyte reactions involving syngeneic cells and the evidence supporting H-2 restriction.

A. T-CELL-MEDIATED CYTOTOXICITY

Murine T-cell-mediated CML reactions have been generated in viuo and in uitro against alloantigens, usually those involving differences between responding and stimulating cells at H-2 (Cerottini et al,, 1970; Henney, 1971; Wunderlich and Canty, 1970; Wagner et al., 1972; Cerottini and Brunner, 1974; Bevan, 1975a). However, if CML is relevant for autologous reactions such as autoimmunity and/or possibly for surveillance against autologous neoplastic cells or against infections, it is important to investigate the genetic parameters and to elucidate the immune mechanisms involved in reactivity against selflike antigens. Until recently some investigations have been reported in which lymphoid cells responded in graft-vs-host reactions (Cohen et al., 1971), mixed lymphocyte cultures (von Boehmer and Adams, 1973; Ponzio et al., 1975), or cytotoxic reactions against differentiation antigens (Ponzio et al., 1975) or tumor-associated antigens (Wagner and Rollinghoff, 1973; Lundak and Raidt, 1973), all which involve nonallogeneic responses. In such studies the possibility of MHC restriction was not investigated. As was mentioned in the intro-


duction, however, a number of more recent examples of CML reactions have been reported in which H - 2 restriction has been tested and observed between stimulating and target cells, effector and target cells, or among responding lymphocytes, stimulating cells and target cells.

1 . Virally Infected

The first published example in which H - 2 restriction was fully realized as being important for T-cell-mediated cytotoxic reactions was the demonstration that T-lymphocytes from lymphocytic choriomeningitis (LCM)-infected mice could lyse LCM-infected targets only if they expressed the same H - 2 haplotype as the donor of the effector cells (Zinkernagel and Doherty, 1974a). Since the LCM virus is a noncytopathic budding RNA virus, other CML studies were performed using the nonbudding cytopathic DNA ectromelia virus. Mice immunized with ectromelia virus also generated cytotoxic T-lymphocytes (CTL), which could lyse only ectromelia-infected targets expressing the same H - 2 haplotype as the immunized donors (Gardner et al., 1974a, b, 1975). Similar H - 2 restriction for cytotoxic effectors was demonstrated in vaccinia virus-infected mice (Kos- zinowski and Thomssen, 1975; Koszinowski and Ertl, 1975), as well as for Sendai virus (Doherty et al., 1976a). In the examples cited above, the CTL were taken from infected mice. A technique has been recently developed in which H - 2 restricted cytotoxic effector cells can be generated in a primary in vitro culture system against LCM- infected peritoneal cells (Dunlop and Blanden, 1977). Cytotoxic effectors have also been detected by culturing lymph node cells from herpes simplex virus-infected mice (Pfizenmaier et al., 1977). These T-cell effectors were both virus specific and H - 2 restricted on a short- term cytotoxic assay of macrophage targets.

In the tumor virus systems, MHC restriction has been demonstrated in the mouse for cytotoxic effectors against the MSV (Plata et al., 1976; Gomard et aE., 1976). Similarly, mice injected with syngeneic Friend virus-infected tumor cells generated cytotoxic effectors which were restricted to lyse Friend-infected, H - 2 matched, cultured target cells (Blank et al., 1976). The CTL generated during the recovery from Friend disease were also found to be preferentially H - 2 restricted in their killing potential (Cheseboro and Wehrly, 1976). An unusual case of H - 2 restriction involving Friend virus has been recently reported (Kumar and Bennett, 1977a, b). In this system Friend virus (FV) activates a population of T-suppressor cells, which, in turn, can suppress mitogen-responsive cells (Kumar and Bennett, 1976; Kumar et al.,


1976). By using cell-mixing experiments, it was demonstrated that mitogen-responsive cells and the suppressor cells that regulate them must share D region H-2 haplotype for suppression to be effective. Surprisingly, the requirement for H-2-restriction was overcome if the suppressor cells were irradiated in uitro, or if they were harvested from either cortisone-treated or infant donor mice (Kumar and Bennett, 1977a, b). It has been postulated for this complicated example of H-2 restriction that the suppressor-mitogen-responsive cell H-2 restriction requirements are determined by a third late-maturing, cortisone- and radio-sensitive cell (Kumar and Bennett, 1977a, b). MHC restriction for T-cell reactivity in such viral systems seems not to be a phenomenon limited to the mouse, since similar results have been reported for Rous sarcoma virus in chickens (Wainberg et al., 1974).

In some of these viral systems mapping studies have been performed using H-2 recombinant and mutant strains to establish whether the H-2 restriction observed was limited to a particular region of the MHC. The K and D regions appear to be the regions relevant for H - 2 homology in T-cell-mediated cytotoxicity to LCM and vaccinia virus- infected targets (Zinkernagel and Doherty, 1975; Zinkemagel, 1976a). The use of H-2 mutant mice, such as the Hzl (H-2ba) (Bailey et d., 1971), which are presumed to be point mutations involving a single cistron at H - 2 K (Klein, 1975), further supports the concept that products of genes mapping within the K region are involved in the restriction of the CML reaction. N o cross-reactivity was observed between H-2K" and H-2K"" for the LCM and vaccinia cytotoxic responses (Zin- kernagel, 1976a). Some cross-reactivity was detected for LCM and ectromelia between the H - 2 K h h mutant and BlO.A(5R) while the latter were the effectors and the former were the virus-infected targets, but not vice versa (Blanden et al. , 1976a; Doherty et al., 1976a). The H - 2 mutant experiments suggest that the determinants responsible for allogeneic graft rejection and the H-2 component recognized in association with the viral modification are coded by the same genetic unit, which would be distinct from the antigenic structure serologically recognized as a private specificity.

It has also been demonstrated that the generation of these H - 2 restricted cytotoxic effector T-cell reactions can be paralleled by in uiuo effects. For example, fatal lymphocytic choriomeningitis could he transferred to host mice with non Ig-bearing lymphocytes only when the cells shared K or D region haplotypes with the host animals (Doh- erty et al. , 1976a). Furthermore, protection against ectromelia infection (Kees and Blanden, 1976) as well as against LCM infection in the spleen (Zinkernagel and Welsh, 1976) was achieved by adoptive cell


transfer of immune cells when these cells shared K or I) with the recipient mice.

2. Chemically Modijied

Concomitant with the work on the H-2 restriction in the LCM system (Zinkernagel and Doherty, 1974a), an in uitro system was developed in which cytotoxic T-lymphocytes were generated against trinitrobenzene sulfonate-(TNP-)modified autologous cells (Shearer, 1974). These effector cells were able to lyse targets modified with the same agent which expressed the same H-2 haplotype as the responding and modified stimulating cells (Shearer, 1974; Shearer et d., 1975a; Forman, 1975). Other chemically modifying agents that covalently bind to cell surfaces have been used to generate effectors in uitro and have also been demonstrated to exhibit H-2 restriction. These include N-(3-nitro-4-hydroxy-5-iodophenylacetyl-~-alanylglycylglycyl-) azide (N-) (Koren et al., 1975; Rehn et al., 1976a),N-(3-nitro- 4-hydroxy-5-iodophenylacetyl) azide, N-(2,4,6-trinitrophenyl)-P-alanylglycylglycyl azide (Rehn et al., 1976b), fluorescein isothiocyanate (Starzinski-Powitz et al., 1976b), and dinitrobenzene sulfonate (For- man, 1975, A.-M. Gilheany et al., unpublished results). Cytotoxic reactions exhibiting self-restriction have been recently demonstrated using TNP-modified human peripheral blood leukocytes (Shaw, unpublished observations).

Mapping studies for H-2 restriction in the TNP-modified CML system indicated that the cytotoxic effectors efficiently lysed only TNP- modified target cells which expressed the same haplotype as the effector and stimulating cell populations at K (or K-plus Z-A) and/or at D (Sheareret al., 1975a; Forman, 1975). In certain mouse strains, particularly those expressing the k haplotype at K through Z-J, lysis was detected on TNP-modified targets expressing the haplotype of the effectors and stimulating cells at the K end, but not on targets expressing effector haplotype at H-2D (Shearer, 1974; Shearer et d., 1975a; Schmitt-Verhulst and Shearer, 1975). The lack of lysis of TNP- modified H-2D-matched targets by effectors from certain mouse strains that are capable of lysing TNP-modified H-2K-matched targets has been attributed to H-2-linked genetic unresponsiveness in the responding cell population (see Section IV) (Schmitt-Verhulst and Shearer, 1975,1976; Schmitt-Verhulst et al., 1976). Restriction of these cytotoxic effector cells to the K and D regions of H-2 has been independently demonstrated by the blocking of lysis by specific antisera directed against H-2K and H-2D products (Schmitt-Verhulst et al., 1976). Blocking withgnti-Z-region sera in the assay had no detectable


effect on lysis (Schmitt-Verhulst et al., 1976). The use ofTNP-modified BIO.A(4R) targets for lysis by B1O.A or B1O.BR effectors does not distinguish between homology requirements at K or at K plus I-A. The complete inhibition of lysis with BIO.A effectors by A.TL anti A.AL serum (strains that appear to differ only at H-2K) (Schmitt-Verhulst et al., 1976), and the lysis of TNP-modified SJUJ targets by A.TL effectors (which appear to share only H-2K) (Schmitt-Verhulst and Shearer, 1976) indicate that products of the I-A region are not required in the CML specificity at the lytic phase. See Table I for a listing of a number of inbred and congenic resistant mouse strains with their respective H-2 haplotypes. Partial mapping studies have been performed in the N-modified CML system with similar results (Rehn et al., 1976a; Shearer et aZ., 1976).

3. Weak Transplantation Antigen Associated

Cultures of lymphocytes differing only at non H-2-linked minor histocompatibility loci do not elicit cytotoxic T-effector cells. However, CML responses against minor H antigens (Bevan, 1975b,c, 1976) and the sex-linked H-Y antigen (Gordon et al., 1975; Gordon and Simpson, 1976) have been generated using cultured lymphocytes from primed

TABLE I

RESISTANT MOUSE STRAINS DISCUSSED IN THIS REVIEW DESICNA~ON OF H - 2 HAPLOTYPES IN THE INBRED AND CONGENIC

H - 2 region"

Mouse strain K I-A I -B I - ] I-E 1C S D

CBA, B1O.BR A, BIO.A BALBk, B10.D2 C57BW10 BlO.A(4R) BlO.A(2R) B1O.HTT A.TH A.TL A.AL SJUJ BlO.A(3R) B lO.A(5R)

k k k k k k k k k k k k k d d d d d d d d d d d b b b b b b b b k k b b b b b b k k k k k d d b $ S S S k k k d S S S S S S s d S k k k k k k d k k k k k k k d

b b b b k d d d b b b k k d d d

S S S S S S S S

" From Klein (1975); Shrefiier et al. (1977).


donor mice. The responding cell donors differed from the priming and challenging immunogenic cells only at the respective minor loci. Sur- prisingly, lysis was detected only on target cells that expressed not only the same minor antigens but also the same H-2 haplotype as the stimulating cells. Mapping studies for CML against the minor H antigens indicated H-2 restriction for the D region and for one or more regions to the left ofl-C (Bevan, 1975~). In the H-Y system the lysis b y female effectors of male targets in the C57BW10 responder strain was restricted to H-2Db (Simpson and Gordon, 1977). When the responding cell donors were F, hybrids, the H-2 restricted specificities were associated with H-2Db, H-2Kd in (B10 x Balb/c)F1, H-2Dk in (B10 x CBA)F,, and H-2Kk in (BlOxA)F1 (Simpson and Gordon, 1977). Part of this H-2 restriction can be accounted for by immune response genes, which will be considered in Section IV. Similar observations have been recently reported involving HLA-associated cytotoxic responses against human H-Y antigens (Goulmy, et al., 1977). A minor histocompatibility antigenic difference, linked to the ninth chromosome in the region which codes for the Thy 1 antigen, has been described in AKWCu and AKWJ mice, which are both H-2k (Zatz, 1977). Similar to the CML reactions observed for other weak histocompatibility antigens the effectors generated against the Thy l-linked antigen(s) were able to lyse only H-2-matched target cell expressing the relevant weak antigen (Zatz, 1977).

The above findings contrast with another report (Peck et al., 1977), in which H-2 restricted CML was described in completely syngeneic secondary in nitro cultures. In these experiments T-cell-enriched lymphocytes were cultured in the presence of fetal bovine serum (FBS) for 5 days and the resulting blast cells were recultured with fresh irradiated spleen cells syngeneic to the responding cell population. Such secondary cultures generated cytotoxic activity that could be detected only on H-2-matched lipopolysaccharide-stimulated spleen cells, whether these targets had ,been cultured in the presence or in the absence of FBS. The latter experiments rule out the possibility that FBS was responsible for the generation of a new antigen on the cell surface (Forni and Green, 1976). When I-region haplotype was shared between target and effector cells, but not K or D haplotypes, no lysis was detected. Although no obvious new antigen was expressed on both “stimulating” and target cells in this system, B-cell-speci fic surface antigens and virus-associated antigens resulting from the activation of endogeneous viruses during cellular stimulation by FBS are possible candidates for the antigenic determinant recognized by the T-effector cell in association with H-2-coded products. H-2 restricted


syngeneic cytotoxicity had been previously reported in long-term lytic assays of mouse lymphocytes cultured on syngeneic fibroblasts (Ilfeld et al., 1975; Shustik et al., 1976).

4. Tumor Immunity

Cytotoxic effector cells can be generated against syngeneic tumors. Indirect evidence has been provided suggesting that H-2 antigens are recognized in addition to the tumor-associated antigens (Germain et al., 1975; Schrader et al., 1975; Schrader and Edelman, 1976). The lytic phase of murine effector cells generated against H-2-matched lymphoid tumors could be blocked with specific anti-H-2 sera directed against antigens controlled by the H-2 complex of normal cells (Ger- main et al., 1975). Similar observations have been reported by others (Schrader et al., 1975; Schrader and Edelman, 1976). These latter investigators have also demonstrated that patching and capping of H-2 antigens on tumor cells resulted in copatching and cocapping of viral antigens (Schrader et al., 1975). Such findings raise the possibility that there may be recognition of self H-2 components in antitumor T- cell-mediated immunity. It would appear, therefore, that products of the MHC may be involved in the T-cell-mediated aspects of antitumor immunity in a manner similar to the other reported phenomena involving H-2 restricted CML.

B. DELAYED HYPERSENSITIVITY

Delayed hypersensitivity (DH) is a type of immune reaction in which T lymphocytes activate an inflammatory response. Characteri- zation of the cell surface antigens on the lymphocytes responsible for DH indicates that this reaction is mediated by a subpopulation of T lymphocytes distinct from cytotoxic effector and suppressor cells (Vadas et al., 1976). Lymphocytes sensitized against soluble protein antigens transferred DH to naive mice only if the donor and recipient were H-2 matched (Miller et al., 1975). Mapping studies revealed that the H-2 region of homology important for this reaction was I-A when the antigen used for sensitization was fowl y-globulin and K , I , or D when dinitrofluorobenzene (DNFB) was the sensitizer (Miller et al., 1976; Vadas et al., 1977).

Tolerance to delayed hypersensitivity against DNFB can be effected by suppressor T cells, which release a soluble suppressor factor (Cla- man et ul. , 1977; Moorhead, 1977). This suppressor factor abrogates the ability of sensitized lymphocytes to transfer contact sensitivity to naive mice. The action of this suppressor factor is both antigen specific and H-2-restricted, since it affects only lymphocytes sensitized against


the same hapten and H-2-matched with the cells from which the soluble suppressor factor was prepared (Moorhead, 1977).

c. PROLIFERATION IN MIXED LYMPHOCYTE REACTIONS

In a primary in uitro mixed lymphocyte reaction against TNP- modified syngeneic cells, only marginal cellular proliferation was detected (Shearer et al., 1975b). However, a strong proliferative response can be detected against TNP-modified syngeneic cells by restimulat- ing primary cultures with TNP-modified cells syngeneic with the responding and modified primary stimulating cells (Schmitt-Verhulst et al., 1977). It was possible to investigate the specificity of the secondary MLR against TNP-modified syngeneic and allogeneic cells, since primary allogeneic MLC cannot be generated in such secondary cultures (Schmitt-Verhulst et aZ., 1977). Intra-H-2 mapping for the secondary MLR indicated that the relevant regions of homology were Z, D, and K or I< plus Z-A (Schmitt-Verhulst et aZ., 1977).

An alternative approach for investigating the specificity of secondary proliferative responses to TNP-modified syngeneic cells in the absence of a primary allogeneic reaction involved the prior elimination of alloreactive clones with bromodeoxyuridine and light (Thomas and Shevach, 1977). Cultures of guinea pig lymphocytes, made unresponsive to allogeneic macrophages, and subsequently primed with TNP-modified syngeneic macrophages, generated secondary proliferative responses only after restimulation with TNP-modified syngeneic and not modified allogeneic macrophages. If, however, the priming was against TNP-modified allogeneic macrophages, secondary stimulation was obtained only with the modified allogeneic macrophages (Thomas and Shevach, 1977). Since the strains 2 and 13 guinea pigs used in this study are considered to differ only in the Z region (B. D. Schwartz et al., 1976), MHC homology involving an Z region product is required between modified primary and secondary stimulating macrophages for such T-cell recognition. Secondary in uitro proliferative responses have been recently obtained using TNP- modified autologous human peripheral blood leukocytes in which proliferation was optimal when primary and secondary modified stimulating cells were from the same donor (Shaw et al., submitted for publication).

111. Fine Specificity of Cytotoxic Effector Cells

The H-2-restricted cytotoxic response for the virally infected and chemically modified systems provides an opportunity to examine the components of immunological recognition. These include the contri-


butions made to the overall specificity of the effector cells by the agents themselves and by the H-2-coded cell surface products involved.

A. MODELS

Two basic models have been proposed to account for the H-2 homology requirements observed for the virally infected, chemically modified, and weak antigen-H-2-restricted autologous CML (Zinker- nagel and Doherty, 197411; Dohertyet al., 1976b). These models essentially differ in that one involves a single receptor recognizing self H - 2 products which are either modified or unmodified, but are in close association with the infecting or modifying agent, whereas the other involves two distinct receptors-one specific for the agent and the second specific for self H-2K or H-2D coded cell-surface products. In the single receptor model, the agent could be modifying cell-surface antigenic structures (altered self). Alternatively, a model involving a single receptor recognizing unaltered-self H-2 products in close association or proximity with the infecting or modifying agent is possible. It has been suggested that H-2 molecules serve as adapters that com- bine with the infecting or modifying antigenic determinants on the cell surface to form an adaptor-antigen complex which is composed of elements of self H-2 coded products and the antigenic determinant (e.g., viral antigen or TNP) (Schrader et al., 1975). For the two-receptor or dual recognition model, one receptor would be specific for the “hapten” or infecting agent. The second receptor would be a responder or effector cell receptor that would function as a responder-stimulator and/or effector-target interaction structure recognizing syngeneic K or D region products. It has also been suggested that genes mapping within the K and D regions of the H-2 complex control the operation of groups of glycosyltransferases (Blanden et al., 1975). Such enzymes would be responsible for attaching sugar residues to the proper sites on syngeneic polypeptides effecting the formatiori of glycopeptides in which the antigenic complex is determined by both the host and viral genome (Blanden et al., 1975). Thus, MHC genes could dictate glycosylation of viral glycopeptides or virus could modify glycosylation of H-2 antigens. However, it is difficult to account for the H- 2-restricted CML observed for the chemically modified and weak transplantation antigen systems by a glycosylation mechanism. It would be necessary, therefore, to postulate that the glycosylation mechanism would be responsible for the H-2 restriction observed for the virally infected C M L , but not for the other H-2-restricted examples.


B. H-2-mSTRICTED SPECIFIClTY

In most of the H-2-restricted CML models investigated, experiments have been performed using F1 hybrid effector cells and virally infected (Zinkernagel and Doherty, 1974b, 1975; Blank et al., 1976), chemically modified (Shearer et al., 1975a; Forman, 1975; Rehn et al., 1976a), or weak antigen-associated (Bevan, 1975b; Gordon and Simpson, 1976) parental stimulating and target cells as a possible approach for distinguishing between the one- and two-receptor models. Without exception, the F! effectors were observed to lyse the target cells of the parental haplotype used for sensitization, but not the targets from the other parental strain. These observations are compatible with the single-receptor model and inconsistent with the dual- receptor model presented in the simplest form. In the altered-self or single-receptor model, clones of effector cells generated by sensitization of responding lymphocytes against modified parental cells of a given H-2 haplotype would lyse modified K or D region matched targets of the same H-2 haplotype only, since the H-2-coded altered- self structures of stimulating immunogen and target antigen would need to be identical. In contrast, the simplest model of dual recognition would predict that either parental haplotype would serve as target cells if H-2 homology is required only between effector and target cells. The restrictions observed in the Fl hybrid-parent experiments would not necessarily exclude the dual receptor model, however, if the following limitations are imposed: First, it must be assumed that receptors for each parental haplotype are clonally expressed in the F1 hybrid responding lymphocytes and that this expression is subject to allelic exclusion; i.e., a given clone of F1 lymphocytes can interact with only a single parental haplotype, both at the responder-stimulator and at the effector-target cell levels. Second, if the H-2 receptor in the dual recognition is required for cell-cell interaction, then this interaction must occur via complementary, but not between like-like, structures, since a like-like interaction through H-2 antigens would neces- sitate allelic exclusion ofH-2 cell-surface products on the F1 cells. It is known that serologically detectable region H-2 antigens are codomin- antly expressed on F, cell surfaces (Cullen et al., 1972). Distinct clones of cells have been shown to react against K and D regions of allogeneic cells (Nabholz et al., 1974; Bevan, 1975a) and autologous modified cells (Schmitt-Verhulst, et aZ., 1976). Thus, in order for a two-receptor model to be valid, it is necessary to postulate four distinct clones, each expressing a receptor for virus, hapten," or weak antigens, plus a complementary receptor for a single parental K - or D-region product to account for full F1 cytotoxic potential.

"


It should be noted in this context that cross-priming for a secondary response to a minor H antigen has been obtained (Bevan, 1976). Lym- phocytes from F, mice immunized in uiuo against cells expressing a weak H antigenic difference from the host (but sharing one H-2 haplotype), challenged in uitro with cells expressing the same minor H antigen and the H-2 haplotype of the other parent-generated effectors that were specific for the in uitro immunization, both with respect to H-2 and the minor antigen (Bevan, 1976). These results do not prove, but are compatible with, the H-2 restriction being expressed in the lytic event between effector and target cells.

Another approach for investigating the relative contribution of self H-2-coded products and the foreign antigenic determinants (e.g., virus, hapten, or weak antigen) in the recognition between effector and target cells is the use of nonradioactive targets for blocking the interaction between the effector and “Cr-labeled target cells (“cold target inhibition”) (Zinkernagel and Doherty, 1975; Shearer et al., 1975a, 1976; Forman, 1975; Bevan, 1975b). In such experiments unlabeled target cells were incubated with the effector cells prior to the addition of S’Cr-labeled target cells to the effectors. The results indicated that blocking cells which were H-2-matched with the effector and “Cr- labeled target cells inhibited the cytotoxic assay only if they expressed the same antigenic determinant (e.g., virus, hapten, or weak antigen) as that used in sensitization. Furthermore, blocking cells that were not H-2-matched with the effector and targets, but expressed the same antigenic determinant used in sensitization, were likewise ineffective inhibitors of cytolysis (Zinkernagel and Doherty, 1975; Shearer et al., 1975a, 1976; Forman, 1975; Bevan, 1975b). Such results tend to favor a single-receptor model in which both the H-2-coded products and the antigenic determinant contribute to the specificity of the cytotoxic effector cell. A dual recognition model would not necessarily be exclued, however, if it is assumed that the occupation of only one receptor by the relevant H-2-coded product or by the antigenic determinant does not compete efficiently with a target expressing both antigenic components.

The lytic phase of allogeneic CML has been shown to b e blocked b y antisera directed against the K or D region antigens of the target cells used for sensitization (Nabholz et al., 1974). In certain of the H - 2-restricted syngeneic CML examples, the involvement in the lytic interaction between effector and target cells of products closely associated on the cell surface with the private antigenic specificity recognized serologically was suggested by experiments using alloantisera directed against H-2 subregions. Thus, although uninfected or unmodified H-2-matched cells do not serve as cold target inhibitors of


cytolysis in these H-2-restricted systems, anti-H-2 sera directed against the H-2 region products involved in sensitization do inhibit the interaction of effector and target cells. In one syngeneic tumor system, antisera directed against products of the entire H-2 region of the target cell inhibited lysis by effectors of the relevant tumor target cells (Schrader et al., 1975). Experiments using alloantisera of either of the two specificities expressed by F1 effector cells sensitized with semisyngeneic tumor cells suggested that blocking of the lytic phase could be achieved only with anti-H-2 sera directed against the target cells, but not with alloantisera directed against the H-2 specificity of the effector cells (Schrader and Edelman, 1976). A similar conclusion had been obtained for allogeneically stimulated effector cells (Schmitt-Verhulst et al., 1976). In another syngeneic tumor system, partial inhibition of lysis was obtained by blocking with antisera directed against products of the D region (Germain e t al., 1975). Block- ing with anti-H-2 sera has also been demonstrated in the H- 2-restricted CML reactions against the vaccinia (Koszinowski and Ertl, 1975) and MSV (Gomard et al., 1976) viruses.

Further analysis of the blocking ofH-2-restricted CML reactions by anti-H-2 sera has been reported using the TNP-modified antigenic cytotoxic model (Schmitt-Verhulst et al., 1976; Burakoff et d., 1976a). The interaction between effectors sensitized against TNP-modified syngeneic lymphocytes and modified target cells was inhibited only by sera directed against the K region of the target cells, but not by sera directed against Z or D region products in a mouse strain for which the TNP-syngeneic CML response is associated only with the K region of H-2 (Schmitt-Verhulst et id., 1976) (lines 1 through 3 of Table 11). When the lysis by effectors from a different mouse strain which was capable of being sensitized against TNP-self associated with H-2D products was tested on TNP-modified target cells sharing only the I-C, S, D regions ofH-2 with the stimulating cells, inhibition was observed only in the presence of sera directed against I-C, S, D products, but not in the presence of sera directed against products of the K region, or an Z-A region product (Schmitt-Verhulst et al., 1976; and lines 4 and 5 of Table 11). Further evidence for the specificity involved in the blocking of target cells with antisera in the TNP-modified syngeneic CML reaction was indicated by the absence of blocking of the lytic interaction by a strong alloantiserum directed against non-H-2 alloantigens expressed on the TNP-modified tumor cells used as targets as well as by the lack of inhibition by anti-Ia sera on target cells enriched for Za- bearing cells (Burakoff et al., 1976b).

Anti-TNP antibodies have been found to block the lysis by effector cells generated either by a TNP-modified aiitologous or by unmodified


TABLE I1 SELECTIVE INHIBITION BY ANTI-ff-2 SERA AND ANTI-TNP SERUM OF THE LYSIS OF

TNP-MODIFIED TARGET CELLS BY EFFECXOR CELLS SENSITIZED WITH TNP-MODIFIEn AUTOI.OGOUS STIMULATING CELLS

Responding Stimulating Target Antiserum Inhibition of cells cells cells directed against cytolysis"

BIO.A B1O.A-TNP B10.A-TNP K k ++ B1O.A B1O.A-TNP B1O.A-TNP ( I-C ,S ,D)d - B 10.A B 1O.A-TNP B 1O.A-TNP (I-A through S)k - B 10.D2 BlO.D2-TNP B1O.A-TNP ( K through I-E)k - BlO.D2 BlO.D2-TNP B1O.A-TNP ( I-C,S,D)d ++ BlO.D2 BlO.D&TNP BlO.D2-TNP All of H-2d ++ BlO.D2 BlO.D2-TNP BlO.D2-TNP ( K through I-E)d + BlO.D2 BlO.D2-TNP B1O.DS-TNP ( I-C,S,D)d + BlO.D2 BlO.D2-TNP BlO.D2-TNP Ia-8 B1O.BR B1O.BR-TNP B10.BR-TNP TNP-KLH ++ B1O.BR C57BL/lO C57BL/lO-TNP TNP-KLH ++ B1O.BR C57BLlO C57BL/lO TNP-KLH -

-

'' + + indicates complete inhibition of lysis; + indicates partial inhibition of lysis; - indicates no inhibition of lysis.

allogeneic stimulators when assayed on TNP-modified targets (Schmitt-Verhulst et al. , 1976; and lines 11 and 12 of Table 11). These observations indicate that some TNP groups are closely associated with unmodified H-2 specificities, since, as shown in line 12 of Table 11, anti-TNP antibodies could block the lysis of TNP-modified targets by effectors that are specific for unmodified alloantigens. The observation that reactivity against or associated with D-end specificities in the TNP-syngeneic CML is blocked only by antiSI-2D-end reagents, but not by anti-H-2K-end reagents suggests that some TNP groups are more closely associated with K or D region products than the K and D region coded antigens are to each other. The blocking studies with anti-H-2 sera directed against the unmodified H-2K or D region products suggest that the lytic interaction involves recognition of products sterically associated with serologically detected H-2 products. How- ever, the CML target cell antigens could be distinct from serologically detected antigens in both the allogeneic and syngeneic H-%restricted systems as suggested by experiments mentioned earlier using H-2 mutant mouse strains that are serologically indistinguishable from wild type, but can generate CML reactions (see Section I).

H-2 hd, H-2 for the generation of CML against TNP-modified syngeneic cells showed an extensive cross-reactivity between effectors generated in the wild- type syngeneic system and TNP-modified mutant target cells (Forman

The data obtained using the H-2Kh mutants H-2


and Klein, 1976). These results contrast with the data obtained with effector cells from the same mutants sensitized against virally infected syngeneic cells, which were found not to be cross-reactive with the wild-type virally infected target cells (Blanden et al., 1976a; Zinker- nagel, 1976a). Both the cross-reactivity found in the TNP-syngeneic system and the absence of cross-reactivity detected in the virally infected system could be obtained by postulating that a single allele at the H-2K locus may control more than one CML determinant on the same molecule. A mutation in this allele would affect only some of these determinants and leave the others unaltered (Forman and Klein, 1976). This interpretation would also account for the observation that although the mutant and wild-type H-2K differ (they react against one another), they still seem to share some determinants in H-2K that can be detected by the stimulation of a strain of a different H-2K haplotype (Forman and Klein, 1975).

The use of cell lines that do not express detectable amounts of H-2 coded cell surface antigens provides another way of investigating the importance of MHC products in H-2-restricted T-cell-mediated lympholysis. A murine teratoma cell line which expresses a T / t locus antigen but does not express detectable amounts of H-2 antigens is not lysed as a TNP-modified target of effector cells generated against TNP-modified syngeneic cells (Forman and Vitetta, 1975). The murine teratoma F9 cell line, which does not express H-2 antigens (Artzt et al., 1975), can be infected with LCM virus resulting in the expression of cell-surface viral antigens (Zinkernagel and Oldstone, 1976). Al- though LCM-infected F9 cells are susceptible to antiviral antibody and complement, they are not lysed by T-cell-mediated cytotoxic effector cells (Zinkernagel and Oldstone, 1976). These results indicate that H-2-coded cell surface products are important in the recognition by T-effector cells leading to lysis of targets. It is unclear from these experiments whether the requirement for H-2 expression is important for LCM-infected target cells only, or for both infected stimulating and target cells.

In order to critically investigate whether the H-2 restriction observed in these syngeneic CML systems is due to H-2 homology (a) between effector and target cells, (b) between stimulator and target cells, or (c) between responder and stimulator as well as between effector and target cells, a series of experiments have been performed in which the potential responding cell pool was made unresponsive. Lymphocytes for this purpose have been prepared by at least three techniques.

First, chimeras in which lethally irradiated FI hybrid mice were repopulated by T-cell-depleted parental stem cells resulted in re-


population of the Fr host mice with parental immunocompetent cells that were selectively unresponsive to the other parental alloantigens (von Boehmer and Sprent, 1976). Effector cells from such chimeras infected with LCM (Pfizenmaier et al., 1976; Zinkernagel, 197613) or vaccinia (Zinkernagel, 1976b) virus lysed infected target cells expressing the alloantigens to which the chimera was unresponsive. Simi- larly, parental lymphocytes that expressed only one H - 2 parental haplotype obtained from F1 recipients, and were unresponsive to the H - 2 antigens of the other parent, could be sensitized against TNP-modified cells to which the lymphocytes were tolerant (von Boehmer and Haas, 1976; Pfizenmaier et al., 1976). The observation that parental lymphocytes that differentiated in the Fr host environment are unresponsive to the relevant alloantigens but are responsive to the virally infected or TNP-modified alloantigens could be interpreted as favoring the altered-self model, since such results would be consistent with the CML being directed against new antigenic determinants created by interaction of self H - 2 products with either viral antigens or TNP. However, these experiments do not formally exclude dual recognition involving a self-recognition structure, since it is possible that such a self-recognition structure could have developed during the differentiation of the parental cells in the F1 host, which could have been associated with development of the unresponsive state in the chimera (Zinkernagel, 1976b).

Another technique that has been recently employed to deplete lymphoid populations of specific alloreactive cells is that of negative selection (Sprent and Wilson, 1977). In this technique, T-cells from strain A mice were injected into irradiated (A x B)FI hybrid recipients, and then recovered 1 day later from the thoracic duct lymph. Such recovered cells were almost entirely of donor origin and were unresponsive to the host alloantigens. Although such lymphocytes were incapable of generating CML against the alloantigens of strain B, they did generate CML responses against TNP-modified cells of strain B (Sprent and Wilson, 1977). These data indicate that the reaction against TNP-modified allogeneic cells does not require the presence of lymphocytes responsive to the same alloantigens. Such results raise the possibility that chemical modification of cell surface proteins can result in the formation of new cell-surface antigens, and that distinct lymphocyte populations exist that can respond to these new antigens. This short-term procedure is considered to involve only cell filtration, not the induction of a tolerant state which avoids the possible compli- cation in interpreting the chimera experiments (discussed above).

A third approach for obtaining a population of cells specifically depleted of alloreactive cells consisted in the selective in vitro elimina-


tion of clones of cells proliferating in response to allogeneic cells (Schmitt-Verhulst, 1977). It was found that when the clones of B1O.BR cells reacting against B10.D2 alloantigens were eliminated from cultured spleen cells after bromodeoxyuridine (BUdR) incorporation and subsequent exposure to light, such cultures were unable to generate cytotoxic effector cells against TNP-modified B 10.D2 allogeneic cells. The same cultures could, however, be sensitized by TNP-modified cells syngeneic to the responder cells. These results indicate that stimulation with BlO.D2 alloantigens in the primary culture activates clones that recognize TNP-modified as well as unmodified B10.D2 stimulators in the secondary cultures. These results do not support the hypothesis that there are separate clones of lymphocytes that are stimulated by modified alloantigens (B1O.DB-TNP) that would not be activated by the unmodified B 10.D2 alloantigens. Similar results have been obtained independently (C. A. Janeway, personal communication). At least two interpretations are possible from these results. First, the clones of cells recognizing TNP-modified alloantigens are among the cells that are sensitized by (and probably bear a receptor for) the alloantigen (favoring a dual-receptor model). Second, the possibility can be raised that a particular class of cells is required for the recognition of “altered self” and that such cells are not generated in primary in uitro CML against TNP-modified alloantigens. The discrepancy between these in uitro data and the results obtained from the in uiuo negative selection experiments (Sprent and Wilson, 1977) outlined above could be due to an artifact inherent to the in uitro culture conditions. Alternatively, this discrepancy could result from a more drastic elimination of reactive cells in uitro, which would include clones of cells of low affinity for the alloantigen that would not have been selected against by the in uiuo depletion, and would not be detected as alloreactive cells, but could interact with the alloantigen in association with a new cell-surface antigen [see two-receptor model of Jane- way et al. (1976)l.

It should be noted that a result different from that obtained for elimination of alloreactive clones for mouse CML reactions using BUdR and light has recently been obtained in the guinea pig for MLR against TNP-modified alloantigens (Thomas and Shevach, 1977). Using two inbred strains of guinea pig that diiler at the I-region (B. D. Schwartzet al., 1976), these investigators found that lymphocytes from which a particular population of alloreactive cells had been eliminated by BUdR and light were still capable of proliferating in response to in uitro priming and boosting by TNP-modified allogeneic macrophages expressing the relevant alloantigens (Thomas and


Shevach, 1977). The latter results involve a strong proliferative population of T lymphocytes distinct from the cytotoxic T-effector cells induced in the previously described CML experiments (Schmitt- Verhulst, 1977). Furthermore, the MHC restriction in the guinea pig MLR system is associated with la alloantigens, not with the analogs of the mouse K and D serologically detected antigens. Therefore, at present no strict contradiction can be claimed between the two types of results. However, one would like to speculate that if MHC-coded products are involved in the cellular interactions leading to differentiated immunocompetent T cells, a common underlying mechanism might be involved irrespective of the T-cell function considered.

c. INFECTING OR MODIFYING AGENT SPECIFICITY

The viruses extensively investigated in the H-2-restricted CML systems have belonged to three general classes, which do not exhibit immunological cross reactivity. Vaccinia and ectromelia which do exhibit CML cross reactivity are both pox viruses. The LCM virus, an arenavirus, and Sendai, a parinfluenza virus, have not exhibited any CML cross-reactivity between them nor with the pox viruses (Doherty et al., 1976a). Recent studies in which mice were immunized with type A influenza viruses, which are serologically distinct, have resulted in the generation of cytotoxic effectors cells which, although H-%restricted, exhibit considerable viral cross-reactivity (Doherty el a/., 1977). This discrepancy between cross-reactivity at the T cytotoxic effector cell level and lack of serological cross-reactivity raises the possibility that the effectors and antibodies are specific for different viral antigenic components. Such differences could account for the observations that virus-specific antibodies did not block viral T-cell lysis (Doherty et al., 1976a; Blanden et al., 1976a) whereas anti-TNP antibodies did inhibit cytotoxicity in the TNP-modified CML (Schmitt-Verhulst et al., 1976; Burakoff et al., 1976a).

One of the potential advantages of the in uith chemically modified syngeneic CML response is that this system can be utilized to investigate the fine specificity of the “haptenic” moiety. No cross-reactivity was detected between effectors generated against TNP-modified stimulating cells when assayed on .i’Cr-labeled N-modified target cells and vice versa (Rehn et al., 1976a). Furthermore, effectors generated by sensitization against TNP-modified syngeneic cells did not lyse H-2-matched target cells modified by the TNP moiety separated from the cell surface by a /3-alanylglycylglycyl tripeptide (Rehn et aZ., 1976b). These results not only demonstrate the high degree of fine specificity contributed b y the “haptenic” or modifying agent, but they

74 GENE M. SHEARGR AND ANNE-MARIE SCHMITT-VERHULST

also tend to argue against dual recognition involving recognition of the TNP moiety and self H-2 by two independent receptors on the T-effector cell (Rehn el al., 1976b). A similar degree of specificity has been reported for proliferative responses against cells modified with a series of nitrophenyl compounds (Janeway, 1976).

It should also be noted that T-effector cells generated against TNP-modified syngeneic stimulating cells do not lyse H-2-matched DNP-modified target cells and vice versa (Forman, 1976; A,-M. Gilheany, et al., unpublished observations). Furthermore, skin paint- ing with dinitrofluorobenzene sensitizes mice for in viuo delayed skin reactions and for in vitro proliferative responses to the dinitrophenyl (DNP) but not to the TNP group (Claman et al., 1977). Thus, a variety of T-cell subpopulations appear to be very specific with respect to the antigens they recognize. It should be noted, however, that early B-cell precursors of antibody-producing cells are also able to distinguish between DNP and TNP (Metcalf and Klinman, 1976), although antibodies against these two haptens are cross reactive (Eisen et al., 1970).

The fine specificity of the TNP-modified syngeneic C M L system has been further investigated by TNP-modification of cells either with trinitrobenzene sulfonate (TNBS), which covalently modifies cell surface proteins, or with TNP-stearolyl-dextran (TSD), an amphipathic molecule, which can be inserted into the lipid bilayer of the cell membrane and binds to cells by noncovalent forces (Henkart et al., 1977). Quantitatively equivalent amounts of TNP were detected on the surface of cells modified by both reagents as measured by fluoresceinated anti-TNP antibody. These cell preparations were compared for their ability to: (a) sensitize syngeneic splenic lymphocytes leading to the generation of cytotoxic effector cells; (b) serve as lysable targets in a 4-hour "'Cr-release assay for effector cells generated in (a); and (c) act as blocking cells in the lysis of TNBS-modified targets lysed by TNP-self effector cells generated in (a). In none of these three experimental systems did TSD-modified syngeneic spleen or H-&matched tumor cells act either as a sensitizing immunogen or as a target antigen. In contrast, TNBS-modified spleen cells sensitized syngeneic lymphocytes to generate effectors against TNBS-modified syngeneic targets. Furthermore, TNBS-modified, H-2-matched cells served as specific lysable targets and as inhibiting cells for such effectors. These results indicate that the manner in which TNP is associated with the cell surface is important in the immunogenicity and antigenicity of hapten-modified syngeneic stimulating cells in generating H- 2-associated CML reactions. These findings raise the possibility that a


covalent or at least a stable linkage with cell surface proteins (possibly H-2-controlled products) is important for immunological function. Fur- thermore, these observations do not favor the dual receptor model for H-2-restricted syngeneic CML if it is assumed in such a model that one receptor is specific for the TNP moiety exclusively, and the second for unmodified-self major histocompatibility products.

Although H-2-restricted specificity has been demonstrated in the CML against TNP-modified syngeneic cells, examples have been published in which some degree of non-H-2-restricted lysis was observed (Shearer et al., 1975a; Burakoff et al., 1976b). Cytotoxic effectors generated against TNP-modified syngeneic cells can lyse TNP- modified non-H-2-matched lymphoid tumor target cells and mitogen- stimulated blasts although the level of lysis is not as high as with TNP-modified H-2-matched targets (Shearer et al., 1975a; Burakoff et al., 197613). These observations could be interpreted as evidence for effector specificity directed: (a) exclusively against the hapten (i.e., TNP group) (Dennert and Hatlen, 1975); (b) against TNP-modified non-H-2-coded cell surface components; (c) against cross-reactive determinants resulting from TNP-modification of different K or D region products (Burakoff et d., 1976b); or (d) against TNP-modified K or D region products, although accompanied by a nonspecific lytic interaction between effector and TNP-modified target (Schmitt-Verhulst and Shearer, 1975; Shearer et al., 1976). It is unlikely that the cross- reactivity can be accounted for by (a), since TNP-modified chicken erythrocytes did not block the cross reactivity when the effectors were generated by sensitization with TNP-modified syngeneic cells (Burakoff et al., 197613). Possibility (b) is unlikely based on the observation that the cross-reactive lysis can be blocked by antisera directed against the H-2 haplotype of the target cells (Burakoff et al., 1976b), although it cannot be excluded that H-2 coded products would be involved even in a nonspecific effector-target cell interaction. At this time results are not available to distinguish between possibilities (c) and (d).

Other unexpected cytotoxic reactions have been detected in which a variety of effectors have been reported to lyse TNP-modified target cells syngeneic with the effectors (Schmitt-Verhulst and Shearer, 1975; Burakoff et al., 1976b; Starzinski-Powitz et al., 1976a). First, effector cells generated by sensitization against H-2 alloantigens are capable of lysing TNP-modified targets syngeneic with the effectors (Schmitt- Verhulst and Shearer, 1975; Lemonnier et al., 1977). Cytotoxic effector cells generated in virally infected mice were also found to lyse not only syngeneic target cells infected with the relevant virus but also


TNP-modified syngeneic cells (Starzinski-Powitz et al., 1976a). The extent of lysis detected on the TNP-syngeneic cells was proportional to the lytic activity detected on the relevant virally infected targets. Furthermore, H - 2 restriction was observed for the lysis by effectors from virally infected mice detected on TNP-modified targets (Starzinski-Powitz et uZ., 1976a). The findings that alloreactive cells and cells reacting against virally infected cells are capable of lysing TNP-modified syngeneic cells have been interpreted as indicating that shared CML specificities exist between TNP-modified K and D region products, allogeneic K and D region products, and syngeneic virally modified K and D gene products (Lemonnier e t aZ., 1977; Starzinski-Powitz et al., 1976a). However, it should be noted that effector cells generated by sensitization with TNP-modified syngeneic stimulators have not been found to lyse unmodified allogeneic target cells of any H - 2 haplotype tested (Schmitt-Verhulst and Shearer, 1975; A.-M. Schmitt-Verhulst and G. M. Shearer, unpublished observations). Furthermore, effector cells generated against TNP-modified syngeneic stimulators were not capable of lysing any of the virally infected target cells tested (Starzinski-Powitz et al., 1976a).

Were these “nonspecific’’ effects due to cross-reactive determinants resulting from TNP-modification of K andfor D region products, one would have expected that the cross-reactivity would be demonstrable in both directions, i.e., that TNP-modified syngeneic cells would generate effectors capable of lysing unmodified allogeneic and virally infected targets. It could also be argued that if, on the one hand, TNP modification of K and D region products of one haplotype generates antigenic determinants that are cross-reactive with TNP-modified K and D gene products of a different haplotype (Lemonnier et al., 1977), and, on the other hand, virally infected targets present determinants shared with TNP-modified cells of the same H - 2 haplotype, one might expect cross-reactivity to occur between determinants of one haplotype and TNP-modified H - 2 products of another haplotype. This prediction is contradicted by the observation that effectors generated from LCM-infected mice are H - 2 restricted in their ability to lyse TNP-modified target cells. The possibility that clones of effector cells generated by sensitization against TNP-modified syngeneic cells are cross-reactive with the antigens expressed on unmodified allogeneic cells has been tested by experiments in which the TNP-modified self- reactive clones of cytotoxic precursors were eliminated by treatment with BUdR and light during primary sensitization.

A summary of results typical of such experiments is presented in Table 111. Exposure of B1O.BR responding cells to BUdR and light


TABLE 111 SELECTIVE ELIMINATION OF CLONES OF BlO.BR RESPONDING CELLS AGAINST EITHER

TNP-MODIFIED SELF OR ALLOANTIGENS

% Specific lysis 2 SE on tumor target”

RDM-4- RDM-4 TNP e14

Elimination of clones Restimulation Section reactive with with (H-2’) (H-2’-TNP) (H-2”)

A B1O.BR-TNP (H-2k-TNP) B1O.BR (H-2’) 10.01 10.01 10.01 B1O.BR-TNP 0.1 f 0.8 5.2 2 0.8 0.4 ? 0.8 C57BL110 (H-2b) 5.3 f 0.6 28.8 2 0.5 65.3 ? 1.3 C57BUlO-TNP 9.2 If: 2.2 39.3 2 2.2 65.6 ? 2.1

B B1O.BR-TNP (H-2’-TNP)” B1O.BR (H-2’) L0.01 10.01 10.01 B1O.BR-TNP 1.1 ? 0.6 72.1 2 1.4 5.3 & 0.9 C57BU10 (H-2’) 6.0 f 0.8 38.0 5 1.0 67.9 f 2.3 C57BL/lO-TNP 8.3 2 0.9 42.9 2 0.5 66.5 ? 2.3

RDM-4- RDM-4 TNP P815

(H-2’) (H-2’-TNP) (H-2d)

C B10.D2 (H-2d) B1O.BR (H-2’) 10.01 10.01 10.01

D B1O.BR (H-2”) B1O.BR (H-2’) ro.01 10.01 10.01

BlO.BR-TNP -0.9 2 1.2 23.8 f 1.4 NT BlO.D2 (H-2d) 2.5 f 1.0 N T 2.02 1.0

BlO.BR-TNP 0.8 f 0.6 27.2 f 2.7 NT B10.D2 (H-2d) 8.2 ? 0.5 NT 44.2? 1.3

“ Primary culture which had not been treated with BUdR and light. * Percentage of specific lysis was obtained by subtracting % lysis by unsensitized cells

(less than 10%) from total % lysis; effector to target ratio was 20: 1 in Sections A and B, 40: 1 in Sections C and D. ‘ NT. not tested.

during primary sensitization against B1O.BR-TNP resulted in the failure to generate cytotoxic effectors which lyse RDM-4-TNP targets after restimulation with B1O.BR-TNP (see Section A, Table 111). In contrast, restimulation of the same BUdR and light-treated cultures with C57BW10 (or C57BWlGTNP) allogeneic cells generated a normal CML response against H - 2 b as detected on the EL-4 targets, as well as a normal “cross-reactive” CML detected on RDM-4-TNP targets (compare with data of Section B, Table 111). Conversely, elimi-


nation of B1O.BR clones reactive against B10.D2 alloantigens in cultures treated with BUdR and light was achieved, as shown by the lack of restimulation with BlO.D2 without affecting the ability of the same cultures to be stimulated by B1O.BR-TNP as detected on RDM-4-TNP targets (Sections C and D, Table 111). If the reaction of allogeneically sensitized effectors, which is detected on TNP-modified cells syngeneic with these effectors, were due to the stimulation by these alloantigens of clones of cells cross-reacting with the clones of cells generated by TNP-modified syngeneic cells, one would have expected that preelimination of clones reactive against TNP-modified syngeneic cells would have abolished this cross-reactive lysis. This, however, was not observed, and suggests that lysis detected on TNP- modified syngeneic cells by effectors sensitized against alloantigens is effected by clones of cells that are not cross-reactive with those clones stimulated by TNP-modified syngeneic cells. The results summarized in Sections C and D of Table 111 show that elimination of all clones of cells reactive against one set of “cross-reactive” B 10.D2 alloantigens (Schmitt-Verhulst and Shearer, 1975; Lemonnier et al., 1977) does not affect stimulation with TNP-modified syngeneic cells. This also suggests lack of extensive cross-reactivity between clones of cells specific for alloantigens and for TNP-self. A similar argument can be deduced from the specificities required to restimulate effector cell precursors in secondary in vitro CML reactions (Schmitt-Verhulst e t . al., 1977). These results obtained in secondary CML cultures do not support the hypothesis that the lack of absolute specificity in these CML reactions can be accounted for by clones of T-lymphocytes which recognize altered-self antigens and are cross-reactive with certain alloantigens. More extensive research will be required in this area to test the hypothesis that the repertory of T-cell specificities is generated in the thymus during ontogeny (Jerne, 1971) by encountering modified MHC products (Lemonnier et al., 1977).

IV. Immune Response Genes for H-2-Restricted Cytotoxicity

It was observed in the initial studies of the CML response against TNP-modified syngeneic cells that effector cells from certain inbred mouse strains lysed TNP-modified targets expressing either the K or D H - 2 haplotype of the responding and modified stimulating cells, whereas other mouse strains generated effectors capable of lysing TNP-modified targets expressing the K haplotype, but not those expressing the D haplotype of the responder and stimulating cells (Shearer, 1974). Mice that exhibited these differences in D-


end-restricted CML differed in the H-2 haplotypes in the left part of the H-2 complex (which included K and a portion of the Z region), although they expressed the same haplotype at H-2D. When the lytic potential of effectors generated by BIO.A and B10.D2 responding cells were compared, the B10.D2 effectors were observed to lyse TNP- modified target cells matched at either the K - orD-end with the cells of the sensitizing phase, whereas B1O.A effectors did not lyse TNP- modified D-end matched targets although they did lyse TNP-modified K-end matched targets (Shearer et al., 1975a; Schmitt-Verhulst and Shearer, 1975; Schmitt-Verhulst et al., 1976).

These strain-dependent preferential response patterns to TNP- modified syngeneic cells restricted at the Dregion appear to be regulated by H-2-linked immune response (Zr) genes, and the ability to respond seems to be a dominant trait, since (BIO.A x B10.D2)F1 mice are high responders to TNP-H-2Dd (Schmitt-Verhulst and Shearer, 1975). That responding cells from the BIO.A strain are defective in their response potential to “NP-H-2Dd has been verified by CML inhibition experiments in whlch region-specific anti-H-2 sera were used to block the interaction between effector cells and TNP-modified syngeneic target cells (Schmitt-Verhulst et al., 1976). The results indicated that BlO.D2 anti-B1O.BR sera (directed against H-2Kk) severely reduced the lysis by BIO.A effectors, whereas B1O.BR anti-BlO.D2 sera (directed against H-2D d, had no effect on the lysis of B1O.A-TNP targets by BIO.A effectors. In contrast, the B1O.BR anti-BlO.D2 sera (directed against H-2Dd) abolished the lysis of B 10.A-TNP targets by B10.D2 effectors (Schmitt-Verhulst et al., 1976). These results indicate that the TNP-self cultures from B10.D2 donors generated effectors specific for H-2Dd-TNP, whereas cultures from BIO.A donors did not generate effector cells specific for H-2Dd-TNP.

Mapping studies for the Zr control of effector cells specific for H-2Dd-TNP have been performed using recombinant mice on the C57BL/10 and A strain backgrounds (Schmitt-Verhulst and Shearer, 1976). As summarized in Table IV, a higher level of responsiveness was found when the s haplotype was expressed to the left of the Z-E subregion as compared with the k haplotype. The A.TH and B1O.HTT were the highest CML responders to H-2Dd-TNP, whereas BIO.A and A.AL were the poorest responders to H-2Dd-TNP, although all four of these strains were high CML responders to their respective TNP- modified H-2K restricted specificities. The A.TL strain, which expresses the high responders haplotype at K and the low responder k haplotype throughout the Z region was an intermediate responder to H-2Dd-TNP, but a high responder to its respective H-2Ks-TNP re-

80 GENE M. SHEARER AND ANNE-MARIE SCHMITT-VEFWULST

TABLE IV I n Vitro CYTOTOXIC RESPONSES TO TNP-MODIFIED A~TOLOGOUS CELLS IN DIFFERENT

INBRED MOUSE STRAINS ON THE C57BL110 AND A GENETIC BACKGROUNDS"

Responding Stimulating H - 2 haplotype at: TNP-modified % Specific Range spleen cells spleen cells K A B J E C S D target cells" lysis 2 SE of lysis

BIO.A B 10.A-TN P k k k k k d d d H - 2 n H - 2 H - 2 k

B1O.HTT B1O.HTT-TNp s s s s k k k d H - 2 d H - 2 % H - 2 k

A.AL A.AL-TNP k k k k k k k d H - 2 d H - 2 = H - 2 '

A.TL A.TL-TNP s k k k k k k d H - 2 d

H-2' A.TH A.TH-TNP s s s s s s s d H - 2 d

H-2'

H-2" H-2'

2.2 2 1.1 1-10 5.2 k 1.0 2-16

30.6 & 1.7 2 2 4 6 38.0 & 1.4 38-53 49.9 * 2.6 50-55 27.3 2 2.1

8.2 * 1.6 2-8 3.1 2 1.9 0-6

32.5 2 4.7 23-33

45.6 2 0.9 10-46 11.1 2 2.5 0-11 33.3 2 1.4 33-44 51.8 2 0.6 1 6 5 2 11.4-C 1.7 5-11

19.9 2 1.1 20-23

" Effector: target ratio = 40 : 1 . I' H - 2 d , H-2': and H-2' target cells were 48-hour phytohemagglutinin-stimulated blast spleen

cells from B10.D2, SJL/J or BlO.S, and B1O.BR donors, respectively.

stricted specificity (Schmitt-Verhulst and Shearer, 1976). These results are compatible with the regulation of CML responsiveness specific for H-2Dd-TNP being under the control of at least two genes, one mapping to the left of the crossover between K and Z-A (in the A.TL strain)-possibly in the K region-and the second mapping inside the Z-A through I-] subregions.

The finding that at least one of the genes controlling responsiveness to H-2Dd-TNP maps to the left of I-A is the first published example of an H-2-linked Zr gene which appears outside of a known Zsubregion. This raises the possibilities that: (a) Zr genes might be randomly distributed so that not all of them would map in the Z region; (b) some Zr genes that are associated with the Z-A subregion are situated to the left of the crossover that occurred in the A.TL strain as a result of this particular recombinant event having actually taken place within Z-A ; or (c) that functionally distinct Zr genes map in different regions of the MHC.

It is noteworthy that this is the only mapped Zr gene which controls the generation of cytotoxic T-effector cells (Schmitt-Verhulst and Shearer, 1975, 1976; Schmitt-Verhulst et al., 1976). Heretofore, Zr


genes mapped within the Z region controlled antibody production (Benacerraf and Katz, 1975), possibly at the level of the interaction between T-helper cells and B cells and/or macrophages. Zr control of in vitro secondary T-cell proliferation in response to soluble antigens also maps within the I region (R. S. Schwartz et al., 1976). The presence of products coded for by the Z-A subregion on soluble factors mediating T-B cooperation (Taussig et al., 1975), and their functional involvement in T-cell proliferation in response to soluble antigens as assessed by antiserum blocking (R. S. Schwartz et al., 1976), suggests that products of genes mapping in the Z region are involved in the cellular interactions required to generate those differentiated lymphoid functions. The differentiation from precursor cells into cytotoxic effector cells could involve different types of cell-surface interaction structures, some ofwhich could be coded for by genes mapping within H-2 in a region distinct from I . It may be significant that this Ir gene for cytotoxic function appears to map within an H-2 region that controls those cell surface products which are associated with the antigens recognized for cytotoxic function.

