Advances in Immunology Volume 37

ADVANCES IN

Immunology VOLUME 37

CONTRIBUTORS TO THIS VOLUME

R. MICHAELBLAESE J. DONALDCAPXA NEILR. COOPER J. DIXON FRANK ROBERTC. GILES M . EDWARDMEDOF GORDOND. Ross ARCYRIOSN. THEOFILOPOULOS GIOVANNA TOSATO CHERYL A. WHITLOCK OWENN. WITTE

ADVISORY BOARD

K. FRANKAUSTEN LEROYE. HOOD JONATHON W. UHR

ADVANCES IN

Immunology EDITED BY FRANK J. DlXON Scripps C h i c and Research Foundation La Jolla, California

VOLUME 37

1985

ACADEMIC PRESS, INC. (Harcourt Brace Jovanovich, Publishers)

Orlando San Diego

New York

London

Toronto Montreal Sydney Tokyo

COPYRIGHT @ 1985, BY ACADEMIC PRESS. INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECIXONIC OR MECHANICAL, INCLUDING PHOTOCOPY. RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC. Orlmdo, Florida 32887

United Kin dom Edition ublished by

ACADEM~CPRESS

dc. (LONDON) LTD.

24-28 Oval R o d , London NWI 7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 6 1-1 70 5 7 ISBN 0 -12 -02 2 4 37 -2 PRINTED IN THE UNITED STATES OP AMERICA

85 86 87 88

9 8 7 6 5 4 3 2 I

CONTENTS

CONTHlMUTORS

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

X i

Structure, Function, and Genetics of Human Class I1 Molecules

ROMENT c. GlLES

AN11

J. DONALD CAPHA

............................... ............................... 111. HLA-DQ Serology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. HLA-DP SerologyiHTC . . . . . . . . . . . . . . . . . . . . . . . . . V. Deletion Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I. Introduction . . . . . . . . . . . . . .

11. HLA-DR SeroIogylHTC . . . . .

VI. VII. VIII. IX.

Monoclonal Antibodies to HLA-D Region Products HLA-DR Biochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HLA-DQ Biochemistry . . . . . . . . .............................. HLA-DP Biochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . X. Supertypic Specificity Localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI. Invariant (Gamma) Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

..................................................... ent Length Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIV. Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV. Conclusion . . . . .................................................. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............

1 5 9

23 31 43 47 50 53 60 63 65 65

The Complexity of Virus-Cell Interactions in Abelson Virus Infection

of Lymphoid and Other Hemotopoietic Cells C I I E R YA. L WHITLOCKA N D OWENN. WITTE

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Neoplastic Transformation by A-MuLV . . . . . . . . . ... 111. Nonneoplastic Changes Induced by v-abl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

IV. The Complexity of Abelson Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. The Complexity of A-MuLV Transformation in Vitro . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

74 79 84 87

95

Epstein-Earr Virus Infection and lmmunoregulation in M a n

CIOVANNA TOSATO A N D R. MICIIAEL BLAESE

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

99

11. Polyclonal B Cell Activation by EBV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Relationship between EBV-Induced Immunoglobulin Production

102

and Immortalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Infectious Mononucleosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

110 112

V

vi

CONTENTS

V. Immunoregulatory Cell Functions in Acute Infectious Mononucleosis VI . Persistent EBV Infection in Normal Individuals . . . . . . . . . . . VII . Selected Disorders Associated with an Abnormal Regulation

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

IX. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References

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

115 122

142 142

The Classical Complement Pathway: Activation and Regulation

of the First Complement Component

NEIL R . Coopeix I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1 . History of the Classical Pathway of the Comple

n ...............

111. The Proteins of the C1 Activation Unit IV . The Complexes of the C l Activation Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . The C1 Activation Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Actions of Activated C1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII . Regulations and Fate of Activated C1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VIII . Comment . . . . . . . . ..... References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

151 153 155 183 194 202 204 206 207

Membrane Complement Receptors Specific for Bound Fragments of C3

GORDOND . Ross

AND

M . EDWARDMEDOF

I . Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Generation of the Ligands for C3 Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Structure and Binding Site Characteristics of the Receptors

.

.........

V . Conclusions . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

217 221 230

261

Murine Models of Systemic Lupus Erythematosus

ARGYHIOSN . THEOFILOYOULOS A N D FAANK J . DIXON

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I1. Derivation of Lupus Mice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111. Natural History and Pathology of L u p ~ sMice . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV . Cellular Abnormalities ....................... .. V . Genetics of Murine SLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Influence of Sex and Sex Hormones on the Pathogenesis of Murine SLE . . . . . VII . Viruses in Murine SLE . . . . . . . . . ........ VIII . Treatment of Murine SLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

269 271 273 301 336 346 348 350

CONTENTS

vii

IX . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

355 358

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

391 385

CONTENTS OF RECENT VOLUMES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

This Page Intentionally Left Blank

CONTRIBUTORS

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

R. MICHAELBLAESE(99), Metabolism Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 J. DONALDCArnA (I), Department of Microbiology, University of Texas Health Science Center at Dallas, Dallas, Texas 75235 NEIL R. CoorEn (151), Department of Irninunology, Scripps Clinic and Research Foundation, Ln Jolla, California 92037

FRANK J. DIXON(269), Department of Immunology, Research Institute of Scripps Clinic, La ] o h , California 920*37 RonEnT C. GILES(l),Department of Microbiology, University of Texas Health Science Center at Dallas, Dallas, Texas 75235 M . EDWARD MEDOF (217), Departments of Pathology and Medicine, New York University Medicul Center, New York, New York 10016 GonDoN D . Ross (217), Division of Rheuinatology-linmunology, Department of Medicine, and the Department of Microbiology-Immunology, University of North Carolina, Chapel Hill, North Carolina 27514 AnGYnIos N . THEOFILOPOULOS (269), Department of Immunology, Re-

search Institute of Scripps Clinic, La Jolla, California 92037 GIOVANNA TOSATO(99), Metabolism Branch, National Cancer Institute, Nationd Institutes of Health, Bethesda, Maryland 20205 A. WHITLOCK (73), Department of Pathology, Stanford University, Pah Alto, C a l i j h i a 94305

CHEnYL

OWENN . WITTE(73),Departinent of Microbiology and Molecular Biology Institute, University of Cal$ornin, Los Angeles, Los Angeles, Cnl$ornia 90024

ix

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY. VOL 07

Structure, Function, and Genetics

of Human Class II Molecules ROBERT C. GILES AND J. DONALD CAPRA Department of Microbiology, University of Texas Health Science Center at Dallas, Dallas, Texas

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IILA-DR Serology/HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HLA-DQ Serology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HLA-DP Serology/HTC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Deletion Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Monoclonal Antibodies to HLA-D Kegion Prodiicts . . . . . . . . . . . . . . . . . . . . . . . . . HLA-DR Biochemistry ..... .......... ......... HLA-DQ Biochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HLA-DP Biochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Supertypic Specificity Localization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Invariant (Gamma) Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Restriction Fragment Length Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fiinction . . . . . ..... .. .... Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11. 111. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. XV.

1

5 9 14 16 18 23 31 43 47 50 53 60 63 65 66

1. Introduction

The human major histocompatibility complex (MHC) or human leukocyte antigen (HLA) complex is located on the short arm of chromosome 6. Molecules encoded within the H L A complex have been implicated in the regulation of T cell and B cell differentiation, and in the ability of the host to mount a humoral andlor cell-mediated response against a myriad of antigens. Additionally, these molecules are thought to be involved in immunologic communication and cell-cell interactions that maintain the integrity of the immunologic system of an individual, including the ability to distinguish self from nonself. At least three classes of molecules are controlled by the H L A region, each functioning in a distinct way to perform immunologic tasks. The class I molecules, HLA-A, -B, and -C, are the classic transplantation antigens. These molecules are responsible for graft rejection and regulate the killing of virus-infected cells. They are composed of two subunits, a 44,000-Da heavy chain and a noncovalently associated 11,500-Da light chain known as Pzmicroglobulin. The heavy chain is an intrinsic membrane glycoprotein which is H L A encoded and structurally polymorphic. P,-Microglol>ulin is an extrin1 Copyright 0 1985 by Academic Preas, Inc. ~ lrlglltn l of rrproductiou in any form reserved. ISBN 0-12-022437-2

2

ROBERT C . GILES A N D J. DONALD CAPRA

sic, nonpolymorphic protein encoded on chromosome 15. The class I molecules are expressed on all cell types except red blood cells. The class I1 molecules, the HLA-D region antigens, are also composed of two subunits but unlike the class I molecules, both subunits are instrinsic membrane proteins and both subunits are encoded within in the MHC. These subunits are noncovalently associated and consist of a heavy or alpha chain of molecular weight 34,000 and a light or @-chainof molecular weight 29,000 (see Fig. 1).The fact that both subunits of the class I1 molecule are encoded within the major histocompatibility complex is somewhat unusual. In most other cases in which multimeric proteins are made u p of dqferent subunits, the genes encoding the separate polypeptide chains are unlinked. Exceptions such as insulin and C4 (the fourth component of complementsee below) arise when a single large precursor polypeptide (polyprotein) is processed to form two or more subunits. The evidence is overwhelming that this form of processing does not occur for the class I1 antigens. Class I1 molecules are involved in mediating mixed lymphocyte reactions (MLR)and communicating between lymphoid cells. While class I molecules have a ubiquitious tissue distribution, class 11 molecules are expressed most abundantly by B lymphocytes, activated T lymphocytes, and antigen-presenting cells including peripheral blood monocytes, macrophages, Langerhans’ cells, and dendritic cells of the lymphoid organs. The early complement components encoded within or near the MHC are referred to as class I11 molecules. Collectively, they represent the C3 converting enzymes of the classical (C2 and C4) and alternative (Factor B)

PLASMA

MEMBRANE a

P

FIG.1. General structure of human class I1 molecules. S-S indicates disulfide bridges. The carbohydrate side chains are depicted as well as a possible phosphorylation site. Adapted from Shackelford et al. (1982).

3

H U M A N CLASS I1 MOLECULES

GLO

0 ’

---

DP I

.. DQI

CLASS

DR I

II

C 2 Bf C4924-0HC4A 21-OH B I

I

I

I

CLASS llI

I

I

I

C

A(Qo,TL)

I..I

CLASS

I

I

Frc. 2. Genetic map of the hunian major histocompatibility complex. The precise order of some of these loci is not known. The centrornere is located to the left. See text for details.

complement pathways (Carroll et al., 1984). These serum proteins participate in cell lysis and mediate inflammatory responses. Initially, it was thought that the class I11 molecules were not as polymorphic as the class I and I1 molecules but additional information indicates that they are at least as polymorphic. In addition, two genes encoding adrenal cytochrome P-450 enzymes, specific for steroid 81-hydroxylation (21-OH) have recently been located within the “class 111 region” (White et al., 1984) (see Fig. 2). In humans, genetic defects in many of the steps of cortisol biosynthesis have been described although only in deficiency of cholesterol side-chain cleavage activity has a defective or deficient P-450 been documented. Of these inborn errors of metabolism, 21-hydroxylase deficiency is by far the most common, occuring in about 1/5000 individuals. It is inherited as a monogenic autosome recessive trait linked to the HLA gene complex. It is likely that this human disease is due to deficiencies or defects of one or both of these two 21-OH genes. Whether there are additional molecules encoded within the major histocompatibility complex remains to be seen although certainly there are suggestions of several others. Figure 2 depicts our current concept of the genetic organization of the human major histocompatibility complex. The class I1 region, the HLA-D region has been divided into three “subregions” and, for simplicity, the subregions within the HLA-D region are shown to encode only a single molecule. However, in all instances, at least two polypeptide chains (an a and a p) are encoded within the subregion and in most instances multiple polypeptide chains are encoded within each subregion of HLA-D. The number of these genes that are actually expressed is still a subject of controversy. The two “x’s’’ in the figure indicate areas which may be recombinational hotspots as many of the recombinations that have been observed both within populations and families occur near these two points (Bodmer, 1984). Nomenclature has been and, at least for the near future, will continue to be a major problem in this area. Recently, at the Ninth International Histocompatibility Workshop, the nomenclature within the HLA-D region was changed considerably. This new nomenclature will be adopted and used throughout this review. While we will explain this in detail later, suffice it to

4

ROBERT C. GILES A N D J. DONALD CAPRA

say that the HLA-D region is now said to be comprised of three subregions referred to as HLA-DP (previously SB), HLA-DQ (previously DS, DC, MB), and HLA-DR (no change). The first hint of what is now known as the human analogs of the murine Ia antigens came in the early 1970s when Yunis and Amos (1971)demonstrated that the antigens responsible for the proliferation of human lymphocytes in the mixed lymphocyte culture were encoded by a locus closely linked to HLA-A, 43, and - C . This mixed lymphocyte culture locus is now known as HLA-D. The serologic characterization of human Ia-like antigens was more difficult than in the mouse system. It was first described by van Rood and coworkers (1975) and depended on the use of selected antisera that inhibited the mixed lymphocyte reaction of cells from HLA-A, - B , and C identical persons. Polymorphism is a hallmark of the molecules encoded by the MHC. Class I, class 11, and class 111molecules all exhibit a high degree of polymorphism. The nature and mechanisms of this diversity are likely to be comprehended rather quickly as our knowledge of their primary structures, both at the DNA and protein levels, is increased. A crucial question concerning this system is, “What was the evolutionary pressure that led to and maintained this polymorphism in the population?” In addition, the close association of some alleles of the HLA-A, -B, -C, and -D genes with human diseases has made this field an especially interesting one. However the precise role(s) of this polymorphism may be somewhat more ellusive until some of the questions regarding the exact immunologic (or other) functions of these MHCencoded molecules have been answered. As our understanding of major histocompatibility complex-controlled immune responsiveness broadens and hybridoma and gene cloning technology advances, specific enhancement of desired immume responses and suppression of deleterious ones will most likely become possible. The use of “state of the art” molecular biological techniques and the advent of highly discriminatory monoclonal antibodies have contributed to an explosion of information regarding the structure and function of this important family of molecules (Krangel et al., 1980; Steinnietz and Hood, 1983; Kaufman et al., 1984). This review will explore the complexity of the HLA-D region, with emphasis on the genetic organization, structure, and function of molecules encoded in this region. Other reviews detailing more of the history, serology, distribution, and function of the DR and Ia antigens can be found in the literature (Ferrone st al., 1978; Bodmer, 1981; Shackelford et al., 1982; Gorzynski and David, 1983; Auffray et al., 1983), and two recent reviews have appeared in Advances in Immunology (Winchester and Kunkel, 1979; Gonwa et al., 1983).

H U M A N CLASS I1 MOLECULES

5

II. HLA-DR Serology/HTC

This review was written immediately after the completion of the Ninth International Histocompatibility Workshop held in Munich, Germany in May 1984. At that meeting, none of the HLA-Dw specificities was upgraded to full status but many new workshop designations were developed. While our understanding of the genetics of class I1 molecules largely comes from the application of classical serology-especially alloantisera, a major development has been the discovery of monoclonal antibodies with polymorphic specificities for HLA-DR products (Hansen et al., 1981; Pierres et al., 1981; Accolla and Pierres, 1983) including some that appear specific for the products of a single allele (Radka et al., 1983). Excellent reviews of the serology, particularly the historical notions of the serology of the HLA-D region and its relationship to DR are available (Winchester and Kunkel, 1979; Ferrone et al., 1978). While some new techniques have improved and simplified HLADR typing (Gruinet et al., 1983) we will only briefly describe some of the new developments particulary as they relate to the structural aspects which will follow. All known human class I1 antigens are encoded in the genetic region centromeric to HLA-B and telomeric to the locus controlling the red cell antigen glyoxylase, GLO. This region is often loosely referred to as the “D” region. Recently, as is evident from Fig. 2, three separate subregions of the HLA-D region have been defined. Historically the first phenotypic trait shown to be controlled by an HLA gene centromeric to HLA-B was its capacity to stimulate strong MLR in uitro. The trait was assumed to result from the product of a single gene designated HLA-D. However, at the present time, the exact contribution of the three separate subregions to the MLR is a matter of controversy. Indeed, inany of these distinctions are just being worked out now that specific alloantisera and monoclonal antibodies are available for each of the subregions. It is likely, however, that the major MLR reactivity is due to disparity at the HLA-DR subregion and the majority of homozygous typing cell (HTC) reactivity will be subject to genes under the control of the HLA-DR subregion. Because it has not yet been unambiguously demonstrated which of the various subregions contributes to most MLR reactivities, the Ninth International Workshop Nomenclature Committee chose not to upgrade the Dw specificities to full status. In addition, (as will be detailed in the next section), “w” has been added to all DQ and DP specificities. The designation HLA-Dw with no further indication of subdivisions, is retained for the MLC defined specifities which have yet to be mapped to a subregion. Tables I to IV summarize our present understanding of the various

6

ROBERT C. GILES AND J. DONALD CAPKA

TABLE I GENE FHEIJLIENCIES FOH HLA-DR” HLA-

Caucasians

Orientals

Negroids

DR1 DR2 DR3 DR4 DR7 DRw8 DRw9 DRwlO DRwll DRwl2 DRwl3 DRw14 DR X

9.5 15.8 12.0 12.7 12.0 3.0 0.8 0.8 12.3 2.0 5.4 5.8 7.9

5.0 15.1 1.8 21.8 2.9 7.3 11.5 0.5 7.2 2.9 6.8 13.2

5.1 15.1 14.9 7.6 13.2 0.8 1.5 2.3 16.5 3.4 3.8 10.7 5.3

Sample sizeb

1926

752

263

4.0

“ From Baur, M. P., Neugebauer, M., Deppe, H., Sigmund, M., Mayr, W. R., Albert, E. D., in “Histocornpatibility Testing 1984.” Number of haplotypes counted. serologic specificities extant in the HLA-DR region. Table I lists the Dw and gene frequencies for HLA-DR reported at the Ninth Workshop. Table I1 indicates new provisional designations for HLA-DH specificities. Note, parTABLE I1 PROVISIONAL DESIGNATIONS FOH HLA-DR SPECIFICITIES~

NEW

~

~~~

New

Previous equivalents

DRwll DRwl2 DRwl3 DRwl4 DRw52 DRw53

LB5 LB5x8, DR5 short, FT23 6.6, 6.1, 6Z 6.9, 6.3, 6X, 901 MT2 MT3

” These

designations (like those in Tables

111 and IV) are derived from the Nomen-

clature Committee of the Ninth Workshop and will be published in “Histocompatibility Testing 1984” (E. D. Alpert, M . P. Baur, and W. R. Mayr, eds.), Springer-Verlag, Berlin and New York, 1984.

H U M A N CLASS I1 MOLECULES

7

TABLE 111 NEWPROVISIONAL DESIGNATIONS FOR HLA-Dw SPECIFICITIES New

Previoiis eqiiivalents

Dw13 Dw14 Dw15 Dw16 Dw17 Dw18 DwlY

UB3 LD40 DYT, YT DB8, B8 7A, (IIw7A) 6A, (DwGA) 6B, (Dw6B)

ticularly, that DR5 has been “split” into DHwll and DRwl2; arid that DRw6 has been similarly split into DKwl3 and DHw14. It will not always be possible in this review to use this newer nomenclature as often the typing which might have revealed the “split” was not done. Table 111 lists new provisional designations for HLA-Dw specificities and their equivalents and, finally, Table IV indicates our present understanding of Dw and DH relationships. At the present time there are still approximately 8% HLA-DR blanks in the Caucasian population and, therefore, it is likely that new specificities exist. Based, however, on what is known at the present time the following compilation of DR specificities is proposed. The DH1 specificity exists on a DR or I-E-like molecule and there is no evidence that there are splits of this specificity. DR2 is split into at least two and possibly a third serologic grouping (Kasahara et al., 1983).There are two separate Dw specificities that correlate with DH2, Dw2, and Dw12 (see Table IV). There is an additional third

Dw

AND

TABLE IV DR RELATIONSHIPS

Dw specificities Dwl Dw2, Dw3 Dw4, Dw5 Dw6, Dw9, Dw7, DW8

Dw12 DwlO, Dw13, Dw14, Dw15 Dw18, Dwl9 Dw16 D w l l , Dw17

Associated DR specificities

DR1 DR2 D R3 DR4 DRwll (5) DRwl3 (w6) DRwl4 (w6) DR7 DRw8

8

ROBERT C. GILES AND J. DONALD CAPRA

grouping that may exist. By restriction fragment length polymorphism (see below), there may be as many as three additional splits. At the present time, however, firm evidence exists for only two splits of DR2. There is no evidence at the present time for splits of DR3 and essentially all DR3 specificities are associated with Dw3. DR4 is extremely complicated and is conveniently divided serologically into a minimum of three separate groupings that have been referred to as 4.1, 4.2, and 4.3. 4.1 includes the LD40 and Dw4 groups. It may include DYT as well. 4.2 includes the DB3 group but includes other specificities as well. The 4.3 group includes the DwlO group. By HTCs there is a minimum of five splits of DR4, Dw4, DwlO, Dw13, Dw14, Dw15 (see Table IV). There is biochemical evidence from several laboratories that different DR4 specificities can be explained by variations in DR (I-E-like) p-chains. The biochemical evidence for this will be presented later but suffice it to say that by both isoelectric focusing in one dimension, by two-dimensional gel electrophoresis and by restriction fragment length polymorphism several groups have demonstrated biochemical variation among DR4 genes and/or their products which correlate with one or more serologic or HTC splits. As mentioned above, DR5 has been split into DRwl1 and DRwl2. DRwll associates with Dw5 and DRwl2 associates with DB6. DRw6 is extremely complicated. It has been split serologically into a minimum of two groups previously referred to as 6.6 and 6.9 as shown in Table 11. The 6.6 and 6.9 terminology has been changed to Dw13 and Dw14 (other equivalents are also noted in Table 11). At the recent Ninth Workshop, evidence was presented utilizing two-dimensional gels that there were four separate biochemical patterns among DRw6 individuals. The alloantisera which define Dw13 and Dw14 correlated with two of these patterns. At the present time, there are no obvious serologic splits of DR7, however by HTCs, DR7s can be divided into two separate groups so it is likely that a minimum of two separate groups of DR7’s exist. DRw8, like DRw6, still does not have full Workshop designation. However, no evidence exists at the present time for serologic splits, although by HTCs there is evidence for a minimum of three separate groups of DRw8 (Mickelson et al., 1983). Despite some attempts DRw9 has not been split serologically or with HTCs. DRwlO seems homogeneous serologically with only a single group. However this has not been studied extensively and there may well be additional splits. Thus DR1, 3, 7, w8, w9, w10, w l l , w12, w13, and w14 presently exist as single serologic entities; DRw9 as two; and DR2 and 4 as three. This makes a

H U M A N CLASS 11 MOLECULES

9

total of 15 serologic specificities that are likely to achieve status as DR alleles. Obviously a goal of modern HLA genetics is to understand the structural basis for each of these specificities and while much of this information is available (see below) we are still ignorant on a number of different issues. Most of these specificities are likely (although not exclusively) encoded in the HLA-DR subregion which encodes a single a-chain (nonpolymorphic) and probably three @-chains(see below). It has been exceedingly difficult to assign specific (3-chain gene products to these polymorphisms. However it seems quite clear that 1-E-like (3-chains are the polymorphic component of class 11 antigens as no variation in a-chains has been observed to explain these specificities. Whether differences in reactivity are explicable based on different p-chain gene products is still a matter of controversy. Finally, since approximately 10% of the population are officially listed as “DR blank’ it is likely that many additional specificities will be discovered. Recently, Wallin et al. (1984) have described by restriction fragment length polymorphism three additional groupings by examining individuals who are hornozygous DR blank/blank. It is likely, with the rapid advances being made utilizing this approach to HLA-DR typing that these HLA-DR blank phenotypes will be dissected in the near future. Thus, the situation in the human HLA-DR subregion is quite reminiscent of the mouse Z-E subregion. While the bulk of evidence suggests that the human DR a-chain is nonpolymorphic, the murine Z-E a-chain is modestly polymorphic. This has been seen both serologically, biochemically, and recently at the DNA level. However, all would agree that the bulk of the polymorphism in both the inurine Z-E subregion and the human HLA-DR subregion relates to the beta chain. in man, it is clear that a minimum of two and probably three (3-chains is expressed. In the mouse, there is some evidence for more than one expressed p-chain. In the genome there are clearly additional I-E-like p-chain genes but evidence for their expression is not definitive (Steinmetz and Hood, 1983). 111. HLA-DQ Serology

As discussed previously many of the early studies which revealed the complexity of the HLA-D region were the result of serological analyses at the cellular level of alloantigenic specificities found on HLA-D region products. By the Seventh Histocompability Workshop not only had clusters of alloantisera defining the HLA-D alleles been described, additional clusters of alloantisera demonstrated defined patterns of cross-reactivity, with each cluster of antisera encompassing two or more of the HLA-DR alleles. By the

10

ROBERT C. CILES AND J . DONALD CAPRA

Eighth Workshop, additional clusters of cross-reactive alloantisera had been described. Due to their ability to recognize several HLA-DR allelic products, these antisera were classified as recognizing “supertypic specificities. ’’ Those recognized to date include the MB, MT, DC, LB, BR and Te series (see Table V). Each of these specificities is found in linkage disequilibrium (associated strongly) with a number of DR specificities. For example, in the MB series, M B l is associated with DH1,2, w6, w8, and w10; MB2 with DR3 and 7; and MB3 with DR4, 5 and w9. Members of a given supertypic series were thought to represent allelic products since they segregate from one another in family studies and, particularly in the case of the MB series, are in Hardy-Weinberg equilibrium. Until recently the relationship between these independently defined supertypic series has not been clear. For example, on the one hand, many MB, DC, LB, and Te specificities appear to have similar, if not identical, DR-associated distributions (i.e., MB1, DC1-see Table V) suggesting that they may represent the same segregant series. On the other hand, the M B and MT series (with the exception of MB1 and MT1) have very different DR-associated distributions and probably represent different series. With this in mind a simplified scheme relating the supertypic specificities to each other is presented in Table V (taken from Hurley et al., 1983a). In this formulation, specificities that have similar distributions are assumed to be identical; slight differences observed by different laboratories in the association of the supertypic specificities with DR haplotypes and the degree of linkage disequilibrium are assumed to be differences in alloantisera used by these groups. In this way, the multiple supertypic series can be condensed into two series of alloantigenic specificities, MB and MT. The DC, LB, and Te series are similar to the M B series, and the BR series is similar to the MT series. One of the more controversial questions involving HLA-D region molecules over the last 5 years has centered around the highly complex patterns of reactivity of these alloantisera recognizing the supertypic specificities. At issue has been the question of the molecular bases for these patterns of TABLE V SUPEKNPICSERIES

MB series Associated

DR specificities MB1 MB2 MB3

DRI, 2, w6, w8, w10 DR3, 7 DR4, 5, w9

MT series Equivalent specificities

DCI, Te21, MT1 DC3, Te24 DC4, Te22, MT4

MT1 MT2 MT3

Associated DR specificities

Equivalent specificities

DR1, 2, wG, w10 DR3, 5, w6, w8 DR4, 7, w9

DC1, Te21, MB1 BR3 BR4

HUMAN CLASS I1 MOLECULES

11

“cross-reactivity” as seen by these alloantisera and more recently by monoclonal antibodies. Controversy has revolved around whether these supertypic specificities represent shared (public) determinants on the associated HLA-DR molecules or specificities in linkage disequilibrium present on molecules distinct from DR. The first real hreakthrough occurred when Tosi et al. (1978) described the specificity DC1 which they argued was on a molecule distinct from DR. This specificity was originally identified by radioimmunoassay using the cell line Daudi (DRw6, blank). By performing both quantitative and qualitative binding studies of a large number of alloantisera on an L251-labeledclass I1 preparation from Daudi, they were able to demonstrate two well-defined class I1 subsets; one subset corresponded to the DRw6-bearing antigens, while the other subset defined the DC1-bearing antigens. This was the first real evidence suggesting the existence of a second HLA-D region locus, although the structural basis of the DC1-bearing molecule was not determined for several years and the crucial nature of the discovery was not widely appreciated at the time (Tosi et al., 1982, 1984see HLA-DQ Biochemistry). Shortly before the Eighth Histocompatibility Workshop, Duquesnoy et al. (1979) defined serologically the MB supertypic specificities. Their data demonstrated that anti-MB1 activity could be removed by absorption with either DR1 or DR2 positive cells, but it could not be lysostripped by antiDR1 or anti-DR2 antisera. During the Eighth Workshop a new polymorphic B cell system, the MT (“multi-specific”) system, was defined (Park et al., 1980). Attempts to understand and reconcile the MT system with MB (DC) led to much of the confusion concerning the second locus and the molecular localization of the supertypic specificities. The studies with DC1 and the MB specificities strongly argued for a second locus, but because DC1, MB1, and MT1 were assumed to be identical specificities (see Table V) and because many believed that the MT series represented an allelic series, the data regarding MB (DC) vs MT became almost impossible to reconcile. Not until studies on the biochemical structure of the second locus product were performed by several laboratories (especially using monoclonal antibodies-see below) did it become apparent that MB (DC) was a specificity on a molecule distinct from HLA-DR. Even with this revelation regarding M B (DC), the molecular basis for the MT (BR) supertypic specificity was not apparent. The questions and some of the answers generated in this debate are dealt with in Section X of this review. The second locus as defined by Tosi et al. (1978) and Duquesnoy et al. (1979) has been called by a variety of names since its discovery. Its oldest nomenclature and probably the most widely accepted was the DC terminology (named for Dora Centis who was instrumental in its discovery in the Tosi laboratory). Each specificity was designated as DC followed by the

12

ROBERT C. GILES AND J. DONALD CAPRA

number of the first DR specificity in strong linkage disequilibrium. For example, DC4 was the specificity controlled by the DC locus in linkage disequilibrium with DR4 and DR5. In some laboratories the locus was referred to as M B (for “More B” or “Milwaukee Blood’) or LB (for “Leiden B”) and in still other laboratories as DS (for “Secondary D”) (Goyert et al., 1982). In fact as it became clear that the molecules bearing the DC (MB) specificities showed considerable variation among the DR alleles (Goyert and Silver, 1983; Giles et al., 1984a-c), the DS 1-7 terminology seemed the most reasonable. The recent HLA nomenclature committee met following the Ninth Workshop and the designation “HLA-DQ” has been adopted for the subregion. Specificities associated with this locus have been given the designation HLA-DQ followed by “w” to indicate the designation is provisional and appropriate numbers, usually in sequence (i.e., HLA-DQwl, etc.) (see Table VI). As several biochemical studies have demonstrated that both the DQ CY and p loci are polymorphic (Giles et al., 1984c), the question arises as to why DQ allele-specific alloantisera have not been readily available. One plausible explanation is that any such alloantisera which do exist would have been designated as “DR-specific alloantisera.” In fact, recent evidence suggests that such alloantisera do exist, either as solely DQ-specific sera or in conjunction with DR allele-specific alloantisera (Stastny et al., 1984). Using fluorescence inhibition with monoclonal antibodies and lysostrip experiments, several alloantisera which recognize “private” specificities on DQ molecules (refered to by the authors as “DS private” or “DSP” specificities) have been described. Thus, in addition to the broad supertypic specificities DQwl (MBl), DQw2 (MB2), and DQw3 (MB3), the DQ molecules also carry allele-specific specificities. Additional evidence for these “splits” of the DQ alleles will be discussed in more detail in the section describing HLADQ biochemistry. The role of the HLA-DQ products in primary and secondary MLR has been difficult to assess due to the extremely high linkage disequilibrium with TABLE VI NEWPROVISIONAL DESIGNATIONS FOH HLA-DQ SPECIFICITIES

New

Previous equivalents

DQwl DQwZ DQw3 DQx

MB1, DCl, MTL, LB-El2, Te21 MB2, DC3, Te24, LB-El7 MB3, DC4, MT4, Te22

~

nCaucasians (sample size: 2016).

Gene frequencies0

~~~~~~

32.3 18.1 23.3 26.3

H U M A N CLASS I1 MOLECULES

13

HLA-DR. However, at the Ninth Workshop a number of attempts were made to dissect the contribution of HLA-DK, HLA-DQ, and HLA-DP to the MLR. These studies are possible now that “subregion specific” monoclonal antibodies and alloantisera can be defined. For example, Stastny’s group (Stastny et al., 1984) demonstrated that many antisera that are referred to as DR typing reagents actually have specificities for the HLA-DQ subregion products. These specificities are not MB (DC) supertypic specificities for if they were the antisera from the beginning would have been considered MB rather than DR typing sera. However, as will be detailed below, it appears that the product of each DQ allele is biochemically distinct from every other. Thus, it is not surprising that many antisera have now been found that recognize, for example, the DQ molecule which is generally associated with DR4 and not the DQ molecule which is generally associated with DR5. By careful dissection of antisera that, for example, recognize the DQ molecule associated most commonly with DR4 but not the DR molecule in DR4 cell lines, Stastny’s group has been able to demonstrate specific contributions of DR subregion products versus DQ subregion products in several systems. These are largely done by blocking reactivities in the MLR with highly defined alloantisera or monoclonal antibodies specific for the products of each of the subregions. By this means the generalization which has emerged is that the vast majority of reactivities are, indeed, due to the HLA-DR subregion product disparities but that in several specific instances, the DQ subregion products can contribute significantly and, indeed, be the sole reactors in an MLR. Similar circumstances pertain for antigen presentation and other reactivities that had previously been referred to as “HLA-D.” In general SB gene products (now termed DP) do not appear to significantly contribute to these reactivities. However, in certain rare instances clear reactivity due to DP differences can be correlated with a primary MLR. One final note on nomenclature needs to be addressed; that is, the obviously awkward description of the “DQ molecule generally associated with DR4” will eventually need a name. At the present time, the nomenclature committee has designated such a molecule as a molecule bearing the DQw3 specificity but, the DQ molecule associated with DR5 cells will likely bear this same DQw3 specificity. Since specificities are being described at the present time that distinguish between two DQw3 molecules (the DQ molecule generally associated with DR4 and that generally associated with DR5), additional terminology is needed. A convenient approach would be to identify such specificities and molecules after a decimal indicating the DR specificity which is generally associated with that DQw specificity. Thus, for example, in the illustration above, the DQ specificity and molecule which is generally associated with DR4 cells, and which bears the DQw3 specificity would be referred to as DQw3.4; while the DQ molecule generally associ-

14

ROBERT C. GILES AND J. DONALD CAPKA

ated with DR5, and which also bears the DQw3 specificity, would be referred to as DQw3.5. Similarly, the DQ molecule associated with a DR1 cell line would be referred to as DQwl. 1, while the DQ molecule and specificity which is generally associated with DR2 cells (and which also contains the DQwl specificity) would be referred to as DQwl.2. As will be appreciated later, each of these molecules (DQwl.l, DQwl.2, DQw1.6, DQw2.3, DQw2.7, DQw3.4, DQw3.5, DQw3.9) appears to be distinct biochemically and as alluded to above, serologically at least in some circumstances. A full dissection of the contributions of these specific DQ molecules to the various reactivities, functions, and indeed disease associations presently attributed to the HLA-D region (and by implication HLA-DR) is an active area of investigation that should prove fruithl in the near future. IV. HLA-DP Serology/HTC

The technique known as the primed lymphocyte test (PLT) has been used to identify an additional series of polymorphic HLA-D region antigens recently referred to as HLA-DP (formerly known as SB or FA). In this technique, responder lymphocytes are selectively sensitized to stimulator lymphocytes which differ from the responder by a few, or a single, stimulatory determinant(s). Such sensitized cells are then restimulated and a “secondary MLR” is performed. Shaw used this technique to demonstrate a secondary B (SB) cell system, HLA-DP, which has been shown to consist of at least six alleles (Shaw et al., 1980, 1982; Pawelec et al., 1982a-c, 1983)(see Table VII for the most recent provisional designations). A recombination between HLA-DR and HLA-DP and two recombinations between HLA-DR and the enzyme marker GU)l mapped the HLA-DP locus centromeric to HLA-DR and telomeric to GLOl (Shaw et al., 1981). Kavathas et al. (1981)also demonstrated that HLA-DP was a distinct locus using HLA deletion mutants. From TABLE VII NEW PROVISIONAL DESIGNATIONS FOR HLA-DP SPECIFICITIES

New

Previous equivalents

DPwl DPw2 DPw3 DPw4 DPw5 DPw6 DPx

SB1, PL3A 5132 sn3 SB4, PL3B SB5 SB6

UCaucasians.

Gene frequenciesa 4.3 11.5 4.0 41.8 4.4 0 40.0

H U M A N CLASS I1 MOLECULES

15

a series of y-ray induced mutant B cell lines (see Section V) which had lost the expression of HLA-DR, -A, and -B, they found two mutants that were not capable of inducing a DR specific secondary stimulation, but were capable of HLA-DP specific stimulation. Attempts to define HLA-DP products serologically have been relatively difficult. DP is apparently a poor immunogen compared to DR and DQ, and alloantisera defining DP are rare. In one study evidence for serological recognition of DP determinants was presented using a two-color fluorescence technique in which a cluster of alloantisera had a pattern of reactivity which correlated almost perfectly with the presence or absence of a particular DP allele (van Leeuwen et uE., 1982). Monoclonal antibodies with polymorphic specificity for DP have also been rather scarce. One DP specific antibody, I-LR1 (Nadler et ul., 1981), has been shown to react with certain DP allelic products (HLA-DPw2, w3, and some DPw4 molecules). The initial biochemical characterization of an HLA-DP molecule was performed using this monoclonal antibody and will be discussed in Section IX. More recently another monoclonal, B7/21, has been described which also recognizes the HLA-DP product (Royston et ul., 1981). Unlike I-LRI, B7/21 is apparently monomorphic, reacting with the product of all known HLA-DP alleles. The antibody, also known as anti-FA, is specific for an antigen which maps identically to D P in H L A haplotype loss mutants (see Section V). Most convincingly, the anti-FA antibody has been used successfully to recognize L cells which have been transfected with the HLA-DP a and p genes (W. F. Bodmer, personal communication). Several other monoclonal antibodies have been described that react with HLA-DP gene products in addition to reacting with either DQ or DR molecules. These will be described in more detail in Section VI. As will be described under HLA-DP biochemistry (Section IX), there is evidence that two HLA-DP genes exist (2 DP CY genes and 2 DP p genes). It has not yet been determined whether both of the genes are expressed and whether any or all of these monoclonal antibodies react with one or both of these gene products. Additionally, if both gene products are expressed, it is not known whether one or both are polymorphic. Although D P is apparently not strongly immunogenic its discovery and subsequent genetic localization was made possible by the relative lack of linkage disequilibrium between DH and DP. Termijtelen et ul. (1983a,b) have examined three HTCs originating from the offspring of first cousin marriages that were heterozygous for DP, but homozygous for their HLA-A -B, -C, and -DR antigens. These HTCs belonged to a group of 15 additional HTCs all originating from first cousin marriages. From an analysis of these HTCs the mieotic distance between the HLA-DP and HLA-DR loci was estimated to be between 1 to 3 cM.

16

ROBERT C. GILES A N D J. DONALD CAPRA

V. Deletion Mutants

The use of somatic cell genetics for the elucidation of the complexity of the HLA-D region resulted in the generation of HLA-deletion cell lines by mutagenesis of lymphoblastoid B cell lines. This technique, originally developed by Pious et al. (1973) and utilized extensively by Kavathas et al. (1980a,b), has provided a powerful approach for the analysis of HLA-D region molecules and their genes. An example of the general technique is drawn from the work in DeMars laboratory. HLA-deletion mutants were produced by treatment of a human

FIG. 3. The derivation of DR-null mutants from human lymphoblastoid cell line LCL 721. Mutations were induced with y-rays. Selections for HLA-antigen loss mutants were imposed with an anti-A2 monoclonal antibody, and anti-B8 alloantiserum and an anti-DR inonoclonal antibody. From DeMars et al. (1983).Reprinted by permission of Elsevier Science Publishing Company, Inc.

HUMAN CLASS I1 MOLECULES

17

FIG.4. Pair of normal appearing No. 6 chroinosoriies (A) and pair of No. 6 chromosomes (B) from a variant that had lost expression ofall HLA antigens encoded tor b y one haplotype and the cis-linked glyoxalase I allele Clo, One of the chromosomes from the variant has an interstitial deletion in 6p. From Kavathas et at. (198011). Reprinted by permission of Elsevier Science Publishing Company, Inc.

heterozygous lymphoblastoid cell line (LCL-72l;-DH1/3,-B5/8,-A2/1) with y radiation and selection with HLA antiserum (see Fig. 3).A large percentage of mutant cell lines resulted from this treatment and after selecting and cloning, several variants found to be expressing the product ofonly a single allele were said to be “hemizygous” (DeMars et al., 1983). Figure 4 illustrates a portion of the karyotype of the cell line mutagenized by Kavathas et al. (1980a,b). This variant lost expression of cis-linked HLAA l , B 8 , and DR3 alleles after y irradiation. The interstitial deletion in 6P of one number 6 chromosome is evident, while the rest of the karyotype was normal. Already these hemizygous cell lines, and additional subregion-loss mutants derived from these lines, have been used to glean several important insights into the genetics of class I1 molecules. As mentioned in Section IV, derivatives of these hemizygous mutants allowed the genetic mapping of the HLA-DP subregion centromeric to HLA-DR and -DQ and telomeric to GLO on the short arm of the sixth chromosome (Kavathas et al., 1981). Several laboratories have used these cell lines to correlate available class I1 DNA probes with serologic specificities. These studies are done by Southern filter hybridization of DNA isolated from these cell lines. These deletion mutants have also been extremely useful in identifying the specificities of monoclonal antibodies to the HLA-D region (Section V I ) . Production of a wider variety of deletion mutants will be invaluable in studying control of expression of HLA-D region antigens. It has already been possible to distinguish deletion mutants from regulatory mutants and thereby allow for precise dissection of the events leading to expression of specific HLA-D region molecules (Accolla, 1983).

18

ROBERT C. GILES A N D J. D O N A L D CAPRA

The importance of the development of HLA deletion mutants cannot be overstated. Unlike mice, there are no inbred humans. In a system that is so extensively polymorphic and contains approximately 10% DR “blanks,” it is virtually impossible to be certain that a particular individual from which a cell line is derived is truly homozygous in the H L A region. This is especially true considering the number of recombinational events that occur, particularly as illustrated in Fig. 2 between DP and DQ. Thus, the development of hemizygous cell lines provided the first unequivocal source of cells that were truly “homozygous. ” That is, serologists, biochemists, and molecular biologists could study these cells and draw conclusions concerning the number of products and/or alleles that were extant. Thus, for example, if two I-E-like DR beta chains are detected in a hemizygous cell line that differ by only a single amino acid, this is primafacia evidence that there are two DR p genes per haplotype rather than entertain the possibility that they represent allelic forms. Appropriate caution needs to be exercised in using these cell lines but fortunately, both the laboratories of Pious and DeMars have monitored most of these lines, and in general they have proven to be extraordinarily valuable. VI. Monoclonal Antibodies to HLA-D Region Products

Most, if not all, areas of the biological sciences have been favorably affected by the discovery and use of hybridoma-derived monoclonal antibodies. The study of the HLA-D region is no exception. Since the discovery of this relatively simple technique of making antigen-specific monoclonal antibodies (Kohler and Milstein, 1975), giant strides have been taken toward understanding the structural and functional complexity of HLA-D region encoded molecules. Structural characterization of HLA-D region antigens lagged behind the characterization of the class I molecules primarily due to the absence of welldefined, high titer alloantisera that could be used to isolate the class I1 antigens. The advent of highly discriminatory monoclonal antibodies simplified the isolation of the class 11 molecules and the description of several HLA-D region encoded molecules followed. Since the biochemical and serological characterization of these molecules is covered extensively elsewhere in this review, this section will describe the advantages, as well as some of the problems, associated with using monoclonal antibodies to study the complexity of the HLA-D region. In addition, some of the more commonly used monoclonals and their specificities will be presented. The advantages of using monoclonal antibodies in this system are immediately obvious. Alloantisera most often are obtained from multiparous women who make antibodies against their genetically disparate children. Usually these antibodies are of low affinity and low titer. These problems can be

HUMAN CLASS I1 MOLECULES

19

overcome with monoclonal antibodies. For instance, repeated immunizations before fusion may lead to production of B cell clones producing antibody of higher affinity that may be preferentially selected postfusion using appropriate screening procedures. By injecting these antibody-producing hybridomas into the peritoneal cavities of recipient mice, high titer ascitic fluid is easily harvested in 2 to 3 weeks. One milliliter of this antibodycontaining ascitic fluid routinely contains from 1 to 10 ing of specific antibo-

dy. Many of the alloantisera that were first used in attempts to characterize the HLA-D region molecules were poorly characterized and often contained antibodies against multiple specificities. For example, most of the antiDQw3 (MB3) alloantisera contain antibodies not only against the DQw3 specificity (which resides on a DQ molecule), but also against the DR specificity of the immunizing cells. If, for instance, the DQw3 alloantiserum was generated by a DR3/3 multiparous woman who has borne several DR4 children, the antiserum would most likely have antibodies against DQw3, as well as DR4. Athough several monoclonal antibodies display similar characteristics (e.g., cross-reactions between DQ and DK), their characterization is generally somewhat easier due to their higher affinity and higher titer. Recently two workshops were held that were devoted primarily to characterizing monoclonal antibodies against human class 11 molecules. These meetings, one held in September 1983, in Edinburgh, Scotland (Steel, 1984), and the other in May 1984, in Munich, Germany (Crumpton et ul., 1984), allowed for exchange of a large number of monoclonal antibodies among several laboratories which has led to a better understanding of their precise specificities. One rather surprising conclusion of both workshops was that few monoclonal antibodies were truly subregion specific. There were many examples of antibodies that reacted strongly with the products of one subregion, (e.g., DR), yet also reacted (less strongly, but significantly), with the products of another subregion, such as DQ andlor DP. This form of “cross-reaction” is not well understood but may be a result of some form of gene conversion of closely related genes. Caution should be used in attributing a specific function to a HLA-D region molecule on the basis of reaction or lack of reaction with a monoclonal antibody unless the exact specificity of the antibody has been carefully determined in the haplotype being examined (many antibodies have been shown to react with one allelic product of one subregion but with other allelic products of another subregion). Several studies have utilized monoclonal antibodies to localize supertypic specificities to a given subregion product. Caution should be used in drawing conclusions from these studies until specificities of the monoclonal antibody can be absolutely correlated with the desired allospecificity. One successful method of correlation has been to block the cytotoxicity of alloantisera with the monoclonal in question.

20

ROBERT C. GILES AND J. DONALD CAPFU

The ability to make reagents which are either subregion (or locus) specific or allele specific is of obvious importance (Makgoba et al., 1983). Such monoclonal antibodies might be useful not only from a serologist’s perspective but may also prove to be useful in assigning certain functions to particular loci or alleles. Due to the present scarcity of such monoclonal antibodies, several innovative techniques have been developed. For example, Radka et al. (1983)have generated some interesting allotype specific (DR5 and DR7) monoclonal antibodies by immunizing rats rather than the conventional BALB/c mouse. Additional monoclonals recognizing polymorphic specificities, such as DQw2 (found in association with DR3 and DR7), might be generated by immunizing animals with a DR3, DQw2 human B cell line, and boosting with a DR7, DQw2 cell line in hopes of expanding DQw2 positive clones. The use of synthetic peptides as immunogens to make monoclonal antibodies with desired specificity has been successfully attempted in other allogeneic systems (Alexander et al., 1983) and would appear to be potentially beneficial in making anti-class I1 antibodies. As more HLA-D region genes are cloned and sequenced, comparisons of these genes are likely to demonstrate areas of difference that could be useful in terms of constructing synthetic peptides for making allele specific monoclonal antibodies. Finally, the technique of producing human anti-class I1 monoclonal antibodies using a human-human hybridoma system is attractive due to the therapeutic value that such monoclonal antibodies might possess. In addition, it is likely that in an allogeneic system the discriminatory properties of human monoclonal antibodies would be even greater than the murine antibodies. Another innovative approach has been the isolation of heavy chain class switch variants of monoclonal antibodies in order to convert noncytotoxic monoclonal antibodies to cytotoxic monoclonal antibodies. This approach has been recently exploited by Parham et al. (1983) who started with the hybridoma Genox 3.53, a monoclonal with DQwl specificity. Spontaneously arising switch variants were selected in the fluorescence activated cell sorter, cloned, and tested. The major advantage of this technique is that both cytotoxic and noncytotoxic antibodies of precisely the same specificity can be developed. These kinds of techniques would provide valuable reagents that could be used by investigators that could be effectively tailormade for specific purposes (precipitation, cytotoxicity, affinity column chrornotography, etc.) but yet with precisely the same variable regions and, therefore, the same specificities. A second issue of major concern in the field of class I1 structure, function, and genetics concerns the use of specific cell lines. Indeed, early studies with what have later turned out to be highly discriminating monoclonal antibodies were clouded by the use of poorly characterized cell lines. One

HUMAN CLASS I1 MOLECULES

21

major problem has been alluded to in Section V, that is, it is exceedingly difficult to be absolutely certain of “homozygosity” under most circumstances. Very often investigators were forced to choose cell lines that grew well, and, more often than not, after the cells had been transformed with EB virus for altogether different reasons. Typing lymphoblastoid cell lines is extremely difficult and several early studies drew conclusions concerning alleles versus loci which were probably wrong because the cell lines being used were not truly homozygous. The deliberate construction of EB transformed cell lines within families in which first cousin marriages are demonstrably clear, is now routine. The two workshops on monoclonal antibodies recognized this extraordinary difficulty with cell lines and attempted to have specific cell lines widely distributed so that similar procedures could be used with the same monoclonal antibody on the same cell line and, thus, maximum amounts of information could be generated. Table VIII gives a list of some of the key homozygous cell lines that are commonly used in class 11 structure, function, and genetics. These cell lines are widely available, have been highly characterized, and provide an important reference source for laboratories working in this arena. Table IX lists some of the more commonly used antibodies and their most probable specificities. It is beyond the scope of this review to detail the contribution of monoclonal antibodies to the study of class I1 gene products in man. However, a short sumtnary is appropriate. It is difficult in retrospect to imagine that the class I1 problem would have been dissected without the use of monoclonal antibodies. Low titer, low affinity alloantisera of limited availability were useful in the early biochemical characterization of certain of the class I1 gene products, but it was with the introduction of monoclonal antibodies that much of the rapid progress in our understanding of the HLA-D region unfolded. As monoclonal antibodies were developed (Lampson and Levy, TABLE VIII COMMONLY USED HOMOZYGOUS B CELLLINES DR haplotype

Cell line

DR1 DR2 DR3 DR4 DR4 (Japan) DR5 DRw6 DR7

HOM2 YGF (3107) wT4Y LEIF-T KT2 LG50 WT46 Mann

22

ROBERT C . CILES A N D J. DONALD CAPRA

TABLE IX REACTIVITYO F SOME WIDELYUSED MONOCLONALANTIBODIES W I T I I SEVERALHOMOZYGOLJS CELLL I N E + ~~~~~

~

Cell lines

Antibody ~

Specificity

DRZ PGF

DR3 WT49

DR4 Leif-T

DK4 KT3

DRw6 WT46

DR7 Maim

~~

IB5 L203 Q2.70 DA6.164 HIE3 17.3.3 Tu22 SDR4.1 IVD12 SG171 Tu39 MHM4 B7/21 ~~~~

DR DR DR DR (except 7) DR3, 4, 7 DR7 only DQ DQwl DQw3 DR, DQ mainly DP DP DP

+ + +

+ + +

+

+

+ + + + + +

+

+

+ + +

+

+

+ + + + + +

+ + + + + + + + + +

~~

A blank space indicates no detectable reaction. Data abstracted froin Cruinpton et ul. (1984). a

1980; Quaranta et al., 1981) they were tested for blocking of alloantisera in order to determine specificity (Grumet et al., 1980; Brodsky, 1984). They were tested on HLA-DR deletion mutants to further refine their specificities and molecules were isolated and studied by two-dimentionsal gel electrophoresis and peptide mapping (Carrel et al., 1981; Accolla et ul., 1981; Tosi et at., 1981; Accola, 1984) and by amino acid sequencing. These procedures led to the delineation of several key monoclonal antibodies relatively early. For example, monoclonal antibody L203, one of the first widely available reagents, recognizes DR gene products in almost all cell lines tested (Lampson and Levy, 1980; Hurley et al., 1982a). There was little, if any, reactivity with other molecules. Monoclonal antibodies such as IVDl2 recognize DQ gene products (specifically DQw3 in this case) in the same cell lines that antibody L203 recognized DR gene products (Giles et al., 1983). Finally, monoclonal antibody I-LR1 with specificity for the HLA-DP subregion (specifically the alleles DPw2, DPw3, and some D P d ) , provided a third monoclonal antibody that isolated a third set of molecules from the same cell line (Hurley et ul., 1982b, 1983b). These findings led to the firm acceptance of the three locus model in the class I1 system (Dick, 1982; Hurley et al., 1983a). Several other sets of monoclonal antibodies were used in other laboratories to establish essentially the same thing. The use of these

H U M A N CLASS I1 MOLECULES

23

monoclonal antibodies, in conjunction with classic serologic techniques, has not only formed the foundation for the biochemical isolation and characterization of class 11 molecules but has also framed the questions that could be better addressed by the molecular biologists. VII. HLA-DR Biochemistry

The extraordinary progress on the biochemistry of class I1 molecules is evident when one appreciates that in the very complete review of the human Ia system done by Winchester and Kunkel (1979) for Advances in Immunology 5 years ago, only four paragraphs were devoted to the structures of these molecules. It was known that the Ia antigens consisted of two noncovalently linked polypeptide chains of approximate molecular weight 29,000 and 34,000. There was considerable controversy at that time as to which of the two polypeptide chains was polymorphic although work was beginning to emerge that strongly suggested that the p- or light chain was variable and that the a-or heavy chain was invariant. There was no clear understanding as to whether the HLA-D region encoded multiple molecules or whether the molecules defined as Ia antigens corresponded to those encoded by the murine I-A or I-E subregion. The early studies on human class I1 molecules involved bulk isolation of molecules in the above molecular weight range and the antigens were often isolated from cell lines that were not class I1 homozygous. Considerable confusion existed as to the relationship between the isolated molecules and the allotypic specificity expressed by the cell line. As mentioned in Section VI, the major breakthrough was the development of monoclonal antibodies. The alloantisera that had been used were simply too weak or not available in enough quantity for the isolation of enough material which reacted with specific antisera of defined characteristics to be of general use. The monoclonal antibodies were extremely important not only in sorting out the genetic complexity of the HLA-D region but also in isolating molecules for biochemical analysis. By and large early studies used monoclonal antibodies that, in retrospect, were very poorly characterized. As has been indicated in the section on monoclonal antibodies above, relatively few monoclonal antibodies are truly subregion specific and even fewer are specific for the products of specific alleles. Thus, inore often than not, investigators were utilizing murine monoclonal antibodies that reacted with molecules in the appropriate molecular weight range and then sorting out from the biochemistry the relationship of these molecules to the known murine structures and genetics. In addition, the extraordinary complexity of the region was not appreciated; that is, while most workers anticipated a human counterpart of the murine I-A and I-E subregion encoded molecules, few ex-

24

ROBERT C. GILES A N D J. DONALD CAPRA

pected that the I-E counterpart, the HLA-DR subregion, would encode multiple P-chains and none had predicted the presence of the HLA-DP locus which had not been defined in the mouse. Be that as it may, significant progress occurred in the period between 1980 and 1982 as monoclonal antibodies were applied to the problem. Kaufmann et al. (1980) and Walker et al. (1980) performed peptide mapping and Charron and McDevitt (1980) two-dimensional gel electrophoresis analysis of several different cell lines with a panel of monoclonal antibodies. The results were fairly conclusive in illustrating that the heavy chains of different HZA-DR allospecificities were virtually identical in all cell lines examined but the light chains of different HLA-DR allospecificities differed and, therefore, must bear the alloantigenic determinant. These studies effectively put to rest the controversy as to which of the two chains bore the polymorphic determinants. These papers were landmarks and set the tone for much of the work that followed. Later investigators would detail that additional p-chains were encoded in the HLA-DR subregion and would more fully document the extent of variability between different @-chains that arose from different alleles. From this point forward all agreed that it was the p-chain that was variable and that the a-chain was largely invariant (Kaufinan and Strominger, 1982). Recently, Hurley et al. (1984) and Giles et al. (1984b) have isolated HLA-DR P-chains from two different homozygous cell lines of each HLA-DR haplotype (DR1-7). The chains were isolated using the DR-reactive monoclonal antibodies L203 and/or IIIE3. Two cell lines of each DR haplotype were used to ensure that any variation demonstrated represented an allelic difference and not a random point mutation. Allelic polymorphism was found among the DR P-chains examining only amino-terminal tyrosine sequences (Table X).

TABLE X AMINO-TEHMINAL AMINOACID SEQUENCES OF DR P-CHAINS Position ~~

~

DR haplotype

10

DR1 DR2 D R3 DR4 DR5 DRw6 DR7

-

_ Y

13

-

_ _

Y y -

-

30

32

-

-

Y Y

-

-

-

26

-

Y

Y Y Y Y -

Y

-

-

-

-

Y Y Y

H U M A N CLASS I1 MOLECULES

25

The next important development occurred when a group of investigators began to appreciate that the complexity of either sequences (Hurley et al., 1982~1,1983b), or two-dimensional gels (Shackelford and Strominger, 1980; deKretser et al., 1982, 1983) was greater than could be anticipated from a single a- and @-chainin the HLA-DR subregion. There were at least three interpretations of these data. First, that variable glycoscylation patterns explained much, if not all, of the heterogeneity. Second, that there were multiple @-chainloci within the HLA-DR subregion and finally, that some of the molecules were encoded in a separate subregion, possibly analogous to the murine I-A subregion. Two kinds of experiments were done to approach these alternative possibilities. For example, Shackelford et al. (1981a) used the monoclonal antibodies L203 and L227 to study precursor-product relationships. L203 is an excellent monomorphic antibody which recognizes essentially all HLA-DR subregion products while L227 recognizes an overlapping population of human la molecules (Lampson and Levy, 1980; Hurley et al., 1982a). On the Ragi cell line, they were able to demonstrate that the determinants for these two monoclonal antibodies were not on the N-linked oligosaccharide since both antibodies recognized nonglycosylated DR antigens from cells treated with tunicamycin. In contrast in another B cell line they found at least two DR light chains. A xenoantiseruin and L203 recognized both light chains whereas L227 precipitated only one of the two. Their interpretation of this result related primarily to variations in high manose and complex type oligosaccharides and that both monoclonal antibodies recognized an H L A DR subregion product. It is now known, of course, that L227 recognizes products of both the HLA-DP and, in some instances, the HLA-DQ subregion and, therefore while most of the conclusions of this paper were well drawn, they are incorrect in that the additional reactivities of these monoclonal antibodies was not appreciated at the time. The work was pivotal, however, in that it illustrated another complexity of the problem which continues to plague investigations in this field. That is, not all cell lines behave the same way and, commonly, observations made within one allospecificity do not necessarily pertain to a second allospecificity. This problem has been addressed by Hurley et al. (1983a) who pointed out that very often monoclonal antibodies will be rnonotnorphic for the products of one subregion but will behave in a polymorphic fashion with the products of another subregion. This has led to great complexity not only from the point of' view of interpreting data and designing experiments but in terms of understanding the evolution of the major histocompatibility complex class I1 genes themselves. Several experiments were performed by Hurley et al. (1982a, 1983c), using the same two monoclonal antibodies that Schackelford and Strominger

26

ROBERT C . GILES A N D J. DONALD CAPRA

used, L203 and L227. Here, by internal radiolabeling certain amino acids, it was deduced that at least three DR @-chainscould be isolated from a cell line. Variations in the yield of specific amino acids in particular positions led these authors to conclude that the HLA-DR subregion consisted of a single a-chain and three @-chaingenes. Unfortunately, the authors did not utilize truly homozygous cell lines, therefore some of these conclusions need to be tempered by the possibility that some of the results were due to allelism rather than separate DR P-chain loci. However, the study was the first to propose the three @-chainstructure of the HLA-DR subregion by amino acid sequence analysis (see Fig. 5). An additional complication of this study was the lack of appreciation that L227 recognized, in addition to HLA-DR subregion products, some products of HLA-DQ and HLA-DP that further complicated the analysis. Both of the above studies had flaws the investigators were only dimly aware at the time. Few fully appreciated the extraordinary complexity of the whole HLA-D region or, indeed, the additional complexities that were introduced by the heteroantibodies, even though they were monoclonal. Important insights then began to derive from monoclonal antibodies to HLA-D region products that were characterized as being either (1)monomorphic for the HLA-DR subregion alone and lacked activity for any other subregion product (these studies were greatly facilitated by studies with deletion mutants), or (2) poEymorphic particularly the development by Johnson et d. (1982) and Radka et al. (1983) of monoclonal antibodies that behaved with HLA-DR subregion products precisely as alloantisera. The development of DR3 and DR5 monoclonal antibodies greatly facilitated structural analysis. Finally, around 1982 it was appreciated that despite the advantages of having cell lines which grew well, a far more crucial determinant was homozygosity and the wide distribution of these cell lines (particularly through the international workshop route) lessened the confusion considerably. In addition, at this time complete structural analysis of DR a-and p-chains emerged from the biochemical analysis from the Hilschmann laboratory (Yang et al., 1982; Kratzin et al., 1983) and from cDNA clones from Lee et al. (1982), and Korman et al. (1982a,b), such that the general structural features of the GLO

0

DP

DQ

DR I

0 ,

FIG. 5. Schematic representation of the HLD-DR subregion derived from biochemical studies. The order of the genes is unknown (see Section XI1 for more details on gene organization). The a-chain is not polymorphic. Probably all three p-chains are polymorphic although this point is not certain.

H U M A N CLASS I1 MOLECULES

27

molecules, as well as allelic variation began to be appreciated. Several papers (Wiman et d . , 1982, Gustafsson et d . (1982), and Larhammer et d . (1982a,b) that refer to DR @ cDNA clones are actually DQ @ structures (see Section VIII). A molecular correlation between cellular reactivities (HLA-Dw) and serologic reactivities (HLA-DR) has long eluded workers in this field. Recently two laboratories, Groner et al. (1983) and Nepom et al. (1983), have made considerable progress toward an understanding of these relationships. Whereas DR antigens are defined by the reactivity of selected alloantisera for “Ia-like” molecules, D antigens are defined by the patterns of reactivity elicited in mixed lymphocyte culture when cells from HLA-D homozygous donors are used as “typing cells” or HTC. As was discussed in Section 11, there are many situations where HLA-D and HLA-DR precisely correlate but in other instances, for example in HLA-DR4, there is a wide variety of different D types that exist within the broad category of HLA-DR4. Thus, less than 70% of DR4 haplotypes in Caucasians are DRw4 positive and this relationship changes from one population to another. Several possible explanations have been proposed for these so-called “splits.” First, HLA-Dw4 and DR4 may be products of different loci, each encoding distinct “Dregion” associated antigens. A second possibility is that Dw4 and DR4 are products of a single gene, the primary product of which is recognized by DR4 alloantisera but through posttranslational modification, significant phenotypic variation occurs that is detected in MLC by T cells as a Dw4 specificity. A third possibility is that the haplotypes expressing the various D specificities (for example, Dw4, DwlO, LD40, etc.) may be encoded by different alleles but share a common cross-reactive or “supertypic” determinant recognized by DR4 alloantisera. The technique used by Groner et al. (1983)and Nepom et al. (1983) was fundamentally the same. Groups of DR4 homozygous cells which differed in their D specificity were studied. One laboratory used two-dimensional gels, the other one-dimensional gel followed by isoelectric focusing. The conclusions of both studies were virtually identical in that the Dw type could be easily predicted by the complexity of the @-chain pattern in the particular analysis. These studies suggest that DR4 is in fact a supertypic allospecifcity that is carried on highly related molecules and the molecular basis for this difference resides on one or more of the beta chains in the DR subregion. The direct correlation between HTC and alloantisera still remains a subject of intense investigation. It is likely that in some instances, HTC (D typing) and DR discrepancies where they exist will be due to differences in different subregions; for example, DQ versus DR. However, in these studies on DR4, it is clear that the difference does not reside in a second locus but from the fact that the specificity DR4 represents a supertypic specificity and that there are several “variant forms” of DR4 that exist in the population.

28

ROBERT C. GILES A N D J. D O N A L D CAPRA

These, have recently been defined a bit better with alloantisera and it is likely that the so-called DR4 cross-reacting group will eventually be expanded to four or five different specificities based on serologic typing. The overwhelming evidence at the present time is that the HLA-DR subregion encodes a single nonpolymorphic a-chain and two or three extensively polymorphic @-chains(Fig. 5). The reason for hedging on a number of 6-chain loci largely derives from the very real possibility that the number of P-chain loci may differ in different haplotypes. Several laboratories have described techniques for the isolation of modest amounts of human class I1 antigens (Walker and Reisfeld, 1982) which were later used to obtain primary structural information (Walker et al., 1983; Wiman et al., 1982a). However, only the Hilschmann laboratory has used this technique to obtain complete primary structures. The structure of an HLA-DR a-chain cDNA is shown in Fig. 6. This amino acid sequence was also deduced by Kratzin et al. (1983). Five laboratories have reported complete a sequences and while there are a few discrepancies they are likely of little consequence. The overall structure of the DR a-chain will be discussed in some detail here and referred to again under the structure of the D Q and DP molecules. The a-chain is composed of 299 amino acids of which 191 are exposed on the outside of the plasma membrane. The membrane imbedded portion of the chain is thought to consist of 23 hydrophobic amino acids. The succeeding 15 amino acids form the cytoplasmically localized hydrophylic tail. The extracellular portion with carbohydrate moieties linked to and Asn118 seems to be organized into two domains. The second domain (which contains the only disulfide bond of the a-chain) displays amino acid sequence homology to immunoglobulin constant regions as well as to the second domain of the beta chain of class I1 antigens (see below), to the third domain of heavy chains of class I molecules and to P,-microglobulin. These observations were made virtually simultaneously by Larhammar et al. (1981, 1982a,b, 1983b), Korman et al. (1982b), Yang et al. (1982), and Lee et al. (1982). The structure(s) of the p-chains has been somewhat more difficult to obtain but as of this writing four different laboratories have obtained complete P-chain sequences. The protein sequence was derived by Kratzin et al. (1980) and cDNA clones were obtained by several laboratories. The coniplete sequence of the cDNA derived from Long et al. (1983a,b) is shown in Fig. 7. The predicted amino acid sequence has 237 amino acid residues. It has two immunoglobulin-like disulfide loops and a 22 amino acid residue membrane integrated segment. Sixteen amino acid residues reside on the cytoplasmic side of the plasma membrane. The single asparagine-linked carbohydrate moeity is attached to Asn’”. The amino terminal 91 residues of the p-chain are homologous to the corresponding region of HLA-A, -B, and -

I

I

O

10

A

E

F

Y

L

N

P

D

O

S

20

G

E

F

M

F

D

F

D

G

30

D

E

I

F

~

V

G l e T G ATC ATE CAG GCC GAG T T C T A T CTG AAT CCT GAC CAA T C A GGC GAG TTT A T E T T T GAC TTT GAT GGT GAT GAG A T 1 T T C CAT 616

40 50 60 D M A K K E T V Y R L E E F G R F A S F E A O G A L A N I A GAT ATG GCA AAG AAG GAG ACG GTC TGG CGG C T T GAA GAA T T T GGA CGL TTT GCC AGC T T T GAG GCT CAA GGT GCA T T G GCC A& ATA GCT 70

100

11c

K

P

V

T

y W L R N GTC ACG TGG C T T CGA AAT

140

130

G

90 P I T N V P P E V T V L T N CCG A T 1 ACC AAT GTA CCT CCA GAG GTA ACT GTG C T C ACG AAC

F I D K F T P P V V P T T C ATC GAC AAG T T C ACC CCA CCA GTG GTC AA

8

S P V E L R E P N V L AGC CCT GTG G M C T G AGA GAG CCC AAC GTC CTC A k

no

*&&8

V D K A N L E I M T K R GTG GAC h A A GCC AAC CTG GAA ATC ATG ACA AAG CGC T i c

T

G

V

S

P S T E D V Y D CCC TCA ACT GAG GAC G T T TAC GAC

'@

E

T

V

F

L

P

150

R

E

D

~

L

F

R

K

F

M

V

L

P

F

GGA AAA CCT GTC ACC ACA GGA GTG T C A GAG ACA GTC T T C C T G CCC AGG GAA GAC CAC CTT T T C CGC AAC T T C CAC T A T CTC CCC TTC C T G 160

P

L

P

E

T

190

T

E

N

V

100

170

R

V

E

~

~

G

L

D

E

P

L

L

K

ACG GTG GAG CAC TGG GGC T T F GAT GAG CCT C T T CTC AAG C i C Ir;dG:F

V

C

A

L

t

L

200

T

L

V

G

L

V

G

I

I

l

G

T!T 210

T

GOAT GT:

C:A

A&

I

I

I

F

CCT CTC CCA GAG A C T ACA GAG AAC G T G GTG T G T GCC C T G GGC C T G ACT GTG GGT CTG GTG GGC ATC A T 1 A T 1 GGG ACC ATC T T C ATC ATC

220

K G V R K S N A A L R R G P 2 t g m AAG GGA GTG CGC AAA AGC AAT GCA GCA GAA CGC AGG GGG CCT C T G TAA GGCACATGGAGGTGATGGTGTTTCTTAGAGAGAAGATCACTGAAGAAACTTCTGC

TTTAATGACTTTACA*AGCTGGCAATATT~AATCCTTGACCTCAGTGAAAGCAGTCATCTTCAGCGTTTTCCAGCCCTATAGCCACCCCAAGTGTGGTT~TGCCTCCTCGATTGCTCC GTACTCTAACATCTAtCTGGCTTCCCTGTCTATTGCCTTTTCCTGTATCTATTTTCCTCTATT~CCTATCATTTTATTATCACCATGCAATGCCTCTGG~TAAAACATACAGGAGTCT

GTCTCTGCTATGGMTGCCCCATGGGGCATCTCTTGTGTACTTATTGTTTAAGGTTTCCTC~~

FIG.6 . Nucleotide sequence and predicted amino acid sequence of a DH a-chain cDNA. Residues in boxes denote the attachment sites for N-linked carbohydrates. C, Cysteine residues. From Larhainmar et al. (1982a,b).

-29 ~ V C L K L P G G S S L A A L T V T lGlnCTCCTCTGGCCCCTGGTCCTGTCCTCTTCTC~GC ATG GTG TGT CTG M G CTC CCT GGA GGC TCC AGC TTG GCA GEG TTG M A GTG ACA CTG ATG 10 1 V L S S R L A F A G D T R P R F L E L L K GTG CTG AGC TCC CGA CTG GCT TTC GCT GGG GAC ACC CGA CCA CCT TTC TTG GAG CTG CTT M G TCT

30 E

R

V

R

F

L

E

R

H

F

~

N

O

E

E

Y

A

40 R

S

E

C

~

F

F

20 N

L 95

G

T

i3G TGT CAT TTC TTC M T CCG M G

485

50 F

D

S

D

V

G

E

Y

R

A

V

R 215

GAG CGG GTG CGG TTC CTG GAG AGA CAC TTC CAT M C CAG GAG GAG TAC GCG CGC TTC GAC AGC GAC GTG GGG GAG TAC CGG GCG GTG A f f i E

L

G

R

P

D

GAG CTG GGG CGG CCT GAT

N MC

A

E

GCC WG

60 70 80 Y Y N S O K O L L E O K R G O V Q N Y C R TAC TGG M C AGC CAG M G GAC CTC CTG GAG CAG M G CGG GGC CAG GTG GAC M T TAC TGC AGA CM

90 100 110 Y G V V E S F T V O R R V ~ P O V T V Y P A K ~ O P TM GGG GTT GTG GAG AGC TTC MA GTG CAG CGG CGA GTC CAT CCT cnt GTG ACT GTG TAT CCT GCA MG MC CAG ccc CTG CAG CAC

120 130 ~ N L L ~ C ~ ~ S G F Y P ~ S X E ~ R U F CAC M C CTC CTG GTC TGC TCT GTG AGT GGT TTC TAT CCA GGC AGC A T 1 G M GTC AGG TGG TTC CGG M C GGC CAG GM GAG

~

365

L O 455

140 R N G GCT GGG

G 545

170 T F P R S G E V Y ACA TTT CCT CGG AGT GGA GAG GTT T M

635

180 190 200 T C O V E H P S V T S P L T V E U S A R S E S ~ O S K M L ACC TGC C M GTG GAG CAC CCA AGC GTA M G AGC CCT CTC AKA GTG GM TGG ACT GCA CGG TCT GM TCT GCA CAG AGC M G ATG CTG ACT

S 725

220 210 230 G V G G F V L G L L F L G A G L F I Y F R N O K G ~ S G L GW GTC GGG GGC TTT GTG CTG GGC CTG CTC TTC CTT EGG GCC GGG CTG TTC ATC TAC TTC AGG M T CAG MA GGA CAC TCT GGA CTT CAG

O 815

237 P T G F L S CCA ACA 0 3 TTC CTG AGC TGA AGTGCAGATGACMTTTAAGGMGMTCTTCTTCCCCAGCTTTGCAGGAT~GCTTTCCCGCCTGGETGTTATTCTTCCMGAi3WG

927

150 V V S T G L I O N G GTG GTG TCC ACG GGC CTG ATC CAG M T GGA

O

GAC

~

160 U T F O T L V M L TGG ACC TTC CAG M C CTG CTG ATG CTA

E

GM

M

~

~

G G C T T T C T U G G A C C T A G T T G T M T ~ T ~ G C ~ T G C A ~ T G T C C T C C C T T G T ~ T T C C T C A G T T C C T ~ C C T T ~ C T G M G T C C C A G C A T T G A T G G C A G C G C C T1046 CATCTT

U A C T T T T G T G C T C C C C T T T G C C T ~ C C C T A T ~ C T C C T G T G C A T C T G T M T C A C C C T G T A C C A C ~ C A C A T T A C A T T ~ T G T T T C T C M A G A T G G A G T T ~ I C I1160 ~

FIG. 7 . Nucleotide sequence of a DR P-chain cDNA. The predicted amino acid sequence is given above the nucleotide sequence. The numbering of amino acids starts with the amino-terminal residue of the mature chain. The 29 preceding amino acids represent the putative signal sequence. The putative poly(A) addition site is underlined. From Long et al. (198313).

E

E

K

A

G

HUMAN CLASS I1 MOLECULES

31

C antigen heavy chains. Residues 92 to 192 of the p-chain display statistically significant homology to members of the immunoglobulin family, &-microglobulin, and the immunoglobulin-like domains of HLA-A, -B, and -C antigen heavy chains. A model depicting the general overall structure of class I1 molecules which was originally described for the EfLA-DR subregion by Korman et al. (1982b)is shown in Fig. 1. It illustrates the general structure ofthe molecule showing the heavy and light chains (a and p) to be similar in size despite their apparent differences by SDS gels. These differences are due to the fact that the a - or heavy chain has two carbohydrate attachment sites whereas the p-chain has but a single site (Shackelford and Strominger, 1983). The disulfide loops are illustrated as well as the portions of the molecule that are cleaved by papain, trypsin, and chymotrypsin, respectfully. The a-chain is depicted to be phosphorylated at its hydrophilic carboxyl terminus (Kaufman and Strominger, 1979), although there is evidence that most if not all phosphoylation of class I1 molecules occurs on the y-chain (P. Peterson, personal communication). Figure 8 shows a comparison of the amino acid sequence of the (Y I1 domain of the heavy and light chain of HLA-DR and HLAB7, pzmicroglobulin, and the Cy3 domain of immunoglobulin G , illustrating the homology between these structures.

VIII. HLA-DQ Biochemistry

The initial evidence demonstrating the possibility of a second locus distinct from HLA-DR was presented in 1978 (Tosi et d . , 1978). The Ia pool from the B cell line Daudi was separated into distinct subsets by means of specifically reacting alloantisera. Using a number of alloantisera in a series of direct and sequential binding tests to an lz5I-labeled Ia preparation from Daudi, Tosi's group demonstrated that the supertypic specificity DQwl (MB1, DC1) resided on a molecule distinct from, but in linkage disequilibrium with, the DR molecule. Although the work was largely dismissed at the time, in retrospect it is clear that this represented the first description of the locus that is now called HLA-DQ. Several studies using a variety of techniques have been responsible for definitively proving the existence of a second locus in the HLA-D region. Shortly after their initial findings, Corte et al. (1981) used a monoclonal antibody (BT3/4) to demonstrate by two-dimensional peptide mapping that the molecule which bears the DQwl specificity was biochemically distinct in both its a and p subunits from those that carry the DR specificities. Shackleford et d.(1981a) used human alloantisera and the monoclonal antibody Genox 3.53 (anti-DQwl) to demonstrate that the supertypic specificity

1co

HLA-DR

a chain

HLA-DR

p

120

110

chain

HLA-A. -8. - C p,-nicroglobul i n

CL CH1

‘2 G E P E

Cd 130

HLA-MI

140

150

160

170

a chain

HLA-DR p c h a i n HLA-A. -8. -C &-nicroglobul i n

‘1 ‘H CH

‘n

FIG. 8. Amino acid sequence comparisons (one-letter code) of the second domain of DR a with the sequence of the second domain of DR p, the third domain of HLA-A, HLA-B, and HLA-C, Pz-microglobulin, the constant domain ofthe K light chain (C,) and the three constant domains ofaii IgC, heavy chain (CH1,CH2,and C,,3). Boxes: residues shared by the sequence of the DR a-chain and any of the other sequences. From Larhamniar et al. (1982b).

HUMAN CLASS I1 MOLECULES

33

DQwl was carried on a two-chain molecule, which as judged by two-dimensional gel electrophoresis, was distinct from the DK molecule. The use of monoclonal antibodies and homozygous cell lines has contributed significantly to understanding the complexity of the HLA-DQ subregion. In addition to the monoclonal antibody BT3/4 used by Corte et al. (1981), biochemical analyses using several other monoclonals aided in dissecting the complex HLA-D region (for a discussion of monoclonal antibodies see Section VI). Genox 3.53 (Brodsky et al. 1980), SDR1.2 (DeKretser et al., 1982), SG171 (Goyert et al., 1982), IVDl2 (Giles et al., 1983), Leu10 (Chen et al., 1984), Tu22 (Pawelec at al., 1982a,b), and CC11.23 (DeMars et al., 1983) are just a few of the DQ-specific monoclonal antibodies which have been reported. Although many of the early reports demonstrating a second locus suggested possible homology of this locus with the murine I-A locus, final proof of this hypothesis awaited amino acid sequence analysis of the DQ molecule. Goyert et al. (1982) first demonstrated homology of the DQ molecule with the murine I-A molecule by amino-terminal amino acid sequence analysis of both the a- and @-chains.Using a monoclonal antibody SG171, which recognizes DK and UQ in DR7 cell lines, and also a rabbit antiserum RB03, which reacts solely with DQ in all cell lines tested, they described the presence of an I-A homolog in at least two DR haplotypes. Subsequently Bono and Strominger (1982, 1983) confirmed that the DQwl specificity resided on the DQ molecule. Amino terminal sequence analysis of the DQwl bearing molecule isolated using the monoclonal antibody Genox 3.53 (anti-DQwl) revealed homology of the a-chain to the murine I-A a-chain. The @-chain isolated in this study appeared to be blocked at the amino terminus. Subsequently, Giles et nl. (1983) using the monoclonal antibody IVD12, demonstrated that the supertypic specificity DQw3 (MB3, DC4) also resides on an HLA-DQ molecule. These data confirmed that the DQwl and DQw3 supertypic specificities reside on HLA-D region molecules which represent allelic products of the HLA-DQ subregion. There is good evidence suggesting that the DQw2 (MB2, DC3) specificity also resides on a DQ molecule (Karr et al., 1983, 1984) further strengthening the original argument that the MBbearing molecules represent an allelic series (Duquesnoy et al., 1979). In addition to the serological and structural variation mentioned previously for the DQ molecules bearing the DQw specificities (i.e., DQwl vs DQwS), there is a growing body of evidence that molecules bearing the same supertypic specificity are distinct from one another. The first example of this distinction was reported by Shackelford et al. (1983), who demonstrated by Go-dimensional gel electrophoresis that the @ subunit of a DQwl-bearing molecule isolated from a DR2 cell line was distinct from the p subunit of a DQwl-bearing molecule from a DRw6 cell line. De Kretser et d.(1983)

34

ROBERT C. GILES AND J . DONALD CAPHA

subsequently detected electrophoretic variation of both a- and @-chains from DQwl-bearing DQ molecules isolated from a DR2 cell line vs a DR6 cell line. Goyert and Silver (1983) showed by two-dimensional gel electrophoresis that DQ p-chains varied in electrophoretic mobility depending upon the haplotype from which the chains were derived. More recently, Giles et al. (1984~)determined the amino-terminal tyrosine sequences for DQ molecules which bear the DQw3 determinant isolated from two DR4 and two DR5 homozygous cell lines and showed that although the distribution of the amino-terminal tyrosine residues in the alpha chains was identical, differences existed between DQ @-chainsisolated from the cell lines of differing DR specificities. This work has since been extended to include amino-terminal tyrosine sequences of DQ molecules isolated from two cell lines of each DR haplotype (DR1-7). (Giles et al., 1984b) (see Tables X and XI). By this limited analysis, a minimum of three allelic forms of DQ a and five allelic forms of D Q p were found. When examining appropriate combinations of DQ a- and p-chain molecules, six out of seven haplotypes examined could be distinguished from each other. These data demonstrate at the primary structural level allelic polymorphism of both the a- and pchains of the HLA-DQ molecule. Shortly after the demonstration of allelic polymorphism of murine class I1 molecules, evidence was presented for the formation of hybrid molecules in F, animals providing a possible molecular mechanism for the phenomenon TABLE XI AMINO-TERMINAL AMINOACIDSEQUENCES OF CC11.23-REACTIVEa-CIIAINS Position DR haplotype DR1 DR2 DR3 DR4 DR5 DRw6 DR7

Cell lines

11

16

19

25

.45.1 MDE 3107 3161 .127 wT49 PRIESS 3164 3105A DHI WT46 LG32 MANN 3163

-

Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Y Y Y Y Y Y Y Y Y Y Y Y

-

Y Y Y

Y Y Y -

Y Y

-

HUMAN C L A S S I1 MOLECULES

35

of gene complementation (Silver et d.,1980; Cook et al., 1980). This trans gene Complementation increases the alloantigen repertoire in heterozygotes. In humans, trans association has recently been reported for DQ molecules by two-dimensional gel electrophoresis (Charron et al., 1984). Recently Giles et al. (1984b) examined two DQ heterzygous cell lines and demonstrated the formation of hybrid DQ molecules within these cells. The DQw3-specific monoclonal antibody IVD12, which by Western blotting analysis reacts with isolated DQ p-chains from DQw3-positive cells, was used to isolate [3H]tyrosine-labeled D Q molecules from two cell lines typed as DQw2/DQw3 heterozygotes. Since the amino-terminal tyrosine sequences of DQw3.4 and DQw3.5 a-chains are distinct from DQw2.7 achains (DQw3.4 and DQw3.5 a-chains possess a tyrosine at position 25, whereas the DQw2.7 a-chain does not-see Table XI) it was possible by subjecting the separated chains of IVD12-reactive molecules to amino-terminal amino acid sequence analysis to compare the results between homozygous and heterozygous cell lines in order to test directly for trans complementation. In these experiments the amount of [3H]tyrosine at position 25 of the DQ a-chains was approximately 50% of the amount expected in comparison to positions 11, 16, and 19 which are invariant between DQw3.4 (or .5) and DQw2.7 a-chains. These data demonstrate at the primary structural level that both the DQw3.4 (or 3.5)and the DQw2.7 a-chains are found associated with the DQw3.4 (or 3.5) @-chainin these heterozygotes. A model depicting cis vs trans complementation is shown in Fig. 9. In addition to the polymorphism found associated with the D Q molecules, further complexity in this family of molecules might exist through the expression of multiple DQ molecules within a single DH homozygous cell line. Although recent findings demonstrate the presence of multiple DQ-like aand @-chaingenes within the genome of a single cell (Auffray et al., 1983), it is not known how many of these genes are expressed. Several studies have attempted to show that at least two DQ molecules are expressed in a cell line. The question of whether two separate DQ like molecules is expressed in a homozygous cell line has been a subject of considerable controversy. Giles et al. (1984a), using a DRl/DQwl cell line demonstrated that after clearing with the monoclonal antibody Genox 3.53 (anti-DQwl), material in the glycoprotein pool continued to react with monoclonal antibody CC11.23 (anti-DQ monomorphic) (Fig. 10). This latter material was subjected to amino acid sequence analysis and clearly represented a DQ molecule. The difficulty with this study was that it proved impossible to perform sequence analysis on the Genox reactive material and/or to compare by peptide map analysis the two immunoprecipitates. Thus, some question remains as to whether these studies definitely prove that two DQ molecules were ex-

36

ROBERT C. GILES AND J. DONALD CAPRA

FIG. 9. Schematic representation of cis- versus trans compleinentatioii i n tlie HLA-DQ subregion. A illustrates the expected combinations with exclusively cis complementation. Only parental chains would recombine and the F, (heterozygote) would have no molecules riot present in the parent (assuming each parent WPS homozygous). In B trans association is illustrated and tlie fbrmation of hybrid “neoantigens” is depicted. In addition to tlie cis moleciiles illustrated in the diagram on the left and right, the arrows depict tlie formation of hyl)rid molecules.

pressed in the homozygous cell line. More definitive studies have been reported by Karr et aZ. in two recent papers. First in a DR5 cell line and later in a DRw6 cell line (Karr et al., 1983, 1984), using a combination of monoclonal antibodies as well as alloantisera to the DR and DQ products, it was demonstrated by two-dimensional gel electrophoresis that in one case, two DQ @-chainsand in the other case, two DQ p- and two DQ a-chains could be distinguished. In addition, they reported that alloantisera, particularly those with specificities for the DQw series of molecules recognized only one of the two DQ molecules that are expressed in each of these homozygous cell lines. Collectively, these studies suggest that all cells expressing class I1 molecules have the potential for expressing two DQ a- and two DQ pchains. Probably only one “set” of these molecules bears the polymorphic determinants that are referred to as the DQw series and that the other set may represent a second allelic series or, indeed, be nonpolymorphic (there is evidence that the DX a-probably the “second” DQ a gene-is relatively invariant by Southern filter hybridization-see Section XII). A fruitful area

37

H U M A N CLASS I1 MOLECULES

of further investigation will obviously be to carefully dissect the specificities of a series of DQ monoclonal antibodies to determine whether they immunoprecipitate a single D Q molecules or two D Q molecules and which of the set of molecules in each instance bears the polymorphic specificities (see Fig. 11). It is of interest that in the study using inonoclonal antibody CC11.23 (Tables XI and XII), Giles et al. (1984b) did not detect differences in amino acid sequence that could be attributed to sequence differences between these two sets of molecules. Thus, the two DQ molecules that are expressed likely are extremely close in structure; indeed, data at the gene level would suggest that this is the case (see Section XII). If, indeed, there are two D Q a- and two D Q p-chains expressed in a homozygous cell, the possibility obviously exists for combinatorial associations leading to a maximum of four distinct molecules in a homozygous cell line (alp1, a l p 2 , a2p1,a2p2)and if full trans association occurs in the H L A DQ subregion, heterozygous individuals could generate as many as 16 distinct molecules. At the time of this writing there are four complete sequences of D Q achains available and three nearly complete sequences of D Q p-chains. Like the DR situation, one set of these structures was developed in the Hilshmann laboratory by classical protein chemical techniques. The structure of 20

4

137.1

Genox 3 53

B

1371

Genox 3.53 5th Poss

C

1321

Genox 3.5. cleored (5

5.0

cc 11.23 1.5

*I)

2 x

I a

1.0

V

0.5

4.0

.:I

3.0

?

g X

2.0

I a v

I .o

w

Fic:. 10. Gel electrophoretic pattern of molecules isolated from the cell line 137.1 using monoclonal antibodies. (A) Genox 3.53; (B) Genow 3.53 after depletion with Genox 3.53; (C) cc11.23 following depletion with Genox 3.53. Both Genox 3.53 and ccll.23-reactive molecule exhibit a bimolecular pattern in the MW range of DR antigens. From Giles et nl. (1984a).

38

ROBERT C. CILES A N D J. DONALD CAPRA

G LO

DP

B

DR

FIG. 11. Scheiiiatic representation of the HLA-DQ subregion derived from biochemical studies. At least one set of molecules (i.e., DQ a1,DQ PI) are polymorphic. The order ofall the genes is not known (see Section XII).

the a-and p-chains of DQwl bearing molecules was derived in this manner by Gotz et al. (1983). Based on the sum total of the peptides isolated from an HLA homozygous cell line, they had concluded previously that there were probably a minimum of two a-chains and seven p-chains present in their pool. The second a-chain proved to be the DQ a-chain and two of the pchains proved to be D Q p-chains along with the two DR p-chains previously described. The three remaining p-chains are still unidentified. The majority of the structures that are available for both the a- and pchains of the HLA-DQ subregion are from cDNA and genomic sequences (largely the former). It is important at the outset in describing these structures to appreciate that at the time of this writing it has not been definitively established (1)which of these chains are conclusively expressed, and (2) in TABLE XI1 AMINO-TERMINAL AMINOACID SEQUENCES O F CCl1.23-REACTIVE P CIlAINS Position ~~

DR haplotype DR1 DR2 DR3 DR4 DR5 DRw6 DR7

0

Not determined

~

Cell lines

9

16

.45.1 MDE 3 107 3161 ,127 wT49 PRIESS 3164 3105A DHI wT46 LG32 MANN 3163

Y Y

Y Y Y Y Y Y Y Y Y Y Y Y Y Y

-

Y Y Y Y Y Y Y Y Y Y

26

~

30

~~

32

~

37 N.D.0 Y Y -

N.D. N.D. Y Y N.1).

H U M A N CLASS I1 M O L E C U L E S

39

each instance, whether the two structures being compared are allelic; that is, as has been mentioned above, there is the likelihood that there are a minimum of two expressed DQ a-chains and two DQ p-chains. There is some evidence that one set of these molecules bears the supertypic specificities DQwl, 2, and 3. It is possible, although not conclusively shown that the socalled DX a- and p-chain represent the second DQ product. It is equally possible that these represent products of a separate subregion. When investigators isolate molecules as have the Hilschmann group by bulk isolation, or when cDNA or genoinic clones are sequenced, it is not known which of the products is being compared as this would require extensive serologic and functional analysis. We are likely close to this goal but as of this writing this information is simply not available. Thus it is possible that one laboratory’s cDNA clone of an a-chain of the DQ subregion isolated, for example, from a DR4 homozygous cell line is not the true allele of the DQ a-chain cDNA isolated from a DR1 homozygous cell line. One of these chains may represent DQ aI and the other represent DQ az.However, for simplicity, our approach in this review will be to deal with the available structures as allelic products and, therefore, describe differences between them as polymorphic differences reflecting the products of a single allelic series. The complete nucleotide sequence and predicted amino acid sequence of the DQ a cDNA clone determined by Schenning et al. (1984) is shown in Fig. 12. Cysteines involved in the intramolecular disulphide bond and the attachment sites for N-linked carbohydrates are within boxes. Figure 13 is a comparison of this sequence with a DQ a-chain derived from a cDNA clone of a DR3/w6 cell line (Raji)and a second DQ a-chain from a cDNA clone ofa DR4w6 cell line (Auffray et al., 1982) and the DQ a sequence derived from a DR2 homozygous cell line by Gotz et ul. (1983). Stars denote amino acid residues not available for comparison. The striking structural homology is evident. The first noteworthy observation is that the three chains are decidedly different in structure. Recall that DR e-chains are essentially identical. Indeed in the DR a-chain, only a single variant position has been found although DR a sequences from six different sources have been described. Although the majority of the D Q a-chain sequences are incomplete (either incoinplete cDNA clones or incomplete amino acid sequence data), alignment of the sequences is easy and the majority of the polymorphism appears to be in the N-terminal or first domain (a1domain)-residues 1-86-see Fig. 13. There is approximately 15% difference between the three chains in this area. The second domain, residues 87-180, shows only four or five differences depending upon the comparison and even these differences tend to be clustered. This relatively uneven distribution of allelic polymorphism was originally described in the murine I region, especially in the a-chain of the I-A subregion by Benoist et ul. (1983). Recall the second domain of class

-23

-1

+1

Leu A5n Ly5 A h Leu net Leu C l y A h Leu A l a Leu Thr Thr Val net Ser Pro cys Cly G l y G l u Asp I l e Val A T CTA M C M A GCT Cn; A X CTC Gcc CCC CTT GCC CTC ACC ACC CTC ATC AGC CCC TGT U;A GCT C M GAC All! CTC

llet I l e rra;Mcu;c ATC

4

91

A5p Hla Val A h Ser Tyr Cly Val A5n Leu Tyr Cln Ser Tyr G l y Pro Ser Gly Cln Tyr Thr Hi5 Glu Phe Asp C l y Asp Glu Gln CAC CAC C K CCC TCT TAT CCX GTA M C T E TAC CAG M TAC CCT CCC T m GCC CAG TAC ACC CAT G M T?T GAT GGA GAT GAG CAC

181

phe n r V a l Asp Leu G l y Arq Lys Clu Thr Val Trp Cys Leu Pro Val Leu Arq Cln Phe Arq Phe A5p Pro Gln Phe M a Leu Thr A5n TTC TAC CTC GAC Cn; u% A m M G GAG ACT CTC TGG m Tpc CCT c1T CTC AGA C M TlT AGA TIT CAC CCG C M TPT CCA CTG ACA M C

271

I l e Ala Vel Leu Ly5 Hi5 Asn Leu A5n Ser Leu I l e Lys Arq Ser A m Ser Thr Ala A h Thr A m Val Pro Glu Val h r Val Phe A T CCP Crc CTA M A CAT M C TR; M C ACT Cn; A T T M A CGC T C C ( M C 1 G C I . GCT ACC M T GAG W CCT GAG CPC ACA CTC TlT

94 361

Ser Lys Ser Pro Val Thr Leu Cly Cln Pro Asn Ilc Leu Ile Cys Leu Val A s p Asn I l e Phe Pro P r o Val Val A5n Ile h r Trp Leu TCC M C TCT CCC Gn; ACA CTC CCX CAG CCC M C ATC CTC A T C M C l T CTC CAC M C A T l T T CCT CCT Cn; CTC&]Ta CR;

451

Ser Asn G1y Hi5 Ser Val Thr G l u C l y Val Ser Glu Thr Ser Phe Leu Ser Ly5 Ser Asp Hi5 Ser Phe Phe Ly5 Ile Ser g r Leu Thr ACC M T CU: CAC T C A CTC ACA G M CCX Crr TIT GAG ACC AGC TTC CTC K C M C ACT GAT CAT X C TIC TTC AAG A X ACP TAC CI'C ACC

154 54 1

Ala CCP

34

64

t Glu

124

t

104 631

Pro A h Pro Met Ser C l u Leu Thr Clu Thr V a l Val Cy5 A l a Leu Gly Leu Ser Val Gly Leu Val Cly I l e Val Val Cly Thr Val Phe CCA CCC CCT ATC TCA GAG CTC ACA GAG ACT CTC CTC TCC CCC CICl CCA TR; TCT GTC U;C CTC GTG GCC ATT GTG CTC u;C ACT Crc TIC

.

t

214 721

'

I l e I l e Arq C l y Leu Arq Ser Val Cly A l a S e r Arq His 41n Gly Pro Leu *** A T A K CCA GGC Cn; CI;T K A W GCT GCT TCC AGA CAC C M G U ; CCC ITC K A A K C C A T C ~ M n ; C M G C A T C C C A ~ A C M ; G

023

~

C94 3T

;

~

A

~

A

C

A

T

C

;

~

~

C

T

A

~

A

~

CC

~

!

A

~

~

C

~

C

C

A

~

T

T

A

X

A

T

A

T

C

C

~

~

231 C

C

~

~

C

~

A

C

~

A

~

~

~

.

~ A C M A ~ ~ ~ ~ ~ ~ M ~ ~ ~ T ~ T C C C ~ ; A ~ ~ A ~ M ~ A C C T A 1063 ~ M G ~ ~ A C P M ~ C A A

~

M

~

C

A

~

C

~

h

T

C

f

f

i

A

T

A AT

A~

~ ~~

~A

C ~ A

M~

~ M ~ A T A T ~ T A C C ~ T A ~ A ~ C A ~ M C A C ~ ~ ~ M C C ~ C C M C T A C ~ A T A ~ C T G A T M C A ~

A

C

C

~

~

M

1103 ~

1261

FIG. 12. Nucleotide sequence and predicted amino acid sequence of a DQ Q cDNA clone. Cysteines involved in the intramolecular disulfide bond and the attachment sites for N-linked carbohydrates are within boxes. From Schenning et al. (1984).

A

G

~

~

M

~

C

M

H U M A N CLASS I1 MOLECULES

PIX-

190

P W l

irl D

200

210

220

41

230

E ~ E I P A P H S E L ~ C CLRSVCASRHQCPL IVVCTY~B ............................................

FIG. 13. Comparison of three available DQ cx amino acid sequences. The pII-ad sequence is derived from a cDNA clone of the DH3,w6 cell line Raji, and pUCHl is from P cDNA clone of a DR4,w6 cell line (Auffray et al., 1982). DCI CY is a protein sequence from a DR2.2 cell line (Cotz et (11.. 1983). Asterisks denote aniino acid residues not available for comparison. Arrows mark exon boundaries. Sites for addition of N-linked carbohydrate and the membrane-spanning segment are within boxes. From Schenning et al. (1984).

I1 molecules is remarkably homologous to immunoglobulin and in this region not only are the D Q a-chains more similar to each other but, as we will discuss later, they are more similar between DH a and D Q a. The three sequences show considerable homology through the transmembrane and cytoplasmic regions although the data here are less complete due to the amino acid sequence structure being incomplete. Peterson’s group has pointed out the possibility that the second domain and transmembrane domains are crucial in the interaction between a-and p-chains of the different heterodiiners and, therefore, have been allowed to diverge less (Schenning et al., 1984). Conversely, the difference between the various types of a-and p-chains, respectfully, may have to be large enough to prevent formation of hybrid antigens such as D Q a with DR P. This region of the molecule would be a likely source for this kind of difference. Figure 14 shows the cDNA sequence and derived amino acid sequence for a DQ p-chain; cystines and the attachment sites for N-linked carbohydrate are within boxes. Arrows mark exon boundaries inferred from the D Q p gene (see Section XII). Figure 15 compares the amino acid sequence of four nearly complete D Q P-chains derived from different cell lines of different DR type. Again, the bulk of the variation between these molecules appears in their first domains. However, three minor clusters of “hypervariability” can be seen between positions 52 and 57, 70 and 77, and 84 and 90. This is more reminiscent of the kind of variation that has been seen in class I molecules in both man and mouse. In no position do all four of the D Q pchains have a different amino acid. However three different amino acid residues occur at 7 positions, 6 of which are located in the amino-terminal domain. This is likely to be far more variation than is due to chance alone and it is likely that the the amino-terminal domain of the p-chain of the DQ molecule is the seat of allelic polymorphic variation which results in various

-21

Asp

t

Leu A r q V a l A l a Thr V a l Thr Leu net Leu A h Ilc Leu Ser Scr Ser Leu A l a Glu Gi:

A; Asp Ser P r o Glu A s p phe vai CCI' GAG CGC AGA GAC TCT CCC GAG CAT Tpc CTC

A CAC CTP CGC &TA GCA A T CTC ACC Tn: ATG CFS CCG ATC C K ACC TCC TCA

8 88

Tyr Gln Phe Lys G l y Leu CYs Tyr P h e Thr Asn Cly Thr Clu A r q V a l Arq G l y V a l Thr A r q H i s l i e Tyr Asn A r q G l u G l u Tyr Val TAC CAC TlT MC GCC C I G R T A C TTC ACC&]GAG CGC Cn; CCC CGT Cn; ACC ACA CAC ATC TAT M C CCA GAG GAG TAC m;

178

A r q phe Asp Ser A s p V a l G l y V a l Tyr Arq A l a V a l Thr Pro C l n Gly A r q Pro V a l A l a C l u Tyr T r p Asn Ser G l n Lys Glu V a l Leu n;C T K GAC AGC CAC Cn; 033 CTC TAC CCC M I A CTC ACG CCC CAG CGC CCC CCT CCC GAG TAC x% M C ACC CAC MC G M CTC I X A

68 26E

Glu Gly A h A r q A l a Ser V a l Asp A r q V a l C g s Arq H i s Asn Tyr G l u V a l A h Tyr A r q G l y Ile Leu Gln A r q A r q Glu Pro Ihr W Gcc CCC CCC GCC TCG CTC GAC ACC C T C M A C A CAC M C TAC GAG Cn; CCG TAC CCC u;C ATC CIC CAG AGC ACA CPC GAG CCC ACA

98 358

dl

38

128 440

V a l A r q Trp Phe Arq A m A s p Gln Glu Glu Thr A h Gly V a l V a l Ser Thr Pro Leu I l e Arq Ann G l y A s p T r p Thr P h e Gln I l e Leu cR1 CCC n;C CGC MT GAT CAC GAG GAG ACA CCC CGC CIT Cn; TCC ACC CCC CTC ATT ACG M C CCT GAC TCG ACC Tn: CAG ATC Cn

mr

158 538

Pro Cln Ary G l y A s p V a l T y r Thr C y s H l s V a l Glu His Pro Ser Leu Gln Ser Pro I l e Thr V a l Glu T r p

188 628

fkq A l l G l n Ser Glu Ser A h Gln Ser Lya U e t Leu Ser Gly V a l G l y Glg P h e V a l Leu Gly Leu I l e Phe Leu G1y Leu Gly Leu I l e CCC Gt.7 CAC T f X GM TCT CCC CAG ACC M C A X ACT u;C GTT CCA CCC TIC CTC CTC CCC CTC ATC Tpc CPT CGC CTP u;C CIT ATC

218 718

V a l met Leu Glu Ckt

Ile Arq C l n A r 9 Ser A r q Lys A y Leu Leu H l s AK

CCT CM AGG

C

~

~

C

C

ACT CCG AM

C

~

T

ccc err C

~

cn; CAC

~

~

A

C~~ACCC-C~GC~ACATCGCCAC&~ACTCAGG

m ***

TGA ~ C

~ A

. ~

229

A A

G C

A C

~ C

~ M

C

~ C

~ A

M C

~ C

T A

~

M A

G ~

A ~

~ A

~

A C

~826 A C 9b6 C

993

FIG. 14. Nucleotide sequence and predicted amino acid sequence of a DQ p cDNA clone. Cysteines and the attachment site for N-linked carbohydrate are within boxes. Arrows mark exon boundaries inferred from a DQ beta gene. From Larhammar et al. (1983a,b).

~ A

C

~

~

C

C

C C

M

43

H UMAN CLASS I1 MOLECULES

COSI I - 102

E lB

PIX- 8-2

U A -

V C -

*****A************

lts 0 200 210 QSPI~AQSESAQSlQLSCVCCFVLGLI HH--0-

FIG. 15. Comparison of DQ p amino acid sequences. The pII-p-2 and pII-p-1 sequences are derived from cDNA clones of the DR3,w6 cell line Raji (Larhammar et ul., 1982b) and cosII-102 from a gene of a DR4, 4 individual (Larhammar et al., 1983b). DCI p is a protein sequence of a DM,2 cell line (Gotz et u l . , 1983). Asterisks denote amino acid residues not available for comparison. Arrows mark exon boundaries. The site for addition of N-linked carbohydrate and the membrane-spanning segment are within boxes. From Schenning et (11. (1984).

functional allospecificities. These data also suggest that the variation is not without limit and that only certain positions may exhibit variation. We will deal with comparisons o f a - and (3-chainsof DQ molecules, a-and (3-chains of DR molecules, and comparisons between DQ and DR in a later section. IX. HLA-DP Biochemistry

The discovery of the monoclonal antibody, I-LR1, shown to be reactive with some of the allelic products of N U - D P led to the initial biochemical characterization of the D P molecule. Using this antibody, the molecule isolated was shown to consist of two chains resembling the a-and (3-chains of the DR antigens in molecular weight as measured by sodium dodecyl sulfate gel electrophoresis (Nadler et al., 1981). Partial amino-terminal amino acid sequence analysis has been performed on I-LR1-reactive molecules (Hurley et al., 1982b, 1983b).These data were the first to demonstrate at the primary structural level that the a- and (3-chains of the D P antigens were distinct from the a- and (3-chains of the DR and DQ antigens (Table XIII). These sequence data were also crucial in allowing others to verify potential cDNA and genomic clones as D P equivalents (see also section XII). A second monoclonal antibody, B7121 (anti-FA), has been described which also recognizes DP (Watson et al., 1983). The antibody has been used to support the findings that DP is distinct from known DR and DQ products. This was accomplished by studying the binding of B7/21 to HLA deletion mutants and by reciprocal immunodepletion experiments. Anti-FA was

44

ROBERT C. CILES A N D J. DONALD CAPRA

TABLE XI11 AMINO ACID SEQUENCE COMPAHISON O F DP A N D I)R ANTKENS ISOLATED FROM T H E SAMECELLLINE a-chains 9

Position DR DP

-11

Y

Position: DR DP (1

9

7 F

Y

-

13

22

24

26

F F

Y -

F F

F F

F F

16

17

18

24

F -

F

-

Y

F

F

-

Y

12

26

28

F

-

-

30

Y

Y

Y

32

Y

-

Indicates absence of the assigned amino acid at that position

shown to bind to a DH, DQ negative mutant cell line and also could not be used to immunodeplete cell lysates of DH- or DQ-reactive material. Two additional findings of particular interest were observed in this study. The authors reported that the D P P-chains isolated from parental cell line LCL721 (DR1, 3) were made up of two closely migrating chains with molecular weights of approximately 25,000 and 27,000. When full-haplotype loss mutants of LCL-721 were used only of the two P-chains was detected depending on the haplotype of the mutant cell line. These data suggest that each DP P-chain on LCL-721 may be an allelic product of one locus. The authors also described the first human class I1 antigen which displays an apparently cryptic, lactoperoxidase-inaccessible,a-chain. When the D P molecule was intrinsically labeled with [35S]methionine and immunoprecipitated with B7/21 the a-chain was labeled strongly. However, a similar isolation of B7121 a-chain was not detected when the cells were 1251-lactoperoxidase labeled. Unlike HLA-DH and -UQ where biochemical analysis has been performed on “cold” material, the only information presently available derived from HLA-DP molecules comes from that reported by Hurley et al. (1982b) using radiolabeled material. Only a single cDNA sequence of UP a has been published (Auffray et al., 1984)and it is shown in Fig. 16. Note, as indicated by stars in the DP a sequence, the concordance of the sequence to the report of Hurley et al. (1982b). Three separate laboratories have reported the isolation of cDNA clones comprising the bulk of the D P P-chain. Unfortunately, however, none of the cDNA clones provide a complete P-chain sequence as one begins at amino

v

v

o

V *

L

v T H E F D C F M F D F M F F F S B G A G A ~ I K A D H V S T V A A F V O T ~ R P T G E F M F E A GUI GCT GGG GCClATC AAG GCG GAC CAT GTG TCA ACT TAT GCC GCG TTT GTA CAG ACG CAT AGA CCA ACA GGG GAG TTT ATG TTT GAP. T i l E

I

D

l

I

A

K

D

E

H

E

A

V

H

V

S

l

C

I

- 1 1.1

-4

G

Q V

V

A

N E

L F F

F N

V P

G D

P O

S S

t

n

G

F

F

I O E D E M P Y V D L O K K E T V Y H L f f F G O A F S F E A GAT GAA GAT GAG ATG TTC TAT GTG GAT CTG GAC M G AAG GAG ACC GTC TGG CAT CTG GAG GAG TTT GGC CM GCC TTT TCC TTT GAG GCT O

G

G

L

A

N

l

A

l

L

N

N

N

L

N

T

L

l

O

R

S

K

H

T

O

A

l

N

I

~

D

181 P

CCC

271

N T L I C H I D K F F P P V AAC ACC CTC ATC TGC CAC A T 1 GAC AAC TTC TTC CCA CCA CTC

361

L N V T Y L C N G E L V T E G V A E S L F L P R T D Y S F H CTC AAC GTC ACG TGG CTG TGC AAC GGG GAG CTG GTC ACT GAG GGT GTC GCT GAG AGC CTC TTC CTG CCC AGA ACA GAT TAC AGC TTC CAC

651

K F H Y L T F Y P S A E D F Y D C R V E H W t L O P P L L K AAG TTC CAT TAC C T C ACC TTT GTG CCC TCA GCA GAG GAC TTC TAT GAC TGC AGG GTG GAG CPC TGG GGC TTG GAC CAG CCG CTC CTC AAG

561

LAG GGC GGG CTG GCT AP.C ATT GCT ATA TTG M C AAC M C TTG AAT ACC TTG ATC CAG CGT TCC MC CAC ACT CAG GCC ACC MCIGAT I

P E V T V F P K E P V E L G O CCT GAG GTG ACC GTG TTT CCC AAG GAG CCT GTG GAG CTG GGC CAG

I V ' G T V L 1 I i K 5 L R S ATC GTG GGC ACC GTC CTC ATC A T A ~ A A G T C T CTG CGT TCT

G GGC

P

CCC

H D CAT GAC

P

R

ccc CCG

A 0 G T L * GCC CAG GGG ACC CTG TGA AATACTGTAAAGGTGACMAATA

726

TCTGAACAGMGAGGACT 1AGGAGAGATCTGAACTCCAGCTGCCCTACAAP.CTCCATCTCAGCTTTTCTTCTCACTTCATGTGNIAACTACTCCAGTGGCTGACTGAATT~CTGACCCT 845 TCAAGCTCTGTCCTTATCCATTACCTCAAAGCAGTCATTCCTTAGTAAAGTTTCCAACAAATAGAAATTAA1GACACTTTGT.TAGCACTAATATGGAGATTATCCTTTCATTGAGCCTT TTAlCCTCTGTTClCCTlltAAGAGCCCCTCACTGTCACCTlCCCGA~ATACCCTAAGACCAAtAAATAClTCA~TATTTCAG-pol~A

964

1

a2

Ic

TM

13'"T

104R

FIG.16. Sequence of cDNA clone of the DP a-chain. The domains are indicated to the right. From AuEray et al. (1984).

96 29

I?! 5t

79

246

sa 206 92

156 I04 2129 31

AS" MT

GGA CAC GAG GM ACA GCT C I Y Cln C l U GlU

lbhr

Ala

M AS"T CGA GlY

ACC

m

CAG

rnr Phs GIn

306

151

:BI

iw 531 229

$31 722 1121 114*

FIG. 17. The nucelotide and predicted amino acid sequence of a D P p cDNA clone and of the first domain exon of a genomic DP p clone (cosII-412).Asterisks denote nucleotide substitutions between the two sequences. Amino acid replacements are underlined. Cysteine residues are boxed, as are two putative attachments sites for N-linked oligosaccharides at Asnlg and Amg8. From Gustafsson et al. (1984a).

H U M A N CLASS I1 MOLECULES

47

acid position 60, one at position 50 and one at position 6. Thus, the information that is available at the present time is lacking in the amino terminal portion of the D P @-chain. Gustafsson et u1. (1984a) have isolated a cDNA clone of D P @ as well as a genomic clone. The composite is shown in Fig. 17. Comparison of the three available (none is complete) sequences of DP @ (Gustafsson et d., 1984a; Long et d., 1984; Roux-Dosseto et al., 1983) is complicated for the same reasons detailed in Section VIII on DQ (Y and @ comparisons, that is, allele vs locus cannot be adequately addressed. However, assuming these are allelic, they are remarkably similar with greater than 95% homology in the regions that can be compared. The basis upon which these structures are called DP derives from four sources of information. First, those that approach the amino terminus, agree with the sequence of Hurley et al. (1982b) that was deduced with the monoclonal I-LR1 that has been shown to be D P specific. Second, Southern filter hybridization among individuals who are homozygous in the HLA-DR subregion but differ in the HLA-DP subregion shows the genes to segregate with HLA-DP genotype not HLA-DR genotype. Third, deletion mutants that have lost expression of DR or DQ but maintain DP expression preserve a restriction fragment length polymorphism pattern consistent with their DP phenotype, and fourth, within informative families, in at least some of these studies, siblings differing only in the HLA-DP subregion, but otherwise being HLA identical, exhibit (using these clones as probes) a restriction fragment length polymorphism which correlates with HLA-DP. Thus, the sum total of the information would suggest that these are, indeed, the true HLA-DP encoded structures. As will be evident later, it is likely that there are two DP cx and two D P @ genes that are closely homologous. Which of the a/@pairs reacts with monoclonal antibodies such as ILRl and B7/21 and which have been isolated by cDNA cloning are subjects of future investigation. At the present time, there are no firm data that both of these genes are expressed and there are some data that one DP (Y and one DP @ gene may be nonfunctional. X. Supertypic Specificity Localization

The molecular bases of each of the supertypic specificities has been an active area of investigation during the last 5 years. This section attempts to summarize a contemporary view on localizing the supertypic specificities to one or more of the three groups of class I1 molecules, DR, DQ, and DP. As previously discussed, the supertypic specificities may be most logically grouped into two series, MB and MT (see Table V). The MB series has been well documented to reside on DQ molecules (Goyert et al., 1982; Giles et al., 1983; Tanigaki et al., 1983a). In the cases of DQwl and DQw3, mono-

48

ROBERT C . GILES AND J . DONALD CAPRA

clonal antibodies have been utilized to isolate molecules whose amino acid sequences are clearly DQ (Bono and Strominger, 1982; Giles et al., 1983). Although the DQw2-bearing molecule has not been verified by primary structural analysis as being DQ, several laboratories have presented evidence suggesting this is the case (Karr et d.,1984). Confirmation of DQw2 localization on DQ awaits the description of a monoclonal antibody specific for DQw2 which can be used for isolation of the molecule for primary structural analysis. The second group of supertypic specificities is the MT series. Two members of the MT series appear to be identical to previously described D Q specificities on the basis of their DR associated distribution. MT1 appears to be indistinguishable from DQwl and MT4 appears to be indistinguishable from DQw3 (see Table V). The nature of the HLA-D region molecules that bear the MT2 and MT3 (recently renamed DRw52 and DRw53 respectively-see Table 11) supertypic specificities has been contorversial. Twodimensional gel electrophoresis and peptide mapping have identified MT determinants on DR molecules (Markert and Cresswell, 1982; Tanigaki et al., 1983a; Goyert et al., 1983; Karr et al., 1982; Koning et d.,1984) on DQ molecules (Goyert et al., 1983; Karr et al., 1982 ) and even on a third set of molecules related to, but distinct from, DR and DQ, known as BR (for “Buffalo-Rome”)(Markert and Cresswell, 1982; Tanigaki et at., 1983a,b). In addition, two studies have shown that there may be a sharing of MT serologic determinants between DR and DQ molecules (Goyert et al., 1983) and between DR and BR molecules (Tanigaki et al., 1983a). Three different constructs have emerged from these studies to explain the MT2 and MT3 specificities: (1)MT2 and MT3 reside on DR molecules, (2) MT2 and MT3 reside on DQ molecules, and (3) MT2 and MT3 reside on molecules that are distinct from DR and DQ and are the products of an additional locus. Since the majority of studies used complex alloantisera in their analyses, additional reactivities directed against specificities other than MT might have added to the number of molecules observed. Recently, Hurley et al. (1984) have analyzed the biochemical bases of the DRw52 and DRw53 serologic specificities using two monoclonal antibodies with MT-like specificity. These monoclonal antibodies, I-LR2(DRw52-like) and 109d6(DRw53-like), were used to define the molecules bearing these specificities from a set of homozygous cell lines. I-LR2- and 109dBreactive molecules were compared to DR, DQ, and in some instances DP molecules isolated from the same cell line by inhibition of cell surface fluorescence or cytotoxicity, as well as amino acid sequence analysis and peptide mapping. Partial amino-terminal amino acid sequences of DR (203-reactive), DRw52 (ILR-2-reactive), and DRw53 (109d6-reactive) molecules were determined (Tables XIV and XV). Table XIV compares the or-chain sequences of MT and DR molecules from DR3, 4, 5, and 7 cell lines. Table XV compares the p-

49

HUMAN CLASS 11 MOLECULES TABLE XIV A M I N O ALIU SEQL~ENCE DATAW C I I A I N ~ AMINO-TEHMINAI. Poai tion

Cell line 127 (DR3) YBE (DH3) 127 YBE 3105 (DRS) DHI (DR5) 3105 DHI PRIESS (DR4) 3164 (DR4) PRIESS 3164 1 1 3 ~ (DR7) 7 3163 (DR7) 113~7 3163

Monoclonal antibody

12

13

22

24

26

32

I-LK2

F

F

F

F

F

L203

F

F

F

F

F

I-LR2

F

F

F

F

F

L203

F

F

F

F

F

109d6

F

F

F

F

F

L203

F

F

F

F

F

109d6

F

F

F

F

F

L203

F

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

F

F

F

F

TABLE XV AMINO-TERMINAL AMINO ACID SEQUENCE

DATA:

p CIlAlNS

Position Cell line

127 (DR3) YBE (DR3) 127 YBE 3105 (DR5) DHI (DR5) 3105 DHI PRIESS (DR4) 3164 (DR4) PRIESS 3164 1 1 3 ~ 7(DR7) 3163 (DR7) 113~7 3163 0

Not determined.

Monoclonal antibody

7

10

17

18

26

30

31

1-LR2

F

L203

F

I-LRZ

F

L203

F

Y Y Y Y Y Y Y Y

F

F F

F

F

F

F

F

F

F

F

F

L203

F

F

F

F

Y Y Y Y Y Y Y Y Y Y Y Y

F

F

Y Y Y Y F

109d6

109d6

F

Y N. 1).

F

F

F

N.D.0 Y

L203

F

Y Y

F

F

F

F

13

32

F Y F F F

Y Y Y Y Y Y Y Y

Y Y

50

ROBERT C. GILES A N D J. DONALD CAPRA

FIG. 18. Schematic diagram of the possible genetic organization of the I I U - D R subregion. Evidence presented to date suggests that the DR inolecules are encoded by a single a-chain gene and multiple p-chain genes within a haplotype. Some but not all ofthese p-chains bear MT (DRw52 and DRw53) serologic determinants. Some of the P-chains also bear DR serologic determinants. All three of these subsets of P-chains may not exist in all cells. From Hurley et al. (1984).

chains from the same preparations. Although only the tryosine and phenylalanine residues are compared, the predominant sequences of the a-and @-chainsof I-LR2 and 109d6-reactive molecules are identical to those derived using the anti-DR monoclonal antibodies. These data show that the major populations of molecules bearing MT2 and MT3 (DRw52 and DRw53) determinants are indistinguishable from DR molecules. The BR molecules have been shown to exhibit homology with DR molecules and could represent one of the multiple DR subsets. Sequence analysis of the BR molecules should establish their relationship to the DR molecules and definitively establish if an additional HLA-D subregion exists. A general model which relates DR and MT specificities is shown in Fig. 18. XI. Invariant (Gamma) Chain

Both mouse and human class I1 antigens are associated intracellularly with a family of basic invariant polypeptides which were first demonstrated by two-dimensional gel electrophoresis (Jones et al., 1978; Owen et al., 1981; Kvist et al., 1982; Machamer and Cresswell, 1982). These polypeptides arise through various processing events and are encoded by a single gene which has recently been mapped to human chromosome 5 (Claesson-Welsh et al., 1984). The invariant (or y) chain, a transmembrane protein, is noncovalently associated transciently with class I1 antigens during their transport to the cell surface (Kvist et al., 1982; Claesson and Peterson, 1983; Claesson et al., 1983). A fraction of these invariant chains is thought to integrate into the plasma membrane independently of the class I1 molecules. There is evidence that the y-chain and DR @-chainsshare common antigenic determi-

5 1 ' Met A s p Aap G l n A r g C T ( ; C A G G C C C C C G G G G G G G ~ G C G G G G ~ A C A ~ G G C ~ C ~ T ~ C ~ ~ G G G G A G ~ G A ~ G ~ A ~ A G G A G G L i G M G C A G G A G C ~ G ~ C ~ G C A A G ACAC ~ C A CAG GAAG CGC C C A G ~ ~114 ~GAT Asp Leu I I c S e r As" GAC c r r ~ r KC c MC

A m

mr

G I Y Phe Ser Ile Leu V a l GGC rrr TCC A r c r r c G r G

Thr V a l Thr S e r G I " A S "

r m crc

ACC

Pro

LCU

Leu

CCG

c n crc

rcc

CAG MC

net

GI"

;A

CAG CCG

Clu C l n Leu P r o net Leu Cly Rrq A r q P r o G l y A 1 1 Pro Clu S e r Lys C y s Ser Arg Gly GAG C A A CTG ccc arc c r c GGC ccc ccc c c r GGC GCC ccc GAG AGC AAG TGC AGC CGC GGA

A l a Leu T y r Shr GCC CTG rac ACA

35 204

T h r L r u L e u Leu A l a G l y G l n A l a T h r T h r A l a TYr Phe Leu T Y r G l n C l n G l n G l y Ar9 Leu Asp L y r Leu crG

65 294

LCU t l n 1.e" G l U A S n L e u A r g l e t LYS L e u P l O LYS Pro P r o Lys P r o Val S e r Lys M e t Arg Met Ala Thr CAG c r c GAC PAC CTC CGC A r G A A G c r r ccc AAG ccr ccc PAC CCT G r G AGC AAG Arc CGC LirG ccc ACC

95 384

hcr c r c

CK

crc

Gcr

ccc

CAC GCC ACC ACC

ccc

rAc Trc

crc

A I ~~ e u p r o ~ e tG

crG ccc Arc

I A~ I ~Leu CCA ccc crr.

G ~ C ACC A r A GAC

GGC CGG

crc

GAC AM

G I ~ I net A ~rhr~ 01"

pro G I " G

ccc

Thr ?let G l u Thr I I e Lisp Trp LyS Val P h e G L u S e r Trp

mrc

cnc

c n

I Pro ~ n e t C I ~ ~ AA SI ~r h r l L y s r y r ccc ATC CAG M r GCC ACC MG r A r GGC MC

rcG

m c rrr

MC

GAG AGC r G G

M e t HLs CRC

Arc

Phe Pro G l u Asn Leu A i r 9

rrc

I25 414

ASP HIS ATG ACA GAG GAC c A r

CAG CGG

V a l M e t H i s Leu Leu C l n A s n A l a A s p Pro Leu Lys Val T y r P r o P r o Leu Lys Gly Ser c r G A r G CAC c r G CTC CAG A A r G c r CAC ccc CTG A A G CTG rCic CCG CCA c r c A A G CGG ACC

ACC

r A c CAC CAG

I55 564

H ~ L S e u Lys A s n

crr

CCG GAG AAC c r G AGA CAC

MG

MC

185 654

H I S Trp Leu Leu P h e G I " Met Ser A r q HIS Ser Leu G l u G l n Lys

c m rGc c r c c r G

rrr

GM

Arc

AGC AGG CAC

rcc r r c

GAG CM

MG

P r o Thr L s p A l n P r o P r o Lys G l u S e r Leu C l u Leu G l u Asp P r o Ser Ser G l y Leu Gly Val Thr Lys G l n Asp Leu G l y Pro V a l P r o ACT GAC G c r CCA CCG MA GAG r c A c r G CAI\ crc GAG GK ccc rcr rcr CCG c r G G G r G r G ACC MC CAC mr crc GGC CCA GK ccc

215 ?44

net

216

ccc

862 C ~ c c ~ O c ~ G L ~ c G C r c r r c ~ ~ c ~ ~ c T ~ ~ c ~ c c c c c ~ c ~ c L i c c r ~ c ~ c ~ r ~ r c r r c c r c c c r r c ~ ~ c c c c cc ~ c r~c c~ cr c c cc cT~cr c G cAc c ~rcrccc~cccrcrt

Arc G

A

C

C

C

+

C

G

T

G

C

C

n

;

C

C

T

G

r

C

A

C

C

~

G

G

A

C

I

A

A

C

~

G

G

G

G

C

G

r

G

A

G

G

K

902 C

C A G G M G T C G C C M M G ~ A G A C A G A r C C C C G W r C ~ A C A r C A C A G C A ~ C C T C C A A C A C M ~ G C r C C A A G A C C r A G G C r C A T G G A C G A G A r G G G A A G G C A C A G G G A G M G G G A r M C C1102 C

r ~ c ~ c c c ~ C ~ c c c c ~ U ; C T c c A c a r c c r c ~ C T c ~ c r c r c c c c r c c ~ ~ c c r r ; c c c r r c c c r r r r c r ~ c c c r ~ r r r ~ c c T ~ c ~ ~ c c r x ~ c c c ~ c r c r c r r c c c ~ rI222 ccccA~ArcAcr CCCCM~AA.CAGCTAA~r-CACCCI7CCCrCC~CC~CCCCCCCCCCCCCCCCCCCrGCAG

I281

FIG. 19. Nucleotide sequence and predicted amino acid sequence of human y-chain. Boxes, in order of appearance, denote the putative initiation codon, the two carbohydrate addition sites, and the stop codon. Amino acid residues forming the putative transmembrane region are underlined. From Claesson et af. (1983).

52

ROBERT C. GILES A N D J. D O N A L D CAPRA

nants (Finn et al., 1983), although there is no amino acid sequence homology between them. The exact function(s) of the invariant chain is undetermined. Some have suggested that it regulates intracellular transport of class I1 antigens, while others postulate that it may prevent the formation of molecules composed of a and p subunits encoded within different subregions (e.g., DR a with DQ p) (Sung and Jones, 1981; Claesson and Peterson, 1983). Recently Peterson has proposed a role for y-chain in the genesis of the high mannose carbohydrate moiety on class I1 a-chains. cDNA clones corresponding to the human invariant chain have been isolated and characterized (see Fig. 19). The nucleotide sequence of the cDNA clone which corresponds to the entire translated portion of the invariant chain demonstrates (1)that the amino-terminus of the y-chain resides on the cytoplasmic side of the membrane (Fig. 20) and (2) that the invariant chain lacks an amino-terminal signal sequence (Claesson et al., 1983; Long et al., 1983). These data together with the distribution of carbohydrate moieties suggest that the invariant chain has a reversed membrane orientation as Coon

a

P

N"a

FIG. 20. Proposed membrane orientation of a-,p-, and y-chains of class 11 molecules. Cysteines (C) and asparagine-linked carbohydrate moieties (CHO) are indicated. The nonglycosylated tails reside on the cytoplasmic side of the membrane. From Claesson et al. (1983).

H UMAN CLASS I1 MOLECULES

53

compared to class I1 molecules. The proposed membrane orientation of a-, p-, and y-chains is shown in Fig. 20. XII. Genes

The utilization of molecular biological approaches in the area of class I1 genetics and biochemistry has had a nionuinental impact in a relatively short period of time. The isolation of cDNA and/or genomic clones encoding HLA-D region molecules, coupled with Southern filter hybridization analysis of human DNA, has contributed greatly to our current understanding of the number and organization of the genes. In addition, DNA sequence studies have been our major source of information concerning the primary structures of class I1 molecules. This section summarizes these studies that have led to the current model of genetic organization within the HLA-D region. The organization of the human class I1 genes presented in Fig. 21 is the result of a number of separate studies. Both the a and p subunit products of each subregion have been shown to be encoded by genes which map to the short arm of the sixth chromosome (Lee et al., 1982; Trowsdale et al., 1983; Auffray et al., 1983; Bohme et al., 1983; Morton et al., 1984). The exact order of these blocks of genes (subregions) comes from studies involving

DPP2

wa2

WBl

+3'

5'

FIG.21. Genomic organization of the human HLA-D region. The arrows indicate direction of transcription where known. Where genes are indicated in dotted rectangles they have not been formally linked by overlapping DNA fragments. This is a consemiis model which involves inany assumptions (see text for detail).

54

ROBERT C. GILES AND J. DONALD CAPRA

deletion mutants (Kavathas et al., 1980a,b, 1981) and recombinations within members of a family. Mapping HLA-DP centromeric to HLA-DR/DQ was firmly established using a series of haplotype loss mutant B cell lines (see Section V ) . Mapping HLA-DQ in relationship to HLA-DR is less certain, but recent studies of a family in which one member possesses a rare crossover event suggests that HLA-DQ is centromeric to HLA-DR ( E . Moller, personal communication). In almost all cases the order of the genes within each subregion may be reversed with regards to the centromere. In general cDNA cloning not only requires a large number of cells synthesizing the product at a high level, but also requires serological reagents which are capable of detecting the product in an in vitro translation of the cDNA clone. Some protein sequence data of the product are generally needed to compare to the cDNA sequence in order to identify the cDNA clone. Because these requirements for cDNA cloning have only recently been worked out for class I1 molecules, the first molecular biologists in the field had to overcome many difficulties before obtaining class II-specific cDNA clones. The initial human class I1 cDNA probe isolated was specific for the HLA-DR a-chain. Lee et al. (1982) used a polyspecific heteroanti-DR serum to precipitate DR a-chains from an in vitro translation system. Sequencing of one cDNA clone permitted its identification by comparison with DR a-chain protein sequence data. Likewise, Wiman et al. (1982b) utilized an antiserum against the p subunit (obtained by immunization of a rabbit with purified P-chains) to isolate cDNA clones specific for the HLA-DQ Pchain. Several innovative approaches for cloning sequences corresponding to proteins expressed at low levels followed shortly thereafter. One approach made use of a DR a-chain-specific monoclonal antibody to purify polysomes which contained the specific mRNA (Korman et al., 1982a). This highly purified mRNA could then be used to clone the cDNA directly or to screen a cDNA library made from poly(A)+ selected mRNA. Another approach that proved successful was the use of synthetic oligonucleotides complementary to portions of the DR a-chain. These oligonucleotides could be subjected to primer extension on B cell membrane-bound poly(A) mRNA templates. Additional sequence data obtained by this method led to the synthesis of a longer oligonucleotide that was then used to screen a cDNA library. Another DR a-chain cDNA was obtained using this approach (Sood et al., 1981). Long et al. (1982) and Wake et al. (1982a) approached the cloning problem by injecting B cell mRNA into frog oocytes which led to the synthesis of the complete DR molecule which was easily recognized by an antiserum raised against the native molecule. This sophisticated assay system could then be used to identify a-and P-chain-specific cDNA clones in a complementation assay in which a and p mRNAs were injected separately. In addition to these (and other) cDNA clones, several genomic clones have +

H U M A N CLASS I1 MOLECULES

55

been isolated from either phage or cosmid libraries of human DNA. The cDNA clones have the advantage of looking only at those loci in the genome which are likely expressed (i.e., not pseudogenes). On the other hand, genomic clones have allowed for the elucidation of the exodintron organization of class I1 genes (Lee et al., 1982; Gorski et al., 1984; Schamboeck et d . , 1983; Das et al., 1983; Larhammar et al., 1983b; Trowsdale et al., 1984). Examples of the sequences of two class I1 genes are shown in Figs. 22 (DR a) and 23 (DQ p). Shown in Fig. 24 is the organization of a prototypic class I1 gene. The structure is very reminiscent of the exodintron organization among class I genes suggesting an evolutionary relationship between the two. The first exon corresponds to the 5' untranslated and signal sequences. The second and third exons encode the two extracellular domains, while the fourth exon encodes a hydrophilic connecting peptide, the hydrophobic transmembrane region, the intracytoplasinic carboxy terminus, and a few nucleotides of the 3' untranslated region. The rest of the 3' untranslated region is present in the fifth exon. Numerous cosmid clones have been isolated which contain a and/or p genes of either DR, DQ, or DP. Examination of these clones by Southern filter hybridization, restriction enzyme mapping, and DNA sequence analysis has been instrumental in determining precise distances between certain genetic loci as well as the orientation of their transcription. For example, Okada et at. (1984) have demonstrated that one of the D Q a and one of the DQ p genes are separated by approximately 10 kb of DNA and are transcribed in opposite directions (shown arbitrarily as DQ pz and DQ a2in Fig. 21). This orientation is identical to that found for the a and p genes of the murine Z-A subregion. Another cosmid isolated carries DQ-like genes (DX a)and another D Q p genes. A large portion of the DX a gene has been sequenced (Auffray et al., 1984). The introdextron organization of the gene is analogous to the other class II genes and apparently contains no unusual features which might prevent its expression. Furthermore, carbohydrate attachment sites are found in identical positions to those found in the DR and DQ a-chains. The DX a gene may demonstrate little or no allelic polymorphism as determined by restriction endonucleases. By comparison, using identical enzymes, the DQ a gene has been shown to display one of three restriction fragment length polymorphisms depending upon the HLA-DR haplotype being examined. Nevertheless, a comparison of the amino acid and nucleotide sequence of two DQ a alleles (DQwl.6 and DQw3.4) and the DX a gene shows that each sequence differs from either of the other two at approximately 25 amino acids and 50 nucleotides. Therefore at the sequence level the two DQ alleles are as similar to DX as they are to each other making it impossible to distinguish between allelic and nonallelic pairs on this basis.

56

rO,.

,^,

ROBERT C . C.II.ES A N D J. DONALD CAPRA

9 .11 1 .

c r r r r r r G I T o r r c r r G G G I G G I l r i r r r G G ~ ~ ~ ~ ~ ~ ~ ~5O.O ~ ~ I ~ ~ ~ G~TGI*TITITCTTITII*11CTIGIITIC

5972

FIG. 22. DNA sequence of the human DH a gene. The amino acid sequence encoded by the exons are shown above the DNA sequence. The 3' untranslated region is underlined. The region ofAlu sequences, the CAT box, the TATAA box (promoter), and the mRNA transcription initiation site (CAP site) are marked. Asterisks, stop codon; polyadenylation signals are also indicated. From Das et al. (1983).

H U M A N CLASS I1 MOLECULES

57

Several cosmid clones containing DP-like genes have also been identified and studied. Okada et al. (1984) have reported one cosmid clone of about 30 kb of DNA that contains two D P (x genes and one D P (3 gene in the order of a-p-a. Larhammer et al. (1984) have similar results with two overlapping cosmids. Likewise Trowsdale et al. (1984) have identified three overlapping cosmid clones which contain coding sequences for two DP a and two DP @ genes (see Fig. 21). In general, within a subregion, the class I1 genes are about 10 kb apart. The subregions, themselves, have not been formally linked so the distance between, for example, DP and DQ is not known. Comparison of the nucleotide and predicted amino acid sequences of these class I1 genes allows for an analysis of their evolutionary relationships. In particular, the predicted amino acid sequences of the a-chains from each of these subregions can be compared to one another and to the murine I-E and I-A a-chains (Table XVI). This kind of comparison clearly demonstrates that the DR and DQ a-chains are the human equivalents of the murine I-E and I-A a-chains, respectively. However, in a similar comparison the D P a subunit appears to be equally similar to each of the other a-chains (54-61% homologous). A domain by domain comparison of DP a with each of the other a-chains does not allow for a simple conclusion as to the evolutionary relationship of D P a to these other molecules. An excellent review of many of these comparisons is given by Gustafsson et al. (1984b). Class I genes of mouse and human have been successfully transfected into mouse L cells. (Goodenow et al., 1983; Lemonnier et al., 1983). In these studies the product of a single transfected gene after association with a second chain, &-microglobulin, already present in the mouse L cells, was expressed on the surface of the L cells and was able to function as a restriction element in T cell-specific cytotoxic responses against viruses. Likewise, class I1 genes of both species have been successfully transfected into mouse L cells (Rabourdin-Combe and Mach, 1983; Malissen et at., 1983). Atthough these expressed molecules were recognized by appropriate monoclonal antibodies they were apparently not able to serve as restricting elements in antigen-specific T cell proliferation assays. Expression of functional class 11 molecules has recently been accomplished by transfection of murine class I1 genes into B lymphoma cell lines (Germain et al., 1983; Ben-Nun et al., 1984). These lymphomas most likely have the distinct advantage of possessing the cellular inachinery necessary to assemble functional class I1 antigens. However, similar attempts to obtain expression of functional HLAD region molecules following transfection has been more difficult. Recently, Gillies et al. (1984) have identified a cell type-specific transcriptional enhancer element associated with the mouse I-E @ gene which is likely to play an important role in the regulated expression of class I1 genes. It is likely that similar elements exist in the human HCA-D region genes.

i20

240

>bO

.

