Annual Review of Immunology Volume 18 2000

Annual Review of Immunology Volume 18, 2000 CONTENTS Annu. Rev. Immunol. 2000.18:1-17. Downloaded from arjournals.annu...

Author: Annual Reviews

43 downloads 1524 Views 8MB Size Report

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

DOWNLOAD PDF

Annual Review of Immunology Volume 18, 2000

CONTENTS

Annu. Rev. Immunol. 2000.18:1-17. Downloaded from arjournals.annualreviews.org by HINARI on 09/01/07. For personal use only.

Discovering the Role of the Major Histocompatibility Complex in the Immune Response, Hugh O. McDevitt Receptor Selection in B and T Lymphocytes, David Nemazee Molecular Basis of Celiac Disease, Ludvig M. Sollid Population Biology of Lymphocytes: The Flight for Survival, Antonio A. Freitas, Benedita Rocha Nonclassical Class II MHC Molecules, Christopher Alfonso, Lars Karlsson Negative Regulation of Cytokine Signaling Pathways, Hideo Yasukawa, Atsuo Sasaki, Akihiko Yoshimura T Cell Activation and the Cytoskeleton, Oreste Acuto, Doreen Cantrell The Specific Regulation of Immune Responses by CD8+ T Cells Restricted by the MHC Class Ib Molecule Qa-1, Hong Jiang, Leonard Chess The Biology of Chemokines and their Receptors, Devora Rossi, Albert Zlotnik Dendritic Cells in Cancer Immunotherapy, Lawrence Fong, Edgar G. Engleman CD8 T Cell Effector Mechanisms in Resistance to Infection, John T. Harty, Amy R. Tvinnereim, Douglas W. White Glucocoricoids in T Cell Development and Function, Jonathan D. Ashwell, Frank W. M. Lu, Melanie S. Vacchio Molecular Genetics of Allergic Diseases, Santa Jeremy Ono Immunology at the Maternal-Fetal Interface: Lessons for T Cell Tolerance and Suppression, A. L. Mellor, D. H. Munn Regulation of B. Lymphocyte Responses to Foreign and Self-Antigens by the CD19/CD21 Complex, Douglas T. Fearon, Michael C. Carroll, Michael C. Carroll Regulatory T Cells in Autoimmunity, Ethan M. Shevach Signal and Transcription in T Helper Development, Kenneth M. Murphy, Wenjun Ouyang, J. David Farrar, Jianfei Yang, Sheila Ranganath, Helene Asnagli, Maryam Afkarian, Theresa L. Murphy The RAG Proteins and V (D) J Recombination: Complexes, Ends, and Transposition, Sebastian D. Fugmann, Alfred Ian Lee, Penny E. Shockett, Isabelle J. Villey, David G. Schatz The Role of the Thymus in Immune Reconstitution in Aging, Bone Marrow Transplantation, and HIV-1 Infection, Barton F. Haynes, M. Louise Markert, Gregory D. Sempowski, Dhavalkumar D. Patel, Laura P. Hale Accessing Complexity: The Dynamics of Virus-Specific T Cell Responses, Peter C. Doherty, Jan P. Christensen The Role of Chemokine Receptors in Primary, Effector, and Memory Immune Responses, Federica Sallusto, Charles R. Mackay, Antonio Lanzavecchia Phosphorylation Meets Ubiquiination: The Control of NF-Kappa-B Activity, Michael Karin, Yinon Ben-Neriah Reservoirs for HIV-1: Mechanisms for Viral Persistence in the Presence of Antiviral Immune Responses and Antiretroviral Therapy, Theodore Pierson, Justin McArthur, Robert F. Siliciano

1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665


Regulation of Antibody Responses via Antibodies, Complement, and Fc Receptors, Birgitta Heyman Multiple Roles for theMajor Histocompatibility Complex Class I--Related Receptor, FcRn, Victor Ghetie, E. Sally Ward Immunobiology of Dendritic Cells, Jacques Banchereau, Francine Briere, Christophe Caux, Jean Davoust, Serge Lebecque, Yong-Jun Liu, Bali Pulendran, Karolina Palucka An Address System in the Vasculature of Normal Tissues and Tumors, E. Ruoslahti, D. Rajotte Genomic Views of the Immune System, Louis M. Staudt, Patrick O. Brown

767

Viral Subversion of the Immune System, Domenico Tortorella, Benjamin E. Gewurz, Margo H. Furman, Danny J. Schust, Hidde L. Ploegh

861

DNA Vaccines: Immunology, Application, and Optimization, Sanjay Gurunathan, Dennis M. Klinman, Robert A. Seder Gamma Delta Cells: A Right Time and a Right Place for a Conserved Third Way of Protection, Adrian C. Hayday

709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:1–17 Copyright q 2000 by Annual Reviews. All rights reserved

DISCOVERING THE ROLE OF THE MAJOR HISTOCOMPATIBILITY COMPLEX IN THE IMMUNE RESPONSE Hugh O. McDevitt Departments of Microbiology and Immunology and Medicine, Stanford University School of Medicine, Stanford, California 94305; e-mail: [email protected]


Key Words immunogenetics, class II MHC, autoimmunity, genetic susceptibility, polymorphism Abstract The discovery that genes in the major histocompatibility complex (MHC) play an important role in the immune response depended on the chance interaction of several unrelated events. The first, and most important, was the decision by Michael Sela to synthesize a series of branched, multichain, synthetic polypeptides based on a backbone of poly-l-lysine. The prototype compound, (T,G)-A–L, was tipped with short random sequences of tyrosine and glutamic acid. This resulted in a restricted range of antigenic determinants composed of only two or three amino acids with a variable length—ideal for binding to the peptide binding groove of MHC class II molecules. The second was the decision by John Humphrey to immunize various strains of rabbits with this synthetic polypeptide. Two of these rabbit strains showed very large quantitative differences in antibody response to (T,G)-A–L. In transferring this system to inbred mouse strains, the third bit of good fortune was the availability at the National Institute of Medical Research, in Mill Hill (London), of the CBA (H2k) and C57 (H2b) strains. The H2b haplotype is the only one mediating a uniform high antibody response to (T,G)-A–L. The fourth critical ingredient was the availability of numerous congenic and H2 recombinant inbred strains of mice produced earlier by Snell, Stimpfling, Shreffler, and Klein. A search for congenic pairs of mice expressing the responder and nonresponder H2 haplotypes on the same background revealed that these strains responded as a function of their H2 haplotype, not of their inbred background. Extensive studies in a variety of inbred strains carrying recombinant H2 haplotypes, as well as a four-point linkage cross, mapped immune response to (T,G)A–L within the murine MHC, between the K and Ss loci. The demonstration that stimulation in the mixed lymphocyte reaction (MLR) mapped to the same region quickly led to attempts to produce antisera in congenic H2 recombinant strain combinations. These antisera identified I-region associated (Ia) antigens. Immunoprecipitation and blocking studies showed that the gene products controlling specific immune responses, the mixed lymphocyte reaction, and the structure of Ia antigens were one and the same—now designated as the I-A MHC class II molecules. These antisera and inbred strains enabled Unanue to demonstrate the peptide binding function of class II MHC molecules. 0732–0582/00/0410–0001$14.00

1

2

MCDEVITT


BEGINNINGS—‘‘as the twig is bent—’’ Chance plays a greater part in our lives than most of us are willing to admit. My father was born in Cincinnati, Ohio, in 1888, into a first generation Irish Catholic family with some pretensions, largely on my grandmother’s part, to have made the transition from shanty Irish to ‘‘lace curtain’’ Irish. My grandfather, who ran a large drayage business in Cincinnati, was a convivial Irishman devoted to trotting horses, the Democratic party, and Irish camaraderie. He could, however, be a stern parent. After graduating from St. Xavier High School, my father enrolled as a freshman at St. Xavier University in Cincinnati. Quite early in his college career, he encountered a Jesuit priest who stated flatly that Darwin’s theory of evolution was wrong. This led to a heated discussion, a threat of corporal punishment by the priest, and forcible physical resistance by my father. Naturally enough for those times, the incident resulted in his expulsion from the University. My grandfather’s response was to conclude that my father’s chance for higher education was at an end. Through friends in the Democratic party (which was then in power in Cincinnati), he found my father a position as a rodman on a surveying gang, laying out the north part of Vine Street. During his year in that position, my father noticed the sad and empty expressions on many of the workers on the surveying team, and the way they looked forward to a lunch box containing a thermos filled with whiskey-laced coffee. The experience stimulated him to look about for other opportunities for further education. At that time, it was possible to enroll in medical school directly out of high school, and accordingly (with his mother’s support), he enrolled in the University of Cincinnati College of Medicine in 1907, graduating in 1911. In medical school, he was exposed to Ehrlich’s side chain theory, which at the time had attracted wide attention and interest (1). Much later, working as a urologist treating syphilis and other venereal diseases, he became acquainted with the uses of arsphenamine and 606 to treat syphilis, and with Ehrlich’s concept of finding ‘‘magic bullet’’ drugs, which would selectively bind to ‘‘side chains’’ on the spirochete and destroy it, while failing to bind to ‘‘side chains’’ on host tissues. As a young child of 8 or 9, I would often sit on our front porch with my father, watching late afternoon summer thunderstorms pass over our small village just outside of Cincinnati. During those discussions, my father would sometimes talk of his tremendous admiration for Ehrlich, the side chain theory, and the many drugs he developed for the treatment of syphilis. He was particularly amused by the fact that Ehrlich allegedly flunked his histology course in medical school but went on to develop many tissue-specific stains and to become an outstanding research scientist. When I asked my father to explain Ehrlich’s side chain theory, he would answer that that was very difficult, but I would understand it once I was in medical school. (Having failed to force my two older brothers to pursue science in high school, or to apply themselves to premedical studies, he switched


ROLE OF MHC IN IMMUNE RESPONSE

3

to furthering his desire to have one of his sons become a physician by indirection. He succeeded in making medicine and research seem like a wonderful career.) These early influences were reinforced by my father’s announcement, while I was in the third grade, that I no longer had to attend church on Sunday (a duty that held little appeal for me). Instead, I was invited to accompany him into the city as he made his rounds at various hospitals, following which we would pick up a quart of ice cream for Sunday dinner at the city’s leading candy and confectionery store. In the summer time, this routine was sometimes changed, and my mother and several of my four older siblings would join us for a steak lunch at Mecklenburg’s German Beer Garden. Mecklenburg’s had a large outdoor dining area covered with a beautiful grape arbor. To me, it seemed like an enchanted garden. The result of all this was that I came to view a career in medicine as fascinating, fun, and certainly worth the effort. During my undergraduate years at Stanford, two experiences made a possible career in research both more appealing and more concrete. The first was a course in evolution taught by Richard Goldschmidt, a visiting professor from the University of California at Berkeley. Goldschmidt had developed a theory of evolution that he termed macroevolution (2, 3). He presented his theory at the end of a fascinating series of lectures reviewing the evidence then available in support of Darwin’s original theory. The second experience was the opportunity to do independent research in genetics under the supervision of Raymond Barratt in the department of biology. The Stanford Department of Biology in the 1940s and 1950s had become prominent due to the work of Beadle and Tatum, who used UV radiation-induced mutants of Neurospora to map out basic biosynthetic pathways. My project was to map the adenineless locus in the first Neurospora linkage group, using a four-point mapping cross. The results were inconclusive because the loci chosen were too far away from the adenineless locus. However, the techniques and concepts were identical to those used to map the Ir-1 gene (controlling the immune response to a synthetic polypeptide antigen), after linkage between that gene and the major histocompatibility complex had been established (see below). These two experiences strengthened an interest in research, but research itself was put aside for several years during medical school, internship, residency, and service in the army, which was then required by the doctor’s draft law and the Berry plan.