A leukemogenic model involving Friend virus has been described in which both H-2-restricted antigenic structures and H-2-linked Zr genes appear to be involved in the generation of a cytotoxic response (Blank et al., 1976). BALB.B mice immunized with a syngeneic leukemic cell line (HFUb) generated cytotoxic effector cells which lysed H F L h targets, but not the HFWk leukemic line, derived from BALB.K mice, which suggested H-2 restriction. In contrast, BALB.K mice immunized with syngeneic leukemic HFWk cells did not generate effectors detectable on either the H-2-matched HFWk or allogeneic HFWb targets (Blank et nl., 1976). However, (BALB.B x BALB.K)FI mice did generate effectors that lysed HFWk targets after immunization with HFWk, indicating the HFL/k cells are immunogenic and that the FI hybrid between the congenic responder and nonresponder is a responder (Blank et al., 1976). Based on the other mapping studies for recovery from Friend disease (Cheseboro et al., 1974), it might be predicted (if the Friend virus H-2-restricted cytotoxic response is important for recovery) that the H-2 specificity relevant in the cytotoxic response will be H-2Dh. Intra-H-2 mapping of the Ir gene(s) controlling response to leukemic HFL/k has not been reported.

A recent study of the cytotoxic effector cells generated by F, hybrids between AKWJ and a number of H-2 congenic strains against an inoculated AKWJ tumor cell line showed that H-2q’k and H-2”’ hybrids failed to respond, whereas H-2’” hybrids did (Meruelo and McDevitt, 1977; Meruelo, 1977). From the differential CML responsiveness ex-


hibited by (AKR x B 10.A[5R])F1 hybrid mice, the locus conferring the ability to generate CML effectors against the AKWJ tumor line could be mapped in either the Z-J or Z-E subregions of the H-2 complex. The finding that H-2 homozygous but not heterozygous animals responded suggested either that responsiveness was determined by a recessively expressed locus, or that a dominantly expressed locus controls suppression of the response (Meruelo and McDevitt, 1977). Another study has been reported in which differential CML responsiveness and tumor susceptibility has been detected against an AKWJ carcinoma cell line in (AKIUJ x C57BL/6)F* and (AKWJ X DBAI2)FI hybrids (Levy et al., 1976). In these experiments, however, a mechanism other than H-2-linked Zr control of response potential appears to be involved (Levy et uZ., 1976). Zn vivo studies indicated H-2-linked strain differences in the ability

of female mice to reject skin grafts from syngeneic males (Gasser and Silvers, 1972). The in vitro generation of an H-&restricted CML by effector cells from female mice stimulated by cells from syngeneic males was found to follow the same H-2-linked pattern of responsiveness (Gordon et aZ., 1975). In other words, females of H-2b haplotype were responders regardless of the non-H-2 background, whereas females of the H-2k and H-2d haplotypes were nonresponders (Gordon et al., 1975). The specificity of the CML effectors generated by H-2b females against cells from syngeneic males was found to be associated with the D region of the H-2 complex. It was recently found (Simpson and Gordon, 1977) that the Fl hybrid between two nonresponder strains (H-2k’d) could generate an in vitro CML against cells from male H-2k’k parents. The H-2 restriction of this CML response was found to be associated with either Kk or Ilk. These results could be interpreted by the complementation between one or more immune-response genes from each nonresponder parental strain, which would confer responsiveness of females to H-Y antigens (Simpson and Gordon, 1977). However, since the specificity of the CML effector cells not only involves the H-Y antigen, but is also restricted by products coded by the K and/or D regions of the H-2 complex, such a phenotypic complementation need not necessarily result from a genotypic complementation. Indeed it is possible that immune response genes contributed by one of the parents (BALB/C, H-Zd) allow recognition of the H-Y antigen in association with products coded by the H-2 locus corresponding to the haplotype of the other parent (CBA, H-2k) despite the fact that these immune-response genes did not allow responsiveness to H-Y association with H-2d.


The mechanism(s) by which these l r genes influence the generation of cytotoxic effector cells is unknown. Since it appears that interacting cell types are involved in the generation of effectors at least for the autologous TNP-modified CML (Hodes et al., 1975), it is possible that these genes exert their influences on distinct cell populations and/or their interactions. The cellular expression of such genetic parameters could be exerted on: (a) a population of cytotoxic precursor cells; (b) on helper cells or cells that selectively augment an H-2D- or H - 2K-restricted response; and/or (c) on suppressor cells that could regulate the level of cytotoxicity generated.

V. Conclusions and Speculation

The published examples for which H-2-restricted CML reactions have been repeatedly observed can be classed into three broad categories, i.e., virally infected, chemically modified, and weak transplantation antigen-associated. The immunogenetic similarities exhibited b y these three systems suggest that they may share certain common mechanistic features as well. However, before making such a conclusion it may be worth discussing the different ways in which the respective antigenic determinants (e.g., virus, hapten, or minor antigen) could be presented in association with H-2-coded products on the cell surface.

In the H-2-restricted viral systems, it is possible that certain viral antigenic determinants could have a selective physical association with H-2-determined antigens on the cell surface. Compatible with this hypothesis is the observation that the vesticular stomatitis virus exhibited H-2 antigenic activity common with the H-2 haplotype of the cells that were infected (Hecht and Summers, 1972). Furthermore, in murine leukemogenic studies viral antigens and private H-2 serological specificities have been observed to modulate simultaneously, suggesting some type of physical interaction between H - 2-coded and viral antigens (Lilly, 1972). It has also been postulated with some supporting evidence that viruses can induce changes in cell surface H-2-coded antigens, which could be accounted for by a mechanism allowing derepression of genes controlling expression of H-2 products corresponding to foreign haplotypes (Garrido et al., 1976). This latter mechanism would appear not to be responsible for the H-2-restricted syngeneic CML considered here, since the H-2 specificities of infected cells appeared to be unchanged and could not account for the specificity observed in these systems (Doherty et al.,

84 GENE M. SHEARER AND ANNE-MARIE SCHMlTT-VERHULST

1976a, b). In this context it should be noted that cytotoxic effector cells against Sendai virus have been generated in vitro from spleen cells of mice immunized against UV-inactivated Sendai (Schrader and Edleman, 1977). These effectors were restricted to lyse H-2-matched tumor cell targets externally coated with inactivated Sendai virus by short-term incubation. The results suggest that viral components can interact with murine cell surface structures in the absence of viral replication and protein synthesis (Schrader and Edelman, 1977).

In the chemically modified system, a random modification of the free amino groups of cell surface proteins (Okuyama and Satake, 1960) results in a nonselective “haptenation” of those proteins including the serologically detectable H-2K and H-2D coded antigens (Forman et al., 1977), which represent only a small portion of the cell surface proteins. An altered-self hypothesis where the actual new antigen consists of modified K or D region products is readily conceivable in such a system (Shearer, 1974), but does not account for the preferential H-2 association of the recognition. In contrast, the concept of physical modification ofH-%coded products is less likely for the system involving the H-2-associated recognition of weak transplantation antigens. More probable is the possibility that the minor transplantation antigens which elicit T-lymphocyte responses are in some way closely associated on the cell surface with the relevant major histocompatibility products.

In the generation of cytotoxic effector cells, the involvement of H- 2-coded products could be important at sensitization and/or during the lytic phase of the CML. An absolute requirement for H-2 homology between effector and target cells or between responding and stimulating cells does not appear to be necessary. This has been shown b y the ability of effector cells from a lymphoid pool tolerized to an alloantigen to generate effectors capable of being sensitized against infected or modified cells expressing the H-2 haplotype of the tolerogen (Pfizenmaier et al., 1976; von Boehmer and Haas, 1976; Zinkernagel, 1976b). These results could be interpreted as indicating that different clones of cells recognize these alloantigens and the new antigens resulting from the interaction of virus or TNP with these alloantigens. This suggests that receptors recognizing alloantigens and antigenic entities associated with these alloantigens are distinct. However, these experiments do not necessarily exclude the possibility that the H-2 homology requirement could function through a complementary interaction structure that might be acquired during the tolerization against the alloantigen, by a process similar to the natural acquisition during ontogeny of tolerance to self H-2-coded products.


Two other experimental approaches involving the selective induction of unresponsiveness have been employed to investigate this problem. The in uiuo results obtained with negative selection of alloreactive clones indicate that sequestration of clones reactive with a given allogeneic haplotype abrogates that allogeneic response without affecting the CML response against that particular TNP-modified alloantigen (Sprent and Wilson, 1977). In contrast, the in uitro elimination from cultured lymphocytes of clones proliferating in response to an alloantigen did not allow sensitization of CML effectors against the same alloantigen modified with TNP (Schmitt-Verhulst, 1977; C. A. Janeway, personal communication). The selective in uiuo sequestration of alloreactive clones in thoracic duct lymph (Sprent and Wilson, 1977) indicates that distinct clones of responding cells recognized alloantigen and TNP-modified alloantigen. However, in order to account for the in uitro elimination of clones reactive against alloantigen and TNP-modified alloantigen (Schmitt-Verhulst, 1977), it might be necessary to postulate affinity diflerences between clones of lymphocytes recognizing alloantigens and modified alloantigens. More explicitly, the in uiuo sequestration of alloreactive cells might filter out only alloreactive lymphocytes of high affinity, leaving behind the lymphocyte clones that can be activated only through the recognition of an additional antigen by a second receptor (dual recognition model). It would have to be presumed that both types of ,clones would be eliminated under the conditions of optimal in uitro allogeneic stimulation. It is worth noting in this context that the generation of an allogeneic and TNP-modified syngeneic CML are dependent upon distinct subpopulation of lymphocytes as determined b y Ly type (Cantor, 1977). This also raises the possibility that the alloreactive cells and cells reacting against modified allogeneic cells would belong to different subpopulations of T-lymphocytes.

One advantage that the chemically modified CML model has over the viral and weak transplantation antigen models is the potential for investigating the antigenic contribution made by the modifying agent to the fine specificity of the T-cell recognition. It would be expected from the dual-recognition model (in its simplest form) that some dissociation could be made between the receptors for the TNP “hapten” and those recognizing unaltered self. In the chemically modified CML system experiments have been designed and performed to test the model by introducing slight changes in the presentation of the “hap ten” on the syngeneic cell surface (Rehn et al., 1976b; Henkart et al., 1977). Cells modified by TNP separated from the cell surface by a tripeptide did not act as targets for effectors generated by sensitization


against TNP-modified cells (Rehn et al., 1976b). Furthermore, cells modified with TNP-stearoyldextran presenting quantitatively similar amounts of TNP groups on the cell surface as TNBS-modified cells did not serve as targets for effectors generated by sensitization with TNBS-modified syngeneic cells (Henkart et al., 1977). If T-cell recognition occurs via two independent receptors-one recognizing the TNP group and the other recognizing self H-2-coded products, either or both of the targets modified by these slight variations in TNP presentation on the cell surface should have been recognized by effector cells specific for TNP-self. These findings are more compatible with recognition by a single receptor that would recognize either “hapten”-modified H-2 products or “hapten” in close association with unaltered-self H-2 products. These results can be most readily interpreted either by the altered-self model or by an “altered” dual- recognition model-since if two receptors do exist, they appear not to be independent from one another. For the TNP-self system the “altered” dual-recognition model could be valid if it assumed that the receptor for hapten recognizes the haptenic moiety plus a portion of the adjacent amino acids.

The chemically modified self CML provides another advantage for investigating the possibility that TNP-modified soluble proteins could be antigenic when presented in association with self H-2 products. Hypothetically, such a system could operate through the presentation of TNP groups by syngeneic macrophages, if sensitization and lysis were to occur through dual recognition. Presumably, this system would involve the recognition of the TNP groups by the T lymphocytes, and the TNP-bearing macrophages would specifically interact with the relevant T lymphocytes via the self H-2-coded products. Fur- thermore, the development of the TNP-modified system for proliferative and cytotoxic responses with human lymphocytes could be useful for elucidating the role of MHC restriction and Zr genes in man (Shaw, unpublished observations). Although the in uitro chemically modified syngeneic system offers some advantages for manipulation and dissection of cell-mediated immune processes, TNP is not a pathogenic agent (Doherty, personal communication). In contrast, certain of the viral systems investigated are pathogenic in man and could be more relevant for elucidation of immune processes potentially relevant for d’ isease.

The final answer to the one-receptor or two-receptor question will probably come from characterization of the T-cell receptor. If it is assumed that all receptors possess idiotypic determinants, recognition should be blocked by either of the two idiotypes (Binz and Wigzell, 1975a; Janeway et al., 1976; Doherty et al., 1976b).


The fact that some of the H-2-restricted syngeneic CML models are under the control of H-2-linked Ir genes as well as being restricted with K and D region products, suggests multiple functional roles for H-2-coded gene products in cell-mediated immunity. If the immune phenomena reviewed here are relevant for natural immunity, MHC products may be simultaneously important for recognition by T lymphocytes (under the control of Ir genes) and for the H-2-restricted structures which they recognize (the K , I , or D region products associated with the infectious agent).

ACKNOWLEDGMENTS We are grateful to Dr. William D. Terry, who has supported our research efforts. We

thank Mrs. Marilyn Schoenfelder for preparation of the manuscript.

REFERENCES Abbasi, K., Demant, P., Festenstein, H., Holmes, J., Huber, B., and Rychlikova, M.

Alter, B. J., Schendel, D. J., Bach, M. L., Bach, F. H., Klein, J., and StimpHing, J. H.

Artzt, K., Bennett, D., Boyse, E. A., and Jacob, F. (1975). Proc. Nutl . Acud. Sci. U.S.A. 72,

Bailey, D. W., Snell, G. D., and Cherry, M. (1971). I n “Imrnunogenetics of the H-2

Benacerraf, B., and Katz, D. H. (1975).Adu. Cancer Res. 21, 121. Benacerraf, B., and McDevitt, H. 0. (1972). Science 175,273. Berke, G., and Amos, D. B. (1973). Nature (London), New Biol. 242,237. Bevan, M. J . (1975a). /. Zrnrnunol. 114,316. Bevan, M. J. (1975b). Nature (London) 256,419. Bevan, M. J. (1975c).J. E x p . Med. 142, 1349. Bevan, M. J. (1976).]. E x p . Med. 143, 1283. Binz, H., and Wiyzell, H. (1975a)./. E x p . Med. 142, 1218. Binz, H., and Wigzell, H. (1975b). Scand. J . Immunol. 4, 591. Blanden, R. V., Dunlop, M. B. C., Doherty, P. C., Kohn, H. I., and McKenzie, I . F. C.

Blanden, R. V., Hapel, A. J., and Jackson, D. C. (1976b). Zmrnunochemistry 13, 179. Blank, K. J., Freedman, H. A., and Lilly, F. (1976). Nuture (London) 260, 2.50. Brondz, B. D., Egorov, I. K., and Drizlikh, G. I. (1975).J. E x p . Med. 141, 11. Burakoff, S. J., Germain, R. N., Dorf, M. E.. and Benacerraf, B. (1976a). Proc. Natl. Acud.

Burakoff, S. J., Gemiain, R. N., and Benacerraf, B. (1976b)./. E x p . Med. 144, 1609. Cantor, H. (1977). Cold Spring Harbor Symp. Quunt. Biol. 41, 23. Cerottini, J.-C., and Bninner, K. (1974). Adu. Zrnrnunol. 18, 67. Cerottini, J.-C., Nordin, A. A., and Bninner, K. T. (1970). Nature (London) 227, 72. Cheseboro, B., and Wehrly, K. (1976)./. E x p . Med. 143,85. Cheseboro, B., Wehrly, K., and StimpHing, J. (1974). /. E x p . Med. 140, 1457. Claman, H. N., Miller, S. D., and Moorhead, J . W. (1977). Cold Spring Harbor Symp.

Cohen, I. R., Globerson, A,, and Feldman, M. (1971).J. E x p . Med. 133, 834.

(1973). Trunsplunt. Proc. 5, 1329.

(1973).]. E x p . Med. 137, 1303.

3215.

System” (A. Lengerova and M. Vojtiskova, eds.), p. 155. Karger, Basel.

(1976a). IninitriioCloicitic.., 3, 541.

Sci. U.S.A. 73, 625.

Quunt. B i d . 41, 105.


Cullen, S. E., Schwartz, B. D., Nathenson, S. G., and Cherry, M. (1972). Proc. NQtl.

Dennert, G., and Halten, L. E. (1975). J . Immunol. 114, 1705. Doherty, P. C., and Zinkernagel, R. M. (1975).J. E x p . Med. 141,502. Doherty, P. C., Blanden, R. V., and Zinkernagel, R. M. (1976a). Transplant. Rev. 29,89. Doherty, P. C., Gotze, D., Trinchieri, G., and Zinkemagel, R. M. (1976b). Zmmunogenet-

Doherty, P. C., Effors, R. B., and Bennink, J. (1977). Proc. Natl. Acad. Sci. USA. 74,1209. Dunlop, M. B. C., and Blanden, R. V. (1977). Cell. Immunol. 28, 190. Eichman, K., and Rajewsky, K. (1975). Eur. /. Zmmunol. 5,661. Eisen, H. N., Michaelides, M. C., Underdown, B. J., Schulenburg, E. P., and Simms,

Erb, P., and Feldmann, M. (1975).J. E x p . Med. 142,460. Ford, W. L., and Atkins, R. C. (1971). Nature (London), New Biol. 234, 178. Forman, J. (1975).J. E x p . Med. 142,403. Forman, J. (1976). Transplant. Reu. 29, 146. Forman, J., and Klein, J. (1975). lmmunogenetics 1,469. Forman, I., and Klein, J. (1976). lmmunogenetics 4, 183. Forman, J., Vitetta, E. S., and Hart, D. A. (1977). J . l m m u d . 118, 803. Forni, G., and Green, I. (1976).]. Immunol. 116, 1561. Gardner, I., Bowern, N. A., and Blanden, R. V. (1974a). Eur.1. Zmmunol. 4,63. Gardner, I., Bowern, N. A., and Blanden, R. V. (1974b). Eur. J. Zmmunol. 4,68. Cardner, I. D., Bowern, N. A., and Blanden, R. V. (1975). Eur. J.Zmmunol.\ 5, 122. Ganido, F., Schinmacher, V., and Festenstein, H. (1976). Nature (London) 259,228. Gasser, D. L., and Silvers, W . K. (1972).Ado. lmmunol. 15,215. Gennain, R. N., Dorf, M. E., and Benacerraf, B. (1975)./. E x p . Med. 142, 1023. Gomard, E. V., Duprez, V., Henin, Y., and Levy, J. P. (1976). Nature (London) 260,707. Gordon, R. D., and Simpson, E. (1976). Proc. Leukocyte Cult. Conf., loth, 1975 p. 521. Cordon, R. D., Simpson, E., and Samelson, L. E. (1975).J. E x p . Med. 142, 1108. Corer, P. A., Lyman, S., and Snell, G. D. (1948). Proc. R. SOC. London, Ser. B 135,499. Goulmy, E., Termijtelen, A., Bradley, B. A., and van Rood, J. J. (1977).Naatre 266,544. Hansen, T. H., Shin, H. S . , and Shreffler, D. C. (1975).]. Exp . Med. 141, 1216. Hauptfeld, V., Klein, D., and Klein, J. (1973). Science 181, 167. Hecht, T. T., and Summers, D. F. (1972). J . Virol. 10, 578. Henkart, P. A., Schmitt-Verhulst, h.-M., and Shearer, G. M. (1977). J . E x p . Med. 146,

Henney, C. S. (1971). /. Zmmunol. 107, 1558. Hodes, R. J., Hathcock, K. S., and Shearer, G. M . (1975).J. E x p . Med. 115, 1122. Hodes, R. J., Schmitt-Verhulst, A.-M., Hathcock, K. S . , and Shearer,G. M. (1976). Scand.

Ilfeld, D., Carnaud, C., and Klein, E. (1975). lmmunogenetics 2,231. Janeway, C. A,, Jr. (1976). Transplant. Reo. 29, 164. Janeway, C. A,, Jr., Wigzell, H., and Binz, H. (1976). Scand. J. Zmmunol. 5,993. Jerne, N. K. (1971). Eur. J . lmmunol. 1, 1. Katz, D. H., Hamoaka, T., and Benacerraf, B . (1973). J . Exp. Med. 137, 1405. Katz, D. H., Graves, M. G., Dorf, M. E., di Muzio, H., and Benacerraf, B. (1975)./. E x p .

Kees, U., and Blanden, R. V. (1976). J. E x p . Med. 143,450. Klein, J. (1975). “Biology of the Mouse Histocompatibility-2 Complex.” Springer-

Klein, J., Chiang, C. L., and Hauptfeld, V. (1977)./. Exp. Med. 145,450.

Acad. Sci. U S A . 69, 1394.

ics 3, 517.

E. S. (1970). Fed. Proc., Fed. Am. SOC. E x p . Biol. 26,78.

1068.

J . lmmunol. 5,369.

Med. 141, 263.

Verlag, Berlin and New York.


Koren, H. S., Wunderlich, J. R., and Inman, J. K. (1975). Transp lan t . Proc. 7, Suppl. 1,

Koszinowski, U., and E d , H. (1975). Nature (London) 255,552. Koszinowski, U., and Thomssen, R. (1975). E u r . J . lmmunol. 5,245. Kumar, V., and Bennett, M. (1976).J. E x p . Med. 143, 713. Kumar, V., and Bennett, M. (1977a). Fed. Proc., Fed. Am. SOC. E x p . B i d . 36, 1202. Kumar, V., and Bennett, M . (197%). Nature (London) (in press). Kumar, V., Caruso, T., and Bennett, M. (1976)l. E x p . Med. 143, 728. Lachmann, P. J., Grennan, D., Martin, A., and Demant, P. (1975).Nature (London) 258,

Lemmonier, F., Burakoff, S. J., Germain, R. N., and Benacerraf, B. (1977). Proc. Natl.

Levy, R. B., Waksal, S. D., and Shearer, G. M. (1976).J. E x p . Med. 144, 1363. Lieberman, R., Paul, W. E., Humphrey, W., and Stimpfling, J . H. (1972). J . E x p . Med.

Lilly, F. (1972). In “FWA Viruses and Host Genome in Oncogenesis” (P. Emmelotet al.,

Lozner, E. C., Sachs, D. H., and Shearer, G. M. (1974)J. E x p . Med. 139, 1204. Lundak, R. L., and Raidt, D. L. (1973). Cell. Zmmunol. 9, 60. McDevitt, H. 0.. and Bodmer, W. F. (1974). Lancet 1, 1269. Melchers, I . , Rajewsky, K., and ShrefRer, D. C. (1973). Eur. J . lmmunol. 3, 754. Meruelo, D. (1977). In “Immunobiology of Bone Marrow Transplants” (G. Cudkowicz,

Meruelo, D., and McDevitt, H. 0. (1977). Fed. Proc., Fed. Am. SOC. E x p . Biol. 36,1202. Metcalf, E. S., and Klinman, N. R. (1976).1. E x p . Med. 143, 1327. Miller, J. F. A. P., Vadas, M. A., Whitelaw, A., and Gamble, J. (1975). Proc. Natl . Acad.

Miller, J . F. A. P., Vadas, M. A., Whitelaw, A,, and Gamble, J. (1976). Proc. Natl . Acad.

Moorhead, J. W. (1977).J. Zmmunol. 119,315. Murphy, D. B., Herzenberg, L. A., Okamura, K., Herzenberg, L. A., and McDevitt, H. 0.

Nabholz, M., Vives, J. , Young, H. M., Meo, T., Miggiano, V., Rynbeck, A., and ShrefAer,

Nabholz, M., Vives, J., Young, H. M., Meo, T., Miggiano, V., Rynbeck, A., and Shreffler,

Okuyama, T., and Satake, K. (1960).]. Biochem. (Tokyo) 47,454. Peck, A., Andersson, L. C., and Wigzell, H. (1977).J. E x p . Med. 145, 802. Pfizenmaier, K., Starzinski-Powitz, A., Rodt. H., Rollinghoff, M., and Wagner, H. (1976).

Pfizenmaier, K., Starzinski-Powitz, A., Rollinghoff, M., Falke, D., and Wagner, H.

Plata, F., Jongeneel, V., Cerottini, J.-C., and Brunner, K. T. (1976). Eur. J . Zmmunol. 6,

Plate, J. (1974).J. E x p . Med. 139, 851. Ponzio, N., Finke, J . H., and Battisto, J. R. (1975).J. immunoi. 114,971. Rehn, T. C., Shearer, C . M., Koren, H. S., and Inman, J. K. (1976a)J. E x p . Med. 143,127. Rehn, T. G., Inman, J. K., and Shearer, G. M. (1976b).J. E x p . Med. 144, 1134. Rich, S. S., and Rich, R. R. (1975).J. E x p . Med. 142, 1391. Rich, S. S., and Rich, R. R. (1976).J. E x p . Med. 143,672. Rosenthal, A. S., and Shevach, E. M. (1973).J. E x p . Med. 138, 1289.

169.

242.

Acad. Sci . U.S.A. 74, 1229.

136, 1231.

eds.), p. 229. North-Holland Publ., Amsterdam.

M. Landy, and G. M. Shearer, eds.). Academic Press, New York (in press).

S c i . U.S.A. 72, 5095.

S c i . U.S.A. 73, 2486.

(1976).J. E x p . Med. 144, 699.

D. C. (1974). E u r . J . Zmmunol. 4, 378.

D. C. (1975). Zmmunogenetics 1,457.

J . E x p . Med. 143, 999.

(1977). Nature (Londnn) 265, 632.

832.


Sachs, D. H., and Cone, J. L. (1973).J. E x p . Med. 139, 1289. Sachs, D. H., David, C. S., Shreffler, D. C., Nathenson, S. G., and McDevitt, H. 0.

Schendel, D. J . , Alter, B. J., and Bach, F. H. (1973). Transplant. Proc. 5, 1651. Schmitt-Verhulst, A.-M. (1977). Zn “Immunobiology of Bone Marrow Transplants” (C.

Cudkowicz, M. Landy, and G. M. Shearer, eds.). Academic Press, New York (in press).

(1975). Zmmunogenetics 2,301.

Schmitt-Verhulst, A.-M., and Shearer, G. M. (1975).J. Exp. Med. 142,914. Schmitt-Verhulst, A.-M., and Shearer, G. M. (1976).J. Erp. Med. 144, 1701. Schmitt-Verhulst, A.-M., Sachs, D. H., and Shearer, G. M. (1976)J. E x p . Med. 143,211. Schmitt-Verhulst, A.-M., Garbarino, C. A., and Shearer, G. M. (1977).J. Zmmunol. 118,

Schrader, J. W., and Edelman, G. M. (1976). J . E x p . Med. 143,601. Schrader, J . W., and Edelman, G. M. (1977).J. E x p . Med. 145,523. Schrader, J. W., Cunningham, B. A., and Edelman, G..M. (1975). Proc. Natl. Acad. Sci.

1420.

U S A . 72, 5066. Schwartz, B. D., Kask, A. M., Paul, W. E., and Shevach, E. M. (1976).J. Erp. Med. 143,

541. Schwartz, R. S., David, C. S., Sachs, D. H., and Paul, W. E. (1976)J. Zmmunol. 117,531. Shearer, G . M. (1974). Eur. J . Zmmunol. 4, 527. Shearer, G. M., Rehn, T. G., and Garbarino, C. A. (1975a).J. E x p . Med. 141, 1348. Shearer, G. M., Lozner, E. C., Rehn, T. G:, and Schmitt-Verhulst. A.-M. (1975b).J. E x p .

Shearer, G. M., Rehn, T. G., and Schmitt-Verhulst, A.-M. (1976). Transplant. Reo. 24,

Shreffler, D. C., and David, C. S. (1975). Adu. Zmmunol. 20, 125. ShrefRer, D. C., and Passmore, H. C. (1971). “Immunogenetics of the H-2 System.”

ShrefRer, D. C., David, C. S., Cullen, S. E., Frelinger, J. A., and Niederhuber, J. E.

Shustik, C., Cohen, I. R., Schwartz, R. S., Latham-Griffin, E., and Waksal, S. D. (1976).

Simpson, E., and Gordon, R. D. (1977). Transplant. Reu. 35,59. Sprent, J., and von Boehmer, H. (1976).J. E x p . Med. 144,617. Sprent, J., and Wilson, D. B. (1977). In “Immunology of Bone Marrow Tmnsplants” (G.

Cudkowicz, M. Landy, and G. M. Shearer, eds.). Academic Press, New York (in press).

Starzinski-Powitz, A., Pfizenmaier, K., Koszinowski, U., Rollinghoff, M., and Wagner, H. (1976a). Eur. J . Zmmunol. 6, 630.

Starzinski-Powitz, A., Pfkenmaier, K., Rollinghoff, M., and Wagner, H. (1976b). Eur. J . Immunol. 6, 799.

Tada, T., Taniguichi, M., and David, C. S. (1976).J. E r p . Med. 144, 713. Taussig, M. J., Munro, A. J., Campbell, R., David, C. S., and Stainer, N. A. S. (1975).J.

Thomas, D. W., and Shevach, E. M. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 2104. Vadas, M. A., Miller, J. F. A. P., McKenzie, I. F. C., Chism, S. E., Sherr, F. W., Boyse,

Vadas, M. A., Miller, J . F. A. P., Whitelaw, A. M., and Gamble, J . R. (1977). Zmmuno-

von Boehmer, H., and Adams, P. (1973).J. Zmmunol. 110,376.

Med. 141,930.

222.

Karger, Basel.

(1977). Cold Spring Harbor Symp. Quunt. Biol. 41,477.

Nature (London) 263, 699.

E x p . Med. 142, 694.

E. A., Gamble, J . R., and Whitelaw, A. M. (1976).j. Erp. Med. 144, 10.

genetics 4, 137.


von Boehmer, H., and Haas, W. (1976). Nature (London) 261, 141. von Boehmer, H., and Sprent, J . (1976). Transplant. Reo. 29,3. Wagner, H., and Riillinghoff, M. (1973). Nature (London), New Biol. 241, 53. Wagner, H., Harris, A. W., and Feldman, M. J . (1972). Cell. Immunol. 4,39. Wagner, H., Gotze, D., Ptschelinzew, L., and Riillinghoff, M. (1975).j. E x p . Med. 142,

Wainberg, M. A., Markson, Y., Weiss, D. W., and Doljanski, F. (1974). Proc. Natl. Acad.

Widmer, M. B., Alter, B. J., Bach, F. H., and Bach, M. L. (1973). Nature (London), New

Wunderlich, J . R., and Canty, T. G. (1970). Nature (London) 228, 63. Zatz, M. (1977). Submitted for publication. Zinkernagel, R. M. (1976a).J. Erp. Med. 143,437. Zinkernagel, R. M. (1976b). Nature (London) 261, 139. Zinkernagel, R. M., and Doherty, P. C. (1974a). Nature (London) 248, 701. Zinkernagel, R. M., and Doherty, P. C. (1974b). Nature [London) 251,547. Zinkernagel, R. M., and Doherty, P. C. (1975). I . Erp. Med. 141, 1427. Zinkernagel, R. M., and Oldstone, M. B. (1976). Proc. Natl. Acad. Sci. U . S A . 73,3666. Zinkernagel, R. M., and Welsh, R. M. (1976).J. Zmmunol. 117, 1495.

1477.

Sci. U.S.A. 71, 3565.

Biol. 242, 239.

I. Introduction

Laboratory stocks of Ruttus noruegicus descended from wild animals that only a few hundred years ago were found exclusively in various parts of Asia. In the first half of the eighteenth century, these animals entered western Europe from the east by way of the Norwe- gian peninsula and were subsequently named after the route by which they had arrived. Large numbers of Norway rats were observed crossing the Volga in the Russian province of Astrakhan in 1727, and they appeared in England by 1728-1730 (Donaldson, 1924).

About 600 years before this migration, the black rat (Rattus ruttus) had arrived in western Europe, where it was responsible for spreading

'This review is dedicated to the memory of Dr. Joy Palm.

en

Current Status of Rat Immunogenetics'

DAVID 1. GASSER

Department of Humon Genetics, Unive&y of Pennsylvania School of Medcine, Philodelphio, Penmylvonia

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

A. Definition of Alleles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 B. Ia-Like Antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . C . Genetic Organization ............ ............................. D. Immunity to MHC Antigens ....................................... 105

- 111. Lymphocyte Alloantigens . . . . . . . . . . . . . .......................... 107 IV. Other Blood Group and tibility Polymorphisms . . . . . . . . . . 112

A. The Ag-C Locus . . . . .................................... 112

. . . . . . . . . . 114

VI. Genetics of the Immune Response . . . . A . MHC-Linked Ir Genes ........................... . . . . . . . . . . . 117

C. Other Genes Regulating Immune Reactions . . . . . . . . . . . . . . . . . . . 123 B. Experimental Aller

. . . . . . . . . . . . . . . . . . 125 B. Alpha-Chain Allo C. Allotypes of the IgC,, Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

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

References . . . . . . . . . . . . . ................................... 134

94 DAVID L. GASSER

the great plagues (Donaldson, 1924), including the plague in London in 1664-1665 (Castle, 1947). Since the Norway rat is the more pug- nacious and powerful of the two species, it became the mortal enemy of the black rat, and within a short time had almost completely exter- minated and supplanted it. The Norway rat apparently reached North America by about 1775, where it once again displaced the black rat, which had preceded it.

Credit for isolating the first mutant goes not to perceptive biologists, but to the perpetrators of the popular sport of rat-baiting. Spectators would bet on the time required for trained terriers to kill their last victim after being placed in a pit with 100-200 recently trapped wild rats. Records exist which indicate that albinos were removed from such collections of rats and retained for breeding and exhibition (Rich- ter, 1954). White rats of European origin became the breeding nucleus for an albino stock at the Wistar Institute in Philadelphia. In the early 19OOs, various investigators began inbreeding these rats as well as a number of other stocks of unknown origin, so that eventually more than 100 inbred lines were developed (Palm and Black, 1971; Festing and Staats, 1973).

Genetic studies on rats have closely paralleled similar studies in mice, but the question is not yet resolvkd as to how closely related these two species might be. Rats have a diploid complement of 42 chromosomes (Committee for a Standardized Karyotype of Rattus noruegicus, 1973), compared to 40 for the mouse. When trypsin- banded karyotypes of mouse and rat chromosomes were compared, the banding patterns were identical for about 40% of the genome (Nesbitt, 1974).

As a result of comparing albumins studied by microcomplement fixation, Sarich (1972) argued that rats and mice are as unrelated as Old and New World monkeys, which he believes to have diverged 3 5 4 0 million years ago. Others believe that rats and mice diverged 10 million years ago, but that a greater degree of DNA divergence has occurred than one might expect because of the short generation time of rats and mice (Laird et aZ., 1969; Kohne, 1970).

The first report in rat immunogenetics was probably that of Fried- berger and Taslakow (1928), who observed four alloagglutinins in the sera of wild rats and two examples of alloantibodies among 2000 cross-tests with domesticated rats.

The next contribution was that of Burhoe (1947), who discovered two erythrocyte agglutinogens, which he called A and M. Some rats supplied to Burhoe by Dr. H. W. Feldman of the University of Michi- gan possessed naturally occurring antibodies that agglutinated the

CURRENT STATUS OF RAT IMMUNOGENETICS 95

erythrocytes of all other rats that he tested. By analogy with the human ABO system, Burhoe named the antigen A, but Castle suggested that Ag would be a more appropriate designation, in order to avoid confusion with the agouti (A) gene. This antigen is now known as Ag-A, and although Burhoe's observations were repeated by Owen ( 1962), none of the inbred strains available today are known to lack the antigen and produce the antibody that defines it (Palm and Black, 1971).

The second antigen discovered by Burhoe was named M, and intensive immunizations were required to develop M antisera (Burhoe, 1947). The animals and reagents that defined the M system have been lost, although an alloantigen designated E, which was described by Owen (1962), may correspond to Burhoe's M antigen.

The number of histocompatibility loci possessed by rats has been studied by the use of the FZ test developed by Little (1914; Snell et al., 1976). Billingham et al. (1962) demonstrated that the Lewis and BN strains differ at a minimum of 14 histocompatibility loci. Ramseier and .Palm (1967) showed that Lewis and DA differ by at least 6 or 7 factors, while Zeiss (1967) obtained an estimate of 8-14 loci using several other strains.

There is reason to believe that all these estimates are far below the actual number of histocompatibility loci. It was first shown in mice that weak incompatibilities frequently escape detection unless very tiny skin grafts are used (Hildemann, 1970). This principle is just as important in rats as in mice (Silvers et al., 1976). Furthermore, the estimated number of histocompatibility loci in mice is as high as 44 when a mating system that increases the proportion of donor genes in the recipients is used, and grafts are observed for as long as 240 days (Brambilla et al . , 1970). The recent identification of the 38th autosomal H locus in mice (Snell et al., 1976) indicates that this estimate is not unreasonable. In fact, because of the weakness of many loci and the occurrence of nests of closely linked loci, loci that d o not segregate in a given cross, and loci whose end products are not represented on skin (Snell et al . , 1976), it has been suggested that the true number of H loci in mice is at least several times as high as the estimate of Bram- billa and colleagues. Since there is no reason to believe that rats are different from mice in these respects, it becomes apparent that all the histocompatibility loci described in this review represent but a small fraction of the complete picture.

Nomenclature problems have plagued rat immunogenetics almost from the time of its inception. A committee' has recently been formed

'This committee consists ofT. J. Gill, 111 (Chairman), B. D. Amos, D. L. Gasser, J. C. Howard, J . Klein, and 0. Stark.

TABLE I DISTRIBUTION OF IMMUNOGENETIC MARKERS AMONG VARIOUS INBRED %RUINS OF RATS*

Strain

ACI ACJ ACP ACUS Albany AS AS2 Atrichis August A990 AVN Axc B D V B D IX BDWCub B H BN BS Buffalo CAPICub Copenhagen DA Fischer Cunn HO (see PVC/c)

4

4

6 1

5 1 4

1 3 1 6 5 4 4 1

a

a

b 1 f

C

1 a

d

d I n 1 b

a a 1

C

2

2

1

1, 2t

1

2 2 1

x- 1 a

a

Not 1 1.1 1.2

x- I a

b Not 1 a

b

Not 1 a X-1 12 a

Not 1 a

A1,B.l'

A1,B.l- A1,B.l-

A 1 ,B. 1+ A1,B.l'

A1,B.l-

A1,B.l' A1,B.l-

A2,B.I-

AI,B.l+ A l.B. 1-

l a SD-1

w- 1 w- 1

1 w- 1

w- 1

3 b w- 1 w- 1

4 b w- 1 w- 1

1 a SD-1 1 a SD-1

lb 1. la Ib 1. l a l a 1. la l a 1. l a lb l a la l a l a la 2+

c l a l a l a 3 U l a l a l a

2 la 1. l a Y Q

2 la 1. l a & 2 Ib 1. l a 9

i4 l a 1. la

l a 1. l a l a 1. la l a

2 la l a la l a la l a

2 la l a l a l a l a la lb 1. l a Ib 1. la Ib l a l a l b 1. l a

Hypodactyle Lewis Long-Evans LOU/C/Wsl LOUIMIWsl LOUIC/lH/ (0KA)Iwsl Marshall MSU N BR OFA OKAMOTO PVGlc S5B Selfed Sherman SH Sprague-Dawley WAC Wistar AG Wistar Fu Wistar R Yoshida Zimmerman

x- 1

x- 1

Not 1

x- 1

Not 1

1

1.2 a

a

b

a a

a

b

A2,B.l-

A2,B. 1-

Al.B.1-

la l a la la l a la la l a l a

2 la la la la la la l a 1. lb

2 la la la 1 1b la la

b w- 1

E l b 1. 1. l a 3

la la 1. la cl 5 l b 1. Ib la la la 2

l a la l a

F S D - 1 1 l b 1. l a 2 la l a la ,+ w- 1

l a la la la la la 2

5 l a la la l b la la la 1. l a $

lb 1. l b

b

m 3 =! 5?

* Alternative designations for what appears to be the same locus are listed in adjacent grouped columns: Ag-B and H - I ; Ly-1 andART-1;

f X-1 indicates cross-reactivity with the Ag-D.1 reagent, but not necessarily possession of the Ag-D' allele (Palm and Black, 1971). t The BN strain for many years carried both alleles at the Ag-C locus (Palm and Black, 1971). Some sublines are now fixed forAg-Cl and

Pto and A p F ; R l - I , W-1 ,SD-I, RL, and I n .

others for Ag-C2, thus making available a congenic pair at this locus.

98 DAVID L. GASSER

within the framework of the International Union of Immunological Societies and will attempt to bring order to the confusion that now exists. At the time of this writing, the committee had not yet met, so all the nomenclature systems in current usage will be described in this review with the understanding that a uniform, more modern tenninol- ogy will soon appear in the literature.

The strain distribution for the various antigenic polymorphisms discussed in this review are listed in Table I. All this information was obtained from references cited in the text.

I I . The Major Histocompatibility Complex (MHC)

A. DEFINITION OF ALLELES

The discovery by Gorer (1937, 1938) and Snell(l953) of the mouse H - 2 histocompatibility system stimulated Bogden and Aptekman (1960) at the Wistar Institute to search for a similar major histocompatibility complex (MHC) in rats. They had previously described an ascites tumor designated AA that originated spontaneously in a female Wistar rat of the inbred PA strain (Aptekman and Bogden, 1955). When this tumor was injected into random-bred Wistar rats, it was lethal in only about 90% of the animals. When they examined the sera of the animals that had survived the tumor challenge, they discovered an antibody that agglutinated the erythrocytes of approximately 90% of the Wistar rats. Since Burhoe had designated the antigen he discovered as “A,” Bogden and Aptekman referred to their antigen as “B” and proposed that it was determined by a genetic locus that they called R-l (Bogden and Aptekman, 1960). Breeding experiments between the B-positive PA strain and the B-negative Lewis strain demonstrated that this antigen was inherited as a single-gene dominant trait (Bogden and Aptekman, 1962). These experiments were continued by Palm, who demonstrated that antigens similar to B could be detected in Lewis and BN rats by reciprocal immunizations. The antigen present in Lewis rats was designated 1, that in BN was called 3, and the antigens were shown to be determined by allelic genes (Palm, 1962). Antigen B was also shown to be determined by an allele at the same locus (Palm, 1964). This locus was then designatedAg-B (Elkins and Palm, 1966), the various alleles being indicated by numerical superscripts. A fourth allele was discovered in DA rats, and it was shown that among various strain combinations the Ag-B locus alone determines antigens that can elicit graft-versus-host (GVH) reactions by spleen cells from unsensitized donors (Elkins and Palm, 1966).


Stimulation in the mixed lymphocyte culture (MLC) reaction was also shown to require incompatibility at the Ag-B locus (Silvers et al., 1967).

While this work was proceeding in Philadelphia, similar experiments were being done in Prague. The first antigen identified by the Prague group was coincidentally named B1 (Frenzl et al., 1960a), but it was subsequently shown that this antigen does not have any histoin- compaGbility effect and is not ljnked to the major histocompatibility locus (Stark and Kr'en, 196%). Stark et al. first reported their characterization of rat M H C antigens at a Paris conference in 1966. In an attempt to observe the nomenclature agreed upon for the mouse H-2 system (Snell et al . , 1964), it was suggested that the rat M H C be designated Rt H - l apd that the various alleles be indicated by small letter superscripts (Stark et al., 1967). The major histocompatibility alleles of AVN, BP, WP, and Lewis rats were named H - l a , H - l b , H - l w,

and H - l ' , respectively. On the basis of typing common rat strains in Philadelphia and Prague, i t is now clear that Ag-B ' corresponds to H - l I ,

Ag-B2 corresponds to H - l w , Ag-B:' corresponds to H - l " , Ag-BJ corresponds to H - l a, Ag-W corresponds to H - l c , and Ag-B" corresponds to H - l b (Giinther and Stark, 1977). The Ag-B7 and Ag-BH alleles described by Kunz and Gill (1974) as yet have no H - l counterparts, and the H-ld, H - l e , H-lf, and N - l alleles have no known parallels in Ag-B terminology.

The complexity of these antigens became apparent when absorption analyses were undertaken. The H - l a a/lele, for example, was found to determine specificities 1 ,2 ,4 , and 7 (Stark and Kkn, 1967a), but this list was soon expanded to include 13, 14, 17, 20, and 21 (Stark et al., 1968). Similar studies were reported by Palm (1971). An u -to-date list

but these authors stress that this listing must be considered. provisional.

The M H C haplotypes and the strains that carry them are listed in Table 11. Also listed in this table are the currently defined antigenic specificities. However, for several reasons this information must be regarded as a very tentative interpretation of rat M H C antigens. First, since Stark and colleagues to a large extent studied a different group of strains from those examined by Palm, the Ag-B specificities reported by Palm do not completely coincide with the H - l determinants re-

orted by Gunther and Stark. Although the listing of Gunther and Etark is more complete, Palm reported specificities which do not coincide with anything reported by Gunther and Stark. For example, antigen 28 in the Ag-B series is present in strains with Ag-B haplotypes 1,

of H - l specificities has been compiled by Gunther and f tark (1977),

100 DAVID L. GASSER

TABLE I1 MAJOR HISTOCOMPATIBILITY COMPLEX (MHC) HAPLOTYPES AND

STRAINS CARRYING THEM ~~~~ ~~ ~~

H-1 H-1 antigenic Ag-B Ag-B antigenic haplotype specificities" haplotype specificities" Positive strains

1

W

n

a

c

b

d

e f

h

5,7,8,16,17,27, 31.32

4,6,9,17,28,34. 35, 36, 37

11, 15, 24,29, 30, 32, 33, 34

1,2,4,7, 13, 14, 17, 30, 33,s

10, 13, 15, 23, 25, 26, 27, 28

2, 3, 8, 9, 18, 23, 24

4, 8, 12, 14, 16, 18, 20,25,29, 36, 37

4, 8, 18, 19, 20 26, 30, 31, 37

22 (others not yet defined)

I

2

3

4

5

6

7" 8"

1, 26, 28 Lewis, F344, S5B, CAR, CAS, NBR, A990, BH, BS, AGUS

WF, OM, BN.B2, WP, L.WP, LOU/C, LOU/M, WAG

2,27, 28, 29, 31

3, 30 BN, L.BN

4,26,27,28,29, DA," BN .B4, AVN, 30 L.AVN, ACI, ACP,

COP

L.AUG, PVG (HO)

L.BUF, M520, ALB, SD

5, 28.31 August, CAP,

6.28, 29 BUF, BP, L.BP,

KG H WKA, OKAMOTO BDV, L.BDV

BDVII AS2, L.AS2

HW, L.HW

" Ciinther and Stark (1977). * Palm (1971). ' Although the MHC haplotype of DA seems to be identical to that of L.AVN ac-

cording to serological analysis and behavior in the MLC, it appears to differ by skin graft incompatibility in the F, test (Giinther and Stark, 1977).

" Kunz and Gill (1974).

2, 4, 5, and 6, but is absent from Ag-B3 rats. None of the H - 1 specificities has a corresponding strain distribution. Such inconsistencies will, it is hoped, be resolved at a series of formal workshops that are to be coordinated with nomenclature discussions (Section I), but until that time the antigenic specificities that have been assigned must be considered provisional.


A second reason for caution in accepting these specificities is that none of them has been studied in rats bearing known MHC recombinant chromosomes. Only after a considerable number of H-2 recombinant mice had been studied was the organization of the H - 2 complex clarified and the importance of cross-reacting specificities evident (Klein and Shreffler, 1971, 1972a; Murphy and Shreffler, 1975).

Finally, the assignment of specificities to the rat MHC antigens is predicated upon the assumption that cross-reactivity involves simple antigenic determinants and complex antibodies (Hirschfeld, 1965). How the data would be interpreted if one assumed that antigens are complex and antibodies are simple has not been fully explored in the rat. Such attempts have been made for the mouse H - 2 system (Thorsby, 1971), but not without strenuous objections involving the specifics of the H - 2 system (Klein and Shreffler, 1972b). There is no argument, however, over the general principle that an MHC system can be interpreted according to either mode and that either approach will result in a somewhat distorted understanding of the true situation (Klein and Shreffler, 1972b; Klein, 1975).

The partial purification ofAg-B molecules from papain digests of Fischer and ACI liver cell membranes has been reported. Antigenic molecules of about 59,000 daltons were dissociated in the presence of sodium dodecyl sulfate (SDS) into a 37,000 dalton fragment with specific Ag-B determinants and an 11,000 dalton fragment that seemed to correspond to &microglobulin (Katagiri et al., 1975a). In addition to the 59,000 dalton fragment, fragments of 35,000 and 25,000 daltons were also obtained by papain digestion. The 35,000 dalton fragment could be split to yield an 11,000 dalton &-microglobulin fragment, but this was not true of the smaller molecule. The 25,000 dalton fragment was devoid of any xenoantigenic activity characteristic of &- microglobulin, but it did carry Ag-B private specificities (Katagiri et al., 1975b). Partial purification ofAg-B antigens has also been reported by Stroehmann and DeWitt (1972a,b). Callahan et al. (1974), and Jones and Feldman (1975).

B. IA-LIKE ANTIGENS

As a result of studies on mice with recombinant H-2 chromosomes, it has been possible to define antigens coded by the immune response region of the mouse MHC (Shreffler and David, 1975). The profound importance of this genetic region has made the study of these Ia (I region-associated) antigens one of the most active areas in immunology. Unfortunately, the rich diversity of MHC recombinants so painstakingly produced in mice is not yet paralleled in rats, so pro-

102 DAVID L. GASSER

gress in this area has not been rapid. However, studies on the first recombinant obtained in laboratory rats have established that Ag-B- linked Ia antigens do occur in rats and that their properties are similar to those described in mice (Butcher and Howard, 1977).

In an experiment designed to search for an MHC recombinant, Butcher and Howard backcrossed (HO x DA)Fl hybrids with HO rats and discovered one product of this mating that typed as an Ag-B4/Ag- B” heterozygote in a lymphocytoxic test with antiserum and complement, but nevertheless, that could respond toAg-B4 cells in the MLC. The recombinant chromosome, which has provisionally been designated lR , has been extracted in the homozygous condition from inter- crosses and is also being backcrossed onto the HO background.

Absorption of an ( A 0 x HO)Fl anti-DA serum with 1R lymphoid tissue removed most but not all of its cytotoxic activity for DA lymphocytes. The residual activity killed only a subpopulation of normal lymphocytes (3540%), but could kill 90% of the lymphocytes from the thoracic duct of T cell-depleted (“B”) rats. Thus the 1R rats lack a B-cell antigen of the parental DA strain, although they possess the DA antigens, which are generally described as “serologically defined” (SD). That the gene for this B-cell antigen is linked toAg-B was shown in extensive backcross tests. Appropriate immunizations involving 1R have yielded antisera that are specific for the two recombinant regions ofAg-B4 and Ag-B’.

Davies and Alkins (1974) apparently raised antisera to rat Ia antigens as part of a study on passive enhancement of graft survival. Al- loantisera raised against WAG (Ag-B‘) spleen cells in AGUS (Ag-B’) rats consistently caused prolonged survival of heart transplants when given intravenously to graft recipients. After absorption by WAG erythrocytes, these sera lost much of their cytotoxic activity for spleen cells, but enhanced the grafts as well as did the originaI sera. These authors concluded that the effective antibody in the absorbed sera was directed against T-cell receptor sites and that it blocked recognition by the lymphocytes of the graft recipient (Davies and Alkins, 1974).

An antiserum that inhibited in vitro lymphocyte function was described by Wekerle et al. (1975). In this system, rat lymph node cells were sensitized by being cultured on mouse fibroblasts for 4 days and were then transferred to ”Cr-labeled target fibroblast cultures in which the extent of lysis was measured. Target-cell recognition was not blocked by antibody against immunoglobulin, but antilymphocyte sera strongly inhibited recognition in this system. These sera were raised in BN rats against Lewis lymphocytes and seemed to be directed to MHC-linked antigens, since they had blocking activity for


congenic strains possessing the same H-1 allele as Lewis. Since the blocking activity could not be absorbed b y kidney or liver cells, the relevant antigens appeared to be Ia-like rather than the SD antigens of the MHC (Wekerle et al., 1975).

An anti-Ia-like antibody that caused passive enhancement of kidney grafts was studied by Soulillou et al. (1976). These authors demonstrated that a single 0.5-ml injection of a hyperimmune Lewis anti-BN serum indefinitely prolonged the survival of (Lewis/BN)F, renal grafts in Lewis recipients. The same antiserum did not enhance the survival of BUFILewis or ACVLewis grafts. When the antiserum was absorbed with BN erythrocytes or platelets, the hemagglutinating activity as well as most of the cytotoxic activity were removed, but the enhancing effect was not diminished. Absorption with thymocytes or splenocytes removed the hemagglutinating, cytotoxic, and enhancing activities of the antiserum. It became apparent that the targets of the enhancing antibodies were not the SDAg-B antigens, and it was suggested that Ia-like antigens were involved. Evidence for this was obtained by demonstrating that unabsorbed antiserum or antiserum absorbed with erythrocytes or platelets could inhibit E A rosette formation, which is dependent on Fc receptors. It had been shown that anti-Ia sera block Fc receptors in the mouse (Dickler and Sachs, 1974); therefore, the rat-enhancing antibodies appeared to be similar to murine anti-Ia (Soulillou et al., 1976).

Another in uitro system that could be inhibited by anti-Ia-like sera was described by Gunther and Bhakdi-Lehnen (1976). Spleen cells from rats immunized against (T,G)-A--L, (H,G)-A--L, or (Phe,G)-A--L were cultured for 5 days in the presence of various sera. The number of direct plaque-forming cells was reduced by the presence of antisera directed to H - l antigens of high responders, a result similar to that obtained in guinea pigs by Shevach et al. (1972).

C. GENETIC ORGANIZATION

The degree to which we understand the subregions of the rat MHC is many years behind the beautiful work that has been done in mice. A number of fully characterized recombinants will be needed before one can determine how the various parts of this complex are or- ganized. This work is hindered to some extent by the lack of any known genetic markers closely linked to the MHC in rats, such as Brachyury, Tufted, and Fused in mice.

Although only one laboratory recombinant, the 1R haplotype discussed above, is available so far, the various parts of the rat MHC may be studied by using rats that possess naturally occurring recombinant

104 DAVID L. GASSER

chromosomes. One such haplotype may be theAg-B allele of the KGH strain (Kunz and Gill, 1974). It appears to be distinctly different from all other inbred strains in hemagglutination testing, so has been designated Ag-B7. However, no consistent MLC reactions occurred between strain KGH lymphocytes and those ofAg-B' strains (Cramer et al., 1974). This suggests thatAg-B' and Ag-B7 chromosomes may differ in the SD portions but be identical in the Ia-like regions.

None of the above experiments provides the crucial information needed to establish the arrangement of genes itJ the rat MHC. The recombinant obtained by Butcher and Howard could be explained by assuming that there is only one locus for the SD antigens or, alternatively, that two or more SD loci are closely linked and that the MLR-determining locus is outside this region (Fig. 1, A and B). It is equally plausible, however, to postulate that the MLR-determining locus lies between two SD loci, and that DA and HO possess the same allele at the second locus, represented as the SD-B" gene in Fig. 1C.

Ag-B alleles that seem to resemble natural recombinants present a special problem, since one does not know the composition of the parental chromosomes from which they were derived, if in fact they are recombinants. The Ag-B7 allele of strain KGH could have been derived by a mechanism similar to any of the three shown in Fig. 1. It is just as likely, however, that Ag-B7 is an SD mutant derived from an Ag-B' chromosome. If so, this would give us no information whatever about the position of the SD and MLR-determining loci or how many of these loci there might be. The same arguments apply to chromosomes isolated from wild animals.

Although we do not have good information about the order of genes in the rat MHC, the list of genes that map in this region is growing. In addition to the loci discussed above and those involved in immunologic responsiveness (Section VI, A and B), there is also at least one locus coding for a minor alloantigen that can be detected when sensitized lymphocytes kill target cells in uitro (Marshak and Wilson, 1976; Marshak et al., 1977). Although Lewis and Fischer are both Ag-€3' strains, cells from a Lewis rat that was sensitized against Fis- cher not only killed Fischer cells in uitro, but those of strains A990, AGUS, BH, AO, WF, BN, MAXX, ACI, DA, AUG, HO, BUF, and ALB as well. Fischer anti-Lewis cells killed Lewis, WF, BN, MAXX, ACI, DA, AUG, BUF, and ALB. In both cases, varying levels of killing occurred, and the level was reproducibly associated with the Ag-B type of the target cells. From these broadly overlapping distributions, it is


SD-A~ MLR' SO# SO-A' MLR) SD-E'

FIG. 1. Possible models for the organization of the rat major histocompatibility complex, and mechanisms by which recombinants could be generated. Model A assumes that there is only one serologically defined (SD) locus and one locus that stimulates in the MLR. Model B assumes there are two SD loci that both map on the same side of the MLR locus. In model C, the MLR locus maps between two SD loci, but the second locus is not detectable in this type of recombination because both parental chromosomes have the same gene at SD-B.

obvious that there must be a number of alleles at the locus controlling this trait and that there is considerable cross-reactivity among the antigenic products of these alleles. Lewis anti-Fischer cells killed target cells from 20/39 (Lewis x Lewis/DA) backcross rats, suggesting that only one locus is involved. All the cells killed were from Ag-B'/ Ag-B' heterozygotes, but none of the Ag-B 'IAg-B' cells were killed, indicating that the locus is closely linked to Ag-B.