. I9 465

561

681 902 912

1042 Llbl 12e2

1402 1521

1641 1762

inel I? 1991 42

1081 12 1171

PI 1270 1190 2510 1b10 1750

2e10 2990 1110

1230

1150 1470

1590 3710

FIG.23. Nucelotide sequence and translated amino acid sequence of a DQ p gene. Underlined in the 5‘ part of the gene are the putative “CAT,” “TATA,” and “cap” elements, with an arrow marking the putative cap site. Cysteine residues and the glycosylation site are within boxes. The nonfunctional cytoplasmic exon is underlined, as is the 3’ untranslated region. The polyadenylylation signal is within a box. From Larharnmar et al. (198313).

60 r 0

ROBERT C. CILES A N D J. DONALD CAPRA

1

I

I

I

5’ UT region and sianal

I I I 1 7 kb ‘Transmembrane First Second and cytoexternal external DIasmic 3‘ UT

peptide

ao

I

2

1 ALU

I

3

1

1

4

I

1

I

5

H LA-DR P 34Y-f amino acids , 2 Sfi lj8‘229 FIG. 24. Organization of a prototypic human class I1 gene. From Das et al. (1983).

Despite these difficulties, expression of human class I1 genes in appropriate cell types promises to simplify current attempts to correlate HLA-D region gene products with their functional roles in the immune response. Much of the work on gene expression has been reported at meetings, but is yet unpublished. Bodmer’s group has transfected cosmids containing the HLA-DP genes in mouse L cells and successfully shown expression by twodimensional gel electrophoresis after immunoprecipitation with specific monoclonal antibodies. Coupled with the techniques of exon shuffling and site directed mutagenesis, transfection of HLA-D region genes into appropriate cells should provide the necessary approaches to localize allodeterminants, functional domains, and potential epitopes which may be directly involved in increased susceptibility to certain diseases. XIII. Restriction Fragment length Polymorphism

The technique of “DNA typing” by examining restriction enzyme fragment length polymorphism (RFLP) has generated considerable attention in many laboratories. Within the confines of class I1 molecular biology, this technique examines structural polymorphism of DNA by Southern filter hybridization of restriction enzyme fragments with class II-specific DNA probes. In general, the use of a battery of restriction enzymes and the appropriate HLA-D region probes permits the identification of haplotypespecific patterns (Wake et ul., 1982b; Trowsdale et al., 1983; Auffray et al., 1983; Bohme et al., 1983). Under conditions of “low stringency” essentially all (Y genes and all p genes cross hybridize-indeed, this is how many of these genes were initially cloned. However, fortunately, conditions can be found such that probes will only detect genes in a single subregion. As discussed previously, there is only approximately 50-60% homology be-

TABLE XVI COMPARISON OF THE AMINOACID

a l (85 amino acids) DQ

I-A

DR

a2

I-E

DQ

SEQUENCES OF

(94 amino acids)

I-A

DR

HUMANA N D MURINECLASS 11 a-CIIAINS' CP/TM/CY (51 amino acids)

I-E

DQ

I-A

DR

I-E

Total (220 amino acids)

~

~~

DP DQ

I-A

DR

40

49 51

48 41 45

43 37 38 65

64

60 68

67 59 55

63 60 56 76

29

29 46

27 30 29

20 19 21 30

DR

I-A

DQ

133

138 165

I-E

~~~

142 140 129

126 116 115 171

~~~~~

The number of identical amino acid positions in each domain is indicated for each combination of sequences. The number of amino acids in each domain is indicated. From Auffray et al. (1984). a

62

ROBERT C . GILES A N D J. DONALD CAPRA

tween subregions but 90-95% homology within a subregion. Thus, a DQ a probe will detect D Q a1and DQ a2but not DP (Y or DR (Y if the conditions are appropriate. An example of the approach is shown in Fig. 25. The probe used is HLADQ p. The number of restriction fragments seen indicates that there must be 2-5 DQ p genes in the genome. Recall this experiment does not address the number of these genes that are expressed. Nonetheless, note that each DRw type (shown across the top of the autoradiogram) has a distinct pattern when both enzyme digests are examined. For example, while Dw4 and Dw8 have an identical pattern with EcoRI (lanes 12 and 16), they are different with BamHI (lanes 4 and 8). Lane 13 contains DNA digested with EcoRI from a 3/4 heterozygote. Note that the pattern is the sum of lanes 12 and 14. Results of “DNA typing’ indicate that in some instances gentoypic polymorphism of class I1 genes is greater than phenotypic variation as evidenced by serological analysis. For example, in one study the DNA from DR3/3 genetically homozygous individuals could be split into two groups. In addition, it is anticipated that this technique will not only aid in classification of those individuals whic are untypeable by present techniques, but will also

FIG.25. Autoradiogram of blot hybridization analysis of homzygous HLA-Dw typing cells with an HLA-DQ p probe. From Owerbach et al. (1983). See text for details.

HUMAN CLASS I1 MOLECULES

63

identify categories of additional polymorphisins which may be biologically relevant. A distinct advantage of “DNA typing” over serological or cellular typing is the ability of this analysis to detect not only polymorphic restriction sites located within the coding region (which may or may not result in the expression of distinct molecules), but also polymorphic sites located within flanking sequences or introns. This kind of polymorphism is better appreciated when examining DNA from numerous individuals with a cDNA probe to a nonpolymorphic molecule such as the DR a-chain. Erlich et al. (1984) have hybridized BgZII-digested DNA with a cDNA probe specific for DR a. Their analysis has revealed three allelic restriction fragment lengths which map near the 3‘ end of the HLA-DR a gene. They propose to utilize these polymorphic restriction sites as genetic markers for the analysis of genetic predisposition to HLA-associated diseases. Caution should be taken in interpretation of data obtained from RFLP analyses as they relate to location of serological specificities andlor “disease genes.” For example Spielman et al. (1984) have reported RFLP patterns using the retriction enzyme EcoRI and the DQ a-chain cDNA probe that “corresponds with the HLA-DR cross-reactive serotypes (DR1, -2, -w6, DR3, -5; DR4, -7) that are associated with variation at the DC (DQ) locus.” Although the patterns obtained do indeed associate with the MTl, MT2 (DRw52), and MT3 (DRw53) specificities it is now known that only the MTl (DQwl) specificity is localized to the HLA-DQ subregion. The MT2 (DRw52) and MT3 (DRw53) specificities have been localized to variations at the DR p loci (Hurley et al., 1984). Correlations such as DQ a polymorphism with the MT specificities must be attributed to close linkage disequilibrium. These same precautions should be taken when attempting to correlate RFLP’s with “disease genes.” Nevertheless, the techniques of “DNA typing,” coupled with the current serological approaches for HLA typing, should prove to be very beneficial. XIV. Function

HLA-D region molecules have been shown to act as primary and secondary stimulators in mixed lymphocyte reactions, as targets for cytotoxic T cells, and as controlling elements in antigen presentation. In the mouse, the immune response to some antigens is restricted to determinants on I-E molecules, while to other antigens the response is restricted to determinants on I-A molecules. Similar delineatioris of function may exist among the multiple HLA-D region molecules as well. This section will review those studies demonstrating the functional roles of the multiple HLA-D region molecules.

64

ROBERT C. GILES AND J. DONALD CAPRA

Products of all three of the currently identified subregions of the HLA-D region can provide the major stimulus for lymphocyte activation in the mixed lymphocyte reaction and in the secondary proliferation of the primed lymphocyte typing assay (reviewed by Termijtelen et al., 1982). In PLTs the responding lymphocytes have been shown to recognize DR antigens (Bach et al., 1979; Inouye et al., 1980; Pawelec and Wernet, 1980; Zeevi et al., 1982), MT determinants (Zeevi et al., 1982), DQ determinants (Zeevi et aZ., 1982), DP antigens (Shaw et al., 1980; Bach and Reinsmoen, 1982; Pawelec et al., 1982a,b), and HLA-D determinants distinct from currently defined serologic specificities (Zeevi et al., 1983; Eckels and Hartzman, 1981; Pawelec et al., 1981, 1982a). Human class I1 molecules were shown to be a cytotoxicity target in a mouse anti-human xenogeneic system by Engelhard et al. (1980). At least two studies have supported this finding using cell-mediated lysis between HLA-A, B-matched, HLA-D mismatched individuals to demonstrate that HLA-D region antigens can serve as targets for CTL (Feighery and Stastny, 1979; Albrechtesen et at., 1979). Krensky et al. (1982) subsequently demonstrated that Daudi, a cell line that expresses DR antigens but no class I antigens, could be used to generate long-term DR-specific human CTL lines. Their findings not only suggested that DR could serve as a target antigen for long-term allogenic CTLs but that these antigens activated OKT4+ cytotoxic cells rather than the traditional OKT8+ effector cells activated by class I molecules. The HLA-D region also encodes restriction determinants utilized by T cells when responding to foreign antigens (Bergholtz and Thorsby, 1977, 1979; Hansen et al., 1978). A number of investigators have described experiments in which HLA-DR molecules served as restricting elements in antigen-specific T cell responses (reviewed by Thorsby et al., 1982). Several groups of investigators have presented evidence suggesting that additional HLA-D region encoded specificities may serve as restriction elements in antigen-specific T cell proliferation. Berle and Thorsby (1982), Qvigstad and Thorsby (1983), and Ball and Stastny (1984a) have all demonstrated that T cells responsive to viral antigens appear to utilize MT-bearing molecules as restriction elements. Eckels et al. (1983) found that some T cell clones utilized HLA-DP-encoded restriction elements in response to certain viral antigens. Ball and Stastny (198413) have provided the initial proof that HLADQ molecules serve as restricting elements in antigen-specific T cell responses. They chose to study the polypeptide antigens GAT and (T, G)-A-L which in the mouse are recognized predominantly by T cells using Z-A-subregion encoded restriction determinants (Schwartz et al., 1976). One human T cell line specific for GAT utilized a restriction determinant on DQ molecules as evidenced by blocking studies with D Q specific monoclonal anti-

H U M A N CLASS 11 MOLECULES

65

bodies. The epitopes on the DQ molecules recognized by the T cell line appeared to be distinct from alloantigenic determinants currently defined by serology.

XV. Conclusion

The human HLA-D region controls the expression of cell-surface antigens involved in communication between lymphoid cells. This communication appears to be critically important in immune responsiveness as suggested by the linkage of disease susceptibility in humans to particular HLA-D region alleles or specificities. In order to understand how these HLA-D region molecules function in cellular collaboration and antigen presentation, it is important to elucidate the assortment of HLA-D region molecules found on the surface of immune response related cells. At least three structurally distinct HLA-D region molecules, DH, DQ, and DP, have been isolated from a single cell line. Coupled with information at the DNA level, this effectively divides the HLA-D region into at least three subregions encoding a minimum of six protein chains. It is very likely that up to 12 subunits are actually expressed. Studies have also attempted to define the biochemical basis for the serologically and functionally defined antigens bearing allospecificities. Based on structural hoinology with the murine I-region antigens, the DR antigens appear to be related to the murine I-E antigens while the DQ antigens are similar to the murine I-A antigens. The number of loci already described in the HLA-D region provide a large repertoire of cell-surface molecules which can be used in communication between immune response related cells. Transassociation of chains within a subregion has been shown to generate additional molecules. Additional mechanisms which generate polymorphism probably exist. The association between susceptibility to particular diseases and certain allelic products of HLA-D subregions suggests that roles played by the multiple HLA-D region molecules in cellular collaboration and antigen presentation are varied.

ACKNOWLEDGMENTS It is a particular pleasure to acknowledge the skilled secretarial assistance of Margaret Wright. Our colleagues at Southwestern Medical Schools, Drs. Peter Stastny, Carolyn Hurley, Gabriel Nunez, and Ted Ball, have provided many stimulating discussions. Speckdl thanks to Dr. Ieke Schreuder for her comments and corrections in Section 11. We are grateful for the skilled technical assistance over the years of Jan Mills, Michelle Firra, Priscilla Presley, and Cindy Clegg. The research in the authors’ laboratory was supported by NIH Grant AI-18922. Robert Ciles is a Fellow of the Arthritis Foundation.

66

ROBERT C . GILES A N D J. DONALD CAPRA

REFERENCES Accolla, R. S. (1983). J. E x p . Med. 157, 1053. Accolla, R. S. (1984).J. E x p . Med. 159, 378. Accolla, R. S., and Pierres, M. (1983).J. Immunol. 130, 283. Accolla, R. S . , Gross, N., Carrel, S., and Corte, G. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 4549. Albrechtsen, D., Arnesen, E., and Thorsby, E. (1979). Transplantation 27, 338. Alexander, H., Johnson, D. A,, Rosen, J , , Jerabek, L., Green, N., Weissman, I. L., and Lerner, R. A. (1983). Nature (London)306, 697. Auffray, C., Korman, A. J., Roux-Dosseto, M., Bono, R., and Strominger, J. L. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 6337. Auffray, C., Kuo, J., DeMars, R., and Strominger, J. L. (1983). Nature (London) 304, 174. Auffray, C., Lillie, J. W., Arnot, D., Grossberger, D., Kappes, D., and Strominger, J. J. (1984). Nature (London)308, 327. Bach, F. H., and Reinsmoen, N. L. (1982). Hum. Immunol. 5, 133. Bach, F. H . , Inouye, H., Hank, J. A,, and Alter, B. J. (1979). Nature (London) 281, 307. Ball, E. J., and Stastny, P. (1984a).Zmmunogenetics 19, 13 Ball, E. J., and Stastny, P. (1984b). Zmrnunogenetics, 20, 547. Ben Nun, A., Glimcher, L. H., Weis, J., and Seidman, J. G. (1984). Science 223, 825. Benoist, C. O., Mathis, D. J., Kanter, M . R., Williams, V. E., and McDevitt, H. 0. (1983). Cell 34, 169. Bergholtz, B., and Thorsby, E. (1977). Scand. J. Zmmunol. 6, 779. Bergholtz, B. O., and Thorsby, E. (1979). Scand. J. Immunol. 10, 267. Berle, E. J., Jr., and Thorsby, E. (1982). Scand. J. Zmmunol. 16, 543. Bodmer, W. F. (1981). Am. J. Hum. Genet. 33, 664. Bodmer, W. F. (1984). In “Histocompatibility Testing 1984” (E. D. Alberts, M . P. Baur, and W. Mayr, eds.). Springer-Verlag, Berlin and New York. Bohme, J., Owerbach, D., Denaro, M . , Lernmark, A., Peterson, P. A,, and Rask, L. (1983). Nature (London) 301, 82. Bono, M. R., and Strominger, J. L. (1982). Nature (London)299, 836. Bono, M. R . , and Strominger, J. L. (1983). Zmmunogenetics 18, 453. Brodsky, F. M. (1984). Zmmunogenetics 19, 179. Brodsky, F. M., Parham, P., and Bodmer, W. F. (1980). Tissue Antigens 16, 30. Carrel, S., Tosi, R., Gross, N., Tanigaki, N., Carmagnola, A. L., and Accolla, R. S. (1981).Mol. Zmmunol. 18, 403. Carroll, M. C., Campbell, R. D., Bentley, D. R., and Porter, R. R. (1984). Nature (London) 307, 237. Charron, D. J., and McDevitt, H. 0. (1980).J. E x p . Med. 152, 18s. Charron, D., Lotteau, V., and Turmel, P. (1984). In “Histocompatibility Testing” (E. D. Albert, W. R . Mayr, and M. P. Bauer, eds.). Springer-Verlag, Berlin and New York, in press. Chen, Y. X., Evan, R., Pollack, M. S., Lanier, L. L., Philips, J., Rousso, C., Warner, N., and Brodsky, F. M. (1984). Hum. Immunol., 10, 221. Claesson, L., and Peterson, P. A. (1983). Biochemistry 22, 3206. Claesson, L., Larhammar, D., Rask, L., and Peterson, P. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 7395. Claesson-Welsh, L., Barker, P. E., Larliammar, D., Rask, L., Ruddle, F. H., and Peterson, P. A. (1984). Zmmunogenetics 20, 89. Cook, R . , Vitetta, E., Uhr, J., and Capra, J. D. (1980).J. Zmmunol. 124, 1594.

HUMAN CLASS I1 MOLECULES

67

Corte, G., Calabi, F., Damiani, G., Bargellesi, A,, Tosi, R . , and Sorrentino, H. (1981).Nature (London) 292, 357. Crumpton, M. J., Bodmer, J. G., Bodmer, W. F., Heyes, J. M., Lindsay, J., and Rudd, C. E. (1984). In “Histocompatibility Testing 1984” (E. 11. Albert, M. P. Baur, and W. R. Mayr, ed.s). Springer-Verlag. Berlin and New York. Das, H. K., Lawrence, S . K., and Weissman, S. M. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 3543. de Kretser, T. A,, Crumpton, M. J., Bodmer, J. G . , and Bodmer, W. F. (1982). Eur. J. Zinmunol. 12, 214. de Kretser, T., Crumpton, M. J . , Bodmer, J. G . , and Bodmer, W. F. (1983). Mul. B i d . Med. 1, 59. DeMars, R . , Chang, C., and Rudersdorf, R. A. (1983). Hum. Immunol. 8, 123. Dick, H. (1982). Zmniunol. Today 3, 200. Dupont, B . , Braun, D. W., Yuuis, E. J., and Carpenter, C. B. (1980). Z n “Histocompatibility Testing 1980” (P. I. Terasaki, ed.), p. 955. UCLA Tissue Typing Laboratory, Los Angeles. Duquesnoy, R. J.. Marrari, M., and Annen, K. (1979). Transplant. Proc. 11, 1757. Eckels, D. D., and Hartzman, R. J. (1981). Hum. Imtnunol. 3, 337. Eckels, D . , Lake, P., Lamb, J., Johnson, A., Shaw, S., Woody, J., and Hartzman, R. (1983). Nature (London)301, 716. Engelhard, V. H . , Kaufman, J. F., Stroniinger, J. L., and Burakoff, S. J. (1980).J. Exp. Med. 152, 545. Erlich, H . , Stetler, D., Sheng-Dong, R., and Saiki, R. (1984). Fed. Proc. 43, 3025. Feighery, C., and Stastny, P. (1979).J . E x p . Med. 149, 485. Ferrone, E. S., Allison, J. P., and Pellegrino, M. A. (1978). Conteinp. Top. Mol. Ztnmunol. 7, 239. Finn, 0. J., Stackhouse, C. J., and Metzgar, R. S. (1983).J. E x p . Med. 158, 1344. Germain, R. N . , Norcross, M. A., and Margulies. (1983). Nature (London) 306, 190. Giles, R. C., Nunez, G . , Hurley, C. K., Nunez-Roldan, A., Winchester, R., Stastny, P., and Capra, J. D. (1983).J. E x ~ JMed. . 157, 1461. Giles, R. C., Chang, C. C., Hurley, C. K., DeMars, R . , and Capra, J. D. (1984a). Dis. Markers, 2, 127. Giles, R. C., Demars, R., Chang, C. C., and Capra, J. D. (19841)). Proc. Natl. Acad. Sci. U . S . A . ,in press. Giles, R. C . , Hurley, C. K., and Capra, J. D. (1984~). J. Immunol. 133, 1. Gillies, S. D., Folsoni, V., and Tonegawa, S. (1984). Nature (London)310, 594. Gonwd, T., A., Peterlin, B. M., and Stobo, J. D. (1983). Adu. Zmmunol. 34, 71. Goodenow, R. S . , Stroynowski, I., McMillan, M., Nicolson, M., Eakle, K., Sher, B. T., Davidson, N., and Hood, L. (1983). Nature (London)301, 1. Gorski, J . , Rollini, P., Long, E., and Mach, B. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 3934. Gorzynski, T. J . , and David, C. S. (1983). Mayo Clin. Proc. 58, 457. Golz, H., Kratzin, H., Thinnes, F. P., Yang, C . , Kruse, T., Pauly, E., Kolhel, S., Egert, G., Wernet, P., and Hilschmann, N. (1983). Iloppe-Seyler’s Z. Physiol. Chem. 364, 749. Goyert, S. M., and Silver, J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 5719. Goyert, S. M., Shively, J. E., and Silver, J. (1982).J. Exp. Med. 156, 550. Goyert, S. M., Moriuchi, J., Winchester, R., and Silver, J. (1983). Ifurn. Immunol. 8, 53. Groner, J. P., Watson, A. J . , and Bach, F. H. (1983).J. Exp. Med. 157, 1687. Grumet, F. C., Charron, D. J., Fendley, B. M., Levy, R., and Ness, D. B. (1980).J. Zmmunol. 125, 2785. Grumet, F. C., Pask, S. B., Ness, D. B., Fendly, B. M., and Engleman, E. G . (1983). Ifum. Zmmunol. 6, 63.

68

ROBERT C . GILES AND J. DONALD CAPRA

Gustafsson, K., Bill, P., Larhammar, D., Wiman, K., Claesson, L., Schenning, L., Servenius, B., Sundelin, J , , Rask, L., and Peterson, P. A. (1982). Scand. J. Zmmunol. 16, 303. Gustafsson, K . , Emmoth, E., Widniark, E., Bohme, J . , Peterson, P. A., and Rask, L. (1984a). Nature (London) 309, 76. Gustafsson, K., Wiman, K., Emmoth, E., Larhammar, D., Bohme, J . , Hyldig-Nielsen, J. J . , Ronne, H., Peterson, P. A., and Rask, L. (1984b). E M B O ] . , 3, 1655. Hansen, G., Rubin, B., Sorensen, S., and Svejgaard, A. (1978). Eur. /. Zminunol. 8, 520. Hansen, J. A,, Martin, P. J., Kamoun, M., Nisperos, B., and Thomas, E. D. (1981). Hum. Zmmunol. 2, 103. Hurley, C. K., Nunez, G . , Winchester, R., Finn, 0. J., Levy, R., and Capra, J. D. (1982a).J . Zmmunol. 129, 2103. Hurley, C. K., Shaw, S., Nadler, L., Schlossman, S., and Capra, J. D. (1982b).1. E x p . Med. 156, 1557. Hurley, C . K., Giles, R. C., and Capra, J. D. (1983a).Zmmunol. Today 4, 219. Hurley, C. K., Shaw, S., Nadler, L., Schlossman, S., and Capra, J. D. (198311). In “IR Genes, Past, Present, and Future” (C. W. Pierce, S. E. Cullen, J. A. Kapp, B. D. Schwartz, and D. C. Shreffler, eds.), p. 197. Humana Press, Clifton, New Jersey. Hurley, C. K., Nunez, G., Winchester, R., Finn, O., Levy, R., Stastny, P., and Capra, J. D. (1983~). In “Hybridomas and Cellular Immortality” (B. H. Tom and J. P. Allison, eds.), p. 83. Plenum, New York. Hurley, C. K., Giles, R. G., Nunez, G., DeMars, R., Nadler, L., Winchester, R., Stastny, P., and Capra, J. D. (1984).1.E x p . Med. 160, 472. Inouye, M . , Hank, J. A., Chardonnens, X., Segall, M., Alter, B. J., and Bach, F. H. (1980).J. E x p . Med. 152, 1435. Johnson, J. P., Meo, T., Riethmuller, G., Schendel, D. J., and Wank, R. (1982).J . E x p . Med. 156, 104. Jones, P. O., Murphy, D. B., Hewgill, D . , and McDevitt, H. 0. (1978).Zminunocheinistry 16, 51. Karr, R. W., Kannapell, C. C., Stein, J. A., Gebel, H. M., Mann, D. L., Duquesnoy, R. J., Fuller, T. C., Rodey, G. E., and Schwartz, B. D. (1982).1. Zmmunol. 128, 1809. Karr, R. W., Kannapell, C. C., Zangara, R. J., Goyert, S. M., Silver, J., and Schwartz, B. (1983).1. Exp. Med. 158, 1374. Karr, R. W., Alber, C., Goyert, S. M., Silver, J., and Duquesnoy, R. J . (1984).J . E x p . Med. 159, 1512. Kasahara, M., Takenouchi, T., Ikeda, H., Ogasawara, K., Okuyama, T., Ishikawa, N., Wakisaka, A,, Kikuchi, Y., and Aizawa, M. (1983). Zmmunogenetics 18, 525. Kaufman, J. F., and Strominger, J. L. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 6304. Kaufman, J. F., and Strominger, J. L. (1982). Nature (London) 297, 694. Kaufman, J. F., Anderson, R. L., and Strominger, J. L. (1980).1. E x p . Med. 152, 37s. Kaufman, J. F., Auffray, C . , Korman, A. J.. Shackelford, D. A,, and Strominger, J. (1984).Cell 36, 1. Kavathas, P., Bach, F. H . , and DeMars, R. (1980a). Proc. Natl. Acad. Sci. U.S.A. 77, 4251. Kavathas, P., DeMars, R., and Bach, F. H. (1980b). Hum. Zmmunol. 4, 317. Kavathas, P . , DeMars, R . , Bach, F. H., and Shaw, S. (1981). Nature (London) 293, 747. Kohler, G . , and Milstein, C . (1975). Nature (London) 256, 495. Koning, F., Schreuder, I., Giphart, M., and Bruning, H. (1984). Hum. Zmmunol. 9, 221. Korman, A. J., Knudsen, P. J., Kaufman, J. F., and Strominger, J. L. (1982a).Proc. Natl. Acad. Sci. U.S.A. 79, 1844. Korman, A. J., Auffray, C., Schamboeck, A., and Strominger, J. L. (1982b).Proc. Natl. Acad. Sci. U.S.A. 79, 6013. Krangel, M. S . , Orr, H. T., and Strominger, J. L. (1980). Scand. 1. Zmmunol. 11, 561.

H U M A N CLASS I1 MOLECULES

69

Kratzin, H., Yang, C. Y., Krusche, J . U . , atid Hilsclimann, N. (1980). Hoppe-Seyler’s Z. Physiol. Chem. 361, 1591. Kratzin, H., Yang, C.-Y., Gotz, H . , Thinnes, F. P., Kruse, T., Egert, G., Pauly, E., Kolbel, S., Wernet, P., and Hilschniann, N. (1983). Hum. Irnmunol. 8, 65. Krensky, A. M., Reiss, C. S., Mier, J. W., Strominger, J. L., and Burakoff, S. J. (1982). Proc. Nutl. Acad. Sci. U.S.A. 79, 2365. Kvist, S., Wiman, K., Claesson, L., Peterson, P. A., and Doberstein, B. (1982). Cell 29, 61. Lampson, L. A., and Levy, R. (1980).J. Itnmunol. 125, 293. Larhammar, D., Wiman, K., Schenning, L., Claesson, L., GustdFsson, K., Peterson, P. A , , and Rask, L. (1981). Scand. J. Irnniunol. 14, 617. Larhammar, D., Schenning, L., Gustafsson, K., Wiman, K., Claesson, L., Rask, L., and Peterson, P. A. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79, 3687. Larhammar, D. , Gustafsson, K., Claesson, L., Bill, P., Wiman, K., Schenning, L., Sundelin, J.. Widmark, E., Peterson, P. A,, and Rask, L. (198211). Cell 30, 153. Larhammar, D., Andersson, G., Andersson, M., Bill, P., Bohme, J.. Claesson, L., Denaro, M., Einmoth, E., Gustafsson, K., Harnmarling, U . , Heldin, E., Hyldig-Nielsen, J. J., Schenning, P. L. L., Servenius, B., Widmark, E . , Rask, L., and Peterson, P. A. (1983a). Hum. Iinmunol. 8, 95. Larhamniar, D., Hyldig-Nielsen, J. J., Serveneius, B., Andersson, G., Rask, L., and Peterson, P. A. (1983b). Proc. Natl. Acad. Sci. U.S.A. 80, 7313. Larhamniar, D., Gustafsson, K., Hyldig-Nielsen, J. J., Hammerling, U., Servenius, B., Rask, L., and Peterson, P. A. (1984). 11-2, HLA Cloning Meet., 3rd, Strasbourg Abstract. Lee, J. S., Trowsdale, J., and Bodmer, W. F. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 545. Lemonnier, F. A , , Le Bouteiller, P. P., Malissen, B . , Colstein, P., Malissen, M., Mishal, Z., Caillol, D. H., Jordan, B. R., and Kourilsky, F. M. (1983). J. Imtnunol. 130, 1432. Long, E. O., Wake, C. T., Strubin, M., Gross, N . , Accolla, R., Carrel, S . , and Mach, B. (1982). Proc. Natl. Acud. Sci. U.S.A. 79, 7465. Long, E. O., Strubin, M., Wake, C. T., Gross, N . , Carrel, S., Goodfellow, P., A c d a , R. S., and Mach, B. (1983a). Proc. Nutl. Acrid. Sci. U.S.A. 80, 5714. Long, E. O., Wake, C. T., Gorski, J.. and Mach, B. (1983b). E M R O J. 2, 389. Long, E. O., Gorski, J.. and Mach B. (1984). Nature (London) 310, 233. Machdnier, C. E. , and Cresswell, P. (1982). J. Immune/. 129, 2564. Makgoba, M. W., Fuggle, S. V., MeMichael, A. J., and Morris, P. J. (1983). Nature (London) 301, 531. Malissen, B., Steinmetz, M., McMillan, M., Pierres, M., and Hood, L. (1983). Nature (London) 305, 440. Markert, M. L., and Cresswell, P. (1982). 1. Inztnunol. 128, 2004. Mickelson, E. M., Nisperos, B., Layrisse, Z., Kim, S. J., Thomas, E. D., and Hansen, J . A. (1983). lrnmunogenetics 17, 399. Morton, C. C., Kirsch, I. R., Nance, W. E., Evans, G. A,, Korman, A. J.. and Strominger, J. L. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 2816. Nadler, L. M., Stashenko, P., Hardy, R., Pesando, J. M . , Yunis, E. J., and Schlossman, S. F. (1981). Hum. Irnmunol. 1, 77. Nepom, B. S., Nepom. G. T., Mickelson, E., Antonelli, P., and Hansen, J. A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 6962. Okada, K., Boxx, J. M., Levy, D., Spies, T., Sonner, J. M., Kappes, D. J., Auffray, C., and Strominger, J. L. (1984). H-2, HLA Cloning Meet., 3rd, Strasbourg Abstract. Owen, M. J., Kossonerghis, A.-M., Lodish, H. F . , and Crumpton, M. J. (1981).J. B i d . Chern. 256, 8987. Owerhach, D., Lernniark, A,, Rask, L., Peterson, P. A., Platz, P., and Svejgaard, A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80. 3758.

70

ROBERT C. GILES A N D J. DONALD CAPRA

Parham, P., Kipps, T. J., Ward, F. E., and Herzenberg, L. A. (1983). Hum. Zmmunol. 8, 141. Park, M. S., Terasaki, P. I., and Bernoco, D. (1980).In “Histocompatibility Testing 1980” (P. I. Terasaki, ed.), p. 572. UCLA Tissue Typing Laboratory, Los Angeles. Pawelec, G., and Wernet, P. (1980). Znimunogenettcs 11, 507. Pawelec, G., Rehbein, A , , Kahle, P., Muller, C., and Wernet, P. (1981). Hum. Zmmunol. 2, 315. Pawelec, G., Kahle, P., Muller, C., Rehbein, A., and Wernet, P. (1982a). Tissue Antigens 19, 269. Pawelec, G . , Shaw, S., and Wernet, P. (1982b). Zmmunogenetics 15, 187. Pawelec, G. P., Shaw, S., Ziegler, A,, Muller, C., and Wernet, P. (1982~). J . Zmmunol. 129, 1070. Pawelec, G., Shaw, S., Schneider, M., Rehbein, A., and Wernet, P. (1983). Zmmunogenetics 17, 179. Pierres, M., Rebouah, J.-P., Kourilsky, F. M., Dosseto, M., Mercier, P., Mawas, C., and Malissen, B. (1981).J. Zmmunol. 126, 2424. Pious, D., Hawley, P., and Forrest, G . (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 1397. Quaranta, V., Pellegrino, M. A., and Ferrone, S. (1981).J. Zn~munol.126, 548. Qvigstad, E., and Thorsby, E. (1983). Scand. J. Zrnmunol. 18, 299. Rabourdin-Combe, C., and Mach, B. (1983). Nature (London) 303, 670. Radka, S. F., Machamer, C., Cresswell, P., Kostyu, D. D., Ward, F. E., and Amos, D. B. (1983).J . Zm~nunol.130, 1863. Roux-Dosseto, M., Auffray, C., Lillie, J. W., Boss, J. M., Cohen, D., DeMars, R . , Mawas, C., Seidman, J. G . , and Strominger, J. L. (1983). Proc. Natl. Acad. Sci. U . S . A . 80, 6036. Royston, I., Omary, M. B., and Trowbridge, I. S. (1981). Transplant. Proc. 13, 761. Schamboeck, A., Korman, A. J., Kamb, A,, and Strominger, J. L. (1983).Nucleic Acids Res. 11, 8663. Scheming, L., Larhammar, D . , Bill, P., Wiman, K., Jonsson, A.-K., h s k , L., and Peterson, P. A. (1984). E M B O J . 3, 447. Schwartz, R., David, C., Sachs, D., and Paul, W. (1976).J . Zmmunol. 117, 531. Shackelford, D. A., and Strominger, J. L. (1980).J. Exp. Med. 151, 144. Shackelford, D. A., and Strominger, J. L. (1983).J. Zmmunol. 130, 274. Shackelford, D. A., and Strominger, J. L. (1984). Hum. Zmmunol. 9, 159. Shackelford, D. A., Lampson, L. A., and Strominger, J. L. (1981a).J. Immunol. 127, 1403. Shackelford, D. A . , Mann, D. L., van Rood, J. J., Ferrara, G. B., and Strominger, J. L. (1981b). Proc. Natl. Acad. Sci. U . S . A . 78, 4566. Shackelford, D. A., Kaufinan, J. F., Korman, A. J., and Strominger, J. L. (1982). Zmmunol. Reo. 66, 133. Shackelford, D. A., Lampson, L. A., and Strominger, J. L. (1983).J . Zmmunol. 130, 289. Shaw, S., Johnson, A. H., and Shearer, G. M. (1980).J. Erp. Med. 152, 565. Shaw, S., Kavathas, P., Pollack, M. S., Charmot, D., and Mawas, C. (1981). Nature (London) 293, 745. Shaw, S., DeMars, R., Schlossman, S. F., Smith, P. L., Lampson, L. A., and Nadler, L. M. (1982).J. E x p . Med. 156, 731. Silver, J., Swain, S. L., and Hubert, J. J. (1980). Nature (London) 286, 272. Sood, A. K., Pereira, D., and Weissman, S. M. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 616. Spielman, R. S., Lee, J., Bodmer, W., Bodmer, J., andTrowsdale, J. (1984).Proc. Natl. Acad. Sci. U . S . A . 81, 3461. Stastny, P., Capra, J. D., Hurley, C. K., Giles, R., DeMars, R., Nunez, G. Myers, L. K., and Ball, E. J. (1984).In “Histocompatibility Testing 1984” (E. D. Albert, M. P. Baur, and W. R. Mayr, eds.) Springer-Verlag, Berlin and New York. Steel, M. (1984). Dis. Markers, 2, 1. Steinmetz, M., and Hood, L. (1983). Science 222, 727.

HUMAN CLASS I1 MOLECULES

71

Sung, E., and Jones, P. P. (1981). Mol. Zminunol. 18, 899. Tanigaki, N., Tosi, R., Duquesnoy, R. J., and Ferrara, 6. B. (1983a).J. E x p . Med. 157, 231. Tanigaki, N., Tosi, R., Sagawa, K., Minowada, J., and Ferrara, G. B. (1983b). Zininunogenetics 17, 371. Termijtelen, A,, van Leeuwen, A , , and van Rood, J . J. (1982). Znmiunol. Reu. 66, 79. Termijtelen, A,, van den Berge, S. J . , and van Rood, J. J. (1983a). Hum. Zmniunol. 8, 11 Termijtelen, A., Khan, P. M., Shaw, S., and van Rood, J. J . (19831)).Zmrnunogenetics 18, 503. Thorsby, E., Berle, E., and Nousiainen, H. (1982). Zrninunol. Reu. 66, 39. Tosi, R., Tanigaki, N . , Gentis, D., Ferrara, G. B., and Pressman, D. (1978).J. Exp. Med. 148, 1592. Tosi, R., Tanigaki, N . , Sorrentino, R., Accolla, R., and Corte, 6. (1981). Eur. J . Zmmunol. 11, 721. Tosi, R., Tanigaki, N., Sorrentino, R., van Mourik, P., and Ivanyi, P. (1982). Zinmunogenetics 16, 187. Tosi, R:, Tanigaki, N., Sorrentino, R., Centis, D., and Ferrara, G. B. (1984).J . Zrnmunol. 132, 277. Trowsdale, J., Lee, J., Carey, J . , Grosveld, F., Bodmer, J., and Bodmer, W. (1983).Proc. Natl. Acad. Sci. U . S . A . 80, 1972. Trowsdale, J., Kelly, A., Lee, J., Carson, S., Austin, P., and Travers, P. (1984). Cell 38, 241. van Leeuwen, A., Termijtelen, A., Shaw, S., and van Rood, J. J. (1982). Nature (London) 298, 565. Van Rood, J. J., van Leeuwen, A., Parlevliet, J . , Termijtelen, A,, and Keuning, J. J. (1975). In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsin, ed.), p. 629. Munksgaard, Copenhagen. Wake, C. T., Long, E. O., Strihin, M . , Gross, N . , Accolla, R., Carrel, S . , and Mach, B. (1982a). Proc. Natl. Acad. Sci. U.S.A. 79, 6979. Wake, C. T., Long, E. O., and Mach, B. (1982b). Nature (London) 300, 372. Walker, L. E., and Reisfeld, R. A. (1982).Journ. Rial. Chem. 257, 7940. Walker, L. E., Ferrone, S., Pellegrino, M. A,, and Reisfeld, R. A. (1980). M a / . Zmrnunol. 17, 1443. Walker, L. E., Hewick, R., Hunkapiller, M. W., Hood, L. E., Dreyer, W. J., and Reisfeld, R. A. (1983). Biochemistry 22, 185. Wallin, J . , Bohme, J., Carlson, B., Moller, E., Peterson, P. A,, and Rask, L. (1984). I n “Histocompatibility Testing 1984” (E. D. Albert, M. P. Baur, and W. Mayr, eds.). Springer-Verlag, Berlin and New York. Watson, A. J., DeMars, R., Trowbridge, I . S., and Bach, F. H. (1983). Nature (London)304, 358. White, P. C., New, M. I., and Dupont, B. (1984). Proc. Natl. Acad. Sci. U . S . A . 81, 1986. Wirnan, K., Claesson, L., Rask, L., Tragardh, L., and Peterson, P. A. (1982a). Biochemistry 21, 5351. Wiman, K., Larhammar, D., Claesson, L., Gustafsson, K., Scheming, L., Bill, P., Bohme, J., Denaro, M., Dobberstein, B., Hammerling, U., Kvist, S., Servenius, B., Sundelin, J., Peterson, P. A,, and Rask, L. (1982b). Proc. Natl. Acad. Sci. U . S . A . 79, 1703. Winchester, R. J . , and Kunkel, H. J. (1979). Adu. Zmmunol. 28, 221. Yang, C.-Y., Kratzin, H., Gotz, H., Thinnes, F. P., Kruse, T., Egert, G . , Pauly, E., Kolbel, S., Wernet, P., and Hilschniann, N . (1982). Zloppe-Seyler’s Z. Physiol. Chem. 363, 671. Yunis, E. J., and Amos, D. B. (1971). Proc. Natl. Acad. Sci. U . S . A . 68, 3031. Zeevi, A , , Scheffel, C., Annen, K., Bass, G., Marrari, M., and Duquesnoy, R. J. (1982). Zmniunogenetics 131, 2777. Zeevi, A,, Annen, K., Scheffel, C., and Duquesnoy, R. J. (1983). Hum. Irnmnuno/. 6, 97. Zeevi, A , , Scheffel, C., Annen, K . , Bass, G., Marrari, M., and Duquesnoy, R. J. (1984). Zmniunogenetics 16, 209.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 07

The Complexity of Virus-Cell Interactions in Abelson Virus Infection of lymphoid and Other Hematopoietic Cells CHERYL A. WHITLOCK* AND OWEN N. W l l l E t *Department of Pathology, Stanford University, Palo Alto, California, and +Department of Microbiology and Molecular Biology Institute, University of California, Los Angeles, Los Angeles, California

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , . , . . . . , , , . , , . . , ,

11.

111.

IV.

V.

A. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. History of A-MuLV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . , , , C. Properties of v-ah1 and the Tyrosine Kinase Transforming Protein . . . . . . . . . D. Use of Anti-A-MuLV Antibodies to Study the c-ahl-Encoded Protein . . . . . . . E. The Pattern of c-abl Expression by Normal and Leukemic Cells . . . . . . . . . . . Neoplastic Transformation by A-MuLV . . . . . . . . . . . . . . . . A. Pre-B Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Techniques for Transformation of Mature B and Plasma Cells . . . . . . . . . . . . . C. A-MuLV-Transformed Cell Lines with Differentiation Potential . . . . . . . . . . . D. Target Cells in the Monocyte-Macrophage Lineage . . . . . . . . . . . . . . . . . . . . . E. Rapid Induction of Thymoinas By A-MuLV F. NIH 3T3 Fibroblasts: A Tool for Virus Stud Nonneoplastic Changes Induced by v-abl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Agar Colony Forination by Fetal Liver Erythroblasts . . . . . . . . . . . . . . . . . . . . B. Resistance of GM-CFC to Leukemia-Associated Inhibition Activity (L C. Lethality of v-abl: A Possible Role in “Hit and Run Transformation” The Coinplexity of Abelson Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Role for Helper Virus in Neoplastic Transformation by A-MuLV . . . B. Proliferating Cells Are the Targets for in Vitro Trailsforination . . . . . . . . . . . . C. Genetics of Susceptibility to Abelson Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . D. A-MuLV Transformation of Cells from Genetically Resistant Mice . . . . . . . . . E. Variations of the in Vioo Disease Process Induced by Mutants of A-MuLV . . F. Use of Site-Directed Antibodies to the Abelson Protein to Study A-MuLV Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Complexity of A-MuLV Transformation i n Vitro . . . . . . . . . . . . . . . . . . . . . . . . A. Early Biological Effects of v-ubl Expression s . . .. . . ... B. Preneoplastic Growth Properties of A-MuL C. Progression of Tumorigenicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Characterization of Cellular Changes Associated with A-MnLV Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Use of Cultured B Cell Lines to Study the Cellular Contribution to Neoplastic Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

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

. ....... . .

74 74 75 76 77 78 79 79 80 82 82 83 83 84 85 85 86 87 87 88 88 89 90 90 91 91 91 92 93 94 95

73 Copyright 0 19% by Academic P r a s , lac. All rights of reproduction in any form resewed. ISBN 0-18-022437-2

74

CHERYL A. WHITLOCK A N D OWEN N . WI'ITE

I. Introduction

A. OVERVIEW The pathway of B lymphocyte differentiation has provided one of the most versatile systems for studying the biochemistry of maturation, cell-cell interaction, and oncogenic transformation. A major reason for its versatility is the availability of continuous cell lines which provide the homogeneous cell populations necessary for biochemical analysis (Raschke, 1980a). The murine retrovirus Abelson murine leukemia virus (A-MuLV) has been particularly useful for studying the early stages of B cell development. Abelson virus rapidly transforms pre-B lymphocytes from mouse bone marrow or fetal liver that are at the stage of development during which the immunoglobulin heavy chain genes are being rearranged. Such cell lines have been used both to analyze the early stages of B cell ontogeny and to prepare and characterize serological reagents which recognize antigens expressed by pre-B and B cells (Coffman and Weissman, 1981a,b; Kincade et aZ., 1981; Coffman, 1982; Kung et al., 1982). Using Abelson virus, a large number of independently derived clonal cell lines can be generated in a short period of time (3 to 4 weeks). This is an important advantage for studying most biological phenomena and is especially important for examining stochastic processes like immunoglobulin gene rearrangements. Unfortunately, most of the A-MuLV-transformed cell lines are locked into a particular stage of maturation and cannot be induced by conventional methods to differentiate. However, rare cell lines have been discovered that retain the capacity to rearrange and sometimes express either the immunoglobulin heavy or light chain genes (Alt et al., 1981a,b; Boss, et al., 1981; Burrows et al., 1981; Lewis et al., 1982; Whitlock et al., 1983a; Akira et al., 1983; Sugiyama et al., 1983).These cell lines can hopefully provide the model systems for unravelling this complex differentiation process. The biological effects of A-MuLV extend well beyond transformation of pre-B lymphocytes. Cells from other stages of B lymphoid maturation as well as from other hematopoietic lineages exhibit a wide-range of biological changes after A-MuLV infection. These include stimulation of growth in the absence of true oncogenic transformation (Waneck and Rosenberg, 1981b), synthesis of growth factors (Twardzik et al., 1982), stimulation of resistance to growth inhibition factors (Broxmeyer et al., 1981), differentiation (Whitlock et al., 1983a; Ziegler et al., 1984), and even cell death (Ziegler et al., 1981; Goff et al., 1982b). An understanding of the complexity of biological effects induced by expression of the A-MuLV transforming gene, v-abl, may provide insight into the function of its normal cellular counterpart, c-abl. Our understanding of the process of A-MuLV transformation has under-

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

75

gone an interesting evolution over the past decade. The initial assumption was that simple expression of the A-MuLV transforming gene (v-ablj in the appropriate target cell was all that was required for transformation to proceed. This may be the case for NIH3T3 fibroblasts, for reasons which will be discussed below. For hematopoietic cell transformation, however, the process is more complex. Deletion mutants of A-MuLV that retain the capacity to transform NIH3T3 fibroblasts fail to transform lymphoid cells (Prywes et al., 1983; Wang, 1983; Wang et al., 1983; Watanabe and Witte, 1983). The helper virus used to package the A-MuLV has a marked effect on lymphoid cell transformation (Rosenberg and Baltimore, 1978; Scher, 1978). Also, for lymphocytes to exhibit a fully oncogenic growth phenotype, changes must occur in expression of cellular genes subsequent to v-abl expression (Teich and Dexter, 1978; Whitlock and Witte, 1981; Lane et al., 1982; Whitlock et al., 1983b). In certain cases, A-MuLV may act only to initiate or promote these cellular changes after which it is no longer needed-a process that has been termed “hit and run transformation” (Newmark, 1983). By adding to the complexity of Abelson virus transformation, these cellular changes provide an in vitro model for studying the progression of oncogenicity and for understanding initiation and promotion of tumors by agents other than oncogenic viruses. Several reviews have been published over the past 5 years on the biology and biochemistry of v-abl, c-abl, and the protein products they encode (Baltimore, 1981; Goff et al., 1981; Risser, 1982; Rosenberg, 1982; Witte, 1983; Wang, 1983; Wang et al., 1983). The purpose of this review is not to duplicate this information but to focus on two major areas concerning the complexity of A-MuLV interaction with the cell. One is the wide range of biological effects induced by A-MuLV as it is expressed in different hematopoietic cell types. The second is the process of oncogenic transformation by A-MuLV and how transformation is augmented or complemented by the expression of cellular genes. Many of the molecular processes in the cell that dictate what biological effect A-MuLV will induce probably are the same as those that determine the degree of oncogenicity the A-MuLV-infected cell will express. A great deal of research is currently focused at understanding the molecular processes involved in interaction of the cell with the v-abl gene. This review will hopefully provide a baseline of the complex biological effects of Abelson virus.

B. HISTORY OF A-MuLV Abelson murine leukemia virus was isolated in 1969 by Abelson and Rabstein (Abelson and Rabstein, 1969, 1970a,b) in an experiment where they infected mice with Moloney MuLV (M-MuLV). Before infection the mice were treated from birth with a corticosteroid, prednisolone. This caused

76

CHERYL A. WHITLOCK A N D OWEN N . WI’ITE

ablation of the thymus which is the normal target of M-MuLV transformation. Five out of 220 mice developed lymphosarcomas, but extracts from only one tumor induced nonthymic lymphomas when injected into normal mice (Abelson and Rabstein, 1969). Unlike M-MuLV, the virus isolated from this tumor (A-MuLV) rapidly induced lymphomas (20 to 40 day latent period) in athymic nu/nu mice (Raschke et al., 1975) and transformed lymphoid cells from bone marrow or fetal liver after in vitro infection (Rosenberg et al., 1975; Rosenberg and Baltimore, 1976; Pratt et al., 1977; Teich and Dexter, 1978). Abelson virus also differed from the parental M-MuLV in that it was a defective virus requiring a replication-competent helper virus for its passage (Scher and Siegler, 1975; Shields et al., 1979).

C . PROPERTIES OF v-abl A N D THE TYROSINE-KINASE TRANS FORMING PROTEIN Soon after isolation of A-MuLV, nucleic acid hybridization and serological analyses showed that A-MuLV resulted from recombination of the M-MuLV genome with a normal cellular gene that has now been designated c-abl (Potter et d., 1973; Parks et al., 1976; Baltimore et al., 1979; Shields et al., 1979; Witte et al., 1979b). Comprehensive reviews of the biochemistry of AMuLV have recently been published (Baltimore, 1981; Goff et d., 1981; Wang, 1983; Wang et al., 1983; Witte, 1983). Abelson virus retains homology with M-MuLV at both the 5‘ and 3’ ends of its genome (Shields et al., 1979). At the 5’ end are the long terminal repeat and a portion of the gag protein sequences that includes p15, p12, and part of p30. At the 3’ end is another M-MuLV derived long terminal repeat. The remainder of the AMuLV genome is homologous to most of the coding sequences of the c-abl gene (Baltimore et al., 1979; Witte et al., 197913; Goff et al., 1980; Wang et al., 1984; Heisterkamp et al., 1983). Only a single polypeptide is known to be translated from v-abl although sequencing data show an additional open reading frame of 789 nucleotides at the 3’ end of the virus (Latt et al., 1983; Reddy et al., 1983).The protein first isolated from A-MuLV-transformed cells (P120) was 120,000 Da and had immuno-crossreactivity with the M-MuLV p12, p15, and p30 (Witte et al., 1978; Reynolds et al., 1978a,b; Rosenberg et al., 1980b). Since its characterization, a larger A-MuLV transforming protein, P160, has been isolated which contains an added internal region that is deleted from P120 (Rosenberg and Witte, 1980; Rosenberg et al., 1980a,b). The ability to precipitate the A-MuLV transforming protein with antiMoloney MuLV antisera has been useful for characterization of its autokinase activity (Witte et al., 1980a). Immunoprecipitates of A-MuLV P120 when mixed with Mn2+ and ATP will transfer the gamma phosphate group to tyrosine residues. Abelson virus-transformed cells also show increased

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

77

levels of tyrosine phosphorylation of cellular proteins suggesting a global role for the kinase activity of PI20 in the biological effects it induces (Sefton et al., 1981). The phosphorylations important for transformation are difficult to determine because of the nonspecificity of the kinase activity. PI20 causes tyrosine phosphorylation of proteins in fibroblasts which are not present in some lymphoid cells (Sefton et al., 1983) and phosphorylation of peptide substrates that have been synthesized or produced by cleavage of large polypeptides (Hunter, 1982; Konopka et al., 1984a,b). In addition, extensive phosphorylation of E . coli-derived proteins is observed when the cloned AMuLV genome is expressed in these cells (Wang et d., 1982). Convincing evidence that the kinase activity is important for transformation comes from analysis of various mutant strains of A-MuLV. All transformation-defective mutants of A-MuLV show either no kinase activity (P92td, P120td) (Reynolds et al., 1980; Witte et al., 198Ob) or barely detectable kinase levels (P155td) (Ponticelli et al., 1982). Other mutants (P90, P1OO) show reduced kinase activities which parallel reduced efficiencies of lymphoid cell transformation (Rosenberg et al. , 1980a). Regretfully, no temperature-sensitive mutants of A-MuLV have yet been isolated that might aid in determining the precise hnction of the kinase activity.

D. USE OF ANTI-A-MuLVANTIBODIESTO STUDY THE C-abl-ENcoDED PROTEIN One strain of mouse (C57L) exhibits a significant degree of resistance to Abelson disease (Witte et al., 1979a). A small portion of the mice that develop tumors after injection with a syngenic A-MuLV induced tumor, eventually reject them. These tumor-regressing mice produce antibodies that precipitate the A-MuLV transforming protein (Witte et d., 1979a). Among the determinants recognized by the antibodies are ones that are in the region of v-abl that is homologous to c-abl. Serum from these mice precipitate a 150,000 molecular weight protein (called NCP150 or P150) from normal cells that is the product of the c-abl gene (Witte et al., 1979b). Analysis of NCP150 expression in A-MuLV-transformed cells using serum from tumor-regressing mice was complicated by the fact that the regions of the molecule that were recognized by the antibodies were regions shared with the A-MuLV protein. This made immunoprecipitation of NCP150 difficult because the amount of A-MuLV P120 or P160 was in great excess to that of NCP150. Studies of NCP150 have been greatly facilitated by production of rabbit antibodies to synthetic peptides and bacterially produced fusion proteins made by utilizing nucleotide sequence data to predict the amino acid sequence (Goff et al., 1980; Srinivasan et al., 1981, 1982; Latt et al., 1983; Reddy et al., 1983; M. Paskind and D. Baltimore, personal communication). Analyses with such antibodies shows that c-abl is expressed at the

78

CHERYL

A.

WHITLOCK A N D OWEN N . W I m E

protein level in many cell types, and levels of c-abl expression are not grossly effected by transformation by A-MuLV P120 (Konopka et al., 1984a,b; J. Konopka and 0. Witte, unpublished). Site-directed antibodies have also been prepared that recognize determinants in the src-homologous kinase region (Konopka et al., 1984a). Precipitation with these antibodies blocks autophosphorylation of P120 and P160. This region is that portion of P120 that is deleted in a transformation-defective mutant of A-MuLV, P92td, that exhibits no kinase activity (Witte et al., 1980b). These data strongly implicate this region of the v-abl gene in the kinase activity of the v-abl molecule. Many of the retrovirus genomes that have regions of homology with the A-MuLV kinase region (v-src, v-fes, v-fps, v-yes) also encode tyrosine kinases (Besmer et al., 1983; Groffen et al., 1983; Reddy et al., 1983; Wang, 1983). The importance of the kinase region for the biological effects of A-MuLV is further supported by studies with mutants of A-MuLV created by site-directed mutagenesis (Prywes et al., 1983). A gag mutant which contains only the 5’ LTR, 300 bp of the gag sequences and the kinase region encodes a protein which retains kinase function and the ability to transform NIH 3T3 fibroblasts. However, this mutant does not efficiently transform bone marrow lymphocytes. This and other deletion mutants are currently being studied in order to understand how the lymphoid and fibroblast transforming functions of v-abl differ.

E. THE PATTERNOF c-abl EXPRESSION BY NORMAL AND

LEUKEMIC CELLS

Serological analyses using tumor-regressor sera and the site-directed antipeptide antibodies have shown NCP150 to be synthesized by all of the normal and transformed cells tested (Konopka et al., 1984a,b; unpublished). This confirms mRNA analyses in mouse and human tissues which showed every cell to synthesize two mRNAs that cross-hybridize to a v-abl probe (Ozanne et al., 1982: Wang and Baltimore, 1983). In mice these mRNAs are approximately 6.5 and 5.5 kb ( Wang and Baltimore, 1983). The greatest amount of c-abl RNA expression in the adult animal is in fibroblasts. During embryological development, c-abl expression is greatly increased during mid-gestation of the 21-day gestation period for mice (Muller et at., 1982) coincident with rapid lymphoid cell development. Genes homologous to c-abl have been found in all vertebrate species that have been examined and in Drosophila melanogaster DNA (Shilo and Weinberg, 1981; Hoffman et al., 1983; Hoffman-Falk et al., 1983). Cellular-abl is located on chromosome 2 in the mouse and chromosome 9 in humans (Goff et al., 1982a; Heisterkamp et al., 1982). Sequencing data show that the most highly conserved region is the kinase region. An obvious conclusion from

COMPLEXITY OF VIHUS-CELL

INTERACTIONS

79

these findings is that c-abl is an oncogene that has been highly conserved in evolution, and it encodes a protein with tyrosine kinase activity. However, careful analyses of P150-c-abl shows no phophorylation of tyrosine residues on the NCP150 molecule (Ponticelli et al., 1982). In addition, NCP150 precipitated with the site-directed antibodies exhibits no autokinase activity (Konopka et al., 1984b). The best evidence that c-abl functions as an oncogene is its association with the Philadelphia chromosome translocation of human chronic myelocytic leukemia. The Philadelphia chromosome, which is characteristic of this disease, is a reciprocal translocation of the long arms of chromosomes 9 and 22. Human c-abl is located very near the breakpoint on chromosome 9 (de Klein et al., 1982; Collins and Groudine, 1983). In one continuous cell line, K562, established from a CML patient with such a translocation, c-abl expression at the level of mRNA is increased 4- to 8-fold over that in normal cells (Collins and Groudine, 1983). Most interestingly, a unique c-abl protein has been identified in this cell line which does express tyrosine kinase activity (Konopka et al., 198413). II. Neoplastic Transformation by A-MuLV

The most dramatic biological effect induced by A-MuLV is rapid neoplastic transformation. When young BALB/c mice are injected with AMuLV, tumors rapidly develop in the lymphoid tissues along the vertebral column (Abelson and Rabstein, 1970b; Siegler et al., 1972; Potter et al., 1973; Risser et al., 1978; Risser, 1982). Transformed cells can also be isolated from the bone marrow, but there is little involvement of cells in the spleen. The cells transformed by this method of infection are basically from two hematopoietic lineages-B lymphocyte and monocyte/macrophage. As noted below, the particular cell type that is transformed can be altered by changing the method of in vivo infection or by controlling the target cell population used for in vitro infection.

A. PRE-BCELLS Pre-B cells with unrearranged immunoglobulin light chain genes and bearing the B lineage membrane marker, B220 (R. Coffman, personal communication) are the major population transformed by A-MuLV (Potter et al., 1973; Sklar et al., 1975; Pratt et al., 1977; Boss et al., 1979; Siden et al., 1979). If adult BALB/c mice are injected intravenously or if bone marrow is infected in vitro, 50 to 60% of the transformed cell lines synthesize cytoplasmic p, heavy chains, but only rare cell lines synthesize light chains. A large proportion (approximately 90%) synthesize terminal deoxynucleotidyltransferase (Baltimore, 1981; Rosenberg, 1982), a known differentiation marker of

80

CHERYL

A.

WHITLOCK A N D OWEN

N.

WITZTE

thymocyte maturation that is also associated with early B lymphocyte maturation (Kung et al., 1976; Silverstone et d.,1978). When fetal liver cells from 13-to 14-day-old fetuses are used as a source of cells for in uitro infection, the transformed cell lines that can be established represent an even earlier stage of B cell development (Waneck and Rosenberg, 1981a). A large percentage synthesize terminal deoxynucleotidyltransferase (64%), but few synthesize p heavy chains. Many of these transformants are at the stage of maturation after which the D and J H segments of the immunoglobulin heavy chain locus have been joined, but are prior to the stage where a V, segment has been selected and rearranged (Ah et al., 198lb). In very recent work R. Phillips and his colleagues (personal communication) have been able to derive Abelson transformed cell lines from SCID (severe combined immune deficient) mice. These cell lines have the phenotype of pre-B cells. This is surprising since SCID mice show no evidence of functional T or B cells or their precursors. Perhaps, an early stem-like cell or transient pre-B cell population can be detected in this transformation system. In spite of a great deal of effort to transform cells that are at even earlier stages of ontogeny (that is with germline arrangements of the heavy chain genes), no such lines have been isolated from transformation of fetal liver or adult bone marrow. In recent experiments, transformation of placental cells has produced cell lines with germline heavy chain genes that express the B lineage marker, B220 (E. Siden, personal communication). A major effort at this point is to isolate such early stage cell lines that have retained the capacity for further maturation. B. TECHNIQUES FOR TRANSFORMATION OF MATURE B AND PLASMA CELLS Although transformation of earlier stages of B cell differentiation has proven difficult, isolation of transformants that represent later stages of development can be easily accomplished by either of two methods. First, by treatment of BALB/c mice with pristane prior to A-MuLV infection (Potter et al., 1973) and, second, by using nontransformed B lymphocyte cell lines as targets for in uitro infection (Whitlock and Witte, 1982; Whitlock et al., 1983a, 1984). A routine method for increasing the efficiency of tumor cell growth is to prime the animal with an intraperitoneal injection of pristane. This treatment alone causes lymphosarcomas to develop in BALB/c mice with a latent period of over 150 days (Potter et aZ., 1973). A-MuLV induces lymphosarcoma in 25 to 40 days (Potter et al., 1973). In the absence of pristane

COMPLEXITY OF VIRUS-CELL INTERACTIONS

81

priming, nearly all of the transformants isolated represent pre-B cells. In the presence of pristane nearly half of the tumors that develop within 50 days are plasmacytomas or plasmacytoid lymphomas (Potter et al., 1973). There is a question, however, whether A-MuLV plays a primary role in transformation of plasma cells. Recent evidence indicates that many of the plasmacytoid lymphoma lines isolated from A-MuLV-infected mice do not express AMuLV (Haschke, 1980b; Mushinski et al., 1983a,b), nor do they have any detectable A-MuLV proviruses in their DNA (Mushinski et al., 1983b). These cell lines do show increased expression of an endogenous oncogene, c-myb, and rearrangement of the c-myb locus can be demonstrated by using Southern blot analysis of the chromosomal DNA (Mushinski et al., 1983b). These patterns of altered expression and arrangement of c-myb are identical to those found for phenotypically similar cell lines derived by mechanisms other than A-MuLV infection. Unlike the plasmacytoid lymphoma lines, the plasmacytoma lines induced by A-MuLV infection of pristane-primed mice retain expression of integrated A-MuLV genomes (Mushinski et al., 1983a,b). Plasmacytomas do not have abnormal arrangement or expression of c-myb. Rather, these cells often show increased expression of another oncogene, c-myc (Mushinski et al., 1983a,b). In a few of the plasmacytomas, the c-myc gene is rearranged and this correlates with synthesis of a new size of c-myc mRNA. It is unclear what effects alteration of the c-myc or c-myb oncogene expression has on cell growth and how A-MuLV infection acts to promote altered expression of these genes. A second method for obtaining A-MuLV transformed B lymphocytes is to infect continuous lines of untransformed B cells. B cell lines can be isolated from long-term cultures of mouse bone marrow that are maintained under conditions which favor only B cell lymphopoiesis (Whitlock and Witte, 1982; Whitlock et al., 1984). Once established, the B lymphocytes can be cloned and continuously passaged if kept in the presence of feeder layers of mouse bone marrow adherent cells. Both pre-B and B cell lines can be isolated in this manner, and both serve as targets for A-MuLV transformation (Whitlock et al., 1983a). Transformation can be easily monitored by the ability to divide in the absence of the feeder cell layers or by growth in agar-containing media. When such culture-adapted cell lines are used for transformation studies, no obvious correlation is found between the stage of maturation and efficiency of transformation (Whitlock et al., 1983a). As will be discussed below, these cell lines represent the unusual case of pure hematopoietic target cell populations for murine retrovirus transformation. As such, they will prove very usehl for studying the molecular changes which occur in lymphoid cells early after A-MuLV infection.

82

CHERYL

A.

WHITLOCK A N D OWEN N. WI'ITE

C . A-MuLV-TRANSFORMED CELLLINES WITH DIFFERENTIATION POTENTIAL Most A-MuLV transformed pre-B cell lines appear to be locked into a single stage of differentiation, As a result, they are useful for characterizing the phenotypes expressed during this stage of maturation and for preparing serological reagents that recognize differentiation proteins expressed by preB cells (Coffman and Weissman, 1981a,b; Kincade et al., 1981; Coffman, 1982; Kung et al., 1982). They fail, however, to provide a useful system for examining the changes that occur as pre-B cells mature. As more A-MuLV transformed cell lines were studied, rare populations that retain the capacity for further maturation were found. These include cell lines that rearrange the immunoglobulin light chain genes (Alt et al., 1981a; Lewis et a!., 1982), lines that exhibit class switching from mu heavy chain synthesis to y2B chain synthesis (Alt et al., 1981a; Burrows et al., 1981; Akira et aE., 1983), and even an early pre-B cell line that rearranged a variety of V, regions onto an already joined D,-J, locus (Sugiyama et aZ., 1982, 1983). In addition, some cell lines can be induced to increase light chain synthesis by agents such as lipopolysaccharide (Boss et al., 1979, 1981). Lipopolysaccharide also increases expression of kappa light chain genes that have been introduced into A-MuLV transformed cells by transfection (Rice and Baltimore, 1982). Another approach to study pre-B lymphocyte maturation is by transformation of the cloned populations of pre-B lymphocytes obtained from long-term cultures of mouse bone marrow (Whitlock et al., 1983a). These A-MuLV transformants retain the capacity for normal differentiation to mature B cells expressing membrane IgM and IgD (Whitlock et al., 1983a). If these pre-B cell lines are infected with A-MuLV and immediately cultured in agar, subclones can be obtained that represent the progeny of a single transformed cell. In a number of cases, the transformed colonies contain B cells expressing a variety of different light chain molecules. Analysis of subclones from these colonies indicate that after A-MuLV transformation the cells had undergone kappa gene rearrangement and even somatic mutation of the expressed kappa segment (Ziegler et al., 1984).

D. TARGETCELLSI N

THE

MONOCYTE-MACROPHAGE LINEAGE

When tumor cells from A-MuLV-infected, pristane-primed mice are passaged in vitro for several weeks, cells with the characteristics of macrophages sometime predominate in the culture (Raschke et al., 1978). Three such cell lines were examined in detail, and all contained rescuable A-MuLV (Raschke, 1980b). When the tumors were first cultured, the predominant cell type was the lymphosarcoma. Recently, Boyd and Schrader (1982) have

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

83

shown that one A-MuLV-transformed pre-B cell line, ABLS-8.1 (Hines and Gragowski, 1983), can be induced to stabily express macrophage characteristics (phagocytic, esterase-positive, Mac-1 ) after treatment with 5azacytidine for 7 days. This underscores the point that the phenotypes expressed by an A-MuLV-transformed cell line may not represent those of the original cell that was infected. Transformed cell lines with the characteristics of promonocytes have also been isolated from in vivo infection with A-MuLV in the absence of pristane priming (Hines and Gragowski, 1983). Two of these lines, AC-5 and AC-8 can be driven to differentiate using conditioned medium from an adherent cell line, 266 AD, that was isolated from mouse bone marrow (Hines, 1983). Supernatants from 266 AD also contain a granulocyte-marcophage colony stimulating activity (Hines, 1983). +

E. RAPID INDUCTION OF THYMOMAS BY A-MuLV Another method for altering the target specificity of A-MuLV is through the route of i n v i m injection. Intraperitoneal or intravenous injections generally lead to tumors in vertebral lymph nodes and in the bone marrow, and there is no involvement of the thymus during the 25 to 40 days before the mouse dies. Intrathymic injection can produce rapid formation of intrathymic tumors (Cook, 1982). All of these tumors produce A-MuLV. Thymic tumors from BALB/mice are Thy-l- and Lyt-1-, 2 - , but those from C57BLlKa are Thy-1+, Lyt-1' 2 - . Some of the C57BLlKa transformants continue to express Thy-1 and Lyt-1 after continuous passage in vitro. The amounts of Thy-1 expressed are well below that of normal T cells. Risser and co-workers (Grunwald et al., 1982a,b; Risser, 1982) have also described Thy-l- cell lines, that were derived by in uivo infection of a C57BL/6 mice with A-MuLV. These cell lines after several passages i n vivo begin to express Thy-1 and lose expression of the A-MuLV genome. These findings provide independent evidence that A-MuLV-transformed cell lines can express Thy-1 and that continued expression of the A-MuLV genome is not required for maintaining the transformed state of all cells. These data also emphasize that the phenotypes expressed by A-MuLV transformed cells can change upon passage even in the absence of known inducing agents. F. NIH 3T3 FIBROBLASTS: A TOOLFOR VIRUSSTUDY A N D MANIPULATION One last target cell population for A-MuLV transformation is NIH 3T3 fibroblasts (Scher and Siegler, 1975). This continuous line derived from mouse embryo fibroblasts has provided a model system for studying growth modulation by a wide variety of DNA viruses, RNA viruses, and cloned oncogenes. Perhaps the reason why it is so useful is that it is well adapted to

84

CHERYL A. WHITLOCK AND OWEN N . WITTE

in uitro culture, and yet it exhibits contact inhibition of growth. Through many years of culture, changes in expression of cellular genes have probably occurred which have brought this cell line nearer to being transformed. As a result, a large variety of exogenously added genes will cause NIH 3T3 fibroblasts to be released from contact inhibition but will not cause similar effects in primary cultures of embryo fibroblasts (v-abl is one such gene). Alone, the ras-1 gene will also not cause transformation of cell lines such as primary rat embryo fibroblasts or baby rat kidney cells, but if cotransfected with other oncogenes such as c-myc, Adenovirus 2 ElA, or polyoma middle T, transformation results (Land et al., 1983; Ruley, 1983). v-abl and the other oncogenes encoding tyrosine kinases have been grouped together as genes which initiate transformation while c-myc and c-myb are thought to promote transformation. Since there is no in uivo correlate of in uitro transformation of NIH 3T3 fibroblasts, the significance of this biological effect of A-MuLV is somewhat reduced. However, a homogeneous cell line that is well adapted to culture, easily transfectable, and transformable with A-MuLV is invaluable. It provides a method for titering virus stocks, for examining the early molecular changes which occur after virus infection, and for preparing virus stocks from cloned v-abl genomes that have undergone site-directed mutagenesis (Krump and van den Berg, 1981; Srinivasan et al., 1982; Prywes et al., 1983; Watanabe and Witte, 1983). NIH 3T3 cells, as well as other transformable fibroblast cell lines, also provide a way of dissecting the complexity of biological effects induced by V-abl. A discordance between fibroblast and lymphoid transforming capacities of A-MuLV mutants has been recognized (Rosenberg et al., 1980a). Certain mutants such as P90 are equal to wild-type P120 in fibroblast transforming capacity, but are poorer at transforming lymphoid cells (Rosenberg et al., 1980a). The gag- mutants, made using site-directed mutagenesis, transform only fibroblast cells indicating that the requirements for transformation of NIH 3T3 cells and bone marrow lymphocytes are different and can be separated (Prywes et al., 1983).

Ill. Nonneoplastic Changes Induced by v-a61

Abelson virus can bind and penetrate into any cell that expresses membrane receptors for the helper virus used to package the replication-incompetent A-MuLV genome. Moloney MuLV is the most common helper virus used to prepare A-MuLV stocks, and this virus can infect nearly every cell type in the animal. Whether following penetration the A-MuLV genome is integrated and expressed is not known for many cell types. Theoretically

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

85

there should be no barrier to integration and expression of the v-abl gene. However, the presence of the A-MuLV transforming protein within certain cell types may interfere with growth and normal functions and actually be lethal (Ziegler et al., 1981).

A. AGARCOLONY FORMATION n Y FETAL LIVERERYTHROBLASTS Evidence for v-abl expression in a wide variety of cell types is supported by findings that show significant biological changes in cells that are not normally targets for neoplastic transformation. One example is erythroid cell precursors (Waneck and Rosenberg, 1981b). When liver cells from young fetuses are infected with A-MuLV and cultured in soft agar, a portion of the colonies become red in color. This is in striking contrast to the consistently white colonies obtained when A-MuLV-infected bone marrow from adult animals is cultured in agar. The red colonies contain erythroid cells at various stages of differentiation, and each colony can be shown to express AMuLV (Waneck and Rosenberg, 1981b). Although A-MuLV induces agar colony formation by erythroid cell precursors, these colonies could not be continuously passaged in culture. In addition, there have been no reports of A-MuLV induced erythroleukemia in mice. One possibility is that A-MuLV expression mimics the normal signals induced by colony stimulating factors, but does not provide a signal for neoplastic transformation. Alternatively, expression of A-MuLV in erythroid cell precursors may stimulate growth and differentiation to a state where A-MuLV has no effect or a deleterious effect on continued growth. OF GM-CFC B. RESISTANCE ACTIVITY(LIA)

TO

LEUKEMIA-ASSOCIATED INHIBITION

A second population of cells that exhibits changes in growth properties following A-MuLV infection is the granulocyte-macrophage colony forming cell in adult BALB/c mouse bone marrow (Broxmeyer et al., 1981). These cells form mixed granulocyte-macrophage colonies in soft agar in the presence of GM-colony stimulating factor. Colony formation is normally inhibited by a leukemia-associated inhibitory activity (LIA) present in the serum of leukemia-bearing mice. If the colony-forming cells are taken from BALB/c mice that have been infected with A-MuLV, they are resistant to the inhibitory effects of LIA. This is not surprising in view of the findings that monocyte precursors are susceptible to transformation by A-MuLV (Hines, 1983; Hines and Gragowski, 1983). It is interesting to postulate that one method by which A-MuLV causes growth stimulation and possibly transformation is by blockage of signals that halt division rather than by production of signals that stimulate growth.

86

CHERYL A. WHITLOCK AND OWEN N . WI’ITE

C. LETHALITY OF v-abl: A POSSIBLE ROLE IN “HITAND RUN TRANSFORMATION” As suggested above, expression of v-abl may not be compatible with longterm proliferation or viability of certain cell types. An example is the lethal effect of A-MuLV on certain fibroblast cells (Ziegler et al., 1981). NIH and BALB 3T3 fibroblasts are equally susceptible to transformation by wild-type and mutant strains of A-MuLV. Upon passage, however, the transformed populations take different courses. NIH 3T3 cells continue to exhibit a transformed phenotype for many months, or indefinitely, and they continuously express the A-MuLV transforming protein. In contrast, within a few days after A-MuLV transformation, BALB-1 cells begin to die. After 3 to 4 weeks of passage, only untransformed cells that have lost the integrated A-MuLV genome remain in the cultures. Although NIH 3T3 cells are more resistant to the lethal effects of A-MuLV, expression of multiple copies of the v-abl gene in a cell can lead to cell death. Cotransfection of v-abl with a selectable gene results in fewer cells expressing the selectable marker than if the nonviral gene is transfected alone (Goff et al., 1982b). The progression of the A-MuLV-infected BALB-1 cultures is an interesting in vitro model of counterselection for non-virally infected cells. Those cells that eventually dominate the culture are those that have lost expression of v-abl and therefore have a growth advantage. By forcing the A-MuLVinfected cells to grow under conditions where only transformed cells will proliferate, a series of lethal-minus mutants of A-MuLV were isolated that retained the capacity for fibroblast and lymphoid cell transformation (Ziegler et al., 1981). The region of v-abl important for the lethal effect has been mapped to the carboxy-terminal third of the protein (Ziegler el al., 1981; Watanabe and Witte, 1983). These studies bring the number of regions of the A-MuLV transforming protein with distinct biological properties to three: (1)src-homologous-the portion of the protein that is necessary for the autokinase activity and transformation of NIH 3T3 fibroblasts, (2) the 5‘ gag sequences that are required in addition to the kinase region for lymphoid cell transformation, and (3) the carboxyl-terminus that is not required for transformation, but the presence of which may complement functions of the kinase domain in producing a lethal effect on infected cells. An appreciation of the possible lethal effects of v-abl on certain cell types is important for understanding the range of biological effects induced by AMuLV infection. The long-term effects will depend upon how the transforming and lethal activities of the A-MuLV protein balance one another within a particular cell type. The longer growth of infected cells can be maintained, the greater the chance that certain of the progeny will undergo changes in cellular genes that will augment the effects of the v-abl gene.

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

87

With this in mind, it is easy to visualize that within a tumor cell population enough cellular changes can occur such that the transforming activity of v-ah1 is no longer necessary for continued proliferation. In this case cells that have lost expression of v-ah1 and its associated lethal effect may have a growth advantage. Thus, expression of v-ah1 may be necessary for the early events of tumor progression but can be selected against as the tumor population expands and progresses toward more autonomous growth. This is one possible model for the “hit and run” role of A-MuLV in production of plasmacytoid lymphosarcomas (Potter et al., 1973; Raschke et al., 1978; Raschke, 1980b; Mushinski et al., 1983a,b; Newmark, 1983). A similar argument could be made to explain why some A-MuLV-transformed cell lines lose the integrated A-MuLV genome after continued in uiuo passage (Grunwald et al., 1982a,b; Risser, 1982). IV. The Complexity of Abelson Disease

Both tumor production in viuo and lymphoid transformation in uitro are complex processes involving more than simple expression of the v-abl gene. The course of Abelson disease in uiuo varies with the age (Abelson and Rabstein, 1970b; Risser, 1982; Rosenberg, 1982) and strain (Abelson and Rabstein, 1970b; Rosenberg and Baltimore, 1976; Risser et al., 1978; Earl and Scher, 1980; Broxmeyer et al., 1981; Risser, 1982; Rosenberg, 1982) of mouse injected, the site of injection (Cook, 1982), and whether the mice are primed with pristane (Potter et al., 1973; Raschke et al., 1978). The ability of A-MuLV to induce tumors in susceptible mice is also affected by the strain of helper virus used to package the replication-incompetent A-MuLV genome (Rosenberg and Baltimore, 1978; Scher, 1978). At least one level of the complexity of the in uiuo disease process can be demonstrated in vitro. This is that cells exhibit restricted in vitro growth potential and low tnmorigenic capacity early after infection but progress with in uitro cultivation to highly neoplastic cells (Teich and Dexter, 1978; Whitlock and Witte, 1981; Whitlock et al., 1983b). A. A ROLE FOR HELPERVIRUSI N NEOPLASTICTRANSFORMATION BY A-MuLV

The helper virus most commonly used to prepare stocks of A-MuLV is MMuLV. This lenkemogenic virus causes thymomas by itself, but with a latency period of greater than 3 months (Moloney, 1960). Since M-MuLV is NBtropic (Hartley et al., 1970; Risser et al., 1978; Rosenberg and Baltimore, 1978),A-MuLVIM-MuLV virus stocks are useful for infecting most strains of mice. Nearly all cell types have membrane receptors that bind M-MuLV,

88

CHERYL A. WHITLOCK A N D OWEN N . WIITE

and this virus is expressed in many cells of infected animals. One assumes that A-MuLV packaged by M-MuLV proteins is capable of binding to and penetrating into a similar range of cell types. The A-MuLV genome can be packaged by a variety of replication competent viruses. All leukemogenic viruses produce A-MuLV virus stocks that efficiently transform NIH 3T3 cells and bone marrow in vitro and produce tumors in BALB/c mice (Rosenberg and Baltimore, 1978; Scher, 1978; Hines and Gragowski, 1983). Nonleukemogenic viruses, such as GrossMuLV, BALB/c endogenous viruses, and Kirsten-MuLV produce virus stocks that are much less efficient at lymphoid cell transformation both in vitro and in vivo (Rosenberg and Baltimore, 1978; Scher, 1978). The precise step of the transformation process at which the helper virus plays a role is unknown.

B. PROLIFERATING CELLSARE THE TARGETS FOR in Vitro TRANSFORMATION Another parameter effecting the ability of A-MuLV to transform is whether the infected cells are in the process of dividing (Rosenberg, 1982). Treatment ofbone marrow cells with amounts of radiolabeled thymidine that are lethal to dividing cells dramatically decreases the number of foci that are formed upon subsequent infection and soft agar culture. This result could be explained by an increased efficiency of recognition or stable integration of A-MuLV by cells that are cycling. Alternatively, it may simply be that the target population for transformation is at a step of B cell differentiation where the cells divide rapidly. Cells that are dividing may exhibit more autonomous growth properties after A-MuLV infection since they are actively expressing genes important for division. Expression of these cellular genes together with expression of v-abl may be the necessary ingredients for neoplastic transformation to occur.

C. GENETICS OF SUSCEPTIBILITY TO ABELSON DISEASE A-MuLV very efficiently produces tumors in adult BALB/c mice and BALB/c congenic mice. In contrast, adult mice of other mouse strains are resistant to tumor production (Abelson and Rabstein, 1970b; Rosenberg and Baltimore, 1975; Risser et al., 1978; Earl and Scher, 1980; Risser, 1982; Rosenberg, 1982). The prototype resistant strain is C57BL/6 (Risser et al., 1978). Genetic analysis of the differences between adult BALB/c and C57BL/6 mice have shown three loci to be involved-Av-1, Av-2, and H-2. Sensitivity alleles at the Av-1 and Av-2 loci are dominant over resistance alleles, and BALB/c mice have sensitivity alleles at both Av-1 and Av-2. The major histocompatibility type of the mouse strain can also effect sensitivity as shown by studies with H-2 congenic mice (Kisser et d., 1978). The

COMPLEXITY OF VIRUS-CELL

INTEHACTIONS

89

effects of the H-2 locus are minor in comparison to Av-1 and Av-2, and sensitivity at either of these latter loci dominates over resistance at the H-2 locus. BALB/c and C57BL/6 mice differ in their ability to mount an anti-MMuLV immune response (Stukart et al., 1981), and this difference may account for part of the H-2 effect. A more active antiviral immune response in C57BL/6 mice could reduce susceptibility to Abelson disease. An active immune response to M-MuLV-infected cells cannot totally explain the resistance of C57BL/6 mice, since lymphoid tissues from these mice express high levels of M-MuLV. However, C57BL/6 fail to express AMuLV which contrasts with high expression of both A-MuLV and M-MuLV by tissues from BALB/c mice (Risser et al., 1978). Other differences in the effects of A-MuLV on different mouse strains include the Thy-1 and Ly-1 phenotypes expressed by thymomas induced by intrathymic injection of A-MuLV into BALB/c and C57BL/Ka mice (Cook, 1982). In addition, A-MuLV induces resistance of granulocyte-macrophage colony-forming cells from BALB/c mice to human leukemia-associated inhibitory activity, but has no effect on such cells from C57BL/6 mice, and induces production of a murine LIA activity in the bone marrow, spleen, and thymus of BALB/c mice, but not of C57BL/6 mice (Broxmeyer et al., 1981). Since A-MuLV does not reduce the number of G-M colony-forming cells in C57BL/6 mice, there must be an intrinsic difference in the ability of AMuLV to induce LIA resistance in cells from these two mouse strains. It is not known if the GM colony-forming cells from these two mouse strains express equal amounts of A-MuLV or whether those from C57BL/6 mice express only the helper virus.

D. A-MuLV TRANSFORMATION OF CELLSFROM GENETICALLY RESISTANTMICE Although adult mice differ in their susceptibility to Abelson disease in uiuo, neonatal mice of all strains are highly susceptible (Abelson and Rabs-

tein, l97Ob; Risser, 1982). Bone marrow cells from almost all adult mice also serve as targets for in uitro transformation. Mouse strains vary in the number of agar foci that develop after in uitro infection, but there is little correlation between in uitro and in uiuo susceptibilities to Abelson transformation (Abelson and Rabstein, 1970b; Rosenberg and Baltimore, 1976; Risser, 1978). The only mouse strain that shows a significant reduction in the number of targets for in uitro transformation is the nulxid mutant mouse strain (Rosenberg, 1982). The discrepancies in the in uitro and in uiuo susceptibilites of various mouse strains make it difficult to develop a coherent model for A-MuLV transformation. These discrepancies may arise due to the different external environments in which the transformants are attempting to proliferate. One important parameter of the in uiuo infection process that

90

CHERYL A. WHITLOCK A N D OWEN N . WITTE

is absent in vitro is the host animal’s immune response to the virally infected cells. Maturation of the immune system is a plausible explanation for the increasing resistance of most mouse strains with age.

E. VARIATIONS OF THE in Vivo DISEASE PROCESSINDUCED BY

MUTANTS OF A-MuLV

A number of mutants of A-MuLV have been isolated that differ in their efficiency of lymphoid cell transformation both in vitro and in vivo. One mutant, PgO, is particularly interesting because of its behavior in vivo (Rosenberg, 1982). The P90 mutant encodes a protein that has a reduced autokinase activity and is less efficient at transforming lymphoid target cells in vitro and in vivo (Rosenberg et al., 1980a,b). The reduced efficiency of P9O transformation may result from the inability of P9O-infected cells to exhibit the autonomous growth properties required for formation of agar colonies in vitro and production of tumors in vivo (Whitlock and Witte, 1981; Whitlock et al., 198313). A low percentage of BALB/c mice infected with P90 develop tumors and, those that do, develop them after a long latent period (>60 days). When virus is isolated from such tumors, it often is found to have a further carboxyl-terminal truncation, encoding a protein of 85,000 molecular weight (Rosenberg, 1982). In striking contrast to P90, the P85 mutant is as efficient as wild-type A-MuLV at inducing tumors in vivo. An inexplicable finding is that P85 is still deficient at stimulating agar colony formation in vitro. These findings again emphasize that data obtained from in vitro growth assays do not allow one to predict the ability of a cell to grow in the complex in vivo environment. They also indicate that the reduced capacity of P90 to induce tumors may operate at more than one level, one of which is removed by the further truncation seen in the P85 mutant.

F. USE OF SITE-DIRECTED ANTIBODIESTO THE ABELSON PROTEIN TO

STUDYA-MuLV DISEASE

In most of the studies which were discussed above, a single question repeatedly arises when trying to interpret the data. Is the failure of A-MuLV to induce a particular biological change due to an inability of the A-MuLVtransforming protein to function in the cellular environment or simply to the inability of the cell to stabily express the v-abl gene. This question is currently being explored by using immunofluorescent staining to examine v-abl expression at a single-cell level (N. Rosenberg, personal communication). Site-directed antibodies will greatly facilitate these studies because of their high affinity for the Abelson transforming protein and their failure to bind helper virus proteins (Konopka et aZ., 1984a). Among the questions to be explored is what frequency of cells in the bone

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

91

marrow can be infected by and express A-MuLV. The agar colony assay for quantitating A-MuLV targets shows a target cell frequency for bone marrow of approximately 1 in lo4 cells. Immunofluorescent staining should indicate whether this low frequency reflects a low percentage of cells in the bone marrow that are capable of being productively infected by A-MuLV. Other questions can be addressed that involve the frequency of cells that can be infected by A-MuLV. For example, can lymphoid cells from adult C57BL/6 mice express v-abl in vivo even though these mice fail to develop tumors, and is the failure of A-MuLV to induce resistance to leukemiaassociated inhibition factor in graunlocyte-macrophage colony-forming cells from C57BL/6 (Broxmeyer et al., 1981) due to decreased expression of v-abl by these cells. V. The Complexity of A-MuLV Transformation in Vitro

Little is known of the early biological effects of v-abl expression in cells. These are difficult studies to do at a molecular level since homogeneous target cell populations are not readily available. As a result of this difficulty, most studies have focused on established A-MuLV-transformed populations that readily grow in culture and can be cloned. These populations require a minimum of 3 to 4 weeks to prepare and always represent the fastest growing subpopulation of cells. As such, the phenotypic characteristics they possess may not accurately represent the phenotypes expressed by most target cells immediately after infection.

A. EARLYBIOLOGICALEFFECTSOF v-abl EXPRESSION The early biological changes induced by A-MuLV include those changes induced in cells that are not destined for neoplastic transformation. In this category are proliferation of erythrocyte precursors from fetal liver (Waneck and Rosenberg, 1981b) and resistance to LIA by granulocyte-marcophage colony-forming cells (Broxmeyer et al., 1981). In addition, the lethal effect of A-MuLV, if it is to occur, should be exhibited early after infection (Ziegler et al., 1981).These effects of A-MuLV (discussed in previous sections) all occur within the first 3 to 5 days following infection. B. PRENEOPLASTIC GROWTHPROPERTIES O F A-MuLV-INFECTED PRE-B CELLS

A second category of early biological changes include the preneoplastic changes that are induced in cells that are destined for neoplastic transformation (Whitlock and Witte, 1981; Whitlock et al., 198313). If bone marrow cells from 3-week-old BALB/c mice are infected with A-MuLV and cultured under conditions where adherent bone marrow cells establish a feeder layer,

92

CHERYL A. WHITLOCK A N D OWEN N . W I m E

a lymphoblast population rapidly expands to confluency by 7 to 10 days. Although these cells continue to proliferate indefinitely if kept in the original culture dish, they do not form colonies in agar, and are slow to form tumors in animals (30 day latent period from lo7 cell innocula). These cells also die in liquid culture if not kept in the presence of bone marrow adherent cells or media conditioned by such feeder cells (Whitlock and Witte, 1981; Whitlock et a l . , 1983b). These characteristics contrast with those of established AMuLV-transformed cell lines which grow autonomously in liquid and soft agar cultures and rapidly form tumors (7 to 10 days after injection of lo6 cells). The restricted growth properties of target cells early after infection may account for the low frequency of A-MuLV-infected bone marrow cells that form agar colonies. Even those few that do have the capacity to form colonies in agar are somewhat growth restricted. Most agar colonies are difficult to propagate as continuous cell lines and must be initially cultivated at high cell density in order to maintain viability. A simpler and more successful way to nurture cells from agar colonies during this early period is to culture them on established bone marrow adherent cell layers (Whitlock et al., 1984). Under these conditions, even small colonies of only a few hundred cells survive and proliferate rapidly (C. Whitlock, unpublished observation). C. PROGRESSION OF TUMORIGENICITY If most A-MuLV-infected cells are preneoplastic in their growth properties, how does progression to full oncogenicity occur? When the A-MuLVinfected lymphoblasts from bone marrow are passed twice weekly (in the presence of adherent feeder cells to maintain viability), the growth properties exhibited by the cells gradually approach those of established A-MuLVtransformed cell lines (Whitlock and Witte, 1981). The time required for this change to occur in uncloned populations is 3 to 6 weeks. In two cloned cell lines, obtained by limited dilution of a recently infected lymphoblast population on feeder layers, the time required to progress to autonomous growth in uitro and high oncogenic potential in uiuo was more than 10 weeks (Whitlock et al., 1983b). This indicates that a component of tumorigenic progression in the uncloned population is subpopulation dominance and that individual cells may vary in their degree of autonomous growth immediately after infection and/or their rate of tumorigenic progression. In a mixed population, those cells with the greatest growth advantage rapidly dominate the population. The process of tumorigenic progression is a late event that is probably mediated solely by changes at the cellular level. Cloned cell lines as they progress shown no changes in v-abl expression or in phosphorylation and kinase activity of the v-abl-encoded protein (Whitlock et al., 1983b). Cells

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

93

early after infection by mutant strains of A-MuLV, such as P90, are more restricted in their growth properties than P120-infected populations, but they eventually progress to fully autonomous growth probably via the same cellular changes (Whitlock and Witte, 1981). A-MuLV may serve only to initiate or promote a series of cellular events, and it may be that these cellular changes are the key events that trigger neoplastic transformation. One piece of evidence is the finding that A-MuLV expression can be lost and the cell still retain the capacity for neoplastic growth (Grunwald et al., 1982a,b; Riser, 1982). This has been shown to occur in an A-MuLV-transformed C57BL/6 tumor cell line that after multiple in oiuo transfers lost v-abl expression. A large percentage of in vivo plasmacytoid lymphosarcomas induced by A-MuLV infection of pristine-primed mice also show no expression of v-abl (Raschke, 1980b; Mushinski et al., 1983a,b). One cannot prove that A-MuLV played a role in the early transforming events of these particular cells, but one can speculate that once the cellular events that enhanced or potentiated transformation occurred, A-MuLV expression may have been superfluous. Loss of v-abl expression may even be advantageous at this stage due to its potential lethal effect. Although the data presented above suggest A-MuLV may play a passive role in transformation maintenance, one piece of evidence argues strongly for A-MuLV playing an active role in long-term maintenance of the transformed state. This is simply that of the hundreds of A-MuLV-transformed cell lines that have been established and studied, almost all retain v-abl expression after several months or years of passage. Actual proof that continued expression of v-abl is necessary awaits isolation of a temperaturesensitive mutant of A-MuLV. D. CHARACTERIZATION OF CELLULAR CHANGES ASSOCIATED WITH A-MuLV TRANSFORMATION Some progress is being made toward characterizing the cellular changes that follow A-MuLV infection. Thus far, five proteins have been found to have altered expression in A-MuLV-transformed cells. A 50,000-Da protein, P50, is synthesized by untransformed cells (both NIH 3T3 and lymphocytes), but its level increases dramatically upon transformation by A-MuLV or by other methods (Rotter et al., 1980, 1981, 1983). Infection of NIH 3T3 cells also causes increased synthesis of a tumor growth factor (TGF) that is released into the culture medium and competes with epidermal growth factor for binding to its membrane receptor (Blomberg et al., 1980; Twardzik et al., 1982). The importance of increased production and/or secretion of TGF in fibroblast transformation is further supported by the inability of AMuLV to transform rat fibroblasts that are resistant to EGF-induced release from contact inhibition (Kaplan and Ozanne, 1983). However, A-MuLV can

94

CHERYL A. WHITLOCK A N D OWEN

N.

WITTE

efficiently transform NIH 3T3 cell mutants that do not express E G F receptors (0.Witte and H. Herschman, unpublished). Two additional proteins have been defined serologically, and these appear to be synthesized exclusively by transformed cells. One is synthesized by both virally induced and spontaneous pre-B and B cell lymphomas, and is recognized by a monoclonal antibody designated 6C3 (Pillemer et al., 1984). Two glycosylated forms of this protein are found on the cell membrane, and they have apparent molecular weights of 125,000 and 160,000. Expression of 6C3 antigen by A-MuLV-transformed cells is being carefully studied because it occurs in parallel with increasing autonomy of in uitro growth (G. Tidmarsh and C. Whitlock, unpublished observations). The second transformation-associated membrane protein is 80,000 molecular weight and is precipitated by antisera from C57BL/6 mice immunized with the B6 A-MuLV-transformed tumor cell line (Grunwald et al., 1982b; Risser, 1982). This line is unique in that after several in viuo transfers it began to express Thy-1 antigens. B6 is also the same cell line that lost expression of v-abl (Grunwald et al., 1982a,b). P80 differs from the 6C3 antigens in that it is expressed a variety of by T cell lymphomas rather than B cell lymphomas. Progress is also being made in characterizing cellular changes at the level of gene expression. Two such changes were mentioned above, namely rearrangement of c-myc in A-MuLV-induced plasmacytomas and c-myb in AMuLV-induced plasmacytoid lymphosarcomas (Mushinski et al., 1983a,b). Transfections of DNA from A-MuLV-transformed cells into NIH 3T3 fibroblasts have been used to detect alterations in other genes (Lane et al., 1982). One would predict that transfection of DNA from A-MuLV-transformed fibroblasts or lymphocytes would transform NIH 3T3 because of transfer of the v-abl gene, but this does not occur at a high frequency because of the lethal effect of v-abl on this cell line. In the case of lymphoid cells, one transfectable gene has been designated tx-2 (Lane et al., 1982). Work is now in progress to clone these and any other genes whose expression may be altered in A-MuLV transformed cells.

E. USE OF CULTURED B CELLLINESTO STUDYTHE CELLULAR CONTRIBUTION TO NEOPLASTICTRANSFORMATION One approach that promises to provide a method for examining the early biochemical changes induced by v-abl interaction with the cell is to use the established in uitro pre-B and B cell lines as targets for infection (Whitlock and Witte, 1982; Whitlock et al., 1983a, 1984). Certain of these lines contain high frequencies of targets as monitored by the ability to form agar colonies after A-MuLV infection. This ability to be transformed by A-MuLV appears

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

95

to increase in some populations as they are maintained in culture for several weeks. The cellular changes that these lymphoid cells undergo with time not only facilitate their transformation by A-MuLV, but also eventually lead to spontaneous transformation (Whitlock et al., 1984). At the preneoplastic stage, the cells still require an adherent feeder cell layer for growth, and this feeder-dependence can be circumvented, in some cases, by A-MuLV infection. Transformation of these cell lines during this preneoplastic state may be a more favored process than transformation of in vivo targets, and it may be analogous to the ease of NIH 3T3 transformation as compared to primary rat embryo fibroblasts (abl will not transform primary embryo fibroblasts). Cell lines that are resistant to A-MuLV transformation can also be obtained. These serve as excellent controls for distinguishing which biochemical changes induced by the Abelson protein actually function in transformation. These pre-B and B cell lines will be useful for characterizing the cellular genes whose expression augment feeder layer dependence. We predict that the cellular changes that lead to feeder-layer independence of the A-MuLVtransformed cells are similar to or identical with those that lead to spontaneous transformation of the non-virally infected cells. Experiments to analyze expression and gene arrangements of known oncogenes in a variety of independently derived preneoplastic and neoplastic cell lines are currently underway. A more promising approach would be to transfect the important genes from the neoplastic cells to the preneoplastic cells. When a method for efficiently having exogenous genes stabily expressed in these cells is developed, then genes important for tumorigenic progression and B cell leukemogenesis can be isolated.

REFERENCES Abelson, H. T., and Rabstein, L. S. (1969). Cancer Res. 10, 1. Abelson, H. T., and Rahstein, L. S. (1970a). Cancer Res. 30, 2208. Abelson, H. T., and Rabstein, L. S. (1970b). Cancer Res. 30, 2213. Akira, S., Sugiyama, H . , Yoshida, N . , Kikutani, H . , Yamamura, Y., and Kishimoto, T. (1983). Cell 34, 545. Alt, F. W., Rosenherg, N. E., Lewis, S., Casanova, R . J., and Baltimore, D. (1981a). In “B Lymphocytes in the Immune Response” (N. Klinman, D. Mosier, C. Scher, and E. Vitetta, eds.), p. 33. Elsevier, Amsterdam. Alt, F . , Rosenberg, N . , Lewis, S . , Thomas, E . , and Baltimore, D. (198lb). Cell 27, 381. Baltimore, D. (1981). Prog. Clin. Biol. Res. 45, 297. Baltimore, D . , Shields, A , , Otto, G . , Goff, S., Besner, P . , Witte, O., and Rosenherg, N. (1979). Cold Spring Harbor Symp. Quant. Biol. 44, 849. Bell, R . , Lillquist, A , , Abelson, H . , and McCaErey, R. (1982). Leuk. Res. 6, 775. Besmer, P., Hardy, W. D., J r . , Zuckerman, E. E., Bergold, P., Lederman, L . , andSnyder, H . W., Jr. (1983). Nature (London)303, 825.

96

CHERYL A. WHITLOCK A N D OWEN N . W I l T E

Blomberg, J., Reynolds, F. H., Jr., van de Ven, W. J., Stephenson, J. R. (1980). Nature (London) 286, 504. Boss, M., Greaves, M., and Teich, N. (1979). Nature (London)278, 551. Boss, M. A., Greaves, M. F., and Teich, N. (1981). Eur. J. Zmmunol. 11, 136. Boyd, A. W., and Schrader, J. W. (1982). Nature (London) 297, 691. Broxmeyer, H. E., Ralph, P., Gilbertson, S., and Margolis, V. B. (1981). Cancer Res. 40,3928. Burrows, P. D., Beck, G . B., and Wabl, M. R. (1981). Proc. Natl. Acad. Sci. U.S.A.78, 564. Coffman, R. L. (1982). Zmmunol. Reu. 69, 5. Coffman, R. L., and Weissman, I. L. (1981a).J. E x p . Med. 153, 269. Coffman, R. L., and Weissman, I. L. (1981b). Nature (London) 289, 681. Collins, S. J., and Groudine, M. T. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 4813. Cook, W. D. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2917. Cory, S., Adams, J. M., and Kemp, D. J. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 4943. Earl, C. D., and Scher, C. D. (1980).J. Cell. Physiol. 104, 153. Goff, S. P., Gilboa, E., Witte, 0. N., and Baltimore, D. (1980). Cell 22, 777. Goff, S. P., Witte, 0. N., Gilboa, E. Rosenberg, N., and Baltimore, D. (1981). J. Virol. 38, 460. Goff, S. P., D’Eustachio, P., Ruddle, F. H., and Baltimore, D. (1982a). Science 218, 1317. Goff, S. P., Tabin, C. J., Wang, J. Y., Weinberg, R., and Baltimore, D. (1982b). J. Virol. 41, 271. Groffen, J., Heisterkamp, N., Reynolds, F. H., Jr., and Stephenson, J. R. (1983). Nature (London)304, 167. Grunwald, D. J., Dale, B., Dudley, J., Lamph, W., Sugden, B., Ozanne, B., and Risser, R. (1982a). J. Virol. 43, 92. Grunwald, D. J., Miller, K., and Risser, R. (1982b).J. E x p . Med. 156, 1195. Hartley, J. W., Rowe, W. P., and Huebner, R. J. (1970). J. Virol. 5, 221. Heisterkamp, N., Groffen, J.. Stephenson, J. R., Spurr, N. K., Goodfellow, P. N., Solomon, E., Carritt, B., and Bodmer, W. F. (1982). Nature (London) 299, 747. Heisterkamp, N., Groffen, J., and Stephenson, J. R. (1983). J. Mol. Appl. Genet. 2, 57. Hines, D. L. (1983). Blood 61, 397. Hines, D. L., and Gragowski, L. (1983). Leuk. Res. 7, 251. Hoffmann, F. M., Fresco, L. D., Hoffman-Falk, H., and Shilo, B. Z. (1983). Cell 35, 393. Hoffman-Falk, H., Einat, P., Shilo, B. Z., and Hoffmann, F. M. (1983). Cell 32, 589. Hunter, T. (1982).1. Biol. Chem. 257, 4843. Kaplan, P. L., and Ozanne, B. (1983). Cell 33, 931. Kincade, P. W., Lee, G . , Watanabe, T., Sun, L., and Scheid, M. P. (1981).J. Zmmunol. 127, 2262. deKlein, A., van Kessel, A. G., Grosveld, G., Bartram, C. R., Hagemeijer, A,, Bootsma, D., Spnrr, N. K., Heisterkamp, N., Groffen, J., and Stephenson, J. R. (1982).Nature (London) 300, 765. Konopka, J. B., Davis, R. L., Watanabe, S. M . , Ponticelli, A. S., SchitT-Maker, L., Rosenberg, N., and Witte, 0. N. (1984a).J. Virol., 51. Konopka, J. B., Watanabe, S. M., and Witte, 0. N. (19841)). Cell 37, 1035. Krump-Konvalinkova, V., and van den Berg, K. J. (1981). Virology 109, 215. Kung, P. C., Gottlieb, P. D., and Baltimore, D. (1976).J. B i d . Chem. 251, 2399. Kung, J. T., Sharrow, S. O., Mage, M . G . , and Paul, W. E. (1982). J. Zmmunol. 129, 81. Land, H., Parada, L. F., and Weinberg, R. A. (1983). Nature (London) 304, 596. Lane, M. A., Neary, D., and Cooper, G. M. (1982). Nature (London) 300, 659. Latt, S. A., Goff, S. P., Tabin, C. J., Paskind, M., Wang, J. Y., and Baltimore, D. (1983).J. Virol. 45, 1195.

COMPLEXITY OF VIRUS-CELL

INTERACTIONS

97

Lewis, S., Rosenberg, N . , Alt, F., and Baltimore, D. (1982). Cell 30, 807. Moloney, J. B. (1960). J. Natl. Cancer Znst. 24, 933. Muller, R . , Slamon, L). J., Treniblay, J. M . , Cline, M. J . , and Verma, I. M. (1982). Nature (~&nd<JfL)299, 640. Mushinski, J. F., Bauer, S. R., Potter, M., and Reddy, E. P. (1983a). Proc. Natl. Acad. Sci. U.S.A. 80, 1073. Mushinski, J. F., Potter, M . , Bauer, S. R., and Reddy, E. P. (1983b). Science 220, 795. Newmark, P. (1983). Nature (London)303, 383. Ozanne, B., Wheeler, T., Zack, J., Smith, C., and Dale, B. (1982). Nature (London)299, 744. Parks, W. P., Hawk, R. S., Aniscowicz, A , , and Scolnick, E. M. (1976). J . Virol. 18, 491. Patschinsky, T., Hunter, T., Esch, F. S., Cooper, J. A., and Sefton, B. M. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 973. Pillenier, E., Whitlock, C., and Weissman, I. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 4434. Ponticelli, A. S., Whitlock, C. A., Rosenberg, N . , and Witte, 0. N. (1982). Cell 29, 953. Potter, M . , Sklar, M. D., and Rowe, W. P. (1973). Science 182, 592. Pratt, D. M., Strominger, J., Parkman, R., Kaplan, D., Scliwaber, J., Rosenberg, N . , and Scher, C. (1977). Cell 12, 683. Premkumar, E., Potter, M., Singer, P. A., and Sklar, M. D. (1975). Cell 6, 149. Prywes, R., Foulkes, J. G., Rosenberg, N., and Baltimore, D. (1983). Cell 34, 569. Raschke, W. C. (1980a). Biochim. Biophys. Acta 605, 113. Raschke, W. C. (1980b). Cold Spring Harbor Synzp. Quant. B i d . 44, 1187. Raschke, W. C., Ralph, P., Watson, J., Sklar, M., and Coon, H. (1975).J. Natl. Cancer Znst. 54, 1249. Raschke, W. C., Baird, S., Ralph, P., and Nakoinz, I. (1978). Cell 15, 261. Reddy, E. P., Smith, M. J., and Srinivasan, A. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 3623. Reynolds, F. H., Jr., Sacks, T. L., Deohaghar, D. N . , and Stephenson, J. R. (1978a). Proc. Natl. Acad. Sci. U . S . A . 75, 3974. Reynolds, F. H., Jr., van de Ven, W. J. M., and Stephenson, J. R. (1978b). J . Virol. 28, 665. Reynolds, F. H. , Jr., van de Ven, W. J., and Stephenson, J. R. (1980).J. Virol. 36, 374. Rice, D., and Baltimore, D. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 7862. Risser, R. (1982). Biochim. Biophys. Acta 651, 213. Risser, R . , Potter, M., and Rowe, W. P. (1978).J. E x p . Med. 148, 714. Rosenberg, N. (1982). Curr. Top. Microbiol. Zmnzunol. 101, 95. Rosenberg, N . , and Baltimore, D. (1976).J. E x p . Med. 143, 1453. Rosenberg, N . , and Baltimore, D. (1978).J. Exp. Med. 147, 1126. Rosenberg, N., and Witte, 0. N. (1980).J . Virol. 33, 340. Rosenberg, N., Baltimore, D., and Scher, C. D. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 1932. Rosenberg, N. E., Clark, D. R., arid Witte, 0. N. (1980a).J . Virol. 36, 766. Rosenberg, N., Witte, 0. N., and Baltimore, D. (1980b). Cold Spring Harbor Symp. Quant. Biol. 44, 859. Rotter, V., Witte, 0. N., Coffnian, R . , and Baltimore, D. (1980).J. Virol. 36, 547. Rotter, V., Boss, M. A , , and Baltimore, D. (1981). J. Virol. 38, 336. Rotter, V., Abutbul, H., and Wolf', D. (1983). Znt. J . Cancer 31, 315. Ruley, H. E. (1983). Nature (London) 304, 602. Scher, C. I). (1978).J. Exp. Med. 147, 1044. Scher, C. D., and Siegler, R. (1975). Nature (London) 253, 729. Sefton, B. M., Hunter, T., and Raschke, W. C. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 1552. Sefton, B. M., Hunter, T., and Cooper, J. A. (1983). Mol. Cell. B i d . 3, 56. Shields, A,, Otto, C., Coff, S., Paskind, M., and Baltimore, D. (1979). Cell 18, 955. Shilo, B. Z., and Weinberg, R. A. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 6789.

98

CHERYL A . WHITLOCK A N D OWEN N. WI lTE

Siden, E. J.. Baltimore, D., Clark, D., and Rosenberg, N. (1979). Cell 16, 389. Siegler, R., Zajdel, S., and Lane, I. (1972).J . Natl. Cancer Znst. 48, 189. Silverstone, A, E., Rosenberg, N., Sato, V. L., Scheid, M. P., Boyse. E., and Baltimore, D. (1978).In “Differentiation of Normal and Neoplastic Hematopoietic Cells” (B. Clarkson, P. A. Marks, and J. E. Till, eds.), Vol. 2, p. 433. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Sklar, M. D., Shevach, E. M., Green, I., and Potter, M. (1975). Nature (London) 253, 550. Srinivasan, A,, Reddy, E. P., and Aaronson, S. A. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 2077. Srinivasan, A., Dunn, C. Y., Yuasa, Y., Devare, S. G., Reddy, E. P., and Aaronson, S. A. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 5508. Stukart, M. J., Vos, A , , and Melitef, C. J. (1981). Eur. J. Immunol. 11, 251. Sugiyama, H., Akira, S., Yoshida, N., Kishimoto, S., Yamamura, Y., Kincade, P., Honjo, T., and Kishimoto, T. (1982).J . Zmmunol. 128, 2793. Sugiyama, H., Akira, S., Kikutani, H., Kishimoto, S., Yamamura, Y., and Kishimoto, T. (1983). Nature (London) 303, 812. Teich, N. M., and Dexter, T. M. (1978). In “Differentiation of Normal and Neoplastic Hematopoietic Cells” (B. Clarkson, P. A. Marks, and J. E. Till, eds.), Vol. 2, p. 657. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Twardzik, D. R., Todaro, G. J., Marquardt, H., Reynolds, F. H., Jr., and Stephenson, J. R. (1982). Science 216, 894. Waneck, G. L., and Rosenberg, N. (1981a). Hematol. Bluttransfus. 26, 467. Waneck, G. L., and Rosenberg, N. (1981b). Cell 26, 79. Wang, J. Y. (1983). Nature (London) 304, 400. Wang, J. Y., and Baltimore, D. (1983). Mol. Cell. B i d 3 773. Wang, J. Y., Queen, C., and Baltimore, D. (1982). J . B i d . Chem. 257, 13181. Wang, J. Y., Prywes, R., and Baltimore, D. (1983). Prog. Clin. Biol. Res. 119, 57. Wang, J. Y. J., Ledley, F., Goff, S., Lee, R., Groner, Y., and Baltimore, D. (1984). Cell 36, 349. Watanabe, S. M., and Witte, 0. N. (1983). J . Virol. 45, 1028. Whitlock, C. A., and Witte, 0. N. (1981).J. Virol. 40, 577. Whitlock, C. A,, and Witte, 0. N. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 3608. Whitlock, C. A., Ziegler, S. F., Treiman, L. J., Stafford, J. I., and Witte, 0. N. (1983a). Cell 32, 903. Whitlock, C. A., Ziegler, S. F., and Witte, 0. N. (1983b). Mol. Cell. B i d . 3, 596. Whitlock, C. A , , Robertson, D., and Witte, 0. N. (1984). J . Zmmunol. Methods, 67, 353. Witte, 0. N. (1983). Curr. Top. Microbiol. Zmmunol. 103, 127. Witte, 0. N., Rosenberg, N., Paskind, M., Shields, A,, and Baltimore, D. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 2488. Witte, 0. N., Rosenberg, N . , and Baltimore, D. (1979a). J . Virol. 31, 776. Witte, 0. N., Rosenberg, N., and Baltimore, D. (1979b). Nature (London) 281, 396. Witte, 0. N., Dasgupta, A., and Baltimore, D. (1980a). Nature (London) 283, 826. Witte, 0. N., Goff, S., Rosenberg, N., and Baltimore, D. (1980b). Proc. Natl. Acad. Sci. U.S.A. 77, 4993. Ziegler, S. F., Whitlock, C. A,, Goff, S. P., Gifford, A,, and Witte, 0. N. (1981). Cell 27, 477. Ziegler, S. F., Treiman, L. J., and Witte, 0. N. (1984).Proc. Natl. Acad. Sci. U.S.A.81,_1529.

ADVANCES IN IMMUNOLOGY. VOL. 37

Epstein-Barr Virus Infection and lmmunoregulation in Man GIOVANNA TOSATO AND R. MICHAEL BLAESE Metabolism Branch, National Cancer Institute, National lnsiiiuies of Health, Beihesda, Maryland

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Polyclonal B Cell Activation by EBV 111. Relationship between EBV-Induced Immunoglobulin Production and Immortalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. Infectious Mononucleosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Immrmoregulatory Cell Functions in Acute Infectious Mononucleosis . . . . . . . . . VI . Persistent EBV Infection in Normal Individuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Selected Disorders Associated with an Abnormal Regulation of EBV Infection . . VIII. Reversal of Infectious Mononucleosis-Associated Suppressor T Cell Activity by D-Matinose: Suppression and Saccharides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........................................... IX. Concluding Remarks References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

99 102

110 112 115 122 129 138 142 142

1. Introduction

Ever since its discovery in association with a relatively unusual malignancy, the African type of Burkitt’s lymphoma, Epstein-Barr virus (EBV) has continued to provide the biologist with interesting and provocative questions. One of the main areas of interest and confusion is related to the complex biology of this virus that has been associated with a variety of different illnesses and demonstrates unique properties in uitro. EBV has been recognized as a ubiquitous virus that infects the majority of adult normal individuals worldwide (G. Henle and Henle, 1979). The virus is harbored within a small number of circulating lymphocytes (Diehl et al., 1968; Gerber et al., 1969; Nilsson, 1979; Tosato et al., 1984a) and periodically in the oral cavity (Gerber et al., 1972; Miller et al., 1973; Chang et al., 1973). Primary infection with this herpes virus is generally asymptomatic, particularly in childhood, but may result in acute infectious mononucleosis (IM) in adolescents or adults (Henle et al., 1968; Niederman et al., 1968; Evans et d., 1968; Sawyer et al., 1971; W. Henle and Henle, 1979). Generally a benign and self-limited acute illness, EBV-induced IM has been associated with the appearance of hypogammaglobulinemia and malignant polyclonal lymphoproliferation (Provisor et al., 1975; Sugden, 1982). In 1974, Bar and co-workers reported of a fatal case of IM in a family with a history of other similar fatal cases among male members. They suggested involvement of EBV, since the serology was compatible and the virus was recovered from lymph nodes and spleen (Bar et al., 1974). Similar cases have 99 Cop)right 0 1985 by Academic Press, Inc AII rig~,trof reproduction in any form rererved ISBN 0-12-022437-2

100

GIOVANNA TOSATO AND 1%. MICHAEL BLAESE

since been reported in other rare families, and an X-linked, recessive, lymphoproliferative syndrome (XLP) has been defined (Purtilo et al., 1977; Hamilton et al., 1980). More recently, isolated, nonfamilial cases of malignant polyclonal lymphoproliferation of EBV-infected B cells have also been described in association with acute EBV-induced IM and in EBV seropositive individuals (Britton et al., 1978; Virelizier et al., 1978; Robinson et al., 1980). Interestingly, some of these patients were undergoing treatment with immunosuppressive agents or anti-T cell monoclonal antibodies (Crawford et al., 1980; Martin et al., 1984). In contrast to these EBV-related lymphoproliferative disorders which are often polyclonal in nature, the African type of Burkitt’s lymphoma is a monoclonal malignancy of B cells usually infected with EBV which have also been found to contain characteristic chromosomal translocations and DNA rearrangements involving the cellular myc oncogene (Klein et al., 1968; Gunvkn et al., 1980; Lindahl et al., 1974; Fialkow et al., 1970; Manalov and Manalava, 1972; Mitelman, 1981; Dalla-Favera et d., 1982; Nee1 et al., 1982; Taub et al., 1982). The precise role of EBV in the etiology and pathogenesis of this lymphoid tumor is still unclear and various hypotheses have been advanced to explain how a number of cofactors, including endemic malaria, could affect infection with EBV and cause this type of cellular transformation (Morrow et al., 1976; de-Th6 et al., 1978). In addition to African Burkitt’s lymphoma, EBV has been associated with a different type of malignancy, nasopharyngeal carcinoma, where the neoplastic EBV-infected cells are of epithelial lineage (Old et al., 1966; Henle et al., 1970b; Desgrandes et al., 1975; Anderson-Anvret et al., 1977; Huang et al., 1978). Recently, a new clinical syndrome, chronic infectious mononucleosis, has been recognized and attributed to a persistent active infection with EBV (Tobi et al., 1982; Straus et al., 1984). This relatively benign but often disabling long lasting illness is characterized by weakness, malaise, recurrent fevers, and is associated with an abnormal antibody reactivity to EBV antigens. Thus, EBV, a ubiquitous virus infecting essentially all of the normal adult population and the causative agent of a self-limited benign illness such as acute I M also has a role in the development of both monoclonal and polyclonal malignant lymphoproliferations of B cells, is associated with a malignancy of epithelial cells, and appears to be involved in a chronic, but benign illness as well. The wide spectrum of clinical syndromes associated with EBV infection has led to the hypothesis that different “strains” of EBV might be responsible for these different conditions (Pizzo et al., 1978; Kieff et al., 1983). This hypothesis has been a central theme in EBV research during the last few years and has led to important clarifications of the structure of this virus

EPSTEIN-BARR VIRUS INFECTION

101

(Kieff et al., 1983). It has failed, however, to demonstrate a unique association between a certain strain and a certain disorder (Given and Kieff, 1978; Raab-Traub et al., 1980; Dambaugh et al., 1980; Fisher et al., 1981; Heller et d., 1981a,b). In addition, with the single exception of P3HR-1, all other EBV “strains” cause normal human B cells to grow indefinitely in uitro (Henle et al., 1967; Pope et al., 1968; Miller et al., 1969; Gerber et al., 1972; Magrath et al., 1983). This transforming property of EBV has been consistently shared by EBV isolates from Burkitt’s tumors, patients with acute IM, and normal virus carriers. In contrast, P3HR-1 EBV, originally isolated from a Burkitt’s tumor, seems to have generally lost on long-term laboratory passage, its capacity to transform cells. It can, however, induce an abortive cycle of viral replication when added to certain EBV-infected B cell lines (Hinuma et al., 1967; Henle et aL., 1970a; Miller et al., 1974; Yajima and Nonoyama, 1976; Menezes et al., 1975). There is extensive antigenic similarity in the viral envelopes of all strains examined so far and no distinct serologic subtypes have been recognized (Coope et al., 1979; Thorley-Lawson, 1979b; Hoffman et al., 1980). Early studies established that there is at least a 90% homology between EBV DNAs from widely different origins (zur Hausen et al., 1972; Kawai et al., 1973; Kieff and Levine, 1974; Pritchett et al., 1975). These studies could not exclude the possibility that 10% of the viral DNA (approximately 10 x lo6 Da) might be different and recent detailed analysis of viral genomic DNAs by nucleic acid hybridization and comparison of restriction endonuclease digests has indicated several differences among various EBV isolates (Bornkamm et al., 1980; Kieff et al., 1983). For example, it has been shown that B95-8 EBV has a deletion involving approximately 10 MDa of its DNA when compared with EBV DNA-derived from Burkitt’s isolates, such as Ag876 or W91, or from a different acute IM isolate, FF41 (Raab-Traub et al., 1980; Fisher et al., 1981). It has also been reported that P3HR-1 DNA is heterogeneous and consists of at least two different sets of molecules (Sugden et al., 1976; Fresen et al., 1977; Delius and Bornkamm, 1978; Heston et al., 1982; Rabson et al., 1983). All these studies, however, have failed to correlate a unique viral DNA sequence with a particular geographic source or type of disease (Sugden, 1982; Kieff et al., 1983). Thus, it appears unlikely that major insertions and deletions of the EBV genome account for differences in disease manifestations associated with EBV infection. Nonetheless, the role of subtle differences in DNA sequence of EBV isolates cannot be excluded since detection of such changes is still beyond the sensitivity of the assays that have been used (Kieff et al., 1983). Alternative hypotheses have been advanced to explain the wide spectrum of clinical manifestations associated with EBV infection. Many believe that even if infection with the same virus occurs in all EBV-related disorders, the

102

GIOVANNA TOSATO AND 13. MICHAEL BLAESE

outcome of this infection is conditioned by a variety of environmental cofactors and by the quality of the immune response of the host (Ziegler, 1981; Fisher and Rabson, 1982; Sugden, 1982). In this review, we will attempt to summarize the known immunoregulatory mechanisms of EBV infection and will demonstrate that this herpes virus has provided us with a powerful experimental tool for the study of B cell development, function, and control. II. Polyclonal B Cell Activation by EBV

It has long been known that EBV induces normal human lymphocytes to proliferate in vitro and to become activated into long-term cell lines that can be propagated in vitro for years (Henle et al., 1967; Pope et al., 1968; Miller et al., 1969; Gerber et al., 1969; Nilsson et al., 1971). The induction of indefinite cellular proliferation in vitro by EBV is termed “transformation” or “immortalization” or “outgrowth.” Characteristic features of these “immortal” cell lines include their B cell phenotype and their ability to secrete all major classes of immunoglobulin (Ig) for months and sometimes years (Nilsson, 1971, 1979; Nilsson and Ponten, 1975; Nilsson and Klein, 1982). Transformed cell lines derived from EBV infection of normal mononuclear cells at nonlimiting viral and cellular concentrations first produce simultaneously Ig of all three major isotypes, although eventually one or a few B cell clones within the line outgrow the remaining cell population and, as a result, Ig of one class may become the only or the predominant secreted product (Bechet et al., 1974; Nilsson, 1979; Nilsson and Klein, 1982). Thus, Ig synthesis represents a characteristic feature of EBV-transformed normal B cells. A similar association between infection with EBV and Ig production was suggested by the characteristic occurrence of hypergammaglobulinemia during acute EBV-induced IM. In this illness serum Ig levels of all major isotypes become generally elevated, and the specific response to EBV antigens is known to account for only a small proportion of this serum Ig elevation (Wollheim and Williams, 1966; Sutton et al., 1973). The property of EBV lines to secrete Ig in uitro and the occurrence of hypergammaglobulinemia during primary EBV infection in vivo led to the hypothesis that this herpes virus was a polyclonal activator of human B cells and, like other mitogens, could induce the production of polyclonal Ig in vitro. Rosen et al. (1977) demonstrated that this hypothesis was correct and showed that, following exposure to EBV, lymphocytes secret polyclonal Ig in culture as evidenced by the presence of p and y heavy chains as well as K and A chains in the supernatants. Since this original observation, the characteristics of polyclonal B cell activation by EBV have been analyzed in detail and our understanding of

EPSTEIN-BARR

VIRUS INFECTION

103

this process is now much more complete. In the presence of EBV (B95-8 strain), human mononuclear cells begin to produce Ig after 3 or 4 days in culture and the number of Ig-secreting cells increases during the following 4 or 5 days of observation (Fig. 1).The B cell response to EBV in uitro is not significantly different from the response to other activators such as pokeweed mitogen (PWM) during the first 7-8 days in culture. Similar to other polyclonal activators, EBV induces the production of all major classes and subclasses of Ig in uitro (Bird and Britton, 1979b; Kirchner et d.,1979; Andersson et al., 1981). Furthermore, lymphocytes from both EBV-immune and nonimmune donors can be activated by EBV in vitro to secrete Ig, and the magnitude of the B cell response is not different between these two groups of donors (Fig. 2) (Tosato et al., 1982~). Infection of the lymphocytes with the virus represents a necessary step to achieve B cell activation and Ig production and is believed to involve a specific receptor for EBV (Jondal et al., 1976; Einhorn and Ernberg, 1978; Simmons et al., 1983; Hutt-Fletcher et al., 1983). Bird et al. have clearly shown that EBV binding to the specific surface receptor is not sufficient to

DAYS

FIG.1. EBV is a polyclonal B cell activator. Mononuclear cells (1.0 X lo6) from 5 normal individuals were cultured in the presence of EBV (BY5-8 strain) or pokeweed mitogen (PWM), and the imrnunoglo1)ulin secreting cell response determined at 2, 4, 6, and 8 days of' culture.

104

GIOVANNA TOSATO AND 1%. MICHAEL BLAESE

a

ti: v) -

1.ooo

(500

c

-Geo

Seronegative

Mean

Seropositive

FIG. 2. Immunity to EBV of the lymphocyte donor is not required for immunoglobulin production in the presence of exogenous EBV. Mononuclear cells (1.0 X lo6) were obtained from normal adult individuals seropositive or seronegative for EBV and were cultured in the presence of EBV (B95-8 strain) for 8 days. At the end of the culture period the immunoglobulinsecreting cell response was determined.

induce B cell activation in vitro (Bird and Britton, 1979a). Using ultraviolet light-inactivated virus to saturate all the available “receptor sites,” they showed that no Ig production or B cell proliferation was induced. Only if biologically active EBV was used, was B cell activation observed. Furthermore, it was reported that the majority of the EBV-induced plaque-forming cells express EBNA, the EBV-related nuclear antigen (Bird and Britton, 1979a). This antigen, detected by iinmunofluorescence techniques, represents a specific marker for EBV infection and appears 10-24 hours after infection with this virus (Reedman and Klein, 1973; Aya and Osato, 1974; Yata et al., 1975; Einhorn and Ernberg, 1978). A related observation was made by Robinson et al. (1981), who looked for evidence of EBV infection in the circulating Ig-secreting cells of patients with acute EBV-induced IM and found that the majority of these cells coexpressed EBNA and cytoplasmic Ig. The continuous presence of exogenous EBV in culture is not necessary for the induction of B cell activation and differentiation. If extracellular EBV is removed from lymphocyte cultures after 1-2 hours of exposure, Ig production at 7 days is similar in magnitude to that observed if the virus had not been removed. Indeed, infection of human lymphocytes in vitro with exogenous EBV is known to occur rapidly and to be completed within an hour after exposure (Thorley-Lawson and Strominger, 1978). Endogenous production of EBV occurs only at a very low level in cultures of human lymphocytes infected with the virus, and cross infection in vitro does not appear to contribute significantly to the activation of B cells not initially infected. In

EPSTEIN-BAKE V I R U S INFECTION

105

different clonal cell lines obtained by EBV transformation of adult lymphocytes, the frequency of cells releasing infectious EBV during a 24-hour period has been reported to be 1 in 10"-10" cells (Sugden et al., 1979; Sugden, 1982). Recent work has indicated that a single EBV particle is sufficient to induce an individual B cell and its progeny to produce Ig in vitro (Yarchoan et al., 1983). Similar results were obtained by others who showed that immortalization of human lymphocytes by EBV follows "one-hit" kinetics (Sugden and Mark, 1977; Henderson et al., 1977). In these studies it was calculated that 1 of every 10-30 EBV particles is competent to infect B cells (Sugden, 1982). EBV-transformed cells, however, usually harbor multiple copies of the viral genome (Kawai et al., 1973; Nonoyaina and Pagano, 1971; zur Hausen et al., 1972; Pritchett et al., 1976; Sugden et al., 1979). Two mechanisms have been reported to account for this finding, either infection with many EBV particles, each contributing one copy of the viral genoines or viral DNA amplification after infection has occurred (Sugden et al., 1979). Unlike most other known human B cell activators, EBV of the B95-8 strain stimulates B cell proliferation, differentiation, and Ig production independent of cooperation by other cells such as helper T cells and monocytes (Kirchner et al., 1979; Bird and Britton, 1979a). As shown in Fig. 3, B cells depleted of T cells respond to EBV but not to the T cell-dependent activator PWM with the generation of Ig-secreting cells. Furthermore, the addition of autologous T cells does not significantly effect, either by enhancing or inhibiting, the B cell response to this virus after 7 days in culture. A similar finding was observed when rnonocytes were removed from the mononuclear cell populations: while the monocyte-dependent responsiveness to PWM was reduced or lost, the EBV-induced responsiveness was not effected (Fig. 4). In addition, production of all Ig isotypes also occurred in the absence of T cells and monocytes (Kirchner et al., 1979). Production of Ig of multiple isotypes can either derive from Ig class switching by individual precursors that can thus produce Ig of different isotypes or alternatively by maturation of individual precursors committed to the production of Ig of one single isotype (Gearhart et al., 1975; Andersson et al., 1977; Benner et al., 1981; Teale et al., 1981). Using cultures of limiting numbers of B cells stimulated with EBV (B95-8) Yarchoan et al. (1983) showed that in the presence of EBV (B95-8)individual precursors and their progeny only produce IgG or IgM, but not both types of Ig (Table I). Thus, single EBV-infected B cells only differentiate to secrete Ig of the isotype they are committed to produce and do not undergo Ig class switch. Not all human B cells can be induced to produce Ig in vitro in response to EBV (B95-8). In the presence of nonlimiting numbers of infectious viral particles, the frequency of circulating B cells that can be activated by EBV in

106

GIOVANNA TOSATO A N D R. MICHAEL BLAESE

aMedium 32,000

w

16,OoO

K

3 t2 3

Y

v)

8rooo

r

t

=

*

PWM

&% EBV 1895.81 ?l

I

Geo. Mean

I SEM 120 Determinations1 T

t

T

-cn

MNC

B CELLS

B

+ T CELLS

FIG. 3. EBV (B95-8 strain) is a T cell-independent B cell activator. Mononuclear cells (MNC, 1.0 x lo"), B cells (0.5 x lo"), and autologous mixtures of B (0.5 X 106) and T (1.0 x lo6) cells were incubated in the presence of EBV (B95-8)or pokeweed mitogen (PWM) for 7 days, and the immunoglobulin-secretingcell response was determined at the end of the culture period.

vitro is low (Sugden and Mark, 1977; Henderson et al., 1977; Yarchoan et al., 1983; Stein et al., 1983; Martinez-Maza and Britton, 1983). In our experiments, on the average, 1 or 2% of the circulating surface Ig-positive B cells could be induced to secrete Ig in vitro (Fig. 5). Other laboratories have reported slightly lower (Martinez-Maza and Britton, 1983) or higher frequencies (Stein et al., 1983), probably depending on the culture conditions and on the assay system used. Similar results were obtained when other EBV-related parameters were studied, including the expression of EBNA and outgrowth of EBV-immortalized B cells (Sugden and Mark, 1977; Henderson et al., 1977). The characteristics of the B cell subset that is susceptible to EBV activation are still poorly defined. Recently, it was reported that only B lymphocytes with high bouyant density can be induced by EBV to produce EBNA and become "immortal," suggesting that resting lymphocytes are the predominant targets for EBV activation (Aman et al., 1984). It is interesting to note that although cellular infection with biologically active EBV is a necessary condition for the expression of EBNA, for B cell proliferation, for Ig production, and for immortalization, infection alone is not sufficient. For example, transfer of EBV receptors to the membranes of EBV receptor-negative cells (mouse fibroblasts, mouse T cells, human T

EPSTEIN-BARR

107

VIRUS INFECTION

0

Medium

lo'oOO 3

F

PWM

D E B V 1895.81

5,000

5 3

9 y

2,500

-1

w

u

MONONUCLEAR CELLS

MONOCYTE-DEPLETED MONONUCLEAR CELLS

FIG.4. EBV (B95-8 strain) is a nionocyte-independent B cell activator. Mononuclear cells (1.0 x lo6)and monocyte-depleted mononuclear cells (1.0 x 106) were cultured with pokeweed mitogen (PWM) or EBV (B95-8), and the immunoglobulin-secreting cell response was determined after 7 days in culture.

TABLE I INDIVIDUAL B CELLPRECURSORS PRODUCEIgG OR IgM BUT NOT BOTIIISOTYPES I N THE PRESENCE O F EPSTEIN-BARR VIRUS Number of wells containingb Exp.

1 2 3

Total number ofculturesa

324 36 288

G-M-

130 15 124

G-M+

G+M-

G+M+

Predicted G+M+C

p valued

139 5 110

26 11 28

29 5 26

28.5 4.4 25.5

>0.9 >0.9 N.9

" Multiple replicate cultures were established containing 60 B cells per well; this number was selected to yield, on the average, less than 1 precursor B cell per culture for immunoglobulin production by EBV. b Number of wells containing either no IgG and IgM, one or both isotypes. c Predicted number of cultures containing IgG and IgM based on the total number of IgG and IgM containing wells and on the assumption that separate precursors produce either IgG or IgM but not both isotypes. d Significance level for the difference between the experimental results and those predicted under the hypothesis that there are no precursors producing both IgM and IgG.

108

GIOVANNA TOSATO AND H. MICHAEL BLAESE

FIG.5. Limiting dilution analysis of immunoglobulin production in the presence of EBV. Multiple replicate cultures containing 105 irradiated (3000 rad) T cells and varying numbers of B cells preincubated with EBV (B95-8) were established. The fraction of cultures not secreting immunoglobulin was determined for each dose of B cells (surface immunoglobulin positive cells) and plotted against the B cell number per culture; the precursor frequency was calculated by Poisson’s statistics. In this experiment approximately 1 in 60 B cells was activated by EBV to secrete immunoglobulin.

cells, etc.) allows the virus to penetrate and this event is associated with the expression of certain virus-related antigens (nuclear, early, and viral capsid antigens) by the infected cells, but does not result in cell immortalization (Graessmann et al., 1980; Volsky et al., 1980, 1981). Similarly, the infection of human placental cells with purified EBV DNA by means of transfection is not associated with the development of evidence of “transformation” in these nonlymphoid cells (Miller et al., 1981). Most of the EBV isolates obtained from different sources (normal individuals, patients with acute EBV-induced IM, patients with the African type of Burkitt’s lymphoma, or other diseases) induce normal human B cells to transform into long-term cell lines. The only exception to this rule is the P3HR-1 strain of EBV (Hinuma et al., 1967). The cell line producing this virus was obtained by cloning of a Burkitt’s-derived cell line (Jijoye) that, unlike its clone, releases virus that has retained the property to induce cellular outgrowth (Ragona et al., 1980). Nontransforming EBV is structurally similar to the transforming virus, but has the unique property of inducing an abortive cycle of viral replication in certain EBV-infected cell lines, such as Raji, with concomitant expression of EBV-related early antigens (Henle et al., 1970a). We have examined EBV P3HR-1 for its ability to induce Ig production in

EPSTEIN-HARK

109

VIHUS INFECTION

vitro. Although some laboratories have failed to obtain Ig secretion with this EBV strain (Tsukuda et al., 1982), others have consistently found it (Tosato et al., 1981a). The reasons for this discrepancy are not as yet clear, but might be related to the well-defined heterogeneous nature of the viral particles produced by the PSHR-1 cell line (Fresen et al., 1977) and to the possibility that, in different laboratory sublines of PSHR-1, the distribution of the heterogeneous particles might be different. Interestingly, conflicting results on other function of this EBV strain have been previously reported (Henle et d., 1967; Gerber et al., 1969; Aya and Osato, 1974; Miller et al., 1974). Ig production by EBV of the PSHR-1 strain in our laboratory has demonstrated strikingly different characteristics from Ig production by EBV of the B95-8 strain. As shown in Table 11, unlike the B95-8 strain which is a T cellindependent activator of B cells, EBV produced by the P3HR-1 cell line required T cells to induce Ig production. Thus, the nontransforming EBV is a T cell-dependent polyclonal R cell activator. However, specific viral immunity to EBV of the lymphocyte donor did not appear to be necessary for Ig production by EBV-PSHR-1, since lymphocytes from EBV sero-negative individuals could also be induced to secrete Ig. Pretreatment of EBVPSHR-1 with neutralizing antibodies, or with ultraviolet light, eliminated the activating capability of this virus, indicating that, similar to B95-8, Ig production by EBV-PSHR-1 requires functionally active virus (unpublished observation). Kinetic analysis of B cell activation by EBV of the PSHR-1 strain showed that Ig production begins on day 4 or 5, reaches a peak by day 7 or 8, and decreases thereafter, being undetectable by day 12 or 13. Unlike EBV of the B95-8 strain, PSHR-1 did not induce purified B cells to proliferate, but had a clear mitogenic effect on purified T cells. These results show that, unlike transforming EBV, B cell activation by EBV of the PSHR-1 strain requires more than simply an interaction between the viral particles and the B cells, and suggests that this virus may have more complex effects TABLE I1 LYMPHOCYTE ACTIVAT~ONBY DIFFERENT STHAINS O F

EPSTEIN-BARR

VlRUS

Immunoglobulin secreting cellsu Cell cultures Mononuclear cells (1 x 10") B cells (0.5X106) B(0.5x 106) T(l.Ox lo6) cells

+

B(0.5~106)+ T2°00K(1.0X106)cells

Medium

B95-8

P3HR-1

253 211 789 636

3242 3104 2749 3917

2234 109 1632 4005

a Immunoglobulin secreting cells produced per culture at the end of a 7 day culture period.