THE FATE OF ANTIGEN AND GENETIC CONTROL OF THE IMMUNE RESPONSE Following two years of service as an ‘‘obligated volunteer’’ in the Army Medical Corps, I spent two years learning the fundamentals of immunology under the supervision of Albert Coons, who had invented and perfected the fluorescent


4

MCDEVITT

antibody technique (4). Looking back now, it requires a major shift in mental perspective to recall how very rudimentary our knowledge of immunology and the mechanics of the immune response was in 1959, at the start of this first postdoctoral fellowship. A 1960 paper by John Humphrey in the Proceedings of the first Antibody Workshop (5) caught my attention. In this paper, Humphrey proposed to test the instructive theory of antibody formation, originally put forward by Pauling and Haurowitz (6). This theory in its simplest form postulated that the polypeptide comprising the antibody molecule was folded around the antigen as a template in the antibody producing cell. Coons and others had shown that antibody producing plasma cells contained and secreted large numbers of antibody molecules. If the instructive theory was correct, this would require the presence of appreciable amounts of antigen in the antibody producing cell. Humphrey proposed to utilize a synthetic polypeptide antigen developed by Michael Sela (7) to obtain a radioactive polypeptide with high specific activity. The antigen, (T,G)-A-L, is a branched multichain synthetic polypeptide with a backbone of poly-L-lysine, side chains of poly-D,L-alanine, and short, random sequences of tyrosine and glutamic acid at the ends of the side chains. Humphrey predicted that by synthesizing this peptide with dehydroalanine, and then reducing the double bonds with tritium gas, it should be possible to detect a very small number of (T,G)-A-L molecules in a single antibody-producing cell (Figure 1). If no antigen were to be found in an antibody-producing cell, it followed that antibody structure must be a function of primary amino acid sequence and that the selective theory of antibody formation (8–10) must be correct. It occurred to me that it might be possible to combine immunofluorescence with autoradiography to detect radioactive (T,G)-A-L in specific antibody-producing plasma cells.

Figure 1 A schematic diagram of the structural pattern of the multichain synthetic polypeptide antigens used in this study. The numbers designate different preparations of each polypeptide.



5

The proposal appealed to John Humphrey, and he accepted me for a two-year postdoctoral fellowship in his laboratory at the National Institute for Medical Research at Mill Hill, outside of London. In carrying out these studies, I was fortunate to enjoy the supervision of Brigitte Askonas as well as John Humphrey. Both taught me a great deal about scientific rigor. We found that there was no detectable antigen in specific antibody producing cells, but that is a separate story (11). After settling in at Mill Hill, my first assignment was to produce a reliable, high-titered antiserum specific for (T,G)-A-L. John Humphrey had already begun this task by immunizing several rabbits of the Sandylop strain with (T, G)-A-L in complete Freund’s adjuvant. However, the antibody response to this polypeptide antigen was almost undetectable. John wrote to Michael Sela saying that (T,G)-A-L appeared to be a poor antigen. Michael replied, with some heat, that it was a perfectly good antigen in his rabbits at the Weizmann Institute for Science in Rehovot, Israel. This remark stimulated John Humphrey to test other strains of rabbits, and accordingly, he had already immunized three strains of rabbits that were not inbred but did breed true for coat color, size, and a few other characteristics. The three strains were the original Sandylop strain, the Dutch strain, and the Himalayan (or Californian) strain. (This was the first of several instances in this experimental saga in which a good idea combined with exceptionally good luck to point the way to the next step.) After some false starts, I decided to do antigen-antibody precipitin curves on the sera from the three strains of rabbits. The results were quite striking. Himalayans uniformly produced 1–1.2 milligrams of antibody/ml of serum following immunization with (T,G)-A-L. Sandylops, on the other hand, produced only 30–50 micrograms of antibody, while Dutch rabbits were variable and produced antibody titers intermediate between these two extremes. The possibility that there might be inherited differences in the ability to produce antibody, or to make an immune response to a specific antigen, was very exciting. I proposed to John Humphrey that we attempt to test this possibility by crossing Sandylops with Himalayans, producing F1s, and then backcrossing the F1s to the two parental strains. Although there was not sufficient animal space to carry out this experiment at Mill Hill, John Humphrey negotiated with Ashley Miles, the head of the Lister Institute, to let us use two large rabbit rooms at the Institute’s animal facilities in Elstree, not far from Mill Hill. When the crosses and backcrosses were complete, immunization of all of the different offspring with (T,G)-A-L showed a clear genetic effect, but with a very complex inheritance pattern. Since these rabbit strains were not inbred, this could have been due to genetic heterogeneity in the parental strains, or to the fact that the trait under examination was determined by multiple genes. The logical next step was to attempt to find the same genetic difference in ability to respond to (T,G)-A-L in inbred animals—the obvious choice being inbred mice. At that time, the only inbred mouse strains available at Mill Hill


6

MCDEVITT

were CBA and (CBA 2 C57 BL/6) F1. Mice of the parental C57 BL/6 strain were only available through Avrion Mitchison, who very kindly provided several animals for immunization as well as for production of F1 and F1 2 parent backcross animals. This was a second stroke of luck. We now know that uniform high response to (T, G)-A-L is seen only in mice expressing the H2b haplotype, such as C57BL/6, while mice of the H2a,k,u,s,f,q haplotypes are low responders. Immunization of CBA, C57, the F1, and the two backcross populations with an optimal dose of (T,G)-A-L in complete Freund’s adjuvant revealed a clear dominant inheritance of ability to respond. CBAs were very low responders, C57s were high responders, and the F1 was intermediate. The two parental backcrosses were low and intermediate for F1 2 CBA and intermediate or high in the F1 2 C57 population (12). In discussing these results with John Humphrey and Michael Sela, I wondered if it might be possible to test the specificity of this genetic control of the immune response by immunizing the same strains with related synthetic polypeptide antigens that had a different amino acid composition. I was delighted to learn from Michael Sela that, as a part of his study on the nature of antigenicity, he had also synthesized the same type of branched synthetic polypeptide antigen in which tyrosine was replaced by either histidine or phenylalanine—to produce (H,G)-A-L and (Phe, G)-A-L. Immunization of these same strains of mice with (H,G)-A-L revealed exactly the opposite pattern—CBA mice were high responders, C57 mice were low responders, and the ability to respond was inherited as a dominant trait (7, 13). These results showed that the phenotype of this particular form of genetic control of the immune response was codominant and specific for the amino acid composition of the antigenic deterimants in the immunizing antigen. However, the mechanism of this genetic control, and the gene or genes, if this were to be a polygenic trait, were a complete mystery. The next several years were the ‘‘wilderness years,’’ devoted to testing various possibilities for the mechanism of this genetic control, combined with efforts to establish linkage of this genetic trait with known genes in the murine linkage map. Numerous mechanisms and linkages were tested, with mostly negative results. The metabolism and degradation of (T, G)-A-L was the same in CBA and C57, and no linkage could be established with immunoglobulin allotype, or with the isotype of specific anti-(T, G)-A-L antibody, coat color, or other easily scored genetic traits (14). In fact, there was no direct evidence to show that this difference in antibody response was due to a gene expressed in immunocompetent cells. It seemed equally possible that this genetic trait could be the indirect result of some gene not directly involved in antibody formation. Michael Edidin, then a postdoctoral fellow in Coons’s lab at Harvard, suggested that it might be possible to transfer the ability to respond from a high responder strain into a low responder strain by transplantation of high responder bone marrow into irradiated low responder recipients. Unfortunately, these experiments, utilizing either neonatal or adult recipients, gave variable and inconclusive results because of the frequent occurrence of graft vs. host disease and/or rejection of the bone marrow transplant.


7


DISCOVERY OF LINKAGE WITH THE MAJOR HISTOCOMPATIBILITY COMPLEX (MHC) In the fall of 1966, I took up a position as an assistant professor in the division of immunology in the Department of Medicine at Stanford University School of Medicine. I quickly became acquainted with Leonard and Leonore Herzenberg in the Department of Genetics. Our discussions covered a wide range of subjects, and I outlined to them how I hoped to show that this genetic control of a specific immune response was expressed in immunocompetent cells by transferring (CBA 2 C57) F1 spleen cells into irradiated CBA low-responder parental recipients. At that time, Len and Lee were also interacting frequently with Marvin Tyan at the Hunter’s Point naval radiation research laboratory. Marv Tyan joined in many of these discussions and offered to carry out the cell transfer experiments, since cell and bone marrow transplantation in mice was his main research interest. Accordingly, I supplied him with F1 and CBA mice to carry out the cell transfers. The results showed quite clearly that transfer of F1 responder spleen cells into irradiated low responder parental recipients transferred the ability to respond to (T,G)-A-L. This clearly established that this genetic control was transferable with immunocompetent cells, although it did not establish whether the gene(s) responsible was expressed exclusively in this cell population. While these experiments were going on, I continued to search for a means of establishing homozygous high responder lymphocyte populations in homozygous low responder recipients. As in the earlier experiments with Michael Edidin, this was difficult because CBA (H2K) and C57 (H2b) differ at the major histocompatibility complex. In one of these discussions, Lee Herzenberg mentioned that congenic strains were available from the Jackson Laboratories in Bar Harbor, Maine. These inbred congenic strains were homozygous for several different H2 haplotypes on several different inbred strain backgrounds. Thus, C3H mice were available with their original H2k MHC haplotype, or as C3H.SW mice, which are congenic with C3H except at the MHC, H2b. (See Table 1.) It was a simple matter to determine that C3H mice like CBAs were low responders, while C57 BL/10 (B10) mice (H2b) were high responders to (T,G)-A-L. Three sets of congenic strain pairs were ordered from the Jackson Laboratories (Table 1) and immunized TABLE 1 Strain

H-2 Type

C3H C3H.SW

k b

A A.BY

a b

B10 B10.BR

b k


8

MCDEVITT

with (T,G)-A-L. This experiment was done simply to establish that both of the C3H strains and both the A strains were low responders, and that both the B10 strains were high responders to (T,G)-A-L. The results of this experiment were exactly the opposite of that predicted (15): C3H.SW, A.BY, and B10 were all high responders, while C3H, A, and B10.BR were all low responders. Thus, these three congenic pairs responded to (T,G)-AL as a function of their H2 haplotype and not as a function of their inbred strain background. These results clearly implied that genetic control of the immune response to (T,G)-A-L and (H,G)-A-L was linked to the major histocompatibility complex. This was quickly confirmed by formal linkage analysis of a cross between CBA and C57 mice. The offspring of a cross between F1 high responder and the CBA low responder parent showed a clear segregation into high responders and low responders, and high response segregated almost perfectly with the H2b haplotype (15). (Progeny testing of the few H2b low responders showed that their offspring were 50% high responders and that the first immunophenotyping was incorrect.) This result, presented at several meetings and seminars in the fall of 1967, stimulated great interest on the part of George Snell, and considerable skepticism on the part of some of the other workers in the H2 field. One criticism was to the effect that the gene had to be located somewhere on the murine linkage map, and it just happened to be linked to the MHC. Snell offered to help in any way he could, and he pointed out that Donald Shreffler, Jan Klein, and Jack Stimpfling (each of whom was more than generous with their mice and reagents) had produced a number of congenic mouse strains on the B10 and A backgrounds that were homozygous for crossover events within the murine MHC (see Figure 2).