D. IMMUNITY TO MHC ANTIGENS

There is evidence from a number of sources that immunity to the MHC antigens differs in at least one important respect from that involving other types of antigens. It has been estimated that the frequency of mouse T cells that are reactive to sheep red blood cells (SRBCs) is about (Kennedy et ul., 1966) or lop4 (Miller et al., 1967). However, the frequency of T cells that are reactive to the antigens coded by one MHC haplotype of the same species is about 1000-10,000 times as great. Although MHC antigens could be more complex than those found on SRBCs, it seems unlikely that they would have 1000-10,000 times as many specificities.

106 DAVID L. GASSER

Szenberg et al. (1962) found the first evidence for this unusual phenomenon in chickens and reported that as few as 40 large donor lymphocytes were sufficient to cause a pock on the chorioallantoic membrane (CAM) of a chicken embryo. Since such a pock indicates a GVH reaction, this suggested that 2.5% of a chicken’s large lymphocytes are reactive to the antigens of one haplotype of the B locus (the MHC of the chicken). In a similar experiment, Simons and Fowler (1966) reported that as few as 30 lymphocytes could produce a pock on the CAM. Simonsen (1967) repeated this type of experiment and estimated that at least 1-2% of the lymphocyte population responded to a given MHC antigen, For several reasons discussed by Simonsen (1967), this is a minimal estimate. Lafferty and Jones (1969) confirmed these findings and also showed that lymphocytes from species unrelated or only distantly related to the domestic fowl failed to initiate a GVH reaction when inoculated into a chicken embryo. Since this type of reactivity is directed against alloantigens rather than xenoantigens, the phenomenon has been designated “alloagression.”

Wilson and his colleagues studied alloaggression in the MLC (reviewed in Wilson et al., 1972). In 1968, they observed that 1 3 % of BN lymphocytes responded to DA antigens in an MLC reaction. This estimate was obtained by enumerating the number of lymphocytes that entered mitosis for the first time during the period of culture (Wilson et al., 1968). It was also shown that although immunization increased the tempo of the proliferative response, it did not increase the number of lymphocytes that reacted to Ag-B antigens in the MLC (Wilson and Nowell, 1971).

Atkins and Ford (1975) estimated the proportion of rat lymphocytes responsive to Ag-B antigens in the GVH reaction. These authors labeled strain A0 thoracic duct lymphocytes (TDL) with L3H1uridine and (A0 x DA)FI TDL with [‘4C]uridine, injected the two groups of cells intravenously into (A0 x DA)F, recipients, and then measured the localization of the 3H-labeled cells relative to that of the 14C- labeled nonreactive F1 cells. The GVH reaction involved a surplus of reactive parental cells in the spleens and a deficit of parental cells in the thoracic duct lymph. The amount of label that was selectively retained in the spleens allowed Atkins and Ford to estimate the proportion of responding lymphocytes and on this basis to calculate that approximately 12% of parental nonimmune lymphocytes react to the products of one Ag-B haplotype (Atkins and Ford, 1975). In a similar experiment, parental donor lymphocytes were labeled with [‘Hluridine and injected into F1 recipients. The number of cells that

CURRENT STATUS OF RAT IMMUNOCENETICS 107

were morphologically transformed in the recipient’s spleen within 24 hours was then measured. On the basis of this experiment, Ford and his colleagues estimated that 4.5-6.0% of the donor T cells were specifically responsive to the recipient’s Ag-B antigens.

Further confirmation of the large percentage of T cells reactive to Ag-B antigens was obtained by the use of anti-idiotypic sera (see Sec- tion VII,E). Purified T and B cells from Lewis and DA rats were exposed to an anti-(Lewis anti-DA) serum labeled with fluorescein isothiocyanate (Binz and Wigzell, 197513). Approximately 6% of Lewis T cells were positive for this idiotypic specificity, and about 1% of Lewis B cells bound this antibody. No T or B cells from DA rats were positive in this test (Binz and Wigzell, 1975b).

On the assumption that a given lymphocyte can react to only one antigen, the unusually large fraction of rat T cells that can react to Ag-B antigens would appear to leave an unexpectedly small proportion of lymphocytes that can react to environmental antigens. That this is not the case was shown in an experiment by Heber-Katz and Wilson (1976). These authors obtained a population of T cells enriched for reactivity for Ag-B2 by growing August strain TDL in bulk cultures with (August x Wistar-Furth)F, spleen cells. This cell population was approximately 10 times more potent in an anti-Ag-B2 GVH reaction than in a GVH reaction againstAg-B3. These cells were then tested for helper activity in an in uitro response against SRBCs. The helper activity of the positively selected population was not significantly different from that of a normal unselected population of August TDL. These authors concluded, therefore, that T cells appear to be reactive to more than one specificity (Heber-Katz and Wilson, 1976). Since there are at least 10Ag-B alleles, and 2-12% of the T cells are reactive to the antigens coded by any one of them, apparently many, perhaps all, rat T cells can react to an Ag-B alloantigen. If immunity to environmental antigens involves T cells that can also react to Ag-B alloantigens, then either the T-cell receptor has more than one specificity or there is more than one type of receptor on each T cell. The possibility that there may be indeed be multiple T-cell receptors has been discussed recently by Wilson et al. (1976).

Ill. Lymphocyte Alloantigens

Some of the most useful antigenic markers on mouse lymphocytes are those that are characteristic of a particular class of cells or a specific stage of differentiation. Since the mouse and rat are so closely related

108 DAVID L. GASSER

that genetic information in one species is frequently paralleled in the other, a brief discussion of mouse lymphocyte antigens seems appropriate.

In mice, a serological distinction can be made between lymphocytes that have not yet left the thymus and those that have on the basis of the thymus leukemia antigen, or Tla, system. The TL.1 antigen, coded by the Tla locus which is closely linked to H-2 on chromosome 17, is present on normal thymus cells but is absent from all other normal cells, including lymphocytes that have migrated from the thymus (Old et al., 1963; Boyse and Old, 1969). The 8, or Thy-1 , alloantigen first described by Reif and Allen (1963, 1964) is present on brain and most T cells and absent from most if not all B cells (Raff, 1971). The concentration of Thy-l antigen appears to be highest on thymocytes, somewhat lower on immature nonrecirculating T cells, and still lower on mature recirculating T lymphocytes (Raff, 1971). The Thy-1 locus is in the second linkage group (chromosome 9), 16.8 centimorgams from dilute (Itakura et al., 1971), and on the opposite side of the dilute locus from Mod-l (Itakura et al., 1972).

Further dissection of the murine T lymphocyte pool has been ac- complished by the study of Ly antigens. Ly-l and Ly-2 antigens are present in high concentrations on thymocytes and in lesser amounts on lymphocytes (Boyse et al., 1968). Ly-l is on chromosome 19 and Ly-2 is on chromosome 6 (Itakura et al., 1972). Another locus specify- ing T-cell antigens is Ly-3, which is in close genetic linkage with Ly-2 (Boyse et al., 1971; Itakura et al., 1972). The Ly-4 locus codes for B-cell antigens (Aoki et al., 1974), but Ly-5 is another T-lymphocyte antigen system and is not linked to Ly-l or Ly-21Ly-3 (Komuro et al., 1975).

The availability of these markers in mice has made it possible to identify three subclasses of peripheral Thy-l+ lymphocytes: (a) Ly- 123+, the first peripheral Thy-l+ cells to appear in ontogeny; (b) Ly-l+, a smaller population of cells that appears later and can give helper activity but are not cytotoxic; and (c) Ly-23+, a late-appearing, small population which can be cytotoxic but not give helper activity (Cantor and Boyse, 1975a). Ly-l+ cells are able to amplify the maturation of Ly-23+ cells to become cytotoxic, but the Ly-l+ cells do not contribute to the killer cell pool (Cantor and Boyse, 197%).

There can be no doubt that this type of study is very promising for elucidating the events in lymphocyte differentiation. One would expect that the mouse genes described above would have homologous counterparts in the rat, and there is evidence that this is indeed the case. Cross-reactivities between mouse and rat T-cell antigens have


been demonstrated (Balch et al., 1976; Ishii et al., 1976), but in these particular studies specific loci were not identified. However, it has been shown that rats possess an antigen that cross-reacts with the Thy-1 antigen of AKWJ mice (Douglas, 1972). This antigen is present on brain, thymus, spleen, and lymph node cells (Douglas, 1972; Micheel et al., 1973; Douglas and Dowsett, 1975). Unfortunately, all strains of rats tested so far have the same antigen, so a genetic analysis has not yet been possible.

A lymphocyte-specific alloantigen system designated Ly-1 was described by Fabre and Morris (1974). Cytotoxic antiserq were raised by immunizing AS and Lewis rats with each other’s t h ~ u s cells. The type possessed by AS was designated Ly-1.1; that of Lewis, DA; and AS2, Ly-1.2. Since the antigen was not present on brain, it did not appear to be homologous to Thy-], and since it was present on normal lymphocytes and could not be modulated by antiserum, it did not seem to be the counterpart of Tla. There is no evidence to suggest which of the mouse Ly loci may be homologous to the rat Ly-1 locus.

A similar antigen, designated ART-1 (antigen of the rat thymus), has been described by Lubaroff (1973, 1976). Cytotoxic antiserum was raised by immunizing Black Hooded rats with Lewis thymocytes. Ninety-five percent of the donor thymus cells were sensitive to this antiserum, compared with 61% of lymph node cells and 3640% of spleen cells, peripheral blood lymphocytes, and TDL, whereas only 6% of bone marrow cells were susceptible (Lubaroff, 1973). This antigen is not present on brain, so it does not appear to be homologous to the mouse Thy-1 antigen. It is also absent from lung, liver, kidney, heart, testis, and erythrocytes (LubaroE, 1977a). The strain distribution for this antigen is included in Table I.

A second antigen was discovered by injecting Marshall 520 rats with Buffalo thymocytes. Evidence suggests that this antigen is coded by an allele of the ART-1 locus, so it has been designated ART-lb (Lubaroff, 197%). This antigen is absent from the thymus cells of Lewis rats and other strains positive for ART-1”.

Antibody against another lymphocyte antigen was found in the M520 anti-Buffalo serum that reacted with Lewis lymph node and spleen cells. That this antigen differed from ART-lb was demonstrated by absorption studies, which may indicate incompatibility at a second ART locus (D. M. Lubaroff, personal communication).

Goldschneider and McGregor (1973) and Goldschneider (1975, 1976) described four rat antigens discovered by the use of rabbit antilymphocyte serum. The rat T-lymphocyte antigen (RTLA) is present on cells of the thymic cortex and medulla as well as T cells in the

110 DAVID L. GASSER

lymph nodes, spleen, blood, and thoracic duct. I t is also present on approximately 13% of bone marrow lymphocytes, a fact that is consistent with the observation of Howard and Scott (1972) that significant numbers of recirculating T cells occur in the normal rat bone marrow.

A second antigen described by Goldschneider and McGregor is characteristic of B lymphocytes, and is now designated rat B-lymphocyte antigen (RBLA) (Goldschneider, 1976). When a rabbit antiserum was prepared against a population of “null” cells which lacked both RTLA and RBLA, a third heteroantigen designated rat bone marrow lymphocyte antigen (RBMLA) was discovered. This antigen can be found on “null” cells of the bone marrow, cortical thymocytes and a subset of T cells in spleen and blood (Gold- schneider, 1976). Lymphocytes which are positive for RBMLA are negative for another antigen previously described by Goldschneider (1975) and called the rat masked thymocyte antigen (RMTA). RMTA is normally expressed on nearly all the T cells of the thoracic duct and lymph nodes and on 58-60% of splenic T cells, but occurs in a masked form on medullary thymocytes. It can be unmasked by neuraminidase treatment or by staining sections of frozen cells. T cells that are positive for RMTA are cortisone-resistant, whereas those which are RMTA- are cortisone-sensitive (Goldschneider, 1975).

On the basis of these findings, Goldschneider has proposed that rat T cells are derived from two ontogenically distinct lines. One of these is antigenically related to medullary thymocytes and comprises a population of recirculating cortisone-resistant T cells. The other population is antigenically related to cortical thymocytes, is cortisone- sensitive, and recirculates poorly (Goldschneider, 1976).

As part of an extensive program to define the bone marrow-derived and thymus-derived components of the immune response, Howard and Scott (1974) discovered a series of alloantigens that occur on peripheral thymus-derived lymphocytes. Cytotoxic antisera were developed by repeated subcutaneous and intraperitoneal injections of spleen and lymph node cells, and assayed by their effects in the presence of complement on TDL as assessed by Trypan Blue staining. Reciprocal immunizations between the HO and August strains produced an antiserum that was highly cytotoxic for TDL obtained from a normal rat of the donor strain, but TDL obtained from T-depleted rats (B rats) were not sensitive to this antiserum. Quite surprisingly, thymocytes were completely resistant to the effects of the antiserum. In fact, the only cell type shown to be positive for these alloantigens was the peripherd thymus-derived cell. The low activity present in lymphoid tissues such as thymus and bone marrow was compatible


with known levels of contamination by peripheral T cells. The antigen was not detected in nonlymphoid tissues, including liver, kidney, testis. brain, erythrocytes, platelets, and skin.

Howard and colleagues were able to define three classes of strains with their original antisera: one type whose cells were sensitive to anti-HO serum; a second type sensitive to anti-August serum; and a third class sensitive, although differentially, to both antisera. The Au- gust anti-HO serum had weaker activity than HO anti-August for cells of the third type of strain. Subsequent work on the antigenic target for August anti-HO serum has not yet demonstrated conclusively whether the antiserum's activity is restricted to thymus-derived lymphocytes. The picture is confused by the fact that this activity is absent from the majority (12116) of August anti-HO sera and that these sera also sporad- ically contain measurable activity against HO B cells. These weak activities are absent from (August x AO)FI anti-HO serum, whose target is completely T-cell restricted.

The locus defined by these sera has been designated Pta, for peripheral thymus-derived cell antigen. The A1 allele at this locus is defined by HO anti-August serum, and the A2 allele by (August x AO)F1 anti-HO. All strains so far tested have been characterized as possessing one or the other of these two alleles (see Table I). In extensive backcross tests, it was shown that Pta is closely linked to the locus for albinism.

A second specificity present on peripheral T cells and absent from B cells is recognized by (A0 x HO)F1 anti-August. This specificity is present in some Pta.Al strains, such as August and DA, but absent from others, such as A0 and Fischer. No Pta.A2 strain has yet been found with this specificity, which has been provisionally designated Pta.Bl. In linkage tests the specificity cosegregates with Pta.Al (J. C. Howard, personal communication).

In an independent experiment, DeWitt and McCullough (1975) also discovered a system of lymphocyte antigens controlled by a locus closely linked to albinism. Cytotoxic antisera were raised b y intraperitoneal injections of lymph node and spleen cells, and it was shown that the antigens detected could not be Ag-A, Ag-B, Ag-C, Ag-D, orAg-E. The new locus was therefore designatedAg-F, and four alleles were demonstrated. Ag-F' is present in Fischer, WistarIFurth, ACI, DA, August, and M520 strains. Ag-F2 is defined by Lewis, Ag-F' by BN, and Ag-F4 by Buffalo. In backcross experiments it was shown that recombination occurred between Ag-F and albinism in 4 out of 49 progeny (DeWitt and McCullough, 1975), which would suggest a recombination frequency of approximately 8%.

Although the locus described by DeWitt and McCullough would

112 DAVID L. GASSER

appear to be the same as Pta, there are some discrepancies in typing. DeWitt and McCullough (1975) reported that August has the same type as Fischer and DA, whereas Howard and Scott (1974) found Au- gust to differ from DA and Fischer. Howard and colleagues found Buffalo to be indistinguishable from Lewis (J. C. Howard and K. F. Mitchell, personal communication) whereas DeWitt and McCullough reported them to be different.

Williams and DeWitt (1976) have isolated and partially characterized the Ag-F' antigen of Fischer lymphocytes. Spleen and lymph node cells were radioiodinated by using a lactoperoxidase method and treated with NP-40. After removing nuclei and particulate matter by centrifugation, the NP-40 was removed and the material was precipitated with polyethylene glycol and run on sodium dodecyl sulfate polyacrylamide gel electrophoresis. Protein was eluted from the gel slices, lyophilized, dissolved in urea, and passed over a microcolumn of Dowex mesh. Eluates were dialyzed and analyzed for specific antigenic activity. By this procedure, a fraction which contained concen- trated Ag-F' antigen was purified. The molecular weight of this peak was estimated to be 37,000 (Williams and DeWitt, 1976).

Determining the phylogenetic relationship between the loci coding for lymphocyte antigens in mice and rats may be extremely difficult until more genetic markers are described in both species. One possible approach would be to determine whether any of the rat T-cell antigen loci are linked to a locus for immunoglobulin light chains. The IB peptide marker of EdeIman and Gottlieb (1970) is known to involve an amino acid substitution at position 23 in the variable region of mouse light chains and is not present on lambda chains. There is an absolute positive correlation between the presence of this marker and the Ly-3.1 antigen, suggesting that the vk locus is closely linked to Ly-3 (Gottlieb, 1974). Even though an exact homolog of the IB allele may not exist in rats, one would expect other light-chain v-region markers to occupy the same position in the genome. It has been shown that Ag-F is not closely linked to the kappa chain allotype locus (K. F. Mitchell, personal communication), but perhaps one of the other T-lymphocyte markers will prove to be kappa-linked.

IV. Other Blood Group and Histocompatibility Polymorphismr

A. THE Ag-C Locus In 1962, Owen reported the development of two absorbed rabbit

antirat sera, designated anti-C and anti-D, which recognized erythrocyte antigenic specificities. Among 5000 rats tested, none whose blood


cells failed to react with one or the other or both of these reagents was found. Extensive breeding tests demonstrated that the C and D antigens were inherited as simple alleles (Owen, 1962). These antigens appear as early as day 10 of gestation (Owen, 1962) and seem to be present on spleen and lymph node cells but absent from liver cells (Palm, 1962). Palm and Black (1971) reported the strain distribution of these antigens, which they referred to as C1 and C2, and recom- mended Ag-C as the designation for the locus determining them.

The discovery of linkage between Ag-C and two loci determining electrophoretic variants of serum esterases suggested that Ag-C may be homologous to the Ea-l locus of mice, since the mouse Ea-l locus is linked to several serum esterase loci (Gasser and Palm, 1972; Gasser et a l . , 1973b). Three additional esterase loci were mapped in the same linkage group, and biochemical characteristics of the rat and mouse esterases have been compared (Womack and Sharp, 1976). The mouse Es-l and rat Es-2 loci code for plasma arylesterases, which migrate very close to the plasma albumins under most electrophoretic conditions. The mouse Es-2 and rat Es- l loci code for esterases which have a preference for butyryl substrates and migrate in electrophoretic fields in a position more anodal than the serum albumins. The linkage relationships of the relevant loci in the two species are as follows (Foster et al., 1969; Robinson, 1972; Gasser et ul., 197313; Womack and Sharp, 1976):

5.6 o M 9.fi cM

5.3 c M Y.6 cM Rat Ag-C Es-2 Es-1

Mouse Ea-l Es-l - Es-2

Although the map intervals (given in centimorgans) shown are so close as to be embarrassing, there are also other esterase loci in the same region of both rats arid mice without known homologs in the other species (Womack and Sharp, 1976). Suffice it to say that the evidence for genetic homology between Ag-C and E a - l is very strong.

Ag-C antigens have been examined in two populations of wild rats (Shonnard et al., 1976). Although the sample sizes were not large, the frequency of the C 1 allele was considerably higher than that of C 2 in both populations (0.78 and 0.65). Whether this allelic distribution will continue to occur in other wild populations is a most interesting problem for future studies. The possibility that selection operates at the Ea-1 locus of mice was suggested by the fact that all inbred strains tested so far have the amorphic E a - l o allele (Foster et al., 1968).

B. THE Ag-D Locus In her earlier studies of the Ag-B antigens, Palm (1962) observed

that Lewis rats possessed an antigen (antigen 2) detectable by

114 DAVID L. GASSER

hemagglutination that was not shared by the WistarIFurth strain. The locus for this antigen was subsequently namedAg-D (Palm and Black, 1971) and shown not to be linked toAg-B (Palm, 1962). Although W F was the only strain with the original antigen 2, now called Ag-D.l, a number of strains showed cross-reactivity with the D1 reagent (Palm and Black, 1971).

C. THE B1 ANTIGEN

A blood group antigen originally designated B or B 1 was identified in Prague (Frenzl et al., 1960a) and seemed to be associated with induction of runt disease (Kkn et aZ., 1960) or erythroblastosis fetalis (Frenzl et al., 1960b). Because of these effects, it was initially believed that B1 was a strong transplantation antigen, but it was demonstrated that the gene coding for B1 was not linked to the MHC (Stark and Kfen, 1967b), and that B1 had no detectable histoincompatibility effect (Frenzl et al., 1972). The locus for this antigen is now being referred to as H-2 (Frenzl et al., 1972).

D. THE H-3, H-4, AND H - 5 LOCI

In the course of transferring the H-2" gene, which, codes for the absence of the B1 antigen, from the WP strain onto BP, another serologically active antigenic system was discovered (Kien et al., 1973). The allele possessed by BP has been designated H-3b, and the null allele of WP is H-3". H-3 is not linked to H - l or H-2 (Kien et al., 1973).

A histocompatibility locus closely linked to albinism has also been reported by Kien et al. (1973). Incompatibility at this locus, designated H-4, leads to rejection of tail skin grafts within 50 days. Since K k n and colleagues were not able to detect serologic activity associated with H-4, it appears that this locus is different from Ag-F and Pta (Section 111).

A histocompatibility locus closely linked to luxate has been discovered during the process of transferring the luxate mutation from Au- gust onto the Lewis background. Incompatibility at this locus, designated H-5, leads to transplantation effects, but not to proven serological activity (Kien et al., 1972, 1973; Kien and Stark, 1972).

E. SEX-LINKED ANTIGENS

The male-specific transplantation antigen first discovered in mice (Eichwald and Silmser, 1955) was shown to have a counterpart in rats when Billingham et al. (1962) observed that about 60% of BN females rejected BN male skin grafts. Although Lewis male skin was not rejected by Lewis females in these experiments, it was rejected by about


40% of the (Lewis x BN)FI hybrid females. It was subsequently shown that when Lewis females were grafted with transplants smaller than 2 cm', 56% were rejected. Furthermore, when BN and Lewis females were sensitized by an injection of spleen cells, the Lewis females rejected their grafts much more effectively than did the BN females (Silvers et al., 1976).

Although the H - 2 locus strongly influences the ability of female mice to reject male skin (Gasser and Silvers, 1971; Bailey and Hoste, 1971), no evidence for a similar Ag-B-linked immune response gene has yet been obtained in rats (Heslop, 1973).

One of the peculiarities of the mouse H-Y system is that the potency of the H-Y antigen is strongly influenced by the H - 2 region (Wachtel et al., 1973). Differences in the potency of the rat H-Y antigen have also been observed, although the role ofAg-B in this effect has not yet been examined. Only 47% of male BN grafts transplanted to sensitized (BN x Lewis)F1 and (Lewis x BN)FI hybrids were rejected, whereas all the Lewis and (Lewis x BN)FI male grafts transplanted to such recipients were destroyed (Silvers et al., 1976).

Evidence for cross-reactivity between the mouse and rat H-Y antigens was first obtained by the demonstration that injection of male rat cells into female mice can sensitize them for accelerated rejection of mouse male grafts (Silvers and Yang, 1973). It was then shown serologically that these antigens cross-react, and in fact, that there is cross- reactivity between the heterogametic antigens of mouse, human, rat, rabbit, guinea pig, leopard frog, South African clawed toad, and chicken (Wachtel et al., 1974, 1975). The fascinating possibility that H-Y is responsible for triggering the development of the indifferent embryonic gonad into a testis has been discussed (Ohno, 1976).

The finding of Bailey (1963) that mice possess an X-linked histocompatibility locus has been extended to rats. Mullen and Hil- demann (1972) reported that the Lewis and Fischer strains differ at an X-linked histocompatibility locus, but that the products of this locus are only weakly immunogenic. The median survival time for grafts of Lewis male skin onto (Fischer x Lewis)F1 male recipients was 335 days, and that of Fischer male skin onto (Lewis x Fischer)F1 males was 259 days (Mullen and Hildemann, 1972).

V. Evidence for Selection at Histocompatibility loci

The A2 substrain of Wistar rats, which had been maintained in the laboratory of Professor M. F. A. Woodruff for 72 generations of brother-sister matings, was observed to continue segregating for at least one pair of histocompatibility genes, designated gl and gz

116 DAVID L. GASSER

(Michie and Anderson, 1966). About one-half of all skin grafts made within the strain were rejected by 3 weeks after grafting. These rats were studied further by Michie and Anderson (1966), who bred the rats for four generations exclusively from brother-sister pairs that had accepted grafts from each other. When this breeding program failed to increase the frequency of two-way “takes,” it became obvious that a powerful selection for heterozygosity was occurring, and the mutually compatible brother-sister pairs were heterozygotes of the gJgz gJgz combination. Michie and Anderson were able to isolate one of the homozygous types and develop an isogenic subline (AYl), which proved that the homozygotes of at least one type (gl/gl) were fertile.

The question of when mortality was occurring was studied by examination of litter sizes. If homozygote deaths occurred after implantation, then mating of the gl/gl x gl/gl type should yield smaller litters than those of the gl/gz x gJg, type. Compensation for preimplantation mortality could have occurred, since more eggs are normally shed and fertilized than implanted. Michie and Anderson found that there was no significant difference in the litter sizes resulting from the two types of matings, and concluded that selective elimination acted before implantation of the fertilized egg (Michie and Anderson, 1966).

The histocompatibility locus for the genes described by Michie and Anderson is currently being designated Ag-E (DeWitt and McCul- lough, 1975).

Evidence for heterozygote advantage at the Ag-B locus was reported by Palm (1969, 1970). In a series of backcross matings that had been set up to study linkage relationships, an excess ofAg-B heterozygous progeny was observed when the mother was of an inbred strain. For example, in the first series of this type, the surviving progeny consisted of 164 heterozygotes and 109 homozygotes. When this work was repeated, Palm observed that a substantial number of the rats were dying in infancy of a condition that resembled GVH disease and that significantly more Ag-B homozygotes than heterozygotes were affected by this condition. One year later, Palm tested a third series of rats in the same way. For some unknown reason, the mortality of these animals was lower than in the first two experiments, and there was no association between mortality and Ag-B type. Palm concluded that, under conditions leading to high mortality, Ag-B heterozygosity was adaptive. Since the affected animals were Ag-B homozygotes and therefore Ag-B-compatible with their mothers, she proposed that non-Ag-B incompatibilities were the targets of this reaction, and that the Ag-B heterozygotes were in some way protected against this disease (Palm, 1970).


VI. Genetics of the Immune Response

A. MHC-LINKED IR GENES

The discovery of specific immune response (Ir) genes in guinea pigs and mice (reviewed in McDevitt and Benacerraf, 1969; Benacerraf and McDevitt, 1972; Gasser and Silvers, 1974) has been extended to a number of systems in the rat. The first antigen studied in this way was poly(Glu"'Lys:':iTyr'") (Simonian et al., 1968). The strain that responded best was ACI, whereas the F344 strain produced the lowest response. The (ACI x F344)F1 hybrid produced about one-third as much antibody as the ACI parent, and breeding studies indicated that more than one gene was involved (Gill et al., 1970). Further data that demonstrated the polygenic nature of this trait as well as a sex influence have been published (Gill and Kunz, 1971; Gill et al., 1971). It has also been shown that the antibodies elicited by poly(G1u"'Lysi"Tyr'J) in high responders have higher binding constants than those produced by low responders and that this trait is determined by an Ag-Blinked gene (Ruscetti et al., 1974).

An especially interesting rat Ir gene involves responsiveness to the A subunits of lactic dehydrogenase (LDH). Sprague-Dawley and Wis- tar rats responded equally well to LDH composed of B subunits (LDH/B4), but at low doses of antigen only Sprague-Dawley rats produced a detectable response to LDH-% (Armerding and Rajewsky, 1970). Wistar rats responded to higher doses of antigen, but their responses were always lower than those of Sprague-Dawley rats. It was shown by Wurzburg (1971) that this gene, which was designated Ir- LDHA, is closely linked to the MHC. Rats with the Ag-B4 (H-la) allele are low responders, whereas all other types tested ( H - l alleles b, c, d, e, f, 1, n, and w) are high responders (Wurzburg, 1971).

The pioneering work of McDevitt and his colleagues on the genetics of response to (T,G)-A--L in mice (McDevitt and Sela, 1965, 1967; McDevitt and Tyan, 1968; McDevittet al., 1972) has been extended to rats. Various inbred strains differ in the level of response to (T,G)-A--L, and these quantitative variations are closely correlated with Ag-B or H - l type (Gunther et al., 1972).

The response of rats to (T,G)-A--L has also been studied by Koch (1974), who demonstrated that the AS and BN strains produce antibodies that are directed to different determinants on this antigen. The anti-(T,G)-A--L produced by AS rats was not capable of binding [I2'I](T,G)-Pro--L, but that produced by BN rats bound this polypeptide to the same extent that it bound [12Y](T,G)-A--L (Koch,

118 DAVID L. GASSER

1974). The AS rats responded very poorly to all doses of a low molecular weight (T,G)-A--L, but the BN rats responded very well to injections of 59 p g or more of this antigen. When a high molecular weight (T,G)-A--L was used, AS responded well and BN responded poorly to doses of 50 pg or less, but the opposite result occurred when a dose of 200 pg was used (Koch, 1974).

The response to (H,G)-A--L is also genetically controlled in rats (Gunther et al., 1973; Gunther and Rude, 1975). The gene for this trait is closely linked to the MHC and also shows the interesting property of genetic complementation, in that heterozygotes of several types produce a higher response than either of the parents from which they were derived (Gunther and Rude, 1975).

A wide variety of rat strains were tested for delayed-type hypersensitivity reactions and antibody production following immunization with GA (L-GIu~O, L-Alaj") and GT (L-GIIP, L-Tyr'O). Some strains produced intense delayed-type responses and high antibody titers, whereas others produced little or no response (Armerding et al., 1974a). WF was a low responder to GT and a high responder to GA whereas ACI was a high responder to GT and a low responder to GA. (ACI x WF)FI hybrids were high responders to both antigens. In the F1 x ACI and F1 x WF backcross generations, the genes for responsiveness to GA and GT were shown to be closely linked to the Ag-B locus. It was also shown that spleen cells from responder rats immunized with GT produced strong DNA synthetic responses when cultured in the presence of GT (Armerding et al . , 1974b).

An Ag-B-linked gene controlling responsiveness to a known sequence polymer, (Tyr-Glu-Ala-Gly),,, was described by Luderer et al. (1976). The strains that were surveyed could be classified as high responders (ACI, WF, AUG), intermediate responder (F344), or nonresponders (Lewis, Buffalo, BN, MAXX). In the (AUG X BN)FI X BN backcross generation, all Ag-BVAg-B" heterozygotes were responders and all Ag-B3/Ag-B3 homozygotes were nonresponders. How- ever, the levels of antibody produced by the rats classified as responders suggested that a second gene not linked to Ag-B was also involved. The intermediate response by the F344 strain, which has the same Ag-B haplotype as Lewis,was attributed to a carrier effect of the mycobacterium in the adjuvant, since F344 was a nonresponder when injected with (T-G-A-Gly),, in incomplete Freund's adjuvant.

Strain differences in susceptibility to autologous immune complex glomerulonephritis have been described. Rats were immunized with a single injection of primary tubular epithelium taken from rat kidneys and emulsified in Freund's adjuvant, and nephritis was assessed by


measurement of protein in their urine as well as histological examination of their kidneys. Under these conditions, the Lewis strain developed severe nephritis, but the BN strain did not. Since the congenic Lew.BN strain was not susceptible, there appeared to be an Ag-B-linked gene involved in susceptibility to this disease (Stenglein et al., 1975).

B. EXPERIMENTAL ALLERGIC ENCEPHALOMYELITIS

Experimental allergic encephalomyelitis (EAE) is the oldest, best- known, and most extensively studied experimental autoimmune disease. By the 1920s it was established that brain tissue possesses organ-specific antigens that may elicit specific immune responses in animals. Since that time, EAE has been used as a model system for studying such problems as the role of adjuvants, transfer of autoimmunity with lymphoid cells, role of cytotoxic serum factors, suppression of disease by serum, and induction of tolerance to self-antigens (reviewed by Paterson, 1966).

The requirements for disease induction vary from species to species, but EAE can readily be induced in rats by the injection of spinal cord from guinea pigs (Lipton and Freund, 1953), even without adjuvants (Bell and Paterson, 1960; Paterson and Bell, 1962; Levine and Wenk, 1961, 1965). Spinal cord of the paca, which is a member of the same infraorder and superfamily as the guinea pig, but of a different family, was almost as encephalitogenic as that of the guinea pig (Levine and Wenk, 1965). The next most active spinal cord for producing EAE in rats came from the agouti, which is also from a different family but the same infraorder and superfamily as the guinea pig (Levine and Wenk, 1965). Spinal cords of cow, dog, human, rabbit, squirrel, mouse, hamster, nutria, chinchilla, and capybara origin were only slightly encephalitogenic in rats (Levine and Wenk, 1965). Rat spinal cord was highly encephalitogenic in some experiments (Lipton and Freund, 1953; Graham et al., 1974; Gonatas et al . , 1974; Levine and Sowinski, 1975), but only slightly encephalitogenic in others (Levine and Wenk, 1965). Homogenized spinal cord from both the salamander and frog is encephalitogenic in at least some strains of rats, but that of the carp is not (Martenson et al., 1972).

The course of the disease may vary according to experimental conditions, but the earliest lesion that has been reported is a generalized increase in vascular permeability specific to the central nervous system (CNS) as early as day 5 or 6 after sensitization (Oldstone and Dixon, 1968). Progressive weight loss is usually observed, as well as tail and hind limb paralysis and loss of bladder and rectal control. In

120 DAVID L. GASSER

the brain and spinal cord, a varying number of capillaries and venules show cuffing with mononuclear cells and small lymphocytes. As the disease progresses, the infiltrating cells spread from the blood vessels to the surrounding areas. The cellular infiltration is followed by primary demyelination, which could result from any one of several mechanisms: (a) myelin could be damaged by circulating or locally produced antibody (Raine and Bornstein, 1970; Bornstein and Iwanami, 1971; Oldstone and Dixon, 1968); (b) sensitized lymphoid cells could cause direct destruction of myelin; or (c) myelin could be damaged by mediators produced by sensitized lymphocytes, or secondarily by activated macrophages influenced by lymphocytes (Gonatas et al., 1965; Wisniewski et al., 1969; Wisniewski and Bloom, 1975). If the animals survive this acute attack (and most of them do if permitted to by the investigator), a recurrent episode frequently occurs about 10 days after the first (McFarlin et al., 1974).

This chain of events is fully dependent upon the presence of thymus-derived lymphocytes. Rats depleted of their T cells are unable to develop EAE, but T-depleted rats develop the disease normally if they have been reconstituted with thymocytes (Gonatas and Howard, 1974). If T-depleted rats are reconstituted with sensitized T cells that were treated with antithymocyte serum, the rats do not get the disease even though they can produce normal amounts of antibody to the inciting antigen (Ortiz-Ortiz et al., 1976). Therefore, T cells are necessary, not as helper cells in a humoral response, but as effector cells involved in producing the lesions.

The presumed target of this reaction is a portion of a basic protein that occurs in myelin, since the injection of a basic protein extracted from myelin induces EAE (Kies and Alvord, 1959; Roboz-Einstein et al., 1962; Kibler et al., 1964). Other constituents of the myelin sheath may also be involved when whole spinal cord is used. Although immunization with whole spinal cord leads to the production of factors that can demyelinate CNS tissue in culture, such factors are not produced if the animals are sensitized with basic protein (Bornstein and Crain, 1971; Kies et al., 1973; Seil et al., 1973, 1975).

It was indeed fortunate that this excellent model system should lend itself to the study of the genetic control of the immune response. It was shown by Gasser et al. (19734 that a single gene linked to the MHC of rats is a major determinant in controlling reactivity to guinea pig spinal cord. Of more than 200 Lewis rats given intradermal injections of guinea pig spinal cord in Freund's complete adjuvant, not one failed to develop EAE. As previousIy shown by Kornblum (1968), rats of the


BN strain did not get the disease when given the same treatment. Lewis x BN F, hybrids were susceptible to the disease, so susceptibility is inherited as a dominant trait. When rats of the (Lewis x BN)F, x BN backcross generation were tested for their Ag-B type and injected with guinea pig spinal cord, the results shown in Table I11 were obtained (Gasser et al., 1973a). None of the Ag-BYAg- B" homozygotes developed any sign of EAE, whereas 12/15 Ag-B'I Ag-B3 heterozygotes developed various degrees of EAE. This finding was confirmed by Williams and Moore (1973), who used guinea pig basic protein in Freund's adjuvant containing Mycobacterium tuberculosis. These authors also observed that none of the Ag-B31Ag-B3

TABLE I11

ENCEPHALOMYELITIS (EAE) SUSCEPTIBILITY WITH Ag-B TYPE AMONG (LEWIS X BN)FI x BN

BACKCROSS RATS"

COMPARISON OF EXPERIMENTAL ALLERGIC

Ag-B ' IAg-B ' Ag-B"IAg-B"

Spinal Spinal Animal cord Animal cord

no. histology" no. histology"

532 535 536 617 664 530 531 728 729 735 737 738 667 663 672

0 3 3 4 4 0 2 0 3 3 1 2 3 2 2

529 538 540 541 542 618 736 665 564 561 669

0 0 0 0 0 0 0 0 0 0 0

" From Gasser et al. (1973a). Copyright 1973 by the American Association for the Advancement of Science.

"The spinal cord ratings are as follows: 0, no infiltration of mononuclear cells; 1, one infiltrate in spinal root; 2, many infiltrates in spinal roots and occasional infiltrates in spinal cord; 3, many root and cord infiltrates; and 4, many confluent root and cord infiltrates.

122 DAVID L. GASSER

homozygotes developed the disease, whereas 2 1/25 heterozygotes developed histological signs of EAE. There is no evidence that the Ag-B 'IAg-B3 heterozygotes which failed to get the disease in either experiment were recombinants. At the present time at least, it appears that this occurred because the gene for susceptibility is incompletely penetrant.

Although the difference between Lewis and BN in their reactivity to guinea pig spinal cord can be explained by the segregation of a single major gene, this is not the case if DA is compared with BN. The DA strain is quite susceptible to EAE when injected with guinea pig spinal cord, but DA x BN F1 hybrids are not nearly as susceptible as Lewis x BN F, hybrids (Gasser et al., 1975). In the (DA x BN)FI x BN backcross generation, only four of 32 animals showed the slightest sign of EAE. The four animals affected (and they were only mildly affected) were Ag-B4/Ag-B3 heterozygotes, suggesting that the A g - B locus may exert a slight effect in this combination. The polygenic nature of this reactivity was further confirmed by the work of Perlik and Zidek (1974), who reported that the AVN strain was not at all susceptible to EAE when injected with guinea pig spinal cord, and that (Lewis x AVN)Fl hybrids had a very low level of susceptibility. This is especially interesting since AVN has the same A g - B allele (Ag-B4) as DA.

An important observation was made by McFarlin et al. (1975a,b) in demonstrating that, although the BN rat was able to respond to the whole basic protein extracted from guinea pig myelin as assessed by either an antibody assay or a skin test, this rat was unable to produce either an antibody or cell-mediated response to a 42-residue encephalitogenic fragment of this molecule. This is consistent with earlier observations that there are at least three mutually exclusive antigenic sites in guinea pig basic protein (Driscoll et al., 1974), that BN rats can react very well against basic protein in the skin test (Williams and Moore, 1973), and that BN rats can produce high levels of circulating antibody against basic protein in the absence of histological evidence of EAE (Gonatas et al., 1974). It therefore appears that the Ir gene present in the Lewis strain but lacking in BN is concerned with recognition of this 42-residue fragment. Levine and Sowinski (1975) demonstrated that if BN rats are injected with rat spinal cord and carbonyl iron adjuvant they can develop severe EAE. This suggests that the BN rat has no constitutional impairment that makes it unable to develop EAE. For example, the suggestion that nonsusceptible animals fail to develop EAE because of a lack of mast cells or vasoactive amines (Gershon, 1976) is not consistent


with the report of Levine and Sowinski as far as the BN rat strain is concerned.

If BN rats can indeed react normally to the encephalitogenic fragment of rat spinal cord, then a comparative study of the amino acid sequences of the basic proteins of various species may give us some insight into the specificity of the immune response gene possessed by Lewis rats. EAE can be induced in guinea pigs by a fully characterized peptide of 10 amino acid residues (Eylar, 1970; Lennon et al., 1970), but the encephalitogenic fragment which is active in rats has not yet been sequenced (McFarlin et al., 1973). Human and bovine basic proteins have been sequenced and are very similar to one another (Eylar, 1970), but the rat differs from these species in that its myelin contains two basic proteins that are encephalitogenic (Cotman and Mahler, 1967; Eng et al., 1968; Martenson et al., 1970). The larger of the two rat proteins (the L protein) is comparable to the basic protein of man, ox, rabbit, and guinea pig in amino acid composition, size (approximately 18,400 daltons) and electrophoretic mobility. The smaller (S) protein is similar to the L protein, except that it is missing 40 amino acid residues within the C-terminal half of the molecule (Dunkley et al., 1972; Martenson et al., 1972). The complete sequence of the S protein has been reported (Dunkley and Camegie, 1974), but that of the L protein is not yet available.

The frequency of thymus and spleen cells in BN rats which can bind labeled rat basic protein was measured by Ortiz-Ortiz and Weigle (1976). Lewis rats were found to possess 17 antigen-binding lymphocytes (ABL) per 10; spleen cells and BN rats had 11 ABLIlOj spleen cells. However, the BN rats had no ABL in their thymus cells, whereas Lewis rats had 8 ABL/10j thymus cells. If the BN rats are completely lacking T cells capable of recognizing rat basic protein, it would appear that EAE induced in these rats by spinal cord emulsions (Levine and Sowinski, 1975) may be dependent not on basic protein, but on other neural antigens.

C. OTHER GENES REGULATING IMMUNE REACTIONS

Rats respond to primary intraperitoneal injections of fresh egg white with hyperemia, pruritis, and edema of the extremities, which has been described as an “anaphylactoid” reaction (Selye, 1937). While studying similar reactions produced by dextran injections, Harris and West (1961) observed that more than one-fifth of the Wistar rats, which they obtained from the Agricultural Research Council’s Field Station at Campton, failed to produce this response. Nonreactivity was found to be inherited as a single gene recessive trait, and the symbols Dx for

124 DAVID L. GASSER

reactivity and dx for nonreactivity were assigned to the genes controlling this response (Harris et al., 1963). The trait did not appear to be immunologically specific, since nonreactors were unresponsive not only to eight different commercial dextrans, but to crude egg albumin, purified ovalbumin, and ovomucoid (Harris and West, 1963). This gene affected only mild reactions, since doses of antigen sufficient to cause lethal anaphylaxis in normal rats were also lethal in “nonreac- tor” rats, although there was a difference between the two groups in the time of survival after injection (Harris and West, 1963).

The response of cultured lymphocytes to mitogenic stimulation is one of the most widely studied parameters of immune function. Estab- lishing genetic variation in the magnitude of this response could be very useful in understanding the mechanisms involved and might even be helpful in defining subpopulations of lymphocytes, since it is known that lymphocyte subclasses can be distinguished by their ability to respond to various mitogens (Stobo, 1972; Schnebli and Dukor, 1972).

The first report of a genetic difference in the ability of rats to respond to mitogens involved a comparison between Lewis and BN strains (Newlin and Gasser, 1973). Peripheral blood lymphocytes of Lewis rats responded to both phytohemagglutinin (PHA) and concanavalin A (Con A) at a level that was approximately five times as great as the BN response. High response was dominant and in F2 and backcross generations it appeared to segregate as though predominantly controlled by a single major gene.

Williams et al. (1973) also studied the difference between Lewis and BN using cultured lymph node cells and to a large extent confirmed the findings of Newlin and Gasser (1973). The only discrepancy involved the response of the (Lewis x BN)F1 hybrids, which Williams and colleagues found to be intermediate between Lewis and BN. Gas- ser and Newlin (1976) later found that when they used culture conditions that were slightly different Prom those employed in their earlier experiments, the response of the FI hybrids was indeed intermediate between the two parental strains. Williams et al. (1973) favored the view that the difference between Lewis and BN was caused by multiple genetic factors, not by a single gene, but did not formally exclude the latter possibility.

In order to determine whether a single gene for high responsiveness could be isolated, we selected high responders from the (Lewis x BN)FI x BN backcross generation and mated them with BN partners. Single-gene segregation continued to occur throughout 10 generations of backcrossing. If high responsiveness were determined


by a number of genes with equal and additive effects, this result would not have been expected. The BC,o high responders on the average incorporate approximately three times as much tritiated thymidine as BN when stimulated with PHA or Con A. Thus the gene that we have isolated accounts for a major part, but not all, of the difference between Lewis and BN. This gene is not closely linked to any histocompatibility locus that we could detect, since 18/19 skin grafts from BC9 high responders survived for more than 100 days on BN recipients (Gasser and Newlin, 1976).

Genetic variation in the response of rats to streptococcal group A carbohydrate was studied by Stankus and Leslie (1975, 1976). Sprague-Dawley rats were immunized with group A streptococcal vaccine, and animals that gave especially high responses were mated with one another to derive a line of high-precipitin responders (HPR). A line of low-precipitin responders (LPR) was also derived. After four generations of selection, most of the LPR rats produced less than 1 mg of precipitating antibody per milliliter, whereas half of the HPR rats produced more than 10 mg of antibody per milliliter (Stankus and Leslie, 1975). The Fs generation of the HPR strain produced a mean precipitin response of 16.0 mg/ml, whereas the F7 LPR rats produced a mean response of 0.2 mg/ml (Stankus and Leslie, 1976). None of the inbred strains examined so far (F344, W/Fu, BN, Cop, Aug, M520, BUF) have produced a response greater than 0.5 mg/ml (Stankus and Leslie, 1976). The mode of inheritance has not yet been elucidated, nor have any genetic linkages been discovered.

VII. Immunoglobulin Genetics

A. KAPPA-CHAIN ALLOTYPES

The first report on rat immunoglobulin allotypes was by Barabas and Kelus (1967), who injected Wistar BB serum proteins into "black and white hooded" recipient rats. After 12 weekly injections, an antiserum resulted that precipitated donor serum in a 1% agar gel. Evidence from immunoelectrophoresis and Sephadex G-200 filtration suggested that the allotypic specificity was on y-globulins, but the class of immunoglobulin involved was not finally determined.

A search for rat allotypes was undertaken by Wistar (1969), who first injected Salmonella adelaide flagella into DA rats and then injected the antigen-antibody complexes into Lewis rats. Antiallotype antibodies were detected both by a microprecipitation assay using "'1- labeled rat IgG and by a type ofCoombs assay using Ig-coated SRBCs.

126 DAVID L. GASSER

An allotypic specificity was discovered that Wistar designated RZ-l for rat immunoglobulin. The antiallotype activity could be inhibited by both IgG and IgM antibodies as well as by purified light chains. Since it was shown by Hood et al . (1967) that the light chains of the rat, like those of the mouse, are predominantly kappa type, it appeared that RZ-I is a genetic marker for the kappa chain.

This work was continued by Gutman and Weissmann (1971), who developed antiallotype sera by injection into rats of rat antisera to pertussis complexed with pertussis. They were able to show that the inbred strains DA, ACI, and Fischer had the same RZ-l allele (designated RZ-l a), yet Lewis, BN, BUF, and W/Fu had an alternative allele designated RZ-l b. They also showed that RZ-l was not linked to Ag-B or to the coat color loci for hooded, agouti, or a I h i s m (Gutman and Weissmann, 1971).

Light-chain allotypes were also examined by Rokhlin et al. (1971), who described an allele designated RL 1 that was present in the MSU strain, and its alternative allele RL 2 in the WAG and August strains. These authors demonstrated that 90-95% of rat immunoglobulins possessed RL determinants, and since the predominant type of light chain in the rat is kappa (Hood et al., 1967), it was concluded that the RL allotypes are kappa-chain markers. When tryptic peptides of reduced light chains of RL 1 and RL 2 allotypes were compared, the results suggested that multiple site differences exist (Vengerova et al., 1972).

Light-chain allotypes were also studied extensively by Armerding (1971), who discovered two alleles that he named W-1 and SD-1. The SD-1 allele was found in Sprague-Dawley rats as well as the inbred strains DA, ACI/f Mai, and F344/f Mai. The W-1 allele was found in the following strains: Wistar AflHan, E3, BDE, Wistar BB, Lewis, August, BICR/M 1 R, BUF/f Mai, AVN/Cub, BN/Cub, CAP/Cub, LEP/Cub, PB/Cub, WAG/Cub, WPICub, AS/Max, ASYMax, BS/Max, HS/Max, and Hw/Max. These allotypes were shown to be on the light chain, and were present on IgM and two subclasses of IgG, immunoglobulins. It appears that the SD-1 allele is the same as the RZ-la gene of Wistar (1969) and Gutman and Weissmann (1971). The strain distribution of the RZ-l allele described by Gutman and Weissmann (1971) corresponds to that of the W-1 gene of Armerding (1971).

The most extensive survey of kappa-chain allotype alleles among inbred strains was reported by Beckers et al. (1974), who also introduced the fourth method of designating these genes in order to be consistent with nomenclature which they had previously proposed for a chains (Section VI1,B). Ik (la) of Beckers et al. (1974) corresponds to RZ-l (Gutman and Weissman 1971), W-1 (Armerding, 1971), and RL 2


(Rokhlin et al, 1971). The Ik (lb) allele of Beckers et al. (1974) corresponds to RZ-I” (Gutman and Weissman, 1971), SD-1 (Armerding, 1971), and RL I (Rokhlin et al., 1971). The results ofthe strain surveys reported by all of these groups are included in Table I.

Rat kappa chains have been sequenced by several groups, and some very interesting conclusions have emerged. The difference between D A (RZ-I”) and Lewis (RZ-I b, kappa chains involves one sequence gap in addition to 11 amino acid substitutions (Gutman et al., 1975), con- firming the conclusion of Vengerova et al. (1972) that these proteins differ at multiple sites. Since one of these genes could not have been derived from the other by a single point mutation, Gutman and colleagues referred to RI-I a and RZ-I as “complex allotypes,” and suggested that these genes exist in tandem rather than as true alleles. This type of mechanism had been proposed by Bodmer (1973), who suggested that serologically detected allelic differences may actually be the products of different closely linked genes and that genetic polymorphism controls which gene is expressed.

A second interesting feature revealed by the kappa sequences involves S211 protein of LOU rats. The LOUlWsl strain, which has been inbred since being separated from a Wistar stock in 1956, is unusually susceptible to the development of tumors that arise in the ileocecal lymph node area (Bazin et al., 1972). More than 10% of the LOU/Wsl rats develop a “leucosarcoma” or “ileocecal immunocytoma,” which usually appears when the animal is about a year old and kills the host a month later. These tumors are transplantable and can be maintained in uitro; 60% of them synthesize immunoglobulins (Bazin et al., 1972; Burtonboy et al., 1973). Most of the light chains produced by these tumors are of the kappa type, although two of 150 monoclonal immunoglobulins were shown to be homologous to the lambda chains of other species (Querinjean et al., 1973).

One of the proteins obtained from this strain, a kappa Bence-Jones protein designated S211, has been sequenced (Starace and Querin- jean, 1975), and compared with the sequences of DA and Lewis kappa chains (Gutman et al., 1975). Although the LOUlWsl and Lewis strains have been typed as having the same allele at the kappa allotype locus (Beckers et al., 1974), there are two amino acid differences between the S112 protein and Lewis kappa chains (Starace and Querinjean, 1975; Gutman et al., 1975). This suggests isotypic variation similar to that which occurs in human lambda chains (Ein, 1968; Gibson et al., 1971; Hess et al., 1971) and rabbit kappa chains (Appella and Inman, 1973). Therefore, there may be two closely linked loci coding for rat kappa chains that are so similar to one another as to be serologically

128 DAVID L. GASSER

indistinguishable, but have differences that can be revealed by amino acid sequencing. Further support for this conclusion was reported by Wang et al. (1976), who sequenced some additional kappa chains obtained from LOUNVsl rats. Two of these, IR 102 and IR 52, had as- paragine at position 138, whereas S211 had lysine at this position.

Starace and Querinjean (1975) also reported that a remarkable degree of similarity exists between the framework residues of the V regions of the rat S211 protein and the human ROY protein, a VKI light chain. The mouse MOPC 41 and rabbit light chains BS-5, BS-1, and 4135 also appear to be homologous to ROY. The similarity between the C regions of S211 and ROY was not especially close, which suggested that selective pressures to conserve the basic structure of the V region were stronger than those operating on the C region (Starace and Querinjean, 1975).

It has been shown by Hunt and Duvall (1976) that in most cases these kappa-chain markers are expressed exclusively by thymus- independent cells. PVG rats were thymectomized, irradiated, marrow-reconstituted, and allowed to recover. They then received an intravenous injection of thymocytes from a donor bearing the opposite allotype. They were subsequently immunized with SRBCs or dinitrophenylated bovine y-globulin. In 24 of 26 rats, there was no detectable allotype derived from the thymus cell donor present in the specific antibodies that were produced. Whether the thymocyte- specified allotype present in two of the rats was derived from contamination by parathymic lymph nodes or from some unorthodox release of the “wrong” allotype by the PVG cells has not been excluded (Hunt and Duvall, 1976).

B. ALPHA-CHAIN ALLOTYPES

As Badn and his colleagues continued their work on the remarkable LOU rats, they selected two histocompatible substrains from the original LOUNVsl stock. The LOU/C/Wsl line was selected for its high incidence of ileocecal immunocytomas and the LOUIMNVsl strain for its high rate of reproduction. By 1973, the incidence of immunocytoma in the LOU/C/Wsl strain was 23%, and that in LOU/M/Wsl was 2% (Bazin et al., 1973). Immunoglobulins of four classes and three subclasses were isolated from the sera of these rats: IgM, IgA, IgE, IgGl, IgGza, IgGzb, and IgGz, (Bazin et al., 1973, 1974a). Immunization of August rats with LOU IgA monoclonal proteins led to the development of antisera that recognized an alpha-chain allotype which was shown to be transmitted as an autosomal single-gene trait (Bazin et al., 1974b). The gene coding for the presence of the LOU allotype was


designated Ia( la), and the allele for absence of this marker was called Ia(1.). In most cases, 100% of the IgA immunoglobulins of Ia(la)/ Ia( la) homozygotes were positive for this marker, whereas about 50% of the IgA proteins of Ia( la)/Ia( 1.) heterozygotes were positive. The locus controlling this allotype was not observed to be linked to any coat color loci (Bazin et al., 1974b) or to the locus controlling kappa allotypes (Beckers et aZ., 1974). The strain distribution of alleles reported by these workers (Beckers et al., 1974) is included in Table I.

c . ALLOTYPES OF THE IGG2b CHAIN

Beckers and Bazin (1975) sensitized OKA rats with immunoglobulins from LOU/C/Wsl and LOU/M/Wsl donors and observed positive precipitin reactions. Since there was activity for light chains, the sera were absorbed with an IgGI-lambda monoclonal protein from a LOU immunocytoma. The absorbed serum had no activity for either kappa or lambda light chains, but was positive for fragments of LOU IgG,,. It had previously been shown that differential susceptibility to trypsin digestion could be used to separate the subclasses of rat IgG,: IgG,, was not digested by trypsin, whereas IgGzb was split into 3.5 S fragments (Nezlin et al., 1973).

An absorbed LOU antiserum specific for IgGzb chains of OKA immunoglobulins was also developed. Forty-six strains of rats could be classified as having the Iy2b( la) marker, precipitated by absorbed LOU anti-OKA serum, or the Iy2b(lb) marker, precipitated by absorbed OKA anti-LOU serum. In Fz and backcross generations, these traits were inherited as though determined by codominant alleles of a single locus that was closely linked to the alpha chain allotype locus, Ia( la). No recombinants were observed among 134 backcross rats (Beckers and Bazin, 1975).

D. IGE-SECRETING IMMUNOCYTOMAS

Although reaginic antibodies had been known to exist in the rat (Binaghi and Benacerraf, 1964) and to have properties similar to those of human IgE (Stechschulte et al., 1970; Jones and Edwards, 1971), this immunoglobulin was not available in large quantity until IgE- secreting tumors were identified (Bazin et aZ., 1974a). Bazin and colleagues reported that about 8% of all LOU/Wsl rats develop IgE- secreting immunocytomas. This discovery was especially important since IgE monoclonal proteins have never been reported in any other laboratory animal, and very few have been discovered in humans. The proteins had a molecular weight of about 183,000 and possessed class- specific determinants not found on IgM, IgA, or IgG. An antiserum

130 DAVID L. GASSER

specific for these determinants abolished the passive cutaneous anaphylactic activity of a rat serum containing reaginic antibodies to antigens of Nippostrongylus brasiliensis worms (Bazin et al., 1974a). Although allotypic variants have not yet been described which would make genetic studies feasible, these myeloma proteins have been useful in studying solubility differences between IgE and IgG (Carson et al., 1975), the number of IgE receptor sites on rat basophile leukocytes and peritoneal mast cells, and the equilibrium constant for the binding of IgE by these cell types (Conrad et al., 1975).