110

GIOVANNA TOSATO A N D R. MICHAEL BLAESE

on human lymphocytes than previously recognized. Particularly interesting are the effects of P3HR-1 on T cells and the role of T cells in B cell activation by this EBV strain. An important unresolved question is the mechanism of T cell proliferation in response to this type of EBV, since T cells are believed to lack “classical” EBV receptors (Jondal et al., 1976). 111. Relationship between EBV-Induced Immunoglobulin Production and Immortalization

It is generally believed that Ig production by EBV (B95-8)-infected B cells occurs only if these cells are also immortalized by the virus (Rickinson and Moss, 1983). Ig production and outgrowth by EBV would thus represent only different manifestations of the transforming activity of this virus. The evidence supporting this view is several fold. Both processes require infection of the B cell by live viral particles (Bird and Britton, 1979a; Miller et al., 1974) and are associated with the expression of EBNA by the Ig-secreting as well as by the transformed cells (Bird and Britton, 1979a; Nilsson and Klein, 1982). Immortalization and Ig secretion by EBV can both be induced by infection of a B cell with a single EBV virion (Henderson et al., 1977; Yarchoan et al., 1983). Both of these processes occur through a direct interaction of EBV with a B cell, without a requirement for cooperation of these B cells with other cells or soluble mediators, irregardless of the immune status of the lymphocyte donor with respect to EBV. Finally, the vast majority of long-term cell lines induced by infection of normal B cells with EBV produce Ig (Nilsson and Klein, 1982). It has been observed, however, that Ig secretion may be dissociated from the occurrence of B cell immortalization by EBV in certain instances. For example, it was noted that not all of the Ig-secreting cells express the EBVinduced nuclear antigen EBNA after infection in vitro with EBV-B95-8 (Bird and Britton, 1979a). This antigen is usually expressed in cells that are infected with transforming EBV (Einhorn and Ernberg, 1978; Takada and Osato, 1979). It was also observed that a small, but definite proportion of the plasma cells found in the circulation of patients with acute EBV-induced IM, do not express EBNA (Robinson et al., 1981). A few cell lines have been obtained by EBV infection of normal lymphocytes that lacked surface, cytoplasmic, or secreted Ig (Fu and Hurley, 1979). In addition, infection of fetal liver lymphocytes with EBV resulted in the outgrowth of cell lines with pre-B characteristics (Hansson et al., 1983) as well as of cell lines that lacked the expression of surface, cytoplasmic, or secreted Ig and failed to demonstrate the typical Ig gene rearrangement normally found in lymphocytes of the B cell lineage (Katamine et al., 1984). In recent experiments (Table 111), we have determined that the frequency

111

EPSTEiN-BARR VIRUS INFECTiON

TABLE 111 IMMUNOGLOBULINPRODUCTION BY EBV Is NOT ALWAYSAssocixrED W I T II IM MOHTAL~ZAT~ON Observed number of positive wellsb Exp.

1 2

Total number of cultures

Ig-/trans-

186 90

51 67

Ig-/trans

3 0

+

Ig+/trans-

Ig+/trans

79 12

53

+

11

Multiple replicate microcultures were established with lo5 irradiated T cells and a small number of autologous EBV-infected B cells, to yield, on the average, less than one precursor per well for immunoglobulin production by EBV. At the end of 6 weeks culture period each well was scored for transformation and immunoglobulin secretion. b Ig-/ +, trans -/+: number ofwells neptiveipositive for immunoglobulin production and for transformation.

of circulating B cells that can be induced to produce Ig in 6itr-oby addition of exogenous EBV (B95-8) is significantly larger, by approximately twofold, than the frequency of cells that are immortalized (Tosato et al., 1 9 8 3 ~ ) . When Ig production and outgrowth of EBV-activated B cells were simultaneously examined by limiting dilution at a single cell level, we confirmed that while most (approximately 95%) of the immortalized cells secreted Ig, only approximately 50% of these single cell derived cultures which had secreted Ig showed morphologic evidence of outgrowth after 6 weeks in culture. Study of the kinetics of EBV-induced Ig production by individual nonimmortalized B cells revealed that the majority of the Ig was produced in the first 6 weeks of culture and that little or no Ig was secreted after that time; in contrast, immortalized B cells continued to secrete Ig throughout a 10-week period of observation. Analysis of the Ig isotypes secreted indicated that B cell precursors committed to IgM production had a significantly greater chance of becoming transformed upon infection with EBV than those committed to IgG or IgA production (unpublished observation). The occurrence of Ig production by EBV-activated B cells that do not eventually give rise to long-term cell lines could be due to a phenomenon of “abortive transformation.” This interpretation would suggest that the cells only capable of a transient production of Ig and those capable of both stable Ig production and continuous growth have undergone an identical type of interaction with the virus, at least initially. Thus, premature death of potentially “immortal” EBV-infected cells would explain the observed dissociation between immortalization and Ig production. Conversely, the absence of Ig secretion by certain B cells immortalized by EBV could be explained on the basis of immaturity of the virally infected cells that are not as yet committed to the production of Ig.

112

GIOVANNA TOSATO A N D R . MICHAEL BLAESE

An alternative explanation for these findings is that Ig production by EBV is the result of a qualitatively different cell-virus interaction from that associated with immortalization by EBV. It could be hypothesized that a certain EBV DNA sequence is responsible for inducing B cell outgrowth, while a distinct sequence is responsible for the induction of Ig secretion. Detailed analysis of viral DNA sequences of transforming and nontransforming EBV strains has demonstrated a number of differences but has so far failed to attribute with certainty the immortalizing property of EBV to a certain sequence of viral DNA (Griffin and Karran, 1984). The different outcomes of B cell infection with EBV could also depend upon the nature or the activation state of the infected cell. In this respect, the isotype commitment of the B cell appears to be an important factor, since we have observed that EBVinfected B cells secreting IgG or IgA are less likely than EBV-infected IgM secreting cells to become stable transformants, and are thus relatively more unstable. In addition, it has been shown that while a proportion of the circulating B cells can be infected with EBV, as demonstrated by the intracellular presence of labeled virus, some of these do not go on to express EBNA (Aman et al., 1984). Furthermore, T cells and placental cells can be infected with EBV and can go on to express EBNA, and yet do not give rise to long-term cell lines (Volsky et al., 1980; Miller et al., 1981). Understanding the relationship between EBV-induced Ig production and induction of long-term cell growth is a challenging problem which will require a detailed analysis of the cell-virus interactions at a molecular level.

IV. Infectious Mononucleosis

Infectious mononucleosis (IM) is a self-limited benign disorder characterized by malaise, sore throat, lymphadenopathy, and lymphocytosis associated with the appearance of atypical cells in the peripheral blood. While a variety of agents have been associated with this syndrome, EBV is the most common cause of IM in adult individuals (Niederman et al., 1968). Apparently the oropharynx is the site of initial infection with EBV. Here the virus infects the epithelium of the upper respiratory tract and the salivary glands where it replicates and is subsequently released into the oral cavity (Epstein and Achong, 1977; Lemon et al., 1977; Morgan et al., 1979; Sixbey et al., 1984). Generalization of the infection appears to be mediated by B lymphocytes originally infected with EBV in the oropharynx that later migrate to distant lymphoid organs and to the peripheral blood. It has been clearly shown that among blood cells only lymphocytes with a B cell phenotype become naturally infected with EBV (Jondal and Klein, 1973; Mizuno et at., 1974; Greaves and Brown, 1975). However, a prominent T cell, rather than

EPSTEIN-BARR

VIRUS INFECTION

113

B cell lymphocytosis is a hallmark of the peripheral blood during IM. The majority of these T cells have been shown to bear a suppressor-cytotoxic phenotype (T8, Leu 2 positive) (Reinherz et aZ., 1980; De Waele et al., 1981). Several unique features characterize primary infection with EBV in man. First, the virus isolated from the saliva of patients with acute IM has the property of inducing normal human B cells to grow indefinitely in uitro, a property shared with most of the laboratory strains of EBV (Miller et al., 1973). A related finding is that the B cells naturally infected with EBV in uiuo can grow continuously in culture giving rise to long-term B cell lines in which each cell is infected with the virus (Gerber et al., 1969; Chang et al., 1971; Nilsson, 1971). It is well established that clinical resolution of acute IM is not associated with a complete clearing of EBV infection from the body. Rather, for years and probably for the life time of the host, the virus can both be isolated from the oral cavity and can be recognized within a small proportion of the circulating B cells which maintain the property of spontaneous outgrowth when cultured in uitro (Miller et al., 1973; Gerber et al., 1969; Chang et al., 1971; Nilsson, 1971). A central theme of research on EBV has been the understanding of how the host controls natural infection with EBV. This is a particularly interesting question since, unlike other infectious agents, EBV has the property of immortalizing B cells conferring the property of continuous and potentially unlimited growth. Neutralizing antibodies that appear during the acute phase of IM and persist for life thereafter play an important role in preventing viremia and subsequent infection of circulating lymphocytes (Henle and Henle, 1979a; Rickinson et al., 1975). It is not clear, however, whether antibodies can prevent the intracellular passage of infectious virus from a productively infected cell to a susceptible cell in contact with it nor whether they can protect epithelial cells in the oropharynx from natural infection (Rickinson et aZ., 1974). These antibodies are specific for an antigen complex composed of three glycoproteins (GP 350, GP 220, GP 85) and one polypeptide (P 140) expressed on the virions and on the membranes of EBV-producing cell lines and have been shown to prevent infection by exogenous virus added to lymphocytes in uitro (Pearson and Qualtiere, 1978; Thorley-Lawson, 1979a,b; Thorley-Lawson and Edson, 1979; Edson and Thorley-Lawson, 1981). After viral absorption and penetration have occurred, however, the virus is protected from the effects of neutralizing antibodies, and the addition of EBV-immune sera containing neutralizing antibodies at this time has no inhibitory effect on immortalization of these EBV-infected cells (Rickinson et al., 1979). Thus, neutralizing antibodies are probably of importance in limiting B cell infection by the EBV that is continuously released in the oropharynx, but seem to have little role in preventing infection derived

114

GIOVANNA TOSATO A N D R . MIC H A EL BLAESE

from cell-cell contact and in controlling B cells already infected with the virus. Considerable evidence supports the view that cellular immunity mediated by T lymphocytes is of primary importance for the control of B cells endogenously infected with EBV during initial infection, although understanding of the mechanism or mechanisms of cellular immune control is incomplete. A major reason for confusion derives from the observation made in several laboratories that T cells obtained from patients with acute IM fail to inhibit in vitro the outgrowth of cell lines from autologous B cells endogenously infected with EBV (Rickinson et al., 1980a; Schooley et al., 1981).Thus, if B cells are purified from the peripheral blood of patients with acute IM and are set up in culture alone or recombined with the autologous T cells at a ratio comparable to that found in the circulation, outgrowth of EBV-infected B cells is usually observed after 4-6 weeks in both types of culture. Similarly, cultures of mononuclear cells from patients with acute IM regularly give rise to long-term cell lines in vitro. We have analyzed this phenomenon in detail and determined the frequency of spontaneous immortalization by limiting dilution cultures of IM B cells alone or in coculture with irradiated or nonirradiated autologous T cells. Our results (Table IV) confirm previous observations that IM T cells have no effect on the spontaneous outgrowth in vitro of autologous B cells endogenously infected with EBV, since the frequencies of spontaneous transformation in cultures of IM B cells alone (not shown), or mixed with TABLE IV OF ACUTEINFECTIOUS MONONUCLEOSIS T CELLSTO FAILURE SUPPRESS SPONTANEOUS OUTGROWTH OF AUTOLOCOUS B CELLS Frequency of spontaneous outgrowth/l06 B cellsb Patient0

B

+ auto T cellsc

B

+ auto irradiated T cells

188

190

337 395

324 327 222

158

Patients with acute EBV-induced infectious mononucleosis. The frequencies of spontaneous B cell outgrowth were determined by Poisson statistics, and are expressed as number of spontaneously transforming B cells per lo6 B cells (surface Ig-positive cells). Multiple replicate microcultures were established containing different numbers of non-T cells (5.0X lo5, 2.5X lo5, 1.25X105,etc.) mixed in culture with autologous T cells (2x lo5)either nonirradiated or irradiated with 2000 rad prior to culture. a

EPSTEIN-BAHH

VIRUS INFECTION

115

irradiated T cells were not different from those observed in cultures of the same B cells mixed with autologous T cells not inactivated by irradiation. These observations have been interpreted to indicate that T cells do not significantly contribute to the control of EBV-infected B cells during acute IM. However, using a variety of other assay systems generally requiring short-term culture, it was clearly shown that T cells during acute IM mediate a number of regulatory functions resulting in an effective control of EBVactivated B cells, as discussed below. How can one reconcile the discrepancy between the apparent inability of IM T cells to control the outgrowth in uitro of autologous B cells naturally infected with EBV in viuo with their effective regulatory function demonstrated in short-term cultures? One possibility is that circulating IM regulatory T cells are terminally differentiated and short lived. Thus, while functionally active T cells are continuously generated in uiuo assuring continuous and prolonged immunoregulation of the infection, most of these immunoregulatory T cells die early in culture. Alternatively, IM T cells are not short lived, but culture conditions are inadequate and result in a premature T cell death. Either of these alternatives would explain a lack of IM T cell control of the spontaneous outgrowth of EBV-infected B cells in uitro, since assays for outgrowth require a 4- to 6-week culture period. Thus, assay involving a long-term culture may not represent an appropriate system for testing T cell immunity during acute IM and may even be potentially misleading. V. lmmunoregulatory Cell Functions in Acute Infectious Mononucleosis

Several investigators have reported that a variable proportion of patients with acute IM develop activated cytotoxic cells in the peripheral circulation that are no longer found during convalescence (Svedmyr and Jondal, 1975; Turz et al., 1977; Lipinsky et al., 1979; Seeley et d.,1981; Klein et d.,1980, 1981; Pate1 et al., 1982). This activity was demonstrated by incubating fresh IM mononuclear cells, or Fcy receptor negative mononuclear cells, or purified T cells with appropriate targets at a high effector-to-target ratio in rather long (6 to 12 hours) chromium-release assays. Only EBV-infected cells could function as targets for IM cytotoxic activity, since a variety of EBVnegative cell lines, including K562, were not killed in these assays (Svedmyr and Jondal, 1975; Lipinski et al., 1979; Seeley et al., 1981). In addition, this cytotoxic interaction was not restricted by the major histocompatibility complex (MHC) since MHC-incompatible targets were killed as well in most instances as autologous cells (Svedmyr and Jondal, 1975; Turz et al., 1977; Lipinski et al., 1979; Seeley et al., 1981). These observations lead to the conclusion that IM cytotoxic cells are EBV specific or at least EBV selective and it was hypothesized that these cytotoxic cells recognize an EBV-encoded

116

GIOVANNA TOSATO A N D R . MICHAEL BLAESE

determinant on the target cells (Svedmyr and Jondal, 1975). More recently, however, it was reported that IM lymphocytes kill EBV-noninfected targets as efficiently as EBV-infected targets (Klein et d., 1980, 1981; Patel et d., 1982). It was also noted that the pattern of cytotoxicity of IM cells was not unique, but was also seen when normal mononuclear cells had been preactivated in vitru by stimulation with antigens and/or mitogens (Klein et al., 1980). These studies concluded that acute IM cytotoxic cells are similar to those that can be activated in vitro by stimulation of normal mononuclear cells and that they can kill a variety of targets irregardless of their expression of MHC and/or of EBV-related antigens. These contrasting results are difficult to reconcile with previous observations and it appears that further work will be necessary to define the determinants recognized by acute IM T cells. Nonetheless, the evidence presented above clearly indicates that during acute IM, cytotoxic cells are often activated in vivo that, in the absence of further stimulation in vitro, are capable of killing a variety of targets, including autologous cells infected with EBV. In addition to enhanced cytotoxic activity, it has been demonstrated in different laboratories that during acute IM, supressor T cells are activated that profoundly inhibit normal lymphocyte proliferation and Ig production (Haynes et aZ., 1979; Tosato et al., 1979, 1982a; Johnsen et al., 1979; Reinherz et al., 1980). This suppressor cell activity has usually been demonstrated by comparing the responses of normal cells cultured alone with the responses of the same normal cells cultured with IM T cells. In this system, IM T cells inhibit normal T cell proliferation in response to antigens and mitogens (Reinherz et al., 1980) and, in addition, profoundly inhibit Ig production induced by PWM (Tosato et al., 1979; Haynes et al., 1979). We have studied the regulation of Ig production by IM T cells in detail (Tosato et al., 1979, 1982a, 1983a). When mononuclear cells from patients with acute IM are stimulated in vitro with the T cell-dependent activator pokeweed mitogen (PWM) or with EBV (B95-8) little or no Ig production is generally observed after 7-8 days in culture (Fig. 6). Different reasons could account for this observation, including the presence of a relatively low number of B cells in the unseparated IM mononuclear cell population, a lack of B cell precursors that can be induced to differentiate into Ig-secreting cells in vitro, or the presence of inhibitory cells that would prevent the B cells from differentiating. Since purified B cell populations obtained from acute IM patients generated Ig-secreting cells in vitro normally, we concluded that IM B cells are functionally normal (Table V). The T cells from these patients, however, were markedly inhibitory to normal B cell differentiation in vitro. Thus, while normal mononuclear cells cultured in the presence of PWM generated large numbers of Ig-secreting cells after 6-7 days in culture, cocultures of the same normal mononuclear cells with T cells from patients

EPSTEIN-BARR VIRUS INFECTION

NORMAL MONO MNC MNC

117

NORMAL MONO T T

+

NORMAL A L L 0 MNC

FIG.6. Inhibited generation of immunoglobulin secreting cells in cultures of normal mononuclear cells (MNC) with T lymphocytes froin patients with acute EBV-induced infectious mononucleosis. Normal and patient MNC (0.5 X 106) were cultured with pokeweed mitogen; in addition, norinal PWM-activated MNC (0.5 x l0cj) were cocultured with 2.0 x 106 normal or patient T cells. The immunoglohulin-secreting cell response was determined at the end of a 7day culture period.

with acute IM resulted in very low Ig-secreting cell responses, indicating the occurrence of suppression by the IM T cells (Fig. 6). This T cell effect was radiation sensitive. In the presence of EBV as a polyclonal B cell activator cocultures of normal allogeneic, HLA incoinpatible B and T cells results in generally low Ig-secreting cell responses when compared to autologous B and T cultures (Tosato and Blaese, 1985). Thus, IM T cells could not be tested for their ability to regulate normal allogeneic B cells activated in oitro with EBV. Nevertheless, since purified B cells obtained fi-om patients with acute IM respond with the normal generation of Ig-secreting cells to EBV stimulation in uitro, it was possible to address the more interesting question as to whether IM T cells can inhibit the activation of autologous B cells infected by EBV (Tosato et ul., 1982a). We observed that IM T cells are generally very effective in inhibiting this autologous response to EBV. Thus, while 1M B cells activated with exogenous EBV (B95-8) in uitro generate large numbers of Ig-secreting cells, autologous mixtures of IM B cells with autologous T cells usually result in an inhibited Ig secreting cell response (Tosato et ul., 1982a). These results indicate that suppression is an important

118

GIOVANNA TOSATO A N D H. MICHAEL BLAESE

TABLE V P U I ~ I F I BE U CELLSFHOM PATIENTS WITII ACUTE EBV-INDUCED 1M ARE FUNCTIONALLY NORMAL NormalU

Exp.

1 2 3 4 5 6

Patient

Medium EBV Medium EBV (1mmunoglol)ulin secreting cellslculture)

1800 200 50 20 70 1600

3100 1800 2500 800 1000 2700

250 20

I00 50 150 20

1700 3800 1500 2500 1100

3000

Purified B cells (0.5X 106) froin 6 patients with acute EBV-induced IM and 6 normal adult individuals were cultured for 8 days with or without EBV (B95-8). The immunoglobulin-secretingcell response was determined at the end of the culture period.

mechanism of control for B cell activation by EBV. In addition, since Igsecreting cells induced by EBV infection include most, if not all of the precursor B cells destined to become immortalized, assay systems of the regulation of Ig production by EBV may represent a suitable, although indirect, way of assessing the control of B cell immortalization by EBV. In certain instances, however, we have observed a clear difference in the degree of suppression manifested by IM T cells in cultures stimulated with either PWM or EBV (unpublished data). In particular, we have noticed that T cells from certain patients with acute IM are very inhibitory for PWMinduced responses by autologous and allogeneic cells while expressing little or no inhibition of EBV-induced responses by autologous cells. Thus, while IM mononuclear cells would fail to respond to PWM stimulation with the production of Ig-secreting cells, they would respond, at least to a certain extent, to stimulation with exogenous EBV in uitro. A similar observation was made by others (Britton, personal communication). The reason for this occasional lack of suppression of autologous EBV-activated cells became apparent during serial studies of patients during the course of their illness. In certain cases of IM, suppressor T cells for EBV-induced B cell activation appear in the circulation later than usual after diagnosis. We hypothesized that since pokeweed, unlike EBV, is also a T cell mitogen, it might provide an additional signal to the IM T cells, inducing them to differentiate into mature suppressor cells. This appears to be the case since the addition of PWM to these EBV-stimulated cultures of IM B and T cells that exhibited

EPSTEIN-BARR

VIRUS INFECTION

119

little or no T cell inhibition revealed a marked suppression in all the cases that we have tested (unpublished observations), while addition of both PWM and EBV to normal cells does not induce suppression. A characteristic, although not necessarily unique feature, of IM suppressor T cells is their apparent ability to suppress only early stages of B cell activation in culture (Tosato et al., 1979). If IM T cells were added to ongoing cultures stimulated with PWM only 24 hours earlier, no significant inhibition of the Ig-secreting cell response was observed at a later point in time. Another feature of IM suppressor T cells is their apparent lack of EBV specificity and restriction by the major histocompatibility complex. Thus, T cell suppression involves both B and T cell responses induced by a variety of antigens and mitogens and includes as targets for suppression HLA incompatible cells. In an attempt to clarify the mechanism of T cell suppression in IM we have looked for a soluble mediator that would mimic the T cell effect. It has been reported that serum obtained from patients with acute IM has the ability to inhibit lymphocyte proliferation in response to a number of antigens, including tetanus toxoid, influenza, and EBV, and to the mitogen PHA (Lai et al., 1977; Wainwright et al., 1979). However, we have consistently failed to demonstrate the production of an inhibitory factor by IM T cells. We have tested the supernatants of unstimulated T cells or T cells stimulated with mitogens, alloantigens, autologous B cells, and EBV-infected B cells; of irradiated or sonicated T cells; and of T cells cultured for different numbers of hours and at different cell densities. This negative result does not rule out the possibility that suppression in this system is mediated by a soluble factor, but certainly suggests that a direct cell-cell interaction between the suppressor T cells and their targets might be necessary. It could be argued that the suppression observed is the result of IM T cells killing the proliferating and differentiating cells in coculture, rather than a distinct suppressor function of IM T cells per se. Both supression and cytotoxicity by IM T cells lack EBV specificity, since EBV noninfected cells were both suppressed and killed at least according to some reports (Tosato et al., 1979; Pate1 1982), and, in addition, both lack MHC restriction, since random normal cells were suppressed and MHC-incompatible targets were killed (Tosato et al., 1979; Lipinski et al., 1979). Clear experimental evidence against this hypothesis is the apparent specificity of suppressor T cells induced during acute IM for the early stages of B cell activation (within 24 hours of infection), a preference certainly not shared by the cytotoxic cells that kill EBV-immortalized and even nonlymphoid targets. No information is yet available on the nature of the stimuli that activate IM T cells to kill or suppress, nor on the precise nature of the cell membrane specificity recognized on the targets that are killed or suppressed. An under-

120

GIOVANNA TOSATO A N D R . MICHAEL RLAESE

standing of these aspects associated with examination at the clonal level of the distribution of cytotoxic and suppressor functions should allow definitive clarification of this issue. The available information suggests that suppressor and cytotoxic functions represent diflerent mechanisms of host defense in patients acutely infected with EBV. Suppressor cells seem to act upon B cells only recently infected with EBV and serve to limit the disease by inhibiting their activation and proliferation. By contrast, IM cytotoxic cells would eliminate the EBV-infected cells that are not susceptible to inhibition by the suppressor mechanism (Fig. 7). This hypothetical scheme of cellular immune defense during acute IM would explain the changes observed in the number of EBV-infected cells detected in the peripheral blood during the acute disease and after recovery (Fig. 8). During acute IM a relatively high proportion of the circulating B cells are infected with EBV, as demonstrated by either enumeration of cells expressing EBNA or by an exam of the frequency of spontaneous outgrowth in vitro. Robinson et al. have reported that between 1 and 20% of the circulating non-T cells express EBNA, while we have found by limiting dilution analysis that, on the average, 130 in lo6 circulating B cells spontaneously give rise to long-term cell lines in vitro. After recovery from acute IM, however, only a very small proportion of the circulating B cells is infected with EBV. Although EBNA-positive cells are not usually recognized by staining of peripheral blood cells, on the average, 3 in lo6 B cells spontaneously give rise to long-term cell lines in vitro, demonstrating that at least 3 in lo6 cells are infected with the virus in normal

Natural Killer Cell

I .

EBV Resting B Cell EBV Infected

Suppressor Effector T Cell

'

B Cell Blast EBV Infected

lg Secreting Cell EBV Infected

Suppressor T Cell Precursor

FIG. 7. Immunoregulation of natural infection with EBV during acute infectious mononucleosis. A schematic representation.

EPSTEIN-HAHH VIRUS INFECTION

121

400

- Geo

Mean

a a

200

a

Y

J

U

u

100

m c3

a a

E

2

u

40

n 0

.

0

0

20

v,

n

0

cn

a

3

:a: a L

6 4

a a

3 2

<1

-e a

-

NORMAL

INFECTIOUS MONO

FIG.8 . Precursor frequencies of spontaneous B cell outgrowth in normal EBV-seropositive individuals and in patients with acute EBV-induced infectious mononucleosis. Multiple replicate cultures containing different numbers of B cells were established and scored at 6 weeks for the presence or absence of transformation. The frequency of spontaneously transforming B cells was determined by Poisson’s statistics and is expressed as number of precursors per 106 cultured B cells (surface immunoglobulin-positive cells).

EBV sero-positive individuals (Fig. 8). We hypothesize that many of the EBV-infected cells persisting in the circulation after recovery from acute IM are the progeny of those cells originally infected with the virus during prirnary infection that have remained under the control of suppressor T cells, and because of this T cell regulation have failed to mature into suitable targets for cytotoxic cells, Most cells infected with the virus during the initial stages of acute IM, probably before suppressor cells have become activated, are eliminated through the cytotoxic mechanism. Moss et al. (1981b) have proposed a different model. They suggested that all EBV-infected cells are invariably killed by specific cytotoxic T cells in

122

GIOVANNA TOSATO AND H. MICHAEL BLAESE

normal individuals. The presence of EBV-infected cells in the circulation of normal EBV sero-positive individuals would not be due to the presistence of cells originally infected with EBV during the acute disease but to a continuous low-grade production of virus in the oropharnyx and subsequent infection of B cells there. These B cells, newly infected with EBV, would be themselves short lived and would be continuously eliminated as a result of specific cytotoxicity. These proposed mechanisms are not necessarily mutually exclusive and it is probable that both suppression and specific cytotoxicity are important. Only when we have a better understanding of the fate of the cells that become naturally infected with EBV during primary infection, of the role of persisting productive infection in the oropharynx in maintaining a low-grade infection of B cells with EBV, and a better understanding in general of the complex mechanisms of EBV infection, will the true contribution of these defense mechanisms be clarified. VI. Persistent EBV Infection in Normal Individuals

Following primary infection, normal individuals continue to harbor EBV in their body for years and probably for life. In these normal individuals EBV can be recognized in the peripheral blood where it infects a small proportion of the circulating B cells and in addition it can periodically be isolated from the saliva. On the average, approximately 3 in lo6 circulating B cells are capable of spontaneously transforming if all regulatory T cells are removed in EBV sero-positive normals (Fig. 8) (Tosato et d . , 1984a). Although rare, these cells, similar to those recognized during accute IM, are capable of unlimited growth in uitro (Gerber et al., 1969; Chang et al., 1971; Nilsson, 1971). Thus, similar to normal cells exogenously infected with transforming EBV in uitro, naturally infected cells can multiply rapidly in uitro and give rise to immortal cell lines. In addition, similar to the characteristics of the virus isolated during acute IM, EBV released in the saliva has the ability to transform normal newborn B cells into long-term cell lines (Miller et al., 1973). An important question is how persistent infection with EBV is kept under control in normal EBV sero-positive individuals. Progressive expansion of EBV-infected cells rarely occurs in uiuo. Thorley Lawson et d . (1977) clearly demonstrated how non-B cells can inhibit the growth of EBV-infected cells in uitro. In these experiments, B cell-enriched populations obtained from normal adults were exogenously infected with EBV in uitro and cultured with or without the addition of autologous non-B cells. Outgrowth of EBV-transformed cells was scored by microscopic examinaton and the results clearly showed that addition of increasing concentrations of autologous adult Ig-negative lymphocytes resulted in a delay of the time of

EPSTEIN-BARR

VIRUS INFECTION

123

outgrowth of EBV-transformed cells. Other laboratories have analyzed this phenomenon in detail and reported that T lymphocytes from EBV-immune individuals profoundly inhibit the growth of B cell exogenously infected with EBV in vitro. Rickinson et a2. have shown that when mononuclear cells from EBV sero-negative adult individuals and from newborn bloods were exposed to EBV in vitro and cultured at various cell densities, foci of EBV-infected (EBNA-positive cells) would appear within 8 to 10 days postinfection and these foci would progressively expand to give rise to a continuous cell line. Similar cultures established with cells from EBV sero-positive individuals demonstrated a completely different pattern of growth. This pattern was characterized by an initial proliferation of EBV-infected cells during the first 8 to 10 days in culture that was followed by the regression of growth, usually apparent by day 14, so that no growth of EBV immortal cell lines was observed by day 28 (Moss et al., 1978). Further experiments revealed that EBV-immune T cells were responsible for this inhibitory effect and that effector cells for regression derived mostly from Ty-depleted T cells (Rickinson and Moss, 1983). In addition, we and others have observed that T cells not only inhibit B cells infected in vitro with exogenous EBV, but they also suppress the growth of B cells naturally infected with EBV in vivo. Thus, spontaneous outgrowth in vitro of cells infected with EBV in vivo is much more easily observed if the T cells have been removed prior to culture or functionally inactivated by irradiation or by the addition of cyclosporin A to the cultures (Bird et al., 1981; Tosato et al., 1984a). Although there is general agreement on the observation that T cells do inhibit the growth of B cells infected with EBV either by virtue of a natural infection in vivo or by addition of exogenous virus in uitro, there is still uncertainty on the mechanisms mediating the T cell inhibition. Moss and Rickinson have described the presence of EBV-specific cytotoxic T cells in regressing cultures that specifically kill autologous EBV-transformed cells (Moss et al., 1979; Rickinson and Moss, 1983). These cells are only present in lymphocyte cultures of EBV sero-positive individuals and are specific for EBV-infected cells, since they do not kill autologous mitogen-activated blasts (Moss et al., 1979, 1981a; Rickinson et al., 1979, 1980a). In addition, they are restricted by MHC since only HLA-A and -B compatible or partially compatible targets are killed (Misko et al., 1980; Rickinson et al., 1980b; Wallace et al., 1981; Moss et al., 1981a). The existence of EBV-specific cytotoxic T cells was later confirmed by growth of cell clones in T cell growth factor-containing media that demonstrated EBV-specific and HLA-restricted killing (Wallace et al., 1982a,b). It was proposed that specific cytotoxic precursors are present in the circulation of EBV sero-positive individuals that, upon in vitro restimulation with autologous or HLA-compatible EBV-infected cells, proliferate and kill the stimulating cells. Rickinson also sug-

124

GIOVANNA TOSATO AND H. MICHAEL BLAESE

gested that the EBV-specific cytotoxic T cell response is directed to an EBVassociated lymphocyte-detected membrane antigen (LYDMA), whose existence was first proposed to explain the EBV-selective T cell response during acute I M (Moss et al., 198lb). However, in spite of extensive research, the molecular nature of LYDMA has not been clarified and no antisera or monoclonal antibodies have been raised to EBV-related antigens capable of blocking EBV-specific cytotoxicity (Wallace et al., 1981; Rowe et al., 1982; Slovin et al., 1982). The in uivo function of EBV-specific cytotoxic cells would be to continuously eliminate EBV-infected cells expressing LYDMA. Since it was proposed that this antigen appears very soon, perhaps within 3 days following infection with EBV, the circulating cells infected with EBV would probably arise from a recent infection with virus produced in the oropharynx (Moss et al., 198lb). In addition to EBV-specific cytotoxic cells, a nonspecific mechanism or mechanisms appear to contribute to the control of EBV-infected cells in vitro. This was suggested by the observation that T cells from EBV seronegative as well as from EBV sero-positive individuals demonstrate an ability to delay, although not to prevent, the growth of autologous B cells infected with EBV in vitro (Shope and Kaplan, 1979; Bardwick et al., 1980; Masucci et al., 1983). This function has been attributed to low-density T lymphocytes that bear an Fcy receptor and have the characteristics of large granular lymphocytes. These cells present in EBV sero-positive and seronegative adult individuals are lacking in newborn bloods. In addition, unlike EBV-specific cytotoxic T cells, they have been reported to be resistant to the effects of irradiation and to treatment with mitomycin C (Depper et al., 1981; Masucci et al., 1983), thus suggesting that this function does not require cellular proliferation. Two distinct mechanisms have been proposed to mediate this nonspecific inhibitory function. Since large granular lymphocytes are cells capable of natural killer activity, inhibition might be due to cytotoxicity directed against EBV-activated cells (Masucci et al., 1983). In this respect, clones of T lymphocytes have been grown in vitro by stimulation with autologous EBV-infected cells and T cell growth factor that have demonstrated a nonspecific cytotoxicity directed against EBV-infected as well as noninfected targets, irrespective of MHC compatibility (Tanaka et al., 1982). Alternatively, the T cells might have a nonspecific antiproliferative effect on the EBV-infected B cells. Thorley-Lawson has reported that a short exposure of EBV-infected B cells to autologous T cells followed by a removal of the T cells resulted in inhibited proliferation of the B cells when measured several days later (Thorley-Lawson, 1980). In addition, since crude preparations of interferon alpha have been reported to mediate similar inhibitory effects, he proposed that the T cells might mediate the antiproliferative effect through the release of interferon (Thorley-

EPSTEIN-BARR

VIRUS INFECTION

125

Lawson, 1981). However, since nonspecific mechanisms of control appear to delay but not to prevent the growth of EBV-infected cells, their role in the in uivo control of residual EBV infection may be limited (Masucci et al., 1983; Rickinson and Moss, 1983). We have described a different assay system to measure EBV-related regulatory functions. We have called it “late suppression” (Tosato et al., 1982a). In this assay, activation by EBV is monitored by measuring the number of Ig-secreting cells produced by the virally stimulated B lymphocytes. This system presents a number of advantages over the others previously described, outgrowth and proliferation. Evaluation of outgrowth induced by EBV is based on morphological criteria which are often very subjective. Most studies have analyzed control mechanisms for outgrowth induced by EBV and have expressed the results as a delay in the appearance of transformation and thus were often based on the difficult determination that the culture was almost, but not as yet, transformed. Use of proliferation to measure B cell activation by EBV is a reliable method if purified populations of B cells are used. If, however, recombined populations of B and T cells are employed, it is dificult to evaluate which cell or cells are proliferating; the B cells because they are infected with EBV; or the T cells because of the occurrence of the autologous mixed lymphocyte reaction; or a specific proliferation of the T cells to the virally infected B cells. In contrast, use of Ig production as a readout for EBV-induced B cell activation is a reliable and objective method that allows us to carefully monitor B cell transformation by EBV, since, by and large, all B cell precursors that will eventually give rise to a long-term cell line produce Ig (Nilsson and Klein, 1982). Other processes that might occur in the culture system, such as T cell proliferation, lymphokine production, etc., are ignored by the readout system unless they directly effect B cells activated by EBV. Human B cells, whether from EBV sero-negative or sero-positive individuals, generate Ig-secreting cells after in uitro infection with EBV and over a 14-day culture period this number increases exponentially (Fig. 9). If T cells from an EBV sero-positive individual are mixed in culture with the autologous B cells infected with EBV at various ratios, including the physiological ratio, the number of Ig-secreting cells is initially similar to that observed in cultures of B cells alone. However, after 8 to 10 days in culture, a characteristic and profound inhibition of the B cell response becomes apparent. As a result, after 14 days in culture, EBV-infected B cells alone produce much higher numbers of Ig-secreting cells compared to the same EBVinfected B cells cultured with autologous EBV immune T cells indicating potent T cell inhibition. By contrast, in EBV sero-negative adult donors, B cell responses to EBV are not suppressed by the autologous T cells during a 14-day culture period, and thus no significant difference is observed in the

126

CIOVANNA TOSATO AND R. MICHAEL BLAESE EBV-SEROPOSITIVE

<10 ,00-

4

6

8

10

12

EBV-SERONEGATIVE

-

14 4 DAYS

6

8

10

12

14

FIG. 9. Purified T cells from a typical EBV-seropositive individual manifest a delayed suppressor activity “late suppression” on EBV-induced autologous B cell differentiation. Purified T cells from a typical EBV-seronegative individual have no effect on autologous B cell activation by EBV. Mononuclear cells (MNC, 1.0 X lo6), B cells (0.5 x lo6), and autologous mixtures of B (0.5 X lo6) and T (2.0 X lo6) cells were cultured for 14 days in the presence of EBV (B95-8) and the immunoglobulin-secreting cell response was determined after 4, 6, 8, 10, 12, and 14 days of culture.

number of Ig-secreting cells produced by the B cells alone or the B and T cocultures. Inhibition of the B cell response is not observed if the EBVimmune T cells are irradiated with 2000 rad prior to culture, or if steroids or cyclosporin A are added to the culture system, indicating that the T cells require in uitro activation and proliferation (Tosato et al., 1982a,b). Since only EBV sero-positive individuals show suppressor activity in this assay, activation of the inhibitory cells is antigen-specific, probably requiring the presence of memory precursors. Furthermore, since activated suppressor T cells from EBV sero-positive donors are not more inhibitory for mitogeninduced Ig production than similarly cultured T cells from EBV sero-negative individuals, suppression appears to be preferentially directed toward EBV-infected cells. In recent experiments, we have observed that depletion of natural killer cells by monoclonal antibody and complement lysis does not significantly affect EBV-specific suppression mediated by EBV-immune T cells after 14 days in culture. Furthermore, suppression appears to be mediated mostly,

EPSTEIN-BARR

127

VIHUS INFECTION

but not exclusively, by T cells expressing the suppressor-cytotoxic phenotype (Table VI). How does late suppression relate to the other previously described systems for T cell control of EBV infection? The characteristic specificity (it is only observed in EBV sero-positive individuals) and the time course of suppression (no initial T cell effect for 8-10 days, followed by profound inhibition by day 14) are features shared with the process described as “regression” (Rickinson and Moss, 1983). It is likely that late suppression and regression describe very similar processes of specific T cell control for autologous EBV-infected cells. It was shown that specific cytotoxic T cells are activated during regression (Moss et al., 1981a; Misko et al., 1980) and felt that killing of the EBV-infected B cells is the mechanism mediating the phenomenon. We have also observed the activation of cytotoxic cells in late suppressed cultures (unpublished observations), but we believe that a noncytotoxic suppressor cell population is also contributing to the phenomenon. The evidence is twofold. First, if the T cells are removed from the suppressed cultures, the residual B cells infected with EBV are capable of rapidly differentiating into Ig-secreting cells (Tosato et al., 1982a) (Table VII). This reversibility suggests that T cell regulation does not occur exclusively through a cytotoxic mechanism and that some of the EBV-infected cells have not been killed. Rather a proportion of the infected cells have been kept in check or suppressed by a iioncytotoxic T cell-mediated mechanism. Second, we have recently been able to grow a continuous T cell line TABLE VI LATEACTINGS U P P R E S S O R T CELLS1)ISPI.AY A S U P r R E S S ( ~ l ~ - C Y T ~ l T O XPIIENOTYPE tc

Immunoglobulin secreting cell response Cell culturesf’

B cells B + auto T cells B + auto T cells (Leu 11b neg.) B + auto T cells (T4 neg.) B + auto T cells (T8 neg.)

Exp. 1

Exp. 2

Exp. 3

22550 4210 3850 1927 16362

38600 15120 15060 -

34200 8160

-

-

3580 22650

Purified B cells (2.5X lo5) obtained from 3 EBV-sero-positive normal individuals were cultured for 14 days in the presence of EBV (B95-8)either alone or mixed with autologous T cells (2.5X1OS)).The T cells were either unfractionated or depleted of Leu 1111, T4, or T8 bearing cells by treatment with the appropriate monoclonal antibody and complement. T4, T8, and Leu I l b monoclonal antibodies recognize, respectively, “helper,” “suppressor-cytotoxic,”and most of the “natural killer” T cells.

128

CIOVANNA TOSATO AND R. MICHAEL RLAESE

TABLE VII SUPPRESSIONMEDIATEDBY EBV IMMUNET CELLSIs REVERSIBLE Immunoglobulin secreting cellsiculture Primary culture

Secondary culture

Culture groupa

Day 14

Day 3

Day 5

Day 7

1. O.5X1O6 B cells 2. O.5x1O6 B + 2.0X106 T cells 3. O.5X1O6 B cells recovered from group 2

16573 283

10491 125

18858 157

8956 40

389

2147

4001

a Purified B cells (0.5X 106) from 3 EBV-sero-positive normal individuals were cultured in the presence of EBV either alone or in the presence of autologous T cells (2.0x106)). At the end of a 14-day culture period (primary culture), T cells were removed from the autologous EBV-activated B and T cell cocultures and the purified B cells were incubated alone for additional 7 days without additional EBV stimulation (secondary culture). The immunoglobulin secreting cell response was determined at the end of the primary culture, and at 3, 5, and 7 days during the secondary culture.

from a sero-positive normal individual that has no detectable cytotoxic activity and, at the same time, can profoundly inhibit the proliferation, outgrowth, and differentiation of autologous B cells induced by EBV (unpublished observation). Thus, we believe that specific T cell control of EBVinfected cells occurs by suppression as well as cytotoxicity. Further clarification of this issue will benefit from the possibility of physically separating the cytotoxic from the suppressor precursors and from a broader understanding of the molecular basis for suppressor-target interactions. The concept that an effective control mechanism for EBV-infected B cells might occur without eliminating the infected cells has been previously proposed. It was observed that under the appropriate conditions adult mononuclear cell cultures infected with EBV do not give rise to long-term cell lines if they are placed on a feeder layer of adult human fibroblasts, but they do transform if no feeder layer is added. Although a proportion of the EBVinfected B cells would express EBNA, these cells would not expand for long periods of time. If the fibroblasts were removed after 1 month of culture, successful outgrowth of EBV-infected cells was then observed (Moss et al., 1976, 1977). This pattern of growth suggests a noncytotoxic control of the EBV-infected cells by the fibroblasts (Rickinson and Moss, 1983). Thus, EBV-infected B cells, destined to continous and progressive growth, can be

EPSTEIN-HAHH

VIRUS INFECTION

129

prevented from proliferating in culture by exogenous regulatory mechanisms, without being destroyed and without losing the property of expanding at a later point in time, once the control mechanisms have been removed. It is interesting to note that the late suppression assay does not appear to monitor those nonspecific mechanisms of control that have been reported to delay the outgrowth of EBV-infected cells. We have consistently failed to observe an effect by nonimmune T cells in this system (Tosato et a l . , 1982a). Similarly, we have found that EBV-immune T cells do not inhibit autologous B cell activation during the first 13-10 days in culture (Tosato et al., 1982a). In addition, depletion of natural killer cells did not significantly effect the magnitude of specific T cell suppression at 14 days. We propose that the control of residual EBV infection in EBV seropositive normals occurs through both suppression and killing. By this hypothesis EBV-infected B cells persist following primary infection and these cells are controlled by a mechanism of specific suppression that prevents their proliferation. These specific inhibitory cells represent a functionally different cell from the nonspecific suppressors detected during the acute phase of IM, but would serve a similar function. In this formulation, cytotoxic T cells represent a backup mechanism of specific defense against EBV that would come into play when suppression had failed to prevent EBV-induced B cell outgrowth. VII. Selected Disorders Associated with an Abnormal Regulation of EBV Infection

The studies performed in vitro with the goal of clarifying the mechanism of control of EBV infection have uniformly indicated the importance of T cells. Directly or indirectly, T cells appear to be always involved in the different pathways that have been proposed for the control of EBV infection. If these conclusions, based on studies performed in vitro, have any relevance to the situation in v i m , then one would predict that individuals with defects in certain T cell functions should reflect these abnormalities with an EBV-related disorder. For example, one would expect to observe the occurrence of a progressive proliferation of EBV-infected B cells in individuals completely lacking functional T cells. Perhaps the best demonstration of the correctness of this hypothesis has been recently provided by patients treated with high doses of an anti-T-cell monoclonal antibody that recognizes an epitope usually referred to as Pl9 (Lanier et a l . , 1983; Hansen et d . , 1984). These were patients with terminal acute myelogenous leukemia who had undergone an ablative chemotherapy regimen followed by bone marrow transplantation. The monoclonal antibody administration was initiated as a form of treatment for severe graft-versus-host disease related to the presence

130

GIOVANNA TOSATO A N D H. MICHAEL RLAESE

of mature T cells in the transplanted bone marrow. This treatment was effective in reducing symptoms related to the graft-versus-host disease, presumably because T cells binding the monoclonal antibody were either destroyed or were functionally inactivated. However, in two reported cases, a polyclonal, progressive, and rapidly fatal lymphoproliferation of EBV-infected donor cells occurred (Martin et al., 1984). In one of these cases, it was probable that the fatal lymphoproliferation followed a recent infection with EBV, since neither the recipient nor the donor of the bone marrow cells had any serologic evidence of prior EBV infection. In the other case, the lymphoproliferative disorder probably arose from latently infected B cells. In these cases, the progressively proliferating cells had a B cell phenotype and were polyclonal; in addition, they lacked chromosomal abnormalities and were infected with EBV. All these characteristics suggest that the malignant lymphoproliferation derived from normal B cells naturally infected with EBV which had progressively expanded in the body since the cells capable of controlling them, the T cells, were either missing or were functionally inactive as a consequence of the treatment. We have looked at the effects of a panel of monoclonal antibodies on EBVspecific late suppression. As mentioned earlier, this assay system permits the monitoring of EBV-specific memory T cells obtained from EBV seropositive individuals in the control of activation of autologous B cells exogenously infected with EBV in uitro. Interestingly, the addition of an anti-T monoclonal antibody, anti-T3 (Kung et al., 1979), that also recognizes the P19 determinant expressed by most T cells, markedly inhibited this EBVrelated control function. A different monoclonal antibody, anti-Tac, that recognizes a surface receptor for T cell growth factor generally expressed by activated T cells (Uchiyama et al., 1981; Leonard et al., 1982) was also inhibitory in this system. By contrast, no effect had anti-T4 (Reinherz et al., 1979) and Leu 11 (Lanier et al., 1983)monoclonal antibodies, that recognize helper T cells and predominantly natural killer cells, respectively, and only a modest effect had anti-T8 monoclonal antibody (unpublished observations). Thus, there is a marked similarity between the effect of anti-T3 monoclonal antibody in uitro, which blocks the ability of T cells to suppress EBVinfected cells, and the effects of a similar monoclonal antibody in uiuo, where its administration was associated with the development of an uncontrolled proliferation of EBV-infected cells. Cyclosporin A has strikingly similar effects on “late-suppression” to those manifested by anti-T3 monoclonal antibody (Tosato et al., 1982b). Cyclosporin A is a recently introduced fungal product that has profound immunosuppressive effects and demonstrates a characteristic selective activity for T cell functions (Britton and Palacios, 1982). It has been shown that many T cell functions are inhibited by cyclosporin A including help, pro-

EPSTEIN-BARH VIRUS INFECTION

131

liferation in response to antigens and mitogens, cytotoxicity, and suppression (White et al., 1979; Hess and Tutscha, 1980; Tosato et al., 198213). By contrast, B cell functions are usually spared (Bore1 et al., 1977; Bird et al., 1981; Tosato et al., 1982b). The molecular basis for the selective effects of cyclosporin A is not completely understood, but there is evidence that it effects interleukin 2 production by T cells (Bunjes et al., 1981; Dos Reis and Shevach, 1982; Miyawaki et al., 1983). We have shown that cyclosporin A at a concentration of 0.5-1 p,g/ml added to EBV-stimulated cultures of EBV sero-positive B and T cells prevents the expected T cell-mediated late suppression (Fig. 10). In addition, if cyclosporin A is added to mononuclear cell cultures from EBV sero-positive individuals, spontaneous outgrowth of EBV-transformed B cells is readily observed, indicating that this drug also blocks T cell control of endogenous EBV infection (Tosato et al., 1984a). Thus, from an operational point of view, the addition of this concentration of cyclosporin A to cultures is equivalent to the physical elimination of EBVspecific regulatory T cells. From these in vitro observations, one would expect the occurrence of EBV-related B cell disorders in patients treated with cyclosporin A at doses comparable to those used in vitro. It has been

W

a

3 I-

$

20,000

@

u, I? -I W

H

0

z CJ I-

8.0001

0

W

a

M

90, 4,000 -

:

I

m B Cells

B+T Cells

B

+ T Cells

+

CYA

FIG. 10. Cyclosporin A inhibits late acting T cell suppression. €3 cells (0.5 X 106) obtained from a group of EBV-seropositive normal individuals were cultured in the presence of EBV (B95-8) either alone or mixed with autologous T cells (2.0 x 10")in medium or with cyclosporin A (1 Fg/ml). At the end of a 14-day culture period the iintnunoglobulin-secreting cell response was determined.

132

GIOVANNA TOSATO A N D R . MICHAEL BLAESE

reported that patients treated with cyclosporin A display progressively rising antibody titers to EBV-related antigens (Navington and Gray, 1980). In addition, in a series of 34 kidney transplant patients given cyclosporin A for prevention of graft rejection three developed a fatal lymphoproliferative process (Calne et al., 1979; Thiru et al., 1981). A detailed study in one of these patients revealed that the proliferating lymphocytes were B cells infected with EBV and expressing EBNA (Crawford et al., 1980). Thus, treatment with agents that profoundly and nonselectively impair T cell function is associated with a disturbing incidence of EBV-related malignant lymphoproliferations. In addition to these cases, other EBV-related lymphoproliferative syndromes have been reported, but their frequency, their precise relationship to EBV, and their pathogenesis are not clear. Some have occurred in kindreds with an X-linked pattern of inheritance and are believed to follow primary EBV infection (Purtilo et al., 1975, 1982); others have affected males and females without a clear genetic predisposition (Fleisher et al., 1982), and others have been reported to follow transplants of thymic epithelium, semi-allogeneic bone marrow, and kidney (Borzy et al., 1979; Hanto et al., 1982). In some of these individuals, humoral and cellular immune defects relating to EBV have been described (Sullivan et al., 1983). Others were reported to have a defect in immune interferon production, or multiple congenital abnormalities including an abnormally located thymus, and, finally, others appeared immunologically intact (Britton et al., 1978; Robinson et al., 1980). The majority of these lymphoproliferative disorders had similar pathological findings and individual well studied cases were associated with a polyclonal proliferation of EBV-infected B cells that characteristically lacked chromosomal aberrations (Robinson et al., 1980). A shift from polyclonality to a dominant clonal type has also been reported (Abo et al., 1982). The many similarities among all these lymphoproliferative processes together with the demonstration in most of some type of immune deficiency suggests a common pathogenesis related to a defect in EBVspecific T cell regulation. In addition to these malignant lymphoproliferative processes, patients with classical rheumatoid arthritis (RA) have been reported to have a variety of immune abnormalities relating to EBV. Most patients with RA demonstrate elevated antibody titers to many EBV-related antigens. In addition, RA patients often have antibodies to RANA which are uncommon in EBV sero-positive normals but are detected in EBV-associated malignancies such as Burkitt’s lymphoma and nasopharyngeal carcinoma (Aspaugh et al., 1978, 1981; Catalan0 et al., 1979; Henle et al., 1979; Ferrell et al., 1981). When cultured in uitro in the presence of exogenous-transforming EBV, RA mononuclear cells give rise to long-term B cell lines more easily than normal cells (Slaughter et al., 1978; Bardwick et al., 1980). This has been shown to result

133

E PSTEIN- UAHH VIRUS INFECTION

from a diminished T cell control of EBV-infected B cells. Thus, RA T cells were less effective than normal T cells in inhibiting EBV-induced B cell proliferation and B cell transformation (Bardwick et al., 1980; Depper et al., 1981). More recent work has attributed this defect to diminished production of interferon by RA T cells (Hasler et al., 1983). Since the T cell function examined in these assays was reported to be resistant to treatment with mitomycin C (Bardwick et al., 1980), it could be that the regulatory function detected involves nonproliferating cells such as natural killer cells and suggests that natural killer activity is abnormally low in patients with RA. We have examined EBV-related late suppressor T cell function in a group of patients with classical RA and found that most of these patients failed to suppress EBV-induced lg production after 12 days in culture (Table VIII) (Tosato et d., 1981b). At this time, normal EBV sero-positive T cells are usually profoundly suppressive. Thus, EBV-immune patients with RA have a profound defect in their ability to regulate B cell activation induced by EBV. We tested for other T cell abnormalities in these patients, including the ability of RA T cells to provide help for Ig production in response to PWM and the ability to suppress Ig production in an allogeneic-induced suppressor system and found that these functions were normal. Thus, RA patients appear to have a marked suppressor T cell defect relating specifical-

TABLE VIII DI.:I'ECTIVE LATEACTING SUPPRESSION IN PATIENTS

WITH

RIIEUMATOIDAHTIIRITIS KA patients

Normalsci

B cells

+ T cells

Suppression

Suppression

(%)

B cells

550 1200 2300 180 2460 1300 400 860

95 81 86 98 88 91 94 90

24000 24600 5860 14680 8420 12560 13320 8190

Mean suppression (%)

90

10600'J 12800 16480 13700 21140 14700 6500 8400

B

€3

+ T cells 6400 20100 4600 7570 1840 9620 15840 8700

("/.I 73 19 21 49 78 32 - 18 -6 16

0 Purified B cells (0.5x 10") from patients with classical rheumatoid arthritis were cultured for 12 days in the presence of EBV either alone or mixed with autologous T cells (2.0x 10"). The immunoglobulin-secretiiig cell response was determined at the end of the culture period. Ininiunoglobulin secreting cells/culture.

'J

134

GIOVANNA TOSATO A N D R . MICHAEL BLAESE

ly to EBV, rather than a global defect in immunoregulatory T cell function (Tosato et al., 198lb). A key question is the relationship between these observations made in vitro and the status of natural EBV infection in t h o . In particular, are these in vitro defects reflected as diminished control of EBV-infected cells in uivo? To address this question we asked whether the number of circulating B cells infected with EBV is higher in patients with RA than in normals. In normal EBV sero-positive individuals less than 3 in lo6 B cells, on the average, give rise spontaneously to long-term cell lines in vitro, while patients with RA have two- to threefold more B cells infected with EBV capable of spontaneous transformation (Fig. 11) (Tosato et al., 1984a). Thus, patients with RA have a significantly higher burden of circulating EBV-infected cells than normals. This finding could not be attributed to an intrinsic defect of RA B

100

.

50

- Geo Mean

I)

20

I 10

i

.. a

5

a

1 .

.

0

t

-

. . 0

2

91

..I

Normals

0.

Rheumatoid Arthritis

FIG. 11. Patients with rheumatoid arthritis (RA)have elevated numbers of EBV-infected €3 cells in the circulation. Precursor frequencies of spontaneous B cell outgrowth were determined in a group of EBV-seropositive normal individuals and of patients with classical rheumatoid arthritis. Multiple replicate cultures containing different numbers of either B cells or mononuclear cells in the presence of 1.0 pg/ml ofcyclosporin A were established and scored at 6 weeks for the presence or absence of spontaneous transformation. The frequency of spontaneously transforming B cells was determined by Poisson's statistics and is expressed as number of precursors per 10" cultured B cells (surface immunoglobulin-positive cells).

EPSTEIN-BARR

VIRUS INFECTION

135

cells, since, at least in vitro, the frequency of B cells transforming in the presence of exogenous EBV was similar in RA patients and normals (Tosato et al., 1984a). As a group, RA patients have both decreased EBV-specific “late suppressor” activity and an increased burden of circulating EBV-infected B cells. However, while essentially all RA patients have this defect in suppressor function, some of the patients have a normal frequency of EBVinfected cells (Tosato et al., 1984a). This suggests that a defect involving one of the regulatory mechanisms can be compensated for in vivo by other control processes. Thus, although these patients do have a significantly increased number of EBV-infected cells in their blood, RA patients have a much lower number than is seen in individuals where the regulatory processes are absent or poorly developed such as early acute IM where the frequency of infected B cells averages 130 per lo6 B cells. It is interesting in this regard that other laboratories have reported that EBV-specific cytotoxic T cell function is normal in patients with RA (Tsoukas et al., 1982). Whether EBV has any role in the pathogenesis of RA, or a subgroup of patients with RA, is unproven at this point and certainly requires further investigation. It is important to remember that not all patients with classical RA have evidence of prior infection with EBV (Elson et al., 1979). It is, however, probably significant that abnormal in vitro regulation of EBV occurs in most patients with RA and that this is associated with an abnormally enlarged pool of EBV-infected cells in the circulation. The existence of elevated numbers of EBV-infected B cells in patients with KA may contribute to the persistent state of enhanced B cell activation observed in this d‘isease. Malignant lymphoproliferation of EBV-infected B cells can result from an absent or a diminished T cell control of B cells infected with EBV. There is also some evidence that exaggerated T cell control of EBV infection can also lead to a disease state. Agammaglobulinemia has been observed following a primary infection with EBV (Provisor et d., 1975; Purtilo, 1980; Greally et al., 1983). In two well documented cases, the hypogammaglobulinemia involved all major classes of Ig, and appeared directly after an episode of clinically and serologically defined EBV-induced IM (Provisor et al., 1975). We have studied a patient with a similar history of acquired hypogammaglobulinemia which followed an illness with clinical and laboratory features of acute EBV-induced IM. In vitro studies revealed that the patient had a normal number of circulating B cells. However, the patient’s mononuclear cells failed to generate an Ig secreting cell response in vitro upon stimulation with PWM; in addition, the patient’s T cells profoundly inhibited Ig production by normal cocultured mononuclear cells (unpublished observation). These findings indicate that the patient had an excess of circulating suppressor T cells and suggested that this excessive suppressor T

136

GIOVANNA TOSATO AND R . MICHAEL HLAESE

cell activity was probably responsible for the impaired Ig production seen in vitro as well as of the patient’s hypogammaglobulinemia. As described earlier, during acute EBV-induced IM, suppressor T cells are generated that nonspecifically suppress Ig production by autologous and normal allogeneic B cells. These suppressor cells usually disappear from the circulation within a few weeks or a few months from diagnosis. It seems very likely that an abnormal persistence of these inhibitory T cells after resolution of primary EBV infection in these patients is associated with their development of late-onset common variable hypogammaglobulinemia. Although this mechanism clearly seems to account for some cases of common variable hypogammaglobulinemia, a definite causal relationship between EBV infection and subsequent humoral immune deficiency is very difficult to establish with present technology in most cases. The mechanism of disappearance of suppressor T cells from the circulation of patients recovering from acute, EBV-induced IM is unknown. In an attempt to clarify this issue, we have studied a group of individuals during acute EBV-induced IM and serially during their covalescence. All demonstrated the characteristic in vitro immune findings during the acute disease and all had a complete immunological recovery 6 weeks to 4 months after diagnosis. At some point during convalescence, we observed in all the patients a unique and characteristic immune response in vitro. While the mononuclear cells or the recombined cultures of the patients own B and T cells responded to PWM with the generation of relatively normal numbers of Ig-secreting cells, the purified T cells from these patients strongly inhibited Ig production by normal allogeneic PWM-activated lymphocytes in coculture (unpublished). A question raised by these experimental results is how a T cell population can inhibit a variety of allogeneic target cells and, at the same time, fail to suppress autologous B cells? A number of possibilities can be raised to explain these data, including the presence in normals but not patients of an appropriate target for suppression or perhaps the presence in the patient of an MHC-restricted contrasuppressor cell population that counteracts the effects of suppression. This unusual pattern of immunoregulatory T cell activity is also seen in another disorder, which appears to be related to EBV infection, chronic infectious mononucleosis. This disorder includes a symptom complex consisting of persistent fatigue and constitutional symptoms with a duration of at least 18 months which frequently, but not necessarily, follows an episode of acute IM. Characteristically these patients’ sera contain antibodies to EBV-related antigens which form a pattern similar to that seen during active EBV infection (Tobi et al., 1982; Straus et al., 1984). This includes the presence of antibodies to EBV early antigens (EA D and/or R), absence of antibodies to EBNA, and generally higher than normal titers of antibodies to the viral capsid antigen of EBV.

137

EPSTEIN-BAHH VIRUS INFECTION

Antibodies to EBV early antigens are usually present during acute EBVinduced IM, but are absent or very low in normal individuals; in contrast, antibodies to EBNA are usually absent during acute IM, but appear 6 inonths to 1 year following primary EBV infection (Henle et d . , 1974; G. Henle and Henle, 1979; W. Henle and Henle, 1979). Serum Ig levels are usually normal in patients with chronic infectious mononucleosis and, in addition, we found that the number of EBV-infected cells in the circulation is generally normal (Tosato et al., 19831-3).Similarly, EBV-related late-acting suppressor functions were also normal in these patients. However, in uitro studies revealed that while mononuclear cells from patients with chronic infectious mononucleosis differentiate normally to become Ig-secreting cells upon PWM stimulation, their T cells suppressed Ig production by normal allogeneic cells in co-culture (Table IX). The observation that patients with chronic infectious mononucleosis present a characteristic pattern of immune reactivity normally found during the recovery phase from acute EBV-induced IM suggests that, from an immunologic viewpoint, patients with this illness are frozen in a state typically found only transiently during the convalescence from acute primary EBV infection. The relationship between the TABLE IX T CELLS O F A PATIENT WIT11 CIIIIONIC: ACTIVE EBV INFECTION SUPPRESS ALLOCENEICBL~TNOT AUTOLOGOUS B CELL RESPONSES: REPRESENTATIVE EXPEHIMENT Iniinunoglobulin secreting cells<( Cultures

Medium

Normal MNCb (0.5X 10") Patient' MNC (0.5X 10") Normal MNC (0.5X10") patient T cells (1.0X10") Normal MNC (0.5X106) patient T cells (0.5x 10") Normal M N C (0.5X 10") patient T c e l l ~ 2 ' ~ ' ~(1.OX ) R 10")

+

279

I ,556 368

PWM 14,334 7,819 3,618

+

319

8,207

+

-

17,144

" Number of immunoglobnlin secreting cells produced per culture at the end of a &day culture period. Mononuclear cells. c Patient with clironic active EBV infection. The diagnosis was based 011 a >2 years history of malaise associated with recurrent fevers and air abnormal pattern of anti1)ody reactivity to EBV (anti-VCA 1280, antiEA 40, anti-EBNAc2). No other illness could be diagnosed after extensive evaluation.

138

GIOVANNA TOSATO AND R . MICHAEL BLAESE

peculiar immune response detected in vitro and the symptomatology presented by patients with chronic IM is unclear at present. It is, however, interesting to hypothesize that an abnormally prolonged physiologic regulatory process may contribute some essential element in the pathogenesis of this syndrome. VIII. Reversal of Infectious Mononucleosis-AssociatedSuppressor T Cell Activity

by

o-Mannose: Suppression and Saccharides

A great deal of experimental evidence has accumulated over the years indicating that cell surface carbohydrates may serve as important recognition and interaction structures (Sharon, 1983). Elegant studies of circulating glycoproteins have clearly demonstrated that their survival in the circulation is dependent upon the type of terminal sugar residue on the glycoprotein molecule. For example, the uptake of galactose-terminated glycoproteins occurs in the liver since hepatocytes possess a lectin-like surface receptor that specifically binds to galactose molecules (Ashwell and Morel], 1974). Similar sugar-specific surface receptors have been recognized on cells of the reticuloendothelial system that can specifically bind and subsequently take up glycoproteins displaying D-mannose, N-acetyl-~-glucosamine,or Lfucose as terminal sugars (Stahl et al., 1978; Shepard et al., 1981). In addition, sugar molecules may provide immunologic specificity to cell surface structures such as blood group determinants (Shen et al., 1968);may be necessary for cell-to-cell contact required for fertilization, differentiation, cell aggregation, and infection with viruses and bacteria (Gelb and Lerner, 1965; Shen et al., 1968;Vacquier and Moy, 1977; Ofek et al., 1977; Grabel et al., 1979; Muramatsu et al., 1979); and may be somehow involved in the interaction between killer cells and their targets (MacDonald and Cerottini, 1979; Stutman et al., 1980; Forbes et al., 1981). Most of the known in vitro immunological functions of lymphocytes such as proliferation, differentiation, Ig production, lymphokine secretion, and activation of suppressor and cytotoxic T cells can be induced by lectins. All the effects mediated by lectins require their binding to lymphocytes and this is dependent upon the presence of specific sugar molecules on the lymphocyte surface membrane. Indeed, binding can be prevented by the addition of specific carbohydrates that will function as competitive inhibitors and block the binding of the lectin-cell surface sugar interaction (Sharon, 1983). The molecular basis for interactions between suppressor cells and their targets is not well understood. One possibility is that some forms of suppression (particularly non-antigen-specific suppression) might depend upon the presence of a lectin-like structure on one cell surface and of a specific carbohydrate molecule on the other interacting cell (Tosato et al., 1983a). If this is

EPSTEIN-BARR

VIRUS INFECTION

139

the case, then the addition of large quantities of a specific sugar to these suppressed cultures could competitively saturate the lectin’s binding capacity and, thus, would interfere with suppression. We tested this hypothesis first in suppressor interactions mediated by T cells from patients with acute EBV-induced IM by adding a variety of saccharides to the cultures containing PWM-activated normal mononuclear cells and T cells obtained from IM patients. A variety of sugars, including D-galactose, N-acetyl-D-glucosamine, N-acetyl-D-galactosamine, L-fucose, L-rhamnose, L-mannose, mannose 1phosphate, cellobiose, and gentiobiose, had no effect on the suppression of Ig production mediated by IM T cells. However, D-mannose and a number of mannose derivatives, including a-methyl-D-mannoside, mannose 6-phosphate, and mannan clearly and consistently inhibited the suppressor effect (Fig. 12). The inhibition of suppression by these sugars was dose dependent and at a concentration of 25 mM they had an effect similar to reducing by about fourfold the number of added suppressor T cells. The sugars capable of reversing suppression did not have a mitogenic or a comitogenic effect on the responding normal mononuclear cells. Thus, D-mannose and certain Dmannose derivatives appear to interfere directly with the process of suppression (Tosato et al., 1983a). Reversal of suppression by these sugars is not due to toxicity. First, toxicity would have to be specific for the suppressor T cells, since the responding normal cells appear to be unaffected. Second, preincubation of the suppressor T cells with sugars capable of reversing the suppression did not decrease their ability to inhibit when washed free of sugar. It is likely that Dmannose and those D-mannose derivatives which were active in these experiments represent either saccharides cross-reacting with those naturally involved in the immunoregulatory interaction, or they represent part of a more complex receptor structure, particularly since high sugar concentrations had to be added in vitro in order to reverse suppression (Tosato et al., 1983a). These findings suggested that specific carbohydrate molecules have an important role in suppression mediated by I M T cells. Recently, we have extended our original observations to other suppressor-cell interactions, and tested the effects of saccharides on several other systems, each involving either a T cell or a non-T cell as the effector of suppression. Interestingly, we have observed that D-mannose and a number of D-mannose derivatives, but not other saccharides, have an ability to consistently reduce the magnitude of the suppression of Ig production mediated by human newborn T cells (Tosato et al., 1982c), by the T cells obtained from the patients with chronic infectious mononucleosis (unpublished), and, finally, by suppressor T cells present in the circulation of a subset of patients with common variable hypogammaglobulinemia (Tosato et al., 1984b). In contrast, no reversal of

140

GIOVANNA TOSATO AND H. MICHAEL BLAESE

1o.ooc f$ 3 Y E T R I C 5 SEM -

8.000

[ I

NORMALMNC COCULTURESOF NORMAL M N C AND I M T CELLS

6.000

Y

a

2

2.4,000 2

i:

2

Lu

0

c3

$ Y

3,000

a

U Y

m

-m

T

2,000

61.500

NO SUGAR

D-MAN

NO SUGAR

D-MAN

FIG. 12. Reversal of infectious mononucleosis (1M)-associated T cell suppression by I>inannose (D-man). Normal mononuclear cells (MNC) were cultured either alone (0.5 X 106)or mixed with T cells (0.5 x 106) from 12 patients with acute IM in the presence of pokeweed rnitogen (PWM) only or PWM and D-Man (25 mM). The immunoglobulin-secreting cell response was determined at 7 days.

“late suppression” was noted upon the addition of D-mannose to autologous cocultures of seropositive B and T cells in the presence of EBV, indicating that not all suppressor T cell interactions involve a mannose-related recognition (unpublished). Similarly, no effect was found with D-mannose or any of the D-mannose derivatives tested on suppression of Ig production mediated by non-T cells

EPSTEIN-BARR

VIRUS INFECTION

141

(probably a monocyte) derived from selected patients with common variable hypogammaglobulineinia. However, marked reversal of this non-T cell-mediated suppression was noted with the addition of a different sugar, N acetyl-D-glucosamine (Fig. 13) (Tosato et al., 1984b). These results indicate that different suppressor-cell interactions may utilize different biochemical mechanisms, each involving different sugar molecules as specific recognition, activation, or interaction signals. Many additional studies will be necessary to definitively prove these points. In particular, it is necessary to determine the site of the endogenous lectin-like receptor structure and of the specific interacting carbohydrate. Most likely these structures are located on the cell membranes or on the soluble mediators participating in the suppression. Sugars have an enormous potential for structural diversity. Unlike peptides and oligonucleotides which depend only on their primary sequence and on the number of constituent monomers to encode information, sugar polymers have an additional potential for diversity because the individual sugar subunits may be differently linked and may form branched structures. Because of these characteristics, polysaccharides have unique qualities as carriers of biological information which we believe play an important role in certain antigen-nonspecific forms of immunoregulation. Normal MNC

A

FIG. 13. Suppression mediated by non-T cells is reversed by N-acetyl-D-glucosamine. Normal inonoauclear cells (MNC) were cultured alone (0.5 x lo6) or mixed with non-T cells (0.5 x I@) from a patient with common variable hypogammaglobulinemid in the presence of pokeweed initogen (PWM) only or PWM and the indicated monosaccharides.

142

GIOVANNA TOSATO A N D R. MICHAEL BLAESE

IX. Concluding Remarks

EBV is a unique viral pathogen for man, in that its target for infection is the immune system itself. Not only are B lymphocytes infected by the virus, but once infected, they become functionally activated and also acquire the property of autonomous growth which introduces a considerably more complicated problem of host defense than encountered with other pathogens. Defense mechanisms include serologic responses to viral and virally associated cellular antigens, but the major protective response involves a multitude of cellular responses. During primary infection, nonspecific suppressor and probably nonspecific cytotoxic mechanisms are activated which regulate and destroy virus-infected cells. However, as with all herpes viruses, EBV persists in the body for life after primary infection and remains capable of inducing polyclonal B cell growth and differentiation which are regulated again by both suppressor and cytotoxic mechanisms. As might be expected with such a complex system of regulation, several distinctive human diseases are associated with EBV infection and disorders of its control. Thus, for example, the suppressor cells activated during acute IM which are critical in preventing expansion of EBV-infected B cells may lead to antibody deficiency and agammaglobulinemia if they themselves are not down-regulated after EBV-specific control mechanisms have fully developed. In addition to teaching us about a fascinating complex series of immunoregulatory control mechanisms in bodily defense, EBV has also proven to be especially useful as a laboratory tool. It immortalizes B cells in uitro, providing exceptionally useful lymphoid cell lines, and it induces B cell differentiation in the absence of cooperating accessory and T cells. This later feature has permitted us to demonstrate, for example, that many immunoregulatory suppressor processes act directly on the B cell rather than exerting their control indirectly via inhibition of helper T cell or accessory cell function. As should be obvious from this review, many, many questions remain unanswered concerning the interaction of the immune system with this ubiquitous virus and its role in diseases ranging the gamut from immunodeficiency to autoimmunity to cancer. New tools are being developed almost daily which offer the prospect for an incredible advance in our understanding over the next decade. REFERENCES Abo, W . , Kamada, M., Motoya, T . , Aya, T., Nakao, T., Takada, K . , Imamura, M., Iwanga, M . , Yano, S., and Osato, T. (1982). Lancet 1, 1272-1276. Aman, P., Ehlin-Hendriksson, B., and Klein, G . (1984). J. E x p . Med. 159, 208-220. Andersson, J., Coutinho, A., and Melchers, F. (1977). J. E x p . Med. 145, 1520-1530. Andersson, V., Bird, A. G . , Britton, S . , and Palatios, R. (1981). fmmunol. Reo. 57, 5-38.

EPSTEIN-BARR

VIRUS INFECTION

143

Andersson-Anvret, M., Forsby, N., Klein, G., and Henle, W. (1977). Znt. J. Cancer 20, 486494. Ashwell, G., and Morell, A. G . (1974). Ado. Enzymol. Relat. Areas Mol. Biol. 41, 99-128. Aspaugh, M. A., Jensen, F. C., Rabin, H., and Tan, E. M. (1978).J. E x p . Med. 147, 10181027. Aspaugh, M. A., Henle, G., Lennette, E. T., and Henle, W. (1981).I. Clin. Inuest. 67, 11341140. Aya, T., and Osato, T. (1974). Znt. J. Cancer 14, 341-347. Bar, R. S . , De Lor, C. J., Clausen, K. P., Hurtubise, P., Henle, W . , and Hewetson, J. F. (1974). N . Engl. J . Med. 290, 363-367. Bardwick, P. A., Bluestein, H. G., Zvaifler, N. J., Depper, J. M., and Seegmiller, J. E. (1980). Arthritis Rheum. 23, 626-632. BCchet, J. M., Fialkow, P., Nilsson, K., and Klein, G. (1974). E x p . Cell Res. 89, 275-282. Benner, R., Coutinho, A,, Rijnbeck, A. M., van Oudenaren, A,, and Hooijkaas (1981). Eur. J. Immunol. 11, 799-804. Bird, A. G., and Britton, S. (1979a). Immunol. Reu. 45, 41-66. Bird, A. G . , and Britton, S. (197911).Scand. J . Immunol. 9, 507-510. Bird, A. G . , McLachlan, S. M., and Britton, S. (1981). Nature (London) 289, 300-301. Borel, J. F . , Feurer, C . , Magnee, C., and Stahelin, H. (1977). Ztnmunology 32, 1017-1025. Bornkamrn, G. W., Delius, H . , Zirnber, U . , Hadenwentz, J., and Epstein, M. A. (1980).J . Virol. 35, 603-618. Borzy, M. S., Hong, R., Horowitz, S . D., Gilbert, E., Kaufinan, D., DeMendonca, W., Oxelius, V.-A., Dictor, M., and Pachman, L. (1979). N . Engl. J. Med. 301, 565-568. Britton, S . , and Palatios, R. (1982). Immunol. Rev. 65, 5-22. Britton, S., Andersson-Anvret, M., Gergely, P.. Henle, W., Jondal, M., Klein, G., Sandstedt, B., and Svedmyr, E. (1978). N . Engl. J. Med. 298, 89-92. Bunjes, D., Hardt, C., Rollinghoff, M., and Wagner, H. (1981).Eur.]. Immunol. 11, 657-661. Calne, R. Y . , Rolles, K . , Thiru, S . , McMaster, Craddock, G. N . , Aziz, S . , White, D. J . G . , Evans, D. B., Dunn, D. C., Henderson, R. G . , and Lewis, P. (1979). Lancet 1, 10331036. Catalano, M. A., Carson, D. A., Slovin, S. F., Richman, D. D., and Vaughan, J. H. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 5825-5828. Chang, R. S . , Hsieh, M. W., and Blankenship, W. (1971).J. Natl. Cancer Inst. 47, 469-476. Chang, R. S . , Lewis, J. P., and Abildgaard, C. F. (1973). N . Engl. J . Med. 289, 1325-1329. Coope, D., Heston, L., Brandsma, J . , and Miller, G. (1979).J. Imrnunol. 123, 232-238. Crawford, D. H., Thomas, J. A., Janossi, G., Sweny, P., Fernando, 0 . N . , Moorhead, J. F., and Thompson, J. H. (1980). Lancet 1, 1355-1356. Dalla-Favera, R., Bregni, M . , Erikson, J., Patterson, D., Gallo, R. C . , and Croce, C. M. (1982). Proc. Natl. Acad. Sci. U . S . A . 79, 7824-7827. Dambaugh, T., Raab-Taub, N . , Heller, M., Beisel, C., Hummel, M., Cheung, A , , Fennewald, S . , King, W., and Kieff, E. (1980). Ann. N.Y. Acad. Sci. 602, 711-719. Delius, H., and Bornkamm, G . W. (1978).J. Virol. 27, 81-89. Depper, J. M., Bluestein, H. G . , and Zvaifler, N. J. (1981). J. Immunol. 127, 1899-1902. Desgrandes, C., Wolf, H., de-ThC, G., Shanrnugaratnam, K., Cammoun, N . , Ellouz, R . , Klein, G., Lennert, K., Munoz, N . , and zur Hausen, H. (1975). Int. J. Cancer 16, 7-15. de-Th6, G . , Geser, A,, Day, N. E . , Tukei, P. M., Williams, E. H., Beri, D. P . , Smith, P. G . , Dean, A. G., Bornkarnm, C . W., Feorino, P., and Henle, W. (1978).Nature (London)274, 756-761. De Waele, M., Thielernans, C., and Van Camp, B. K. J. (1981).N . Engl. J. Med. 304,460-461. Diehl, V., Henle, G., Henle, W., and Khon, G . (1968).J. Virol. 2, 663-669.

144

GIOVANNA TOSATO AND R. MICHAEL BLAESE

Dos Reis, G. A,, and Shevach, E. (1982).J. Immunol. 129, 2360-2367. Edson, C. M., and Thorley-Lawson, D. A. (1981).J. Virol. 39, 172-184. Einhorn, L., and Ernberg, I. (1978). Znt. J. Cancer 21, 157-160. Elson, C. J., Crawford, D. H., Bucknall, R. C., Allen, C., Thompson, J. L., Epstein, M. A., Hall, N. D., and Bacon, P. A. (1979). Lancet 1, 105. Epstein, M. A,, and Achong, B. G. (1977). Lancet 2, 1270-1272. Evans, A. S., Niederman, J. C., and McCollum, R. W. (1968). N. Engl. J . Med. 279, 11231127. Ferrell, P. B., Aitcheson, C. T., Pearson, G . R., and Tan, E. M. (1981).J . C h . Znoest. 67, 68 1-687. Fialkow, P. J., Klein, G., Gartler, S. M., and Clifford, P. (1970). Lancet 1, 384-386. Fischer, D. K., and Rabson, M. S. (1982). N. Engl. J. Med. 306, 111. Fisher, D., Miller, G., Gradoville, L., Heston, L., Westrate, M., Maris, W., Wright, J., Brandsma, J., and Summers, W. (1981). Cell 24, 543-554. Fleisher, G., Starr, S., Koven, N., Kaniiya, H., and Henle, W. (1982).J. Pediatr. 100, 727730. Forbes, J. T., Brettauer, R. K., and Oeltrnann, T. N. (1981).Proc. Natl. Acad. Sci. U.S.A.78, 5797-5801. Fresen, K. O., Merkt, B., Bornkarnm, G. W., and zur Hausen, H. (1977). Znt. J. Cancer 19, 317-323. Fu, S. M., and Hurley, J. N. (1979). Proc. Natl. Acad. Sci. U . S . A . 76, 6637-6640. Gearhart, P., Sigal, N. H., and Klinnian, N. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 17071711. Gelh, L. D., and Lerner, A. M. (1965). Science 147, 404-405. Gerber, P., Whang-Peng, J., and Monroe, J. H. (1969).Proc. Natl. Acad. Sci. U.S.A. 63,740747. Gerber, P., Nonoyama, M., Lucas, S., Perlin, E., and Goldstein, L. I. (1972). Lancet 2, 988989. GivBn, D., and Kieff, E. (1978).J. Virol. 28, 524-542. Grabel, L. B., Rosen, S. D., and Martin, G. R. (1979). Cell 17, 477-484. Graessmann, A., Wolf, H., and Bornkamm, G . W. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 433-436. Greally, J. F., Lawlor, E . , Lyons, J., Rickinson, A., Sakanioto, K., and Purtilo, D. (1983)./. Clin. Zmmunol. 3, 207-211. Greaves, M. F., and Brown, G . (1975). Clin. Zmmunol. Zmmunopathol. 3, 514-524. Griffin, B. E., and Karran, L. (1984). Nature (London) 309, 78-82. GunvBn, P., Klein, G., Klein, E., Norin, T., and Singh, S. (1980).Znt. J. Cancer 25, 711-719. Hamilton, J. K., Paquin, L. A., Sullivan, J. L., Maurer, H. S., Cruzi, F. G., Provisor, A. J., Steuber, C. P., Hawkins, E., Yawn, D., Cornet, J. A., Clausen, K., Finkelstein, G. Z., Landing, B., Grummet, M., and Portilo, D. (1980).J . Pediatr. 96, 669-673. Hansen, J. A., Martin, P. J., Beatty, P. G., Clark, E. A., and Ledbetter, J. A. (1984). In “Human Leukocyte Markers Detected by Monoclonal Antibodies” (G. S. Eisenbarth and B. Haynes, eds.), pp. 195-212. Springer-Verlag, Berlin and New York. Hansson, M., Falk, K., and Ernberg, I. (1983).J. E x p . Med. 158, 616-622. Hanto, D. W., Frizzera, G., Gajl-Peczalska, K. J., Sakamoto, K., Purtilo, D. T., Balfour, H. H., Simmons, R. L., and Najarian, J. S. (1982). N . Engl. J . Med. 306, 913-918. Haler, F., Bluestein, H. G., Zvaifler, N. J., and Epstein, L. B. (1983).J. E x p . Med. 157, 173188. Haynes, B. F., Schooley, R. T., Payling-Wright, C. R . , Grouse, J. E . , Dolin, R., and Fauci, A. S. (1979).J. Zmmunol. 123, 2095-2101.

EPSTEIN-UAHH

VIRUS INFECTION

145

Heller, H., Gerber, P., and Kieff, E. (1981a).J. Virol. 37, 698-709. Heller, M., Danibaugh, T., and Kieff, E. (198111). J. Virol. 38, 632-648. Henderson, E., Miller, G . , Robinson, J., and Heston, L. (1977). Virology 76, 152-163. Henle, G., and Henle, W. (1979). In “The Epsteiii-Barr Virus” (M. Epstein and B. Achong, eds.), pp. 297-320. Springer-Verlag, Berlin and New York. Henle, G., Henle, W., and Diehl, V. (1068). Proc. Natl. Acad. Sci. U . S . A . 59, 94-101. Henle, G . , Lennette, E. T., Aspaugh, M. A,, and Henle, W. (1979). Clin. Exp. Imtnunol. 36, 415-422. Henle, W., and Henle, G. (1979). In “The Epstein-Barr Virus” (M. Epstein and B. Achong, eds.). pp. 61-78. Springer-Verlag, Berlin and New York. Henle, W., Diehl, V., Kohn, G . , zur Hansen, H., and Henle, 6. (1967). Science 157, 10641065. Henle, W., Henle, G., Zajac, B., Pearson, G . , Waubke, R., and Scriba, M. (1970a). Science 169, 188-190. Henle, W., Henle, G., Ho, H.-C., Burtin, P., Cachin, Y., Clifford, P., de Schryrer, A,, deTh6, G., Diehl, V., and Klein, 6. (1970b). J. Natl. Cancer Znst. 44, 225-231. Henle, W., Henle, G . , and Horwitz, C. A. (1974). ZIuni. Pathol. 5, 551-565. Hess, A. D., and Tutscha, P. J. (1980).J . Zmmunol. 124, 2601-2608. Heston, L., Rabson, M., Brown, N., and Miller, 6. (1982). Nature (London) 295, 160-163. Hinuma, Y . , Konn, M., Yarnaguchi, J., Wudarski, D., Blaskeslee, J . , and Grace, J. (1967).J. Virol. 1, 1045-1051. Hoffman, G . J., Lazarowtiz, S. G., and Hayward, S. D. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 2979-2983. Huang, D. P., Ho, H. C., Henle, W., Henle, G . , Saw, D., and Lui, M. (1978). Znt. J. Cancer 22, 266-274. Hutt-Fletcher, L. M., Fowler, E . , Lambris, J. D., Feighny, K. J., Simmons, J. G . , and Ross, G. 11. (1983). 1.Irnniunol. 130, 1309-1312. Johnsen, H. E. , Madsen, M . , and Kristensen, T. (1979). Scund. 1. Zmrnunol. 10, 251-255. Jondal, M., and Klein, G. (1973).J. E r p . Med. 138, 1365-1378. Jondal, M., Klein, G., Oldstone, M. B. A,, Bokish, V., and Yefenof, E. (1976). Scand. J . Irnrnritio/. 5 , 401-410. Katamine, S . , Otsu, M., Tada, K., Tsuchiya, S., Sato, T., Ishida, N., Hanjo, T., and Ono, Y. (1984). Nature (London)309, 369-372. Kawai, Y . , Nonoyama, M., and Pagano, J. (1973).J . Virol. 12, 1006-1012. . U . S . A . 71, 325-358. Kieff, E., and Levine, J. (1974). Proc. Natl. A c Q ~ Sci. Kieff, E., Dambaugh, T., Hummel, M . , and Heller, M. (1983). In “Advances in Viral Oncology” (G. Klein, ed.), pp. 133-182. Raven, New York. Kirchner, H. , Tosato, G . , Blaese, R. M., Broder, S., and Magrath, I. T. (1979).J . Zrninunol. 122, 1310-1313. Klein, E., Klein, G., Nadkarni. J. S., Nadkarni, J. J . , Wigzell, H., and Clifford, P. (1968). Cancer Res. 28, 1300-1310. Klein, E., Masucci, M. G., Berthold, W., and Blazar, B. A. (1980). Cold Spring Harbor Conf. Cell Proli& 7 , 1187-1197. Klein, E., Ernberg, I., Masucci, M. G . , Szigeti, R., Wu, Y. T., Masucci, G., and Svedmyr, E. (1981). Cancer Res. 41, 4210-4215. Kung, P. C., Goldstein, G . , Reinherz, E. L., and Schlossnian, S. F. (1979). Science206, 347349. Lai, P. K., Alpers, M. P., and MacKay-Scollay, E. M . (1977). Znt. J. Cancer 20, 21-29. Lanier, L. L., Englernan, E. G., Gatenby, P., Babcock, G. F., Warner, N. L., and Herzenberg, L. A. (1983). Zmrnunol. Reo. 74, 143-160.

146

GIOVANNA TOSATO AND R. MICHAEL BLAESE

Lemon, S. M., Hutt, L. M., Shaw, J. E., Li J.-L. H., Nonoyama, M.,and Pagano, J. S. (1977). Nature (London) 268, 268-270. Leonard, W., Depper, J. H., Uchiyama, T., Smith, K. A., Waldmann, T. A., and Greene, W. C. (1982). Nature (London) 300, 267-269. Lindahl, T., Klein, G., Reedman, B. M., Johansson, B., and Singh, S. (1974).Int. J. Cancer 13, 764-772. Lipinski, M., Fridman, W. H . , Turz, T., Vincent, C., Pious, D., and Fellous, M. (1979).J. E x p . Med. 150, 1310-1322. MacDonald, H. R . , and Cerottini, J. C. (1979).J. Zmmunol. 122, 1067-1072. Magrath, I., Erikson, J., Whang-Peng, J., Sieverts, H., Armstrong, G., Benjamin, D., Triche, T., Alabaster, O., and Croce, C. (1983). Science 222, 1094-1098. Manalov, G., and Manalava, Y. (1972). Nature (London) 237, 33-34. Martin, P. J., Shulman, H. M., Schulbach, W. H., Hansen, J. A,, Feter, A., Miller, G., and Thomas, E. D. (1984). Ann. Znt. Med. 101, 310-315. Martinez-Maza, 0..and Britton, S. (1983).J. E x p . Med. 157, 1808-1814. Masucci, M. G., Bejarano, M. T., Masucci, G., and Klein, E. (1983). Cell. Ztnniunol. 76, 311321. Menezes, J., Leibold, W., and Klein, G. (1975). E x p . Cell Res. 92, 478-484. Miller, G., Enders, J. F., Lisco, H., and Kohn, H. I. (1969). Proc. SOC. Exp. Biol. Med. 132, 247-252. Miller, G., Niederman, J. C., and Andrews, L. L. (1973). N . Engl. J . Med. 288, 229-232. Miller, G., Robinson, J., Heston, L., and Lipman, M. (1974).Proc. Natl. Acad. Sci. U.S.A.71, 400-4010. Miller, G., Grogan, E . , Heston, L., Robinson, J., and Smith, D. (1981). Science 212,452-655. Misko, I. S., Moss, D. J., and Pope, J. H. (1980).Proc. Natl. Acad. Sci. U.S.A.77,4247-4250. Mitelman, R. (1981).Ado. Cancer Res. 34, 141-167. Miyawaki, T., Yachie, A., Ohzeki, S., Nagaoki, T., and Taniguchi, N. (1983).J. Immunol. 130, 2737-2742. Mizuno, F., Aya, T., and Osato, T. (1974). Br. Med. J. 3, 689. Morgan, D. G., Niederman, J. C., Miller, G., Smith, H. W., and Dowalibi, J. M. (1979). Lancet 2, 1154-1157. Morrow, R. H . , Kisuule, A., Pike, M. C., and Smith, P. G. (1976)./. Notl. Cancer Znst. 56, 479-483. Moss, D. J., Pope, J. H., and Scott, W. (1976). Med. Microbial. Zmmunol. 162, 159-167. Moss, D. J ~ Scott, , W., and Pope, J. H. (1977). Nature (London) 268, 735-736. Moss, D. J., Rickinson, A. B., and Pope, J. H. (1978). Znt. J. Cancer 22, 662-668. Moss, D. J., Rickinson, A. B., and Pope, J. H. (1979). Int. /. Cancer 23, 618-625. Moss, D. J., Wallace, L. E., Rickinson, A. B., and Epstein, M. A. (1981a). Eur. J. Zmmunol. 11, 686-693. Moss, D. J., Rickinson, A. B., Wallace, L. E., and Epstein, M. A. (1981b). Nature (London) 291, 664-666. Muramatsu, T., Gachelin, G., Damonneville, M., Delahre, C., and Jacob, F. (1979). Cell 18, 183-191. Navington, J., and Gray, J. (1980). Lancet 1, 536-537. Neel, B. G., Jhanwar, S. C., Chaganti, R. S. K., and Hayward, W. S. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 7842-7846. Niederman, J. C., McCollum, R. W., Henle, G., and Henle, W. (1968).J. Am. Med. Assoc. 203, 205-209. Nilsson, K. (1971). Znt. J. Cancer 8, 432-441.

EPSTEIN-HARH

VIRUS INFECTION

147

Nilsson, K. (1979). In “The Epstein-Barr Virus” (M. A. Epstein and B. G. Achong, eds.), pp. 225-281. Springer-Verlag. Berlin and New York. Nilsson, K . , and Klein, G. (1982). Adv. Cancer Res. 37, 319-380. Nilsson, K . , and Ponten, J. (1975). Znt. J . Cancer 15, 321-341. Nilsson, K . , Klein, G . , Henle, W.. and Henle, G. (1971). Int. J. Cancer 8, 443-450. Nonoyania, M., and Pagano, J. S. (1971). Nature (Zmidun) New Biol. 233, 103-106. Ofek, I . , Mirelnian, D., and Sharon, N. (1977). Nature ( L u n c h )265, 623-625. Old, L. J . , Boyse, E. A., Oettgen, H. F., De Harven, E . , Geering, G., Williamson, B., and Clifford, P. (1966). Proc. Natl. Acad. Sci. U . S . A . 56, 1699-1704. Patel, P. C., Dorval, G., and Menezes, J. (1982). Infect. Z?nmun. 38, 251-259. Pearson, G. R . , and Qualtiere, L. F. (1978).]. Vit-01. 28, 344-351. Pizzo, P. A., Magrath, I. T., Chattopadhyay, S . K., aiggar, R . J.. and Gerber, P. (1978).Nature (London) 272, 629-631. Pope, J. H., Horne, M. K.,and Scott, W. (1968). Int. J. Cancer 3, 857-866. Pritchett, R. F . , Hayward, S. D., and Kieff, E. (1975).J. Virol. 15, 556-569. Pritchett, R. F . , Pedersen, M., and Kieff, E. (1976). Virulugy 74, 227-231. Provisor, A. J., Iacuone, J. J., Chilcote, R. R., Neiburger, R. G . , Crussi, F. G., and Baehner, R. L. (1975). N . Engl. J. Med. 293, 62-65. Purtilo, D. T. (1980).J. Pediatr. 96, 669-673. Purtilo, D. T., Yang, J. P. S., Cassel, C. K., Harper, P., Stephenson, S. R., Landing, B. H., and Vawter, G. F. (1975). Lancet 1, 935-941. Purtilo, D. T., DeFlorio, D., Hutt, L. M., Bhawan, J., Yang, J. P. S., Otto, R.. and Edwards, W. (1977). N . EngI. J. Med. 297, 1077-1081. Purtilo, D. T., Sakamoto, K., Barnabei, V., Seeley, J.. Bechtold, T., Rogers, G., Yetz, J., and Harada, S. (1982). Am. J. Med. 73, 49-56. Raab-Traub, N., Danibaugh, T., and Kieff, E. (1980). Cell 22, 257-267. Rabson, M., Heston, L., and Miller, G. (1983). Proc. Natl. Acad. Sci. U . S . A . 80, 2762-2766. Ragona, G., Emberg, I., and Klein, G. (1980). Virology 101, 553-557. Reedman, B. M., and Klein, G. (1973). Int. J. Cancer 11, 449-520. Reinherz, E. L., Kung, P., Goldstein, G . , and Schlossnian, S. F. (1979).Proc. Natl. Acad. Sci. U.S.A. 76, 406-4065. Reinherz, E. L., O’Brien, C . , Rosenthal, P., and Schlossman, S. F. (1980).J. Inimunol. 125, 1269- 1274. Rickinson, A. B., and Moss, D. J. (1983). 111 “Advances in Viral Oncology” (G. Klein, ed.), pp. 213-238. Raven, New York. Rickinson, A. B., Jarvis, J. E., Crawford, D. H . , and Epstein, M. A. (1974).Int. J. Cancer 14, 704-715. Rickinson, A. B., Epstein, M. A,, and Crawford, D. H. (1975).Nature (London)258,236-238. Rickinson, A. B., Moss, D. J., and Pope, J. H. (1979). Int. J. Caiicer 23, 610-617. Rickinson, A. H . , Moss, I).J., Pope, J. H., and Ahlberg, N. (1980a).Znt.J. Cancer 25, 59-65, Rickinson, A. B., Wallace, L. E., and Epstein, M. A. (19801,). Nature (London) 283, 865-867. Robinson, J. E., Brown, N., Andiman, W., Halliday, K., Francke, U., Robert, M. F., Andersson-Anvret, M., Hortsman, D., and Miller, G . (1980). N . Engl. J. Med. 302, 12931297. Robinson, J., Smith, D., and Niederman, J. (1981).J. E x p . Med. 153, 235-244. Rosen, A. M., Gergely, M., Jondal, M., Klein, G., and Britton, S. (1977).Nature (London)267, 52-54. Rowe, M., Hildreth, J. E. K., Rickinson, A. B., and Epstein, M . A. (1982). Irit. J. Cancer29, 373-381.

148

GIOVANNA TOSATO A N D R. MICHAEL HLAESE

Sawyer, R. N . , Evans, A. S., Niedernian, J. C . , and McCollum, R. (1971).J. Infect. Dis. 123, 263-270. Schooley, R. T., Haynes, B. F., Grouse, J., Payling-Wright, C., Fauei, A. S.. and Dolin,R. (1981). Blood 57, 510-517. Seeley, J., Svedmyr, E., Weiland, O . , Klein, G., Mnller, E., Ericksson, E., Andersson, K., and Van der Waal, L. (1981).J. Znmunol. 127, 293-300. Sharon, N. (1983). Adv. Zmmunol. 34, 213-298. Shen, L., Grollman, E. F., and Ginsburg, V. (1968).Proc. Nutl. Acad. Sci. U . S . A .59,224-237. Shepherd, V. L., Lee, Y. C . , Schlesinger, P. H . , and Stahl, P. D. (1981).Proc. Natl. Acad. Sci. U.S.A. 78, 1019-1022. Shope, T. C., and Kaplan, J. (1979).J. Znirnunol. 123, 2150-2155. Simmons, J. G., Hntt-Fletcher, L. M . , Fowler, E., and Feighny, R. J. (1983).J. I7nrnunol. 130, 1303- 1308. Sixbey, J. W., Nedrud, J. G . , Rub-Trauh, N . , Hanes, R. A., and Pagano, J. S. (1984).N . Engl. 1. Med. 310, 1225-1230. Slaughter, L., Carson, D. A., Jensen, F. C . , Holbrook, T. L., and Vaughan, J. H. (1978).J. E x p . Med. 148, 1429-1434. S h i n , S. F., Frisman, D. M . , Tsoukas, C. D., Royston, I., Bird, S . M . , Wormsley, S. B . , Carson, D. A,, and Vaughan, J. H. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 2649-2653. Stahl, P. D., Rodnian, J. S., Miller, M. J., and Schlesinger, P. H. (1978).Proc. Natl. Acad. Sci. U . S . A . 75, 1399-1403. Stein, L. D . , Ledgley, C. J , , and Sigal, N. H. (1983). Irnmunol. 130, 1640-1645. Straus, E. S., Tosato, G., Armstrong, G . , Lawley, T., Preble, 0. T., Henle, W., Davey, R., Brus, I., and Blaese, R. M. (1985). Ann. Znt. Med. (in press). Stutman, O., Dieu, P . , Wisun, R. E . , and Lattime, E. C. (1980).Proc. Natl. Acad. Sci. U.S.A. 77, 2895-2898. Sugden, B. (1982). Reo. Infect. Dis. 4, 1048-1061. Sugden, B., and Mark, W. (1977).J. Virol. 23, 503-508. Sugden, B., Summers, W. C., and Klein, G. (1976).J. Virol. 18, 765-775. Sugden, B., Phelps, M., and Domoradzki, J. (1979).J. Virol. 31, 590-595. Sullivan, J. L., Byron, K. S., Brewster, F. E . , Baker, S. M., and Ochs, H. D. (1983).J. Clin. Invest. 71, 1765-1778. Sutton, R. N . , P., Reynolds, K., Almond, J. P., Marston, S. D., and Edmond, R. J. D. (1973). Clin. E x p . Iinnzunol. 13, 359-366. Svedmyr, E., and Jondal, M. (1975). Proc. Nutl. Acad. Sci. U.S.A. 72, 1622-1626. Takada, K., and Osato, T. (1979). Znteroirology 11, 30-39. Tanaka, Y., Sugamura, K., and Hinuma, Y. (1982).J. Zinrnunol. 128, 1241-1245. Taub, R . , Kirsch, I., Morton, C., Lenoir, G., Swan, D., Tronick, S., Aaronson, S . , and Leder, P. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 7837-7841. Teale, J. M., Lafrenz, D., Klinman, N . R., and Strober, S . (1981).J. Zmrnunol. 126, 1952-1957. Thiru, S . , Calne, R. Y., and Navington, J. (1981). Transplant. Proc. 13, 359-364. Thorley-Lawson, D. A. (1979a).Nature (London)281, 486-488. Thorley-Lawson, D. A. (1979h). Cell 16, 33-42. Thorley-Lawsn, D. A. (1980).J. Zitmunol. 124, 745-751. Thorley-Lawson, D. A. (1981).J . Zmnunol. 102, 829-833. Thorley-Lawson, D. A . , and Edson, C. M. (1979).J. Virol. 32, 458-467. Thorley-Lawson, D. A . , and Strominger, J. L. (1978). Virology 86, 432-441. Thorley-Lawson, D. A., Chess, L., and Strominger, J. L. (1977).J. E x p . Med. 146, 495-508. Tobi, M., Ravid, Z . , Feldman-Weis, V., Ben-Chetrit, E., Morag, A., Chowers, I., Michaeli, Y., Shalit, M., and Knobler, H. (1982). Lancet 1, 61-63.

EPSTEIN-BAHH VIRUS INFECTION

149

Tosato, G . , and Blaese, R. M . (1985). In “Ataxia-telangectasia: Genetics, Neuropathology and Itnnrunology of a Degenerative Disease of Childhood” (R. A. Gatti and M. Swift, eds.). Liss, New York, in press. Tosato, G . , Magrath, I . , Koski, I . , Dooley, N . , and Blaese, R. M . (1979).N . Engl. J. Med. 301, 1133-1137. Tosato, G . , Magrath, I. T., and Blaese, R. M. (1981a).I n “Mechanisms of Lymphocyte Activation” (K. Resh and H. Kirchner, eds.), pp. 621-623. Elsevier, Amsterdam. Tosato, C., Steinberg, A. D., and Blaese, R. M. (1981b). N . Engl. J. Med. 305, 1238-1243. Tosato, C., Magrath, I . T . , and Blaese, K. M. (1982a). J. Ztnn~unol.128, 575-579. Tosato, G . , Pike, S. E . , Koski, I. R., and Blaese, R. M . (1982b).J. lmtnunol. 128, 1986-1991. Tosato, G., Pike, S., De Seau, V., and Blaese, R. M. (1982~). Fed. Proc., Fed. Am. Soc. E x p . B i d . 41, 799. Tosato, C., Pike, S. E., and Blaese, R. M. (1983a). J . EX),. Med. 158, 1048-1060. Tosato, C., Straus, S., Butler, J., Pike, S. E., De Seau, V., Brus, I., and Blaese, R. M. (198%). Clin. Res. 31, 522. Tosato, C . , Yarchoan, R., Pike, S., De Seau, V., and Blaese, R. M. (1983~). Fed. Proc., Fed. Am. Soc. E x p . B i d . 42, 411. Tosato, G., Steinberg, A. D., Yarchoan, K., Heilman, C. A., Pike, S. E., De Seau, V., and Blaese, R. M. (1984a).J. C h . Invest. 73, 1789-1795. Tosato, G . , Pike, S., Muchmore, A. V., and Blaese, R. M. (19841,). Clin. Res. 32, 361a. Tsoukas, C. D., Carson, D. A , , Fong, S., Slovin, S. F., Fox, R. I . , and Vaughan, J. H. (1982). Clin. Immunol. Iniinuiiopathol. 24, 8-14. Tsukuda, K., Volsky, D. J., Shapiro, I. M., and Klein, 6. (1982). Eur. J. lmniunol. 12, 87-90. Turz, T., Fridman, W. H . , Senik, A , , Tsapis, A,, and Fellow, M. (1977).Nature (London) 269,

806-808. Uchiyanra, T., Broder, S., and Waldman, T. (1981). Iminunol. 126, 1393-1397. Vacquier. V. D., and Moy, G. W. (1977). Proc. Natl. Acad. Sci. U . S . A . 74, 2456-2460. Virelizier, J. L., Lenoir, G., and Criscelli, G. (1978). Lancet 2, 231-234. Volsky, D. J , , Shapiro, I . M., and Klein, G. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 54535457, Volsky, D. J., Klein, G., Volsky, B . , and Shapiro, I. M . (1981).Nutitre (London)293, 399-401. Wainwright, W. H . , Veltri, R. V., and Sprinkle (1979).J. Infect. Dis. 140, 22-32. Wallace, L. E., Moss, D. J., Rickinson, A. B., MeMichael, A. J., and Epstein, M. A. (1981). Eur. J. Iinmunol. 11, 694-699. Wallace, L. E., Rickinson, A. B., Rowe, ki., and Epstein, M. A. (19824. Nature (London)297, 413-415. Wallace, L. E . , Rickinson, A. B., howe, M., Moss, D. J.. Allen, D. J . , and Epstein, M. A. (1982b). Cell Immunol. 67, 129-140. White, I). J. G . , Plumb, A. M., Pawelec, C., and Brons, 6. (1979).J . Znimund. 27, 55-58. Wollheim, F. A,, and Williams, R. C . , Jr. (1966). N . E n g l . J. Med. 274, 61-67. Yajima, Y . , and Nonoyania, M . (1976).J. Virol. 19, 187-194. Yarchoan, R., Tosato, G., Blaese, R. M., Simon, R. M., and Nelson, D. (1983).J. E x p . Med. 157, 1-14. Yata, J., Desgranges, C . , Nakagawa, T., Favre, M. C., and de-The, G . (1975). Znt. /. Cancer 15, 377-384. Ziegler. J. L. (1981).N . Engl. J. Med. 305, 735-745. zur Hauren, H., Diehl, V., Wolf, H., Schulte-Holtlrauseir. H., and Schneider, V. (1972). Nature (London) New Biol. 237, 189-190.

This Page Intentionally Left Blank

ADVANCES IN IMMLINOLOGY. VOL. 37

The Classical Complement Pathway: Activation and Regulation of the First Complement Component’ NEIL R. COOPER Department of Immunology, Scripps Clinic and Research Foundation, l a Jolla, California

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. History of the Classical Pathway of the Complement System ... ..... 111. The Pro A. C l q ... .............. B. C l r ....................................................... C. C1 Inhibitor (Cl-In) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. The Complexes of the C1 Activation Unit . . . . . . . . . . . . . A. The ClrzClsz Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. C1 . . . . . . . . . . V. The C l Activation

C. Regulation of C1 Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI. Actions of Activated C1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Regulation and Fate of Activated C1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Inhibition of Activated C1 by Cl-In . . . . . . . B. Dissociation of Activated C1 C. C l q Receptor Interactions ............................. VIII. Coinmetit . ....................................................... ....................................................... References

151

165 178 183

200 202 204

206 2m 207

1. Introduction

The history of studies of the complement system goes back approximately 100 years to the first investigations of the nature of immunity and includes, through the years, many of the major investigators in immunology: Buchner, Bordet, and Ehrlich to name several of the earliest. The studies have progressed at an uneven pace-a series of seminal discoveries initiating phases of rapid discovery with intervening periods of modest growth. The first phase of growth began at the end of the last century with the discovery of the destructive actions of complement on bacteria and sensitized erythrocytes and progressed through the first decade of this century; it was followed by a period of steady but slow growth in knowledge. The next phase of rapid 1

Publication number 3541 I M M 151 Copyright 0 1985 by Acadeniic Press. Inc. All rights of reproduction in any form reserved. ISBN 0-12-022437-2

152

NEIL R . COOPER

discovery was initiated with the discovery that complement consists of several sequentially reacting factors. Reagents to detect and measure the four known complement components were elaborated and used for the analysis of reaction mechanisms and the assessment of complement involvement in disease. After a period of modest advance, the approximately simultaneous elaboration of a mathematically based theory of complement action and the initial applications of the techniques of modern biochemistry to the study of the complement system in the late 1950s and early 1960s permitted the isolation of the factors in a high state of purity and initiated the analysis of their reaction mechanisms and biological reactions in molecular terms. The period of rapid advance following these discoveries and their application was again followed by a period of slower progress. Another phase of rapid discovery is likely to be initiated with the applications of the techniques and approaches of molecular genetics to the complement system. It is appropriate at this juncture to evaluate the current state of knowledge of the complement system. The present review concentrates on the initial events occurring during the activation of the classical pathway. The review aims to present a comprehensive picture of the current state of knowledge of the events involved in the activation and regulation of the first complement component. Space precludes inclusion of studies of C1 biology and biochemistry which are not directly pertinent to the C1 activation process and its control mechanisms. These include evolving information about the genes which code for the proteins of the C1 activation complex and investigations of the biosynthesis of the C1 proteins. Also not included because of space limitations and lack of information concerning relevance to the activation process are studies of the C l q inhibitor and other factors which interact with C1 such as fibronectin. A series of investigations which indicates that mernbrane associated C l q represents an intrinsic Fc receptor in the plasma membrane of certain cells as well as other investigations which suggest that intercellular interactions such as platelet adhesion and phagocytosis and intracellular activation reactions such as stimulation of oxidative processes and mediator release have also not been considered in detail for the same reasons. The overall perspective of this review is historical; emphasis has been placed on major discoveries and confirmations in each area pertaining to the subject so as to present a comprehensive overview; limitations of space have precluded citing many relevant papers. Reviews within the last 10 years dealing with various aspects of the classical pathway of the complement system are by Cooper (l), Loos (2), Muller-Eberhard (3), Porter and Reid (4,5), Reid and Porter (6), Sim (7), and Ziccardi (8). Earlier good reviews by Mayer (9), Muller-Eberhard (lo), Osler ( l l ) , and Osborn (12) are also recommended.

T H E CLASSICAL C O M P L E M E N T PATHWAY

153

II. History of the Classical Pathway of the Complement System

In 1888, Nuttall (13) observed that blood destroyed anthrax bacilli in vitro in a reaction which did not require leukocytes. It was thus distinguished from phagocytosis which was under considerable study at that time since its discovery slightly earlier by Metchnikoff (14). Buchner (15) showed that cellfree serum possessed the ability to destroy typhoid bacilli in 1889; both Nuttall and Buchner were aware that this activity was relatively labile and lost after heating at 55°C for 30 minutes or on storage. Buchner termed the activity alexin. He also found (16) that the ability to heinolyze erythrocytes of another species, now known as immune hemolysis, was also lost on comparable heating of the serum. Bordet (17) made the important discovery in 1895 that the bactericidal activity of serum from an immunized animal was lost after heating and restored by the addition of normal serum. He suggested that the bactericidal activity was due to two factors: a stable specific substance, termed the preventive substance and later the sensitizer, amboceptor or antibody, which increased with immunization and a heat labile factor, termed alexin or addiment and later complement, which did not increase during immunization. Both factors were necessary for inactivation of the bacteria. Bordet also showed in 1898 (18) that the same phenomenon occurred with erythrocytes. Bordet stated: “We regard the sensitizer as uniting with the corpuscle and so modifying it as to allow a direct absorption of the alexin.” Ehrlich and Morgenroth (19-21) obtained similar results in investigating immune hemolysis and showed also that antibody would combine with the erythrocytes in the absence of complement, or in its presence, but that complement would only react after antibody had done so, thus establishing the reaction sequence for these two factors. H e went on to formulate his famous lock and key hypothesis for the action of antibody and complement and drew diagrams showing the antibody linking the complement to the antigen (22). The complexity of the complement system was also apparent from the earliest days. Buchner (16) observed that dialysis against distilled water destroyed Complement activity. Ferrata in 1907 (23) made the same observation but in addition showed that Complement activity was restored by recombining the redissolved precipitate together with the supernatant fraction. Sachs and Terruchi (24) and Brand (25) confirmed the work the same year; furthermore, Brand showed that the redissolved precipitate functioned before the soluble fraction. This represented the discovery of the first and second components and established their sequence of action. The precipitated fraction was known for some time as the midpiece and the soluble

154

NEIL R. COOPER

fraction as the endpiece; other names for these fractions were globulin and albumin fractions and euglobulin and pseudoglobulin. The third component was identified in 1911 by Omorokow, Sachs, and collaborators (26-28) as a heat-stable functional complement activity destroyed by treatment of serum with cobra venom. Coca (29) showed shortly thereafter that the same factor was destroyed by treatment of serum with yeast. Weil (30) found that the third component acted after the first and second components. C4 was not discovered until some years later in 1926 by Gordon et al. (31). It was identified as a heat-stable factor destroyed by treatment of serum with ammonia. A number of studies of the complement system appeared during the 30year period from 1910 to 1940. Osborn in his comprehensive review (12) cites over 400 papers published in the complement field before 1935. These papers dealt with various methods for preparing complement deficient sera, studies of the physiochemical properties of the components, and numerous biological activities involving the participation of the complement system. The involvement in numerous bacterial and viral diseases and many other clinical conditions was also studied. The use of sera selectively treated with various reagents such as cobra venom or ammonia together with euglobulin, pseudoglobulin, and heated serum in various combinations permitted the determination of the amount of each of the four components in an unknown sample of serum. Hegedus and Greiner (32), Bier et al. (33), and Ecker et al. (34) pioneered the use of such reagents and established criteria for their preparation and use. They were termed “R reagents” by Bier et al. (33) and this terminology was utilized until the onset of modern biochemical approaches to the study and analysis of complement in the 1960s. In this terminology, R1, R2, R3, and R4 are reagents for the titration or measurement of the concentrations of C1, C2, C3, and C4, respectively. Several crucial discoveries were made in the late 1950s and early 1960s. In 1958 Mayer (35) elaborated the one-hit theory of immune hemolysis. As discussed later, this furnished a quantitative approach to the study of complement action and reaction mechanisms. Two independent lines of investigation carried on independently by Becker (36-39) and by Lepow and colleagues (40-44) culminated with the discovery that the first complement component was a proenzyme which became activated during complement action. The first applications of chromatographic and other developing analytical approaches to the study of complement by Lepow and co-workers (45,46) and Miiller-Eberhard (47,48) permitted the identification and biochemical analyses of the proteins of the system. These discoveries provided the tools, technologies, and conceptual approaches which ushered in the phase of modern research on the complement system.

THE CLASSICAL C O M P L E M E N T PATHWAY

155

111. The Proteins of the C1 Activation Unit

A. C l q 1. Discovery and History

C l q was initially identified by its ability to precipitate soluble IgG aggregates. Earlier studies beginning with Heidelberger (49) had shown that fresh serum contributed protein to, and enhanced the precipitation of antigenantibody complexes in the quantitative precipitation reaction (50-52). This effect was abrogated by heating the serum at 56°C for 30 minutes (49,50,53) or by removal of divalent cations (54-56); precipitation of C l q also depleted complement activity from the serum (49,50,54). A serum factor possessing the ability to precipitate soluble IgG aggregates was independently identified by Muller-Eberhard and Kunkel (47) and by Taranta et al. (57) in 1961. The serum factor was heat-labile (56"C, 30 minutes), sedimented in gradients with a rate of 11 S, and was essential for complement activity. Since it was required for immune hemolysis (47,57) but reacted directly with soluble IgG aggregates in the absence of divalent cations, which were known to be required for the action of the the first complement component (58,59), it was presumed to act prior to C1. C l q was identified as one of the three constituent proteins of C1 by Lepow and colleagues in 1963 (45). Shortly thereafter this research group showed that the three constituent activities of C1, termed Clq, C l r , and C l s recombined with each other in free solution in the presence of calcium to form a macromolecule which possessed C1 activity (60). The reassembled C1 macromolecule sedimented in sucrose density gradients at approximately the same rate as C1 activity in unseparated serum (60). C l q from several other species, in so far as has been analyzed, has a similar chemical composition and bears a strong resemblance to human C l q in properties and morphology. This includes not only rabbit (61-63), bovine (69), porcine (65), rat (66), and horse (67) Clq, but also nonmammalian C l q from the frog (68,69). This protein has obviously been highly conserved in evolution. C l q is one of the most cationic proteins in human serum as it migrates as a slow y,-globulin. It is a euglobulin which precipitates with only slight reduction in ionic strength of the serum. The extremely cationic charge and low solubility of C l q at reduced ionic strength coupled with the ability to reversibly bind to aggregated IgG and to DNA and other polyanions have provided a number of approaches to the isolation of C l q (47,61,70-73). Currently C l q is generally purified in most laboratories by the method of Assimeh et al. (74) or a modification of that method (75-77) or by the techniques of Reid (78), Tenner et al. (79), or Ziccardi (SO). The Tenner procedure eliminates IgG and the C l q inhibitor from the preparation.

156

NEIL R. COOPER

2 . Composition and Properties C l q is one of the most cationic proteins in human serum with a slow y2 electrophoretic mobility (Table I). Values ranging between 70 (81) and 180 pg/ml have been reported for the concentration of C l q in human serum (81). Measurements in this laboratory by radial immunodiffusion with attention to various potential sources of error have yielded an average value of 70 pg/ml (81). A similar value, 63 pg/ml, was obtained by determining the hydroxyproline content of the euglobulin fraction of serum (82). C l q is a large structurally complex molecule. Human C l q has a molecular weight of 410,000 as determined by sedimentation equilibrium (6,83) or 393,000, calculated from S (S20,w = 11.1)and D ( D = 2.5 X lOW'cm 2/second) using a partial specific volume of 0.73 (83).The frictional ratio is 1.7 (83).The carbohydrate content as ascertained by various authors (61,83,84,85) has ranged from 7.4 to 9.8% by weight and averages 8.3% (5). Human C l q is TABLE I PHYSICOCIIEMICAL PHOPERTIES OF HUMAN Clq" Extinction coefficient

;rm) Sedimentation coefficient

6.82

(E

11.1

(SZ0.W)

Diffusion coefficient (D) Frictional ratio Carbohydrate content (by weight, average) Carbohydrate types and amounts Molecular weight (S and D ) Molecular weight (sedimentation equilibrium) Polypeptide chains and molecular weights (dissociated) Polypeptide chains, molecular weights, and ratios (SDSPAGE, reduced) Noncovalent subunits, molecular weights, and ratios (SDSPAGE) Polypeptide chain composition Electrophoretic mobility Concentration in serum a

2 x 107 cmVsecond 1.7 8.3% Glucosyl galactosyl disaccharides-5.6%. complex type carbohydrate-2.5% 393.000 410,000 A: 24,000 B: 23,000 c: 22,000 A: 27,500 B: 31,600 C: 34,000 in 1:1:1ratio A-B dimers: 69,000 C-C dimers: 54,000 in 2:l ratio 6A 6B 6C

+

+

Yz 70 pg/ml

Table assembled from data in references 5, 61, 73, 78, 81, 83-87, and 233.

THE CLASSICAL COMPLEMENT PATHWAY

157

approximately 8.3% carbohydrate by weight (5,61,73,84-86). Approximately 70% of the carbohydrate content (5.6%by weight of Clq)is accounted for by glucosylgalactosyl disaccharide units coupled to hydroxylysine residues. There are also a few galactose monosaccharides per Clq molecule (86).In addition, Clq contains glucosamine, mannose, fucose, and sialic acid (5,61,73,84,85) which together account for approximately 2.5% of the molecule per weight. These latter sugars are located exclusively in the noncollagenous portion of the Clq molecule and are of the complex type as determined by Mizuochi et al. (87).Six asparagine linked chains per Clq molecule were found. Clq has an extinction coefficient of 6.82 at 280 nm (61,88). It has also been reported as 2.9 x lo5 M - ' (89). Clq contains three types of polypeptide chains termed A, B, and C in order of increasing mobility on SDS-PAGE analyses performed under reducing conditions (6)(Table I). The three chains, each of which is about 200 amino acids in length, have molecular weights of 22,OO to 27,000as ascertained by sedimentation equilibrium or gel filtration or 27,500to 34,800as estimated by SDS-PAGE under reducing conditions (61,78,90). As the three chains are present in equiinolar amounts (61,90),Clq is composed of 6A, 6B, and 6C polypeptide chains. The A, B, and C polypeptide chains contain 12, 8, and 4% carbohydrate, respectively (61). There was initially considerable confusion as to the number of subunits and the arrangement of the three types of Clq polypeptide chains into subunits since SDS-PAGE studies performed in the absence of reducing agents, by different laboratories, yielded divergent results. Yonemasu and Stroud (91)and Heusser et al. (92)found two noncovalently linked subunits in a 2:l ratio of the larger (apparent MW 65,000)to the smaller (apparent MW 42,000)while Reid et al. (61)and Assimeh et al. (74)found only the larger. The confusion was resolved by Reid (6,90)who showed that the migration of the smaller but not the larger subunit was influenced markedly by the salt concentration of the sample; in high salt it aggregated and failed to enter the gel and thus was not visualized. The identification of the larger noncovalent subunit as composed of disulfide linked A and B polypeptide chains and the smaller as disulfide linked C-C polypeptide chains (90,91) together with the 2:l ratio of the A-B to C-C subunits and the equimolar concentration of the individual subunits indicates that the 6A, 6B, and 6C polypeptide chains of Clq are assembled into 6 A-B and 3 C-C d'imer subunits. The disulfide bonds linking together the polypeptide chains of Clq are located very close to the N-terminal end of the molecule. The A-B polypeptide chains are held together by a single disulfide bond linking adjacent cysteine residues situated four residues from the N-terminal end of each chain. The C-C polypeptide chains are also linked by a single disulfide bond also located four residues from the N-terminal end (90,93,94) (Fig. 1).

NEIL R. COOPER

158

I "Hinge" or

A8 Chain Chain C Chain

E L S

B

20

10

30

R A p 0 G K K*G E A G R P*G R R G R P*G L K*G E

q

NTGi

P'G

K D G V D G L P'G

P P'G

I

40

G E p*b A P*G I R m G I

T G P P*A I P'G I P G I P G T P G P N G 4 P*G T P'G I V G I P*G M P*G L P'GA

"Kink" Region

K*b E

K*G L P*-

E P b I a

1

q

- - - G L P+

G L A

tf- G

I R

mG E F

--

50

D Q

/G

- GP

FIG. 1. Partial primary amino acid sequence of the collagenous portion of Clq. Disulfide cross-links are shown. The asterisk denotes hydroxylated amino acids. The region of the "hinge" or "kink' as seen in electron micrographs is bracketed. Data are from (5,78,93,94,97).Single letter amino acid code is used: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N , Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; X, unknown; Y, Tyr.

The chemical composition of C l q is most unusual for a serum protein. As initially noted by Miiller-Eberhard (95,96), the molecule contains the hydroxylated amino acids hydroxylysine and hydroxyproline and a considerable amount of glycine. Similar findings were made shortly thereafter by other research groups (6,84). Quantitative studies have shown the molecule to contain 17 glycine, 2 hydroxylysine, and 5 hydroxyproline residues per 100 amino acids (6,73,84). The carbohydrate composition of C l q is also distinct as it largely consists of neutral hexoses including glucose (6,73) which are not present in other serum glycoproteins. The glucose and galactose are linked as a disaccharide to the hydroxyl group of hydroxylysine (73,84). The presence of high concentrations of glycine together with hydroxylysine, hydroxyproline, neutral hexoses, and hydroxylysine-linked glucosygalactosyl disaccharide residues suggested that C l q was collagen like (95) as these features are characteristic of collagen and basement membrane proteins. This impression was substantiated by the demonstration by Reid et al. (61) that the molecule was susceptible to partial degradation by collagenase. Prolonged collagenase treatment reduced the molecular weights of the A, B, and C polypeptide chains by approximately 40% indicating that each chain contained a collagen-like region(s) (61). This impression was confirmed by primary sequence determinations and numerous other studies by Reid and his colleagues which revealed that approximately 40% of each of the three polypeptide chains contained the repeating triplet Gly-X-Y where the X residue was often proline and the Y residue was frequently hydroxylysine or hydroxyproline (5,78,93,94,97) (Fig. 1).The repeating sequences begin with the ninth, sixth, and third residues from the N-terminal end of the A, B, and C polypeptide chains, respectively, and extend for approximately 80 residues. There are, however,

E

K*

Q

k?

T H E CLASSICAL COMPLEMENT PATHWAY

159

minor exceptions to the Gly-X-Y triplet sequence. These include the substitution of alanine for glycine in positions 9 of the B chain and 36 of the C chain and the insertion of an “extra” amino acid, threonine, between residues 38 and 39 of the A chain and an extra “triplet” between the 39th and 44th residues (5,78,93,94,97). Most of the hydroxylysine residues of each chain bear the glucosylgalactosyl disaccharide. Collagen fibrils contain three polypeptide chains, each of which is in a minor helix which associate with each other to form a major helix (98).The dimensions and structure of the associated fibrils and resulting helical collagen fiber are characteristic. Support for the presence of collagen triple helical structure in C l q came from electron micrographic studies which revealed not only that C l y exhibited a strikingly unusual structure, as further described below, but also that the six thin strands connecting the globular peripheral subunits to the central stalk were rectangular in shape, linear, and rarely bent (85,99-103). Furthermore, the diameter of the connecting strands (approximately 1.5 nm) was consistent with the cross sectional diameter of a collagen triple helix (98). These strands were also left intact (by electron microscopy) after pepsin digestion at pH 4.4 at low temperature whereas most proteins and the non-triple helical regions of collagen are resistant to such treatment (5,103,104). The length of the strands together with the stalk region (23 nm) were also consistent with a stretch of approximately 80 residues of triple helical structure (5, 6, 90). Further support for the presence of a triple helical structure came from circular dichroism studies which showed a positive band at 230 nm which disappeared after collagenese treatment (5,98). The presence of a positive band between 220 and 230 nm is strongly suggestive of collagen-like triple helical structure. The noncollagenous portion of the C l q polypeptide chains includes the 8, 5, and 2 N-terminal residues of the A, B, and C chains, respectively, and the 110 C-terminal residues of each of the chains. The three chains have been completely sequenced by Reid and colleagues (4,105). Although different in primary sequence, there are areas of marked homology between the chains (4,5,105).

3. Molecular Architecture The earliest electron micrographs taken by Svehag and colleagues (1OO,1O1), Polley (1O2), and Shelton et al. (99) showed that C l q possessed a most unusual structure. Subsequent electron microscopic studies have amply confirmed the distinctive structure and revealed additional fine details of the structure (85,103,106), C l q from a “side” view resembles a bunch of six tulips (4). Each of the globular heads or “flowers” are 7 nm long by 5

160

NEIL R. COOPER

nm wide and appear to be subdivided or to have a cleft at the end [dimensions are averages from Shelton et al. (99) and Knobel et al. (85)l. The 6 connecting strands or “stems” average 11.5nm in length and are 1.5 nm in width while the central portion representing the gathered “stems” has average dimensions of 11.2 nm in length by 4.5 nm in width. As noted above, the collagen-like portions of C l q are resistant to prolonged digestion with pepsin at pH 4.5 (94). Although the A, B, and C polypeptide chains are truncated by this treatment, the collagen-like portion of Clq, representing approximately 40% of the molecule by weight, remains intact and disulfide bonded (94); it also exhibits the positive band in the far UV range between 220 and 230 nm in circular dichroism studies exhibited by triple helical collagen structures and by C l q although the peak is increased in magnitude and shifted slightly (103). In the electron microscope, pepsin digested C l q lacks the peripheral globular heads. The remainder of the molecule including the connecting strands and central portion remains intact (103). The only apparent alteration in structure of this portion of C l q observed is a reduction in the average angle between adjacent strands from 41 to 23°C. The earlier described findings concerning the covalent and noncovalent structure of C l q and physicochemical composition and characteristics together with the electron microscopic studies led Reid and Porter to propose a model for the structure of C l q consistent with all findings (6,90) (Figs. 2 and 3). Earlier models of C l q structure (61,91,92) were not consistent with amino acid sequence and other studies. In the model, which is now generally accepted, each of the six heads, connecting strands, and a portion of the central subunit consists of A, B, and C polypeptide chains (6) which extend the entire length of the C l q molecule (90). Each single “flower” then consists of an A-B heterodimer noncovalently associated with a C polypeptide chain. The A and B polypeptide chains are linked by a single disulfide bond situated near the base of the central subunit or “stem” and the C polypeptide chain is disulfide bonded to the C polypeptide chain of an adjacent subunit. Clq thus consists of three disulfide bonded pairs of A, B, and C chain units. In the model the central subunit or “stalk’ at the N-terminal end of C l q consists of six A, B, C triplets lying parallel and in contact with each other (Fig. 3). The six A, B, C units diverge from each other and ultimately form the six connecting strands and become the heads visualized in the electron microscope at the C-terminal end of the molecule. The stalklike central portion, which measures 11.2 nm in length and 4.5 nm in diameter, together with the six connecting strands, each 11.5nm long and 1.5 nm in width, comprises the collagen-like portion of the molecule. Reid and Porter have suggested on the basis of the several lines of evidence, noted earlier, that the collagen-like portions of the A, B, and C polypeptide chains

THE CLASSICAL COMPLEMENT PATHWAY

161

s I

C

N

i+C

Collagen-like sequence

Non-collagen-like sequence

Collagenous portion Globular portion

FIG.2. Molecular model of C l q structure. Two subunits, each consisting of an A-€3 heterodimer and one C polypeptide chain are shown. The C polypeptide chains ofadjacent subunits are joined by a disulfide bond. Adapted from 6 and 90.

are each in a minor helix with each triplet of the three chains being arranged together into a major helix (6,90)in analogy to the structure of collagen (98). Using collagen measurements as a guide, arrangement of the entire 78 residue stretch of collagen-like sequence of the A, B, and C polypeptide chains into a triple helical structure could yield a fibril with a length of 22 nm. This

FIG.3. Model of C l q structure. Adapted from 6 and 90 with measurements from 85 and 99.