THE MURINE MAJOR HISTOCOMPATIBILITY COMPLEX IN 1968 It was immediately apparent that mapping the gene or genes determining genetic control of this specific immune response offered the most direct means of ultimately identifying the gene or genes involved. Good fortune once again was at

Figure 2 The map of the murine MHC on chromosome 17. The only definitely established loci in the murine MHC in 1968 were the K and D loci encoding the serologically identified gene products (which are now known to be the two major class I MHC molecules), and the Ss locus, identified by a serological polymorphism in a serum protein, now known to be the fourth component of complement.

9



hand, this time in the form of a large number of congenic strains of mice homozygous for crossover events known to have occurred either between the K and Ss loci or between the Ss and D loci. These various strain combinations are listed in Table 2. They had been produced over the preceding several years by Shreffler, Klein, and Stimpfling, with the goal of creating a series of crossover MHC haplotypes that could be used for further detailed mapping of the murine MHC. Two approaches were taken to determine the map position of the genetic control of the immune response to (T,G)-A-L and (H,G)-A-L. (Because this was the first genetic control of the immune response to be linked in the murine linkage map, it was given the gene designation Ir-1, for Immune response – 1.) The first approach utilized 11 inbred mouse strains and lines, which were homozygous for a known crossover event within the MHC. Nine of these H2 recombinants, with the position of the crossover event, are listed in Table 2. They are derived from crossovers between H2a and H2b, between H2k and H2d, between H2s and H2al, and between H2q and H2a. Immunization of these strains with (T,G)-A-L, (H,G)A-L, and (Phe,G)-A-L showed that 9 of the 11 strains responded as did the donor of the K locus end of the recombinant H2 chromosomes. This would map Ir-1 near or to the centromeric side of the K locus. However, two of these recombinants, H2tl and H2y, gave the opposite result; they responded as did the donor of the D locus end of the recombinant H2 chromosomes. Both of these latter two H2 recombinants arose from crossovers between the K and Ss loci. These results would be compatible either with the mapping of the Ir-1 gene within the MHC, between the K and Ss loci, or they could have been due to a double crossover event. However, the H2y recombinant was derived from a known single crossover event because of outside markers in the cross giving rise to this H2 recombinant. Thus, this result almost certainly mapped Ir-1 within the murine MHC, lying between the K and Ss loci. The results described above were confirmed by the second approach, utilizing a four-point mapping cross. This cross utilized a strain developed by Margaret TABLE 2 Strain

H2 Haplotype

B10.A (1R) B10.A (2R) B10.A (3R) B10.A (4R) B10.A (5R) H2al H2tl H2th AQR

h h i h i (not inbred) al (not inbred) tl (not inbred) th (not inbred) Y

Crossover at K-Ss

Crossover at Ss-D x x

x x x x x x x


10

MCDEVITT

Green, carrying an H2q chromosome with additional markers approximately 10 centimorgans (brachyury) and 5 centimorgans (tufted) centromeric to the H2 complex. This mapping cross is diagrammed in Figure 3. Complete analysis of the offspring of this mapping cross required the testing of more than 1500 mice because the offspring of this cross were typed and confirmed by progeny testing. In 484 offspring, only 2 recombinants were detected. Both were confirmed by progeny testing, which showed that the crossover event had occurred within the H2 region itself. Thus, the two approaches were concordant in mapping Ir-1 within the murine MHC, between the K and Ss Loci (16) (Figure 3). The map position of the Ir-1 gene within the MHC suggested very strongly that the gene product, then completely unknown, might in some way be related to the known genes within the major histocompatibility complex. However, the striking antigenic specificity of the Ir-1 gene effect, combined with the recent demonstration by Mitchell and Miller (17, 18) that T cells had separate, distinct regulatory functions over B cells, raised the possibility, first put forward by Benacerraf, that the Ir-1 gene(s) might encode an antigen-specific receptor on T cells. I, and many others (see below), found this hypothesis very appealing, particularly because of the sharp discrimination of the Ir-1 genetic control, which could readily differentiate between (T,G)-A-L (H2b) and (Phe, G)-A-L. Mice carrying the H2q haplotype are capable of responding to (Phe, G)-A-L but do not respond to (T,G)-A-L, although the antibodies are quite cross-reactive. (At this point, a warning and a disclaimer about the accuracy of memory versus the written record is appropriate. For example, the description given above of the mapping of the Ir-1 gene is a reasonable statement of my memory of the process. In looking up the original papers on this subject, I was stunned to find a paper (19) presented at a 1969 symposium on immunogenetics of the H2 system, at Liblice, Czechoslovakia, in which I stated that the results to date indicated that Ir-1 was either within the H2 locus or near the left-hand margin and that ‘‘complete resolution may not be possible on the basis of the present experiments.’’ Continued analysis of that mapping cross ultimately produced the result described above, but I had little memory of my gloomy assessment of the progress of that experiment in late 1969. Later in this discussion, I mention one or two other instances in which my memory of the progression of thinking about these experiments and their implications does not completely fit the written record. I am now much more in agreement than I was previously with those historians of science who insist that it is only the written record of science that can be relied upon.)

FINDING THE IR-1 GENE PRODUCT In the spring of 1972, Maurice Landy enlisted my help in organizing the last of a series of five NIAID-supported symposia on immunology entitled ‘‘Genetic Control of Immune Responsiveness.’’ The symposium covered a wide variety of topics, with particular emphasis on MHC-linked genetic control of immune

11


Figure 3 T is the short tail allele at the brachyury locus, while tf is the tufted phenotype allele at the tufted locus. These are linked to H2q and the mapping cross was carried out with an F1 between this linkage testing strain and the H2s haplotype from A.SW mice.


12

MCDEVITT

responses, immunoglobulin allotype–linked immune response genes, and several related topics. In my memory, the most exciting new finding was the presentation by Fritz Bach and Jan Klein. These colleagues had noticed a 1971 study by Yunis & Amos (20) showing that ability to stimulate in a mixed lymphocyte reaction (MLR) was linked to HLA, but separable by recombination from the genes encoding HLA A and B in three families with one recombinant offspring in each family. Bach and Klein used mice carrying recombinant H2 haplotypes to search for the same phenomenon in the mouse. Their results showed that a gene or genes mapping in the same position as Ir-1 encoded structures on the surface of lymphocytes that were responsible for the strongest stimulation in the mixed lymphocyte culture reaction (MLR) (21). Since MLR stimulation involved a recognition by T cells of alloantigenic structures on other lymphocytes, the implication was clear that at least the structures responsible for stimulation in MLR were expressed on the surface of lymphocytes, and that it should therefore be possible to produce antisera to these molecules (later called Ia antigens) by immunizing one inbred strain with lymphocytes from another inbred strain that was genetically identical except for differences in the Ir-1 region of the MHC. (For example, A.TL and A.TH are two inbred strains with identical backgrounds and identical MHCs, except for Ir-1 and Ss). These MLR studies used many of the same H2 recombinant strains that had been used to map Ir-1 and raised the possibility that the Ir-1 gene product and the MLR stimulation gene product might be similar or identical gene products. In fact, a review of this study (21) again jogged my memory because I had forgotten that Bach and Klein had suggested that ‘‘it is possible that the Ir product is the T cell receptor and that it is this same molecule which can act as the stimulatory agent in MLC.’’ In any event, the identical map position of Ir-1 and MLR stimulation immediately stimulated several laboratories to attempt to produce antisera to the gene products of this region, by the method described above. There were two other very important findings presented at this conference. The first is the set of experiments by Gene Shearer and Edna Mozes using limiting dilution cell transfer techniques, which indicated that Ir-1 gene products are expressed on B cells. As is know well known, this ultimately proved to be correct, and my refusal to accept Shearer and Mozes’ results at face value, and my stubborn belief that Ir-1 encoded some form of T cell receptor, remain a source of embarrassment to this day. The Irish may be lucky, but they can also be stubborn (22). The other exciting new finding at this Brook Lodge conference was the report from Ira Green that guinea pig anti-MHC alloantisera inhibited antigen-induced proliferation of responder lymphocytes in an Ir gene–specific manner. Green also presented preliminary data which suggested that T cells could only respond to an Ir gene–controlled antigen when the antigen was presented by cells sharing the responder Ir genotype (22). These results are discussed further below. Immediately following the Brook Lodge conference in the fall of 1972, a number of laboratories undertook the production of antisera specific for I region



13

gene products by immunizing in strain combinations, such as A.TL and A.TH, which differed only at the I region, or at the I region and the Ss locus. In my laboratory, these sera were produced by Beverly Deak and Gunter Hammerling. By the late spring of 1973, Gunter and Bev were finding that these antisera exhibited complement-mediated cytotoxicity primarily for B cells of the immunizing genotype. These results were contrary to the expectation of many of us, and Gunter and I had many extensive discussions about all of the proper controls that needed to be done. However, by the fall of 1973, the results showed beyond any doubt that anti-I region sera identified a polymorphic protein(s) that was expressed primarily on the surface of B cells and macrophages and was not expressed on T cells (23). Although both Shreffler and Klein’s initial results suggested that I region gene products were expressed on T cells, further testing, as well as the independent findings of David Sachs, confirmed that I region antigens were expressed primarily on B lymphocytes (see references in Reference 23). It is beyond the scope of this essay to describe the numerous experiments carried out over the next several years, which finally established beyond any doubt that the I region contains several genes encoding class II MHC molecules (also called Ia antigens) that are responsible for the observed genetic control of specific immune responses and for stimulation in the mixed lymphocyte culture reaction. Among the most notable findings in this regard were the results presented by Green at Brook Lodge (see above). These findings led to several studies by Shevach and Rosenthal (24, 25) showing that interaction between antigen-presenting macrophages and T lymphocytes occurred only when the two cell types were histocompatible for at least one MHC haplotype. When the antigen under study was also under Ir gene control, (responder 2 non-responder) F1, T cells could only respond when the antigen was presented on macrophages of responder genotype. Presentation of the antigen by nonresponder macrophages failed to stimulate F1 responder T cells (25). In a sense, these observations were ahead of their time. In retrospect, these observations, particularly the latter experiments (25), are a clear demonstration of the expression of Ir genes in macrophages and of MHC restriction on the ability of macrophages and T cells to interact. However, at the time these observations were made, the Ir gene products—the Ia antigens or class II MHC molecules—had not been described; little was known about the guinea pig MHC; and the necessary but complex terminology made it difficult to completely understand the now obvious implications of these experiments. An even earlier demonstration, using anti-H2 antisera, of the dependence of T-B cooperation upon products of the MHC was published by Kindred and Shreffler in 1972 (26). However, the antigenic system used was not one known to be under Ir gene control, and the full implication of this finding, as with the other findings noted above, was not widely appreciated. The clear evidence from my own and other laboratories that molecules encoded in the I region were expressed on B cells and macrophages was difficult to reconcile with an equally persuasive set of results, also from my laboratory, obtained