E. ANTI-RECEPTOR SITE ANTIBODIES

One of the most exciting developments in immunology within the last few years has involved the study of anti-idiotypic antibodies. The term “idiotypic specificity” was introduced by Oudin (1967) to designate an antigenic specificity that is unique in two respects: it is peculiar to antibodies against one given antigen and to the individual (or group of individuals that are closely related genetically) capable of producing it. By the use of highly inbred rat strains that can be made immune to rather well-characterized histocompatibility antigens of other inbred strains, it has been possible to raise antibodies ‘against idiotypic determinants characteristic of a given strain and a given anti-Ag-B antibody specificity.

Ramsier and Lindenmann (1969) first reported that the injection of parental strain lymphoid cells into F1 recipients induces the production of anti-receptor site (RS) antibodies. Anti-RS antibodies could also be produced against antisera. A (Lewis x DA)FI anti-(DA anti- Lewis) serum was produced by injecting into the F1 hosts serum from DA rats that had been grafted with Lewis skin (Ramsier and Linden- mann, 1971). All these experiments were dependent upon inhibition of polymorphonuclear cell migration mediated by a factor obtained from spleen cell cultures referred to as PAR, or product of antigenic recognition. These results have been confirmed by the use of other assays. Binz and Lindenmann (1971) utilized alloantibody-coated red blood cells as well as direct binding of iodinated anti-RS antibodies to lymphoid cells to demonstrate that alloantibodies and normal lymphoid cells compete for the same labeled antialloantibodies. Binz et al. (1973) demonstrated that such anti-RS antisera could inhibit local GVH reactions as assessed by the popliteal lymph node assay. McKearn et al. (1974) demonstrated that Fl rats can be made resistant to GVH disease by sensitization with parental lymphoid cells, that this resistance is age-dependent and radiosensitive, and that anti-RS antibodies can be obtained from the sera of these resistant rats.

These experiments did not establish whether the T-lymphocyte receptor was the same as the B-cell receptor. Evidence for this was presented in a subsequent paper by Binz and Wigzell (1975a). They


obtained an unusually high titer of anti-Lewis anti-DA antibody after 4-7 intraperitoneal inoculations of 25 x 10" column-purified Lewis T cells into (DA x Lewis)F1 recipients. This antiserum precipitated Lewis anti-DA in gel and had a very high activity against SRBCs coated with IgG from Lewis anti-DA serum. By indirect radioimmunoassay, the antiserum was shown to bind normal Lewis lymphocytes, but not DA or BN lymphocytes. This antiserum was able to inhibit both the MLC and GVH reactions of DA lymphocytes against (Lewis x DA)FI cells, but reactivity against (Lewis x BN)Fl cells was not significantly affected.

Lewis anti-DA serum covalently bound to Sepharose beads absorbed the capacity of this antiserum to react with T cells as measured by radioimmunoassays or by MLC or GVH reactions. Lewis T cells could also absorb the capacity of this antiserum to agglutinate SRBCs coated with Lewis anti-DA antibodies. These experiments demonstrate very convincingly that the T-lymphocyte receptor has the same idiotypic specificity as that which appears on anti-idiotypic serum antibodies (Binz and Wigzell, 1975a).

VIII. The Current linkage Map

The current linkage map of Rattus noruegicus is shown in Fig. 2. Some of the linkages shown were discussed in this review, and data on the remaining ones were taken from the following sources: Robinson

Et- 1

Ea-2. Es-4

€1-3

I( Ao-C

G1-1

h

a 7*b

-11 -n-5

FIG. 2. The current linkage map of Rattus noruegicus. Approximate distances in centimorgans are shown at the left. Linkages of uncertain order are bracketed. MHC, major histocompatibility complex.

132 DAVID L. GASSER

(1972), Gasser (1972), Gasser et al. (1973b), and Womack and Sharp (1976). The figure reflects several uncertainties that are yet to be resolved. In the first linkage group, H - 4 is shown as being distinct from the Pta or Ag-F locus, when in fact, it may be the same locus. In the seventh linkage group, only two regions of the MHC are shown, an SD region and an I region. Undoubtedly, there are as many loci in the MHC of the rat as in that of the mouse, but conclusive evidence so far available allows us to recognize only one SD locus and one I-region locus that stimulate MLC reactions (Section I1,C) (Butcher and How- ard, 1977).

Although the map looks very scanty now, one would hope that within a few years many additional rat genes will be mapped. Progress in this direction should be helped considerably by using mouse linkages as a guide. In her comparison of the mouse and rat karyotypes, Nesbitt (1974) specifically identified those portions of the mouse and rat genomes that appeared to be very similar cytologically. In the first two columns of Table IV are listed those portions of mouse and rat chromosomes that Nesbitt found to have the greatest similarity. In the third column are listed some of the mouse genes known to map in these regions. The homologous rat linkages established so far are shown in the last column.

The study of the comparative genetics of rats and mice is in its earliest stages. To what extent homologous linkages will mimic cytogenetic similarities remains to be seen. The classic studies on this problem in Drosophila have demonstrated that within cytologically homologous regions of chromosomes of D. melanoguster and D . pseudoobscura, the gene arrangements are very different indeed (Stur- tevant, 1940; Sturtevant and Novitski, 1941; Dobzhansky, 1951). One might therefore expect that representatives of two different genera, Mus and Rattus, would differ a great deal in their genetic linkages. Nevertheless, the segment of the rat linkage group V(Ag-C, Es-2, Es-1) that appears to be homologous to a portion of the eighth mouse chromosome is about 15 centimorgans long. Furthermore, the genes for pink-eyed dilution, albinism, and hemoglobin beta chains are linked in both species, even though the corresponding chromosomal regions do not appear to be very similar cytologically (Nesbitt, 1974). Further linkage information in the rat will be extremely useful in assessing the degree of chromosomal rearrangement that apparently occurred in the derivation of mammalian species.

ACKNOWLEDGMENTS I am deeply grateful to Drs. H. Bazin, D. Gotze, J . C. Howard, D. M. Lubaroff, K . F.

Mitchell, and W. K. Silvers for critically reading the manuscript and making helpful suggestions.

TABLE IV REGIONS OF CYTOLOGICAL HOMOLOGY BETWEEN MOUSE AND RAT CHROMOSOMES: COMPARISON WITH GENETIC LINKAGES

Rat chromosome Mouse chromosome and region" and region"

Mouse genes in this region"

Homologous linkages in the rat

R2, distal 68% M3, A2-H4 R3, distal 60% M2, C3-H3 Ea-6,1r-2, H-3, H-13, a , S o p a , Sop' R4, distal 72% M6, A3-G3 Ly 2,3; l g , Ldr-1 R5, distal 90% M4, A1-El Lo, Mup-1, b,Pgm-2, H-15, H-16, Gild-1 R6, distal 88% R8, entire chromosome M9, entire Thy-I, Mpi -1 , d , Mod-1, Ttf

M 12, M-F2

chromosome R2, proximal 32% M7, Al-C Es-8, H-22, Gpi-1, p . H 4 I

R7, all except proximal 8% and distal 17%

M14, A3-E5

R10, all except proximal M8, B3-D1 Ea-1, Es-1, Es-6, GOT-2, Es-2, Es-7, Ag-C, Es-2, Es-1 31% and distal 23% Es-5, H-19

R14, long arm M 17, A2-E 1 H-2 ( K , 1 , S , D); Tla SD and I subregions of MHC

" Nesbitt (1974); Nesbitt and Francke (1973). "Th' IS information . was compiled by comparison of recent linkage information published in the Mouse News Letter with the chromo-

'Gasser (1972). " The c,p,Hbb linkage group in the rat undoubtedly corresponds to the same group in the 7th mouse chromosome, but these genes

somal banding patterns of Nesbitt and Francke (1973).

are distal to the segments in the two species that appear similar cytologically.

P 3

c1 w w

134 DAVID L. GASSER

REFERENCES Aoki, T., McKenzie, I. F. C., Sturm, M. M., and Liu, M. (1974). Zmmunogenetics 3,291. Appella, E., and Inman, J. K. (1973). Contemp. Top. Mof. Zmmunol. 2, 51. Aptekman, P. M., and Bogden, A. E. (1955). Cancer Res. 15, 89. Armerding, D. (1971). Eur. J . Zmmunol. 1, 39. Armerding, D., and Rdewsky, K. (1970). Protides Biol. Fluids, Proc. Colloq. 17, 185. Armerding, D., Katz, D. H., and Benacerraf, B. (1974a). Zmmunogenetics 1,329. Armerding, D., Katz, D. H., and Benacerraf, B. (1974b). Zmmunogenetics 1,340. Atkins, R. C., and Ford, W. L. (1975).J. Exp Med. 141, 664. Bailey, D. W. (1963). Science 141, 631. Bailey, D. W., and Hoste, J . (1971). Transplantation 11,404. Balch, C. M., Dagg, M. K., and Cooper, M. D. (1976).]. Zmmunol. 117,447. Barabas, A. Z., and Kelus, A. S. (1967). Nature (London) 215, 155. Bazin, H., Deckers, C., Beckers, A., and Heremans, J . F. (1972). Znt. J . Cancer 10,568. Bazin, H., Beckers, A., Deckers, C., and Moriami, M. (1973).J. Natl. Cancer Znst. 51,

Bazin, H., Beckers, A., and Querinjean, P. (1974a). Eur. J. Zmmunol. 4 ,44. Bazin, H., Beckers, A., Vaerman, J. P., and Heremans, J. F. (1974b).J. Zmmunol. 122,

Beckers, A., and Bazin, H. (1975). Zmmunochemistry 12, 671. Beckers, A., Qubrinjean, P., and Bazin, H. (1974). Zmmunochemisty 11,605. Bell, J., and Paterson, P. Y. (1960). Science 131, 1448. Benacerraf, B., and McDevitt, H. 0. (1972). Science 175,273. Billingham, R. E., Hodge, B. A., and Silvers, W. K. (1962). Proc. Natl. Acad. Sci. U.S.A.

Binaghi, R. A., and Benacerraf, B. (1964). J . Zmmunol. 92, 920. Binz, H., and Lindenmann, J . (1971).J. E x p . Med. 136, 872. Binz, H., and Wigzell, H. ( 1 9 7 5 a ) ~ . E x p . Med. 142, 197. Binz, H., and Wigtell, H. (1975b).J. Exp. Med. 142, 1218. Binz, H., Lindenmann, J., and Wigzell, H. (1973). Nature (London) 246, 146. Bodmer, W. F. (1973). Transplant. Proc. 5, 1471. Bogden, A., and Aptekman, P. A. (1960). Cancer Res. 20, 1372. Bogden, A., and Aptekman, P. A. (1962). Ann. N.Y. Acad. Sci. 97,57. Bornstein, M. B., and Crain, S. M. (1971).]. Neuropathol. E x p . Neurol. 30, 129 (abstr). Bornstein, M. B., and Iwanami, H. (1971).J. Neuropathol. Exp. Neurol. 30,240. Boyse, E. A., and Old, L. J . (1969). Annu. Reo. Genet. 3, 269. Boyse, E. A., Miyazawa, M., Aoki, T., and Old, L. J. (1968). Proc. R. SOC. London, Ser. B

Boyse, E. A., Itakura, K., Stocked, E., Iritani, C. A., and Miura, M. (1971). Transplanta-

Brambilla, G., Cavanna, M., Parodi, S., and Baldini, L. (1970).Experientia 26, 1140. Burhoe, S. 0. (1947). Proc. Natl. Acad. Sci. U.S.A. 33, 102. Burtonboy, G., Bazin, H., Deckers, C., Beckers, A., Lamy, M. E., and Heremans, J. F.

Butcher, G. W., and Howard, J. C. (1977). Nature 266, 362. Callahan, G. N., Stroehmann, I., and DeWitt, C. W. (1974). Zmmunology 27, 1141. Cantor, H., and Boyse, E. A. (1975a).J. Exp. Med. 141, 1376. Cantor, H., and Boyse, E. A. (1975b).J. Exp. Med. 141, 1390. Carson, D., Metzger, H., and Bazin, H. (1975).J. Zmmunol. 115,561. Castle, W. E. (1947). Proc. Natl. Acad. Sci. U.S.A. 33, 109.

1359.

1035.

48, 138.

170, 175.

tion 11, 351.

(1973). Ear. J . Cancer 9,259.


Committee for a Standardii~d Karyotype of Rattus norwegicus. (1973). Cytogenet. Cell

Conrad, D. H., Bazin, H., Sehon, A. H., and Froese, A. (1975).J. Zmmunol. 114, 1688. Cotman, C. W., and Mahler, H. R. (1967). Arch. Biochem. Biophys. 120,384. Cramer, D. V., Shonnard, J . W., and Gill, T. J . (1974).J. Immunogenet. 1, 421. Davies, D. A. L., and Alkins, B. J . (1974). Nature (London) 247,294. Dewitt, C. W., and McCullough, M. (1975). Transplantation 19,310. Dickler, H. B., and Sachs, D. H. (1974).J. E x p . Med. 140,779. Dobzhansky, T. (1951). “Genetics and the Origin of Species,” 3rd ed. Columbia Univ.

Donaldson, H. H. (1924). “The Rat,” Mem. Wistar Inst. Anat. and Biol., No. 6. Wistar

Douglas, T. C. (1972). J. E x p . Med. 136, 1054. Douglas, T. C., and Dowsett, A. P. (1975).J. Immunol. 115, 283. Driscoll, B. F., Kramer, A. J., and Kies, M. W. (1974). Science 184, 73. Dunkley, P. R., and Camegie, P. R. (1974). Biochem. J . 141, 243. Dunkley, P. R., Coates, A. S., and Carnegie, P. R. (1972). Proc. Aust. Biochem. SOC. 5,37. Edelman, G. M., and Gottlieb, P. D. (1970). Proc. Natl. Acad. Sci. U.S.A. 67, 1192. Eichwald, E. J . , and Silmser, C. R. (1955). Transplant. Bull. 2, 148. Ein, D. (1968). Proc. Natl. Acad. Sci. U.S.A. 60,982. Elkins, W. L., and Palm, J. (1966). Ann. N.Y. Acad. Sci. 129,573. Eng, L. F., Chao, F.-C., Gerstl, B., Pratt, D., and Tavaststjema, M. G. (1968). Biochemis-

Eylar, E. H. (1970). Proc. Natl. Acad. Sci. U.S.A. 67, 1425. Fabre, J . W., and Morris, P. J . (1974). Tissue Antigens 4, 238. Festing, M., and Staats, J . (1973). Transplantation 16, 221. Foster, M., Petras, M. L., and Gasser, D. L. (1969). Proc. Int. Congr. Genet., 12th, 1968

Frenzl, B., Kien, V.,and Stark, 0. (1960a). Folia B w l . (Prague) 6, 121. Frenzl, B., men, V., Stark, O., Smetana, K., and Kraus, R. (1960b). Folia B i d . (Prague) 6,

Frenzl, B., =en, V., and Stark, 0. (1972). Anim. Blood Groups Biochem. Genet. 3,

Friedberger, E., and Taslakowa, T. (1928). Z. Immunitaetsforsch. 59,271. Gasser, D. L. (1972). Biochem. Genet. 6, 61. Gasser, D. L., and Newlin, C. M. (1976). Unpublished observations. Gasser, D. L., and Palm, J . (1972). Genetics 71, s19 (abstr.). Gasser, D. L., and Silvers, W. K. (1971).J. Zmmunol. 106, 875. Gasser, D. L., and Silvers, W. K. (1974). Adu. Zmmunol. 18, 1. Gasser, D. L., Newlin, C. M., Palm, J., and Gonatas, N. K. (1973a). Science 181, 872. Gasser, D. L., Silvers, W. K., Reynolds, H. M., Black, G., and Palm, J. (1973b). Biochem.

Gasser, D. L., Palm, J., and Gonatas, N. K. (1975).]. Immunol. 115, 431. Gershon, R. K. (1976). In “The Role of Products of the Histocompatibility Gene Com-

plex in Immune Responses” (D. H. Katz and B. Benacerraf, eds.), p. 318. Academic Press, New York.

Genet. 12, 199.

Press, New York.

lnst. Press, Philadelphia, Pennsylvania.

try 7,4455.

Vol. 1, p. 245 (abstr.);

135.

(Suppl. l), 24.

Genet. 10,207.

Gibson, D., Levanon, M., and Smithies, 0. (1971). Biochemistry 10,3114. Gill, T. J., 111, and Kunz, H. W. (1971).]. Zmmunol. 106,980. Gill, T. J., Kunz, H. W., Stechschulte, J., and Austen, K. F. (1970)J. Immunol. 105, 14. Gill, T. J., 111, Enderle, J., Germain, R. N., and Ladoulis, C. T. (1971).J. Immunol. 106,

1117.

136 DAVID L. GASSER

Goldschneider, I. (1975). Cell Immunol. 16, 269. Goldschneider, I . (1976). Cell. Zmmunol. 24,289. Goldschneider, I., and McCregor, D. D. (1973).J. E x p . Med. 138, 1443. Gonatas, N. K., and Howard, J. C. (1974). Science 186,839. Gonatas, N. K., Levine, S., and Shoulson, R. (1965). Ann. N.Y. Acad. Sci. 122,6. Gonatas, N. K., Gonatas, J. O., Stieber, A., Lisak, R., Suzuki, K., and Martenson, R. E.

Corer, P. A. (1937).J. Pathol. Bacteriol. 44, 691. Corer, P. A. (1938).J. Pathol. Bacteriol. 47, 231. Gottlieb, P. D. (1974).J. E x p . Med. 140, 1432. Graham, D. I., Gonatas, N. K., and Gasser, D. L. (1974). Lab. Inoest. 31, 24. Giinther, E., and Bhakdi-Lehnen, B. (1976). Proc. Leukocy. Cult. Conf, loth, 1975 p.

Giinther, E., and Rude, E. (1975).J. Immunol. 115, 1387. Gunther, E., and Stark, 0. (1977). In “The Major Histocompatibility System in Man and

Giinther, E., Rude, E., and Stark, 0. (1972). Eur.J. Zmmunol. 2, 151. Giinther, E., Rude, E., Meyer-Delius, M., and Stark, 0. (1973). Transplant. Proc. 5,

Gutman, G. A., and Weissman, 1. L. (1971). J . Zmmunol. 107, 1390. Gutman, G. A., Loh, E., and Hood, L. (1975). Proc. Natl. Acad. Sci. U S A . 72,5046. Harris, J. M., and West, G . B. (1961). Nature (London) 191,399. Harris, J. M., and West, G. B. (1963). Br. I. Phurmucol. Chemother. 20, 550. Harris, J. M., Kalmus, H., and West, G. (1963). Genet. Res. 4,346. Heber-Katz, E., and Wilson, D. B. (1976).J. Exp . Med. 143, 701. Heslop, B. F. (1973). Transplantation 15,31. Hess, M., Hilschmann, N., Rivat, L., Rivat, C., and Ropartz, C. (1971). Nature (London)

Hildemann, W. H. (1970). Transplant. Reu. 3,5. Hirschfeld, J. (1965). Science 148,968. Hood, L., Gray, W. R., Sanders, B. G., and Dreyer, W. J . (1967). Cold Spring Harbor

Howard, J. C., and Scott, D. W. (1972). Cell. Immunol. 3,421. Howard, J. C., and Scott,,D. W. (1974). lmmunology 27,903. Hunt, S. V., and Duvall, E. (1976). Biochem. SOC. Trans. 4,38. Ishii, Y., Koshiba, H., Yamaoka, H., and Kikuchi, K. (1976). J . Immunol. 117,497. Itakura, K., Hutton, J. J., Boyse, E. A., and Old, L. J. (1971). Nature (London), New Biol.

Itakura, K., Hutton, J. J., Boyse, E. A., and Old, L. J. (1972). Transplantation 13,239. Jones, J. M., and Feldman, J. D. (1975). Transplantation 19,219. Jones, V. E., and Edwards, A. J. (1971). Zmmunology 21,383, Katagiri, M., Natori, T., Tanigaki, N., Kreiter, V. P., and Pressman, D. (1975a).

Katagiri, M., Tanigaki, N., and Pressman, D. (1975b). Transplantation 20, 135. Kennedy, J. S., Till, J. E., Siminovitch, L., and McCullough, E. A. (196S).J. Immunol.

Kibler, R. F., Fox, R. H., and Shapira, R. (1964). Nature (London) 204, 1273. Kies, M. W., and Alvord, E. C., Jr. (1959). In “Allergic Encephalomyelitis” (M. W. Kies

Kies, M. W., Driscoll, B. F., and Seil, F. J. (1973). Science 179,689. Klein, J. (1975). “Biology of the Mouse Histocompatibility-2 Complex.” Springer-

(1974).Am.]. Pathol. 76, 529.

405.

Animals” (D. Cotze, ed.), p. 207. Springer-Verlag, Berlin and New York.

1467.

234,58.

Symp. Quant. B i d . 32, 133.

230, 126.

Transplantation 19,230.

96, 975.

and E. C. Alvord, Jr.. eds.), p. 293. Thomas, Springfield, Illinois.

Verlag, Berlin and New York.


Klein, J., and Shreffler, D. C. (1971). Transplant. Reo. 6, 3. Klein, J., and ShrefRer, D. C. (1972a). J. E x p . Med. 135, 924. Klein, J . , and Shreffler, D. C. (1972b). Tissue Antigens 2,78. Koch, C. (1974). Zmmunogenetics 1, 118. Kohne, D. E. (1970). Q. Rev. B i o p h y s . 3, 327. Komuro, K., Itakura, K., Boyse, E. A,, and John, M. (1975). Zmmunogenetics 1,452. Komblum, J. (!968).J. Zmmunol. 101, 702. Kien, V., and Stark, 0. (1972). Anim. Blood C r o u p s Biochem. Genet. 3, Suppl. 1, 85

Kien, V., Vesely, K., Frenzl, B., and hark, 0. (1960). Folio Biol. (Prague) 6, 333. Kien, V., Bi l i , V., Kriiakova, M., and Kienovi, D. (1972). Anim. Blood C r o u p s Biochem.

Kien, V., Stark, O., Bila, V., Frenzl, B., Kienova, D., and Krgiakovi, M. (1973).

Kunz, H. W., and Gill, T. J. (1974). J . Zmmunogenet. 1,413. Lafferty, K. J., and Jones, M. A. S. (1969). Aust. J . E x p . Biol. Med. Sci. 47, 17. Laird, C. D., McConaughy, B. L., and McCarthy, B. J. (1969). Nature (London) 224,

Lennon, V. A., Wilks, A. V., and Camegie, P. R. (1970).J. Zmmunol. 105, 1223. Levine, S., and Sowinski, R. (1975).J. Zmmunol. 114,597. Levine, S., and Wenk, E. J. (1961). Am. J . Pathol. 39, 419. Levine, S., and Wenk, E. J. (1965). Ann. N.Y. Acad. Sci. 122,209. Lipton, M. M., and Freund, J. (1953).j. Zmmunol. 71,98. Little, C. C. (1914). Science 40, 904. Lubaroff, D. M. (1973). Transplant. Proc. 5, 115. Lubaroff, D. M. (1976). Fed. Proc., Fed. Am. SOC. E x p . Biol. 35, 215 (abstr.). Lubaroff, D. M. (1977a). Cell. Zmmunol. 29, 147. Lubaroff, D. M. (1977b). Manuscript in preparation. Luderer, A. A., Maurer, P. H., and Woodland, R. H. (1976).J. Zmmunol. 117, 1079. McDevitt, H. O., and Benacerraf, B. (1969). Ado. Zmmunol. 11,31. McDevitt, H. O., and Sela, M. (1965).J. E x p . Med. 122, 517. McDevitt, H. O., and Sela, M. (1967).J. E x p . Med. 126, 969. McDevitt, H. O., and Tyan, M. L. (1968).J. E x p . Med. 128, 1 . McDevitt, H. O., Deak, B. D., ShrefRer, D. C., Klein, J., Stimpfling, J . H., and Snell,

McFarlin, D. E., Blank, S. E., Kibler, R. F., McKneally, S., and Shapira, R. (1973).

McFarlin, D. E., Blank, S. E., and Kibler, R. F. (1974).J. Zmmunol. 113,712. McFarlin, D. E., Hsu, S. C.-L., Slemenda, S. B., Chou, F. C.-H., and Kibler, R. F.

McFarlin, D. E., Hsu, S. C.-L., Slemenda, S. B., Chou, S. C.-H., and Kibler, R. F.

McKearn, T. J., Hamada, Y., Stuart, F. P., and Fitch, F. W. (1974). Nature ( L a d o n ) 251,

Marshak, A., and Wilson, D. B. (1976). Fed. Proc., Fed. Am. Soc. E x p . B io l . 35, 354

Marshak, A., Doherty, P. C., and Wilson, D. B. (1977). J . E x p . Med. (in press). Martenson, R. E., Diebler, G. E., and Kies, M. W. (1970).J. Neurochem. 17, 1329. Martenson, R. E., Diebler, G . E. Kies, M. W., Levine, S., and Alvord, E. C., Jr. (1972)J.

Micheel, B., Pasternak, G., and Steuden, J. (1973).Nature (London), New Biol. 241,221. Michie, D., and Anderson, N. F. (1966).Ann. N.Y. Acad. Sci. 129, 88.

(abstr.).

Genet. 3, Suppl. 1, 84 (abstr.).

Transp lan t . Proc. 5 , 1463.

149.

G. D. (1972).J. E x p . Med. 135, 1259.

Science 179,478.

(1975a).J. E x p . Med. 141, 72.

(1975b). J . Zmmunol. 115, 1456.

648.

(abstr.).

Zmmunol. 109,262.

138 DAVID L. GASSER

Miller, J. F. A. P., Mitchell, G. F., and Weiss, N. S. (1967). Nature (London) 214, 992. Mullen, Y., and Hildemann, W. H. (1972). Transplantation 13, 521. Murphy, D. B., and Shreffler, D. C. (1975).J. E x p . Med. 141,374. Nesbitt. M. N. (1974). Chromosoma 46,217. Nesbitt, M. N., and Francke, U. (1973). Chromosoma 41, 145. Newlin, C. M., and Gasser, D. L. (1973).J. Zmmunol. 110,622. Nezlin, R. S., Krilov, M. Yu., and Rokhlin, 0. V. (1973). Zmmunochemistry 10,651. Neilsen, H. E., and Koch, C. (1975). Scand. J . Zmmunol. 4,31. Ohno, S. (1976). Cell 7, 315. Old, L. J., Boyse, E. A., and Stockert, E. (1963).]. Natl. Cancer Znst. 31,977. Oldstone, M. B., and Dixon, F. J. (1968). Am. J. Pathol. 52,251. Ortiz-Ortiz, L., and Weigle, W. 0. (1976).J. E r p . Med. 144,604. Ortiz-Ortiz, L., Nakamura, R. M., and Weigle, W. 0. (1976).]. Zmmunol. 117,576. Oudin, J . (1967). Proc. R. Soc. London, Ser. B 166,207. Owen, R. D. (1962). Ann. N.Y. Acad. Sci. 97,37. Palm, J . (1962). Ann. N.Y. Acad. Sci. 97, 57. Palm, J. (1964). Transplantation 2,603. Palm, J. (1969). Transplant. Proc. 1, 82. Palm, J. (1970). Transplant. Proc. 2, 162. Palm, J. (1971). Transplantation 11, 175. Palm, J., and Black, G . (1971). Transplantation 11, 184. Paterson, P. Y. (1966). Ado. Zmmunol. 5, 131. Paterson, P. Y., and Bell, J. (1962).J. Zmmunol. 89, 72. Perlik, F., and Zidek, Z. (1974). Z. Zmtnunitaetsforsch. 147, 191. Querinjean, P., Bazin, H., Starace, V., Beckers, A,, Deckers, C., and Heremans, J. F.

(1973). Zmmunochemistry 10, 653. Raff, M . C. (1971). Transplant. Reo. 6, 52. Raine, C. S., and Bornstein, M. B. (1970).J. Neuropathol. E x p . Neurol. 29, 177. Ramsier, H., and Lindenmann, J. (1969). Pathol. Microbiol. 34,379. Ramsier, H., and Lindenmann, J. (1971).J. E x p . Med. 134, 1083. Ramsier, H., and Palm, J . (1967). Transplantation 5, 721. Reif, A. E., and Allen, J. M. V. (1963). Nature (London) 200, 1332. Reif, A. E., and Allen, J. M. V. (1964).J. E x p . Med. 120,413. Richter, C. P. (1954).J. Natl. Cancer Znst. 15, 727. Robinson, R. (1972). “Gene Mapping in Laboratory Mammals,” Part B, p. 198. Plenum,

Roboz-Einstein, E., Robertson, D. M., DiCaprio, J. M., and Moore, W. (1962). J .

Rokhlin, 0. V., Vengerova, T. I., and Nezlin, R. S. (1971). Zmmunochemistry 8, 525. Ruscetti, S. K., Kunz, H. W., and Gill, T. J., 111. (1974).J. Zmmunol. 113, 1468. Sarich, V. M. (1972). Biochem. Genet. 7 , 205. Schnebli, H. P., and Dukor, P. (1972). Eur. J . Zmmunol. 2,607. Seil, F. J., Rauch, H.C., Einstein, E. R., and Hamilton,A. B. (1973).J. Zmmunol. 111,96. Seil, F. J., Smith, M. E., Leiman, A. L., and Kelley, J. M. (1975). Science 187,951. Selye, H. (1937). Endocrinology 21, 169. Shevach, E. M., Paul, W. E., and Green, I. (1972).J. E x p . Med. 136, 1207. Shonnard, J. W., Cramer, D. V., Poloskey, P. E., Kunz, H. W., and Gill, T. J., 111. (1976).

Shreffler, D. C., and David, C. S. (1975).Ado. Zmmunol. 20, 125. Silvers, W. K. and Yang, S. L. (1973). Science 181,570. Silvers, W. K., Wilson, D. B. and Palm, J . (1967). Science 155, 703. Silvers, W. K., Murphy, G. and Poole, T. W. (1977). Zmmunogenetics 4, 85.

New York.

Neurochem. 9,353.

Zmmunogenetics 3, 193.


Simonian, S. J., Gill, T. J., 111, and Gershoff, W. N. (1968).J. Zmmunol. 101, 730. Simons, M. J., and Fowler, R. (1966). Nature (London) 209, 588. Simonsen, M. (1967). Cold Spring Harbor Symp. Quant. Biol. 32, 517. Snell, G. D. (1953).J. Natl. Cancer Inst. 14,691. Snell, G. D., Hoecker, G., Amos, D. B., and Stimpfling, J. H. (1964). Transplantation 2,

Snell, G. D., Dausett, J., and Nathenson, S. (1976). “Histocompatibility.” Academic

Soulillou, J.-P. Carpenter, C. B., d’Apice, A. J . F., and Strom, T. B. (1976).J. Erp. Med.

Stankus, R. P., and Leslie, G . A. (1975). Zmmunogenetics 2,29. Stankus, R. P., and Leslie, G. A. (1976). Zmmunogenetics 3 , G . $tarace, V., and Querinjean, P. (1975).J. Zmmunol. 115, 59. Stark, 0.. and Kien, V. (1976a). Folia Biol. (Prague) 13,93. Stark, O., and Kien, V. (196713). Folia Biol. (Prague) 13,299. Stark, O., Kien, V., and Frenzl, B. (1967). In “Polymorphismes biochemiques des

Stark, O., Frenzl, B., and Kien, V. (1968). Folia Biol. (Prague) 14, 169. Stechschulte, D. J., Orange, R. F., and Austen, K. F. (1970).J. Zmmunol. 105, 1082. Stenglein, B., Thoenes, G. H., and Giinther, E. (1975).J. Immunol. 115, 895. Stobo, J. D. (1972). Trunsplant. Reo. 11,60. Stroehmann, I., and DeWitt, C. W. (1972a). Zmmumlogy 23,921. Stroehmann, I., and DeWitt, C. W. (1972b). Zmmunology 23,929. Sturtevant, A. H. (1940). Genetics 25,337. Sturtevant, A. H., and Novitski, E . (1941). Genetics 26, 517. Szenberg, A., Warner, N. L., Bumet, F. M., and Lind, P. E. (1962). Br.J. E x p . Pathol. 43,

Thorsby, E . (1971). Eur. J. Zmmunol. 1,59. Vengerova, T. I . , Rokhlin, 0. V., and Nezlin, R. S. (1972). Zmmunochernistry 9, 1239. Wachtel, S. S., Gasser, D. L., and Silvers, W. K. (1973). Science 181,862. Wachtel, S. S., Koo, G. C., Zuckerman, E. E., Hammerling, U., Scheid, M. P., and Boyse,

Wachtel, S. S., Koo, G. C., and Boyse, E . A. (1975). Nature ( L a d o n ) 254,270. Wang, A.C., Fudenberg, H. H., and Bazin, H. (1976). Biochem. Genet. 14,209. Wekerle, H., Eshhar, Z., Lonai, P., and Feldman, M. (1975).Psoc. NatLAcad. Sci. U.S.A.

Williams, P. B. C., and DeWitt, C. W. (1976).J. Zmmunol. 117, 33. Williams, R. M., and Moore, M. J. (1973).J. E x p . Med. 138, 775. Williams, R. M., Moore, M. J.. and Benacemaf, B. (1973).J. Zmmunol. 111, 1571. Wilson, D. B. (1967). J. E x p . Med. 126, 625. Wilson, D. B., and Nowell, P. C. (1971).J. Erp. Med. 133, 442. Wilson, D. B., Blyth, J. L., and Nowell, P. C. (1968).J. Erp. Med. 128, 1157. Wilson, D. B., Howard, J . C., and Nowell, P. C. (1972). Transplant. Rev. 12, 3. Wilson, D . B., Heber-Katz, E., Sprent, J., and Howard, J. C. (1976). Cold Spring Harbor

Wisnieski, H. M. and Bloom, B. R. (1975).J. E x p . Med. 141, 346. Wisniewski, H. M., Prineas, J., and Raine, C. S. (1969). Lab. Znoest. 21, 105. Wistar, R. (1969). Zmmunology 17, 23. Womack, J. E., and Sharp, M. (1976). Genetics 82, 665. Wiirzburg, U. (1971). Eur. J. Zrnmunol. 1,496. Zeiss, I . M. (1967). Transplantation 5, 1393.

777.

Press, New York.

143, 405.

animaux,” p. 501. lnst. Natl. Rech. Agron., Paris.

129.

E. A. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 1215.

72, 1147.

Symp. Quant. Biol . 41,559.

Antigen-Binding Myeloma Proteins of Mice

MICHAEL POTTER

lobomtory of Cell Biology, National Cancer Institute, Bethedo, Maryland

1. Introduction 11. Structures of

B. VL: Classification of Sequences; CDR ................................ 147 C. V,: Classification of Sequences; CDR

A. Dinitrophenyl (DNP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 B. Phosphocholine (PC) . . . .

D. Clucan (a3C) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192

C . N-Acetylglucosamine . . ............... H. Flagellin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

F. Fructan-Inulin (GPlF) and Crass Levan (CP6F) ..................... 196

I . Lipopolysaccharides . . . . . . . . . . . . J . N-Acetylmannosamine .

VI. Concluding Remarks ....................... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

I. Introduction

The immunoglobulins are a large family of structurally related proteins secreted by plasma cells. Plasma cells are normally differentiated so that each cell is usually restricted to making immunoglobulin with only a single VL and VH region. Many thousands of possible VL-VH pair differentiations may develop in one organism. Each differentiated plasma cell can undergo normally a limited clonal proliferation. Thus, in normal serum no single molecular form with a single VL-VH pair is found in high enough concentration to permit isolation and structural characterization. Our knowledge of the chemical nature of immunoglobulin V regions has depended upon experimental methods for ex- panding clones to the point where sufficient yields of homogeneous immunoglobulin are available for analysis, or upon tumors of immunoglobulin-secreting cells. Some forms of immunization are associated with monoclonal responses of sufficient magnitude to permit isolation of homogeneous immunoglobulins (Krause, 1970, 1971; Braun and Jaton, 1974; Haber, 1971; Hamburg et al., 1976). Askonas et al. (1970) developed a method for propagating clones of memory cells in sublethally irradiated recipients, and when antigen was in-

141

142 MICHAEL POTTER

jected it was possible to expand clones sufficiently for isolation of limited quantities homogeneous antibody (see also Eichmann, 1972). The major quantitative source of homogeneous immunoglobulins (the myeloma proteins), however, are from spontaneous or induced plasmacytomas. This review will be concerned almost exclusively with myeloma proteins from one experimental and genetic origin, the plasmacytomas that are induced in BALBlc mice (Potter et al., 1975) by the intraperitoneal implantation of solid plastic materials or mineral oils and related substances (see Potter, 1975, for review and references of plasmacytomagenesis in the mouse).'

The homogeneous immunoglobulins (myeloma proteins) produced by plasmacytomas induced in one inbred strain reflect the immunoglobulin-producing potential of the individual. The biological question then arises: Are a series of sequentially collected myeloma proteins a stochastic, readout, of the total immunoglobulin-producing potential or are they a selected sample? If plasmacytomagenesis begins in B-lymphocytes and averts the normal differentiation process (in which antigen plays a role), then plasmacytomas could represent the immunoglobulin-producing potential of the virgin B-lymphocyte pool. If, on the other hand, plasmacytomagenesis begins in immunoglobulin-secreting cells whose precursors have interacted with antigen, then the myeloma proteins are only a sample of the total Ig-producing potential. Two approaches to characterizing BALB/c myeloma proteins have been made. First, myeloma proteins have been screened for binding activity with a series of selected antigens (Cohn, 1967; Eisen et al., 1968; Potter and Leon, 1968; Potter, 1970, 1971; Vicari et al., 1970; Sher and Tarikas, 1971). These antigens were available in the laboratory for one reason or another or because the antigens had been used to study normal immune responses. Many of these screening studies produced the active proteins that are the basis of this review. Screening studies have suggestively revealed selective trends, chiefly because groups of proteins that bind the same antigen or hapten have been found. These repetitions, appearing in what many would regard as still a relatively small sample,2 suggest that the myeloma proteins are in fact a nonstochastic sample of the Ig-

' Plasmacytomas can be induced in one other standard inbred stain, NZB, by the same methods (Warner, 1975), mice congenic to BALB/c that carry other allotypes: CB-20, CAL-20 and from recombinant inbred strains derived from BALB/c (see Potter et al., 1975). ' It is difficult to estimate the total number of myeloma proteins that have been

screened to all antigens, since different screens have used different groups of myelomas and different antigens. Possibly the true number is between 1000 and 1500.

ANTIGEN-BINDING MYELOMA PROTEINS OF MICE 143

producing potential. The second approach to the systematics of myeloma proteins from a single source is chemical characterization by amino acid sequencing.

As homogeneous immunoglobulins with antigen-binding activity became available, they were used in different types of studies. Im- munochemists have used them to study the specificity of antigen binding, and by the methods of X-ray crystallography have determined the structure of the antigen-binding site of one of the PC-binding proteins, McPC603. Advances in the comparative structure of V domains have made it feasible to propose reasonable hypothetical models of binding sites of other homogeneous immunoglobulins from primary structural data. As will be described in the section on DNPBMP, the combined use of primary structures, nuclear magnetic resonance, and electron spin resonance spectroscopy may make it possible to defhe very accurately the chemical structure, size, and shape of binding sites in solution of immunoglobulins that cannot be resolved by X-ray crystallography (Dwek et al., 1977).

Biologically, many AgBMP have provided prototypes of molecular species in normal antibody of the corresponding hapten-binding specificity. Similarities of AgBMP to corresponding antibody were first recognized by idiotypes and similar IEF properties but have recently been extended by primary structural studies. Comparative studies of immune responses to some of the antigens (that are recognized by AgBMP) in different inbred strains of mice have revealed polymorphisms in the antibody populations. The genes controlling some of these polymorphisms are linked to the allotype of the IgcH complex locus (for reviews, see Eichmann, 1975; Weigert et al., 1975; and Weigert and Potter, 1977). The AgBMP are useful in analyzing the nature of genes that control these polymorphisms.

The literature on antigen-binding myeloma proteins reflects the divergence of interests and the emergence of different areas of study. This review compromises with the different approaches and attempts to summarize relevant immunochemical information available on the most extensively studied groups of antigen-binding myeloma proteins; included are studies of the biology and chemistry of the antigens involved; of the primary structures of the binding sites in V regions; and, finally, of the biological relationship, when known, of the antigen-binding myeloma proteins to antibodies of the same specificity.

The review will begin with two structural aspects of myeloma proteins: ( I ) the three-dimensional structures of V-domains and antigen- binding sites and (2) the primary structures of VL- and VH-domain

144 MICHAEL POTTER

isotypes in the mouse. This material is essential to the discussion of the V regions of specific groups of hapten-binding myeloma proteins. Glaudemans (1975) has reviewed the AgBMP that bind carbohydrate antigens.

Comment. Very recently, Kohler and Milstein (1975, 1976) have greatly extended the potential for retrieving homogeneous antigen- binding immunoglobulins in the BALB/c mouse by fusing normal antibody-producing cells to plasmacytoma cells. The hybrid cells utilize the quantitative immunoglobulin-producing potential of the parent plasmacytoma and now produce both the myeloma protein and the homogeneous immunoglobulin, which can usually be readily separated by aflinity chromatography. The availability of these proteins will greatly expand our knowledge of Ig V regions and of antigen-binding activities not found in the myeloma population.

Abbreviations

1. Myeloma proteins that bind haptens are referred to by a prefix (= hapten) and a suffix -BMP (= binding myeloma protein). Prefixes used: DNP (dinitrophenyl); PC [phosphocholine (official term) or phosphorylcholine (commonly used term)]; a3G (a1 + 3 linked glucans); a6G (a1 + 6 linked glucans); P6Gal ( P l 3 6 linked galactans); GPlF (P2+ 1 linked fructans that have a terminal glucosyl residue); GP6F (P2 + 6 linked fructans that have a terminal glucosyl residue). Example: DNPBMP = dinitrophenyl-binding myeloma protein.

2. Miscellaneous antigens: Ag, antigen; Fla, flagellin; LPS, lipopolysaccharide. OA, ovalbumin.

3. Immunoglobulin structures: V-region polypeptides (about 120 amino acids) are classified by amino acid sequence as isotypes (see arbitrary definition in text). Three groups of isotypes are VK (V-kappa), VX (V-lambda), and VH (V-heavy); the isotypes are identified by an arabic number alone or with a letter. There are 28 VK isotypes VK-1, VK-2, etc., or VK-SlA, VK-21B. Many isotypes now listed by a number alone will subsequently be subdivided into other isotypes by complete sequences, as with the VK-21 isotype; hence the use of letters as well as numbers to designate an isotype.

4. The Wu-Kabat antigen complementarity-determining regions (CDR) of V regions are L1 (C23 to WSs), L2 (I48 to GS,), and L3 (C88 to F9$) for VL domains; and H1 (G26 to W,,), H2 to 161), and H3 (Ca to C,,,) for the VH domains using the sequence locations in McPC603. The remaining portions of the V regions are called the framework sequences, since they determine the folding, interaction, and conformation of the domain.


5. Idiotypes: Idiotypes are antigenic determinants located in the Fv (a globular part of the Ig molecule containing the VL + VH domains). Operational descriptive definitions used here are: IdI, individual idiotype (individual antigenic specificity unique to a specific protein of a given VL-V,-domain isotype combination); IdX, cross-speci fic or shared idiotype, found on more than one protein of the same VL-VH domain isotype combination or proteins containing the same VL- or V,-domain isotype. Example 1: Members of a group of proteins such as the 11 GPlFBMP all of the same isotype composition (VK-11 VH-5) may each be identified by an Id1 but also share one or more IdX determinants (see text). Example 2a: IdX determinants may be specific for only VL or V,; for instance, antisera specific for VK-21 have been developed, and these may identify V, determinants associated with different VH-domain isotypes. Example 2b: IdX determinants may be associated with closely related isotypes; e.g., the S63-Tl5 idiotype is an IdX found on TEPC15 (VK-22 VH-4A) and MOPC511

6. Single-letter amino acid code: A = Ala, B = Asx, D = Asp, E =

Glu, F = Phe, G = Gly, H = His, I = Ile, K = Lys, L = Leu, M = Met, P = Pro, Q = Gln, S = Ser, T = Thr, V = Val, W = Trp, Y = Tyr, and Z = Glx. Pca = pyrollidone carboxyl acid.

7. The sequence from amino terminus to CysZ3 in V L = NHz-CysZ3 peptide.

8. IEF = isoelectric focusing; NMR = nuclear magnetic resonance; ESR = electron spin resonance.

(VK-24 VH-4D).

II. Structures of BALB/c Mouse V Regions

A. V-REGION DOMAIN STRUCTURE: MODEL BUILDING

Understanding the V-region structure has been greatly advanced by the availability of three-dimensional models of immunoglobulin V domains and Fab fragments (Poljak et d., 1974; Segal et al., 1974; Epp et al., 1974; Edmundson et al., 1974). These models produced two valuable working principles: (1) the location of the antigen-binding surface, and (2) the conformational homology of the framework.

1 . Antigen-Binding Surface

Much of the V-region polypeptide sequence is involved in establish- ing the globular stability, the shape of the folded domain, and the critical interdomain interaction sites. Collectively, these parts of the sequence form the framework of the domain.

146 MICHAEL POTTER

Wu and Kabat (1970) and Kabat and Wu (1971) in an analysis of primary structures of V regions from different species identified subregions in the primary structures that varied the most in structure in immunoglobulins both between species and within species. They predicted that these regions supplied the antigen contacting amino acids. Three complementarity-determining regions (CDR) were defined in each V domain, H1, H2, and H3 in the V, and L1, L2, and L3 in V, (see Figs. 1 and 2). The three-dimensional models of V regions revealed that discontinuous CDR in the primary structure converted to form a continuous highly contoured surface roughly 30 x 40 A, facing the solvent (Poljak et al., 1974; Segal et al., 1974). Homologous CDR from different V-region isotypes may differ from each other in secondary structure. Secondary structural differences are caused by differences in length. The two most structurally variable CDR are L1 and H3 (see Tables V and VIII). The variations in these structures modify the reactive surface of the Fv, but also shape the interface region between the VL and VH domains.

2. Con formational Homology

Structural studies of antigen-binding myeloma proteins in the mouse have been greatly enhanced by the successhl solution of the three-dimensional structure of the phosphorylcholine-binding myeloma protein McPC603 (Segal et al., 1974). This structure provided a model of a mouse Fab region and a basis for comparison with a human Fab fragment (Poljak et al., 1974) and V domains (Epp et al., 1974). In a detailed comparison of the electron density maps of the V, domain of the mouse McPC603 at 3.1 A resolution, and the human V, domain of the human RE1 at 2.8 A resolution, Padlan and Davies (1975) demonstrated that the average relative displacement of a-carbon atoms of homologous amino acids was less than 1.5 A for residues in the framework regions. This result, which can be readily appreciated by comparing the a-carbon skeletons of the two domains (see Davies et al., 1975a,b), has revealed the important principle of the conformational homology of V, domains. It will be important to extend these observations with more examples to establish whether the principle applies to Vi and VH as well and whether conformational homology applies to all V,-domain isotypes in the mouse.

3 . Model Building The three-dimensional homology of the framework of Ig V domains

between distantly related species (such as mouse and man) has en- couraged X-ray CrystalIographers to build hypotheticaI models of V

ANTIGEN-BINDING MYELOMA PROTEINS O F MICE 147

domains from primary structural data. The major remodeling essential to building these hypothetical structures involves the Wu-Kabat CDR. Basically a new array of CDR as determined by primary structures and known CDR structures are added on the proved framework structure (for details see legend of Fig. 1). A basic assumption has been made (E. A. Padlan, personal communication) that two CDR of similar length have a strong probability of having a similar secondary structure. Second, many CDR can be hypothetically remodeled from longer homologous structures by excising amino acids and rejoining the chain. In many cases this is very feasible, since the differences involve loops or bends.

4 . Primary Structures of BALBIc Mouse V regions

The primary structures of the V,, and VH domains of the antigen- binding myeloma proteins in the BALB/c mouse provide information useful in the structure-function studies of antigen binding sites and in the understanding of the genetic control of immunoglobulins.

Fortuitously, the mouse is an especially advantageous species in which to study the genetic control of immunoglobulin structure. First, the inbred strains provide genetic uniformity and thus minimize (but may not completely eliminate) polymorphisms that could complicate the defining of an isotype. Second, V, chains in the mouse display a great variety of sequences in the accessible amino-terminal (framework) peptide extending to Cysz3 (Hood et al., 1970, 1973). The V,-domain isotypes comprise a large segment of the mouse VL chains since 97% of mouse L chains are of the kappa type (McIntire and Rouse, 1970). Characteristic NHZ-CysZ3 sequences thus provide a convenient means for identifying and classifying kappa-chain isotypes and may lead to a classification of corresponding structural genes. Third, many antigen-binding myeloma proteins are structurally and idiotypically similar to normal antibodies of the same binding specificity, which makes these myeloma proteins biologically relevant.

B. VL: CLASSIFICATION OF SEQUENCES; CDR

1. v, The NHz-Cysz3 sequences from V, chains derived from the BALB/c

mouse are so diverse that a characteristic sequence provides a useful identification of an isotype. Thus, all available V, sequences (Hood et al., 1976; Appella and Inman, 1973; Kabat et aZ., 197613; Potter et al., 1976a) were compared to a common hypothetical sequence (Tables I

148 M

ICH

AE

L

PO

TT

ER

................ cn

cncn

cni

v)c

nv

) ................

cl E

EE

................ c

nc

nm

cn

; cn

cn ................ 4

44

v) E

................ o

n

nio

cl

nn

n

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

i4c

ci4

ic

c4

4;4

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

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

............. b

iw

uc

li

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

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

Jc

lc

lJ

;

mc

nv

)icn

v)v

) v

);

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

c3 b

b

4

................ b

bH

b;

................ N

U

............. >;J cl cl;

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

cl iQQQ

+Q

QQ

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

................ >

>>

>;

2>

............. ;w w w

;

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

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

TABLE I BALB/c MOUSE VL REGIONS CLASSIFIED BY NHTCys,, PEPTIDE SEQUENCE"

VK Isotype:

............. iVK-1 i

iVK-2 ; iVK-26 i

;.x:3.. ; iVK-4 i iVK-5 i iVK-6 i VK-7 VKS

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

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

D I V M T Q S P A S L S V S L G E R V T I T C r) ....... ...... ...... ...... ..... ......

iTi I Li T i D i A i S i r

E; iEi i Li i I i T A iAi V K 4

i L i i I i M i A; P M 2 i E i V i Li Z i I i M i A; L S

n iv i ivi L iTi i Li P iDi Q A i S i . . j s i r

iVi V G iLi P

. . . .

v iVi i Ti iLi T T I i i p A is; L is ; i s ; ...... i D i ............ ...... ...... ........... ...... ...... ...... M

iE: . N . . ...... ...... .....

S M v "A- I K S S A R M S

A M S ........... ...... S ..... ...... i n - 9 i i Q i . . is i i A i S L

i s ; i Ai D I M i VK-I1 i

:A; D D S

i VK-10 i i Q i T T i Q i T T i A! . . . . . .

............ ...... ..... A V VK-12 Q

0 V T Q V

VK-13 Q P B Y A V T

I i. P .i

............ i VK-14 i i VK-15 i ............ VK-16 ............ i VK-17 i

VK-20 VK-21 VK-22 VK-23 VK-24

VK-25

N

...... ...... ...... T iKi F iM! T 0 S

;K; :Mi M iv; L ...... ...... ...... S E V G

...... . . 0 L V G ...... ...... ...... * z S Q i S i F iMi T V i D i i S i D

v I Z is i F A =I

7

i D i Z i si S N H K F M T V B S S E T T V M A 1 K

I A v A S T F A T A S K K S

L T T P D S S L S I B E L D P S S S I B E L K P S S S

...... ...... I ...... $

8 8 Q

g I 4

E A A F N P T T S A S S m

VA P A V V T Q Q S A o L T T S P G E T V T L T C 6 Isotype: F VA-1 VA-2 E E G

m " ( ) = amino acid not identified; 0 = no amino acid at this position; VK-isotypes circumscribed by dotted lines are related (see text).

The continuous letter sequences across the table are hypothetical sequences based on the most common amino acid found at that position in mouse kappa and lambda chains.

2 WJ

8

150 MICHAEL POTTER

and 11), and any sequences that differed from all others b y three or more residues were arbitrarily assigned a number. From this list 26 V,-domain isotypes (VK-1, VK-2, etc.) were defined. The purpose of assigning a number to a characteristic V, domain sequence is eventually to use it to enumerate and define V, structural genes.

The availability of complete V, sequences can further expand the number of isotypes. When complete VL-region sequences are available, differences of 5 or more amino acids will arbitrarily constitute the basis for designating a new isotype. For example, Gray et aZ. (1967) sequenced MOPC'IOE (VK-21). Subsequently, McKean et al. (1973a,b) sequenced 3 other VK-21 proteins, MOPC321, TEPC124, and MOPC63, and found each different from MOPC'IOE by 21 or 22 positions, which occurred in both framework and CDR segments. MOPC321 and TEPC124 and MOPC63 were more closely related to each other than MOPC70E b y having the same amino acids in 14 different positions (Table 111). MOPC321 and TEPC124, while differing from each other at only 3 positions, differed from MOPC63 at 8 positions. Thus, MOPC7OE is designated a VK-21A isotype, MOPC63 is a VK-SlB, and MOPC321 and TEPC124 are VK-21C.

The two proteins MOPC321 and TEPC124, which differ from each other at 3 positions, could both be synthesized from the same germ line structural gene; if so, it must be postulated that the differences arise from (1) somatic mutations or (2) a cryptic polymorphism now existent in the BALB/c inbred strain. The genetic basis for defining structural V genes is not yet available; hence the use of the term isotype.

Recently, Weigert and Riblet ( 1976) have found IdX-idiotypic antisera that are specific for VK-21 isotypes. These antisera can detect that IdX on whole Ig in both normal serum and myeloma proteins. This method may make it possible to quantitate the proportion of immunoglobulins in normal serum expressing the isotype. Examination of the partial sequences in Table I reveals structural relationships among different VK isotypes; e.g., (1) VK-1, VK-26, VK-2, and VK-3 share Vz, T7, Ls, D17, S20, and SZ2; (2) VK-4, VK-5, and VK-6 share El, Ls, Ilo, and Alo; (3) VK-9, VK-10, VK-11 share Q2, Ss, A13; (4) VK-14 and VK-15 share Ky, Mll, and Vls; and (5) VK-17 and VK-18 share Ss, Mll, D17, and Szo. These may represent relationships that are somewhat analogous to those found in the human V, subgroups.

The listing of myeloma protein L chains according to V, and Vh domain types is given in Table 11; as may be seen in 17 of the cases, more than one protein with the same sequence has been found. Hap- ten- or antigen-binding activities associated with V, and Vh are also


TABLE I1 V,-DOMAIN ASSIGNMENTS

~ ~~ ~ ~ ~ ~~~

Iso- An tigen-binding Refer- type Myeloma protein activity ences

VK- 1 VK-26 VK-2 VK-3 VK-4

Vk-5 VK-6 VK-7 VK-8

VK-9 VK- 10 VK-11

VK-12 VK- 13 VK-14 VK- 15 VK-16 VK-17 VK- 18 Vk-19 VK-20 VK-21A VK-2 1 B VK-21C VK-22 VK-23 VK-24 VK-25 VA- 1 VX-2

MOPC460, MOPC467 SAM M368G McPC843, McPC674 W3129, W3434 S117,J539, JPC1, SAPClO.TEPC601,

MOPC29 TEPC29 LPC 1 McPC603, McPC870, MOPC384

TEPC191, XRPC24, XRPC44

MOPC41, MOPC320 MOPC173, UPCIO, ABPC48, Y5476 EPClO9, J606, ABPC4, UPC61,

W3082, AMPC1, TEPC957 MOPC31C, TEPC173, XRPC23 MOPC149 HOPC5 MOPC2l MOPC47A, MOPC47B, McPC611 MOPC157 MePC600 MPCll McPC773, MOPC265 MOPC7OE MOPC63, ABPC22 TEPC124, MOPC321 HOPC8, TEPC15, S107 MOPC46, MOPC172, S23, XRPC25 MOPC167, MOFC5ll SAMM368A MOPCI04E,/558, HOPC1, etc. MOPC315

DNP, Fla - - a6G pGGal, GlcNac

- PC, Salmonella tel

aviv LPS - Cp6F GplF

- PC DNP PC, c

a3G, - DNP

-

" Hood et ul . (1973). bMorse et al. (1977). ' Appella and Inman (1973). ' I Hood et al. (1976).

' Schiff and Fougereau (1975).

I' Svasti and Milstein (1972).

Rudikoff et al . (1973).

Potter et a l . (1976a).

'Smith (1973) [see also Kabat et al. (1976b)J. (MPC11 has 12 residues that precede sequence shown here.)