162

NEIL R. COOPER

is obviously in complete agreement with the length of the connecting strands together with the central stalk. The diameter of a collagen-like fibril (1.5nm) is also identical to that of the connecting strands, and six such A, B, C triplets or fibrils lying parallel to each other in hexagonal arrangement forming the stalk would yield a structure with the diameter (4.5 nm) observed for the central stalk of Clq. Each of the six A, B, C subunits has a “bend” or “kink’ approximately in the middle of the collagen-like portion where the connecting strands diverge from the central stalk (Figs. 1 and 3). Reid (97) and Porter and Reid (107) have pointed out that the Gly-X-Y triplet sequence of the C chain is interrupted at residue 36 where an alanine is substituted for a glycine. Such a replacement would interfere with triple helix formation as there is no room for an L-amino acid side chain in a collagen-like triple helical structure (98). This interruption in triple helical structure would occur approximately 10.5 nm from the N-terminus of each of the A, B, C triple helical strands, a value consistent with the position of the observed bend (11.2 nm from the Nterminus). In addition, two other features are likely involved in the formation of the bend. First, there is the insertion of a threonine between two adjacent Gly-X-Y-triplets in the A chain at position 39 and second, alignment of the sequences of the A, B, and C polypeptide chains to give maximum homology yields an “extra” Gly-X-Y triplet in the B chain between positions 43 and 45 (4,s). It is likely that there is some flexibility in the collagen-like strands of C l q at the “bend” which acts as a kind of hinge. Earlier studies suggested some flexibility (85,99,101) and, as observed by Brodsky-Doyle et al., there was a dramatic reduction in the angle between adjacent connecting strands from 41 to 23”, as visualized by electron microscopy after removal of the heads by treatment of C l q with pepsin (103). Measurements by low angle neutron scattering, a technique less susceptible to distortion than adherence to a grid as in electron microscopy revealed an average angle of at least 60” of the connecting strands with the axis of the central stalk (108). Schumaker et al. (106) have made a careful study of the angle of inclination of the connecting strands to an axis drawn through the central stalk-like portion of C l q . A preferred angle of 50” was observed but the range of angles found, 28 to 80°, led these authors to conclude that there exists limited flexibility about this point and that it does represent a “semi-flexible joint” (106). Two other techniques, intrinsic protein fluorescence and fluorescence secondary to binding of the fluorescent probe TNS (toluidinylnaphthalene-6-sulfonate) to apolar sites on C l q revealed considerable alterations as a consequence of changes in temperature well below the levels of denaturation or loss of C l q activity (109). Such low temperature thermally induced structural alterations are also indicative of flexibility of the molecule. Certain additional findings can be related to the model for C l q . Among

THE CLASSICAL COMPLEMENT PATHWAY

163

these is the observation by Heusser et al. (92) that the C polypeptide chain is virtually selectively labeled when the molecule is radiolabeled with iodine, a finding confirmed by others (79,85,110). Enzymatic digestion studies indicate that the tyrosine(s) which are labeled are in the head portion of the C l q molecule (79). The 6 complex-type polysaccharide units found on the C l q heads are most likely distributed equally with one asparagine-linked sugar chain per globular region. Radiolabeling studies confirm the localization to the globular heads and indicate that the complex type sugar chain is attached to the A polypeptide chain (68,79,87). This additional carbohydrate undoubtedly accounts for the more intense carbohydrate staining of the A chain on SDS-PAGE in comparison to B and C polypeptide chains. Other physiochemical properties of C l q include the ability to bind 3 bivalent calcium atoms per molecule with a K , of about 76 F M (111). Although Ca2+ is needed for integrity of C1 and for its activation, the function of calcium in relation to C l q is unknown. It may serve to stabilize the molecule as it has been found to facilitate reoxidation and restoration of activity after reduction (112) but other interpretations are possible. 4 . Domains, Interactions, and Functional Correlates As earlier noted, C l q was discovered by its ability to bind to soluble IgG aggregates (47,57). The role of C l q in serving as the recognition unit of C1 mediating binding of the macromolecule to immune complexes, IgG aggregates and various nonimmunoglobulin containing C1 binding substances has been well documented (2,3,45,113,114). Immunoglobulins exhibit specificity in their ability to bind C l q and also C1 (10,115-118) with IgG, being most efficient followed by IgG, and IgG,. IgG, interacts poorly (115)although its Fc fragment possesses this ability (119). IgM also binds C1 but IgA, IgD, and IgE do not (116,118). Monomeric immunoglobulins of the IgG and IgM classes also interact weakly with C l q as first shown by Muller-Eberhard and Calcott (117) and with C l (116). At full saturation the valence of C l q for IgG was found to be 5-6; subsequent comprehensive studies showed the valence could be a multiple of 6 (115). The valence of C l q for immunoglobulins at saturation clearly suggests that each of the 6 globular heads in Reid and Porter’s model possess a binding site for immunoglobulins. The earliest experiments supporting this interpretation come from the studies of Knobel et al. (120). In their studies, collagenase digested radiolabeled Clq, although devoid of functional hemolytic activity, had some ability to bind to immune complexes. The authors showed that a radiolabeled fragment of C l q with an approximate molecular weight of 43,000 retained the ability to bind. The authors correctly attributed this binding activity to the peripheral subunits although the model for the polypeptide chain composition of this subunit which they were utilizing was not

164

NEIL R. COOPER

correct. Prior (61) and subsequent (12,20) attempts to isolate C l q heads remaining after collagenase treatment were complicated by the insolubility of the digestion products. Paques et al. (122) carefully examined the digestion conditions necessary to obtain soluble products. The soluble heads had an S rate of 4.45 and a molecular weight of 57,000 (sedimentation equilibrium) and consisted of three chains, with apparent molecular weights under dissociating conditions but without reduction of 23,000, 17,000, and 16,000. Comparison by Porter and Reid (5) of the amino acid composition of the 57,000 product (122)with Reid’s values (123) suggests that the fragments are the C-terminal 110 residues of the A, B, and C chains along with a short stretch of the collagen-like sequence. The 57,000 fragment did not bind to the Fc fragment of IgG and did not interact with C l r and C l s to form C1 but it did inhibit binding of intact C l q to antibody coated erythrocytes (122). Hughes-Jones and Gardner (124) also examined the products and properties of collagenase digested C l q . Prolonged digestion with collagenase yielded a product with a molecular weight of 37,000 (gel filtration). On SDS-PAGE under dissociating conditions in the absence of a reducing agent two bands migrating 21,000 and 17,000 were observed; the 17,000 band was more diffuse and stained more intensely than the 21,000 band and thus may have contained two polypeptide chains. Amino acid composition studies confirmed the noncollagenous nature of the soluble product. The soluble 37,000 product inhibited the interaction of radiolabeled C l q with IgG containing immune complexes. The functional affinity constant describing the interaction of the 37,000 product with the IgG containing complex (1.8-5.8 x lo4 M - l) was comparable to that previously observed for the reaction of intact C l q with monomeric IgG, further supporting the identification of the 37,000 product as the IgG binding portion of C l q . C l q also interacts with certain other noniminunoglobulin C 1 binding substances through the globular heads. This includes DNA (125)and glutaraldehyde-treated erythrocytes (124). There is some evidence that the same sites are involved in these interactions with C l q as with the IgG interaction with C l q (110,125,126). The C l r and C l s subunits of C1 interact with C l q through sites located in the collagenous portion of the molecule. The first documentation of this came from the studies of Reid et al. (127). In these experiments, C l q was digested with pepsin at pH 4.5, a procedure previously demonstrated by Reid (94) and Brodsky-Doyle et al. (103) to yield the intact collagen-like portion of Clq. In this procedure the globular heads are degraded to small peptides. The collagen-like fragments remaining after peptic digestion remain intact and disulfide bonded (94,103). In SDS-PAGE analyses under nonreducing conditions, two bands representing the truncated disulfide linked A-B heterodimer and the truncated disulfide linked C-C dimer are seen (94). The apparent molecular weights of these two noncovalently associ-

T H E CLASSICAL COMPLEMENT PATHWAY

165

ated subunits were 35,800 and 28,500; after reduction, estimated molecular weights were 16,900, 16,900, and 15,000 (94); these contrast with the values of 20,000 and 18,200 for the A-B and C-C dimers and 11,400, 8,800, and 9,600 for the shortened A, B, and C chains respectively when the amino acid compositions are used to estimate molecular weights (94). The intact collagen-like fragment utilized by Heid et d . (127) competed with intact C l q in interacting with C l r and C l s although it was hemolytically inactive and, as earlier observed, did not bind to antibody (128). Electron microscopic studies of chemically cross-linked C1 by Strang et al. (75)and Poon et al. (129) indicate that the Clr,Cls, tetramer interacts with the connecting strands of the collagenous portion of C l q . These findings have been confirmed by direct studies by Siege1 and Schumaker (130) with the isolated collagen-like portion of C l q . Direct interaction of this portion of C l q with the ClrCls, tetramer was demonstrated employing an ultracentrifugal technique (130). The measured association constant, 2 x lo7 M - l , was similar to that describing the interaction of the Clr2Cls, tetramer with intact C l q (3.6 to 6.7 x lo7 M - l ) (130,131). The collagenous portion of the C l q molecule is also involved in the interaction of C l q with fibronectin (132) and platelets (133).Studies indicate that a specific C l q receptor present on human B lymphocytes, null cells, monocytes, and macrophages selectively recognizes the collagenous portion of the C l q molecule (134,135). This was ascertained not only by showing competition with collagen but also by demonstrating that C l q but not C1 triggers the receptor (136).

B. C l r

AND

Cls

1. Discovery and History Lepow and colleagues showed that the enzymatic activity they (40,41,44) and Becker (36-38) had previously identified with C1 was a property of the C l s subunit of the molecule (45,60). With the isolation of the activated form of C l s by Haines and Lepow (46,137), and its identification as a distinct protein able to inactivate C2 and C4 and to hydrolyze synthetic amino acid esters, it became clear that the activities of activated C1 were largely, if not entirely, synonomous with those of the C l s subunit. The isolation of the proenzyme form of C l s was accomplished independently by Okamura et al. (138), Sakai and Stroud (139), and Valet and Cooper (140). C l q was appreciated early to be the subunit of C1 which mediated the binding of C1 to immune complexes as noted earlier. Although C l r was shown to be essential for C1 action, its role as an activator of C l s was first demonstrated by Naff and Katnoff (141). Employing impure preparations, they clearly showed that C l r activated C l s and also possessed the ability to hydrolyze certain amino

166

NEIL H. COOPER

acid esters; they also described several inhibitors of C l r action. Activated C l r was first isolated in highly purified form in 1971 by deBracco and Stroud (142) and proenzyme C l r several years later by Gigli et al. (143), Sim and Porter (144), and Ziccardi and Cooper (145). With these isolations it became possible to describe C1 activation in biochemical and physicochemical terms. As first suggested by their virtually indistinguishable amino acid compositions (144,146), C l r and C l s are very similar proteins. They are of similar size and are activated by limited proteolytic cleavage into two disulfide linked fragments with the active site located in the smaller, C-terminal subunit. Each, in activated form, is a serine protease type enzyme. The impression that they arose by gene duplication is supported by recent sequence studies by Arlaud and Gagnon (147) which show that 22 of the 50 Nterminal, amino acids of the smaller, active site-containing, light chains of each are in identical positions (Fig. 4). As anticipated, this area and the primary amino acid sequence around the serine residue of the active site also possess homology with other serine proteases including bovine chymotryp-

Elastase ( p i g ) C h y m t r y p s i n A (cow) T r y p s i n (cow) Plasmin (human) Factor X (cow) Thrombin (cow)

1 10 20 30 40 V V G G T E A Q R N S W P S Q I S L Q Y R S G S S W A H T C G G T L I R Q N W V - FH FC G G S L I NE N W v I V N G E E A V P G S W P W Q V S L Q D K T G I V G G Y T C G A N T V P Y Q V S L N - - S G - - - Y H F C G G S L I N S Q W V

--

V V G G C V A H P H S W P W Q V S L R T R F G - - - M H F C G G T L I S F E W V

I V G G R D C A E G E C P H Q A L L V N E E N - - - E G F C G G T I L N E F Y V I V E G Q D A E V G L S P W Q V M L F R K S P Q - - E L L C G A S L I S D R W V

A

G S

wv

Conserved residues

I V G G

C l r b chain

I I G G Q K A K M G N F P W Q V F T N I H G R G - - - - - - G G A L L G D R W I

Cls b chain

I I G G S O A D I K N F P W Q V F F D N P W A - - - - - - - G G A L I N E Y W V

P W Q V S L

50

41

HFCGG

S G

60

L I

70

80

Elastase ( p i g ) C h y m t r y p s i n A (cow) T r y p s i n (cow) Plasmin (human) F a c t o r X (cow) Thrumbin (cow)

M V V L L L

Conserved residues

L T A A H C

C l r b chain

L T A A H - T L Y P F C A G H - - P S L K Q - O A C Q G D S G G V F A X X D P N

Cls b c h a i n

L T A A H - V V E G

T T S T T T

A A A A A A

A A A A A A

H C H C H C H C H C H C

V O G V Y K L E L H L L

R E V C A G - - - G N G V R - S G C Q G D S G G P L H T T I C A G - A S G V S S C M G D S G G P L V S G F C A G Y - - L E G G K - D S C Q G D S G G P V V K S L C A G H - - L A G G T - D S C Q G D S G G P L V Q A F C A G Y - - D T Q P E - D A C Q G D S G G P H V Y P F C A G Y K P G E G K R G D A C E G D S G G P F V

--

C A G Y

--

G G

D

- C L V C K K - C S G - C F E - - T R F M K S P Y

C Q G D S G G P L V

--

C

ACGKDSGEGR

FIG.4. Primary sequence of portions of the light chain of activated Clr, Cls, and homologous serine protease enzymes. Adapted from 147-152. Gaps left to maximize homology are denoted by (-). Residue numbering is from 147 and is arbitrary; it has been adjusted to maximize homology of all of the proteases listed. Conserved residues and invariant conserved residues are from 150. Residues are in the single letter code: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; X, unknown; Y, Tyr.

THE CLASSICAL COMPLEMENT PATHWAY

167

sin trypsin, thrombin and Factor X, porcine elastase, and human plasmin (Fig. 4) (147-151). Overall homology with these enzymes is approximately 25% (152). C l r and C l s were first isolated by ion exchange and gel filtration approaches. More recently, a number of affinity-based approaches utilizing the ability of C l r and C l s to form a reversible complex in the presence of calcium and to form C1 on addition of C l q have been used in combination with various chromatographic and precipitation based methods. Currently C l r and C l s are purified in most laboratories by one of the following methods: Arlaud et al. (153), Chapuis et al. (154), Gigli et al. (143), Siege1 et al. (155),or Ziccardi (SO). The same or similar methods can be used to purify the activated forms of C l r and C l s since most of their physicochemical properties are not changed by activation. However, prior to purification of the activated forms, C1 must be activated, usually by incubation with aggregated IgG or immune complexes. Furthermore, such activation must be carried out after an initial step to remove C 1 inhibitor. Since C l r and C l s were first isolated from human plasma, the human proteins have been most thoroughly characterized. Rabbit (156,209), guinea pig (208), and bovine C l r and C l s (157)have since been isolated. In so far as studied, the properties are quite similar to those of their human counterparts. A wide range of values has been reported for the concentrations of C l r and C l s in human serum. Deterininations by various groups einploying radial iminunodiffusion have yielded values as low as 34 kg/ml and as high as 110 pg/ml. In a study of the source of this variability, Ziccardi and Cooper (81) found that C1 in human serum has a tendency to activate if diffusion is carried out at room temperature rather than in the cold; furthermore, with activation, C1-inhibitor binds to both proteins producing aberrant diffusion patterns as discussed later. The determination must also be carried out in agarose as the supporting inediuin since agar has a tendency to activate C1. The third and largest source of variability is the tendency of the purified proteins, especially C l r , to aggregate with storage. Since aggregated proteins produce falsely small rings in radial immunodiffusion studies, their use as internal standards in such studies leads to the overestimation of the concentration of the proteins in human serum. In our experience (Sl), determinations of the concentrations of C l r and C l s in human serum yield values of 34 &ml for C l r and 31 kg/ml for C l s . As the molecular weights of the single chains of both proteins are very similar, the concentrations of C l r and C l s are equimolar in serum.

2. Composition and Properties C l r and C l s properties are shown in Table 11. The amino acid compositions of C l r and C l s are unremarkable for plasma proteins and as noted

168

NEIL R. COOPER

TABLE 11 PIIYSICOCHEMICAL PROPERTIES OF H U M A N Clr Clr Extinction coefficient (E::,,) Sedimentation coefficient IS %An.) Frictional ratio Carbohydrate content (by weight, average) Carbohydrate types

Molecular weight (S and D) Molecular weight (sedimentation equilibriiun) Molecular composition Molecular weight (monomer) (SDS-PAGE) Molecular weight (activated monomer) (SDS-PAGE) Electrophoretic mobility Concentration in serum a

ANV

Clsu

Cls

11.2- 11.7

9.3- 11.7

6.7

4.3

1.41 9.4%

1.46 7.1%

Mannose, galactose, N-acetylglucosamine, sialic acid 188.000 170,000 Noncovalent dimer 95,000 Disulfide linked 60,000 and 35,000 fragments

Mannose, galactose, N-acet ylglucosamine, sialic acid 86.200 85,000 Monomer, dinierizes in calcium 87,000 Disulfide linked fragments, 59,000 28,000 Da

P

(Y

34 kg/ml

31 pg/ml

Table assembled from data in references 145, 149, 153, 154, 161-165, 167, and 173-175.

above, are virtually identical. The molecules do not contain y-carboxyglutamic acid (149,158). The most comprehensive chemical evaluation of C l r and C l s is that of Sim et at. (149). C l r contains 9.4% carbohydrate by weight and Cls 7.1% which is composed of mannose, galactose, N-acetylglucosamine, and sialic acid. All are present in equal amounts except for a substantial additional increment of N-acetylglucosamine in the light chain portion of Clr. The carbohydrate is present in both the heavy and light chains of the activated forms of both molecules (7,149,159,160). Partial specific volumes, calculated from the amino acid and carbohydrate contents, are 0.714 and 0.717 cm3/g for CIr and Cls, respectively (149). The extinction coefficient for C l r (EiFm)has been reported as 11.3 (154), 11.7 (149), and 11.2 (161); it has also been reported as 9.6 X 104 M - ' cm-l (89). Kecent estimates of the extinction coefficient for Cls (Eiz,,,) cluster around 10.0, i.e, 9.3 (162), 9.5 (149), 11.6(161), and 11.7 (163);the extinction coefficient has also been reported as 7.9 x lo4 M - ' cm-' (89). Stokes radii of 5.0 nm

T H E CLASSICAL C O M P L E M E N T PATHWAY

169

for C l r and 4.1 nm for C l s and frictional ratios of 1.41 for C l r and 1.46 for Cls have been calculated (149). C l r sediments in the presence or absence of calcium with a sedimentation rate of 6.7 2 0.4s (S20,w)(164) a value similar to measurements of 7.0 S (145) 7.5 S (165), 7.9 S (164), 7.1 S (153) obtained by sucrose density gradient ultracentrifugation. Isolated proenzyme C I r has a molecular weight of 170,000 as determined by sedimentation equilibrium (164), a figure consistent with an earlier calculation from S and V of 188,000 (165), and values of 168,000 (142), 185,000 (165), 198,000 (145), 166,000 (149), and 200,000 (154), estimated by gel filtration (Table 11). However, SDS-PAGE analyses of proenzyme C l r performed under reducing or nonreducing conditions have given values of 83,000 to about 100,000 (145,149,153,154,167); gel filtration and sedimentation studies in dissociating solvents have provided comparable molecular weights (149,164). These data indicate that C l r is a noncovalent dimer. Sequence studies (149,151,152) have confirmed earlier impressions (145,149)that the two subunits forming the dimer are identical. The C l r dimer is not calcium dependent. It can, however, be reversibly dissociated by slightly reducing the pH to 5.0 (160). A lower pH (4.0) is required to dissociate the two subunits of activated C l r ; alkaline pH does not do so. Dissociation occurs in the presence or absence of calcium. Dissociation at acid pH has been utilized to derive an affinity based purification procedure for C l r (167). Takahashi et aE. (168), Sim and Porter (144), and Ziccardi and Cooper (159) independently showed that activated C l r is composed of two disulfidelinked fragments (Fig. 5). SDS-PAGE studies have yielded values for the molecular weights of the two major disulfide linked fragments of 60,000 and 35,000 (159,161), 58,000 and 36,000 (149), 68,000 and 47,000 (168), 57,000 and 35,000 (167). The sum of these values approximates the molecular weights of a C l r monomer subunit. Lack of release of other fragments has been confirmed by amino acid composition studies and more recently by primary sequence analyses. The larger chain of activated C l r has been termed the heavy, H, A or a chain and the smaller the light, L, B, or b chain by various investigators. C l r and activated C l r migrate in SDS-PAGE under nonreducing conditions with the same mobility which is slightly faster than the reduced proenzyme. Prolonged incubation of activated C l r at 37°C is accompanied by two additional successive proteolytic cleavages (153,169,170). As carefully studied by Arlaud et (11. (153), non-disulfide-linked 35,000 and 7,000-10,000 fragments are successively removed from the N-terminal end of the heavy chain. This leaves a truncated dimeric C l r molecule which retains antigenic activity and has a molecular weight of 110,000 and a sedimentation rate of 6.1 S . The secondary and tertiary cleavage reactions are, like the initial C l r

170

NEIL R. COOPER

FIG. 5. Activation of Clr. SDS-PAGE analyses under reducing conditions. The upper panels show the patterns obtained with radiolabeled Clr, labeled as the precursor molecule, and then activated; the lower panels show the corresponding stained gels.

activation reaction and as described further below, characterized by a high activation energy (36.8 kcal/mol) and a lack of concentration dependence; they are inhibited by calcium, NPGB (nitrophenylguanidinobenzoate),and DFP (diisopropylfluorophosphate) (153). It is therefore likely that these cleavages are mediated by C l r and not by a contaminating protease. The truncated dimeric activated C l r molecule contains the active site and is able to cleave and activate Cls; it is, however, unable to interact with Cls, to form a Clr,Cls, complex, in part because of its lack of calcium binding ability (153).The physiological meaning of this cleavage is not clear. Studies of the calcium binding ability of C l r by Villiers et al. (111)indicate that C l r binds one calcium atom per C l r dimer with a Kd of 17 pM while activated C l r binds 2.6 0.4 atoms of calcium per dimer with a K , of 33 pM. It is not known whether the calcium ions merely strengthen the Clr,CIs, interaction which does occur in the absence of calcium (160) or fulfill other functions. Calcium has effects on C l r solubility, especially after activation when calcium facilitates precipitation of Clr, thus it might be anticipated to influence C l r conformation. Calcium also retards C l r autoactivation (153,159) as discussed below, and activation of C l s by activated C l r (145,171,172). C l s sediments in EDTA-containing buffers with a sedimentation coefficient of 4.3 % 0.3 S (Sz0,J (164) which is similar to earlier values of 4.7 S (149), 4.3 S (140), and 4.5 S (140) obtained by sucrose density gradient ultracentrifugation. Isolated proenzyme Cls has a molecular weight of 85,000 as determined by sedimentation equilibrium in the presence of

*

THE CLASSICAL COMPLEMENT PATHWAY

171

EDTA (164), a value consistent with the earlier figure of 86,200 calculated from S and D (174) and with estimates of 83,000 (149) to as high as 110,000 (139) obtained by SDS-PAGE analyses performed under reducing conditions (Table 11). In more recent studies, values close to the molecular weight (83,000-87,000) have been obtained in SDS-PAGE analyses by most laboratories (149,173,175). Cls isolated in the presence of chelating agents is a single chain 87,000 protein as discussed above. Upon addition of calcium, however, Cls under- I goes dimerization as first demonstrated in sucrose density gradient and gel filtration studies by Valet and Cooper (174). The dimer is readily dissociated upon addition of EDTA. Although Sim et al. and Porter and Reid (4,5,149) failed to observe dimerization a number of other laboratories have documented and studied the dimerization phenomenon employing a number of different techniques (76,153,173,176). Barkas et al. (148), Takahashi et al. (177), Sakai and Stroud (178),and Valet and Cooper (140) independently showed that activated Cls is composed of two large disulfide-linked fragments (Fig. 6). The molecular weights of the two major disulfide linked fragments as determined by SDS-PAGE are 59,000 and 28,000 (175), 56,000 and 27,000 (149), 57,000 and 28,000 (173), 68,000 and 34,000 (177).The sum of these values approximates the molecular weight of proenzyme Cls and thus no large peptides are lost. This has been confirmed by amino acid composition studies and more recently by primary sequence determinations. The larger chain has been termed the heavy, H, A or a chain and the smaller the light, L, B or b chain by various investigators. Takahashi et al. (177) first showed that the activation cleavage site was in the

FIG. 6. Activation of Cls. SDS-PAGE analyses under reducing conditions. The upper panels show the patterns obtained with radiolabeled Cls, labeled as the precursor molecule, and then activated; the lower panels show the corresponding stained gels.

172

NEIL R. COOPER

C-terminal end of CIS. Proenzyine and activated C l s have the same mobility in SDS-PAGE analyses performed under nonreducing conditions and migrate slightly faster than the zymogen. Cls, like C l r , undergoes further hydrolytic cleavage (156,179,180). The additional cleavages in the heavy chain did not result in loss of enzymatic activity but impaired ability to interact with C l q and C l r to form C1 (180). The biological significance of these additional cleavages is not known. Studies of the calcium binding ability of C l s by Villiers et al. (111)showed that diineric proenzyme or activated C l s binds 2 divalent calcium atoms per dimer with a rC, of 27 to 40 p M . Other ions with comparable crystal ionic radii, i.e., Tb3+, Ba2+, and Sr2+ also bind well whereas Mg2+ and Mn2+ do not. Calcium does not have any major effect on the enzymatic activity of Cls; it probably strengthens the Clr,Cls, interaction. It could also potentially stabilize the C l s dimer.

3. Activation and Properties of the Activated Forms Ziccardi and Cooper (159)found that highly purified proenzyme C l r spontaneously fragmented into 60,000 and 35,000 fragments when incubated alone at 37°C in buffers containing EDTA. This process results in C l r activation as it occurs at approximately the same rate as C l r acquires the ability to cleave and thereby to activate C l s (159). Several lines of evidence favor this cleavage reaction as an intrinsic property of C l r reflecting the normal C l r activation process as represented during C1 activation and not the consequence of the action of a contaminating protease present in the purified preparations. First, purified uncleaved radiolabeled C l r incorporated into macromolecular C1 is also hydrolyzed into disulfide-linked fragments of the same size with activation, (156,159), and activated C l r isolated from serum following activation by immune complexes or aggregated IgG possesses the same chain structure. In contrast, C l r acted upon by proteases such as trypsin (159)and plasmin (163,181) is also readily cleaved but at site(s) either near the middle (trypsin, 159)or heavy chain ends (plasmin, 181)of the C l r molecule. Thus such trypsin-like proteases cleave C l r at different sites than the site characteristic of C l r activation. Second, C l r purified directly from serum either by direct adsorption to immobilized C l s attached to an affinity column and then eluted with EDTA (179) or by adsorption to C l s bound via anti-Cls to Sepharose and subsequently eluted at pH 5.0 (167) autoactivates in a comparable manner. Somewhat similarly, purified C l r mixed with purified C l q and C l s to form C1 and then reisolated from the 16 S C1 peak after sucrose density gradient ultracentrifugation also autoactivates (179). Contaminating proteases, unless firmly bound to C l r , would be separated by these procedures. Third, C l r purified by numerous other methods in multiple laboratories also exhibits the phenomenon of autoactivation, although to

THE CLASSICAL COMPLEMENT PATHWAY

173

different extents. (7,153,159,167,169,18%185).It is obviously unlikely that the multiple C l r purification methods would all yield C l r contaminated by the same proteases. Fourth, C l r autoactivation is independent of concentration over wide pH (7-13) and concentration ranges (50-350 pg/ml)(153,179,185) although some concentration dependence is noted below and also above these values (179,185). These properties are not characteristic of protease contamination which would be anticipated to have a pH optimum and to be bimolecular and thus quite concentration dependent. Fifth, spontaneous C l r autoactivation has been found by several groups to be retarded by concentrations of >10 mM D F P and by 0.5 to 5 mM calcium (145,153,167,179,184,185), polyanethol sulfonate (159), C1 inhibitor (159,167,186)and 1mM NPGB (153,179,184,185). Inhibition by these agents was found to be reversible and, in the case of NPGB and DFP, not associated with detectable binding to the proenzyme or activated molecules. These inhibitors, with the exception of calcium, also block C1 activation (171,184,187,188). It is unlikely that a contaminating protease would have the same inhibition spectrum; it is also improbable that DFP, NPGB, and C1 inhibitor would bind reversibly. Sixth, C l r autoactivation is blocked on forming ClrzClsz by addition of calcium and C l s (8,153,159,188). This property is also not likely to affect the activity of a contaminating protease, particularly in view of the susceptibility of Clr, in Clr,Cls2 to plasmin and other proteases (163,181). Although some C l r preparations may obviously contain proteases, the preponderance of evidence noted above favors C l r autoactivation as an intrinsic property of isolated C l r and not the consequence of the action of an unrelated contaminating protease. However, it is difficult to reconcile this hypothesis with the findings that C I r preparations isolated by different investigators apparently differ markedly in their relative susceptibility to autoactivation (7,143,184,185). Values from 10% or less autoactivation in 2 hours at 37°C in the presence of EDTA (184) to 50% autoactivation in 20 minutes (153) to 95% activation in 10 minutes at 37°C (159) have been reported. Other intermediate levels both with and without an initial lag period prior to activation have also been observed (153,184,185,189). Although there is a tendency for the more rapidly autoactivating C l r preparations to have been purified in the absence of protease inhibitors (145,169) and the relatively more stable C l r preparations in the constant presence of DFF, often together with other inhibitors (153,183,184), there are exceptions (167,185). Since C l r does undergo conformational rearrangement during activation and C1 activation, as further described below, we feel it likely that the difference in susceptibility to autoactivation of the various C l s isolates reflect different conformational states. Such conformationally different forms could possibly be produced by the purification procedures, for example, partial isolation as

174

NEIL R. COOPER

Clr,Cls,, presence or absence of EDTA, calcium, or enzyme inhibitors, binding to proteins on affinity resins, length of the procedure, etc. Differences among various Clr preparations consistent with protease contamination include the previously reported inability of C l r purified by the procedure of Ziccardi and Cooper (145) form C1. This however is due to the propensity of purified C l r to aggregate, especially in the presence of calcium (7,153,179), since most preparations in our hands do form C1. Other concerns, such as minor relative concentration differences in the ability of DFP to retard autoactivation (7,153,179,184,185), have other explanations as does the variable possible presence of contaminating enzymes in C l r preparations (185) and concentration dependence of autoactivation at very low C l r protein levels (185). Although activation of C l r by activated C l r molecules has been reported (167,183), most (153,159,189), but not all (194) recent studies fail to confirm such intermolecular activation (153,159,184). Furthermore, activated C l r acting on precursor C l r is not responsible for C l r autoactivation because activated C l r is readily detected by various techniques and also irreversibly binds DFP, NPGB, and C1 inhibitor (153,159,184). C l r is also activated by plasmin (163,181). Previous reports of activation by trypsin (165) were found to be erroneous (145). Each of the polypeptide chains of the C l r dimer is cleaved during the activation process, Each polypeptide chain of C l r contains an active site, thus the dimer contains two active sites. Although the mechanism of C l r activation is not clear, whether considered as activation within C1 or as autoactivation, it is likely that zymogen site in each zymogen polypeptide chain cleaves a susceptible bond in the other polypeptide chain. A number of zymogens have been reported to have intrinsic proteolytic activity which is augmented following cleavage (190,191). Approximation of the zymogen active site with the susceptible bond in the adjacent chain would initiate activation; this would occur in C1 with binding of C1 via C l q to activators. This mechanism for C l r activation was postulated by Ziccardi and Cooper (159). In this conception, C l r undergoing autoactivation would be in the appropriate conformation. This hypothesis implies that C1 has considerable flexibility and ability to alter its conformation. Consistent with this interpretation, conformational changes during C l r activation have been detected by several different experimental approaches. These include iodination of the surface of the protein with lactoperoxidase (167,179,192,193). Clr, labeled as the precursor and then autoactivated, is found to contain a moderate amount of radioactivity in the light chain portion of the molecule. In contrast, C l r labeled following activation contains little radioactivity in the light chain. The increased ability of the activated C l r molecule to bind calcium (2.6 atoms per mole) as compared to the proenzyme (one atom per

T H E CLASSICAL C O M P L E M E N T PATHWAY

175

mole) ( I l l ) , and changes in antigenicity (194)also are indicative of conformational changes. Similarly, autoactivation of C l r is associated with and exactly parallels an increase in intrinsic fluorescence intensity (194). Circular dichroism studies show changes in the far UV range (194). The conformational changes, probably in their evolutionary stage, are clearly documented by binding of a fluorescent hydrophobic probe (195). This probe detects rapidly occurring conformational changes in C l r which are completed prior to significant autoactivation (195). Activated C l r cleaves very few synthetic amino acid esters. Only acetylarginine methylester and acetyl-L-glycine methylester (170,171,196,197)and several peptide thioesters (161) are significantly hydrolyzed. N-Carbobenzoxy-L-tyrosine p-nitrophenylester has been reported to be a substrate ofactivated C l r (197)but this cleavage reaction has been found to be inhibited by some, but not other C l r inhibitors. The spectrum of esters which are cleaved is distinct from that of activated C l s and other similar enzymes (171,196). Kinetic studies and measurements of enzymatic characteristics and constants have been performed by several investigators (161,170,197). Activated C l r cleaves and activates Cls. It is inhibited in this action by PMSF (171,197), DFP (143,145,168,172,184), and NPGB (184), p-tosyl-Llysine chloromethylketone (197), C1 inhibitor (145,172,198), and calcium (143,145,171,172). N,N-Dimethylamino p-(p’panidino benzoyloxybenzylcarbonyloxy glucolate) (170), liquoid or polyanetholsulfonate (145,170- 172), pentosan polysulfate (172), and several amidines and guanidines (199,200) also inhibit; leupeptin has been reported to inhibit (170)but others have not found this (197). Clr, following activation, cleaves its natural substrate C l s (140,145, 171,172,177,178). C l s is activated by this cleavage. The enzymatic cleavage of Cls by activated C l r is rather unusual as the enzyme and substrate, C l r and Cls are normally present in C1 as a firm calcium-dependent protein-protein complex containing two polypeptide chains each of C l r and Cls, as more completely discussed below. C l s activation by C l r is second order and strongly temperature dependent (172).The Clr2Cls2tetramer remains intact following C l r and C l s activation under physiologic conditions until dissociated by C1 inhibitor (see below). There is, therefore, no turnover of the substrate, Cls, by C l r in the complex. In the fluid phase in the presence of EDTA, there is modest turnover ofCls by C l r . Activated C l r cleaves a single Arg-Ile or Lys-Ile bond in proenzyme C l s (7,149). Cleavage of C l s is highly temperature dependent with an activation energy of 23-32 kcalimol. (7). C l s was earlier reported to autoactivate (139,201). Later studies have shown that this is not the case and the molecule is quite stable to prolonged incubation at 37°C (143,172,173,178). Trypsin (140,202,203) and lysosomal enzymes (203) also activate C l s but

176

NEIL R. COOPER

further degrade the molecule into smaller fragments. Since these fragments retain certain properties of activated Cls, they will undoubtedly find use in structure-function studies. The ability to activate C1 in plasma was attributed to plasmin some years ago (43,202,204) although this remained controversial with other workers denying this possibility (205,206). In semipurified and purified systems also some workers have found Cls activation by plasmin (41,43,202,207)while others have failed to observe activation or have attributed this property to other contaminants (205,206,210). We have reinvestigated this question with highly purified reagents and found that plasmin degrades the C h monomer into small fragments and produces only slight activation in this process (163,181). The calcium-dependent dimer is resistant to plasmin cleavage and activation. Although C1 is activated by plasmin, this is via direct attack on the Clr subunit which, following activation, cleaves and activates Cls in the usual manner which is discussed later. Activated Cls, alone, or in the C1 molecule cleaves and thereby activates its two natural substrates C2 (40,137,211-213) and C4 (40,137,214,215). In addition to these two substrates, which are the only natural protein substrates, activated Cls cleaves a number of synthetic esters and amides. The esterase activity of activated C1, a property of the activated Cls subunit, was appreciated prior to the isolation of Cls, and its characterization as a constituent of C1; Cls was known as C1 esterase for a number of years. It was also appreciated early that activation was a prerequisite for the expression of enzymatic activity directed at either C2 and C4 or synthetic esters (38,40,43). The spectrum of amino acid esters and amides cleaved is unusual in that compounds possessing either basic (38,171)or aromatic (44,200)amino acids are cleaved. Among the arginine and lysine-containing esters readily cleaved by activated Cls are N-acetylglycyl-L-lysine methylester, N-acetyl-L-lysine methylester, benzoyl-L-arginine methylester, tosyl-L-arginine methyl ester, a-N-carbobenzyloxycarbonyl-L-lysine paranitrophenyl ester (38,44,46, 200,216). Particularly susceptible aromatic esters include N-acetyl-L-tyrosine ethylester, N-acetyl-L-tyrosine methylester, N-acetyl-L-phenylalanine ethylester, and a-N-carbobenzyloxycarbonyl-L-tyrosinep-nitrophenyl ester (38,44,46,200,216,217). N-Benzyloxycarbonyl-L-phenylalanyl-L-valyl-L-arginine p-nitroanilide is also cleaved (218) as are certain thioesters (161). The relative activity on these different compounds clearly separates activated Cls from other such enzymes such as trypsin, chymotrypsin, thrombin, and plasmin. Kinetic parameters of the cleavage of synthetic substrates by activated Cls have been analyzed and Michaelis constants (KJ, maximal velocity (Vma), catalytic constants, and other parameters determined for many substrates (46,149,173,200,219).These studies together with numerous inhibition approaches have permitted the design of active site inhibitors for activated Cls (162)and have facilitated initial studies of certain characteristics of

T H E CLASSICAL COMPLEMENT PATHWAY

177

the active site of the enzymes. The active site appears to contain a highly hydrophobic area (200,219)and an anionic binding site (200). Activated Cls is inhibited by DFP (137),and other phosphonate esters as well as by NPGB (184).One mole of DFP is bound per mole (87,000Da) of activated Cls (143,144,178). C1 inhibitor also blocks the fractional activity of activated Cls. Several anidines and guanidines inhibit (199,200).In addition, M-[o-(2[chloro-s-fluorosulfonylpheylureido)phenoxy butoxy] benzamidine (162)is a potent inhibitor.

4 . Domins, Interactions, and Functional Correlates The light chain of the activated Clr and Cls molecules contains the enzymatic site as shown by the binding of DFP to this portion of the molecule (50,143,145,146,148,168,177,178) (Fig. 7).The multispecific protease inhibitor, C1 inhibitor binds also to the light chain to this portion (145,220)ofeach molecule. The entire light chain of activated Clr has been sequenced by Arlaud and collaborators (147,151,152) confirming earlier partial sequences (149,168). The portion of the active site of Cls surrounding the active serine was determined some years ago (148);other portions of the light chain of Cls

"50 2,500

10,000

5,000

10

20

30

Segment Number

FIG. 7. Binding of DFP to the light chain of activated Clr, Cls, and trypsin. SDS-Page analyses under reducing conditions.

178

NEIL R . COOPER

have been sequenced (147,149,177). These studies allow, in the case of C l r and partially so with Cls, the localization within the molecules of the active site serine and other residues involved in the “charge-relay system” (Fig. 4). Both molecules lack the histidine loop disulfide bond common to other serine proteases (147,150). The functional consequences of this difference are not known. Both C l r and C l s normally exist as dimers under physiological conditions in the presence of calcium. The location of the intersubunit binding sites within C l r and C l s involved in formation of the dimers has been examined employing surface radiolabeling with lactoperoxidase. Since accessible tyrosine residues, which are labeled by this procedure, are present in comparable amounts in the heavy and light chains of both C l r and C l s (149,159), changes in labeling pattern are not explainable by differences in tyrosine content. It is reasonable to interpret such changes as an indication of exposure of the area to the solvent. However, since conformational changes may alter exposure of tyrosine residues at sites which are distant from the polypeptide chain or contact point being studied, such data should be interpreted with caution. Radiolabeling studies suggest that the contact sites facilitating dimerization of the Cls polypeptide chains reside in the heavy chain of each member of the dimer (160), it is likely also that the calcium binding sites are located in this portion (111). Similar studies with C l r suggest the two C l r monomers also interact via the heavy chain (160) while studies with partially hydrolyzed C l r also suggest that the calcium binding sites are present in this portion of C l r (153). C. C1 INHIBITOR (Cl-In)

1 . Discovery and History The ability of normal human serum to inhibit the enzymatic activity of C1, first noted by Ratnoff and Lepow (44), was investigated by Levy and Lepow and the active principle termed C1 esterase inhibitor (221). The protein was first isolated by Pensky et al. (222). C1-In was independently purified and characterized without knowledge of its complement function by Schultze et al. (223) and termed a,-neuraminoglycoprotein; some years later identity of this protein with C1-In was ascertained (224). Shortly after the first purification, C1 inhibitor was found to be lacking in patients with the disease termed hereditary angioneurotic edema by Donaldson and Evans (225). The previous year an inhibitor of kallikrein and a permeability factor had been shown to be lacking in this disease by Landerman et al. (226); the protein with these various functions is now known to be C1-In. Shortly thereafter, two forms of this disease were identified by Rosen et al. (227): in the more common, C1 inhibitor levels determined antigenically or functionally are

T H E CLASSICAL CO MP LEMENT PATHWAY

179

low (530% of normal) while in the second or variant form, normal or even elevated levels of C1-In protein are found but the protein is devoid of functional activity. The disease is inherited as an autosomal dominant trait (227,228) and has been extensively studied (227,229,230). The reduced levels of C1-In are responsible for the disease as shown not only by abrogation of attacks by inhibitor infusion (231,232) but also attacks are reduced and frequently eliminated by the administration of synthetic androgens, which increase the synthesis of normal C1-In protein (230,233,234). The mediation systems producing the disease, the permeability increasing factor, and the exact pathogenesis of the disease are not yet known. Complement has been implicated (235) as has the fibrinolytic system (236) and bradykinin (237). More study is needed. The protein was shown early to not only inhibit activated Cls, but also C l r (159,187), plasinin (187), kallikrein (187,226,238), activated plasma thromboplastin antecedent (XIa) (239), activated Hageman Factor (XIIa) (239), and fragments (240). It thus is a multispecific inhibitor of the enzymes of the plasma effector systems-the complement, coagulation, kinin-forming, and fibrinolytic systems. Its activity against these enzymes differs, however, and all of the proteases inhibited by C1-In except those of the complement system have other, more potent inhibitors in plasma; some also have the ability to degrade and thereby inactivate C1-In (220). It is nevertheless likely to be of major importance in regulating kallikrein, Factor XIa and Factor XIIa fragments (240-242). It is the only plasma inhibitor of activated C l r and C l s (243,244). C1-In does not inhibit trypsin (220,245) and is degraded by trypsin, plasmin and leukocyte elastase (220,246). Many methods have been evolved to purify C1-In (220,247-249). At the present time, most laboratories purify the protein by the method of Reboul and collaborators (249,250). Nonfunctional forms of the molecule isolated from the serum of patients with hereditary angioneurotic edema have been studied by several investigators including ourselves (227,251,252). It appears that minor structural changes lead to loss of activity and that different patients exhibit various alterations. No single alteration leading to a loss in activity has thus been identified. C1-In purified from guinea pig serum has been reported to have a molecular weight of 170,000 and to sediment with a 6.1 S rate and to also have functional properties quite different from its human counterpart (2,253). Rabbit C1-In (254) as well as C1-In from other species has been partially characterized (255). Different values have been reported for the concentration of C1-In in human plasma ranging to over 200 pg/ml. We have obtained a value of 137 5 11 pg/ml for the concentration of C1-In normal human sera by radial

180

NEIL R. COOPER

*

immunodiffusion and 128 10 pg by a newly developed immunochemical assay which measures C1-In function (256). 2 . Composition and Properties

C1-In contains somewhat low levels of glycine, and sulfur containing and aromatic amino acids (245,247). Primary sequence data are not available. It is one of the most heavily glycosylated proteins in serum as it is approximately 35% carbohydrate by weight (245) (Table 111). Of this 17% is sialic acid and 12% hexose (224,245). The structure of the carbohydrate is unknown but one study suggests that galactose is penultimate (257). The protein has an extinction coefficient (EiZm)at 280 nm of 4.50 (245). The partial specific volume is 0.667. C1-In sediments with a rate of 3.7 (245). It migrates as an 01 globulin. The molecular weight, as ascertained by sedimentation equilibrium, is 104,000 (245). SDS-PAGE analyses show that C1-In is composed of a single polypeptide chain and give approximately the same molecular weight value as that obtained by sedimentation equilibrium (220,249). In some purification schemes a minor second band with an apparent molecular weight of about 96,000 is seen (220). Since other methods do not yield C1-In with this second band (249), it is likely that it arises by a degradative process occurring during the procedure; and can be prevented by inclusion of protease inhibitors during purification (250). In this regard C1-In is susceptible to degradation by various enzymes as noted earlier. Although the 95,000 derivative form is functionally active, further degradative products are not (220). TABLE III PWSICOCAEMICAL PHWERTIES OF HUMAN C1 INHIBITOR^ Extinction coefficient

(EEJ Sedimentation coefficient (s20,w)

Carbohydrate content (by weight, average) Carbohydrate types and amounts Molecular composition and subunits Relative charge Concentration in serum

4.50 3.7

34% 17% sialic acid 12% hexose Single polypeptide chain a 137 pg/ml

Table assembled from data in references 220, 224, 245, 249, and 256.

THE CLASSICAL C O M P L E M E N T PATHWAY

181

3. lnhibition of Activated C l r and C l s C1-In binds to activated C l r and C l s forming an extremely stable complex which is not dissociated by SDS together with a reducing agent (159,220,258) or by other dissociating agents such as guanidine, urea, or low pH (250). The complexes do not undergo enzymatic degradation as is the case with a number of other enzyme-inhibitor complexes. The protease-inhibitor complexes contain one molecule of C1-In and one polypeptide chain of activated C l r or Cls. They thus have an apparent molecular weight by SDS-PAGE analysis in the absence of reducing agents which approximates the sum of the molecular weights of C1-In and one polypeptide chain of either activated C l r or C l s (159,220). Complex formation blocks the enzymatic activity of both activated C l r and C l s (159,198,220). Although the affinities of C1-In for activated C l r is (Kd = 1.2 x 10-7 M ) and activated C l s ( K , = 9.6 x M ) (259) are similar, C1-In reacts four times more slowly with C l r (159,259). The second order rate constants for the reaction of activated C l r and C l s with C1-In in EDTA were found to be k, = 2.8 x 103 M - 1 sec-l and k, = 1.2 x 104 M - l sec-l, respectively (259). The affinity of C1-In for activated C l r and C l s is not altered by the presence or absence of calcium. The speed of the reaction of the inhibitor with activated C l s is also not affected by EDTA; however, the reaction of C1-In with activated C l r is slowed two-fold by EDTA (k = 1.5 x lo3 M-' sec- l ) (259). The activation energy for the binding of activated C l r with C1In is also greater (44.3 kcal/niol) than for the activated Cls-C1-In interaction (11.7 kcal/mol) (258). In a noncomplexed mixtures in Ca2+ or EDTA of activated C l r and CIS, C1-In thus effectively inhibits only activated C l s (249). The reaction with either activated enzyme has a pH optimum of 7.58.2 (250). Inhibition of the enzymatic activity of activated C l s by C1-In is directly related to the concentration of activated Cls-C1-In complexes (249). Full inhibition corresponds to a 1:l molar ratio of the two proteins. Although the ability of activated C l r to activate Cls, a measure of protease activity, is effectively blocked by C1-In, the titration curve was found to be linear only at low ratios of activated C l r to C1-In (249). Full inhibition required a molar excess of C1-In over activated C l r (145,249). Sucrose density gradient ultracentrifugal studies revealed that the activated Clr-C1-In complex sediinented in EDTA-containing buffers with a rate of 6.5 S (260). Although the activated C l r dimer and isolated C1-In sediment with S rates of 7.1 and 4.3, respectively, by this technique, it is not possible from these data alone to determine whether the complex contains two activated C l r polypeptide chains and two C1-In molecules; i.e. (acti-

182

NEIL R. COOPER

vated C1r-C1-Inz) or one of each, although the S rates render the latter more likely, The complex is obviously extremely assymetric. Sucrose density gradient ultracentrifugal studies showed that the activated Cls-C1-In complex sedimented with a rate of 4.8 S in EDTA and 7.7 S in the presence of calcium (260). Activated C l s radiolabeled in the presence of EDTA, which is unable to form the calcium dependent C l s dimer, sedimented with a 6.1,s rate in the presence of C1-In (260). These findings suggest that the calcium-dependent C l s dimer remains associated following binding of C1-In. The complex formed between activated C l s and C1-In in the presence of calcium thus most likely has the composition: (activated C ls-C l-In)z. The light chain of activated C l r and C l s is involved in the interaction with C1-In as clearly shown by SDS-PAGE analyses of the complexes under reducing conditions. After reduction, the stained SDS-gels show complexes with molecular weights of approximately 135,000 and 130,000 for the activated Clr-C1-In and the activated Cls-C1-In complexes, respectively; in each case also a band with the molecular weight of the heavy chain of activated C l r or C l s is also visualized (159,220,243). Prior treatment of activated C l s with DFP blocks binding of C1-In (46,198,200,261). These data imply that C1-In binds at or close to the active site of the enzymes. The bond with C1-In is likely to be covalent. In studies of the nature of the interaction, sialic acid was not found to be important in C1-In activity but did affect in vivo clearance (257). Modification of lysine residues impairs cornplexing (248). The bond is also slowly disrupted at strongly alkaline p H (243,250). Most telling is the ability to hydroxylamine to dissociate complexes as shown by Sim et al. (243), a finding which suggests the involvement of an ester bond. Formation of an ester bond with the reactive serine in the active site is a possibility. Inhibition by C1-In is considerably enhanced by heparin (259,262-264). This is due, at least in part, to acceleration of the reaction of C1-In with activated C l s (259,263) and, to a lesser extent, with activated C l r (250). Since heparin alters the electrophoretic mobility of C1In, it is likely that it directly binds to C1-In and that this interaction is responsible for increased activity by an unknown mechanism (263). Heparin also has numerous other effects on the complement system including the C1 step (265, 266). Nevertheless, it appears that the major anticomplementary activity of heparin results from the potentiation of C1-In function since heparin did not have this property in serum in which the C1-In had been rendered functionally inactive by the presence of antibody to C1-In but reacquired it after addition of purified C1-In (263). 4 . Masking of the Antigenicity of Activated C1r and C l s In addition to the above noted effects of C1-In on activated C l r and CIS, binding of C1-In to both proteins alters the reactivity of the enzymes with

THE CLASSICAL COMPLEMENT PATHWAY

183

specific antibodies. This is particularly dramatic in the case of activated C l r . Most antisera to C l r fail to react with the activated protein, in striking contrast to their ready reactivity with the precursor protein as shown by Ziccardi and Cooper (267). Several lines of evidence clearly showed that the loss of anitgenicity was due to simple physical masking of the relevant antigens of activated C l r by the C1-In molecule. Thus, most of the antisera were directed to the light chain portion of the molecule. Others have since noted the same phenomenon (198). Although most antisera to C l s failed to detect changes with activation, certain antisera also possessed this property; furthermore, antisera to C l s could be readily rendered specific for the light chain portion of the activated molecule by appropriate absorption (267). C1 activation in human serum as a consequence of complement activation can be quantitated by assessing the disappearance of C l r by immunochemical approaches (268). This assay has found use clinically in measuring C1 activation in diseases (269,270). We have developed a somewhat different imniunochemical assay, also based on the loss of C l r reactivity with specific antibody after activation in the presence of Cl-In, and used it to quantitate C1-In function in human serum (256). This assay has proved to be a very useful simple assay to diagnose hereditary angioedema in sera from patients. Both forms of the disease (see below) as well as the acquired form of C1-In deficiency can be diagnosed by the assay.

IV. The Complexes of the C1 Activation Unit

A. THE Clr,Cls, COMPLEX 1 . Composition and Properties

Valet and Cooper (140,165) first demonstrated by sucrose gradient ultracentrifugation that proenzynie C l r and C l s form a stable calcium-dependent complex while Laurel1 and colleagues (271) showed the presence of C l r C l s complexes in human serum. Nagasawa and co-workers (272) and subsequently others (154,168) also observed that the affinity of C l r for C l s was higher than the affinity of either for C l q and evolved a purification procedure for the activated C l r C l s complex. The demonstration that the complex comprised of the zyinogen forms of C l r and C l s contained four polypeptide chains, two each of C l r and C l s followed soon thereafter (81,143). The composition of Clr2Cls, has been confirmed by others (76,160,273) employing diverse approaches. The sedimentation rate of the complex was initially estimated at 10 S (165); later studies indicate that it is 8.7 & 0.5 (160,176), Tschopp et al. (164) determined the molecular weight of the complex as to be 340,000 by sedimentation equilibrium, a value which is consistent with the postulated com-

184

NEIL H. COOPER

position. From the failure of the complex to dissociate at concentrations of 1 p M , these workers estimated the dissociation constant of the Clr,Cls, complex in the presence of calcium to be at least lo9 M - l . The Stokes radius of the complex is 84-90 nm (7). The substantial molecular weight together with the relatively small S rate indicates that the complex is assymetric. This is amply confirmed by the electron micrographs (165) as discussed in the next section. The activated C l s subunits also form a stable Clr,Cls, complex with the same physicmhemical properties. Arlaud and co-workers (160) have shown that a complex is formed between a Clr, dimer and a single monomer C l s molecule when the purified components are mixed together in the absence of calcium. This Clr,Cls complex sediments with a rate of 7.7 S in contrast to the 8.7-8.8 S rate of the tetrameric complex (160). The ability to form this artificial complex indicates that calcium-independent as well as calcium-dependent forces are involved in the formation of the Clr,Cls, complex. The Clr,Cls, complex binds 4.0 f 0.4 atoms of calcium per mole of protein with a Kd of 15 2 1 pM, a value greater by one atom than the sum of the calcium binding abilities of the constituent C l r and C l s dimers (111).Whether this additional calcium atom facilitates bridging between the subunits cannot be answered at present. The activated complex binds 5.0 -+ 0.4 calcium atoms/mol with a Kd of 32 t 2 pM, a value equal to the sum of the binding abilities of the individual activated dimeric subunits.

2. Molecular Architecture The first electron micrographs of the Clr,Cls, complex were those of Tschopp et uZ. (176). In their studies and in later work by Strang et al. (75), the complex is visualized as a linear chain of 6 to 8 globular domains having an approximate width of 3-4 nm and a contour length of 51 2 (176) to 59 (75) nm. In examining 100 images, 60% were found to have a reversed “S” or question-mark shape; only 1%had a normal “S” or “nonreversed S” configuration. The authors suggested, on the basis of the fact that the (Clr), dimer is more stable than the (Cls), dimer, that (Clr), forms a core to which a C l s monomer is attached at each end (176).

*

3. Domains, Interactions, and Functional Correlates Certain C l r and C l s interactions and properties are altered in the Clr,Cls, complex. For example, the Clr,Cls, complex does not spontaneously activate (153,159,188) unlike isolated C l r in most investigator’s hands. Also, the Clr,Cls, complex interacts with C l q in free solution to form C1(81,165) whereas neither C l r nor C l s alone binds to C l q in solution to any significant extent (140,165). The interaction, although weak with an association constant of 3.6 to 6.7 x lo7 M - l , is functionally significant as it

T H E CLASSICAL COMPLEMENT PATHWAY

185

allows C1 activation. The surface labeling pattern is also altered in the Clr,Cls, complex. Whereas C l r and C l s incorporate equivalent amounts of radioactivity if individually radiolabeled 70% of the label is incorporated into C l r and only 30% into C l s on labeling the Clr,Cls, complex, suggesting that C l r is more exposed (192). Other radiolabeling studies carried out by several groups (160,192,193) suggest that the heavy chains of both C l r and Cls are involved in the intersubunit interactions forming the Clr,Cls, complex. In the presence of immune complexes there are other changes. For example, C l r will bind strongly to C l q attached to immune complexes but not to C l q or immune complexes alone (145,184). Clr,Cls2 also binds more firmly to C l q when the latter is attached to an immune complex (127,131). The activated Clr,Cls, complex exhibits a somewhat different lactoperoxidase catalyzed surface activation pattern than the complex formed with the two proenzymes (160,192). The proportion of label incorporated into the larger chain of the activated proteins decreases and the proportion in the light chain increases (160,192). The changes are most pronounced with C l r . There are undoubtedly additional functional sites and/or domains in the Clr,Cls, complex. For example, it is likely that the activated Clr,Cls, complex has binding sites for C2 and C4 which stabilize the molecules and thus facilitate cleavage. Modulation of C2 and C4 cleavage by C1 constituents has been reported (274,275); this may be a reflection of some of these interactions. Also C1 inhibitor, which binds irreversibly to the light chain of activated C l r and Cls, probably interacts at several locations on the tnolecule. In addition, there may be other functional sites. Although earlier work suggested that C l r also bound directly to immune complexes (276) an interpretation later found to be erroneous (189), it is quite possible that the Clr,Cls, tetramer interacts with the activator as noted in the section on C1 activation. The nature and location of these various interaction sites remain to lie studied.

B. C1 I . Discovery and History As noted earlier, C1 was discovered in 1907 by Ferrata (23), Sachs and Teruuchi (24), and Brand (25). The earliest quantitative measurenients employed hemolytic techniques; initially pseudoglobulin was used as the source of the other components. Beginning around 1940, the pseudoglobulin was supplemented with heated serum to titrate C1; this reagent was known as R1 (33). Although extensively used for many years for titrations and measurements, R1, and “R” reagents in general suffer from both practical and theoretical limitations. For example, they lack complete specificity as C1 is not

186

NEIL 13. COOPER

always the component in lowest titer in R1, an obvious requirement for specificity as the immune hemolytic reaction is limited by the component in lowest functional concentration. Such reagents are also generally lacking in specificity. Finally, and most important, they represent functional detecting reagents and do not permit accurate kinetic analyses or the expression of results in molecular terms. These various problems were overcome by two types of developments in the late 1950s and early 1960s which were first, a mathematical formulation of complement action and second, the first applications of modern biochemical approaches and technology to the study of complement action, particularly the action of C1. Mayer formulated a theory of Complement action in 1961 termed the “one-hit theory of immune hemolysis” (35). In the theory, cornplement-induced hemolysis is viewed as a noncumulative process in which a single molecule of at least one of the complement components at some stage of the reaction sequence suffices for production of a lesion in the erythrocyte membrane which in turn leads to lysis of the cell. The theory was shown early to apply for C1 and C2 (277,278) and subsequently for other components (279). Application of the Poisson distribution function provides a method for quantitation based on the one hit theory since the number of hits or lesions per cell is represented by the negative natural logarithm of the proportion of lysed cells, i e . , In (1-y). A -In (l-y) value of unity indicates that a successful reaction of an average of one molecule of a given component per cell has occurred. This provides the means to calculate the number of molecules of a complement component in absolute terms. However, as molecules of components used in nonhemolytically fruitful reactions, i. e., turned-over and inactivated by a complement enzyme in solution or bound at inappropriate places on the cell surface, do not register in such titrations, minimal estimates are obtained (280,281). For this reason results are expressed as “effective molecules” (35).Nevertheless, such measurements permitted the analysis and quantitation of C l action in molecular terms and greatly aided the interpretation of kinetic studies and the study of reaction mechanisms. The initial biochemical studies of C1 included those from Becker’s laboratory (36-39,58) and from Lepow and colleagues, initially with Pillemer (4044). They indicated first, that C1 could be eluted and reattached to immune complexes; second, that C1 was a proenzyme which was converted to an active enzyme in the course of complement action; third, that the activated C1 enzyme was able to cleave synthetic substrates and inactivate C2 and C4; and fourth, that the activated enzyme was inhibited by DFP. Numerous other studies around the same time dealt with analyses of the physicochemical properties of C1 and its metal requirement. These studies together with the demonstration that C1 consisted of three activities, Clq, Clr, and CIS, separable by ion exchange chromatography by Lepow et al. (45) and NaE et

THE CLASSICAL COMPLEMENT PATHWAY

187

uZ. (60) permitted, and initiated the use of modern biochemical approaches for the analysis of C1. Over the next 20 years, numerous studies of the properties of the individual C1 components were carried out as described in the preceding sections. However, except for functional measurements of C1 hemolytic activity, few studies of the nature of macromolecular C1 have been conducted, largely because of the lack of approaches to study the properties of the intact native molecule. Studies of C1 were also impeded by the tendency of C1 to dissociate with only slight dilution of serum. This problem has made purification of the C1 macromolecule impossible. Published methods to do so, while extremely useful, yield either impure C1(282,283) or a mixture of C l q with the Clr,Cls, complex (143,284). Most studies of C1 properties and characteristics have thus used C1 reconstituted from mixtures of purified Clq, Clr, and C l s (45,74,175) or C l q and Clr,Cls, (75,176,273). In so far as can be ascertained, i. e., sedimentation characteristics, molecular and molar composition, functional activity, activation characteristics, etc. reconstituted C l is functionally and physicochemically indistinguishable from native C1 in serum (81,89,174,175,273). In addition to C l q , C l r , and Cls, a fourth component of C1, termed C l t , was claimed to be present in C 1 preparations prepared by affinity chromatography on Sepharose IgG columns (74). The protein appeared to interact with C1 components and enhance certain C1 related functions (276,285). The presence of C l t , since identified as the P component of amyloid (286,287), in C1 fractions was adventitious and a function of the ability of this pentameric molecule to interact in a calcium-dependent manner with IgG and with the Sepharose (81,143,288,289). Native C1 as found in plasma or serum is a calcium-dependent complex of three distinct proteins as considered earlier. The macromolecular C 1 complex can be demonstrated in normal human serum by the Ouchterlony technique as shown by Ziccardi and Cooper (290). A continuous precipitin line, indicating immunologic identity, is observed on reacting human serum with antibody to the three C1 components, a finding which indicates that the C l q , C l r , and C l s antigens are on a single entity, the C1 complex in normal serum (Fig. 8) (290). C1 dissociation following removal of calcium by addition of EDTA can be demonstrated by the same technique (Fig. 8). Simple modifications and the use of radial iminunodiffusion permit this observation to be used for quantitation of C1 in human serum by comparison with a normal serum pool as well as for quantitation of free C1 subunits in the presence of macromolecular C1 (290). Since virtually all of the C l q , C l r , and C l s in serum is in C1, estimates of the C I concentration in serum have been obtained by a summing together the concentrations of the individual subunits, as determined by radial immu-

188

NEIL R . COOPER

FIG. 8. Demonstration of macromolecular C1 in human serum by the Ouchterlony technique. Antibodies to Clq, C l r and C l s are in the outside wells and normal human serum is in the center. The macromolecular complex is evident when the assay is performed in the presence of calcium (A) since a continuous line of immunologic identity is seen with the different antisera. Dissociation of the complex occurs in the presence of EDTA (B) as indicated by the crossing precipitin lines.

nodiffusion. As noted in the respective sections, different values have been obtained for the concentrations of C l q , C l r , and C l s in human serum. In this laboratory, average values of 70, 34, and 31 pg/ml have been obtained for C l q , C l r , a n d C l s , respectively, in normal sera yielding a C1 concentraM (277). tion of 135 pg/ml (81). This corresponds to 1.8 x C1 activity is present in the serum of many species including not only mammals (255), but amphibia (68), primitive fish (291), and birds (292). Few studies of the physicochemical characteristics of C1 from these sources have been performed.

2. Composition and Properties C1 is a macromolecular complex of three distinct proteins: Clq, C l r , and Cls. However, C1 can be viewed as two weakly interactingproteins, C l q and Clr,Cls,, under physiologic conditions in free solution in undiluted serum since it dissociates into these subunits with slight dilution (81,130,273,293295) or on increasing the ionic strength (81,295). The association constant for the Clr,Cls, complex is approximately lo9 M-I but only about 2 to 8 x lo7 A4-l for the reaction C l q Clr2Cls, S C l q Clr,Cls, (89,130,131,295); calculations from activity measurements on native C l in serum yield4.5 x lo8 M-l(294). Since the concentration of C1 in human serum is 1.8 x 107 M-1 (81),it can be calculated from the association constant that approximately 70 to 90% of the C1 in serum is in the functionally active macromolecular C 1 complex and the remainder is free C l q and the Clr,Cls, complex. Activated

+

T H E CLASSICAL COMPLEMENT PATHWAY

189

C1 in free solution has a 10-fold lower association constant (130,296) in free solution indicative of a lower affinity complex; thus activated C1 is largely dissociated under physiologic conditions in serum. This most likely accounts in part for the presence of free subunits in pathological sera (297) and with activation o f C l in serum (268). Despite the firm association between Clr, and Cls, in the C 1 complex or in Clr,Cls,, some exchange between C l s in the complex and free C l s does exist under physiologic conditions since a proportion of labeled C l s added to serum slowly finds its way into C1 (298). Ultracentrifugal studies have been utilized to determine the sedimentation rate of C1. Estimates of the sedimentation rate of C1 ranged from 15 to 19 S (45,60,276,285,293,299,300).In our studies employing sucrose density gradient ultracentrifugation, we found that C1 sediments with a rate of 16 S whether determined for C1 in whole serum or for equimolar mixtures of purified C l q , C l r , and C l s (81,174). Later studies employing analytical ultracentrifugation have confirmed this value and indicate that the S20,wfor C1 is 15.9 (273) to 16.3 S (176). Studies carried out to determine the reasons for this variability (81,273) showed that the S rate of C l varied with experimental conditions, notably alterations in ionic strength and serum dilution as has also been observed by others (293,299). Artifactual values ranging from 12 to 23 S are due to (1)aggregation of C 1 which occurs on mixing very high concentrations of the reactants together, or (2) the use of a disproportionate excess of C l q over Ch,Cls,, or (3)carrying out the sedimentation analyses at low ionic strength. Reduced values result from C1 dissociation which occurs when the concentrations of C l q and Clr,Cls, are below physiologic levels. Combinations of these effects also occur. The molar composition of the C1 complex has been ascertained by determining the concentrations of the three C 1 subunits, C l q , C l r , and C l s in serum, in the 16 S peak of sucrose density gradients performed on serum, and in C1 reconstituted from combinations ofpurified C l q , C l r , and C l s by radial immunodiffusion. In such studies we found that C l q , C l r , and C l s were present in a 1:2:2 molar ratio when molecular weights of 410,000, 95,000, and 87,000 were utilized for C l q and the single polypeptide chains of C l r and C l s , respectively (81);the same ratio was obtained in euglobulin by others (143). Although C l r and C l s form a firm complex in solution as noted earlier, neither C l r nor C l s interacts significantly with C l q in solution (81,165,174); C l q interacts effectively only with the Clr,Cls, complex. These data together with the observation, by double immunodiffusion studies, that virtually all of the C l q , C l r , and C l s in serum or in reconstituted C 1 was in C 1 (Fig. 8) with only minor concentrations of free C1 subunits (81,297) indicate that C 1 is composed of one molecule of C l q and two polypeptide chains each of C l r and Cls. Although a multiple thereof is possible, the 16 S sedimentation rate makes this unlikely. Thus the composi-

190

NEIL R . COOPER

tion of C1 is C l q , Clr,Cls, (81).The molecular weight of such a complex, obtained from the sum of the molecular weights of the constituent proteins, is 774,000. This composition has been confirmed by various approaches (176,273). The molecular weight, determined by sedimentation equilibrium, is 739,000 (273). C1 requires calcium for its action and for the integrity of its structure as known for many years (43,45,54). The Clr,Cls2 complex binds four calcium atoms with a Kd of 15 pM and C l q three atoms with a Kd of 76 pM as earlier noted (111). Whether these calcium ions are involved in maintaining the integrity of the C1 is not known. In this regard, only the Clr2Cls, complex is calcium dependent (111,165,176). Studies with purified C1 reconstituted from equimolar mixtures of the components by Ziccardi (301) showed that other metals could replace calcium in C1. All divalent cations from the first transition period of the periodic table (Ca2+, Mu2+, Co2+, Ni2+, and Zu2+) together with Cd2+ and Tb3+ could mediate the formation of functionally active 16 S C1 which sedimented at 16 S. Mg2+ and Ba2+ were not active. The active metals exhibited funtional affinity constants of about 5.5 x lo4 M - l , a value similar to that obtained for calcium binding to Clr,Cls2 (6.7 x lo4 M + ) (111). However, since the calcium concentration in serum far exceeds that of any other active metal ion, it is probably the only physiologic cofactor (301).

3. Molecular Architecture and Proposed Model for the Structure of C1 Although excellent electron micrographs of C1q and clear pictures of the Clr,Cls, complex were obtained some years ago, native C1 has not been visualized in the electron microscope, apparently because of dissociation under the condition used to prepare samples for study. Strang et al. (75) and subsequently Poon et al. (129) in Schumaker’s laboratory circumvented this problem by using a water soluble carbodimide to cross-link C l q and Clr,Cls, in C1. Although exhibiting some SDS stable inter and intrasubunit cross-links, the cross-linked C1 reacted with antisera to C l q , C l r , and Cls; furthermore there was no change in the normal 16 S sedimentation rate although minor proportions sedimented faster and slower. These findings allay somewhat concerns that the basic structure of C1 has been altered by cross-linking. The cross-linked C1 failed to dissociate on addition of EDTA and also remained intact under the conditions used to prepare the grids for electron microscopy. In electron micrographs of cross-linked C I , the central stalk of C l q up to and including the “bend,” “kink,” or angle was well visualized and not altered in apparent structure or dimensions (75,129). The globular heads were also clearly seen. However, there was a poorly defined extra mass located between the central bundle or “stalk” and the globular heads along

T H E CLASSICAL COMPLEMENT PATHWAY

191

the diverging arms of the C l q molecule. Although these workers observed the same reversed “S” configuration described by Tschopp et al. (176) for the Clr, Cls, complex when equimolar mixtures of C l r and C l s were examined in the absence of C l q , this structure was not seen in images of C1 indicative that the Clr,Cls, complex had assumed a more compact structure. Analyses of projections of lateral views showed the additional mass to be located roughly in the center of the arms; views from the “end” of C l q molecules indicated that the extra mass was also centrally located with the globular heads projecting beyond. Strang et al. constructed a potential model for C1 consistent with the electron micrographs (75). In the model, the Clr,Cls, “string” of “beads” or domains was wrapped around and between the diverging arms of the C l q molecule. A subsequent refinement of this model by Poon et al. (129) envisages the middle portion of the linear Clr,Cls, complex as passing completely through the cone or cage formed by the diverging arms of C l q with the two ends of the complex emerging between opposite arms of C l q and wrapping back around the outside of the C l q arms. This model is consistent with the electron microscopic appearance and dimensions of C1, C l q , and the linear 50-59 nm long Clr,Cls, complex. It also addresses the concerns of symmetry involved in attaching the string-like reversed “S”-shaped Clr,Cls, subunit which has 2-fold symmetry, to C l q which exhibits %fold symmetry. Each identical half of the “reversed S”-shaped Clr,Cls, tetramer, defined with regard to a rotational axis passing through the center of the string like molecule, would have to possess the same number of sites for binding to C l q . The problems of maintaining symmetry come from a need to have these binding sites on the two symmetrical halves of the Clr,Cls, complex oriented or pointed in the same direction in order for them to be able to attach to complementary sites on the C l q arms. The model proposed by Poon et al. (129) fits these considerations of symmetry; a model in which the Clr,Cls, complex wraps around the outside would not because the identical C l q binding sites on each half of the symmetrical Clr,Cls, complex would then point in opposite directions. The model of Poon et al., although compatible with the electron micrographs and consistent with the size and shape of C l q and Clr,Cls, as well as the concerns related to symmetry, has two major problems, as they also noted (129). These are, first, the ready reversibility of the C l q Clr,Cls, interaction indicated by the affinity constant of only about 5 X lo7 M - ’ (130,131,295). It is difficult to envisage how Clr,Cls, could readily dissociate from C l q if the complex passed between the C l q arms through the center of the molecule. Second, C1 inhibitor rapidly binds to activated C l r and C l s and rapidly dissociates the activated molecules from C l q . Furthermore, it is the interaction with C l r (261) which initiates the dissociation of ClrCls-C1-In complexes from

+

192

NEIL R . COOPER

C1 (198,259,261,302). Not only access of CI-In to activated C l r but also dissociation of the complex would be difficult in the proposed model. In addition, further limitation of C1-In access and even more impaired dissociation would be anticipated if C l q were also bound to the activator via the heads. However, access to C1-In to activated C l r is in fact, greatly increased in C1 bound to immune complexes (259,261). This is not compatible with the model of Poon et al. (129). Also, surface radiolabeling studies employing lactoperoxidase yield a comparable distribution and proportion of radioactivity in C l r and C l s regardless of the presence or absence of C l q or the binding of C1 to immune complexes (192). This would not be likely to occur if the model of Poon et al. (129) were correct. Furthermore, if the linear or string-like structure of the Clr,Cls, tetramer visualized in the electron microscope and also evident on hydrodynamic measurements (176) is the structure of the native Clr,Cls, complex, it is most difficult to imagine how the two C l r chains activate each other and even more difficult to envisage how the same active sites gain access to two C l s subunits in the complex. Finally, the string-like Clr,Cls, structure, if native, does not allow for the calcium-dependent Cls, dimer, a structure generally found by most workers (7,174,176). A model of C1 in which the Clr,Cls, subunit is wrapped around the outside of the C l q arms allows unimpeded dissociation of Clr,Cls, from C l q and permits free access of Clr,Cls, to C1 inhibitor and to C2 and C4. Poon et al. (129)were deterred from postulating this model for the reasons of symmetry noted above. The model for C1 structure for C1 proposed in Fig. 9 places the Clr,Cls, subunits on the outside of the C l q arms; it also

c1

Clq

FIG. 9. Proposed model for C1 and its dissociation products.

Clr,Clsp

T H E CLASSICAL C O M P L E M E N T PATHWAY

193

I

FIC. 10. Proposed model for the native ClrpClsp complex as found in C1. Rotation around the axis shown and some “stretching” would produce the “reversed S”-shaped ClrzClsz structure found in hydrodynamic and electron micrographic studies.

addresses the concerns of symmetry. In the model, two identical “C”shaped C l r C l s subunits, each containing one polypeptide chain of C l r and one of Cls, are stacked together vertically and wrapped around the C l q arms. Such a C l r C l s complex would have a contour length of 25-39 nm. This length is sufficiently long to wrap around the diverging C l q arms. The model proposed in Fig. 9 has the further advantage of placing the C l r and Cl s subunits in contact with each other, an important consideration because of the mutually interdependent C l r and CIS activation events; the model also allows for not only a CIr, diiner but also the calcium-dependent CIS, dimer. It also would allow for binding of Clr,Cls, to the activator, if this occurs. This model, like that of Poon et al. (129), places the ClrCls, complex on the arms where movement of the C l q arms would readily be transmitted to Clr,Cls,. Both models are compatible with the electron micrographs. In the proposed model, there is the assumption that the string-like Clr,CIs2 molecule studied in hydrodynamic (176) and electron micrographic approaches (75,129,176) is not the native form of the molecule as found in C1 but is an altered form which is only found after the complex has dissociated from C1. The model proposes that the identical “C”-shaped C l r C l s subunits rotate about a vertical “hinge-like” bond located between the two noncovalent C l r subunits as shown in Fig. 10. Two “C”-shaped subunits, if

194

NEIL R. COOPER

opened by rotation around an axis at one end and slightly “stretched out,” can produce an elongated string-like structure with a “reversed S” shape (Fig. 10). Various approaches to test the proposed model are possible. V. The C1 Activation Process

A. C1 COMPLEXING AGENTSA N D ACTIVATORS C l q and C1 are bound and activated by immune complexes or aggregates containing IgG or IgM but not by those containing IgA, IgD, or IgE (3,5,115,116,303). Among IgG subclasses IgG, is most reactive followed by IgG, and IgG,; intact IgG, is minimally reactive although its Fc region binds C1 (304). Within IgG, the C1 binding site has been localized within the Cy2 domain of the Fc portion of the molecule (305)and the Cp4 portion of the Fc region of IgM (306,307). Further localization to the last two (C-terminal) pstrands of the Cy2 domain has been postulated on the basis of several lines of evidence (308,309). This area does not correspond to the region of IgG earlier thought to be the C l q binding region. Synthetic peptides duplicating the primary sequence in this region inhibit the Clq-IgG interaction (338,339). Other studies suggest the presence of C l g binding regions in other areas of the IgG molecule (340). In addition, C1 is bound and activated by an amazing variety of other substances (1,2,113,310,311). These include lipids, certain bacteria, viruses, parasites, mycoplasma, transformed cells, and subcellular membranes, and other structures as well as several proteins, carbohydrates, lipids, polyions, and other substances (Table IV). Many of these activate more efficiently than immunoglobulins. There are no obvious, nor is there likely to be, biochemical or structural features shared by all of these diverse substances. This is further discussed below. Bacteria thus far found to activate are certain Klebsiella, SaZmonella, and E . coli strains (311-315), while C1 activating viruses include all retroviruses thus far tested (316-320). Parasite structure binding and possibly activating C1 include antigens from immature Schistosomes (321) and Trypansoma brucei (322). Mycoplasma pneumoniae has also been found to bind C1 (323). Certain transformed cells (113), cytoskeletal intermediate filaments (324), some mitochondria1 membranes (325,326), as well as C-reactive protein in complex with phosphorylcholine (327, 328), p15E envelope protein of retroviruses (329), myelin (330,331), some polysaccharides (332,333), lipid A of lipopolysaccharides (333,334), polyanions like heparin (265,335,336), dextran sulfate (303,337), polyvinyl sufate (265,337), and polyanethol sulfonate (265,337), polynucleotides (265,341) including DNA (342), nitrophenylated molecules (343), polyglutanic acid (344), polylysine (344), monosodium urate

THE CLASSICAL COMPLEMENT PATHWAY

195

TABLE IV SUBSTANCESTHATBIND A N D ACTIVATE C1 Bacteria

E . coli strains Klebsiella strains Salmonella straiirs M ycoplasma Mycoplasma pneumoniae Viruses Retroviruses Parasite structures from Schistosomo mansoni Trypanosotnu brucei Proteins Immunoglobulins CRP (phosphorylcholine) complexes p15E retroviral surface protein Myelin basic protein Carbohydrates Ant venom polysaccharide Certain di- and trisaccharides Dextran sulfate Lipids Lipid A Polyions Heparin Polyvinyl sulfate Polyanethol sulfonate Polynucleotides Miscellaneous Monosodium urate crystals Mitochondria1 membranes Certain cellular membranes Nitrophenylated molecules

crystals (345) all bind Clq or C1 and most activate C1 (Table IV). It is highly likely that numerous other substances will be found which bind and activate

c1. B. C1 ACTIVATIONREQUIREMENTSAND

THE

ACTIVATION PROCESS

C1 activation by the large number of substances noted above requires integrity of the C1 macromolecule. Binding to C1, as investigated for most of the nonimmune C 1 activators, is via the C l q subunit. Although C1 activation by either immunoglobulin-containing or nonimmune C1 activators requires binding of C1 to the activator, probably with farily high affinity, C1 binding is not synonomous with C1 activation. Borsos

196

NEIL H. COOPEH

et ul. (346) first showed some years ago that C I binding and C1 activation were two separate events with binding preceding activation but not invariably leading to it. Binding without activation has been demonstrated in numerous experimental situations by immunoglobulin-containing (347-350) and nonimmune (334,345,351-353) C1 binders and activators. C 1 binding is nevertheless a requirement for C1 activation. The strength of C l q and thus C I binding increases, as would be anticipated, as the number of contacts between a single C l q or C 1 molecule and the activator increases. This has been examined in multiple systems utilizing a number of techniques. In the case of immunoglobulin activators, C l q or C 1 binding affinity increases markedly with the size of the complex, i.e, dimer 2 trimer 2 tetramer, etc. (115,116,354-358). Measurements of affinity binding constants by Hughes-Jones (358)ranged up to 10x M - for the binding of C l q to heavily aggregated IgG. Measurements of the binding constants for IgG dimers have yielded values around 9 x lo5 (19,354) and for IgG trimers around 9 X lo6 (19,354). Most IgG-containing immune complexes as well as aggregated IgG gave values around 2 x lo8 M - ' (344,358). It was initially thought that a monomer of IgM but a dimer or larger oligomer of IgG was needed to bind and activate C l (354,359-362). However, monomeric IgG binds C l q and C 1 (110,116,174,344,355) and activates C 1 (355,356,363) although poorly, i. e., 10- to 100-fold less well than the dimer (356). Several workers have estimated the functional affinity constant describing the binding of C l q to monomeric IgG at around 2.5 x lo4 M - l (115,358,363). Tschopp et ul. (356) have utilized a series of chemically cross-linked IgG polymers to examine the role of antigen in C 1 binding and activation. The IgG molecules were cross-linked such that the antigen combining sites remained accessible to antigen. In their studies, the binding affinity and C 1 activation were not affected by the presence or absence of antigen. Occupancy of antigen binding sites also did not markedly influence C l q binding to immune complexes (364). These data suggest that occupancy of the antigen binding site is unimportant and imply that antigen serves to aid in the formation of IgG clusters which in turn increase the binding affinity. This is the associative model for complement activation (366). These data also suggest that the conformational changes in immunoglobulins secondary to antigen binding, which have been found in some systems (365), are not sufficient to induce complement activation. Other work by Jaton et al. (357), who also observed conformational changes in the Fc region, also failed to show correlations with complement binding. These various studies therefore do not support the allosteric theory of complement activation (366) which postulates that antigen binding triggers conformational change in the Fc region leading to enhanced binding and activation. Obviously the role of antigen becomes a moot point when one considers

THE CLASSICAL COMPLEMENT PATH WAY

197

C1 binding and activation by noiiiininunoglobulin-containing substances. Although a number of these substances bind C1 poorly and some do not activate, many bind C1 more avidly than aggregated IgG and some activate C1 more efficiently. Most and probably all of the substances noted earlier are inultimeric in structure and thus provide the possibility of multiple binding sites. Hughes-Jones and Gardner deterinined that Clq binding increased with size of various polyionic substances including polyglutamine and polylysine and dextran sulfate (344).Large polyions, over IOO,OOO, gave values of approximately lo8 M - l, comparable to immune complexes (344). Glutaraldehyde treated erythrocytes also gave values of lox M - l (358). There is no question from the literature that aggregation of IgG and the use of polymeric C1 binding and activating substances as noted above increase the affinity of C1 binding. C1 binding is an obvious prerequisite for normal C1 activation. There is some confusion, however, as to whether C1 binding affinity directly correlates with extent of activation and the activation rate. Tschopp et al. (356)have found a direct correlation for monomeric, dimeric, and trimeric IgG. Doekes et al. (355),although observing also that activation as well as binding increased with the extent of IgG aggregation, failed to find that these events were well correlated. Others have also failed to find a direct correlation between C1 binding and activation using immunoglobulin and noniiiiinunoglobulin activators (349,353),although these workers have not used polymers of defined sizes. In terms of activation rate, Tschopp has also found that activation rate correlates directly with C l q binding affinity (363).Folkerd et al. (353)have also found that the rate of activation does not correlate with the C1 binding affinity. Although it is thus not clear whether the rate and exent of C1 activation correlate with the binding affinity of C1 for the activator it is evident that high binding affinity alone is not sufficient for, or completely responsible for activation. This is further supported by the finding of a number of avid Clq and C1 binding substances which fail to activate. Other factors are clearly involved. Although a number of aspects of C1 activation are not yet understood, other considerations relating to activation involve C1 dissociation, the regulatory role of C1 inhibitor and other influences which are considered below. C1 activation kinetics have been found by several investigators to be sigmoidal(89,184,363,367). The same kinetics have been observed for either Clr or Cls activation in C1 (184).The lag phase is rendered very short or may not be evident when high concentration of aggregates or highly crosslinked aggregated IgG are present (89,184,363). Whether this explains the failure of others (295)to observe a lag phase is not known. Regardless of whether first or second order kinetics prevail, C1 activation is a multistep reaction sequence. Tschopp (363)postulated a C1 activation mechanism

198

NEIL R. COOPER

involving two intermediate steps. This hypothesis, supported and extended by Kilcherr et al. (89), generally fits the activation kinetics observed under various experimental conditions and concentrations of the reactants. The first step in the model is the rapid reversible binding of C1 via the C l q heads to the activator. In this model, no minimum number of sites must be occupied, a feature which allows for C1 activation by monomeric IgG and some nonimmune activators which bind poorly. Activators with additional attachment sites for C l q would activate more rapidly. Binding, in Tschopp’s model, is followed by a slower phase in which rearrangement of the C l r dimer occurs so as to permit autolytic C l r activation between the two adjacent C l r protomers. The third phase is the rapid activation of C l s by the newly activated C l r . This model is valid if the rate of association-dissociation between C l q and Clr,Cls, and any other changes or events are fast compared to the rate of the activation step(s) (89).This is possible as activation involves steps anticipated to be slow, i. e., binding, association-dissociation of C l q and Clr,Cls,, movement within C l q secondary to binding to the activator, etc. In addition, some stability of the binding of Clr,Cls, to C l q is probably necessary to permit these steps to occur; this may be a time consuming process. It is also unlikely that the C l r internal rearrangements leading to autoactivation occur automatically once initiated; more likely continued stable binding to C l q is necessary until C l r is fully activated. In this regard, the strength of the interaction between C l q and the Clr,Cls, subunit increases with binding of C1 to activators as first noted by Reid et al. (127). Hughes-Jones and Gorick (131)showed that the association constant describing the interaction between C l q and Clr,Cls, increased from 3.6 X lo7 M - ’ in the fluid phase to 3.6 x lo8 M - I for C1 bound to immune complexes. A similar conclusion can be obtained by comparing values for fluid phase C1 obtained by Siege1 and Schumaker (130) and Kilcherr et al. (89), which are 6.7 x 107 and 2 x 107 M - l, respectively, to Ziccardi and Tschopp’s (294)value of 4.5 x lo8 M - for immune complex bound C1. In another study, Ziccardi obtained similar results as he observed a fluid phase association constant of 8 x lo7 M - which increased to 2.6 X lo8 M - after binding to an immune complex, reflecting the increased strength of interaction (295). The rate constant describing the activation reaction also increased 6- to 7-fold for C1 bound to an immune complex indicating that C1 activation occurs much more rapidly when immune complex bound (295). Ziccardi’s recent studies (186,188,295) indicate that the strength of association between C l q and Clr,Cls, is a primary factor in determining whether activation occurs. He has observed (186,188), as have other workers, that C l reformed from purified components (131,175,284,363,367) as well as partially purified C1 (370) activates spontaneously. In examining the mechanism of spontaneous C1 activation, he observed that increasing the

THE CLASSICAL COMPLEMENT PATHWAY

199

strength of association between C l q and Clr,Cls, in the fluid phase by decreasing the ionic strength (295) led to an increased rate of spontaneous activation. The increase in the C1 association constant describing the Clq CIr,CIs, interaction exactly paralleled the increased kinetics of spontaneous C1 autoactivation as measured by an increase in the rate constant (295). In face, at half physiologic ionic strength, the association constant in the fluid phase equalled that for C I bound to an immune complex at normal ionic strength and rates of C1 activation were comparable (295). Ziccardi concluded from these data that a major role of immune complexes in C 1 activation is to increase the strength of association between C l q and Clr,Cls, so as to allow internal “spontaneous” activation to proceed. He views C1 activation as conforming to the model:

In this conception, the increased rate of C1 activation with reduction in ionic strength or with C1 binding to an immune complex is viewed as the result of a change in the k,lk, ratio due to a decrease in k, and not to a change in the intrinsic activation rate or k,. This hypothesis of C1 activation (295), which primarily emphasizes the role of association-dissociation and those of Dodds et aE. (1841, Tschopp (363), and Kilcherr et al. (89) which emphasize intramolecular events are not mutually exclusive. Neither represents a full explanation of the C1 activation process. The mechanism responsible for the increased strength of association between activator bound C l q and Clr,Cls, in immune complex induced activation as reflected in the binding constant and activation rate constants is not known. Several groups have postulated that the Clr,Cls, complex may form additional bonds with Clq, or with the activator (89,131,363,368). Such interactions have not been found in studies with isolated C I subunits with two exceptions: the retroviral activation system in which C l s as well as C l q bind to the p15E envelope protein of these viruses (208) and C1 activation by a univalent hapten antibody complexes (369). With other activators labile bonds could be formed between CIr,Cls, in intact C I and the activator which disrupt with dissociation from the activator. The possibility of additional attachment sites has not been adequately investigated. Alternatively multiple weak interactions may occur. Further study is clearly indicated. The mechanism of C l r activation in the C1 molecule is not entirely understood. It is most likely, however, similar to that postulated by Ziccardi and Cooper (159). In this concept, an active site located in the zymogen form of each of the two identical polypeptide chains of the C l r dimer cleaves a susceptible bond in the other member of the dimer. An alternative explanation is that the zymogen active site cleaves the same monomer rather than the adjacent monomer. In either interpretation, the cleavage could only take

200

N E I L R. COOPER

place after molecular rearrangement and alignment of the susceptible cleavage site with the zymogen active site. This could be induced, for example, by binding of C1 to an immune complex. Zymogens of a number of proteases have intrinsic enzymatic activity which is potentiated after conformational changes or peptide bond cleavage (190,191). Other workers have since postulated the same or a very similar mechanism for C l r activation in C1 (2,4,153,184). It is likely that the C l r zymogen form with proteolytic activity is inhibited by NPGB (184). This corresponds to the Clr* of Dodds et al. (184). Keversible inhibition of activation of zymogens by inhibitors of the active site of the activated form, as in this case for C l r inhibition by NPGB, DFP, and C1 inhibitor, is consistent with findings with other zymogens having intrinsic proteolytic activity (371,372). Regardless of the precise mechanism involved, during C1 activation each of the two polypeptide chains of C l r in a C1 molecule is cleaved by an active site in the zymogen of each polypeptide chain. These C l r active sites then cleave each of the two polypeptide chains of C l s in the same C1 molecule. This clearly must involve rearrangements within the C1 molecule in order for the active site of zymogen C l r to first come into proximity to an uncleaved C l r polypeptide chain and subsequently, after C l r cleavage and activation, to be adjacent to the susceptible peptide bond in Cls. Understanding of the molecular processes involved requires further study. C1 activation, to be ultimately comprehended in molecular terms, requires an understanding of the types of interactions between the individual components and probably also the activator. The biochemical events occurring during and responsible for the activation events must also be fully described and integrated with physicochemical and other approaches. Much further study is clearly necessary to achieve these aims. C. REGULATIONOF C1 ACTIVATION In normal human serum, C1 is stable and does not activate spontaneously. However, as noted in the previous section, Ziccardi (188)has found that C1 reconstituted from purified complement components activates spontaneously by an intramolecular autocatalytic mechanism. He also observed that normal human serum blocks C 1 autoactivation by C1 reconstituted from mixtures of purified Clq, C l r , and C l s (188). The principle in normal human serum which blocks spontaneous activation of purified C1 and prevents spontaneous activation of C 1 in normal serum was identified as C1-In (186). Thus, in the presence of purified C1-In, spontaneous cleavage of C l r or C l s in purified reconstituted C1 does not occur. In addition to spontaneous C1 activation, physiologic concentrations of C1-In blocked C1 activation by several weak, nonimmune C1 activators including DNA and

THE CLASSICAL C O M P L E M E N T PATHWAY

20 1

heparin. Although C1-In does not block C1 activation induced by immune complexes at 37°C in most workers’ hands (186,284), Folkerd et al. (153) have reported it to have this property. In Ziccardi’s hands, CL-In also inhibits C1 activation by immune complexes when the reaction is slowed by reducing the temperature to 20°C (186). Such inhibition of immune complex induced activation was not accompanied by the SDS-stable firin binding which is characteristic of the interaction of C1-In with activated C l r and Cls. Only weak reversible binding of C1-In to the immune complexes could be demonstrated. As activation was blocked, cleavage of C l r or C l s also did not occur. C1-In apparently interacts weakly with unproteolyzed, nonactivated C1. It is likely that C1-In binds reversibly to a traiisitional or intermediate forin of C1 which is not yet fully activated and cleaved (186). Such a forin would be akin to, and possibly identical with the transitional or active zymogen forin of C l r postulated b y Dodds et al. (184) to explain the reversible inhibition of the activation of immune complex bound C 1 by NPGB. The ability of C1-In to prevent C1 activation by weak activators may be physiologically significant as suggested by Ziccardi (8,186) since this would prevent low level, nonspecific or undesired C 1 activation. This may well be of considerable importance considering the large number of substances which bind and activate C1. Prevention, by physiologic levels of C1-In, of the inherent tendency of C1 to autoactivate is likely to be an important homeostatic function. This finding may have relevance for the understanding of hereditary angioedema. C1-In concentrations greater than 0.35 times serum levels were able to block spontaneous activation of physiologic concentrations of C1; concentrations of 0.25 times serum levels or less failed to do so (186). This furnishes a possible explanation for the finding that sera of patients with symptomatic hereditary angioedema have less than 0.3 times physiologic concentrations of C1-In (227). Only moderate increases in circulating C1-In levels induced by androgen treatment prevents attacks in previously symptomatic patients (373).This finding together with the dernonstration of the phenomenon of spontaneous C1 activation and the inhibitory effects of C1-In on the process (186,188) suggested to Ziccardi (8) the possibility that increased C1 autoactivation resulting from the reduced levels of C1-In may underlie the marked classical pathway activation exhibited by such patients. As noted above, C1-In does not efficiently regulate C1 activation induced by immune complexes. It thus appears that immune complex dependent activation is not under positive host regulation. This is apparently largely true. However, C1 activation by small immune complexes or complexes formed with nonavid antibody or with ratios of antigen to antibody far from equivalence, all of which are poor C 1 activators, may well be regulated by C1-In. Further study in this area is needed.

202

NEIL R . COOPER

VI. Actions of Activated C1

Activated C1 and activated C l s in C1 inactivate the hemolytic activity of C1 and C4 (43). Characterization of this phenomenon on a molecular basis revealed that activated C l s is an enzyme which cleaves C2 (374,375) and C4 (214) as noted earlier. Both C2 and C4 are thereby activated and the union of the larger cleavage fragment of each molecule into a protein-protein complex, C4b,2a, in free solution or on the surface of activators is facilitated (214). This complex is a proteolytic enzyme which has the ability to cleave and activate C3; it thus represents the next functional unit of the classical pathway. Later studies with highly purified C2 (375-378) and C4 (379,380) have provided detailed data concerning the structural changes and enzymatic parameters of these activation reactions. An additional important facet is the requirement for C4 for efficient cleavage of C2 by activated C l s when contained within the activated C1 macromolecule (275,376,381). For example, Strunk and Colten’s studies (381)of the mechanism of enhancement of C2 cleavage by C4, earlier demonstrated by Gigli and Austen (275), indicated that C4 provided a site for C3 deposition. More recently, Thielens et al. (376) made a careful study of the enzymatic constants and other characteristics employing highly purified proteins. The K , , V,,,, and turnover number for C2 cleavage by activated C l s were all markedly and progressively reduced when C l s was incorporated into the Clr,Cls, complex and into C1. Under physiologic conditions, little cleavage of C2 would thus occur in the absence of C4b (376). C4 cleavage, in contrast, is equivalent regardless of whether activated C l s is free or in activated C1 and neither the K,,, for C4 cleavage nor the C4 turnover numbers are altered as shown by Ziccardi (382). Many earlier studies of C2 and C4 activation by activated C1 were carried out under nonphysiologic conditions and thus may not accurately reflect the situation prevailing during C1 activation in serum. Several different processes occasionally having opposite effects may confuse the issue. For example, under even slightly dilute conditions, C1 dissociates into C l q and the Clr,Cls, complex as considered in detail earlier; activation under such conditions has different requirements and does not reflect the physiologic events. For example, C1 activation under physiologic conditions is complete in 3 minutes (382); in diluted systems it is much slower (346); furthermore turnover of Clr,Cls, by antibody-bound C l q is also amplified in the case of dissociated C1. Another example is the omission of C1-In in earlier studies employing purified components; CI-In not only regulates C1 activation under certain conditions as considered in the previous section but also eficiently regulates the action of activated C1 on C4 and C2. The action of C1In is also dilution sensitive. Under physiologic conditions it occurs in less

T H E CLASSICAL COMPLEMENT PATHWAY

203

than 1 minute as described next, in dilute systems it is extremely inefficient in this property (383). In a study of C1 activation by several different immune complexes, aggregated IgG and nonimmune C1 activators in undiluted human serum and in mixtures of purified C1 and C1-In at physiologic concentrations of all reactants, Ziccardi found (382) that C1-In very rapidly and effectively controlled the actions of activated C1. Following addition of limiting amounts of the various activators to serum or the purified component mixtures composed of C1, C1-In, C2, and C4, binding of C1-In was very rapid; 59%of the newly activated C1 was complexed by C1-In within 13 seconds and 90% in 55 seconds. Because of this, activated C1 turned over few C2 and C4 molecules, on an average of 4 C2 and 35 C4 molecules per C1 molecule activated by any of the immune activators. Despite the low numbers of C2 and C4 molecules which are activated under such physiologic conditions, the components are efficiently utilized as the C4,2 enzyme is formed (382). However, successful triggering of the classical pathway at the C1 step by an immune or nonimmune activator does not automatically lead to assembly of the C4,2 enzyme and progression of classical pathway activation. Many aspects of the mechanisms involved in classical pathway activation and its regulation remain to be studied and understood. Among these are the repeated observations that not all C1 activation events lead to formation of the C3 converting enzyme. A major factor in this failure is the lack of C2 activation. Thus, some activators or activation events efficiently activate C4 but cleave and activate much less C2. For example, the efficient C1 activating lipopolysaccharide from Salmonella minnesota, H595, triggers the same amount of C4 activation per activated C1 molecule as a number of immunological activators, but this is coupled with little C2 activation (382). Certain immunological activators also exhibit the same general phenomenon; thus some immune complexes are more efficient in mediating C2 activation than others (382) and there are differences in regard to the same property between aggregated IgG, aggregated Fc, fragments and immune complexes (384,385). However, the failure to form the C4,2 enzyme is not always at the C2 step; C1 activation by certain substances is not followed by effective C4 activation (382,384). In part such results with C4 may be a reflection of the ability of C4 to not only bind to multiple surfaces but also to the Fab portion of antibody molecules as shown by Goers and Porter (385). C4b bound to antibody is probably more efficient in accepting activated C2 because of the very limited life-span of activated C2 and the short distance it would have to diffuse from its site of cleavage by activated C1 situated on the Fc portion of the same antibody molecule (385).This alone, however, is not likely to be the entire explanation for the variable activation of C2 and formation of functionally

204

N E I L H. COOPER

effective C4,2 complexes. Another factor of unknown relative importance is the observation that C1 elutes following activation from some but not other immune complex activators (349). One variable in this elution is size, as C1 or more likely activated Clr,Cls, and potentially also C l q elutes more readily from small than large complexes (386,387). Activated dissociated C1 or Clr,Cls, would not efficiently form C4,2 in the fluid phase. Any C3 convertase, which is formed, would not be located on the activator or on a surface and would thus not activate the latter portion of the complement reaction or, in fact, any of several other surface dependent biological phenomena. C1-In also has differing efficiencies in regulating C1 activated by immune complexes of different size (387). Considerable additional study is needed to understand these many complexities of classical pathway activation. VII. Regulation and Fate of Activated C1

A. INHIBITION OF ACTIVATED c1 BY Cl-IN C1-In binds to and blocks the activity of the activated Clr,Cls, complex (260) and activated C l (198,260,261,267).Although activated C l s reacts at a similar rate with C1-In regardless ofwhether the enzyme is free or contained within activated C1, the reactivity of the C l r subunit is considerably enhanced when present in activated C1 attached to an immune complex although C l s still reacts more rapidly (198,259,261). In addition, the binding of C1-In to activated C l s and C l r in activated C1 which is itself attached to an immune complex leads to dissociation of C1 as discussed next.

B. DISSOCIATION

OF

ACTIVATED Clr,ClS,

AND

c1

C1 has a tendency to dissociate following C1 activation. This is clearly shown by the 10-fold decrease in the association constant after activation of C1 in the fluid phase (130,296). The normal concentrations of C l q , Clr , and Cl s in serum together with the binding constant indicate that spontaneously activated C1 in serum is largely dissociated. This may account in part for the presence of free C1 subunits in pathological sera (297) and after C1 activation in serum. Activated Clr,Cls, free in solution is not likely to efficiently mediate formation of C4,2, the C 3 cleaving enzyme in free solution, thus dissociation would effectively stop progression of the complement sequence. The decrease in association constant following activation is likely also to pertain for activated C 1 bound to immune complexes following activation although the resulting tendency to dissociate is counterbalanced by the increased affinity of the C l q Clr,Cls, interaction characteristic of immune complex bound C1 (131,294). However, activated C1 bound to im-

THE CLASSICAL COMPLEMENT PATHWAY

205

mune complexes is effectively dissociated by Cl-In under physiological conditions. Laurell and colleagues were the first to observe C1 dissociation products in pathological human serum (297) and later were able to generate such products following treatment of normal human serum with activators of the classical complement pathway (206). We also observed dissociation of C1 in human serum by an immunochemical test following incubation with aggregated IgG (268). Studies by Laurell and associates (206) employing crossed immunoelectrophoresis indicated that complexes containing C l r , Cls, and Cl-In were generated by activation of the clasical pathway in normal human serum. Ziccardi and Cooper (302) and Sim, Arlaud, and coworkers (198,261) independently demonstrated the essential role of C1-In in inducing dissociation, analyzed the reaction, and determined the molecular composition of the released complexes. The released complexes contained activated C l r , Cls, and C1-In with each activated C l r and C l s molecule being complexed to a C1-In molecule (302). The release complexes sedimented in sucrose density gradients with a rate of 9 S and had a diffusion coefficient of 2.3 x cm"/second. This gives a calculated molecular weight of 330,000, a value consistent with the value of 382,000 obtained by addition of the subunit molecular weights for a Clr-Cls-C1-In complex (302). These data thus show that two activated Clr-Cls-(Cl-In), complexes are released per activated C1 molecule. Identical results were obtained with purified components and with serum and with both immune and nonimmune classical pathway activators (302). C l q remained bound to the activator (261). Other studies of the reaction mechanism employing combinations of DFP treated activated C l r and C l s in activated C1 indicate that C1-In binds first and most readily to the activated C l s subunit as noted earlier (261); however it is the binding to the activated C l r subunit that is most important in inducing dissociation (261). Binding of activated C1 to an immune complex is not required for C1-In to induce dissociation since it also dissociates the activated C h - C l s , complex in free solution with release of two Clr-Cls-(Cl-In), complexes per activated Clr,-Clsz complex (260). Of some interest also is the insensitivity of the complex to dissociation with EDTA (260,302) whereas the Clr,Cls, complex is readily disrupted by the chelating agent as discussed earlier. By an unknown mechanism therefore, C1-In binding strengthens the interaction between activated C l r and C l s and siinultaneously weakens the interactions between the two subunits of the C l r dimer and the C l s dimer in activated C1. Activation of the classical pathway in human serum can be ascertained and quantitated by measuring the concentration of the activated Clr-Cls-(ClIn), complex. This has proved to be a useful technique in studying classical pathway activation mechanisms as well as for assessing activation of the pathway in human diseases. Several techniques for quantitating the complex

206

N E I L R. COOPER

have been utilized. The activated Clr-Cls-(Cl-In)2 complex can be quantitated in human serum by immunodiflusion (268), crossed immunoelectrophoresis (206), electroimmunoassay (388), a radioimmunoassay termed the C1-inhibitor complex assay (389), and by an enzyme-linked immunosorbent assay or ELISA (390,391). C. C l q RECEPTOR INTERACTIONS Free C l q binds to several types of human cells as noted first by Dickler and Kunkel (392) for lymphocytes. This reaction has been amply confirmed and studied not only for lymphocytes (134,393,394) and specifically B lymphocytes, (135)but also for other cell types including monocytes (135), null cells (135), polymorphonuclear leukocytes (135), platelets (133,395), and endothelial cells (396). Studies of the nature of binding to B lymphocytes, polymorphonuclear leukocytes, and monocytes by Tenner and Cooper (134,135) show that the reaction is specific, reversible, saturable, and of moderately high affinity, indicating that it is receptor mediated. Inhibition studies indicate that the collagenous region of the C l q molecule is involved in the binding to the cell surface (134,136). Furthermore the binding site on C l q is masked in C 1 because native C1 does not have the ability to bind to the cell surface. Functional studies indicate that the interaction of particlebound C l q with polymorphonuclear leukocytes via the C l q receptor triggers an oxidative metabolic response in the cells (136). This reaction sequence beginning with C1 binding and leading to C1 activation and C1 inhibitor-mediated dissociation furnishes a mechanism to concentrate C1 activators on the surface of immunologically reactive cells through a C l q bridge. If this reaction pathway is of biological significance one would expect it to be most important in the early stages of host defense before antibody formation and thus involved with host defense against nonimmune activators such as certain viruses, bacteria, and parasites that activate C1 in the absence of antibody.

VIII. Comment

This review has assessed the current state of knowledge of C1 structure, activation, and regulation. As is apparent, much has been learned and the overall framework of the activation process has been elucidated. Although gaps in knowledge remain, particularly in the precise details at the molecular level, it is likely that a comprehensive understanding of these processes will be achieved in the near future. This information will provide an invaluable background for future insights arising from the applications of the techniques of molecular genetics to the study of the first component.

T H E CLASSICAL C O M P L E M E N T PATHWAY

207

ACKNOWLEDGMENTS I am particularly grateful to Drs. Wipreclit Augener, Peter Harpel, Andrea Tenner, Gunter Valet, and Robert Ziccardi who are responsible for the C1 contributions which have come from this laboratory. I also would like to particularly acknowledge innumerable valuable discussions with Dr. Ziccardi. Special thanks are due to Bonnie Bradt for the art work and to Bonnie Weier for efficient and patient assistance with all phases of the preparation of this manuscript. Work from this laboratory was supported by United States Public Health Service Grants A1 17354 and CA 14692.

REFERENCES 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. 28. 29. 30. 31. 32. 33. 34. 35.

Cooper, N. R. (1983). Fed. Pruc., Fed. Am. Soc. E r p . Biol. 42, 134. Loos, M. (1983). Prog. Allergy 30, 135. Miiller-Eberhard. H. J. (1975). Annu. Reu. Biocheni. 44, 697. Porter, R. R., and Reid, K. B. M. (1978). Nature (London) 275, 699. Porter, R. R., and Reid, K. B. M . (1979). Adu. Protein Chem. 1, 1. Reid, K. B. M., and Porter, R. R. (1975). Contemp. Top. Mol. Zmmunol. 4, 2. Sim, R. B. (1981). In “Methods in Enzymology” (T. Lorand, ed.), Vol. 80, p. 26. Academic Press, New York. Ziccardi, R. J. (1983). Springer Seniin. Ztntnunopathol. 6, 213. Mayer, M. M. (1958). Prog. Allergy 5, 215. Muller-Eberhard, H. J. (1969). Annu. Reo. Biochem. 38, 389. Osler, A. (1961). Adu. Znmtrnol. 1, 132. Osborn, T. W. (1937). “Complement or Alexin.” Oxford Univ. Press, London and New York. Nuttall, G. Z. (1888). H y g . ZnfektKr. 4, 353. Metchnikoff, E. (1883). B i d . Zentralbl. 3, 560. Buchner, H. (1889). Zentralbl. Bakteriol. 6, 1. Buchner, H. (1893). Arch. Hyg. (Berlin) 17, 179. Bordet, J . (1895). Ann. Inst. Pasteur 9, 462. Bordet, J. (1898). Ann. Inst. Pasteur 12, 688. Ehrlich, P., and Morgenroth, J. (1899). Klin. Wuchenschr. (Berlin) 36, 1. Ehrlich, P., and Morganroth, J. (1900). Klin. Wochensclzr. (Berlin) 37, 453, 681. Ehrlich, P., and Morgenroth, J. (1901). Klin. Woclienschr. (Berlin) 38, 256, 569. Ehrlich, P. (1906). “Studies in Immunity.” Wiley, New York. Ferrata, A. (1907). Klin. Wochenschr. (Berlin) 44, 366. Sachs, H., and Teruuchi, Y. (1907). Klin. Wochenschr. (Berlin) 44, 467 and 520. Brand, K. (1907). Klin. Wochenschr. (Berlin) 44, 1075. Ornorokow, L. (1911). 2. ZaununForsch. 10, 285. Rib, H., and Sachs, H. (1911). Zentralhl. Bakteriol. 50, 43. Sachs, H., and Omorokow, L. (1911). Z. ZmmunForsch. 11, 710. Coca, A. (1914). Z. Zmmunitaetsforsch. 21, 604. Weil, E. (1913). Biochemistry 48, 347. Gordon, J., Whitehead, H. R . , and Wormall, A. (1926). Biochem. J. 20, 1028. Hegedus, A , , and Greiner, H . (1938). Z.ZmmtrnForsch. 92, 1. Bier, 0. G . , Leyton, G . , Mayer, M. M., and Heidelberger, M. (1945).J. E x p . Med. 81, 449. Ecker, E. E., Pillemer, L., and Seifter, S. (1943).J. Znzmunol. 47, 181. Mayer, M. M. (1961). “Immunochemical Approaches to Problems in Microbiology,” p. 268. Rutgers Univ. Press, New Brunswick, N.J.

208 36. 37. 38. 39. 40. 41. 42. 43.

44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76.

77. 78. 79.

NEIL R. COOPER

Becker, E. L. (9155). Nature (London) 176, 1073. Becker, E. L. (1956). J . Zmmnunol. 77, 462. Becker, E. L. (1956).J. Zmmunol. 77, 469. Becker, E. (1960). J . Zmmunol. 84, 299. Lepow, I. H., Ratnoff, 0. D., Rosen, F. S., andPillemer, L. (1956). Proc. SOC. E x p . B i d . Med. 92, 32. Lepow, I. H., Ratnoff, 0. D., and Levy, L. R. (1958).J. E x p . Med. 107, 451. Lepow, I. H., Ratnoff, 0. D., and Pillemer, L. (1956).Proc. Soc. E x p . Biol. Med. 92, 111. Lepow, I. H., Wurz, L., Ratnoff, O., and Pillemer, L. (1954).J . Zmmunol. 73, 146. Ratnoff, 0. D., and Lepow, I. H. (1957).J . Exp. Med. 106, 327. Lepow, I. H., NaK, G. B., Todd, E. W., Pensky, J., and Hinz, C. F. (1963).J . Exp. Med. 117, 983. Haines, A. L., and Lepow, I. H. (1964).J. Zmmunol. 92, 456. Miiller-Eberhard, H. J., and Kunkel, H. (1961). Proc. Soc. E x p . Biol. Med. 106, 291. Miiller-Eberhard, H. J., and Nilsson, U. R. (1960).J . E x p . Med. 111, 217. Heidelberger, M. (1940). Science 92, 534. Heidelberger, M . , Bier, 0. G., and Mayer, M. M. (1942).J . Exp. Med. 75, 285. Maurer, P. H., and Talmage, D. W. (1953).J . Zmmunol. 70, 435. Weigle, W. O., and Maurer, P. H. (1957).J . Zmmunol. 79, 211. Heidelberger, M. (1941). J. E x p . Med. 73, 681. Mayer, M. M. (1961). In “Experimental Immunochemistry” (E. A. Kabat and M. M. Mayer, eds.), 2nd Ed., p. 133. Thomas, Springfield, 1.; Levine, L., Cowan, K. M., Osler, A. G., and Mayer, M. M. (1953).J . Zmmnunol. 71, 359. Maurer, P. H., and Weigle, W. 0. (1953).J . Zmmunol. 71, 284. Taranta, A,, Weiss, H. S., and Franklin, E. C. (1961). Nature (London) 189, 239. Becker, E. L. (1959). J . Zmmunol. 82, 43. Laporte, R., Hardre de Looze, L., and Sillard, R. (1957).Ann. Znst. Pasteur 92, 15, 28, 33 N d , G. B., Pensky, J . , and Lepow, I. H. (1964).J. E x p . Med. 119, 593. Reid, K. B. M., Lowe, D. M., and Porter, R. R. (1972). Biochem. J. 130, 749. Lowe, D. M., and Reid, K. B. M. (1974). Biochem. J. 143, 265. Volanakis, J. E., and Stroud, R. M . (1972).J . Zmmunol. Methods 2, 25. Campbell, D., Booth, N. A , , and Fothergill, J. E. (1978). Biochem. J. 177, 531. Ziola, B., Salmi, A,, and Penttinen, K. (1982).J . Zmmunol. Methods 49, 247. HoRken, K., McLaughlin, P. J., Price, M. R., Preston, V. E., and Baldwin, R. W. (1978). Znimunochemistry 15, 409. McDonald, T. L., and Burger, D. (1983). Zmmunology 37, 517. Alexander, R. J., and Steiner, L. (1980). J . Zmmunol. 124, 1418. Slayter, H. S., Alexander, R. J., and Steiner, L. (1983). E u r . J. Zmmunol. 13, 102. Yonemasu, K., and Stroud, R. M. (1971).J. Zmmunol. 106, 304. Agnello, V., Winchester, R. J.. and Kunkel, H. G. (1970). Zmmunology 19, 909. Sledge, C. R., and Bing, D. H. (1973). J. Zmmunol. 111, 661. Calcott, M. A., and Muller-Eberhard. H. J . (1972). Biochemistry 11, 3443. Assimeh, S. N., Bing, D. M., and Painter, R. H. (1974).J . Zmmunol. 113, 225. Strang, C. J., Seigel, R. C., Phillips, M. L., Poon, P. H., and Schumaker, V. N. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 586. Arlaud, G. J., Sim, R. B., Duplaa, A. M., and Colomb, M. G. (1979). Mol. Zmmunol. 16, 445. Kolb, W. P., Kolb, L. M., and Podack, E. R. (1979).J. Zmmunol. 122, 2103. Reid, K. B. M . (1974). Biochem. J. 14, 189. Tenner, A. J., Lesavre, P. H., and Cooper, N. R. (1981).J. Zmmunol. 127, 648.

T H E C L A S S I C A L C O M P L E M E N T PATHWAY

80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99.

209

Ziccardi, R. J. (1985). Zintnunol. Tech. (in press). Ziccardi, R . J.. and Cooper, N. R . (1977).J . Znimnnol. 118, 2047. Rosano, C. L., Parhami, N., and Honvitz, C. (1979).J. Lab. Clin. Med. 94, 593. Calcott, M. A., and Miiller-Eberhard, H. J. (1972). Biocheniistrtt 10, 1166. Yonemasu, K . , Stroud, R . M., Niedernieier, W., and Butler, W. T. (1971). Biochem. Biophys. Res. Commun. 43, 1388. Knobel, H. R., Villger, W., and lsliker, H. (1975). Eur. J . Zn~munol.5, 78. Shinkai, H., and Yonemasu, K. (1979). Biochent. J. 177, 847. Miziiochi, T., Yoneinasu, K., Yamashita, K., and Kobata, A. (1978).J. B i d . Cheni. 253, 7404. Liberti, P. A,, and Paul, S. M. (1978). Biochen~istry17, 1952. Kilcherr, E., Fuchs. H., Tschopp, J., and Engel, E. (1982). Mol. Immunol. 19, 683. Reid, K. B. M., and Porter, R. R. (1976). Biochem. J . 155, 19. Yonemasu, K., and Stroud, R . M. (1972). Znimunochemistrll 9, 545. Heusser, C., Boesman, M., Nordin, J. H . , and Isliker, H. (1973).J. Zmtnunol. 110, 820. Reid, K. B. M., and Thompson, E. 0. P. (1978). Biochem. J. 173, 863. Reid, K. B. M. (1976). Biochem. J. 155, 5. Muller-Eberhard, H. J. (1968). Ado. Zmmunol. 8, 1. Muller-Eberhard. H. J. (1971). Prog. Zmmunol. 1, 553. Reid, K. B. M. (1977). Biochetn. J. 161, 247. Traub, W., and Piez, K. A. (1971). Adu. Protein Chem. 25, 243. Shelton, E., Yonemasu, K., and Stroud, R . M. (1972). Proc. Natl. Acad. Sci. U.S.A. 69, 65.

Svehag, S. E., and Bloth, B. (1970). Acta Pathol. Microbiol. Scand. Sect. B 78, 260. Svehag, S. E., Manhem, L., and Bloth, B. (1972). Nature (London) New Biol. 238, 117. Polley, M. J. (1971). Prog. Zmmunol. 1, 597. Brodsky-Doyle, B., Leonard, K. R., and Reid, K. B. M. (1976). Biochem. J. 159, 279. 104. Miller, E. J. (1972). Biochetnistry 11, 4903. 105. Reid, K. B. M., Gagnon, J , , and Frampton, J. (1982). Biocheni. J . 203, 559. 106. Schumaker, V. N., Poon, P. H., Seegan, 6. W., and Smith, C. A. (1981).J . Mol. Biol. 148, 191. 107. Porter, R . R., and Reid, K. B. M. (1978). Nature (London) 275, 699. 108. Gilmour, S., Randall, J. T., Willan, K. J.. Dwek, R. A , , and Torbet, J. (1980). Nature (London) 285, 512. 109. Honvitz, P. M., Wasylewski, Z., and Kolb, W. P. (1980). Biochem. B b p h y s . Res. Commun. 96, 382. 110. Lin, T., and Fletcher, D. S. (1978). Znimunocheinistry 15, 107. 111. Villiers, C. L., Arlaud, G . J , , Painter, R. H., and Colomb, M . G. (1980). FEBS Lett. 117, 289. 112. Heusser, C. H., Boesman, M . , Knobel, H. R., Jacot-Gnillarmod. H., and Isliker, H. (1975). Zmrnunochemistry 12, 213. 113. Cooper, N. R. (1973). Contemp. Top. Mol. Ztmnunol. 2, 155. 114. Muller-Eberhard, J. H . , Nilsson, U. R., Dahnasso, A. P., Polley, M. J., and Calcott, M. A. (1966). Arch. Pathd. 82, 205. 115. Schumaker, V. N . , Calcott, M. A,, Spiegelberg, H. L., and Miiller-Eberhard, H. J. (1976). Biochemistr!y 15, 5175. 116. Augener, W., Grey, H. M., Cooper, N. R., and Muller-Ebehard. H. J. (1971). Inmunochemistry 8, 1011. 117. Miiller-Eberhard, H. J , , and Calcott, M. A. (1966). Zmniunochemistry 3, 500. 118. Ishizaka, T., Ishizaka, K., Salmon, S., and Fudenherg, H. (1967). J . lmmunol. 99, 82. 100. 101. 102. 103.

210

NEIL R. COOPER

119. Isenman, D. E., Dorrington, K. J., and Painter, R. H. (1975). J. Zrnmnunol. 114, 1726. 120. Knobel, H. R., Heusser, C., Hodrick, M. L., and Isliker, H. (1974). J . Zmmunol. 112, 2094. 121. Fletcher, D. S., Williams, H. R . , and Lin, T. Y. (1978). Biochim. Biophys. Acta 540,270. 122. Paques, E. P., Huber, R., Priess, H., and Wright, J. K. (1979).Hoppe-Seyler’sZ.Physiol. Chem. 360, 177. 123. Reid, K. B. M. (1979). Biochem. J. 179, 367. 124. Hughes-Jones, N. C., and Gardner, B. (1979). Mol. Zmmunol. 16, 697. 125. Van Schravendijk, M. R., and Dwek, R. A. (1982). Mol. Zmmunol. 19, 1179. 126. Easterbrook-Smith, S. B., Zavodszky, P., Willan, K. J., Gettin, S. P., and Dwek, R. A. (1978). Biochem. Soc. Trans. 6, 1126. 127. Reid, K. B. M., Sim, R. B., and Faiers, A. P. (1977). Biochem. J. 161, 239. 128. Wiger, D., and Natvig, J. B. (1972).J. Zmmunol. 109, 933. 129. Poon, P. H., Schumaker, V. N., Phillips, M. L., and Strang, C. J. (1983).J. Mol. B i d . 168, 563. 130. Siegel, R. C., and Schumaker, V. N. (1983). Mol. Zmmunol. 20, 53. 131. Hughes-Jones, N. C., and Gorick, B. D. (1982). Mol. Zmmunol. 19, 1105. 132. Menzel, E. J., Smolen, J. S., Liotta, L., and Reid, K. B. M. (1981).FEBS Lett. 129, 188. 133. Wantier, J. L., Souchon, H . , Reid, K. B. M., Peltier, A. P., and Caen, J. P. (1977). Zmmunochemistry 14, 763. 134. Tenner, A. J., and Cooper, N. R. (1980).J. Zmmunol. 125, 1658. 135. Tenner, A. J., and Cooper, N. R. (1981).J. Zmmunol. 126, 1174. 136. Tenner, A. J., and Cooper, N. R. (1982).J. Zmmunol. 128, 2547. 137. Haines, A. L., and Lepow, I. H. (1964).J. Zinmunol. 9, 468. 138. Okamura, K., Muramuta, M . , and Fujii, S. (1973). Biochim. Biophys. Actu 295, 252. 139. Sakai, K., and Stroud, R. M. (1973).J. Zmmunol. 110, 1010. 140. Valet, G., and Cooper, N. R. (1974).J. Immunol. 112, 339. 141. NafT, G. B., and Ratnoff, 0. D. (1968).]. E x p . Med. 128, 571. 142. deBracco, M., and Stroud, R . M. (1971).J. Clin. Invest. 50, 838. 143. Gigli, I., Porter, R. R., and Sim, R. B. (1976). Biochem. J. 157, 541. 144. Sim, R. B., and Porter, R. R. (1976). Biochem. SOC. Trans. 4, 127. 145. Ziccardi, R. J., and Cooper, N. R. (1976).J. Zmmunol. 116, 496. 146. Cooper, N. R., and Ziccardi, R. J. (1976). Miami Winter Symp. 11, 167. 147. Arlaud, G. J., and Gagnon, J. (1981). Biosci. Rep. 1, 779. 148. Barkas, T., Scott, G. K., and Fothergill, J. E. (1973). Biochefn. SOC. Trans. 1, 1219. 149. Sim, R. B., Porter, R. R . , Reid, K. B. M., and Gigli, I. (1977). Biochem. J. 163, 219. 150. Young, C. L., Barker, W. C., Tornaselli, C. M., and Dahoff, M. 0. (1978). “Atlas of Protein Sequence and Structure” (M. 0. Dayhoff, ed.), Vol. 5 , Suppl. 3, p. 73. Natl. Biomedical Foundation, Silver Spring, Maryland. 151. Arlaud, G. J., Gagnon, J., and Porter, R. R. (1982). Biochem. J. 201, 49. 152. Arlaud, G. J., and Gagnon, J. (1983). Biochemistry 22, 1758. 153. Arlaud, G. J. Villiers, C. L., Chesne, S., and Colomb, M. G . (1980). Biochim. Biophys. Actu 616, 116. 154. Chapuis, R. M., Isliker, H., and Assimeh, S. N. (1977). Zminunochemistry 14, 313. 155. Siegel, R. C., Schumaker, V. N., and Poon, P. H. (1981).J. Zmmunol. 127, 2447. 156. Ishizaki, E., Yoshioka, Y., Mori, Y., and Koyama, J. (1976). J. Biochem. 80, 1423. 157. Campbell, R. D., Booth, N. A., and Fothergill, J. (1979). Biochem. J. 183, 579. 158. Ziccardi, R. J . , Cooper, N. R., and Hugli, T. E. Unpublished observations. 159. Ziccardi, R. J., and Cooper, N. R. (1976).J. Zmmunol. 116, 504. 160. Arlaud, G . J., Chesne, S., Villiers, C. L., and Colomb, M. G . (1980).Biochim. Biophys. Acta 616, 105.

T H E CLASSICAL COMPLEMENT PATHWAY

211

McRae, B., Lin, T. Y., and Powers, J. C. (1981).J. B i d . Chem. 356, 12362. Bing, D. H . , Laura, R . , Andrews, J. M . , and Cory, M. (1978). Biochemistry 17, 5713. Cooper, N. R., Miles, L. A,, and Griffin, J. H. (1980).J. Zmmunol. 124, 1517. (Abstr.) Tschopp, J., Villiger, W., Fuchs, H., Kilcherr, E . , and Engle, J. (1980).Proc. Natl. Acad. Sci. U.S.A. 77, 7014. 165. Valet, G., and Cooper, N. R. (1974).J . Zmmunol. 112, 1667. 166. Reid, K. B. M., and Porter, R. R. (1976). Biochem. J. 155, 19. 167. Villiers, C. L., Duplaa, A. M., Arlaud, G. J., and Colomb, M. G. (1982). Biochim. Biophys. Acta 700, 118. 168. Takahashi, K., Nagasawa, S., and Koyama, J. (1975). FEBS Lett. 55, 156. 169. Assimeh, S. N . , Chapuis, R. M., and Isliker, H. (1978). Zrnmunochemistry 15, 13. 170. Okamura, K., and Fuji, S. (1978). Biochim. Biophys. Acta 534, 258. 171. Naf€, G. B., and Ratnoff, 0. D. (1968).J. E x p . Med. 128, 571. 172. Arlaud, 6 . J., Reboul, A,, and Colomb, M. G. (1977).Biochim. Biophys. Acta 489, 227. 173. Arlaud, G. J., Reboul, A,, Meyer, C. M., and Colomb, M. G. (1977). Biochim. Biophys. Acta 485, 215. 174. Valet, G., and Cooper, N. R. (1974).J. Zmmunol. 112, 339. 175. Cooper, N . R., and Ziccardi, R. J. (1977).J. Zmmunol. 119, 1664. 176. Tschopp, J . , Villiger, W., Fuchs, H . , Kilcherr, E . , and Engel, J. (1980).Proc. Natl. Acad. Sci. U.S.A. 77, 7014. 177. Takahashi, K., Nagasawa, S . , and Koyama, J. (1975). F E E S Lett. 50, 330. 178. Sakai. K., and Stroud, R. M. (1974). Io,tnunochemistry 11, 191. 179. Ziccardi, R. J., and Cooper, N. R. Unpublished studies. 180. Mori, Y., Ueda, E., Takenchi, T., Tanuichi, S., and Koyama, J. (1980).J . Biochem. 87, 1762. 181. Cooper, N. R., Miles, L. A , , and Griffin, J. H. (1984). To be submitted. 182. Kasahara, Y., Takahashi, K., Nagasawa, S., and Koyama, J. (1982). FEBS Lett. 141, 128. 183. Takahashi, K., Nagasawa, S., and Koyama, J. (1976).FEBS Lett. 65, 20. 184. Dodds, A. W., Sim, R. B., Porter, R. K.,and Kerr, M. A. (1978). Biochem. J. 175, 383. 185. Bauer, J., and Valet, 6. (1982). Mol. Immunol. 19, 487. 186. Ziccardi, R. J. (1982).J. Znimunol. 128, 2505. 187. Ratnoff, 0. D., Pensky, J , , Ogston, D., and N d , G. B. (1969).J. Exp. Med. 129, 315. 188. Ziccardi, R. J. (1982). J. Immunol. 128, 2500. 189. Taylor, P. A,, Fink, S., Bing, D. H., and Painter, R. H. (1977). J. Zmmunol. 118, 1722. 190. Kassell, B., and Kay, B. (1973). Science 180, 1022. 191. Londsdale-Eccles, J., Kerr, M. A , , Walsh, K. A , , and Neurath, H. (1979). FEBS Lett. 100, 157. 192. Villiers, C. L., Chesne, S., LaCroix, M. B., Arlaud, 6. J., and Colomb, M. G. (1982). Biochem. J. 203, 185. 193. Bauer, J., and Valet, G. (1981).Biochim. Biophys. Acta 670, 129. 194. Villiers, C. L., Arlaud, G. J., and Colomb, M. G. (1983). Biochem. J. 215, 369. 195. Kasahara, Y., Takahashi, K . , Nagasawa, S., and Koyama, J. (1982). FEBS Lett. 141, 128, 196. Volanakis, J . E., Schrohenloher, R. E., and Stroud, R. M. (1977). J. Zmmunol. 119, 337. 197. Andrews, J. M., and Baillie, R. D. (1979). J. Immune!. 123, 1403. 198. Arlaud, R. J., Retioul, A,, Sim, R. B., andColomb, M. G. (1979).Biochim. Biophys. Acta 576, 151. 199. Ashgar, S. S., Pondman, K. W., and Cormane, R. H. (1975).Biochim. Biophys. Actu 317, 539. 200. Bing, D. H. (1969).Biochemistry 8, 4503. 201. Morgan, P. H., and Nair, I. G. (1975). Biochem. Biophys. Res. Cotmnun. 66, 1037. 161. 162. 163. 164.

212 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241.

NEIL H. COOPER

Ratnoff, 0. D., and Naff, 6. B. (1967).J. Exp. Med. 125, 337. Taubman, S. B . , and Lepow, I. H. (1971). Zmmunochemistry 8, 951. Lepow, I. H., Pillemer, L., and Ratnoff, 0. D. (1953).J . Exp. Med. 98, 277. Laurell, A. B., Lundh, L., and Malmquist, J. (1965). Acta Pathol. Microhiol. Scand. 86, 318. Laurell, A,, Johnson, U., Martensson, U., and Sjoholm, A. G. (1978). Acta Pathol. Microbiol. Scund. 86, 299. Taubinan, S. B., and Lepow, I. H. (1971). Zmmunochertiistrzj 8, 951. Bartholomew, 1%. M., and Esser, A. F. (1980). Riocheniistry 19, 2847. Mori, Y., Koketsu, M., Abe, N., and Koyama, J. (1979). Biochetn. J. 85, 1023. Donaldson, V. H. (1968).J. Appl. Physiul. 25, 416. Polley, M. J., and Miiller-Eberhard, H. J. (1968).J. Exp. Med. 128, 533. Stroud, R. M., Austen, K. F., and Mayer, M. M. (1965). Imrnunochet~iistry2, 219. Stroud, R. M . , Mayer, M. M., Miller, J. A,, and McKenzie, A. T. (1966). Zmmunochemistry 3, 163. J. E x p . Med. 121, 819. Miiller-Eberhard, H. J . , and Lepow, I. H. (1965). e 7217. ~~~stry Patrick, R. A,, Taubman, S. B., and Lepow, I. H. (1970). Z ~ ~ ~ ~ 7 i u n f ~ c h7, Cooper, N. R. (1971). Prog In~munol.1, 567. Nagaki, K., and Stroud, R. M. (1969). J. Zmr7iunol. 102, 421. Morgan, P. H., and Nair, I. 6. (1977). J. Immunol. 119, 19. Canady, W. J., Westfall, S., Wirtz, G. H., and Robinson, D. A. (1976). Zrninunocheinistry 13, 229. Harpel, P. C., and Cooper, N. R. (1975).J. Clin. Invest. 55, 593. Levy, L. R . , and Lepow, I. H. (1959). Proc. Soc. Exp. B i d . Med. 101, 608. Pensky, J., Levy, L. R., and Lepow, I. H. (1961). J. Biol. Chem. 236, 1674. Schultze, H. E., Heide, K., and Haupt, H. (1962). Klin. Wochenschr. 40, 427. Pensky, J., and Schwick, H. G. (1969). Science 163, 698. Donaldson, V. H., and Evans, R. R. (1963). Am. J. Med. 35, 37. Landerman, N. S., Webster, M. E., Becker, E. L., and Ratcliffe, H. E. (1962).J . Allergy 33, 330. Rosen, F. S., Alper, C. A,, Pensky, J., Klemperer, M . R., and Donaldson, V. H. (1971).J. Clin. Invest. 50, 2143. Alper, C. A , , and Rosen, F. S. (1976). A d v . Hum. Genet. 7, 141. Frank, M. M., Gelfand, J. A,, and Atkinson, J. P. (1976). Ann. Znt. Med. 5, 580. Gelfand, J. A., Sherins, R. J., Alling, D. W., and Frank, M. M. (1976). New Engl. J. Med. 295, 1444. Bergamaschini, L., Cicardi, M., Tucci, A., Gardinali, M., Frangi, D., Valle, C., and Agostoni, A. (1983). Allergy 38, 81. Gadek, J. E., Hosea, S. W., Gelfand, J. A,, Santaella, M . , Wickerhauser, M., Triantaphyllopoulos, D. C., and Frank, M. M. (1980). New Engl. J. Med. 302, 542. Rothback, C., Green, R. L., Levine, M. I., and Fireman, P. (1979).Am. J . Med. 66, 687. Gadek, J. E., Hosea, S. W., Gelfand, J. A., and Frank, M. M. (1979).J. Clin. Invest. 64, 280. Donaldson, V. H., and Rosen, F. S. (1964).J. Clin. Invest. 43, 2204. Donaldson, V. H., Rosen, F. S., and Bing, D. H. (1977). Trans. Am. Assoc. Phys. 60, 174. Curd, J. G., Prograis, L. J., and Cochrane, C. A. (1980).J . Exp. Med. 152, 742. Gigli, I . , Mason, J. W., Colman, R. W., and Austen, K. F. (1970).J. Ztnmunol. 104, 574. Forbes, C. D., Pensky, J., and Ratnoff, 0. D. (1970j.J. Lnh. Clin. Med. 76, 809. Schreiber. A. D., Kaplan, A. P., and Austen, K. F. (1973). J . Clin. Invest. 52, 1402. Darnus, P. S., Hicks, M., and Rosenberg, R. D. (1973). Nature (London) 246, 355.

T H E CLASSICAL C O M P L E M E N T PATHWAY

213

242. Saito, H., Goldsmith, G. H . , Moroi, M., and Aiko, N. (1979). Proc. Nutl. Acurl. Sci. U . S . A . 76, 2013. 243. Sim, R. B., Reboul, A., Arlaud, G . J . , Villiers, C. L., and Colomb, M. G . (1979). FEBS Lett. 97, 111. 244. Ziccardi, R. J. (1981).J . Znanunol. 126, 1769. 245. Haupt, H., Heitnlmrger, N . , Krnnz, T., and S c h i c k , H. 6. (1970).Eur. J. Biochem. 17, 254. 246. Brower, M. S., and Harpel, P. C. (1982).J. B i d . C h e m 257, 9849. 247. Harpel, P. C . (1976). In “Methods in Enzymology” (L. Lorand, ed.), Vol. 80, p. 43. Academic Press, New York. 248. Minta, J. O., and Aziz. E. (1980).J . Iminunol. 126, 250. 249. Reboul, A,, Arlaud, G. J., Siin, R. B., and Colomb, M . 6. (1977). FEBS Lett. 79, 45. 250. Sim, R. B., and Reboul, A. (1981).In “Methods in Enzymology” (L. Lorand, ed.), Vol. 80, p. 43. Academic Press, New York. 251. Harpel, P. C . , Hugli, T. E., and Cooper, N . R. (1975).J. Clin. Znoest. 55, 593. 252. Curd, J. G . , Yelvington, M., Ziccardi, R. J., Mathison, D. A,, and Griffin, J. H. (1981). Clin. Exp. Zmmunol. 45, 261. 253. Loos, M., Opferkuch, W., and Hingelmann, R. Z. (1971).Med. Mikrobiol. Imrnunol. 156, 194. 254. Ishizaki. E., Mori, Y., and Koyama, J. (1977).J. Biochem. (Tokyo)82, 1155. 255. Gigli, I., and Austen, K. F. (1971).Annu. Reo. Microbiol. 25, 309. 256. Ziccardi, R. J , , and Cooper, N. R. (1980). Clin.Imn,unol. Zrntnunoputhol. 15, 465. 257. Minta, J. 0. (1981).J. Zrnmunol. 126, 245. 258. Nagaki, K., Iida, K . , and Inai, S. (1974). Int. Arch. Allergy A p p l . Znmiunol. 46, 935. 259. Sim, R. B., Arlaud, G. J . , and Colomb, M . G. (1980).BiochirnBiophys. Actu 612, 433. 260. Chesne, S . , Villiers, C. L., Arlaud, G . J , , LaCroix, M. G., and Colomb, M. G. (1982). Biochem. J. 201, 264. 261. Sim, R. B., Arlaud, 6. J., and Colomb, M. G. (1979).Biochem. J. 179, 449. 262. Rent, R., Myrrman, R., Fiedel, B. A , , and Gewurz, H. (1976).Clin.Exp. Immunol. 232, 264. 263. Caughman, G. B., Boackle, R. J . , and Vesely, J. (1982). M u / . Irnmunol. 19, 287. 264. Nagaki, K., and Inai, S. (1976). Znt. Arch. Allergy A p p l . Imrnunol. 50, 172. 265. Raepple, E., Hill, H. V., and Loos, M . (1976). Immunochernistry 13, 251. 266. Loos, M., Volanakis, J . E., and Stroud, R. M. (1976). Inimunocherni.stry 13, 251. 267. Ziccardi, R. J., and Cooper, N . R. (1978).J. Immunol. 121, 2148. 268. Ziccardi, R. J., and Cooper, N . R. (1978).J. E x p . Merl. 147, 385. 269. Yan, D., GU, X., Wang, D., and Yang, S. (1981).1. Ininizrnol. 127, 1064 270. Cooper, N. R., Nemerow, G. R., and Mayes, J. T. (1983). Springer Senzin. Itnmunopathol. 6, 195. 271. Laurell, A. B., and Mrtensson, U. (1974).Actu Pnthol. Microbiol. Scund. 82, 585. 272. Nagasawa, S., Takahashi, K., and Koyama, J. (1974).FEBS Lett. 41, 280. 273. Siegel, R. C., Schumaker, V. N., and Poon, P. H. (1981).J. Im?nunol. 127, 2447. 274. Gigli, I., and Austen, K. F. (1969).J. Exp. Med. 129, 679. 275. Gigli, I., and Austen, K. F. (1969).J. Exp. Med. 130, 833. 276. Assimeh, S. N . , and Painter, R. H. (1975).J. Zmmunol. 115, 488. 277. Borsos, T., and Rapp, H. J. (1963).J. Imrnunol. 91, 851. 278. Borsos, T., Rapp, H. J., and Mayer, M. M. (1961).J. Immunol. 87, 310. 279. Rapp, H. J . , and Borsos, T. (1970). “Molecular Basis of Coinplenient Action.” Appleton, New York. 280. Cooper, N . R., and Miiller-Eberhard, H. J. (1968).lintnunochemistry 5, 155.