14

MCDEVITT

by Katie Bechtol and John Freed using tetraparental embryo fusion chimeras (27), which indicated that Ir-1 was expressed in T cells and not in B cells. Resolution of this paradox came when Joan Press repeated these experiments using tetraparental bone marrow chimeras assayed by a plate-binding antibody assay (28). Antisera from both the tetraparental embryo fusion and bone marrow chimeras showed that almost all of the specific antibody was of responder allotype. Our resulting publication (28) was described by Jan Klein as the ‘‘Oops, we goofed’’ paper (29). The resolution of the tetraparental results, as well as findings from many other laboratories, finally established that the Ir-1 gene product, Ia antigens, and the stimulating molecules in the MLR are all one and the same entity—namely class II MHC molecules. However, the mechanism by which class II MHC molecules influenced the immune response to specific antigens remained a mystery for several more years. The solution to that mystery has required the contributions of a large number of laboratories, the cloning and sequencing of many class II MHC molecules by my own (30, 31) and many other laboratories, and most importantly, the demonstration by Unanue and colleagues that MHC class II molecules can bind a large number of different peptides 9 to 25 amino acids in length—which will undoubtedly soon be the subject of another essay in this series.

THE MHC AND DISEASE Long before class II MHC molecules were identified either in mouse or in human, it seemed reasonable to test the possibility that MHC genotype might play a role in susceptibility or resistance to infectious diseases or to autoimmune diseases. Because of my position in a Division of Immunology that was primarily concerned with rheumatologic diseases, and because of the striking autoimmunity manifested in systemic lupus erythematosus (SLE), it seemed reasonable to study the MHC genotypes in patients with this disease, compared to a normal control population. Fortune again smiled on this enterprise, because Stanford was also home at that time to Rose Payne and Julia and Walter Bodmer. Walter readily agreed to an exploratory study of HLA genotype in patients with SLE, and the study was accordingly carried out with the help of Carl Grumet. The results showed that patients with SLE had an increase in HLA B8 and B15 compared to the control population (32). These results stimulated us to publish a brief article describing the implications of these findings and suggesting that similar findings might be found in a number of autoimmune diseases (33). In the ensuing years, studies from many laboratories around the world revealed that susceptibility to most autoimmune diseases is determined in part by HLA genotype. Particularly strong associations between specific HLA haplotypes, and in some cases HLA sequence polymorphisms, and susceptibility to rheumatoid arthritis (34), Type I insulin-dependent diabetes mellitus (35, 36), and several other diseases were dis-


15

covered. It is beyond the scope of this essay to describe these results in detail. It is now clear that most autoimmune diseases are polygenic traits in which MHC genotype, and in particular class II MHC genotype, can play a major role in some instances, or a relatively minor role in other instances. One of the most important challenges today is to understand the mechanism by which particular sequence polymorphisms mediate susceptibility or resistance to specific autoimmune diseases (37).


ENDINGS—the twig is bent Preparing a reminiscence such as this has its good and bad sides. It has been wonderful fun to look up and reread some of the original studies that led to the discovery of linkage between genetic control of the ability to mount an antibody response to synthetic peptides and genes in the major histocompatibility complex. At the outset, the meaning of this genetic linkage was a mystery. Within a very few years, it became apparent that MHC genes play a major role in the development of the immune system and in the initiation of the cellular and humoral immune response. The progress that has been made in our understanding of the genetics, structure, and function of the MHC class I and class II molecules is awe-inspiring. All of us who have worked in this field over the years have been extremely fortunate to be a part of a series of major advances in our understanding of the immune system and of vertebrate biology. The bad side of preparing such a memoir lies in the similarity to preparing one’s own obituary. While it gives you an opportunity to put the best possible face on the matter, it does inspire some morbid thoughts—another notable trait of the Irish character. Fortunately, science in general, immunology in particular, and especially the study of the MHC continue to present us with mysteries and challenges that have yet to be met. The extraordinary degree of genetic polymorphism in MHC class I and class II genes, the many structurally similar genes with as yet undiscovered functions, the evolutionary origin of MHC genes, and the mechanisms by which MHC molecules determine disease susceptibility, and the ability to respond or not respond to particular peptides are all matters that have yet to be worked out in complete detail. And for many of these problems, the devil is in the details. One problem that continues to elude us is the understanding of the molecular mechanism by which MHC class I or class II sequence polymorphisms determine susceptibility to a wide variety of autoimmune diseases. That problem continues to tantalize, and there may yet be time to see it through to its solution. The twig, once bent, stays that way.

Visit the Annual Reviews home page at www.AnnualReviews.org.

16

MCDEVITT


LITERATURE CITED 1. Ehrlich P. 1900. The Croonian Lecture to the Royal Society. Proc. R. Soc. London 66:424 2. Goldschmidt R. 1940. The Material Basis of Evolution. New Haven, CT: Yale Univ. Press 3. Gould SJ. 1977. The return of hopeful monsters. Nat. Hist. 86:22–30 4. Coons AH, Creech HJ, Jones RN. 1941. Proc. Soc. Exp. Biol. Med. 47:200 5. Humphrey JH. 1960. Methods for detecting foreign antigen in cells, and their sensitivity. In Mechanisms of Antibody Formation. Proc. Symp., Prague, May 27–31, 1959. Prague: Czech. Acad. Sci. 6. Pauling L. 1940. J. Am. Chem. Soc. 62:2643 7. Sela M, Fuchs S, Arnon R. 1962. Studies on the chemical basis of the antigenicity of proteins 5. Synthesis and characterization of some multichain and linear polypeptides containing tyrosine. Biochem. J. 85:223 8. Jerne NK. 1955. The natural selection theory of antibody formation. Proc. Natl. Acad. Sci. USA 41:849 9. Talmage DW. 1959. Immunological specificity. Unique combinations of selected natural globulins provide an alternative to the classical concept. Science 129:1643 10. Lederberg J. 1959. Genes and antibodies. Do antigens bear instructions for antibody specificity or do they select cell lines that arise by mutation? Science 129:1649 11. McDevitt HO, Askonas BA, Humphrey JH, Schechter I, Sela M. 1966. The localization of antigen in relation to specific antibody-producing cells. I. Use of a synthetic polypeptide [(T,G)-A—L] labeled with iodine-125. Immunology 11(4): 337–51 12. McDevitt HO, Sela M. 1965. Genetic control of the antibody response. I. Demonstration of determinant-specific differ-

13.

14.

15.

16.

17.

18.

19.

ences in response to synthetic polypeptide antigens in two strains of inbred mice. J. Exp. Med. 122(3):517–31 McDevitt HO, Sela M. 1967. Genetic control of the antibody response. II. Further analysis of the specificity of determinant-specific control, and genetic analysis of the response to (H,G)-A–L in CBA and C57 mice. J. Exp. Med. 126(5):969–78 McDevitt HO. 1968. Genetic control of the antibody response. III. Qualitative and quantitative characterization of the antibody response to (T,G)-A–L in CBA and C57 mice. J. Immunol. 100(3):485– 92 McDevitt HO, Tyan ML. 1968. Genetic control of the antibody response in inbred mice: transfer of response by spleen cells and linkage to the major histocompatibility [H-2] locus. J. Exp. Med. 128(1):1–11 McDevitt HO, Deak BD, Shreffler DC, Klein J, Stimpfling JH, et al. 1972. Genetic control of the immune response. Mapping of the Ir-1 locus. J. Exp. Med. 135(6):1259–78 Miller JFAP, Mitchell GF. 1968. Cell to cell interaction in the immune response. I. Hemolysin-forming cells in neonatally thymectomized mice reconstituted with thymus or thoracic duct lymphocites. J. Exp. Med. 128:801–20 Mitchell GF, Miller JFAP. 1968. Cell to cell interaction in the immune response. II. The source of hemolysin-forming cells in irradiated mice given bone marrow and thymus or thoracic duct lymphocytes. J. Exp. Med. 128:821–37 McDevitt HO, Shreffler DC, Snell GD, Stimpfling JH. 1970. Genetic control of the antibody response: genetic mapping studies of the linkage between the H-2 and Ir-l loci. In Immunogenetics of the H2 System, Proc. Symp., Liblice, Czech.,


20.

21.

22.


23.

24.

25.

26.

27.

28.

1970, ed. M Vojtiskova, A Lengerova, pp. 69–75. Basel, Switzerland: Karger Yunis EJ, Amos DB. 1971. Three closely linked genetic systems relevant to transplantation. Proc. Natl. Acad. Sci. USA 68:3031–35 Bach FH, Widmer MB, Bach ML, Klein J. 1972. Serologially defined and lymphocyte-defined components of the major histocompatibility complex in the mouse. J. Exp. Med. 136:1430–44 McDevitt HO, Landy M. 1972. Genetic Control of Immune Responsiveness. New York/London: Academic Hammerling GJ, Deak BD, Mauve G, Hammerling U, McDevitt HO. 1974. Blymphocyte alloantigens controlled by the I-region of the major histocompatibility complex in mice. Immunogenetics 1(1):68–81 Rosenthal AS, Shevach EM. 1973. Function of macroophages in antigen recognition by guinea pig T lymphocytes. I. Requirement for histocompatible macrophages and lymphocytes. J. Exp. Med. 138:1194–1212 Shevach EM, Rosenthal AS. 1973. Function of macroophages in antigen recognition by guinea pig T lymphocytes. II. Role of the macrophage in the regulation of genetic control of the immune response. J. Exp. Med. 138:1213–29 Kindred B, Shereffler DC. 1972. H-2 dependence of cooperation between T and B cells in vivo. J. Immunol. 109:940 Bechtol KB, Freed JH, Herzenberg LA, McDevitt HO. 1974. Genetic control of the antibody response to polyL(Tyr,Glu)-poly-D,L-Ala—Lys in C3H in equilibrium with CWB tetraparental mice. J. Exp. Med. 140(6):1660–75 Press JL, McDevitt HO. 1977. Allotypespecific analysis of anti-(Tyr,Glu)-AlaLys antibodies produced by Ir-1A high

29.