Barstad et 01. (1974b). Appella (1971).

' D. J . McKean, personal communication.

"' Weigert et ul. (1970). " Schulenberg et al. (1973).

TABLE I11 VK-21 ISOTYPES"

L1 L2 L3 M70 M63 -N S-Y-B-H Y G G Z R-T-T-LV-A- A-T B-B-B M321 K-T-Y-N-Z Y R-LZ---I-R-T-T-LV-A-V-T B-B- S T124 W-Y-B-Z Y R-L-Z- -1-R-T-T-B-V-A-V-T B-A S

VK 21A 21B 21c 2 1 c

" Data of McKean et ol. (1973a,b) summarized from Kabat et ol. (1976b) and personal communication from McKean,


shown. In the VK-1, VK-4, and VK-8 isotypes, different hapten- or antigen-binding specificities are associated with the same VK isotype. In VK-10 and VK-23 only the italicized listing has the hapten-binding specificity. Similar hapten-binding specificity may be associated with different V,.; DNP-binding proteins may be VK-1, VK-23, or Vh-2; phosphorylcholine (PC) proteins may be VK-8, VK-22, or VK-24.

2. VA

Several VA regions have been completely sequenced (Appella, 1971; Schulenberg et ul., 1971; Dugan et al., 1973; Cesari and Weigert, 1973). Many others have been tentatively sequenced by analysis of the amino acid composition and chromatographic properties of the tryptic and chymotryptic peptides (Weigert et al., 1970; Cesari and Weigert, 1973). Unlike the kappa system in the mouse, which has thus far been associated with only single C, gene, there are clearly two separate CA genes: CAI and CA2 (Schulenberg et al., 1971; Dugan et al., 1973). The CAI and C, peptides differ from each other in 29 of 108 positions (Dugan et al., 1973). It is not known whether the 2 CA genes are linked. Further, the available sequences also do not reveal whether known VX isotypes can be joined to either the Ch isotypes during differentiation.

Eighteen VA regions are listed by Kabat et aZ. (197615). Weigert and Riblet (1976) have recently added others, including several from NZB mice. In contrast to the many VK isotypes, there appear to be only two Vh isotypes: VA-1 and Vh-2. Each of these isotypes is associated with the corresponding Ch-1 and CA-2 isotype. Vh-2 is thus far represented only by the MOPC315 Vh (Tables I and 11). Partial sequence analysis of the 21 Vh-1 isotype regions has revealed 12 that have an identical sequence called the Vh, sequence; the others differ from this prototype by no more than 3 amino acids, and all of these occur in CDR regions (Weigert et al., 1970, Cesari and Weigert 1973, Weigert and Riblet 1976) (Table IV).

The MOPC315 Vh-2 is the only sequence available in this isotype, and this differs from the common Vh-1 in 14 positions in both framework and CDR (see Table I11 for comparison of CDR differences). The N H ~ - C Y S ~ ~ Vh sequences are given in Table I, and the assignments of myeloma proteins in Table 11. Antigen binding activities thus far associated with Vh isotypes are: DNP with MOPC315 Vh-2; a3G with Vh- 1; and very recently 4 hydroxy-3-nitrophenylacetyl (NP) with Vh-1 (Jacket al., 1977; 0. Makela, M. Potter and M. Weigert, unpublished observations). Only three whole proteins (MOPC104, 5558, and UPC102) have a3G binding activity; others containing the A,,

12 h." R20 S 176 S 178 Y5606 YFi444 HN20

M31LE-E-G

VA-1

$ L N N N n m

r \r

8 T> 4

m F-R---F-V-Vh-2 Y

C 4

I S-D \'-D-D-M


sequence are inactive (Cesari and Weigert, 1973). The evidence then indicates Vh-1 isotype interacts with other VH to form Ig with different binding specificities.

3 . V, CDR

A list of all available VL Wu-Kabat CDR are given in Table V. These are listed according to the VK or Vh isotype. The L1 CDR is the most variable, both in primary and secondary structure. L2 and L3 regions differ thus far only in primary structure. The size and presumably the secondary structure of L1 is the same for proteins within a Vh isotype. Primary structural variations do occur within the same Vh isotype, and these can be nonconservative, i.e., involve polar versus nonpolar amino acid side chain groups. The association of characteristic L1 length with the NH2-Cy~23 framework sequence is evidence in favor of coevolution of both features as modifications developing from mutations affecting a single structural gene (Potter et al., 1976a).

C. VH: CLASSIFICATION OF SEQUENCES; CDR

VH-domain isotypes have less variety in the amino-terminal sequence than V, ones. The segment NH2-PheZ7 is part of the VH framework and therefore is used as a basis for classification (Table VI). Three differences to Phe27 are arbitrarily used as the minimum for assigning a VH isotype number. Some confidence in this assignment is obtained when more than one protein can be found with the characteristic sequence. Fortunately, many complete VH-domain sequences are available and provide additional evidence for defining isotypes. Five differences in the complete VH sequence arbitrarily comprise the basis for a new isotype, and this is done with a letter. Seventeen different VH isotypes have thus far been defined. (Table VI); the assignments of myeloma proteins to these isotypes is given in Table VII. In addition, the hapten-binding activity of the myeloma protein is also shown in Table VII. Hapten-binding activities appear to b e associated with a given isotype or to cluster about closely related isotypes. For example, VH-1B is associated with P6Gal binding; VH-1C with GP6F binding; VH-4A,-4B7-4C,-4D,-4E with PC binding; VH-5 with GPGF; VH-7 with a3G; and VH-8 and VH-9 (which appear to have some relationship to each other) with DNP binding. These relationships suggest a dominant role for the VH in determining antigen-binding specificity. In cases where there are specific VL and VH isotypes associated with a specific binding activity, e.g., VH- 1B-VK-4; VH-5-VK- 11, formation of antibody of that specificity may depend upon precise VL-VH pairing, and possibly no other gene products can create Ig of

Y

TABLE V WU-KABAT COMPLEMENTAFUTY-DETERMINING REGIONS OF MOUSE VL DOMAIN

Iso- L1 L2 L3 ......................................... ................ . . . . . . . . . . . . . . . . . . . type c*3.. .w,, I , s . . Gs7 C88. , F,,

VK-1 M460 R S S Q L V H S T B G B . . Y L H p VK-26 S368G R S S Q S I V H S B G B T . Y L Z

J539 S A S S S V S S . . . L H X24 S A S S S V S Y . . . . . . . M H

M . . . . VK-4 S117 A A ? ? S V S Y . . . L B r . . . . 3

X44 S A S S S V S Y . . . M H 3 . . . . M

VK-8 M603 K S S Z S L L B S G B Z K B F L A !a VK-9 M41 R A S Q B I G S L . . . . . . S B VK-10 M173 S A S Q S I G N . . . . . . Y L B Y Y T S S L H S Q Q Y S K L P R T

5476 R A S Z B I S B . . . . . . Y L B U10 R A S Z B I S B . . . . . . Y L B

VK-11 El09 Q A S Q G T N I N . . . . . . L N AM1 Q A S Q G T N I N . . . . . . L N A4 Q A S Q G T N I N . . . . . . L N U61 Q A S Q G T S I N . . . . . . L N A47N Q A S Q G T S I N . . . . . . L N 3082 Z A S K G T S I B . . . . . . L B JSOS Z A S Z G T B I B . . . . . . L B

VK-15 h121 K A S E N V V T . . . . . . Y V S Y G A S N R Y T G Q G Y S Y P Y T VK-19 M11 K A S Q B V S T T . . . . . . V A Y S A S Y R Y T Q Q H Y S T P ? T VK-20 M265 I T S T B I B G . . . . . . . L H

M773 I T S T B I B G . . . . . . . L H VK-2lA M70 R A z S Q S V B B S G I S . . F M N Y A A S B Q G S Z Z S K E V P W T * VK-21B M63 R A S Z S V B S Y G B S . . F M Z Y L A S N L Z S Z Z B B Z B P W T 3 VK-21C M321 R A S K S V N T Y G N S . . F M Z Y R A S N L Z S Z Z S B Z B P W T '

T124 R A S Q S V B W Y G B S . . F M Z Y R A S N L Z S Z Z S B B A P W T 7 H 8 T A S Z S L Y S S K H K V H Y L A S107 T A S Z S L Y S S K H K V H Y L A

X25 R A S Q S I S N N . L H

R M511 R S S K S L L Y K D G K T . Y L N Y L M S T R A S Q E L V E Y P L T r

VK-22 T15 T A S Z S L Y S S K H K V H Y L A

VK-23 M46 R A S Q B I S N N . . . . . . L H c1 . . . . . 5

VK-24 M167 R S S K S L L Y K D G K T . Y L N

VK-25 368A R S S K S L L H S B G ? T . Y -- VA-1 M104' R S S T G A V T T S N . . . Y A N G G T N N R A P A L W Y S N H W V

S178 R S N T G A V T T S N . . . Y A N C N T N N R V P A L W Y S B R W V 5606 R S N T G A V T T S N . . . Y A N G G T N N R A P A L W Y S N H W V 5444 R S S T G A V T T S N . . . Y A N G G T N N R A P A L W C S N H W V 3 2020 R S T T G A V T T C N . . . Y A N G G T N N R A P A L W Y S N H W V 5

vl Vh-2 M315 R S S T G A V T T S N . . . Y A N G G T N N R A P A L W F R B H F V

TABLE VI V"-DOMAIN ISOTYPES, BALB/c MOUSE"

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Isotype: E V K L E S G G G L V Q P G G S L K L S C A A S G F VH

VH-1A VH-1B VH- 1C VH-1D VH-2 VH-3 VHlA V H l B VH-4C VH-4D VH-4E VH-5

L

L i5 $ k 8

F

L P L

V Z R T V R T V R T V R T V R T

v v R T E P M V

- - v V T K

r

4 4

R ..... p ..... ivi i q Q iP i E i K i A V M K Y

..... ..... V Y .......... ..... M A ..... K .......... ivi iUi .9. :pi E i K i

VH6 .9.. V

VH-8 D '..'. i Q i *..** i Q i i Pi i K j is; Q IS: iTi ' S i iV i F Y .... ..... : S ; iTi : S i ivi T S i K i Z T : ..... ..... .......... VH-9 i9.i ;pi i P i s

VH- 10 v v T V A R ..... S

..... A V ..... .....

VH-7

M

a The continuous letter sequence across the top of the table is hypothetical and based on the most common amino acid found at that position. Dash means no sequence determined.

TABLE VII ASSIGNMENT OF PROTEINS TO VH ISOTYPES

I so- w e

~

Myeloma proteins

Antigen- binding Refer- activity ences

VH-1A VH-1B

VH-1C VH-1D VH-2 VH-3 VH-4A VH-4B VH-4C VH-4D VH-4E VH-5

VH-6 VH-7 VH-8 VH-9 VH-10

S117 JPC1, SAPC10, CBPC4, TEPC191, J539,

UPC10, Y5476, ABPC48 MOPC173 SAPCl5 MOPC47A TEPC15, HOPC8, S107, S63 W3207 McPC603 MOPC511 MOPC167 W3082,J606, UPC61, ABPC47N, AMPCl,

MOFC21 MOPC104E, J55S MOPC315 MOPC460 (XRPC25) S 176

TEPC601, XRPC24, XRPC44

EPC109, ABPC4

GI cN ac PGGal

Gp6F -

PC PC PC PC , c PC,C

Gp6 F

- a3G DNP DNP 5AU

a b c

d e f g a,h a i j h k

1 a m n a

" Hood et al. ( 1976).

' Potter et al. (1976h).

' Bourgois et al. (1972); Rocca-Serra et

f S. Rudikoff, personal communication.

Rudikoff et al. (1973).

Barstad et al. (1974a).

al. (1975).

Robinson et al. (1974).

" Rudikoff and Potter (1976). Rudikoff and Potter (1974).

j E. Appella, in Kabat et al. (1976h). ' Vrana et al. (1977). ' Milstein et al. (1974). See, also, Kabat et

"I Francis et al. (1974). " Barstad et al. (1974a).

al. (1976b).

similar binding specificity. In such cases as PC and DNP binding, several different VL structures can participate with closely related VH isotypes.

The CDR of all VH regions thus far sequenced are listed in Table VIII. Differences in length are found in all three VH CDR, but clearly the most striking variations in CDR structure in VH occur in H3. The L1 and H3 regions appear to be the most important for determining the variations in surface contour in the Fv region; however, more subtle changes in H1 and H2 are also contour determining. The effect of secondary structural changes can be visualized by examining the stereo tracings of the a-carbon backbones of the McPC603 and two other hypothetical models in Figs. 1 and 2.

TABLE VIII COMPLEMENTARXTY-DETERMININMINING REGIONS OF VH DOMAIN F m

6 Iso- H1 H2 H3 type c,. .............. ..Wae I* ..................................... I61 c,. .......................... CI,,

VH- L4 VH-1B

VH- 1C

VH-1D VH-4A

VH-4B

VH-4C

VH-4D

VH-4E

VH-5

VH-6 VH-7

VH-8 VH-9

S117 J539 x24 X44 u 10 5476 M173 T15 S 107 H8

3207

M603

M511

M 167

3082 J- A47N E 109 U6 1 M21 M 104 J558 M315 M460

F B F S R Y W M E . F D F S K Y W M S . F D F S R Y W M S . F D F S R Y W M S . F D F S G Y ? M S . F D F A T Y W M S . F D F S R Y W M S . F T F S D Fi'YiMiEi. F T F S B FIYiMiEi.

..........

F T F s B F~Y~M~E;. F T F s B F~Y~M!E!.

F T F s B F~Y~M~E:. F T F s B F ~ Y ~ M ~ E ; . F T F s D F:Y~M~E~.

. . . .

. . . .

. . . .

. . . .

.......... F T F S B Y W M B . F T F S B Y ? M B . F T F S N Y W M N . F T F S N Y W M N . F T F S N Y W M N . F T F S S F G M H . Y T F T ? Y Y M T . Y T F T B Y Y M K . Y S I T S G Y F W N S B I T N G Y M B

G E I H P D S G T I . . N Y T P S A R L H Y Y G Y N A . . Y W

G E I D P N S S T I . . N Y T P S A R S P Y Y A . . . . . . M

A A S i R i N K A N D Y T T i E i Y S A S A R D . Y Y G S S Y . i W i Y F A A SiRiNKANDYTTiEiY S A S A R D . Y Y G S N Y . i W i Y F

...... 5 A A s ~ R ~ N K A N D Y T T ~ E ~ Y s A s

A A S ~ R ~ N K A N D Y T T : E ~ Y S A S A R N . Y K Y D L V . ~ W ~ Y F . . p

A A S ~ R ~ N K G N K Y T T ~ E ~ Y S A S A R N . Y Y G S T . .iwiu~ . . ;4

8 A A S ~ R ~ D K A N D Y T T ~ Z : Y S A S A R D G D Y G S S Y . ~ W ~ Y F

A A S ~ R ~ S K A H D Y R T ~ E ~ Y s A s T R D A D Y G D s Y F ~ G ~ Y F ia

A R D . Y Y G s s Y .iwiy F

. . . . M

. . . . . . . .

. . . . . . . . cl . . . . . .

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

A E I R L K S H N Y A T H Y A E S S T G . . . . . . . . . . F A E I R L K S H N Y A T H Y A E S S T G . . . . . . . . . . F A E I R L K S H N Y A T H Y A E S S T G . . . . . . . . . . F A Y I S S G S S T L . . D Y A H T A R H C N Y P W Y A . . . M

C F I K Y D G S B . . . Y G B P S A G D N D H L Y . . . . . F

FIG. 1. Stereo drawing of the carbon skeletons of the Fv regions (V, and VH domains) of McPC603 (top), MOPC315 (center), EPClO9 (bottom). The McPC603 structure was determined from a 2.8 8, electron-density map (Segal et al., 1974); the MOPC315 and EPClO9 structures are hypothetical models. This figure should be viewed with a pocket-size stereoscope which may be obtained from Hubbard Scientific Co., P.O. Box 105, Northbrook, I l l . 60002, or from Abram Instrument Corp., 606 E. Shiwassee St., Lansing, Mich. 48901, or other sources.

The MOPC315 structure was constructed as follows: The complementarity- determining regions (CDR) of McPC603 V, were used as prototypes for the MOPC315 with the exception of the L1 structure, which has taken from the Mcg lambda chain (Schiffer et al., 1973). The L1 region of lambda chains has a helixlike structure unlike that in McPC603. Since the MOPC315 Hchain CDR all differ from McPC603 in length, remodeling of McPC603 was done. The MOPC315 H1 is one residue longer than McPC603; the additional Gly residue was added at the beginning of H1 (see Fig. 2) in the part of the segment exposed to the exterior; this permitted the Fa and N,, side chains to project into the cavity. H2 of McPC603 is 3 residues longer than MOPC315; 3 amino acids from the amino-terminal end of the McPC603 loop were excised (see Fig. 2), and the remaining segment was remodeled to correspond with the NEW (Poljak et al., 1974) H2 region. H3 of MOPC315 is 2 amino acids shorter than McPC603; the C-terminal part of H3 was made to correspond to the structure in McPC603, which allowed the bulky hydrophobic residues to project into the binding cavity. The part of H 3 that was exposed in McPC603 was excised.

The hypothetical model of the GplF-binding myeloma protein ABPC47N was built by E. Padlan and M. Vrana, NIH, who kindly made this available. The VL region was that of McPC603 except for L1 for which the human V, RE1 (Epp et ol., 1974) structure was used, which has the same length as ABPC47N L1. The VH region of McPC603 was again used; only H 3 had to be remodeled to form the structure created by the deletion of the sequence of 8 amino acids, between Amss and Phelw in McPC603; this was ac- complished by joining Amss to Phe,,, by a simple bridge. The deletion exposes the framework of the bend from which H 3 is formed. A large cleftlike groove is created across the binding surface, which is formed by H 1 and H2 on one side and L1, L2, and L3 on the other. L2, which is not involved in McPC603 or MOPC315, can potentially participate in binding.

162 MICHAEL POTTER

FIG. 2. Binding-site region of McPC603 (above) and hypothetical construction if the MOPC315 binding site (below) showing only the six complementarity-determining regions (CDR): L1, L2, L3, H1, H2, and H3. 0 = amino acids whose side chains contact hapten. @ = affinity site-labeled amino acid. The McPC603 binding site is a com- puterized printout of the carbon skeleton.

Ill. Groups of Myeloma Proteins That Bind the Same Haptens

A. DINITROPHENYL (DNP)

1 . General Characteristics, Background

Prior to 1967 a considerable body of immunochemical information was available on the binding properties of anti-DNP antibodies of several mammalian species. The structural basis of DNP binding, however, was not known, owing largely to the unavailability of a homogeneous DNP binding immunoglobulin. In most species the humoral response to DNP-protein immunization was highly heterogeneous, particularly among the late antibodies. Thus Eisen et al. (1967) began screening human myeloma proteins for DNP-binding activity


and found the BRY protein. This led to a similar search in the mouse system, resulting in the finding of two DNPBMP: MOPC315 (Eisen et al., 1968) and MOPC460 (Jaf€e et al., 1969). Initial screening was done by testing proteins for a spectral red shift (difference absorption spectrum at 450 nm) upon ligand binding (Eisen et al., 1968). The MOPC315 and MOPC460 proteins were also found to form very intense precipitin lines in Ouchterlony tests, and this provided a rapid screening test. Precipitin tests, however, had to be subjectively evalu- ated, as some proteins formed diffuse precipitin bands with antigens. Many of the latter type of reacting proteins were subsequently shown by fluorescence quenching, equilibrium dialysis, and other methods to have such low bindingaffinities, e.g., for E-DNP-L-lysine (K, = 10' M - ' or less), that they did not resemble induced DNP antibodies (Shubert et al., 1968; Eisen et d., 1970; Parker and Osterland, 1970). Only 4 DNP-binding myeloma proteins from the mouse which have binding constants higher than 5 x lo-' M - ' have been studied in any detail (Table IX).

2. Sequence Studies

The MOPC315 V regions have been completely sequenced-the VL by Schulenburg et al. (1971) and Dugan et al. (1973), and the VH and CHI by Francis et al. (1974). The availability of these sequences have been chiefly instrumental in our current understanding of the structure of the DNP binding site. The MOPC315 primary structures pro-

TABLE IX DINITROPHENYL (DNP) BINDING MYELOMA PROTEINS

K, ( M - I )

2,4- Di-

VH VI. 2-Methyl- nitro- iso- iso- 6-DNP-L- naphtho- naph-

Protein Strain IgC, type type lysine quinone tho1

MOPC315' C.B N, A VH-8 Vh-2 1 x lo7 5 x 10" 1 X 1Oj MOPC460" BALB/c A VH-9 VK-1 3 x 10' 1.6 x l W 5 x 10" XRPC25" BALB/c A VH-9? VK-23 2 X 10; - - HPC-3" NZB G2a ? ? 8 x I W ? ?

" Eisen et al. (1971). Jaffe et al. (1971). Sutton et al. (1977), Sharon and Givol (1976).

" Martin et al. (1972). No appreciable binding.

164 MICHAEL POTTER

duced some unexpected findings. Goetzl and Metzger (1970) began sequencing affinity-labeled peptides from MOPC315 and provided the first indications that the L chain was lambda. Complete sequence analysis by Schulenburg et al. (1971) and Dugan et u1. (1973) demonstrated that the MOPC315 A chain was a new class of mouse A (CA2). Only partial amino acid sequences have been determined for MOPC460 V, (Haimovich et al., 1972; Hood et al., 1976), MOPC460 VH (Hood et al., 1976; Lifter et al., 1974), and XRPC25 V, (Potter et al., 1976a; S. Rudikoff, personal communication). The VL are derived from different V, and Vk and also have different lengths in the L1 CDR (Potter et al., 1976a). An affinity-labeled peptide with the sequence S-K-I-R-Y-A-Z-B-T-Y-Y-B has been isolated from MOPC460 VH (Lifter et al., 1974); its location has not been clearly determined by homology, but it is sufficiently different from MOPC315 VH as to be clear that the two VH regions are different though related isotypes. No structural data are available for the IgG2, myeloma protein HPC-3 of NZB origin (Warner and Ovary, 1970; Martin et al., 1972).

An extensive literature on MOPC315 and MOPC460 forms the basis for the material discussed herein. Four major problems will be considered: the specificity of the binding site; conformational changes associated with hapten binding; the primary and tertiary structure of the binding site; the relationship of the DNP-binding myeloma proteins to antibodies with similar specificity.

3. Specijcity of the DNP Binding Site

a. MOPC315. Early studies of MOPC315 were involved with estab- lishing the location of the binding site in the Fab region. Several problems arose because MOPC315 was an IgA immunoglobulin and little was known about homogeneous antibodies of this heavy-chain class. Further, equilibrium dialysis studies indicated only 1.2 DNP binding sites per 7 S monomeric unit (Eisen et al., 1968); Underdown et a1. (1971) studied the subunit structure of MOPC315 in detail in search of an unusual structural basis for this finding. MOPC315 monomer proved, however, to have the conventional 2L + 2H chain structure, and the low yield of sites was attributed to denaturation of the protein in diluted solutions. For example, equilibrium dialysis run at concentrations of 0.03 mg/ml yielded 1.4 sites per 7 S unit whereas 1.9 sites per 7 S unit were obtained at concentrations of 0.52 mg/ml.

Electron microscopy revealed that most of the MOPC315 was in the form of dimers that were joined antiparallel end-to-end at the distal ends of the Fc regions (Dourmashkin et al., 1971). In the presence of bifunctional haptens, MOPC315 formed tetramers composed of two


dimers superimposed on each other along the axis of the Fc region and linked at the Fab via the bifunctional hapten (Green et al., 1971).

MOPC315 can be degraded proteolytically to produce the usual types of Fab fragments: pepsin Fab’, (Fab’)2, papain Fab. The Fc fragment from papain digestion is obtained in low yield. A careful study of the molecular size of the pepsin degradation products revealed that MOPC315 Fab’ gave two peaks in Sephadex G-75. One of these components was the conventional Fab‘, and the other was a smaller fragment, Fv, that was found to contain the combining site (Inbar et al., 1971). The molecular weight of the Fv was estimated to be 30,000 and to consist of only the variable regions. Fv fragments have only occasionally been produced in high yield from other IgA myeloma proteins.

Inbar et al. (1971) prepared the pepsin Fab’ fragment of MOPC315 and found that it readily crystallized into long needles in low ionic strength 0.002 M sodium acetate, pH 4.7, at 22°C. Unfortunately these crystals were unsuitable for X-ray diffraction studies.

The MOPC315 myeloma protein polymer, 7 S monomer, or fragments (pepsin, papain Fab’, Fv) have been shown to have strong binding activity for 2,4-~-DNP-~-lysine ( K , = 1 x lo7 M - ’ ) and various derivatives by equilibrium dialysis (Eisen et al., 1968; Pecht et al., 1972), fluorescence quenching and inhibition of fluorescence quenching (Eisen et al., 1968, 1970; Eisen, 1971; Michaelides and Eisen, 1974); and temperature-jump chemical relaxation spectrometry (Pecht et al., 1972; Haselkorn et al., 1974). The findings indicated that the DNP ring group was oriented in the binding site by a stacking interaction with an available tryptophan (see below) and that interactions between the protein and the nitro groups located at the 2 and 4 PO- sitions were essential for strong binding.

Eisen et al. (1970) found by the inhibition of fluorescence quenching of 2,4-DNP-~-lysine that 2,4,6-TNP-~-Lysine was the most potent inhibitor and 2,6-DNP-~-lysine and 4-NO2-pheny1-~-lysine were weak or noninhibitory, respectively. MOPC315 was found by equilibrium dialysis (Michaelides and Eisen, 1974) to have a K, 1.1 x 1Oj M-‘, for 2,4-dinitronaphthol, a compound with a larger ring structure.

The most extensive comparative study of derivatives of E-

DNP-L-lysine has been made by Haselkorn et al., who studied 38 different nitrophenyl ring structures. They induced rapid temperature jumps in mixtures of MOPC315 and haptens and studied spectrometrically their association and dissociation. This method can establish association and dissociation rate constants, and dynamic equilibrium constants for a variety of ligands in a range of concentra-

166 MICHAEL POTTER

tion between 1 x low7 and 2 x M. Because of its dynamic nature, this method can also detect intermediate encounter complexes that form between the ligand and amino acid side chains from the antibody as the ligand enters the site. Thus conformational changes on antigen binding can be detected.

Haselkorn et al. (1974) showed that maximal binding depended upon the location of two nitro groups in the 2 and 4 positions. Binding of the mononitro and 2,6-dinitro compounds was weak or negligible. The study of C-1 derivatives also revealed three other subsites on the hapten that contribute to binding: (1) first hydrophobic region, (2) second hydrophobic region (both located 6-8 t% from the DNP ring site, and (3) a positively charged subsite that reacts with the COO- group of an amino acid attached to the hydrophobic connecting structure. The high speed spectrometric studies of MOPC315~-~-lysine interaction revealed only a single relaxation time, which was evidence against a local conformational change associated with binding.

The picture that emerged from these studies was that the MOPC315 had a highly complementary binding site for 2,4-dinitrophenyl ring structures. However, the MOPC315 was found to bind other ligands that were chemically unrelated to 2,4-DNP structures: (1) Shubert et al. (1968) found that MOPC315 bound 5-acetyluracil bovine serum albumin (BSA) and purine BSA; (2) Michaelides and Eisen (1974) and Eisen et al. (1970) found that MOPC315 had a “strange” and unexpected cross-reaction with 2-methyl- 174-naphthoquinone (menadione, or vitamin &). By equilibrium dialysis, MOPC315 had a K , of 5.1 x 1Oj M-’ for menadione. In this same study a great variety of other compounds were screened for binding activity, and the only other compounds with suggestive activity were 5-acetyluracil caproate, which had a K , of 2.44 x 104 M-’, tetrahydofolic acid ( K , 5.8 x lo4 A!-’), pyridoxol (K1 = 2.1 x lo4 M-’) riboflavin ( K , 3.7 x lo4 A!-’), and flavin mononucleotide (K, 6.8 x lo4 M-’).

The “strange” cross-reaction with menadione was considered to be due to similarities in the ring structures between DNP and the naphthoquinone, both of which could potentially stack on an available tryptophan in the binding site. Further, both DNP and naphthoquinone rings have similar charge distributions and each has a positively charged complex capable serving as an acceptor in a charge transfer complex (Michaelides and Eisen, 1974). Rabbit anti-2,kDNP and 2,4,6-TNP and guinea pig anti-2,g-DNP antibodies had “substantial affinity” for menadione (Michaelides and Eisen, 1974), and antibodies to menadione may also have DNP-binding activity (Johnston and Eisen, 1974). Strange cross-reactions have also been noted in con-


ventional anti-DNP and 5 A U antibodies (Underdown and Eisen, 1971).

b. MOPC460. The binding site of the IgA(k)MOPC460 protein differs considerably from that in MOPC315. Hapten-binding studies (JaEe et al., 1969) using equilibrium dialysis gave an association constant (K,) for e-DNP-L-Lysine of 3 x 1Oj M-' and Ki values as determined by inhibition of binding €-DNP-L-Lysine in equilibrium dialysis of 1 x 10'MM-' for 2,4-dinitronaphthol, 1.2 x lo5 M - ' for 2,4,6- TNP amino caproate, and 1.6 x lo4 M - ' for menadione. Other weak inhibitors were 2,6-aminocaproate, 4-MNP-aminocaproate, and purin-6-oyl aminocaproate (JaEe et al., 1971). Thus the M460 protein in contrast to the M315 protein had a higher affinity for 2,4- dinitronaphthol and a lower affiity for 6-DNP-L-lysine, suggesting that a larger area of the molecule was involved in binding and that the large ring structure was better accommodated (see also Willan et al., 1977).

Rosenstein and Richards (1972) found that MOPC460 competitively bound 2,4-~-DNP-~-lysine and 2-methyl-174-napthoquinone thio- glycolate (MenTG). The MenTG, but not DNP, binding was decreased in the papain Fab fragment. The MOPC460 protein contains a reactive Cys-SH in the heavy chain adjacent to the sequence -V- L-S-G-(E-A)-C-R-P- (Jackson and Richards, 1973). When this SH is blocked, the binding of MenTg, but not DNP, decreases (Rosenstein and Richards, 1976). The spatial separation of DNP and Men (menadione or vitamin k) sites has been demonstrated by nanosecond fluorometry (Manjulaet al., 197613) and by the binding ofthe protein to DNP and Men groups attached at various distances to an insoluble substratum, i.e., Sepharose (Rosenstein and Richards, 1976). In the first method, the free SH group was allowed to react with a fluorescent probe that can donate energy to the bound hapten. The distance between the probe and the hapten can be calculated by nanosecond fluorometry. In the second method, haptens with different spacers were attached to Sepharose and the binding of MOPC460 was determined. The shortest distance between the Men group and the Sepharose that still permitted binding was 22 A, and the shortest distance between the DNP and Sepharose was 8.8 A. It was concluded from this finding that the Men and DNP sites were separated by 13.2 A. The MOPC315 site was also studied; the depth ofthe Men site was again 22 A, and the DNP site was 11 A. It is very difficult to see how both menadione and DNP can compete for the same site and still be so far apart. These findings await structural confirmation.

Although the structure of the M460 V regions has not yet been com-

168 MICHAEL POTTER

pleted, available amino acid sequences of the CDR L1, L2, and H1 (Haimovich et al., 1972; Lifter et al., 1974) have been determined and indicate that both primary and secondary structural differences will be found in the combining site surface. H1, for example, is one residue shorter than H1 in M315, and L1 is one residue longer. It will be of great interest to the study of protein structure to determine the secondary structure of L1 in MOPC460.

c. X W 2 5 . Only a few studies are available on XRPC25 (Sharon and Givol, 1976). This protein has a high binding affinity for E-

DNP-L-lysine, but does not bind menadione or the bulky 2,4-DNP- naphthol (Willan et al., 1977). An Fv fragment has been prepared from XRPC25 (Sharon and Givol, 1976).

4 . Conformational Changes Associated with Hapten Binding

Several methods have been employed to study conformational changes in MOPC315 during hapten binding. Changes in fluorescence as detected by circular dichroism spectra in the far UV region did not provide evidence for a conformational change (Inbar et al., 1973). In this study some quantitative changes were noted in the near UV region, but these could not be unequivocally interpreted as reflect- ing a conformational change. Pecht et al. (1972), using the temperature-jump procedure, found only a single relaxation effect, which was also evidence against a conformational change associated with hapten binding. Dwek et al. (1975b) have allowed the M315 protein to interact with dinitrophenyl nitroxide spin label compounds of different lengths and ring structure and concluded that the binding site in solution has considerable structural rigidity. It has been shown from NMR studies that conformational changes on hapten binding are limited to the binding-site region (Dwek, 1977) and hence are localized to one or very few amino acid side chains.

Using high speed T-jump-chemical relaxation kinetic studies of the binding of the M460 protein with E-DNP-lysine, Lancet and Pecht (1976) found by spectroscopic analysis during the relaxation time two interconvertible conformational states in the M460 protein. This is the first firm evidence with a homogeneous myeloma protein that a local conformational change occurs during hapten binding. Whether this local change will have effects on other domains remains to be demonstrated. This evidence again clearly points out another remarkable difference between the M460 and M315 DNP binding proteins.

5. Structure of the DNP Binding Site

Several different techniqiies have been used to study the structure of the DNP binding site: affinity-site labeling; model building from


primary structures; and spectroscopic methods using hapten spin- label compounds. Affinity-site labeling has identified residues close to the hapten binding site, but, thus far, with the myeloma proteins has not identified contact residues. Model building has provided an approximation of the site of MOPC315, and this has been refined by nuclear magnetic resonance (NMR) studies using hapten spin-labeled probes.

a. A$nity-Site Labeling Agents. Two classes of affinity-site labeling agents have been used to label amino acids near the binding sites of MOPC315. Metzger and Potter (1968) used two mononitrodiazoniurn compounds (p- and rn-nitrobenzenediazonium fluoroborate) which have a predilection to label tyrosines. These compounds are bound by MOPC315 despite the presence of only a single NO, group, and, when bound, Tyrsr of the light chain (Goetzl and Metzger, 1970) was diazotized to azotyrosine. Haimovich et al. (1970, 1972) and Givol et al. (1971) have used a series of DNP-bromoacetyl compounds that have the ability to label a variety of amino acids. Two compounds BADE (N-bromoacetyl-N'-DNP-ethylenediamine) and BADL (E-N- bromoacetyl-e-N-DNP-lysine) had different labeling properties; 96% of the BADE label appeared on Tyrs4 of the light chain, and 95% of the BADL label appeared on Lys52 of the heavy chain (Table X). In the hypothetical three-dimensional model, both of these residues appear at the entrance of the cavity and close to the site where the DNP is bound.

Affinity-site labeling studies of M460 protein have been made with the bromoacetyl reagents (Haimovich et al., 1972). The BADE compound labeled Lyss4 in VL, but no labeling was obtained in the VH with the BADL compound (Table 8). Using photoaffinity labeling reagents 2,4-dinitrophenylalanyl diazoketone and 2,4 din itro- l-azobenzene, Lifter et a2. (1974), Yoshioka et al. (1973), Hew et al. (1973), and Richards et al. (1974) labeled LysS4,, and TYr33~ and Tyr?88H. The exact location of the second Tyr is not yet clear because of incomplete sequence data.

The lateral dimensions of the MOPC460 site have been estimated by electron spin label mapping to be 10 A x 11 A as compared to 8.5 A x 11 A for MOPC315 and 7.5 A x 8 A for XRPC25 (Willan et al., 1977). While the chemical structure ofthe MOPC460 site has not been determined and awaits the availability of amino acid sequence analysis, there is disagreement on the nature of the menadione and DNP binding sites. One group proposes two separate sites (Richards et al., 1975; Manjula et al., 1976b; Rosenstein and Richards, 1976); the other (Willan et al., 1977) propose a single site.

b. Model Building. None of the nitrophenyl binding myeloma pro-

TABLE X AFFINITY LABELING OF MOFC315 BY BROMOACETYL COMPOUNDS"

% Label found on amino acid in

Structure Name" MO FC3 15L MOPC315H w

DNP-NH-CH~-CHZ-NH--CO--CH2-Br BADE 96% Tyr3, 4% Lyszz g YOOH

DNP-NH-CH,-kH2PH-NH-CO-CH,-Br DNP-NH-CHz-CHz-CH~-CH-NH-CO-CHz-Br

LOOH YOOH

r

4

s

BADB 87% TyrS1 13% LysSz B A D 0 66% Tyr3, 34% Lyssz g

8 DNP-NH-CHz--CH~PH2-CH,-~H-NH-C0-CHZ-Br BADL 5% Tyr= 95% LysM

" From Givol et al. (1971). I, BADE = N-bmmoacetyl-N'-DNP = ethylenediamine; BADL = E-N-bromoacetyEeN-DNP-lysine.

ANTIGEN-BINDING MYELOh4.A PROTEINS OF MICE 171

teins has yet been crystallized in suitable form to make possible a three-dimensional structure determination by the techniques of X-ray crystallography. Nonetheless, the structure of the MOPC315 binding site is well characterized, and we have in good probability a very accurate concept of the three-dimensional structure. This is possible because crystallographic studies of V domains from different species have exhibited a remarkable degree of conformational homology, in particular in the framework regions. This makes it possible then to construct a framework and add on the 6 Wu-Kabat CDR. Model building encounters its chief problems with arranging the secondary structures of the CDR. A number of CDR have been determined, and in many cases it is possible to borrow a structure of similar length.

Four prototype structures determined by X-ray crystallography are available: the human NEW Fab (Poljak et al., 1974), Mcg L-chain dimer (Schifferet al., 1973), and RE1 (K) chain dimer (Eppet al., 1974) and the McPC603 mouse ( K ) Fab (Segal et al., 1974).

In other situations where (as is often the case) the CDR with a new length is encountered, remodeling is necessary. This requires hypothetical assumptions. Other methods, however, can reinforce model building. Four such approaches are: (1) mapping of contact residues in solution with ESR-hapten probes; (2) determination of the 270 mHz NMR spectrum of protons near the probes (with identification of amino acid side chain); (3) specific chemical blocking of amino acid side chains; (4) affinity site labeling.

Poljak et al. (1974) suggested that the MOPC315 site could be constructed from the known sequences and existing models of the human NEW Fab (a lambda L chain-containing protein). Recently, Padlan et al. (1976b) have built the MOPC315 Fv. In the latter study two models were built independently, one at the Weizmann Institute and the other in Bethesda. The two models were brought together, and a final model based on both studies was constructed (Padlan et al., 1976b).

The MOPC315 model was constructed by Padlan et u Z . (1976b) using the McPC603 framework backbone and CDR structures from other crystals or remodeled CDR. A stereo drawing of the MOPC315 Fv is given in Fig. lb, and the legend to this figure gives specific details on the structural alterations made to construct the site. A comparison display of the CDR of MOPC315 and McPC603 is given in Fig. 2.

The most satisfying features of the MOPC315 model once constructed was a pocket formed in the middle of the hypervariable surface at the exposed interface between VL and VH. The pocket was lined on one side by H1, H2, and on the other by L1. The roof of the cavity

172 MICHAEL POTTER

was formed by H3, and the floor by L3. The pocket itself differed from the comparable one in McPC603 in that it contained many aromatic amino acid side chains. It was then relatively straightforward to place the E-DNP-L-lysine in the cavity. The ring structure was oriented in the cavity in a plane over TrpgSL with two NO2 groups facing inward. It was proposed that the NO, groups formed hydrogen bonds with ASN36L and AS^^^^. [As will be discussed below, Dower e t al. (1977) and Dwek et al. (1977) suggest that the NO2 groups interact with ASn36~ and TrysIL.] Hydrophobic interactions with other ring structures, e.g., PheS4", PhegHL, and aliphatic are feasible. The entrance to the pocket contained T y 1 - 3 ~ ~ and Lys5zH, the two amino acids that were identified by affinity-labeling studies (see below).

Using a series of DNP hapten spin-labeled nitroxide compounds (Fig. 3) and NMR, Dwek et al. (1975a,b, 1977), Sutton e t al. (1977), Wain-Hobson et al . (1977), and Dower et al. (1977) studied the DNP- binding site of MOPC315. The spin-label compounds (Fig. 3) differ in length between the hapten and spin label. Using ESR spectroscopy, Dwek et al. (1975b) and Sutton et al. (1977) calculated the dimensions of the binding site to be 10-11 A in depth with lateral dimensions of9 19 X 6 19. The data revealed that the site had considerable structural

NO2

+O.% nm-

FIG. 3. Dinitrophenyl (DNP)-hapten spin-label compounds used in electron spin resonance and nuclear magnetic resonance spectroscopy of DNP binding sites by Dwek et al. (1975a,b, 1977).


rigidity. By means of five-membered ring spin labels instead of the six-membered ring structures, the sides of the side could be probed in a more sensitive fashion by ESR spectra, and it was found that the site was asymmetric. In an analysis of the NMR spectral changes that occur in the MOPC315 Fv on binding with DNP, Dower et al. (1977) and Dwek et al. (1977) were able to show that the DNP ring structure lies in an “aromatic box” enclosed by Trp,,, Phe34H, and Tyr34L. With NMR spectroscopy, Wain-Hobson et al. (1977) located one of the 3 histidines in the MOPC315 at the edge of the binding site. The orientation of the DNP ring structure in the site, according to Dwek et al. (1977) and Dower et al. (1977), favors the formation of hydrogen bonds between MOPC315 and the nitro groups to involve and Tyr34L. As pointed out, this is in contrast to the proposal of Padlan et al. (1976b). It should be noted here that Tyr34L is an affinity site-labeled amino acid. It is not always possible to draw firm structural conclusions from affiity-labeling studies. The interpretation of Dwek et al. (1977) (Fig. 4) is favored.

The details of the interactions of E-DNP-L-glycine within the MOPC315 site according to Dwek et al. (1977) are shown in Fig. 4, reproduced from Dwek et al. (1977). Another method of potential help in determining the fine structure of the binding site are the circular dichroism spectra of hapten-antibody complexes (Rockey et al., 1971; Rockey and Freed, 1976; Freed et al., 1976). Characteristic near-UV circular dichroism peaks have been observed in MOPCSlSTNP aminocaproate interactions, suggest the participation of tryptophan side chains (Rockey and Freed, 1976; Orin et al., 1976).

According to Willan et al. (1977), the XRPC25 site is the narrowest and apparently is too small to admit bulky ring structures, such as menadione and 2,4 dinitronaphthol, whereas the MOPC460 is at the other extreme and is the largest of the three sites. It is notable too that the depth of the DNP binding site in the cavity is roughly the same for all three molecules. The three proteins are able to precipitate with DNP-BGG and DNP-BSA; this suggests that the protrusions on some globular proteins are long enough to interact and enter restricted cavitary binding sites on immunoglobulins.

Another feature of the three myeloma proteins is the apparent difference in the contours of the antigen binding site region. As reflected by the partial sequence data, there are remarkable differences in the L1 regions of these molecules. This suggests that a functional feature of the antibodies may be associated with a large surface of the protein. Specific contours may favor binding haptens on different mac- romolecular carriers. Here the complementarity of antigen antibody surfaces may play a role.

25 61

FIG. 4. MOPC315 hapten-binding site with amino acid side chains and dinitrophenyl (DNP)-glycine hapten (atoms in filled circles), kindly provided by Dr. R. A. Dwek, Oxford University (see Dwek et al., 1977). The view is from the solvent looking directly into the site region; all six Wu-Kabat complementaritydetermining regions (CDR) are shown. The following side chains are drawn in: L1: Tyr,, Amss; L3: TrpV3, Phe,,; H1: Phew; H2: Pheba; H3: LeuloS and TyrIM. The proposed actions of amino acid side chains are tabulated:

CDR Amino location acid Proposed action

L1 TYr, H-bound to NO,, forms part of aromatic box

L3 TTY~ DNP stacks on planar ring, part of aromatic box, AmJB H-bound to NO, group

H-bound to NO2 Hisy7 At binding site entrance

H1 Phe3, Part of “aromatic box,” van der Waals contact with

Not shown, but Padlan et al. (1976b) suggest H-bound to hapten

NO, group Asnss

H 2 PheJo - H 3 Leula3 Points out of pocket

Points in, interacts with TyrwL


There is insufficient evidence available to warrant conclusions about the absolute requirements for DNP binding. The three molecules so far examined appear to share cavities or depressions into which the extended DNP group can enter. The basic structural features of the cavities are (1) a Trp onto which the planar DNP ring can stack, (2) side chain groups that can form hydrogen bonds with NO2 groups, and (3) an aromatic pocket formed by protruding groups, such as Phe.

6. Relationship of DNPBMP to Anti-DNP Antibodies

The humoral immune response to poly-DNP-substituted antigens in mice, as in many other mammalian species, is heterogeneous (e.g., Pink and Askonas, 1974). By the use of special immunogens or immunization procedures, relatively restricted DNP-antibodies can be obtained (Montgomery and Pincus, 1973; Pink and Askonas, 1974). However, the comparison of clones derived from different individuals in the same inbred strain induced by immunization with DNP-OA and boosted with DNP-OA revealed a remarkable heterogeneity as detected by isoelectric focusing comparisons of clonal products (Pink and Askonas, 1974). These findings suggest that it may be difficult to associate MOPC315, MOPC460, and XRPC25 with antibodies generated in responses to DNP-protein immunizations, although in one isolated study (Granato et al., 1974) idiotypic relationships between MOPC315 and anti-DNP antibodies were found. However, systematic testing of DNP-producing clones, produced by the method of Pink and Askonas (1974) for idiotypic cross specificities, have not been reported.

The binding affinities of mouse DNP-antibodies for e-DNP-L-lysine generated by intensive immunization of CBA mice are often two or- ders of magnitude higher than MOPC315 (K, 1 x 10' M - ' ) (North and Askonas, 1974). A homogeneous antibody with a K , of 1 X log to 1 x 10"' M - ' for e-DNP-L-lysine probably resembles MOPC315, MOPC460, and XRPC25 by having a pocket or cavity. However, the amino acid side chains projecting into such a pocket must have a different arrangement and composition to produce the stronger binding affinity. It is an interesting exercise to modify the MOPC315 site to increase the binding affnity. The addition of one or two more nonpolar amino acid side chains might be sufficient to enhance hydrophobic interactions to allow for binding affinities of 10' or greater. One candi-

FIG. 4 legend (continued): In the proposed model by Padlan et al. (1976b), the NO, groups are H-bonded to Asn,

of the L chain and AmSB of the H chain, a different orientation from that shown here. Padlan et ul. (1976b) prefer the later orientation because Tyr, is an affinity-labeled residue that is presumably noncontacting.

176 MICHAEL POTTER

date substitution might involve Asn36L, which could be changed to a Phe, a Leu, or even a Trp. If such a relationship is possible, then it is conceivable that a somatic mutation involving the MOPC315L chain in a MOPC315 clone could increase or decrease binding affinity. It would be of great interest to determine whether the clones produced by Pink and Askonas (1974) carry the unusual VA-2 chain found in MOPC315.

7. Comments

The DNPBMP have been studied most extensively by im- munochemists. The interest in the DNP hapten itself was no doubt the major impetus for the work. It is surprising that so little has been learned about the biology of DNPBMP, in particular the relationship to natural antibodies and natural immunogenic stimuli. Progress is complicated by the general finding that normal DNP-binding Ig is very heterogeneous, and by the question of nonspecific interactions. V-region isotyping of more DNPBMP with high K, in mice might clarify whether specific VL-VH pairs are essential in forming DNP sites and how many different combinations are involved.

B. PHOSPHOCHOLINE (PC)

Melvin Cohn (1967) was the first to discover an antigen-binding myeloma protein in the mouse when he demonstrated that S63 (IgA,k) myeloma protein precipitated pneumococcal C polysaccharide (PnC). Since that time 13 myeloma proteins that precipitate PnC have been found (Table XI). Cohn et al. (1969) attempted to identify the haptenic determinant in this polysaccharide based on existing compositional data, but failed to find a potent inhibitor, although they did note that choline at very high concentrations was a weak inhibitor. Leon and Young (1971) reexamined this problem and deduced from the findings of Tomasz (1967) and Brundish and Baddiley (1967) that choline in PnC was joined by a phosphodiester bond, and hence that phosphorylcholine was a potential hapten. They then made the important observation that the zwitterion phosphorylcholine was a potent inhibitor of all the proteins that bound PnC. In addition they studied the inhibitory ability of choline and choline derivatives and found differences in the relative inhibitory activity of choline, L-a-glycerophosphoryl- choline, and phosphorylcholine (see Table XI). The precipitation of PnC by MOPC167 and subsequently MOPCS11 (Potter and Lieberman, 1970) was inhibited by relatively low concentrations of choline and acetylcholine. MOPC167 also agglutinated sheep red blood cells (SRBC) coated with human P-lipoprotein (Leon and

TABLE XI CHARACTERISTICS OF PHOSPHORYLCHOLINE-BINDING MYELOMA PROTEINS

V" domain (No. of amino

differ- iso-

* acid VK U

Q 5

ences type ? 5

Year Strain K , x 1O;M-I T15:S63 dis- Ig of phosphoryl- idio- from se- IEF'

Protein covered class origin choline" type" G PA : PC : C' J' TlS!) quence" pattern U z

1 Q

z r

U

S63 1967' CXK BALBIc + 10:1:5% - - S 107 1968" CXK BALBIc + 8.5 : 1 : 620 0 22 1 MOPC299 1968 CXK BALB/c + 7.5: 1: 430 - - - McPC603 1968' CXK BALBIc 1.6 2 0.4 - 10: 1:870 7 8 2 MOPC167 1968 CXK BALBIc 1.2 2 0.2 - 0.8 : 1 : 10 13 24 3 MOPC511 1969'' ~ K ( A ) BALBIc 0.16 + ' I 5 24 3 TEPC15 1969t CXK BALBlc 23 5 05 + 11~1550 0 22 1 HOPC8 1969d CXK BALBIc 24 2 05 + 1 22 1

CAL-9 - ALPC43" 1974 Y 3 K

CBPC2" 1974' CXK CB-20 - 8.2: 1:520 2(36) 22 W3207 197 1" BALBlc -

Y5236 197 1" a BALBIc + Y5170 1971" a BALBlc +

s E 'd

H - - -

1 E? 5 a 2 - 0

- - - 0 %

- - -

I' Chesebro and Metzger (1972), Metzger et al. (1971). '' Potter and Lieberman (1970), Sakato and Eisen (1975).

'! See Table X. (I See Tables 1-111. 'Claflin (1976a,b), Rudikoff and Claflin (1976).

I' Tumor no longer produces myeloma protein.

Leon and Young (1971).

CPA = L-0-glycerophosphoryylcholine; PC = phosphocholine; C = choline.

k m

178 MICHAEL POTTER

Young, 1971). The exact structure of the PnC has not yet been established (Gotschlich and Liu, 1967; Watson and Baddiley, 1974), and larger haptenic structures including sugars or other residues attached to phosphorylcholine have not been isolated from PnC.

Other antigens from a variety of sources have been identified that contain phosphorylcholine: (1) an internal membrane antigen from Lactobacillus acidophilus species No. 4 isolated from the mouse gastrointestinal flora (Potter, 1970,1971) (this antigen was isolated by Mr. Charles K. Mills, American Type Culture Collection); (2) an antigen liberated into culture media by Trichoderma and Fusarium (of mouse gastrointestinal origin) (Potter, 1970, 1971); and (3) nematode- associated antigens. Crandall and Crandall (1969, 1971) found that mice infected with Ascaris suum larvae developed a macroglobulin in the serum that reacted with PnC. In subsequent studies Brown and Crandall ( 1976) demonstrated that these macroglobulins were antigenically (idiotypically) related to the idiotype of the TEPC15 myeloma proteins. Pery et al. (1974) showed that homogenates of eggs, infective larvae, fourth-stage larvae, and secretory and excretory products of the adult nematodes Nippostrongylus brasiliensis (a rat parasite) and Hemonchus contortus (a sheep parasite) had a phosphorylcholine-containing antigen. Bennett and Bishop (1977) elucidated the structure of the capsular polysaccharide of type XXVII Streptococcus pneumoniae (pneumococcus) (Fig. 5). The repeating unit of this polysaccharide contains 1 mol of phosphorylcholine attached to L-rhamnose. This antigen is precipitated by PCBMP (Ben- nett and Bishop, 1977). APCBMP protein (McPC603) precipitates with a lipopolysaccharide antigen from a Proteus morganii species of mouse origin (Potter, 1971). Immunization with these Proteus morganii organisms induces antiphosphorylcholine antibodies carrying TEPC 15 idiotypes and also antibodies that have a structural relationship to McPC603 (L. Claflin, personal communication).

The phosphorylcholine hapten is widely distributed in nature.

1 . Sequence Studies

Nine VH sequences are available (Rudikoff and Potter, 1974, 1976; Hood et al., 1976). These are summarized in Table XII. As may be seen, the framework sequence of VH-4A (e.g., TEPC15) is common to all; single framework substitutions are found in VH-4C (W3207), two in VH-4D (MOPC511), and five are located in VH-4E (MOPC167). Amino acid substitutions are found in two of the CDR, a few in H2, but the most dramatic differences occur in H3, where both length and primary structural differences are observed (Rudikoff and Potter,


HO Ck,

\&

FIG. 5. Repeating unit of type XXVII Streptococcus pneurnoniae capsular polysaccharide determined by L. G . Bennett and C. T. Bishop, National Research Coun- cil, Ottawa.

1976). A possible explanation for the origin of these differences is that the VH-4 proteins are a family that have expanded by gene duplica- tion. The oldest divergence is between the MOPC167 and TEPC15. The size or length differences in H3 possibly reflect selective pressures favoring the H3 secondary structure and the contour of the antigen-binding surface.

The V, sequences of the PCBMP have been determined only through the first invariant Trp,, (Barstad et al., 1974b; Potter et al., 1976a). There are three V, domain isotypes that differ markedly from each other (Tables I, 11, V): VK-22 (T15, H8, S107), VK-8 (McPC603), and VK-24 (M167, M511). All three basic V, sequences from PCBMP have very long L1 regions, which may be an important functional structural feature of PCBMP VL. Claflin et al. (1975) and Rudikoff and Clafiin (1976) have analyzed the L chains of the BALB/c PCBMP by isoelectric focusing and have found three basic banding patterns for each of the prototype sequences: (1) the M511-M167 pattern; (2) M603-W3207 pattern; and (3) the T15, H8, and S107 pattern.

The V regions of the mouse PCBMP reflect a somewhat complex evolution. First, VH domains are clearly all related through extensive similarities in the framework sequences. Sequence differences within the VH domains involving the framework and H3 suggest that the group of proteins is controlled by more than one structural gene, and, using the arbitrary definitions for V-region isotypes (Table VII), five different isotypes were proposed. It will be of interest to determine whether all the VH-domain isotypes can form PC-binding proteins

TABLE XI1 VH-4 PHOSPHORYLCHOLINE-BINDING MYELOMA PROTEINS"."

80 H3 120 I I I I I D . Y Y G S S Y . WI-

NH, HI 40 H2 I !

S 107 1 1 I 1 ID . Y Y G S S Y . Wi- TEPCl5

S63 ; I I I ID . Y Y G S S Y . Wi- l i I I ID . Y Y G S S Y . WI-

HOPC8 I ; I

Mc603 i** l I G*-K* 1 [ N . Y Y G S T . . W ; * -

I F A

MOPC 167--V 1 , T~S-H-RIY T-TJD A D Y G D s Y F GI-

I I

I D . Y Y G S N Y . W / - I Y5236

I I I N . Y K Y D L V . W I - i ! I

i ! SAD I G ID G D Y G S S Y . W / - I i

I- i I I I I

I

I

W3207 MOPC511

(' Sequences by Rudikoff and Potter (1974, 1976) and Hood et al. (1976). '' * = Contact residues; 0 = no amino acid corresponding to longer sequence.


when combined with each of the V,-domain isotypes, and further, whether all the VL-VH combinations are expressed in the B-cell pool.

2. The McPC603 Phosphorylcholine Binding Site

The McPC603 pepsin Fab’ crystallizes in 42% saturated (NH4)2S04-0.1 M imidazole solutions at pH 7.0 (Rudikoff et al., 1972; Segal et al., 1974). The crystals of M603 Fab’ as well as isomorphous replacements were suitable for X-ray diffraction studies, and a high- resolution (3.1 A) electron density map was obtained that permitted tracing the polypeptide chain. In addition, a mercurated phosphorylcholine compound was synthesized by T. F. Spande, NIAMD (see Padlan et al., 1973), 2(5’acetoxymercury-2’-thienyl)ethylphosphoryl- choline (AMTEPC), which was bound by the crystals and permitted localization of the hapten binding site in the crystal. No change in the electron density map b y hapten binding was found. This suggested that no large conformational change occurred during binding.

The Fv contains the VL and VH domains, which are oriented to each other at roughly a 90” angle. The hapten-binding site is located in a wedge-shaped cavity formed by both VL and VH. This cavity (see stereo Fig. 1) is part of a continuous contoured surface 30 A x 40 A that faces the solvent at the end of the Fv most distant from its junction with CHI. In the McPC603 Fv, the hapten (10 A x 2 A) occupies only a small portion of this surface and the hapten- binding site is formed by parts of L1, L3, H1, H2, and H3 and lies in the interface between the VL and VH domains. L2 is shielded from the hapten-binding site by L1 and H3 (see stereo drawing). It is apparent then that much of the remaining structure is available for interaction with the molecular carrier of the phosphorylcholine hapten.