214

N E I L R. C O O P E R

281. Cooper, N. R., Polley, M. J., and Miiller-Eberhard. H. J. (1970). Zmmunochemistry 7, 341. 282. Colten, H. R., Bond, H. E., Borsos, T., and Rapp, H. J. (1969). J. Zmmunol. 103, 862. 283. Bing, D. H. (1971).J. Zmmuttol. 107, 1243. 284. Medicus, R. G., and Chapuis, R. M. (1980).J. Immunol. 125, 390. 285. Assimeh, S. N., and Painter, R. H. (1975).J. Zmmunol. 115, 482. 286. Katz, A , , Weicker-Thorne, J., and Painter, R. H. (1977). Am. J. Pathol. 88, 679. 287. Pepys, M. B., Dash, A. C., Munn, E. A,, Feinstein, A,, Skinner, M., Cohen, A. S., Gewurz, H., Osmond, A. P., and Painter, R. H. (1977). Loncet 1, 1029. 288. Cooper, N. R., and Ziccardi, R. J. (1979). Mol. Zmmunol. 16, 821. 289. Painter, R. H. (1977).J. Zmmunol. 119, 2203. 290. Ziccardi, R. J., and Cooper, N. R. (1978). Science 199, 1080 291. Ross, G. D., and Jensen, J. A. (1973).J. Zmmunol. 110, 175. 292. Benson, H. N., Brumfield, H. P., and Pomeroy, B. S. (1961).J. Zmmunul. 87, 616. 293. Loos, M., Hill, H. U., Welleck, B., and Heinz, H. P. (1976). FEBS Lett. 64, 342. 294. Ziccardi, R. J., and Tschopp, J. (1982). Biochem. Biophys. Res. Cornmun. 107, 618. 295. Ziccardi, R. J. (1984).J . Zmmunol. 132, 283. 296. Kilcherr, E., Fuchs, H., Tschopp, J., and Engel, J. (1982). Mol. Zmmunol 19, 683. 297. Laurell, A. B., Mrtensson, U . , and Sjoholm, A. G. (1976).Acta Pathol. Microbiol. Scand. 84, 455. 298. Bartholomew, R. M., and Esser, A. F. (1977). J , Zmmunol. 119, 1916. 299. Colten, H. R., Borsos, T., and Rapp, H. J. l1968). J. Zmmunol. 100, 808. 300. Loos, M., Laurell, A. B., Sjoholm, A. C., Martensson, U., and Berkel, A. I. (1980).J . Zmmunol. 124, 59. 301. Ziccardi, R. J. (1983).J , B i d . Chem. 258, 6187. 302. Ziccardi, R. J . , and Cooper, N. R. (1979).J . Zmmunol. 123, 788. 303. Ishizaka, T., Ishizaka, K., Borsos, T., and Rapp, H. (1966).J . Zmmunol. 97, 176. 304. Isenman, D. E., Dorrington, K. I.. and Painter, R. H. (1975).J. Zrnmunol. 114, 1726. 305. Colomb, M., and Porter, R. R. (1975). Biochem. J . 145, 177. 306. Hurst, M. M., Volanakis, J. E . , Hester, R. B., Stroud, R. M., and Bennett, J. C. (1974).J . Exp. Med. 140, 1117. 307. Plaut, A. G., Cohen, S., and Toniasi, T. B. (1972). Science 176, 55. 308. Burton, D. R., Boyd, J., Brampton, A. D., Easterbrook-Smith, S. B., Emanuel, E. J., Novotny, J., Rademacher, T. W., van Schravendijk, M. R., Sternberg, M. J., and Dwek, R. A. (1980). Nature (London) 288, 338. 309. Emanuel, E. J., Brdmpton, A. D., Burton, D. R., and Dwek, R. A. (1982). Biochem. J . 205, 361. 310. Esser, A. F. (1982).In “Biological Membrane” (D. Chapman, ed.), Vol. 4, p. 277. Academic Press, New York. 311. Loos, M. (1982).Ann. Zmmunol. (Znst. Pasteur) 133C, 165. 312. Loos, M., Wellek, B., Thiesen, R.,and Opferkuch, W. (1978). Infect. Zmmun. 22, 5. 313. Clas, F., and Loos, M. (1981). Infect. Immun. 41, 1138. 314. Betz, S. J., and Isliker, H. (1981).J . Zmmunol. 127, 1748. 315. Tenner, A. J.. Ziccardi, R. J., and Cooper, N. R. (1984).J. Zmmunol. 133, 886. 316. Cooper, N. R., Jensen, F. C., Welsh, R. M., and Oldstone, M . B. A. (1976).J . Exp. Med. 144, 970. 317. Welsh, R . M., Cooper, N. R . , Jensen, F. C., and Oldstone, M. B. A. (1975). Nature (London)257, 612. 318. Welsh, R. M., Jensen, F. C., Cooper, N. R., andoldstone, M. B. A. (1976).Virulogy74, 432.

T H E CLASSICAL C O M P L E M E N T PATHWAY

215

319. Sherwin, S . A., Benveniste, R. E., and Todaro, G. J. (1978). Int. J. Cancer 21, 6. 320. Gallagher, R. E.. Schrecker, A. W., Walter, C. A., and Gallo, R . C. (1978). J. Natl. Cancer lnst. 60, 677. 321. Santoro, F., Ouaissi, M., Pestel, J., and Capron, A. (1980).J . Immunol. 129, 2886. 322. Musoke, A. J., and Barbet, A. F. (1977). Nature (London) 270, 438. 323. Bredt, W., Wellek, B., Brunner, H., and Loos, M. (1977). Infect. Immun. 15, 7. 324. Linder, E., Lehto, V. P . , and Stenman, S. (1979). Nature (London) 278, 176. 325. Storrs, S. B., Kolb, W. P . , Pinckard, R. N . , and Olson, M. S. (1981).J. Biol. Chem. 256, 10924. 326. Giclas, P. C., Pinckard, R. N., and Olson, M. S . (1979).J. Immunol. 122, 146. 327. C h s , D. R., Siegel, J . , Petras, K., Osmond, A. P., and Gewnrz, H. (1977).J . Immunol. 119, 187. 328. Kaplan, M. H., and Volanakis, J. E. (1974)./. Immunol. 112, 2135. 329. Bartholornew, R. M., Esser, A. F . , and Miiller-Eberhard, H. J. (1978).J. E x p . Med. 147, 844. 330. Cyong, J. E., Witkin, S. S . , Rieger, B., Barbarese, E., Good, R. A., and Day, N . K. (1982).J. E x p . Med. 155, 587. 331. Vanguri, P., Koski, C. L., Silverman, B., and Shin, M. L. (1982). Proc. Natl. Acad. Sci. U.S.A. 79, 3290. 332. Schultz, D. R., and Arnold, P. I. (1981).J. Zmmunol. 126, 1994. 333. Loos, M., Bitter-Sverman, D., and Dierich, M . P. (1974).J. Immunol. 112, 935. 334. Cooper, N . R., and Morrison, D. C. (1978).J. Zmmunol. 120, 1862. 335. Rent, R., Ertel, N . , Einstein, R., and Gewnrz, H. (1975).J. Immunol. 114, 120. 336. Fiedel, B. A,, Rent, R., Myhorman, R . , and Gewurz, H. (1976). Immunology 30, 161. 337. Loos, M . , and Bitter-Sverman, D. (1976). Immunology 31, 931. 338. Boackle, R. J , , Johnson, B. J., and Caughrnan, G. B. (1979). Nature (London) 282, 742. 339. Lucas, T. J., Munoz, H., and Erikson, B. W. (1981).J. Zmmunol. 127, 2555. 340. Isenman, D. E., Ellerson, J. R . , Painter, R. H., and Dorrington, K. J. (1977). Biochemistry 16, 233. 341. Yachnin, S., Rosenbluin, D., and Chatman, D. (1964).J. Immunol. 93, 542. 342. Peltier, A. P., Cyna, L., and Dryll, A. (1978). Immunology 35, 779. 343. Loos, M., and Thiesen, R. (1978).J. Immunol. 121, 24. 344. Hughes-Jones, N. C., and Gardner, B. (1978). Immunology 34, 459. 345. Giclas, P. C., Ginsberg, M. H . , and Cooper, N. R. (1979).J. Clin. Invest. 63, 759. 346. Borsos, T., Rapp, H. J., and Walz, V. L. (1964). J. bnmunol. 92, 108. 347. Curd, J. G., and Cooper, N. R . (1978). J. Immunol. 120, 1769. 348. Allen, R., and Isliker, R. (1974). Itnmunochemistry 11, 243. 349. Fiist, G . , Medgyesi, G. A,, Rajnavolgyi, E., Csecsi-Nagy, M., Czikora, K., and Gergely, J. (1978). Immunology 35, 873. Immunochemistry 6, 461. 350. Colten, H. R . , Borsos, T. and Rapp, H. J. (1969). 351. Fust, G . , Bertok, L., and Juhasz-Nagy, S. (1977). Infect. Immun. 16, 26. 352. Bartholomew, R. M., and Esser, A. F. (1978).J. Immunol. 121, 1748. 353. Folkerd, E. J., Gardner, B., and Hughes-Jones, N. C. (1980). Immunology 41, 179. 354. Wright, J. K., Tschopp, J., Jaton, J. C., and Engel, J . (1980). Biochem. J. 187, 775. 355. Doekes, G., Van Es, L. A,, and Daha, M. R. (1982). Iinmunokgy 45, 705. 356. Tschopp, J., Schulthess, T., Engel, J., and Jaton, J. C. (1980). FEBS Lett. 112, 152. 357. Jaton, J. C., Huser, H., Riesen, W. F., Schlessinger, J., and Givol, D. (1976).J . Immunol. 116, 1363. 358. Hughes-Jones, N. C. (1977). Immunology 32, 191. 359. Cohen, S. (1968). I. Zmmunol. 100, 407.

216

NEIL R. COOPER

360. Hyslop, N. E., Dourmashkin, R. R . , Green, N. M., and Porter, R. R. (1970).J. E x p . Med. 131, 783. 361. Borsos, T., and Rapp, H. J. (1965). Science 150, 505. 362. Ishizaka, K., Ishizaka, T., and Banovitz, J. (1965). J. Zmmunol. 93, 1001. 363. Tschopp, J. (1982). Mol. Zmmunol. 19, 651. 364. Liberti, P. A,, Bausch, P. M., and Schoenberg, L. M. (1982). Mol. Zmmunol. 19, 143. 365. Schlessinger, J., Steinberg, I. Z., Givol, D., Hochmann, J., and Pecht, I. (1975). Proc. Nut!. Acud. Sci. U.S.A. 72, 2775. 366. Metzger, H. (1978). Contemp. Top. Mol. Zmmunol. 7, 119. 367. Lin, T. Y., and Fletcher, D. S. (1980).J. B i d . Chent. 255, 7756. 368. Siegel, R. C., Schumaker, V. N., and Poon, P. H. (1981).J. Zmmunol. 127, 2447. 369. Goers, J. W., Ziccardi, R. J , , Schumaker, V. N., and Glovsky, M. M. (1977).J. Zmmunol. 118, 2182. 370. Lepow, I. H., Ratnoff, 0. D., and Levy, R. (1958).J. Exp. Med. 107, 451. 371. Morgan, P. H., Walsh, K. A., and Neurath, H. (1974). FEBS Lett. 41, 108. 372. Collen, D. (1980). Thromb. Huemostasis 43, 77. 373. Fields, T., Ghebrehiwet, B., and Kaplan, A. P. (1983).J. Allergy Clin. Zmmunol. 72, 54. 374. Miiller-Eberhard, H. J., Polley, M. J., and Calcott, M. A. (1967).J. Exp. Med. 125, 359. 375. Polley, M. J., and Miiller-Eberhard, H. J. (1968).J . Exp. Med. 128, 533. 376. Thielens, N. M., Villiers, M. B., Reboul, A., Villiers, C. L., and Colomb, M. G . (1982). FEBS Lett. 141, 19. 377. Kerr, M. A., and Porter, R. A. (1979). Biochetn. J. 171, 99. 378. Nagasawa, S., and Stroud, R. M. (1977). Proc. Nutl. Acud. Sci. U . S . A . 74, 1998. 379. Schreiber, R. D., and Muller-Eberhard, H. J. (1974).J. E x p . Med. 140, 1324. 380. Bolotin, C., Morris, S . , Tack, B., and Prahl, J. (1977). Biochemistry 16, 2008. 381. Strunk, R., and Colten, H. (1974).J , Zmmunol. 112, 905. 382. Ziccardi, R. J. (1981).J. Zmmunol. 126, 1771. 383. Gigli, I., Ruddy, S., and Austen, K. F. (1968).J. Zmmunol. 100, 1154. 384. Dodds, A. N., and Porter, R. R. (1979). MoZ. Zmmunol. 16, 1059. 385. Goers, J. W., and Porter, R. R. (1978). Biochem. J. 175, 675. 386. Doekes, G . , VanEs, L. A., and Daha, M. R. (1982). Zmmunology 45, 705. 387. Doekes, G., VanEs, L. A., and Daha, M. R. (1983). Immunology 49, 215. 388. Laurell, A. B., Martensson, U., and Sjhholm, A. G. (1979). Acta Puthol. Microbiol. Scand. Sect. C 87, 79. 389. Hack, G. E., Hannema, A. J . , Eerenberg-Belmer, A. J., Out, T. A., and Arlberse, R. C. (1981).J. Zmmunol. 127, 1450. 390. Harpel, P. C., and Cooper, N. R. (1982). Clin. Res. 30, 563A. (Abstr.) 391. Harpel, P. C., and Cooper, N. R. (1984). Submitted for publication. 392. Dickler, H. B., and Kunkel, H. G. (1972).J . Exp. Med. 136, 191. 393. Sobel, A. T., and Bokisch, V. A. (1975). Fed. Proc., Fed. Am. SOC. E x p . B i d . 34, 965. 394. Cabay, Y., Perlrnann, H., Perlman, P., and Sobel, A. T. (1979).Eur. J. Zmmunol. 9,797. 395. Suba, E. A., and Csakao, G. (1976). J. Zmmunol. 117, 304. 396. Shadworth, M. F., Cunningham, P. H., and Andrews, P. S. (1979). Fed. Proc. Fed. Am. SOC. E x p . B i d . 38, 1075.

ADVANCES IN IMMUNOLOGY. VOL. 37

Membrane Complement Receptors Specific for Bound Fragments of C3 GORDON D. ROSS* AND M. EDWARD MEDOFt 'Division of Rheumatology-Immunology, Department of Medicine, and the Department of Microbiology-Immunology, University of North Carolina, Chapel Hill, North Carolina, and tDepartments of Pathology and Medicine, New rork University Medical Center, New York, New York

217 217 . . . . . . . . . . . . 219 221 221 222 223 225 226 111. Structure and Binding Site Characteristics of the Receptors . . . . . . . . . . . . . . , . . , 230 A. Complement Receptor Type One . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 B. Complement Receptor Type Two . . . . . . . 235 C. Complement Receptor Type Three , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 1). Complement Receptor Type Four . . . . 240 IV. Functions of C3 Receptors . . . , , . , . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . , , , 242 A. Functions of Erythrocyte C3 Receptors . , . . . . . . . . . . . . . . . . . . . . . . . . . . 242 B. Functions of Neutrophil C3 Receptors . . . . . . . . . . 25 1 C. Functions of Monocyte/Macrophage C3 rs . . . . . . . . . . . . . . . . . . . . . . . 255 D. Functions of Lymphocyte C3 Receptors 257 E. Functions of Kidney Podocyte C3 Receptors . . . . . . . . . . . . . . . . . . . . . . . . . , . , 260 F. Functions of Mast Cell C3 Receptors . . . . . . . . . 261 V. Conclusions . . . . . . ................................................ 261 References . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1

1. Background

A. INTRODUCTION One of the most important functions of the complement system is to coordinate interactions of host inflammatory cells with pathogenic substrates. This coordination is required for substrate recognition by appropriate cell types, initiation of relevant surface or intracellular processes, and cooperation between cell populations. Both afferent and efferent limbs of the inflammatory process are dependent upon this coordinating function of complement. Organization of host cell responses by complement is achieved through 217 Copyright 0 1985 b v Academic Prusa, Inc All rights of reproduction in any form reserved ISBN 0-12-022437-2

218

GORDON D . ROSS AND M . EDWARD MEDOF

the interaction of cell membrane receptors with activation products of serum complement proteins. Although a number of different complement components are involved in this complex process, C3l plays a pivotal role. During complement activation, C3-derived fragments are both liberated into the fluid phase and covalently bound to the substrate. The fluid-phase C3 fragments function to induce leukocytosis (C3e), increase vascular permeability (C3a), and suppress antibody synthesis (C3a), whereas the bound C 3 fragments participate in assembly of convertase enzymes that amplify C3 activation and further progression of the complement cascade. Because the bound C3 fragments also form ligands which attach the substrate to receptors of particular cell types, they are critical in determining the eventual fate of the substrate. In particular, fixation of bound C3 fragments onto bacteria leads to enhanced phagocytosis and bacteriacidal activity via phagocyte C3 receptors, and this is probably the most important function of the complement system. The membrane receptors for bound fragments of C3 are the subject of this review. Data concerning all aspects of structure, ligand specificity, and function are discussed. The mechanisms for generation and degradation of C3receptor ligands via the classical and alternative pathways are also reviewed. Data are included on the biosynthesis and genetic regulation of receptor expression where available. Finally, abnormalities of expression and/or function that have both clinical relevance and contribute to an understanding of the function of complement receptors in vivo are discussed. Additional information concerning these and other types of complement receptors can be found in other recent reviews (Fearon and Wong, 1983; Schreiber, 1983; Weigle et d.,1983; Fearon, 1983, 1984).

Abbreviations: B, D, H, and I, alternative pathway complement components factor B, factor D, factor H (previously PlH), and factor I (previously C3b-inactivator); BSA, bovine serum albumin; C4bp. C4-binding protein; C3, native third component of complement; C3i, C3 inactivated by disruption of internal thiolester bond; C3b, major fragment derived from C3 activation; iC3b. C3b that has been cleaved by factor I at two sites; C3dg, 41K M , (41,000 , M,. fragmolecular weight) fragment derived from iC3b by cleavage wit11 factor I; C ~ C140K ment product of factor I cleavage of iC3b; CRI, complement receptor type one, specific for c region in bound C3b, C4b, and iC3b; CR2, complement receptor type two, specific for d region site in C3b, iC3b, C3dg, and C3d; CRS, complement receptor type three, specific for iC3b, baker’s yeast, rabbit erythrocytes, and certain bacteria; DAF, decay accelerating factor; E, sheep erythrocyte; EAC, erythrocyte-antibody-complement complex, containing C3 fixed by way of the classical pathway; EBV, Epstein-Barr virus; EBV-R, EBV receptor; EC3, erythrocytes coated with C3 fixed by way of the alternative pathway; IA, immune adherence; NADG, N-acetyl-D-glucosaimine; NP40, Nonidet P40, a nonionic detergent; PMA, phorhol myristate acetate; PNH, paroxysmal nocturnal hemoglobinuria; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; SLE, systemic lupus erythematosus.

MEMBRANE COMPLEMENT RECEPTORS FOR

C3

219

B. HISTORY Historically, the study of cellular receptors for bound C3 fragments can be divided into three phases. The existence of such receptors was inferred in the first phase from observations that, under certain conditions both in uitro and in vivo, microorganisms adhered to blood cells (reviewed in Nelson, 1963). Studies by early investigators (Laveran and Mesnil, 1901; Levadite, 1901; Govaerts, 1920a,b; Wallace and Wormall, 1931) showed that such adherence reactions occurred in immune serum, and demonstrated that complement was required. The description (Nelson, 1953) of the immune adherence (IA) reaction in which immune complexes or microorganisms sensitized with antibody and complement attached to primate, but not nonprimate, erythrocytes demonstrated that human red cells could mediate the adherence, and that specific membrane factors (IA receptors) were involved (Nelson, 1956; Nelson and Nelson, 1959). Bacteria or soluble antigen-antibody complexes treated with complement remained adhered even after washing steps, indicating that the responsible complement factor was bound to the substrate. Subsequent in vitro studies with purified components showed the essential role of C3 (Nelson, 1956; Mills and Levine, 1959; Siqueira and Nelson, 1961; van Loghem and van der Hart, 1962). The observation (Nelson, 1953, 1956) that IA could enhance phagocytosis suggested that the phenomenon could be physiologically relevant. One major impact of this first phase in the investigation of C3 receptors was the exploitation of the specificity of the adherence reactions for quantitation of antigens, antibodies, or complement in the laboratory. In the second phase of investigation, quantitative techniques for evaluating receptor-ligand reactions were adopted, and studies were performed with individual cell types (reviewed in Bianco and Nussenzweig, 1977; Bianco, 1977; Ross, 1980). Of particular relevance was the introduction of the rosetting technique (Lay and Nussenzweig, 1968). This methodology permitted analysis of the specificity of receptor interaction with defined complement components that were fixed to either sheep (erythrocytes) E or other types of particles. Using rosetting techniques, it was found that fixed C4b, as well as C3b, could mediate IA (Cooper, 1969). It was also demonstrated that small lymphocytes derived from spleen and lymph nodes, but not thymus, rosetted with C3-coated sheep E (Bianco et al., 1970; Bianco and Nussenzweig, 1971). These complement receptor lymphocytes (CRL) coincided with the B lymphocyte population. Also of importance during this period was the characterization (Tamura and Nelson, 1967; Lachinann and MullerEberhard, 1968; Ruddy and Austen, 1969, 1971) of the serum enzyme C3b/C4b inactivator (factor I or I) and its serum cofactors plH-globulin (factor H or H) (Whaley and Ruddy, 1976a,b; Weiler et d . ,1976; Pangburn

220

GORDON D . ROSS A N D M . EDWARD MEDOF

et al., 1977; Pangburn and Miiller-Eberhard, 1978; Nagaki et al., 1978) and C4-binding protein (C4bp) (Ferreira and Nussenzweig, 1977; Nagasawa and Stroud, 1977; Scharfstein et al., 1978; Gigli et al., 1979). These control proteins were shown to play a central role in the breakdown of C3b and C4b into well-defined subfragments. The existence and distribution among various cell types of distinct receptors specific for certain C3b degradation products as well as for C3b itself were also demonstrated in this second period by rosette assay methods. One key observation was that normal B lymphocytes bound cell intermediates bearing either the large hemolytically active C3b fragment or the small C3 degradation fragment known as C3d (Ross et al., 1973; Eden et al., 1973; Okada and Nishioka, 1973), whereas leukemic lymphocytes from patients with chronic lymphatic leukemia primarily bound only C3d-bearing intermediates (Ross et al., 1973; Ross and Polley, 1975). Later polymorphonuclear and mononuclear phagocytes were shown to bind sheep E bearing another intermediate product of C3 degradation, iC3b, but not C3d (Ross and Rabellino, 1979). Another important finding was that tissue cells of the glomerular epithelium also expressed C3b receptor activity (Gelfand et al., 1975). During this second historical period, studies of the functional effects of receptor engagement also were undertaken. It was reported that engagement of C3b receptors on phagocytes was associated with opsonic activity manifested by a respiratory burst and induction of a number of intracellular processes including lysosomal enzyme release, generation of superoxide radicals, and phagocytosis (reviewed by Muller-Eberhard and Schreiber, 1980). It was subsequently shown that these biological responses were initiated primarily by virtue of the enhanced binding of substrate to the cells afforded by C3 and C3 receptors (Lay and Nussenzweig, 1969), and that the actual triggering of cellular functions was mediated by other receptor types (i.e., Fc receptors) (Ehlenberger and Nussenzweig, 1977; Newman and Johnston, 1979). The third and current phase of C3 receptor research emerged out of the ability to isolate, purify, and characterize the respective receptor molecules themselves. An important event in the inception of this phase was the isolation of the C3b/C4b receptor (Fearon, 1979) in the course of studies to identify membrane surface factors that restricted complement activation on host erythrocyte membranes. The solubilized 205K M , glycoprotein (gp205), initially detected by its ability to inhibit the alternative pathway C3-convertase, was subsequently identified (Fearon, 1980) as the C3b/C4b receptor when antibodies against it were shown to block CSb-dependent rosetting. Other key developments have been (1)the isolation of C3d receptors (initially in fragmented form) from culture supernatants of B lymphoblastoid

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

22 1

cells (Lambris et al., 1981) and subsequently in its native state by affinity chromatography of solubilized B-lymphocyte membranes, (Iida et al., 1983; Weis et a l . , 1984), and (2) the identification, in the course of an analysis of differentiation antigens on blood cell membranes, of a family of two-chain molecules (Sanchez-Madrid et al., 1983a,b), one of which is CR, and functions as an iC3b receptor (Beller et al., 1982). An additional key development was the demonstration that IA receptors (CR,) on primate erythrocytes were intimately involved in the process of immune complex clearance (Medof and Oger, 1982; Medof et a l . , 1982a,b, 1983a; Cornacoff et al., 1983), and played a unique role in degradation of C3b and iC3b (Medof et al., 1982qd; Ross et al., 1982; Medicus et al., 1983) (as well as C4b and iC4b) to CSdg (and C4d) (Medof and Nussenzweig, 1984b). A conceptual advance arising from these findings has been the recognition that membrane factors can participate in extracellular reactions, as well as intracellular events, and thus had to be regarded as components of the complement system and included in any analyses that would attempt to fully understand C3 and C4 metabolism and function. C. DEFINITION OF THE RECEPTORS Four distinct types of receptors specific for substrate-bound C3 fragments have so far been described. Rather than naming them according to their respective ligands as in the past, these receptors have been termed CR,, CR,, CR,, and CR, (with numbering according to their order of discovery). This new nomenclature is more appropriate because their interactions are not limited to single ligands, and in particular because iC3b binds to all four receptor types. Table I lists the specificity, structure, cellular distribution, and distinguishing monoclonal antibodies for these four types of receptors. Of the four receptors, CR,, CR,, and CR, have been studied in some detail and are now characterized both structurally and with respect to their binding properties. The fourth, CR,, has very recently been proposed as a result of some new observations. Although postulated in several laboratories, its existence is not yet uniformly accepted. Its structure is not yet unequivocally established and data that are available concerning its specificity are tentative. II. Generation of the Ligands for C3 Receptors

The C3 receptor ligands are derived from C3b fragments that become covalently bound to substrates in the course of complement activation. It should be emphasized that this C3b fixation process does not involve C3 receptors. As will be discussed in the following sections, C3b molecules first attach to the target of complement activation in a reaction independent of C3

222

G O R D O N D. ROSS A N D M . E D W A R D M E D O F

TABLE I MEMBRANERECEPTOIISFOR BOUND FRAGMENTS OF C3: CR1, CR2, CR3, A N D CRI STRUCTURE, SPECIFICITY, CELLTYPEDISTRIBUTION, A N D DISTINGUISIIING MONOCLONAL ANTIBODIES Receptor type

Cell type distribution

Monoclonal antibodies

Erythrocytes, granulocytes, monocytes, B and some T cells, kidney podocytes, dendritic reticulum cells B lymphocytes

44DU, 57FO. 57H, 31D C3RTo5, Ell Anti-B2IJ, HB-5"

Two chains: 165K a-chain 95K p-chain

Monocytes and macrophages, granulocytes, N K cells

Two chains: 150K a-chaine 95K P-chain"

Monocytes and macrophages, granulocytes, (NK cells?)

Anti-Mac-l' OKMl", OK M 9", OKMlO", MN-41, anti-Mol, antiLeu-1s" Anti-LeuM5n.C

Specificity

Structure

CR I

C3b>C411 >iC3b C3i, C3c

Four allotypes: 160K, 190K, 220K, 250K

CR2

iC3b=C3dg >C3d %C311 iC3b, S.

140K

CR3

ceriuiseae,

rabbit erythrocytes, S. epidermidis

CR4

iC3b=CJdg >C3d

Becton-Dickinson Co. Coulter Immunology Boeringer-Mannheim. Ortho Pharmaceutical Corp. If the putative CR4 turns out to be p150,95 (see discussion on CR4), then it would have a 150K a-cha and 95K &chain. Preliminary data suggest that anti-Leu-M5 may be specific for the a-chain of ~ 1 5 0 , ' (C. D. Ross, M. A. Arnaout, and L. L. Lanier, unpublished observation). f1

b

receptors. The C3b molecules that have fixed to the substrate may then interact with C3 receptors either directly or after normal degradation processes that expose other receptor binding sites not available in C3b.

A. COVALENT BINDINGOF C3b Fixation of C3b occurs as a result of uptake of nascent C3b fragments by the substrate surface. The nascent C3b is generated when C3 is cleaved by a C3-convertase that was previously assembled on the substrate. The metastable nascent C3b forms covalent bonds with hydruxy andlor amino groups present in protein, carbohydrate, and/or lipid moieties in the substrate (reviewed in Law, 1983; Tack, 1983). Hydroxyester and amidoester bonds are formed when acyl groups of the activated internal thiolester within the

M E M B R A N E COMPLEMENT RECEPTOHS FOR

C3

223

cw '-chain of nascent C3b encounter nucleophilic oxygen or nitrogen atoms within the hydroxy or amino radicals. At physiological pH, hydroxyester bond formation occurs preferentially (Law et al., 1984). The acyl groups on nascent C3b can also react with nucleophilic oxygen atoms in water and hydrolyze. Because of the high relative concentration of water, only 5-20% of the nascent C3b that are generated bind to the substrate. The short halflife of nascent C3b resulting from competition by water for the metastable binding site allows only those C3b generated in the immediate vicinity of the C3-convertase to be bound, and is responsible for the fact that the C3b uptake occurs in a clustered fashion around C3-convertases.

B. ACTIVATIONOF C3 BY

THE

CLASSICAL A N D ALTERNATIVE PATHWAYS

C3-convertase can form on substrates in two ways. Classical pathway C3convertase complexes, C4b2a, can assemble when complement fixing antibodies directed against the substrate are present (reviewed in Porter and Reid, 1978). Alternative pathway C3-convertase, CsbBb, can assemble in the absence of antibody when the surface characteristics of the substrate are such that deposited C3b fragments are not subject to immediate inactivation (see below) and are thus able to interact with factors B and D (reviewed in Muller-Eberhard and Schreiber, 1980). The reactions involved in formation of the two C3-convertases are shown schematically in Fig. 1.

1 . C3b Deposition by the Classical Pathway A key step in assembly of the classical pathway C3-convertase on the substrate is the deposition of C4b fragments. The C4b fragments are deposited upon interaction of C4 with substrate-bound C1 (reviewed in Porter and Reid, 1978; Goers and Porter, 1978; Kerr, 1980). Binding of a single IgM or an IgG antibody doublet is sufficient for uptake and activation of C1. Upon interaction with antibody by the C l q component of C1, the C l r and C l s components are converted from proenzymes to active serine esterases and acquire the ability to cleave C4 (reviewed in Ziccardi, 1983). Following cleavage by the activated C1, an internal thiolester bond in the a-chain of C4 (similar to that of C3) is activated, generating nascent C4b. Nascent C4b, like nascent C3b, binds covalently via hydroxyester and amidoester bonds. The deposited C4b fragments serve as binding sites for C2, that then may be cleaved by C l s such that the C2 zymogen is converted to enzymatically active C2a. Classical pathway C3-convertase activity is manifested by the C4b2a complex. 2. C3b Deposition by the Alternative Pathway The deposition of C3b fragments constitutes a critical step in the assembly of the alternative pathway C3-convertase that is analogous to the uptake of

A t c3

c4tc2

\ A R

R

-

R

R

R

R

R

B

C3tBtij

R

I

R

t

c3

R

R

R

R

FIG. 1. Assembly of C3-convertase and deposition of C3b on substrates. (A) Classical pathway: Interaction of C4 with activated C1-antibody complexes (Cl-Ab) generates nascent C4b. The nascent C4b fragments form hydroxylester and amidoester bonds with nucleophilic acceptor groups (R) in the antibody or on the substrate surface in close proximity to their site of generation. The bound C4b provide the required binding sites for C2. The same C1-esterase that cleaves C4 also cleaves C2. Cleavage of C2 to C2a and formation of the classical pathway C3-convertase C4b2a complex can only occur on bound C4b fragments that are positioned such that C2 is in close proximity to Ab-C1 complexes. The C2a portion of the C4b2a complex is a serine esterase, but it is only active when complexed with C4b. The C4b2a complex spontaneously and irreversibly dissociates with a half-life of 5 minutes at 37"C, and this dissociation is accelerated by C4bp and DAF. Interaction of C3 with the substrate-bound C3-convertase complex (C4b2a) generates nascent C3b. The nascent C3b fragments, in turn, form covalent bonds with substrate acceptor groups (R) in the immediate vicinity of their site of formation adjacent to C4b2a, but do not bind to C4b2a. R groups for C3 (or C4) may be carboxyl groups of carbohydrates or amino groups of proteins, and are not specific C3 receptors. (B) Alternative pathway: The inherent instability of the internal thiolester in the native C3 molecule spontaneously and continously generates small amounts of fluid-phase and bound C3* (C3i). This C3* forms a magnesium-dependent complex with factor B (B), such that the B in the complex becomes labile to cleavage by the serine protease factor D (D). Cleavage of B in the complex generates bound C3*Bb complexes that form the initial C9convertase of the alternative pathway. Interaction of additional C3 with C3*Bb leads to cleavage of C3 into C3a fragments and nascent C3b, followed by covalent attachment of a proportion of the nascent C3b to substrate acceptor groups (R) in the immediate vicinity of the C3-convertase. These bound C3b then form C3bBb C3-convertase sites that cleave more C3. As with the C4b2a complex, spontaneous decay dissociation of C3bBb (half-life 2 minutes at 37°C) is accelerated by a control protein, in this case H rather than C4bp. However, properdin (not shown) associates with C3bBb bound to activating surfaces, generating a C3bBbP complex with extended half-life of 20 minutes at 37°C. This action of properdin that stabilizes bound C3-convertase, as well as the action of H that rapidly dissociates both fluid-phase C3bBb and C3bBb bound to host membranes (see text), works to focus the C3-convertase on the substrate surface rather than in the fluid phase. Further interaction of B and more C3 with these complexes extends C3b deposition outward to acceptor sites progressively more distant from the site of the initial C3-convertase formation.

M E M B R A N E C O M P L E M E N T RECEPTORS FOR

C3

225

C4b. In contrast to the classical pathway process, discrimination between appropriate and inappropriate sites of activation occurs after C3b deposition rather than before (reviewed in Muller-Eberhard and Schreiber, 1980). The initial deposition is not induced or focused but rather is “spontaneous.” It occurs as a consequence of continuous low grade conversion of the C3 thiolester bond into a metastable state. It has been proposed that metastable C3 is first hydrolyzed to C3i, which then reacts with factors B and D to form a fluid phase C3 convertase. This fluid phase convertase then generates nascent C3b, which, in turn, can bind to substrate (reviewed in Law, 1983). It has alternatively been proposed that metastable C3 condenses directly with substrate acceptor groups to form covalently bound C3i. Irrespectively, once the initial C3b or C3i (C3*) is deposited, it serves as a binding site for factor B. Factor D, which is in the fluid phase, rather than cell bound like Cls, cleaves the C3*-bound factor B zymogen to enzymatically active Bb, forming the alternative pathway C3-convertase, C3* Bb. Once assembled on the substrate, both classical and alternative pathway C3-convertases induce accumulation of fixed C3b clusters in their immediate vicinity. The alternative pathway enzyme differs from the classical pathway enzyme in two important ways. First, although both C4b2a and C3bBb decay by “spontaneous” dissociation, binding of properdin (P) to C3b in the C3bBb complex stabilizes the alternative pathway enzyme. This stabilization greatly enhances the ability of the enzyme to augment further C3b uptake. Second, because factor D is in the fluid phase and exists in an activated state, conversion of the C3b-bound factor B zymogen can occur anywhere, rather than only in the proximity of bound Cls, as in the case of C4b-bound C2. Therefore, deposition of C3b is not limited to the immediate vicinity of the first site of spontaneous C3b fixation, and this permits alternative pathway mediated C3b uptake to propogate topographically and extend outward from the site of initial C3-convertase assembly. This ability to incorporate C3b into progressively more distant acceptor sites markedly enhances the number of C3b fragments that can be deposited and the extent to which saturation of the substrate surface with C3b can occur. C. FACTORS CONTROLLING C3 ACTIVATION Several mechanisms focus the assembly of C3-convertases on substrates and not on host cells. In the case of the classical pathway C3-convertase, C4b2a, formation of the enzyme is governed by substrate-bound specific antibody. With the alternative pathway C3-convertase, C3bBb, enzyme assembly is controlled by the relative affinities of B and H for deposited C3b. These affinities and thus the outcome may vary on different surfaces (reviewed by Muller-Eberhard and Schreiber, 1980). For example, the absence of sialic acid or sulfated mucopolysaccharides on certain cell or bacterial

226

GORDON D. ROSS A N D M . EDWARD MEDOF

surfaces is associated with a lower affinity binding of H (1 x lo6 M - l) than of B (5 x lo6 M - 1 ) for fixed C3b (Kazatchkine et al., 1979). This results in an ineffective restriction of B binding to C3b, permitting efficient assembly of C3-convertase on the surface. By contrast, the presence of other as yet unidentified entities in certain bacterial capsules (Brown et al., 1983) is correlated with a reduced binding affinity of B and consequent inefficient enzyme formation. Additionally, host blood cell membranes contain a 70K M , protein termed decay accelerating factor (DAF), which markedly enhances decay dissociation of C3-convertases (Nicholson-Weller et al., 1981, 1982; Pangburn et al., 1983a,b). As will be discussed in more detail in the Section IV,A, this membrane control protein plays a central role in preventing assembly of convertases on host cells. D. DEGRADATION OF C4b

AND

C3b

BY

CLEAVAGE WITH FACTOR I

When accumulated on the substrate in sufficient quantity, fixed C3 fragments become ligands for C3 receptors. As soon as C3b is fixed, it becomes subject to the proteolytic action of the serum factor I enzyme, so that degradation of C3b can occur to a variable extent concurrently with uptake. The Imediated degradation of substrate bound C3b (and C4b) are diagrammed in Figs. 2 and 3. The extent of I-mediated cleavage prior to receptor engagement depends upon several factors. Important among these is the accessibility (discussed above) of the deposited C3b to H . If the surface characteristics are such that H binding is restricted, then I-mediated breakdown is impaired and, as a result, C3b accumulates rapidly and can remain intact. If on the other hand, the microenvironment of the C3b permits H binding, then I-mediated breakdown is augmented and C3b is converted to iC3b. Bound C4b is also subject to the proteolytic action of factor I. Similarly, C4bp, which is a functional analog of factor H, promotes the I-mediated breakdown of C4b. In practice, different substrate surfaces may express a heterogeneity of microenvironments, so that a combination of C3b and iC3b (and C4b and iC4b) may be generated in various proportions. I-mediated conversion of C3b to iC3b involves cleavage of the C3b a'chain at two sites (Harrison and Lachmann, 1980). A 3 K M , fragment termed C3f is liberated, and the resulting 3-chain iC3b fragment remains covalently bound via the aminoterminal portion of its a'-chain ( ( ~ ' 2to) the substrate. A change in conformation may occur in the molecule, since following this cleavage, binding sites become exposed for conglutinin (Lachmann and Muller-Eberhard, 1968), monoclonal anti-C3g (Lachmann et al., 1982), and CR, (Ross et al., 1983a). I-mediated conversion of C4b to iC4b parallels that of C3b to iC3b. Only a single cleavage site in iC4b has been described so far however (Nagasawa et aZ., 1980). Although factor H and C4bp can greatly augment formation of substrate-

MEMBRANE COMPLEMENT RECEPTORS FOR

C3b R

C3b

H a’

NH2-

Q

~

R

4 s p

Q‘

I

+11 I S

i C3b

-P

a’

NH2-

I S

R

l

Q’

I

I

S

S

vp

I t cu1

NH2-

lL I

S

c3c

C4b

I

s - Y

C3d (g 1

~

s s - Y I + C R 1 or C4bp

I*CR1 or H

NH2

227

C4b

N

4 -P

C3

iC4b

s

I t CR1 (or C 4 b p )

I a‘

d

I P

NH2

C4d

lL +P

a‘

7

I

I

c4c I

s -Y

I

s

FIG. 2. Schematic representation of I-mediated breakdown of substrate-bound C3b and C4b fragments. Hydroxyester or amidoester bonds link the C3b and C4b fragments via their a’chains to substrate acceptor groups (R). C3h (left side of diagram): In the presence of factor H (H) or CRI, &tor I (I) cleaves the a’-chain of C3b at two closely spaced sites. This cleavage releases a 3KM, fragment termed C3f, and generates the 3-chain iC3b fragment (a‘268K, a’l 43K, and p 76K Mr). In the presence of CRI (but not H or C4bp). I can cleave the a’2 68K chain of iC3b, releasing the 140KM, C3c fragment into the fluid-phase and leaving the 41K M , C3dg fragment bound to the substrate. Although the 8K M , C3g fragment may be cleaved from the C3dg fragment by various proteases forming fluid C3g and bound C3d (33K MJ, this final breakdown usually does not occur in blood. It is unknown whether leukocyte elastase generated at sites of inflammation cleaves fixed C3dg down to C3d. C4b (right side of diagram : In the presence of C4bp or CR1, I cleaves the a’-chain of C4b into the 4-chain iC4b intermediate (a‘ 75K, a’ 17K, p 75K, and y 31K). In the presence of CR,, the a’75K chain of iC4b is cleaved hy factor I, releasing the 140K M , C4c fragment and leaving the 45K M , C4d fragment bound to the substrate. This reaction is greatly enhanced when C3b are clustered around iC4b (see also Fig. 3). High concentrations of C4bp* can also promote breakdown of iC4b into C4c and C4d if C4b is deposited in high density clusters.

bound iC3b and iC4b in vitro, their roles in the breakdown of fixed C3b and C4b in viva are less well understood than in the control of fluid-phase C3b and C4b. In the case of fluid-phase C3b and C4b, these control proteins must be present to prevent unfocused complement activation. The large amount of nascent C3b that is hydrolyzed in the course of C3 activation could react with B and D in solution to form fluid-phase C3bBb. If some activated C1 escapes control by C1 inhibitor, C4b could likewise enter into

228

G O R D O N D . ROSS A N D M . E D W A R D M E D O F

R

R

R

R

FIG. 3. Formation of C3 receptor ligands. C3b clusters ran either interact with H or with CR1, displacing Bb. Interaction of C3b3b with H results in I-mediated conversion to iC3bi3b (releasing 3K M , C3f fragments), but does not permit further iC3b breakdown, probably because of the low affinity of H for fixed iC3b (Ross et d.,1983a). Interaction of C3b3b with CR1 also results in conversion of the C3b3b to iC3bi3b, but can additionally lead to further Imediated breakdown of the iC3bi3b to C3dg3dg. This latter cleavage releases C3c fragments into the fluid phase. The subsequent interaction of I and H-generated iC3bi3b with CR, can similarly result in production of C3dg3dg and C3c. The ability of CHI to support iC3b fragmentation may be a consequence of its capacity to cluster in the membrane and form multipoint bonds with the iC3b. C4b3b clusters can interact with C4bp and H, or with CRI, releasing C2a, and permiting I-mediated cleavage to iC4bi3b. *C4bp alone may be able to support this reaction, because C4bp may serve as an I-cofactor with both C3b and C4b to iC3b and iC4b, both in the fluid-phase (Fugita and Nussenzweig, 1979) and bound (Gottlieb and Medof, unpublished observation). Interaction with CR, results in I-mediated conversion of iC4bi3b to C4d3dg. At high C4b densities, C4bp can also support cleavage of iC4b to C4d, perhaps because it exists as a multimer of 7 covalently bonded subunits (Dahlback et al., 1983). that may be able to mediate multipoint binding.

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

229

formation of fluid-phase C4b2a. Concomitant C3 turnover in the fluid-phase is prevented by serum C4bp, H, and I. Because of their high concentrations in serum (500 and 250 p,g/ml, respectively), H and C4bp can rapidly bind to fluid C3b and C4b, and prevent assembly of the fluid-phase enzymes. Following assembly of C3-convertase on the substrate, fixation of additional C3b depends upon maintenance of focused C3 activation at the substrate surface. Indeed, in the absence of H (Fearon and Austen, 1977; Schreiber et al., 1978; Fujita et al., 1981) or C4bp (Gigli et al., 1979), C3 is consumed in the fluid-phase and C3b fixation is markedly impaired. Furthermore, in patients with genetic deficiencies of H or I, little C3 or B is available in the blood because of continuous fluid-phase consumption. In the case of the substrate-bound C3b and C4b, as will be discussed below, the I-mediated conversion of C3b to iC3b and of C4b to iC4b can be supported by membrane CH,, in addition to serum H and C4bp. Moreover, further degradation of iC3b and iC4b is mediated principally by CR,. The relative participation of H and C4bp as compared to CR, in formation of fixed iC3b and iC4b is unknown. With C3b on sheep erythrocytes, CR, can be >103-fold more efficient on a weight basis than H in promoting the conversion of C3b to iC3b (Medof and Nussenzweig, 1983). CR, can likewise be more efficient than C4bp in the breakdown of cell bound C4b to C4c and C4d (Medof and Nussenzweig, 1984a,b). It is possible that the relative eEciencies of membrane associated CR, and of the serum factors in these conversions may vary however, depending upon the substrate. The effect of the surface microenvironment on the accessibility of bound C4b to C4bp has not been studied. The 68K M , 01'2 peptide of iC3b is highly sensitive to further proteolysis in the region between the cysteine residue which participates in the thiolester bond and that which disulfide bonds this peptide to the p chain (Lachmann and Muller-Eberhard, 1968; Pangburn et al., 1977; Natsuume-Sakai et d., 1978; Gaither et al., 1979). Cleavage at one or more sites in this region results in release of three-chain C3c fragments (140K M,) into the fluid-phase, leaving either 41K M , C3dg or 33K M,. C3d fragments (both single-chain) remaining bound to the substrate (Ruddy arid Austen, 1971; Bokisch et al., 1975; Law et at., 197%; Lachmann et al., 1982; Ross et al., 1982). In blood, this fragmentation probably occurs only upon cell interaction (with CR,), since when serum alone is used it requires 18-24 hours at 37°C (Lachmann et al., 1982). As will be discussed below, factor I can mediate this further breakdown and release when fixed iC3b is attached to CR, (Medof et al., 1982e; Ross et al., 1982; Medicus et al., 1983). Other enzymes including trypsin (Gitlin et al., 1975; Lachmann et al., 1982), serum plasmin (Nagasawa and Stroud, 1977),and notably leukocyte elastase (Johnson et al., 1976;Taylor et al., 1977;Carlo et al., 1979, 1981), and cathepsin G (Spitznagel et al., 1974;

230

GORDON D . ROSS AND M . EDWARD MEDOF

Carlo et aZ., 1981), also efficiently cleave iC3b into C3dg and then C3d. However, the proteolysis with these other enzymes is markedly inhibited in serum (Carlo et al., 1981; Lachmann et al., 1982). Moreover, C3dg fragments rather than C3d fragments are usually detected in vivo on blood cells from patients with autoimmune disease (Lachmann, 1981; Ross et al., 1984~). Whether these proteases or serum kalikrein, which has recently been shown to degrade fluid phase iC3b into a C3dg like fragment termed C3d-K (Meuth et al. , 1983), normally participates in the fragmentation of substrate-bound iC3b and generation of physiologically relevant fixed C3dg (or C3d) ligands remains to be determined. A complete discussion of the binding specificities of each C3 receptor for the above described bound fragments of C3 (and C4) is contained in the sections following on the individual receptor types. Briefly (see Table I), CR, binds with high affinity to C3b and with lower affinity to C4b and iC3b. CR, binds with high affinity to iC3b and CSdg, with somewhat lower affinity to C3d, and with very low affinity to C3b. CR, binds only to fixed iC3b among C3 fragments. Interestingly, recent studies (Ross et al., 1984a) indicate that CR, also binds directly in the absence of fixed C3 to yeast, rabbit erythrocytes, and certain bacteria. CR, appears to have a specificity very similar to CR,, except that binding to C3b has not been detected.

111. Structure and Binding Site Characteristics of the Receptors

The CR,, CR,, and CR, membrane proteins have been examined using a number of physicochemical methods, and their ligand interactions studied in both the purified and membrane-associated state. Information available concerning the structure of each of the proteins, and of the as yet unisolated putative CR,, is summarized together with data that has been obtained on the chemistry of their interaction with respective ligand(s). In the cases of CR, and CR,, some data about their biosynthesis are available, and have been included.

A. COMPLEMENT RECEPTOR TYPEONE (CR,)

1 . Structure When CR,, extracted from human erythrocytes, is analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), one major band is observed, indicating that it is a single-chain molecule (Fearon, 1979). The M , is -30K greater under reducing conditions (Dykman et al., 1983a,b), suggesting that internal disulfide bonds are present. The M , of the reduced molecule was initially reported to be 205K. CR, binds to lentil

MEMBRANE COMPLEMENT RECEPTORS FOR

C3

231

lectin and elutes with a-methylmannoside showing that it contains carbohydrate (Fearon, 1979). On gel filtration in the presence of 0.1% Nonidet P40 (NP40) detergent, it elutes with a much higher M,. of 1-1.2 x lo6, suggesting that it can exist in multimers of 4-6 subunits. Although CR, was initially thought to be homogeneous with respect to size, recent studies have demonstrated that it is polymorphic. Most of the earlier studies employed pooled erythrocytes as a source of CR,. The polymorphism of the molecule was subsequently observed when CR, in erythrocytes from a number of individual subjects was compared by surface labeling, immunoprecipitation (or affinity chromatography), SDS-PAGE, and autoradiography (Dykman et al., 1983a; Wong et al., 1983). In initial analyses, two CR, species differing by -30K M , were noted. The smaller 190K M , species was found to be present in 70% of individuals, the larger (220K M,) present in 3%, and both species present in about 27%. Family studies suggested that expression of these three phenotypes is regulated by codominant alleles at a single autosomal locus. The more common allele (termed A in one study and F in another) has a gene frequency of 0.83 and the less common (termed B and S ) a frequency of 0.16. In subsequent studies, two additional CR, species have been detected (Dykman et al., 1984a,b). These species of 160K and 250K M,., respectively, are determined by rarer alleles designated C and D , with gene frequencies of 0.01 and -0.001, respectively. The 160K and 250K CR, variants have been observed so far only in association with the type A or B variants. A curious feature of this polymorphism is that, in the heterozygous state, the relative expression on erythrocytes of different CR, species may vary markedly (9O:lO to 10:9O) between different individuals, but is constant in a given subject (Dykman et al., 1983a, 1984a,b). Also, on leukocytes there is characteristically greater expression in heterozygotes of the high-molecular-weight variant (Dykman et al., 1983b). This polymorphism is unique among complement proteins because of the large variations in M,, substantially exceeding those of the a-chain of C4, which are due to differences in carbohydrate content (Karp et al., 1982). Treatment with endoglycosidase-F decreases the M , of individual CR, species by about 10K (Wong et d.,1983; Dykman et d.,1984b), providing further evidence for a high content of sugars in CR, structure, but does not alter the relative mobilities on gels of the two common forms, indicating that variations in N-linked oligosaccharides cannot account for the size difference. Whether variations in O-linked oligosaccharides play a role has not been established. Another observation made in the course of investigation of this polymorphism is the presence, in association with each erythrocyte CR, species, of an accompanying “minor CH, band” that is 15K M , larger in size (Dykman et al., 1983a). The relative proportion or increment in size of the

232

GORDON D . ROSS AND M . EDWARD MEDOF

minor band is not diminished by reduction and alkylation nor is it altered by treatment with endoglycosidase-F.

2 . Biosynthesis Studies of the biosynthesis of CR,, performed with the HL-60 promyelocytic cell line, have provided some information concerning intracellular processing of the molecule (Atkinson and Jones, 1984). These cells, when stimulated, express CR, on the cell surface detectable by rosette assay with C3b-coated erythrocytes, When CR, synthesis was induced in the presence of [35S]methionine and the biosynthetically labeled CR, analyzed, an intracellular precursor that was 22K M , smaller (188K M,) than the surface CR, (210K M,) was detected. Treatment with endoglycosidases showed that the intracellular molecule had a high mannose content while the surface CR, had a N-linked complex-carbohydrate composition. The half-time for newly synthesized receptor to attain an M , of 210K is -45 minutes.

3. Variation on Dqferent Cell Types Earlier studies with polyclonal antibodies raised against erythrocyte CR, had found that CR, was immunochemically identical on different blood cell types (Fearon, 1980). More recent studies in which CR, from different cell types has been isolated and examined directly by SDS-PAGE support structural identity of erythrocyte CR, with the CR, of monocytes and B cells, but indicate that the CR, of polymorphonuclear leukocytes, and perhaps also that of T cells, is -5K M , larger. However, no antigenic differences between the different sized CR, molecules have been detected with either monoclonal or polyclonal antibodies (Dykman et al., 1984a; Wilson et al., 1983). The significance of this observation is not yet known. CR, molecules present on the various blood cell types have not been directly compared physicochemically in other ways, e.g., by electrophoresis, nor has CR, from renal podocytes or tissue macrophages been isolated and examined. 4 . Binding Site Characteristics Intact C3b binds to CR, with high-affinity and has been presumed to serve as the principal CR, ligand. The percentage of CR, cells that rosette with C3b-bearing cellular intermediates increases as a function of C3b density (Cooper, 1969). Clustering and multivalent presentation of the C3b is important for the CR, interaction. C3b dimers bind with higher affinity than do C3b monomers (Arnaout et al., 1981, 1983a). Also, complement-activating immune complexes or particles bind to CR, in blood, despite the presence of simultaneously generated (monomeric) fluid-phase C3b. Fixed C4b can also bind to CR, as evidenced by its ability to promote rosetting with CR,-bearing cells. Very high C4b densities are required, +

MEMBHANE COMPLEMENT RECEPTORS FOH

C3

233

however, and the binding is weaker than that mediated by fixed C3b (Cooper, 1969). When C3b is deposited by the classical pathway, and is thus clustered around C4b2a complexes, rosetting with CR, cells is enhanced by the adjacent C4b (Bokisch and Sobel, 1974). This finding suggests that cooperation with C3b could be the major ligand function of C4b physiologically. Fixed iC3b also serves as an important ligand for CR, (Ross et al., 1983a; Medof and Nussenzweig, 1984a). As with C3b, the binding of iC3b to CR, increases with iC3b density (Ross et al., 1983a; Medof and Nussenzweig, 1984a). Because the binding affinity of iC3b is lower than that of C3b, the importance of iC3b clustering is greater and the enhancing effect of nearby C4b more pronounced. Either C4b or iC4b can participate with classical pathway-derived iC3b in the formation of an effective CR, ligand, and either C3b or other nearby iC3b fragments with alternative pathway-derived iC3b (Medof and Nussenzweig, 1984a). The interaction of these various ligands with CR, is physiologically relevant, because it represents the primary mechanism for both clearance of CR,-associated circulating immune complexes and the degradation of fixed iC3b into fixed C3dg and fluid C3c (see below). If complexes are not cleared and fixed iC3b is not degraded, then the complexes may bind avidly to neutrophil CR, by way the fixed iC3b in the complexes, and degranulation that produces tissue injury may result from the interaction of the CR,-attached complexes with neutrophil Fc receptors (see Sections IV,A and B). Although monoclonal anti-CR, has permitted more convenient CR, isolation, quantitation, and structural analysis, its ability to block selectively different epitopes of CR, has not yielded important insights concerning the nature of the CR, ligand binding site(s). The inaliility of certain monoclonal anti-CR, antibodies to compete with each other (Iida et al., 1982)for binding to CR, has indicated that mice recognize at least 4 distinct CR, epitopes. By contrast, analysis of uptake of C3b dimers versus rabbit polyclonal anti-CR, has suggested that as many as 9 rabbit anti-CR, binding sites were available per CR, molecule (Wilson et al., 1982). Rabbit anti-CR, totally block C3b binding to CR,. Although some monoclonals block more effectively than others, their abilities when used individually to inhibit EC3b rosetting to CR,-bearing cells are highly dependent on C3b density (Jida et ul., 1982; Gerdes et nl., 1982; Hogg et al., 1984). Furthermore, the ability of certain monoclonals to block CR, activity varies in different systems, e.g., rosetting of EAC1423b with tonsil lymphocytes versus immune complex binding to erythrocytes (Iida, Medof, and Nussenzweig, unpublished observations). Because a ligand-induced organization of individual CR, into clusters is required for efficient receptor function (Dierich and Reisfeld, 1975; Abrahamson and Fearon, 1983), antibodies that do not block the CSb-binding site +

234

GORDON D . ROSS AND M. EDWARD MEDOF

may still be able to inhibit particle binding to cells by preventing this required clustering of receptors. Both hydrophobic and charge interactions may participate in the binding of ligands to CR,. Elution of solubilized CR, from C3b-Sepharose requires both high salt and nonionic detergent, and does not occur with either salt or detergent alone (Fearon, 1979). Also, binding of either the isolated receptor or receptor-bearing cells to either C3b, iC3b, or C4b-bearing intermediates or C3b dimers is stronger at low ionic strength (Dobson et al., 1981; Arnaout et al., 1981; Ross et al., 1983a). Several lines of evidence indicate that the ligand binding sites of C3b and C4b reside in the c region of the fragment. Both iC3b and iC4b retain affinity for CR,, but C3dg and C4d do not. In addition, fluid-phase C3c inhibits EAC14b and EAC1-3b rosettes (Ross and Polley, 1975), and C3c coated microspheres form rosettes inhibitable by anti-CR, (Ross and Lambris, 1982). The affinity of fluid-phase monomer binding to CR, is very low, since fluid-phase (monomeric) C3c is not retained by CR,, and the interaction of it or fluid-phase C3b with CR, can be demonstrated only by inhibition of binding studies (Arnaout et al., 1981). The higher affinity of a C3b dimer apparently results from the presence of two active binding sites per dimer, as C3b conjugated to IgC binds to CR, with lower affinity than does a C3b dimer. This latter finding also argues against the importance of conformation changes introduced into C3b as a result of covalent association with substrate (Arnaout et al., 1983a). Although the binding of C4b to CR, appears to be weaker than that of Cab, the relative abilities of bound C3b and C4b to interact with CR, are not precisely quantitated. One problem in this analysis has been the preparation of particles with not only the same densities of the two ligands but also the same organization of the ligands on the particle. Since sites of deposition of C4b are in principle constrained to the location of antibody-bound C l s (reviewed in Porter, 1983), it may be more difficult to form CR, reactive C4b clusters than C3b clusters. Recent findings that, at neutral pH, the C4 isotype derived from the C4 A locus preferentially transacylates onto amino, but not hydroxyl, group nucleophiles (Isenman and Young, 1984; Law, 1984; Law et al., 1984) could further limit the ability of C4b, relative to that of C3b, to cluster. Studies that have attempted to equalize C4b and C3b deposition on intermediates and address this problem (Cooper, 1969) have found that 3-4 times more deposited C4b than C3b was required for comparable immune adherence. In addition, inhibition of CR,-dependent rosetting required more fluid-phase C4b than C3b (Ross et al., 1978a). While C3b may be the principal CR, ligand and C4b a secondary ligand limited to a supportive role in the presence of intact complement, C4b has been shown to function as an opsonin in C3-deficient individuals.

M E M B R A N E C O M P L E M E N T RECEPTORS FOR

C3

235

The influence of the microenvironinent of deposited C3b (or C4b) on the ability of the C3b (or C4b) to interact with CR, has received only limited attention. Initial studies found that, in contrast to the interaction of factor H with cell-bound C3b, the interaction ofisolated CR, with bound C3b was not altered by the prior removal of membrane sialic acid (Fearon, 1979). It has subsequently been reported that interaction of membrane-associated CR, with cell-bound iC3b on sialic acid-poor rabbit erythrocytes is much less efficient than with bound iC3b on sialic acid-rich sheep erythrocytes (Medicus et al., 1983). CR, interaction with fixed iC3b on zymosan is similarly inefficient compared to that with iC3b on sheep cells. Variation in the pattern of C3b deposition on different substrates arising from differences in the nature of nucleophilic C3b acceptor sites or differences in C3b clustering could play roles in these effects. These differences in C3b deposition on different substrate could be important for CR, function and require future studies for clarification (also see Section IV, B, 1).

B. COMPLEMENT RECEPTOR TYPETwo (CR,) 1 . Structure The C3d receptor of B lymphocytes was identified in 1973 by Ross et al., Eden et al., and Okada and Nishioka. It is now known that intact CR, consists of a single glycoprotein chain of 140K M,. (gp140). The molecule was first isolated from Raji B-lymphoblastoid cells by Bare1 et al. (1981), and thought to be a C3b receptor because of its binding and elution from C3bagarose. In the same year, Lambris et aZ. (1981) isolated a 72K M , glycoprotein (gp72) from Raji cell spent culture media on C3d-agarose, demonstrated that rabbit antibody to gp72 inhibited all CR,-mediated rossetting of B lymphocytes, and concluded that gp72 was CR,. Iida et al. (1983), and later Weis et nl. (1984), demonstrated that isolated gp140 bound to C3d-agarose and ECSd, and concluded that gp140 was CR,. At this time however, it was uncertain if membrane gp140 had the same CR, activity as isolated gp140. In particular, the two gpl40-specific monoclonal antibodies used in these studies (anti-B2 and HB-S), caused only partial inhibition of intact cell CR, activity. Frade et al. (1984) later demonstrated that rabbit anti-gpl40 inhibited all CR, activity of Raji cells and normal B-lymphocytes. In addition, it was also shown that absorption with purified gp72 blocked all rosette-inhibiting activity of the anti-gp140, and also inhibited the uptake of radiolabeled anti-gpl40 onto Raji cells by 55%. It is therefore probable that gp72 represented a fragment of CR, containing the C3d-binding site. Details of gp140 and gp72 isolation procedures have been recently reviewed (Ross and M yones, 1984).

236

G O R D O N D. ROSS A N D M . EDWARD M E D O F

2 . Binding Site Properties Elution of solubilized CR, from C3-agarose required both high-salt and nonionic detergent, suggesting that the activity of the binding site involves both charge and hydrophobic interactions (Iida et al., 1983; Weis et al., 1984). The importance of charge interactions was also indicated by the enhancement of CR,-dependent rosette formation observed in low-ionic strength buffers (Ross et al., 1983a). The binding specificity of CR, has been primarily characterized from tests of EC3 rosette formation with Raji cells, other B cell lines, and normal B cells. Raji cells, that express large amounts of CR, and no other type of C3 receptor, formed rosettes with EC3bi or EC3dg bearing as few as 1000 molecules of the respective C3 fragment per E. Removal of the 8K M, C3g fragment from EC3dg caused an apparent reduction in the binding affinity of EC3dg for CR,, as Raji cell CR,-dependent binding to EC3d required at least 2000 molecules of C3d per E (Ross et al., 1983a). This may indicate a secondary binding site for CR, in the g region of C3dg or iC3b. However, because saturating amounts of monoclonal anti-C3g had no effect on the binding of EC3dg to Raji cell CR, (Ross et aZ., 1983a), an alternative possibility is that either the conformations of C3d and C3dg may differ or that the strong anionic charge of C3g (Lachman et al., 1982) may contribute to the binding affinity of C3dg by interaction with a cationic region of CR,. In addition to its high-affinity for iC3b, C3dg, and C3d, several observations suggested that CR, also had a low-affinity for C3b: (1) lymphocyte EAC13d rosettes were inhibited by fluid-phase C3b (Ross and Polley, 1975; (2) a low-affinity binding site for fluid-phase C3b was identified on Raji cells (Frade and Strominger, 1980), despite the absence of detectable Raji cell CR, (Lambris et al., 1980; Iida et al., 1982; Tedder et al., 1983); and (3)CR, was successfully isolated on C3b-agarose (Bare1et al., 1981). At first, however, it was unclear that such “C3b” reactions with CR, were not actually due to very small amounts of undetected contamination of C3b reagents with iCSb, C3dg, or C3d. For example, at least 100,000 molecules of fixed C3b per E were required to produce weak rosettes with 40-60% of Raji cells (Weis et al., 1984; Frade et al., 1984), whereas only 1000 molecules of fixed iC3b or C3dg per E produced a similar degree of Raji rosetting (Ross et al., 1983a; Frade et al., 1984). Thus, it appeared possible that such EC3b rosetting might have been produced by as little as 1% contamination with fixed iC3b or C3dg. Further complicating interpretation of the specificity of the Raji cell-EC3b rosette reaction was the finding that Raji cells could convert rosette-negative EC3b into rosette-positive EC3bi with an intrinsic secreted factor I (Lambris et al., 1980). The question of fixed C3b binding to CR, was finally resolved by use of a special monoclonal anti-C3d that had a specificity

MEMBRANE COMPLEMENT RECEPTORS FOR

C3

237

very similar to CR, (Frade et al., 1984). This monoclonal anti-C3d bound with high affinity to EC3bi, EC3dg, and ECSd, but its binding affinity for EC3b was estimated to be 100-fold lower. Very small amounts of this antiC3d completely inhibited Raji cell rosette formation with EC3bi, ECSdg, or EC3d, but had no effect on Raji cell rosettes with EC3b. Thus, the EC3b rosettes could not have been due to small amounts of contaminating iC3b, C3dg or C3d, as the binding activity of such contamination would have been blocked by the anti-C3d. Because anti-CR, did inhibit all Raji cell EC3b rosettes (as well as all rosettes with ECSbi, ECSdg, and EC3d) it was concluded that CR, had a very low &nity for fixed C3b that was on the order of 100-fold less than its affinity for fixed iC3b or CSdg. Studies of CR, binding of fluid-phase radiolabeled C3 fragments have only been reported for fluid-phase C3b binding to Raji cells (Frade and Strominger, 1980). These studies estimated approximately 5 x lo4 C3b-binding sites per cell and a relatively low affinity of 1 x lop6 M . Complete exposure of the CR,-binding site in C3 probably results from a conformational change that occurs when C3b is cleaved by factor I. Because CR, rosettes are inhibited by fluid-phase C3b and soluble CR, readily binds to C3b-agarose, the CR,-binding site may be better exposed in fluid-phase C3b than in fixed C3b. It appears unlikely however, that CR, reacts with fluid-phase C3b in oioo because fluid-phase C3b has such a short half-life in blood. Evidence has also been presented that CR, may serve as the B cell receptor for Epstein-Barr virus (EBV). Both CR, and EBV receptors (EBV-R) were shown to be coexpressed on normal B lymphocytes and B cell lines Uondal et al., 1976; Yefenof et al., 1976). Initial studies demonstrated inhibition of EBV-R by fluid-phase C3, EBV inhibition of EC3d rosettes, and cocapping of EBV-R and CR, (Yefenof et al., 1976). Later however, a putative EBV-R was isolated and compared to isolated gp72 (now recognized to be a fragment of CR, containing the C3d-binding site), and it was concluded that CR, was not the EBV-R (Hutt-Fletcher et al., 1983). The isolated EBV-R, though of similar size to intact CR, (145K M,.), did not have detectable C3d-binding activity. In addition, the isolated gp72 did not have the same EBV-R inhibiting activity as did the isolated EBV-R. Finally, a rabbit antibody to gp72, that inhibited EC3d rosettes, did not detectably inhibit EBV-R activity. Not reported at that time was the finding that a rabbit anti-EBV-R (that was not monospecific and reacted with other B cell membrane antigens besides the EBV-R) did have very potent CR,-inhibiting activity (G. D. Ross and J. G. Simmons, unpublished observation). Recently, with the recognition that gp72 represented only a fragment of CR,, isolated intact CR, was examined for EBV-R activity (Fingeroth et al., 1984). These studies demonstrated that when isolated CR, was linked to S. aureus by way of a noninhibitory IgG monoclonal antibody (HB-5), the resulting

238

GORDON D. ROSS A N D M. EDWARD MEDOF

CR,-bearing bacteria bound 'Z5I-labeled EBV. As a control, it was shown that bacteria coated in a similar manner with HLA antigen did not bind 1251labeled EBV. Thus, isolated CR, appears to have EBV-R activity resembling intact B cells and may represent the B cell EBV-R. Additional experiments are required to determine whether rabbit antibody to intact CR, inhibits EBV-R activity, since antibody to the CR,-binding site (rabbit anti-gp72) did not inhibit EBV-R. Perhaps the EBV-binding site of CR, is different from the Cad-binding site, and located at the opposite end of the molecule. C. COMPLEMENT RECEPTORTYPETHREE(CR,)

1 . Structure CR, is an iC3b receptor on phagocytic cells and the lymphocytes that function in natural or antibody-dependent cytotoxicity (NK/K cells). CR, was initially identified as a receptor on phagocytic cells that was distinct from lymphocyte CR, because it bound EC3bi but not EC3d (Ross and Rabellino, 1979; Carlo et al., 1979; Ross and Lambris, 1982). Subsequently, the monoclonal antibody known as anti-Mac-1 was shown to inhibit phagocyte rosettes with EC3bi but not with EC3b (Beller et al., 1982). This finding suggested that anti-Mac-1 was specific for either CR, or for an antigen that was positioned near enough to CR, on the cell membrane that anti-Mac-1 could sterically block the CR, binding site. Evidence that CR, was the same as Mac-1 came from the identification of three patients who had an apparent genetic deficiency of both the Mac-1 antigen and CR, activity (Ross et al., 1983b, 1984b). Finally, another monoclonal antibody specific for the Mac-1 antigen, known as OKM1, that did not inhibit iC3b binding to CR,, was used to link detergent-solubilized OKMl antigen to S. aureus bacteria. These OKM 1 antigen-coated bacteria were then shown to bind to EC3bi in much the same way as did CR,-bearing phagocytic cells (Wright et al., 1983a). Thus, the Mac-1/OKM1 antigen is probably the same as CR, because (1)several monoclonal antibodies to the antigen (anti-Mac-1, antiMol, OKM10, MN-41, and anti-Leu-15) inhibit CR,-dependent EC3bi rosettes (Beller et d., 1982; Arnaout et d., 1983; Wright et d., 1983a; Eddy et al., 1984; Ross et al., 1984a); (2) CR,-deficient leukocytes do not express this antigen (Ross et al., 1984b; Klebanoff et al., 1984); and (3) the isolated antigen binds to EC3bi but not to EC3b (Wright et al., 1983a). CR, is a member of a family of three membrane antigens that have structurally identical 95K M , P-chains, each linked noncovalently to one of three distinct a-chain types (Sanchez-Madrid et al., 1983b). The a-chain of CR, is 165K M,, and the a-chains of the other two antigen family members, LFA-1 and p150,95, are 185K and 150K M,, respectively. Both the a-and p-chains contain carbohydrate (Kurzinger and Springer, 1982), and both chains are

MEMBRANE COMPLEMENT RECEPTORS FOR

C3

239

exposed on the membrane surface to radioiodination and reaction with chain-specific antibodies. When analyzed by either tryptic mapping or reaction with polyclonal antibodies, the P-chain showed no structural homology with the a-chains of either CR, or LFA-1. In addition, there were no similarities detected between the a-chains of CR, and LFA-1 (Sanchez-Madrid et al., 1983a). Finally, biosynthetic studies showed that a-and P-chains were synthesized separately and then joined together after glycosylation (Ho and Springer, 1983).

2. Binding Site Properties CR, appears to have a complex lectin-like binding site with specificity not only for iC3b, but also for baker’s yeast (Saccharomyces cervisiae), rabbit erythrocytes (Ross et al., 1984a), S. epidermidis (Ross et al., 1984b), and possibly also S. aureus (Klebanoff et al., 1984). A sugar specificity of the CR, binding site was suggested by the inhibition of CR, by N-acetyl-D-ghcosamine (NADG), and the binding of CR, to the protein-free yeast cell wall extract known as zymosan (Ross et al., 1984a). NADG inhibition of CR, required use of EC3bi bearing small amounts of fixed iC3b, and little inhibition was observed with EC3bi bearing large amounts of fixed iC3b. However, despite the finding that 50-100 mM NADG was required to inhibit CR,dependent EC3bi rosettes, concentrations of NADG as high as 200 mM were found to have no detectable effect on CR,, CR,, or Fc receptors (Ross et al., 1984a). In comparing CR, to typical lectins such as Con A or lentil lectin, it should be noted that elution of glycoproteins from these other lectins can require concentrations of a-methylmannoside as high as 2 M . Additional experiments are required to demonstrate the individual specific sugars in zyinosan and staphylococcus that bind to CR,. CR, resembles bovine serum conglutinin (K) in that it binds to both fixed iC3b and zymosan, and is inhibited by either EDTA or NADG (Ross et al., 1983a, 1984a). Because the origin and function of bovine K were unknown, it was of interest to determine if CR, might be the human homolog of K. However, the structure of K (six 48K M , subunits) was found to be distinct from that of CR, (Davis and Lachmann, 1983),and surface-bound K did not serve as the iC3b receptor of bovine leukocytes (Ross et al., 1984a). Thus, CR, is probably not the human homolog of bovine K, and the origin and function of K remain unknown. All four C3 receptor types can bind to iC3b, so each is in fact an iC3b receptor. With large amounts of iC3b per ECSbi, neutrophils and monocytes can bind EC3bi via CR,, CR,, and CR, simultaneously. Likewise, lymphocytes can bind EC3bi via CR, and/or CR, (B and some T cells) or via CR, (NK/K cells). For this reason, measurement of CH,-specific EC3bi rosettes may require prior blockade of CR, and/or CR, with receptor-specif-

240

GORDON D . ROSS AND M. EDWARD MEDOF

ic antibodies (Ross et al., 1983a). CR, differs from the other C3 receptor types in that it binds only to iC3b and not to other types of C3 fragments, and that its activity is blocked completely by EDTA. Expression of full CR, activity requires both calcium and magnesium divalent cations (Wright and Silverstein, 1982; Ross et al., 1983a). The possibility that CR, might be able to bind directly to unopsonized zymosan was initially suggested by the similarity of CR, and K (Ross et al., 1983a), and the identification of patients with a genetic deficiency of the CR3/LFA-l/p150,95 antigen family whose leukocytes did not give a phagocytic or respiratory response to zymosan (Ross et a!., 1984b). Because treatment of normal neutrophils with any of 4 different anti-CR, a-chain specific monoclonals, but not with monoclonal anti-LFA-1 a-chain, inhibited responses to zymosan, the deficiency of CR,, and not the deficiencies of LFA-1 or p150,95 was probably responsible for the absent response to zymosan. Although macrophages synthesize and secrete C components (Whaley, 1980) that could potentially opsonize particles (Ezekowitz et al., 1984; Johnson et al., 1984), several lines of evidence indicate that the binding of zymosan to CR, does not require phagocyte-derived fixed iC3b on the zymosan (Ross et al., 1984a). First, there was no effect on zymosan binding to neutrophil CR, when assays were performed in the presence of amounts of Fab anti-C3c that were 10-times greater than those required to block EC3bi binding to CR, completely. Second, the zymosan-binding site in CR, appears to be distinct from the iC3b-binding site in CR,. Three monoclonal antibodies to the a-chain of CR, (anti-Mac-1, anti-Mol, and Mn-41) blocked binding of both EC3bi and zymosan to CR3. However, anti-Leu-15 selectively inhibited EC3bi binding to CR,, whereas OKM 1 selectively inhibited zymosan binding to CR,. Binding of rabbit E was also blocked by OKM1, suggesting that these particles bind to the same site in CR, that binds to baker’s yeast or zymosan (Ross et al., 1984a).

D. COMPLEMENT RECEPTORTYPEFOUR(CR,) 1 . Structure CR,, a C3d receptor of phagocytic cells, is structurally distinct from CR,, but has similar ligand specificity. Neither polyclonal nor monoclonal antibodies to CR, bound to CR,-bearing cells, and a polyclonal anti-CR2 that completely inhibited CR, activity had no effect on CR, activity (Frade et al., 1984). Two preliminary findings suggest that CR, may be p150,95, the third member of the CR,/LFA-l antigen family. First, EC3d binding activity was absent on macrophages that had been allowed to adhere to culture dishes

MEMBRANE COMPLEMENT HECEPTORS FOR C 3

24 1

coated with monoclonal anti-@-chain(Wright et a l . , 1984). This treatment presumably caused complete modulation of all membrane antigens containing the common @-chain(CR,, LFA-1, and p150,95). Because adherence of macrophages to either anti-CR, a-chain or anti-LFA-1 a-chain coated plates caused no inhibition of EC3d rosettes, the anti-P-chain-mediated inhibition of EC3d rosettes may have resulted from modulation of p150,95. A second reason that CR, may be p150,95 is that CR, activity (EC3dg rosettes) was undetectable on neutrophils from a patient with genetic deficiency of the CR3/LFA-l/p150,95 antigen family (G. D. Ross, unpublished observation). Because neither anti-CR, nor anti-LFA-1 inhibited normal neutrophil EC3dg rosettes (Frade et al., 1984), the deficiency of p150,95 may explain the CR, deficiency of this patient’s cells. Proof of the identity of p150,95 and CR, will require demonstration that isolated p150,95 has CR, activity (i.e., binds to EC3d), and that antibodies to the a-chain of p150,95 inhibit the CR, activity of intact cells.

2 . Binding Site Properties Neutrophil and monocyte CR, bind EC3bi and EC3dg bearing relatively large amounts of fixed C3 per E. Although CR, also binds to EC3d, rosettes have only been observed with in uitro differentiated monocytes that are presumed to express greatly increased numbers of CR, per cell. Formation of EC3dg rosettes with neutrophils or monocytes required >40,000 C3dg molecules per E. However, even with EC3d bearing 100,000 C3d per E, no EC3d rosettes were observed with normal blood neutrophils and monocytes. The differentiation of isolated blood monocytes into macrophrage-like cells capable of binding EC3d required cultivation in media supplemented with either fetal bovine serum (Inada et al., 1983) or phorbol myristate acetate (PMA) (Wright et d.,1984). Because the CR, of stimulated monocytes did bind to EC3d, the primary specificity of CR, may be C3d. The C3d specificity of neutrophil EC3dg rosettes is supported by the demonstration that EC3dg rosettes were inhibited by fluid-phase C3d complexes prepared from detergent-solubilized EC3d (Ross et al., 1983a). In addition, neutrophil uptake of 1251-labeledfluid-phase CSdg was inhibited by excess unlabeled fluid-phase C3d (Vik and Fearon, 1984). Both fixed and fluid-phase iC3b also bind to CR,. With EC3bi bearing 10,000 iC3b molecules per E, all neutrophil rosetting activity is CR, dependent and inhibitable by anti-CR,. As the amount offixed iC3b is increased from 10,000 to 40,000 molecules per E, neutrophils begin binding EC3bi to both CR, and CR, (Ross et al., 1983a). With EC3bi bearing >45,000 iC3b molecules per E, neutrophils whose CR, and CR, have been blocked with inhibitory amounts of both anti-CR1 and anti-CR, form apparent CR,-dependent rosettes (Frade et al., 1984). CR, binding of fluid-phase iC3b was confirmed by demonstration that fluid-phase

242

GORDON D. ROSS A N D M. EDWARD MEDOF

iC3b inhibited the uptake of radiolabeled fluid-phase C3dg (Vik and Fearon, 1984). EDTA inhibited the CR,-dependent rosetting activity of both EC3dg with neutrophils (Ross et al., 1983a) and EC3d with macrophages (Wright et al., 1984). However, this may not indicate a binding site requirement for divalent cations, as EDTA did not inhibit neutrophil uptake of either fluidphase C3dg (Vik and Fearon, 1984) or C3dg-coated microspheres (Ross et al., 1983a). EDTA can also cause membrane shape perturbations by chelation of membrane calcium, and perhaps such membrane rearrangements diminish the accessibility of CR, to large C3dg-bearing particles. IV. Functions of C3 Receptors

Analysis of the functional properties of the receptors for bound C3 fragments has permitted a better understanding of the mechanisms of a number of cellular processes, and has led to new insights into the roles of different cell types in inflammatory responses. Investigation of CR, function has been facilitated by its presence alone (without other C 3 receptor types or Fc receptors) in erythrocytes, and by the ease of working with this cell type. In contrast, studies of the function of CR,, CR,, and putative CR, have been complicated by the fact that these receptors are present only in combinations on normal cells, and by the greater difficulty in working with nucleated cells that can alter their surface expression, as well as release internal factors. Moreover, because methods for CR, purification have been available longer and erythrocytes may be obtained in large quantities, there have been many more functional studies of isolated CR, than of the other receptors in isolated form. More is known, therefore, about the biological properties of the CR, glycoprotein than about those of the other C3 receptor molecules.

A. FUNCTIONS OF ERYTHROCYTE C3 RECEPTORS CR, is the only type of C3 receptor that has been detected on erythrocytes. Although immune adherence was first described over 30 years ago in the initial phase of C 3 receptor research (see above), the significance of this phenomenon was unknown. The realization that erythrocyte CR, accounted for >90% of the total CR, content in blood led to speculations that red cells had an important nonrespiratory role in certain immune functions (Siegal et al., 1981). The central role of the IA receptor remained unclear, however, until the demonstration that erythrocytes, by way of CR,, had a major role in the processing and clearance of circulating immune complexes.

1 . Function of Erythrocyte CR, The detection in the red cell membrane of an H-like activity (Fearon, 1979), in the course of studies of alternative pathway inhibition, demon-

M E M B R A N E COMPLEMENT RECEPTORS FOR C 3

243

strated that the CR, molecule, like H, could displace Bb from bound C3b and supply similar cofactor activity in the I-mediated cleavage of C3b to iC3b. These observations led to further investigation of the effect of CR, on regulation of the complement cascade. Purified CR, bound to fixed C4b, as well as fixed C3b, on sheep E (Dobson et al., 1981). It also resembled C4bp (Iida and Nussenzweig, 1981) in that it could displace C2a from C4b, and provide 1-cofactor activity in the cleavage of C4b. The ability of CR, to inactivate classical pathway C3 and C5 convertases, as well as the alternative pathway enzyme, further supported the notion that CR, was a complement inhibitor. Studies with soluble immune complexes suggested that CR, could be important in the “processing” of fixed C3b (and C4b), and that the red cell surface could constitute an important site of this processing (Medof and Oger, 1982). When soluble Ag:Ab:C complexes, prepared with bovine serum albumin (BSA), anti-BSA, and diluted serum (as a source of complement), were added to unfractionated blood cells, the immune complexes bound predominantly to erythrocytes. Similar competition by erythrocytes for immune complexes was found with other antigedantibody systems (Medof et al., 1982b; Hekmatpanah et al., 1982), and red cells were employed successfully for quantitation of immune complexes of various types (Aikawa et al., 1979; Tsuda et al., 1980; Pederson et al., 1980), indicating that the phenomenon was a general one. When erythrocytes were isolated after addition of immune complexes, and the red cells bearing complex-associated C3b in their receptors were reincubated in serum reagents, the complexes rapidly dissociated from the cells (Medof et al., 1982b). Studies using purified factors and Ag:Ab:C prepared with 12sI-labeled C3 revealed that the release of complexes was mediated by factor I and was associated with generation of C3c (Medof et al., 1982a). After release, the immune complexes showed enhanced reactivity with cells bearing CR, (lymphocytes), but reduced binding to PMN (bearing CR,, CR,, and CR,) (Lam and Medof, 1982). The reaction occurred independently of H. Moreover, the red cell binding could not be prevented by preincubation of the Ag:Ab:C with high concentrations of either I alone or I and H (Medof et al., 1982c, 1983b). These findings suggested that iC3b in immune complexes, as well as C3b, was a ligand for erythrocyte CR, and that interaction with CR, was required for I-mediated breakdown of iC3b and generation of C3c. In other studies (Medicus and Arnaout, 1982), factor I was shown to cleave iC3b on human erythrocytes much more efficiently than iC3b on sheep erythrocytes. When the kinetics of soluble immune complex interaction with red cells in undiluted serum were analyzed, a series of reactions were observed (Medof et al., 1983a,b). Following initial binding of the Ag:Ab:C to erythrocyte CR, via classical pathway generated C3b and C4b, additional C3b was incorporat-

244

GORDON D. ROSS AND M. EDWARD MEDOF

ed into the Ag:Ab:C via the alternative pathway. As this C3b accumulated, it was repetitively bound to the receptor and was concurrently degraded by factor I. The C3 was bound covalently to both antibody (Takahashi et d., 1976; Gadd and Reid, 1981) and antigen (Medof et al., 1982d; Takata et d., 1984). During this process, immune complexes remained closely associated with the erythrocyte membrane (Medof et al., 1983b) and were sequentially modified. One important effect of this modification was to enhance their reactivity with cells of the reticuloendothelial system. In vivo studies in Rhesus monkeys (Cornacoff et at., 1983)showed that, during transit through the liver and spleen, Ag:Ab:C that were initially bound to erythrocytes were subsequently removed and the erythrocytes returned to the circulation. Although CR,-mediated cleavage of C3b by factor I provides a mechanism for transfer of the immune complexes to these tissues, the cell type and receptors involved have not yet been identified. Nevertheless, in demonstrating that the competition for immune complexes by erythrocytes occurs in uiuo, these studies in monkeys (Cornacoffet al., 1983)have lent important support to the concept that the red cell interaction is a physiologically relevant process. Additional evidence for the importance of this mechanism of immune complex clearance may come from the study of red cells from patients with systemic lupus erythematosus (SLE) (see Section IV,A,3). Studies with purified CR, and cellular intermediates bearing well-defined C3 fragments have permitted precise characterization of the mechanism of the red cell reactions and the red cell factors involved. They have shown that the effect of erythrocytes on I-mediated C3b and iC3b cleavage resides entirely in the CR, molecule (Medof et al., 1982e; Ross et al., 1982; Medicus et al., 1983), and have demonstrated that CR, binds to iC3b (Medof and Nussenzweig, 1984). Intact-cell CR, binding to iC3b (Ross et nl., 1983a) has also been demonstrated. If C3b on cellular intermediates is first converted to iC3b by I and H, C3c is released upon subsequent exposure to I and CR,. The cells may also be treated with CR, alone and then washed before addition of I. Likewise, these studies have permitted comparison of the efficiencies of CR, and H in supporting various C3b cleavages by factor I. Although H can promote fragmentation of fixed iC3b (into fixed CSdg and fluid C ~ Csee , Section II,D, and Fig. 2) at low ionic strength (Ross et at., 1982), CR, promotes this I-mediated breakdown >104-foldmore efficiently, and the reaction occurs at isotonic, as well as reduced ionic strength (Medof and Nussenzweig, 1983). CR,, on a weight basis, can also promote the conversion of C3b to iC3b on sheep EAC >103-fold more efficiently than H (Medof and Nussenzweig, 1983). Although this may vary with the substrate, the resistance of C3b andlor iC3b to H and I would allow interaction of complexes with CR, on cells in the blood. Recent studies have shown that CR, interaction with substrate-bound C4b

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

245

fragments and I results in degradation to C4c and C4d (Medof and Nussenzweig, 1984, and see Fig. 3). Although it was previously shown that I plus C4bp could promote this fragmentation (Gigli et al., 1979), very high concentrations of I and C4bp, and high densities (>3000 molecules/cell) of deposited C4b were necessary to demonstrate the C4bp enhancing effect (Fujita and Tamura, 1983). The cleavage by CR, occurs 103-fold more eficiently on a weight basis than that by C4bp and, perhaps of more significance, is augmented >lO-fold by neighboring C3b fragments. Likewise, the I plus CR, cleavage of bound C3b or iC3b is reciprocally enhanced by neighboring C4b (Medof and Nussenzweig, 1984b). These findings are consistent with the notion (see above) that the ligand for CR, can be C4b3b. The lesser efficiency of inactivation of C4b prior to C3b uptake could provide a mechanism of relative protection of the classical pathway C3-convertase, C4b2a, until sufficient C3b had accumulated for efficient CR, binding. It has been generally held that C3b receptors were present only on primate, and not on nonprimate, erythrocytes (reviewed in Nelson, 1963). The ability of platelets (Nelson, 1956; Nelson and Nelson, 1959) to mediate immune adherence in nonprimates led to the suggestion that nonpriinate erythrocytes do not function in C3b processing. Recently, however, rabbit erythrocytes have been shown to mediate immune adherence if rabbit C3 is used (Horstmann et al., 1984). Moreover, rabbit red cells promoted I-mediated cleavage, and thus possessed CR,-like cofactor activity. CR, isolated from mouse spleen cells has been shown to be a single protein of 190K M , (Kinoshita and Nussenzweig, 1984). The isolated CR, molecule, but not mouse H , promotes cleavage of cell-bound mouse iC3b to (140K M,) C3c and (40K) C3d(g), and otherwise appears to be functionally equivalent to CR, of human origin. These findings raise the possibility that CR,-mediated processing of immune complexes may be a general phenomenon in all mammalian species.

2. The Number of C R , per Erythrocyte An intriguing characteristic of CR,, that has relevance to an understanding of both CR, function and the genetic regulation of its synthesis, is that the number of CR, expressed per erythrocyte appears to be highly variable among individuals. CR, levels on erythrocytes from different donors have been shown to vary by as much as 10- to 100-fold whether assessed functionally by immune adherence (Miyakawa et al., 1981)or binding of dimericC3b (Wilson et al., 1982), or antigenically using polyclonal or inonoclonal antibodies (Iida et al., 1982; Wilson et al., 1982; Walport et al., 1984; Ross et al., 1984~).Although different values for the mean level were initially reported using different techniques, there now appears to be agreement on a

246

GORDON D. ROSS A N D M . EDWARD MEDOF

more consistent value of -500 sites per erythrocyte (640 ? 157, 423 ? 180, 480 & 160) using monoclonal antibodies (Ross et al., 1984c; Fearon, 1984; Medof and Nussenzweig, 1984b). Individuals who are apparently healthy have been observed occasionally with erythrocyte levels <50, while others have been noted with levels >1200. Although CR, levels on erythrocytes gradually decrease with erythrocyte age (Sim and Sim, 1984), analyses of serial samples from several normal individuals have indicated that erythrocyte CR, levels remain relatively constant (-+lo%) in a given subject. Some variation has been observed in healthy women, although the effect of intermittent iron deficiency, either subclinical or clinical, noted in some subjects has not been evaluated. Erythrocyte CR, levels can also vary in healthy women during pregnancy (Ouelette, Nussenzweig, and Medof, unpublished observations). Family studies have indicated that the level of erythrocyte CR, expression in healthy subjects is in part genetically controlled. Low levels are more frequently found in offspring of parents with low levels and vice versa. One study reported a trimodal frequency distribution of erythrocyte CR, levels in the population (Wilson et al., 1982). This study proposed that an autosomal dominant mode of inheritance, regulated by two codominant alleles (designated H and L) at a single locus, accounted for the distribution. According to this model, high CR, levels would be determined by the H H genotype, intermediate levels by the HL genotype, and low levels by the LL genotype. In contrast, two other studies of normal individuals in England (Walport et al., 1984) and in North Carolina (Ross et al., 1984c) found a bell-shaped frequency distribution of CR, levels with no indication of high or low subgroups, implying that the genetic control might be more complex, perhaps involving more than two alleles. There is no apparent relationship between the genetic regulation of CR, levels and that of CR, phenotypes constituting the size polymorphism discussed above (Wong et al., 1983; Medof et al., 1983b). An earlier study also reported an individual who had erythrocytes with defective immune adherence but normal CR,-dependent rosetting of leukocytes with C3b-bearing intermediates (Rothman et al., 1975). Studies of variation in the population of CR, levels on leukocytes have not yet been systematically performed nor have correlations been made in individual subjects between erythrocyte CR, levels and CR, levels on other cell types. Although earlier studies of C3b particle binding had suggested that CR, genes might be linked to the major histocompatibility complex (MHC), one of these studies (Curry et al., 1976) employed Raji cells which are now known to possess CR, rather than CR,, and the other (Gelfand et al., 1974) was done in mice and was later proven to be wrong (Ferreira and Nussenzweig, 1976). Recent studies (Hatch et al., 1984) have found that there is no relationship between the inheritance of CR, and of HLA types, indicating

MEMBRANE COMPLEMENT RECEPTOKS FOR C 3

247

that the genes for CR, and HLA are not closely linked. Interesting, on the other hand, in humans the gene for CR, is linked with that for the serum control protein, C4bp (Rodriguez de Cordoba et al., 1984). Although the location of these genes is unknown, this association may constitute a linkage group of functionally related molecules similar to that of C2, C4, and B, but outside of the MHC. 3. Diminished C R , Expression on Erythrocytes from Patients with Systemic Lupus Erythematosus (SLE) and Other Diseases

A number of studies have shown that expression of CR, on erythrocytes is diminished in patients with SLE (Miyakawa et al., 1981; Iida et al., 1982; Wilson et d . , 1982; Inada et al., 1982; Taylor et d.,1983; Horgan and Taylor, 1984; Ross et al., 1984c; Walport et al., 1984) and some other diseases thought to be associated with immune dysfunction. These studies are not only of interest clinically, but also are of interest because of their relevance to an understanding of the regulation and function of CR, in vivo. In initial studies (Miyakawa et aZ., IYSl), absent immune adherence activity was observed with erythrocytes from 66% of 56 patients with SLE in Japan. This erythrocyte abnormality persisted in 3 patients who were tested serially, despite treatment and remission of symptoms. Additionally, low or absent immune adherence was found with increased frequency among the relatives of patients with low immune adherence activity, suggesting that the defect might be of genetic origin rather than a manifestation of disease processes. No increased uptake of lZ51-labeled anti-human IgG could be demonstrated by the abnormal erythrocytes, arguing against blockage of receptors by immune complexes. Subsequent studies employing anti-CR, showed that CR, antigen, as well as CR, function, was deficient, indicating that the number of CR, molecules on the erythrocyte surface was diminished. In one study (Iida et al., 1982), erythrocytes from 34 unselected SLE patients in New York City were found to contain an average of 60% less CR, per cell than did erythrocytes from normal individuals. This difference was observed whether CR, antigen was quantitated either by uptake of 1251-labeledanti-CR, antibodies onto intact red cells, or by radioimmunometric assay after extraction of the cell membranes with NP40. This finding provided further evidence that the defect was probably not due to an inaccessibility of receptors caused by occupation with immune complexes. Diminished erythrocyte CR, levels correlated with reduced C4 hemolytic titers and increased C l q binding activities of serum, and in 2 of 4 individuals whose erythrocytes were tested longitudinally, significant increases in CR, occurred and coincided with remission. In another study (Wilson et al., 1982), erythrocytes from 38 patients with inactive SLE from Boston were found to contain CR, levels that averaged

248

GORDON D. ROSS AND M . EDWARD MEDOF

45% of normal. Control experiments could not demonstrate blockade of receptors by either immune complexes or autoantibodies to CR, in the patients’ plasma, and comparison of uptake of C3b dimers and anti-CR, again indicated that patients’ erythrocytes contained fewer CR, than did normal erythrocytes. In these patients, correlations were not found between diminished CR, and complement consumption. Moreover, when the distribution of the CR, levels in SLE erythrocytes was compared with that in normal erythrocytes, the most striking difference was the near absence of a patient group with high CR, levels (putative H H phenotype, see above), and a predominance of patients in a low CR, number group (putative LL phenotype). In addition, when families of the patients were analyzed, diminished erythrocyte CR, in patients correlated in all cases but one, with diminished erythrocyte CR, in parents. Thus, even though the studies of both Iida et al. (1982) and of Wilson et al. (1982) each directly demonstrated the SLE-related absence of erythrocyte CR, molecules, the first study suggested that acquired factors might contribute to reduced CR, levels whereas the second study concluded that genetic factors, and not acquired factors, were principally involved. As will be discussed below, studies of kidneys of SLE patients (Kazatchkine et at., 1982; Emancipator et al., 1983) have shown that the CR, in other sites is also effected. In view of the genetic regulation of CR, size, and in addition, the initial findings of a possible genetically determined CR, functional defect associated with SLE, an investigation was carried out to compare the functional properties of the individual CR, size phenotypes. All four of the known CR, size variants (A, B , C and D)were found to accelerate decay of C4b2a and promote I-mediated fragmentation of cell-bound C3b into C3c and C3dg with comparable efficiency (Medof et al., 1983b). Identical C3 fragments were generated when any of the four variants were added to fluid-phase C3b in the presence of factor I. In other studies, erythrocytes bearing type F o r S CR, were shown to bind dimeric C3b with the same affinity (Wong et al., 1983). When the frequencies of the various CR, phenotypes among SLE patients were examined, no significant differences from the general population were observed (Dykman et al., 1984b). An interesting, but unexplained, observation was that SLE patients that were heterozygotes for the types A and C CR,, expressed higher proportions of the type C variant relative to the type A variant than did normal individuals (Dykman et al., 1984b). More recent studies have indicated that both genetic and disease factors contribute to the regulation of erythrocyte CR, levels (Walport et al., 1984; Ross et al., 1984~).Even though the distribution of CR, levels among 138 normal individuals appeared to follow a nearly bell-shaped distribution, condominant inheritance of CR, numbers was suggested in studies of CR, levels

M E M B R A N E C O M P L E M E N T RECEPTORS FOR C 3

249

among family members. In the 14 normal families examined, there was a good correlation ( r = 0.81) between the mean CR, levels in parents versus their offspring. However, this same relationship of CR, levels between parents and their offspring (and also among siblings) was not observed in parallel studies of families of patients with SLE. When CR, levels on erythrocytes from 18 clinically-active and -inactive SLE patients in England (Walport et al., 1984; Ross et al., 1984c) were compared with erythrocyte CR, levels in their first degree relatives, no correlation was observed ( r = -0.19). In the individual families, SLE patients frequently had CR, levels that were lower than predicted by the codomiiiant inheritance patterns observed in normal families. I n one key family whose relationships were confirmed by HLA typing of lymphocytes, the SLE patient expressed only 379 CR, per erythrocyte (putative LL phenotype) whereas both of the patient’s parents and three normal siblings expressed >900 CR, per erythrocyte (putative H H phenotype). Similar findings were subsequently obtained in studies of SLE families in France (Wilson et al., unpublished observation) employing the same polyclonal anti-CHI reagent utilized earlier in the Boston study (Wilson et al., 1982). Studies of 79 SLE patients followed serially in North Carolina have indicated a negative correlation of CR, levels and disease activity. CR, levels decreased with increasing disease activity and CR, levels increased with remission (Ross et al., 1 9 8 4 ~ ) .

4 . Acquisition of Erythrocyte C3dg in Disease An interesting observation arising out of the above studies is that C3dg fragments can be detected on CR,-deficient erythrocytes from patients with SLE, and that erythrocyte CR, and C3dg levels are inversely correlated. Radioimmune assay of patients’ erythrocytes with monoclonal antibodies to C3c, C3g, a i d C3d indicated that disease activity and low CR, levels were associated with the presence of 200-500 molecules of fixed C3dg per erythrocyte (normal <80 C3dg per erythrocyte). With flares of disease activity, CR, levels decreased as much as SO%, and C3dg appeared on erythrocytes. Conversely, when disease activity indices indicated an improvement in disease activity status, then CR, and C3dg levels returned to normal values (Ross et al., 1984~).Similar findings of disease-associated loss of CR, and appearance of erythrocyte C3dg were also made in patients with a variety of other disorders associated with autoantibodies and/or complement activation. The changes were most striking in chronic cold agglutinin disease (Parker et al., 1984b), but also occurred to a lesser extent in autoimmune hemolytic anemia, paroxysmal nocturnal hemoglobinuria, acute glomerular nephritis, and mycoplasin pneumonia (Ross et al., 1 9 8 4 ~ ) . The mechanism of the relationship between diminished CR, levels and the appearance of C3dg on erythrocytes is not yet clear. In in uitro studies,

250

GORDON D. ROSS AND M. EDWARD MEDOF

erythrocytes in normal blood incubated with immune complexes (prepared with PM2 double-stranded DNA and SLE plasma-derived IgG anti-DNA) were shown to acquire -200 molecules of fixed CSdg during the process of immune complex binding and release from erythrocyte CR, (Ross et al., unpublished observation). Fixation of C3 to red cells occurred while CR, sites were occupied with complement-activating immune complexes, and the majority of C3 that was fixed to the red cells was bound to the immune complexes, as it was released at the same time as the complexes. However, -5% of the fixed C3 remained bound to the erythrocytes after the complexes had undergone I-mediated release from CR,. This suggests that nascent C3b that are locally generated by complexes attached to CR, can also bind to acceptor groups on the adjacent erythrocyte surface. The erythrocyte-bound C3b could then be subsequently converted to C3dg by I and CR,. Of greatest importance was that the amount of erythrocyte CR, was unchanged by the binding and release of immune complexes in this in uitro whole blood system, suggesting that the loss of CR, observed in patients might not occur in blood. It was hypothesized that the CR, of SLE patients’ erythrocytes might be lost during the process of immune complex transfer from erythrocyte CR, to the liver macrophage phagocytic system (Ross et al., 1 9 8 4 ~ ) .

5. Dqferential Roles of Erythrocyte CR, and DAF in Convertase Regulation Recent investigations of the respective functions in erythrocytes of CR, and of decay accelerating factor (DAF) (that is also present in erythrocyte membranes) provide further support for the primary role of CR, in processing C3b and C4b associated with immune complexes or other substrates. Although antibodies to DAF efficiently block its ability to accelerate decay of C3-convertases assembled on cellular intermediates, these same anti-DAF antibodies have no effect on the ability of intact erythrocytes to augment decay of bystander-cell-associated enzyme (Medof and Nussenzweig, 1984a). In contrast, anti-CR, antibodies completely block this intact erythrocyte activity. These findings suggest that CR, functions extrinsically on C3b (and C4b) associated with targets of complement activation, whereas DAF functions only intrinsically within the host cell membrane. If this concept is correct, DAF and not CR, may constitute an important membrane component responsible for the protection of host cells from complement-mediated damage, particularly CR,-bearing cells that must come into intimate contact with immune complexes or other complement activators. Indeed, DAF has been shown to be deficient in erythrocytes of patients with paroxysmal nocturnal hemoglobinuria (PNH) (Pangburn et al., 1983a,b; Nicholson-Weller et al., 1983), a condition associated with heightened susceptibility of blood cells to alternative pathway-mediated lysis. PNH erythrocytes have

MEMBRANE C O M P L E M E N T RECEPTORS FOR

C3

25 1

also been shown to have abnormalities in membrane glycophorin-a (Parker et al., 1984a) that may contribute to an increased efficiency of membrane bound C3-convertases (Parker et al., 1982).

B. FUNCTIONS OF

NEUTROPHIL

c3 RECEPTORS

Neutrophils express CR,, CR,, and CR,, and use these receptors primarily to generate firm binding to bacteria, yeast, or soluble immune complexes prior to either ingestion or cytotoxic reactions required for clearance.

1 . CR, Although peripheral blood neutrophils normally express only 3000 CR, per cell, this number is increased 20-fold in 30 minutes after neutrophil activation. Two different stimuli have been shown to cause this rapid increase in CR, per neutrophil: (1)treatment of neutrophils in whole blood with chemotactic factors such as C5a or f-Met-Leu-Phe, and (2) incubation of isolated neutrophils in buffer at 37°C (Fearon and Collins, 1983). Isolated neutrophils maintained in Hanks’ buffer at 37°C for 30 minutes exhibited a parallel increase in both CR, and CR, (M. J. Walport and G . D. Ross, unpublished observation), whereas no change in the numbers of receptors for C5a or f-Met-Leu-Phe occurs under these conditions (Fearon and Collins, 1983). Despite avid CR,-dependent binding of EC3b to neutrophils, little or no phagocytosis or respiratory burst is observed (Ehlenberger and Nussenzweig, 1977; Newman and Johnston, 1979; Ross et al., 1984a). In addition, no respiratory burst is observed after treatment of neutrophils with saturating amounts of polyclonal F(ab’)2-anti-CR, (Ross et d.,1984a). Treatment of neutrophils with phorbol myristate acetate (PMA) activates CR, in some way so that bound EC3b are ingested. However, PMA-treated neutrophils are still unable to mount a respiratory burst to EC3b (Wright and Silverstein, 1983). Undoubtedly, CR, does play a major role in phagocytosis by promoting a firm binding of particles to neutrophils that facilitates recognition by other receptors that do trigger ingestion such as Fc receptors (Ehlenberger and Nussenzweig, 1977; Newman and Johnston, 1979) or CR, (Ross et al., 1984a). However, the relative role of CR, versus the other types of receptors for fixed C3 probably varies greatly with different types of particles depending upon the relative amounts of fixed C3b, iC3b, and C4. Particles that activate the alternative pathway contain a variable amount of fixed C3b that is resistant to H and I breakdown into iC3b. Preliminary studies however, suggest that H- and I-resistant C3b on rabbit E may have a relatively low affinity for CR, (G. D. Ross, unpublished observation) that parallels its low affinity for H (Pangburn and Muller-Eberhard, 1978; Kazatchkine et al.,

252

GORDON D . ROSS A N D M . EDWARD MEDOF

1979). Thus, despite their coating of fixed C3b, alternative pathway activating particles may bind primarily to CR,. For example, with serum-opsonized zymosan (an alternative pathway-activating particle that contains both fixed iC3b and resistant C3b), responses by neutrophils were inhibited by anti-C3 or anti-CR,, but not by anti-CR, (Ross et al., 1984a). The inhibition by antiC3 demonstrated that fixed C3 was an active opsonin on the zymosan particles. However, the lack of inhibition by anti-CR, demonstrated that fixed C3b binding to CR, did not contribute significantly to the neutrophil response to serum-opsonized zymosan. It is assumed that the opsonic role of fixed C3 on zymosan comes from the enhanced binding mediated through iC3b binding to CR, (see Section IV,B,2). The most important neutrophil CR, ligands may be fixed C4b and iC3b present on antibody-coated particles and immune complexes that activate the classical pathway. Recent studies have also suggested that clusters of fixed C4b and iC3b molecules may provide enhanced binding to CR, as compared to fixed iC3b alone (Medof and Nussenzweig, 1984b). Although the CR, of unstimulated neutrophils does not promote true phagocytosis (as defined by formation of a membrane vesicle around the ingested particle), small particles and soluble complexes may be endocytosed via neutrophil CR, into clathrin-coated pits (Fearon et al., 1981; Schreiber et al., 1982). This was demonstrated with rabbit (polyclonal) antiCR, and CSb-membrane protein complexes prepared from detergent-solubilized EC3b (C3b-OR). Monovalent Fab anti-CR, and C3b-OR each bound avidly to neutrophil CR,, but remained on the membrane surface without being ingested. Cross-linkage of these surface-bound CR, ligands with either goat F(ab’)2-anti-rabbit IgG or F(ab’)2-anti-C3, respectively, caused the resulting CR,-bound complexes to be endocytosed. By contrast, bivalent F(ab’),-anti-CR, was endocytosed without need of further crosslinkage. Thus, endocytosis by CR, required a small soluble complex that bound to several CR, simultaneously (Fearon et al., 1981). C3b-coated E . coli were also apparently endocytosed by way of neutrophil CR, (Schreiber et al., 1982). However, bivalent monoclonal anti-CR, was not endocytosed by neutrophils until cross-linked by F(ab’),-anti-mouse Ig (Hogg et al., 1984). Thus, ligand binding to more than two CR, molecules may be required to trigger endocytosis. The rabbit The rabbit F(ab’),-anti-CR, that was endocytosed had been shown to recognize 8-9 different CR, epitopes (Wilson et al., 1982), and thus could potentially cross-link many cell surface CR,. Even though most immune complexes in the blood are probably cleared by way of erythrocyte CR, transport (Medof and Oger, 1982) to liver and spleen (Cornacoff et al., 1983), neutrophil CR,-mediated endocytosis may be important for elimination of soluble immune complexes at tissue sites of inflammation.

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

253

CR, has also been reported to stimulate neutrophil release of histaminase (Melamed et al., 1982a). Fluid-phase C3b dimers and rabbit ECSb, both of which bind to CR,, triggered histaminase release inhibitable by anti-CR,. However, sheep ECSb, that also binds avidly to CR,, did not trigger histaminase release (Melamed et al., 1982b). This latter finding suggests that histaminase release may be triggered by another receptor that works synergistically with CR,. For example, both rabbit EC3b and sheep EC3b bind to neutrophils by way of CR,, but only rabbit EC3b and not sheep EC3b is ingested. In this case, CR, serves only to promote strong binding of the rabbit EC3b to the neutrophil surface, and ingestion is triggered by CR, binding to some component of the rabbit E membrane (Ross et al., 1984a). Because Fab anti-Mol (anti-CR,) inhibits neutrophil release of histaminase triggered by opsonized zymosan (Arnaout et ul., 1983b), CR, is probably the receptor that triggers histaminase release in response to rabbit EC3b. 2 . CR,

As with CR,, the amount of CR, expressed on resting neutrophils is low (4000 CR, per cell), but can be rapidly increased by more than 20-fold by exposure to the same chemotactic factors and isolation conditions that stimulate increased expression of CR,. In addition, PMA and calcium ionophore have been shown to induce the expression of as many as 600,000 CR, per neutrophil without stimulating an increase in the expression of LFA-1 per neutrophil (Arnaout et al., 1984). However, the physiologic significance of such massive receptor expression stimulated with nonphysiologic agents such as PMA or calciuin ionophore is unknown. Parallel radioimmune assay of CR, and CR, on isolated neutrophils stimulated by incubation in Hanks’ buffer at 37°C for 30 minutes revealed only slightly more CR, (65,000) than CR, (55,000) per neutrophil (Ross et al., 1984b). One puzzling aspect about CR, function was that avid CR,-dependent binding of EC3bi did not trigger significant ingestion or a respiratory burst, whereas weak CR,-dependent binding of zymosan and rabbit E triggered both ingestion and a respiratory burst. Mapping of functional epitopes of CR, with monoclonal antibodies indicated that one possible explanation was that the a-chain of CR, contained two distinct binding sites: (1) an iC3bbinding site that provided firm binding of iC3b-coated particles but did not trigger functions, and (2) a function-triggering site that bound to zymosan or rabbit E but did not bind to fixed iC3b. As mentioned previously, antiLeu-15 inhibited EC3bi rosettes but had no effect on zymosan binding or ingestion, whereas OKM 1 inhibited zymosan binding and ingestion but had no effect on EC3bi rosettes (Ross et ul., 1984a). Rabbit E are bound and ingested by way of CR, (Ross et al., 1984a) but sheep E are not ingested unless sialic acid is first removed by treatment with

2.54

G O R D O N D. ROSS A N D M . E D W A R D M E D O F

neuraminidase (G. D. Ross, unpublished observation). Even though normal sheep E are not ingested, CR, appears to be able to synergistically enhance the Fc receptor-mediated ingestion of sheep E sensitized with IgG (EAI,c), perhaps by way of CR, binding weakly to some component of the sheep E membrane (Ross et al., 1984a). First, treatment of neutrophils with anti-CR, (either IgG antibody or Fab fragments) cuased partial (20-40%) inhibition of EAI,, ingestion (Arnaout et al., 1983; Ross et al., 1984a). Second, EAI,, ingestion by neutrophils from CR,-deficient patients was reduced to 20-40% of that observed in neutrophils from normal controls (Arnaout et al., 1982). In either case, reduced EAIgGingestion was only observed under conditions that were suboptimal for Fc receptor-mediated ingestion (either suboptimal IgG sensitization or reduced incubation time). The binding site of CR, was required because anti-CR, only inhibited EA,,, ingestion in buffers containing divalent cations that allowed activity of the CR,-binding site, and not in buffers containing either EDTA or NADG (Ross et al., 1984a). Firm binding of sheep E to neutrophils was insufficient to stimulate CR,-mediated ingestion, as sheep EC3b were not ingested despite firm binding to neutrophils by way of CR,. Thus CR,-mediated enhancement of EAfgF ingestion required that ingestion be first induced by attachment of sensitizing IgG to Fcreceptors. The p-chain of CR, probably does not have a primary role in the binding of iC3b or zymosan. First, monoclonal anti-p-chain did not inhibit neutrophil binding of either EC3bi (Sanchez-Madrid et al., 1983a; Wright et al., 1983a) or zymosan (Ross et al., 1984a). Second, T cells that express p-chains in surface LFA-1 do not bind EC3bi or yeast. On the other hand, the pchains of both CR, and LFA-1 may have a similar role in triggering cell functions following ligand binding to each respective a-chain type. The monoclonal anti-p-chain known as TS 1/18 (Sanchez-Madrid et al., 198315) inhibited both CR,-dependent neutrophil responses to zymosan (G. D. Ross, J. A. Cain, and T. A. Springer, unpublished observation) and LFA-1mediated T cell cytotoxicity (T. A. Springer, personal communication). Because CR,-mediated ingestion of zyinosan was also blocked by neutrophil treatment with cytochalasin B (Ross et al., 1984a), microfilaments are probably involved in CR,-mediated ingestion, and these may be linked to the pchain of CR,. Accordingly, it might be hypothesized that linkage of antibody to the p-chain of CR, might prevent ingestion by inhibiting the normal pchain interaction with microfilaments. Alternatively, particle binding to the function-triggering site in the a-chain might induce an aggregation of pchains that opened a calcium ion channel required to trigger cell function. If p-chains functioned by this latter mechanism, then attachment of anti-pchain antibody might inhibit cell functions by interfering with p-chain aggregation.

M EM B R ANE COMPLEMENT RECEPTORS FOR

c3

255

3. C R , Little is known of the function of neutrophil CR,. Because EC3dg that bind to CR, are not ingested, CR, is unable to induce ingestion of red blood cells. However, it is unknown if the CR, of PMA-activated neutrophils might acquire the ability to induce particle ingestion. Even though relatively high densities of fixed iC3b or C3dg are required for sheep E binding to CR,, much higher densities of fixed C3 have been detected on bacteria opsonized with pooled normal serum (Newman and Mikus, 1984). On the other hand, some types of soluble immune complexes apparently do not bind to CR,. Radiolabeled BSA/anti-BSA immune complexes in whole blood that have nndergone binding and I-dependent release from red cells contain primarily fixed C3dg and bind to tissue spleen cells and CR, bearing cells more than to blood neutrophils (Lam and Medof, 1982). The situation may be different with other types of immune complexes such as DNA/anti-DNA that do not get reduced in size by complement activation and which acquire a dense coating of fixed C3dg. Immune complexes generated in blood with PM-2 DNA (approximately 1 X lo6 M,) and IgG anti-DNA (isolated from a patient with SLE) have been shown to contain 100-200 molecules of fixed C3 and an equivalent number of fixed C4 molecules (G. D. Ross and R. P. Taylor, unpublished observation). The mechanism by which such complexes are cleared after release from E CR, is unknown but may involve CR,. OF MONOCYTE/MACROPHAGE c3 RECEPTORS c. FUNCTIONS

Monocytes and macrophages express the same CR,, CR,, and CR, expressed by neutrophils, and the functions of these receptors are for the most part similar. A major difference is that macrophages are long-lived cells that have a greater range of differentiation possibilites than do neutrophils. Various stimuli cause changes in either C3 receptor density or function, and some of these changes appear to be reversible. 1. CR,

As with neutrophils, the CR, of monocytes is unable to induce phagocytosis of sheep EC3b (Newman et al., 1980, 1981, 1984; Wright and Silverstein, 1982; Newman et al., 1984), but is capable of promoting endocytosis of small C3b-OR complexes or polyclonal F(ab’),-anti-CR, (Fearon and Wong, 1983). When monocytes are cultured under several conditions, both CR, and CR, acquire the ability to induce EC3 ingestion. Monocytes grown adherently to plastic develop phagocytically active CR, and CR, after 7 days culture with media supplemented only with autologous serum (Newman et

256

GORDON D . ROSS AND M. EDWARD MEDOF

al., 1984). On the other hand, monocytes grown in Teflon-coated chambers in suspension do not acquire phagocytically active CR, and CR, unless the media is supplemented with PMA, fibronectin, or serum amyloid P component (Wright and Silverstein, 1982; Wright et al., 198313). Because adherent monocytes lay down a coating of intrinsic fibronectin, it has been proposed that this intrinsic fibronectin stimulates the differentiation of adherently grown monocytes. Unlike other stimuli that promote the differentiation of both CR, and CR,, a T cell derived lymphokine has been described that causes the selective differentiation of CR, (Gresham and Griffin, 1984). After treatment with this lymphokine, monocytes ingested EC3b and little EC3bi. The small amount of EC3bi ingested was shown to be mediated by CR,, because it was inhibited by anti-CR,. It is presumed that the differentiated monocytes produced in vitro have their counterparts in uiuo that are differentiated in a similar manner. In the mouse for example, resting peritoneal macrophages have phagocytically inactive C3 receptors, whereas thioglycollate elicited (inflammatory) peritoneal macrophages have phagocytically active C3 receptors (Bianco et al., 1975). Recent studies of human peritoneal macrophages have demonstrated that a high proportion of these cells have the ability to ingest both EC3b and EC3bi, and thus resemble the monocyte-derived macrophages produced by 7 day adherent cultures (Newman and Becker, 1983). 2. CR, Monocyte CR, resembles neutrophil CR, in that avidly bound EC3bi are not ingested (Newman et al., 1981, 1984; Wright and Silverstein, 1982), whereas weakly bound yeast, zymosan, and rabbit E are ingested (Czop et al., 1978; Ezekowitz et al., 1984; Ross et al., 1984a). As with neutrophils, the binding of unopsonized zymosan to monocytes triggered a superoxide burst (S. L. Newman, personal communication). With either culture-derived adherent macrophages or fibronectin-activated macrophages, 2-3 times more EC3bi than EC3b are ingested (Newman et al., 1981, 1984; Wright and Silverstein, 1982). The reason for this is unknown. Both CR, and CR, are active and function independently, because blockage of either CR, or CR, with receptor-specific antibodies inhibit only EC3b or only EC3bi ingestion, respectively. In this type of assay, little or no binding of EC3bi to CR, occurs because F(ab'),-anti-CR, produced
M E M B R A N E C O M P L E M E N T RECEPTORS FOR C 3

257

3. CR, Little fiinctional data are available on monocyte/macrophage CR,. Unlike CR, and CR,, the CR, of activated marcophages does not have the ability to trigger the- ingestion of EC3. Either adherent macrophages cultured in bovine serum or PMA-activated macrophages bind EC3d to CR,, but neither ingests the bound E (Inada et al., 1983; Wright et al., 1984).

D. FUNCTIONS OF LYMPHOCYTE C3 RECEPTORS Lymphocytes are heterogeneous in their expression of C3 receptors. The majority of peripheral blood B cells express CR,, and 7 5 4 5 % of peripheral blood B cells also express CR, (Ross et al., 197%). CR, has never been detected on T cells, but a minor proportion of T cells does express small amounts of CR, (Wilson et al., 1983). CR, is expressed on the non-T non-B lymphocyte subset that functions in natural killing (NK cells) and antibodydependent cellular cytotoxicity (ADCC, K cells), and is not expressed on the majority of B and T lymphocytes (Ault and Springer, 1981). CR, has been shown to be expressed at a very early stage in B cell differentiation (Tedder et al., 1983), but then is lost during blastogenesis prior to differentiation into plasma cells (G. D. Ross, unpublished observation). CR, is acquired on developing B cells after the acquisition of CR,, so that mature B cells express both CR, and CR, (Tedder et al., 1984). Because the surface phenotype of lymphocytes from patients with chronic lymphatic leukemia (CLL) is thought to mirror some discrete stage of normal B cell development, it may be significant that a high proportion of CLL lymphocytes characteristically express CR, and lack CR, (Ross et d . ,1973; Ross and Polley, 1975; Hogg et al., 1984). There may be a late stage of normal B cell development when cells express CR, and have lost CR,.

1. CR, Affinity-purified rabbit F(ab’),-anti-CR, was not mitogenic for lymphocytes by itself, but did enhance synergistically the blastogenesis induced by pokeweed mitogen (Daha et al., 1984). Furthermore, with submitogenic doses of pokeweed mitogen, blastogenesis was triggered by the secondary addition of F(ab’),-anti-CR,. Because pokeweed mitogen stimulation requires B cells, T cells, and monocytes, all of which express CR,, it is unknown which cell types are affected by the anti-CR, treatment. Diverse results have been reported in attempts to characterize the effects of C3b on lymphocyte responses. Initially, human C3b was reported to stimulate mouse splenocyte blastogenesis (Hartmann and Bokisch, 1975).

258

GORDON D . ROSS AND M . EDWARD MEDOF

However, the mechanism of this stimulation was unclear, because human C3b was known to be unreactive with murine CR, (Dierich et al., 1974; Rabellino et al., 1978). Subsequently, it was reported that contaminating H in human C3b preparations triggered murine lymphocyte responses, and that human C3b preparations depleted of H were inactive (Hammann et al., 1981). Human lymphocyte responses have been reported to be inhibited by fluid-phase C3b or denatured native C3 (C3i) (Berger and Fleisher, 1983), but to be enhanced by fixed C3b on sheep E (Lobo and Burge, 1982). The stimulatory effects of fixed C3b were thus similar to the effects of anti-CR,. On the other hand, the inhibitory effects of fluid-phase C3 or C3b may be due to either the binding of C3b/C3i to CR,, or the binding of contaminating C3a to C3a receptors. CR, ligands (see below) and C3a (Morgan et al., 1982) are known to suppress lymphocyte responses. Because both lymphocytes and monocytes secrete factor I that can convert C3b into iC3b (Whaley, 1980; Lambris et aE., 1980), “pure” C3b added to cultures may be converted to iC3b and bind to CR, and CR, in addition to CR,. Thus, anti-CR, is probably the only CR,-specific ligand useful in these types of in vitro experiments. Another function of lymphocyte CR, may be to promote antigen trapping or presentation by macrophages or dendritic cells. One such mechanism that has been previously proposed was termed “complement-dependent bridge formation between cells” (Dierich and Landen, 1977). In this scheme, the antigen presenting cell might express a C3-convertase that would allow it to fix C3b onto its surface that would bind (“form bridges with”) CR1-bearing cells. Alternatively, IgM/antigen/C3b complexes might be bound to macrophage or dendritic cell CR,, and then concentrated on the surface of these cells without being ingested. The surface-bound and concentrated C3b/iC3bcoated antigen molecules might then link CR,-bearing B lymphocytes (or T lymphocytes) to the antigen presenting cell, thus promoting antigen recognition. When lymphocytes or any other CR,-bearing cells bind C3b-bearing complexes, the complexes may trigger complement activation at the cell surface that could make the lymphocytes targets of bystander lysis. Thus, it is probably significant that the CR, of lymphocytes has been shown to inhibit both the C3 and C5-convertases of the classical pathway (Iida and Nussenzweig, 1983). Nascent C3b that is generated and binds to the lymphocyte membrane despite the CR, inhibition, could be controlled by lymphocyte associated DAF. Lymphocytes active in natural and antibody-dependent cytotoxicity (NK/K cells) are not believed to express CR, (Tedder et al., 1983). Yet in a K cell assay for lysis of IgG-coated chicken E, fixed C3b on the chicken EA,,, caused a significant enhancement of hemolysis (Perlmann et al., 1975, 1981;

MEMBRANE COMPLEMENT KECEPTORS FOR C 3

259

Ghebrehiwet et al., 1979; Wahlin et al., 1983). Either some K cells do express CR, or, alternatively, the lymphocytes in the assay may be able to secrete an intrinsic I (Lambris et al., 1980) that converts target cell C3b into iC3b that binds to NK/K cell CR,. These questions can only be answered with antibodies to CR,, by determining if anti-CR, reverses the enhancement provided by fixed C3b on target cells. 2. C R , Because CR, is expressed exclusively on B lymphocytes, it might be predicted that CR, would be shown to have some role in regulation of B cell function. Rabbit F(ab’),-anti-CR, (anti-gp72) was not mitogenic, but inhibited blastogenesis triggered by pokeweed mitogen, a mixed lymphocyte reaction, or tetanus toxoid (Lambris et al., 1982). F(ab’),-anti-CR, also inhibited Ig synthesis assayed by reverse hemolytic plaque formation. By contrast, anti-CR, did not inhibit mitogenesis stimulated by either PHA or Con A, and on most occasions, enhanced blastogenesis stimulated by these T cell-specific mitogens by 20-40%. As this anti-CR, was not affinity-purified, there was some concern that it might contain natural antibodies to human lymphocytes that might be responsible for the responses observed. It was thus important that either C3d (Schenkein and Genco, 1979) or C3d complexed with affinity-purified F(ab’),-anti-C3d (Lambris et al., 1982) produced inhibition resembling the anti-CR,. Furthermore, a similar inhibition profile of lymphocyte responses was reported for a C3dg-like fragment generated with kallikrein termed C3d-K (Meuth et al., 1983). Other fluid-phase fragments that bind to CR,, such as iC3b, C3b, and “C3b-like” C3i, may inhibit lymphocyte responses in a similar manner (Berger and Fleisher, 1983). The mechanism by which ligand binding to B cell CR, might inhibit T cell responses to tetanus toxoid or a mixed lymphocyte reaction is unknown. Although it might be hypothesized that CR, was expressed on activated T cells, CR, was undetectable on pokeweed mitogen-stimulated T cell blasts (Tedder et al., 1984). Further characterization of the function of CR, may require the use of monoclonal antibodies to CR, that completely block the C3d-binding site. Unfortunately, the two monoclonal anti-CR, antibodies that are commercially available (anti-B2 and HB-5) do not inhibit EC3d rosettes. 3. CR, Although CR, is believed to be expressed primarily on K/NK cells and not on B or T cells (Ault and Springer, 1981), a small subset of suppressor T cells that express CR, has been identified by double staining with anti-Leu-2 and anti-Leu-15 (same as D12, Landay et al., 1983).

260

GORDON D . ROSS AND M. EDWARD MEDOF

Several studies have suggested that CR, may have some role in both K cell lysis of bovine EAI,, and NK cell killing of K562 cells. Bovine red cells coated with human C3b, iC3b, or C3d were not lysed by human lymphocytes, but fixed C3 fragments did enhance lysis of bovine red cells coated with IgG (Perlmann et al., 1981; Wahlin et al., 1983). Target cell-bound C3 fragments did produce lysis when present on bovine red cells bearing small amounts of sensitizing IgG that was insufficient in itself to trigger K cell lysis. Although fixed C3b, iC3b, and C3d were all active, fixed iC3b was much more active in the enhancement of K cell hemolysis than was fixed C3d or C3b (iC3b>C3d>C3b). The larger enhancement by fixed iC3b as compared to fixed C3b supports the findings that CR,, but not CR,, are present on K/NK cells. The enhancement by fixed C3b may result from conversion of fixed C3b to iC3b by lymphocyte-secreted I. The mechanism of the enhancement by fixed C3d is unclear, as CR, is not believed to be expressed by K/NK cells. K/NK cells may express p150,95, and this putative CR, may be the source of K cell C3d receptor activity. A role for CR, in NK cell activity was first suggested by the absence of N K activity in lymphocytes from patients with a genetic deficiency of CR, (Ross et al., 1983b). Later, however, it appeared that there might be no causal relationship between the deficiencies of CR, and NK activity when these patients were shown to be deficient in LFA-1 and p150,95 as well as CR, (Ross et al., 198413). Furthermore, anti-LFA-1, but not anti-CR,, inhibited normal lymphocyte NK activity (Ault and Springer, 1981; Krensky et al., 1983). Subsequently, however, treatment of normal lymphocytes with a mixture of OKMl (anti-CR,) and anti-LFA-1 was shown to produce more inhibition of N K activity than did treatment with anti-LFA-1 alone (Kohl et al., 1984). In addition, normal lymphocyte NK activity against K562 target cells was shown to be enhanced by prior fixation of iC3b to the K562 cells. By contrast, there was no enhanced killing of K562 cells bearing an equivalent amount of fixed C3b (G. D. Ross, R. H. R. Ward, and P . J. Lachmann, unpublished observation). Because K562 cells and other NK-sensitive target cells are frequently activators of the alternative pathway of complement, NK target cells may be coated with fixed iC3b in uivo. Finally, because CR, has been shown to bind to rabbit E, yeast, and bacteria that lack fixed iC3b, it appears possible that CR, may also be able to bind directly to certain components of NK target cells without need of fixed iC3b. Clarification of the possible role of CR, in NK cell function will probably require a better understanding of the lectin-like binding specificity of CR,. OF KIDNEY PODOCYTEC3 RECEPTORS E. FUNCTIONS

Although kidney podocytes bind both EC3b and EC3bi (Carlo et at., 1979), they apparently express only CR, (that binds both EC3b and EC3bi) and lack detectable CR, entirely (Kazatchkine et al., 1982; Kazatchkine and

M E M B R A N E C O M P L E M E N T RECEPTORS FOR

C3

261

Fisher, 1984). Studies of the tissue distribution of CR, have suggested that kidney podocytes may have a greater surface density of CR, than does any other type of cell (Hogg et nl., 1984). Characterization of the function of renal CR, has proven difficult because isolated kidney podocytes have quickly lost cell surface CR, during attempts to maintain these cells in culture. Recent studies, however, have indicated that podocyte CR, does have I-cofactor activity for cleavage of C3b, and can mediate absorptive endocytosis of CR, ligands (M. D. Kazatchkine, personal communication). In diseases with immune complex-mediated damage of glomeruli, podocyte CR, may have some role in inactivation or clearance of C3b/iC3b-bearing complexes that traverse the basement membrane. In surveying kidney sections from 75 patients with nephritis for CR, by immunofluorescence, renal CR, was found to be present in all forms of kidney disease except proliferative nephritis of SLE (Kazatchkine et al., 1982; Emancipator et al., 1983). It is unknown whether renal CR, is either lost as a result of nephritis or low prior to onset of nephritis because of genetic inheritance. Other forms of nephritis associated with deposition of C3 and IgG did not show a reduction in CR, detectable by immunofluorescence.

F. FUNCTIONS OF MAST CELLC3 RECEPTORS Only peritoneal mast cells from the rat have been analyzed for complement receptors. Rat mast cells bound and ingested both EC3b and EC3bi, suggesting the expression of phagocytically active CR, and CR, (Vranian et al., 1981). Because EC3bi were ingested more avidly than were ECSb, it is likely that EC3bi were not ingested by way of CR,, and thus that mast cells express CR, in addition to CR,. V. Conclusions

Characterization of the function of complement receptors represents one of the final frontiers of complement research that has previously detailed the functions of the plasma complement components. Much remains unknown about the reactions of complement at the cellular level, particularly with regard to the role of complement and complement receptors in the immune response. Finally, even though it has been long recognized that the major function of complement was clearance of bacteria by way of opsonization, there are still many unknown features about the functions of phagocyte complement receptors.

REFERENCES Abrahamson, D. R . , and Fearon, D. T. (1983). I d . Invest. 48, 162. Aikawa, T., Mitaniura, T., Tammonio, K . , and Horiuchi, Y. (1979)./. Lab. Clin. Med. 94, 902. Arnaout, M . A , , Melamed, J., Tack, B. F . , and Colten, H. R . (1981).J . Imrnunol. 127, 1348.

262

GORDON D. ROSS AND M. EDWARD MEDOF

Arnaout, M. A , , Pitt, J., Cohen, H. J . , Melamed, J., Rosen, F. S., and Colten, H. R. (1982). New Engl. J . Med. 306, 693. Arnaout, M. A,, Dana, N., Melamed, J . , Medicus, R., and Colten, H. R. (1983a).Zmniunology 48, 229. Arnaout, M. A,, Todd, R. F., 111, Dana, N., Melamed, J . , Schlossman, S. F., and Colten, H. R. (1983b).J. Clin. Invest. 72, 171. Arnaout, M . A., Todd, R. F., 111, and Dana, N. (1984).Fed. Proc., Fed. Am. Soc. E x p . B i d . 43, 1655. Atkinson, J. P., and Jones, E. A. (1984).J. Clin. Znoest. 74, 1649. Ault, K. A., and Springer, T. A. (1981).J. Zmmunol. 126, 359. Barel, M., Charriaut, C., and Frade, R. (1981). Fed. Eur. Biochem. Soc. 136, 111. Beller, D. I . , Springer, T. A,, and Schreiber, R. D. (1982).J. E x p . Med. 156, 1000. Berger, M., and Fleisher, T. A. (1983).I . Zmmunol. 130, 1021. Bianco, C. (1977). In “Comprehensive Immunology” (R. A. Good and S. B. Day, eds.), Vol. 2, p. 60. Plenum, New York. Bianco, C., and Nussenzweig, V. (1971). Science 173, 154. Bianco, C., and Nussennveig, V. (1977). Contemp. Top. Mol. Zmmunol. 6, 145. Bianco, C., Patrick, R., and Nussenzweig, V. (1970).J. Exp. Med. 132, 702. Bianco, C., Griffin, F. M., and Silverstein, S. C. (1975).J . E x p . Med. 141, 1278. Bokisch, V. A., and Sobel, A. T. (1974).J . E x p . Med. 140, 1336. Bokisch, V. A., Dierich, M. P., and Miiller-Eberhard, H. J. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 1989. Brown, E. J . , Joiner, K. A., Gaither, T. A., Hammer, C. H., and Frank, M. M. (1983). J. Zmmunol. 131, 409. Brown, H. C., and Boom, J. C. (1983). Trans. A. Soc. Trop. Med. Hyg. 32, 209. Carlo, J. R., Ruddy, S., Studer, E. J., and Conrad, D. H. (1979).J . Zmmunol. 123, 523. Carlo, J. R., Spitznagel, J . K., Struder, E., and Ruddy, S. (1980). J. Zmmunol. 124, 1516. Carlo, J. R . , Spitznagel, J. K., Struder, E. J . , Conrad, D. H . , and Ruddy, S. (1981). Zmmunology 44, 381. Cooper, N. R. (1969). Science 165, 396. Cooper, N. R. (1975).1. Exp. Med. 141, 890. Cornacoff, J. B., Hebert, L. A., Smead, W. L., Vanaman, M. E., Birmingham, D. J., and Waxman, F. J. (1983).J. Clin. Invest. 71, 236. Curry, R. A,, Dierich, M. A., Pellegrino, M. A,, and Hoch, J. A. (1976). Zinmunogenetics 3, 465. Czop, J. K., Fearon, D. T., and Austen, K. F. (1978).J. Zmmunol. 120, 1132. Daha, M. R., Bloem, A. C., and Ballieux, R. E. (1984).J. Zmmunol. 132, 1197. Dahlback, B., Smith, C. A,, and Miiller-Eberhard, H. J. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 3461. Davis, A. E., and Lachmann, P. J. (1983). Zrninunobiology 164, 231. Davis, A. E., Harrison, R. A., and Lachmann, P. J. (1984).J. Zmmunol. 132, 1960. Dierich, M. P., and Landen, B. (1977).J . E x p . Med. 146, 1484. Dierich, M. P., and Reisfeld, R. A. (1975).J . E x p . Med. 142, 242. Dierich, M. P., Pellegrino, M. A., Ferrone, S., and Reisfeld, R. A. (1974).J. Znimunol. 112, 1766. Dohson, N. J., Lambris, J. D., and Ross, G. D. (1981).J . Zmmunol. 126, 693. Dykman, T. R . , Cole, J. L., Iida, K., and Atkinson, J. P. (1983a).Proc. Natl. Acad. Sci. U.S.A. 80, 1698. Dykman, T. R., Cole, J. L., Iida, K., and Atkinson, J. P. (1983b).J. E x p . Med. 157, 2160. Dykman, T. R., Hatch, J. A., and Atkinson, J . P. (1984a).J. E x p . Med. 159, 691.