30.

31.

32.

33.

34.

35.

36.

37.

17

and low responder chimeric mice. J. Exp. Med. 146(6): 1815–20 Klein J. 1986. Natural History of the Major Histocompatibility Complex. New York: Wiley. p. 542 Mathis DJ, Benoist CO, Williams VE II, Kanter MR, McDevitt HO. 1983. The murine Ea immune response gene. Cell 32(3):745–54 Benoist CO, Mathis DJ, Kanter MR, Williams VE II, McDevitt HO. 1983. Regions of allelic hypervariability in the murine Aa immune response gene. Cell 34(1):169–77 Grumet FC, Coukell A, Bodmer JG, Bodmer WF, McDevitt HO. 1971. Histocompatibility (HLA) antigens associated with systemic lupus erythematosus. A possible genetic predisposition to disease. N. Engl. J. Med. 285(4):193–96 McDevitt HO, Bodmer WF. 1972. Histocompatibility antigens, immune responsiveness and susceptibility to disease. Am. J. Med. 52(1):1–8 Nepom GT, Erlich H. 1991. MHC class II molecules and autoimmunity. Annu. Rev. Immunol. 9:493–525 Acha-Orbea H, McDevitt HO. 1987. The first external domain of the non-obese diabetic mouse classII I-Ab chain is unique. Proc. Natl. Acad. Sci. USA 84(8):2435–39 Todd JA, Bell JI, McDevitt HO. 1987. HLA-DQb gene contributes to susceptibility and resistance to insulin-dependent diabetes mellitus. Nature 329(6140): 599–604 Chao CC, Sytwu HK, Chen EL, Toma J, McDevitt HO. 1999. The role of MHC class II molecules in susceptibility to type I diabetes—identification of peptide epitopes, and characterization of the T cell repertoire. Proc. Natl. Acad. Sci. USA 96:9299–9304


Annu. Rev. Immunol. 2000. 18:19–51 Copyright q by Annual Reviews. All rights reserved

RECEPTOR SELECTION IN B AND T LYMPHOCYTES David Nemazee Department of Immunology, Scripps Research Institute, La Jolla, California 92037; e-mail: [email protected]


Key Words immunoglobulin genes, T cell receptor, immune tolerance, receptor editing, gene rearrangement Abstract The process of clonal selection is a central feature of the immune system, but immune specificity is also regulated by receptor selection, in which the fate of a lymphocyte’s antigen receptor is uncoupled from that of the cell itself. Whereas clonal selection controls cell death or survival in response to antigen receptor signaling, receptor selection regulates the process of V(D)J recombination, which can alter or fix antigen receptor specificity. Receptor selection is carried out in both T and B cells and can occur at different stages of lymphocyte differentiation, in which it plays a key role in allelic exclusion, positive selection, receptor editing, and the diversification of the antigen receptor repertoire. Thus, the immune system takes advantage of its control of V(D)J recombination to modify antigen receptors in such a way that self/non-self discrimination is enhanced. New information about receptor editing in T cells and B-1 B cells is also discussed.

INTRODUCTION Lymphocytes often express a single antigen receptor despite their genetic potential to simultaneously express many. This fact has focused interest on how expression of multiple receptors is limited. It has recently come to light that cells bearing antigen receptors may actively attempt to express new ones, which can lead to expression of multiple receptors, inactivation of previously expressed receptors, or replacement of old receptors with new ones—a process called receptor editing. Receptor editing occurs through ongoing or renewed antigen receptor gene rearrangement. In this overview, we discuss features of both T cell receptor (TCR) and immunoglobulin (Ig) loci that promote secondary receptor gene rearrangements, and we review the evidence for receptor editing in B and T cells. Receptor editing has recently been the subject of several brief reviews (1–4).

0732–0582/00/0410–0019$14.00

19

20

NEMAZEE


SECONDARY REARRANGEMENTS V(D)J recombinase rearranges and ligates together dispersed variable (V), diversity (D), and joining (J) minigene elements and brings the newly assembled V(D)J exon in proximity to the constant (C) exons of these genes (see accompanying article by Schatz, ‘‘Evolution and Mechanisms of VDJ Recombination,’’ pp. 495– 97). An important feature of some antigen receptor genes is the ability to undergo so-called secondary, or nested, rearrangements, which can salvage the functionality of loci with primary out-of-frame rearrangements and favor receptor editing, i.e., the modification of already functional genes. This process is affected by several factors including the number of gene segments in the locus, their organization along the DNA, and the organization and types of recombination signal sequence (RSSs) adjacent to the gene segments in question. Some typical organizations of receptor genes and their influence on secondary rearrangements are shown schematically in Figures 1 and 2. As we discuss below with respect to the well-studied mouse and human systems, some antigen receptor gene loci appear to be particularly specialized to carry out secondary rearrangements, whereas other loci may disfavor them. The organizations of mouse and human antigen receptor gene loci are shown schematically in Figure 3.

Loci That Favor Secondary Rearrangements Ig-j Rearrangements in the mouse involve initial joining of one of ;140 Vj elements to one of four functional Jj elements (5) (Figure 3A) (6–12); a similar organization is seen in humans, who have about 66 Vj elements and 5 Jjs (13) (Figure 3B) (14). The j locus lacks D gene segments; consequently, upon primary VjJj joining on one allele, secondary rearrangements between remaining upstream Vjs and downstream Jjs can occur in a single step (shown schematically in Figure 1B). In the mouse, Jj1 and Jj2 rearrangements are preferred (15), which retains downstream Js available for secondary rearrangement (16). Furthermore, because many Vj genes are placed in a transcriptional orientation opposite to the Jj elements, j loci often rearrange by inversion, retaining thereby the entire repertoire of Vjs for subsequent rearrangements (shown schematically in Figure 2B, bottom). When the V gene segments are in the same transcriptional orientation as the segments to which they rearrange, deletional rearrangements excise intervening DNA, which is permanently lost from the chromosome (Figure 2B, top). Analysis of the Ig-j loci in mouse or human B cell lines has shown that a single allele can undergo two or more successive V-J recombinations (16–23). Isolation of the excised circular DNA from sequences intervening in the V/J recombination revealed that they often contained independent V/J joins, indicating that secondary j rearrangements are common in the mouse splenic B cells analyzed (24, 25). B cells of mice lacking the j-locus on one chromosome have increased usage of downstream Jjs on the remaining allele, consistent with dis-


RECEPTOR EDITING

21

Figure 1 Role of gene organization in facilitating or inhibiting receptor editing. V, D, and J coding elements are flanked by recombination signal sequences carrying 12-bp spacers (open triangles) or 23-bp spacers (black triangles) that constrain the range of possible rearrangements in cis. Recombinase joins elements with 12-bp spacers to those with 23 bp. Because of their overall organization, loci vary in their abilities to support receptor editing type rearrangements. (A) Cartoon of one type of gene organization similar to mouse and human Ig-H loci (see Figure 3 for details). The presence of D elements along with V genes in the same transcriptional orientation as the J/C cluster forces deletional rearrangements. Primary VDJ assembly cannot be replaced by recombination using conventional signal sequences. (B) In contrast, in loci without D elements, sequential rearrangements are often possible. In this example, a primary (10) V4-to-J join is replaced by a subsequent secondary (20) rearrangement between V2 and the downstream J. Such secondary rearrangement permits the replacement of potentially functional V4-to-J joins, i.e., receptor editing. This type of organization is also seen in mammalian TCRa loci.

placement of out-of-frame VJ joins by nested rearrangement (23). Thus, j locus structure permits replacement reactions, allowing receptor editing. TCRa At the TCRa locus, any one of about 100–200 Va segments can be joined to one of 50 Ja elements (Figure 3A, B) (26–29). A primary VJ rearrangement may yield an out-of-frame join that can be replaced by subsequent rearrangement of an upstream Va segment with one of the remaining downstream Ja loci (30). T cells may undergo many nested Va-Ja recombinations (31, 32). In addition, thymocytes rearranging TCRa loci retain excised DNA circles containing the TCRd locus, usually in a rearranged form (30, 33). Since these episomal DNAs are lost with cell proliferation (30, 34, 35), TCRa rearrangements


22

NEMAZEE

Figure 2 Gene organizations that inhibit or facilitate receptor editing. (A) Cluster type receptor gene organization is used in many lower vertebrates and is retained in certain mammalian receptor gene loci, such as mouse Ig-k. Rearrangements occur within clusters but not between adjoining clusters, preventing editing and potentially posing problems for isotype exclusion. (B) Inversional rearrangements are dictated by gene orientation. Variable gene segments in inverted transcriptional orientations relative to J/C clusters are indicated by upside-down Vs. Such elements join though inversion rather than excision of intervening DNA. Hypothetical V3 and V4 elements must undergo deletion during primary rearrangement to Js, whereas V1 and V2 elements rearrange by inversion. Note that subsequent secondary rearrangement can again occur through either deletion or inversion, but inversional rearrangements retain more V genes and change the orientations of V elements intervening the break points.

functionally exclude TCRd expression. Furthermore, TCRa rearrangements favor initial use of Ja’s at the 58 end of the locus (36), possibly suggesting progressive use over time of 58-to-38 Ja’s. The TCRa locus organization constrains rearrangement to occur by deletion rather than inversion (27) (shown schematically in Figure 1B) (28, 29).

Loci That Disfavor Secondary Rearrangements Ig-H In contrast to the Ig-j and TCRa loci, some loci are so constructed that they tend to suppress secondary rearrangements, at least through conventional recombination signal sequences. VH genes are arranged in the same transcriptional orientation as the D and J elements (37, 38), forcing VDJ assembly at the Ig-H locus to occur by deletion of the intervening DNA (38) (Figure 3). As a consequence, correction of nonfunctional VDJ rearrangements by secondary rearrangement is prevented because no more D regions are available for recombination to upstream VHs. While unrecombined upstream VHs and downstream JHs are typically retained, they both have recombination signal sequences carrying


Figure 3 Antigen receptor loci of (A) mouse and (B) human. Note that Ig-j, TCR-a, TCR-d, and TCR-b have structures that are compatible with secondary, replacement rearrangements in both mouse and human. Conventional V(D)J recombination disfavors receptor editing at the Ig-H locus of mouse or human because of the 12/23 rule and the arrangement of VH elements in the same transcriptional orientation as the JH/CH cluster. In the mouse Ig-k and TCR-c loci, functionally rearranged genes cannot efficiently be altered by secondary rearrangements because of their cluster type organization, whereas editing is possible in the human versions of these loci provided that the 38 most J’s are not initially used. In the TCR-d locus of both mouse and human, TCR-a rearrangements exclude TCR-d expression.