Details of the hapten-binding interactions with amino acid side chains of McPC603 (Padlan et al., 1976) are shown in Fig. 6. Some contributions may come from VL, but this sequence has not yet been completed in the L3 and the role of these amino acids cannot be assessed as yet. Since the phosphorylcloline hapten is joined to the antigen carrier by a phosphoester (see Fig. 5 ) group, the positively charged trimethylammonium group is most interior: the negatively, charged GlU35~ and GlusgH react electrostatically with it. The negatively charged phosphate group is nearer the exterior of the binding pocket, and the positively charged guanidinium groups of Arg52~ and (possibly) LysXH interact with it. In addition, hydrogen bonds between the phosphate and Tyr33H, and Arg,, contribute to binding. Other hapten interactions are formed by van der Waals contacts with Tyr3aH, T Q ~ ~ H , and ?=. The important hapten contacting residues (Tyr=,

182 MICHAEL POTTER

FIG. 6. Amino acid side chain in interactions in the McF'C603 site from Padlan et al. (1976a). The tabulated residues interact with the hapten.

Amino Location acid Interaction via: With PC group

H1 Tyr3, H bond, van der Waals contacts 0 in PO, H 1 Glu,, coo- Trimethylammonium+ H 2 Guanidinium PO,-

H 2 Gluu coo- Trimethylammonium+ H3 Trp,,, van der Waals contacts Hapten L3 96 van der Waals contacts Hapten

H bond


Glu35, Arg5,, LysM, Glu59) are preserved or conserved in all the VH of PCBMP (Padlan et al., 1976). MOPC167 does have a difference affecting contacting residues, a Phe-Gly doublet replaces TrpIo4 in H3. This alteration affects the binding surface, but may have only a slight effect on the binding of phosphorylcholine or choline. For a detailed discussion of the basis of choline binding in MOPC167, see Padlan et al. (1976).

Grossberget al. (1974) blocked arginyl groups in HOPC8 b y glyoxy- lation and inactivated the phosphocholine-binding activity. Also partial esterification of carboxylate groups drastically decreased binding. Both of these procedures independently support the X-ray studies by implicating arginyl and Glu or Asp groups in binding.

Kabat et al. (1976a) conducted a computer search of all available V regions for pairs or amino acids that occurred nonrandomly and found that three significant pairs in H1, Phe32-Tyr33, Phe32-Glu35, and

occurred in all PCBMPs. Further, the sequence Phe3,- Tyr3,Met3,-Glu3, was found only in VH of PCBMP. Two of the residues, Tyr33 and G ~ U ~ ~ , are contacting, the other two are essential for the conformation of the secondary structure. A human IgM PCBMP Fr has been described by Riesen et al. (1975). There are two structural similarities between the mouse PCBMP and Fr (Riesen et al., 1976; Riesen and Jaton, 1976). First, the H1 region sequence between positions 32 and 35 is the PheTyrMetAsp. This is very similar in structure to the PheTyrMetGlu sequence in H1 identified by Kabat et al. (1976a) as being unique to mouse PCBMP. Second, the L1 region of Fr is long, as in all three mouse V, isotypes (Table V). Both of these features suggest these will be common features in the human and mouse PC-binding sites.

Using the affinity-site labeling reagent p-diazoniumphenylphospho- choline (DPPC), Chesebro and Metzger (1972) and Metzger et al. (1971) attempted to label 5 PCBMPs. No labeling was obtained with MOPC511. The H chains of MOPC167 and McPC603 were labeled, but no single labeled amino acid could be recovered in high enough yield to permit identification. Labeling of HOPC8 and TEPC15 with DPPC gave two tyrosines in sufficient yield to permit identification of the sequence source. One labeled was Tyr3* in L1 and the other was identified only by its associated sequence AzoTyrSerTyrProLeuGlx. Presumably this sequence is near the binding site and possibly partly in L3.

Several PCBMPs (TEPC 15, HOPC8, McPC603) give significant enhancement of fluorescence upon binding the PC hapten (Pollet and Edelhoch, 1973; Pollet et al., 1974). Using fluorescence titration (see

184 MICHAEL POTTER

section on P6GalBMP) it has been possible to determine the K, of the PCBMP for PC (Pollet and Edelhoch, 1973; Glaudemans et al., 1977). Glaudemans et al. (1977) have compared the K , of McPC603 and HOPC8 for the low-molecular-weight PC hapten and two large polysaccharide PC-antigens, the pneumococcal C polysaccharide and the S27 capsular polysaccharide (Fig. 5) and found no increase in K, with the whole polysaccharide. This result suggests that most of the interaction between C and S27 polysaccharides and PCBMP occurs at the hapten-binding site. The hapten-binding site represents only a small portion of the potential binding surface of the Ig.

3. Relationship to Phosphorylcholine Antibodies

The immunization of mice with antigens containing phosphocholine (PC) produces humoral antibodies to phosphocholine that are idiotypically and or structurally related to PCBMP (Cohn et al., 1969; Sher and Cohn, 1972; Cosenza and Kohler, 1972; Lieberman et al., 1974; Kohler, 1975; Cosenza et al., 1977; Claflin, 1976a,b). PC-specific antibodies in mice of various strains have been induced by immunization with PnC polysaccharide (Sher and Cohn, 1972); R36A pneumococci (Cosenza and Kohler, 1972; Lieberman et al., 1974; Cla- flin et d., 1974); PC-carrier conjugates, e.g., flagellin, Limulus hemocyanin (Lee et al., 1974; Gearhart et al., 1975), infection with Ascaris suum eggs (Brown and Crandall, 1976). Using these antigens there is general agreement that the antibodies elicited to the PC hapten are restricted (Sher and Cohn, 1972; Cosenza and Kohler, 1972; Lee et al., 1974; Claflin, 1976a,b).

Comparisons of anti-PC to PCBMP have been made serologically using anti-idiotypic antisera (Cohn et al., 1969; Cosenza and Kohler, 1972; Sher and Cohn, 1972; Lieberman et at., 1974; Claflin et al., 1974; Claflin and Davie, 1974; Gearhart et al., 1975) or structurally using isoelectric focusing banding patterns (Claflin et al., 1975; Cla- flin, 1976a,b) or more directly with partial amino acid sequences (Rudikoff and Claflin, 1976).

Structural analysis (IEF, partial amino acid sequences) of PC antibodies has revealed restricted heterogeneity of VL (Rudikoff and Claflin, 1976; Claflin, 1976a,b). Banding patterns characteristic of VK-22 (TEPC15, HOPC8, S107); VK-24 (MOPC167, MOPC511), and VK-8 (McPC603, W3207?) have been found in anti-PC from most conventional strains (C58, CBA, C57BU6, DBAI2, SWR, A, CE), but only VK-22 predominates in anti-PC of BALB/c. The A/J PC antibody VL chains have been partially sequenced, and the three amino-terminal sequences of VK-22, VK-24, and VK-8 have been identified (Rudikoff

ANTICEN-BINDING MYELOMA PROTEINS OF MICE 185

and Claflin, 1976) in roughly equivalent yields, thus supporting the IEF findings. Claflin (1976a,b) has found a genetic polymorphism of the VK-22 chains and has described two phenotypic forms that can be distinguished by IEF banding patterns. The variant K-PC8-A is found in AKWJ, C58/J, RF/J, PUJ; K-PCS-B is found in most other strains, including BALB/c, C57BL, and A/J. The phenotypes are detected as a slight difference in the characteristic VK-22 I E F banding pattern.

Rudikoff and Claflin (1976) identified the characteristic VH-4 amino-terminal sequences in A/J PC antibody. In contrast to the VH-4A BALB/c sequence, two amino acids were found at position 23: Glx and Ala. One PCBMP of CB-20 origin, the CBPC2 myeloma, has been partially sequenced, and two differences from the BALB/c VH-4 sequence have been found: at position 14 there was a Ser instead of Pro, and at position 16 there was an Arg instead of Gly (Claflin et d. , 1975). These structural differences may reflect strain-specific polymorphisms in the VH-4A structural genes, but more data are needed to c o n h this.

Anti-idiotypic antisera to PCBMP have been extensively used in cellular and genetic studies of the immune response to PC antigens in mice, and because of this some discussion of the complexity of these idiotypes is given. PC antibodies from BALB/c mice and other inbred strains regularly share idiotypes with certain PCBMP (Sher and Cohn, 1972; Cosenza and Kohler, 1972; Claflin et al., 1974; Claflin and Davie, 1974; Lieberman et al., 1974; Lee et al., 1974; Gearhart et al., 1975). Thus far the most readily demonstrable idiotypes in PC antibody are those associated with VK-22-VH-4A PCBMP (S63, S107, TEPC15, and HOPC8, etc.; see Tables 11, VII). Several types of anti- idiotypic antisera are elicited by immunization with VK-22-VH-4A proteins. Anti-idiotypic sera can be produced relatively easily in appropriate inbred strains such as A/He, A/J, and CE/J (Potter and Lieberman, 1970; Sher and Cohn, 1972; Sakato and Eisen, 1975) or in heterologous hosts (Claflin and Davie, 1974). Similar antisera can also be produced in BALB/c mice, but only when germ-free hosts are used (Sakato et al., 1976). The determinants identified by the various antisera are defined by cross-specificities and by whether they can be inhibited by hapten. As yet the determinants have not been allocated to specific L or H chains or specific secondary structures. Also it is not clear that any of the antisera distinguish alleles of VK-22 or VH-4A isotypes.

Sakato and Eisen (1975) produced antisera to TEPC15 in A/J and CE/J mice and demonstrated they cross-reacted with the VK-24-VH- 4D protein MOPC511. The idiotype T15 (M511) then is shared by

186 MICHAEL POTTER

VK-22-VH-4A and VK-24-VH-4D proteins. Since these have very different L-chain subunits, the antisera may be specific for structures on VH-4A and VH-4D chains. Heterologous rabbit idiotypic antisera have been prepared to the VK-22-VH-4A protein HOPC8 (Claflin and Davie, 1974). The antisera have been specifically purified by absorption onto a solid-phase absorbant containing HOPCS protein and elution with hapten. The antibody thus obtained is site specific (inhibitable by hapten) and is also specific for only VK-22-VH-4A PCBMP and some species of PC antibody (Claflin and Davie, 1974). Antisera produced in germ-free BALB/c mice to VK-22-VH-4A PCBMP are also hapten inhibitable (Sakato et al., 1976). Some antisera produced in heterologous hosts identify specificities that are common to all VH-4 isotypes (Claflin and Davie, 1974).

Antisera to the PCBMP MOPC511 (VK-24-VH-4D), MOPC167 (VK-24-VH-4E), and McPC603 (VK-8-VH-4C) can be produced in BALBlc mice (Sakato and Eisen, 1975) as well as other appropriate inbred strains (Potter and Lieberman, 1970). Most of these antisera recognize only individual specificities. It should be noted that most Id1 determinants in the P6GALBMP (Mushinski and Potter, 1977) and GPlFBMP (Lieberman et al., 1975, 1976) are difficult to demonstrate in antibodies of the corresponding binding specificity.

There has been some difficulty in demonstrating the MOPC167 and McPC603 idiotypic determinants with homologous antisera in PC antibodies (Lee et al., 1974; Rudikoff and Claflin, 1976). This may be due to the fact that these antibodies recognize individual antigenic specificities (IdI) unique to the myeloma protein. The availability of IdX determinants for VK-24-VH-4E and VK-8-VH-4C and VK-8- VH-4B or other combinations may reveal similarities between PC antibody in PCBMP.

The T15 (M511) and or S63-Tl5 idiotypes are useful reagents for the study of PC-specific precursor B-lymphocytes (Cosenza and Kohler, 1972; Sigal et al., 1975; Gearhart et al., 1975). B-lymphocytes can be converted to Ig-secreting cells in uitro and the secreted Ig assayed with the anti-idiotypic antibody. Also, B-lymphocytes expressing either of the above-mentioned idiotypes can be suppressed by anti- idiotypic antibody, and this leads to the appearance of other types of PC antibody-producing clones (Cosenza et al., 1977). It has been estimated that the frequency of PC-specific lymphocytes in the mouse spleen is 1.6 to 52 per 1O6splenic B-lymphocytes and that about 20% of these cells do not express the T15 (M511) or S63-Tl5 idiotypes (Gearhart et al., 1975; Sigal et al., 1975). There is considerable variation in the percentage of T15 idiotypes among PC-specific


B-lymphocytes, and this may range from 2 to 50% (Gearhart et al., 197.5).

Strain differences in the expression of the VK-22-VH-4A idiotypes have been reported (Sher and Cohn, 1972; Lieberman et aZ., 1974). At the present time it seems probable that these reflect different levels of idiotypes in PC antibody rather than polymorphic differences determined at the structural gene level. For example, the prototype strains that express relatively low amounts of TEPC15 (M511) idiotypes are: (1) C.57BU6 (Lieberman et al., 1974) and (2) N J (Sher and Cohn, 1972), both of which express all three (VK-24, VK-22, and VK-8) V,, isotypes in antibody; by contrast strain BALBlc expresses predominantly VK-22 (Claflin, 1976a,b).

4 . Comments

In contrast to the immunochemically oriented studies with the DNPBMP, the PCBMP have proved to be both immunochemically and biologically interesting. In mice there are natural antibodies to PC. The PCBMP of mice appears to be derived from this group of noimal antibody-forming cells. IEF, amino acid sequence, and idiotypic data clearly show the close structural similarities of the VK- 22-VH-4A PCBMP and anti-PC. The specific natural PC antigens are not known, although several PC antigens have been isolated from the microflora of the mouse. Further studies on natural immunity are needed to clarify the role of antigen in the pathogenesis of plasmacytomas producing PCBMP.

The PC response in the mouse is restricted. Thus far, three VK isotypes and four very closely related VH-4 isotypes are involved in forming PCBMP. All possible VL-vH combinations have not yet been isolated.

Strain-specific polymorphisms in PC immune responses appear thus far to be caused by regulatory differences involving differential expression of 3 clonotypes VK22-VH4A, VK24-VH4D, VK8-VH4C.

C. GALACTAN (P6Gal)

Sher and Tarikas (1971) found that the 5539 IgA myeloma protein from the Salk collection precipitated P-azophenyl /3-D-galactoside bovine y-globulin. Inhibition of precipitation was achieved with p-NOz-phenyl /3-D-gahctoside, isopropyl thiogalactoside, and weakly, with a-D-fucose. Seven GalBMP have been found in the NIH collection (Table XIII). These were identified as precipitins for one or more antigens containing /3( 1,6)-~-1inked galactopyranose groups such as Larchwood arabinogalactan, gum ghatti, galactans isolated from wheat

188 MICHAEL POTTER

TABLE XI11 BGGALBMP

Fluores-

changes" Isotype cence

Myeloma Strain IgcH VK VH (AFmaX) Reference

JPCl CBPC4 SAPC 10 XRPC24 XRPC44 TEPC19I J539 TEPC601

BALBIc

BALBfc BALBJc BALBlc BALBfc BALBlc BALBfc

CB-20 A 4 18 A 4 1B A 4 1B A 4 1B A 4 1B A 4 1B A 4 1B A 4 1B

None

None +30.2 -8.3 None +28.3 + 20.8

Rudikoff et al. (1973) Mushinski and Potter (1977) Potter et al. (1972) Rudikoff et al. (1973) Rudikoff et al. (1973) Potter et al. (1972) Sher and Tarikas (1971) Manjula et al. (1975)

" See Jolley et al. (1973, 1974), Manjula et al. (1975).

germ (used in the mouse diet), and galactans obtained from the hardwood shavings used in both conventional and germfree mouse bedding (Potter, 1971; Potter et at., 1972). Using the S10, T191 GalBMP and arabinogalactan as an antigen in inhibition of precipitation tests, /3( 1,6)-~-linked galactan oligosaccharides @6Gall to /36Ga14) increased in inhibitory efficiency with progressively larger oligosaccharides. The /36Ga14 oligosaccharide was the largest available inhibitor.

These findings, though, clearly indicated that the specificity was directed to /3( 1 -+ 6) D-linked galactan polymers. Glaudemans et al. (1973) demonstrated that six of the GalBMPs precipitated with beef lung galactan, a polysaccharide of mammalian origin containing /3( 1 + 6) D-galactan side chains. Eichmann et al. (1976) demonstrated that the 5539 P6GalBMP precipitated with (snail Helix pomatia and Lymnaea stagnalis) galactans. Galactan-containing material was also identified in Sepharose 2B after treatment with propionic acid (Eichmann et al., 1976).

1 . V-Region Sequences of P6GalBMP

Partial amino acid sequences of the GalBMP have revealed that all VL domains belong to the VK-4 isotype and all VH belong to the VH-1B isotypes (Rudikoff et al., 1973; Potter et al., 1976a,b, 1977) (Tables I, 11, VI, and VII). The VH-lA, -1B, and - l C isotBes share a similar amino-terminal sequence but differ in the remainder of the sequences. They are tentatively divided because of the associations with different hapten-binding specificities (Tables VI and VII). The CDR and much


of the framework of J539 has been sequenced by Dr. N. Rao, NCI (see Potter et al., 1976b). The isotype restrictions seen in the V regions of the P6GalBMP suggest that only a few genes can interact to form P6Gal- binding sites.

2. Galactan Binding Site

P6GalBMP have been purified by immunoabsorption on Sepharose-bovine serum albumin-pazophenyl P-D-thiOgalaCt0- pyranoside columns. Purified antibodies interacted with Gal oligomers, and changes in the fluorescence spectra were quantitated in the Perkin-Elmer MPF-3 fluorescence spectrophotometer (Jolley et aZ., 1973, 1974; Jolley and Glaudemans, 1974; Manjula et al., 1976a). In the fluorescence titration method, both enhancement and quenching of the intrinsic fluorescence of the protein (preferably Fab) can be used to measure ligand binding and determine K, (Jolley and Glaudemans, 1974). In the fluorescence spectrophotometer, the protein is irradiated at 280 nm or 295 nm, and the fluorescence is monitored between 330 and 350 nm. When excited at 280 nm, the fluorescence is due to tryptophanyl residues, and the energy emitted can be changed by bound ligands. Thus a change in the microenvi- ronment of a tryptophanyl residue can be reflected in a change in the fluorescence. Changes in AF,,, range from -8 to +30, depending on the protein (Jolley et ul., 1973; Pollet et al., 1974; Streefierk and Glaudemans, 1977). When the absolute value of AF,,, exceeds 2 3 % , a K , can be suitably derived.

Three of the p6GalBMP showed maximal increases in fluorescence of 28.3% (J539), 30.2% (XRPC24), and 20.8% (TEPC601); one protein (XRPC44) showed a quenching of fluorescence (-8.3%); and three were unchanged (S10, T191, and J l ) .

The plot of the percentage increase in tryptophanyl fluorescence intensity (in the GalBMP) obtained with different hapten-protein molar ratios permitted calculation of ij (the fraction of sites occupied by hapten). The association constant (K,) for polymers, monomers, or Fab of those GalBMP that were associated with fluorescence changes could be obtained by the slope of the plot of d c vs 5 (Jolley and Glaudemans, 1974; Manjula et al., 1976a). Using this approach, straight-line Ulc vs U Scatchard plots were obtained (Jolley et al., 1974). The percentage contribution of Gal2, Gal3, and Gal, oligosaccharides was determined for XRPC24 and 5539. In all cases studied, the terminal nonreducing galactosyl residue contributed most of the binding energy, the second galactosyl residue contributed more than a third, and residue 4 contributed very little (Table XIV). Thus the hap-

TABLE XIV BINDING PROPERTIES OF MYELOMA PROTEINS THAT BIND NEUTFWL POLYSACCHARIDES

Percent contribution to total Maximal Olige in binding oligomer No.: binding KO

Myeloma saccharide (M, (M-' Refer- protein series Method' 1 2 3 4 5 6-7 cal/mole) x 10') ence

~~ _ _ _ _ ~

XRPC24 J539 W3129 QUPC52 W3434 w3082 UFC61

ABE4 ABE47N EPC 109 MOFC104E J558

~~ ~

P6CAL S6CAL &G &G &C G6, PIFU 6G, PIFh alC, PIFh a lC , PIFh a l G , PlF" a lC , PIFh a3G a3G

FT FT E D ED IP IP IP FT FT FT FT IP IP

~~

47 69 96 54 73 93 56 61 91 5 5 72 - 74 94 - 43 95 - 52 97 - 51 75 - 56 75 - 61 78 - 53 72 - 0.39 0.94'

(a3G 5-7 > ff3G4

100 100 91 88 98

100 100 94 98 100 95 1'

> a3G3

- - - -

loo 95 96 100 - - - - - - loo - loo - 100 - 100 -

1' - > a3G2)

-7430 - 7530 -7180 -5340 -7500 -7560 - 7560 -7470 - 7400 -6950 - 7630 - -

2.93 3.44 1.9 0.084

3.6 3.6 2.9 2.7 1.2 3.8 0.29

-

-

a a b b C

C

C

d d d d e e

" Jolley et al. (1973, 1974). " Cisar et al. (1975). Cisar et al. (1974).

" D. Streefkerk and C. P. J. Glaudemans, unpublished results. Leon and Young (1971). FT, fluorescence titration; ED, equilibrium dialysis; IP, inhibition of precipitation.

I' Series of oligosaccharides having a general structure (1 PFruf 2). + lPFruf 2 + 6 Glc (n = 0, 1, 2). The number in column 4

" Series of oligosaccharides having a general structure (lPFruf2).+ lPFrufl+ 1 A l c p (n = 0, 1, 2, 3). Further, see remark in

i Relative inhibitory power.

applies to the total number of sugar residues; there always being only one glucose moiety.

footnote g.


ten for the X24 and J539 proteins appears to be the terminal trisaccharide. In a detailed comparison of X24 and J539, small differences in binding were noted for the first and second sugars (Jolley et al., 1974; for a detailed discussion, see review by Glaudemans, 1975). Jolley et al. (1974), have made an extensive study of the binding of some thirty sugar derivatives with XRPC24 and J539. From these studies, the contributions of individual groups could be assessed (Fig. 7). In the disaccharide 6-@P-D-ga~actopyranosyl-D-galactose, the most important atoms involved in binding are localized on one side of the disaccharide. The ring oxygen atom and the hydroxyls of the first galactosyl group contribute 54.5% of the binding energy for J539 and 47% for XRpC24 (Table XIV). Glaudemans et al. (1975) extended this study by preparing bulky substituents of the C'-6, C-1, C-2, C-3, and C-4 in the disaccharide, and they found that these did not interfere with effective binding. These data support the hypothesis that the galactose disaccharide is bound by contacts through groups on only one of its surfaces and apparently does not penetrate into a pocket or cavity.

3. GalBMP Zdiotypes

Each of the PGGalBMP has its own characteristic individual antigenic specificity (IdI) that can be demonstrated by hemagglutination inhibition (Rudikoff et al., 1973; Mushinski and Potter, 1977). Thus it can be predicted that there will be some sequence differences among the eight available PGGdBMP. Recently, cross-reactive idiotypic determinants (IdX) have been found within the PGGalBMP group. These occurred only on the GalBMP and did not cross-react with over 100 non-GdBMP (Potter et al., 1976b; Mushinski and Potter, 1977). Such group-specific idiotypes can be detected with both mouse and rabbit idiotypic antisera. They provide further evidence that the PGGalBMP are a family of structurally related proteins, which are controlled by

0.1 02.l5 0.1 /-

/ '.

\ I

FIG. 7. p-DGaIactosyl-( 1 + 6)-D-gdactopyranose showing the fractional contributions of OH to binding as determined by increase in fluorescence intensity on hapten binding. Adapted from Jolley et u Z . (1974).

192 MICHAEL POTTER

closely related VL and VH structural genes. The P6GalBMP and InuBMP (Lieberman et al., 1975, 1976; Potter et al., 1976b) families appear to reflect a similar type of genetic control in this respect. If further structural findings reveal only a small number of amino acid differences among the closely related VK and VH chains then the idiotypic differences observed may emanate from a few discrete changes in amino acids that proejct out to the surface of the domain.

The relationship of the Id1 and IdX in the P6GalBMP to idiotypes on antigalactan antibodies is currently being explored. Antibodies to galactans have been raised in mice by immunization with gum ghatti and by Gal-KLH (keyhole limpet hemocyanin) (Mushinski and Potter, 1977). The IdX have been thus far identified on the anti-gum ghatti antibodies (Mushinski and Potter, 1977; Potter et al., 1976b). Only one Id1 has been found, and this is the XRPC24. The failure to find the other Id1 in antibodies suggests that some GalBMP proteins may be idiotypically unique, or that they occur infrequently or stochastically. If so, these phenotypes might result from somatic mutation.

Manjula et al. (1976a) have made recombinant molecules with H and L chains from P6GalBMP and found every recombinant to retain Pl,G-D-galactan specificity. It is interesting that one recombinant showed higher binding affinity for a P6Gal ligand than either the donor of the H or L chains. These authors also found that the idiotype of any recombinant always resembled that of the H-chain donor.

D. GLUCAN (a3G)

Three myeloma proteins-IgM MOPC104E (Leon et aZ., 1970), IgA J558 (Blomberg et al., 1972, 1973), and UPClO2 (Cisar et al., 1974)- that bind a1,3 dextrans were all tentatively identified as precipitins for the B 1355S1,3 dextran. B1355S dextran was isolated and characterized by Dr. Allene Jeanes (1968). The MOPC104E (and probably the others as well) agglutinate SRBC coated with palmitoyl B 1355s dextran prepared by the Tsumita and Ohashi (1964) method. The IgM MOPC104E also fixes guinea pig complement upon binding to B1355S173 dextran (Leon et al., 1970).

The specificity of MOPC104E for a1,S-linked glucans (nigerose oligosaccharide) was demonstrated by Leon et aZ. (1970) using inhibition of precipitation. The more sensitive inhibition of compelment fixation could not be used, as these glucans were anticomplementary (Leon et aZ., 1970). The relative inhibition of precipitation ratios for a3G5:a3G4:a3G3:a3G2:a-CH3-~-Glc were 1 : 1 : 0.94 : 0.39 : 0.0018, suggesting that the combining site was complementary for a3G (Table XIV). Lundblad et al. (1972) studied the specificity of the binding site of 5558 using the B1498S dextran (which contains about


27% a1,3 linkages). In this study a3G5 and a3G4 were better inhibitors than a3G3, suggesting that the binding site was complementary for a hapten of the size of a pentasaccharide. Cisar et al. (1974) found that the binding site of UPC102 was similar to MOPC104E. The association constant of MOPC104E a-D-glucopyranosyl( 1 + 4) for a-D-glucopyranosyl( 1 + 3)-D-glucitol was 3.6 X lo4 it-’ (Young et al., 1971).

Leon et al. (1970) noted that MOPC104E also precipitated different dextrans that contain varying proportions of a1,6-, a1,4-, a1,2-, and a1,G-linked glucose linkages. The dextrans B1299S, B1254, B1141, B1396, and B1399, which lacked a1,3 linkages, were also precipitated. Lundblad et al. (1972) found that J558 precipitated optimally with dextrans containing a relatively high content of a1,3 linkages, but some precipitation was obtained with dextrans with low or no a1,3 content. The chemical basis for these reactions is not explained. A natural antigen containing a1,Slinked glucans has not yet been identified in the mouse environment.

Structural studies of the binding sites of a3GBMP have not yet been made, despite the fact that the complete sequence of the MOPC104E and 1558 VA-1 L chains has been determined (Weigert et al., 1970; Appella, 1971). The VH-7 region sequence up to the first invariant Trp,, has recently been made available (Hood et al., 1976). The MOPC104E and J558 chains differ from each other in this partial sequence in three, or possibly four, positions, suggesting that these chains will be controlled by two different VH-7 structural genes.

Idiotypic antibodies to the J558 IgA myeloma proteins have been prepared in the strain N H e mice. Using an NJ anti-J558 : J558 system in a solid phase radioimmunoassay, Carson and Weigert (1973) demonstrated that the binding was inhibited by B1355S dextran. The idiotypic antibody was therefore directed to a determinant near or in the antigen-binding site.

Carson and Weigert (1973) also studied the competitive ability of recombinant myeloma proteins with different VA- 1 chains to compete with 5558 for its anti-idiotypic antibody. A series of recombinant molecules were made by recombining the 5558 H chain with a series of A - 1 chains. Recombinants with S104E, MOPC104E, H2020, S176, and RPC2O VA-1 were as effective as 5558 in binding the anti-idiotypic antibody. The S178 A chain, which differs from MOPC104E at three positions, namely, Ser for Asn at position 25 (in Ll), Gly for Asn at position 52 (in L2), and His for Arg at position 97 (in L3), was one-sixth as effective as J558 in competing for the binding site. MOPClO4E by itself was 20 times less inhibitory than 1558 on a mole for mole basis, thus implicating the VH domain in determining the structure of the

194 MICHAEL POTTER

idiotype, since both have the same VA-1 L chains. It will be recalled from a previous discussion that the MOPC104E and 5558 myeloma proteins have different sites (Lundblad et al., 1972; Cisar et aZ., 1974) and that MOPC104E is more complementary for a3G3 whereas 5558 is more complementary for a3G3.

Binding of a3G oligosaccharides apparently involves contributions from both VH-7 and VA-1 CDR. The J558 idiotype appears to be structurally determined by VH-7, whereas the a3G binding depends upon both VH-7 and VA-1. The VH-7 H3 structure will be particularly interesting to determine.

Blomberg et aZ. (1972, 1973) have found that antibodies induced in several strains of mice, including BALB/c, by the injection of B 13558 dextran can also inhibit the A/J anti J558 : J558 system, demonstrating for the first time structural similarities in induced antibodies and myeloma proteins with similar hapten-binding properties. The ability to form antibodies with the 5558 idiotype differs greatly among inbred strains. Some strains, such as C57BL, SJL, CBA, AKR, A/He, and NZB, do not produce antibodies to B1355S dextran that cross-react with the 5558 Id system. The responses of congenic strains and of recombinant inbred strains derived from nonresponding C57BL and responding BALB/c strains indicate that the genes controlling the response to dextran B1355S are linked to the IgCH allotype-locus in the mouse.

Immunization with B1355 polysaccharide does not produce a large amount of antibody, and in this respect the immunization resembles many T-independent responses in mice by having relatively few PFC and low antibody levels (0.6-0.9 mg/ml) (Hansburg et al., 1976). Re- cently, Hansburg et al. (1976) have discovered that E. coli B contains an antigen that cross reacts with dextran B1355S; i.e., E. coli B organisms can absorb out J558 myeloma protein. Immunization with B1355S and E. coli B produces in appropriate strains of mice, a large amount of antibody (about 14.5 mg/ml) that is spectrotypically similar to B1355S antibody. This antibody strongly reacts with idiotypic sera produced in rabbits to MOPC104E, UPC102, and J558. Recently, Hansburg et al. (1977) have shown that B1355S dextran-E. coli B immunization of C57BW6 mice induces in occasional individuals antibody with a3G specificity but does not share idiotypic specificities with a3GBMP of BALB/c.

Since the VA-1 genes are not absent in most of these strains of mice (Weigert and Riblet, 1976), the evidence suggests the possibility of a VH-7 gene defect, i.e., a lack of VH-7 genes in negative strains, or a regulatory inability or inefficiency to differentiate VH-7 and VA-1 in the same cell, or it could indicate the inability to activate B cells expressing VH-7 and VA-1 genes. Whatever the lesion, it potentially


appears to involve three VH-7 structures: VH-7 (MOPC 104E), VH-7 (J558), and VH-7 (U102) (Hansburg et al., 1977). The need for VH-7 structural studies in this system is apparent. Also, it would be of interest to determine whether C57BL mice or other negative strains have B-lymphocytes with the MOPC104, J558 idiotypes.

E. GLUCAN (a6G)

The myeloma proteins W3129, W3434 (from the Salk Institute), and QUPC52 (from the NIH) precipitate dextrans with a high content of a(l + 6) D-glucose linkages, such as the B512, N236, B1141, B1399, B1299, B1299S3, and B1424 dextrans (Cisaret al., 1974). Recently two NZB a( 1 + 6) dextran-binding proteins (NZB3936 and NZB3858) have been discovered (Hood et al., 1976). The availability of linear and branched oligosaccharides of the isomaltose (a6G) series has provided the means for approximating the size of the complementarity antigen-binding sites (Cisar et al., 1974, 1975). Precipitin inhibition, equilibrium dialysis, and fluorescence titrations were used to determine the size of the binding site and the cbntribution of the successive sugar residues in the oligosaccharide to the total binding energy. These studies are summarized in Table XIV. As may be seen there are striking differences between the W3129 and QUPC52 proteins. Near-maximal binding of the W3129 protein was achieved with 6G3; the disaccharide in the series contributed 61% of the total binding energy (Cisar et al., 1975). By contrast, the QUPC52 protein bound the disaccharide using only 5% of the total binding energy, and near- maximal binding is first achieved with the pentasaccharide. The QUPC52 is another example of a protein (see 5558 above) that is not binding its antigen by the terminal end of the oligomer chain. Thus Cisar et al. (1975) predict that the binding site of QUPC52 is a long groove that can bind the linear dextran anywhere along the chain.

Structural Studies, Zdiotypes

The W3129 and W3434 VK chains are both in VK-3; no data are available on QUPC52. The VH regions are available only for 20 residues; W3434 appears to be VH-1, and W3129 has three differences suggesting that it will be placed in a new VH isotype. The two NZB a1 + 6 dextran-binding proteins have a similar amino-terminal sequence in the VH but striking differences from those of BALB/c origin, suggesting that there may be considerable polymorphic differences if it is subsequently demonstrated that these VH domains are homologous (Hood et al., 1976) to W3129 or W3434. This, however, remains to be done by immunochemical studies.

Idiotypic antibodies to W3129 have been prepared in strain NHeJ

196 MICHAEL POTTER

mice b y Weigert et al. (1974). In a solid-phase radioimmunoassay, the A/He anti-W3129 : W3129 was inhibited by isomaltose oligosaccharides as follows: 06G6 = &G5 >> a6G4 > a6G3 >>> a6G2.

The trisaccharide, which had nearly the maximal binding energy (Cisar et al., 1975), was very close to a maximal inhibitor of the anti- idiotypic system. Thus the anti-idiotypic antibody was directed to determinants near the binding site. The W3434 protein partially inhibited the W3129 idiotype, revealing a cross-specific idiotypic relationship. The QUPC52 failed to inhibit the W3129 idiotypic system, suggesting that this protein may be composed of V regions that are structurally unrelated to W3129.

F . FRUCTAN-INULIN (GplF) AND GRASS LEVAN (Gp6F)

After the demonstration by Grey et al. (1971) that the IgG, J606 myeloma protein bound several levans, a relatively large number of myeloma proteins that bind the heterolinked (2 + 6 and 2 + 1) polyfructan from Aerobacter laevanicum have been identified (Cisar et al., 1974; Lieberman et al., 1975). These proteins can then be subdivided into two distinct hapten-binding groups: the larger group of 11 inulin-binding myeloma proteins (Gp 1FBMP) having specificity for p2 + 1-linked fructans (e.g., inulin), and a smaller group of 4 proteins having specificity for p2 + &linked fructans (grass levan) the GP6FBMP (Table XV).

In nature, most fructans have terminal glycosyl groups (Aspinall, 1970). Thus, these polysaccharides are not true homopolymers in a

TABLE XV BACTERIAL LEVAN-BINDING MYELOMA PRWEINS

Inulin (GplF) Grass levan (GP6F) V regions: VK-11-VH-5 VK-10-VH-1

Strain C, class

BALBIc kc J606(C3) UFC 10(GzJ I gA W3082 1’54 76

UPC109 ABPC48 ABPC47N TEPC957 ABPC45 ABPC4 TEPC803 MOFC702

NZTEPC6906 NZTEPC3660

CXBH IgA CXBHPC2 NZB I gA


TABLE XVI VH-5 REGIONS OF INULIN-BINDING MYELOMA PROTEINS (GBIFBMP)

H 3 I I I H2,

I ' I H ? ' I I 1 I I I 1 I 1

NH, I I 48 I 80 92 97 ; 115

ABPC4 1 I V I I Y G Y T I : T

I-)--- G Y T f I P

EPC 109 ;-1--- I : I I 1 G H T I ; T

+

I S-T U PC6 1 I V I I

ABPC47N I I v I I Y A Y + I I I I I I

strict sense. Furthermore, oligosaccharides isolated from polyfructans can have a terminal glucosyl group, in which case they will be called Gpl- or GP6-fructans. The availability and purification of two linear polyfructans-inulin, which is composed of p2 + 1 fructan linkages, and grass levan, which is cornposed of p2 + 6 linkages-makes it possible to make immunoabsorbents and isolate the myeloma proteins (Vrana et al., 1976; J. Tomasic and C. P. J. Glaudemans, unpublished, 1977).

A limited number of studies have been carried out on this group of proteins to characterize the hapten-binding sites.

Three-dimensional space-filling models of p2 + 1- and p2 + 6- linked fructans, as pointed out by Cisar et aZ. (1974), show that p2 + &linked fructan has an extended linear structure, while the p2 + l-linked fructan has a twisted structure. Thus, the binding sites for these haptens must have different shapes and contours. The V regions of the GplF-binding proteins are composed of VK-11-VH-5 while the GP6F binding proteins are composed of VK-10-VH-1 group (Tables I, 11, VI, and VII).

1. G p l F B M P

Comparative hapten-binding studies using inhibition of precipitation have been completed on W3082 and UPC61 (Cisar et aZ., 1974), two of the GplF-binding proteins. They have similar sites, and the maximum contribution to binding is contributed by the terminal three sugars (Table XVI).

In that work, the binding of two murine meyloma proteins with anti-inulin specificity were studied with oligomers having a structure PF2 + lPF2 +, but terminated by a &linked reducing glucose moiety linked to fructose. In inulin, the polysaccharide chains are nonreducing and are terminated by an a-D-glucopyranose unit as in suc-

198 MICHAEL POTTER

rose. Thus, the ligands studied by Cisar et al. are unlike those of the (supposed) homologous antigen.

Streefierk and Glaudemans (1977) have studied the 2 --* 1 linked fructofuranosyl-containing oligomers terminated, as in sucrose, by an a-D-glucopyranose residue. Thus, those oligomers truly represent the structure of inulin. In their studies on A4, A47N, E 109, and U61, these workers showed that maximum complementarity was achieved with the tetrasaccharide (F32G11) in the case of A47N and A4, and the pentasaccharide (Fd2G1') in the case of El09 and U61. As the glucose moiety itself apparently contributes little to the binding of these ligands, the corresponding immunoglobulin combining sites were computed to have a maximum dimension of 14 x 14 x 7 a and 15 x 14 x 10 %i, respectively.

It is interesting that the unusually oblong combining site in the hypothetical model of El09 would very nicely accommodate a compact determinant of that size (Fig. 1C).

They also found that methyl a-D-glucopyranose itself appears not to bind in the combining site, thereby indicating that these four myeloma proteins bind along the antigen chains. An interesting finding, furthermore, was that the K,'s of whole inulin and the Fab' fragments of these four proteins were identical to those obtained with the best binding oligosaccharides. It should be remembered that inulin is a rather small antigen (molecular weight 5000).

The idiotypes of the GplF-binding BALBlc IgA proteins have been studied by Lieberman et al. (1975). Each of the proteins thus far, as was the case for the PGGalBMP, has a characteristic individual idiotype (IdI). Groups of GplFBMP share cross-specific idiotypes (IdX). The IdX idiotypes are highly specific for GPlFBMP; 120 non- GPlF binding proteins failed to inhibit these reactions. Most IdX idiotypic systems are inhibitable by haptens, and hence these idiotypes are located in the active site region (Lieberman et al., 1975). A remarkable complexity of IdX determinant associated with the Gp lFBMP suggests there will be a corresponding heterogeneity in the primary structures. However, complete VH sequences on four of the proteins have revealed only a few amino acid replacements (Table XVI) (Vrana et al., 1976, 1977) and the preliminary indications on the most variable part of the V,, the L1 region, have again revealed only small differences (Table V). Available partial sequences have not indicated striking sequence differences, implying that the idiotypic differences (IdI) will be controlled by subtle changes in primary structure.,

The sequence differences of GPlFBMP thus far available are summarized in Tables I, 11, VI, VII. The VH-5 regions have a large dele-


tion in the H 3 region, as compared to McP603, the deleted region being between Asn,, and Phelos. The available sequences, however, have been used to build a hypothetical model (E. Padlan, M. Vrana, M. Potter) of the ABPC47N binding site region (Fig. 1) (Potter et al., 1977). In the model, the secondary structures of H1, H2, L2, and L3 were borrowed from McPC603. The H 3 region was remodeled by bridging the antiparallel p-structures between Thr,, and Phe,,, with Gly. The L1 structure was borrowed from RE1 V, (Epp et al., 1974), which has a similar length. As may be seen, the surface of the G p l F protein is drastically different from those in DNPBMP or PCBMP (see Fig. 1). Instead of a pocket or a small cavity, there is a large cavity extending across the binding surface. The CDR regions are rich in amino acids such as Gln, Asn, Glu, Asp, Thr, and Ser, which could potentially form H bonds with sugars in the antigens if brought into close proximity. Streefkerk and Glaudemans (1977) have shown that the best binding oligosaccharide Fi2G,' has a compact, almost helical, structure. Such an antigen would require a widely spaced combining site.

2. G P 6 F B M P

Less information is currently available on the GP6FBMP (Table XV). These proteins can be specifically isolated by an immunoabsorbent made with grass levan (J. Tomhsic and C. P. J. Glaudemans, unpublished observations). The L chains belong to VK-10 and have the same length in L1 as do VK-1, but they differ in primary structure. The VH chains are VH-lC, and the structures of H 2 and H 3 are not known. Difficulties in isolating Gp6F oligosaccharides have made it impossible to determine relative binding specificities (Cisar et al., 1974). The G P l F oligosaccharides do not inhibit precipitation of these proteins with levan P6 (Cisar et al., 1974).

G. N-ACETYLGLUCOSAMINE

In a screening program involving 275 mouse meyloma sera, Vicari et al. (1970) identified one IgA myeloma protein (S117) which precipitated with BGG-p-azophenyl-P-N-acetylglucosamide; S 117 also precipitated with group A streptococcal polysaccharide; with blood group H substance after the first and third stages of periodate oxidation and Smith degradation, and with p-teichoic acid. Inhibition of precipitation was optimal with ~-D-GNAc( 1 --* 3)-[p-D-GNAc( 1 + 6)]-~-Gal. This inhibitor was twice as effective as ~-D-GNAc( 1 + 3)-P-~-Gal, and 3 to 4 times more efficient than ~-D-GNAc( 1 + 6)-~-Gal .

Structural studies of S 117 have revealed that the L chains belong to

200 MICHAEL POTTER

VK-13. There are structural differences in L1, however, which distinguish S117 from the P6GALBMP VK-13L chains. It is also surprising that the VH sequence of S 117 is remarkably similar to the VH sequences of VH of the GalBMP.

The immune response to streptococcal A polysaccharide is heterogeneous, but one of the regularly elicited clonotypes has idiotypic similarities to the S117 myeloma protein. Using a guinea pig anti- idiotypic antibody to S 117, Berek et at. (1976) demonstrated that some strains of mice produced antibody capable of inhibiting the precipitation of S117 with the anti-idiotypic serum whereas others did not. Thus the S 117 myeloma protein contains a polymorphic idiotypic determinant, and this is a useful genetic marker. The marker is linked to the IgcH allotype locus (Berek et al., 1976).

H. FLAGELLIN The MOPC467 IgA myeloma protein was originally found to pre-

cipitate with heat extracts prepared from many of the major Sal- monella serogroups. Pasteurella pneumotropica, and Herellea uag- inicola (Potter, 1970,1971). The antigen in heat extracts was destroyed by treatment with trypsin or phenol. Smith and Potter (1975) identified the antigen in cultures of Salmonella milwaukee; the antigen was located on the flagella. Polymerized flagellin of S. milwaukee was also precipitated by MOPC467.

Immunization of strain BALB/c mice with flagellin from S. milwaukee raises antibodies which share idiotypic determinants with the MOPC467 myeloma protein (Smith et al., 1977). In addition, isoelectric focusing patterns of isolated L chains from MOPC467 and anti- flagellin antibodies have revealed remarkably similar binding patterns indicating that MOPC467 is closely related to antibodies of the same binding specificity. The MOPC467 is probably an antiprotein antibody and the deteimination of the structure of its binding site region may provide information on the properties of binding sites for proteins.

I. LIPOPOLYSACCHARIDES

Two IgA ( K ) myeloma proteins McPC870 and MOPC384 precipitate the lipopolysaccharides isolated by the Westphal method from Sul- monella tel auiu, S . tranoroa, and Proteus mirabilis sp 2. The lipopolysaccharides from S. tranoroa and S . tel auiu have not been chemically defined as yet, and the basis for the common antigen recognized by the two myeloma proteins is unknown. However, both polysaccharides contain rare trideoxy sugars (Luderitz et al., 1967,


1968). These proteins agglutinate SRBC coated with the lipopolysaccharides. Inhibition of precipitation or agglutination can be achieved with a-methyl-D-galactoside for the MOPC384, but not the McPC870, polysaccharide. The Proteus mirabilis species, which is the source of the lipopolysaccharides, was isolated from the gut flora of the BALB/c mouse. These organisms were used to immunize BALB-c mice, but the antibodies so induced did not share idiotypes with either MOPC384 or McPC870. Idiotypic antibodies raised to MOPC384 or McPC870 in A/J mice did not show any cross-reactivity by the precipitin method. The McPC870 and MOPC384 V, chains are VK-8, the same isotype as McPC603. The MOPC384 and McPC870 form another group of myeloma proteins, which may be related to natural antibodies in the mouse and are combinations of related VH and VL groups .

J. N-ACETYLMANNOSAMINE

The IgA MOPC406 myeloma protein precipitates with the lipopolysaccharide of Salmonella weslaco (Potter, 1970), a lipopolysaccharide that contains N-acetyl-D-mannosamine (Luderitz et aZ., 1967, 1968). N-Acetyl-D-mannosamine was found to inhibit the agglutination of S. weslaco lipopolysaccharide by MOPC406 (Potter, 1970). Rovis et al. (1972) demonstrated that the specificity of MOPC406 was for the P-pyranoside of 2-acetamid0-2-deoxy-~-mannosamine. A number of other D-mannosamine-containing polysaccharides failed to precipitate with MOPC406.

VI. Concluding Remarks

The emphasis in this review was given to the immunochemical propeities of the antigen-binding myeloma proteins (AgBMP) of BALB/c mice. The AgBMP, besides their immunochemical interest, have offered insights into several aspects of the biology of antibodies, in particular the genetic basis of diversity and the genetic control of antibody formation. This discussion will summarize some of the major points that are related to specific areas.

MyeZoma Proteins. It should be stressed that in the BALB/c myeloma system only a small fraction of proteins have a known antigen-binding activity. Binding activity as measured by most conventional immunochemical methods to a considerable variety of haptens and antigens has been demonstrated, however. For certain haptens, more than one myeloma protein with the same binding specificity occurs: e.g., for GPlF, 13 proteins; for PC, 11 proteins; for

202 MICHAEL POTTER

PGGAL, 8 proteins; and for GPGf, a3G, and a6G, 3 to 4 proteins. Many of the haptens are carried on antigens that can be isolated from the immunological environment of the mouse (microbial flora, diet, bedding).

No actual estimates of the relative number of AgBMP were given, because, first, the screening tests were often performed with only a fraction of available or accumulated serum samples and, second, only a limited number of antigens have been tested. The number of AgBMP found in the BALB/c mouse makes this nonetheless a potentially valuable system for asking further questions about myeloma proteins. Should all myeloma proteins be expected to have a reasonable binding specificity for a relevant antigen? Most myeloma proteins so far tested (about 90%) have no known binding activity. This lack of reactivity could be explained in several ways. First, the testing system has not yet employed the correct antigens or the correct method for detection. For example, there are very few studies in which myeloma proteins have been tested by radioimmunoassays for protein antigens. A second reason for not finding binding specificity in a screening might be attributed to a possible degeneracy in the immunoglobulin-forming mechanism. Possibly many VL-VH pairs are “misfits” and cannot bind any relevant antigen. Possibly too, somatic mutational mechanisms (particularly accentuated in neoplastic cells) have altered binding sites, so that they can no longer bind relevant antigens.

It is beyond the intended scope of this review to discuss in detail the pathogenesis of plasmacytomas in BALB/c mice and its relationship to normal plasma cell development. The data suggest that plasmacytomas in BALB/c mice are derived from cells that have matured to a late stage in B-lymphocyte development (see Potter and Cancro, 1977). If this reasoning is correct, most myelomas might be expected to be derived from cells that have interacted with and been selected by antigen. It has not been ruled out that the plasmacytomagenic process can bypass the influence of antigen in vivo and convert B-lymphocytes to plasma cells and plasmacytomas. If this latter mechanism does occur, one might expect that the myeloma proteins reflect the types of Ig formed in the pre-antigen selected B-cell population. In any case, it will be useful to study binding-site regions of myeloma proteins with no known antigen-binding activity with models for possible binding activities.

Antigen-Binding Sites. Immunochemical studies of AgBMP have provided considerable data on the detailed structure of antigen- binding sites. The three-dimensional structure of the PCBMP McPC603 has been determined by X-ray crystallography with the hap-


ten in the site. Contact amino acids have been identified and are po- sitioned in a small region so that the hapten can be effectively bound. X-Ray crystallographic studies of McPC603 and several human myeloma proteins have revealed the principle of conformational homology of the V-region framework. This has provided the basis for model-building hypothetical binding sites by adding known or remodeled Wu-Kabat CDR structures onto domain frameworks. Models may be tested by other independent means, such as ESR and NMR spectroscopy using hapten-spin label probes (see discussion in Section II,A), affinity-site labeling, and other techniques. Two hypothetical models constructed by Dr. E. Padlan and colleagues were shown in stereo displays of a-carbon skeletons. This technique, when refined, may provide a catalog of binding sites and obviate the need for X-ray crystallography of every Fv fragment. However, the three- dimensional structure of antibodies has been derived from only a few crystallographic studies; more will be required before we can be as- sured hypothetical model building is indeed reliable.

Hapten-binding sites studied thus far with both actual and hypothetical models appear to be located in the more external portions of the interface region of the VL and VH domains. Variations in the length and secondary structure of CDR drastically modify the detailed topology of this region and may form small pockets or large cleftlike cavities. A broad surface (30 A x 40 A) facing the solvent is formed by the convergence of the 6 CDR. While hapten-binding sites thus far appear to occupy only a portion of the potential antigen-binding surface created by the CDR, much more ofthis surface may be implicated when new crystal structures become available.

Genetic Basis of Antibody Diuersity. Myeloma proteins derived from the same genetic background, such as the inbred BALB/c mouse, can be used to deduce the genetic basis of antibody formation and diversity. The picture that appears to emerge is that antibody diversity is created to a large degree by the multiplicity of germ-line V-region structural genes, as reflected by the number of isotypes that have been arbitrarily defined in the BALB/c by amino-acid sequence data. A beginning catalog of mouse V-isotypes was named by a new system for convenience of discussion. An isotype was defined by three or more amino acid differences in the partial amino-terminal sequence of VL extending to Cysz3 or VH extending to Phez7, or five or more differences in the V region. Owing to incomplete sequence data and inadequate sampling, a complete VK- and VH-domain isotype listing is not available, and awaits further intensive sequencing. Thus far there are 28 VK isotypes; 2 VA isotypes (Table 11) and 17 VH isotypes (Table VII).

204 MICHAEL POTTER

The V-region isotypes associated with specific hapten-binding activities show several patterns. The simplest pattern is for the GPlF- and PGGALBMP, where 1 VK and 1 VH isotype appear to be involved in creating antibodies of the respective specificity. Amino acid sequences of 4 GPlFBMP VH regions revealed that the four proteins were indeed derived from the same isotype (as defined). There were, however, three or fewer sequence differences involving the framework that appeared to be randomly distributed. Partial sequences of the V, of the two groups and the VH of the P6GALBMP are consistent thus far with the assignment to a single respective V isotype. The a3GBMP may also be tentatively included in this category. In the second pattern there are 2 or more isotypes for VL and VH. The PCBMP have 3 VK isotypes and 4 VH isotypes. While the VK isotypes have many primary structural differences in the framework, the VH isotypes have very few, indicating a close relationship. The DNPBMP also reflect a similar degree of heterogeneity, but here only a limited number of proteins have been studied. There is fragmentary evidence that one isotype group may be involved in more than one binding activity; the best case concerns VA-1, which is found in both a3GBMP and proteins with no demonstrable binding activity.

The multiplicity of V-isotypes in the BALB/c mouse raises a problem that has just begun to be appreciated: How is the process of V-region pairing developed during differentiation? Is there some process that activates a single VL gene in a cell with a single VH gene in such a way that all the possible pairs are regularly produced? Or is there some controlled way in which specific VL and VH genes are activated in the same cell?

While there is no conclusive answer at the present time, the myeloma system provides a source of materials that can be tested for VL-VH pairs through the use of VL and VH isotype-specific antisera.

The complexity of the Ig-gene differentiation process is reflected in part in the problems that many workers are now having in interpreting polymorphisms in humoral responses that occur in different inbred strains of mice to the same haptens (Eichmann, 1975; Weigert et al., 1975; Weigert and Potter, 1977). Are these due to structural gene differences that have occurred in the species Mus musculus, or do they reflect polymorphisms in the process that regulate the differentiation of V-region structural genes?

A detailed comparative study of homologous V isotypes in different strains of mice is not yet available. Hood et al. (1976) are beginning such a study by obtaining sequences from the NZB myelomas. Thus we do not know how much variation has occurred within the species Mus in the complex loci that govern VL and VH structures.


While much of the heterogeneity of V-region structures in the mouse can b e plausibly explained as arising from multiple genes, there is substantial evidence from primary structural data (Tables 111, IV, and XV) and individual antigenic specificities (Mushinski and Potter, 1977) of a microheterogeneity involving a very few amino acids (4 or fewer per V region) that might be explained as arising from mutational processes affecting differentiated germ line genes (somatic mutations). This process appears to involve both the CDR as well as the framework of the V region.

Idiotypes. Idiotypes are antigenic determinants that are located on V domains. Several kinds of idiotypes are further defined by the to- pological assignment of the idiotype (restriction to VL or V,; dependence on both VL and V,; relation to the binding site) or the distribution of the idiotype on other myeloma proteins or immunoglobulins (found on only a unique species of homogeneous Ig-i.e., individual antigenic specificity or IdI; cross-specific determinant found in more than one myeloma protein composed of the same VL and VH isotypes, such as the IdX determinants of the GPlF- or PGGalBMP; or a cross- specific determinant shared by myeloma proteins or immunoglobulins with different isotypes, such as the T15(M511) idiotype or the U10- M173 idiotype (Bosma et al., 1977). The basic conclusion from the present review concerning idiotypes is that a single myeloma protein may contain more than one idiotype, as might be expected; thus it will be important in the future to define idiotypes as precisely as possible so that workers in different laboratories can have antisera of the same specificity when attempting to duplicate experimental procedures.

ACKNOWLEDGMENTS The author gratefully acknowledges the help, discussions, and contributions to this

manuscript of Drs. Stuart Rudikoff, NCI; Eduardo Padlan and C. P. J. Glaudemans, NIAMD. I also thank Drs. Latham Claflin, University of Michigan, Raymond Dwek, Oxford University, Daniel Hansburg, Washington University, David McKean, Mayo Clinic, and Dirk Streetkerk, NIAMD, for reading and correcting parts of the manuscript. I thank Ms. Linda Brunson and Mrs. Rosalyn Joftes for their patience and help in preparing this manuscript.

REFERENCES Appella, E. (1971). Proc. Natl. Acad. Sci. U . S A . 68, 590. Appella, E., and Inman, J. K. (1973). Contemp. Top. Mol. Immunol. 2, 51. Askonas, B. A., Williamson, A. R., and Wright, B. E. G. (1970). Proc. Natl. Acad. Sci.

Aspinall, G. 0. ( 1970). “Polysaccharides.” Pergamon, Oxford. Barstad, P., Farnsworth, V., Weigert, M., Cohn, M., and Hood, L. (1974a). Proc. Natl.

Barstad, P., Rudikoff, S., Potter, M., Cohn, M., Koningsberg, W., and Hood, L. (1974b).

U S A . 67, 1398.

Acad. Sci. U S A . 71,4096.

Science 183,962.