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

263

Dykman, T. R., Hatch, J. A., Aqua, M. S., and Atkinson, J. P. (1984b). Submitted for publication. Eddy, A., Newman, S. L., Cosio, F . , LeBien, T., and Michael, A. (1984). Clin. Itnnzut~ol. Zminunopatho~.31, 371. Eden, A., Miller, G. W., and Nussenzweig, V. (1973).J. Clin.Itwest. 52, 3239. Ehlenberger, A. G., and Nussenzweig, V. (1977).J. E x p . Med. 145, 357. Emancipator, S. N . , Iida, K., Nussenzweig, V., and Gallo, G. R. (1983). Clin. Immunol. Imntunopathol. 27, 170. Ezekowitz, R. A. B., Sim, R. B., Hill, M., and Gordon, S. (1984). J. Exp. Med. 159, 244. Fearon, D. T. (1979). Proc. Natl. Acad. Sci. U . S . A . 76, 5867. Fearon, D. T. (1980). J. Exp. Metl. 152, 20. Fearon, D. T. (1983). Springer Semiti. Inmzunopathol. 6, 159. Fearon, D. T. (1984). Imniunol. Today 167, 4919. Fearon, D. T., and Austen, K. (1977).J. Exp. Med. 146, 177. Fearon, D. T., and Collins, L. A. (1983). J. Imnmnol. 130, 370. Fearon, D. T., and Wong, W. W. (1983). Annu. Reu. Imnwnol. 1, 243. Fearon, D. T., Kaneko, I., and Thomson, G . G. (1981).J. Exp. Med. 153, 1615. Ferreira, A , , and Nussenzweig, V. (1976). J. Z m n i u t d 117, 771. Ferreira, A,, and Nussenzweig, V. (1977).J. Exp. Med. 146, 1001. Fingeroth, J. D., Weis, 1. J., Tedder, T. F . , Strominger, J. L., Biro, P. A., and Fearon, D. T. (1984). Proc. Natl. Acud. Sci. U.S.A. 81, 4510. Frade, R., and Stominger, J. (1980).J. Zmmunol. 125, 1332. Frade, R., Myones, B. L., Barel, M., Krikorian, L., Charriaut, C., and Ross, 6. D. (1984). Submitted for publication. Fujita, T., and Nussenzweig, V. (1979).J. E x p . Med. 150, 2670. Fujita, T., and Tamura, N. (1983).J . E x p . Med. 157, 1239. Fujita, T., Takata, Y . , and Tdlnllrd, N . (1981). J. E x p . Med. 154, 1743. Gadd, K. J., and Reid, K. B. M. (1981). Riochern. J. 195, 471. Gaither, T. A,, Hammer, C. H . , and Frank, M. M. (1979).J . Ztnmunol. 123, 1195. Gelfand, M. C., Sachs, D. H., Lieherman, R., and Paul, W. E. (1974).J. Exp. Med. 139, 1142. Gelfand, M. C., Frank, M. M., and Green, I. (1975).J . E x p . Med. 142, 1025. Gerdes, J., Naiem, M., Mason, D. Y., and Stein, H. (1982). Immunology 45, 645. Ghebrehiwet, B., Medicus, R. G., and Miiller-Eberhard, H. J. (1979).J . Immunol. 123, 1285. Gigli, I., and Fearon, D. T. (1981). Fed. Proc., Fed. Am. Soc. E x p . Bid. 40, 1172 (Abstr.) Gigli, I., Fujita. T., and Nussenzweig, V. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 5696. Gitlin, J. D., Rosen, F. S., and Lachmanii. P. J. (1975). J . Exp. Med. 141, 1221. Goers, J. W. F., and Porter, R. R. (1978). Biochem. J. 175, 675. Govaerts, 1’. (1920a). C . R . SOC. Riol. 83, 196. Govaerts, P. (l920b). C.R. Soc. Biol. 83, 197. Gresham, H. D., and G r i f h , F. M., Jr. (1984). Clin. Res. 32, 504A. Hammann, K. P., Raile, A., Schmitt, M., Mussel, H. H., Peters, H., Scheiner, O . , and Dierich, M. P. (1981). Ininiunobiobgy 160, 289. Harrison, H. A . , and Lachmann, P. J. (1980). Mol. Ininiunol. 17, 9. Hartmann, K. U., and Bokisch, V. A. (1975).J . Eq,. Med. 142, 600. Hatch, J. A., Atkinson, J. P., Suarez, B. K., and Dyknian, T. R. (198.1).J. Immunol. 132, 2168. Hekmatpanah, D. S., Nies, K. M., Louis, J . M., and Medof. M. E. (1982). Clin.Res. 30, 53A. Ho, M.-K., and Springer, T. A. (1983).J. B i d . Cheni. 258, 2766. Hogg, N., Ross, 6 . D., Jones, D. B . . Slusarenko, M., Walport, M . J., and Lachmann, P. J. (1984). Eur. J . In~rnunol.14, 236. Horgan, C., and Taylor, R. P. (1984). Arthritis Rheum. 27, 320.

264

GORDON D . ROSS A N D M . EDWARD MEDOF

Horstmann, R. D., Pangburn, M. K., and Miiller-Eberhard, H. J. (1984).Fed. Proc., Fed. Am. Soc. Exp. Biol. 43, 1765. Hutt-Fletcher, L. M., Fowler, E., Lambris, J. I)., Simmons, J. G . , and Ross, G . I). (1983).J. Zmmunol. 130, 1309. Iida, K., and Nussenzweig, V. (1981).J. E x p . Med. 153, 1138. Iida, K., and Nussenzweig, V. (1983).J. Zmtnunol. 130, 1876. Iida, K., Mornaghi, R . , and Nussenzweig, V. (1982).J. E x p . Med. 155, 1427. Iida, K., Nadler, L., and Nussenzweig, V. (1983).J. E x p . Med. 158, 1021. Inada, S., Brown, E. J . , Gaither, T. A , , Hammer, C. H., Takahashi, T., and Frank, M. M. (1983). Proc. Natl. Acad. Sci. U.S.A. 80, 2351. Inada, Y., Kamiyama, M., Kanernitsu, T., Hyman, C. L., and Clark, W. S. (1982). Clin. Exp. Zmmunol. 50, 187. Isenman, D. E., and Young, J. R. (1984).J. Immunol. 132, 3019. Johnson, E., Eskeland, T., and Bertheussen, K. (1984). Scand. J . Zmmunol. 19, 31. Johnson, V., Ohlsson, K . , and Olsson, I. (1976). Scand. J. Immunol. 5, 421. Jondal, M., Klein, G . , Oldstone, M. B. A,, Bokish, V., and Yefenof, E. (1976). Scand. J. Zmmunol. 5, 401. Karp, D. R . , Atkinson, J. P., and Schreffler, D. C. (1982).J. B i d . Chem. 257, 7330. Kazatchkine, M. D., and Fisher, E. (1984). Fed. Proc.. Fed. Am. Soc. Exp. Biol. 43, 1766. Kazatchkine, M . D . , Fearon, D. T., and Austen, K. F. (1979).J . Ztnmunol. 122, 75. Kazatchkine, M. D., Fearon, D. T., Appay, M. D., Mandet, C., and Bariety, J. (1982).J. Clin. Inuest. 69, 900. Kerr, M. A. (1980). Biochem. 1. 189, 173. Kinoshita, T., and Nussenzweig, V. (1984). Fed. Proc., Fed. Am. SOC. Exp. B i d . 43, 1491. Klebanoff, S. J., Beatty, P. G . , Schreiber, R. D., Ochs, H. D., and Waltersdorph, H. M. (1984).J. Immunol., in press. Kohl, S., Springer, T. A., Schmalsteig, F. S., Loo, L. S., and Anderson, D. C. (1984). J. Zmmunol. 133, 2972. Krensky, A. M., Sanchez-Madrid, F., Robbins, E., Nagy, J. A,, Springer, T. A,, and Burakoff, S. J. (1983).J. Zmmunol. 131, 611. Kurzinger, K., and Springer, T. A. (1982).J. B i d . Chem. 257, 12412. Lachmann, P. J. (1981). Immunol. Today 2, 144. Lachmann, P. J., and Miiller-Eberhard, H. J. (1968).J. Immunol. 100, 691. Lachmann, P. J., Pangburn, M. K., and Oldroyd, R. G . (1982).J. E x p . Med. 156, 205. Lam,T., and Medof, M. E. (1982). Fed. Proc., Fed. Am. Soc. E r p . B i d . 41, 850. Lambris, J. D., Dobson, N. J., and Ross, G . D. (1980).J. Exp. Med. 152, 1625. Lambris, J. D . , Dobson, N . J., and Ross, G. D. (1981).Proc. Nut/. Acad. Sci. U.S.A.78, 1828. Lambris, J. D., Cohen, P. L., Dohson, N. J., Wheeler, P. W., Papamichail, M., and Ross, G . D. (1982). Clin. Res. 30, 514a. Landay, A,, Gartland, G . L., and Clement, L. T. (1983).J . Immunol. 131, 2757. Laveran, A,, and Mesnil, F. (1901). Ann. Inst. Pusteur 15, 673. Law, S.-K. A. (1983). Ann. N.Y. Acad. Sci. 421, 246. Law, S.-K. A. (1984). E M B O J . , in press. Law, S.K., and Levine, R. P. (1977). Proc. Natl. Acud. Sci. U.S.A. 74, 2701. Law, S. K., Fearon, D. T., and Levine, R. P. (1979a).J . Immunol. 122, 7.59. Law, S. K . , Lichtenberg, N. A,, and Levine, R. P. (1979b).J . Zmmunol. 123, 1388. Law, S.-K. A., Minich, J. M., and Levine, R. P. (1984). Biochemistry, in press. Lay, W. H., and Nussenzweig, V. (1968)./. E x p . Med. 128, 991. Lay, W. H., and Nussenzweig, V. (1969).J . Immunol. 102, 1172. Levadite, C. (1901). Ann. Inst. Pasteur 15, 894. Lobo, P. I . , and Burge, J. J. (1982). Eur. J . Immunol. 12, 682.

MEMBRANE COMPLEMENT RECEYIOKS FOR

C3

265

Medicus, R. G., and Arnaoiit, M. A. (1982). Fed. Proc., Fed. Ant. Soc. E x p . B i d . 41, 848. Medicus, R . G . , Melamed, J . , and Arnaout, M . A. (1983).Eur. J . I i i m i u n d . 13, 456. Medof; M. E., and Oger, J. J.-F. (1982). J. Clin. Lub. Zmnutiol. 7, 7. Medof, M. E., and Nussenzweig, V. (1983).Imniirnobiology 164, 275. Medof, M. E., and Nussennveig, V. (1984a). Fed. Prcrc., F e d A m . Soc. EX),.Biol. 43, 1666. Medof, M. E., and Nussenzweig, V. (19841)).J. E x p . Med. 159, 1669. Medof, M . E., and Prince, G. M. (1983a). Zntniutiology 50, 1. Medof; M. E., and Prince, G. M. (19831,). Itnmtrnology 50, 11. Medof; M. E . , Prince, G. M., and Mold, C . (1982a). Fed. Proc., Fed. Am. Soc. E x p . Riol. 41, 996. Medof, M. E., Prince, G . M., and Oger, J. J. F. (19821)). Clin. E r p . Zrnniunol. 48, 715. Medof, M. E., Prince, 6. M . , and Mold, C. (1982~).Proc. Nutl. Acud. Sci. U . S . A . 79, 5947. Medof, M. E., Froh, D., Prince. G . M., and Mold, C. (1982d). Clin. Hes. 80, 236A. Medot, M. E., Iida, K., Mold, C., and Nusseiizweig, V. (1982e).J. E x p . Med. 156, 1739. Medof, M. E., Lam, T., Prince, 6. M., aiid Mold, C. (1983a).J . Imtnunol. 130, 1336. Medof, M. E., Iida, K., Nussenzweig, V., Dykman, T., Dixit, R., and Atkinson, J. P. (198:3b). Clin. Hes. 31, 733A. Melanied, J . , Ariiaout, M . A., and Colten, H. R. (1982a). J. Iinmunol. 128, 2313. Melamed. J . , Arnaout, M. A., and Medicus, R. G. (1982b). Fed. Proc., Fed. Am. Soc. Ex),. B i d 40, 966. Meuth, J. L., Morgan, E. L., DiScipio, R. G.. and Hugli, T. E. (1983).J . Zm~ntcwl.130, 2605. Mills, S . E., and Levine, L. (1959). Zrnrtiunology 2, 383. Miyakawa, Y., Yamada, A , , Kosaka, K., Tsuda, F., Kosugi, E., and Mayumi, M. (1981). Luncet 2, 493. Morgan, E. L., Weigle, W. O., and Hugli. T. E. (1982). J. E x p . Med. 155, 1412. Miiller-Eherhard. H. J., and Schreiber, R. I). (1980). Ado. Irnitiwiol. 29, 1. Nagaki, K., Iida, K., Okubo, M., and Inai, S. (1978). Znt. Arch. Allergy A p p l . Ztnnzunol. 57, 221. Nagasawa, S., and Stroud, R. M. (1977). Zi,zi,irriiocheniistry 14, 749. Nagasawa, S., Ichihara, C . . and Strorid, R. M. (1980). /. Ztninunol. 125, 578. Natsuiime-Sakai, S., Hayakawa, J. I., and Takahashi, M. (1978).J. Zintnuriol. 121, 491. Nelson, R. A. (1953). Science 118, 733. Nelson, R. A. (1956). Proc. R. Soc. Med. 49, 55. Nelson, R. A., and Nelson, I). S. (lSS9). YaleJ. Bid. Med. 31, 201. Nelson, D. S. (1963). Ado. Ifninunol. 3, 131. Newinan, S. L., and Becker, S . E. (1983). Fed. Proc., Fed Am. SOC. E r p . B i d . 42, 428. Newman, S. L., and Johnston, R. B., Jr. (1979). J. Imniutiol. 123, 1839. Newinan, S . L., and Mikus, L. K. (1984). Fed. Proc., Fed. Am. Soc. E r p . R i d . 43, 1765. Newman. S. L., Mrisson, H. A . , and Henson, P. M . (1980). J . Ztnniunol. 125, 2236. Newman, S . L., Dobson, N. J., Lainbris, J. D., Ross, G. D., and Henson, P. M. (1981). Fed. Proc.. Fed. Am. Soc. E x p . B i d 40, 1017. Newman, S . L., Devery-Pocius, J., Ross, G. I > . , and Henson, P. M. (1984). Sulmitted For publication. Nicholsoii-Weller, A , , Burge, J., and Aristen, K. F. (1981). J. Inin~uncil.127, 2035. Nicholson-Weller, A , , Burge, J., Fearon, D. T., Weller, P. F., and Austen, K. F. (1982).J. Imrrtunol. 129, 184. Nicholson-Weller, A., March, J. P., Rosenfeld, S. I., and Aiisten, K. F. (1983). Proc. Ntatl. A d . Sci. U.S.A. 80, 5066. Okada, I*., and Nishioka, K . (1973). J . Zrntnunol. 111, 1444. Okuda, T., and Tachibana, T. (1981). Proc. Jpn. Soc. Zmnwnol. 11, 1. Pangliurn, M. K., and Miiller-Eherhard. H. J. (1978). Proc. Nutl. Acutl. Sci. U.S.A. 75, 2416.

266

G O R D O N D. ROSS A N D M . E D W A R D M E D O F

Pangburn, M. K., and Miiller-Eberhard, H. J. (1980).J. Exp. Med. 152, 1102. Pangburn, M. K., Schreiber, R. D., and Miiller-Eberhard. H. J. (1977).J. E x p . Med. 146, 257. Pangburn, M. K., Schreiber, R. D., and Miiller-Eberhard, H. J. (1983a).Proc. Natl. Acad. Sci. U.S.A. 80, 5430. Pangburn, M. K., Schreiber, R. D., Trumbold, J. S.,and Muller-Eberhard, H. J. (1983b). J. E x p . Med. 157, 1971. Parker, C. J., Baker, P. J., and Rosse, W. F. (1982).J. Clin.Znuest. 69, 337. Parker, C. J., Soldato, C. M., and Rosse, W. F. (1984a).J. Clin.Znuest. 73, 1130. Parker, C. J , , Soldato, C. M., and Telen, M. J. (1984b). J. Clin. Invest., in press. Pederson, S. E., Taylor, R. P., Morley, K. W., and Wright, E. L. (1980).J. Zmmunol. Methods 38, 269. Perlmann, P., Perlmann, H., and Muller-Eberhard. H. J. (1975). J . Exp. Med. 141, 287. Perlmann, H., Perlmann, P., Schreiber, R. D., and Miiller-Eberhard. H. J. (1981). J. E x p . Med. 153, 1592. Porter, R. R. (1983). Mol. Biol. Med. 1, 161. Porter, R. R., and Reid, K. B. M. (1978). Nature (London) 275, 699. Rabellino, E. M . , Ross, G. D., and Polley, M. J. (1978).J. Zntmunol. 120, 879. Rodriguez de Cordoba, S . , Dykman, T. R., Ginsberg-Fellner, F., Ercilla, G., Aqua, M., Atkinson, J. P., and Rubinstein, P. (1984). Proc. Natl. Acad. Sci. U . S . A . , in press. Ross, G. D. (1980).J. Zmmunol. Methods 37, 197. Ross, G. D., and Polley, M. J. (1975). J. Exp. Med. 141, 1163. Ross, G. D., and Lambris, J. D. (1982). J . Exp. Med. 155, 96. Ross, G. D., and Myones, B. L. (1984). Zmmunol. Tech., in press. Ross, G. D., and Rabellino, E. M. (1979). Fed. Proc., Fed. Am. SOC. E x p . Biol. 38, 1467. Ross, G. D., Polley, M. J., Rabellino, E. M., and Grey, H. M. (1973).J. Exp. Med. 138, 798. Ross, G. D., Jarowski, C. I., Rabellino, E. M., and Winchester, R. J. (197th).J. Exp. Med. 147, 730. Ross, G. D., Winchester, R. J., Rabellino, E. M., and Hoffman, T. (1978b).J. Clin. Znuest. 62, 1086. Ross, G. D., Lambris, J. D., Cain, J. A., and Newman, S . L. (1982).J. Zmmunol. 129, 2051. Ross, G. D., Newman, S. L., Lambris, J. D., Devery-Pocius, J . E., Cain, J. A., and Lachmann, P. J. (1983a).J. Exp. Med. 158, 334. Ross, G. D., Thompson, R. A., Ward, R. H. R., Walport, M. J . . Harrison, R. A., and Lachmann, P. J. (1983b). Proc. Workshop Clin. Aspects Complement Mediated Dis., Bellagio, Ztaly p. 15. Ross, G. D., Cain, J. A , , and Lachmann, P. J. (1984a). Submitted for publication. Ross, G. D., Thompson, R. A., Walport, M. J., Springer, T. A., Watson, J. V., Ward, R. H. R., Lida, J., Newman, S. L., Harrison, R. A., and Lachmann, P. J. (1984b). Submitted for publication. Ross, G . D., Walport, M. J., Parker, C. J., Lentine, A. F., Fuller, C. R., Yount, W. J., Myones, B. L., Winfield, J. B., and Lachmann, P. J. (1984~).Arthritis Rheum. 27, S28. Rothman, I. K., Gelfand, J. K., Fauci, A. S., and Frank, M. M. (1975).J. Zmmunol. 115, 1312. Ruddy, S., and Austen, K. F. (1969). J . Zmmunol. 102, 533. Ruddy, S., and Austen, K. F. (1971). J. Zmmunol. 107, 742. Sanchez-Madrid, F., Simon, P., Thompson, S., and Springer, T. A. (1983a).J. Exp. Med. 158, 586. Sanchez-Madrid, F., Nagy, J. A,, Robbins, E., Simon, P., and Springer, T. A. (1983b).J. E x p . Med. 158, 1785. Scharfstein, J., Ferreirn, A,, Gigli, I., and Nussenzweig, V. (1978). J. Exp. Med. 148, 207. Schenkein, H. A,, and Genco, R. J. (1979).J . Zmmmol. 122, 1126. Schreiber, R. D. (1983). Springer Setnin. Zmmunopathol., in press.

MEMBRANE COMPLEMENT RECEPTORS FOR C 3

267

Schreiber, R. D., Pangburn, M. K . . Lasavre, P. H., and Muller-Elwrhard, H. J. (1978). Proc. Natl. Acad. Sci. U.S.A. 8, 3948. Schreiber, R . D., Pangburn, M. K . , Bjornson, A. B . , Brothel-s, M . A , , and Miiller-Eberhard, H . J. (1982). Clin. Immunol. Zmrnunopathol. 23, 335. Siegal, I . , Liu, T. L., and Gleicher, N. (1981). Lancet 2, 556. Sim, E., and Sim, R. B. (1984). Biochem. J , , in press. Siqueira, h i . , and Nelson, R. A. (1961).J . Immunol. 86, 516. Spitznagel, J. K . , Dalldorf, F. G., Lefell, M. S . , Folds, J. D., Welsh, I. R. H., Cooney, M . H., and Martin, L. E. (1974). Lab Znuest. 30, 774. Tack, B. F. (1983). Springer Semin. Zmmunopathol. 6 , 259. Takahashi, M . , Czop, J., Ferreira, A . , and Nussenzweig, V. (1976). Trunsplant. Reu. 32, 121. Takata, Y . , Tamura, W., and Fujita, T. (1984).J . Immunol. 132, 2531. Tamura, N . , and Nelson, R. A. (1967).J . Zmmunol. 99, 582. Taylor, J . C., Crawford, 1. P . , and Hugli, T. E. (1977). Biochemistry 16, 3390. Taylor, R. P., Horgan, C . , Buschbacher, R., Brtrnner, C. M., Hess, C. E., O’Brien, W. M., and Wanebo, H . J. (1983). Arthritis Rheum. 26, 236. Tedder, T. F., Fearon, D. T., Gartland, G. L., and Cooper, M. D. (1983).J . Zmmunol. 130, 1668. Tedder, T. F., Clement, L. T., and Cooper, M . D. j1984). J . Immunol. 133, 678. Tsuda, F., Miyakawa, Y., and Mayumi, M. (1980). Zmmiinology 37, 681. Turk, J. L. (1962). Immunol. Methods C.I.O.M.S. Symp. p. 10. Van Loghern, J. J . , and Van der Hart, M. (1962). Vox Sang. 7, 539. Vik, 1).P., and Fearon, D. T. (1984). Fed. Proc., Fed. Am. SOC. Exp. B i d . 43, 1665. Vranian, G . , Jr., Conrad, D. H., and Ruddy, S. (1981).J . Immunol. 126, 2302. Wahlin, B., Perlmann, H . , Perlniann, P . , Schreiber, R. D., and Miiller-Eberhard, H. J. (1983). 1. Zmniunol. 130, 2831. Wallace, J. M., and Wormall, A. (1931). Parisitdoggy 23, 346. Walport, M. J . , Ross, G. D., Mackworth-Young, C., Watson, J. V., Hogg, N . , and Lachmann, P. J. (1984). Clin. E x p . Zmmunol., in press. Weigle, W. O., Goodman, M. C . , Morgan, E. L., and Hugli, T. E. (1983). Springer Semin. Zmmunopathol. 6, 173. Weiler, J. M., Daha, M. R., Austen, K. F., and Fearon, D. T. (1976). Proc. Natl. Acarl. Sci. L7.S.A. 73, 3268. Weis, J. J., Tedder, T. F . , and Fearon, D. T. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 881. Whaley, K. (1980).J . E X ~ JMed. . 151, 501. Whaley, K . , and Ruddy, S. (1976a).J . E x p . Med. 144, 1147. Whaley, K.,and Ruddy, S. (1976b). Science 193, 1011. Wilson, J. G . , Wong, W. W., Schur, P. H . , and Fearon, D. T. (1982). New Engl. J . Med. 307, 981. Wilson, J. G., Tedder, T. F., and Fearon, D. T. (19813).J . Imtnunol. 131, 684. Wong, W., Wilson, J., and Fearon, D. (1983).1. Clin. Znuest. 72, 685. Wright, S. D., and Silverstein, S. C. (1982).J . E x p . Med. 156, 1149. Wright, S. D., and Silverstein, S. C. (1983).J . E x p . Med. 158, 2016. Wright, S. D., Rao, P. E., van Voorhis, W. C . , Craigmyle, L. S., Iida, K . , Talle, M. A., Westberg, E. F., Goldstein, G . , and Silverstein, S. C. (1983a). Proc. Natl. Acad. Sci. U . S . A . 80, 5699. Wright, S. D., Craigmyle, L. S., and Silverstein, S. C. (1983b).J . E x p . Med. 158, 1338. Wright, S. D., Licht, M. R., and Silverstein, S. C. (1984).Fed. Proc., Fed. Am. Soc. E x p . Biol. 43, 1487. Yefenof, E., Klein, G . , Jondal, M., and Oldstone, M. B. A. (1976). Znt. J . Cancer 17, 693. Ziccardi, R. J. (1983). Springer Semin. Zntmunopathol. 6 , 213.

This Page Intentionally Left Blank

AIWANCES IN IMMUNOLOCY. VOL 37

Murine Models of Systemic lupus Erythematosus' ARGYRIOS N. THEOFILOPOULOS AND FRANK J. DIXON Department of Immunology, Research Institute of Scripps Clinic, La Jolla, California

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Derivation of Lupus Mice . . . . . . . . . 111. Natural History and Patholog ...................... '4. Survival and Body Weight B. Morphologic Manifestations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Serologic Manifestations ...................................... IV. Cellular Abnormalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Transfer of Autoimmune Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Surface Characteristics and Numerical Abnormalities of T and B Cells . . . . . . C. Functional Ahnorma~itiesof B Cells, T Cells, Macrophages, and Related

269 27 1 273 274

276 284 30 1 301 303

. . . . . . . . . . . . . . . 309 I). Defects in Tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . E. Thymic Defects . . . . . . . ...................................... V. Genetics of Murine SLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... A. Inheritance of the Aittoimmune Traits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Relationship among Autoirnmune Traits and Their Association with Disease . C. Complementarity of Genetic Backgrounds among Lupus Mice and the Role of Accelerators and Other Autosonial Genes . . . . ............... D. Relationship between the Traits, H-2, and Ig Genes . . . . . . . . . . . . . . . . . . . . . VI. Influence of Sex and Sex Hormones on the Pathogenesis of Murine SLE . . . . . . VII. Viruses in Murine SLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... VIII. Treatment of Murine SLE ...................................... IX. Conclusions ........................ .......... References ......................................

330 332 336 337 338 34 1 345 346 348 350 355 358

I. Introduction

The etiopathogenesis of autoimmune diseases cannot readily be analyzed without appropriate animal models. Inbred mice that spontaneously develop a disease similar to human systemic lupus erythematosus (SLE) have been an important model for elucidating the pathogenesis of this disease, and an enormous help in the development and testing of various immunologic con1 This is Publication No. 3665IMM from the Department of Immunology, Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, California 92037. Our work cited herein was supported, in part, by National Institutes of Health Grants AI-07007, AM-31023, AM-33826, National Cancer Institute Grants CA-27489 and AG-01743 and the Cecil H. and Ida M. Green Endowment Fund.

269 Copyright 0 19x5 try Academic Prcsc. Inc. All rights or reproduction in any form reserved. ISBN 0-12-022437-2

270

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

cepts. Murine SLE is a good model of human SLE since most of the immunologic abnormalities apparently fundamental to the human disease also appear to be operative in the mouse. Thus, it is felt that understanding the pathogenesis and etiology of murine lupus should lead to a better comprehension of this human disease, and of autoimmunity and immunoregulation in general. New Zealand Black (NZB) mice and the F, hybrids produced by the mating of NZB with New Zealand White (NZW) mice [(NZBXW)F,] were the first described models of spontaneous lupus-like autoimmune disease (1,2). Two and a half decades of study of this disease in New Zealand (NZ) mice provided a reasonable understanding of its pathogenesis, and enumerated a variety of immunologic and virologic peculiarities, but contributed little real knowledge about the etiologic mechanisms involved. Moreover, the availability of only one strain of mouse and its hybrid in the study of this complex immunologic disorder made it impossible to determine whether the unusual immunologic and virologic features observed in these mice were causal to disease, secondary epiphenomena, or merely idiosyncracies of the NZ murine strains. The recent availability of two new strains of mice (MRL, BXSB) that also spontaneously develop a SLE-like syndrome (3-5)has greatly enhanced the potential value of murine SLE as an experimental model. Now this disease can be observed in mice of three quite different genetic backgrounds and immunopathologic constellations, and essential immunopathologic common denominators of the disease can be identified. Other murine strains mentioned as developing SLE-like characteristics are the Palmerstrom-North (6,7) and the Motheaten (8-10) mice, but studies on these strains are very limited and, therefore, will not be considered further in this review. This article will review pertinent information and recent observations regarding the natural history and salient histoimmunopathologic features of SLE as it occurs in NZ, BXSB, and MRL mice, and subsequently analyze their serologic, histologic, immunologic, virologic, genetic, and hormonal abnormalities. Therapeutic approaches utilized for prevention and/or treatment of murine SLE will also be outlined. One of the most interesting results emerging from the study of these several SLE strains is that this disease is caused by genetically determined abnormalities of the hematopoietic stem or lymphoid precursor cells, and not by abnormal autoantigens. These abnormalities are commonly expressed as a generalized B cell hyperactivity leading to hypergammaglobulinemia, early IgM to IgG antibody switching, production of autoantibodies against a wide array of endogenous antigens, formation of immune complexes (ICs), and a variety of histopathologic manifestations of which glomerulonephritis is the most prominent. The immunologic abnormalities leading to B cell

MURINE MODELS OF SLE

27 1

hyperactivity have not yet been fully elucidated. Nevertheless, recent studies suggest that major mechanisms may be hyperresponsiveness to and/or hyperproduction of B cell proliferation/differentiation promoting interleukins derived from accessory cells, particularly T-helper cells. Furthermore, these recent studies have demonstrated that whatever the basic iminunologic defects, the development of disease may not occur until late in life unless independent accelerating factors are superimposed. These accelerating factors may be endogenous (abnormal genes, hormones) or exogenous (viral, bacterial). The effects of the accelerators are not restricted to a particular strain, but do differ depending on the genetic background on which they act. II. Derivation of Lupus Mice

The NZB (H-2c’) strain, which inherits hemolytic anemia, was bred for coat color from an unknown background in the late 1950s by M. Bielschowsky at the Otago University Medical School, Dunedin, New Zealand (1). Some of the NZB lines are positive for a lymphocyte surface antigen that is maternally transmitted (Mta), whereas others are not (11,12). Those being Mta+ have probably been contaminated by “old i n b r e d strains, as shown by the similarity of their mitochondria1 DNA restriction endonuclease patterns with that of “old inbred” strains (13). Such NZB mice have a substantially modified autoimmune disease, including delayed onset and low titers of autoantibodies and prolonged survival compared to the noncontaminated Mta- NZB lines (13,14). The phenotypically normal NZW (H-2”)was developed by W. H. Hall at the Otago Medical School Animal Breeding Station (cited in 15). In 1963, Helyer and Howie (2) described how the original hemolytic disease pattern in the parent NZB changed to one of florid renal dysfunction and failure in the (NZBx W)F, hybrids. These hybrid mice underwent changes remarkably similar to, if not identical with, those of human lupus nephritis. The short lifespan of the female NZBx W mouse and its relatively uniform pathologic behavior immediately offered considerable advantages as an experimental model in the study of autoimmune disease and have subsequently been studied by a large number of investigators. The MRL and BXSB SLE strains were developed in 1976 by Murphy and Roths at the Jackson Laboratory (reviewed in 3,4). The MRL (H-2k) strain originated as a by-product of a series of crosses involving inbred strains AKR/J (H-2k), C57BL/6J (H-2’>),C3H/Di (H-2k), and LG/J (H-2d’?. The series of crosses began in 1960 to transfer a mutation for achrondroplasia (cn) from the high leukemic background of strain AKR to a background without early incidence of leukemia. A new problem of dental malocclusion developed, and was not eliminated until the final backcrosses to strain LG/J. From this point,

272

AHGYHIOS N . THEOFILOPOULOS A N D FRANK J. DlXON

rigid inbreeding was carried out. It was estimated froin their natural history that the new composite genome of the MRL mouse derived 75% from LG, 12.6% from AKR, 12.1% from C3H, and 0.3%from C57BL/6. In the twelfth generation of inbreeding, some of the offspring developed massive generalized lymph node enlargement early in life, while the others did not. These two types of mice were separated and, with subsequent in-breeding, became two sublines of the MRL mouse: one termed MRL/l that expressed lymphadenopathy, and the second termed MRL/n that did not. The two substrains shared at least 89% of their genome. Breeding tests involving F,, F,, and reciprocal backcrosses between MRL/1 and MRL/n inice indicated that the massive lymphoproliferation was controlled by a single autosomal recessive gene, termed lymphoproliferation (Zpr). Genetic linkage of Zpr has not yet been established despite the fact that approximately 47% of the autosoinal genome has been tested and 27 linkage markers, including H-2, on various chromosomes have been analyzed (4). The isolation of Zpr in substrain MRLA permitted the ready development of a pair of congenic inbred lines by transfer of the mutant gene from MRL/l mice to mice derived from the MRL/n stem line. The original substrain MRWn was redesignated MRL/ Mp-+/+ at F28. By early 1978, the Zpr gene had been transferred by 5 cycles of cross-intercross matings to MRL/Mp-+ / with an estimated residual difference between the lines MRL/Mp-Zpr/Zpr and MRL/Mp-+/+ of less than 1%of the genome. By 1980,lO cycles had been completed, reducing the residual difference to less than 0.1%. Recent tryptic peptide map analysis demonstrated that the H-2k haplotype of MRL mice is derived from the AKR and not the C3H ancestors (16). Monocytic (Ly-1 -, MAC-1 +, MAC-2 + ) cell lines expressing small amounts of surface IAk, IEk, and inappropriate H-2", IA", IE" cell surface antigens have been isolated from MRL/I mice (16a). Such cells, representing a minute subset of MRL/l lymph node monocytes which derive their H-2 from the LG/J (H-2") ancestor, have been hypothesized to play a significant role in the lymphoid hyperplasia of this mouse by inducing Ly-l+ , H-2k T cells to proliferate. Transfer of the Zpr gene to other standard inbred strains used in cancer and iminunogenetic research was recently accomplished. These strains include AKR, BALB/c, C3H/HeJ, C57BL/6, C57BL/10, and SJL (4,17). These new congenic inbred strains with Zpr do develop lymphoproliferation and autoantibodies, but the histologic manifestations of SLE are much less evident than in the MRL-lprllpr substrain (see Section V). The availability of these mice makes it possible to begin to separate accelerating gene effects from background modifications and to study interactions with other identified genes controlling the immune system. In addition to these strains, a new autosomal recessive mutation gZd (generalized lymphoproliferative disease) which occurred in uninanipulated C3H/HeJ mice produces early-onset autoimmune responses and massive

+

MUHINE MODELS OF SLE

273

lymphoid hyperplasia (4,18). However, this gene is not allelic with Zpr and appears to be located on chromosome 1 between the genes P e p 3 and Lp. Mutant gld mice live only half as long (12 months) as normal controls (23 months). By 20 weeks, lymph nodes of such mice are 50-fold heavier than those of coisogenic CSH/HeJ-+/+ mice with increased B cells, T cells, and particularly null cells. Serologically, gld mice develop antinuclear, antithymocyte autoantibodies and hypergammaglobulinemia, but only about 14% of the autopsied mice have significant lupus-like nephritis and none have vascular disease (18). The recombinant inbred strain BXSB/Mp (H-2”) derives from a cross between a C57BL/6J (H-2”) female and a SB/Le (H-2“) male (3,4). Inbred strain SB/Le, originally developed from a noninbred stock, is homozygous for the linked mutant genes satin (sa) and beige (bg) on chromsome 13. Homozygous beige mice show pigment dilution and giant lysosomal granules homologous to those of the Chediak-Higashi syndrome in man. Initially, an association of spontaneous pneumonitis with the homozygous beige genotypes of backcrosses of [(C57BL/6JxSB/Le)F,] hybrids to SB/Le and of a typical lymphoproliferation in nonbeige male (B6x SB)F, hybrids and SB x (B6x SB)F, backcrosses was observed. To study the lymphoproliferative disease, matings were designed to produce recombinants separating satin from beige, starting with an SBILe-sa bglsa hg male and a C57BL/6J female. Following 4 generations and subsequent brother x sister matings with forced heterozygocity at the sa locus, the recombinant inbred strain BXSB/Mp was derived. The sa gene did not influence autoimmune expression, and was eliminated by further breeding. In contrast, a trait or gene(s) on, or linked to, the Y chromosome of BXSB mice and apparently derived from SB/Le (3), was found to profoundly enhance autoimmunity in this strain of lupus mice. The coat color genes, polymorphic isoenzyme and biochemical loci, H-2, lymphocyte surface alloantigens and IgH-1 (IgG,,) allotype of all the main lupus strains described above are given in Table I. 111. Natural History and Pathology of lupus Mice

The natural history of the various SLE strains has been described by Murphy and Roths (3, 4), Theofilopoulos and Dixon (19), Warner (20), and Howie and Helyer (21). These accounts from different laboratories show considerable agreement. Of the histopathologic and serologic abnormalities of SLE mice, some are common to all strains and others are not, and considerable differences in disease onset exist. The major focus of this review relative to the pathology of SLE mice, particularly of the new strains, is primarily based on our detailed studies of several thousands of these mice over the last 6-7 years.

274

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

TABLE I DERIVATION A N D GENETIC MARKERS OF SLE MICE NZB

NZW

BXSB

Derivation

Unknown

Unknown

50% C57BLl6JO 50% SB/Led

Coat color H-2 I&-1 (IgG,,) allotype K light chain Vregion allotype Lymphocyte surface alloantigens Polymorphic isozyme and biochemical loci

Black H-2d

White H-2’ e

Agouti (brown) H-2L b

e

75% LGlJ 13% AKWJ 12% C3HlDi 0.3% C57BU6J White H-Zk j

Igk-EBa

Igk-EBb

Thy-1.2, Lyt-1.2, Lyt-2.2, Lyt-3.2, Qa-lla, Mlsa Car-la, Car-ea, Es-lL, Es-Z”, E s - ~ Es-G’, ~, Es-loh, Gpi-ls, Gpd-1‘1, Gr-la, Hbbd, Id-la, Ldr-la, Mod-l”, Mod-2‘, Pgm-Zd, Acf-lS, Gpt-ll), Lap-lb, Hco, Pre-la

MRL

-

-

Thy-1.2, Lyt-1.2, Lyt-2.2, Lyt-3.2, TL-, Qa-lL Amy-la, Car-Zn, Dip-I”, Es-la, Es-Zb, E s - ~ c , Es-lOa, Got-la, Got-Zb, Gpi-la, Gpd-l”, Gr-la, Hbbs, Id-la, Mod-la, Mod-211, Np-la, Pgm-la, Pgm-2“

Thy-1.2, Lyt-1.2, Lyt-2.1, Lyt-3.2, TL-, Qa-l’j Amy-le, Car-2”. Dip-lb, Es-lh, Es-Zb, E s - 3 ~ , Es-10.1, Got-la, Got-2b. Gpi-la, Gpd-lb, Gr-la, Hbba, Id-la, Ldr-la, Mod-la, MOd-ZL, Np-lH, Pgm-1”. Pgm-2”

The most consistent abnormality is the development of immune complexmediated glomerulonephritis. However, this dominant feature is associated with several other disorders such as thymic atrophy, lymphoid system hyperplasia, and different types of vasculitides, some associated with myocardial infarction. Thus, the health of the animals, as measured by survival, is determined by the interaction of several pathologic manifestations. The histologic and serologic abnormalities associated with the various lupus strains are listed in Table 11. A. SURVIVAL AND BODYWEIGHT In each kind of SLE mouse, the disease occurs in two forms depending upon the presence or absence of accelerating factors (3,19,21,22): a late-life variety in which typical SLE develops in the second year of life, and an early acute form in which the immunologic abnormalities become observable as early as 1or 2 months of age with significant mortality in the first 6 months of life (Fig. 1).Thus, the mean survival time observed in our breeding colony

275

MUHINE MODELS OF SLE

TABLE I1 H I S T O L ~ ~A; N IC D SEHOLOC:IC: ABNO~WALITIES OF SLE MIC:E: Histoiniinunopathologic characteristics

Strain NZB NZBxW MRW1 BXSB

GN

+ +++ +++

+++

Thymic atrophy

Lymphoid hyperplasia

+

+ + +++ ++

+ + +

Arteritis 0 0

+ 0

Myocardial infarcts

+ + + +

Arthritis

0 0

+ 0

Serologic characteristics Common: Hy~~ergarnmaglobu~ine~nia, ANA, anti-ds- and ssDNA antibodies, antihapten antibodies, high levels of retroviral gp70, immune complexes (gp70-anti-gp70, DNA-anti-DNA, and others), reduced complement levels (NZB is C5-deficient), cryoglobulins Uncornnion: High incidence of inomclonal or oligoclonal IgGs in MRWI, anti-Sm antibodies in M R U 1 and MRL/n, IgM and IgG rheumatoid factors in MRWl (and other lpr homozygous substrains), antierythrocyte antibodies in NZB mice, NTA primarily in NZB and NZBX W mice

for the various lupus strains are NZB, females 430 days and inales 469 days; NZB x W, females 245 days and males 406 days; BXSB, females 574 days and males 161 days; MRL/l, females 143 days and inales 154 days; MRL/n, females 476 days and inales 546 days. The phenotypically normal NZW mice have a mean survival time of 675 days for females and 717 days for males, which is very similar to those of normal mice. The mean survival figures quoted by Howie and Helyer for NZ mice (21) and by Murphy and Roths for BXSB and MRL mice (3) are in close agreement with our values.

FIG. 1. Mortality curves for late-life and accelerated forms of SLE occuring in NZBXW, BXSB, and MRL mice.

276

AHGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

The average body weight for each of the above strains obtained at 5 months of age is as follows: NZB, 40 g; N Z B x W , 41 g; BXSB, 21 g; MRL/n, 43 g; MRL/l 41-48 g (males and females, respectively). B. MORPHOLOGICMANIFESTATIONS 1 . Glomerulonephritis

The major cause of death in all lupus mice is glomerulonephritis, which ranges from an exudative and proliferative acute form in the BXSB male to a largely subacute proliferative form in the MRL/l male and female, and a more chronic obliterative form in the N Z B x W female. Only in the BXSB male mice are polymorphonuclear leukocytes a significant element in the glomerular lesions. In the MRL/I mice, glomerular lesions involve the accumulation of monocytes and proliferation of both endothelial and mesangial cells with occasional crescent formation and basement membrane thickening. The obliterative lesion in the NZB x W female is accompanied by heavy mesangial and, at times, intravascular proteinaceous deposits, moderate proliferation of all glomerular cellular elements, and crescent formation. In a relatively small proportion of NZB mice that appear to develop particularly severe hemolytic disease and in which death results principally from profound anemia, the kidneys have normal dimensions and texture except for hemosiderin deposits. The majority of older NZB mice (>12 months) manifest mesangial proliferation in their glomeruli and a chronic form of glomerulonephritis with hyalinization of the glomeruli. Heavy proteinuria and associated anasarca are most frequent in the NZBx W females. At 5-6 months of age, these mice have a mean urinary protein value of 6.6 mg/day, and a 40% incidence of advanced anasarca at the terminal stages of the disease. MRL/I males and females and BXSB males between 3 and 6 months of age have urinary protein values from 2.6 to 3.8 mg/day with an incidence of terminal anasarca approximately half that of NZB x Ws. Immunofluorescence studies (5,21-24) reveal granular deposits of mouse IgG and C3 of variable intensity in one or more locales, including glomerular capillary walls, mesangial and tubulointerstitial sites in the above 3 early-life SLE developing substrains. In the N Z B x W mice, deposits are present largely in the glomerular mesangium at or before 5 months of age. With disease progression and widening of the glomerular mesangium, heavy mesangial deposits occur, generally with accompanying glomerular capillary wall deposits. A predominant or exclusive glomerular capillary wall pattern of deposits is also occasionally seen. Extraglomerular renal deposits of IgG and C3 are present in the peritubular tissue and arterioles, and increase in frequency with age. Striking glomerular deposits of IgG and C3 involving

MURINE MODELS OF SLE

277

both the mesangium and the glonierular capillary wall of the hypercellular BXSB glomerulus are observed as early as 3 months. Extraglomerular deposits similar to those in the NZBxW mice increase in frequency with age. Granular IgG and C3 deposits in the MRL/I strain also increase from 2 to 5 months. The deposits are present in both the capillary wall and the mesangium, often predominating in the former where the finely granular deposits are, in general, closely approximated. Immunofluorescent studies also reveal the presence of DNA and of serum gp70 in the nephritic glomeruli of all SLE strains in patterns similar to those seen for IgG and C3; elution of the kidney-bound IgG with low pH or high salt buffers prior to staining with anti-DNA may be necessary for the DNA deposits to be observed (23). Murine retroviral envelope gp70 antigen is detectable with an anti-Rauscher gp70 antibody in granular deposits on glomerular and tubular basement membranes (5,24). Such deposits, however, are often much less striking than the IgG and C3 deposits; some glomeruli of mice with impressive IgG accumulations have no gp70. The variability of gp7O staining and its disparity with IgG and C3 are greatest in the MHL/I strain. The nature and specificities of renal Ig deposits of these mice have not been fully characterized. Studies by Lambert and Dixon (23) demonstrate only IgG is in NZBxW kidney eluates, the vast majority of the IgG,, subclass. More recently, Slack et uZ. (25) tested kidney eluates from SLE mice to determine the IgG subclasses present in the form of immune complexes. IgG eluted from the BXSB male kidney is dominated by IgG,,, antibody, with smaller amounts of IgG,, IgG,,, and IgG,. Similarly, IgG eluted from MRL/l kidneys is predominantly of the IgG,, and IgGZbsubclasses, whereas in NZB and NZBXW, IgG,, exceeds the smaller amounts of remaining subclasses. Initial studies by Lambert and Dixon (23) in NZBxW mice and subsequently by Andrews et aZ. (5) in all SLE strains show that concentrations of antinuclear antibody (ANA) in IgG eluted from kidneys vs the concentration of ANA in serum IgG are from 2 to 10 times higher in eluates from NZBX W kidneys and 2 to 4 times higher in eluates from both MRL/l and BXSB kidneys. Antibodies against double-stranded (ds)DNA are considerably more concentrated in NZB x W renal eluates, being 25-31 times greater than in serum, whereas MRL/l eluates show up to 6 times the concentration and BXSB eluates u p to 12 times. Antibodies to single-stranded (ss)DNA are also more concentrated in eluted renal IgG than in serum IgG with values 513 times greater for NZBx W and 11-21 times for BXSB eluates. In some studies, specific concentrations of antiretroviral antibodies in such kidney eluates have also been observed (26). The isoelectric spectra of the IgG anti-DNA antibody in renal eluates have

278

ARGYHIOS N . THEOFILOPOULOS AND FRANK J. DIXON

been examined by Ebling and Hahn (27), who observed that eluates of MRL/I and NZBx W mice contain a restricted number of DNA-binding bands, all ofwhich focus at pH 8.0-9.0. It was suggested that subpopulations of IgG anti-DNA antibodies, i.e., those with an alkaline pl, are more pathogenic than others. Several studies in experimental animal systems (28-31) have shown that cationic (positive charge), but not anionic (negative charge) ICs can give rise to glomerular basement membrane deposits in subendothelial and subepithelial sites and emphasize the potential importance of charge as a determinant of IC localization.

2 . Thymic Atrophy A second consistent feature of SLE pathology in all strains is severe cortical thymic atrophy (5,19). The initial lesion appears to consist of cortical thymocyte loss with or without later medullary degeneration, often cystic. While 90% of the mice show medullary atrophy, 5-10% in each group reveal what appears to be medullary or stoma1 hyperplasia which maintains or even increases the normal thymic size in spite of a cortical loss. This medullary hyperplasia does not correlate consistently with any other clinical or pathologic feature. The thymic atrophy associated with abnormal fine structure appears by the fourth month in female NZBxW mice, which at 6-7 months of age have lost 70-90% of their cortexes. In BXSB males and MRL/I mice of both sexes, thymic atrophy and cystic necrosis appears by 2 months of age and progresses to a complete loss of the cortical areas by 4.5 and 3.5 months of age, respectively. The histologic lesions of the NZ thymus have been studied in detail by several investigators (reviewed in 21). Lymphoid hyperplasia and sometimes follicular aggregations of lymphocytes in the medulla, numerous mast cells, epithelial degeneration and vaculization, and abnormal age-dependent evolution of intrathymic lymphocyte populations have been observed. Moreover, abnormalities in thymic epithelial cells correlate with a decrease in thymic hormonal levels in serum occurring with age in such mice (32,33). However, the relationship of both phenomena to autoimmune expression remains uncertain. Studies to date have failed to show significant improvement by treatment of NZ mice with thymic extracts (34,35). 3. Vascular Disease and Myocardial lnfarction Fifteen to 30% of each of the above types of mice have, at autopsy, old and/or acute myocardial infarcts involving either ventricle and judged extensive enough to be a contributing cause of death. Studies by Accinni and Dixon (36) demonstrated that medium and small coronary arteries and arterioles of such animals have focal degenerative lesions consisting of periodic acid-Schiff (PAS)-positive or eosinophilic material deposited in the intima

M U H I N E MODELS OF SLE

279

(and, to a lesser extent, in the media), degenerative changes in the media without accompanying cellular inflammation, and occasional intimal cell proliferation or swelling. These lesions, together with platelet aggregation, occasionally occlude the vascular lumens. Granular deposits of mouse immunoglobulin, C3, and occasionally gp70 are present in the walls of medium and small arteries, arterioles, and venules of myocardia both with and without infarcts. Moreover, dense deposits of foreign material are found by electron microscopy in areas corresponding to the immune deposits. These findings are consistent with the interpretation that the noninflainmatory coronary lesions are caused by local deposition of antigen-antibody complexes. The IC-mediated injury appears to lead to thrombotic and/or obliterative vascular changes contributing to decreased coronary blood flow and development of myocardial infarction. In contrast to the relatively low incidence of degenerative coronary disease in these mice, studies by Hang et al. (37) of male F, offspring from crosses of NZW females and BXSB males demonstrate an incidence of degenerative cornonary vascular disease and associated myocardial infarcts as high as 80%.The majority of vascular and myocardial lesions of these hybrids are located in the right ventricle, right atrium, and left subendocardium. Systemic hypertension, secondary to autoimmune disease and IC glomerulonephritis, appears to predispose to the increased cardiovascular disease

(38). In addition to the degenerative vascular disease in early life SLE mice, approximately 75% of older MRL/I mice have a necrotizing exudative polyarteritis involving mostly medium-sized arteries of the kidney, genital organs and heart (5). In spite of the coronary arteritis in this strain, the incidence of myocardial infarcts is the same as for BXSB and NZBxW, suggesting that arteritis per se does not significantly predispose to myocardial infarction. Detailed serologic and histologic studies by Berden et al. (39) indicate that the noninflammatory vascular disease of the (NZWx BXSB)F, males is associated with an early onset of minimal autoantibody production, which gives rise to persistent low levels of circulating ICs. The necrotizing polyarteritis of MRL/1, however, is associated with sudden development of high autoantibody levels and circulating ICs.

4 . Lymphoid Hyperplasia Marked splenic and lymph node hyperplasia exists in all murine lupus strains, and lymphoid infiltrates may also occur in the lungs, kidneys, liver, salivary glands, and bone marrow. The degree of lymph node hyperplasia varies considerably among strains, resulting in lymph node sizes ranging from normal to 2 or 3 times normal in

280

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

older NZBx W females; 10 to 20 times normal in older BXSB males, and up to 100 times normal in 4- to 5-month-old MRL/l mice (4,5,19). Lymph node and splenic changes in NZB mice have been studied extensively. Two phases of lymphoproliferative change have been described (4042). The first occurs from 3 to 11 months of life, and consists primarily of extensive development of large lymphoid follicles with multiple germinal centers in the white pulp of the spleen and lymph node cortex. These changes may be seen even in the absence of positive autoimmune markers, and gradually shade into a second phase of significant plasma and reticulum cell hyperplasia in the white pulp of spleens and medullas of lymph nodes, as well as in other lymphoid tissues. This second phase is consistently associated with overt autoimmune disease. Lymphoid infiltration of nonlymphoid organs is most prominent in the lung, where peribronchovascular aggregation of lymphoid tissue is frequent, particularly in the hilar region. This may be relatively infrequent in young animals, however, 80% of all NZB mice beyond 12 months of age are so affected. As the infiltrate may be extensive enough to occupy a considerable volume of pulmonary parenchyma, the lesion may cause marked respiratory distress. All MRL/l mice develop massive generalized lymph node enlargement, commencing by 8 weeks of age and progressing to over 100 times control weights by 16-18 weeks (4,s).The degree of lymph node enlargement is somewhat smaller in males than females. Often a single mesenteric node is larger than any of the other massively enlarged peripheral nodes (up to 2200 mg compared to 50 mg for similar normal nodes) and can yield from S X10s to 1x lo9 cells. The proliferation is predominantly lymphocytic, with admixtures of histiocytes, plasma cells, and immunoblasts. The diffuse proliferation of small lymphocytes obliterates normal nodal architecture. There is a normal plasma cell density at 2 months of age, but by 3 months, increasing numbers of small clusters of plasma cells are evident. By 4 to S months of age, large groups of plasma cells, many of which contain prominent Russell bodies, develop in the medulla concurrent with overt lymphoproliferation. Plasma cells also are seen singly and in groups among the massive areas containing the proliferating small lymphocytes. Iminunoperoxidase analysis of spleen and lymph node sections indicated that all murine isotypes and IgG subclasses are represented in the infiltrating plasma cells, but IgG, and IgG,, containing cells are most prominent. In one-third to one-half of older, terminally ill MRL/I mice, the larger nodes show extensive hemorrhagic and cystic necrosis which is probably responsible for the frequent clinically evident terminal reduction (sometimes one-third of the maximum) in lymph node size (4,s). The cell cycle analysis by Raveche et al. (43) of lymph node cells from older MRL/I mice indicates that fewer than 5% of such cells are proliferat-

M U R I N E MODELS OF SLE

28 1

ing, which is similar to values in control mice of several strains and substantially less than proliferation in spleens (17-25%) and bone marrows (20-24%) of MRL/l mice. Since the majority of these lymph node cells appear to be in a resting state, it was suggested that most lymphocytes accumulate as a result of migration rather than in situ proliferation. The proliferative signals, therefore, appear to be delivered prior to the cells' appearance in lymph nodes, possibly in the spleen or thymus. Because old MRL/l mice, however, have marked numerical increases in lymph node cells, the absolute number of proliferating cells is actually increased. Furthermore, [3H]thymidine uptake studies by Isakov et al. (unpublished observations) demonstrated that within the first 24 hours of culture, MRL/l lymph node cells incorporated 7to 8-fold more thymidine than normal cells; this higher uptake ceases after this point. In addition to lymph node enlargement, there is an approximate 7-fold enlargment in the spleens of older MRL/I mice (735 vs 145 mg in agematched MRL/n mice). MRL/1 mice also develop clusters of perivascular infiltrates (plasma cells, lymphocytes, and histiocytes) prominent in the lung, salivary glands, kidneys, liver, and synovium. Both perivascular and peribronchial lymphocyte infiltrates are observed in the lung, whereas infiltrates appear as clusters in periarteriolar or pelvic regions in the kidney. Lymphoid infiltrates of salivary glands and of MRL/I synovia are discussed further in Sections II,B,6 and 7 . The massive lymphoproliferation, blurring of lymph node architecture, and extensive perivascular infiltration in many MRL/l organs raise the question of a malignant lymphoma, although the following observations would appear to contradict this possibility. (1) No dissemination beyond the perivascular area into the parenchyma of major organs has been observed. (2) There is no leukemic blood picture. (3) Attempts to transplant lymphoid masses from older MRL/l to very young MRL/l, MRL/n, or F, hybrids have been negative. (4) Owing to extensive hemorrhagic and cystic necrosis, the lymph node size, as stated above, decreases at a very advanced stage of disease, These findings make it unlikely that the MRL/1 lymphoproliferative process is malignant, but the possibility of it being prelyinphomatous has not been ruled out. MRL/n mice do not develop this type of lymphoproliferative syndrome. Older BXSB mice of both sexes also exhibit moderate lymphadenopathy and splenomegaly (50%incidence, lymph nodes about 10-20 times normal size, female spleens 4 times normal and male spleens 8 times normal size.) At autopsy, the enlarged peripheral and abdominal lymph nodes in 4- to 5month-old males retain their architecture, but are characterized by a predominantly lymphocytic proliferation with admixed plasma cells and histiocytes. The enlarged spleens show moderate white pulp hyperplasia and

282

AHGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

increased erythropoiesis. The lymphoproliferative process is seen much later in female BXSB mice (around 15-17 months of age) but is histologically similar to that of the males.

5 . Neoplasia There is considerable variation in the reported incidence of lymphoid neoplasms in NZB mice (21,41,44) (2-3% to a high of 50%). Some of these variations may result from the relatively small or selective nature of some of the series, but, as discussed by Holmes and Burnet (45), there is considerable uncertainty in distinguishing between profound lymphoid hyperplasia and genuine neoplasia. In our studies of NZB, NZBxW, BXSB, and MRL/1 mice, the incidence of lymphoid tumors was very low (
MURINE MODELS OF SLE

283

processes. Thus, by light and electron microscopy, the inflamined synovium of MRL/I mice has a very similar cellular composition to that of humans with rheumatoid arthritis (47-49). Since the MRL/I mouse represents the only animal with noninfectious, spontaneous arthritis accompanied by rheumatoid factor (RF) (see Section III,C,8), and closely resembles humans with rheumatoid arthritis (RA) in manifestations, i.e., vascular lesions, synovial inflammation, perisynovial and periarticular tissue mononuclear infiltrates, pannus formation, cartilagenous joint surface destruction, and joint fusion as a terminal event, this is an excellent experimental model for RA as well as for SLE.

7. Other Signijcant Histopathologic Characteristics Certain histologic characteristics of MRL/I mice have suggested that these mice might also serve as models of Sjogren’s syndrome. Our studies (46) have shown that approximately 60% of older MRL/l mice exhibit salivary gland periductal mononuclear infiltrates which sometimes extend beyond the periductal region and are associated with acinar destruction. However, identical but less frequent salivary gland infiltrates can be found in N Z B x W and MRL/n mice. Comparable infiltrates are also found in lung parenchyma and renal pelvis of these strains. Hoffman et al. (50) report conjunctivitis in 85% of MRL/l and 50% of MRL/n mice, and severe mononuclear infiltrates in lacrimal glands of 100% of MRL/I and 95% of MRL/n mice. Conjunctivitis and mononuclear infiltrates in lacrimal glands are also present in the other SLE strains. However, using Schirmer’s tests, only N Z B x W have dry eyes. The same authors (51) have reported the presence of band keratopathy with bilateral calcium corneal deposits in approximately 90% of MRL/l and MRL/n mice, and a posterior uveitis in 35% of adult MRL/l mice. Other SLE strains are free of this complication, but the incidence is 40% in normal DBA/2 mice. This band keratopathy of MRL mice is associated with significant hypercalcemia, striking elevation of parathyroid hormone binding by serum, and some parathyroid gland histologic abnormalities. Furthermore, others (52) have observed Ig and DNA deposits within the ciliary process and choroid in the eyes of a majority (60-90%) of older N Z B x W mice; electron-dense deposits are visible between the basement laminae and pigmented epithelium of the ciliary process. Studies by L. M. Hang et al. (unpublished observations) also identify arteritis in the sclerae of MRL/l mice associated with lymphoid infiltrates in the choroid, but not the ciliary process. These findings raise the possibility of primary autoimmune endocrine disorders as a component of murine SLE and suggest that lupus mice might be interesting models of spontaneous immunologically mediated ocular diseases. In NZB mice, a high incidence (up to 50%) of peptic ulcers has been

284

AHGYRIOS N . THEOFILOPOULOS A N D F R A N K J. DIXON

reported (53), occurring most coinmonly immediately distal to the pyloric ring and striking in histologic similarity to chronic peptic ulcers in man. The ulcers show no correlation with hemolytic or renal disease markers, and their etiology and pathogenesis are uncertain. Lambert and Oldstone (54) have found IgG and complement (C) deposits without associated histopathologic changes in the choroid plexuses of up to 80% of older NZBXW mice. Furthermore, Alexander et al. (55) report prominent mononuclear cell infiltrates restricted to the choroid plexus and meninges of approximately 50% of older MRL/1 mice, whereas MRL/n mice show lower incidence and delayed but widespread inflammatory infiltrates involving cerebral vessels and meninges but not the choroid plexus. These findings in mice raise the possibility that complications in the central nervous systems of humans with SLE may be caused by such immunohistologic changes. Accinni et al. (56) have demonstrated IC deposits on oocytes of lupus mice. According to this study, approximately 35% of NZBxW, 72% of MRL/I, and 20% of BXSB female autoimmune mice have granular deposits of murine IgG with less frequent C3 and occasional ssDNA and retroviral gp70 in the zona pellucida (ZP) of mature and atretic follicles. The ICs associated with ZP may be derived from the circulation or formed in situ. The significance of this IC-mediated oophoritis in ovulation and fertilization remains to be determined. Spontaneous endometrial deposition of IC in aging NZB X W mice has also been observed (57). In addition, subepithelial Ig deposits in apparently normal skin of NZBx W mice, similar in location and appearance to those seen in human SLE, have been reported (58,59). C. SEROLOGIC MANIFESTATIONS As with histologic abnormalities, some of the serologic characteristics (elevated serum Ig concentrations, antinuclear antibodies, anti-ds- and antissDNA antibodies, antiretroviral envelope gp70 antibodies, ICs and reduced complement levels) are common in all SLE strains. Other factors, however, such as anti-Sm, anti-y-globulins, and anti-red blood cell autoantibodies are expressed uniquely or at different levels in one or another of these strains, and are associated with related, unique clinicopathologic manifestations (Table 11). A detailed description of the main serologic abnormalities follows.

1 . Immunoglobulins Serologic analyses have indicated that the allotype of the IgG,, subclass a , whereas that of the BXSB strain is b, and for the NZB and NZW strains is e. However, Southern blotting experiments with MRL/l-derived monoclonal anti-DNA autoantibodies (60) and MRL/1 as well as MRL/n liver DNA (61, R. Kofler and A.

(Igh-l locus) for the MRL/l and MRL/n mouse is

285

MUHINE M O D E L S OF SLE

N . Theofilopoulos, unpublished observations) suggest that the IgH haplotype of MRL/l is o f t h e j type, derived either from the LG/J (more likely) or the C3H ancestors. The discrepancy between the serologic typing of MRL mice and the conclusions derived from the restriction enzyme polymorphisms might be explained by the lack of appropriate antisera to distinguish the closely related a a n d j allotypes at the time of the initial testing. In addition, with regard to the NZB mouse, recent findings (62,63) suggest that its haplotype should be designated n instead of e since its Igh-G(p), Igh-5(6), and Igh-4(yl) loci are of the d haplotype, whereas the downstream loci (Igh-3,y2b; Igh-l,yBa, I g h - 7 , ~Igh-2,a) ; are of the e haplotype. The levels of IgM and IgG subclasses in young and older diseased SLE strains compared to normal strains of mice are shown in Table I11 (ref. 15 and unpublished observations). All SLE strains have significantly higher total polyclonal IgG concentrations at preclinical or clinical stages than normal controls. The highest increase is observed in male and female MRL/I mice: a 4-fold increase over normal at 2 months of age, and a 10- to 13-fold increase at 4 to 5 months of age. In all SLE mice, there is also an approximately 3-fold increase in IgM concentrations compared to controls. Among the IgG subclasses, increases are predominantly IgG,, IgG,,, and IgG,I, in older MRL/l mice, IgG,, in older NZ mice, IgG,I, and IgG, in male BXSB mice. MRL/n mice devoid of the lpr gene do not show the profound hypergammaglobulinemia of the congenic lpr homozygous MRL/l mice, whereas BXSB females are within normal range. The most striking feature of the serum protein upon electrophoresis is the TABLE I11 SERUMIMMUNOGLOBULIN LEVELS(pdml) IN SLE MICE IgC I

IRG3

W Z I >

Strain

Y"

0"

Y

0

Y

0

Y

0

Y

BXSB d BXSB 0

200 170 600 310 280 250 NT 210 390

628 215 1600 405 750 610 NT 240 410

600 290 2700 480 340 720 270 200 500

950 520

1000 600 1800 80 200 130 490 580 800

1400 980 19000 1100 4600 6100 1800 980 1600

1000 450 1100 270 320 680 240 780 800

4200

890 25 400 20 70 60 43 30 50

MRWl MRUn

NZB NZB X W NZW C57BW6 BALBIc

2500 1600 2500 2400 720 600

~

1100 3000 269 1100 800 1100 1800

O

2400 210 250 300 1460 360 156 120 68

a The concentrations of IgM and IgG subclasses in sera (pdnil) were determined by isotype and subclass specific radioimmunoassays. b Y, Young mice of 1-2 months of age. 0, Old mice of 5 months of age for BXSB and MRL mice, and 7-12 months of age for NZ and normal strains of mice. Female mice from all strains were tested, except the BXSB mice, where both males and females were analyzed.

286

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

relatively high incidence of monoclonal y-globulins in MRL/I mice, particularly in their last month of life (5). Studies in our laboratory (unpublished observations) demonstrate that 53% of 5- to 6-month-old female and 29%of MRLll males have monoclonal Ig, with two bands detected in a few animals. This monoclonal Ig is of the IgG isotype with K light chains, and the majority are ofIgG, or IgG,, subclasses. No significant correlation is apparent between the presence of restricted IgG bands and levels of polyclonal IgG in serum, or levels of autoantibodies and cryoglobulins. However, the presence of such restricted IgG correlates somewhat with the degree of plasmacytosis, glomerulonephritis, and necrotizing vasculitis. A few of these monoclonals exhibit autoantibody activity against ssDNA or IgG (RF). Repeated efforts by nitrocellulose acetate electrophoresis froin seropositive mice failed to demonstrate such monoclonal y-globulins in kidney eluates. In contrast to MRL/I mice, congenic non-Zpr MRL/n and (MRL/lx MRL/n)F, hybrids not homozygous for the Zpr gene are devoid of such monoclonal y-globulins in their sera. These results strongly indicate that expression of monoclonal or restricted IgG (as well as exceedingly high levels of polyclonal IgG) is related to the lpr gene and associated T cell proliferation (see Section IV). A low incidence of monoclonal Ig has been observed by Andrews et aZ. (5)in the other lupus strains (BXSB, NZB, NZBXW). Only 23% of older BXSB male mice, 13% of NZB, and 17%of NZB x W females have such restricted bands, which are much less intense than those of the MRL/I mice. However, Sugai et aZ. (64) report incidences of restricted y-globulins as high as 30% in NZBX W mice over 11months ofage. In contrast to those in MRL/I mice, the NZB X W monoclonal y-globulins isolated by these investigators are all of the IgM isotype, and do not exhibit autoantibody activity. In NZBXW mice, the expression of these IgM monoclonals is often associated with a generalized lymphoproliferative disorder that, in some mice, shows features of malignancy.

2. Autoantibodies to Nuclear Antigens Antibodies against a variety of nuclear constituents are characteristic of SLE in humans and mice (65). Among SLE mice, ANA titers are highest in MRL/I mice, followed by NZBx W females, and BXSB males (5).In all strains tested, a peripheral or rim pattern of nuclear fluorescence is always present at the highest positive serum dilution, whereas the homogeneous pattern is sometimes seen at lower dilutions. Studies of humans and mice with SLE indicate that these ANAs are primarily directed against DNA, RNA-protein complexes (the ribonucleoproteins, RNP and Sm) or histones. Anti-DNA antibodies fall into four broad classifications (reviewed in 65). (1) Antibodies reactive with native or double-stranded DNA (dsDNA) which are

MURINE MODELS OF SLE

287

not cross-reactive with denatured or single-stranded DNA (ssDNA); this antibody appears to be rare. (2) Antibodies reactive only with ssDNA which do not cross-react with dsDNA. Such anti-ssDNA antibodies which appear to be directed against purine or pyrimidine bases do not react with dsDNA since the bases are buried within the double helix and are sterically inaccessible. (3) Antibodies reactive with determinants coininon to ds- and ssDNA which appear to recognize the sugar-phosphate backbone. Some of these antibodies also react with polynucleotides and phospholipids, including cardiolipin, which contains two phosphodiester linked side chains separated by three carbon atoms (66-68). (4)Antibodies reactive with a form of DNA whose bases have a different orientation to the backbone than that of right-handed DNA; this form of DNA has been termed “left-handed” or Z-DNA (69,70). Recent studies suggest that Z-DNA is formed in enhancer segments which turn on gene transcription (71). Anti-dsDNA antibodies develop relatively late in the course of murine SLE, with none of the affected strains showing significant levels at 2 months. At 4-5 months, all SLE strains have significant levels of anti-dsDNA, the highest being found in the 4- to 5-month-old MRL/I and the 9-month-old female NZBx W. Anti-ssDNA antibodies are found at low concentrations in immunologically normal mice. However, greater levels of these antibodies appear in NZB, NZBx W, and MRL/l mice at 2 months and even higher levels at 4-5 months, at which time BXSB males also have abnormal levels. Anti-ZDNA autoantibodies have been sought only in MRL/l mice; sera from 5month-old animals bind significant amounts ofZ-DNA as well as ds- or ssDNA (70). The number of anti-ssDNA secreting cells in spleens of lupus mice also increases with age, and is much higher than in age-matched normal controls. The number is highest in older female NZBXW and MRL/I mice, followed by NZB and BXSB male mice (72,73). In older MRL/l and NZ mice, approximately 1 of every 5-10 spontaneous Ig-secreting cells (IgSC) produces antissDNA antibodies (0.05-0.09% of the total spleen cell population). In older BXSB male mice, 1of 29 IgSC secretes anti-ssDNA antibodies (0.009% of the total spleen population). Although anti-DNA is at a very low level in serum and is secreted spontaneously by few spleen cells, several studies (74-78) have demonstrated high levels of such antoantibodies induced by normal lymphocytes stimulated in uiuo or in vitro with niitogens such as bacterial lipopolysaccharide (LPS). Moreover, limiting dilution experiments (79) with LPS-stimulated splenocytes of lupus and normal mice have shown similar numbers of precursor frequencies for IgM anti-DNA producing cells. Furthermore, experiments by Datta et al. (80) with rabbit antiidiotypic antibodies against monoclonal idiotypes of anti-DNA antibodies derived from MRL/l mice demon-

288

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

strate that such idiotypes are expressed not only in SLE strains of mice (MRL/l, MRL/n, NZB) but also in several normal strains (BALB/c, C57BL/6, AKR, SWR, CBA). The idiotypes in normal mice were barely present or undetectable in supernatants of nonstimulated spleen or fetal liver cells, as well as sera, but they became detectable and of equal levels to lupus mice upon stimulation of spleen or fetal liver cells with LPS. Conversely, antiidiotypic antibodies against monoclonal anti-dsDNA antibodies derived from normal (BALB/c) mice recognized idiotopes present in the serum of MRL/I mice (81). These findings suggest that (1) the existence of anti-DNA clones is not limited to autoimmune strains, (2) autoantibody idiotypes are related to a conserved family of antibody variable regions, and (3) autoantibody clones are expanded only in autoimmune mice. The antigenic structure and/or its mode of presentation involved in the induction of anti-DNA autoantibodies in SLE mice remain unclear at present. Immunization by nucleic acid antigens released from cell disintegration or derived from chemical or physical modifications of normal DNA has been proposed (82-86). Although naturally occurring nucleic acids are weak immunogens, some forms can serve as high-molecular-weight haptens when complexed to carrier proteins (86). Thus, autoantibodies to ssDNA-methylated bovine serum albumin (BSA) complexes are easily induced in most animals. However, repeated attempts by several investigators to induce antibodies to dsDNA in experimental animals have generally met with little success. The only exception is the study of Pancer et al. (87) who reported that low-molecular-weight DNA (approximately 150 bp) harvested from supernatants of cultured lymphocytes induces anti-dsDNA antibodies in normal and autoimmune mice, as does sheared calf thymus DNA. In this regard, it is of interest that recent studies (88) have shown that multiples of an approximately 200 bp DNA subunit are generated upon lysis of target cells by cytotoxic T lymphocytes, perhaps via activation of a specific endonuclease. Furthermore, Sano and Morimoto (89) have found that DNA isolated from DNA-anti-DNA complexes of human lupus sera contain two classes of DNA: one very small fragment with a molecular weight of 25,000 and 30 to 40 bp, and a larger fragment with a molecular weight of 100,000 and 150 bp; these small fragments have a content of guanine-cytosine (G-C) higher than that of the bulk of mammalian DNA. It was suggested that such unusually high G-C content can conformationally alter DNA structure, which may serve as the target for antibody production. However, deliberate immunization of experimental animals with such altered forms of DNA fails to develop anti-dsDNA antibodies (90). Recent studies by Schwartz, Stollar, and associates (90)raise doubts about the significance of DNA immunization in inducing the spontaneous antiDNA antibodies found in murine and human lupus. These investigators

M U R I N E MODELS OF SLE

289

propose that epitopes on nonnucleic acid structures, which may include bacterial cell wall phosphoglycolipids, can activate a large family of B cells, some of which form antibodies that can cross-react with the nucleic acids. This concept is based on two main findings. (1) Approximately 25% of murine or human hybridoma anti-DNA autoantibodies react broadly with a wide range of polynucleotides and phospholipids, including cardiolipin, all of which share a common characteristic, diester-linked phosphate groups in their backbones (66-68). (2) Antibodies induced in experimental animals after immunization with various nucleic acids are highly selective and react only with the immunogen (90). In contrast, spontaneous lupus autoantibodies and those induced in normal animals by adjuvants alone cross-react with multiple nucleic acid antigens. Further studies are needed to determine whether the serologic diversity of lupus autoantibodies is real or is, as proposed by the above studies, an epiphenomenon caused in part by an autoimmune response against an epitope such as phosphate residues present in a variety of biologic molecules. The in uivo and in uitro dependence of spontaneous and induced antiDNA antibodies on T cells and/or macrophages has been investigated. The responsiveness of autoimmune and normal mice to injected ssDNA was found by Fournie et al. (91) and Pancer et al. (92) to be T dependent. Although spontaneous anti-DNA antibody production by cultured splenocytes of lupus mice (NZBXW, BXSB, MRL/I) was reported to be macrophage and T cell dependent by Sawada and Tala1 (77), others (74,93)failed to observe such dependence in uitro. The lack of dependency by spontaneous anti-DNA production in uitro on T cells is not totally unanticipated, since B cells and plasma cells of lupus mice secreting such autoantibodies might have already received the necessary help in uiuo, and required no further assistance in uitro. The anti-DNA autoantibody repertoire of lupus mice has been examined by several investigators using the development of hybridoma autoantibodies and xenogeneic antiidiotypic antibodies against them. Studies by Andrzejewski et al. (94) and Rauch et al. (95) in MRL/I mice as well as Marion et al. (96), Hahn and Ebling (97), and Tron et al. (98) with NZBxW mice have indicated that although the anti-DNA (reactive with ssDNA only or both ssand dsDNA) responses are the products of a large number of different antibody producing clones, many of the clones produce antibodies with similar idiotypes not only within a given strain but also among genetically unrelated lupus (and normal) mice. Some of the in uitro or in uiuo produced Ig that share a given anti-DNA idiotype, however, do not display DNA binding activity (80,95), indicating the presence of the idiotypes on a parallel set of antibodies with different antigen-binding specificities. Whether the related clones producing anti-DNA autoantibodies are derived from a few clones by

290

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

extensive somatic germ line variable region gene mutation during clonal expansion, or from a large number of clones expressing related and conserved variable region genes is unknown. It should be noted, however, that studies by Tron et al. (99) with monoclonal antibodies specific for dsDNA derived from NZBXW mice have challenged this view because of their failure to demonstrate cross-reactive idiotypes with such monoclonal autoantibodies or to find a major representation of any of the idiotypes examined in the sera of these animals. The isotypic characteristics of anti-DNA autoantibodies have been investigated in NZBXW mice by Tala1 and associates (100,101), who find that antibodies to DNA detected early (1 month of age) are of IgG and IgM isotypes. Antibodies to DNA increase slowly as animals mature from 1 to 4 months of age, and the Ig isotype distribution remains relatively unchanged until 5 to 6 months, when, in females, there is a sudden marked increase in DNA binding by IgG antibodies. The same sequence of events is repeated in males, but approximately 4 months later. Similar changes have been reported by Steward and Hay (102) in which age-related switching of IgM antidsDNA antibodies to IgG,, and IgG,,, subclasses evolves. Thus, this switch from IgM to IgG appears to herald the onset of severe disease, which occurs first in females and later in males. IgG, subclass antibodies may be more pathogenic than other subclasses. For example, IgG,, has the longest halflife among IgG subclasses, consumes complement (C) with the greatest efficiency, displays the lowest C-dependent solubilization when complexed with antigen (103), activates mouse mast cells (104), and induces stronger immune responses and memory when cornplexed with antigen than does antigen alone (103,105). Such time course studies have not been performed in other lupus strains, but it is clear that serum isotypes and autoantibodies in older MRL/l and BXSB mice at the preclinical or clinical stage are predominantly IgG (25,106). The importance of autoantibody affinity, such as that of anti-DNA, in the development of glomerular disease has also been studied in both humans and mice with SLE. Steward et al. (107) has presented evidence supporting the view that overt renal disease in NZ mice is associated with low avidity anti-DNA antibodies in serum. Presumably such low avidity antibodies favor the persistance in the circulation of antigen excess ICs, which then deposit in the tissues. These findings may be misleading, since antibody of whatever affinity or avidity in serum may not be representative of the antigen-complexed antibody deposited in the tissues. In fact, when Winfield et al. (108) evaluated the anti-dsDNA antibodies in SLE patients with and without active nephritis, sera from patients with active nephritis contained antibodies of lower affinity. However, anti-dsDNA antibody of uniformly high

M U R I N E MODELS OF SLE

29 1

affinity was found by these investigators in the corresponding glomerular eluates of SLE patients with nephritis. There is substantial evidence that renal deposits of DNA and its reactive antibody in lupus inice play an important role in the pathogenesis of their renal disease. First, ANA and anti-DNA autoantibodies are found concentrated in their kidney eluates. Second, induction of tolerance to ssDNA has been reported to prevent expression of murine lupus nephritis (109,110). Third, genetic studies in F, generation lupus mice (111) and in appropriate F, crosses and backcrosses (112) have shown, with a few exceptions, a direct relationship between anti-DNA antibody levels, glomerulonephritis, and mortality. Fourth, DNA-anti-DNA complexes, albeit not in large quantities (113-116), have been found in human and murine lupus sera. In addition to anti-DNA autoantibodies common to all lupus strains, other types of ANA have been observed only in particular lupus mice, e.g., antiSin autoantibodies were detected only in MRL mice (117). Others have not been studied comprehensively and their preseiice or not in all lupus strains remains unknown, e.g., antihistone autoantibodies have been sought and found thus far only in sera of NZBxW mice (118). The presence of anti-Sm-autoantibodies in lupus inice has been surveyed with a double iinmunodiffusion assay performed by Eisenberg et al. (117), who find that only MRL/I and MRL/n contain such autoantibodies, whereas NZB, NZBXW, and BXSB mice are uniformly negative. Anti-Sm autoantibodies appear in MRL/l inice after the onset of disease at around 4-5 months of age, with 37% of the males and 10% of the females positive. In MRL/n mice, anti-Sm autoantibodies appear at approximately 4-5 months with a frequency similar to that in the MRL/l, progressing thereafter to 83% in females and 57% in males 9-12 months old. A monoclonal anti-Sm antibody derived from MRL/l mice by Lerner et al. (119) precipitates five sinall nuclear ribonucleoproteins (snRNP) containing U1, U2, U4, U5, and U6 RNAs, identical to those precipitated by human anti-Sin autoantibodies. Evidence has been presented that siiRNP particles containing U1 RNA are involved in processing transcribed RNA (120,121). Time course studies by Pisetsky et al. (122) demonstrate an independent expression of anti-Sm and anti-DNA antibodies in individual MHL/l and MRL/n mice. Moreover, although anti-Sm levels are similar in both types of mice, anti-DNA is much more prevalent in MRL/l than MRL/n mice. These findings indicated that different iminunoregulatory disturbances may be involved in anti-DNA and anti-Sm expression, and that the Ipr gene, which affects anti-DNA and total Ig levels, has minimal, if any, effect on anti-Sin antibody levels. Pisetsky et al. (122) also find, despite anti-DNA levels of comparable value among individual age-matched MRL mice, large variability in anti-Sin responses among

292

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

positive animals. This variability may be related to the paucity or absence of anti-Sin precursor cells in contrast to a much larger number of anti-DNA precursors. Without genetically determined specificities, somatic mutation would be required for the generation of a repertoire of anti-Sm autoantibodies. Such random mutational events may occur at various times throughout the lives of individual animals. Further studies by Cohen and Eisenberg (123) suggest that only splenocytes from MRL mice with circulating anti-Sm can generate anti-Sm-secreting cells in uitro, whereas cells from seronegative animals, even in the presence of LPS, fail to do so. These investigators observe that generation of anti-Sin by MRL cells in vitro depends on the presence of T cells, but the ability of cells from individual MRL mice to generate anti-Sm is limited by the availability of Sm-specific B cell precursors and not by the relative absence of T cells capable of providing specific help for the anti-Sin response. Additional studies by these authors (124) on Sm-primed MRL/n mice demonstrate the presence of T (Thy-l+ , Lyt-1 +) cells specifically recognizing Sm antigens, but the presence or absence of serum anti-Sm in individual mice had no bearing on their capacity to manifest in uitro T cell reactivity to Sm after in viuo priming. The importance of T cells in regulating anti-Sm responses is further suggested by Eisenberg et al. (106)who finds that serum and cell-secreted anti-Sm antibodies in MRL mice are almost exclusively of the IgG,, subclass, which is highly T dependent. The clinical significance of the anti-Sm autoantibody is questionable since the disease is equally severe in anti-Sm-positive and antiSm-negative mice. Moreover, the anti-Sin autoantibody appears in MRL/l mice very late in the disease process. Antihistone antibodies have been demonstrated in the sera of NZB x W mice by Gioud et al. (118).Serum antihistone antibody levels are very low in animals younger than 4 months of age, progressively increase from 4 to 8 months of age (100% positive by 8 months), and show an age-related maturation from IgM to IgG. The predominant antibody activity in older mice involved the individual histones, H2B and H3, a histone reactivity pattern somewhat different than that reported for human SLE sera (125). Splenocytes from NZBx W mice also spontaneously produce in vitro antihistone antibodies, which switched from predominantly IgM in young mice to IgG in older mice. Such autoantibodies were undetectable or of very low value in splenocyte cultures or sera of normal mice. Although no serologic studies have been performed on the other lupus strains, monoclonal antihistone antibodies have been obtained not only from NZBXW but also from MRL/n splenocytes following fusion with a plasmacytoma (Rubin, Balderas, and Theofilopoulos, unpublished observation). As discussed below, some anti-yglobulin monocloiial antibodies cross-reactive with histones have also been obtained from MRL/I mice (126,127). Although antihistone antibodies in

MURINE MODELS OF SLE

293

human SLE appear to correlate with disease activity and responses to therapy (128,129), their pathologic significance in SLE remains unclear. Autoantibodies to Z-DNA have been reported in sera of older MRL/l mice (70). Levels of such autoantibodies correlated with levels of anti-ss- or antidsDNA antibody titers. Such autoantibodies against left-handed Z-DNA have also been described in humans with SLE, and correlate with clinical manifestations (130). Finally, studies by Eilat et al. (131-133) demonstrate the presence of antiRNA antibodies in sera and supernates of cultured spleen cells from NZB x W mice. Furthermore, monoclonal anti-RNA antibodies have been derived from NZBxW (134) and MRL/l mice (119). The antigenic determinant recognized by the NZBx W sera and their monoclonal anti-RNA antibody is largely composed of a trinucleotide sequence of single-stranded RNA containing G, C, and U residues (134,135). The monoclonal anti-RNA derived from MRL/I mice recognizes an antigenic determinant on the large RNA moiety of the ribosomes rather than on a ribosomal protein (119).

3. Antiretroviral gp70 Autoantibodies Relatively large amounts of the glycoprotein gp70, so designated because of its 70,000 molecular weight as well as its structural and immunologic similarity to retroviral envelope protein, circulate in the blood of virtually all strains of mice free from any viral particles (136-138). Although multiple immunologically and structurally related gp70s are found in all mice, the serum gp70 molecule is similar in all mice and closely resembles (but by peptide fingerprint analysis is not identical with) the xenotropic virus coat protein isolated from NZB and NZB x W hybrid mice (138,139). Lupus mice have significant serum gp70 levels, but similar levels are also apparent in several normal strains (5,140). The origin of serum gp70 and mechanisms controlling its expression have recently been investigated. Although lymphoid cells express surface, virionfree gp70 (141,142), they do not seem to be a major serum gp70 source, since these levels are not affected by thymectomy or splenectomy (143). Recent studies (144) demonstrate that one injection of bacterial lipopolysaccharide (LPS) greatly enhances serum gp70 expression, but is not accompanied by increases in endogenous xenotropic virus or other viral constituents. Further studies (145) indicate that increased serum gp70 expression is unrelated to the activation and/or presence of lymphocytes and that the site of serum gp70 synthesis is the hepatic parenchymal cells. Enhancement of serum gp70 by LPS also correlates with induction of other liver synthesized proteins, such as haptoglobin, collectively defined as acute phase reactants (APR), and the expression of gp70 in serum is enhanced not only by LPS, but also by other inducers of APR. As a whole, these findings suggest that serum

294

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

gp70 is an APR and behaves as a normal host constituent and not a product of the viral genome. Serum gp70 production has been found to be controlled by multiple genes, some of which are recessive. Recent studies suggest that three genes, tentatively designated Sqp-la, It), and 1" (146) are involved in gp70 production. This system appears to be closely linked to, but not in, the H-2 region on chromosome 17. S q p - l a correlates with relatively large amounts of serum gp70, Sqp-Ib corresponds with low serum gp70 and increased gp70 production after LPS, while Syp-lc relates to small amounts of serum gp70 that are unresponsive to LPS. Although high levels of serum gp70 can be found in both lupus and nonlupus mice, only lupus strains spontaneously produce antibodies against this molecule and contain gp70-anti-gp70 complexes in serum (147). Th'is complexed gp70 varies in molecular size from 9 S to 19 S on sucrose density gradient untracentrifugation. The antibody bound gp70 increases not only in incidence but also in amounts with disease progression in all the SLE strains, though total serum gp70 levels do not usually change greatly after the animals reach 1-2 months of age. The appearance of antibody-complexed gp70 varies with age among the SLE strains, but parallels the onset of renal disease and persists throughout the disease course. The association of antibody to serum gp70 with the renal disease of lupus mice has also been demonstrated by genetic studies (111,112,148). Moreover, treatment of some lupus strains with prostaglandin E, (149) or low-calorie diet (150) results in prolonged survival and retarded gloinerulonephritis, both of which correspond well with a specific reduction of circulating gp7O-anti-gp7O complexes. Furthermore, as indicated above (Section III,B), a variety of studies have shown gp70 deposits together with immunoglobulins and C in mesangial and capillary sites of diseases glomeruli and in vascular lesions of lupus mice (5,24). Because multiple immunologically related gp70s are expressed in every mouse, and antibodies to one or more of them may be made, Izui et al. (151) have analyzed the nature of gp70 and anti-gp7O antibodies in the circulating immune complexes of SLE-prone mice. They find that antibody complexed gp70 is identical to the NZB xenotropic type gp70 free in serum, and differs from that of several other retroviruses. Furthermore, in exhaustive specificity tests, anti-gp70 antibodies isolated from circulating immune complexes preferentially bind NZB-xenotropic gp70 over other types of retroviral gp70 [BALB/c, AKR, NIH Swiss xenotrope; AKR ecotrope; AKR recombinant; Rauscher inurine leukemia virus (MuLV), ampliotrope]. These results indicate that the high serum levels of xenotropic gp70 in themselves apparently do not cause murine nephritis, but the critical factor is the unique ability of SLE mice to respond to this autoantigen.

MURINE MODELS OF SLE

295

Stndies by Izui et al. (152) with NZBXW F, mice suggested that gp70anti-gp70 IC is the most important contributor to the pathogenesis of fatal glomerulonephritis. However, studies by Dixon et al. (111) on F, generations of crosses between lupus mice, including the NZB x W, show a significant correlation between anti-DNA and anti-gp7O responses and fatal glomerulonephritis. Furthermore, preliminary studies by Dixon and associates (unpublished data) on sublines of MRL/l mice characterized by low gp70 and related IC levels establish that the autoimmune disease, including glomerulonephritis, is essentially unchanged. Thus, although gp70-antigp70 deposits are involved in the pathogenesis of lupus nephritis, they may not be the major participant. It is more accurate to say that levels of gp70anti-gp7O and DNA-anti-DNA complexes show the best correlation with G N and mortality in lupus mice.

4 . Complement Levels In all types of mice with lupus-like syndromes, the concentrations of hemolytic complement fall with disease onset or progression, manifested by the presence of autoantibodies and related ICs (5).The NZB mouse is C5 deficient. However, studies of the C5-deficient F, generation mice resulting from crosses between NZB and BXSB reveal no relationship between the presence or absence of C 5 and disease (F. J. Dixon and P. J. McConahey, unpublished observations).

5 . C ryoprecipitates Increased levels of cryoglobulins compared to normal sera are found in all lupus strains and correlate with age and disease severity. The most striking concentration of cryoglobulins is found in MRL/I mice of both sexes, and values increase from 170 pg/ml at 2 months to >2000 pg/ml at 5 months (5). 6. Antierythrocyte Antibodies

The most consistent abnormality in NZB mice is the development of an autoimmune type of Coomb’s positive hemolytic anemia (hematocrit <40) associated with the appearance of erythrocyte autoantibodies (21,153). The incidence of this autoantibody in both sexes is 5-10% at 3 months, 25-35% at 6 months, 60-80% at 9 months, and > 90% by 15 months of age. This hemolytic anemia has several characteristics in common with its human counterpart, e.g., marked reticulocytosis (>25%), spherocytosis, increased osmotic fragility of erythrocytes, decreased erythrocyte survival in v i m , and splenomegal y. Antierythrocyte autoantibodies of NZB mice are directed predominantly to two erythrocyte autoantigens (153,154). Anti-X autoantibody, unique in that it is spontaneously produced only by the NZB strain, reacts with an

296

AHGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

exposed erythrocyte surface autoantigen (X) and mediates the autoimmune hemolytic anemia. Anti-HB autoantibody, by contrast, is ubiquitous in aging mice of all strains and directed to cryptic autoantigen (HB) exposed by limited proteolysis of mouse erythrocytes with bromelin. The anti-X autoantibody is polyclonal and represented within all IgG subclasses, whereas the anti-HB is predominantly of the IgM isotype (155). NZB-derived IgM monoclonal anti-HB autoantibodies have been developed (156,157). One of these monoclonal antibodies cross-reacts with trimethylammonium-containing molecules such as choline and phosphorylcholine (158). These compounds may, therefore, represent the erythrocyte determinant that is uncovered by the proteolytic activity of bromelin or at least constitute part of this determinant. Klotz et al. (159) also reported IgG monoclonal antierythrocyte antibodies that react with an antigen which is exposed on the erythrocyte surface. Holborow et d.(160) have described a third type of predominantly IgG autoantibody in NZB mice detectable by indirect immunofluorescence and directed to antigens (HOL) present on both murine and human erythrocytes; the HOL antigen on intact erythrocytes is inaccessible to autoantibody but is readily exposed in dried red cell smears and on erythrocyte stromata. It has been suggested that this antigen may be related to spectrin, a component of the internal aspect of the erythrocyte membrane (161). Finally, a fourth type of cold (4°C) reactive IgM autoantibody directed against the cryptic autoantigen I has been described (162). Among these antierythrocyte autoantibodies, only the anti-X is restricted to NZB mice; the others are found in several murine strains. The restrictive anti-X antibody expression in NZB mice is not attributed to the exclusive presence of X-antigen binding B lymphocytes, which are found in all strains examined, but to the presence of X-binding helper T cells unique in the NZB strain (163). In contrast to X antigens, B and T cells reactive with the ubiquitous HB antigen have been found in all strains examined. These data suggest that circumvention of immunologic tolerance to the X-antigen in NZB mice may occur at the T lymphocyte level and/or that T cell factors are necessary for induction of this IgG type autoimmune response. In addition to the erythrocyte-bound X and HB autoantigens, soluble analogues, termed SEA-X and SEA-HB, respectively, exist in murine plasma (164,165). SEA-X and SEA-HB have been found in the plasma of all murine strains, regardless of their H-2 background. In contrast to the high incidence of antierythrocyte autoantibodies in NZB and female NZBX W mice, male and female BXSB and NZBXW male mice have a moderately low incidence (20-40%) and MRL/l mice of both sexes a very low (10%)incidence (5). These low incidences are comparable to those of immunologically normal mice. It is not known, however, whether the

MUKINE MODELS OF SLE

297

erythrocyte autoantibodies of lupus stains other than NZB recognize the Xantigen or not.

7. Natural Thymocytotoxic Antibody (NTA) Shirai and Mellors (166-168) were the first to describe the occurrence of a thyinocytotoxic autoantibody in NZB sera. NTA is in the IgM class of immunoglobulins and shows optiinal reactivity with mouse thyrnocytes at 4"C, but is also slightly reactive at 37°C. This autoantibody appears early (2-3 weeks of age) in NZB mice, but is not present in newborns. Approximately 50-60% of such animals are NTA-positive at 1 month of age, an incidence that increases to nearly 100% by 3 months. In the same studies, it was observed that 20% of NZBxW mice are positive for NTA at 2 months of age, 50% at about 3-6 months of age, and almost 100% after 12 months of age. NTAs of NZ mice have been found cytotoxic for syngeneic thyinocytes as well as thymocytes of all murine strains tested. Absorption tests indicate that the reactive antigen is distributed only in the thymus, spleen, lymph nodes, and brain of adult mice, and only in the thymus of newborn mice. This distribution suggests that the antigen may be similar to the T cell Thy 1 antigen. However, since NTA reacts with, and is absorbed by, thymocytes of all strains examined, regardless of their Thy 1 allelle (Thy 1.1and Thy 1.2), the conclusion is that the antigen differs from the Thy 1 alloantigen system. More recent studies by Imai et al. (169) have indicated that NTAs of NZB mice consist of two types of autoantibodies in terms of their target specificities. NTA-2 is strongly cytotoxic only against desialized lymphocytes, whereas NTA-1 is cytotoxic against both intact thymocytes and asialolymphocytes. N-Acetyllactosamine and a so-called complex-type oligosaccharide prepared from porcine thyroglobulin unit B glycopeptide were potent NTA-2 inhibitors, whereas NTA-1 was only slightly inhibited by the various simple sugars tested. This result suggested that NTA-2 preferentially recognizes terminal P-galactosyl residues of sugar chains that are masked by sialic acid residues on the cell surface. The biologic and pathologic significance of NTA is unknown. Initially, it was suggested that NTA might be involved in the thymic atrophy and immune dysregulation seen in old SLE mice (166). In fact, certain studies report that NTAs, particularly NTA-2, have a preferential reactivity against concanavalin A (Con A)- induced or antigen-specific suppressor T cells, thus implying a primary role in autoimmune induction (170,171). This implication, however, has been questioned since (1) the two new lupus strains (BXSB and MRL/l), when examined by Eisenberg et al. (1972)were found to have very low NTA levels despite their acute severe SLE; (2) detailed studies of F, lupus-prone mice by Dixon et al. (111) showed no correlation

298

ARGYRIOS N. THEOFILOPOULOS A N D FRANK J. DIXON

between NTA levels and lupus expression; (3) studies of congenitally asplenic (Dh/+) NZB mice indicated that such mice do develop antierythrocyte autoantibodies and ANA despite a delayed onset and very low NTA titers (173); (4) several authors (172,174-177) reported NTA in many immunologically normal strains of mice, although the fine specificities of NTA in such mice have not yet been compared to those in NZ mice; (5) studies of NZBXC58 recombinant mice and NZBxW backcross mice showed independent segregation of NTA and other autoimmune traits (178180); and (6) Con A-activated spleen cells from in vivo NTA-treated and untreated normal mice equally suppressed anti-SRBC antibody production in vitro, suggesting that NTA treatment does not affect the direct precursors of suppressor T cells (181). The above findings would argue against a primary role of NTAs in systemic autoimmune disease expression, but their possible secondary role together with other autoantibodies in the disease of NZ mice still needs to be considered.

8. Autoantibodies to IgG (Rheumatoid Factors) Unique among lupus strains of mice, the MRL/l substrain develops a polyarthritis characterized by pannus formation, cartilage and bone erosions, and histology similar to human RA (46). In accordance with histologic findings, these mice develop IgM and IgG RF at approximately 3 to 4 months of age (5,46). The serum polyclonal IgMRFs react with all murine IgG subclasses as shown by solid phase radioimmunoassays, but the degree of reactivity varies depending on the substrate subclass (126). The majority of the sera react most strongly with mouse IgG,,. Nevertheless, inhibition experiments indicate that MRL/l polyclonal IgMRFs react mostly with distinct epitopes in each IgG subclass. Thus, significant inhibition of binding to a given IgG subclass substrate is achieved when the competing aggregated IgG subclass added to the serum is the same as that used to coat microtiter wells. Most of the polyclonal IgMRFs in MRL/1 mice, therefore, consist of distinct non-cross-reactive antibody populations, each relatively specific for one IgG subclass. With regard to heterologous IgG, extensive cross-reactions with human, cow, rat, rabbit, and goat IgG are noted (126). A reactivity by double immunodiffusion was shown to occur among sera of MRL/1 mice (182). On the basis of this unusual reactivity among themselves, sera of older MRL/l mice have been divided into two groups: Group A sera form a precipitation line in gels with at least one member from Group B. Group B sera, in addition, interacts with at least one member of its own group. Intermediate IgG complexes (sedimenting between 6.6 S and 19 S and presumably containing IgG RF) were found to be the reactants causing these precipitation interactions. No explanation for these precipitation interactions between the Group A and B sera has been given, but self-associat-

M U R I N E MODELS O F SLE

299

ing IgG RFs have been observed in such mice (183). Moreover, IgG RF complexes with anti-DNA antibodies, particularly those of the IgG,, subclass, have been found in the sera of MRL/l mice (184). Monoclonal IgMRFs derived after fusion of MRL/I spleen cells with a nonsecretor plasmacytoma were described by Theofilopoulos et al. (126). These monoclonal IgMRFs only bound murine IgG and not other murine isotypes. Eight inurine IgG subclass-specific clonotypes were identified, the majority of which reacted with multiple IgG subclasses or IgG,, alone. A few clones reacted solely with IgC,,,, IgG,, or IgG,. The preferential reactivity of serum polyclonal and monoclonal RF with IgG,, might be explained by a larger IgG,, autoantibody representation in the ICs abundantly present in sera of such mice (5,25,106). Recent experiments strongly suggest that RFs represent a secondary manifestation of IC formation during immune responses (185,186). Monoclonal IgMRF with exclusively anti-IgG,, activity exhibit allotypic specificity which, with few exceptions, reacts with a, c, and e, but not b, d, or j IgG,, allotypes. Four clonotypes can be distinguished by cross-reactivity with IgG from species other than mice; inonoclonals possessing activity against several inurine subclasses cross-react extensively with heterologous IgG, including all human IgG subclasses without allotypic restrictions, whereas monoclonal IgMRFs specific for murine IgG,, or IgGZb do not cross-react with heterologous IgG. All of the monoclonal RFs bound to Fc but not to the Fa11 portion of mouse IgG, conclusively demonstrating that these monoclonals had the sine qua non qualities of RF. The monoclonal RFs are not directed against the Clq binding site at the CH2 domains or the protein-A hinding site at the CH2-CH3 junction. Using recombinant and mutant murine IgG subclasses, we have determined that the IgG,, or IgG,,, specific RFs of MRL/l origin react with an epitope located at the CH3 domain of the respective iininunoglobulins (A. N. Theofilopoulos, unpublished observation). Furthermore, the monoclonal RFs are devoid of anti-dsDNA, anticollagen, or antistreptococcal peptidoglycan cross-reactivity, but some monoclonals cross-react with histones and/or both histones and ssDNA. Cross-reactivity with histones and/or ssDNA is mediated by the same binding site since interactions with these substrates (IgG, DNA, histones) is reciprocally inhibited by each of the antigens (127). Moreover, a few RFs react with cytoskeletal elements on Hep-2 cells (127). It is of interest that cross-reactions of human and mouse autoantibodies with dissimilar molecules do not appear to be isolated events. Thus, some human polyclonal and monoclonal IgMRF have been found to display reactivity with histones and/or DNA (187-189). Cross-reactions of mouse monoclonal anti-DNA antibodies with cytoskeletal elements such as vimentin, actin, and spectrin (190,191)and monoclonal multiple organ-reactive autoantibodies (192,193) have also been described. The nature of these cross-reactions has not been

300

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

defined, but might be explained by (1) sharing of a structural epitope, (2) sharing of a conformational epitope, or (3)accommodation of more than one epitope by an antibody combining site. These findings raise important questions regarding the extent of antibody diversity, repertoire of autoantibodies, and the means by which systemic autoiininune diseases are caused. The clinical significance of RFs in the pathogenesis of RA remains controversial. Studies in MRL/l mice indicate a close correlation between presence of circulating IgMRFs and demonstrable synovial and/or joint pathology, i.e., 95% of mice with significant levels of IgMRF have synovitis and/or arthritis (46) and the congenic MRL/n mice lacking the Zpr gene and serum RFs do not develop arthritis. Conversely, 80% of mice with significant arthritis have elevated levels of IgMRF. However, the possibility that RFs themselves are responsible or sufficient for the arthritic process seems unlikely in view of the following observations. Two normal mouse strains (C3H, C57BL/6) in which the Zpr gene has been introduced (C3H-Zpr/lpr and C57BL/6-lpr/zpr) have serum IgMRFs in levels that frequently exceed those in MRL/l mice, yet no arthritis is observed (194,195). However, polyclonal and monoclonal IgMRFs of C57BL/6-Zpr/lpr appear to differ from those of the MRL/I mice in that they show preferential reactivity with IgG, instead of IgG,a (196). This difference in subclass specificity of RFs in the two IprlZpr substrains may have relevance to the arthritis expression, or lack thereof. For example, RFs of MRL/l mice may have better complementarity with the predominantly IgG,, autoantibodies (anti-DNA, anti-Sm, etc.) found in ZprlZpr substrains, thereby allowing formation of pathogenic complexes. The need for other autoantibodies that are complementary to R F specificity for RA-like manifestations to appear is further suggested by the fact that (1) several normal mice injected with LPS do develop IgMRFs with IgG, specificity, but not arthritis (194), and (2)the 129/Sv mice, which spontaneously develop IgM and IgA anti-IgG,, R F but not other types of autoantibodies, are free from arthritis (197, and personal observations). Perhaps IgG subclass specificities of RFs as well as background genes and abnormalities (i.e., other types of autoantibodies, ICs, and/or sensitized T cells or unique joint susceptibility) may all be required for the final expression of arthritis. The influence of genetic background on histologic and serologic effects of autoimmune accelerators, such as the Zpr gene, will be discussed further in Section

v,c.

A recent study indicated the presence of antibodies to native or denatured type I and I1 collagen in sera of MRL/I, but not other lupus strains (198). These antibody levels increased with age, and there was a significant proliferative response of young MRL/l T cells to type I, but not to type 11, collagen. The primary or secondary contributions, if any, of anticollagen autoantibodies in the arthritic process of MRL/I mice remain unknown.

M U R I N E MODELS OF SLE

301

IV. Cellular Abnormalities

Because the major characteristic of lupus is autoantibody production, a variety of studies have been performed to determine the exact nature of the defect. Below, lymphoid cell transplantation studies among lupus and normal strains of mice, and numerical as well as functional analyses of lymphoid cells in such mice are presented.

A. TRANSFER OF AUTOIMMUNE DISEASE

The best means of demonstrating that lupus is caused by lymphoid system abnormalities is by transplantation experiments wherein lymphoid cells are exchanged between early and late-life disease lupus mice, and between SLE and H-2 compatible normal mice. Transfer of spleen cells from NZB mice with established disease was found to be capable of inducing antierythrocyte antibodies in young syngeneic and H-2 compatible allogeneic normal recipients (199-203). However, the first transplantation experiment that clearly established induction of lupus by lymphoid cells was that of Morton and Siege1 (204,205), who lethally irradiated NZB and H-2 compatible normal mice and reciprocally transferred fetal liver, spleen, or bone marrow cells. They then followed the recipients for expression of autoimmunity, and observed that transfer of NZB cells into normal recipients induced autoantibodies and glomerulonephritis, whereas transplantation of lethally irradiated NZB mice with normal cells was not followed by autoantibody production in the recipients. On the basis of these experiments, the authors assigned a primary role to the NZB hemopoietic stem or precursor lymphoid cell population in the etiology of this autoimmune disease. Similar experiments were subsequently performed by Akizuki et uZ. (206), who observed that NZBXW bone marrow cells treated with anti-Thy 1.2 serum complement to eliminate differentiated T cells were capable of transferring disease to H-2 incompatible nonautoimmune lethally irradiated recipients. Similarly, in more recent experiments, Jyonouchi et al. (207) observed that the increased capacity of NZB bone marrow cells to form B cell colonies in vitro could be transferred into lethally irradiated normal H-2 compatible recipients and, conversely, lethally irradiated NZB mice given normal bone marrow had all of the characteristics of the normal donor. These experiments were then extended by Eisenberg et aZ. (208) and Theofilopoulos and Dixon (19,209) to the new lupus strains, BXSB and MRL, respectively. The availability of BXSB mice, which show substantial sexlinked differences in mortality and disease expression, and of MRL mice with similar differences between the congenic MRL/l and MRL/n substrains makes possible the perforinance of histocompatible cell transfer experiments. Thus Eisenberg et al. (208), by transferring male and female BXSB

+

302

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

bone marrow cells from 1-month-old mice into male and female lethally irradiated recipients of the same age in all four combinations, shows that the pace of disease expression is indeed a property of the transplanted lymphoid cells. As shown in Table IV, only male bone marrow induces early-life disease (GN and death) in both male and female recipients; female marrow induced late disease in males and females. Similar results have been obtained by transferring spleen cells from 4-month-old donors to 2-month-old lethally irradiated recipients. Once more, the sex of the spleen cell donor rather than the sex of the recipient has determined the pace of disease development, and the recipients express serologic markers similar to those of the donors, i.e., recipients of male bone marrow and spleen cells form higher concentrations of polyclonal IgG, anti-DNA antibodies, and gp70anti-gp70 complexes than recipients of female bone marrow or spleen cells. Also observed is that spleen cell transfer from old male mice with clear-cut disease does not produce disease in recipients any faster than transfer of bone marrow from premorbid mice. Therefore, the active cells in these transfers seem to be stem cells-not differentiated autoantibody secreting B cells-and the development of BXSB disease does not appear to result from an accumulation of defects at the stem cell level, which appears equally abnormal throughout the animals' lives. Similar bone marrow and spleen cell transfer experiments were performed by Theofilopoulos and Dixon (19,209) between the MRL/l and MRL/n mice. In each instance, 1-month-old lethally irradiated MRL/n mice receiving bone marrow or spleen cells derived from 1-month-old MKL/l mice as well as those receiving MRL/1 fetal liver cells from 17-day-old fetuses develop a severe wasting syndrome at 3 to 5 months posttransplantaTABLE IV BXSB BONEMARROW CIIIMEWAS

50% survival (months)

Group11

8')

M+F M-M

8 > 18 > 18

F-, F F+ M ~

~~

Iks

Anti-ssDNA % binding

(mg/mI)

2023 18+3 7.6k0.3 6.8t0.6

8.220.9 9.320.9 2.820.2 6.220.6

Ig bound gP70

GN

4624.3 35.1e5.3

4+ 4f

9.0k2.6

If

6.721.8

-

I+

,I M + F, female recipients of male cells; M + M, male recipients of male cells; F F, female recipients of female cells; F M, male recipients of female cells. Serologic assays and degree of GN were assessed 4 and 6 months posttransfer, respectively. Recipients were 1.5 months of age at the time of transfer and donor cells were obtained from 1-month-oldanimals. The actual 50% survival at approximately 8 months in M + F and M + M corresponds to 6.5 months posttransfer and compares to 5 . 1 month siirvival for unmanipulated male BXSB.

M U R I N E MODELS OF SLE

303

tion accompanied by a severe depletion of all lymphoid cells. Pretreatment of the MRL/l bone marrow inoculum with anti-Thy 1.2, anti-Thy 1.2 complement or certain agglutinins to remove differentiated T and B cells is unsuccessful in inhibiting this rejection phenomenon, despite the fact that such treatments have been reported by others to eliminate graft-vs-host (GvH) reaction among mice differing at the major or minor histocompatibility antigens (210-212). The basis of this phenomenon among mice that are near congenic (99.9% of identical genome) and differ only in that segment of the genome containing the Zpr gene is unclear at this time. It is of interest that, in contrast to the MRL/l + MRL/n chimeras exhibiting the rejection phenomenon, the MRL/n MRL/l chimeras show significantly prolonged survival (>10 months) as well as suppression of lymphadenopathy and splenomegaly, indicating irrelevance of intrinsic nonlymphoid factors to disease expression. Also, serum polyclonal IgG and autoantibody levels are significantly lower in MRL/n + MRL/I than MRL/l += MRLfl recipients: the latter die within 5-6 months of transfer. The above experiments, as a whole, clearly demonstrate that (1) inurine lupus hematopoietic stem cells or lymphoid precursor cells are inherently abnormal and express all that is necessary for the expression of autoimmunity, and (2) autoantigens per se or other nonlymphoid cell associated environmental factors have little influence on autoimmune disease. To further verify this, Francis et al. (213) transplanted male or female BXSB mice with liver cells from male or female BXSB fetuses (segregated by histologic examination of gonads), and observed that male fetal liver cells never exposed to a male environment were still capable of inducing male-type early-life disease in both male and female lethally irradiated recipients similar to adult male spleen or bone marrow cells. Following the demonstration that murine lupus is caused solely by lymphoid system abnormalities and not abnormal autoantigens or other nonlymphoid cell associated factors, a variety of studies have been introduced to define the numerical and functional abnormalities of these cells that might elucidate its pathogenesis. The diverse abnormalities uncovered are summarized in Table V and discussed below in detail.

+

-

B. SURFACECHARACTERISTICS A N D NUMEHICAL ABNORMALITIES OF T AND B CELLS Analysis of the phenotypic expression of various surface markers on lymphocytes of lupus mice have been performed by several investigators. Initially, Stobo et aZ. (214) demonstrated a low B cell frequency and increased frequency of “nu!l” cells in NZB and NZB X W mice. They postulated that these changes indicated either (1)an abnormality in the capacity of Ig negative progenitor cells (possibly through a defect in the T cell compartment) to

304

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

TABLE V PIIENOTYPIC AND FUNCTIONAL ABNORMALITIESOF LYMPIIOID CELLSI N SLE MICE"

B cells Numerical Decreased numbers of sIg+ cells in NZ mice, increased numbers in BXSB and MRWl Decreased frequency with age of C3 and/or Fc receptor-bearing cells Increased sIgM/sIgD ratios in NZ mice Increased incidence of Ly-1 B cells in NZ mice Functional Elevated numbers of spontaneous Ig secreting and/or containing cells (hypergammaglobulinemia) Increased numbers of spontaneous antihapten antibody secreting cells Increased numbers of mitogen-induced clonable B cells Higher numbers of activated and/or proliferating cells Increased response to antigens, mitogens, and accessory helper T cell-derived signals in young NZ and BXSB mice Most of the polyclonal and autoantibody secreting cells belong to the Lyb-3+,5+ subset, and further in NZ mice to the Ly-l+ subset Defects in tolerance induction +

T cells Numerical Increased numbers ofThy-1.2+, Ly-1.2+, B220+, 9F3+ cells in M R W l mice. Decreased Thy-1.2+ cells in NZ mice Decreased numbers of T3-T4 + cells in MRWl mice Decreased numbers of Ly-123+ cells in NZB mice Functional Increased polyclonal helper activity in MRWl mice; spontaneous secretion of a B cell differentiation factor by the proliferating T cells Increased helper activity for cytotoxic responses in NZ mice Defective production and response to IL-2 (primarily in MRWl mice) Decreased spontaneous and mitogen-induced production of IFN Defects in AMLR responding cells Defects in tolerance induction Monocytes-macrophages Monocytosis in BXSB mice Increased frequency of Ia+ resident peritoneal macrophages in NZB and M R W l mice aUnless otherwise specified, the listed defects apply to all early-life SLE developing strains.

progress along normal pathways that involve the acquisition of surface Ig determinants, or (2) unusual numbers of B cells in transition between the precursor state and the state of active Ig secretion and, therefore, low sIg levels. In more recent studies by Cohen et al. (215), B cells of 8- to 10-weekold NZB mice had an increased cell surface IgM/IgD ratio compared to normal controls. Both of these findings (reduced numbers of B cells and

MUHINE MODELS OF SLE

305

increased IgM/IgD ratio per cell) were subsequently confirmed by Theofilopoulos et al. (216). The increase in IgM/IgD ratio was considered by Cohen et al. to indicate advanced B cell maturity, since LPS-activated B cells display a similar characteristic (217). Theofilopoulos et al. (216) find that male BXSB mice have a moderate proliferation of sIg+ cells, which represent the main cell population of the enlarged nodes of this strain, and MRL/l mice have a reduced percentage, but not absolute number of sIg+ cells. Furthermore, detailed studies by Theofilopoulos et al. (140) on the surface characteristics of B cells from lupus mice demonstrate that (1)the developmental Ig isotype diversity in all lupus strains is normal; sIgM is present on spleen cells obtained immediately after birth, whereas sIgD+ cells appear 3 days after birth. In older lupus mice, there is an increased frequency of IgG expression, which may be related to age-associated switching of total antibody and autoantibody from IgM to IgG; (2) B cells from newborn BXSB and MRL/I mice, as in immunologically normal mice, do not reexpress sIg after modulation with anti-Ig. In contrast, B cells of newborn NZ mice do express sIg after anti-Ig-induced modulations; (3) the rates at which sIg-anti-Ig complexes cap and become endocytosed in all SLE strains are within normal limits, (4) there is a higher frequency of C 3 receptor bearing cells in young lupus mice than normal mice, suggesting hastened B cell maturation, because the C3 receptors appear late in B cell development. However, the frequency of these C 3 receptor bearing cells falls markedly with advanced age and disease. This decrease may be caused by occupation of C3 receptors by complement-fixing ICs. It should be noted that recent studies have shown decreased C3b (CRI) receptors of red cells on human lupus patients (218,219), but whether this defect is genetic (220)or an epiphenomenon caused by C3 receptor occupation by ICs (221) or other ligands is debatable. CRI receptors are believed to play an important role (together with factor H and I of the complement system) in the degradation of IC associated C3b to iC3b and subsequently to C3c and C3dg/C3d (222, 223), (5) IgGFc receptor bearing cells decline with age, perhaps again due to IC occupation of the receptors or because these cells leave the circulation after IC binding. Abnormal Fc receptor function have also been described in humans with lupus and other disorders as well as normal individuals, particularly those of the HLA-DR2, -MTI, or B8/DR3 haplotype (224-226), (6) ontogenic development of Ia+ and Lyb 5+ B cells is normal, with normal or slightly elevated numbers of positive cells and alloantigen density; anti-Lyb 5 serum, which is known to suppress in vitro immune responses to thymus-independent (TI) antigens class I1 such as TNP-Ficoll (227), exerts equal suppressive activity in lupus and normal splenocyte cultures, and (7) SLE-prone strains, all characterized by high serum gp70 levels, have percentages of gp70+ cells in spleens (25-3096) and

306

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

thymuses (14-29%) comparable to those in normal mice (DBA/2, SJL, 129GIZ+) with similarly high serum gp70 levels. Recent studies by Manohar et al. (228) and Hayakawa et al. (229) reported increased numbers of large (activated) Lyt-1 , Thy-1 -, high sIgM , low sIgD + , and Iaf B cells in NZB and NZBx W mice. Smith et al. (230) found similar cells in thymectomized BXSB male mice. The Lyt-1 alloantigen, originally thought to be expressed only on helper T cells, was subsequently shown to be expressed, albeit to a lesser density, in all murine T cell subsets, including suppressor/cytotoxic T cells (231). More recently, it was observed that Lyt-l+ and Thy-1- lymphocytes can be found in the B cell area of lymph node germinal centers of normal mice and in some murine B cell lymphomas (228,232). Because of its presence in a small population of B cells, it has been suggested that this alloantigen be renamed Ly-1, instead of Lyt-1 (229). Although the mean frequency of Ly-l+ B cells in most normal mice (including nude and CBA/N) is approximately 2% of the total spleen cells, the frequencies in NZB and, more prominently, in NZBxW, can reach 5-15%. Such cells can be found in spleens, but not nodes or thymuses. Despite the higher numbers of such cells in NZ mice, the pattern of ontogenic development does not differ from normals; the appearance of Ly-1 B cells coincides with the appearance of the first IgD-bearing B cells in spleens of 3-day-old mice, and rapidly reaches adult levels. NZB-xid mice, which fail to develop typical NZB autoimmunity (see Section IV,C, 1)apparently have normal numbers of Ly-l+ B cells, which implies that autoantibody secreting cells belong in this subset of B cells. Of considerable interest, therefore, is the observation of Hayakawa et al. (229) that FACS-sorted Ly-l+ B cells from NZB mice secrete approximately 40 times as much IgM as sorted Ly-1 B cells from BALB/c mice cultured in the absence of added antigens or mitogens. Ly-1- B cells, in contrast, did not secrete detectable levels of IgM except when sorted from recently immunized animals. Thus, spontaneous IgM secretion by spleen cells cultured from nonimmunized NZB mice appears to be due entirely to the activity of the lymphoblastoid-like Ly l + B cells. More recent studies by Hayakawa et al. (233) have indicated that IgM autoantibody secreting cells (anti-DNA, NTA, anti-red cell) of NZB and normal mice, after mitogenic stimulation, belong in this Ly-l+ B cell subset. These investigators found that a different B cell subpopulation (IgM+, IgD-, Ly-1-) secretes most of the IgM antibodies produced in response to exogenous antigens and thus conclude that Ly-l+ B cells contribute a functionally distinct B cell population destined to respond to autoantigens only. These findings, although provocative, must be cautiously interpreted, since (1)MRL/I mice, which develop a different type of autoimmune disease than NZB but are characterized by autoantibodies and hypergammaglobulinemia, have normal Ly-l+ B cell numbers; (2) Okumura et al. (234) in in vitro experiments with unprimed Ly-l+ B cells and Shimamura +

+

+

307

M U R I N E MODELS OF SLE

et (11. (235) in adoptive transfer experiments with sheep red blood cell (SRBC) primed Ly-1 B cells have shown that such cells can exert feedback suppression (suppressor-inducers), a finding that does not fit well with the increased numbers in NZB mice, despite the presence of hypergammaglobulinemia and autoantibodies. The proportion and absolute numbers of T cells have also been studied by several investigators. Initially, decreased numbers of T cells in NZ mice were described and attributed by some authors to the effects of NTA and selective loss of suppressor T cells (170,171,214,236). Subsequently, Cantor et al. (237)determined the proportion of Ly T cell subclasses in the spleens of NZB mice during the first 12 months of life. These investigators demonstrated that during this time, NZB mice develop inordinately high proportions of Lyt-I+ and Lyt-23 cells, and substantially reduced concentrations of Lyt-123+ cells compared to age and sex-matched normal BALB/c mice. On the basis of additional functional in oitro experiments, these authors concluded that these alterations are compatible with reduced feedback-suppressive activity in NZB spleen cells. They also noted that Lyt-123+ cells and associated feedback-suppressive activity are substantially reduced in aged mice of several inbred strains, suggesting that the reduction may be a manifestation of immunologic senescence which apparently occurs earlier in NZB than normal mice. Theofilopoulos et al. (216) also enumerated Thy-1 cells in lymphoid organs of all lupus strains and found a decrease in NZB, female NZBXW, and BXSB male mice with advanced age and disease. MRL/n mice show a modest increase in frequencies and absolute numbers with advanced disease. In contrast, more than 90% of the cells in the enlarged lymph nodes of the congenic MRL/l mice are positive for Thy-l+ antigen, indicating that their lymphadenopathy is largely due to T cell proliferation. These investigators have also used cytotoxicity assays to examine the various Ly subsets in lymph nodes of MRL/l females, and find that advanced age and lymphoproliferation is accompanied by an extreme reduction in Lyt-123+ and Lyt-23+ cells compared to values in young animals, and a concomitant increase in the percentage and numbers of T h y - l + , Lyt-, or null cells. Subsequently, the more sensitive FACS technique has found that these Lyt- T cells actually express low amounts of Lyt-1 antigen (238). The proliferating Thy-1 +, Lyt-l+ cells are Ia- according to immunofluoresence techniques, but a large proportion of the Thy-1 splenocytes also express I-J determinants (140,216). In addition, the proliferating T cells in MRL/l lymph nodes possess lower electronegative surface charge (239) and a higher number of potassium channels (240) than the MRL/n cells. Although MHL/I lymph node cells display similar amounts of surface receptors for peanut agglutinins as MRL/n lymph nodes, the expression of surface receptors for other lectins is either lower (Con A) or higher (Helix pomatia) (239). In addition, MRL/l but not MRL/n, lymph node cells exhibit altera+

+

+

+

+

308

ARGYRIOS N. THEOFILOPOULOS A N D FRANK J. DIXON

tions in terminal carbohydrate structure of glycolipids and glycoproteins (240a). Further analysis of the proliferating T cells of MRL/I mice (and of other mice that carry the Zpr mutation) has been performed by Morse et al. (241,242). Utilizing a battery of monoclonal and polyclonal antibodies for T and B cells, these authors found that the proliferating cells were low Thy-1 +,low Lyt-l+, Lyt-2- , sIg- but the majority also express an alloantigen [Ly-5(B220)]normally detected only on cells of the B lineage. Similarly, a xenogeneic monoclonal antibody against MRL/l T cells, although stained brightly 80-90% of cells in enlarged nodes of MRL/l mice but weakly about 30% of T cells in MRL/n nodes, it also stained strongly B cells from both types of these mice as well as B cells of normal mice (243). In addition, a monoclonal antibody (anti-Ly-6.2, 34-10-7) which reacts with Thy 1.2+ as well as bone marrow cells of normal mice was found to stain the majority of the proliferating Thy-1.2+, Lyt-1 cells in lymph nodes of Zpr homozygous mice (244). Finally, another monoclonal antibody (mAb 9F3) was observed to stain brightly 90-98% of lymph node cells from lpr homozygous strains, but 10 to 50 times less intensely 55-70% of T cells from control non-lpr congenic mice (245). This mAb also stains resting or mitogen-activated B cells (as well as macrophages, granulocytes, and erythrocytes) in all lpr homozygous strains and their congenic non-lpr counterparts as well as normal strains of mice. These results raise questions with regard to the T or B cell lineage of the proliferating cells in lpr homozygous mice. However, molecular genetic studies have shown that (1)the proliferating T cells in MRL/I lymph nodes do not have rearranged Ig heavy chain genes as in pre-B and B cells (241), (2) they express mRNAs corresponding to the recently cloned P-chain of the murine T cell receptor (241a), and (3) they are positively stained with a monoclonal antibody directed against an allotypic receptor determinant present on roughly 25% of normal cells (M. Bevan and A. N. Theofilopoulos, unpublished observations). Whether all the proliferating cells express the idiotype related antigen needs to be investigated further. Studies by Wofsy et al. (246) have, however, shown that the proliferating cells in MRL/I mice lack the surface antigen L3T4, which is present on normal helper T cells. In toto, the studies cited above indicate that although some of the proliferating cells in Zpr homozygous mice aberrantly express certain B cell surface markers, these cells are clearly of the T cell lineage and that this aberrant expression may simply reflect the state of activation or differentiation of these cells. Indeed, normal T cells, upon mitogenic stimulation, have been found to express certain B cell antigens that are identified by the mAb 9F3 and mAb 34-10-7 (244,245). The abnormal T cell proliferation in Zpr homozygous mice is also associated with high expression of the oncogene c-myb, but such increases were not detected in lymphocytes of other SLE strains nor in humans with SLE (246a). +

MUHINE MODELS OF SLE

309

C. FUNCTIONAL ABNORMALITIES OF B CELLS,T CELLS, MACROPHAGES, A N D RELATED INTERLEUKINS The most notable immunologic marker of murine (and human) SLE is B lymphocyte hyperactivity manifested by spontaneous polyclonal antibody production and secretion of various autoantibodies. Many investigators have, therefore, studied the functional status of the various immunocytes that participate in and regulate immune responses in an attempt to gain information on the development of the B cell abnormality which leads to disease expression.

1 . B Cell Abnormalities a. Polyclonal B Cell Activation. Moutsopoulos et al. (247) using a sensitive assay initially observed that splenic lymphocytes of NZ mice, in contrast to normal strains, spontaneously produce IgM at birth. By 6 to 10 weeks of age, spleen cells from NZ mice produce 20- to 40-fold more IgM than normal strains. The authors concluded that, unlike normal B cells, NZ B cells are activated at birth. This conclusion was subsequently confirmed by Manny et al. (248) who attributed the IgM hyperproduction to two independently segregating genetic mechanisms: (1) increased IgM-containing cells, and (2) increased IgM secretion per cell. On the basis of crosses with normal mice, it was tentatively concluded that the NZB mouse had one or more dominant genes determining the number of IgM-containing cells, and that the increased amount of IgM secreted per cell was controlled by a recessive gene (or genes). Subsequently, other investigators confirmed the early B cell maturity and polyclonal activation in NZ mice, and extended these findings to the BXSB male and the MRL/I strain. Thus, Izui et al. (249) found that all lupus strains spontaneously produce more antihapten antibody secreting cells in spleens, and greater concentrations of antihapten antibodies in sera than age-matched immunologically normal strains. This increased B cell nonspecific antibody production correlates well with the spontaneous development of anti-DNA antibodies. Subsequently, Theofilopoulos et al. (72) and Slack et al. (25)demonstrated B cell hyperfunction as a common characteristic of all lupus strains, manifested by the inordinate numbers of Igcontaining/secreting splenocytes. In NZB and NZB X W mice, the high frequency of Ig-secreting cells (IgSC) is detectable as early as 1 month of age, and increases somewhat thereafter. In contrast, the high frequency of IgSC in BXSB male and MKL/l mice is first observed at or a little before the clinical onset of disease (approximately 3 months of age). The number of spontaneous IgSC correlates with the number of Ig-containing cells, but the number of the latter is larger than the former in each strain examined (approximately 6-7% of the total spleen cells contain Ig, whereas 0.4-0.6% secrete Ig). With advanced clinical disease, NZ mice have 5- to 10-fold, and

310

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J . D I X O N

MRL/l mice a 30-fold higher number of Ig-containing cells than younger, syngeneic animals. Of interest, spleen cells from young autoimmune mice secrete predominantly IgM, but switch to predominantly IgG with age and disease onset. In contrast, spleen cells from normal mice show minimal agerelated polyclonal Ig switching. Slack et d . (25) also observed that the agerelated enhanced IgG production in lupus mice is selective for IgG subclasses. IgG,- and IgG,,,-secreting cells increase 6- and %fold in BXSB males from 2 to 5 months of age, while IgG, and IgG,, show less increase (3- to 6fold). MRL/1 females from 2 to 5 months of age have an &fold increase in IgG,,-producing cells, a 6-fold increase in IgG,,,-producing cells, and smaller increases in IgG,- and IgG,-secreting cells. A distinct pattern of isotype preference is common to NZBx W and NZB, with IgG,,-secreting cells being selectively stimulated as these mice approach the time of renal disease onset. The expanded IgG subclasses are the primary antibodies found to be involved in serum and kidney IC of each mouse. Further evidence for a spontaneous polyclonal B cell activation in all early-life lupus strains was obtained by Slack et nl. (250), who observed an increased degree of somatic mutations or diversity of the highly conserved murine AlV region in such mice compared to age-matched late-life lupus and normal strains. The levels of mutated XlV correlated well with the degree of hypergammaglobulinemia being highest in MRL/1 followed by male BXSB and female NZBx W. Similarly, increased levels of hypermutated XlV can be induced in normal mice after mitogenic (LPS) or antigenic (TNP-OVA) stimulation. Raveche et al. (43,251) performed cell cycle analysis of lymphocyte activation in normal and lupus mice and found that spleens of NZB and MRL/l, but not male BXSB, have higher percentages and absolute numbers of spontaneously proliferating cells at disease onset than age-matched normal mice. The increased proliferation was not found in thymuses or bone marrows. Increased absolute numbers (but not percentages) of proliferating cells are also found in the enlarged nodes of older MRL/l mice. Percoll discontinuous density-gradient centrifugation of anti- Thy 1-2 complement treated spleen cells shows increased numbers of large in v i m activated B cells in all lupus mice, particularly NZB and MRL/I; less in male BXSB. In addition to increased IgSCiIg-containing cells and large activated B cells, studies by Ohsugi and Gershwin (251a) Kincade and associates (207,252), and Theofilopoulos et al. (140) also demonstrated increased spontaneous and LPS-induced clonable B cells in semisolid agar cultures of spleen and bone marrow cells in all lupus strains. This is considered another measure of polyclonal B cell activation. More recently, Jyonouchi and Kincade (253, 254) reported an earlier emergence (15-16 days of gestation) and larger numbers of colony-forming B cells in fetal livers of NZB compared to normal mice. Furthermore, these authors found that hyperactive regulatory

+

M U R I N E M O D E L S OF S LE

311

(adherent) cells in young NZB marrow and fetal livers may promote fimctional maturation of normal pre-B cells. In addition, potent substances (perhaps not Ig) in the serum of young, but not old, NZB mice had the same inducing effects. b. Defects in B Cell Signul Requirements. In the preceding sections, we have stressed the importance of polyclonal B cell activation and increased Ig secretion in murine lupus. Such polyclonal Ig secretion could be primary or secondary to a number of immune defects, including decreased suppressor or increased helper T cell function and abnormalities in other accessory cells such as macrophages as well as in the signals necessary for B cell maturation, Before considering which of the above may be responsible for abnormal B cell maturation in lupus, a brief description of the normal mechanisms of B cell activation, proliferation, and differentiation is outlined. A resting B cell is activated and induced to clonal expansion and Ig secretion by complex mechanisms, and calls for antigen-specific and/or antigennonspecific T cell help. The work of several authors suggests that a resting B cell must first be activated before undergoing any other response (255-260). However, the possibility that some subsets of small “resting” B cells can respond to maturation factors cannot be completely excluded. For example, Andersson and Melchers (259) reported that although small B cells (isolated on the basis of cell density) could not proliferate in response to B cell replication and maturation factors, they could differentiate into Ig-secreting cells. B cell activation is characterized by a transition from the G,, to the G , phase of the cell cycle accompanied by cellular enlargement. T-independent (TI) antigens have the unique ability to activate B cells following binding to the B cell antigen receptor (signal 1)without the need for the T cell-derived activating signal (signal 2). This second signal may be delivered by the TI antigens either by their polyclonal activity (TI-class 1 antigens) or by their polyvalency and ability to cross-link antigen receptors (TI-class 2 antigens) (261,262). On the other hand, most antigens are so-called “T-dependent” (TD) and require both signal 1 and 2 to induce B cell activation (262,263). Signal 2 consists of a direct antigen-specific Ia-restricted T cell-B cell interaction (264). Once the B cell is activated by TI or TD antigens, it expresses acceptor sites for lymphokines or interleukins which mediate proliferation [interleukin 1 (IL-I), interleukin 2 (IL-2) and B cell growth factors (BCGF)], and differentiation [T cell replacing factors (TRF), B cell differentiation factor (BCDF)] (255-265a). Multiple BCDFs might exist which can induce IgM secretion (BCDFp) (266), IgG secretion (BCDFy) (267), IgA secretion (BCDFa) (268), or IgE secretion (BCDFE)(269). The exact sequence of lymphokine receptor expression on B cells is still under investigation; however, following activation, receptors or acceptor sites for growth factors appear to be expressed earlier than those for differentation factors. Under normal circumstances, a fraction of all B cells are in the activated or

312

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

Ig-secreting state. These in vivo activated B cells obviously contribute to basal Ig secretion, which maintains normal immunoglobulin levels in serum. Immunization is accompanied by increased circulating activated B cells (270). Do the B cell responses of lupus mice differ from the normal responses outlined above? The answer is clearly yes in some strains, but these differences are mainly quantitative rather than qualitative. To assess the signal requirements of B cells from young lupus mice, Prudhomme et al. (271) made use of the fact that in cultures of very low density (lo4cells/250 pl), normal B cells do not proliferate or secrete Ig after LPS stimulation unless accessory signals found in supernatants of Con Astimulated T cells (CAS) are added. Similarly, B cells do not proliferate in response to anti-p in low-density cell cultures, and do not secrete Ig at any cell density unless T cell-derived factors are provided. LPS and anti-p presumably prime the B cells by inducing them to express receptor or acceptor sites for antigen nonspecific T cell-derived helper factors. These investigators also assessed the in vitro SRBC response of autoimmune strains, since B cells require several signals to respond to this antigen. In fact, in low cell density cultures, even in the presence of LPS, at least three signals (antigen, LPS, T helper cell products) are required to induce an anti-SRBC response (258). Using the above in vitro systems, it was found that B cells of young (3week-old) BXSB and NZBx W mice, but not of young MRL/I mice, are more responsive than normal B cells to almost every signal tested. Thus, BXSB and N Z B x W B cells proliferate in vitro at a 3- to 5-fold higher rate than normal B cells when exposed to polyclonal activators such as anti-p or LPS. Nevertheless, these hyperactive B cells cease to respond to LPS (or anti-p) in low density cultures (lo4 B cells/well). The proliferative response was partially restored by adding CAS of normal splenocytes, and the response is higher in BXSB and NZBXW than in MRL/l and normal mice (Table VI). Similarly, differentiation of LPS-stimulated B cells at very low cell density, or anti-p stimulated B cells at any cell density is dependent on the presence of soluble T cell-derived factors in all autoimmune and normal strains tested, but again Ig production is highest in BXSB and N Z B x W mice (Table VI). Furthermore, low-density cultures of B cells from all normal or autoimmune strains tested require three signals to respond to a thymus-dependent (TD) antigen (SRBC), i. e., antigen, polyclonal activator (LPS) (thought to replace the Ia-restricted signal normally delivered by T cells), and conditioned medium (CAS). This antigen-specific response is again higher (2- to &fold) in BXSB and N Z B x W than in the other strains (Table VI). Interestingly, these investigators found that, although BXSB and N Z B x W cells produce predominantly IgG, in response to LPS (in contrast to normal B cells for which

TABLE VI SIGNAL REQUIREMENT A N D DEGREE OF RESPONSEBY B CELLSOF SLE MICE Proliferative response" ([3H]thyrnidine uptake-Acpm)

B cell origin BXSB NZBXW

MRWl BALBIc C567BW6

LPS

8 -42 41 - 176 -665

LPS

+ CAS

3072 1845 983 926 647

DifferentiationL (IgM ndrnl) LPS

28 30 18 12 19

LPS

+ CAS 192 180 82 50 86

Antigen-specific responsec (indirect anti-SRBCiculture) Medium

13 12 12 10 NT

SRBC

+ CAS

84 82 75 72 NT

SRBC

+

94 81 62 68 NT

LPS

CAS

+

LPS

SRBC

118 114 82 80 NT

+ CAS + LPS 1880 1784 802 792 NT

DAssayedat day 3 of culture. LPS was added to cultures of lo4 B cells in 250 pl at a concentration of2.5 pg/nil. Con A supernatant (CAS)was derived from C57BL/6 spleen cells incubated for 2 hours with 4 p,g/rnl of Con A, washed, and recultured for 24 hours in the absence of Con A. CAS was added at a concentration of 40% (viv). Minus valnes indicate less uptake than that of similar cultures i n medium alone (no niitogen added). 61gM was nieasured by radioiininunoassay after 6 days of culture of lo-' B cellsl250 pl in the presence of LPS (2.5 ~ F / m l or ) LPS CAS (40% viv). r2x 105 B cellsiinl from each of the listed strains were cultured for 5 days. SRBCs were added at a concentration of 2 x 10Vni1,LPS was added at 2.5 pg/ ml, and CAS at 40% (v/v).