24

NEMAZEE


23bp spacers, and therefore they cannot be joined together by recombinase (see Figure 1A). Despite this, secondary rearrangements can sometimes occur on the Ig-H locus. Early in mouse B cell development, secondary D-to-J joining can replace primary DH-to-JH rearrangements by deletion or inversion (39), but this apparently does not occur in humans (40). Furthermore, VH elements can sometimes undergo secondary recombination to preformed VDJ elements through VH replacement by targeting noncanonical, but conserved, RSS-like sequences that are present in the 38 end of the rearranged VH coding sequence (41). VH replacement is discussed in a later section. Ig-L k Loci The lambda locus in the laboratory mouse is composed of two linked miniclusters spanning about 200 kb (42–47). Vk2, Vkx/Jk2 Ck2 lies upstream of Vk1/Jk3-Ck3, Jk1–Ck1, but most rearrangements occur between V and J regions of the same cluster (Figure 3A). All rearrangements involve DNA deletions. The recombined Vk1-Jk1-Ck1 gene is the most common type of k rearrangement (48). Thus, combinatorial diversity is extremely limited in the mouse k locus, and receptor editing is precluded (shown schematically in Figure 2A). The human k-locus has an array of ;70 Vk genes upstream of seven JkCk clusters; each J-C cluster has a single J, and only three of the clusters are functional (Figure 3B) (49). Unlike in the mouse k-locus, this organization allows nested recombination and receptor editing (50).

Suppressive Rearrangements RS/kde An additional mechanism facilitated by receptor gene organization is the ability to preclude function of certain loci by nested rearrangements that can themselves have no functional product. One important example of a rearranging, but noncoding, gene element that illustrates this mechanism is the RS/kde element (‘‘recombining sequence’’ in mouse, ‘‘kappa deleting element’’ in human) (51). This element inactivates the Ig-j locus by deletional recombination (52). RS is found ;25 kb downstream of the mouse Cj exon (53) (Figure 3A). (RS should not be confused with the RSS, recombination signal sequence, carried by all rearranging receptor genes). The kde is in a similar position in the human j-locus (54) (Figure 3B). Both RS and kde contain an RSS sequence with a 23-bp spacer that can recombine through a V(D)J recombinase–dependent process either to unrearranged Vjs or to noncanonical RSS sites in the Jj-Cj intron (25, 51, 55). RS/kde rearrangements occur in ;75% of mouse B cells that express k L-chain (56) and in an even higher fraction of human k-producing cells (13, 57). RS rearrangements also occur on the second allele of approximately 12% of mouse j-expressing cells (58). Because the RS/kde encodes no protein, these joins are nonfunctional (55, 59). It appears that the RS/kde elements have no other purpose than to inactivate j-genes, many of which were previously functional (57). Since RS rearrangement appears concurrently with Ig-k locus recombination, they may

RECEPTOR EDITING

25

be predicted to clear the way for k expression (59, 60). Thus, j locus structure favors replacement reactions that allow receptor editing.


wJa /drec A similar process of deletional recombination can occur in the TCRd locus, where the wJa element joins to drec, eliminating the entire TCRd locus that is embedded in the TCRa locus (Figure 3A,B) (61, 62). This type of nonprotein coding recombination is reported to occur in ;70% of human T cells that go on to express TCRa (63). Like the wJa element, RS and kde have homology to J regions (64), but unlike most antigen receptor pseudogenes, they retain consensus RSSs (51, 52, 55), implying evolutionary conservation of rearrangement function.

QUASI-ORDERED REARRANGEMENTS IN DEVELOPMENT B Cells Rearrangements generating antigen receptor genes generally occur in a particular temporal sequence regulated by control of gene accessibility. In mouse B cell development, H-chain genes are assembled first. PreB cells bearing in-frame, functional IgM H-chains (l-chains) are selected for clonal expansion and differentiation, whereas cells lacking l-chains are eventually targeted for elimination (65). Cells lacking l-chains often attempt rearrangements on the second allele, which, if functional, can rescue the cells. This selection process is a major developmental checkpoint that involves the testing of l-chain based on its ability to associate with surrogate L-chain components k5 (66) and VpreB (67) and to mediate transmembrane signals through associated Ig-a/b signal transducers (68– 71). Only after a proliferative burst of approximately six cell divisions (72) do cells exit cell cycle; then recombination activity is redirected to the Ig L-chain loci generating sIgM` B cells (73). Gene Assembly and Lineage Commitment in T Cells In abT cell development, TCR loci are similarly rearranged in a temporal sequence, beginning with DJ rearrangements at the TCRb loci, followed by Vb-to-DJ rearrangements, a proliferative burst of those cells expressing functional b-chains, and subsequent redirection of recombinase to the TCRa loci (74–77). In a further parallel to B cell development, testing of the b-chain in T cell development involves protein association with a surrogate chain, pTa (78), and signaling through a preTCR complex (79). Thymocytes lacking b-chains, pTa, or molecules involved in preTCR signaling fail to develop further (80–82). Enforced expression of b-chain transgenes blocks endogenous b-chain gene assembly (83) and promotes development to TCRa gene rearrangement and expression (84).


26

NEMAZEE

Expression of a complete a/bTCR is clearly insufficient for RAG downregulation (31, 32, 85, 86). a/b T cells are subject to the process of positive selection (87), in which cells with a receptor biased to weak MHC self-reactivity are selected to advance in development. This maturation is coincident with the cessation of TCRa rearrangement (88, 89). RAG downregulation can be experimentally induced in thymocytes by antibody cross-linking of the TCR (90). The abundance of V genes and J genes in the TCRa locus allows cells to undergo multiple rearrangement attempts, which may be necessary for efficient positive selection (31, 32, 86). T cell development in the thymus is further complicated by two factors. First, common precursors can give rise to either cd or ab T cells, a situation that is regulated in part by the functionality of rearrangements. And, although the antigen receptors of these cells use different protein chains, the TCRd locus is embedded in the TCRa locus (Figure 3) and is physically excised in the process of TCRa rearrangements (27, 29, 91). The d and c loci rearrange prior to a and b loci (92, 93), and an intact cd receptor likely promotes development in much the same way as b/pTa heterodimers. Lineage determination is also under control of the Notch pathway (94). Likewise, it has been proposed that cd- or b/pTa-mediated signals may affect the Notch pathway itself (95). Regardless of the details of ab vs. cd lineage commitment, it is of interest to observe patterns common to the organization and rearrangement of different antigen receptor genes. First, in all lymphocytes, one chain partner gene is preferentially rearranged and expressed before the other. Two waves of RAG gene expression correspond to the points in development in which each locus is recombined (96, 97). After expression of the first chain, cells are selected for proliferative expansion (78, 96). The initially expressed Ig-H, TCR-b, and TCR-d loci invariably include D elements, whereas loci rearranging later lack D regions (Figure 3). Absence of D regions facilitates receptor editing (Figure 1) but limits potential receptor gene diversity, suggesting that receptor editing may be an important selective force in evolution. The initially rearranging chains are each encoded in a single locus, as is the TCR-a locus, which should theoretically limit dual expression. However, in humans and mice the Ig-L chain is encoded by two or three loci, respectively, each with two alleles that rearrange independently of one another, compounding the problem of allelic exclusion with one of isotypic exclusion.

ALLELIC EXCLUSION The vast majority of mammalian B cells bear a single antibody heavy and light chain (98–103), despite the ability to produce two, one encoded by each allele. In the mouse, most cells express a single light chain (104, 105), despite the ability to express six (2 alleles 2 3 isotype loci: j, k2/kx, k1/k3). As discussed below, allelic and isotypic exclusion are ensured by a variety of active processes. The

RECEPTOR EDITING

27

evolutionary selective forces that make allelic exclusion desirable are not particularly clear. It has been argued that B cell monospecificity is important for both immune self-tolerance and efficient antibody effector function. In T lymphocytes, several mechanisms ensure that cd receptors are excluded from abT cells and vice versa, despite the ability of both types of receptor to rearrange in the same cells (95). Furthermore, in abT cells the lack of allelic exclusion (allelic inclusion) of TCRb chains is rare. On the other hand, in these same cells coexpression of two TCRa chains is common (86, 106, 107). Allelic inclusion is also observed in TCRd (108) and TCRc (109). Why exclusion should be strict in some situations but not in others is unclear. The presumed risk of autoimmunity that may be caused by relaxed allelic exclusion remains to be proved.


Mechanisms of Allelic Exclusion Stochastic Factors Several mechanisms in B cells have been proposed to explain the phenomena of allelic and isotypic exclusion (e.g., expression of only Ig-j or Ig-k, but not both). One major mechanism is inherent to the recombination machinery itself. Because recombination introduces small deletions and insertions at the coding join, about two thirds of rearrangements are out of frame. In some cases, the V, D, or J elements themselves may harbor stop codons, or such codons may be created in the process of recombination. In theory, these stochastic mechanisms alone reduce allelic inclusion of a single locus to less than 20%. This frequency drops further if a time limit is imposed on the rearrangement process, and further still if rearrangement is intrinsically preferred in one allele over another. Some evidence suggests both of these factors play a role in L-chain allelic and isotypic exclusion. In the mouse j-light chain locus, DNA demethylation and concomitant accessibility to recombination is, at least initially, monoallelic (110), though this is clearly not a strict rule since about half of all B cells have rearrangements on both j alleles (104). Furthermore, most j-expressing cells lack k rearrangements, whereas virtually all k-producing cells bear j rearrangements, usually on both alleles (13, 56, 111–113). During B cell development in the bone marrow, k-bearing B cells emerge 24 h later than j` B cells (114), and analysis of k-producing hybridomas revealed that most k` cells had rearranged a single k-gene, suggesting that k gene recombination is both inefficient and time limited (57). While these stochastic factors reduce allelic and isotypic inclusion, they are complemented by more active regulatory processes. Selective Factors A second mechanism that could enforce allelic exclusion in mature cells is counter selection against double producing cells. As originally conceived, excess H-chain expression was thought to be toxic to cells (115). Evidence suggesting this possibility came from analysis of myeloma cells, in which L-chain loss is toxic if H-chains are produced. While this toxicity explanation is not supported by experiments in which double H-chain expressing B cells were generated by transgenesis in vivo (116), it is nevertheless possible that

28

NEMAZEE

double producers are counterselected for reasons other than toxicity. If random H/L pairs are frequently self-reactive, as has been predicted on a priori grounds (117), cells bearing two different receptors would be much more likely to be autoreactive than cells with a single H/L pair. Consequently, counterselection of double producers may occur as a consequence of tolerance induction. This might explain the curious finding that in k5-deficient mice, which manifest poor allelic exclusion of H-chains in the preB compartment (118), allelic exclusion is restored in peripheral B cells (66).