206 MICHAEL POTTER

Bennett, L. G., and Bishop, C. T. (1977). Can.J. Biochem. 55,8. Berek, C., Taylor, B. A,, and Eichmann, K. (1976).J. Erp. Med. 144, 1164. Blomberg, B., Geckler, W. R., and Weigert, M. (1972). Science 177, 178. Blomberg, B., Carson, D., and Weigert, M. (1973). Specijc Recept. Antibodies, Antigens

Bosma, M., Dewitt, C., Hausman, S. J., Marks, R., and Potter, M. (1977).]. Erp. Med. (in

Bourgois, A., Fougereau, M., and dePreval, C. (1972). Eur. J . Biochem. 21,446. Braun, D. G., and Jaton, J. C. (1974). Curr. Top. Microbiol. Immunol. 66,32. Bridges, S . H., and Little, J. R. (1971). Biochemistry 13,2525. Brown, A. R., and Crandall, C. A. (1976)J. Immunol. 116, 1105. Brundish, D. E., and Baddiley, J. (1967). Bi0chem.J. 105,306. Capra, J . D., Berek, C., and Eichmann, K. (1976).J. Immunol. 117,7. Carson, D., and Weigert, M. (1973). Proc. Natl. Acad. Sci. U S A . 70,235. Cebra, J. J., Gearhart, P. J., Kamat, R., Robertson, S. M., and Tseng, J. (1977). Cold

Cesari, I. M., and Weigert, M. (1973). Proc. Natl. Acad. Sci. U.S.A. 70,2112. Chesebro, B., and Metzger, H. (1972). Biochemistry 11, 766. Cisar, J., Kabat, E. A., Liao, J., and Potter, M. (1974).J. Erp. Med. 139, 159. Cisar, J., Kabat, E. A., Dorner, M. M., and Liao, J. (1975).J. E x p . Med. 142,435. Claflin, J. L. (1976a). Eur. J . Immunol. 6 , 666. Claflin, J. L. (1976b). Eur. J . Immunol. 6,669. Claflin, J. L., and Davie, J. M. (1974).J. Erp. Med. 140, 673. Claflin, J. L., Lieberman, R., and Davie, J. M. (1974). J . Immunol. 122, 1747. Claflin, J. L., Rudikoff, S., Potter, M., and Davie, J. M. (1975).J. E r p . Med. 141, 608. Cohn, M. (1967). Cold Spring Harbor Symp. Quant. Biol. 32,211. Cohn, M., Notani, G., and Rice, S. A. (1969). Immunochemistry 6, 111. Cosenza, H., and Kohler, H. (1972). Science 176, 1027. Cosenza, H., Quintans, J., and Lefiovits, I. (1973)J. Immunol. 5, 343. Cosenza, H., Julius, M. H., and Augustin, A. A. (1977). Transplant. Rev. 34, 3. Crandall, C. A., and Crandall, R. B. (1969).J. Parasitol 5!5, 1108. Crandall, C. A., and Crandall, R. B. (1971). E x p . Parasitol 30,426. Davies, D. R., Padlan, E. A., and Segal, D. M. (1975a).Annu. Rev. Biochem. 44,639. Davies, D. R., Pedlan, E. A., and Segal, D. M. (1975b). Contemp. Top. Mol. Immunol. 4,

Dourmashkin, R. R., Vivell, G., and Parkhouse, R. M. (1971).J. Mol. Biol. 56,207. Dower, S. K., Wain-Hobson, S., Gettins, P., Givol, D., Jackson, R., Perkins, S. J., Sunder-

Dugan, E. S., Bradshaw, R. A., Simms, E. S., and Eisen, H. N. (1973). Biochemistry 12,

Dwek, R. A. (1976). Contemp. Top. Mol. Immunol. 6, 1. Dwek, R. A., Jones, R., Marsh, D., McLaughlin, A. C., Press, E. M., Price, N. C., and

Dwek, R.A., Knott, J. C. A., Marsh, D., McLaughlin,A. C., Press, E. M., Price,N. C.,and

Dwek, R. A., Wain-Hobson, S., Dower, S., Gettins, P., Sutton, B., Perkins, S. J., and

Edmundson, A. B., Ely, K. R., Girling, R. L., Abda, E. E., Schiffer, M., Westholm, F. A.,

Eichmann, K. (1972). Eur.J. lmmunol2,301. Eichmann, K. (1975). Immunogenetics 2,491.

Cells Int. Convocation Immunol. [Proc.], 3rd, 1972 p. 285.

press).

Spring Harbor Symp. Quant. Biol. 41,201.

127.

land, C. A., Sutton, B., Wright, C., and Dwek, R. A. (1977). Biochem. J . 165, 207.

5400.

White, A. I. (1975a). Philos. Trans. R. SOC. London, Ser. B 272,53.

White, A. I. (1975b). Eur. J . Biochem. 53,25.

Givol, D. (1977). Nature (London) 266,31.

Fausch, M. D., and Deutsch, H. F. (1974). Biochemistry 13,3816.


Eichmann, K., Uhlenbruck, G., and Baldo, B. A. (1976). Zmmunochemistry 1 3 , l . Eisen, H. N. (1971). Prog. Zmmunol., Congr. Zmmunol., Zst, 1971 p. 244. Eisen, H. N., Little, J . R., Osterland, K., and Simms, E. S. (1967). Cold Spring Harbor

Eisen, H. N,, Simms, E. S., and Potter, M. (1968). Biochemistry 7,4126. Eisen, H. N., Michaelides, M. C., Underdown, B. J., Schulenberg, E. P., and Simms,

Epp, O., Colman, P., Fehlhammer, H., Bode, W., Schiffer, M., Huber, R., and Palm, W.

Francis, S. H., Graham, R., Leslie, Q., Hood, L., and Eisen, H. N. (1974). Proc. Natl.

Freed, R. M., Rockey, J. H., and Davis, R. C. (1976). Zmmunochemistry 13,509. Gearhart, P. J., Sigal, N. H., and Klinman, N. R. (1975).J. E x p . Med. 141, 56. Givol, D., Strausbauch, P. H., Hunvitz, E., Wilchek, M., Haimovich, J., and Eisen, H. N.

Glaudemans, C. P. J. (1975).Ado. Carbohydr. Chem. 31,313. Glaudemans, C. P. J., Jolley, M. E., and Potter, M. (1973). Carbohydr. Res. 30,409. Glaudemans, C. P. J., Zissis, E., and Jolley, M. E . (1975). Carbohydr. Res. 40, 129. Glaudemans, C. P. J., Manjula, B. N., Bennett, L. G., and Bishop, C. T. (1977). Zm-

Goetzl, E. J., and Metzger, H. (1970). Biochemistry 9,3862. Gotschlich, E. C., and Liu, T.-Y. (1967).]. Biol. Chem. 242,463. Granato, D., Braun, D. G., and Vassalli, P. (1974).]. Zmmunol. 113,417. Gray, W. R., Dreyer, W. J., and Hood, L. (1967). Science 155,465. Green, N. M., Dourmashkin, R. R., and Parkhouse, R. M. E. (1971).J. Mol. Biol. 56,203. Grey, H. M., Hirst, J. W., and Cohn, M. (1971).J. Erp. Med. 133,289.

Symp. Quant. Biol. 32, 75.

E. S. (1970). Fed. Proc., Fed. Am. Soc. E x p . Biol. 29, 78.

(1974). Eur. J . Biochem. 45, 513.

Acad. Sci. U.S.A. 71, 1123.

(1971). Biochemistry 10,3461.

munochemistry (in press).

Grossberg, A. L., Krasz, L. M., Rendina, L., and Pressman, D. (1974).J. Zmmunol. 113, 1807.

Haber, E., (1971). Ann. N. Y. Acad. Sci. 190, 283. Haimovich, J., Eisen, H. N., and Givol, D. (1970). Proc. Natl. Acad. Sci. U S A . 67,1656. Haimovich, J., Eisen, H. N., Hunvitz, E., and Givol, D. (1972).Biochemistry 11,2389. Hansbnrg, D., Briles, D. E., and Davie, J. M. (1976).J. Zmmunol. 117, 569. Hansburg, D., Briles, D. E., and Davie, J. M. (1977).J. Zmmunol. (in press). Haselkorn, D., Friedman, S., Givol, D., and Pecht, I . (1974). Biochemistry 13,2210. Hew, C.-L., Lifter, J., Yoshioka, M., Richards, F. F., and Koningsberg, W. H. (1973).

Hood, L. E., Potter, M., and McKean, D. J. (1970). Science 170, 1207. Hood, L. E., McKean, D., Fansworth, V., and Potter, M. (1973). Biochemistry 12, 741. Hood, L., Loh, E., Hubert, J., Barstad, P., Eaton, B., Early, P., Fuhrman, J., Johnson, N.,

Kronenberg, M., and Schilling, J. (1976). Cold Spring Harbor Symp. Quant. Biol. 41, 817.

Biochemistry 12,4685.

Inbar, D., Rotman, M., and Givol, D. (1971).J. Biol. Chem. 246,6272. Inbar, D., Hochman, J., and Givol, D. (1972). Proc. Natl. Acad. Sci. U S A . 69,2659. Inbar, D., Givol, D., and Hochman, J. (1973). Biochemistry 12,4541. Jack, R . S. , Imanishi-Kari, T., and Rajewsky, K. (1977). Eur. J . Immunol. 7, 559. Jackson, P., and Richards, F. F. (1973).J. Immunol. 112,96. Jaffe, B. M., Eisen, H. N., Simms, E. S., and Potter, M. (1969).J. Zmmunol. 103,872. Jaffe, B. M., Simms, E. S., and Eisen, H. N. (1971). Biochemistry 10, 1693. Jeanes, A. (1968). Encycl. Polym. Sci. Technol. 8,693. Johnston, M. F. M., and Eisen, H. N. (1974). Biochemistry 13,5547. Johnston, M. F. M., Barisas, B. G., and Shirtevant, J. M. (1974). Biochemistry 13,390.

208 MICHAEL POTTER

Jolley, M. E., and Glaudemans, C. P. J. (1974). Carbohydr. Res. 33,377. Jolley, M. E., Rudikoff, S., Potter, M., and Glaudemans, C. P. J . (1973). Biochemistry 12,

Jolley, M. E., Glaudemans, C. P. J., Rudikoff, S.,and Potter, M. (1974). Biochemistry 13,

Kabat, E. A., and Wu, T. T. (1971). Ann. N.Y. Acad. Sci. 190,382. Kabat, E. A., Wu, T. T., and Bilofsky, H. (1976a). Proc. Natl. Acad. Sci. U.S.A. 73,617. Kabat, E. A., Wu, T. T., and Bilofsky, H. (1976b). “Variable Regions of Immunoglobulin

Chains: Tabulations and Analyses of Amino Acid Sequences.” Medical Computer Systems, Bolt, Beranak, and Newman. Cambridge, Mass.

3039.

3179.

Klinilran, N. R., and Press, J. L. (1975).]. E r p . Med. 141, 1133. Kohler, G. (1975). Transplant. Reu. 27,24. Kohler, G., and Milstein, C. (1975). Nature (London) 256,495. Kohler, G., and Milstein, C. (1976). Eur. J. Immunol. 6, 511. Krause, R. M. (1970). Fed. Proc., Fed. Am. Soc. E x p . B i d . 29,59. Krause. R. M. (1971). Ado. Immunol. 1 2 , l . Lancet, D., and Pecht, I. (1976). Proc. Natl. Acad. Sci. USA. 73,3549. Lee, W., Cosenza, H., and Kohler, H. (1974). Nature (London) 2 4 7 , s . Leon, M. A., and Young, N. M. (1971). Biochemistry 10, 1424. Leon, M. A., Young, N. M., and McIntire, K. R. (1970). Biochemistry 9, 1023. Lieberman, R., and Potter, M. (1976). lmmunogenetics (abstr.) 4,417. Lieberman, R.,Potter, M., Mushinski, E. B., Humphrey, W., Jr., and Rudikoff, S . (1974)J.

Lieberman, R., Potter, M., Humphrey, W., Jr., Mushinski, E. B., and Vrana, M. (1975).].

Lieberman, R., Potter, M., Humphrey, W., Jr., and Chien, C. C. (1976).J. Immunol. 117,

Lifter, J., Hew, C.-L., Yoshioka, M., Richards, F. F., and Konigsberg, W. H. (1974).

Luderitz, O., Ruschmann, E., Westphal, O., Raff, R., and Wheat, R. (1967).]. Bacteriol.

Luderitz, O., Gmeiner, J., Kickhofen, B., Mayer, H., Westphal, O., and Wheat, R. W.

Lundblad, A., Steller, R.,, Kabat,’E. A., Hirst, J. W., Weigert, M. G., and Cohn, M.

McIntire, K. R., and Rouse, A. M. (1970). Fed. Proc., Fed. Am. Soc. E r p . Biol. 29,704

McKean, D., Potter, M., and Hood, L. (1973a). Biochemistry 12,749. McKean, D., Potter, M., and Hood, L. (1973b). Biuchemistry 12,760. Manjula, B. N., and Glaudemans, C. P. J. (1976). lrnmunochemistry 13,469. Manjula, B. N., Glaudemans, C. P. J., Mushinski, E. B., and Potter, M. (1975). Car-

Manjula, B. N., Glaudemans, C. P. J., Mushinski, E. B., and Potter, M. (1976a). Proc.

Manjula, B. N., Richards, F. F., and Rosenstein, R. W. (197613). Immumchemistry 13,

Margulies, D. H., Kuehl, W. M., and Schad, M. D. (1976). Cell 8,405. Martin, N., Warner, N. L., Roeder, P. E., and Singer, S. J. (1972). Biochemistry 11,4999. Melchers, F. (1969). FEBS Symp. 15, 169. Metzger, H., and Potter, M. (1968). Science 162, 1398. Metzger, H., Chesebro, B., Hadler, N. M.. Lee, J., and Otchin, N. (1971). Prog. Im-

E x p . Med. 139,983.

E x p . Med. 142, 106.

2105.

Biochemistry 13,3567.

93, 1681.

(1968). J . Bacteriol. 95,490.

(1972). lmmunochemistry 9,535.

(abstr.).

bohydr. Res. 40, 137.

Natl. Acad. Sci. U.S.A. 73,432.

929.

munol., Congr. Immunol., l s t , 1971 p. 253.


Michaelides, M. C., and Eisen, H. N. (1974)./. E x p . Med. 140, 687. Milstein, C., Adehigbo, K., Cowan, N . J., and Secher, D. S. (1974). Proc. Immunol. 2,

Montgomery, P. C., and Pincus, J. H. (1973).]. lmmunol. 111,42. Morse, H. C., 111, Goode, H. J., Jr., and Rudikoff, S. (1977)./. lmmunol. 119, <361. Mushinski, E. B., and Potter, M. (1977)./. lmmunol. In press. North, J. R., and Askonas, B. A. (1974). Eur. /. Immunul. 4,361. Olander, J.. and Little J. R. (1975). lmmunochemistry 12,383. Orin, G. B., Davis, R. C., Freed, R. M., and Rockey, J. H. (1976). lmmunochemistry 13,

Padlan, E. A., and Davies, D. R. (1975). Proc. Natl. Acad. Sci. U S A . 72,819. Padlan, E. A., Segal, D. M., Spande, T. F., Davies, D. R., Rudikoff, S., and Potter, M.

Padlan, E. A., Davies, D. R., Rudikoff, S., and Potter, M. (1976a). lmmunochemistry 13,

Padlan, E. A., Davies, D. R., Pecht, I., Givo, D., and Wright, C. (1976b). Cold Spring

Parker, C . W., and Osterland, C. K. (1970). Biochemistry 9, 1074. Pecht, I., Givol, D., and Sela, M. (1972). J. Mol. Biol. 68, 241. PBry, P., Petit, A., Poulain, J., and L d a u , G. (1974). Eur. J. lmmunol. 4, 637. Pink, J . R . L., and Askonas, B. A. (1974). Eur.1. lmmunol. 4,426. Poljak, R. J., Amzel, L. M., Chen, B. L., Phizackerley, L. P., and Saul, F. (1974). Proc.

Pollet, R., and Edelhoch, H. (1973)./. Biol. Chem. 248,5443. Pollet, R., Edelhoch, H., Rudikoff, S., and Potter, M. (1974)./. Biol. Chem. 249,5188. Potter, M. (1970). Fed. Proc., Fed. Am. SOC. E r p . Biol. 29, 85. Potter, M. (1971). Ann. N.Y. Acad. Sci. 190,306. Potter, M . (1975). In “Cancer: A Comprehensive Treatise” (F. F. Becker, ed.), Vol. 1,

Potter, M., and Cancro, M. (1977). In 30th Ann. Symp. Fundumentul Cancer Res. Univ.

Potter, M., and Leon, M. A. (1968). Science 162,369. Potter, M . , and Lieberman, R. I. (1970).J. Erp. Med. 132,737. Potter, M., Mushinski, E. B., and Glaudemans, C. P. J. (1972)./. Zmmunol. 108,2. Potter, M., Pumphrey, J. G., and Bailey, D. W. (1975). /. Natl. Cancer lnst. 54,

Potter, M., Padlan, E. A., and Rudikoff, S. (1976a)./. lmmunol. 117,626. Potter, M., Rudikoff, S., Vrana, M., Rao, D. N., and Mushinski, E. B. (1976b). Cold

Spring Harbor Symp. Quunt. Biol. 41,661. Potter, M., Rudikoff, S., Padlan, E. A,, and Vrana, M. (1977). In “Antibodies in Human

Diagnosis and Therapy” (E. Haber and R. M. Klause, eds.). Raven, New York. Richards, F. F., Lifter, J. L., Hew, C.-L., Yoshioka, M., and Koningsberg, W. H. (1974).

Biochemistry 13,3572. Richards, F. F., Konigsberg, W. H., Rosenstein, R. W., and Varga, J. M. (1975). Science

187, 130. Riesen, W. F., and Jaton, J.-C. (1976). Biochemistry 15,3829. Riesen, W., Rudikoff, S., Oriol, R., and Potter, M. (1975). Biochemistry 14, 1052. Riesen, W. F., Braun, D. G., and Jaton, J.C. (1976). Proc. Natl. Acad. Sci. USA. 73,

Robinson, E. A,, Smith, D. F., and Appella, E. A. (1974).]. Biol. Chem. 249,6605. Rocca-Serra, J., Milili, M., and Fougereau, M. (1975). Eur. 1. Biochem. 59, 511. Rockey, J. H., and Freed, R. M. (1976). Scand. /. Zmmunol. 5,655.

157.

517.

(1973). Nature (London), New Biol 245, 165.

945.

Harbor Symp. Quant. Biol. 41,627.

Natl. Acad. Sci. U.S.A. 71,3440.

p. 161. Plenum, New York.

Houston, Texas (in press).

1413.

2096.

210 MICHAEL POTTER

Rockey, J. M., Dorington, K. J., and Montgomery, P. C. (1971). Nature (London) 232, 192.

Rosenstein, R. W., and Richards, F. F. (1972).J. Zmmunol. 108, 1467. Rosenstein, R. W., and Richards, F. F. (1976). lmmunochemistry 13,939. Rovis, L., Kabat, E. A., and Potter, M. (1972). Carbohydr. Res. 23,223. Rudikoff, S., and Claflin, J. L. (1976).]. Exp. Med. 144, 1294. Rudikoff, S., and Potter, M. (1974). Biochemistry 13,4033. Rudikoff, S., and Potter, M. (1976). Proc. Natl. Acad. Sci. U.S.A. 73,2109. Rudikoff, S., Potter, M., Segal, D. M., Padlan, E. A., and Davies, D. R. (1972).Proc. Natl.

Rudikoff, S., Mushinski, E. B., Potter, M., Glaudemans, C. P. J., and Jolley, M. E. (1973).

Sakato, N., and Eisen, H. N. (1975)~ . E x p . Med 141, 1411. Sakato, N., Janeway, C. A., Jr., and Eisen, H. N. (1976). Cold Spring Harbor Symp.

Schechter, I., Ziv, E., and Licht, A. (1976). Biochemistry 15,2785. Schiff, C., and Fougereau, M. (1975). Eur. J . Biochem. 59,525 Schiffer, M., Girling, R. L., Ely, K. R., and Edmundson, A. B. (1973). Biochemistry 12,

Schulenburg, E. P., Simms, E. S.,Lynch, R. G., Bradshaw, R. A.,andEisen, H. N. (1971).

Segal, D. M., Padlan, E. A,, Cohen, G. H., Rudikoff, S., Potter, M. and Davies, D. R.

Sharon, J., and Givol, D. (1976). Biochemistry 15, 1591. Sher, A., and Cohn, M. (1972). Eur. 1. Zmmunol. 2,319. Sher, A., and Tarikas, H. (1971).J. Zmmunol. 106,1227. Sher, A., Lord, E., and Cohn, M. (1971).]. Zmmunol. 107, 1226. Shubert, D., Jobe, A,, and Cohn, M. (1968). Nature (London) 220, 882. Shubert, D., Roman, A., and Cohn, M. (1970). Nature (London) 225, 154. Sigal, N. H., Gearhart, P. J., and Klinman, N. R. (1975). J . Zmmunol. 68, 1354. Smith, A. M., and Potter, M. (1975).]. Zmmunol. 114, 1847. Smith, A. M., Slack, J., and Potter, M. (1977). Eur. J . Immunol. 7, 497. Smith, G. P. (1973). Science 181,941. Streefkerk, D., and Glaudemans, C. P. J. (1977). Biochemistry 16,3760. Sutton, B., Gettins, P., Givol, D., Marsh, D., Wain-Hobson, S., William, K. J., and Dwek,

Svasti, J., and Milstein, C. (1972). Biochem. J . 128,427. Tomasz, A. (1967). Science 157,694. Tsumita, T., and Ohashi, M. (1964).J. Erp. Med. 119, 1017. Underdown, B. J., and Eisen, H. N. (1971). /. Zmmunol. 106, 1431. Underdown, B. J., Simms, E. S., and Eisen, H. N. (1971). Biochemistry 10,4359. Vicari, G., Sher, A., Cohn, M., and Kabat, E. A. (1970). lmmunochemistry 7,829. Vrana, M., TomAsic, J., and Glaudemans, C. P. J. (1976).J. Immunol. 116, 1662. Vrana, M., Rudikoff, S., and Potter, M. (1977). Biochemistry (in press). Wain-Hobson, S., Dower, S. K., Gettins, P., Givol, D., McLaughlin, A. C., Pecht, I.,

Warner, N. L. (1975). lmmunogenetics 2, 1. Warner, N. L., and Ovary, Z. (1970)./. Zmmunol. 105,812. Watson, M. J., and Baddiley, J. (1974). Biochem. J . 137,399. Weigert, M., and Potter, M. (1977). lmmunogenetics 4,401. Weigert, M., and Riblet, R. (1976). Cold Spring Harbor Symp. Quant. Biol. 41, 837.

Acad. Sci. U.S.A. 69, 3689,

J . Exp. Med. 138, 1095.

Quant. Biol. 41, 719.

4620.

Proc. Natl. Acad. Sci. U.S.A. 68, 2623.

(1974). Proc. Natl. Acad. Sci. U.S.A. 71,4298.

R. A. (1977). Biochem.J. 165, 177.

Sunderland, C. A., and Dwek, R. A. (1977). Biochem. J . 165,227.

ANTIGEN-BINDING MYELOMA PROTEINS OF MICE 21 1

Weigert, M., Cesari, I. M., Yonkovich, S. J., and Cohn, M . (1970). Nature 228,2045. Weigert, M., Raschke, W. C., Carson, D., and Cohn, M. (1974).J. Exp. Med. 139, 13. Weigert, M., Potter, M., and Sachs, D. (1975). Zmmunogenetics 1,511. Weinstein, Y., Givol, D., and Strausbausch, P. H. (1972). Eur. J . Zmmunol. 2, 186. Willan, K., Marsh, D., Sunderland, C. A., Sutton, B. J., Wain-Hobson, S., Dwek, R. A.,

Wu, T. T., and Kabat, E. A. (1970).J. E x p . Med. 132, 211. Yoshioka, M., Lifter, J., Hew, C.-L., Converse, C. A., Armstrong, M. Y. K., Konigsberg,

Young, N. M., Jocius, I. B., and Leon, M. A. (1971). Biochemistry 10, 1971.

and Givol, D. (1977). Biochem.J. 165, 199.

W. H., and Richards, F. F. (1973). Biochemistry 12,4679.

Human Lymphocyte Su bpopu lations'

1. CHESS' AND S. F. SCHLOSSMAN

Division of Tumor Immnology, Sidney Faher Concur Institute, Hamrrd M c a l school, Barton. MasnrdruseUs

1. Introduction ........................................................... 11. Classical Cell Surface Determinants on Human Lymphocytes .............

Blood T-cell Subclasses (THI) ...........................................

V. Purification of Lymphocyte Subclasses ................................... A. Affinity Chromatography on Columns ................................ B. Fluorescence-Activated Cell Sorting . . . . . . . . . . . C. Rosette Depletion Techniques ....................................... D. Other Separation Techniques . . . . . . . .............................

VI. The Functional Analysis of Isolated Human Lymphocyte Subpopulations . . A. General Considerations ..............................................

111. Antigens Distinguishing Human Thymocytes (HTL) and Peripheral

IV. Human B-Cell Specific Antigens ..............................

B. Proliferation in Response to Soluble and Cellular Antigens . . . . . . . . . . . . C. Mediator Production by Subclasses of Human Lymphocytes ........... D. Cell-Mediated Lympholysis .................................. : . . . . . . E. Cell-Mediated Destruction of Syngeneic Tumor Cells ................. F. Analysis of Human Peripheral Blood B-Cell Function ................. G. Characterization and Isolation of Regulator Cells in Human

Peripheral Blood.. .................................................. H. The Functional Heterogeneity of Null Cells ............. References .............................................................

213 214

2 16 220 222 222 223 225 226 226 226 227 228 230 231 233

236 237 238

I. Introduction

The precise dissection of the cellular mechanisms and interactions involved in the generation of the human immune response has been faciliated by recent advances in three interrelated areas: (1) the development of in uitro methods for the characterization and identification of human lymphocyte classes by cell surface markers; (2) the development of new techniques for the isolation of highly purified subclasses of human lymphocytes and monocytes; and (3) the development of in uitro techniques to discriminate the functional properties and interactions of the isolated subsets of lymphocytes. These advances have occurred during the last few years despite the obvious limitations inherent in the study of the human immune re-

' This work was supported in part by A1 12069, CA-19589, N01-CB-43964, and N01-

Present address: Dept. of Medicine, Columbia University College of Physicians and CB-53881 from the National Institutes of Health.

Surgeons, New York, New York.

2 13

2 14 L. CHESS AND S. F. SCHLOSSMAN

sponse. In particular, one limitation has been the lack of genetically defined strains that have been so important in both the generation of alloantisera to unique subclasses of cells and in the genetic analysis of T, B, and macrophage cell interactions. In addition, in uivo studies of human immune finction are necessarily limited and have relied for the most part upon the direct in uitro and in uiuo study ofpatients with

experiments of nature” that arise in the congenitally immunodefi- cient patient or in patients with autoimmunity. Even these studies have been limited both by the rarity and heterogeneity of the individual disorders. In the present review we focus our attention on some approaches to overcoming difficulties inherent in the investigation of the human system with particular attention to the three areas noted above. Emphasis is placed on studies undertaken in the author’s laboratory over the past 4 years, and no attempt is made to cite exhaus- tively many excellent studies carried out elsewhere using similar approaches.

“

I I . Classical Cell Surface Determinants on Human lymphocytes

During the early 1970s enormous excitement was generated in clinical immunology by the finding that human T and B cells could be readily distinguished by cell surface markers. These surface structures initially included intrinsically bound surface membrane Ig (Froland and Natvig, 1970; Grey et al., 1971; Siegal et al., 1971), the receptor for sheep erythrocytes (E receptor) (Brain et al., 1970; Coombs et al., 1970; Lay et al., 1971; Jondal et al., 1972), receptors for the complement components (Bianco et al., 1970), and the receptors for the Fc fragment of antibody molecules (Dickler and Kunkel, 1972; Basten et al., 1972). Many studies indicated that the subset of human lymphocytes forming rosettes with sheep erythrocytes (E+) were T cells (Fro- land, 1972; Jondal et al., 1972; Wybran et al., 1972). The evidence for this conclusion stemmed from studies demonstrating that E+ cells represent a population distinct from B lymphocytes as identified by membrane-bound Ig; patients with profound hypogammaglobu- linemia and B-cell deficiency have normal or increased numbers of E-rosetting cells; most human thymocytes are E rosette positive; and, few E-rosetting cells are found in patients with congenital thymic aplasia. In contrast, a reciprocal subset of human B cells was defined, which, like their counterparts in rodents, had intrinsically bound surface Ig and contained the receptor for the complement components C3d and C3b (see Moller, 1973). To support this view, it was shown that patients with infantile X-linked agammaglobulinemia lacked cells bearing surface Ig, whereas patients with congenital thymic ap-

HUMAN LYMPHOCYTE SUBPOPULATIONS 215

lasia had normal or higher percentages of Ig+ cells (Grey et al., 1971; Cooper et al., 1971; Geha et al., 1973). More recently, and as will be discussed in detail below, the evidence for the E-rosette marker as a distinguishing characteristic of human T cells and the evidence for the Ig and C receptor determinants for a distinguishing characteristic of B cells have been shown more directly in in vitro functional studies of isolated populations of T and B cells.

With respect to the Fc receptor, initial studies indicated that the predominant cells bearing receptors for IgG in the human peripheral blood were surface Ig-bearing B lymphocytes and monocytes. Recent studies now indicate that a significant percentage of human T cells have receptors for the Fc fragment of IgG, and an even larger percentage have receptors for the Fc fragment of IgM (Moretta et al., 1976, 1977). In addition, there exists a third population of cells in peripheral blood which, for want of a better term, have been referred to as a null-cell population (Jondal et al., 1973b; Greenberg et al., 1973; Chess et aZ., 1974~). This population of cells is E rosette negative and surface Ig negative, but heterogeneous with respect to both complement and Fc receptors. The subset of null cells bearing both the complement and Fc receptors has been isolated (see below) and shown to be the predominant lymphocyte effecting antibody- dependent cellular cytotoxicity (Perlmann et al., 1975; MacDermott et al., 1975; Brier et al., 1975). The functional properties of Fc-bearing cells, which are E rosette negative, Ig negative, and lack complement receptors, are currently under active study. Taken together, the evidence suggests that the Fc receptor, although perhaps useful in the discrimination of subclasses of T or B cells, is not particularly useful for the initial characterization of human lymphocytes into T or B subpopulations.

In addition to the four markers discussed above, a number of investigators have reported additional determinants that may distinguish human T and B cells. Thus, human T cells have been shown to form rosettes with rhesus monkey red cells (Lohrmann and Novikovs, 1974), human red cells (Kaplan and Clark, 1974), and with human B lymphoblasts (Jondal et al., 1975). In addition, they may contain receptors for measles-induced surface antigens (Valdimmarsson et al., 1975) and lectins such as the Helix pomatia ones (Hammarstrom et al., 1973). In contrast, human peripheral B cells have receptors for EB viral antigens (Jondal and Klein, 1973) and have been reported to have receptors for mouse red blood cells. These additional cell surface markers of human lymphocytes have been recently reviewed (Bloom and David, 1976). Since they have been less extensively analyzed then the four classic determinants, there is less unanimity on the proportion

2 16 L. CHESS AND S . F. SCHLOSSMAN

of cells bearing either of these determinants and, in fact, evidence still is sparse with respect to functional studies.

As suggested above, the ability to characterize distinct subpopulations with the conventional techniques has allowed for their quantita- tion in peripheral blood and in other lymphoid tissue in both normal individuals and patients. Moreover, these cell surface markers have permitted the development of methods that allow for the isolation of distinct receptor-bearing lymphocyte populations for functional studies. Despite the usefulness of conventional cell surface markers, it should be emphasized that they are still relatively crude and inexact. In particular, the E- and EAC-rosetting tests have been difficult to quantitate and the results vary because of a number of technical factors; these include serum source, contamination with red cells, incubation temperature, and either the method of contrifugation or the force used in resuspension of the rosettes. In addition, the analysis of surface Ig on human B cells is complicated by the fact that B cells and other cell populations bear receptors for the Fc portion of immunoglobulins. Therefore, it has been emphasized that techniques involving identification of surface immunoglobulins with fhoresceinated anti-Ig reagents require prior pepsin digestion (Winchester et al., 197513). The analysis of Fc receptors on lymphoid cells depends on rosetting techniques using IgG- and IgM-coated red cells, binding of labeled aggre- gated IgC or antigen-antibody complexes. These methods have also been fraught with technical difficulties that have not as yet been completely resolved (Arbeit et aZ., 1976). The finding now of Fc receptors on human T cells adds considerably to these difficulties.

Perhaps of greater biological importance, the techniques outlined above do not adequately deal with the extraordinary heterogeneity that is known to exist within the T, B, and null-cell populations. Dis- section of this heterogeneity is critical for studies directed at the hnc- tional properties of lymphocytes and for an understanding of the maturation and differentiation of human lymphocytes. For these reasons, considerable attention has been directed at defining more precise quantifiable surface determinants on human lymphocyte subclasses and relating these determinants to states of differentiation and functional properties of cells.

111. Antigens Distinguishing Human Thymocytes (HTL) and Peripheral Blood T-cell

Subclasses (THJ

In mice, the relationship between thymocytes and T-cell surface antigens with the distribution, life history, stages of differentiation, and, perhaps most interestingly, function of T-cell subclasses have

HUMAN LYMPHOCYTE SUBPOPULATIONS 2 17

been extensively analyzed (Shiku et al., 1975; Cantor and Boyse, 1975a,b; Cantor and Weissman, 1976). These phenotypic deterini- nants on the surface of thymus-derived cells for the most part are recognized by alloantisera prepared in mice bearing the appropriate genotypes (Shen et al., 1975). Whereas murine alloantisera to theta, TL, Lyl, Ly2, Ly3, and 1 region deteiminants have proved to be extremely valuable, comparable reagents for human cells are just be- coming available. For example, heteroantisera directed toward human T-cell determinants have been prepared using a variety of immunization schedules, absorption procedures, and methods for assay. The antigens used for preparing these antisera have included human thymocytes, soluble extracts of thymocytes, brain, leukemic lymphoblasts bearing E-rosette markers, peripheral lymphocytes from patients with X-linked agammaglobulinemia, and, recently, monkey thymocytes (Aiuti and Wigzell, 1973; Brown and Greaves, 1974; Greaves and Janossy, 1976; Balch et ul., 1976). Virtually all these heteroantisera require extensive absorption with combinations of cells including allogeneic cultured B-cell lymphoblastoid lines, B-cell chronic lymphatic leukemias, bone marrow cells, erythrocytes, and fetal cells to render them specific. Most of the heterologous antihuman T-cell reagents prepared to date do not discriminate between thymocytes and peripheral T cells. In addition, many of the anti-T-cell sera are specific only by complement-dependent lytic assays, but not when analyzed by immunofluorescence, especially with extremely sensitive fluorescence techniques. In addition, only a few ofthe resulting anti-T-cell reagents appear to discriminate functionally distinct T-cell subclasses (Woody et al., 1975; Brouet and Tobin, 1976).

In an attempt to overcome some of these difficulties, we have prepared antisera to highly purified T-cell and/or thymocyte populations in rabbits and absorbed the resulting antilymphocyte sera with autologous lymphoblastoid cell lines (Schlossman et al., 1976; Chess and Schlossman, 1977). We have found that the use of autologous B lymphoblastoid cell lines for absorption of heteroantisera, as well as the use of purified populations of T cells and thymocytes for immunization, are of considerable importance with respect to the development of specific antibodies with a high degree of specificity. These approaches have facilitated the removal of species, differentiation, MHC or fetal antigens to which heteroantisera are often directed. For example, several antisera have been prepared that identify unique cell surface determinants on human thymocytes, but are not detected on more mature normal human peripheral T cells or on B cells. These antisera (anti-HTL) are prepared by immunizing rabbits with E-rosette-positive, acute lymphoblastic leukemic cells and absorbing

218 L. CHESS AND S. F. SCHLOSSMAN

the antisera with autologous B lymphoblastoid cell lines derived from the same patient when in clinical remission. The HTL antisera react exclusively with thymocytes and with leukemic cells from those patients with the T-cell variety of acute lymphoblastic leukemia (ALL). Since the anti-HTL sera react only with thymocytes and T-cell ALLs, they are analogous to the geneticaIly defined anti-TL sera developed in the mouse. Of importance is the failure to detect HTL antigens on mature normal peripheral T cells, which suggests that these antisera define a state of differentiation within the T-cell pool. Interestingly, the sub- group of ALL patients whose blast cells bear the HTL antigen frequently present with high white counts and thymic masses, and do poorly clinically with respect to chemotherapy (Sallan et al., 1977). The HTL unreactive patients, on the other hand, which account for approximately 80% of ALLs in childhood, bear Ia-like determinants on their blast cell surfaces (see below). Of additional interest is the fact that some of these anti-HTL antisera prepared by immunizing rabbits with T-cell leukemic blasts can be absorbed with thymocytes and subsequently shown to react with leukemic blast cells, but not with either thymocytes or purified T cells (unpublished observations).

Although the HTL antiserum defines differentiation antigen(s) on thymocytes, its failure to react with peripheral T cells precludes its usefulness in distinguishing functionally unique subclasses of peripheral T cells. In order to develop antisera capable of distinguishing antigens on more differentiated T cells, we have used a similar approach to that outlined above. Highly purified T cells were isolated by anti-F(ab)* immunoabsorbent column chromatography, nylon passage, and rosette depletion of complement-bearing cells prior to immunization of rabbits. The resulting rabbit antisera was then absorbed with an autologous B-cell line (Evans et al., 1977). Again the use of the autologous B-cell line was shown to be more efficient for absorption of these antisera than were allogeneic B cell lines or chronic lymphocytic leukemia (CLL) cells. For example, antisera can be rendered T-cell specific by absorption with as few as 2 x 1W autologous B-lymphoblastoid cells. Their activity on peripheral T cells and thyrnocytes plateaus generally at 60 and 90%, respectively, and cannot be hrther absorbed by as many as 1 x lo9 B cells. At maximal concentrations of antisera, 5040% of purified T cells are lysed in the presence of complement. In contrast, 90-100% of thymocytes are lysed (Fig. 1). No further increase in T-cell lysis is obtained by a second treatment with antibody and complement, suggesting that a discrete subclass of T cells is recognized. The antigen(s) recognized on T cells is designated THI, and the absorbed heteroantiserum as anti-TH1. Simi-


lo 0 - - .,

Ill60 l/aO 1/40 1/20 1/10

SERUM MLUTW

FIG. 1. Complement-dependent lysis of lymphoid subpopulations using serial dilu- tions of anti-THI absorbed with 1 x 10” autologous lymphoblastoid cells. ., Fetal thymocytes; 0, peripheral T cells; 0, unfractionated lymphocytes; 0, peripheral B cells.

lar results were obtained with anti-T,, serum when analyzed by immunofluorescence using a sandwhich technique with fluoresceinated goat) antirabbit F(ab)* on the fluorescence-activated cell sorter. Of more importance, as will be shown below, the anti-TH1 serum iden- tifies a functionally distinct subclass of human T cells. Taken together, the HTL and TH1 antigens allow a preliminary view of the differentiation of human T cells from thymocytes (Fig. 2). Thus, thymocytes bear receptors for sheep erythrocytes, HTL, and TH1 antigens. With further differentiation, the sheep cell receptor is retained, whereas the HTL antigen is lost. In contrast, the TH1 antigen, which is detected on all thymocytes, is found in only 40-60% of peripheral T cells. Further

DIFFERENTIATION ANTIGENS ON HUMAN T CELLS

FIG. 2. DiEerentiation antigens on human thymus-derived lymphoid cells.

220 L. CHESS AND S . F. SCHLOSSMAN

support for this differentiation scheme was obtained by analysis of T leukemic cells. In the childhood T-cell leukemias, the lymphoblasts are usually E rosette positive, HTL positive, and T H , negative. In contrast, the T lymphoblasts found in a variant to chronic lymphatic leukemia are E rosette positive, HTL negative, but bear the T H 1 antigen. Thus, the morphologically more mature leukemic cells in CLL, like the more differentiated T cells in normal individuals appear to have lost the HTL antigen, while maintaining the T H 1 antigen. We have not as yet identified an E rosette positive, HTL negative, TH1 negative subclass in chronic lymphatic leukemia, but one would predict that such leukemic cells, exist.

IV. Human 8-Cell Specific Antigens

As indicated above, a considerable number of cell surface determinants have been used in defining and quantitating circulating human B cells. These include surface Ig, the complement receptor, the Fc receptor, EB viral receptors, and heteroantisera directed at B-lymphocyte antigens. Recently, a new group of polymorphic antigens has been characterized which are linked to the major histocompatibility locus and are expressed on B cells and monocytes, but not on T cells (van Rood et al., 1975; Jones et al., 1975; Mann et aZ., 1975; Winchester et al., 1975a). These groups of antigens are distinct from the gene products of the HLA-A, B, and C locus, which are expressed on essentially all types of cells with the exception of erythrocytes. The major histocompatibility complex (MHC) gene products expressed predominantly on B cells are closely linked to and include at least some of the HLA-D region determinants, which control the mixed lymphocyte reaction (MLR). It is of interest that these B cell determinants appear to be analogous to the murine Ia antigens even though they are found outside the HLA-A and HLA-B regions, whereas in the mouse, the genes controlling the expression of MLC and Ia antigenic determinants are mapped between the H2-D and H-2K loci.

The human Ia-like determinants were first detected using human alloantibody obtained from multiparous females which had been absorbed with platelets to remove antibody directed toward HLA-A, B, and C gene products. Furthermore, Ia-like molecules of 23,000 and 30,000 molecular weight (p23,30) have now been isolated and purified from human B lymphoblastoid cell membranes and used to generate rabbit anti-p23,30 antibody (Humphreys et aZ., 1976). The p23,30 antigen was shown to be a B lymphocyte-specific cell surface polypeptide complex and immunochemically different from HLA-A,


B, and C antigens, immunoglobulins, and p2 microglobulins. In addition, the antigen complex was shown to be easily identified on the surface of B lymphocytes and a subclass of null lymphocytes which bear the receptor for complement. The evidence that the p23,30 antigen complex is analogous to murine Ia antigens includes (a) chemical and structural similarities; (b) tissue distribution; (c) linkage to the MHC; and (d) biological function as judged b y effects of antisera to p23,30 in a variety of systems including the inhibition of the MLC reactivity and ADCC as well as its effects on the differentiation of B cells triggered by products of activated human T cells (Strominger et al., 1976; Humphreys et al., 1976; Chess et al., 1976; Friedman et aZ., 1977).

Similar Ia-like molecules have been isolated from other human B-cell lines. For example, an antigen complex very similar to the p23,30 antigen was isolated from the Yoder cell line, which is homozygous for HLA-2 and -7 (Fuks et d., 1977). The isolated p23,30 complex from these cells inhibits the cytotoxicity of B-cell alloantisera specific for Yoder cells while not inhibiting anti-HLA-2 or -7 sera. More recently, a similar Ia-like complex has been derived from the cultured human lymphoblastoid cell line BRI-8 (Snary et al., 1977). These antigens comprise two glycoproteins of 33,000 and 28,000 molecular weight and are therefore very similar to the p23,30 molecular complex derived from the IM-1 and Yoder cell lines. Anti- body to these determinants blocks the MLC reaction, and the isolated antigens inhibit specifically B-cell alloantisera.

Recently, we found b y complement-dependent lysis studies and by immunofluorescence that the p23,30 antigen is detected on a hnction- ally important subclass of human monocytes (Breard et al., 1977) and, perhaps more interestingly, appears on highly purified populations of T cells after in uitro sensitization to alloantigen (Evans et aZ., 1977). These latter results suggest that Ia-like determinants may be masked and expressed on distinct subsets of human T cells and detected only by functional studies directed at the inhibition of selected T-cell functions. Alternatively, the expression of Ia on T cells may result from the binding of Ia-like molecules to putative Ia receptors on the surface of activated T cells.

In Table I we have summarized the cell surface markers of human peripheral T, B, and null lymphocyte classes and thymocytes with respect to the conventional and more recently described cell surface determinants. These surface antigens provide the framework by which methods have been developed for the isolation and functional characterization of distinct human subclasses as described in the sections


TABLE I SURFACE PROPERTIES OF HWAN PEIUPHERAL

LYMPHOCYTE SUBSETS AND THYMOCYTES

Lympho- Surface determinants

subpopu- E EAC Ia-like cyte

lations rosette rosette SmIg HTL TI, (~23,301

" Null cells are heterogeneous with respect to EAC receptors and p23,30, the same

'ITH, is present on 50-60% of peripheral T cells and approximately 90% of

' T cells activated by alloantigen express the p,23,30 antigen.

2 0 3 0 % of cells reacting with each.

th ymocytes.

below. We would emphasize that, given the overlap between surface determinants on various subclasses of cells, it is important to analyze lymphocyte subpopulations not only by their surface determinants, but, in addition, their functional properties.

V. Purification of Lymphocyte Subclasses

A. AFFINITY CHROMATOGRAPHY ON COLUMNS Affinity chromatography has been especially useful for the separa-

tion of large numbers of cells and provides an initial fractionation procedure for subsequent isolation techniques and functional studies. Most of the techniques reported have relied on the use of solid-phase supports to which are bound, either by adherence or by covalent linkage, a variety of molecules that have affinity for cell surface receptors, antigens, or other determinants. The solid supports most commonly used have included cross-linked dextrans (Sephadex G-200) (Schlossman and Hudson, 1973; Chess et al., 1974a), glass and ac- rylamide plastic beads (Degalon) (Wigzell and Anderson, 1969; Jon- dal et al., 1972). A primary factor in the choice of the solid support is the nonspecific retention of the cells. For example, it is known that B cells and monocytes will nonspecifically adhere to glass or nylon. Moreover, certain B and T cells will nonspecifically absorb to Dega- lon. For this reason we have chosen in our laboratory to use Sephadex G-200, which is a near-perfect filter for human lymphocytes and

T B Null Th ymocyte


monocytes when specific antibodies are not covalently linked. During the last few years we have had experience with Sephadex G-200 covalently linked to antihuman F(ab)2 as a cellular immunoabsorbent for the primary isolation and separation of surface Ig+ and Ig- lymphocytes. As shown in earlier studies, human B cells bind to Sephadex anti-F(ab), as a consequence of either intrinsic or cell surface Ig, whereas non-Ig-bearing cells (T plus null cells) do not bind. In addition, unlike anti-F(ab)2 columns made on Degalon supports, Sephadex G-200 anti-F(ab)2 allows most Ig- Fc receptor-bearing cells to pass through. One of the other major advantages of these columns is that the bound surface Ig-bearing cells can be recovered by competitive elution with human Ig. Quantitatively both the passed and recovered cells account for greater than 90% of the starting population of lymphocytes. Cells passing directly through the anti-F(ab)2 columns contain less than 2% Ig-bearing lymphocytes, whereas those eluted from the column contain greater than 98% Ig+ cells and less than 2% E-rosetting cells. The passed population is heterogeneous with respect to E rosetting, since only 7040% of this population form E rosettes. Of importance is the fact that the separated populations remain functionally intact, can be recovered with minimal loss, and maintain their surface properties in vitro. In addition, the Sephadex anti-F(ab)2 column techniques can be modified to specifically isolate and deplete cells bearing unique cell surface antigens which are recognized by specific antisera (Chess et al., 1976). Thus, Ig- cells can first be coated with rabbit anti-p23,30 and then passed over a Sephadex G-200 goat antirabbit Ig column. Cells passing through such columns can be shown to be p23,30. By competitive elution with rabbit y-globulin, the bound cells can be recovered and shown to be p23,30+. Similarly, anti-TH1 coated Ig- cells can be passed over similar columns and both THI+ and TH1- subclasses recovered. This general approach can obviously be utilized for any heteroantisera that recognize distinct differentiation antigens or other determinants on lymphocyte surfaces.

B. FLUORESCENCE-ACTIVATED CELL SORTING

The reader is referred to publications by Hertzenberg and his colleagues (Bonner et d., 1972; Loken and Herzenberg, 1975) describing detailed methodology and analysis of cell separation capabilities of the Becton-Dickinson fluorescence-activated cell sorter (FACS). This instrument analyzes and separates cells on the basis of either cell surface fluorescence or size and can provide a histogram of the number of positive fluorescent cells obtained against the intensity of fluores-


cence. Individual subclasses of cells as defhed by either their scatter of fluorescence profile can be charged and specifically deflected and recovered for subsequent studies. Both direct and indirect fluorescent techniques can be used on the cell sorter. Utilizing the direct approach, cells are stained by resuspending approximately 2 to 3 x loe cells in 0.2 ml of fluoresceinated antibodies and incubated for 1 hour at 4°C. After 3 washes, the labeled cells can be processed on the FACS at approximately 500-1000 cells per second, and the intensity of fluorescence is recorded for each individual cell on the pulse height analyzer. Background fluorescence is determined by analyzing appropriate negative controls including cells labeled with fluoresceinated normal serum. Cell-sorter analysis of purified populations of cells isolated by immunoabsorbent chromatography has been particularly useful. For example, 98% of human B cells isolated by Sephadex anti-F(ab)z immunoabsorbent columns are surface Ig+ as detected on the cell sorter. In contrast, less than 1% of the isolated T-cell populations bound surface Ig. It was also shown that the null-cell population was heterogeneous with respect to the quantity of surface Ig and that a small subclass of cells had a relatively low degree of fluorescence as detected on the FACS. This small population of cells, when analyzed by either direct or indirect techniques utilizing the same polyvalent anti-F(ab), reagents with fluorescence microscopy was not detectable as surface Ig+. Perhaps of more importance, with fluoresceinated rabbit anti- p23,30 sera it was shown that greater than 90% of B cells isolated by immunoabsorbent columns were intensely stained whereas less than 2% of the T cells were reactive (Schlossman et al., 1976). A subset of cells within the null-cell population accounting for approximately 2040% of the population is p23,30+. These data were important with respect to the analysis of the purity of the populations isolated by column chromatography since the method used for detection of cells was distinct from the antibody used for primary isolation procedures.

In more recent studies, purified T-cell populations were analyzed using indirect fluorescence on the cell sorter with the rabbit anti-THl antibody (Evans et al., 1977). The developing reagent was a hores- ceinated pepsin digested goat antirabbit F(ab)z. Anti-T", binding to B cells was not significantly different from that obtained with normal rabbit sera. The distribution of fluorescence binding to T cells with anti-TH1 sera was Gaussian. Thus, greater than 90% of peripheral T cells reacted with anti-TH1 sera, contrasting with a restricted number (5040%) of cells lysed by the same antisera using antibody and complement. It was therefore important to isolate low density-staining cells from high density-staining cells to ascertain whether cells bind-


ing few antibody molecules correspond to TH1- subclasses as detected by complement-mediated lysis. Thus, both low-density (lowest 25%) and high density binding (highest 25%) T cells were collected and analyzed by complement-dependent lytic assays, and it was shown that the TH1- subclass, i.e., the low density-staining cells were not lysed by treatment with anti-TH1+ complement, whereas the high density-straining cells were lysed. As will be described below, these two subclasses were then analyzed for functional properties and shown to be distinct.

C. ROSETTE DEPLETION TECHNIQUES

The simple E- and EAC-rosetting techniques adapt themselves to widespread use as both analytic and cell separation tools (Mendes et al., 1973; Greaves and Brown, 1974; Wahl et d., 1974; MacDermott et aZ., 1975). The problems involved in utilizing these techniques for separation are similar to those encountered in analytical rosetting assays. In addition, the volume of lymphoid cells that can be handled is limited, since repeated rosetting is required before highly purified populations are obtained. With repeated rosettings, there are losses of cells, and the recovered cells may be contaminated with cells that influence the results obtained in functional assays. Further procedures directed at the removal of sheep red blood cells or EAC cells bound to lymphocytes in the rosetted population, such as lysis with ammonium chloride, may alter the functional capacity of the recovered lymphocytes. Nevertheless, rosetting techniques are particularly useful as a second step in the isolation of subclasses of cells. Thus, E- and EAC- rosette depletion techniques have been used to subfractionate the Ig- population obtained from anti-F(ab), columns or on the FACS into E rosette-positive and E rosette-negative subclasses. After E depletion of Ig- cells, the resulting population is predominantly Ig- and E rosette negative. This population, which has been termed “null cells” is still heterogeneous with respect to other cell surface markers, including the p23,30 antigen, the Fc receptor, and the complement receptor. Functional properties of the p23,30+ Ig- E rosette-negative null-cell subclass will be described below. In addition, EAC depletion of the Ig- population of cells can be used to great advantage. The population remaining at the interface of Ficoll-Hypaque under these conditions comprises a class of cells which is greater than 92% E rosette positive and represents a highly purified population of T cells. T cells isolated by these methods are Ig-, E rosette positive, and EAC rosette negative and contain less than 1% p23,30+ cells. Further, only 5040% of these cells are lysed by the anti-TH1 sera.


D. OTHER SEPARATION TECHNIQUES

Several other techniques are available for the separation of complex mixtures of viable cells into their component subsets. These techniques rely on biophysical differences that exist in various lymphocyte subclasses including plasma membrane electrical charge differences, density, and cell size. Thus, a number of methods have been developed utilizing cellular electrophoresis (Ambrose, 1965; Hayry et al., 1975), sedimentation over a variety of density gradients (Shortman et al., 1975; Raidt et al., 1968; Geha and Merler, 1974b), and velocity sedimentation (Miller, 1976) for separation of different cell populations. The reader is referred elsewhere for the details involved in these techniques. As a general rule, there remains considerable overlap in the functional and cell surface characteristics of cells isolated by these techniques, and it appears that their greatest usefulness will be in the definition of the heterogeneity of cells initially fractionated by other methods.

VI. The Functional Anolyrir of lroloted Humon Lymphocyte Subpopulotionr

A. GENERAL CONSIDERATIONS

The functional properties of isolated T, B, and null cells have been analyzed extensively with respect to a number of in vitro assays of cell-mediated immunity. The data from our laboratory are summarized in Table I1 and point out the functional heterogeneity of the cells isolated by the methods described above. It is clear that a number of the assays of cell-mediated immunity in man discriminate functionally unique subpopulations of cells. Unique properties of T cells include their capacity to proliferate in response to specific soluble antigens, to recognize foreign HLA-D determinants in mixed lymphocyte culture (MLC), to recognize HLA-A and B determinants as the effector cells in cell-mediated lympholysis (CML), and to secrete some lymphokines, such as lymphocyte mitogenic factor (LMF). The only distinct functional property of B lymphocytes is their capacity to secrete Ig. However, it should be noted that B cells, in addition, are the only cells that have receptors for Epstein-Barr (EB) virus and allow EB viral replication. The characteristics of null cells that distinguish them from B cells are their capacity to function as the effector cell in antibody-dependent cellular cytotoxicity, and, perhaps equally interesting, their capacity to differentiate into granulocyte-forming colonies and erythrocyte-forming colonies. The latter two functions reflect the heterogeneity of the null-cell population. In contrast to the assays that appear to distinguish T, B, or null cells, a number of assays of cell-mediated immunity (Table 11) are subserved by more than one


TABLE I1 1MMUNOLOGIC FUNCTIONS O F HUMAN LYMPHOCYTE SUBSETS in Vitro

Immune functions T B Null

Pro1 i ferative responses 1. Soluble antigen-triggered proliferation 2. Response to alloantigens in mixed lymphocyte

3. Stimulating capacity in MLC

1. Migration inhibitory factor (MIF) 2. Leukocyte inhibitory factor (LIF) 3. Lymphocyte mitogenic factor (LMF)

1. Cell-mediated lympholysis (CML) of allogeneic cells 2. Antibody-dependent cellular cytotoxicity (ADCC) 3. Mitogen-induced nonspecific cytotoxicity

1. Capacity for Ig synthesis in cell culture 2. Plaque-forming cells

1. Precursors of granulocyte- and erythrocyte-forming

2. Proliferative response to Epstein-Ban virus

culture MLC

Mediator production

Cytotoxic responses

Antibody production

Miscellaneous functions

cells, B cells, and T cells

+ +

+/-

+ + +

+ - +

++

+ + -

+ +

+

++

N T ' N T

- + +

+ N T

+

" NT. not tested.

subclass of cells. For example, migration inhibitory factor (MIF) is made by both T and B cells, and it appears that B cells quantitatively account for most of the MIF produced by unseparated populations. Whether small numbers of T cells are required for B-cell MIF production remains controversial. Another mediator of cell-mediated immunity, leukocyte inhibitory factor (LIF), also appears to be produced by both T and B subpopulations, whereas LMF is produced exclusively by T cells. With respect to null cells, it is important to point out that a subclass of these cells can differentiate into Ig-secreting cells in uitro. This subclass bears the Ia-like determinant, p23,30, and thus may represent a subpopulation of B cells at a different stage in differentiation. In a subsequent section of this review we will focus attention on studies pertaining to the functional properties of subpopulations of lymphocytes and describe further the heterogeneity that exists within isolated T, B, and null-cell populations.

B. PROLIFERATION IN RESPONSE TO SOLUBLE AND CELLULAR ANTIGENS

It is known that purified T cells but not B cells from individuals previously sensitized to relatively complex antigens [purified protein


derivative (PPD), mumps, tetanus toxoid, etc.] respond by proliferation and incorporation of tritiated thymidine to specific soluble antigens in vitro (Geha et al., 1973; Chess et al., 1974a,b). Using the ability to subfractionate T cells into TH1+ and THI- subclasses, evidence has been obtained that the population of cells proliferating to specific soluble antigens are THI-, while those proliferating in response to alloantigens in MLC are TH1+ (Evans et d., 1977; Chess and Schlossman, 1977).