+

314

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

IgG, predominated), they nevertheless require T cell-derived soluble factors to secrete high levels of IgG,. IgM to IgG, switching is known to be highly T cell dependent (267). Notably, non-mitogen-activated splenic B cells of young BXSB and NZB x W mice produce abnormally high numbers of IgSC when exposed to BCDFs. This differentiation is thought to depend on the presence of B cells activated in uiuo and suggests that, prior to disease onset, even young lupus mice already have an increase in spontaneously activated B cells. In conclusion, the studies of Prud’homme et aZ. (271) demonstrated that for proliferation and differentiation, B cells from normal mice are dependent on the same signals as B cells from normal mice. However, the B cells of two autoimmune strains (BXSB, NZBx W), once triggered by the usual signals, give abnormally high responses which might be the basis of their autoimmune diseases. On the other hand, the B cells of a third SLE strain, MRL/l, respond normally in the above-described in vitro assays. This does not exclude the possibility, however, that MRL/l B cells have more subtle defects which become manifest later in life. Using FACS isolated B cells from NZB and limiting dilution experiments, Pike et al. (272) reached conclusions essentially similar to those of Prud’homme and associates. These investigators found that, in comparison to normal B cells, NZB B cells exhibited the following: (1) higher spontaneous conversion into antibody-forming cell clones in the absence of antigen or mitogen, (2) substantially higher cloning efficiency at optimal antigen or mitogen concentration, and (3) a lower antigen concentration optimum. These differences argue for a heightened excitability of the NZB B cell to triggering stimuli as a dominant factor in the etiology of their disease. c. Effects of the xid Gene on Murine Lupus. Certain mice, such as the prototype CBA/N strain, bear an X-linked gene (xi4 and are deficient in a subset of mature B cells (Lyb-3+, Lyb-5+) responsible for immune responses to the so-called TI-class 2 antigens such as polysaccharide antigens with repeating structures (reviewed in 273). Studies by Steinberg and associates (274-278) demonstrated that introduction of this gene into lupus mice significantly retards or arrests disease manifestations. Such mice derived from a series of crosses and backcrosses between lupus and CBA/N mice bear predominantly lupus strain autosomal genes, but have X chromosomes derived only from the xid gene bearing CBA/N. NZB.xid mice fail to manifest the autoimmune syndrome characteristic of NZB mice, have a much reduced quantity of polyclonal Ig (particularly IgM, IgG,) and autoantibodies, and live almost normal life spans (274). The same is true for NZBxW.xid (275), BXSB.xid (278,279), and (NZBxBXSB)F,.xid (278). Although originally reported otherwise (280,281), recent studies indicate a beneficial effect of the xid gene in the disease of MRL/1 mice as well (277). Expression of the Zpr gene is unaffected in MRL/I.rid mice, since

MURINE MODELS OF SLE

315

these mice develop lymphadenopathy and other T cell defects such as reduced production of interleukin-2 (IL-2). Since the main effects of the xid gene appear to be deletion, retardation, or change in homing patterns of Lyb-3 , 5 B cells, the conclusion is that the main subpopulation of B cells responsible for autoantibody production belongs in this subset, irrespective of autoantibody specificity. This conclusion is supported by limiting dilution experiments indicating reduction in the anti-DNA B cell precursors in xid-bearing inice (282), and reduction of anti-red blood cell antibody in immunized mice treated with anti-Lyb-3 antibodies (283), as well as the finding that spontaneously hyperactive B cells of NZB mice belong to the xid gene affected B cell subset (284). However, the requirement of the Lyb-3 5 B cell subset for autoantibody production appears not to be absolute, since (1) stimulation of intact NZB.xid with polyclonal activators (LPS) for prolonged periods of time led to eventual production of large amounts of autoantibodies (285), (2) spontaneous autoantibodies to red cells and DNA occurred in a sinall percentage (20-40%) of NZB,xid, and late-life NZBx W.xid developed autoantibodies such as NTA, anti-red blood cell and anti-DNA antibodies (286,287), and (3) several (CBA/Nx NZB)F, males, despite expressing the CBA/N-associated iinmunologic defects, were still able to produce autoantibodies, although to a lesser degree than similar F, females (286). Whether or not cells of the Lyb-3,5 lineage are the only cells that produce autoantibodies, it is generally agreed that introduction of the rid gene retards inurine SLE. Since B cells of lupus mice were found to require accessory signals in order to proliferate and differentiate, and since Lyb-3,5+ B cells are considered to be the primary target of antigen-nonspecific MHC-nonrestricted helper signals (264,288), Fieser et al. (289) examined whether B cells from xid lupus mice could respond to accessory factors. They found that B cells from NZB.xid and MRL/I.xid had drastically reduced responses to BCGF- and BCDF-like activity in conditioned media froin EL4 cells compared to non-xid B cells. These results demonstrate that introduction of the xid mutation into SLE strains does indeed render their B cells unresponsive to the proliferation and differentiation signals provided by T cell-derived helper factors, as has already been described for CBA/N B cells (264,288,290). Thus, the reduction in disease severity in lupus background inice afforded by the xid locus is probably due to prevention of development of B cells which hyperrespond to or are the targets of accessory helper signals. d. Acceleration of SLE by Polyclonal B Cell Activators. Transfer ofautoimmunity with bone marrow or spleen cells of SLE inice into histocompatible normal mice and the absence of autoimmunity after reciprocal transfer of normal lymphoid cells into SLE strains suggest the irrelevance of nonlymphoid cell-associated polyclonal activators as primary causative agents of inurine SLE. Nevertheless, experimental data discussed below indicate that +

+

+

+

+

316

AHGYRIOS N. THEOFILOPOULOS A N D FRANK J. DIXON

in a genetically predisposed individual, polyclonal B cell activators may serve secondarily as accelerators of autoimmunity. Such polyclonal B cell activators may be found in a variety of bacterial, parasitic, and viral products (291). Bacterial LPS has been repeatedly shown to induce in vitro and in vivo a variety of autoantibodies, including anti-DNA (74-76,292), anti-y-globulin (293,294), antithyinocyte (295), and antierythrocyte autoantibodies (296), primarily of the IgM class. Although several studies (75,292) have shown that injections of LPS into normal mice can lead to production of autoantibodies, none of them has documented induction of significant histologic or clinical IC disease, i. e., glomerulonephritis and vascular lesions. Recently, Hang et al. (194,297) performed experiments in mice to determine (1) can chronic B cell activation alone induce significant autoimmune disease, and (2) what role(s), if any, might a genetically determined host predisposition play in the development of such disease? These investigators found that chronic stimulation with a nonantigenic but mitogenic lipid A fraction of LPS, beginning early in life, can greatly accelerate the onset of late-life SLE disease of MRL/n, BXSB females, and NZW females, as evidenced by the accelerated mortality (from the second year of life in unmanipulated mice to 6-7 months in LPS-treated mice), polyclonal IgG levels, autoantibodies, and IC-mediated glomerulonephritis. However, similar chronic mitogen stimulation has very little effect on survival and development of glomerulonephritis in immunologically normal mice (BALB/c, C57BL/6) despite the similar increases in polyclonal IgG and autoantibody levels. Thus, it appears that genetic predisposition is of paramount importance in determining the immunopathologic effects of stimulation by polyclonal B cell activators such as LPS. It is difficult at present to explain why normal mice, despite the expression of autoantibodies, fail to develop the severe autoimmune disease of lupus-prone mice. Two possibilities seem likely: (1) either the fine specificities (Ig class, affinities, epitopic specificities) of the autoantibodies induced differ in the two groups, or (2) additional complementary autoantibodies or other factors arising spontaneously in SLE-prone mice are needed to sustain any adverse effects induced by the mitogen.

2 . T Cell Abnormalities As indicated above, murine lupus appears to be caused by abnormalities of the lymphoid system manifested by advanced B cell maturity. Moreover, when compared to normal mice, NZB or NZB x W mice under 10 weeks of age show excessive antibody responses to a variety of antigens (298-304). Recent studies by Park et al. (305) on immune responses of lupus strains to thymus-dependent (TD) antigens show stronger responses in all SLE strains compared to age- and allotype-matched immunologically normal controls. The IgG subclass profiles in all SLE strains differ from normals; thus, in response to TD antigens, all SLE strains make responses characterized by

317

M U R I N E MODELS OF SLE

IgC,,? IgG,,> IgG,> IgG,, thereby deviating from a normal pattern of IgG,> IgC,,> IgG,,> IgG,. Ontogenically, Tl-2 responses appear earlier in NZ mice than in other SLE and normal strains. Additionally, in these and other studies (306), some SLE strains (NZB, MRL) sustained immune responses to antigens for longer periods than normal mice. As discussed earlier, spontaneous autoantibodies of lupus mice switch isotype (IgM + IgG) later in life, and many of the autoantibodies, as well as the predominant isotypes in kidney deposits, are of the IgC, subclass (23,25,1OO,101). Collectively, these findings should be interpreted relative to the experiments mentioned above in which it was found that T cells and their products exert profound effects on TI and TD responses as well as on the polyclonal B cell activation and the Ig subclass switch at the B cell level (265-269,307-310). Therefore, a variety of studies have been performed in both humans and mice with SLE to determine whether B cell hyperactivity and autoantibody production are caused by numerical and/or functional defects of the various regulatory T cell subsets. a. Suppressor T Cells. An attractive hypothesis for the observed B cell hyperactivity of lupus mice is lack of or deficiencies in suppressor T cells. However, despite some initial findings supporting this concept, more recent detailed studies have failed to disclose any significant abnormality in this regulatory T cell subset. Krakauer et al. (311), using pokeweed mitogen (PWM)-driven IgM biosynthesis in uitro as an indicator system, found that Con A-activated spleen cells of adult NZB x W mice have decreased suppressor potential compared to Con A supernatants of cells from adult normal or young NZBxW mice. Despite the deficiency in production of T cell derived suppressor factors, these animals’ B cells are responsive to normal T cell derived suppressor factors. Furthermore, the in uiuo administration of supernatants of Con A-activated normal spleen cells to NZB X W mice results in decreased Ig levels, antinuclear antibodies, proteinuria, and renal pathology (312,313). Subsequent investigations by Cantor et al. (237) suggest that the suppressor defect in NZB mice is caused by a numerical decrease in the Ly-123+ cells. These investigators have suggested that other types of cell circuit defects also occur in BXSB and MHL/I mice (314); in BXSB mice, the claim is that their Lyt-l+ cells are unable to induce Lyt-123+ cells to express feedback suppression, whereas in MRL/l mice, Lyt-1 helper cells (abundant in these mice) are insensitive to suppressor effects by Lyt-23 cells. However, supportive data for these claims are scant. A related assertion is that NTA in the sera of NZ mice are responsible for the decreased numbers and/or function of suppressor T cells, since these autoantibodies are reportedly to have preferential reactivity with the suppressor Lyt-23f T cell subset (169,170). It was also reported that daily injections of NTA-containing sera into very young NZBXW mice induce an earlier loss of Con A-induced suppressor T cell function than in uninjected +

+

318

ARGYRIOS N . THEOFILOPOULOS A N D FRANK

1. DIXON

controls (171). Similar studies have been performed in humans with SLE (Con A-induced suppression, thymocytotoxic antibody effects i n vitro, subsets of suppressor T cells defined by monoclonal antibodies, or IgG Fc binding) and defects were reported that sometimes correlate with disease activity (315-319). Thus, all these findings were interpreted to indicate that systemic autoimmunity, hypergammaglobulinemia, and autoantibody production in lupus are caused by generalized defects in suppressor T cell components. This proposition, however, has been questioned. Studies by Theofilopou10s et al. (72) fail to disclose any significant differences in spontaneous and LPS-induced Ig secretion between Con A-stimulated spleen cells from young and old autoimmune mice and normal controls. Moreover, B cells from autoimmune and normal strains are equally receptive to Con A-induced suppressor signals. The studies of Creighton et nl. (320) on IgG and IgE responses to exogenous antigens by lupus and normal mice reveal no significant suppressor T cell differences in these two groups of mice. Others have also reported normal suppressor T cell function in NZ mice (248,321). Furthermore, as discussed above (Section 11,C,7), a variety of studies have questioned the primary importance of NTA in the expression of systemic autoimmunity. To recapitulate: (1) NTA can be found in immunologically normal mice; (2) MRL/l and BXSB mice, despite their fulminant SLE, express very low levels and incidence of NTA; (3) hereditary asplenic ( D h / + ) NZB mice develop autoimmune disease without NTAs; (4)recombinant NZB inbred strains express NTAs in the absence of other autoantibodies or, conversely, express anti-DNA and antierythrocyte autoantibodies without NTA; and (5) no relationship exists between NTA levels and other autoantibodies or mortalities in F, generations of lupus mice. The concept of a generalized suppressor T cell defect causing human lupus has also been questioned, since several recent studies assessed these cells by a variety of means and found them to be within normal limits (322,323), and antilymphocyte antibodies do not react preferentially with any particular type of immunocyte (324). The reasons for the discrepancy between studies that find suppressor T cell defects and those that do not are not clear at present. It would appear that the initial conclusion that SLE is caused by suppressor T cell defects was premature, although the issue is still open. Experiments addressing the status of autoantigen specific suppressor T cells have not yet been performed. Furthermore, one cannot exclude the development of late, secondary suppressor T cell abnormalities that might contribute to disease expression. b. Helper T Cells. Because Ig hypersecretion, autoantibody production, and hyperresponsiveness to exogenous antigens may be caused by increased helper T cell activity instead of a decreased suppressor T cell activity, experi-

MUHINE M O D E L S OF SLE

319

nients have been done to determine the status of the former cells in lupus mice. A systematic analysis of helper T cell function in lupus mice was performed b y Theofilopoulos et al. (72) and Prud’homme et (11. (325). Initial experiments examined the degree of help provided by increasing numbers of isolated T cells from young and old animals to a standard number of mitogen(LPS) activated syngeneic and allogeneic, but H-2 compatible, B cells isolated from spleens of young animals (72). Increments of T cells from young and old NZB and BXSB autoimmune inice added to a standard number of B cells from syngeneic young mice at all doses and ages show equal help in enhancing the IgSC frequency after LPS stimulation. Moreover, the T cell help from these two autoimmune strains to their own B cells is not significantly different from that of T cells from young and old normal mice of the same H-2 haplotype (BALB/c for NZB, C57BL/6 for BXSB). Conversely, when T cells from young and old NZB and BXSB mice are added to B cells from young, normal counterparts, the help is not significantly greater than that of T cells from normal strains. The only notable exception is the MRL/I strain, in which T cell-enriched populations from old animals added at a 4:l ratio to B cells from syngeneic young animals provide 2-3 times the help offered by equal numbers of T cells of young syngenic animals or T cells from young and old normal mice of the same haplotype (CSH/St). We have demonstrated (216) that the T to B cell ratio in MKL/l mice is approximately 30:l in uiuo. Moreover, MKL/I mice have up to a 100-fold excess of total T lymphocytes compared to normal mice. Therefore, the enhanced helper activity seen in uitro with maximum T to B ratios of 4:1 should be far greater in the intact MRL/I mouse. These and other experiments (326) suggest that advanced B cell maturity in MRL/I may be the result of heightened T cell helper activity exerted by the proliferating Lyt-l+ cells. Helper T cells secrete factors (BCGF, BCDF) that can induce proliferation and differentiation of activated B cells. Since the above described cellmixture experiments suggest that the proliferating T cells of MRL/l mice exert excessive help, Prud’hoinme et (11. (325) attempted to determine whether this effect is mediated by soluble factors and, if so, the nature of the factor produced. Cultured lymph node and spleen cells from this substrain of lupus inice indeed spontaneously produces (in the absence of mitogenic stimulation) abnormally high levels of a factor(s) that causes mitogen- or anti-p-activated isolated B cells froin mice of various H-2 backgrounds to differentiate into IgSC. Normal levels of this activity are produced by the congenic MRL/n substrain lacking the 1pr gene and by normal murine strains. However, such a factor is hypersecreted in vitro, albeit to lower levels than MKL/l, by lymphoid cells of the C57BL/6-lpr/lpr and C3H/ HeJ-lprIIpr mice, which develop autoimmune and T cell proliferative syndromes. It appears that the presence of the lpr gene is sufficient for the expression of a T cell subset that spontaneously secretes a BCDF-like fac-

320

A R G Y R I O S N . THEOFILOPOULOS A N D F R A N K J. D I X O N

tor(s). Cells of MRL/l mice begin producing BCDF in uitro as early as 1 month of age, but levels increase with age and onset of lymphoproliferation. Cell depletion studies reveal that this factor is produced by T cells of the Lyt-l+ phenotype, which is the dominant cell in the enlarged MRL/l lymph nodes and spleens. Because of its association with the lpr genotype, this factor is termed L-BCDF. No IL-2, BCGF, interleukin 1 (IL-l), inacrophage activating factor (MAF), interferon, or macrophage Ia recruiting factor (MIRF) has been detected in unstimulated MRL/l splenic or lymph node-derived supernatants expressing BCDF activity. L-BCDF notably enhances both IgM and IgG production by LPS-activated B cells. Interestingly, IgG production is more responsive to L-BCDF than IgM production. LPS by itself induces mostly IgG, and IgG,,, secretions in normal B cells. The addition of MRL/l lymph node supernatants together with LPS has little effect on IgG, secretion, but considerably enhances production of the T cell dependent subclasses IgG,, IgG,,, and IgG,b (Table VII). The high incidence and levels of polyclonal and monoclonal IgG, and IgG,, and of IgG,, autoantibodies in these animals (5,25,106,327) provide in uiuo analogies for the in uitro effects of L-BCDF. Thus, increased numbers of T helper phenotype cells and increased production of an antigen-nonspecific BCDF-like factor secreted by such cells could readily cause the B cell hyperactivity and early severe disease in MRL/l mice. c. Defects in ZL-2.Interleukins are a family of molecules transmitting growth and differentiation signals between various types of leukocytes and thus presumably are major effectors of immune regulation. Among the various interleukins, IL-2 produced by T cells is believed to provide a universal signal for proliferation of antigen or mitogen-activated T cells through its binding to specific cell surface receptors (328-331). Because of the central role believed to be played by IL-2 in regulating T cell responses, and in view of the immunoregulatory abnormalities of SLE-prone mice, several investigators have performed detailed analyses of IL-2 production and consumption by T cells of autoimmune mice and their relationship to the disease process. Studies by Altman and associates (332) demonstrated that age-dependent, reduced Con A-induced mitogenic response and IL-2 production are coinmon features of all SLE strains. This defect appears at 3-6 weeks of age in the early, severe SLE-developing MRL/l and inale BXSB strains and progresses thereafter. The most severe abnormality in IL-2 production is found in MRL/l mice. Similar defects appear at a later stage in NZ, MRL/n, and female BXSB mice. Detailed analysis of cells from the enlarged lymph nodes and spleens of older MRL/l mice demonstrate that such cells (1) respond poorly to Con A or allogeneic stimulator cells, even in the presence of exogenous IL-2, (2) do not suppress IL-2 production by normal spleen cells,

32 1

MURINE MO DELS OF SLE

PRODUC:TION O F

Strain of origin of B cells cultured with LPS" ~~

A

TABLE VII B CELL DIFFERENTIATION FACTOH T C~1.i.a OF MRL/l MICE

BY TIlE h O L I F E H A T I N C :

Isotypes secreted (8 of total IgG) Supernatant added

IgM (ng/ml)

IgC" (ng/niI)

IgCl

IgGB,

IgG21,

IgGB

-

230 603

33 277

6.1 18.5

18.8 25.6

25.0 49.3

50.1 6.6

613 1211

55 560

7.3 16.1

45.4

30.9 27.5

16.4 2.8

~

BALB/c

MRL/l SNc MRLll

-

MRWl S N

53.6

+

OAnti-Thy-1.2 C-treated spleen cells were cultured at a concentration of 5 X 104 cells/well for 6 days in the presence of 2 . 5 kg/ml of LPS. "Total IgG is the s u m of the values obtained with the four IgC subclasses. IgM and IgG subclasses were determined by radioimmunoassays. CMRL/l lymph node supernatant (SN) was prepared by culturing the cells from 4-month-old mice (2x106 cells/crn2 per 0.62 in1 of medium) for 24 hours. Such supernatants were added at 40% (dv).

and (3) are relatively incapable of adsorbing or inactivating exogenously added IL-2. These results indicate that T cells of MRL/I mice are severely defective in their responses to mitogenic stimuli, IL-2 production, and IL-2 receptor site expression. Similar results were reported independently by Dauphinee et a2. (333)and extended to other lpr homozygous mice, such as the C57BL/6J-lpr, by Wofsy et a2. (334). These investigators also found that the defect in Con A-induced IL-2 production is not corrected by the presence of IL-1, a macrophage product, or its presumed comitogen analog, phorbol myristic acetate (PMA). Santoro et al. (335,336)also failed to correct the defect by the addition of IL-1, but, in contrast to the studies of Wofsy, they reported restoration of the IL-2 defect by Con A + PMA; presumably PMA renders the T cells more receptive to Con A by extending the G, phase of the cell cycle and reduces IL-2 adsorption by retarding the S phase. It is notable that in all of the above studies, the IL-2 decrease could be detected prior to the massive T cell proliferation of Ipr homozygous mice, suggesting that this abnormality (1)may involve T cells phenotypically different from those causing lymphoproliferation, and (2) the IL-2 deficiency observed in bulk cultures is not caused by dilution of normal IL-2-producing cells with abnormal, proliferating T cells. It is of interest that T cell proliferation occurs in such mice despite the significantly decreased IL-2 production. As hypothesized by Tala1 and associates (333,334), possibly controlled Lyt-1 T cell proliferation requires IL-2-dependent T cell differentiation, i.e., uncontrolled proliferation of Lyt-l+ cells might arise as a consequence +

322

ARGYWIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

of the IL-2 deficiency, Alternatively, T cell proliferation in MRL/l mice may be IL-2 independent. The above conclusions and the postulated involvement of an IL-2 defect in murine lupus have recently been challenged by Hefeneider et aZ. (337) and Simon et al. (338)who, by limiting dilution studies, found that the number of IL-2 producer cells, Con A responder cells, and amount of IL-2 secreted per cell in MRL/l mice are within normal limits or even increased. These authors attributed the paucity of IL-2 production in bulk cultures to the dilution of normal T cells by the proliferating T cells. However, since Altman et al. (332), Dauphinee et aZ. (333)and Davidson et aZ. (242)found that the IL-2 defect of MRL/l mice is detectable early in life prior to lyinphoproliferation, and that such a defect also appears in other lupus strains that are not characterized by lymphoproliferation, one would be inclined to believe that the defect is real. Further studies with more accurate assays of the amount of IL-2 produced and numbers of acceptor sites for this molecule on lymphoid cells from autoimmune and normal animals may provide more conclusive answers. The causes of decreased IL-2 production and response in lupus mice are unknown, but they do not appear to involve increased suppressor T cell function or inefficient IL-1 production. Impaired cell maturation or occupation of IL-2 receptors by passive in uivo absorption of IL-2 has not been excluded. The relationship of the IL-2 defect to the disease process also remains unclear. Recent studies by Bocchieri et al. (339) in lines of recombinant NZB x C58 mice suggest that low IL-2 levels are not necessarily associated with high autoantibody levels. It should be noted that studies in humans with SLE or RA also found defective IL-2 production (340,342), and one study showed defective IL-1 production (341). However, such low IL-2 levels are not unique to autoimmune diseases, but have also been observed in other conditions such as aging (343,344), primary and aquired immunodeficiency syndromes (345,346), parasitic infections (347), and other disorders. The availability of (1)cloned human (348-351) and, expected in the near future, murine IL-2 as well as (2) the recent development of monoclonal antibodies to conventional IL-2, recombinant IL-2, and chemically synthesized IL-2 (352,353) necessary for purification of large quantities of this lymphokine may allow the precise definition in uitro and in uivo of the defects observed and of the potential usefulness of IL-2 in manipulating systemic autoimmune and other diseases. d . Defects in Interleukin 3 (IL-3). T cell-derived IL-3 is a new lymphokine that is distinct biochemically and biologically from IL-2 (354). It is involved in regulation of hematopoietic stem cell development into all of the major blood cell types, and it appears to play a role in promotion of early T and B cell differentiation (355,356).

MURINE MODELS OF SLE

323

Levels of IL-3 or its synonymous colony-stimulating factor (CSF) were recently assessed in murine strains with lupus. The limiting dilution studies of Hefeneider et al. (337) indicate that the number of CSF producer cells present in spleens of MRL/l mice is similar to the number found in normal control mice. In addition, CSF producer T cells from MHL/I mice make similar levels of lymphokine activity, as do producer T cells from normal mice. These findings contrast with a recent report by Palacios (357) who examined IL-3 production in MRLA mice and found it to be increased. Spontaneous release of IL-3-like activity is detected, according to this study, in supernatants frotn spleen cells of 6 week old MHL/I mice, and the titers increase with age. The IL-3 producing cells were shown to be Thy- l+, Lyt-1 , and the response of such cells to exogenous IL-3 was within normal limits. Importantly, IL-3-sensitive Thy-1 +,Lyt-1 cell lines from spleens of MRL/l mice induce, in the presence of Con A, small “resting” syngeneic or H-2 compatible normal B cells to differentiate into antibody-producing cells secreting predominantly IgG,, IgG,, and IgA. The author concludes that abnormal production of IL-3 may account for the outgrowth of Thy-1+, Lyt-l+ cells as well as the B cell hyperactivity in the MRL/I mouse. The results of Palacios are somewhat reminiscent of those of Prud’homme et al. (325), although the former investigator identifies the B cell differentiationinducing factor as IL-3. However, such identification should be considered tentative, since (1) as cited above, Hefeneider et al. (337) found normal levels of CSF (synonymous to IL-3) in MRL/I mice, and (2) experiments by Fieser et al. (3374 with a different IL-%dependent line than that of Palacios (DA-1 versus Ea3 cells) as the indicator system do not show spontaneous IL-3 production by spleen and lymph node cells of MRL/l mice. Furthermore, a decrease in IL-3 production over that of MRL/n and normal mice is observed after Con A stimulation in vitro. e . Intederon Abnormalities. Interferon (IFN) is classified into three groups on the basis of its antigenic properties and functions (reviewed in 358): (1) I F N a is produced mainly by leukocytes in response to viral and nonviral stimuli, and is stable at pH 2; there are at least 20 distinct genes for human IFNa, (2) IFNP, synthesized predominantly by fibroblast-like cells and, to a much lesser extent, by leukocytes, is also acid-stable, and (3)IFNy, or “immune IFN,” is released by lymphocytes following exposure to mitogens or specific antigens. IFNy is inactivated by incubation at pH 2 and is generally more heat-labile than than IFNa or IFNP. Both I F N a and IFNy may be involved in the regulation of immune responses in uiuo. IFNy in mice has recently been shown to be identical to macrophage activating factor (MAF), and perhaps to macrophage Ia inducing/recruiting factor (MIRF) (359,360). Mouse IL-2 appears to be required in the induction of IFNy secretion by predominantly Lyt-2 T cells (361,362). Furthermore, cloned +

+

+

324

ARGYHIOS N . THEOFILOPOULOS A N D FRANK J. D I X O N

murine IFNy was recently shown to substitute for a late-acting helper factor, and synergize with other helper factors in the stimulation of B cell antibody responses in vitro (363,364). Studies of Hooks et al. (365,366) demonstrated high levels of undefined species of IFN in SLE patients, particularly those with active disease. These findings were, in part, confirmed by Preble et al. (367), who also found elevated IFN levels in SLE patients, but no correlation with clinical activity. Preble further identified this IFN as IFNa, although it was acid labile like the immune IFNy. A deficient in vitro production by peripheral blood mononuclear cells of SLE patients in response to several IFN inducers, and an increase in IFN-induced 2’-5’-adenylate synthetase in these cells was also identified (368). With regard to murine SLE, Santoro et a!. (369) reported normal or increased levels of IFNy in Con A-stimulated MHL/l lymphocytes despite the reduced IL-2 levels in similar cultures. Unlike normal mice, wherein most of the IFNy is produced by Lyt-2+ cells, IFNy in MRL/l mice is produced mainly by Lyt-1 cells. In contrast to these findings, the studies of Kofler et a2. (370) demonstrated that Con A-stimulated lymphocytes from all early-life lupus developing strains produce lower levels of IFNy than normal controls, and coordinate with reduced levels of MAF, MIRF, and IL-2. The significance of these findings remain unknown. It should be noted, however, that NZ mice treated with IFNy inducers (371,372) or IFNy (373,374) itself show accelerated disease manifested by earlier mortality and glomerulonephritis. +

3. Macrophage Defects Essential in the cellular and molecular events that underlie immune competence is the mononuclear phagocyte, which processes and presents antigen to lymphocytes and generates interleukins, such as IL-1, that influence lymphocyte activity. Moreover, the phagocytic function is important in disposal of immunologically undesirable materials such as ICs. Surprisingly, relatively little information is available on the numbers and function of mononuclear phagocytic cells in murine SLE. A recent study by Wofsy et al. (375)found a dramatic, progressive increase in peripheral blood monocytes in BXSB males, beginning as early as 2 months of age, and accounting for 50-90% of peripheral blood mononuclear cells by 6 months of age. Although these cells were designated as monocytes by virtue of the presence of the Mac-1 antigen, Fc receptors, and other morphologic characteristics of monocytes, they were atypical since they lacked Ia antigens as well as nonspecific esterase and myeloperoxidase. Thus, their characterization as monocytes is tentative. Such so-called monocytosis has not been observed in other SLE strains. The significance of this finding in the BXSB

M U R I N E M O D E L S OF S L E

325

disease is unknown, but studies on individual BXSB male mice (L. M. Hang, unpublished observations) have failed to correlate the degree of “monocytosis” and disease severity. Most reports on the phagocytic activity of NZB-derived macrophages have shown heightened phagocytosis of antigen (376,377), although some data suggest that the NZB cells are relatively unable to degrade the ingested antigen (377). Studies on the clearance of inert particles or ICs by macrophages of NZ mice are inconclusive; some show a reduction in clearance (378), others portray clearance as increased or normal (379,380), whereas still others claim normal clearance but weak binding affinity for Fc receptors of Kupffer cells and therefore easy rerelease into the circulation (381). Decreased in v i m clearance of antibody-sensitized red blood cells has been described in human autoimmune disorders as well (224,382). It was initially thought that this defect is caused by a primary defect in the number or function of receptors for Fc, or secondarily by occupation of the receptors by circulating ICs (224). Recent studies, however, indicate that the numbers of IgG Fc receptors on mononuclear cells of lupus and rheumatoid arthritis patients are normal or increased (383,384), and that there is no correlation between levels of circulating ICs and degree of clearance defect (383,385). As stated above, most T lymphocyte activities require that macrophages take up and present the antigen in the context of a MHC-I region product on the cell surface. Macrophages from neonatal normal mice, tested at an age when immune responsiveness is low and tolerance is easily induced (386, 387), present antigen poorly (388).This defect has been correlated with the small number of macrophages that bear Ia antigens in spleens of neonatal mice, compared with adult mice (389). In lupus mice, ontogenic studies by Theofilopoulos et al. (140) demonstrated that Ia mononuclear cells appear in the spleens at the same time as in normal mice. Moreover, the frequencies and absolute numbers of Ia+ macrophages early in life (1month) is the same in both groups. However, the studies of Lu and Unanue (390), Kelley and Roths (391), and Kofler et al. (370) have shown an increased frequency of Ia+ resident peritoneal macrophages later in the lives of MRL/l (70% IA+ by 5 months) and NZB (50% IA+ by 10 months), but not in NZBxW or BXSB lupus mice (approximately 5-10% IA+). Whether the amount of Ia expressed per cell is also increased is not clear. Since Lu and Unanue were able to induce Ia+ macrophages in normal mice upon repeated injections of supernatants derived from non-mitogen-stimulated MRL/I spleen (but not lymph node) cells, these authors conclude that Ia+ cell recruitment was induced by a MIRF (1FN)-like substance secreted by MRL/I proliferating T cells (390). However, Kelley and Roths (391) and Kofler et al. (370) observed that CSH-lpr and C57BL/6-lpr do not show the age-related increases in Ia+ macrophages despite the presence of Ipr-driven lymphoid hyperplasia. Fur+

326

ARGYRIOS N . THEOFILOPOULOS A N D F U N K J. DIXON

thermore, Kofler et al. (370) indicated that the increase in IA+ -1E resident peritoneal macrophages in NZB and MRL/I mice was not caused by increased spontaneous or Con A-induced production of or response to exogenous cloned T cell derived Ia-recruiting lymphokines. These authors found that no MIRF, MAF, or IFN (all perhaps representing one type of lymphokine but assessed with different assay systems) is elaborated spontaneously by splenocytes of lupus mice. Although such lymphokines can be induced upon culturing in the presence of Con A, cells from lupus mice produce no more of such factors than those from normals and, in fact, splenocytes from lupus mice with early-onset disease (MRL/l, female NZBXW, male BXSB) elaborate lower levels than late-life lupus and nonautoimmune mice. In addition, no increased response of NZB and MRL/l macrophages to exogenous cloned T cell-derived Ia recruiting lymphokines has been observed. MRL/n and (MRL/lX MRL/n)F, mice without homozygous Zpr genes do not show an increased frequency of Ia+ macrophages with age, indicating that MRLll mice need the Zpr gene to express increase in Ia+ macrophages. However, as stated above, C3H/HeJ and C57BL/6 mice of normal background made homozygous for the lpr gene have no such increases. Furthermore, neonatal thymectomy of MRL/l mice, which inhibits lyinphoproliferation and retards disease expression, does not diminish the increased frequency of Ia macrophages. To summarize, the age-related increase in Ia+ peritoneal macrophages in lupus mice is not a universal characteristic nor an absolute requirement for lupus. Furthermore, phenotypic expression of the lpr gene by lymphoid hyperplasia is not required, and the effect of the Zpr gene in this phenomenon is dependent on or modified by the genetic background on which it acts. The significance of this phenomenon in the disease pathogenesis of NZB and MRL/l mice remains unclear, although a hypothesis is advanced by Rosenberg et al. (392) that overproduction of interferon and increased T cell response to Ia, as well as enhanced Ia expression, may be responsible for MRL/l lymphoproliferation. The evidence reviewed above, however, does not support this hypothesis. Humoral immune responses of MRL/1 mice to TD, but not to TI, antigens have been found greatly diminished with age (305,393). However, this does not appear to be caused by defects in antigen-presenting cells, since MRL/lpresenting cells mixed with normal primed T cells give a brisk response. In contrast, no significant response is seen when normal antigen presenting cells and MRL/I splenocytes are mixed (393a). The results indicate abnormalities in antigen-responding T cells of MRL/I mice or, more likely, dilution of few antigen-specific T cells by the large numbers of the abnormal Lyt-1 proliferating cells. A close proximity between antigen-presenting (macrophages, B cells) and antigen-responding T cells is clearly required for efficient TD antigen responses in vitro (394). +

+

+

MUHINE MO DELS OF SLE

327

4 . Nnttrral Killer Cells It has been proposed that natural killer (NK) cells play important roles in iininune surveillance mechanisms, such as tumor cell killing, bone marrow transplant rejection, and GvH disease. Furthermore, they appear to regulate cellular development, to produce and effect the action of IFN, to inhibit microbial infections, and, inore recently, to regulate huinoral iininune responses (395,396).Because of these diverse functions, the status of N K cells has been investigated in both human and rnurine SLE, but the findings are inconclusive. In general, NK activity has been reported to be depressed in humans with SLE (397-401) and shows somewhat reduced responsiveness to IFN (399-403). This impaired NK cell function does not appear to be caused by cell-mediated suppressor mechanisms or deletion of effector cells, but rather by an impaired release of a soluble cytotoxic factor (399). With regard to inurine lupus, the studies of Croker et nl. (404) with NZB spleen cells showed that NK activity is within the same range observed in several normal strains. On the other hand, based on increased bone inarrow resistance of NZBx W mice to parental cells (405), heightened NK activity in such mice has been postulated. Treatment with %r, which decreases NK activity b y its localization in bone marrow, was reported to reduce anti-DNA levels in NZB x W mice (406); despite this reduction, however, the H9Srtreated mice died earlier than the untreated controls, in large part due to the appearance of poorly differentiated sarcomata.

5 . Other Cellular and Humord Abnormalities Additional cellular and huinoral abnorinalities have been described in lupus strains of mice, the significance of which is not yet clear. Several reports have described an age-dependent decline in certain T cell functions of NZ mice such as allograft rejection, GvH reactions, and killer cell activity against allogeneic cells (407-409). A more recent comprehensive study, however, by Zinkernagel and Dixon (410) on the age-dependent capacity of NZ mice to generate T cell-mediated immune responses against alloantigens and virus-infected target cells, as well as immune protection against intracellular parasites, failed to disclose any significant defects. Botzenhardt et al. (411) described the development of a significant unidirectional primary in vitro T cell-mediated lyinpholytic reaction by lymphoid cells of very young NZB mice against H-2 identical allogeneic cells. Most other strains of mice make detectable cytolytic responses to H-2 compatible cells only after prior in uioo priming. Studies by Theofilopoulos et al. (412)confirm this finding, but fail to reveal such abnorinal behavior in the other lupus strains. Rich et nl. (413)and Fischer-Lindahl and Hausmann (414) presented evidence that the antigens recognized by NZB mice are

328

ARGYHIOS N . THEOFILOPOULOS AND FRANK J. DIXON

Qa-l”-coded determinants, which are expressed in all H-2d strains typed for Qa-1 thus far except NZB (which carry the Qa-la allelle). Concurrent and subsequent studies, however, by Theofilopoulos et al. (415), Davidson et al. (416), Stockinger and Botzenhardt (417), and Smith et a1. (14) show that this abnormal cytolytic response is not solely dependent on antigens coded for by genes to the right of H-2D, such as Qa-1, but might involve other antigens, including a maternally transmitted antigen (11,14) and retroviral gp70 (415). This enhanced primary cytolytic response of NZB cells against H-2 compatible strains as well as enhanced secondary cytolytic responses studied by Huston and Steinberg (418,419) suggest that accelerated proliferation and differentiation, which characterize NZB B cells, also apply to NZB T cells. This T cell abnormality is determined by different sets of genes from those governing abnormalities of B lymphocytes, despite the fact that both develop in parallel (416). Experimental evidence suggests that the increased cytolytic responses of NZB cells are caused by an unusually high helper potential, and not suppressor T cell defects (417). The contribution of this T cell hyperactivity to NZB disease is unknown. It appears, however, that this abnomrality is not a general requirement nor a common etiologic factor in the development of murine SLE, since MRL and BXSB mice are devoid of such activity (412,415). Cytolytic responses to alloantigens and viral antigens have also been examined in MRL/I mice. Proliferative in uitro allogeneic responses of MRL/l spleen and lymph node cells are diminished in older animals, and this defect is not corrected by the addition of exogenous IL-2 (332). In contrast, a dichotomy of primary cytolytic responses against allogeneic cells between MRL/I spleen and lymph node cells has been observed. Thus, spleen cells of young and aged MRL/I mice are efficient effectors, whereas those from lymph nodes of older animals were defective (332,412). Limiting dilution experiments by Simon et al. (338)showed a normal CTL precursor frequency of cytotoxic T lymphocytes in spleens, but a greatly diminished frequency in lymph nodes. However, because of the enormous cellular increase in the lymphoid compartment, the total absolute numbers of cytolytic precursor cells in lymph nodes of aging MRL/l mice is higher (up to 30-fold) than in normal mice. A simultaneous determination of the frequencies of IL-2 precursor cells, cytolytic precursors, and proliferating precursors in lymph nodes of 4-month-old MRL/l mice showed frequencies of 1/16500, 1/4085, and 1/900, respectively, i. e., reduced frequencies but increased absolute numbers. In contrast to this conclusion, Scott et al. (420) showed broad defects in cell-mediated responses of spleen cells from older MRL/l mice, including the inability to generate primary allospecific and hapten-specific cytolytic T lymphocytes, or secondary hapten- and virus-specific cytotoxic T cells. Additionally, proliferative responses to hapten and natural antigens were deficient and the delayed-type hypersensitivity response was weak.

M U H I N E MODELS OF SLE

329

Cell-mediated responses of BXSB inale mice have been examined by Creighton et al. (421); quantification of cytotoxic T cell respoiises to alloantigens and viruses (lyinphocytic choriomeningitis, vaccinia) shows no difference in the kinetics of appearance or relative activity of cytotoxic T cells per spleen of these inice compared to control mice. Furthermore, induction of lymphocytes iininune to Listeria inonocytogenes is only slightly increased. The autologous inixed lyinphocyte reaction (MLR), i.e., proliferation of T lymphocytes upon coculturing with autologous or syngeneic non-T cells, is decreased in both human (422) and niurine (423-426) SLE, but no cellular basis for the defect is known. In human SLE, abnormalities of the stimulating cells (self B cells or rnacrophages expressing Ia) or of both responding and stiinulator cells have been reported, whereas in SLE mice, the defect has been placed primarily on the responder cell population. Yet, the studies of Bocchieri et al. (426)with (NZBX C58) recombinant inbred lines fail to correlate the autologous MLR deficiency with autoantibody production. Kinetically, immune responses to TI-class 2 antigens tested by Park et al. (305) seem to be defective in MRL and NZB but not NZBxW or BXSB mice, in that the IgG responses do not decrease at the same rates as normal strains. Instead, the responses are sustained or increased with time. Since N Z B x W inice behave normally in this respect, the authors conclude that the defect of the parental NZB strain is recessive. Moreover, since MRL/n mice exhibit the defect, albeit to a lesser extent than the congenic MRL/I mice, it was concluded that the lpr gene is not responsible. Similar prolonged immune responses in NZB inice have been observed by others (299). Studies by Cowdery and Steinberg (306) suggest that the defect in NZB and MRL/l mice is refractoriness of target cells to classical antibody-mediated suppression, not differences in the quality of antibody produced coinpared to normal mice. In contrast, studies by Goidl (427) in NZB inice imply that the cause is reduced autoantiidiotypic antibody levels, which normally down-regulate immune responses. Another possibility is the presence in such mice of increased numbers of spontaneously primed B cells, which are not as easily suppressed by antibody as virgin B cells (428). Lack of suppressor T cells cannot be involved, if one considers that suppression of responses to TI-2 antigens has been found to be T-independent (429). Whatever the mechanism, such a defect in down-regulation of iminune responses may be an important contributor to the development of autoiininunity in some SLE strains. Park et al. (305)found that although most SLE strains show higher magnitudes of responses to exogenous antigens than normal mice, the avidities of antibodies produced are not significantly different from normal. Similar findings have been reported by McKearn et nl. (299) and Naor ct al. (304). However, Goidl et al. (430)observed restricted heterogeneity of avidities of antibody to TNP produced by SLE strains compared to normals. The re-

330

ARGYRIOS N . THEOFILOPOULOS A N D FRANK

1. DIXON

sults, as a whole, fail to support the notion that murine SLE results from polyclonal expansion of low-avidity receptor-bearing B cells that might escape tolerance, or from the presence of low-avidity products of inherited germ line V genes. D. DEFECTSIN TOLERANCE Spontaneous autoantibody development might result from resistance to, termination of, or breakdown of self-tolerance. Studies have been designed to examine the ability of SLE mice to be tolerized to a variety of exogenous antigens, particularly deaggregated heterologous y-globulins. Initial studies by Staples and Tala1 (431) demonstrated that adult NZ mice are relatively resistant to tolerance induction by deaggregated human yglobulin (HGG) as shown by immune elimination and hemagglutination assays. However, NZ mice under 3 weeks of age are easily tolerized, although this tolerance is rapidly lost. Subsequent studies (19,432,433) demonstrated defects in the induction of tolerance to heterologous y-globulins not only in young NZ mice but also in male BXSB and, to a lesser extent, in MRL/I mice; female BXSB and MRL/n mice are normal. Of the strains not generally developing the lupus-like disease, SJL, DDD, and BALB/c strains also show relative resistance to tolerance induction (19,434,435). Macrophage hyperphagocytic activity might account for the failure of deaggregated y-globulins to induce tolerance in SLE mice. BALB/c and SJL mice, as noted above, are relatively resistant to tolerance induction with ultracentrifuged HGG preparations, but such mice can be rendered tolerant if challenged with biofiltered HGG, a process which removes small amounts of aggregates from the centrifuged preparation (435,436).These experiments suggest that BALB/c and SJL mice are hyperphagocytic and are immunized by trace amounts of microaggregated IgG in certain preparations of ultracentrifuged IgG. This finding is consistent with the work of Parks et al. (437), who found that tolerance to HGG is dependent on the preparation and degree of HGG deaggregation as well as HGG sources. However, biofiltered HGG still could not induce tolerance in SLE-prone mice, ruling out the possibility that microaggregates in ultracentrifuged HGG interfered with tolerance induction (19,436). The cellular basis for this defect in tolerance induction to a T D antigen such as y-globulins has been examined by several investigators. The conclusion from the studies of Amagai and Cinader (432), and of Laskin et a2. (433,438), in NZB mice is that the primary defect in tolerance induction is at the pre-T or T cell level. The NZB T cells manifest this defect independent of maturation in the NZB or normal thymic microenvironment (438), and they can actively interfere with the expression of tolerance by normal cells (439). The transplantation studies of Hang et al. (440)have also suggested that the cellular basis for resistance to induction of tolerance to HGG in BXSB male mice rests on a bone marrow stem cell

MURINE MODELS OF SLE

33 1

population that has been depleted of differentiated T cells by anti-Thy-1.2 + complement treatment. A inacropliage-independent defect in tolerance induction at the B cell level in all lupus strains of mice has also been described by Golding et al. (441,442). Interestingly, these investigators note that low epitope density conjugates of HGG with TNP must be used for the tolerance defect to become apparent, indicating that resistance to tolerogenesis is by no means absolute. They observed that all lupus strains, like normal mice, can easily be tolerized with TNP,,HGG and TNP,,HGG. However, when a tolerogen with a lower epitope density is used (TNP,HGG), several control strains are all rendered tolerant; NZ B cells are resistant to all concentrations of TNP,HGG tested, whereas B cells from BXSB male and MKL/l mice are resistant to low concentrations of this tolerogen. These investigators suggest that resistance to B cell tolerance may be a consequence of prior polyclonal activation, and that a similar loss in B cell susceptibility to tolerance induced to self-antigens with low epitope density may lead to autoantibody formation and disease. It should be noted that B cells of newborn NZ mice, unlike those of normal or other lupus mice, were found by Theofilopoulos et al. (140) to regenerate surface Ig within short periods of time following anti-Ig modulation (the universal antigen). Inability to regenerate surface Ig by normal neonatal B cells after interaction with anti-Ig (antigen?) has been proposed as a means by which tolerance to self is induced at early stages of B cell ontogeny (443). Some investigators failed to observe any profound defect of tolerance induction in NZB mice. For example, Pike et ul. (272) found only a marginal difference in B cell tolerance induction in oitro to fluorescein conjugated HGG (FLU,HGG) and no difference from normals to FLU,HGG. Furthermore, Purves and Playfair (444) record identical levels at which tolerance develops to TI antigens (such as pneumococcal polysaccharide type 3 and bacterial levan) in NZB inice and in nornial strains. Moreover, although NZB inice fail to show tolerance at the level of priinary response after antigen feeding, they are normally tolerant when a secondary response to a lower dose of antigen is evaluated (445), and MRL/I and BXSB mice are nearly normal in induction of tolerance by the intragastric route (446). As in NZ mice, variant studies with MRL/I mice have also appeared. Thus, Santoro et al. (447) found that tolerance induced to haptcn modified syngeneic splenocytes was within normal limits in young and old MKL/l mice. The significance of this rather uniform defect of lupus mice in disease pathogenesis is unknown at present, nor do we know whether it is primary or secondary in nature, or applies to all or only a few antigens with unusual characteristics (i.e., IgG for which not only antigen receptor binding but also Fc receptor binding occurs). Studies of Dixon and associates (111) on F,

332

ARGYRIOS N . THEOFILOPOULOS A N D F R A N K J. DIXON

generations of lupus mice show no predictive value or correlation with mortality, autoantibody levels, and defects in tolerance induction to HGG. Recent studies (447a) clearly indicate that the defect in tolerance induction to heterologous IgG is not required for the development of an SLE-like syndrome, and the induction or enhanced production of autoantibodies by the Ipr gene is not related to this cellular abnormality E. THYMICDEFECTS

By virtue of its epithelial microenvironment, its giant nursing cells, and its hormone-like substances, the thymus is essential for T cell differentiation and helper, suppressor, and cytotoxic subsets. Therefore, the role of the thyinus in the pathogenesis of autoimmune diseases such as murine SLE has been scrutinized by several investigators, with four main areas of research in mind: (I) thymic histopathologic characteristics, (2) state of the thymic microenvironment, (3) thymic hormone levels, and (4) presence or absence of the thymus. In regard to histopathologic characteristics, as discussed in Section 111,B,2, all lupus strains develop early severe thymic atrophy, particularly involving the cortex and, to a lesser extent, the medulla. The means by which this atrophy occurs and its causal relationship to the disease process are unknown at present, but NTA has been considered a contributor (166, 167). There are, however, several immunologically normal strains of mice that exhibit incidences and titers of NTA similar to lupus mice, but without early thymic involution (172). Whether this early thymic atrophy is of primary importance in the immune dysregulation associated with lupus, or is a secondary epiphenomenon of an early aging process in lupus mice, is unknown. In aged humans, some theorize that age-associated thymic atrophy is responsible for senescence of the immune system and subsequent development of autoantibodies and tumors at high incidences (448). Is there any unique intrinsic abnormality in the thymic microenvironment of lupus mice that might contribute to the disease process? Histologic observations and culturing of thymic epithelial cells suggest that such intrinsic abnormalities exist, but thymic transplantation experiments do not support this conclusion. Initially, Holmes and Burnet (449) and de Vries and Hijmans (450,451), who performed comparative histologic examinations of thymuses from NZ and normal mice, noticed significant early changes in the large medullary epithelial cells involved in thymic hormone secretion, and formation of Hassal’s corpuscles in the NZ strains. In the NZB mice, the large epithelial cells severely decrease in number in the first weeks after birth. In contrast, extensive hyperplasia of the large epithelial cells and Hassall’s corpuscles is observed in the NZW and NZB x W F, mice, apparent even in newborn animals. Many of the epithelial aggregates had been invaded by lymphoid cells, and both of these cell types show a variety of degener-

333

M U R I N E M O D E L S OF SLE

ative changes. Depletion of epithelial cells occurs at approximately 8 months for NZBxW, and at 1 year for NZW. Neither epithelial hyperplasia nor depletion of epithelial cells is obvious in normal strains. Subsequently, based on experiments with thymic explants of NZ mice in uitro, Gershwin et al. (32)concluded that there is a preprogrammed intrinsic defect of NZ thymus epithelial cells. Although thymic epithelial cells of normal mice proliferate in vitro, explants of NZB thymus age 1 to 52 weeks fail to produce growth ofany cellular elements except fibroblasts. In contrast to normal mice in which long-term cultures of epithelial cells can be established, the epithelial growth of NZBXW mice is transient, and degenerates within a few weeks. Similarly, whereas isolated thymic epithelial cells from BALB/c or young NZBxW mice induce elevated amounts of Thy-1.2 on spleen cells and increase the responsiveness of nude spleen cells to the T cell mitogens, cultures from older NZB or NZB X W mice are ineffective. On the basis of these findings, Gershwin et aE. suggest that such intrinsic abnormalities and functional alterations of thymic epithelial cells in NZ mice may be critical to the development of autoimmunity. Thymic transplants between late-life and early-life disease lupus substrains, as well as between lupus and normal mice, have also been performed to determine whether the thymus of a lupus-prone mouse can confer disease. Theofilopoulos et al. (452) demonstrated that the thymic genotype is irrelevant to the pace and characteristics of disease in the two MRL substrains (MRL/I, early-disease substrain; MRL/n, late-disease substrain) (Table VIII). Thus, MRL/l mice thymectomized when 1 day old and transplanted at 1 month of age with MRL/n thymuses retained the disease phenotype of unmanipulated MRL/l mice, including lymphoproliferation and a 50% mortality at 5-6 months of age. MRL/n mice are similarly unchanged TABLE VIII

RELATIONSIITPI3ETWEEN TIIYMIC ORIGIN

AN13 SLTRVIVAl., D E C R E E O F LYMPIIOPHOLIFERATION,

A N D SEROLOGIC: C1IAKAC:TERISTICS O F

MRL MIC:E

Number of cells

Groups

MRL/1 (unmanipulated) MRWn (unmanipulated) MRWn thymus + M R U l Tx” MRWl thymus + MRWn Txb MRWl Tx (no implant)

5 0 8 survival (days)

160 510 186 498 >390

MLNa Spleen ( x 10-6)

679 ND 590 NII 13

287 81 310 110 109

IgG (inglml)

Anti-ss 11N A (% binding)

31.0 13.4 30.7 10.8 11.6

77.8 51.7 82.2 54.6 23.6

“ M L N , Mesenteric lymph node ”Animals were neonatally thymectomized and transplanted with thymuses at I month of age

334

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. D I X O N

by thyinectoiny and transplantation with MHL/I thymuses; these mice do not develop lymphoid hyperplasia, and 50% remain alive at 17 months of age, which is not appreciably different froin unrnanipulated controls. Therefore, differentiation of MRL/n stem cells under the hormonal or microenvironmeiital influences of a thymus possessing the lpr genotype does not lead to abnormal T cell differentiation or early autoimmunity. In other experiments related to bovine y-globulin (BGG) tolerance resistance in NZB mice, Laskin et al. (438) demonstrated that this defect, caused by an abnormal pre-T cell, can be expressed in both NZB or normal thymic microenvironments. Conversely, normal pre-T cells that have matured in an NZB thymic microenvironment show little change in their ability to be tolerized. In contrast to these recent findings, which imply irrelevance of NZ thymus in the expression of autoimmunity-related abnormalities, early studies by Helyer and Howie (21,453)showed that normal nonautoimmune mice, when neonatally thymectoinized and immediately grafted with thymus glands from newborn NZB or N Z B x W hybrids, developed strongly positive autoimmune markers and renal changes characteristic of lupus nephritis. With regard to thymic hormonal defects, initial reports indicated that all SLE strains show a premature decline (as early as 1 month of age) in the production of a circulating thymic hormone, thymulin, as assessed by its rapid disappearance from the serum (33,454) and a sharp diminution of the number of thymulin-containing cells in the thymus (455). In addition, thymulin inhibitory molecules (perhaps antithyinulin autoantibodies) are detectable in the sera of NZB X W mice (456). Furthermore, it was claimed that administration of thymic hormones to NZB mice (457) or transplantation of thymuses or of thyinocytes from young NZB mice to older syngeneic mice (458,459) can temporarily prevent some of the immunologic defects, and delay the onset of autoimmunity. However, others have failed to inhibit the disease of autoimmune mice treated repeatedly with thymocytes from young counterparts (460,461). Furthermore, attempts to confirm the therapeutic efficacy of thyinocin (bovine fraction V) in the NZB and NZB x W disease have also failed to disclose any significant differences in autoantibody levels and survival between treated and untreated mice (34,35). Further, the synthetic pentapeptide, thymopoietin, although showing some effect on autoantibody levels, fails to increase survival time in the treated group (462). The need for a thymus (and T cells) in the development of murine lupus has also been examined by several investigators. Helyer and Howie (453) reported that neonatal thymectomy within 24 hours of birth in both NZB and N Z B x W mice does not prevent the onset of disease and that, following this procedure, the autoimmune process develops precociously and with more acute manifestations. They also showed that immediate grafting of a neonatal thymus from a nonautoimmune strain fails to prevent disease. East and

MURINE MODELS OF SLE

335

associates (463) confirmed that neonatal thymectomy does not prevent autoantibody production in NZB mice and appears, in some instances, to accelerate its onset. Holmes and Burnet (449), although confirming that neonatal thymectomy does not prevent Coomb’s conversion in NZB mice, indicated that it induced a 2 to 3 month delay in the onset of the disease. More recent studies by Roubinian et al. (464) showed that neonatal thymectomy (2- or 3day-old mice) has the opposite effect on survival of female and male NZB x W mice; thymectomy significantly prolongs survival of the females, but increases mortality in the males. The thymectomized males have an early and persistent increase in anti-DNA antibodies associated with an accelerated switch to IgG occurring at 4 rather than at 9 months. In contrast to these findings, Steinberg et al. (465) reported that neonatally thymectomized NZBxW female mice show accelerated disease. Similarly, Hang et al. (466) found that neonatal thymectomy has a slight accelerating effect on the NZB x W female disease. With respect to male BXSB mice, Smith et al. (230) reported that neonatal thymectomy (within 48 hours after birth) causes markedly increased yglobulin and autoantibody titers, a dramatic increase in lymphadenopathy, and worsening of renal disease compared to the unmanipulated counterparts. In addition, such mice have a loss of Lyt-2+ cells and an increased proportion of L y - l + , Thy-1.2- (? B) cells. However, the studies of Hang et al. (466) disclose no significant effects of neonatal thymectomy on the earlylife SLE of male BXSB or the late-life SLE of female BXSB mice, i.e., no enhanced lymphoproliferation, accelerated glomerulonephritis, or marked increases in serum IgG and autoantibody levels. The differences in results between these two studies are not easily explained, but one possibility might be differences in the BXSB colonies used. The unmanipulated BXSB male mice used by Smith et al. had a 50% glomerulonephritis-associated mortality at 8 months of age with low incidences of lymphadenopathy (10%). In contrast, the unmanipulated BXSB males used by Hang et al. followed the typical picture described originally by Murphy and Roths (3) with 50% GNassociated mortality at 5 to 6 months and a 40-50% incidence of moderate lymphadenopathy. The development of autoimmunity in neonatally thymectomized NZB x W and BXSB mice implies that they have B cells and precursors with the potential for proliferation and differentiation to autoantibody-secreting cells without much T cell contribution. We should point out, however, that because neonatally thymectomized mice have considerable numbers of residual T cells, some of the accessory signals required for B cell proliferation and differentiation could be generated even in the absence of a thymus. Furthermore, as discussed above (Section IV,C, l,b), activated B cells from NZB x W females and BXSB males, like those of normal mice, cannot prolife-

336

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

rate or differentiate in vitro unless accessory signals from T cells and/or macrophages are present. In contrast to normal B cells, however, B cells of these two lupus strains hyperrespond to accessory signals (271). Based on these findings, the development of autoimmune disease in thymectomized BXSB and NZB x W mice could result from signals provided by residual T cells not influenced by thymectomy and the inherent hyperresponsiveness of the B cells of such mice to these signals. In this regard, it is of considerable interest that lethally irradiated female BXSB mice cotransplanted with female whole or T-enriched spleen cells together with male spleen cells show considerably less disease than females transplanted with male cells only (467). In contrast to differentiated female spleen T cells, neither female bone marrow nor thymic cells can delay or modulate the early male disease in similar cotransfers. These results suggest that the delay of disease in unmanipulated female BXSB mice may be determined by regulatory T cells. With respect to MRL/1 mice, there is a consensus that neonatal thymectomy prevents lymphoproliferation as well as serologic and histologic manifestations of autoimmunity (452,465) (Table VIII). Not yet determined is whether the thymic effect necessary for T cell proliferation is exerted within the thymic microenvironment or extrathymically via thymic hormones, or both. Time-course experiments have shown that the beneficial effects of thymectomy in this lupus substrain of mice are progressively diminished with time elapsed between birth and the time of thymectomy, i.e., little effect is observed if thymectomy is delayed beyond 3 weeks postnatally (465, 466). This finding indicates that, once a sufficient number of abnormal cells has received the thymic influence, they proceed irreversibly to proliferate and induce the autoimmune process. V. Genetics of Murine SLE

Since the several murine strains predisposed to SLE differ in their origins, MHC, Ig allotypes, and other immunogenetic features, it does not seem likely that all SLE is the product of one or a few common genetic determinants. The pathogenesis of SLE, i.e., multiple autoimmune responses, circulating ICs, and IC disease involving particularly the kidneys and blood vessels, generally is similar in all affected strains. However, the assortment of autoimmune responses and disease manifestations vary from one strain to another, suggesting a different genetic background underlying the particular constellation of immunopathologic features in each strain. Although the uncomplicated expression of these genetic backgrounds produces a late-life lupus in each SLE strain, the superimposition of genetically determined accelerating factors can change the disease to an acute early-life form. These accelerating factors vary from strain to strain, and are linked to the Y chro-

M U H I N E MODELS OF SLE

337

mosome in the BXSB mouse, an autosomal recessive lymphoproliferative gene (Ipr)in the MKL mouse, and the female endocrine environment in the NZB X W. This section analyzes the genetic basis of the specific autoimmune traits, the genetic relationship or lack thereof among these traits, and their pathogenic significance. Also considered is a comparison of the immunogenetic backgrounds of the several lupus mice and the nature and mode of operation of their accelerating factors.

A. INHERITANCE OF THE AUTOIMMUNE TRAITS Since most autoantibodies observed in lupus mice can also be observed in normal mice at late age or after mitogenic or antigenic stimulation, it is important to indicate that genetic defects may be regulatory rather than structural. In regard to anti-DNA antibodies, the studies of Raveche et al. (468) with (NZB x DBA/2)Fl hybrids and backcrosses to both parental strains suggest that NZB mice contribute only a single dominant gene to the spontaneous production of anti-ssDNA, the quantity of which appears to be under additional control either by a gene dosage effect or, more likely, by some kind of regulatory gene or genes. However, studies by Yoshida et al. (469) on (NZB x W) x NZW backcrosses indicate that the production of anti-ssDNA antibodies is under multifactorial control, and that two or three independent dominant genetic systems of NZB mice are likely to be involved. These investigators further indicate that the occurrence of anti-dsDNA is determined by the interaction of two dominant unlinked genes from NZB mice (Ah-1, Ads-2) and that the amount of antibody produced and conversion from IgM to IgG are influenced by two dominant genes (Ah-3, Ads-4) from NZW mice (cited in 470,471). The genes of NZB mice involved in the appearance of anti-ss- and anti-dsDNA antibodies seem to be the same (469). With respect to the inheritance of anti-gp7O autoantibodies and of related gp70-anti-gp70 complexes, the studies of Nakai et al. (148) and Maruyama et al. (472) with NZB, NZW, the F, hybrid and the F,xNZW backcross mice have suggested that NZB mice contribute a single dominant gene (Agp-1) to the appearance of this type of IC, and that additional unlinked or very loosely linked dominant NZB and NZW gene(s) (Agp-2, Agp-3) operate with it to control the degree of the anti-gp70 response. In regard to NTA, genetic studies by two groups of investigators (468,469) using appropriate backcrosses and recombinant NZB inice revealed that the spontaneous appearance of this autoantibody is controlled by a single dominant gene or a cluster of closely linked genes. However, since the F, hybrids of NZB have low titers and relatively late onset of NTA compared to NZB, it was suggested that a gene dosage effect or another genetic locus or loci

338

ARGYRIOS N . THEOFILOPOULOS A N D F R A N K J . DIXON

modifies NTA production levels. The genes controlling the presence and quantity of NTA and anti-ssDNA are not linked to each other (468). With respect to antierythrocyte autoantibody, an incidence of 100% has been observed in NZB mice and around 7040% in F,s of NZB mice with other New Zealand strains (NZW, NZC) (20). However, in crosses of NZB with normal, non-NZ strains of mice, the incidence drops to 0-20%. On the basis of these findings, the following two models of genetic regulation for this autoantibody have been proposed (471). (1) Two genetic loci (Aia-1, Aia-2) are involved in the full expression of antierythrocyte autoantibody production, of which one is dominant and unique to NZB (Ah-l), and the other (Aia-2)is recessive and present in several NZ strains such as NZB, NZW and NZC, but not in other strains. (2) Alternatively, one NZB dominant gene contributes the full expression of Coomb’s positivity, but the effect of this gene can be modified to varying degrees by a second dominant gene present in all other strains except those of the NZ background. Another genetic abnormality, namely the polyclonal B cell activaction of NZB mice expressed by high numbers of Ig-containing and/or -secreting cells, has been suggested to be controlled by at least two independent genetic loci (248). According to this concept, a single dominant gene determines the number of IgM-containing cells, and a second recessive gene regulates the amount of IgM secreted per cell. Finally, the studies of Knight and Adams (473,474) and subsequently of others in (NZB X W) x NZB backcrosses have suggested the involvement of at least three dominant genes or clusters of closely linked genes in the development of renal disease, two of which (Lpn-2, Lpn-3) are contributed by the NZW, and one (Lpn-1) by the NZB mice. Other genes modifying the expression of lupus nephritis have been suspected, since F, offspring of NZB mice with mouse strains other than NZW manifest a later onset and milder form of disease than the NZBxW cross. Such modifying genes may be expressed dominantly in virtually all mouse strains except NZB and NZW. B. RELATIONSHIP AMONG AUTOIMMUNE TRAITS A N D THEIRASSOCIATION WITH DISEASE The possible existence of linkage or interactions among the various abnormal genes expressed in SLE has been investigated by using backcrosses, F, generations and recombinant lupus mice. Studies by Shirai and associates (148,469,471,472,475) of (NZBX W)x NZW backcrosses suggest that the genes for anti-dsDNA, anti-ssDNA, anti-gp7O antibodies, and renal disease are closely linked. They also reported a linkage between these traits and NTA, albeit to a lesser degree. According to their studies, although there is a linkage between amounts of serum IgM and antihapten antibodies, these

MURINE MODELS OF SLE

339

aspects of polyclonal B cell activation do not correlate with other autoimmune traits. Detailed studies with large numbers of F, lupus mice were conducted by Dixon and associates (lll), who examined both the genetic relationship among the various immunologic traits and their association with disease in crosses of BXSB males with feinales of three other lupus strains (MRL, NZW, and NZB). This breeding scheme provides male mice with early acute lupus (50% mortality from 121 to 167 days) and 25-45% female mice with chronic late-life lupus and little or no autoimmune disease in the remainder (50% mortality for all females between 361 and 445 days). F, females derived from the well-studied NZB x W were also examined. Various immunologic parameters were analyzed periodically throughout the lives of these mice. Early rises in levels of serum IgM, IgG, and spontaneous antihapten antibodies were used as indicators of nonspecific immunologic hyperactivity (B cell polyclonal activation). The other factors examined prior to and during the course of clinical disease were ANA, antibody to ss- and dsDNA, NTA, total IC levels (determined by the Raji assay), and gp70 ICs, all of which are considered measures of specific autoimmune responses. It is clear that in the six groups of BXSB derived F, offspring (males and females), the magnitude of IgG, ANA, anti-DNA, and IC (total and gp70-anti-gp70) even at 2 to 3 months of age correlates well with the survival not only of male mice with early acute lupus but also of female mice that develop lupus in the second year of life (Table IX). At 2 to 3 months of age, the NZB x W females reveal no prognostic immunologic indicators, but shortly before the onset of their clinical disease (5 months), all of the above specific immunologic parameters correlated closely with survival. Interestingly, there is little or no prognostic value or disease association with NTA, spontaneous antihapten antibodies, and resistance to tolerance induction measured between 2 to 6 months of age. The association between early IgM levels and survival is good only in NZBXW F, mice, while early increases in IgG levels have predictive value in all F, crosses except NZBXW. In general, in these F, studies, reasonably good correlations are evident among the various autoimmune traits clearly associated with survival (ANA, anti-DNA, anti-gp70) as well as early elevated IgG levels. In contrast, those parameters not well associated with survival (IgM, NTA, antihapten) are not, in general, closely related with each other or with any other tests. These findings as a whole suggest either genetic linkage between or a pleotropic gene action affecting the disease-associated autoimmune traits. Eastcott et al. (476) also performed a genetic analysis of B cell hyperactivity as manifested by numbers of IgM-containing cells and amounts of IgM secreted per cell in NZBXSWR crosses (Fl, F,, and backcrosses) and found that, although the severity and incidence of renal lesions is influenced to some

340

ARGYRIOS N . THEOFILOPOULOS AND F U N K J. DIXON

TABLE IX IMMUNOLOGIC PAHAMETERS: PHEIlICTIVE A N I l NONPHEDICTIVEO F DISEASE I N Fz CROSSES OF SLE MICE Correlation with survival

NZBxBXSB Parameters Predictive Anti-dsDNA Anti-ssDNA

Age (months)

2-3 5 2-3

5 ANA gp70-ICS Total ICs Partially predictive IgC

IRM Nonpredictive NTA Anti-DNP Tolerance resistance

3 6 2-3 6 2-3 4

M

F

M

+') +

+ + + 0 +

+ + + + + + + + t

+

+ + + + 0 + + +

2 6 2 5

+ 0 + +

+ +

4 6

+

2 2-3

NZWxBXSB

0 0 0

0 0

+ + +

0

+ 0 +

0

0

0 0 0 0

0

0 0 0

F

OC

+ + + + + 0 + 0 + 0

+ 0 + 0 0 0

+

MRWlxBXSB M

F

+ + +

+ + + + + + + + + + +

0

+ + + + + + + + 0 + + 0 0 0

0

NZBxNZW F

0

+ 0 + 0 + 0 + 0 + 0 0

+ +

+

0 0 0 0

0 0

0

0 0

OThe ages at which the mice were bled for serologic analyses. 6 + , Correlation with survival at a p value <0.05. "0, No significant correlation.

extent by the presence of B cell hyperactivity, a proportion of F, and backcross progeny mice that do not show IgM hyperproduction eventually develop autoantibodies and autoimmune disease. The relationship of various autoimmune traits has also been studied by using recombinant inbred lines derived from matings of NZB mice with nonautoimmune mice. According to these studies, certain humoral and cellular abnormalities of lupus mice are each inherited as independent traits controlled by unlinked genes. Thus, Raveche et uZ. (179,276) found that NTA, IgM serum levels, anti-ssDNA, antierythrocyte antibodies, splenome-

MUHINE MODELS OF SLE

34 1

galy, hyperdiploidy and increased numbers of cells in S phase, all characteristic of NZB mice, are inherited independently in 27 recombinant lines from crosses of NZB with ALN or N F S normal mice. Bocchieri et al. (178) studied NZB recombinant lines derived from crosses with C58 mice and concluded that the expression of deficiency in an autologous MLR, antierythrocyte autoantibody, high levels of sex-limited protein (Slp) in male NZB mice, and NTA all segregate independently. These recombinants were also studied by Datta et al. (477) who found additionally that IgM hypersecretion and high immune responses (both measures of B cell hyperactivity) as well as increased levels of retrovirus expression also segregate independently. Thus, these investigators conclude that the autoimmune phenotype is not the result of a single “autoimmunity gene” giving rise to the various autoimmune traits, but rather results from faulty regulation of a number of independently segregating genes.

c. COMPLEMENTARITY OF GENETIC BACKGROUNDS AMONG LUPUS MICE A N D

THE

ROLE OF ACCELERATORS A N D OTHERGENES

As discussed above, the one factor that constitutes a convenient phenotypic marker for genetic analysis is the variable age of onset in the different sexes and substrains of lupus mice. This difference in the pace of disease development has been used b y Dixon et al. (19,22,111) to assess complementarity among late-life and early-life lupus backgrounds, as well as the role of accelerators and other autosomal genes in disease expression. As depicted in Table X, when BXSB females (late life disease) are crossed with MRLA male (early-life disease), offspring of both sexes are relatively healthy (50% mortality at approximately 2 years of age), suggesting little complementation between the parental disease-prone genetic backgrounds. Similarly, crosses of male or female MRL/l with NZB, NZW, and other strains of the opposite sex result in either late-life or no SLE, except for the MRLA x NZB cross, which produces females expressing significant disease in the first year of life. One possible explanation for the lack of disease in MRL/I F, hybrids could be that the MRL/l SLE is primarily determined by the homozygous lpr recessive gene, the expression of which is lost in F, hybrid crosses with strains lacking this recessive gene. On the other hand, when female BXSB are crossed with NZB or NZW, the female offspring develop severe SLE and die young from severe glomerulonephritis. The character and timing of SLE in BXSBXNZB or BXS B X N Z W F, are very similar to that in the classic N Z B x W model. Thus, the genetic background yielding late-life SLE in BXSB females can complement or add to the NZB and NZW backgrounds and produce typical SLE in young females. These data indicate recessive trait complementation between BXSB and NZB or NZW strains, similar to that between NZB and

342

AHCYHIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

GENE.rIc:

TABLE X COMPLEMENTATION BETWEEN SLE MICE

F1 crosses Female

50% mortality (months) Male

Female 12 16 17 19 22 >24

MRLI1 MRUl NZB NZW BXSB MRWl

X

X

NZB NZW MRLII MRLII MRUl C57BLl6

NZB

X

NZW

BXSB BXSB NZB NZW MRWl MRWn BXSB BALBlc

X

NZB NZW BXSB BXSB BXSB BXSB BALBIc BXSB

X X X X

X

X X

X X

X X

Male 20

>24 >24 18

>24 >24

9

14

10 9 8 9 16 22 >24 >24

21 16 6 5 6 7 >24 >24

NZW. Whether this contribution is qualitatively the same as that seen in the NZBXW, or only quantitatively similar is not yet known. However, when BXSB females are crossed with normal strains such as BALB/c, the F, offspring develop little or no disease, regardless of sex. It is of interest that when BXSB males (early-life disease) carrying the Ylinked accelerating factor are crossed with any lupus strain, an acute, early disease quite similar to that of the BXSB male is observed in F, males. Thus, the late-life backgrounds of all SLE mice are accelerated by, or allow the expression of, the BXSB Y-linked trait. This is true even with the MRL/l and MRL/n mice, which do not complement the late-life SLE background of the BXSB female. As expected, the F, females of lupus background mice crossed with male BXSB develop a disease similar to that of the F, offspring of crosses with female BXSB, since the Y chromosome is not involved. In crosses of BXSB males with normal strains (C3H, BALB/c, C57BL/6), there is no late-life disease to accelerate, and the Y-linked factor is relatively ineffectual. Practical value may lie in understanding the early SLE of both male and female F, hybrids from crosses of BXSB males with NZB or NZW, since the females’ disease is clinically and kinetically quite similar to that in the NZB X W cross, and importantly, the males developed SLE much earlier

343

MUHINE MODELS OF SLE

than the females. These NZB x BXSB or NZW x BXSB crosses enable study of early SLE in both sexes, which is not possible in the NZBXW hybrids. To investigate further whether the Y-linked accelerating gene of BXSB is in itself sufficient to induce early disease, Kosenberg and Steinberg (478) recently developed partially inbred congenic NZB. BXSB and NZW. BXSB mice, and studied levels of IgSC as a measure of their autoimmune state. In such mice, the BXSB contributes the male chromosome, and 3% of the autosomal genes. These researchers found that full expression of B cell hyperactivity required not only the Y chromosome but also one or more BXSB autosomal genes. Additional studies by Izui et al. (479)with (NZW X SB/Le)F, mice demonstrated that male mice, as in the case of(NZBx BXSB)F,, develop early-life severe lupus. These investigators thus concluded that all or almost all of the genetic abnormalities expressed in the BXSB male strain are contributed by the ancestral SB/Le strain. Transfer of the lpr gene to other standard inbred strains of mice (C3H/HeJ, C57BL/6, C57BL/10, SJL, AKR) by successive cross-intercross matings has recently been accomplished (4). The availability of such mice has allowed evaluation of Zpr gene function independent of most other genes in the MRL background, which in itself leads to autoantibody production and a late-life lupus syndrome (MRL-+/+ mice). Initial studies by Pisetsky et aZ. (480) with C57BL/G-lpr/lpr mice demonstrated that the Zpr gene can stimulate autoantibody (anti-DNA) production in mice other than the MKL strain, and does not require abnormalities unique to this background to potentiate autoreactivity. However, subsequent studies by these and other investigators (17,194,195,242,481) indicate that despite the induction of lymphoproliferation and autoantibodies in lpr homozygous mice with normal backgrounds, the histopathologic manifestations and mortality rates are considerably less than those in lpr homozygous mice with the lupus background. Thus, unmanipulated SLE-prone MRL-+/+ mice have a 50% mortality at 17 months of age, but introduction of the Zpr gene accelerates the 50% mortality to 5 months of age. In contrast, introduction of the Zpr gene into the normal B6 or C3H mice has much less effect; 50% mortality occurs at around 11-12 months of age compared to beyond 18 months for their unmanipulated counterparts. Furthermore, with regard to histopathologic manifestations (Table XI), Zpr induces early-life severe glomerulonephritis as well as arteritis and arthritis only in MRL-+/+ mice which have the SLE background. At 4 months of age, C3H ZprlZpr mice have about the same degree of lymphoid hyperplasia as the MRL-Zpr/Zpr, whereas lymphoid hyperplasia in BG-Zpr/Zpr is delayed, attaining magnitudes similar to those of the other strains only by 8 months ofage. In all Ipr substrains, the proliferating cells are of the Thy-1 +,Lyt-1 phenotype and secrete in vitro a BCDF-like factor, but levels of the latter are highest in MRL-Zpr lymph node cells. Serologically, hypergammaglobulinemia is great+

344

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J . DIXON

TABLE XI INFLUENCE OF GENETICBACKGROUND O N TIlE HISTOLOGIC A N D SEROLOGIC EFFECTS OF TIIE lpr GENE Histologic parameters"

Mouse strain MRL-lpr/lpr MRL- +/ + C57BL/6-lpr/lpr C57BLl6 C3H/HeJ-lpr/lpr CeH/HeJ

Serologic p a r a m e t e d

Anti-ssDNA Total 50% Total IgG IgM RF ICs IgM survival lymphoid (months) GN massr Arteritis Arthritis IgG (mg/ml) (mg/ml) (pg/mI) (pg/ml

5 17 12 >18 12 >18

3.6 3.1 1.3 0 0.5 0

1.9 ND 0.6 ND 1.7 ND

2.5 1.6 0 0 0

0

2.2 0.5 0 0 0 0

34.6 8.0 19.4 4.9 20.9 4.2

0.27 0.15 0.17 0.04 0.29 0.04

0.66 0.16 0.24 0.03 0.26 0.02

78 16 208 <5 167 <5

>401 1: 2,

a Determined at the age corresponding to 50% survival. bMRL mice tested at 5 months of age. C57BL16 and C3H/HeJ strains tested at 12 months of age. CSum of axillary, inguinal, mesenteric, retroperitoneal, and submandibular nodes.

est in MRL-lpr/lpr mice, but in all lpr substrains, IgG, and IgG,, predominate. Although the lpr gene markedly increases the total and gp70 related ICs when introduced into the lupus background MRL-+/+ mice, it induces very small increases in the two strains with normal backgrounds. Furthermore, anti-DNA antibodies are highest in the MRL-lpr/lpr, and the IgM to IgG switch occurs much earlier than in the other lpr substrains. Yet, the lpr gene induces serum polyclonal IgMRFs early in all substrains, and to higher levels in the normal mice than in those with the lupus background, although this may be caused by removal of RFs from the sera of the lupus background mice by complexing with the abundant autoantibodies or ICs present in their sera. Further analysis provides evidence that autoantibodies induced by the lpr gene in mice of lupus and normal backgrounds may differ not only quantitatively but also qualitatively. For example, whereas IgMRFs of MRL-lpr/lpr mice have a predominant IgG,, specificity, those of the normal background mice react predominantly with IgG,. As a whole, the findings of these studies indicate that, although autoantibodies can be induced by the lpr gene in various mice, the background genome has a profound influence on the pathogenicity of the autoantibodies induced, perhaps in part by influencing the type of B cell clones engaged in autoantibody production as well as the kinetics and levels of the abnormal responses. Furthermore, the amount of autoantigen available for immune complex formation (for example, gp70) may differ in various genetic backgrounds. As discussed in a previous section, similar modifying effects of background genes on disease expression have

I

21

345

M U H I N E M O D E L S OF SLE

been observed following in uivo injections of an exogenous immunostimulator, i.e., LPS (194, 297). To summarize, the results show that (1)a background genetic complementation exists between BXSB and NZ mice, but not between BXSB and MRL/I, (2) there is little autoimmune disease with crosses between MRL/1 and NZ mice or between either BXSB males or MRL/I with normal mice, (3) the male BXSB Y-linked accelerator gene or factor can accelerate disease of other lupus mice, but cannot induce disease in normal mice; this effect requires some autosomal BXSB genes in addition to the Y chromosome, but the nature of these additional genes and their interaction with the Y-linked trait are not yet known, and (4) the background genome on which an accelerator gene, such as the lpr, has been introduced, has profound influence on the pathogenicity of the abnormal responses induced, i.e., lupus-prone, but not normal, mice allow or complement expression of such pathogenicity. D. RELATIONSHIP BETWEEN

THE

TRAITS,H-2,

AND

Ig GENES

As indicated, lupus strains differ in H-2 haplotypes and Ig allotypes. This finding, however, does riot exclude the involvement of immune response and immune regulation gene abnormalities in disease pathogenesis. Studies by Bocchieri et al. (178) on recombinant lines from crosses of NZB with normal C58 mice suggest that neither H-2 or Ig genes are involved in the pathogenesis of the NZB disease. For example, the major controlling genes for expression of anti-X erythrocyte antibody the NTA are not structural genes for heavy or light antibody chains, since certain lines produce C58 heavy and light chains, and yet are Coomb’s or NTA positive. Conversely, others had NZB heavy and light chains and do not produce antierythrocyte or NTA autoantibodies. Furthermore, certain lines that have the recombinant Igh haplotype with Igh-C genes from NZB and Igh-V genes from C58 still express high autoantibody levels. Subsequent studies on the same recombinant lines by Datta et al. (477) also fail to demonstrate linkage between retrovirus expression, B cell hyperactivity (spontaneous or SRBCinduced IgM hypersecretion), reduced autologous MRLs, H-2 haplotype, Ig markers, and a variety of isozyme and other biochemical markers (Pep-3, Mup-1, Pgm-1, Igk, Hbb, Lap-1, Igh-C, Es-10, Gpt-1, H-2) located on various chromosomes. Studies by Dixon and McConahey (unpublished observations) on F, generations of BXSB males with MHL/l, NZW, or NZB females also fail to disclose any association between Igh-1 allotype, autoantibody levels (ANA, anti-DNA, NTA, anti-gp70), polyclonal B cell hyperactivity (high levels of IgM, IgG, and antihapten antibodies), and mortality rates. Moreover, no correlation of such parameters with H-2 has been observed in NZBxW F, mice. The absence of any association between H-2 and Ig genes is also suggested by the studies of Adams and Knight (482) who

346

ARGYRIOS N. THEOFILOPOULOS AND FRANK J. DIXON

found that, of the three genes controlling expression of glomerulonephritis in NZBxW and the two genes governing autoimmune anemia (Aia-1, Aia-2) in NZB and NZBXNZC F,, only Aia-2 is tightly linked to the MHC. According to these investigators (482,483), none of the three genes governing lupus nephritis can be a heavy chain Ig gene because these are on murine chromosome 12 (484), whereas Lpn-1 and Lpn-2 appear to be on chromosome 17, and Lpn-3 is not linked to the heavy chain allotype, as clearly shown. Moreover, Lpn-1 and Lpn-2 are not related to K light chain genes, since these are located on chromosome 6 (485). The finding that Ada-1 (governing autoimmune anemia) is on chromosome 4 again precludes its being a heavy chain or K light chain gene. Similarly, these genes are not related to the murine A light chain gene, since this gene is located on chromosome 16 (486). Studies ofHirose et al. (470) in (NZBXW) x NZB backcrosses suggesting an association of anti-DNA and anti-gp70 levels with the H-2" haplotype are at some variance with the above observations. These investigators recently developed the ZWDI8 strain (a NZW congenic line carrying the H-2d haplotype of NZB), produced (NZBxNZWD/8)F1 mice, and examined the difference in several immunologic abnormalities between the H-2d/H-22 heterozygous NZBx W F, and the H-2d/H-2c1homozygous NZBxWD8 F, mice. As with the results on recombinant lines of NZB mice cited above, no immunologic abnormalities were observed in ZWD/8 mice which express the H-2d haplotype of NZB. However, in comparison with NZBxW F, mice, the NZBxWD8 F, mice show markedly lower serum levels of the anti-DNA and anti-gp70 antibodies, and a lower incidence of proteinuria and early mortality, although there is no significant difference in the incidence and the amount of NTA, antierythrocyte antibody, and serum Ig levels between these two hybrids. Thus, in this system, it seems that a gene(s) within or closely linked to the H-2" complex of the NZW strain specifically acts to intensify the levels and quality (IgM + IgG) of anti-DNA and antigp70 autoantibodies, thus promoting the severity of renal disease in NZB x W mice. VI. Influence of Sex and Sex Hormones on the Pathogenesis of Murine SLE

Sex hormones and X- or Y-chromosome-linked genes may influence the expression of autoimmune disease. It is well known that hormones of the hypophysis, thyroid, parathyroid, adrenals, and gonads affect the lymphoid system's homeostasis and immune responses by mechanisms that have not yet been defined. Within the intricate homeostatic role played by hormones in lymphocyte function, the effects of the gonads on the immune response and autoimmune

MURINE MODELS OF SLE

347

disease are particularly apparent. In general, females make larger immune responses and are more susceptible to most autoimmune diseases than males, e.g., the incidence of SLE in women after puberty is nine times that in men (487). No explanation is readily apparent for this sex difference, but experimental and clinical studies in man and animals tend to incriminate, at least in part, female sex hormones rather than X-chromosome-associated genes. Females and castrated males in both lower animals and humans have consistently higher levels of Ig and immune responses than normal males, although the direct immunosuppressive effects of testosterone or immunoenhancing effects of estrogens have not been shown conclusively (488). Recent findings of elevated estriol levels in SLE patients with manifestations of Klinefelter’s syndrome (489,490) and of the absence of SLE in the castrated female monozygotic twin of a lupus victim (491) suggest further that chronic estrogenic stimulation may play an important role in the prevalence of SLE in females. Although the total amount of estrogens recovered from female SLE patients is normal, estradiol activity may be enhanced due to abnormalities in female hormone metabolic patterns (492). As in human SLE, studies primarily conducted by Tala1 and associates (493-495) and Steinberg and associates (496) in NZBx W F, mice implicate female sex hormones as autoantibody accelerating factors and as participants in the overall earlier mortality in females than males. Castrated NZB x W males resemble females in their accelerated autoimmune disease, detectable by 6 months of age. However, testosterone or dihydrotestosterone inhibits the onset of autoimmune disease in females or castrated male NZB x W mice following subcutaneous implantation of the androgens in silastic tubes. On the other hand, although prepubertal castration of female NZBXW mice is without effect, estrogen administration accelerates overt disease in both males and females. The modes by which sex hormones modify the NZB x W disease are explained variously as effects on antigen presentation and handling by the immune system, and as androgen-induced enhancement of suppressor T cell activity (497) or of tolerance inducibility (433). The accelerating effects of female hormones such as estrogens are by no means uniform in all murine SLE strains. For example, in mice homozygous for the Zpr gene, the sex influence appears to be strain dependent. Thus, although administration of androgens has been reported to retard female MRL/I disease (465), in this strain of lupus mice, sex appears to have little effect, since females die only slightly earlier than the males (5). In C57BL/G-lpr/lpr mice, the females exhibit significantly higher autoantibody levels (RF, anti-DNA, anti-gp70) than males (498), but in CSH-ZpdZpr and AKR-Zprllpr mice, no differences in autoantibody levels exist between the two sexes (194,195). These results suggest that the modifying effects of sex hormones are dependent on background genes.

348

AHGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

Additionally, in contrast to NZBxW mice, the BXSB male develops disease much earlier than females (3-5). In this strain, the male sex-determined accelerated autoimmunity is Y-chromosome-linked and not horinonally mediated. This conclusion is based on the following: (1)castration of males has no effect on the disease course (499), (2) the disease is inherited in a Y-linked fashion (father to son) in F, crosses of BXSB males with other autoimmune strains (19), and (3) transfer of early, severe SLE by male BXSB bone marrow or spleen cells is independent of the irradiated BXSB recipient’s sex (208). Interestingly, a human counterpart of BXSB male-predominant disease was recently described by Lahita et al. (500) in familial studies of SLE patients. VII. Viruses in Murine SLE

There has been considerable debate regarding the role of viruses in the etiology and/or pathogenesis of human and murine SLE. An initial report (501) that NZB disease could be transferred to normal mice with cell-free extracts and filtrates of NZB splenocytes was taken as a strong indication for the involvement of a transmissible agent such as virus in the etiology of murine lupus. However, studies by others (502,503) failed to confirm this finding. Subsequent research demonstrated that from NZB mice of any age, retroviruses can be isolated in high titers. Detection of this xenotropic virus demands cocultivation of NZB tissue homogenates with cells from species other than mice. Recent studies (138,139) have shown that two distinct xenotropic viruses can be isolated from NZB mice, one of which (NZB-X1) is spontaneously produced by NZB and NZB hybrid cells in culture, and the second (NZB-X2) is induced by treatment of NZB fibroblasts with IdU. NZB-X1 is distinct in that it possesses a unique surface glycoprotein (gp70) similar to gp70 found free in the serum of all mouse strains tested, but only associated with a virion derived from NZ mice. By crossing NZB to SWR, two independently segregating autosomal dominant loci (Nzv-1, Nzv-2) control the expression of infectious xenotropic virus (NZB-X1) (504,505). Nzv-1 alone or in association with Nzv-2 is responsible for high-grade virus expression. Low titers of virus are expressed when Nzv-2 is present in the absence of Nzv-1. A series of studies by Datta and associates with F,, F,, and backcross progeny of NZBX SWR mice (506, 507) and recombinant NZB XC58 lines (477) have seriously questioned the involvement of this virus in the pathogenesis of lupus. The results can be summarized as follows. (1)Some progeny whose tissue homogenates express titers of xenotropic virus as high as those of the NZB parent fail to develop signs of autoimmunity, (2) virus-negative offspring from these crosses still

MUHINE MODELS OF SLE

349

developed autoantibodies, (3) the phenotypic expression of virus does not correlate with the incidence of glomerular lesions, and (4) levels of the viral antigen gp70 do not correlate with the development of nephritis in these crosses. Although infectious retroviruses do not appear to contribute to the etiology and pathogenesis of murine lupus, the presence of anti-gp70 antibodies and serum retroviral gp70-anti-gp70 complexes, as discussed in Section 111,C,3, does correlate with disease and appears to contribute to glomerulonephritis (147,152). Although the gp70 involved in such complexes biochemically resembles xenotropic NZB gp70, as stated, such nonvirion-associated gp70 is also found in sera of normal mice at levels sometimes similar to those of lupus mice. The bulk of this gp70 is synthesized in the liver (144,145). However, only lupus mice mount an antibody response to this glycoprotein (147). Chronic viral infections, however, may have a secondary modifying role in murine lupus. For example, lymphocytic choriomeningitis (LCM) and polyoma viruses (508), and induced retrovirus (509) infections all cause or elevate ANA and SLE-like disease in mice. Although these viruses probably act, in part, by causing formation of antiviral antibodies and ICs, parallel polyclonal stimulation of ANA and other autoantibodies, or alterations of the regulatory T cell subsets, must be considered as potential enhancers in appropriate genetic backgrounds. Neonatal LCM virus infection changes the 50% mortality paint caused by SLE-like disease from 16 to less than 5 months in NZB females from over 2 years to 6 months in NZW, from 18 to 9 months in BXSB females, and from 18 to 12 months in MRL/n mice. In contrast, normal C3H and SWR mice injected neonatally with LCM virus and examined from birth to 2 years ofage do not develop the fatal SLE-like disease. The effects of LCM virus infection are reminiscent oftliose induced by chronic LPS stimulation in mice (Section IV,C, 1,d), which accelerates disease in late-life lupus substrains, but has markedly less effect on similarly treated normal mice, a finding that points to the importance of the genetic background in determining the pathogenicity of autoantibodies induced by immunostimulation. Consequently, although viruses may not be involved in the pathogenesis of murine lupus, they may induce or enhance aberrant responses and autoaggression with subsequent development of autoimmune manifestations via their polyclonal B cell activating potential, their cytolytic capacity, their possible tropisms for certain subpopulations of regulatory lymphocytes, and their possible capacity to associate with and modify autoantigens (510). It should be noted, however, that not all viruses accelerate lupus. For example, infection of NZ mice with lactate dehydrogenase virus was reported to inhibit anti-DNA antibodies and ameliorate the disease (511).

350

ARGYRIOS N . THEOFI1,OPOULOS AND FRANK J. DIXON

VIII. Treatment of Murine SLE

Because of their histologic and serologic similarities to human systemic autoimmunity, the murine models are invaluable in assessing the efficacy of various therapeutic modalities. Most of these approaches are not specific, and attempt to intervene and modify disease-associated inflammatory processes and/or hyperactivity of the immune system. However, a few specific means of treatment, such as those aiming at the inhibition of autoantibody production, have also been tried. Numerous immunosuppressive or immunomodulating agents have been tested. Corticosteroids, such as cortisone acetate (512), betamethazone phosphate (513), methylprednisolone (514), and hydrocortisone (515), slightly increase life expectancy and reduce disease symptoms in NZ mice. Methyprednisolone pulse therapy is currently the treatement of choice in SLE patients (516-518). Cytostatic or cytotoxic agents, such as 6-mercaptopurine (519), azathioprine (514), cyclophosphamide (514,520-525), frentizole (526), and dactinomycin (527,528) have all (particularly cyclophosphamide) been found to prevent, and in some instances retard or arrest, the disease progression in NZ mice. However, long-term treatment with some of these agents can markedly increase malignant tumors in these animals (515,520,521,529). Our studies (Hang, Theofilopoulos, and Dixon, unpublished observations) indicate that cyclophosphamide prevents disease in MHL/I and BXSB male mice when initiated before the clinical onset of renal disease, without an increase of malignant lymphoma. However, once renal damage occurs, cyclophosphamide cannot reverse the disease process, although Smith et al. (530) has claimed such a reversal, including lymphoid hyperplasia, in MRL/l disease. Combined regiments of cytostatic drugs and corticosteroids are superior to either drug alone in preventing deaths from renal disease (514,531). Cyclosporin A, a fungal cyclic endecapeptide with powerful immunosuppressive actions (532), inhibits in vitro spontaneous and LPS-induced anti-DNA autoantibody production (Gozes and Theofilopoulos, unpublished data) by lymphocytes of lupus-prone mice. Suppression following in vivo administration also reportedly occurs (533).Although cyclosporin treatment of a few SLE patients relieved arthralgia, the drug produces nephrotoxic side effects and angio-edema accompanied by reduced levels of C1 esterase inhibitor, and thus had to be discontinued (534). An antiviral agent, ribarivin, also reportedly prolongs survival, reduces anti-DNA antibodies, and reverses proteinuria in NZB x W mice (535). Other nonspecific treatments have been studied in SLE mice. NZB mice treated with antilymphocyte globulin show suppression of the hemolytic anemia (536); however, no favorable effects on antinuclear antibodies, renal disease, or elevated Ig levels in NZB x W mice are noted (537).A rat mono-

MURINE MODELS OF SLE

351

clonal antibody to Thy 1.2 antigen reportedly partially prevents MRL/I T cell proliferation and renal disease when given in large amounts (538).Furthermore, it was recently shown that large doses of monoclonal anti-IA" antibodies prolong survival in NZBx W mice and produce remission of established renal disease (539); anti-IAd was less efficacious. The possible use of such anti-IA antibodies in the treatment of this and other human autoimmune diseases is questionable since, in mice, it results in prolonged B cell depletion and severe long-term impairment of primary, secondary, and tertiary responses to priming antigens (540). Thus, despite the beneficial effects of anti-IA antibodies in murine spontaneous (SLE) or induced [experimental encephalitis (541), experimental myasthenia gravis (542)] diseases, caution is needed before initiating this method with humans. Another approach for nonspecific immunosuppression is total lymphoid irradiation (TLI), initially one of the treatments for malignant hematologic disorders. TLI consists of repeated lymphoid organ irradiation (150-220 radldose to a total of 2-3500 rads) over a 2- to 3-week period with bone marrow shielding. TLI increases survival and decreases proteinuria in NZ (543,544) and MRL/l (545,546) mice, and also prevents lymphoid hyperplasia in the latter if performed early in life. This treatment is also beneficial in some cases of human intractable rheumatoid arthritis or lupus (547, 548). The mechanisms by which TLI inhibits autoimmunity have not yet been fully elucidated. Although it was initially believed that TLI results in suppressor cell induction, studies in MRL/l mice (545) and more recent human studies (549,550) suggest that the main effect is helper T cell reduction. The dramatic effects on human and murine systemic autoimmune diseases suggest that TLI has potential in the treatment of fulininating SLE. However, the desirability of such radiotherapy in humans remains questionable due to occasional severe side effects and the marked reduction of T cell-dependent antibody responses in mice and humans after TLI. Other attempts to modify the hyperactive immune system and restore defective T cell function include the use of soluble immune response suppressor factors, multiple transplants of young syngeneic thymuses or thymocytes, and administration of thymic hormones. Thus, multiple injections of crude supernatants from Con A-activated normal lymphocytes reportedly prolong survival and retard GN in NZBxW mice (312,313). Because such crude supernatants contain many soluble factors (some with opposing effects), their mode of action has not yet been accurately defined. Future availability of cloned suppressor T cell lines, preferably with autoantigen specificity, may allow use of this approach in human autoimmune disorders. In fact, basic myelin-specific (551)and mycobacteria-specific (552,553) T cell lines have been used for the treatment or prevention of experimentally induced allergic encephalomyelitis and adjuvant arthritis, respectively, in

352

ARCYHIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

rats. Multiple transplants of young syngeneic thymocytes reportedly increase the longevity of NZ mice (459), although repeated attempts by others do not confirm this (460,461). On the premise that the underlying autoimmune defect in SLE is in the T cell compartment, probably caused by defective processing of thymus-derived cells, various thymic hormones have also been given to lupus-prone mice, with little success. Such preparations included thymosin (34,35),thymopoietin, and ubiquitin (462),facteur thymique serique (457), and thymulin (533). Both pharmacologic and physiologic amounts of prostaglandin E (PGE) modulate immune responses (554). Initial studies by Zurier et al. (555-557) and subsequently by others (558)show that pharmacologic quantities of PGE modulate the expression of SLE in NZBxW mice. Development of GN is suppressed, and survival is greatly prolonged. Maximal beneficial effects occur when treatment begins at 2 months of age. Further studies by Izui et al. (149) and Kelley et al. (559) demonstrate that PGE also prevents the development of fatal GN and the massive lymphoproliferation that accompanies the SLE of MRL/l mice, but does not protect BXSB male mice from similar disease. In MRL/l mice, PGE suppresses the immune response to retroviral gp70, but has little effect on the quantity or quality of anti-DNA antibodies, although it reduces the deposition of anti-DNA containing ICs in the kidneys (149). The mode by which these beneficial effects are mediated is unknown, but may include increased clearance of ICs by PGE-activated mononuclear phagocytes, suppression of the antibody responses to certain autoantigens, such as the retroviral envelope antigens, or direct inhibitory effects on proliferation of certain abnormal lymphocyte subsets, as in MRL/I mice. Prostaglandin synthesis is catalyzed by the enzyme fatty acid cyclooxygenase, and inhibition of this enzyme prevents the formation of PGEs and thromboxane. However, inhibition of endogenous PGE with ibuprofen, a fatty acid cyclooxygenase inhibitor, does not modify survival or autoantibody levels in NZB x W mice (560).This result might be explained on the inability of this compound to influence local production of thromboxanes in the kidney, despite its inhibitory effects on platelet-derived thromboxane (V. E. Kelley, personal communication). Calorie intake alone or variations in different dietary components influence immunological function (561,562). One might therefore expect that appropriate modification of the diet could affect the expression of autoimmune disease. Studies, particularly by Good, Fernandes, and associates (563-570) unequivocally demonstrate that low-calorie or low-protein diets reduce autoantibody and immune complex levels, and correct various abnormal immunologic parameters, including IL-2 levels. Subsequent studies by others also demonstrate that essential fatty deficient diets (571) and diets

MURINE MODELS OF SLE

353

(menhaden oil or marine lipids) enriched with the polyunsaturated fatty acid eicosapentaenoic acid (572,573), a 20-carbon analog of arachidonic acid, prevents proteinuria and prolongs survival in NZB x W mice. Conversely, diets enriched in lipid but with similar caloric content as the low-fat diets reportedly accelerate lupus nephritis (574,575). However, more recent studies by Kubo et al. (576) suggest that the most important factor in dietary effects on autoimmune diseases is not the source of energy, but the total calorie intake, i.e., low calorie intake, regardless of the energy source, is beneficial. Beach et al. (577) and Kubo et al. (578) also found that restriction in food intake greatly prolongs the life span of MRL/l mice, and reduces lymphoproliferation despite minimal anti-DNA and immune complex level effects. Optimal results are achieved by initiating dietary manipulations of lupus mice early in life, but considerable effects are also seen when treatment is initiated later. This finding underscores the possibility of using dietary manipulation of immunologic function as a means of therapeutic intervention in established autoimmune disease. The modes by which diets modify autoimmunity are not well defined, but may include effects on lymphocyte functions, reduction in production of certain autoantigens and antibodies, such as gp70-antigp70 complexes (150), and/or modification of the lipoxygenase or cyclooxygenase pathways, particularly of thromboxane in the kidneys (reviewed in 579). Restriction in certain trace minerals, particularly zinc, also reduces autoimmunity in NZ and MRL/l mice, including lymphoproliferation in the latter (577,580,581). The mechanism of this effect could be partly due to the effects of zinc on T and B cell functions, and its interactions with hormones, vitamins, and other essential nutrients, and/or to the zinc dependency of thymic hormones. Certain diets can induce, rather than suppress, lupus-like syndromes. Thus, it has recently been shown that 40-45% of monkeys fed diets high in alfalfa seeds develop serologic abnormalities similar to human lupus. The mechanisms for this effect are again unknown, but L-canavanine (a nonprotein arginine analog) that is highly concentrated in alfalfa seeds and sprouts is considered the responsible agent. An enhancing effect of alfalfa diets on humans with SLE has recently been observed (583). Moreover, diets deficient in vitamin A enhance autoimmunity in NZB mice (584). Amelioration of autoimmune diseases by specific suppression of pathogenic immune responses has been a highly sought therapeutic goal. Bore1 et al. (109,110) attempted to achieve this by inducing tolerance in NZBxW mice injected with nucleic acids coupled to isologous IgG and, initially reported attenuation of the disease by such treatment; continuous administration of tolerogen was required for this effect (585). These authors also demonstrate that intravenous administration of nucleosides coupled to iso-

354

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

geneic spleen cells into normal and NZB mice generates suppressor cells that diminsh the formation of nucleoside-specific antibody-forming cells (586). However, in more recent studies, these investigators report that young adult NZB x W female mice receiving either nucleoside-IgG or nucleoside-spleen cells fail to show significant differences from control mice in levels of proteinuria, blood urea nitrogen, anti-DNA antibodies or survival (587). In fact, such treatment of young mice leads to a more rapid production of anti-DNA antibodies and more severe renal disease. Thus, this type of specific treatment does not seem promising. An alternative means proposed by these investigators is the specific elimination of nucleoside-specific B cells by nucleosides coupled to plant toxin ricin A chain (588). Although such conjugates may reduce antinuclear antibody production by lupus B cells in vitro, their in vivo usage and effects are questionable and perhaps deleterious since the toxin-nucleoside conjugates would be expected to form complexes with the circulating antinuclear autoantibodies and thus exert toxicity on the retriculoendothelial system as well as other cell types expressing Fc and/or C receptors. Another means proposed for specific treatment of lupus is the use of autoantibody dominant idiotypes (Id) and/or respective antiidiotypes. Antiidiotypes may suppress immune responses by interaction with Id-bearing B cells, induction of Id-bearing suppressor T cells, or inactivation of Id-expressing helper T cells (589,590). Thus, initial studies by Hahn and Ebling (97) report suppression of circulating antibodies to dsDNA, diminished GN, and delayed mortality in premorbid NZB x W female mice following repeated administration of an IgG,, monoclonal anti-DNA antibody. The authors attribute the suppressive effects to the induction of antiidiotypic antibodies that result in the elimination or reduction of the respective dominant idiotype. The same investigators (501) found that administration of a monoclonal IgG,,k antiidiotypic antibody directed against a major cross-reactive idiotype of NZBX W IgG antibodies to DNA also diminishes anti-DNA levels and slightly prolongs survival. However, the results are transient and suppression is followed by the emergence of new anti-DNA clones that do not bear this major idiotype. Within a few weeks, clinical nephritis and death follow. Similarly discouraging results on this treatment are reported by Teitelbaum et al. (592) who indicate that a xenogeneic antiidiotypic antibody against the dominant idiotype of MRL anti-DNA antibodies does not result in suppression of the idiotype and, in fact, in some instances results in an augmented production of autoantibodies. These results are not surprising since antiidiotypic antibodies, depending on their concentration, may induce not only suppression but also enhancement of immune responses (590). These findings point to the difficulties in tailoring antiidiotypic treatments to

MUHINE MODELS OF SLE

355

induce long-term suppression, but not enhancement or induction of new idiotypes. IX. Conclusions

A number of different strains of mice spontaneously develop an autoimmune disorder with features of human SLE, although the clinical syndromes vary substantially among the various murine models. We have reviewed herein the histopathologic, serologic, lymphocytic, virologic, hormonal, and genetic characteristics of these models. It is obvious from the foregoing that the pathogenetic mechanisms underlying murine SLE are highly complex, apparently well-programmed genetically, but still diverse and their bases not as yet well defined. The major common feature is the ability of lymphoid cells from these mice to carry the information necessary to transmit the disease. Furthermore, significant serologic and cellular experimental data support the statement that the final immunopathologic perturbation in murine (and human) SLE is a B lymphocyte hyperactivity with corresponding enhancement of serum antibodies and autoantibodies, particularly IgG, and consequent formation of pathogenic antigen-antibody ICs. On the basis of the available data, it appears that this B cell hyperactivity is polyclonal (but not necessarily panclonal) in nature, since not only antibodies against a wide array of autoantigens, but also antibodies against incidental antigens, such as haptens, can be detected. Evidence, however, has been presented that the serologic diversity of lupus autoantibodies may not be as wide as previously thought, but, in part, an artifact caused by autoantibodies that cross-react with a variety of biologic molecules sharing similar structural or conformational epitopes. The B cell component that exhibits this hyperactivity appears to belong primarily (but not exclusively) in the Lyb-3+ ,5+ subset, as demonstrated by the significantly reduced Ig and autoantibody levels in lupus mice rendered homozygous to the xid gene. In NZ, but not MRL/l mice, this hyperactive B cell population was further shown to express the Ly-1 alloantigen. The presence of autoantibodies and of their idiotypes in (1)some recombinant lupus x normal murine lines expressing the normal partner’s V genes, (2) normal mice in which an exogenous (LPS or other polyclonal activator) or endogenous (Ipr gene) immunostimulator has been introduced, and (3) unmanipulated normal mice, all indicate that lupus mice are not unique in their structural genetic Ig elements whose presence determines the development of their autoimmune disease. Furthermore, recent nucleotide (592a) and protein (593) sequence data on a few monoclonal autoantibodies derived

356

ARGYliIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

from lupus mice would suggest that autoantibodies do not use different structural genetic elements than those for antibodies to exogenous antigens, rather autoantibodies and antibodies to exogenous antigens appear to utilize the same germline pool of structural Ig elements. Thus, as indicated in Fig. 2, the heavy chain 5’ untranslated region, leader peptide, and amino acids 1 to 94 of a MRL/1 derived monoclonal anti-DNA antibody were found by Kofler and associates to be encoded by a V gene closely related to the C57BL/6 nitrophenyl antibody family. These findings do not exclude the possibility that autoantibodies may preferentially utilize certain V gene families over others. It therefore seems likely that specificity for self antigens results from certain combinations of individual elements and of entire Ig chains or, alternatively, from somatic mutations of antibodies originally elicited by exogenous antigens. However, if one agrees that idiotypes sterically may represent mirror images of antigens, it would be reasonable then to assume that the immune system accepts structural diversity in the universe as nothing new or strange, since it “sees” these structures continuously in its complementary idiotype-antiidiotype circuits. Accordingly, “foreign determinants” are, in fact, never foreign (594,595). Since genetic rearrangement, H and L chain combinations, and somatic mutations are generally thought to occur randomly, autoantibody producing B cell clones should arise continuously during the entire life span of an individual. If, for whatever reason, the B cell component is hyperactive, then more mutations would be expected, particularly in IgG V regions, and far less, if any, in the IgM isotype. In lupus mice, in fact, not only B cell hyperactivity, but also a high mutation rate in antibody molecules such as the AlV regions have been demonstrated, with the two events paralleling one another. Similarly hypermutated AlV regions can be induced in normal mice after repeated mitogenic or antigenic stimulation. The cause of B cell hyperactivity in lupus has not yet been fully elucidated. Autonomous B cell maturation does not appear likely, since B cell proliferation and difierentiation in lupus mice was found to be dependent on the same number of accessory signals as in normal mice. Lack of, or reduced, generalized suppressor T cell function has also recently been questioned as the cause of B cell hyperactivity, since no consistent abnormalities in this regulatory T cell subset can be demonstrated. The evidence thus far suggests that B cell hyperactivity in murine lupus is caused either by heightened excitability to triggering stimuli (antigens, mitogens, helper T cell derived interleukins), as in NZ and BXSB mice, or to the hyperproduction of helper signals, as in MHL/l mice (596,597). In the latter, a massive T helper cell proliferation occurs, and expression of lymphoproliferation and disease is thymus dependent. Whatever the basis of the genetically determined B cell hyperactivity, this

FIG.2. Nucleotide sequence and deduced amino acid sequence of a MRLil-derived monoclonal anti-DNA antibody (MRL-DNA10). MRLDNA10 H chain V region is compared with the C57BL/6 nitrophenyl (NP) V I Igene 6 (Ighh)and the BALB/c "NP-equivalent" Vks gene 3 (Igh"). Dotted lines signify identity with MRL-DNA10 nucleotide sequence. Amino acids differing from a deduced V,, gene 6 sequence (replacing the deletion in codon 22 by T) are indicated by asterisks. For clarity, differences from the deduced VI, gene 3 amino acid sequence are not marked. The putative D and J H gene-encoded sequences and corresponding amino acids of MRL-DNA10 are also depicted. The single base difference from the BALBlc JIi3 sequence and the resulting amino acid change is indicated by an asterisk. Complementarity deterniining regions are boxed. Amino acid numbering according to Kabat et al. ("Sequences of Proteins of Inininnological Interest," U. S. Dept. of Health and Human Services Publication No. 1-323).

358

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

defect may not become overt until late in life unless accelerating factors are superimposed. These accelerating factors may be endogenous and genetically determined, as are the Y-chromosome-linked accelerating gene(s) of BXSB mice, and the Zpr gene of MRL mice. Accelerating factors of autoimmunity may also include female hormones, and some exogenous factors such as viruses and bacterial products that may activate B cells polyclonally. However, genetic and other studies clearly document that the background genome on which an accelerator or immunostimulator is introduced is the determining factor on the pathogenicity of the abnormal responses induced, i.e., lupus, but not normal, background mice allow or compliment expression of such pathogenicity. It is conceivable that background genes may influence the type of B cell clones engaged in autoantibody production as well as the kinetics and levels of the abnormal responses. Regarding the possible analogies between murine and human SLE, the NZ mice parallel humans in the sex, distribution, and immunopathologic characteristics of this disease, and BXSB males are the counterpart of a recently described SLE afflicting primarily fathers and sons. In contrast, no human correlate of SLE in the MRL/l mouse has been identified, yet this mouse can be considered to represent a global model of systemic autoimmunity, with a wide range of clinical and serologic characteristics that mimick Sjogren’s syndrome, rheumatoid arthritis-, mixed connective tissue-, and SLE-like disease, and syndromes sometimes encountered in humans undergoing graft-versus-host reactions. Future molecular studies in these models on the structure and regulation of Ig genes, structure of autoantibodies, the T and B cell regulatory circuits, and cell-associated or soluble immune recognition and regulation molecules should provide a more definitive picture on the origin and pathogenesis of these complex disorders. REFERENCES 1. Bielschowsky, M., Helyer, B. J ~ and , Howie, J. B. (1959). Spontaneous anemia in mice of the NZBlBl strain. Proc. Unit. Otago Med. School 37, 9. 2. Helyer, B. J., and Howie, J. B. (1963). Renal disease associated with positive lupus erythematosus in cross-bred strains of mice. Nature (London) 197, 197. 3. Murphy, E . D . , and Roths, J. B. (1979). Autoimmunity and lymphoproliferation. Induction by mutant gene Zpr and acceleration by a male-associated factor in strain BXSB mice. In “Genetic Control of Autoinimune Disease” (N. R. Rose, P. E. Bigazzi, and N. L. Warner, eds.), pp. 207-220. Elsevier, Amsterdam. 4. Murphy, E. D. (1981). Lymphoproliferation (Ipr) and other single-locus models for murine lupus. In “Immunologic Defects in Laboratory Animals” (E. M. Gershwin and V. Merchant, eds.), Vol. 2, pp. 143-173. Plenum, New York. 5. Andrews, B. S . , Eisenberg, R. S . , Theofilopoulos, A. N., h i , S., Wilson, C. B., McConahey, P. J., Murphy, E. D., Roths, J . B., and Dixon, F. J. (1978). Spontaneous murine lupus-like syndromes. Clinical and immunopathological manifestations in several strains. J . Exp. Med. 148, 1198.

MUHINE M O D E I S OF SLE

359

6. Walker, S. E., Gray, R. H., Fulton, M., Wigley, R. D.,and Schnitzer, B. (1978). Palinerston North mice, a new animal model of systemic lupus erythematosus. J . Lab. Clin. Med. 92, 932. 7. Davidson, W. F. (1982). Immunologic abnormalities of the autoimmune mouse, Palinerston North. J . Imnmunol. 129, 751. 8. Green, M. C., and Schultz, L. D. (1975). Motheaten, an immunodeficient mutant of the mouse. I. Genetics and pathology. J. Hered. 66, 250. 9. Schultz, L. D., and Green, M. C. (1976). Motheaten. an immunodeficient mutant of the mouse. 11. Depressed immune competence and elevated serum immunoglobulins. J . Immunol. 116, 936. 10. Davidson, W. F., Morse, H . C . , 111, Sharrow, S. O., and Chused, T. M. (1979). Phenotypic and functional effects of the motheaten gene on inurine B and T lymphocytes. J . lnimunol. 122, 884. 11. Fischer-Lindahl, K., Bocchieri, M., and Riblet, R. (1980). Maternally transmitted target antigen for unrestricted killing by NZB T lymphocytes. J . E x p . Med. 152, 1583. 12. Fischer-Lindahl, K., and Burki, K. (1982). Mta, a maternally inherited cell surface antigen of the mouse, is transmitted in the egg. Proc. Natl. Acad. Sci. U.S.A. 70, 5362. 13. Huston, M. M., Smith, R., Houston, D. P., and Rich, R. R. (1983). Differences in maternal lineages of New Zealand Black mice defined by restriction endonuclease analysis of mitochondrial DNA and by expression of maternally transmitted antigen. J. E x p . Med. 157, 2154. 14. Smith, R., Huston, D. P., and Rich, R. R. (1982). Priiiiary cell-mediated lympholysis response to a maternally transmitted antigen. J. E x p . Med. 156, 1866. 15. Bielschowsky, M . , and Goodall, C. M . (1970). Origin ofinbred NZ mouse strains. Cancer Res. 30, 834. 16. Callahan, C . , Dixon, F. J . , and Theofilopoulos, A. N. (1984). Ia antigens of AKR and C3H mice are not identical, and the I region o f t h e MHC of MRL mice is derived from AKR mice. J . Immutiogenet. (submitted). 1Ga. Cronin, P. S., Sing, A. P., Glimcher, L. H., Kelley, V. E . , and Reinisch, C. L. (1984).J. lmmunol. 133, 822. 17. Roths, J. B., Murphy. E . D., h i , S., and Kelley, V. E . (1983). Modification ofexpression of lpr by background genome. Fed. Proc., Fed. Am. Soc. E x p . Biol. 42, 1075 (Abstr.). 18. Koths, J. B., Murphy, E. I)., and Eicher, E. M. (1984). A new mutation, gld, that produces lymphoproliferation and autoimmunity in C3H/HeJ mice. J . E r p . Med. 159, 1. 19. Theofilopoulos, A. N., and Dixon, F. J. (1981). Etiopathogenesis of murine SLE. lmmurkol. Rev. 55, 179. 20. Warner, N. L. (1977). Genetic aspects of autoimmune disease in animals. I n “Autoinimunity” ( N . Talal, ed.), pp. 33-62. Academic Press, New York. 21. Howie, J. B., and Helyer, F. J. (1968).The immunology and pathology in NZB mice. Adu. lmnkunol. 9, 215. 22. Dixon, F. J. (1981). Murine systemic lupus erythematosus. lmmunol. Today 2, 145. 23. Lamhert, P. H . , and Dixon, F. J. (1968). Pathogenesis of the gloinerulonephritis of NZB/W mice. J . E x p . Med. 127, 507. 24. Yoshiki, T., Mellors, R. C.. Strand, M., and August, J. T. (1974). The viral envelope glycoprotein of murine leukemia virus and the pathogenesis of immune complex glomerulonephritis of New Zealand mice. J . Exp,. Med. 140, 1011. 25. Slack, J. H . , Hang, L. M . , Barkley, J., Fulton, R. J., D’Hoostelaere, L., Robinson, A , , and Dixon, F. J. (1984). Isotypes of spontaneous and mitogen-induced autoantibodies in SLE-prone mice. J . ~ m m z r n d .132, 1271. 26. Dixon, F. J., Oldstone, M. B. A,, and Tonietti, G. (1971). Pathogenesis of immune complex glomerulonephritis of New Zealand mice. J . E x p . Med. 134, 65-71.

360

ARGYHIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

27. Ebling, F., and Hahn, B. H. (1980). Restricted subpopulations of DNA antibodies in kidneys of mice with systemic lupus. Arthritis Rheum. 23, 392. 28. Caulin-Glaser, T., Gallo, G. R . , and Larnm, M. E. (1983). Nondissociating cationic immune complexes can deposit in glomerular basement meml)rane. J. E x p . Med. 158, 1561. 29. Gallo, G. R . , Caulin-Glaser, T., and Lamm, M. E. (1981). Charge of circulating immune complexes as a factor in glomerular basement membrane localization in mice. J . Clin. Znoest. 67, 1305. 30. Oite, T., Batsford, S. R . , Mihatsch, M. J., Takamiya, H., and Vogt, A. (1982). Quantitative studies of in situ immune complex glomerulonephritis in the rat induced by planted, cationized antigen. J. E x p . Med. 155, 460. 31. Gauthier, V. J., Mannik, M., and Striker, G. E. (1982). Effect ofcationized antibodies in preformed immune complexes on deposition and per lice in renal glomeruli. J. E x p . Med. 156, 766. 32. Gershwin, M. E . , Ikeda, R. M., Krose, W. L., Wilson, F., Shifrine, M., and Spangler, W. (1978). Age-dependent loss in New Zealand mice of morphological and functional characteristics of thymic epithelial cells. J . Zminunol. 120, 971. 33. Bach, J. F., Dardenne, M., and Salomon, J. C. (1973). Studies on thymus products. Presence of serum “thymic activity” in adult NZB and (NZBXNZW)F, mice. Clin. Erp. Zminunol. 14, 247. 34. Gershwin, M. E., Steinberg, A. D., Ahmed, A,, and Derkay, C. (1976). Study of thymic factors. 11. Failure of thyinosin to alter the natural history of NZB and NZBINZW mice. Arthritis Rlaeum. 19, 862. 35. Mehta, J.. Knotts, L., Craig, C., Burkholder, S., Miller, C., and Hahn, B. (1978). Effect of altered lymphocyte fuaction on immunologic disorders in NZB/NZW mice: 111. Acceleration of disease by thymosin. Arthritis Rheum. 21, 196. 36. Accinni, L., and Dixon, F. J. (1979). Degenerative vascular disease and myocardial infarction in mice with lupus-like syndrome. Am. J. Pathol. 96, 477. 37. Hang, L. M., h i , S., and Dixon, F. J. (1981). (NZBXBXSB)Flhybrid: A model ofacute lupus and coronary vascular disease with myocardial infarction. J. E x p . Med. 154, 216. 38. Hang, L. M . , Stephens-Larson, B. A., Henry, J. P., and Dixon, F. J. (1983). The role of hypertension in coronary vascular disease and myocardial infarcts associated with SLE. Arthritis Rheum. 26, 1340. 39. Berden, J. H. M., Hang, L. M., McConahey, P. J., and Dixon, F. J. (1983). Analysis of vascular lesions in inurine SLE. J. Zrnmunol. 130, 1699. 40. Denman, A. M., and Denman, E. J. (1970). Proliferative activity in the lymphatic tissues of germ-free New Zealand Black mice. Znt. J. Cancer 6 , 108. 41. East, J. (1970). Immunopathology and neoplasms in New Zealand Black (NZB) and SJL-J mice. Prog. E x p . Tuinor Res. 13, 84. 42. East, J., deSousa, M. A. B., and Parrott, D. M. (1965).Immunopathology ofNew Zealand Black mice. Transplantation 3, 711. 43. Raveche, E. S . , Steinberg, A. D., DeFranco, A. L., and Tjio, J.-H. (1982). Cell cycle analysis of lymphocyte activation in normal and autoimmune strains of mice. J. Zmmunol. 129, 1219. 44. Talal, N. (1974). Autoimmunity and lymphoid malignancy in New Zealand Black mice. Prog. Clin. Zmmunol. 2, 101. 45. Holmes, M. C . , and Burnet, F. M. (1963). The natural history of autoimmune disease in NZB mice. A comparison with the pattern of human autoimmune manifestations. Ann. Intern. Med. 59, 265.

36 1

MURINE MODELS OF SLE

46. Hang, L. M., Theofilopoulos, A. N., and Dixori, F. J. (1982). A spontaneous rheumatoid arthritis-like disease in MRL/I mice. J. Exp. M e d . 155, 1690. 47. Janossy, G., Panayi, G., Duke, O., Bofill, N . , Poulter, L. W., and Goldstein, G . (1981). Rheumatoid arthritis: A disease of T lymphocyte/macrophage immunoregulation. Lancet 2, 839. 48. Ishikawa, H . , and Ziff, M. (1976). Electron microscopic observations of immunoreactive cells in the rheumatoid synovial nieinhrane. Arthritis Rheum. 19, 1. 49. Kurosaka, M., and Ziff, M. (1983). Innnunoelectron microscopic study of the distribution of T cell subsets in rheumatoid synovium. J. E x p . Med. 158, 1191. 50. Hoffman, R. W., Alspangh, M. A., Waggie, K. S., Durham, J. B . , and Walker, S . E. (1984). Sjiigren's syndrome in MRL/I and MRL/n mice. Arthritis RAeum. 27, 157. 51. Hofliuan, R. W., Yang, H. K., Waggie, K. S . , Durham, J. B., Burge, J. R . , and Walker, S . E. (1983). Band keratopathy in MRL/I and MRL/n mice. Arthritis Rheum. 26, 645. 52. Pertschuk, L. P., Szabo, K., Vuletin, J. C . . and Rainford, E. A. (1977). Ocular immunopathology in inurine lupus. Atti. J. Pathol. 101, 122. 53. Williams, A. W., Howie, J. B., Helyer, 8.J., and Simpson, L. 0. (1967). Spontaneous peptic ulcers in mice. Aust. J. E x p . Biol. Med. Sci. 45, 105. 54. Lambert, P. W., and Oldstone, M. B. A. (1972). Host imniunoglobulin G and complement deposits in the choroid plexus during spontaneous immune complex disease. Science 180, 408. 55. Alexander, E. L., Murphy, E. D., Roths, J. B., and Alexander, 6. E. (1983). Congenic autoimmune murine models of central nervous system disease in connective tissue disorders. Am. J. Neurol. 14, 242. 56. Accinni, L., Albini, B., Andres, G., and Dixon, F. J. (1980). Deposition of immune complexes in ovarian follicles of mice with lupus-like syndrome. Am. J. Pathol. 99, 589. 57. Natah, P. G . , Correani, A , , and DeMartino, C. (1980). Spontaneous endometria] deposition of immune complexes in aging NZB/W F, mice: Possible role of oncornavirus antigens. Chn. Inlmunol. Itnmunopathol. 15, 11. 58. Gillian, J. N., Hurd, E. R., and Ziff, M. (1975). Subepidermal deposition of iminiinoglobulin in NZB/NZW F1 hybrid mice. J . Inzniunol. 114, 133. 59 Sontheimer, R. D., and Gillian, J. N. (1979). Regional variation in the deposition of subepidermal immunoglobulin in NZB/W F, mice: Association with epidermal DNA synthesis. J . Invest. Dennutol. 72, 25. 60. Barrett, K. J . , and Trepicchio, W., Jr. (1984). The MRL/Ipr and MRL/+ strains ofinice are haplotype J at the heavy chain variable region locr~s.Arthritis Rhenm. 27, S55 (distract). 61. Tsonis, P. A,, Gautsch, J. W., Kofler, R., Dixon, F. J., and Theofilopoulos, A. N . (1984). Structural polymorphisins in Cmu switch regions oflu~iusand lion-lupus mice. Fed. Proc., Fed. Am. Soc. E x p . Biol. 43, 1614 (Abstr.). 62. Stall. A. M . , and Loken, M. R. (1984). Allotypic specificities of murine IgD and IgM recognized by monoclonal antibodies. J . Z ~ i i t ~ i u n o132, l. 787. 63. Parsons, M., Cazenave, P.-A,, and Herzenherg, L. A. (1981). Igh-4d, a new a h t y p e at the mouse IgGl heavy chain locus. Iinmunogenetics 14, 341. 64. Sngai, S . , Pillarisetty. R., and Talal, N . (1973). Monoclonal inacrog~obulinemia in NZB/NZW F1 mice. J. E x p . Med. 138, 989. 65. Tan, E. M . (1982). Autoantibodies to nuclear antigens (ANA): Their iminunobiology and medicine. Ado. Immunol. 33, 167. 66. Lafer, E. M., Rauch, J . , Andrzejewski, C., Mudd, D., Furie, B., Furie, B., Schwartz, R. S . , and Stollar, B. D. (1981). Polyspecific monoclonal lupus autoantibodies reactive with both polyuucleotides and phospholipids. J. E x p . Med. 153, 897.

+

362

AHGYHIOS N . THEOFILOPOULOS A N D FRANK J . DIXON

67. Koike, T., Tomioka, H., and Kimagai, A. (1982).Antibodies cross-reactive with DNA and cardiolipin in patients with systemic lupus erythematosus. Clin. Exp. Zmmunol. 50, 298. 68. Shoenfeld, Y . , Hsu-Lin, S. C. Gabriels, J. E., Silberstein, L. E., Furie, B. C., Furie, B., Stollar, B. D., and Schwartz, R . S. (1982). Production of autoantibodies by hnman-human hybridomas. J. Clin. Invest. 70, 205. 69. Moller, A,, Gabriels, J. E., Lafer, E. M., Nordheim, A., Rich, A., and Stollar, B. D. (1982). Monoclonal antibodies recognize different parts of Z-DNA. J . B i d . Chem. 257, 12081. 70. Lafer, E. M., Moller, A,, Nordheim, A., Stollar, B. D., and Rich, A. (1981). Antibodies specific for left-handed Z-DNA. Proc. Natl. Acad. Sci. U.S.A. 78, 3546. 71. Kolata, G. (1983). Z-DNA moves toward “real biology.” Science 222, 495 (editorial). 72. Theofilopoulos, A. N . , Shawler, D. L., Eisenberg, R. A., and Dixon, F. J. (1980).Splenic immunoglobulin secreting cells and their regulation in autoimmune mice. J. Exp. Med. 151, 446. 73. Bell, D. A,, Clark, C., Blomgren, S. E., and Vaughan, J. H. (1973). Anti-DNA antibody production by lymphoid cells of NZBlW mice and human systemic lupus erythematosus (SLE). Clin. Zmmunol. Znimimopathol. 1, 293. 74. Fish, F., and Ziff M. (1982). The in oitro and in vioo induction of anti-double-stranded DNA antibodies in normal and autoimmune mice. J. Immunol. 128, 409. 75. Fournie, 6.J., Lambert, P. H., and Miescher, P. A. (1974). Release of DNA in circulating blood and induction of anti-DNA antibodies aRer injection of bacterial lipopolysaccharides. J . Exp. Med. 140, 1189. 76. Izui, S., Kobayakawa, T., Zryd, M. J., Louis, J., and Lambert, P.-H. (1977). Mechanism for induction of anti-DNA antibodies by bacterial lipopolysaccharides in mice. 11. Correlation between anti-DNA induction and polyclonal antibody formation by various polyclonal B lymphocyte activators. J. Zmmunol. 119, 2157. 77. Sawada S., and Talal, N. (1979). Characteristics of in oitro production of antibodies to DNA in normal and autoimmune mice. J. Zmmunol. 122, 2309. 78. Roder, J. C., Bell, D. A., and Singhal, S. K. (1978). Regulation of the autoimmune plaque-forming cell response to single-strand DNA (ssDNA) in oitro. J . Zmmunol. 121,38. 79. Pisetsky, D. S., and Caster, S. A. (1982).The B cell repertoire for autoantibody: Frequency of precursor cells for anti-DNA antibodies. Cell. Zmmunol. 72, 294. 80. Datta, S. K., Stollar, B. D., and Schwartz, R. S. (1983). Normal mice express idiotypes related to autoantibody idiotypes of lupus mice. Proc. Natl. Acad. Sci. U.S.A. 80, 2723. 81. Monier, J. C., Brochier, J., Moreira, A., Sault, C., and Roux, B. (1984). Generation of hybridoma antibodies to double-stranded DNA from non-autoimmune BALB/c strain: Studies on anti-idiotype. Zmmunol. Lett. 8, 61. 82. Talal, N. (1970). Iinmunologic and viral factors in the pathogenesis of systemic lupus erythematosus. Arthritis Rheum. 13, 887. 83. Lambert, P. H . , and Dixon, F. J . (1970). Genesis ofantinuclear antibody in NZB/W mice: Role of genetic factors and of viral infections. Clin. Exp. Zmmunol. 6, 829. 84. Dubroff, L. M., and Reid, R. J., Jr. (1980). Hydralazine-pyrimidine interactions may explain hydralazine-induced lupus erythematosus. Science 208, 404. 85. Tan, E. N. (1976). Immunopathology and pathogenesis of cutaneous involvement in systemic h ~ p u serythematosus. J. Inoest. Dematol. 67, 360. 86. Stollar, B. D. (1981). The antigenic potential and specificity of nucleic acids, nucleoproteins, and their modified derivatives. Arthritis Rheum. 24, 1010. 87. Pancer, L. B . , Milazzo, M. F., Morris, V. L., Singhal, S. K., and Bell, D. A. (1981). Irnmunogenicity and characterization of supernatant DNA released by murine spleen cells. J . Immunol. 127, 98.

MURINE MODELS OF SLE

363

88. Duke, R. C., Chervenak, R., and Cohen, J. J. (1983). Endogenous endonuclease-induced DNA fragmentation: An early event in cell-mediated cytolysis. Proc. Natl. Acad. Sci. U . S . A . 80, 6361. 89. Sano, H . , and Morimoto, C. (1982). DNA isolated from DNAianti-DNA antibody immune complexes in systemic lupus erythematosus is rich in guanine-cytosine content. J. Zmmuno/. 128, 1341. 90. Madaio, M. P . , Hodder, S., Scbwartz, R. S . , and Stollar, B. D. (1984). Responsiveness of autoimmune and normal mice to nucleic acid antigens. J. Zmmunol. 132, 872. 91. Fournie, G . J., Lambert, P. H . , Bankhurst, A. I)., and Miescher, P. A. (1976). Features of the immune response to DNA in mice. 11. Participation of B and T cells. Clin. E x p . Zmmunol. 26, 52. 92. Pancer, L. B., Bell, D. A,, and Singhal, S. K. (1980). Induction ofanti-ssDNA antibodies in normal and preautoimniune mice in t h o . J. Imtnunol. 124, 939. 93. Roder, J. C . , Bell, D. A,, and Singhal, S . K. (1978). Loss of self-tolerance to singlestranded deoxyribonucleic acid (ssDNA) in oitro. J. Initnunol. 121, 29. 94. Andrzejewski, C., Rauch, J., Lafer, E . , Stollar, B. D., and Schwartz, R. S . (1980). Antigen-binding diversity and idiotypic cross-reactions among hybridoma autoantibodies to DNA. J. Immunol. 126, 226. 95. Rauch, I., Murphy, E., Roths, J. B . , Stollar, B. D., and Schwartz, R. S. (1982). A high frequency idiotypic marker of anti-DNA autoantibodies in MRL//pr/lpr mice. J . Zmmuno/. 129, 236. 96. Marion. T. N., Lawton, A. R., 111, Kearney, J. F . , and Briles, 1). E. (1982). Anti-DNA autoantibodies in (NZBXNZW)Fl mice are clonally heterogeneous, but the majority share a common idiotype. J. Zmmunol. 128, 668. 97. Hahn, G. H., and Ebling, F. M. (1983). Suppression of NZB/NZW murine nephritis by administration of a syngeneic monoclonal antibody to DNA. J. Clin. Invest. 71, 1728. 98. Tron, F . , Guern, C . , Cazenave, R. A., and Bach, J. F. (1982). Intrastrain recurrent idiotypes among anti-DNA antibodies of (NZBx NZW)FI hybrid mice. Eur. J. In&munoZ. 12, 761. 99. Tron, F., Jacob, L., and Bach, J.-F. (1983). Murine monoclonal anti-DNA antibodies with an absolute specificity for DNA have a large amount of idiotypic diversity. Proc. Natl. Acad. Sci. U . S . A . 80, 6024. 100. Talal, N. (1977). Autoimmunity and Lymphoid Malignancy: Manifestations of Immunoregulatory Disequilibrium. In “Autoimmunity - Genetic, Immunologic, Virologic, and Clinical Aspects” (N. Talal, ed.), pp. 183-206. Academic Press, New York. 101. Papoian, R., Pillarisetty, R . , and Talal, N. (1976). Immunological regulation of spontaneous antibodies to DNA and RNA 11. Sequential switch from IgM to IgG in NZB/NZW F1 mice. Znimunology 32, 75. 102. Steward, M. W., and Hay, F. C. (1976). Changes in immunoglobulin class and subc1ass of anti-DNA antibodies with increasing age in NZB/W Fl hybrid mice. Clin. E x p . Immunol. 26, 363. 103. Farkas, A. I., Medgyesi, 6. A , , Fust, G . , Miklos, K . , and Gergely, J. (1982). Immunogenicity of antigen coniplexed with antibody. I. Role of diflerent isotypes. Immunology 45, 483. 104. Daeron, M. J , , Couderc, J , , Ventura, M., Liakopoulos, P., and Voisin, G . A. (1982). Anaphylactic properties of mouse monoclonal IgG2a antibodies. Cell. Zmn~unol.70, 27. 105. Kunkl, A,, and Klaus, 6 . G. B. (1981). The generation of memory cells. V. Preferential priming of IgGl B memory cells by immunization with antigen-IgG2a antibody complexes. Zniinunology 44, 163.