Feedback Suppression of Recombination Ig-H While selective forces obviously play a major role in lymphocyte biology in general, a considerable body of evidence argues that instructive mechanisms have a dominant role in establishing allelic exclusion. In preB cells, functional l-chains actively block further H-chain rearrangements (66, 119–121). This inhibition occurs through assembly of the surrogate L-chain components k5 (120, 122) and VpreB (67) with the membrane-bound form of l-chain (121, 123–125), followed by signaling via the associated Ig-a/b complex (125–127). Forced expression of a heavy chain by transgenesis substantially blocks VDJ assembly on the H-chain locus (119, 123, 128, 129). Disruption of the preB cell receptor (preBCR) signaling complex inhibits both B cell development and allelic exclusion at this stage (103, 118). H-chains that fail to associate with surrogate Lchain, such as certain chains that include VH81X sequences, are essentially absent from the peripheral B cell pool (72, 130, 131). In normal B cells, many cells have incomplete, DJ rearrangements on the nonfunctional H-chain gene allele, suggesting that the rate of V-to-DJ recombination is slow compared to the ability of the cell to perceive functional H-chain assembly in a preBCR complex (132, 133). In addition, in the mouse, otherwise nonfunctional Dl chains generated from partial D-J rearrangements that involve D reading frame 2 can terminate recombination and block further development (65, 134, 135). Overall, there is compelling evidence for active feedback regulation of H-chain gene recombination that contributes to allelic exclusion. TCRb An analogous feedback process in thymocytes regulates allelic exclusion of the TCRb locus. Pairing of TCRb chain with the invariant preTCRa chain generates a signaling complex with CD3 that signals cessation of V(D)J recombination and developmental progression (78, 79, 136, 137). Enforced expression of a b-chain transgene (Tg) blocks endogenous rearrangements and promotes developmental progression (83, 138).

Ig-L-Chain Allelic Exclusion That Ig-L chain loci must similarly be subject to feedback suppression was predicted on a priori grounds, but the evidence for feedback suppression in light chain gene rearrangements is much less compelling or clear than the evidence for


RECEPTOR EDITING

29

H-chain or TCR-b allelic exclusion. B cells from normal animals manifest substantial allelic exclusion (102), but sensitive techniques are able to identify double producing cells above background (139–141). In contrast to the ability of H-chain transgenic constructs to suppress H-chain rearrangements, enforced expression of L-chain transgenes led to mixed results. Initial experiments by Storb and colleagues using the MOPC 21 j-chain Tg were consistent with a strict feedback regulation model, as hybridomas generated from these mice appeared to lack endogenous j expression when H-chain was also expressed (142). Curiously, in this study many hybridomas lacked H-chain expression but expressed endogenous j-chains (142). It was later found that k-chain-producing hybridomas from these mice not only coexpressed the j Tg but also often rearranged and expressed endogenous j-loci (139). Some other transgenic mouse lines apparently failed to exclude endogenous rearrangements because of insufficient protein expression (119), but this quantitative effect could not explain how in the MOPC 21 mice endogenous rearrangements were excluded in some B cells but not in others. Similarly, in MOPC-167 j-gene transgenic mice, endogenous j expression was suppressed in only a subset of B cells (123). It is unlikely, but not formally excluded, that the incomplete allelic exclusion observed in these j-transgenic mouse experiments was solely the result of defective Tg expression. Certain conventional j transgenic mice manifest excellent allelic exclusion in a substantial proportion of cells (143, 144). Several lines of evidence suggest that in some cells even normal L-chain gene expression is not sufficient to suppress further L-chain gene rearrangement. Several sIg ` B cell tumor lines and long-term IL-7-dependent bone marrow B cell lines express new receptor chains through secondary rearrangements (22, 145– 149). In an analysis of episomal DNA present in mouse spleen cells, Harada & Yamagishi found that of 16 clones containing VjJj joins excised by nested j rearrangements, 5 were in-frame and theoretically functional (24), suggesting that in these B cells the natural genes were frequently incapable of suppressing further rearrangements. Perhaps more convincingly, in mice in which functional VjJjgenes were targeted to the natural locus, endogenous j rearrangements were inhibited in some, but not all, B cells (144, 150). Again, only a subset of cells, which varied in frequency depending upon the variable region gene used, appeared subject to feedback suppression. A criticism of these experiments is that even the targeted j-genes may be aberrantly expressed owing to the premature juxtaposition of a Vj promoter near the Cj locus. This could conceivably target recombination to the j locus abnormally early in development, perhaps in cells that lack H-chains and that therefore are unable to perceive the presence of light chain by BCR signaling. Conversely, if DNA accessibility in the j-locus is stochastically controlled (110), the targeted gene may at first be transcriptionally silent in some cells that make the other allele accessible. The targeted allele might then be coexpressed at a later time, resulting in allelic inclusion of a significant fraction of cells, an unexpected but well-documented, property of targeted Lchain mice (144, 150, 151).

30

NEMAZEE


In contrast to these results, it appears that when expressed as transgenes, certain H/L pairs are extremely efficient at suppressing endogenous L- and H-chain protein expression and RAG expression in the bone marrow (152), suggesting that BCR specificity plays a critical role in development. One interpretation of these results is that, like T cells, B cells are positively selected by self-antigens that downregulate RAG expression. Some studies with cell lines are consistent with this idea (145, 146). However, if antigen-specific positive selection was associated with cessation of V(D)J recombination in B cells, then allelic inclusion would be predicted to occur frequently among normal B cells in vivo (as is the case for TCRa exclusion in T cells). On the contrary, as we discuss below, there is substantial evidence from the study of autoantibody transgenic mice that negative selection, i.e. tolerance mediated by encounter with self-antigens, blocks developmental progression and prevents recombinase downregulation. These results can be reconciled with the above-mentioned results by taking into account the influence of immune tolerance on feedback suppression of recombination.

RECEPTOR EDITING Receptor Editing Monitored In Vivo in Transgenic Models of Immune Tolerance The fact that H`L antibody transgenes could be used to generate mice in which most B cells had a defined specificity (119) stimulated studies analyzing immune tolerance in B and T cells (153). Transgenic mouse models using autoantibodies to HEL, MHC class I alloantigens, DNA, erythrocytes, and other antigens have been useful in defining a number of ways that self-reactivity is controlled (154). In transgenic autoantibody models in which developing B cells were confronted with antigen in a multivalent form, autoreactive cells were absent from the peripheral lymphoid system, but a population of cells carrying a low level of receptor was present in the bone marrow (155–157). Two lines of evidence suggested that receptor editing, i.e. autoantigen-induced secondary Ig gene rearrangements, might be a mechanism of immune tolerance in these models. In mice transgenic for the 3–83 antibody (anti-H-2Kk,b), B cells were numerous and they almost exclusively expressed the Tg-encoded antibody as a consequence of feedback suppression of recombination (158). But when cognate antigen was introduced by breeding the 3–83 transgenes to the appropriate MHC background, spleen and lymph node B cells were reduced in number, and those cells that remained lacked self-reactivity. B cells in the peripheral lymphoid organs retained surface expression of the 3–83 H-chain, but not the 3–83 L-chain, and an extremely high percentage of cells expressed endogenous k-chain. In the bone marrows of antigen-expressing mice, transgenic sIg ` B cells expressed high levels of recombinase mRNA and manifested rearrangements at the endogenous L-chain loci. Based on these findings, it was suggested that autoantigen binding


RECEPTOR EDITING

31

by immature bone marrow B cells could reinduce or prolong L-chain gene rearrangements, allowing a cell to alter specificity and escape death (158). It was further postulated that the ability to undergo receptor editing was limited to early stages of B cell development because 3–83 mice that expressed cognate antigen only in the periphery, under control of a liver-specific Kb Tg, underwent profound B cell deletion without appreciable receptor editing. Independent evidence for receptor editing was obtained in transgenic mice carrying H ` L genes of dsDNA antibody 3H9 (159). Spleen cells of H ` L Tg mice and hybridomas generated from them were analyzed for antibody specificity and endogenous rearrangements. While dsDNA-specific B cells were lacking from the hybridoma sample, indicating that specific tolerance was induced, splenic B cell numbers were surprisingly normal, particularly in older mice. Splenic B cells retained transgenic H-chain expression on the cell surface but lacked 3H9 L-chain expression as detected with specific anti-idiotypic antibody. The hybridomas expressed endogenous L-chains that apparently altered B cell specificity and extinguished dsDNA binding. Suppression of dsDNA specificity was reversible in one hybridoma, when the endogenous j was lost, suggesting that binding of the transgenic H-chain with the endogenous L-chain outcompeted transgenic L-chain. In addition to this evidence for ‘‘phenotypic’’ allelic exclusion, among these hybrids an excessive use of endogenous Vj12/13 genes associated with downstream Jjs was observed. In control mice transgenic for just the 3H9 L-chain, only 6/25 hybrids excluded endogenous L-chain expression, but the endogenous L-chains that were expressed had the typical distribution of Vj and J usage. Radic et al (160) and Prak et al (23) extended this work to test for toleranceinduced receptor editing in 3H9 H-chain-only Tg mice. This test was possible because 3H9 H-chain binds DNA in association with diverse L-chain partners (161), but spleen hybridomas from 3H9 H-chain Tg mice lack reactivity to dsDNA (160, 162). Hybridoma analysis again revealed a strikingly limited usage of Vj genes in splenic B cells with excess use of Vj12/13 and skewing to downstream Jj’s, particularly Jj5. These studies also documented that in many cells multiple rearrangement attempts were made on a single chromosome. Assuming that additional attempts were required to replace j-chains that conferred autoreactivity, these results provide an explanation for skewing of Jj usage. This would predict that many VJ joins displaced by secondary rearrangements were in-frame and functional, an assumption that was not directly tested. An independent study by Eilat and colleagues of a different set of VH11 anti-DNA H-chain mice yielded similar results (163). Since in these studies no L-chain Tg was present, the later stages of B cell development should have been relatively normal. Furthermore, because simple apoptosis of autoreactive cells from an initially random population would be unlikely to account for the observed skewing in Jj usage, these results suggested that autoreactive cells were rescued by ongoing Igj recombination.