For example, purified T cells from individuals sensitized to tetanus toxoid, PPD, and mumps antigen were treated in the presence of complement with media alone, normal rabbit serum, or anti-TH, sera. After treatment, T cells were washed extensively and cultured either in the presence or in the absence of soluble antigen and pulsed with ['Hlthymidine after 6 days. Although anti-THl sera lysed 60% of the cells, it had no significant effect on the antigen-induced proliferative response. More important, when identical cultures were cocultivated with mitomycin C-treated allogeneic cells in the MLC assay, the anti-TH1, but not the normal rabbit serum, eliminated the proliferative response. These data suggested that the THI+ cell, but not the THI- cell, contained the MLC responsive cells, whereas the TH1- cells were capable of responding to specific soluble antigens. To determine whether T H 1 + cells could also respond to soluble antigens, the T H 1 +

and THI- cells were isolated as described above on the FACS. The cells binding anti-THl were divided into weak-binding (lowest 25%) or strong-binding (highest 25%) cells. It was shown that the lowest- binding cells (THI- by complement-mediated lysis) developed an excellent proliferative response to soluble antigens but did not react appreciably to allogeneic cells. In contrast, the isolated T H I + subclass were MLC responsive but did not proliferate in response to soluble antigens. These data support the view that the TH1 differentiation antigen was dissecting two unique subclasses of T cells, each subserving different functions; one programmed to respond in MLC, the other triggered by specific soluble antigens.

c. MEDIATOR PRODUCTION BY SUBCLASSES OF HUMAN

The distinct cell surface determinants have permitted a reevaluation of mediator production by individual lymphocyte subclasses. Lym- phocytes from individuals exhibiting delayed hypersensitivity are triggered in vitro by antigen to produce factors (lymphokines) with distinct biological properties including MIF (David, 1966; Bloom and Bennett, 1966; Rocklin et aZ., 1970), LMF (Maini et al., 1969), and LIF

LYMPHOCYTES


(Rocklin, 1974). Since the production of these factors correlated with the delayed skin reaction, it was generally assumed that the antigen- induced stimulation of these factors was mediated by T cells. The development of cellular separation techniques has allowed for hrther dissection of the cell populations which interact and produce these factors. In studies using Ig- and Ig+ populations, it was shown that MIF was produced both by T and B cells (Rocklin et al., 1974). Further, it was shown that MIF produced by B cells was indistinguishable from that produced by T cells and that B cells, despite their failure to proliferate in response to antigen, produced quantitatively more MIF than T cells. In these earlier studies, the possibility remained that small numbers of contaminating T cells could be contributing to the production of MIF or are themselves responsible for its production. This question was of particular importance, since many B cell functions are T-dependent. Moreover, in the guinea pig, MIF production by B cells appears to be T-cell dependent (Wahl and Rosenstreich, 1976).

To further explore the cell populations involved in mediator production, cells were analyzed after treatment with anti-TH1 sera or p23,30 sera (Evans et al., 1977; Chess and Rocklin, 1977). T cells from a number of individuals previously shown to exhibit hypersensitivity to either streptokinase-streptodornase (SKSD), PPD, or Candida antigens were treated with either normal rabbit serum (NRS) or anti-TH1 in the presence of complement. The remaining T-cell population ( T H I - )

were tested for their capacity to elaborate MIF after specific triggering by antigen. Treatment of T cells with NRS and complement had no significant effect on MIF production. In contrast, anti-TH1 and complement abrogated the production of MIF, indicating that the MIF was produced by the T H I + cells. It should be emphasized that the same T H 1 + subset ofcells did not proliferate in response to soluble antigens. To investigate the possibility that the TH1 + cells were contaminating isolated B-cell populations, B cells were treated with either NRS or anti-TH1 and complement prior to triggering with antigen and assaying for MIF production. Neither treatment had any effect on B-cell production of MIF. These results demonstrate that the T-cell subclasses primarily responsible for MIF production play no demonstrable role in the production of MIF by B cells. However, these studies do not formally exclude the possibility that contaminating T H i - cells, which by themselves do not produce MIF, may influence the production of MIF by B cells.

In similar studies, treatment of unfractionated populations of lymphocytes with anti-TH1 and complement did not affect MIF secretion. This finding is in accord with previous studies showing that B cells


quantitatively produce more MIF than T cells. Moreover, treatment of unfractionated populations of cells with anti-p23,30 and complement significantly reduced the MIF production but had no effect on the production of MIF by purified T cells.

The data presented on MIF can be viewed with respect to clinical observations that lymphoid populations from some patients with immunodeficiency states contain cells that undergo antigen-induced proliferation but do not make MIF (i.e., patients with sarcoidosis and chronic mucocutaneous candidiasis) (Rocklin et al., 1971) whereas cells from other patients are capable of MIF production but do not proliferate (Levin et al., 1970; Kirpatrick et al., 1972; Whitcomb and Rocklin, 1973), i.e., patients with miliary tuberculosis, Wiskott-Aldrich syndrome, and candidiasis treated with transfer factor). We would suggest that patients whose cells are capable of being triggered by antigens to elaborate MIF but not to proliferate have functionally intact T H 1 + cells or B cells and an absence or functional impairment of the THl- subclass. Similarly, patients whose lymphocytes proliferate but do not secrete mediators may have intact TH1- cells and have impaired function at least with respect to mediator production in both TH1+ cells and B cells. In these patients it is entirely possible, and even likely, that other functions of THl+ cells and B cells may remain normal, since mediator-producing cells may still represent subsets of THl+ cells and B cells.

Lymphocyte mitogenic factor (LMF), unlike MIF, has been shown to be produced exclusively by T cells (Geha et aZ., 1973; Rocklin et aZ., 1974; Breard et al., 1977). To determine whether the same subclass of T cells produces LMF as makes MIF, populations of T cells were treated with anti-TH1 plus complement and cultured with antigen for 48 hours, and LMF production was measured. Lysis of THl+ bearing cells did not abolish the proliferative response of these cells to a variety of antigens; nevertheless, the supernatants from these cultures did not contain LMF activity (Evans et aZ., 1977). In contrast, treatment with normal rabbit serum plus complement had no effect. These results support the view that TH1+ cells produced both LM F and MIF.

D. CELL-MEDIATED LYMPHOLYSIS

The allograft reaction involves a complex sequence of cellular immune events that have been studied in detail using a variety ofin uitro model systems. These studies have demonstrated that one critical pathway in allograft rejection is the sensitization and subsequent differentiation of T cells to cytotoxic cells which recognize cell surface determinants genetically controlled by the major histocompatibility complex (Cerrottini and Brunner, 1974). A number of investigators


have demonstrated during in vitro sensitization in one-way mixed leukocyte cultures (MLC) that effector T cells are generated which can specifically kill target cells bearing alloantigens in common with the sensitizing cell (Hayry and Defendi, 1970; Solliday and Bach, 1970; Lightbody et al., 1971).

This phenomenon has been termed cell-mediated lympholysis (CML). Several aspects of the allogeneic (ML response are important for an understanding of the mechanism by which T cells kill either autologous or syngeneic tumors. The generation of cytotoxic cells requires a proliferative response which can be blocked by antimitotic agents (Bach et al., 1972; Cantor and Jandinski, 1974). The majority of proliferating cells appear to recognize HLA-D determinants on the stimulating cell. Evidence for the requirement of interacting T cells has been suggested from studies which have shown that HLA-D reactive cells do not by themselves mediate killing, but instead provide helper interactions with a differentiating cytotoxic cell (Eijsvoogel et al., 1973; Schendel et al., 1973). Moreover, the cytotoxic cell recognizes serologically defined determinants (HLA-A and HLA-B) that are genetically distinct from the HLA-D locus. In the mouse, CML effector cells and proliferative helper cells are T lymphocytes and represent a functionally distinct subclass of Ly23 and Lyl cells, respectively (Cantor and Boyse, 1975a). In man, only the THI+ subset of cells pro- liferates in response to alloantigens; anti-Ia sera, including anti-p23,30 and B cell alloantisera, are potent inhibitors of the MLC response (Wernet et al., 1975; Humphreys et al., 1976).

Analysis of the T cells that differentiate into killer cells was performed using anti-TH1 sera. Treatment of killer cell populations generated after sensitization with anti-TH1 antisera substantially, but not completely, reduces CML activity, suggesting that both MLC responsive and killer cells are T,, +. Further experiments will be needed to completely resolve this point.'

Alloactivated T cells are not only effector cells in CML, but develop the capacity to mediate antibody-dependent cellular cytotoxicity (ADCC). Lysis of alloactivated T cells with anti-p23,30 plus complement eliminated this function. Additional insights into the heterogeneity of T cells will be defined when other functions of activated T cells are investigated.

E. CELL-MEDIATED DESTRUCTION OF SYNCENEIC TUMOR CELLS

Recently, in murine systems the principles of CML have been applied to the study of mechanisms involved in the generation of cytotoxic cells with specificity directed toward syngeneic cells that


have been virally altered, chemically altered, or associated with tumors. Both the recognition and specificity of the effector phase of syngeneic killing appear to be dependent on either alteration of cell surface determinants controlled within the MHC or, alternatively, on antigens recognized in association with MHC determinants (Zinker- nagel and Doherty, 1975; Gardner et al., 1975; Shearer et al., 1975; Schrader and Edelman, 1976). These examples of murine cytotoxicity against altered H-2 identical tissue have proved to be useful in the study of both tumor immunity and autoimmunity. In recent studies some mechanisms involved in the destruction of human acute leukemic cells by MHC-identical lymphocytes were examined (Son- del et al., 1976). Activation of human cytotoxic T lymphocytes (CTL) directed against serologically defined (SD) HLA-A and B determinants requires an active collaborative response to Ia surface molecules. A similar collaborative response may be important in the generation of cytotoxic cells to weak antigens in uitro. Thus, the addition of an allogeneic LD stimulus might provide the required helper signal needed to induce the cell-mediated killing of MHC-identical leukemic cells. These studies demonstrated that human lymphocytes in the presence of a third-party signal could be sensitized in uitro to recognize and destroy fresh leukemic blasts obtained from MHC identical siblings while not mediating destruction of normal lymphocytes or normal PHA blasts from other healthy MHC identical siblings.

The antigens recognized on the leukemic cells should not be considered tumor-specific until further analysis. However, the MHC iden- tity between the patients and siblings studied in these experiments demonstrated that the cytotoxic recognition of leukemic blasts was not a consequence of recognition of genetically controlled foreign MHC antigens. In addition, previous studies have demonstrated that healthy MHC identical siblings cannot generate CML against presumptive minor loci (non-MHC antigens) on each other’s lymphocytes even when sensitized in uitro in the presence of an unrelated third party (Sondel and Bach, 1976). Thus, the antigens recognized on the leukemic blasts are not conventional histocompatibility antigens (either MHC- or minor locus-controlled) that would be recognized as foreign on PHA blasts derived from normal lymphocytes. Whether these antigens are expressed on embryonic or other differentiated tissues remains open for further investigation. This question may be approached experimentally by using recently developed antigen- specific CML blocking techniques (Sondel and Bach, 1976).

Certain aspects of these studies in relation to the cellular mechanism of in uitro sensitization and killing of MHC identical human


leukemic cells are important. As noted above, murine studies by several groups have demonstrated that chemically or virally altered syngeneic cells can sensitize lymphocytes to become specifically cytotoxic to these altered cells. The antigenic alteration has been shown to specifically involve products of the MHC. In uitro destruction of H-2 identical tumor cells can also be directed at altered H-2 antigens. Thus, if spontaneously arising human leukemias were induced by interaction with a putative antigen-altered oncogenic virus, one would have predicted that the CML response to human leukemias would parallel those reported in murine systems. However, it was found that human leukemic cells alone rarely induce detectable antileukemic CML. The generation of antileukemic CML was enhanced by the addition of MHC-unrelated stimulating cells. In the murine system chemically or virally altered syngeneic cells alone are sufficient to stimulate CML; whether the addition of an allogeneic trigger would augment the cytotoxic response has not yet been determined. The necessity for and role of the third-party stimulation in augmenting sensitization to human leukemic blasts remains to be clarified. Al- logeneic stimulating cells may provide the required proliferative helper cell stimulus analogous to that observed in the generation of CTLs to alloantigens. Alternatively, allogeneic lymphocytes may b e necessary to overcome the action of suppressor T cells.

In more recent studies of the killing of syngeneic leukemic blasts by siblings’ lymphocytes, we have demonstrated that the effector cell in this system is a T lymphocyte; studies are now under way to see whether the subclass of cells required for the generation of these killer cells is similar to the ones generated against allogeneic target cells.

F. ANALYSIS OF HUMAN PERIPHERAL BLOOD B-CELL

The one unequivocal function of B cells is their capacity to synthesize and secrete immunoglobulins. It is clear from studies in a number of mammalian species that clones of antibody-forming precursor cells are genetically programmed prior to interaction with antigens to synthesize antibody molecules of highly restricted specificity. Mecha- nisms by which these clones are triggered to differentiate into antibody-forming cells depend on the maturational state of the B cells, on their location within a variety of lymphoid compartments and, perhaps equally important, on the nature of signals delivered to the cell surface. Thus, although some antigens appear to be capable of triggering precursor B cells alone (“T independent”) the majority require additional signals that in many instances can be shown to origi-

FUNCTION


nate from antigen-triggered T cells. The nature of the specific T cell- generated signals is poorly understood at present. Some can be nonspecifically bypassed by: (1) macrophages and their products (Calderon et al., 1975); (2) nonspecifically activated T cells (Mond et al., 1972; Rubin and Coons, 1972; Amerding and Katz, 1974); and (3) polyclonal activators, such as lipopolysaccharide (LPS) or pokeweed mitogen (PWM) (Andersson et al., 1972; Parkhouse et al., 1972; Coutinho and Moller, 1975).

The relationship between nonspecific and antigen-specific T-cell helper effects is critical to our understanding of human B-cell functions. Most studies investigating human B-cell Ig synthesis have utilized PWM, since human cells respond poorly if at all to LPS or soluble anti-Ig. In a number of studies, human B cells have been found to respond to PWM by blast transformation, thymidine uptake, plasma cell development and Ig secretion. Optimal differentiation of B cells can be shown to be dependent on the presence of either stimulated T cells, macrophages, or their products.

Although these studies have provided important information concerning B cell function and have allowed the analysis of B cell immunodeficiency disorders, defining the mechanisms of human B cell activation has been limited by the difficulty in developing specific in vitro systems. In particular, plaque-forming assays have facilitated our understanding of the mechanisms and nature of signals required for specific antibody production in other species. Several investigators have recently reported the primary in vitro induction of antibody synthesis to hapten-conjugated T-dependent carriers or sheep erythrocytes (Watanabe et al., 1974; Dosch and Gelfand, 1976; Fauci et al., 1976). The methods used so far lack the ease and reproducibility observed in murine systems. The difficulties encountered may reflect in part the fact that human peripheral B cells may be less differentiated and more easily tolerized than splenic and lymph node cells. Human peripheral blood B lymphocytes triggered nonspecifically by polyclonal mitogens to secrete antierythrocyte antibodies in vitro continue to bear Ia determinants (Friedman et al., 1977). Since it is known that Ia determinants are not readily detected on plasma cells, it has been suggested that peripheral blood lymphocytes may be more immature than B lymphocytes found in other lymphoid compartments (Schlossman et al., 1976).

In this regard we and others have shown that B cells are capable of producing a significant plaque-forming response to sheep erythrocytes when triggered by PWM in the absence of antigen. In contrast, unfractionated populations and purified populations of T cells are not.


The PWM activated plaque-forming response represents true antibody synthesis since: (1) it can be inhibited by rabbit antihuman Ig; (2) it requires complement to lyse the plaques; (3) each plaque contains at least one, and up to ten central lymphoid cells; and (4) treatment with cycloheximide at doses known to inhibit protein synthesis inhibits plaque formation (Friedman et al., 1977). These controls are essential since many artifacts arise in human B-cell plaque-forming systems to heterologous erythrocytes.

A more “physiologic” trigger for B cells is the soluble products of activated T cells. Supernatants from antigen-triggered T cells containing lymphokines can induce B cells to differentiate into antibody- synthesizing cells. A requirement for optimal triggering of B-cell plaque formation is prior binding and elution from anti-F(ab)2 columns. B cells recovered from nylon wool columns or by other techniques are not as readily triggered by either PWM or activated T-cell supernatants. These studies suggest that partial activation of B cells by anti- F(ab), columns may provide an initial signal and stimulate antigen triggering or allow for subsequent triggering by PWM or T-cell products. Further, the addition of sheep erythrocytes to PWM-triggered B cells does not enhance the number of plaque-forming cells, but instead suppresses the response. The suppression is specific, since the response to unrelated erythrocytes is not diminished (Friedman and Chess, 1977). These data support the notion that peripheral blood B cells may be easily tolerized and, together with the observation that antibody-secreting B cells retain Ia antigens, provide support for the view that circulating human B cells are immature.

The pokeweed-induced plaque-forming system is less ideal than testing with antigen-specific T cell-dependent plaques. Nevertheless, these systems have allowed investigation of a number of the features of B-cell differentiation in man. We have studied the effects of the anti-p23,30 sera on plaque-forming cells following stimulation with PWM and T cell supernatants. The following observations were made: (1) anti-p23,30 but not normal rabbit serum markedly reduces the generation of PFCs, but only partially suppresses the proliferative response to PWM; (2) the induction of both PFC and B cell proliferation b y soluble products of activated T cells is markedly inhibited by anti-p23,30; (3) the inhibitory effect of anti-p23,30 in the mitogen- induced PFC response can be demonstrated only when the antiserum is present in the early phases of cell culture-its effects are minimal at later stages; and (4) the p23,30 antigen is retained by the more “differentiated” antibody-secreting cells after 6 days of culture (Friedman et al., 1977). These results suggest that the Ia-like determinant, p23,30,


although present on mature PFCs, plays its major role during the early differentiation of activated precursor B cells into antibody-forming cells.

G. CHARACTERIZATION AND ISOLATION OF REGULATOR CELLS

It is apparent in murine systems that both the type and intensity of the cellular and humoral immune responses to antigen can be homeo- statically regulated by a complex series of interactions involving distinct classes of lymphocytes and macrophages. Distinct subclasses of T cells are capable of helper, amplifier, and suppressor effects on the development of T cell-mediated functions such as cytotoxicity and B-cell secretion of Ig (Gershon, 1974; Cantor and Boyse, 1976). Two interesting aspects of these complex cellular interactions are (1) that the regulatory influences of T-cell subclasses (both helper and suppressor) are under relatively strict genetic control by genes linked to the MHC (reviewed by Katz and Benacerraf, 1975) and (2) that a number of regulatory functions of these subclasses are mediated by factors that influence other T-cell subclasses, B cells, or macrophages (Tada et al., 1976; Taussig and Munro, 1974; Rich and Pierce, 1974).

In man, it is now clear that regulatory cells exist and for the most part are found in the Ig- subclass of cells and within adherent cell populations (Waldmann et al., 1974; Siegal et al., 1976; Friedman et al., 1977). For example, it has been shown that the formation of PWM- induced B-cell plaques can be suppressed by the addition of autologous Ig- cells. Furthermore, this suppression can be augmented by prior addition of concanavalin A (Con A). Treatment of Con A-activated Ig- cells with anti-TH1 serum plus complement, while eliminating approximately 50% of the cells, had no effect on the suppressor cell function. Thus, one subclass of human suppressor cells may exist within the T H I - subclass. In addition, other subpopulations of human peripheral mononuclear cells can also suppress the differentiation of B cells triggered by polyclonal mitogens. Adherent cells within the Ig- population, e.g., like the nonadherent population, can also suppress B cells. Whether adherent Ig- cells are identical to nonadherent suppressor T cells remains to be determined.

In addition, evidence has been presented by a number of investigators using purified human T cells that at least one subclass of T cells has the capacity to augment B-cell differentiation into Ig- synthesizing cells. These putative helper T-cell subclasses have been recognized both by heteroantisera directed at unique subclasses of T cells and by alloantisera present in patients with juvenile rheumatoid arthritis and are directed against specificities of subclasses of T cells

IN HUMAN PERIPHERAL BLOOD


that suppress B-cell secretion of Ig (Strelkauskas et al., 1977). Moreover, anti-TH, antibody eliminates that subclass of T cells which secrete factors capable of augmenting B-cell differentiation (Breard et al., 1977). Thus, the concept is emerging that in man there are distinct subclasses of T cells that can initiate, help, and suppress the differentiation of B cells. Clarification of the specificity of these effects will have to await more precise antigen-specific B cell functional assays.

H. THE FUNCTIONAL HETEROGENEITY OF NULL CELLS

As indicated initially, considerable evidence exists that there are subclasses of lymphocytes which lack the precise surface properties of the predominant differentiated classes of T and B lymphocytes. In man, this null-cell subclass is distinguished from T cells by its failure to form spontaneous rosettes with sheep erythrocytes and from B cells by its lack of detectable surface Ig. Further, the null cell is heterogeneous both with respect to the complement and Fc receptors, and, perhaps of greater importance, the null-cell subclass is heterogeneous with respect to the Ia determinant, p23,30. Approximately 2040% of the null cells isolated b y immunoabsorbant anti-F(ab)2 chromatography followed by E-rosette depletion, are both complement- receptor positive and p23,30 positive (Chess et a/ . , 1976).

Functional studies of the null-cell subclass have indicated that null cells are distinct from mature T cells (MacDermott et ul., 1975). Null cells do not proliferate in response to soluble or allogeneic cell surface antigens, whereas T cells do; and they do not kill allogeneic lymphocytes by direct CML. In addition, null cells, but not T cells, are the effector cells in ADCC. In contrast, the p23,30-positive subclass of null cells is functionally similar to B cells in that both subclasses spontaneously secrete Ig. However, the null-cell subclass, unlike circulating surface Ig+ B cells, effect ADCC. Treatment of null cells with anti-p23,30 plus complement depletes both their capacity to synthesize Ig in cell culture and their capacity to effect ADCC. The simplest interpretation of these findings is that the ADCC effector cell (i.e., the E-, Ig-, p23,30+, complement receptor-positive) in man and Ig- synthesizing cell (E-, Ig+) bear close relationships and may represent different stages of B-cell differentiation. However, it is important to point out that the null cell bearing p23,30 is distinct from most peripheral B cells in being nonadherent, non-Ig bearing, and capable of eliciting the ADCC reaction. Whether there still exists a heterogeneity of cell types within the p23,30+, Ig-, EAC+ subclass that would distinguish the ADCC-reactive cell from the Ig- synthesizing cell cannot be resolved with the available reagents.

It is important to emphasize that approximately 70% of the null-cell


subclass is p23,30- and the function of most of these cells remains largely unknown. Some of these cells have the capacity to differentiate, after appropriate triggering, in cell culture into granulocyte colonies (Richman and Chess, 1977), erythrocyte colonies (Nathan and Chess, 1977), and E+ T cells. Thus, the null compartment of peripheral human mononuclear cells appears to contain precursor cells capable of generating the entire spectrum of hematopoietic and lym- phopoietic cells. Whether all these precursors will be distinguished with antigens on their surface analogous to p23,30 remains to be determined.

REFERENCES Aiuti, F., and Wigzell, H. (1973). Clin. E x p . Immunol. 13, 171. Ambrose, E. J., ed. (1965). “Cell Electrophoresis.” Churchill, London. Amerding, D., and Katz, D. H. (1974).]. E x p . Med. 140,19. Anderson, J., Sjoberg, O., and Moller, G. (1972). Eur. J . Zmmunol. 2,349. Arbeit, R. D., Henkart, P. A., and Dickler, H. B. (1976). In “In Vitm Methods in Cell

Mediated Immunity” (B. R. Bloom and J. R. David, eds.), p. 143. Academic Press, New York.

Bach, M. L., Alter, B. J., Lightbody, J. J., and Bach, F. J. (1972). Transplant. Proc. 4,205. Balch, C. M., Lawton, H. R., and Cooper, M. D. (1976). In “In Vitm Methods in Cell

Mediated and Tumor Immunity” (B. R. Bloom and J. R. David, eds.), p. 105. Academic Press, New York.

Basten, A., Miller, J. R. A. P., Sprent, J., and Pye, J. (1972).J. E x p . Med. 135, 610. Bianco, C., Patrick, R., and Nussenzweig, V. (1970).J. E x p . Med. 132, 702. Bloom, B. R., and Bennett, B. (1966). Science 153, 80. Bloom, B. R., and David, J. R., eds. (1976). “In Vitro Methods in Cell Mediated and

Tumor Immunity.” Academic Press, New York. Bonner, W. A., Hullet, N. R., Sweet, R. C., and Herzenberg, L. A. (1972). Reu. Sci .

Instmm. 43,404. Brain, P., Cordon, J.. and Willetts, W. A. (1970). Clin. E x p . Immunol. 6, 681. Breard, J. M., Chess, L., Fuks, A., and Schlossman, S. F. (1977). Fed. Proc., Fed. Am. SOC.

Brier, A. J., Chess, L., and Schlossman, S. F. (1975).J. Clin. Inuest. 56, 1580. Brouet, J. C., and Tobin, H. (1976).]. Immunol. 116, 1041. Brown, C., and Greaves, M. F. (1974). Scand. J . Immunol. 3 , 161. Calderon, J. J., Kiely, M., Lefko, J. L., and Unanue, E. R. (1975).J. E x p . Med. 142,151. Cantor, H., and Boyse, E. A. (1975a).J. E x p . Med. 141, 1376. Cantor, H., and Boyse, E. A. (1975b). J . E x p . Med. 141, 1390. Cantor, H.. and Boyse, E. A. (1977). Cold Spring Harbor Symp. Quant. B w l . Vol. XLI (in

Cantor, H., and Jandinski, S. (1974).J. Exp. Med. 140, 1712. Cantor, H., and Weissman, I. (1976). Prog. Allergy 20, 1. Cerottini, J. C., and Bmnner, K. T. (1974). Adu. Zmmunol. 18,67. Chess, L., and Rocklin, R. E. (1977). In preparation. Chess, L., and Schlossman, S. F. (1977). Contemp. Top. Zmmunol. (in press). Chess, L., MacDermott, R. P., and Schlossman, S. F. (1974a).J. Immunol. 113,1113.

Exp. Biol. 36, 1262.

press).


Chess, L., MacDermott, R. P., and Schlossman, S. F. (1974b).J. Zmmunol. 113, 1222. Chess, L., MacDermott, R. P., Sondel, P., and Schlossman, S . (1974~). Prog. Zmmuml. 3,

Chess, L., Evans, R., Humphreys, R. E., Strominger, J. L.,and Schlossman, S. F. (1976).

Coombs, R. R., Burner, B. W., Wilson, A. B., Holm, B., and Lindgren, B. (1970). Znt.

Cooper, M. D., Lawton, A. R., and Bockman, D. E. (1971). Lancet 2,791. Coutinho, A., and Moller, G. (1975). Ado. Immunol. 21, 114. David, J. R. (1966). Proc. Natl. Acad. Sci. U.S.A. 56, 72. Dickler, H. B., and Kunkel, H. G. (1972).J. E x p . Med. 136, 191. Dosch, N. M., and Gelfand, E. W. (1976).]. Zmmunol. Methods 11, 107. Eijsvoogel, V. P., duBois, M. J., Meinesz, A., Bierhorst-Eijlander, Zyelemaker, W. P.,

Evans, R. L., Breard, J. M., Lazarus, H., Schlossman, S. F., and Chess, L. (1977).J. E x p .

Fauci, A. S., Pratt, K. R., and Whalen, G. ( 1 9 7 6 ~ 1 . Zmmunol. 117,21a’. Friedman, S. M., and Chess, L. (1977). In preparation. Friedman, S. M., Breard, J. M., Humphreys, R. E., Striiminger, J. L., Schlossman, S. F.,

Froland, S. S. (1972). Scand. J. Zmmunol. 1, 269. Froland, S. S., and Natvig, J. B. (1970). Znt. Arch. Allergy Alwl . Zmmunol. 39, 121. Fuks, A. et al. (1977). In preparation. Gardner, J. D., Bowern, W. A., and Blanden, R. J. (1975). Eur.1. Zmmunol. 5, 1222. Geha, R. S., and Merler, E. (19744. Eur. J. Zmmunol. 4, 193. Geha, R. S., and Merler, E. (197413). Cell. Zmmunol. 10, 86. Geha, R. S., Merler, E., and Rosen, F. S. (1973).]. Clin. Znoest. 52, 1726. Gershon, R. K. (1974). Contemp. Top. 3, 1 . Greaves, M. F., and Brown, G. (1974).J. Zmmunol. 112,420. Greaves, M. F., and Janossy, G. (1976). In “In Vitro Methods in Cell Mediated and

Tumor Immunity” (B. R. Bloom and J. R. David, eds.), p. 89. Academic Press, New York.

Greenberg, A. H., Hudson, L., Shen, L., and Reitt, I. (1973). Nature (London) 242,111. Grey, H. M., Rabellino, E., and Pirofsky, B. (1971).J. Clin. Znoest. 50, 2368. Hammarstrom, S., Hellstrom, U., Perlmann, P., and Dillner, M. L. (1973).J. Erp. Med.

Harrison, M. R., and Paul, W. E. (1973). J. E x p . Med. 138, 1602. Hayry, P., and Defendi, V. (1970). Science 168, 133. Hayry, P., Anderrson, L. C., Gamberg, C., Roberts, P., Rank, A., and Nordling, S. (1975).

Humphreys, R. E., Chess, L., Schlossman, S . F., and Strominger, J. (1976).J. E x p . Med.

Jondal, M., and Klein, G. (1973). J . E x p . Med. 138, 1365. Jondal, M., Holmm, G., and Wigzell, H. (1972). J. E x p . Med. 136, 207. Jondal, M., Wigzell, H., and Aiuti, F. (19734. Transplant. Rev. 16, 163. Jondal, M., Wigzell, H., and Aiuti, F. (1973b). Transplant. Reu. 16, 169. Jondal, M., Klein, E., and Yefenof, E. (1975). Scand. J. Zmmuml. 4,259. Jones, E . A., Goodfellow, P. N., Bodmer, J. G., and Bodmer, W. F. (1975). Nature

Kaplan, 1. F., and Clark (1974).J. Zmmunol. Methods 5, 131. Katz, D. H., and Benacerraf, B. (1975). Trcinsplnnt. Reo. 22. 17S.

125.

J. E x p . Med. 144, 114.

Arch. Allergy Appl. Zmmunol. 39, 658.

and Schellekens, P. T. (1973). Transplant Proc. 5, 1675.

Med. 145, 221.

and Chess, L. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 711.

138, 1270.

Zsr. J . Med. Sci. 11, 1299.

144,98.

(London) 256, 650.


Kirkpatrick, C. N.,, Rich, R. R., and Smith, T. R. (1972).]. Clin. Znoest. 51, 2948. Lay, W. H., Mendes, N. F., Bianco, C., and Nussenzweig, V. (1971). Nature (London)

Levin, A. S., Spitler, L. E., Stites, D. P., and Fudenberg, H. H. (1970). Proc. Natl. Acad.

Lightbody, J. J., Bernoco, D., Miggiano, V. L., and Ceppellini, R. (1971). G. Batt. Virol.

Lohrmann, H. P., and Novikovs, L. (1974). Clin. Immunol. Zmmunopathol. 3,99. Loken, M. R., and Herzenberg, L. A. (1975). Ann. N.Y. Acad. Sci. 254, 163. MacDermott, R. P., Chess, L., and Schlossman, S. F. (1975). Clin. Immutwl. Zm-

Maini, R. W., Bryceson, D. M., Wolstencroft, R. A., and Dumonde, D. C. (1969). Nature

Mann, D. L., Abelson, L., Harris, S. D., and Amos, D. B. (1975).J. E x p . Med. 142, 84. Mendes. W. F., Tolnai, M. E., Silveira, R. P., Gilbertson, R. P., and Metzger, R. S. (1973).

Miller, R. G. (1976). In “In Vitro Methods in Cell-Mediated and Tumor Immunity”

Moller, G. (1973). Transplont. Reo. 16, 1. Mond, J. J., Takahashi, T., and Thorbeck, G. J. (1972).]. E x p . Med. 136, 715. Moretta, L., Ferrarini, M., Mingeri, M. C., Morretta, A., and Webb, S. (1976).J. ImmunoI.

Moretta, L. S., Webb, S. R., Gross, C. E., Lydyard, P., and Cooper, M. D. (1977).J. E x p .

Nathan, D., and Chess, L. (1977). In preparation. Parkhouse, B. M., Janossy, G., and Greaves, M. F. (1972). Nature (London), New B i d .

Perlmann, P., Perlmann, H., and Miiller-Eberhard, H. J. (1975).J. E r p . Med. 141,287. Raidt, D. J., Mishell, R. J., and Dutton, R. W. (1968).J. E x p . Med. 128, 681. Rich, R. R., and Pierce, C. W. (1974).J. Zmmutwl. 112, 1360. Richman, C., and Chess, L. (1977). In preparation. Rocklin, R. E. (1974).J. Immunol. 122, 1461. Rocklin, R. E., Meyers, 0. L., and David, J. R. (1970). 104,95. Rocklin, R. E., Reardon, G., Sheffer, A., Churchill, W. H., and David, J. R. (1971). Proc.

Rocklin, R. E., MacDermott, R. P., Chess, L., Schlossman, S. F., and David, J. R. (1974).

Rubin, A. S., and Coons, H. H. (1972).J. E x p . Med. 135, 437. Sallan, S. E., Chess, L., Frei, E., O’Brien, C., Nathan, D. G., Strijminger, J. L., and

Schlossman, S. F. (1977). Submitted for publication. Schendel, D. J., Alter, B. J., and Bach, F. H. (1973). Trunsplant. Proc. 5, 1651. Schlossman, S. F., and Hudson, L. (1973).J. Immunol. 107,725. Schlossman, S. F., Chess, I+ Humphreys, R. E., and Strominger, J. L. (1976). Proc. Natl.

Schrader, J. W., and Edelman, G. M. (1976).J. E r p . Med. 143,601. Shearer, G . M., Rehn, T. G . , and Garbarino, C. A. (1975).J. E x p . Med. 141, 1348. Shen, F. W., Boyse, E. A., and Cantor, H. (1975). Zmmunogenetics 2,591. Shiku, H., Kisielow, P., Bean, M. A., Takahash, T., Boyse, E. A., Oetigen, N . F., and Old,

Shortman, K., von Boehmer, H., Lipp, J., and Hopper, K. (1975). Transplant. Reo. 25,

230, 531.

Sci. U S A . 67, 821.

64,243.

munopathol. 4,415.

(London) 224,43.

J . Zrnmunol. 111, 860.

(B. R. Bloom and J. R. David, eds.), p. 283. Academic Press, New York.

117,2171.

Med. 146, 184.

235,21.

Leukocyte Cult. Conf., Sth, 1970 p. 639.

J . EX^. Med. 140, 1303.

Acad. Sci. U.S.A. 73, 1288.

L. J. (1975). J . E x p . Med. 141,227.

163


Siegal, F. P., Pemis, B., and Kunkel, H. G . (1971). Eur. J . Immunol. 1,482. Siegal, F. P., Siegal, M., and Good, R. A. (1976).J. Clin. lnoest . 58, 109. Snary, D., Barnstable, C . J., Bodmer, W. F., Goodfellow, P. N., and Crumpton, M. J .

Solliday, S., and Bach, F. H. (1970). Science 170, 1406. Sondel, P. M., O’Brien, C., Porter, L., Schlossman, S. F., and Chess, L. (1976).J. lm-

munol. 117, 2197. Strelkauskas, A. J., Schauf, V., Wilson, B. S., Schlossman, S. F., and Chess, L. (1977). In

preparation. Striiniinger, J . L.. Chess, L., Humphreys, R. F., Mann, D., Parham, P., Robb, R.,

Schlossman, S. F., Springer, T., andTerhorst, C. ( 1976). In “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (D . H. Katz and B. Benacerraf, eds.), Academic Press, New York.

(1977).J. E x p . Med. (in press).

Tada, T., Taniguchi, M., and Takemori, T. (1976). Tranqdant. Reo. 26, 311. Tausig, M. J . , and Munor, A. J. (1974). Nature (London) 251, ci3. Valdimarsson, H . , Agnersdottir, G., and Lochman, P. J . (1975). Nature (London) 255,

van Rood, J . J., van Leeuwen, A., Keuning, J . J.. and Termigtelen, 0. D. (1975).

von Boehnier, H. (1974). J . Zmmunol. 112, 70. Wahl, S. J . . Iverson, G. M., and Oppenheim, J . (1974).J. EX?,. Med. 140, 1631. Wahl, S. M., and Rosenstreich, D. L. (1976).J. Exr,. Med. 144, 1175. Waldmann, T. A., Dunn, V., Broder, S., Blackman, M., Blaese, R. M., and Stroller, W.

Watanabe, T., Yoshizaki. K., Yagnra, T., and Yamamara, Y. (1974)J. Zmmunol. 113, 608. Wernet, P. R., Winchester, R., Kunkel, H. G., Wemet, D., Giphart, M.,von Leeuwen, A.,

and van Rood, J . J . (1975). Transplant. Proc. 7, 193. Whitconib, M., and Rocklin. R. E. (1973). Ann. Intern. Med. 79, 161. Wigzell, H., and Andersson. B. (1969).J. E x p . Med. 129,23. Winchester, R. J. , Fu, S. M., Wernet, P., Kunkel, H . G., Dupont, B., and Jersild, C.

Winchester, R. J., Fu, S. M., HoEman, T., and Kunkel, H. G. (1975b)J. lmmunol. 114,

Woody, J. W., Ahmed, A,, Knudsen, R. C., Strong, D. M., and Sell, K. W. (1975)J. Clin.

Wyliran, J., Carr, M. C., and Fudenberg, H. H. (1972). Zinkernitgel, R. M., and Doherty, P. C. (1975)./. E x p . Med. 141, 1427.

554.

Transplant. Proc. 7, .3l.

(1974). Lancet 2, 609.

(1975a).J. E x p . Med. 141, 924.

1210.

lnoest . 55, 956.

Subject Index

A Diffusion chambers

45-46

47-49

corneal, immunological privilege and,

millipore, immunological privilege and,

Antigens human B-cell specific, 220-222 Iiuman thymocytes and peripheral

blood T-cell subclasses, 216-220 E

B

Bcells

Blood group and histocompatibility human, specific antigens, 220-222

polymorphism

Ag-C locus, 112-113 Ag-D locus, 113-114 B 1 antigen, 114 H-3, H-4 and H-5 loci, 114 sex-linked antigens, 114-1 15

rat

Bone marrow space, immunological privilege and,

24-25 Brain, immunological privilege and,

11-15

C

Cheek pouch hamster, immunological privilege and,

15-22 Cornea, immunological privilege and,

Cytotoxic effector cells 8-1 1

fine specificity, 64-65 H-2-restricted, 66-73 infecting or modifying agent, 73-78 models, 65

Cytotoxicity

78-83 H-2 restricted, immune response genes,

T-cell mediated, 57-63

D

Delayed hypersensitivity, T-lymphocy tes and, 63-64

Eye anterior chamber, immunological

privilege, 3-8

ti

Hair follicle

23-24 matrix, immunological privilege and,

Hamster cheek pouch, immunological privilege

and, 15-22 Histocompatibility loci

rat, evidence for selection at, 115-116

I

Immune response rat genetics

experimental allergic encephalitis,

MHC-linked Ir genes, 117-119 other genes regulating immune

Immune response genes, H-2 restricted

Immunoglobulin genetics

119-123

reactions, 123-125

cytotoxicity and, 78-83

rat allotypes of IgG,, chain, 129 alpha chain allotypes, 128-129 anti-receptor site antibodies, 130-131 IgE-secreting immunocytomas,

kappa chain allotypes, 125-128 129-1 30

Immunologically privileged sites anterior chamber of the eye, 3-8

24 3

244 SUBJECT INDEX

artificial alymphatic skin flaps, 34-41 corneal diffusion chambers, 45-46 millipore diffusion chambers, 47-49 muscle, 42-43 scars, 46-47 skin islands, 43-45 traumatized panniculus carnosus

muscle, 41-42 bone marrow space and, 24-25 brain, 11-15 cornea, 8-1 1 eye lens, 11 hamster's cheek pouch, 15-22 liver and, 29-32 matrix of hair follicle, 23-24 prostate and, 27-29 subcutaneous tissue, 22-23 testicle and, 25-27 uterus and, 32-34

1

Lens, immunological privilege and, 11 Liver, immunological privilege and,

Lymphocytes, see also Bcells, T-cells 29-32

functional analysis of isolated human

cell-mediated destruction of tumor

cell-mediated lympholysis, 230-231 general considerations, 226-227 heterogeneity of null cells, 237-238' mediator production and, 228-230 peripheral blood Bcell function,

proliferation in response to antigen,

regulator cells in blood, 236-237

subpopulations

cells, 231-233

233-236

227-228

human, classical cell surface

subclass purification determinants on,. 214-216

affinity chromatography on columns,

fluorescence-activated cell sorting,

rosette depletion techniques, 225 other separation techniques, 226

222-223

223-225

Lymphocyte alloantigens, rat, 107-112

M

Major histocompatibility complex rat

definition of alleles, 98-101 genetic organization, 103-105 Ia-like antigens, 101-103 immunity to MHC antigens, 105-107

restriction for distinct T-lympocyte functions

delayed hypersensitivity and, 63-64 proliferation in mixed lymphocyte

T-cell-mediated cytotoxicity, 57-63 MHC, see Major histocompatibility

complex Mixed lymphocyte reactions, T-cells and,

64 Mouse

reaction, 64

V-region structures V-region domain structure: model

V,: classification of sequences; CDR,

VL: classification of sequences; CDR,

building, 145-147

155-162

147-155 Muscle, immunological privilege and,

Myeloma proteins 42-43

groups binding the same hapten N-acetylglucosamine, 199-200 N-acetylmannosamine, 201 dinitrophenyl (DNP), 162-176 flagellin, 200 fructan-inulin (GplF) and grass levan

(GpGF), 196-199 galactan (p6Gal), 187-192 glucan (a3G), 192-195 glucan (a6G), 195-196 lipopolysaccharides, 2DO-201 phosphorylcholine (PC), 176-187

P

Panniculus carnosus muscle traumatized, immunological privilege

and, 41-42 Prostate, immunological privilege and,

27-29

SUBJECT INDEX 245

R

Rat current linkage map, 131-132 genetics of immune response

experimental allergic encephalitis,

MHC-linked Ir genes, 117-119 other genes regulating immune

reactions, 123-125 histtrompatibility loci, evidence for

selection, 115-1 16 immunogenetics, historical, 93-98 immunoglobulin genetics

allotypes of IgC,, chain, 129 alpha chain allotypes, 128-129 anti-receptor site antibodies, 130-131 IgE-secreting immunocytomas,

kappa chain allotypes, 125-128 lymphocyte alloantigens, 107-1 12 major histocompatibility complex

definition of alleles, 98-101 genetic organization, 103-105 Ia-like antigens, 101-103 immunity to MHC antigens, 105-107

histocompatibility polymorphisms

119-123

129-130

other blood group and

Ag-C locus, 112-113 Ag-D locus, 113-114 B1 antigen, 114

H-3, H-4, and H-5 loci, 114 sex-linked antigens, 114-115

S

Scars, immunological privilege and,

Skin 46-47

islands, immunological privilege and, 4 3 4 5

Skin flaps alymphatic, immunological privilege

and, 34-41 Subcutaneous tissue, immunological

privilege and, 22-23

1

T-cells subclasses, distinguishing antigens,

Testicle, immunological privilege and,

Thymocytes

216-220

25-27

human, antigens distinguishing peripheral blood T-cell subclasses, 2 16-220

U

Uterus, immunological privilege and, 32-34

CONTENTS OF PREVIOUS VOLUMES

247

Volume 1

Transplantation Immunity and Tolerance M. HASEK, A. LENGEROVA, AND

T. HRABA

Immunological Tolerance o f Nonliving Antigens

RICHARD T. SMITH

Functions o f the Complement System ABRAHAM G. OSLER

In Vitro Studies of the Antibody Response ABRAM B. STAVITSKY

Duration o f Immunity i n Virus Diseases J. H . HALE

Fote and Biological Action o f Antigen-Antibody Complexes

WILLIAM 0. WEICLE

Delayed Hypersensitivity to Simple Protein Antigens

P. G. H. CELL AND B. BENACERRAF

The Antigenic Structure o f Tumors P. A. CORER

AUTHOR INDEX-SUBJECT INDEX

Volume 2

Immunologic Specificity and Molecular Structure

FRED KARUSH

Heterogeneity o f y-Globulins JOHN L. FAHEY

The Immunological Significance o f the Thymus

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

Cellular Genetics of Immune Responses G. J. V. NOSSAL

Antibody Production by Transferred Cells CHARLES G. COCHRANE AND

FRANK J. DIXON

Phagocytosis DERRICK ROWLEY

Antigen-Antibody Reactions i n Helminth Infections

E. J . L. SOULSBY

Embryological Development o f Antigens REED A. FLICKINCER


Volume 3

In Vitro Studies o f the Mechanism o f Anaphylaxis

K. FRANK AUSTEN AND

JOHN H . HUMPHREY

The Role of Humoral Antibody i n the Homograft Reaction

CHANDLER A. STETSON

immune Adherence D. S. NELSON

Reaginic Antibodies D. R. STANWORTH

Nature of Retained Antigen and i ts Role in Immune Mechanisms

DAN H. CAMPBELL AND

JUSTINE S. GARVEY

Blood Groups in Animals Other Than Man W. H. STONE AND M. R. IRWIN

Heterophile Antigens and Their Significance i n the Host-Parasite Relationship

C . R. JENKIN


248 C O N T E N T S OF PREVIOUS V O L U M E S

Volume 4

Ontogeny and Phylogeny o f Adaptive Immunity

ROBERT A. GOOD AND

BEN W. PAPERMASTER

EMANUEL SUTER AND

HANSRUEDY RAMSEIER

Cellular Reactions i n Infection

Ultrastructure of Immunologic Processes JOSEPH D. FELDMAN

Cell Wa l l Antigens of Gram-Positive Bacteria

MACLYN MCCARTY AND

STEPHEN I. MORSE

Structure and Biological Activity of Immunoglobulins

SYDNEY COHEN AND RODNEY R. PORTER

Autoa n ti bodies and Disease H . G . KUNKEL AND E. M. TAN

Effect o f Bacteria and Bacterial Products on Antibody Response


J. MUNOZ

Volume 5

Natural Antibodies and the Immune Response

STEPHEN V. BOYDEN

Immunological Studies with Synthetic Polypeptides

MICHAEL SELA

Experimental Allergic Encephalomyelitis and Autoimmune Disease

PHILIP Y. PATERSON

The Immunology of Insulin C. G . POPE

Tissue-Specific Antigens D . C. DUMONDE


Volume 6

Experimental Glomerulonephritis: Immunological Events and Pathogenetic Mechanisms

EMIL R. UNANUE AND

FRANK J. DIXON

Chemical Suppression o f Adaptive Immunity

ANN E. GABFUELSON AND

ROBERT A. GOOD

Nucleic Acids as Antigens OTTO J. PLESCIA AND

WERNER BRAUN

In Vitro Studies o f Immunological Responses o f lymphoid Cells

RICHARD W. DUTTON

Developmental Aspects o f Immunity JAROSLAVSTERZLAND ARTHUR M. SILVERSTEIN

Anti-antibodies PHILIP G. H . CELL AND

ANDREW S. KELUS

Conglutinin and lmmunoconglutinins P. J. LACHMANN


Volume 7

Structure and Biological Properties o f Immunoglobulins

SYDNEY COHEN AND CESAR

M ILSTEIN

Genetics o f Immunoglobulins i n the Mouse MICHAEL POTTER AND

ROSE LIEBERMAN

Mimetic Relationships between Group A Streptococci and Mammalian Tissues

JOHN B. ZABRISKIE

lymphocytes and Transplantation Immunity DARCY B. WILSON AND

R. E. BILLINCHAM

C O N T E N T S OF PREVIOUS VOLUMES 249

Human Tissue Transplantation JOHN P. MERIULL


Volume 8

Chemistry and Reaction Mechanisms of Complement

HANS J . MULLER-EBERHARD

Regulatory Effect of Antibody on the Immune Response

JONATHAN w. UHR AND

GORAN MOLLER

The Mechanism o f Immunological Paralysis D . w. DRESSER AND

N. A. MITCHISON

In Vitro Studies o f Human Reaginic Allergy ABRAHAM G. OSLER, LAWRENCE M. LICHTENSTEIN, AND DAVID A. LEVY


Volume 9

Secretory Immunoglobulins THOMAS B. TOMASI, JR., AND

JOHN BIENENSTOCK

Immunologic Tissue Injury Mediated by Neutroph i l i c leukocytes

CHARLES G . COCHRANE

The Structure and Function of Monocytes and Macrophages

ZANVIL A. COHN

The Immunology and Pathology o f NZ8 Mice

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


Volume 10

Cel l Selection by Antigen i n the Immune Response

GREGORY w. SlSKlND AND

BARUJ BENACERFMF

Phylogeny of Immunoglobulins HOWARD M. GREY

Slow Reacting Substance o f Anaphylaxis ROBERT P. ORANGE AND

K. FRANK AUSTEN

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

OSCAR D. RATNOFF

Antigens o f Virus-Induced Tumors KARL HABEL

Genetic and Antigenetic Aspects of Human Histocompatibility Systems

D . BERNARD AMOS


Volume 11

Electron Microscopy o f the Immunoglobulins

N. MICHAEL GREEN

Genetic Control of Specific Immune Responses

HUGH 0. MCDEVITT AND

BARUJ BENACERFMF

The lesions i n Cell Membranes Caused by Complement

JOHN H. HUMPHREY AND

ROBERT R. DOURMASHKIN

Cytotoxic Effects of lymphoid Cells In Vitro PETER PERLMA” AND

GORAN HOLM

H. S. LAWRENCE Transfer Factor

Immunological Aspects o f Malar ia Infection

IVOR N. BROWN


Volume 12

The Search for Antibodies with Molecular Uniformity

RICHARD M. KRAuSE

250 CONTENTS O F PREVIOUS VOLUMES

Structure and Function o f yM Macrog I obu li ns

HENRY METZGER

Transplantation Antigens R. A. REISFELT AND B. D. KAHAN

The Role of Bone Marrow i n the Immune Response

NABIH I . ABDOU AND

MAXWELL RICHTER

D. W. TALMAGE,

H. HEMMINGSEN

GERALD WEISSMANN AND

PETER DUKOR

Cell lnteroction i n Antibody Synthesis

J. hDOVICH, AND

The Role of Lysosomes in Immune Responses

Molecular Size and Conformation of Immunoglobulins

KEITH J. DORFUNGTON AND

CHARLES TANFORD


Volume 13

Structure and Function o f Human Immunoglobulin E

HANS BENNICH AND

S. CUNNAR 0. JOHANSSON

individual Antigenic Specificity of Immunoglobulins

JOHN E. HOPPER AND

ALFRED NISONOFF In Vitro Approaches to the Mechanism of Cell-Mediated Immune Reactions

BARRY R. BLOOM

Immunological Phenomena in leprosy and Related Diseases

J. L. TURK AND

A. D. M. BRYCESON

Nature and Classification o f I mmedi ate-Type Allergic Reactions

ELMER L. BECKER


Volume 14

lmmunobiology of Mammalian Reproduction

ALAN E. BEER AND

R. E. BILLINGHAM

Thyroid Antigens and Autoimmunity SIDNEY SHULMAN

Immunological Aspects o f Burkitt's lymphoma

GEORGE KLEIN

Genetic Aspects o f the Complement System CHESTER A. ALPER AND

FRED S. ROSEN

The Immune System: A Model for Differentiation in Higher Organisms

L. HOOD AND J. PRAHL


Volume IS

The Regulatory Influence of Activated T Cells on B Cell Responses t o Antigen

DAVID H. KATZ AND BARUJ BENACERRAF

The Regulatory Role o f Macrophages i n Antigenic Stimulation

E. R. UNANUE

Immunological Enhancement: A Study o f Blocking Antibodies

JOSEPH D. FELDMAN

Genetics and Immunology of Sex-linked Antigens

DAVID L. GASSER AND

WILLYS K. SILVERS

Current Concepts of Amyloid EDWARD C. FRANKLIN AND

DOROTHEA ZUCKER-FRANKLIN


CONTENTS OF PREVIOUS VOLUMES 25 1

Volume 16

Human Immunoglobulins: Classes, Subclasses, Genetic Varionts, and Id i oty per

J. B. NATVIG AND H. G . KUNKEL

lmmunolog ica l Unresponsiveness WILLIAM 0. WEIGLE

Participation o f lymphocytes i n Viral Infections

E. FREDERICK WHEELOCK AND

STEPHEN T. TOY

Immune Complex Diseases in Experimental Animals and Man

C. G. COCHRANE AND D. KOFFLER

The Immunopathology o f Joint Inflammation in Rheumatoid Arthritis

NATHAN J. ZVAIFLER


Volume 17

Antilymphocyte Serum EUGENE M. LANCE, P. B. MEDAWAR, AND ROBERT N. TAUB

In Vitro Studies of Immunologically Induced Secretion o f Mediators from Cells and Related Phenomena

ELMER L. BECKER AND

PETER M. HENSON

KEITH M. COWAN Antibody Response to Viral Antigens

Antibodies to Small Molecules: Biological and Clinical Applications

SAM M. BEISER VINCENT P. BUTLER, JR., AND


Volume 18

Genetic Determinants o f Immunological Responsiveness

DAVID L. GASSER AND WILLYS K. SILVERS

Cell-Mediated Cytotoxicity, Al lograf t Rejection, a n d Tumor Immunity

JEAN-CHARLES CEROTTINI AND

K. THEODORE BRUNNER

Antigenic Competition: A Review o f Nonspecific Antigen-Induced Suppression

HUGH F. P~oss AND DAVID EIDINGER

Effect o f Antigen Binding on the Properties of Antibody

HENRY METZCER

lymphocyte-Mediated Cytotoxicity and Blocking Serum Activity to Tumor Antigens

KARL ERIK HELLSTR~M AND

INCEGERD HELLSTR~M


Volume 19

Molecular Biology o f Cellular Membranes with Applications to Immunology

S. J. SINGER

Membrane Immunoglobulins and Antigen Receptors on B and T lymphocytes

NOEL L. WARNER

Receptors for Immune Complexes on lymphocytes

VICTOR NUSSENZWEIC

Biological Activities o f Immunoglobulins of Different Classes and Subclasses

HANS L. SPIEGELBERC

SUBJECT INDEX

Volume 20

Hypervariable Regions, Idiotypy, and Antibody-Combining Site

J. DONALD CAPRA AND

J . MICHAEL KEHOE

MARIAN ELLIOTT KOSHLAND Structure and Function of the J Chain

252 CONTENTS OF PREVIOUS VOLUMES

Amino Acid Substitution and the Antigenicity o f Globular Proteins

MORRIS REICHLIN

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

DONALD c. SHREFFLER AND

CHELLA S. DAVID

ALFRED J. CROWLE Delayed Hypersenshivty i n the Mouse

SUBJECT INDEX

Volume 21

X-Ray Diffraction Studies o f Immunoglobulins

ROBERTO J. POLJAK

Rabbit Immunoglobulin Allotypes: Structure, Immunology, and Genetics

THOMAS J. KINDT

Cyclicul Production o f Antibody as a Regulatory Mechanism i n the Immune Response

WILLIAM 0. WIEGLE

Thymus-Independent B-Cell Induction a n d Paralysis

ANTONIO COUTINHO AND

GORAN MOLLER

SUBJECT INDEX

Volume 22

The Role of Antibodies i n the Rejection and Enhancement of Organ Allografts

CHARLES B. CARPENTER,

ABUL K. ABBAS

HARVEY R. COLTEN

STEPHEN C. GREBE AND

J. WAYNE STREILEIN

ANTHONY J. F. D'APICE, AND

Biosynthesis o f Complemeni

Graft-versus-Host Reactions: A Review

Cellular Aspects of Immunoglobulin A MICHAEL E. LAMM

YA. S. SHVARTSMAN AND

M. P. ZYKOV

Secretory Anti-Influenza Immunity

SUBJECT INDEX

Volume 23

Cellular Events i n the IgE Antibody Response

KIMISHIGE ISHIZAKA

Chemical a n d Biological Properties of Some Atopic Allergens

T. P. KING

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

B o DUPONT, JOHN A. HANSEN, AND EDMOND J. YUNIS

lmmunochemical Properties o f Glycolipids and Phospholipids

DONALD M. MARCUS AND

GERALD A. SCHWARTING

SUBJECT INDEX

Volume 24

The Alternative Pathway of Complement Activation

0. G ~ T Z E AND H. J. MULLER-EBERHARD

Membrane and Cytoplasmic Changes in B lymphocytes Induced by Liga ndCSurface Immunoglobulin Interaction

GEORGE F. SCHREINER AND

EMIL R. UNANUE

HOWARD B. DICKLER Lymphocyte Receptors for Immunoglobulin

Ionizing Radiation and the Immune Response

ROBERT E. ANDERSON AND

NOEL L. WARNER

SUBJECT INDEX

A ?

c 9 B B

D O E l F 2 6 3 H 4 1 5 J 6

advances in immunology volume 25

Documents