364

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

106. Eisenberg, R. A,, Winfield, J. B., and Cohen, P. L (1982). Subclass restriction ofanti-Sm antibodies in MRL mice. J. Immunol. 129, 2146. 107. Steward, M. W., Katz, F. E., and West, N. J. (1975).The role oflow affinity antibody in immune complex disease. I. Quantity of anti-DNA antibodies in NZB/W F I hybrid mice. Clin. Exp. Zmmunol. 21, 121. 108. Winfield, J. B., Koffler, D., and Kunkel, H. G. (1975). Specific concentration of polynucleotide immune complexes in the cryoprecipitates of patients with systemic lupns erythematosus. J . Clin. Invest. 56, 563. 109. Borel, Y., Lewis, R. M., Andre-Schwartz, J., Stollar, B. D., and Diener, E. (1978). Treatment of lupus nephritis in adult (NZBX NZW)FI mice by cortisone-facilitated tolerance to nucleic acid antigens. J. Clin.Invest. 61, 276. 110. Borel, Y . , Lewis, R. M., and Stollar, B. D. (1973). Prevention of miirine lupus nephritis by carrier-dependent induction of immunologic tolerance to denatured DNA. Science 182, 76. 111. Dixon, F. J , , Theofilopoulos, A. N., McConahey, P., and Prudhomme, G. J. (1983). Murine systemic lupus erythematosns. Prog. Zmmunol. 5, 1115. 112. Yoshida, H., Kohno, A,, Ohta, K., Hirose, S . , Maruyama, N., and Shirai, T. (1981). Genetic studies of autoimmunity in New Zealand mice. 111. Associations among anti-DNA antibodies, NTA, and renal disease in (NZBxNZW)F,x NZW backcross mice. J . Zmmunol. 127, 433. 113. Tron, F., Letarte, J., Barreira, M. C. R. A , , and Lasavre P. (1982). Specific detection of circulating DNA:anti-DNA immune complexes in human systemic lupus erythematosus sera using inurine monoclonal anti-DNA antibody. Clin. Exp. Itnmunol. 40, 481. 114. Bruneau, C. D., and Benveniste, J. (1979). Circulating DNA:anti-DNA complexes in systemic lupus erythematosus. J. Clin. Invest. 64, 191. 115. Koffler, D., Agnello, V., and Kunkel, H. G. (1974). Polynucleotide immune complexes in serum and glomeruli of patients with systemic lupus erythematosus. Am. J. Pathol. 74, 109. 116. Izui, S., Lambert, P. H., and Miescher, P. A. (1977). Failure to detect circulating DNAanti-DNA complexes by four radioimmunological methods in patients with systemic lupus erythematosus. Clin. E x p . Immunol. 30, 384. 117. Eisenberg, R. A , , Tan, E. M . , and Dixon, F. J . (1978). Presence of anti-Sm reactivity in autoimmune monse strains. J. Exp. Med. 147, 582. 118. Gioud, M., Kotzin, B. L., Rubin, R. L., Joslin, F. G . , andTan, E. M. (1983).Znoivoand in oitro production of antihistone antibodies in NZB/NZW mice. J . Zmmunol. 131, 269. 119. Lerner, E. A., Lerner, M. R., Janeway, C. A., Jr., and Steitz, J. A. (1981). Monoclonal antibodies to nucleic acid-containing cellular constituents: Probes for molecular biology and autoimmune disease. Proc. Natl. Acad. Sci. U.S.A. 78, 2737. 120. Yang, V. W., Lerner, M. R., Steitz, J. A., and Flint, S. J. (1981). A small nuclear ribonucleoprotein is required for splicing of adenoviral early RNA sequences. Proc. Natl. Acad. Sci. U.S.A. 78, 1371. 121. Lerner, M. R . , Boyle, J. A , , Mount, S. M., Wolin, S. L., and Steitz, J. A. (1980). Are snRNPs involved in splicing? Nature (London) 283, 220. 122. Pisetsky, D. S . , McCarty, G. A., and Peters, D. V. (1980). Mechanisms of autoantibody production in autoimmune MRL mice. J . Exp. Med. 152, 1302. 123. Cohen, P. L., and Eisenherg, R. A. (1982). Anti-Sm autoantibodies in MRL mice: In vitro detection and generation of antibody-forming cells. J . Zmmunol. 129, 2682. 124. Cohen, P. L., and Eisenberg, R. A. (1982). T cell recognition of the Sin nuclear antigen: Induction of T cell proliferative responses in MRL/Mp-+/+ mice. J . Zmmunol. 129,2142. 125. Hardin, J., and Thomas, J. 0. (1983). Antibodies to histones in systemic lupus erythema-

M U R I N E M O D E L S OF S L E

365

tosus: Localization of proniinent autoantigens on histones H I and H2B. Proc. Nntf. Acucf. Sci. U . S . A . 80, 7410. 126. Theofiloporilos. A. N., Balderas, R . S., Hang, L. M., and Dixon, F. J. (1983). Monoclonal IgM rheumatoid factors derived from arthritic MRL/Mp-lpr/lpr mice. J . E x p . Med. 1.58, 90 1. 127. Ruliin, R. L., Balderas, R. S., Tan, E. M . , Dixon, F. J.. and Theofilopoulos, A. N. (1984). Multiple autoantigen binding capalAities of mouse inonoclonal antibodies selected for rheumatoid factor activity. J. Exp. Med. 159, 1429. 128. Cioud, M., Aitkaci, M., and Monirr, J. C. (1982). Histone antibodies in systemic lupus erythematosus: A possible diagnostic tool. Arthritis Rlaeum. 2.5, 407. 129. Fishl)ein, E., Alarcon-Segovia, D., and Vega, J . M. (1979). Antibodies to histones in systemic lupus erythematosus. C h . E x p . Zrnntunol. 36, 145. 130. Lafer, E. M., Valle, R . P. C., Moller, A,, Nordheim, A., Schur, P. H., Rich, A., and Stollar, B. D. (1983). Z-DNA-specific antilmdies in human systemic lupus erythematosus. J . C h . Znoest. 71, 314. 131. Eilat, D., Schechter, A. N., and Steinberg, A. D. (1976). Antibodies to native tRNA in NZR/NZW mice. Nature (London) 259, 141. 132. Eilat, D., DiNatale, P., Steinberg, A. I]., and Schechter, A. N. (1977). Properties of tRNA-specific antibodies from NZB/NZW mice. J. Zrnmunol. 118, 1016. 133. Eilat, D., Patton Reeves, J . , Steinberg, A. D., and Schechter, A. N. (1978). Measurement of antigen-specific IgG autoantibody prodnction in oifro. J. Itnmtrnol. 121, 310. 134. Eilat, D., Sasson, S. A. B., and Laskov, R. (1980). A ribonucleic acid-specific antibody produced during autoinnnune disease: Evidence for nucleotide sequence specificity. Eur. 1. ItnmUn<J/. 10, 841. 135. Eilat, D. (1982). Monoclonal autoantibodies: An approach to studying autoimmune disease. Mol. lmmuno!. 19, 943. 136. Lerner, R. A,, Wilson, C. B., DelVillano, B. C., McConahey, P. J., and Dixon. F. J. (1976). Endogenous oncornaviral gene expression in adult and fetal mice: Quantitative, histologic, and physiologic studies of the major viral glycoprotein, gp70. J. E x p . Med. 143, 151. 137. Hino, S., Stephenson, J. R . , and Aaronson, S. A. (1976). Radioiininunoassays for the 70,000-molecular-weight glycoproteins of endogenous mouse type C viruses: Viral antigen expression in normal mouse tissues and sera. 1.Viral. 18, 933. 138. Elder, J. H., Jensen, F. C., Bryant, M. L., and Lerner, R. A. (1977). Polymorphism ofthe major envelope glycoprotein (gp70)of inurine C-type viruses: Virion associated and differentiation antigens encoded by a multi-gene family. Nature (London) 267, 23. 139. Elder, J. H., Gautsch, J. W., Jensen, F. C., Lerner, R. A,, Chused, T. M., Morse, H. C., Hartley, J. W., and Rowe, W. P. (1980). Differential expression of two distinct xenotropic viruses in NZB mice. C/in. Zmmunol. Ztmnunoputhd. 15, 493. 140. Theofilopoulos, A. N . , Balderas, H. S . , Gozes, Y., Fidler, J. M., Liu, F. T., Aliined, A., and Dixon, F. J. (1982). Surface and functional characteristics of B cells from lupus-prone murine strains. Clin. Zmnzunol. Zmniunoputhol. 23, 224. 141. Del Villano, 8. C., Nave, B., Croker, B. P., Lerner, R. A,, and Dixon, F. J. (1975). The oncornavirus glycoprotein gp69/71: A constituent of the surface of normal and malignant thymocytes. J. E x p . Merf. 141, 172. 142. Obata, Y., Ikeda, H., Stockert, E., and Boyse, E. A. (1975). Relation of G,, antigen of thytnocytes to envelope glycoprotein of murine leukemia virus. J . Exp. Med. 141, 188. 143. Obata, Y., Stockert, E . , Yamaguchi, M., and Boyse, E. A. (1978). Source and hormonedependence of Gix-gp70in mouse serum. J. E x p . Med. 148, 793. 144. Hara, I., h i , S., McConahey, P. J., Elder, J. H., Jensen, F. C., and Dixon, F. J. (1981).

366

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. D I X O N

Induction of high serum levels of retroviral eno gene produts (gp70) in mice by bacterial lipopolysaccharide. Proc. Natl. Acad. Sci. U.S.A. 78, 4397. 145. Hara, I., h i , S., and Dixon, F. J. (1982). Murine serum glycoprotein gp70 behaves as an acute phase reactant. J . E x p . Med. 155, 345. 146. Maruyama, N., Lindstrom, C., Sato, H., and Dixon, F. J. (1983). H-2 linked regulation of serum gp70 production in mice. lmrnunogenetics 18, 365. 147. Izui, S., McConahey, P. J., Theofilopoulos, A. N., and Dixon, F. J. (1979). Association of circulating retroviral gp70-anti-gp70 immune complexes with murine systemic lupus erythematosus. J. E x p . Med. 149, 1099. 148. Nakai, Y., Maruyama, N . , Ohta, K., Yoshida, H., Hirose S., and Shirai, T. (1980). Genetic studies of autoimmunity in New Zealand mice. Association of circulating retroviral gp70 immune complexes with proteinuria. Zmmunol. Lett. 2, 53. 149. Izui, S., Kelley, V. E., McConahey, P. J., and Dixon, F. J. (1980). Selective suppression of retroviral gp70-anti-gp70 immune complex formation by prostaglandin El in niurine systemic erytheniatosus. J. E x p . Med. 152, 1645. 150. Izui, S., Fernandes, G . , Hara, I., McConahey, P. J., Jensen, F. C., Dixon, F. J., and Good, R. A. (1981). Low-calorie diet selectively reduces expression of retroviral envelope glycoprotein gp70 in sera of NZBXNZW F1 hybrid mice. J. E x p . Med. 154, 1116. 151. Izui, S., Elder, J. H . , McConahey, P. J.. and Dixon, F. J. (1981). Identification of retroviral gp 70 and anti-gp70 antibodies involved in circulating immune complexes in NZBXNZW mice. J . E x p . Med. 153, 1151. 152. Izui, S., McConahey, P. J., Clark, J. P., Hang, L. M . , Hara, I. and Dixon, F. J. (1981). Retroviral gp70 immune complexes in NZB X NZW Fz mice with murine lupus nephritis. J. E x p . Med. 154, 517. 153. DeHeer, D. H., and Edgington, T. S. (1976). Cellular events associated with the immunogenesis of anti-erythrocyte autoantibody response of NZB mice. Transplant. Reo. 31, 116. 154. Linder, E. , and Edgington, T. S. (1972). Antigenic specificity of anti-erythrocyte autoantibody responses by NZB mice: Identification and partial characterization of two erythrocyte surface autoantigens. J. Zmmunol. 108, 1615. 155. DeHeer, D. H., and Edgington, T. S. (1974).Clonal heterogeneity of the anti-erythrocyte autoantibody responses of NZB mice. J . Zmmunol. 113, 1184. 156. Pages, J. M . , and Bussard, A. E. (1978). Establishment and characterization of a permanent inurine hybridoma secreting monoclonal autoantibodies. Cell. Zmmunol. 41, 188. 157. DeHeer, D. H., Pages, J. M., and Bussard, A. E. (1980). Specificity of antierythrocyte autoantibodies secreted by a NZB-derived hybridoma and NZB peritoneal cells. Cell. Zmmunol. 49, 135. 158. Serban, D., Pages, J. M . , Bussard, A. E., and Witz, I. P. (1981). The participation of trimethylammonium in the mouse erythrocyte epitope recognized by monoclonal autoantibodies. Zrnmunol. Lett. 3, 315. 159. Koltz, J. L., Phillips, M. L., Miller, M. M., and Teplitz, R. L. (1981). Monoclonal autoantibody production by hybrid cells lines. Clin. Immunol. Zmrnunopathol. 18, 368. 160. Holborow, E. J., Barnes, R. D. S., and Tuffrey, M. (1965).A new red-cell autoantibody in NZB mice. Nature (London) 207, 601. 161. Linder, E. (1977). Autoantibodies against the inner aspect of erythrocyte membranes in NZB mice. Clin. E x p . Immunol. 27, 531. 162. Costea, N., Yakulis, V . , and Heller, P. (1970). Cold reactive antibody in NZB/B1 mice. Blood 35, 583. 163. DeHeer, D. H . , and Edington, T. S. (1976). Specific antigen-binding and antibody-

M U R I N E MODELS OF SLE

164.

165.

166. 167. 168.

169. 170.

171. 172.

173.

174. 175. 176. 177. 178.

179.

180.

181.

367

secreting lymphocytes associated with erythrocyte antoantibody responses of NZB and genetically unrelated mice. J. Zmmunol. 116, 1051. Linder, E . , and Edgington, T. S. (1973). Iminunobiology of the autoantibody response. I. Circulating analogues of erythrocyte autoantigens and heterogeneity of the autoimmune response of NZB mice. Clin. Exp. Zmmunol. 13, 279. DeHeer, D. H . , Linder, E. J., and Edgington, T. S. (1978). Delineation of spontaneous erythrocyte autoantihody responses of NZB and other strains of mice. J. Zmmunol. 120, 825. Shirai, T., and Mellors, R. C. (1971). Natural thymocytoxic autoantibody and reactive antigen in New Zealand Black and other mice. Proc. Natl. Acad. Sci. U.S.A. 68, 1412. Shirai, T., and Mellors, R. C. (1972). Natural cytotoxic autoantihody against thymocytes in NZB mice. Clin. E x p . Immunol. 12, 133. Shirai, T., Yoshiki, T., and Mellors, R. C. (1972). Thymus dependence of cells in peripheral lymphoid tissues and in the circulation sensitive to natural thymocytotoxic autoantibody in NZB mice. /. Zmintrnol. 109, 32. Imai, Y., Nakano, T., Sawada, J.-I., and Osawa, T. (1980). Specificity of natural thyniocytotoxic autoantibody developed in New Zealand Black mice. J . Zmmunol. 124, 1556. Shirai, T., Hayakawa, K., Okuinura, K., and Tada, T. (1978). Differential cytotoxic effect of natural thyinocytotoxic autoantibody of NZB mice on fmctional subsets of T cells. J. Zmmunol. 120, 1924. Klassen, L. W., Krakauer, R. S., and Steinberg, A. D. (1977). Selective loss of suppressor cell function in New Zealand mice induced by NTA. J. Zmmunol. 119, 830. Eisenberg, R. A., Theofilopoulos, A. N . , Andrews, B. S . , Peters, C. J., Thor, L., and Dixon, F. J . (1979). Natural thymocytotoxic autoantibodies in autoimmune and normal mice. J. Iinmunol. 122, 2272. Gershwin, M. E., Castles, J. J.. Ikeda, R. M., Erickson, K., and Montero, J. (1979). Studies of congenitally immunologic mutant New Zealand mice. I. Autoiinniune features of hereditarily asplenic (Dh/+) NZB mice: Reduction of naturally occurring thymocytotoxic antibody and normal suppressor function. J , Zmmunol. 122, 710. Schlesinger, M. (1965). Spontaneous occurrence of autoantibodies cytotoxic to thymus cells in the sera of mice of the 129 strain. Nature (London) 207, 429. RaR, M. C. (1971). Naturally-occurring autoantibodies to mouse thymocytes. Zsr. J. Med. Sci. 7, 724. Aver, I. O., Tomasi, T. J., Jr., and Milgrorn. F. (1974). Natural thymocytotoxic autoantibodies in NZB and other strains of mice. Ce//. Immunol. 10, 404. Martin, W. J., and Martin, S. E. (1975). Thymus reactive IgM autoantibodies in normal mouse sera. Nature (London) 254, 716. Bocchieri, M. H., Cooke, A,, Smith, J. B., Weigert, M., and Riblet, R. J. (1982). Independent segregation of NZB immune abnormalities in NZB X C58 recombinant inbred mice. Eur. J. Zmmunol. 12, 349. Raveche, E. S . , Novotny, E. A , , Hansen, C. T., Tjio, J. H., and Steinberg, A. D. (1979). Genetic studies in NZB mice. V. Recombinant inbred lines demonstrate that separate genes control autoimmune phenotype. J. E x p . Med. 153, 1187. Hirose, S . , Maruyama, N., and Shirai, T. (1980). Polyclonal B cell activation and autoinimunity in New Zealand mice. I. Natural thymocytotoxic autoantibody (NTA).J . Zmmunol. 125, 610. Nakanishi, A,, Imai, Y., Nakano, T., and Osawa, T. (1982). Induction of autoimmune phenomena in normal mice treated with natural thymocytotoxic autoantibody. J . Zmmunol. 128, 2137.

368

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

182. Eisenberg, R. A., Thor, L. T., and Dixon, F. J. (1979). Serum-serum interactions in autoimmune mice. Arthritis Rheum. 22, 1074. 183. Nardella, F. A., Teller, D. C., h i , S., and Mannik, M. (1984). Self-associating IgGrheumatoid factors in MRL/I autoimmune mice. Arthritis Rheum. 27, 1165. 184. Izui, S . , and Eisenberg, R. A. (1980).Circulating anti-DNA-rheumatoid factor complexes in MRL/l mice. Clin. Zmmunol.Imnzunopatho1. 15, 536. 185. Nemazee, D. A , , and Sato, V. L. (1983). Induction of rheumatoid antibodies in the mouse. Regulated production of autoantibody in the secondary humoral response. J. Exp. Med. 158, 529. 186. Nemazee, D. A. (1984). Immune complexes can trigger specific, T cell dependent, autoanti-IgG antibody production in mice. J . Exp. Med. (submitted). 187. Hannestad, K., and Stollar, B. D. (1978). Certain rheumatoid factors react with nucleosomes. Nature (London) 275, 671. 188. Agnello, V., Arbetter, A., Ibanez, G., Powell, R., Tan, E. M., and Joslin, F. (1980). Evidence for a subset of rheumatoid factors that cross-react with DNA-histone and have a distinct cross-idiotype. J . Exp. Med. 151, 1514. 189. Theofilopoulos, A. N . , Rubin, R., Balderas, R., Tsokos, G., Tan, E., and Dixon, F. J. (1984). Monoclonal mouse and human rheumatoid factors with multiple autoantigen specificities. Fed. Proc., Fed. Am. SOC.E x p . B i d . 43, 1735 (Abstr.). 190. Andre-Schwartz, J., Datta, S. K., Shoenfeld, Y., Isenberg, D. A., Stollar, B. D., and Schwartz, R. S. (1984). Binding of cytoskeletal proteins by monoclonal anti-DNA lupus autoantibodies. Clin. Zmmunol. Zmnzunopathol. 31, 261. 191. Dighioro, G., Lymberi, P., Mazie, J. C., Rouyre, S., Butler-Browne, G. S., Whalen, R. G . , and Avrameas, S. (1983). Murine hybridomas secreting natural antibodies reacting with self-antigens. J . Zmmunol. 131, 2267. 192. Haspel, M. V., Onodera, T., Prabhakar, B. S., McClintock, P. R., Essani, K., Ray, U. R., Yagihashi, S., and Notkins, A. L. (1983). Multiple organ-reactive monoclonal autoantibodies. Nature (London) 304, 73. 193. Satoh, J., Prabhakar, B. S., Haspel, M. V., Ginsberg-Fellner, F., and Notkins, A. L. (1983).Human monoclonal autoantibodies that react with multiple endocrine organs. New Engl. J . Med. 309, 217. 194. Hang, L. M., Aguado, M. T., Dixon, F. J.. and Theofilopoulos, A. N . (1985). Induction of severe autoimmune disease in normal mice by simultaneous action of multiple immunostimulators. J. Exp. Med. (in press). 195. Warren, R. W., Caster, S. A., Roths, J . B., Murphy, E. D., and Pisetsky, D. S. (1984). The influence of the lpr gene on B cell activation: Differential antibody expression in lpr congenic mouse strains. Clin. Zmmunol. Zmmunopathol. 31, 65. 196. Aguado, M . T., Balderas, R., Dixon, F. J., and Theofilopoulos, A. N. (1984). Monoclonal rheumatoid factors from lupus and normal background lpr homozygous mice. J . Zmmunol. (Submitted). 197. VanSnick, J. L., and Masson, P. L. (1979). Age-dependent production of IgA and IgM autoantibodies against IgG2a in a colony of 129/Sv mice. J . Exp. Med. 149, 1519. 198. Phadke, F. R., Fouts, R., Parrish, J., and Baker, R. S. (1984).Autoreactivity to collagen in a murine lupus model. Arthritis Rheum. 27, 313. 199. Holmes, M. C., Gorrie, J., and Burnet, F. M. (1961).Transmission by splenic cells of an autoimmune disease occurring spontaneously in mice. Lancet 2, 638. 200. Kaye, D., and Hooke, E. W. (1964). Influence of autoimmune hemolytic anemia on susceptibility to Salmonella infection. Proc. Soc. E x p . B i d . Med. 117, 20. 201. Allman, V., Ghdar, A., Playfair, J. H. L., and Roitt, I. M. (1969). Transfer of autoantibody formation by NZB bone marrow cells. Transplantation 8, 899.

MURINE MODELS OF SLE

369

202. Lindsey, E. S., and Woodruff, M. F. A. (1967). The effect of irradiation and infusion of marrow and spleen cells of autoimmune anemia in NZB/BI mice. Proc. R. SOC. Med. 60, 826. 203. Denman, A. M., Russell, A. S., and Ilenman, E . J. (1969). Adoptive transfer of the disease of New Zealand Black mice to normal mouse strains. Clin. E x p . Immunol. 5 , 567. 204. Morton, J. I . , and Siegel, B. V. (1974). Transplantation of autoimmune potential. I. Development of antinuclear antihodies in H-2 histocompatible recipients of bone marrow . U . S . A . 71, 2162. from New Zealand Black mice. Proc. Natl. A c Q ~Sci. 205. Morton, J. I . , and Siegel, B. V. (1974). Early autoantihody formation in lethally irradiated or drng-treated H-2-compatible recipients of pre-autoimmune NZB bone marrow or fetal liver cells. Transplantation 17, 624. 206. Akizuki, M., Reeves, J. P., and Steinherg, A. 1).(1978). Expression of autoimmunity by NZB/NZW marrow. Immunol. lmmunopatho/. 10, 247. 207. Jyonouchi, H., Kincade. P. W . , Good, R. A,, and Fernandes, G . (1981). Reciprocal transfer of abnormalities in clonahle B lymphocytes and inyeloid progenitors between NZB and DBA/2 mice. J. Ztnrnunol. 127, 1232. 208. Eisenberg, R. A,, h i , S., McConahey, P. J., Hang, L. M., Peters, C. J., Theofilopoulos, A. N., and Dixon, F. J. (1980). Male determined accelerated autoimmune disease in BXSB mice: Transfer by bone marrow and spleen cells. J. Zmmunol. 125, 1032. 209. Theofilopoulos, A. N. (1982). Role of the thymus in murine SLE and cellular transfer of the disease. Arthritis Rheum. 25, 726. 210. Onoe, K., Fernandes, G . , and Good, R. A. (1980). Humoral and cell-mediated immune responses in fully allogeneic bone marrow chimera in mice. J. E x p . Med. 151, 115. 211. Korngold, R., and Sprent, J. (1978). Lethal graft-versus-host disease after bone marrow transplantation across minor histocompatibility barriers in mice. J. E x p . Med. 148, 1687. 212. Reisner, Y., Ikehara, S., Hodes, M. Z., and Good, R. (1980). Allogeneic hematopoietic stem cell transplantation using mouse spleen cells fractionated by lectins: In uitro study of cell fractions. Proc. Natl. Acad. Sci. U . S . A . 77, 1164. 213. Francis, S., Hang, L. M., Gozes, Y., Dixon, F. J.. and Theofilopoulos, A. N. (1984). BXSB male disease phenotype can he transferred by fetal liver cells. Fed. Proc., Fed. Am. Soc. E x p . B i d . 43, 1831 (Ahstr.). 214. Stoho, J. D., Talal, N., and Paul, W. E. (1972). Lymphocyte classes in New Zealand mice. I. Ontogeny and mitogen responsiveness of thyinocytes and thymus-derived lymphocytes. J. Iinmunol. 109, 692. 215. Cohen, P . , Ziff, M., and Vitetta, E. S. (1978).Characterization o f B cell defect in the NZB niouse manifested by an increased ratio of surface IgM to IgD. J. Zmmtmol. 121, 973. 216. Theofilopoulos, A. N., Eisenherg, R. A., Bourdon, M., Crowell, J. S., Jr., and Dixon, F. J. (1979). Distribution of lymphocytes identified hy surface markers in murine strains with SLE-like syndromes, J. Exp. Med. 149, 516. 2i7. Bourgois, A,, Kitajima, K., Hunter, I. R., and Askonas, B. A. (1977). Surface immunoglobulins of lipopolysaccharide-stimulated spleen cells. The behavior of IgM, IgD and IgG. Eur. J. Immunol. 7, 151. 218. Miyakawa, Y., Yamada, A , , Kosaka, K., Tsuda, F., Kosugi E., and Mayumi, M. (1981). Defective immune-adherence (C3h) receptor on erythrocytes from patients with systemic lupus erythematosus. Luncet 2, 493. 219. Iida, K., Mnrnaghi, R., and Nussenzweig, V. (1982). Complement receptor (CRJ deficiency in erythrocytes from patients with systemic lupus erythematosus. J. E x p . Med. 155, 1427. 220. Wilson, J. G . , Wong, W. W . , Schur, P. H., and Fearon, D. T. (1982). Mode of inheri-

370

221.

222.

223. 224.

225.

226.

227.

228. 229.

230.

231.

232. 233.

234.

235.

236. 237.

ARGYRIOS N. THEOFILOPOULOS AND FRANK J . DIXON

tance of decreased C3b receptors on erythrocytes of patients with systemic lupus erythematosus. New Engl. J. Med. 307, 981. Ross, G. D., Walport, M . J., Parker, C. J., Lentine, A. F., Fuller, C. R.,Yount, W. J., Myones, B. L., Winfield, J. B., and Lachmann, P. J. (1984). Acquired loss of erythrocyte CRL (C3b-receptor) in systemic lupus erythematosus and other diseases with autoantibodies and/or complement activation. Arthritis Rheum. 27(Suppl. 4), S28 (Abstr.). Medof, M. E., Iida, K., Mold, C., and Nussenzweig, V. (1982). Unique role of‘ the complement receptor CRI in the degradation of C3b associated with immnne complexes. J. Exp. Med. 156, 1739. Medof, M. E., and Nussenzweig, V. (1984). Unique role ofthe complement receptor CR, in the degradation of C3b associated with immune complexes. J. Exp. Med. 156, 1739. Frank, M. M., Hamburger, M. I., Lawley, T. J . , Kimberly R. P., and Plotz, P. H. (1979). Defective reticuloendothelial system Fc-receptor function in systemic lupus erythematosus. New Engl. J. Med. 300, 518. Kimberly, R. P., Gibofsky, A., Salmon, J. E., and Fontino, M. (1983). Impaired Fcmediated mononuclear phagocyte system clearance in HLA-DR2 and MT1-positive healthy young adults. J. E x p . Med. 157, 1698. Lawley, T. J., Hall, R. P., Fauci, A. S.,Katz, S. I., Hamburger, M. I., and Frank, M. M. (1981). Defective Fc-receptor functions associated with the HLA-B8./DRW3 haplotype: Studies in patients with dermatitis herpetiformis and normal subjects. New Engl. J. Med. 304, 185. Subbarao, B., Mosier, D. E., Ahmed, A., Mond, J. J., Scher, I., and Paul, W. E. (1979). Role of a nonimmunoglobulin cell surface determinant in the activation of B lymphocytes by thymus-independent antigens. J . Exp. Med. 149, 495. Manohar, V., Brown, E., Leiserson, W. M., and Chused, T. M. (1982). Expression of Lyt-1 by a subset of B lymphocytes. J . Zmmunol. 129, 532. Hayakawa, K., Hardy, R. R., Parks, D. R., and Herzenberg, L. A. (1983). The “Ly-1 B” cell subpopulation in normal, imniunodefective, and autoimmune mice. J. Exp. Med. 157, 202. Smith, H. R., Chused, T. M., Smathers, P. A., and Steinberg, A. D. (1983). Evidence for thymic regulation of autoimmunity in BXSB mice: Acceleration of disease by neonatal thymectomy. J. Zmmunol. 130, 1200. Ledbetter, J. A., Rouse, R. V., Micklem, H. S., and Herzenberg, L. A. (1980). T cell subsets defined by expression of Lyt-I, 2, 3 and the Thy-1 antigens: Two parameter immunofluorescence and cytotoxicity analysis with monoclonal antibodies modifies current views. J . Exp. Med. 152, 280. Lanier, L. L., Warner, N. L., Ledbetter, J. A., and Herzenberg, L. A. 1981. Expression of Lyt-1 antigen on certain inurine B lymphomas. J. Exp. Med. 153, 998. Hayakawa, K., Hardy, R. R., Honda, M., Herzenberg, L. A., Steinberg, A. D., and Herzenberg, L. A. (1984). Ly-1 B cells: Functionally distinct lymphocytes that secrete IgM autoantibodies. Proc. Natl. Acad. Sci. U.S.A. 81, 2494. Okumura, K., Hayadawa, K. ,and Tada, T. (1982). Cell-to-cell interaction controlled by immunoglobulin genes. Role of Thy-I-, Lyt-I+, Ig+ (B’) cell in allotype-restricted antibody production. J. E x p . Med. 156, 443. Shimamura, T., Habu, S., Hashimoto, K., and Sasaki, S. (1984). Feedback suppression of the immune response in oioo. 111. Lyt-l+ B cells are suppressor-inducer cells. Cell. Zmmunol. 83, 221. Waksman, B. H., Raff, M. C., and East, J. (1972). T and B lymphocytes in New Zealand Black mice. Clin. Exp. Irnrnunol. 11, 1. Cantor, H., McVay-Boudreau, B., Hugenberger, J., Naidorf, J., Shen, F. W., and Gershon, R. K. (1978). Immunoregulatory circuits among T-cell sets. 11. Physiologic role of feedback inhibition in oioo: Absence in N Z B mice. /. Exp. Med. 147, 1116.

MUHINE MODELS OF SLE

37 1

238. Lewis, D. E., Giorgi, J. V., and Warner, N. L. (1981). Flow cytometry analysis ofT cells and continuous T-cell lines from autoimmune MRLll mice. Noture (London) 289, 298. 239. Dumont, F., and Habbersett, R. G. (1982). Unusual cell surface properties of the T lymphocyte population expanding in MRL/Mp-/pr/lpr mice. Itmtiunology 47, 271. 240. DeCoursey, T. E., Chandy, K. G., Fischbach, M., Talal, N., Cahalan, M. D., and Gupta, S. (1984). Digerences in ion channel expression in T lymphocytes from MRL-lpr and MRL-+I+ mice. Fed. Proc., Fed. A m . Soc. E x p . Biol. 43, 1736 (Abstr.). 240a. Katagiri, T., Mori, T., Nakano, T., Veno, K., and Kano, K. (1984). Aberrant expression of Forssman and Paul-Bunnell antigens on lymph node cells of MRL/Mp-/pr//pr mice. J . Immunol. 133, 3143. 241. Morse, H. C., 111, Davidson, W. F., Yetter, R. A., Murphy, E. D., Roths, J. B., and Coffinan, R. L. (1982). Abnormalities induced by the mutant gene Ipr: Expansion of a unique lymphocyte subset. J. Immunol. 129, 2612. 241a. Singer, P., and Theofilopoulos, A. N. (1984). In preparation. 242. Davidson, W. F., Roths, J. B., Holmes, K. L., Rudikoff, E., and Morse, H. C., 111 (1984). Dissociation of severe lupus-like disease from polyclonal B cell activation and IL2 deficiency in C3H-/pr//pr mice. J . Immunol. 133, 1048. 243. Dumont, F. J., Habbersett, R. C., Nichols, E. A., Trefinger J. A , , and Tung, A. S. (1983). A monoclonal antibody (l00C5) to the Lyt-2- T cell population expanding in MRL/Mp-/pr//pr mice detects a surface antigen normally expressed o n Lyt-2+ cells and B cells. Eur. J. Immunol. 13, 455. 244. Melino, M. R., Dumont, F. J . , Hahbersett, R. C., and Hensen, T. H. (1983). Expression of a hone marrow-associated Ly-6 determinant on the T cell population expanding in the lymph nodes of the autoimmune mouse strain MRL/Mp-lpr//pr. J . Ittimtrnol. 130, 1843. 244a. Takei, F.(1984). Unique surface phenotype of T cells in lymphoproliferative autoimmune MRL/Mp-lpr//pr mice. J . Zmtnunol. 133, 1951. 245. Dumont, F. J., Habbersett, R. C . , and Nichols, E. A. (1984). A new lymphocyte surface antigen defined by a monoclonal antihody (9F3) to the T cell population expanding in MRL/Mp-lpr//pr mice. J. Zinttiund 133, 809. 246. Wofsy, D.. Hardy. R. R., and Seaman, W. E. (1984). The proliferating cells in autoinimune MRLllpr mice lack L3T4, an antigen on “helper” T cells that is involved in the response to class I1 major histocompatibility antigens. 1. Immunol. 132, 2686. 246a. Mountz, J. D., Steinberg, A. D., Klinman, D. M., Smith, H. R., and Mushinski, J. F. (1984). Autoimmunity and increased c-myb transcription. Science 226, 1087. 247. Moutsopoulos, H. M., Boehm-Truitt, M . , Kassan, S. S . , and Chused, T. M. (1977). Demonstration of activation of B lymphocytes in New Zealand Black mice at birth by an immunoradiometric assay for murine IgM. J. Immunol. 119, 1639. 248. Manny, N . , Datta. S. K., and Schwartz, R. S. (1979). Synthesis of IgM by cells of NZB and SWR mice and the crosses. J. Znimunol. 122, 1220. 249. Izui, S., McConahey, P. J., and Dixon, F. J. (1978). Increased spontaneous polyclonal activation of B lymphocytes in mice with spontaneous autoimmune disease. J. Zmmunol. 121, 2213. 250. Slack, J. H., Hdng, L. M., Atnundson, C., Clevinger, B. L., and Dixon F. J. (1984). Spontaneous initogen and antigen induced diversity of murine lambda 1variable region. J. Imnirrnol. (submitted). 251. Raveche, E. S., Alabaster, O., Tjio, J. H., Taurog, J., and Steinberg, A. D. (1981). Analysis of NZB hyperdiploid spleen cells. J. Ztnmunol. 126, 154. 2 5 1 ~ .Ohsugi. Y. arid Gershwin, M. K. (1979). Studies of mngcntially inimunnlogic mutant New Zealand mice. 111. Growth of B lymphocyte clones in congenitally athymic (nude) and hereditarily asplenic (Dh/ +) mice: A primary B cell defect. J . Imftlunol. 123, 1260.

372

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

252. Kincade, P. W., Lu, G . , Fernandes, G . , Moore, M. A. S., Williams, N . , andGood, R. A. (1979). Abnormalities in clonable B lymphocytes and myeloid progenitors in autoimmune NZB mice. Proc. Natl. Acad. Sci. U . S . A . 76, 3464. 253. Jyonouchi, H . , Kincade, P. W., Good, R. A,, and Gershwin, M. E. (1983). B lymphocyte lineage cells in newborn and very young NZB mice: Evidence for regulatory disorders affecting B cell formation. J. Zmmunol. 131, 2219. 254. Jyonouchi, H., and Kincade, P. W. (1984). Precocious and enhanced functional maturation of B lineage cells in New Zealand Black mice during embryonic development. J. Exp. Med. 159, 1277. 255. Howard, M., and Paul, W.E. (1983). Regulation of B cell growth and differentiation by soluble factors. Annu. Reo. Zmmunol. 1, 307. 256. Parker, D. (1980). Induction and suppression of polyclonal antibody responses by anti-Ig reagents and antigen-nonspecific helper factors: A comparison of the effects of anti-Fab, anti-IgM, and anti-IgD in murine B cells. Zmmunol. Reo. 52, 115. 257. Pike, B., Vaux, D., Clark-Lewis, I., Schrader, J., and Nossal, G. (1982). Proliferation and differentiation of single hapten-specific B lymphocytes is promoted by T cell factor(s) distinct from T cell growth factor. Proc. Natl. Acad. Sci. U.S.A. 79, 6350. 258. Zubler, R. H . , and Glasebrook, A. L. (1982). Requirements for three signals in “Tindependent” (lipopolysaccharide-induced) as well as T-dependent B cell responses. J . E x p . Med. 155, 666. 259. Andersson, J., and Melchers, F. (1981). T cell-dependent activation of resting B cells. Requirement for both nonspecific unrestricted and antigen-specific Ia-restricted soluble factors. Proc. Natl. Acad. Sci. U.S.A. 78, 2497. 260. Melchers, F . , and Andersson, J. (1984). B cell activation: The steps and their variations. Cell 37, 715. 261. Snow, E. C., Vitetta, E. S., and Uhr, J. W. (1983).Activation ofantigen-enriched B cells. I. Purification and response to thymus-independent antigens. J . Zmmunol. 130, 607. 262. Snow, E. C., Noelle, R. J., Uhr, J. W., and Vitetta, E. S. (1983). Activation ofantigenenriched B cells. 11. Role of linked recognition in B cell proliferation to thymusdependent antigens. J. Zmmunol. 130, 614. 263. Roehm, N. W., Marrack, P., and Kappler, J. W. (1983). Helper signals in the plaqueforming cell response to protein-bound haptens. J. E x p . Med. 158, 317. 264. Singer, A . , Asano, Y., Shigeta, M., Hathcock, K. S., Ahnied, A., Fathman, C. G., and Hodes, R. (1982). Distinct B cell subpopulations differ in their genetic requirements for activation by T helper cells. Zmmunol. Reu. 64, 137. 265 Swain, S. L., Wetzel, G. D., Soubiran, P., and Dutton, R. W. (1982). T cell replacing factors in the B cell response to antigen. Zmmunol. Rev. 63, 111. 265a. Zubler, R. H . , Lowenthal, J. W., Erard, F., Hashimoto, N., Devos, R . , and MacDonald, H . R. (1984).Activated B cells express receptors for, and proliferate i n response to, pure interleukin 2. J. Exp. M e d . 160, 1170. 266. Pure, E., Isakson, P. C., Kappler, J. W., Marrack, P., Krammer, P. H . , andvitetta, E. S. (1983). T cell-derived B cell growth and differentiation Factors. J. E x p . Med. 157, 600. 267. Isakson, P. C., Pure, E., Vitetta, E. S., and Krammer, P. H. (1982).T cell-derived B cell differentiation fiactor(s).Effect on the isotype switch of niurine B cells. J. E x p . Med. 155, 734. 268. Mayer, L., Shu, M.-F., and Kunkel, H. G . (1982). Human T cell hybridomas secreting factors for IgA-specific help, polyclonal B cell activation, and B cell proliferation. J . E x p . Med. 156, 1860. 269. Kishimoto, T., and Ishizaka, K. (1975). Immunologic and physicochemical properties of enhancing soluble factors for IgG and IgE antibody responses. J . Imrnunol. 114, 1177.

M U R I N E MODELS OF SLE

373

270. Kehrl, J. H., and Fauci, A. S . (1983). Identification, purification, and characterization of antigen-activated and antigen-specific human B lymphocytes. J . E x p . Med. 157, 1692. 271. Prud’homme, G. J., Balderas, R. S . , Ilixon, F. J., andTheofilopoulos, A. N. (1983). B cell dependence on and response to accessory signals in murine lupus strains. J. Exp. Med. 157, 1815. 272. Pike, B. L., Layton, J. E . , Rose, N. R., and Nossal, G. J. V. (1981). Surface imniunogkhdin isotype, activation, and tolerance susceptibility of B lympocytes of New Zealand Black mice. Cell. Zmmunol. 63, 385. 273. Scher, I. (1982). CBA/N immune defective mice; evidence for the failure of a B cell subpopulation to be expressed. Zmmunol. Hew. 64, 117. 274. Taurog, J. D., Moutsopoulos, M., Rosenberg, Y. J., Chused, T. M., and Steinberg, A. I>. (1979). CBA/N X-linked B-cell defect prevents NZB B-cell hyperactivity in F I mice. J . E x p . Med. 150, 31. 275. Steinberg, B. J., Smathers, P. A., Frederiksen, K., and Steinberg, A. D. (1982).Ability of the .rid gene to prevent autoimmunity in (NZBxNZW)Fl mice during the course of their natural history, after polyclonal stimulation, or following immunization with DNA. J. Clin. Invest. 70, 587. 276. Steinberg, A. D., Huston, D. P., Taurog, J. D., Cowdery, J. S . , a d Raveche, E. S. (1981). The cellular and genetic basis of murine lupus. Imrnunol. Rev. 55, 121. 277. Steinberg, E. B., Santoro, T. J., Chused, T. M., Smathers P. A., and Steinberg, A. 11. (1983). Studies of congenic MRL-/pr/lpr.xid mice. J. Zmmunol. 131, 2789. 278. Smith, H. R . , Chused, T. M., and Steinberg, A. 11. (1983). The effect of the x-linked immune deficiency gene (xid) upon the Y chromosome-related disease of BXSB mice. J . Immunol. 131, 1257. 279. Golding, B., Golding, H., Foiles, P. G., and Morton, J. I. (1983).CBA/N X-linked defect delays expression of the Y-linked accelerated autoimmune disease of BXSB mice. J . Zmmunol. 130, 1043. 280. Cowdery, J. S . , Jr., Taurog, J. D., and Steinberg, A. D. (1980). Effect of CBA/N xid on spontaneous production of antibodies to DNA in MRL/I and NZB backcross mice. Scond. J. Immunol. 12, 499. 281. Kemp, J.. Cowderey, J. S., Steinberg, A. D., and Gershon. R. (1982). Genetic control of autoimmune disease: Interactions between xid and lpr. J. Imniunol. 128, 392. 282. Pisetsky, D. S . , Caster, S . A,, and Steinberg, A. D. (1983). Effect of xidon anti-DNA Bcell precursors. Cell. Zntnzunol. 78, 326. 283. Kemp, J , , Dohrman, L., Jayawardena, A,, Huber, B., and Gershon, R. K. (1980). B cell subsets in autoimmunity: An autoimmune hemolytic response to mouse erythrocytes is modulated by the DBA/N defect and anti-Lyb3 antisera. Cell. Znzmunol. 53, 84. 284. Nakajima, P. B., Datta, S . K., Schwartz, R. S., and Huber, B. T. (1979). Localization of spontaneously hyperactive B cell of NZB mice to a specific B cell subset. Pmc. Nutl. Acud. Sci. U . S . A . 76, 4613. 285. Smathers, P. A,, Steinberg, B. J . , Reeves, J. P . , and Steinberg, A. D. (1982). Effects of polyclonal immune stimulators upon NZB.xid congenic mice. J . Znimunol. 128, 1414. 286. Romain, P. L., Cohen, P. L., Fish, F., Ziff, M., and Vitetta, E. S. (1980). The specific B cell subset lacking in the CBA/N mouse is not required for the production of autoantibody in (CBA/NXNZB)F1 male mice. J. Zmmtrnol. 125, 246. 287. Ohsugi, Y., Gershwin, M. E., Ahmed, A , , Skelly, H.R . , and Milich, I>. R. (1982).Studies of congenitally immunologic mutant New Zealand Mice. VI. Spontaneous and induced autoantibodies to red cells and DNA occur in New Zealand x-linked immunodeficient (wid) mice without phenotypic alterations of the xid gene or generalized polyclonal B cell activation. J . Zmmunol. 128, 2220. 288. Delovitch, T. L., Kaufinan, K., Gorczynski, R. D., and Sinclair, G. 11. (1983). In oitro

374

289.

290. 291. 292.

293. 294. 295.

296. 297.

298. 299.

300. 301. 302. 303.

304.

305.

306.

307.

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

analysis of allogeneic lymphocyte interaction. VIII. Characterization of helper cotnponents of allogeneic effect factor (AEF) that activate Lyb5+ and Lyb5- B cells to respond to thymus-dependent and thymus-independent antigens. J . Zmwunol. 131, 2246. Fieser, T. M., Gershwin, M. E., Steinberg, A. D., Dixon, F. J., and Theofilopoulos, A. N. (1984).Abrogation of murine lupus by the xid gene is associated with reduced responsiveness of B cells to T cell helper signals. Cell. Zmmunol. 87, 708. DeFranco, A. L., Kung, J. T., and Paul, W. E. (1982).Regulation ofgrowth and proliferation in B cell subpopulations. Zmmunol. Reo. 64, 161. Goodman, M. G., and Weigle, W. 0. (1981). Role of polyclonal B-cell activation in self/non-self discrimination. Zmniunol. Today 2, 54. Izui, S . , Lambert, P.-H., Fournie, G. J., Turler, H., and Miescher, P. A. (1977). Features of systemic lupus erythematosus in mice injected with bacterial lipopolysaccharide. Identification of circulating DNA and renal localization of DNA-anti-DNA complexes. 1. E x p . Med. 145, 1115. Izui, S., Eisenberg, R. A,, and Dixon, F. J. (1979). IgM rheumatoid factors in mice injected with bacterial lipopolysaccharides. J . Zmmunol. 122, 2096. Dresser, D. W. (1978). Most IgM-producing cells in the mouse secrete auto-antibodies (rheumatoid factor). Nature (London) 274, 480. Izui, S., Kobayakawa, T., Louis, J., and Lambert, P.-H. (1979). Induction of thymocytotoxic autoantibodies after injection of bacterial lipopolysaccharides in mice. Eur. ]. Zmmunol. 9, 338. Meredith, P. J., Kristie, J. A , , and Walford, R. L. (1979). Aging increases expression of LPS-induced autoantibody-secreting B cells. J. Zmrnunol. 123, 87. Hang, L. M., Slack, J. H . , Amundson, C., Izui, S . , Theofilopoulos, A.N., and Dixon, F. J. (1983).Induction of murine autoimmune disease by chronic polyclonal B cell activation. 1. E x p . Med. 157, 874. Elkerbout, E. A., and Hijmans, W. (1974). The long-term antibody response of New Zealand mice to sheep red blood cells. Immunology 26, 893. McKearn, J. P., Miller, G. W., and Quintans, J. (1978). The immune response in NZB mice of different ages to thymus-dependent and thymus-independent phsophorylcholine antigen. Znimunology 34, 1063. Baum, J. (1969).Increased 7s antibody response to sheep erythrocytes in the 2-month-old NZB mouse. Clin. Exp. ~mrnunol.5, 251. Playfair, J. H. L. (1968). Strain differences in the immune response of mice. I. The neonatal response to sheep red cells. Immunology 15, 35. Salomon, J. C., and Benveniste. J. (1969). The immnne response in NZBXNZW F, hybrid mice. Clin. E x p . Zmrnunol. 4 , 213. Evans, M. M., Williamson, W. G . , and Irvine, W. J. (1968). The appearance of immunological competence at an early age in New Zealand Black mice. Clin. E x p . Zmnmunol. 3, 375. Naor, D., Bonavida, B . , Robinson, R. A., Shibata, I. N., Percy, D. E., Chia, D., and Barnett, E. V. (1976). Immune response of New Zealand mice to trinitrophenylated syngeneic mouse red cells. E u r . J . Zmmunol. 6, 783. Park, C. L., Balderas, R. S . , Fieser, T. M., Slack, J. H ., Prud’homme, G . , Dixon, F. J., and Theofilopoulos. A. N. (1983). Isotypic profiles and other fine characteristics of immime responses of SLE mice to exogenous TI)- and TI-antigens. J. Zmmunol. 130, 2161. Cowdery, J. S., and Steinberg, A. D. (1982).Regulation of primary, thymus-independent, anti-hapten responses of normal and autoimmune mice by syngeneic antibody. J . Immunol. 129, 1250. Mongini, P. K. A., Paul, W. E., and Metcalf, E. S. (1982). T cell regulation of immu-

M U R I N E M O D E L S OF S L E

308.

309. 310.

311. 312.

313.

314.

315.

316.

317.

318. 319.

320.

321.

322. 323.

324.

375

noglobtihi class expression in the antibody response to trinitrophenyl-ficoll. Evidence fbr T cell enhancetnent of the iinmiinoglol)uliii class switch. J . Exp. Med. 155, 884. Mongini, P. K. A , , and Paul, W. E. (1982). T cell regulation of the IgC2a response to TNP-Ficoll: Evidence that allotype congenic mice contain both helper cells that preferentially enhance IgG2a synthesis and suppressor cells that specifically suppress IgG2a synthesis. J. Zmniunol. 128, 2405. Augustin, A. A,, and Coutinho, A. (1980). Specific T helper cells that activate B cells polyclonally. In oitru enrichment and cooperative fiinction. J. Exp. Med. 151, 587. IIeFranco, A. L., Ashwell, J. D., Schwartz, H. H., and Paul, W. E. (1984). Polyclonal stimulation of resting B lymphycotes by antigen-specific T lymphocytes.J . Exp. Med. 159, 861. Krakauer, R . S., Waldmann, T. A., and Strober, W. (1976). Loss of sl~ppressorT cells in adult NZB/NZW mice. J . Exp. Metl. 144, 662. Krakauer, H. S., Strober, W., Hippeon, D. L., and Waldmann, T. A. (1977).Prevention of autoimmune disease in experimental systemic lupiis erythematosus by the administration of soluble immune response suppressor. Science 196, 56. Steinberg, A. D., Krakaner, H., Reinertsen, J. L., Klassen, L. W.. Infeld, I)., Reeves, J. P., Williams, 6. W., and Antonovych. T. (1978). Therapeutic studies i n NZBlNZW mice. VI. Age-dependent effects of concanavah A stiinnlated spleen cell supernate. Arthritis Rheuni. 21, 204. Cantor, H . , and Gershon, H. K. (1979).Immunological circuits: Cellular composition. Fed Proc., Fed. Am. Soc. Ex},. R i d . 38, 2058. Breshnihan, B., and Jasin, H. E. (1977). Suppressor function of peripheral blood mononuclear cells in normal individuals and patients with systemic lupus erytheinatosns. ]. C h i . Inoest. 59, 106. Abtlou, N . I., Sagawa, A,, Pascual, E., Hebert, J., and Sadeghee, S. (1976). Suppressor T eel1 al)normality in idiopathic systemic lupus erytheinatosus. Cfin. Inmunot. Imnitrnoputhol. 6, 192. Sakane, T.. Steinberg, A. I>., and Green, I. (1978). Studies of immune hnctions of patients with systemic lupus erythematosus. I. Failure of suppressor T cell activity related to impaired generation of, rather than response to, srippressor cells. Arthritis Rheutn. 21, 657. Sagawit, A , , and Abdou, N. I. (1978). Snppressor-cell dysfunction in systemic lupus erythematosus. J . C h . Intiest. 62, 789. Alarcon-Segovia, D., arid Huiz-Arguelles, A. (1978). Decreased circulating thymus-derived cells with receptors for the Fc portion of iiiiinuiio~lobulirrG in systemic lupus erytheinatosus. J . Clin. Inoest. 62, 1390. Creighton, W. D., Katz, I). H., and Dixon, F. J. (1979). Antigen-specific iinmrrnocompetency, B cell function, and regulatory helper and suppressor T cell activities in spontaneously autoimmune mice. J . I t n n ~ u t i o l .123, 2627. Hoder, J. C., Bell, A. D., and Singhal, S. K. (1977). Hegulation ofthe immune response in aritoimmune NZB/NZW F, mice. I. The sptmtaneoiis generation of splenic suppressor Cells. Cell. I??L?>lUttfJ/.29, 272. Tsokos, C. C., and Balow, J. E. (1983). Cellular immune responses i n systemic lupus erytheniatosus. C h i . Zmnunol. Inwmunopathol. 26, 267. Nakamura, Z.-I., Asano, T., Yano, K . , and Oftiji, T. (1982). Heevaluation of slippressor cell fiinction in systemic hipus erythematosns. <:/in. ~ ! ~ I ~ I I I It,lrllfitlo}Jathol. ~ ~ J / . 24, 72. Okada, J., and Stastny, P. (1981). Antibodies against T cells, B cells and nlonocytes in patients with systemic lupus erythematosus. Arthritis Hheutn. 24, S76 (Abstr.).

376

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

325. Prudhoinnie, G . J., Park, C. L., Fieser, T. M . , Kofler, R., Dixon, F. J., and Theofilopoulos, A. N. (1983). Identification of a B-cell differentiation Cactor(s) spontaneously produced by proliferating T cells in murine lupus strains of lpr/Lpr genotype. J. E x p . Med. 157, 730. 326. Sawada, S., and Talal, N . (1979). Evidence for a helper cell promoting anti-DNA antibody production in murine lupus. Arthritis Aheum. 22, 655 (Abstr.). 327. Fisher, C. L., Eisenberg, R. A , , and Cohen, P. L. (1984). Antichromatin autoantibodies in SLE mice: Subclass distribution and epidemiology. Fed. Proc., Fed. Am. Soc. E x p . Biol. 43, 1833 (Abstr.). 328. Smith, K. A. (1980). T-cell growth factor. Znzmunol. Rev. 51, 337. 329. Cantrell, D. A,, and Smith, K. A. (1983). Transient expression of interleukin 2 receptors. Consequences for T cell growth. J. Exp. Med. 158, 1895. 330. Robb, R. J., Munck, R. A., and Smith, K. A. (1981). T cell growth factor receptors. Quantitdion, specificity, and biological relevance. J . E x p . Med. 154, 1455. 331. Smith, K. A., and Ruscetti, F. W. (1981). T cell growth factor and the culture of cloned functional T cells. Ado. Zmniunol. 31, 137. 332. Altnian, A,, Theofilopoulos, A. N . , Weiner, R., Katz, D. H., and Dixon, F. J. (1981). Analysis of T cell function in autoimmune murine strains. Defects in production of, and responsiveness to, interleukin 2.1. E r p . Med. 154, 791. 333. Dauphinee, M. S., Kipper, S . B., Wofsy, D., andTalal, N . (1981). Interleukin2 deficiency is a common feature of autoimmune mice. J . Intmunol. 127, 2483. 334. Wofsy, D., Murphy, E. D., Roths, J. B., Dauphinee, M. J . , Kipper, S. B., and Tala1 N. (1981). Deficient iiiterleukin 2 activity in MRLiMp and C57BL/GJ mice bearing the lpr gene. J. Exp. Med. 154, 1671. 335. Santoro, T. J., Luger, T. A,, Raveche, E. S., Smolen, J. S . , Oppenheim, J. J., and Steinberg, A. D. (1983). In uitro correction of the interleukin 2 defect of autoimmune mice. Eur. J . Zminunol. 13, 601. 336. Santoro, T. J., Malek, T. R., Rosenberg, Y. J., Oppenheim J. J . , and Steinberg, A. D. (1984). The induction of interleukin 2 (IL2) responsiveness in autoimmune MRL-!pr/lpr mice. Fed. Proc., Fed. Am. Soc. E x p . B i d . 43, 1736 (Abstr.). 337. Hefeneider, S. H., Conlon, P. J., Dower, S. K., Henney, C. S., and Gillis, S. (1984). Limiting dilutiou analysis of interleukin 2 and colony-stimulating factor producer cells in normal and autoimmune mice. J . Imwunol. 132, 1863. 337a. Fieser, T., Ihle, J. N., Dixon, F. J . , and Theofilopoulos, A. N . (1984). In preparation. 338. Simon, M. M., Prester, M., Nerz, G., and Kuppers, R. C. (1984). Quantitative studies on T cell functions in MRL/Mp-Zpr/lpr mice. I. Frequency analysis of precursor cells of proliferating, cytotoxic and T cell growth factor-secreting T lymphocytes reveals an increase in absolute numbers with a concomitant decrease in percentage of immuno competent cells. Clin. Itmunol. Iniinunopnthol. 33, 39. 339. Bocchieri, M. H., Knithveis, L., and Seaton, D. S. (1984). Cytokine production by NZB, C58 and NZB and C58 recombinant inbred mice. Cell. Zmmunol. 88, 453. 340. Alcocer-Varela, J., and Alarcon-Segovia, D. (1982). Decreased production and response to interleukin-2 by cultured lymphocytes from patients with systemic lupus erythematosus. J. Clin. Znoest. 69, 393. 341. Linker-Israeli, M . , Bakke, A. C., Kitridou, R. C., Gendler, S., Gillis, S., and Horwitz, D. A. (1983). Defective production of interleukin 1and interleukin 2 in patients with systemic lupus erythematosus (SLE). J. Zmtnunol. 130, 2651. 342. Nakainura, T., Miyaraka, N., Talal, N., and Russell, I. J. (1982). Interleukin-2 related regulatory defects in autoimmune disease. Arthritis Rheum. 25, S83 (Abstr.).

M U R I N E MODELS OF SLE

377

343. Thomiin, M. L., and Weigle, W. 0.(1982). Cell-mediated immunity in aged mice: A n underlying lesion in IL2 synthesis. 1. Iwnunol. 128, 2358. 344. Gillis, S . , Kozak, R . , Durante, M . , and Weksler, M. (1981). Immunological studies of aging. Decreased production of and response to T cell growth factor by lymphocytes froin aged humans. I . Clin. Itaoest. 67, 937. 345. Flomenberg, N . , Welte, K., Mertelsinann, K . , Kertnan, N . , Giohanu. N . , Venuta, S., Feldman, S., Kruger, G . , Kirkpatrick, D., Dupont, B., and O’Reilly, R. (1983). Immunologic effects of interleukin 2 in primary iiiimiiiiodeficiency diseases. I . Intinuno/. 130, 2644. 346. Ciobanu, N . , Qwelte, K., Kruger, G . , Vemita, S., Cold, J . , Feldman, S. P., Wang, C. Y., Koziuer, B., Moore, M. A. S.. Safai, B., and Mertelsmann, R. (1983). Defective T-cell response to PHA and mitogenic monoclonal anti1)odiesin inale homosexuals with acquired iminunodeficiency syndronie mid its in citro correction by interleukin 2. /. Clin. Ztnmunol. 3, 332. 347. Reiner, N . E., and Finke, J. H. (1983). Interleukin 2 deficiency in murine Leishrnuniusis donooani and its relationship to depressed spleen cell responses to phytoliemagglutiniii. 1. Immunol. 131, 1487. 348. Taniguchi, T., Matsui, H., Fujita, T., Takaoka, C., Kashima, N . , Yosliinioto, R . , and Hamuro, J. (1983). Structure and expression of a cloned cDNA for human interleukin-2. Nutitre ( ~ ~ n d o t302, i ) 305. 349. Fujita, T., Takaoka, C., Matsui. H., and Taniguchi, T. (1983). Structure of the huinan interleukin 2 gene. Proc. Nutl. Acad. Sci. U . S . A . 80, 7437. 350. Holl)rook, N. J., Smith, K. A,, Fornace, A. J . , Coineau, C. M., Wiskocil, H . L., and Crabtree, G . R . (1984).T cell growth fkictor: Complete nucleotide sequence and organiziition of the gene in normal and malignant cells. Pnrc. N n d . Acud. Sci. U . S . A . 81, 1634. 351. R o b l , R. J., Kutny, R. M., and Chowdry, V. (1983). Purification and partial sequence analysis of huinan T-cell growth factor. Proc. Nutl. Acud. Sci. U.S.A. 80, 5990. 352. Smith, K. A , , Favata, M. F . , and Oroszlan, S . (1983). Production and characterization of inonoclonal antibodies to human interleukin 2: Sti-ategy and tactics. 1. Zmmunol. 131, 1808. 353. Altman, A., Cardenas, J. M., Houghten, R. A , , Ilixon, F. J., and Theofilopoulos, A. N . (1984). Antibodies to human interleukin-2 inducetl by syllthetic peptides. Proc. N u t / . Acad. Sci. U.S.A. 81, 2176. 354. Ihle, J. N . , Keller. J., Henderson, L.. Klein, F., and Palaszynski, E. (1982). Procedures for the purification of interleukin 3 to homogeneity. /. Ittmunol. 129, 2431. 355. Ihle, J. N . , Rebar, L., Keller, J., Lee, J. C., and Hapel. A. (1981). Interlellkin 3: Possilde roles in the regulation of lymphocyte differentiation and growth. Immunol. Rev. 98, 101. 356. Palacios, R . , Henson, C., Steinmetz, M., and McKearn, J. P. (1984). Interleukiii-3 supports growth of mouse pre-B-cell clones in oiiro. Nature (London) 309, 126. 357. Palacios, R. (1984). Spontaneous production of interleukin 3 by T lyinpliocytes from autoiinniune MRLIMp-lprllpr mice. Eur. /. Imnruno/. 14, 599. 358. Friedman, R. M., and Vogel, S. N. (1983). Interferons with special emphasis on the immune system. Ado. Imniutiol. 34, 97. 359. Schreiber, R. D., Pace. J. L., Russell, S. W.. A h a n , A . , and Katz, D. H. (1983). Macrophage-activatiiig factor produced by a T cell hyhridoma. Physiochoclieniical and biosynthetic resemblance to gamma-interferoli. 1. Znlrnund. 131, 826. 360. Zlotnik, A , , Roberts, W. K . , Vasil, A., Bluinenthal. E., Larosa, F . , Leihson, H . J.. Endres, R . O., Graham. S . D.,White, J., Hill, J . , Henson, P., Klein, J. R . , Bevan, M. J., Marrack, P . , and Kappler, J. W.(1983). Coordinate production by a T cell hybridoina of

378

AHGYRIOS N . THEOFILOPOULOS A N D PRANK J. DIXON

gamma-interferon on three other lymphokine activities: Multiple activities of a single lymphokine? 1. Zmmunol. 131, 794. 361. Klein, J . R . , and Bevan, M. J. (1983). Secretion of immune interferon and generation of cytotoxic T cell activity in nude mice are dependent on interleukin 2: age-associated endogenous production of interleukin 2 in nude mice. 1. Zmmunol. 130, 1780. 362. Torres, B. A,, Farrar, W . L., and Johnson, H. M. (1982). Interleukin 2 regulated immune interferon (IFN?) production by normal and suppressor T cell cultures. 1.Znimunol. 128,

2217. 363. Leibsnn, J. H., Gefter, M., Zlotnik, A., Marrack, P., and Kappler, J. W. (1984).Role ofyinterferon in antibody-producing responses. Nature (London) 309, 799. 364. Johnson, H. M., and Torres, B. A. (1983). Recombinant mouse interferon-? regulation of antibody production. Infect. Zmnaun. 41, 546. 365. Hooks, J. J., Moutsopoulos, H. M., Geiss, S. A,, Stahl, N. I., Decker, J. L., and Notkins, A. L. (1979). Immune interferon in the circulation of patients with autoimmune diseases. New Engl. 1. Med. 301, 5. 366. Hooks, J. J . , Jordon, G. W., Cnpps, T . , Moutsopoulos, H. M., Fauci, A. S., and Notkins, A. L. (1982). Multiple interferons in the circulation of patients with systemic lupus erythematosus and vasculitis. Arthritis Rheum. 25, 396. 367. Preble, 0. T., Black, R. J., Friedman, R . M., Klippel, J. H., and Vilcek, J. (1982). Systemic lupus erythematosus: Presence in human serum of an unusual acid-labile leukocyte interferon. Science 216, 429. 368. Prehle, 0. T., Rothko, K., Klippel, J. H . , Friedman, R. M., and Johnston, M. I. (1983). Interferon-induced 2'-5' adenylate synthetase in vivo and interferon production in uitro by lymphocytes from systemic lupus erythematosus patients with and without circulating interferon. 1. Exp. Med. 157, 2140. 369. Santoro, T. J., Benjamin, W. R . , Oppenheim, J. J., and Steinherg. A. D. (1983). The cellular basis for immune interferon production in autoimmune MRL-Zprllpr mice. 1. Zmmunol. 131, 265. 370. Kofler, R., Schreiber, R. D., Dixon, F. J., andTheofilopoulos, A. N . (1984). Macrophage Ia expression, production of and response to macrophage-stimulating lymphokines in murine lupus strains. Cell. Immunol. 87, 92. 371. Carpenter, D. F., Steinberg, A. D., Schur, P. H . , and Talal, N. (1970).The pathogenesis of autoimmunity in New Zealand mice. 11. Acceleration of glomerulonephritis by polyinosinic.polycytody1ic acid. Lab. Znuest. 23, 628. 372. Walker, S. E. (1977). Accelerated mortality in young NZB/NZW mice treated with the interferon inducer Tilorone hydrochloride. Clin. Zmmunol. InLmunopathol. 8, 204. 373. Adam, C . , Thoua, Y., Ronco, P . , Verroust, P . , Torey, M., and Morel-Marioger, L. (1980). The effect of exogenous interferon: Acceleration of autoimmune and renal diseases in (NZB/W)FI mice. Clin. E x p . Zmmunol. 40, 373. 374. Heremans, H . , Billiau, A,, Colomhatti, A , , Hilgers, J., and DeSomer, P. (1978). Interferon treatment of NZB mice: Accelerated progression of autoimmune disease. Infect. Zmmun. 21, 925. 375. Wofsy, D., Kerger, C . E., and Seaman, W. E. (1984). Monocytosis in the BXSB model for systemic lupus erythematosus. 1. Erp. Med. 159, 629. 376. Braverman, I. M., and Slesinski, J. (1970). Study of autoimmune disease in New Zealand mice. 1V. Immunologic reactivity and reticuloendothelial function. 1. Znuest. Dermatol. 55, 317. 377. Thomas, H. I . , and Weir, D. M. (1972). The in uitro activity of peritoneal exudate cells from New Zealand Black mice (NZB/BI) in the presence of 125I-BSA (bovine serum albumin). Clin. E x p . Zmmunol. 12, 263. 378. Morgan, A. G., and Steward, M. W. (1976). Macrophage clearance function and immune

MURINE MODELS OF SLE

379

complex disease in New Zealand BlacklWhite F, hybrid mice. Clin. E x p . Immunol. 26, 133. 379. Finbloom, D. S., and Plotz, P. H. (1979). Studies of reticuloendothelial function in the mouse with model immune complexes. 11. Serum clearance, tissue uptake, and reticuloendotlielial saturation in NZB/NZW mice. J. Imniunol. 123, 1600. 380. Morton, 1. I., and Siegal, B. V. (1970). Transplantation of autoimmune potential: IV. Reversal of the NZB autoimmune syndroine by bone marrow transplantation. Proc. Soc. E x p . B i d . 133, 1055. 381. Magilavy, D. B., Rifai, A., and Plotz, P. H. (1981). An abnormality of immune complex kinetics in murine lupus. J. lrnmunol. 126, 770. 382. Hamburger, M . I . , Moutsopoulos, H. M., Lawley, T. J.. and Frank, M. M. (1979). Sjogren’ssyndrome: A defect in reticuloendothelial system Fc-receptor-specific clearance. Ann. Intern. Med. 91, 534. 383. Fries, L.F., Mullins, W. W., Cho, K. R., Plotz, P. H . , and Frank, M. M. (1979). Monocyte receptors for the Fc portion of IgG are increased in systemic lupus erythematoslls. J. rlnmunoi. 132, 695. 384. Katayama, S., Chia, D.. Nasu, H., and Knutson, D. W. (1981). Increased Fc receptor activity in monocytes from patients with rheumatoid arthritis: A study of monocyte binding and catabolism of soluble aggregates of IgG in oitro. J . lrnmunol. 127, 643. 385. Parris, T. M., Kimberly, R. P., Inman, R. D., McDougal, J. S., Gibofsky, A , . and Christian, C. L. (1982). Defective Fc receptor-mediated finiction of the mononuclear phagocyte system in lupus nephritis. Ann. Intern. Met!. 97, 526. 386. Klinman, N . R . , Wylie, D. E., and Teak, J. M. (1981). B cell development. Imniunol. Today 2, 212. 387. Metcalf, E. S . , Schrater, A. F., and Klinman, N . R . (1979). Murine models of tolerance induction in developing and niatuw H cells. Iniinunol. Rel;. 43, 142. 388. LII, C. Y . , Calamai, E. G., and Unanue. E . R . (1979). A defect in the antigen-presenting function of macrophages from neonatal mice. Nature (London) 282, 327. 389. Lu, C:. Y., Beller, D. I., and Unanue, E. R . (1980). During ontogeny Ia-bearing accessory cells are found early in the thymus but late in the spleen. Proc. Nutl. Acad. Sci. U.S.A. 77, 1597. 390. Lri, C. Y . , and Unanue, E. R. (1982). Spontaneous T-cell lyinphokine production and enhanced macrophage Ia expression and tuinoricidal activity i n MRL-/pr mice. Clin. lniniunol. Iniinttnopatho/. 25, 213. 391. Kelley, V. E., and Rotlis, J. 8 . (1982). Increase in macrophage Ia expression in autoimmune mice. Role of the Ipr gene. J. Imniunol. 129, 923. 392. Rosenberg, Y. J , , Steinberg, A. D., and Santoro. T. J. (1984). The basis ofautoimmunity in MHL-lpr/lpr mice: a role for self la-reactive T cells. Inmuno/. Toduy 5 , 64. 393. Wilson, D. A , , and Braley-Mullen, H. (1982). Iininiinoregulatioii in MRL/Mp-lpr/lpr mice: Evidence for decreased helper T-cell and increased snppressor T-cell function with age. Cell. Intnmol. 74, 72. 393a. Fischl~ach,M. (1984). Defective T cell response to presented antigen in autoimmune mice. J. Inzmino/. 133, 2365. 394. Martinez-Ahso, C., and Coutinho, A. (1981). B cell activation hy helper cells is a twostep process. Nature (London)290, 60. 395. Herherman, R. B. (1983). Natural killer cells. In “The Bidogy of Iinmunohgic Disease” (F. J. Dixon and D. W. Fisher, eds.), pp. 75-85. Sinauer, Sunderland, Mass. 396. Abruzzo, L. V . , and Rowley, I). A. (1983). Homeostasis of the antihody response: Immunoregulation by N K cells. Science 222, 581. 397. Karsh, J , , Dorval, G., and Osterland, C. K . (1981). Natural cytotoxicity in rheumatoid arthritis and systemic lupus erythematosus. Cliti. Inmiunol. Jmniunopathol. 19, 437.

380

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

398. Goto, M . , Tanimoto, K., and Horiuchi, Y. (1980). Natural cell mediated cytotoxicity in systemic lupus erythematosus. Suppression by anti-lymphocyte antibody. Arthritis Rheum. 23, 1274. 399. Sibbitt, W. L., Jr., Mathews, P. M., and Bankhurst, A. D. (1983). Natural killer cell in systemic lupus erythematosus. J . Clin. Invest. 71, 1230. 400. Strannegard, O., Hermodsson, S., and Westberg, G. (1982). Interferon and natural killer cells in systemic lupus erythematosus. Clin. Exp. Imrnunol. 50, 246. 401. Tsokos, C. G . , Rook, A. H., Djeu, J. Y., and Balow, J. E. (1982). Natural killer cells and interferon responses in patients with systemic lupus erythematosus. Clirt. E x p . Immunol. 50, 239. 402. Fitzharris, P., Akocer, J., Stephens, H. A. F., Knight, R. A., and Snaith, M. L. (1982). Insensitivity to interferon of NK cells from patients with systemic lupus erythematosus. Clin. Exp. ImmunoE. 47, 110. 403. Neighbor, P. A,, and Grayzel, A. I. (1981). Interferon production in oitrn by leukocytes from patients with systemic lupus erythematosus and rheumatoid arthritis. Clin. Exp. lmmunol. 45, 576. 404. Croker, B. P., Zinkernagel, R. M., and Dixon, F. J. (1979). Natural cytotoxiccells in NZB mice: Spontaneous in oioo and in uitro primed activity and structural characteristics. Clin. lmmunol. lnimunopathol. 12, 410. 405. Lotzova, E., and Gudkowicz H. (1973). Resistance of irradiated F1 hyhrid and allogeneic mice to bone marrow grafts of NZB donors. J . Immunol. 110, 791. 406. Seaman, W. E., Blackman, M. A,, Greenspan, J. S., andTalal, N. (1980). Effect ofHgSron immunity and autoimmunity in NZB/NZW F1 mice. J . Imrnunol. 124, 812. 407. Gelfand, M. C., and Steinberg, A. D. (1973). Mechanisms of allograft rejection in New Zealand mice. I. Cell synergy and its age-dependent loss. J . Immunol. 110, 1652. 408. Stutman, O., Yunis, E.J., and Good, R. A. (1968). Deficient immunologic functions of NZB mice. Proc. SOC. Exp. Biol. Med. 127, 1204. 409. Falkoff, R. M., Scavulli, J. F., and Dutton, R. W. (1978). Independent analysis of the T helper and T killer cell function in young and old NZB mice. J . Immunol. 121, 897. 410. Zinkernagel, R. M., and Dixon, F. J. (1977). Comparison of T cell-mediated immune responsiveness of NZB, (NZBxNZW)F] hybrid and other murine strains. Clin. E x p . Immunol. 29, 110. 411. Botzenhardt, V., Klein, J.. and Ziff, M. (1978). Primary in uitro cell-mediated lympholysis reaction of NZB mice against unmodified targets syngeneic at the major histocompatibility complex. J. Exp. Msd. 147, 1435. 412. Theofilopoulos, A. N., Shawler, D. L., Katz, D. H . , and Dixon, F. J. (1979). Patterns of immune reactivity in autoimmune murine strains I. Cell-mediated immune responses induced by H-2 identical and H-2 incompatible stimulator cells. J. Imrnunol. 122, 2319. 413. Rich, R. R., Sedberry, D. A , , Kastner, D. L., and Chu, L. (1979). Primary in oitro cytotoxic response of NZB spleen cells to Qa-lb-associated antigenic determinants. J . E x p . Med. 150, 1555. 414. Fischer-Lindahl, K., and Hausmann, B. (1980). Qed-I-a target for unrestricted killing by T cells. Eur. J. Immunol. 10, 289. 415. Theofilopoulos, A. N., Shawler, D. L., Balderas, R., Elder, J. H., Katz, D. H., and Dixon, F. J. (1981). Specificities of NZB anti-H-2" CML reactions: Role of Qa-1 and retroviral gp70 antigens. J. Imnwnol. 126, 1154. 416. Davidson, W. F., Chused, T. M., and Morse, H. C . , 111 (1981). Genetic and functional analysis of the primary in oitro CTL: Response of NZB lymphocytes to H-2-compatible cells. lmmunogetietics 12, 445. 417. Stockinger, B., and Botzenhardt N . (1980). On the T cell hyperreactivity of NZB mice

MUHINE MODELS OF SLE

418. 419.

420.

421.

422.

423.

424.

425. 426.

427. 428. 429.

430.

431. 432. 433. 434.

435. 436.

38 1

against H-2-identical cells. Evidence for primary response characteristics and an increased helper potential. J. E x p . Med. 152, 296. Huston, D. P., and Steinberg, A. D. (1980). NZB cytotoxic lymphocyte responses. Kinetic analyses. J . E x p . Med. 152, 748. Huston, D. P., Rich. R. R., Steinherg, A. D., and Triiitt, C. A. (1979). Regulation of cytotoxic lymphocyte responses in New Zealand mice by alloantigea-activated spleen cells. J. Zmmuno/. 122, 1476. Scott, C. F., Jr.. Tsurufuji, M., Lu, C . Y . , Finberg, R., and Sy, M . 4 . (1984). Compdrison of antigen-specific T cell responses in autoiinniune MRL/Mp-lpr/lpr and MRLIMp-+I+ mice. J . Zmmunol. 132, 633. Creighton, W. D., Zinkernagel, R. M . , and Dixon, F. J. (1979). Tcell-mediated immune responses of Inpus-prone BXSB mice and other murine strains. Clin. E x p . Zmmunol. 37, 181. Sakane, T., Steinberg, A. D . , and Green, I. (1978). Failure of autologous mixed lymphocyte reactions between T and non-T cells in patients with systemic lupus erythematosus. Proc. N o d . Acad. Sci. U . S . A . 75, 3464. Glimcher, L. H., Steinberg, A. D., House, S. B., and Green, I . (1980). The autologous mixed lymphocyte reaction in strains of mice with autoimmune disease. J . Immunol. 125, 1832. Smith, J. B . , and Pasternak, R. 11. (1978). Syngenic mixed lymphocyte reaction in mice: Strain distribution, kinetics, participating cells and absence in NZB mice. J . Znctnunol. 129, 1889. Hom, J. T., and Talal, N. (1982). Decreased syngeneic mixed lymphocyte reaction in autoimmune susceptible mice. Scand. J. Zmmunol. 15, 195. Bocchieri, M. H., Rihlet, R. J., and Smith, J. B. (1981). Autologous mixed lymphocyte reactions in NZB mice: Analysis in recombinant inbred lines shows a T cell defect unrelated to autoantihody production. Eur. J. Zmmunol. 11, 159. Goidl, E. A. (1981). Decreased production of autoanti-idiotypic antibody by the NZB strain of mice. ZCN-UCLA Meet. XX, 1. Speck, N. A , , and Pierce, S. K. (1982). The induction of B cells refractory to antibodyspecific immunoregulation. Eur. J. Zmmunol. 12, 449. Brodeur, P. H., and Wortis, H. H. (1980). Regulation of thymus-independent responses: Unresponsiveness to a second challenge of TNP-Ficoll is mediated by hapten-specific antibodies. J . Zmmunol. 125, 1499. Goidl, E. A , , Fernandes, G . , Weksler, M. E . , Siskind, 6. W., and Good, R. A. (1983). Studies of immune responses in mice prone to auto-immune disorders. I. Heterogeneity of the affinities of anti-hapten antibodies produced by NZB, NZW and related strains of mice. Cell. Zmmunol. 80, 20. Staples, P. J., and Talal, N. (1969). Relative inability to induce tolerance in adult NZB and NZBXNZW F, mice. J . Ex)). Med. 129, 123. Amagai, T., and Cinader, B. (1981). Resistance of MRL/Mp-lpr/lpr mice to tolerance induction. Eur. J. Zmmtrnol. 11, 923. Laskin, C. A,, Taurog, J. D., Smathers, P. A., and Steinberg, A. D. (1981). Studies of defective tolerance in murine lupus. J. Zmmunol. 127, 1743. Hosono, M . , and Fujiwara, M. (1979). Studies on the resistance to tolerance induction against human IgG in DDD mice. 11. Tolerogen-resistant T-cell population in the spleen. Zm7nunology 37, 353. Fujiwara, M., and Cinader, N . (1974). Cellular aspects of tolerance. Vl. The effect ofage on responsiveness and tolerance inducibility of SJL mice. Cell. Zmmunol. 12, 205. Staples, P. J., Steinberg, A. D., and Talal, N. (1970). Induction ofiinmunologic tolerance

382

ARGYRIOS N . THEOFILOPOULOS AND FRANK J. DIXON

in older New Zealand mice repopulated with young spleen, bone marrow, or thymus. J.

E x p . Med. 131, 1223. 437. Parks, D. E . , Doyle, M. V . , and Weigle, W. 0 . (1978). Induction and mode of action of suppressor cells generated against human gammaglobulin. I. An immunologic unresponsive state devoid of demonstrable suppressor cells. J . E x p . Med. 148, 625. 438. Laskin, C. A , , Smathers, P. A., Heeves, J. P., and Steinberg, A. D. (1982). Studies of defective tolerance induction in NZB mice. Evidence for a marrow pre-T cell defect. J. E x p . Med. 155, 1025. 439. Laskin, C. A , , Smathers, P. A , , Lieberman, R . , and Steinberg, A. D. (1983). NZB cells actively interfere with the establishment of tolerance to BCG in radiation chimeras. J. lmnaunol. 131, 1121. 440. Hang, L. M . , Izui, S . , Slack, J. H., and Dixon, F. J. (1982). The cellular basis for resistance to induction of tolerance in the BXSB SLE male mice. J . Inirnunol. 129, 787. 441. Colding, E. A,, Cohen, P. L., McFadden, S. F., Ziff, M., and Vitetta, E. S . (1980). Defective B cell tolerance in adult (NZBxNZW)FI mice, J . E x p . Med. 152, 730. 442. Golding, E. A. (1983). Defective B cell tolerance induction in New Zealand Black mice. I. Macrophage independence and comparison with other autoimmune strains. J. Immrcnol. 131, 2630. 443. Sidman, C. L., and Unanue, E. R. (1975). Receptor-mediated inactivation of early B lymphocytes. Nature (London) 257, 149. 444. Purves, E. C., and Playfair, J. H. L. (1973). Neonatal tolerance characteristics of the antibody-forming cell precursors of the NZB mouse. Clin. E x p . Immunol. 15, 113. 445. Cowdery, J. S . , Curtin, M. F., Jr., and Steinherg, A. D. (1982).Effect ofprior intragastric antigen administration on primary and secondary anti-ovalbulim responses of C57BL/6 and NZB mice. J. E x p . Med. 156, 1256. 446. Miller, M . L., Cowdery, J. S., Laskin, C. A., Curtin, M. F., Jr., and Steinherg, A. D. (1984). Heterogeneity of oral toleronce defects in autoimmune mice. Clin. Immunol. Inmtunopathol. 31, 231. 447. Santoro, T. J., Scott, D. W., and Pisetsky, D. S. (1981). The induction of tolerance and suppression in autoimmune MRL mice using hapten-modified self. J. Immunol. 127, 690. 447a. Izui, S., and Masucla, K . (1984). Resistance to tolerance induction is not prerequisite to development of niurine SLE. J. Immtrnol. 133, 3010. 448. Weksler, M. E. (1983). Senescence of the immune system. In “The Biology of Immunologic Disease” (F.J. Dixon and D. W. Fisher, eds.), pp. 295-306. Sinauer, Sunderland, MA. 449. Holmes, M. C., and Burnet, F. M. (1964). Experimental studies of thymic function in NZB mice and their F, hybrids with C3H. Aust. J . E x p . Riot. Med. Sci. 42, 589. 450. DeVries, M. J.. and Hijmans, W. (1966). A deficient development of the epithelium and autoimmune disease in NZB mice. J . Patho/. Bacterial. 81, 487. 451. DeVries, M. J , , and Hijnians, W. (1967). Pathological changes of thymic epithelial cells and autoimmune disease in NZB, NZW and (NZBxN2W)Fl mice. Inmunology 12, 179. 452. Theofilopoulos, A. N . , Balderas, R. S . , Shawler, D. L., Lee, S . , and Dixon, F. J. (1981). The influence of thymic genotype on the SLE-like disease and T cell proliferation of MRL/Mp-lpr/lpr mice. J. E x p . Med. 153, 1405. 453. Helyer, B. J . , and Howie, J. B. (1963). The thymns and autoimmune disease. Lancet 2, 1026. 454. Bach, M. A , , and Charreire, J. (1979). Role of circulating thymic factor in self-recognition and self-tolerance. Ann. N . Y . Acad. Sci. 332, 55. 455. Savino, W., Dardenne, M., and Bach, J. F. (1983).Thymic hormone containing cells. 11. Evolution of cells containing the serum thymic factor (FTS or thymulin) in normal and autoimmune mice, as revealed by anti-FTS monoclonal antibodies. Relationship with Iahearing clls. Clin. E x p . Immztnol. 52, 1.

MUHINE MODELS OF SLE

383

456. Dardenne, M . , Savino, W., arid Bach, J. F. (1984). Autoimmune mice develop anti1)odies to thymic hormone. Production of anti-thyniulin monoclonal autoantibodies from diabetic (cll)/dl)) and B/W mice. J. Inrinutiol. 133, 740. 457. Talal, N., Daupliinee, M., and Pillarisetty, R. (1975). Efl:ct of thymosine on tliymocyte proliferation and autoimniunity i n NZB mice. Ann. N . Y . Acad. Sci. 249, 438. 458. Steinlxrg, A. D., Law, L. D., and Talal, N . (1970). The role of NZBlNZW F l thymus in experimental tolerance and autoimmunity. Arthritiy Rheum. 13, 369. 459. Kysela. S., and Steinberg, A. D. (1973). Increased survival ofNZB/W mice given multiple syngeneic young thymiis grafts. Clin. Immune/. Ittimunopatho/. 1, 133. 460. Knight. J. C., and Adams, D. D. (1978). Failure of transferred tliyinus cells to suppress or prevent autoantibody production in NZB and NZB X NZW mice. J. Clin. Lab. Itnrnunol. 1, 151. 461. Hoffnian, A . A , , and Harllek, R . J. (1979). Immunoregulation in New Zealand mice: I. Failure of the transfer of syngeneic spleen or thymus cells to influence the natural disease in New Zealand mice. Arthritis Rheum. 22, 412. 462. Gershwin. M. E.. Kruse, W., and Coldstein, 6. (1979).The effect of tliymopoietin 32-36 and iibiquitin o n spontaneous iininune pithology of New Zealand mice. J . R h t n . 6, 610. 463. East, J.. deSousa, M. A. R . , Parrott, D. M. V., and Jacpet, H. (1967). Consequences of neonatal thymectomy in New Zealand black mice. Clin. E x p . Zititnuno/. 2, 203. 464. Roubinian, J. R., Papoian, R., and Talal, N . (1977). Effects of neonatal tliyinectotny aiid splenectomy on survival and regulation of autoantilmdy formation in NZB/NZW F, mice. J. Zittinuno/. 118, 1524. 465. Steinlwrg, A. D., Rotlis, J. B., Murphy, E. D., Steinberg, R. T., and Raveche, E. S. (1980). Effects of thymectomy or androgen atlniinistration upon the autoimmune disease of MRLiMp-tprilrtr miccl. J . I m t t ~ r r i i o l .125, 871. 466. Hang. L. M., Theofilopoulos, A. N . . Balderas, R. S., Francis, S. J . , and Dixon, F. J. (1984). Eflect of thymectomy on several kinds of SLE mice. J . Zniinutiol. 132, 1809. 467. Hang, L. M . , h i , S., Theofilopoulos, A. N . , and Ilixon F. J. (1982). Suppression of transferred BXSB niale SLE disease by female spleen cells. J. Irnn~ztnol.128, 1805. 468. Raveche, E. S., Steinberg, A. I)., Klassen, L. W., aiidTjio, J. H. (1978). Genetic studies in NZB mice. I. Spontaneous autoantibody production. J . E x p . Med. 147, 1487. 469. Yoshida, H., Kohno, A , , Ohta, K., Hirose, S.. Maruyama, N . , and Shirai, T. (1981). Genetic studies of autoimmunity i n New Zealand mice. 111. Associations arnoiig anti-DNA antil)otlies, NTA, and renal disease in (NZBx NZW)FIX NZW backcross mice. J. Imnirrno/. 127, 433. 470. Hirose, S..Nagasawa, R . , Sekikawa, I., Hamaoki, M., Isliida, Y . , Sato, H.. and Shirai, T. (1983). Enhancing effect of H-2-linked NZW gene(s) on the autoimmime traits of (NZBXNZW)FI mice. J , E x p . Med. 158, 228. 471. Sliirai, T. (1982). The genetic hasis of autoiiiiinunity i n murine lupus. Zmmuiiol. Today 3, 187. 472. Maruyama, N., Furukawa, F., Nakai. Y . , Sasaki, Y . , Ohta, K., Ozaki, S., Hirose, S., and Shirai, T. (1983). Genetic studies of autoimmunity i n New Zealand mice. IV. Contribution of NZB and NZW genes to the spontaneous occurrence of retrovird gp70 inimune complexes i n (NZBXNZW)F, hybrids aiid the correlation to renal disease. J. Zmrnuno!. 130, 740. 473. Knight, J. G . , Adams, D, I)., and Purves, H . 11. (1977). The genetic contribution of the NZB moiise to the renal disease of the NZBXNZW 1iyl)ritl. C h i . E x ~ JZitmuno/. . 28, 352. 474. Knight, J. G . , and Adarns, D. 1). (1978).Three genes for hipus nephritis in NZBXNZW mice. J. E x p Med. 147, 1653. 475. Maruyama, N., Olita, K . , Hirose, S., and Shirai, T. (1980). Genetic studies of autoim-

384

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. D I X O N

munity in New Zealand mice. Association among natural thymocytotoxic autoantibody (NTA), renal disease, and H-2 complex. Immunol. Lett. 2, 1. 476. Eastcott, J. W., Schwartz, R. S., and Datta, S. K. (1983). Genetic analysis of the inheritance of B cell hyperactivity in relation to the development of autoantibodies and glomerulonephritis in NZBXSWR crosses. J. Zmmunol. 131, 2232. 477. Datta, S. K., Owen, F. L., Womack, J. E . , and Riblet, R. (1982).Analysis of recombinant inbred lines derived from “autoimmune” (NZB) and “high leukemia” (C58)strains: Independent multigenic systems control B cell hyperactivity, retrovirus expression, and autoimmunity. J. Zmmun{~l.129, 1539. 478. Rosenberg, Y. J , , and Steinberg, A. D. (1984). Influence of Y and X chromosomes on B cell responses in autoimmune prone mice. J. Inmunol. 132, 2161. 479. Izui, S . , Masudoe, K., and Yoshida, H . (1984). Acute SLE in F1 hybrids between SB/Le and NZW mice: Prominently enhanced formation of gp70 immune complexes by a Y chromosome-associated factor from SBiLe mice. J. Zmmunol. 132, 701. 480. Pisetsky, D. S., Caster, S. A,, Roths, J . B., and Murphy, E. D. (1982).lpr gene control of the anti-DNA antibody response. J. Zmmunol. 128, 2322. 481. Izui, S . , Kelley, V. E., Masuda, K., Yoshida, H., Roths, J. B., and Murphy, E. D. (1984). Induction of various autoantibodies by mutant gene Ipr in several strains of mice. J . Zmmunol. 133, 2267. 482. Adams, D. D., and Knight, J. G. (1980). H gene theory of inherited autoimmune disease. Lancet 1, 396. 483. Adams, D. D., Adams, Y., Knight, J. B., andvanloghem, E. (1981).Genetics ofautoimmune disease in the New Zealand mice and man. In Menarini Series on Iminunopathology” (P. A. Miescher, L. Bolis, and G. Torrigiani, eds.), pp. 135-146. Schwabe, Basel. 484. D’Eustachio, P . , Pravtcheva, D., Marcu, K . , and Ruddle, F. H . (1980). Chromosomal location of the structural gene cluster encoding murine immunoglobulin heavy chains. J . E x p . Med. 151, 1545. 485. Susan, D. D’Eustachio, P., Leinwand, L., Seidman, J . , Keithley, D., and Ruddle, F. H. (1979). Chromosomal assignment of the mouse k light chain genes. Proc. Natl. Acad. Sci. U.S.A. 76, 2735. 486. D’Eustachio, P., Bothwell, A. L. M., Takaro, T. K., Baltimore, D., and Huddle, F. H. (1981). Chromosomal location of structural genes encoding murine immunoglobulin lambda light chains. J . E x p . Med. 153, 793. 487. Dubois, E. L. (1974). The clinical picture of systemic lupus erythematosus. In “Lupus Erythematosus” (E. L. Dubois, ed.), 2nd Ed., pp. 232-242. Univ. of South. California Press, Los Angeles. 488. Cohn, D. A. (1979). High sensitivity to androgen as contributing factor in sex differences in the immune response. Arthritis Rheum. 22, 1218. 489. Stern, R . , Fishman, J.. Brusman, H., and Kunkel, H. G . (1977). Systemic lupus erythematosus associated with Klinefelter’s syndrome. Arthritis Rheum. 20, 18. 490. Michalski, J. P., Snyder, S. M., Leod, R. L., and Talal, N. (1978).Monozygotic twins with Klinefelter’s syndrome discordant for systemic lupus erythematosus and symptomatic myasthenia gravis. Arthritis Rheum. 21, 306. 491. Yocum, M . W., Grossman, J., Waterhouse, C., Abraham, G. N., May, A. G., and Condemi, J. J, (1975). Monozygotic twins discordant for systemic lupus erythematosus. Arthritis Rheum. 18, 193. 492. Lahita, R. G., Bradlow, H. L., Kunkel, H. G., and Fishman, J. (1979). Alterations of estrogen metabolism in systemic lupus erythematosus. Arthritis Rheum. 22, 1195. 493. Roubinian, J. R . , Talal, N . , Siiteri, P. K., and Sadakiam, J. A. (1979). Sex hormone modulation of autoimmunity in NZB/NZW mice. Arthritis Rheum. 22, 1162.

M U R I N E MODELS OF SLE

385

494. Roubiniiin, J. R . , Papoian, R . , and Talal, N. (1977). Androgenic hormones moderate antoantibody responses and improve survival i n murine lupus. /. Clin. Invest. 59, 1066. 495. Roul~inian,J. R., Talal, N., Greenspan, J. S., Coodnian, J. R . , atid Siiteri, P. K . (1978). Effects of castration and sex hortnone treatnient on survival, anti-niicleic acid antibodies, and glomerulonephritis in NZB/NZW F, mice. J . E x p . Med. 147, 1568. 496. Steinberg, A. D., Melez, K. A., Raveche, E. S . , Reeves, J. P . , Boegel, W. A,, Sniathers, 1’. A., Taurog. J. D., Weinkin, L., and Duvic. M. (1979). Approach to the study of the role of sex hormones in autoininiunity. Arthritis Rheum. 22, 1170. 497. Weinstein, Y., and Berkovich, Z. (1961). Testosterone effect on bone marrow, thymus, and suppressor T cells in the (NZBx NZW)F, mice: Its relevance to autoimmunity. /, Itntntrnol. 126, 998. 496. Warren, R . W., Roths, J. B., Murphy, E. D . , and Pisetsky, D. S. (1984). Mechanisms of polychal B-cell activation in autoimmune B6-lpr/!pr mice. Cell. Iminunol. 84, 22. 499. Eisenberg, H. A,, Lee, S., and Dixon. F. J. (1980). Effect of castration on male-determined acceleration of autoimmune disease in BXSB mice. J . Zn~munol.125, 1959. 500. Lahita, R. C., Chiorazzi, N., Gihofsky, A,, Winchester, R . J . , and Kunkel, H. G. (1983). Familial systemic lupus erytheniatosus in males. Arthritis Rheum. 26, 39. 501. Mellors, R. C . . and Huang, C. Y. (1967). Ininiunopathology of NZBiBL mice. VI. Virus separahle from spleen and pathogeneic for Swiss mice. /. E x p . Med. 126, 53. 502. Russell, P. J., Hicks, J. D., Boston, L. E., and Abbott, A. (1970). Failure to transfer haeinolytic anaemia or gloinerulonepllritis with cell-free inaterial from NZB mice. Clin. Ex)). Immunol. 6, 227. 503. Schaap, 0. L., deGroot, E. R., and van Loghem, J. J. (1975). Failure to induce antoimtnune phenomena in mice by administration of highly purified C-type particles from NZB inice. Pathol. Microbiol. 42, 171. 504. Datta, S. K . , and Schwartz, R. S. (1977). Mendelian segregation of loci controlling xenotropic virus production in NZB crosses. Virology 83, 449. 505. Ilatta, S. K., and Schwartz, R. S. (1976). Genetics of expression ofxenotropic virus and autoimmunity in NZB mice. Nature (London)263, 412. 506. Datta, S. K., Manny, N., Andrzejewski, C . , Andre-Schwartz, J.. and Schwartz, R. S . (1978). Genetic studies of autoirnmunity and retrovirus expression in crosses of New Zealand Black mice: I. Xenotropic virns. /. E x p , Med. 147, 854. 507. Datta, S. K., McConahey. P. J., Manny, N., Theofilopoulos. A. N . , Dixon, F. J . , and Schwartz, R. S. (1978). Genetic studies of autoimmunity and retrovirus expression in crosses of New Zealand Black mice: 11. The viral envelope glycoprotein gp70. /. E x p . Med. 147, 872. 508. Tonietti, G., Oldstone, M. B. A,, and Dixon, F. J. (1970). The effect of induced chronic viral infections on the immunological diseases of New Zealand mice. J . E x p . Med. 132, 89. 509. Crocker, B. P., DelVillano, B. C . , Jensen, F. C . . Lerner, R. A . , and Dixon, F. J. (1974). Iminiinopathogenicity and oncogenicity of murine leukemia viruses: I. Induction ofimmunologic disease and lymphoma in (BALB/c x NZB)F, mice by Scripps leukemia virus. J. E x p . Med. 140, 1028. 510. Ahmed, R . , and Oldstone, M. B. A. (1984). Mechanisms and I d o g i c a l iinplications of virus-induced polyclonal B cell activation. In “Concepts i n Viral Pathogenesis” (A. L. Notkins and M. B. A. Oldstone, eds.), pp. 231-238. Springer-Verlag. Berlin and New York. 511. Oldstone, M. B. A , , and Dixon, F. J. (1972). Inhibition ofantibodies to nuclear antigen and to DNA in New Zealand mice infected with lactate dehydrogenase virus. Science 175, 784.

386

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J. D I X O N

512. Giltinan, P. J . , Holmes, M. C., and Burnet, F. M. (1965). Cortisone acetate treatment of haemolytic anaemia in NZB mice. Aust. J. E x p . B i d . Med. Sci. 43, 523. 513. Casey, T. P. (1968). Systemic I U ~ L I erythematosus S in NZB X NZW hybrid mice treated with the corticosteroid drug betamethasone. J . Lab. Clin. Med. 71, 390. 514. Gelfand, M. C., and Steinberg, A. D. (1972). Therapeutic studies in NZB/W mice. 11. Relative efficacy of azathioprine, cyclophosphamide and methylprednisolone. Arthritis Rheum. 15, 247. 515. Walker, S. E., Anver, M. R . , Schecliter, S. L., and Bole, G. G . (1978).Prolonged lifespan and high incidence of neoplasms in NZB/NZW mice treated with hydrocortisone sodium succinate. Kidney Znt. 14, 151. 516. Fan, P. T., Yu, D. T. Y . , Fowlston, S., Eisman, J., and Bluestone, R. (1979). Effect of corticosteroids on the human immune response: Comparison of one and three daily Igm intravenous pulses of methylpredinsolone. J. Lab. Clin. Med. 91, 625. 517. Fessel, W. J. (1980). Megadose corticosteroid therapy in systemic lupus erythematosus. J . Rheumatol. 7, 486. 518. Isenberg, 13. A , , Morrow, W. J. W., and Snaith, M. L. (1982). Methylprednisolone pulse therapy in the treatment of systemic lupus erythematosus. Ann. Rheum. Dis. 41, 347. 519. Casey, T. P. (1968). 6-Mercaptopurine administration to NZB mice. Aust. J . E x p . B i d . Med. 46, 327. 520. Walker, S. E . , and Bole, 6. G . , Jr. (1973). Augmented incidence of neoplasia in NZB/ NZW mice treated with long-term cyclophosphamide. J. Lab. Clin. Med. 82, 619. 521. Hahn,B. H . , Knotts, L., Ng, M., and Hamilton, T. R. (1975).Influence ofcyclophosphainide and other immunosuppressive drugs on immune disorders and neoplasia in NZB/NZW mice. Arthritis Rlzeum. 18, 145. 522. Casey, T. P. (1968). Immunosuppression by cyclopliospliarnide in NZBX NZW mice with lupus nephritis. Blood 32, 436. 523. Chia, D., Levy, L. Barnett, E. V., and Carnes, W. H. (1981). Effects of pulse cyclophosphamide on NZB/W disease. Immunology 43, 57. 524. Hurd, E. R.,and Ziff, M. (1977). The mechanism of action of cyclophosphamide on the nephritis of (NZBxNZW)F, hybrid mice. Clin. E x p . Zmmunol. 29, 132. 525. Morris, A. D., Esterly, J., Chase, G., and Sharp, 6. C. (1976). Cyclophosphamide protection in NZB/NZW disease. Arthritis Rheum 19, 49. 526. Walker, S. E., Solsky, M., and Schnitzer, B. (1982). Prolonged lifespans in female NZBiNZW mice treated with the experimental immunoregulatory drug frentizole. Arthritis Rheum. 25, 1291. 527. Gabrielsen, A. E., Lubert, A. S., and Olsen, C. T. (1979). Suppression of murine lupus erythematosus by Dactinomycin. Nature (London) 264, 439. 528. Gabrielsen, A. E., and Olsen, C. T. (1979). Late treatment of murine lupus erythematosus with dactinomycin (actinomycin D). Clin. E x p . Irnmunol. 35, 33. 529. Casey, T. P. (1968). The development of lymphomas in mice with autoimmune disorders treated with azathioprine. Blood 31, 396. 530. Smith, H. R . , Chused, T. M., and Steinberg, A. D. (1984). Cyclophosphamide-induced changes in the MRL-lpr/Zpr mouse: Effects upon cellular composition, immune function, and disease. Clin. lmmunol. Immunpathol. 30, 51. 531. Hahn, B. H., Bagby, M. K., Hamilton, T. R . , and Osterland, C. K . (1973).Comparison of therapeutic and immunosuppressive effects of Azathioprine, Predinisolone and combined therapy in NZB/NZW mice. Arthritis Rheum. 16, 163. 532. Klaus, G. G . B. (1981). The effects of cyclosporin A on the immune system. Zrnnaunol. Today 2, 83.

M U H I N E MODELS OF SLE

387

533. Israel-Biet, I]., Noel, L.-H., Bach, M.-A,. Dardenne, M., and Bach, J. F. (1983). Marked reduction of DNA antibody producing aiid glonirrulopathy in thymuliri (FTS-Zn) or cyclosporin A treated (NZBX NZW)F, mice. Clin. E x p . Immunol. 54, 359, 534. Isenberg, D. A., Snaith, M. L., Morrow, J. W., Al-Khander, A. A., Cohen, S . L., Fisher, C., and Mowhray, J. (1981). Cyclosporin A for the treatment of systemic lupus erythematosus. Znt. J . Immunophannacol. 3 , 163. 535. Klassen, L. W., Budman, D. R., Williams, G. W . , Steinherg, A. D . , and Gerber, N. L. (1977). Ribavirin: Efficacy in the treatment of inurine autoimmune disease. Science 195, 787. 536. Denman, A. M., Russell, A. S., Loewi, G., and Ilenman, E. J. (1971). lmmunopathology of New Zealand Black mice treated with anti-lymphocyte globulin. Imnzunology 20, 973. 537. Denman, A. M., Russell, A. S., andDennian, E. J. (1970).Renaldiseasein(NZBXNZW)F, hybrid mice treated with antilymphocyte a n t i l d y . Clin. Erp. Ztnrnunol. 6, 325. 538. Seaman, W. E . , Wofsy, D., Greenspan, J. S., and Ledbetter, J. A. (1983). Treatment of autoimmune MRLllpr mice with monoclonal antibody to Thy-1.2. A single injection has sustained eflects on lymphoproliferation and renal disease. J . Ztnmunol. 130, 1713. 539. Adelman, N. E., Watling, 11. L.. and McDevitt, H. 0. (1984). Treatment of (NZBX BZW)FI disease with anti-I-A monoclonal antibodies. J . E x p . Med. 158, 1350. 540. Waldor, M. K., Hardy, R. R., Hayakawa, K., Steinman, L., Herzenberg, L. A,, and Herzenberg, L. A. (1984). Disappearance and reappearance of B cells after in oioo treatment with monoclonal anti-I-A antibodies. Proc. Natl. Acad. Sci. U .S.A. 81, 2855. 541. Steinman, L., Rosenbaum, J. T., Sriram, S . , and McDevitt, H. 0. (1981). In O ~ O Oeffects of antibodies to immune response gene products. Prevention of experimental allergic encephalitis. Proc. Natl. Acad. Sci. U . S . A . 78, 7111. 542. Waldor, M . K., Sriram, S., McDevitt, H. O., and Steinman, L. (1983). Zn oiuo therapy with monoclonal anti-I-A antibody suppresses immune response to acetylcholine receptor. Proc. Natl. Acad. Sci. U . S . A . 80, 2713. 543. Kotzin, B., and Strober, S. (1979). Reversal of NZB/NZW disease with total lymphoid irradiation. J . Exp. Med. 150, 371. 544. Slavin, S. (1979). Successful treatment of autoinnnune disease in (NZB/NZW)FI female mice by using fractionated total lymphoid irradiation. Proc. Natl. Acad. Sci. U .S.A. 76, 5274. 545. Theofilopoiilos, A. N . , Balderas, R., Shawler, I).L., h i , S . , Kotzin, B. L.. Strober, S . , and Dixon, F. J. (1980). Inhibition ofT cell proliferation and SLE-like syndrome of MRL/I mice by whole body or total lymphoid irradiation. 1. Iinmunol. 125, 2137. 546. Moscovitch, M . , Rosenmann, E . , Neeinan, Z . , and Slavin, S. (1983). Siiccessful treatment of autoimmune manifestations in MRL/I and MRL/n mice using total lymphoid irradiation. E x p . Mol. Pathol. 38, 33. 547. Kotzin, B. L., Stroher, S . . Engleinan, E . G . , Calin, A., Hoppe, R. T., Kansas, G . S.. Terrell, C . P . , and Kaplan, H. S . (1981).Treatment of intractable rheumatoid arthritis with lymphoid irradiation. New Engl. J. Med. 305, 969. 548. Trentharn, I). E., Belli, J. A , , Anderson, R. J., Buckley, J. A., Goetzl, E. J . , David, J. R.. and Austen, K. F. (1981). Clinical and immunologic effects of fractionated total lymphoid irradiation in refractory rheumatoid arthritis. New Engl. J . Med. 30, 976. 549. Stroher, S . (1984). Overview: E&ct of total lyinphoid irradiation o n autoirninnne disease and transplantation. J . Itnmunol. 132, 968. 550. Brahn, E., Helfgott, S. M., Belli, J. A., Anderson, R. J., Reinherz, E. L., Schlossman, S . F., Austen, F. K., and Trentham, D. E . (1984). Total lymphoid irradiation therapy in refractory rheumatoid arthritis. Arthritis Rheum. 27, 481.

388

ARCYRIOS N . THEOFILOPOULOS A N D FRANK J. DIXON

551. Ben-Nun, A,, Wekerle, H., and Cohen, I. R. (1981). Vaccination against autoimmune encephalomyelitis with T-lymphocyte line cells reactive against myelin basic protein. Nature (London) 292, 60. 552. Holoshitz, J.. Naparstek, Y . , Ben-Nun, A,, and Cohen, I. R. (1983). Lines ofT lymphocytes induce or vaccinate against autoimmune arthritis. Science 219, 56. 553. Cohen, I. R . , Ben-Nun, A,, Holoshitz, J., Maron, R . , and Zerubavel, R. (1983).Vaccination against autoimmune disease with lines of autoimmune T lymphocytes. lminunol. Today 4, 227. 554. Goodwin, J. S., and Webb, D. R. (1980). Regulation o f t h e immune response by prostaglandins. Clin. Immunol. Immunopathol. 15, 106. 555. Zurier, R. B . , Darnjanov, I., Sayadoff, D. M., and Rothfield, N. F. (1977). Prostaglandin E l treatment of NZB/NZW F, hybrid mice. 11. Prevention of glomerulonephritis. Arthritis Rheum. 20, 1449. 556. Zurier, R. B., Darnjanov, I., Miller, P. L., and Biewer, B. F. (1978). Prustaglandin El treatment prevents progression of nephritis in murine lupus erythematosus. J . Clin. Lab. 11nmunol. 1, 95. 557. Zurier, R. B., Sayadoff, D. M., Torrey, T. B., and Rothfield, N. F. (1977). Prostaglandin El treatment of NZB/NZW F1 hybrid mice. I. Prolonged survival offemale mice. Arthritis Rheum. 20, 723. 558. Kelley, V. E., Winkelstein, A,, and Izui, S. (1979). Effects of Prostaglandin E on immune complex nephritis in NZB/W mice. Lab. Inuest. 41, 531. 559. Kelley, V. E., Winkelstein, A., Izui, S . , and Dixon, F. J. (1981). Prostaglandin E l inhibits T-cell proliferation and renal disease in MRL/I mice. Clin. Zmniunol. Inzrnunopathol. 21, 190. 560. Kelley, V. E., Izui, S., and Halushka, P. V. (1982). Effect of Ibuprofen, a fatty acid cyclooxygenase inhibitor, on murine hipus. Clin. linmunol. lininunopathol. 25, 223. 561. Levy, J. A. (1984). Effects of nutrition on the immune response. Zn"Basic and Clinical Immunology" (H. H. Fudenberg, D. P. Stites, J. Stobo, and J. V. Wells, eds.), 5th Ed., pp. 227-305. Lange Medical Publ., Los Altos, CA. 562. Good, R . A. (1981). Nutrition and immunity. J . Clin. Iininunol. 1, 3. 563. Fernandes, G., Yunis, E . J., Smith, J., and Good, R. A. (1972). Dietary influence 011 breeding behavior, hemolytic anemia, and longevity in NZB mice. Proc. Soc. E x p . Biol. Med. 139, 1189. 564. Fernandes, G., Yunis, E. J.. and Good, R . A. (1976). Influence ofprotein restriction 011 immune fiinetions in NZB mice. J . Zmntzrnot. 116, 782. 565. Fernandes, G . , Yunis, E. J., Jose, D. G., and Good, R. A. (1973). Dietary influence 011 antinuclear antibodies and cell-mediated immunity in NZB mice. Znt. Arch. Allergy Appl. lminunol. 44, 770. 566. Fernandes, G., Friend, P., Yunis, E. J., and Good, R. A. (1978). Influence of dietary restriction on immunologic function and renal disease in (NZBX NZW)Fl mice. Proc. Natl. Acad. Sci. U . S . A . 75, 1500. 567. Fernandes, G., Yunis, E . J., and Good, R. A. (1976).Influence of diet on survival of mice. Proc. Natl. Acad. Sci. U.S.A. 73, 1279. 568. Friend, P. S., Fernaudes, G . , Good, R. A , , Michael, A. F., and Yunis, E . J. (1978). Dietary restriction early and late: Effects on the nephropathy of the NZBXNZW mouse. Lah. Invest. 38, 629. 569. Safai-Kutti, S . , Fernandes, G . , Wang, Y . , Safai, B., Good, R . A., and Day, N. K. (1980). Reduction of circulating immune complexes hy calorie restriction in (NZB x NZW)Fl mice. Clin. linmunol. Irnniunopathol. 15, 293. 570. lung, L. K. L., Palladino, M. A., Calvano, S., Mark, D. A., Good, R. A . , and Fernandes,

M U H I N E MODELS OF SLE

571.

572.

573.

574. 575.

576. 577.

578.

579. 580. 581. 582.

583. 584.

585.

586. 587.

588.

389

G. (1982). Effect of calorie restriction on the prod~ictionand responsiveness to interleukin 2 in (NZBX NZW)Fl mice. Clin. Znmiunol. Zmrnutiopothd. 25, 295. Hurd, E. R . , Johnston, J. M . , Okita, J. R . . MacDonald, P. C., Ziff, M., and Gilliam, J. N. (1981). Prevention of glomerulonepliritis and prolonged survival in New Zealand Black/New Zealand White FI hybrid mice fed an essential fatty acid-deficient diet. J. Clin. fntiest. 67, 476. Prickett, J. D., Robinson. D. R.,and Steinlierg, A. D. (1981). Dietary enrichment with the pnlyunsaturated Fatty acid eicosapentaenoic acid prevents proteinuria and prolongs survival in NZBXNZWlF, mice. J. Clin. Intiest. 68, 556. Prickett, J. D., Robinson, 1).I t , and Steinberg, A. D. (1983). Effects of dietary enrichment with eicosapentaenoic acid upon autoimmune nephritis i n feniale NZBX NZWIF, mice. Arthritis Rheum. 26, 133. Kelley, V. E . , and Izui, S. (1983). Enriched lipid diet accelerates lupus nephritis in NZBXW mice. Am. J. Pathol. 111, 288. Levv, J. A,, Ibrahini, A . B . , Shirai, T., Ohta, K . , Nagasawa, R . , Yoshida, H., Estes, J . , and Gardner, M . (1982). Dietary fdt effects ininiune response, production of antiviral Factors, and irnmnne complex disease in NZB/NZW mice. Proc. Notl. Acad. Sci. U.S.A. 79, 1974. Kubo, C., Johnson, B. C., Day, N . K., and Good, R. A. (1984). Calorie source, calorie restriction, iminiinity and aging. J. Nutr. 114 (in press). Beach, R. S . , Gershwin, M . E., and Hurley, L. S. (1982). Nutritional factors and autoiinnirinity. 111. Zinc deprivation versiis restricted food intake in MRLA mice-The distinction hetween interacting dietary influences. J. Irnniunol. 129, 2686. Kubo, C., Day, N . K., and Good, R . A. (1984). Influence ofearly or late dietary restriction on life span and immunological parameters i n MRL/Mp-lpr/lpr mice. Proc. Natl. Acod. Sci. U . S . A . 81, 5831. Levy, J. A,, and Marrow, W. J. W. (1983). Dietary regulation of the autoimmune process in niurine lupus. Imnunol. Today 4, 249. Beach, R. S., Gershwin, M . E., and Hurley, L. S. (1981). Nutritional factors and autoimmunity. I. Iinniunopatliology of zinc-deprived New Zealand mice. J . Zmniunol. 126, 1999. Beach, R. S . , Gershwin, M . E., and Hurley, L.S. (1982). Nutritional factors and autoimmunity. 11. Prolongation of survival in zinc-deprived NZB/W mice. J. Znlnlunol. 128, 308. Malinow, M. R , , Bardana, E. J., Pirofsky, B., Craig, S., and McLaiighlin, P. (1982). Systemic lupus erytheinatosus-likc syndrome in inonkeys fed alfalfa sprouts: Role 0 1 a nonprotein amino acid. Science 216, 415. Roberts, J. L., and Hayashi, J. A . (1983). Exacerbation of SLE associated with alfalfi ingestion. New Engl. J. Med. 308, 1361. Gershwin, M. E . , Lentz, D. R . , Beach, R. S., and Hurley, L. S. (1984). Nutritional factors and autoimmunity. IV. Dietary vitamin A deprivation induces a selective incr-ease i n IgM autoantihodies and Iiyperjiammagloliiilinemia in New Zealand Black mice. J. Zmoiunol. 133, 222. Lewis, R . M . , Smith, C. A., Killyam, L., and Borel, Y. (1979). Sex differences i n the maintenance of immunologic tolerance to endogenous nucleic acid antigens and the prevention of niurine lupiis nephritis. Clin. Zmmunol. lmtnunopathol. 13, 92. Borel, Y., and Young, M . C. (1980). Nucleic acid-specific suppressor T cells. Proc. Natl. Acnd. Sci. U.S.A. 77, 1593. Borel, Y., Borel, H., Schneeberger, E., and Lewis, R. M. (1984). Influence of intravenous administration of oucleoside-coupled spleen cells on inurine lupus nephritis i n female (NZBXNZW)FI mice. C h . Imrntrno~.Inununopathol. 30, 451. Morimoto, C., Mashuho, Y., Borel, Y . , Steinlierg, A., and Schlossman, S. F. (1983).

390

ARGYRIOS N . THEOFILOPOULOS A N D FRANK J . DIXON

Selective inhibition of antinucleoside-specific antibody production by nucleoside-ricin A conjugate. J . Znmunol. 131, 1762. 589. Theofilopoulos, A. N., and Dixon, F. J. (1982). Autoimmune disease: Immunopathology and etiopathogenesis. Am. J. Puthol. 108, 321. 590. Eichmann, K. (1978). Expression and function of idiotypes on lymphocytes. Ado. Zrnmunol. 26, 195. 591. Hahn, B. H., and Ebling, F. M. (1984). Suppression of murine lupus nephritis by administration of an anti-idiotypic antibody to anti-DNA. J . Znafzunol. 132, 187. 592. Teitelbaum, D., Rauch, J., Stollar, B. D., and Schwartz, R. S. (1984). In vioo effects of antibodies against a high frequency idiotype of anti-DNA antibodies in MRL mice. J. Zinmunol. 132, 1282. 592a. Kofler, R., Noonan, D. J., Levy, I).E . , Wilson, M. C., Mdler, N. P. H., Dixon, F. J., and Theofilopoulos, A. N. (1985). Genetic elements utilized for a murine lupus anti-DNA autoantibody are closely related to those for antibodies to exogenous antigens. J. E x p . Med. (submitted). 593. Eilat, D., Hochberg, M., Pumphrey, J . , and Rudikoff, S. (1984). Monoclonal antibodies to DNA and RNA from NZB/NZW F1 mice: Antigenic specificities and NHz-terminal amino acid sequences. J. Imrnunol. 133, 489. 594. Coutinho, A. (1980). The self-nonself discrimination and the nature and acquisition of the antibody repertoire. Ann. Zmmunol. (Paris) 131D, 235. 595. Kohler, H., Levitt, D., and Bach, M. (1981). A non-galilean view ofthe immune network. Zinniunol. Toduy 2, 58. 596. Theofilopoulos. A. N., Prud’homme, C . J., Fieser, T. M., and Dixon, F. J. (1983). B cell hyperactivity in lupus. I. Immunological abnormalities in lupus-prone strains and the activation of normal B cells. Zinmunol. Toduy 4, 287. 597. Theofilopoulos, A. N., Prud’hoinme, C . J., Fieser, T. M., and Dixon, F. J. (1983). I3 cell hyperactivity in lupus. II. Defects in response to and production of accessory signals in lupus-prone mice. Zmmunol. Toduy 4, 317.

Index A Abelson disease complexity of genetics of susceptibility to, 88-89 proliferating cells as targets for in uitro transformation, 88 role for helper virus in neoplastic transformation, 87-88 transformation of cells from genetically resistant mice, 89-90 use of site-directed antibodies, 9091 variations in disease process induced by mutants, 90 complexity of transformation in vitro characterization of associated cellular changes, 93-94 early biological effects of o-abl expression, 91 preneoplastic growth properties of infected pre-B cells, 91-92 progression of tumorigenicity, 9293 use of cultured B cell lines to study, 94-95 history of, 75-76 neoplastic transformation by NIH 3T3 fibroblasts and, 83-84 pre-B cells, 79-80 rapid induction of thyrnoinas and, 83 target cells in inonocyte-inacrophage lineage, 82-83 techniques for transforination of mature B and plasma cells, 80-81 transfornmed cell lines with diffcrentiation potential, 82 nonneoplastic changes induced by, 84-85 agar colony formation from fetal liver erythroblasts, 85 lethality of, 86-87 resistance of GM-CFC to leukemia-associated inhibition activity, 85 overview, 74-75

pattern of c-ah1 expression by normal and leukemic cells, 78-79 properties of u-ahl and tyrosine kinase transforming protein, 7677 use of anti-A-MuLv antibodies to study c-abl-encoded protein, 77-78

C C 1,activated actions of, 202-204 regulation and fate of C l q receptor interaction, 206 dissociation of activated CIr,Cls, and C1, 204-206 inhibition of activated C1 by C1In, 204 C1 activation, process of C1 activation requirements and process, 195-200 C1 coinplexing agents and activators, 194- 195 regulation of, 200-201 C1 activation unit complexes of C1, 185-194 Clr,Cls,, 183-185 proteins of C1 inhibitor, 178-183 Clq, 155-165 C l r and Cls, 165-178

c3 hackground, 217-218 definition of receptors, 221 history, 219-221 functions of receptors erythrocyte, 242-251 kidney podocyte, 260-261 lymphocyte, 257-260 mast cell, 261 monocyte macrophage, 255-257 neutrophil, 251-255 generation of ligands for receptors, 221-222 391

392

INDEX

activation of C3 by classical and alternative pathway, 223-225 covalent binding of C3b, 222-223 degradation of C4b and C3b by cleavage with factor 1, 226-

HLA-DR biochemistry, 23-31 serology/HTC, 5-9 HLA-D region products monoclonal antibodies to, 18-21

230 factors controlling C3 activation,

I

225-226 structure and binding site characteristics of receptors type four, 240-242 type one, 230-235 type three, 238-240 type two, 235-238 Complement system, classical pathway, history of, 153-154

Infectious mononucleosis acute, immunoregulatory cell functions in, 115-122 associated suppressor T cell activity, reversal by D-mannOSe and saccharides, 138-142 Epstein-Barr virus and, 112-115

E Epstein-Barr virus abnormal infection, selected disorders associated with, 129-138 irnrnunoglobulin production induced by, relationship to immortalization, 110-112 infectious mononucIeosis and, 112-

115 persistent infection in normal individuals, 122-129 polyclonal B cell activation by, 102-

Lupus mice, see also Murine SLE cellular abnormalities defects in tolerance, 330-332 functional abnormalities of B cells, T cells, macrophages and related interleukins, 308-330 surface characteristics and numerical abnormalities of T and B cells, 303-308 thymic defects, 332-336 transfer of autoimmune disease,

301-303

110

H HLA-D deletion mutants, 15-18 function of, 63-65 genes, 53-60 invariant (gamma) chain, 50-53 restriction fragment length polymorphism, 60-63 supertypic specificity localization,

47-50 HLA-DP biochemistry, 43-47 serology/HTC, 14-15 HLA-DQ biochemistry, 31-43 serology, 9-14

derivation of, 271-273 natural history and pathology of,

273-274 morphologic manifestations, 276-

284 serologic manifestations, 284-300 survival and body weight, 274-276

M Murine SLE, see also Lupus mice functional abnormalities in B cells, 309-316 macrophage defects, 324-326 natural killer cells, 327 other cellular and humoral abnormalities, 327-330 T cells, 316-324

393

INDEX

genetics of, 336-337 complementarity of genetic hackgrounds among lupus mice and role of accelerator and other autosoinal genes, 341345 inheritance of au toimm iinc traits, 337-338 relationship among autoimmune traits and their association with disease, 338-341 relationship between the traits, H-2 and Ig genes, 345-346 influence of sex and sex hormones on pathogenesis of, 346-348 rriorphologic rnanifcstations arthritis, 282-283 glomerulonephritis, 276-278 lymphoid hyperplasia, 279-282 neoplasia, 282 other significant histopathologic characteristics, 283-284 thymic atrophy, 278

vascular disease and myocardial infarction, 278-279 serologic manifestations antierythrocyte antibodies, 295297 antiretroviral gp70 autoantibodies, 293-295 autoantibodies to IgC, 298-300 autoantibodies to nuclear antigens, 286-293 coinpleinent levels, 295 cryoprecipitates, 295 immunoglobulins, 284-286 natural thyrnocytotoxic antibody, 297-298 treatment of, 350-355 viruses in, 348-349

T Tyrosine kinase, transforming protein, properties of ~ - 2 h and, l 76-77

This Page Intentionally Left Blank

CONTENTS OF RECENT VOLUMES Volume 20 Hypervariable Regions, Idiotypy, and Antibody-Combining Site J. DONALD CAPRA A N D J. MICIIAEL KEliOE Structure and Function of the J Chain MARIAN ELLiom KOSHLAND Amino Acid Substitution and the Antigenicity of Globular Proteins MORRIS REICHLIN The H-2 Major Histocompatibility Complex and the I Immune Response Region: Genetic Variation, Function, and Organization

DONALD c. SFIREFFLER S. DAVID

A N D CllELLA

Delayed Hypersensitivity in the Mouse ALFRED J. CROWLE

SUBJECT INDEX

Volume 21 X-Ray Diffraction Studies of Immunoglobulins ROBERTO J. POLJAK Rabbit Immunoglobulin Allotypes: Structure, Immunology, and Genetics THOMAS J. KINDT Cyclical Production of Antibody as a Regulatory Mechanism in the Immune Response WILLIAM 0.WIEGLE

Thymus-Independent B-Cell Induction and ParoIy sis ANTONIO COLITINHO A N D GORAN MilLLER

SUBJECT INDEX

Volume 22 The Role of Antibodies in the Rejection and Enhancement of Organ Allografts CHARLES B. CARPENTER, ANTHONY J. F. D’APICE, A N D ABUL K. ABBAS Biosynthesis of Complement HARVEYR. COLTEN Graft-versus-Host Reactions: A Review STEPHENC . GREBEA N D J. WAYNE STREILEIN

Cellular Aspects of Immunoglobulin A MICHAEL E. LAMM Secretory Anti-Influenza Immunity YA. S. SHVARTSMAN A N D M . P. ZYKOV

SUBJECT INDEX

Volume

23

Cellular Events in the IgE Antibody Response KIMISHICE ISHIZAKA Chemical and Biological Properties of Some Atopic Allergens T. P. KING

395

396

CONTENTS O F RECENT VOLUMES

Human Mixed-Lymphocyte Culture Reaction: Genetics, Specificity, and Biological Implications BO DUPONT, J O H N A. HANSEN, AND

Antigen-Binding Myelomo Proteins of Mice

MICIIAELPOTTER Human Lymphocyte Subpopulations L. CHESSA N D S . F. SCHLOSSMAN

EDMOND J. YUNIS lmmunochemical Properties of Glycalipids and Phospholipids DONALDM. MARCUSAND GERALDA. SCHWARTlNC

SUBJECT

INDEX

Volume 26 Anaphylatoxins: C3a and C5a

SUBJECT INDEX

TONYE. HUCLIA N D HANSJ MULLER-EBERHARD

Volume 24 The Alternative Pathway of Complement Activation 0. GOTZE A N D H. J.

M ti LLER-EBERHARD Membrane and Cytoplasmic Changes in B Lymphocytes Induced by Ligand-Surface Immunoglobulin Interaction GEORGER. SCHREINERA N D EMILR. UNANUE Lymphocyte Receptors for Immunoglobulin HOWARD

JAN

KLEIN

The Protein Products of the Murine 17th Chromosome: Genetics and Structure ELLENS. VITETTA A N D J. DONALD CAPRA

Expression and Function of ldiotypes on Lymphocytes

K.

EICHMANN

The B-Cell Clanotype Repertoire

NOLAN H.

B. DICKLER

Ionizing Radiation and the Immune Response R O B E R T E. ANDERSON A N D NOEL

H-2 Mutations: Their Genetics and Effect on Immune Functions

SICAL A N D N O R M A N

R.

KLINMAN SUBJECT

L.

WERNER SUBJECT I N D E X

Volume 25 Immunologically Privileged Sites

F. B A K E R A N D R. E. BILLINGHAM

CLYDE

Maior Histocompatibility Complex Restricted Cell-Mediated Immunity GENEM. SHEARER A N D ANNE-MARIE

INDEX

Volume

27

Autoimmune Response to Acetylcholine Receptors in Myasthenia Gravis and Its Animal Model JON L I N D S T R O M

MHC-Restricted Cytotoxic T Cells: Studies on the Biological Role of Polymorphic Major Transplantation Antigens Determining T-cell Restriction-Specificity, Function, and Responsiveness

ROLF M.

ZINKERNAGEL AND PETER

D o H E n n

SCNMIIT-VERHULST

Current Status of Rat lmmunogenetics

DAVIDL. GASSER

Murine Lymphocyte Surface Antigens IAN F. c. MCKENZIEA N D T E R R Y

Pmmn

c.

CONTENTS OF RE:CENT VOLUMES

The Regulatory and Effector Roles of Eosinophils

PETERF. WELLERA N D EDWARD J. GOETZL SUBJECT INDEX

397

Antibody-Mediated Destruction of VirusInfected Cells J. G . PATRICK SISSONSA N D MICHAEL

B. A. OLDSTONE Aleutian Disease of Mink

DAVIDD. PORTER,AUSTINE. LARSEN, HELENG. PORTER

AND

Volume 28

Age Influence on the Immune System

The Role of Antigen-Specific T Cell Foctors in the Immune Response

TOMIOTADAA N D KO OKUMURA The Biology and Detection of Immune Complexes ARCYRIOS N . THEOFILOPOULOS A N D

FRANK J. DIXON The Human la System R. J. WINCHESTER A N D H. G. KUNKEL Bacterial Endotoxins and Host Immune Responses

DAVIDC. MORRISONA N D RYAN

JOHN

L.

Responses to Infection with Metazoan and Protozoan Parasites in Mice

GRAHAM F. MITCHELL

TAKASHIMAKINODANAND MARCUERRE M. B. KAY SUBJECT INDEX

Volume 30 Plasma Membrane and Cell Cortex Interactions in Lymphocyte Functions

FRANCISLOOR Control of Experimental Contact Sensitivity

HENRYN. CLAMAN, STEPHEND. MILLER, PAULJ. CONLON,A N D JOHN W. MOORHEAD Analysis of Autoimmunity through Experimental Models of Thyroiditis and Allergic Encephalomyelitis WILLIAM 0. WEICLE

SUBJECT INDEX The Virology and lmmunobiology of Lymphocytic Choriomeningitis Virus Infection

Volume 29 Molecular Biology and Chemistry of the Alternative Pathway of Complement

HANSJ. MULLER-EBERHARD AND ROBERT D. SCHREIBER Mediators of Immunity: Lymphokines and Monokines

Ross E. ROCKLIN,KLAUS BENDTZEN, DIRKGREINEDER

AND

M. J. BUCHMEIER, R. M. WELSH,F. J. DUTKO,A N D M. B. A. OLDSTONE INDEX

Volume 31 The Regulatory Role of Macrophages in Antigenic Stimulation Part Two: Symbiotic Relationship between Lymphocytes and Macrophages

EMILR. UNANUE Adaptive Differentiation of Lymphocytes: Theoretical Implications for Mechanisms of Cell-Cell Recognition and Regulation of Immune Responses

DAVIDH.KAIZ

T-cell Growth Factor and the Culture of Cloned Functional T Cells KENDALLA. SMITHAND FRANCIS W.

RUSCETII

398

C O N T E N T S OF R E C E N T VOLUMES

Formation of B Lymphocytes in Fetal and Adult Life PAULW. KINCADE Structural Aspects and Heterogeneity of Immunoglobulin Fc Receptors JAY C. UNKELESS, HOWARD FLEIT, A N D IRAS. MELLMAN The Autologous Mixed-Lymphocyte Reaction MARC E . WEKSLER, CHARLES E. MOODY, JR., A N D ROBERT W. KOZAK

INDEX

Volume 32 Polyclonal B-Cell Activators in the Study of the Regulation of Immunoglobulin Synthesis in the Human System T H O M A S A. W A L D M A N N A N D SAMUEL BRODER Typing for Human Alloantigens with the Prime Lymphocyte Typing Technique N. MORLING, B. K. JAKOBSEN,P. PLATZ, L. P. RYDER, A. SVEJCAAHD, A N D M . THOMSEN Protein A of Stophylococcus aureus and Related Immunoglobulin Receptors Pioduced by Streptococci and Pneumonococci JOHN J. LANCONE Regulation of Immunity to the Azobenzene-arsonate Hapten MARKI. GREENE,MITCHELLJ. NELLES, MAN-SUN SY,AND ALFRED NISONOFF Immunologic Regulation of Lymphoid Tumor Cells: Model Systems for Lymphocyte Function ABUL K. ABBAS

INDEX

Volume 33 The CBA/N Mouse Strain: An Experimental Model Illustrating the Influence of the X-Chromosome on Immunity IRWIN SCHER The Biology of Monoclonal Lymphokines Secreted by T Cell Lines and Hybridomas AMNONALTMAN A N D D A V I D H. KATZ Autoantibodies to Nuclear Antigens (ANA): Their lmmunobiology and Medicine E N G M . TAN The Biochemistry and Pathophysiology of the Contact System of Plasma CHARLES G. C O C t I R A N E AND JOHN H. GRIFFIN Binding of Bacteria to Lymphocyte Subpopulations MARIUSTEODORESCU A N D EUGENE P. MAYER

INDEX

Volume 34 T Cell Alloontigens Encoded by the IgT-C Region of Chromosome 12 in the Mouse F . L. OWEN Heterogeneity of H-2D Region Associated Genes and Gene Products TEDH. HANSEN, KEIKO OZATO,A N D DAVIDH. SACHS Human Ir Genes: Structure and Function T H O M A S A. GONWA, B. MATIJA PETERLIN, A N D J O H N D. STOBO Interferons with Special Emphasis on the Immune System ROBERT M . FRIEDMAN A N D STEFANIE N. VOCEL Acute Phase Proteins with Special Reference to C-Reactive Protein and Related

CONTENTS OF RECENT VOLUMES

Proteins (Pentaxins) and Serum Amyloid A Protein M. B. PEPYSA N D MARILYNL. BALTZ

399

The Influence of Histamine on Immune and Inflammatory Responses DENNISJ. BEER, STEVEN M . MATLOFF. A N D ROSS E. ROCKLIN

Lectin Receptors as Lymphocyte Surface Markers NATHANSHARON

INDEX

INDEX

Volume 36

Volume 35

Antibodies of Predetermined Specificity in Biology and Medicine RICHARDALAN LEHNEH

The Generation of Diversity in Phosphorylcholine-Binding Antibodies ROGER M . PERLMUT~ER, STEPHENT. CREWS,RICHARDDOUGLAS, GREG SOHENSEN,NELSON JOHNSON,NADINE NIVERA,PATRICIA J. GEARHART,A N D LEROYHOOD

A Molecular Analysis of the Cytolytic Lymphocyte Response STEVENJ. BURAKOFF,OFRA WEINBERGER,ALAN M. KHENSKY,A N D CAROLS. REISS

Immunoglobulin RNA Rearrangements in B Lymphocyte Differentiation JOHN ROGERS A N D h N D O L P l f WALL Structure and Function of Fc Receptors for IgE on Lymphocytes, Monocytes, and Macrophages HANS L. SPIEGELBEHC The Murine Antitumor Immune Response and Its Therapeutic Manipulation ROBERTJ. NORTH Immunologic Regulation of Fetal-Maternal Balance DAVID R. JACOBY,LARSB. OLDING, A N D MICHAELB. A. OLDSTONE

The Human Thymic Microenvironment BARTONF. HAYNES Aging, ldiotype Repertoire Shifts, and Compartmentalization of the Mucosal-Associated Lymphoid System ANDREWW. WADEA N D MYRONR . SZEWCZUK A Major Role of the Macrophage in Quantitative Genetic Regulation of Immunoresponsiveness and Antiinfectious Immunity CUIDOBIOZZI, DENISEMOUTON, CLAUDE STIFFEL, A N D YOLANDE BOUTHILLIER

INDEX

This Page Intentionally Left Blank