32

NEMAZEE


Analysis of Receptor Editing in Gene Targeted Antibody Gene Mice VH Replacement Gene targeted autoantibody Tg mice were generated to more closely mimic the natural genomic context and to test for VH replacement in vivo. In theory, gene targeting to the natural locus should facilitate editing because secondary rearrangements on the targeted allele, should they occur, can silence or modify antibody expression. In conventional Tg mice, the suppression of Tg expression is highly inefficient because transgenes are typically inserted on chromosomes without access to potential rearrangement partners. A complicating feature of mice with targeted H-chain loci is that they retain upstream D regions, which are normally absent after VDJ assembly. These dangling Ds can rearrange to remaining J elements of the constructs or to the heptamers embedded in the targeted VH region. Early studies (164) failed to identify use of the conserved heptamer embedded in the expressed VH, but destructive rearrangements caused by D joining to other heptamer-like sites in the gene were observed. Another study (165) found frequent editing by D-regions, but also found VH-to-VDJ replacement events that rescued H-chain function, a result predicted from prior work with cell lines (41, 166). This and subsequent results from studies of other H-chain targeted mice (116, 163, 167, 168) have verified that such replacement events at the Ig-H locus occur at detectable frequency in these animals, indicating that nested rearrangements alter specificity in vivo. But there is debate about the frequency, stage specificity, and tolerance inducibility of the VH replacement reaction (169). In vitro analyses indicate that the embedded heptamer is extremely inefficient at targeting V(D)J recombination relative to conventional RSSs (170). When VH-to-VDJ replacements have been observed in mouse B cells, they almost always include additional ‘‘N’’ nucleotides derived from TdT activity, indicating that most VH replacements occur at the proB stage, prior to L-chain gene rearrangements (116, 165, 167). This in turn suggests that tolerance signaling through intact surface immunoglobulin (Ig) does not drive the VH-to-VDJ replacement response. Editing at the j-Locus The possibility that editing on the H- or L-chain loci could be tolerance induced was more directly tested in models of H`L autoantibody gene targeted mice. Chen and colleagues analyzed mice coexpressing heavy and light chain targeted constructs encoding the 3H9H/Vj4 anti-dsDNA antibody or the 3H9 H-chain with a different targeted L-chain (Vj8) that conferred reactivity to ssDNA (152). The 3H9/Vj4 dsDNA reactive combination was subject to extensive j-chain editing in the absence of editing on the H-chain locus. Ninety-eight percent of cells retained the expression of the 3H9 H-chain, and the vast majority of these failed to express the other H-chain allele, but at the L-chain locus extensive rearrangements occurred on both the targeted and untargeted jalleles. Only 4% of hybridomas lacked detectable secondary j-rearrangements. L-chain editing occurred by inversional rearrangements in 57% of the hybrid-


RECEPTOR EDITING

33

omas, deletional type rearrangement events in 28% of the cells, and inclusional rearrangements (i.e. on the opposite allele with an unaffected targeted allele) in only 11% of cells. In the 3H9H/Vj8 mice, B cells bearing this receptor are often seen in the periphery, in an anergic state (162). But suggestive evidence of receptor editing was also observed in the 3H9H/Vj8 B cells because in both conventional and 3H9H/Vj8 gene-targeted mice a subset of B cell hybridomas showed exaggerated usage of Vj12/13 along with skewed Jj usage on the untargeted allele (152). (In the targeted 3H9H/Vj8 mice the Vj8 gene was joined to Jj5, providing no additional functional rearrangement possibilities on that allele.) These results suggested that tolerance-induced receptor editing was focused on the L-chain loci and could be induced by both strong and weak toleragens, dsDNA and ssDNA, respectively. In a second study, targeted j- and H-chain gene mice were generated encoding the 3–83 antibody, specific for self-MHC molecules H-2Kk and H-2Kb, and light chain editing in B cells was monitored (151). Quantitative southern blotting assessed the extent of recombination in the j-loci and indicated that in the autoreactive combination at least 30% of targeted alleles and 59% of wild-type alleles were rearranged, while in B cells of j-only mice, these percentages were 10% and 27%, respectively. As a control, the 3–83 j mice were bred with H-chain gene mice of an irrelevant, nonautoreactive specificity. (Unfortunately, in this study, it was not possible to compare 3–83 H`L targeted mice with and without antigen because the targeted H-chain gene was linked to a cross-reactive antigen locus contributed by the embryonic stem cell strain–129.) Most control cells appeared to retain expression of the targeted L-chain, whereas at least 85% of the autoreactive B cells lost expression of the autoreactive specificity on the cell surface. RS-type recombinations inactivating the j-loci were 60-fold more prevalent in the autoreactive context than in the presence of an innocuous receptor. In the bone marrows of mice with an innocuous H`L receptor, small preB cells were absent owing to accelerated developmental progression. However, in mice with autoreactive receptors that manifested receptor editing at the DNA level, this compartment was very large, suggesting that receptor editing specifically caused a retrograde step in development from sIg` to sIg1 stages. This is as one might predict from the high frequency of nonfunctional secondary rearrangements associated with inactivation of the targeted j-gene. In both this study and that of Chen et al (152), it appeared that receptor editing could allow a rather efficient rescue of previously autoreactive cells because peripheral B cell numbers were relatively normal.

Central B Cell Tolerance Is Associated with Developmental Block It has been known for a long time that B cell production can be suppressed by anti-IgM treatment (171). Early transgenic mouse studies on B cell tolerance to membrane MHC molecules indicated that self-reactive B cells were absent from


34

NEMAZEE

the peripheral lymphoid organs, but large numbers of newly formed autoreactive cells bearing a low density of sIg were retained in the bone marrow (155, 156). Because these data were superficially consistent with the clonal selection hypothesis and classical studies on B cell tolerance (172), they were interpreted to mean that autoreactive cells were eliminated at the preB-to-B cell transition. It now appears most likely that autoreactivity initially prevents B cell developmental progression and promotes receptor editing that allows some cells to change their specificities. While cells failing to alter their receptors eventually die, this tempo of cell death is essentially identical to the rate of turnover of preB cells that fail to generate any receptor at all. Thus, a cell death program succeeds a receptor editing phase. An early indication that this might occur was provided in studies performed in the 1970s, in which bone marrow B cells were challenged with anti-IgM antibodies (173, 174). In these cultures, immature B cells rapidly lost sIg but were not reduced in number, as assessed by quantitation with anti-MHC class II antibodies; after removal of the anti-Ig stimulus, only a small subset reexpressed sIg. In contrast, mature splenic B cells rapidly reexpressed sIg after similar treatment. This ‘‘irreversible receptor modulation,’’ as it was called at the time, was probably the first evidence for editing, which is predicted to frequently lead to nonfunctional secondary rearrangements. More recently, in experiments in which highly purified bone marrow B cells, rather than unpurified bone marrow preparations, were challenged with anti-IgM antibodies, this stimulus was found to result in apoptosis (172, 175, 176). A possible resolution of the discrepancy between the two types of study is that the presence of bone marrow accessory cells may be required for the editing response (177). Experiments studying antigen-induced apoptosis in immature B cells of autoantibody transgenic mice are consistent with a model in which cell death is slow and delayed. Culture of anti-HEL Tg bone marrow B cells with membrane-bound HEL blocked developmental progression at the fraction E (immature B cell) stage but failed to cause rapid cell death (178). The block was reversible upon antigen removal during the first 1–2 days, but at later times B cell death occurred. Enforced expression of Bcl2 in immature autoreactive B cells was shown to further prolong survival at this developmental stage (178) but was unable to relieve the developmental block. The receptor editing model provided a rationale for this delayed and reversible death program in immature B cells. In short-term cultures of 3–83 (anti-MHC) B cells, anti-BCR antibodies failed to accelerate B cell death during the initial 48 h of culture but stimulated secondary L-chain gene rearrangements in a considerable proportion of cells, estimated to be from 25% to 50% of all cells (179). This same study reported that anti-BCR treatment could also induce receptor editing in normal, nontransgenic bone marrow B cells. In another model system, immature bone marrow B cells of 3–83 Tg mice were expanded in IL-7 and challenged with antigen. Under these conditions, autoantigen blocked developmental progression and promoted receptor editing but did not appreciably accelerate cell death (180, 181). Consistent with this idea, Bcl-2 overexpression did not allow autoreactive receptor-bearing B cells to escape from the bone marrow


RECEPTOR EDITING

35

but appeared to improve the efficiency of their escape by receptor editing, probably by prolonging the time window of secondary rearrangement (182). In an independent study, Bcl-2 overexpressing immature B cells spontaneously switched from j to k L-chain expression in vitro (146). Somewhat analogous results were obtained in Bcl-xl-overexpressing anti-HEL/mHEL double Tg mice, in which central tolerance was apparently disrupted, yielding enhanced peripheralization of autoreactive B cells with an anergic phenotype along with enhanced receptor editing (183). Curiously, however, the same Bcl-xl Tg had minimal effects in the 3–83 model of central tolerance (J Lang, D Nemazee, unpublished data). Overall, these data suggest that receptor editing is associated with a developmental checkpoint. The notion that through secondary recombination autoreactive cells can overcome a developmental block implies that inhibition of secondary recombination in autoantibody Tg mice should prevent B cells from maturing to populate the periphery. This idea was directly tested in both the anti-H-2Kk,b and anti-DNA models by breeding Ig-H`L transgenes to a RAG-1- or RAG-2-deficient background (184, 185). B cell development did not advance past the late bone marrow stages, and B cells did not populate the peripheral lymphoid organs. The developmentally arrested RAG-deficient autoreactive cells underwent apoptosis in situ (185). In contrast, on an antigen-free (H-2d), RAG-deficient background anti-H2Kk,b B cells did develop and populated the spleen. These studies are consistent with a model in which developmental arrest is coupled to receptor editing, and cells that fail to appropriately modify their receptors die.

Locus Specificity of B Cell Receptor Editing Several lines of evidence suggest that tolerance-induced receptor editing in immature B cells is focused primarily on the Ig-L chain loci, although some editing at the H-chain locus may also occur. One indication of this is that most autoantibody Tg mice exhibit excellent H-chain exclusion, but poor L-chain allelic exclusion (152, 158, 159). Other evidence has been gleaned from cell culture models of receptor editing. Bone marrow B cells of 3–83 Tg mice cultured in IL-7 maintain excellent H and L-chain exclusion in the absence of cognate antigen, but jrearrangements could be induced in over two thirds of cells upon anti-BCR treatment (181). In these same cells, VDJ assembly on the H-chain locus was not detectable by a sensitive PCR assay. However, as discussed above, conventional H-chain transgenic mice do not allow an assessment of VH replacement events, and when targeted H-chain autoantibody mice have been analyzed, VH-to-VDJ replacement occurred infrequently and usually included N-region additions indicative of recombination at the proB stage prior to sIgM expression (152).

Developmental Stage Specificity of Receptor Editing Immature bone marrow B cells sensitive to tolerance-induced receptor editing rapidly lose this sensitivity as they mature, and that stage is succeeded by an apoptosis-sensitive one. These cells can be distinguished by a series of surface


36

NEMAZEE

markers (186) and by the ability of the latter cells to migrate to the spleen (187). More detailed analysis has been facilitated by the ability to selectively expand receptor editing–sensitive B cells from bone marrow using IL-7 and to follow their differentiation after IL-7 withdrawal (146, 188). In these cultures, IL-7 withdrawal is required for recombinase expression, possibly because cell cycle events in proliferating cells suppress recombinase expression (189). RAG-2 protein is selectively degraded in cycling cells (189). Upon IL-7 withdrawal, cells express increasing amounts of sIg, and the ability to respond to antigen challenge by editing decreases and apoptosis sensitivity increases concomitantly. Since more mature cells presumably also must retain the option to fix their receptors and to participate in the immune response, the inability of these cells to be induced to receptor editing by antigen alone makes sense. Is there a role for receptor editing in receptor diversification? Besides providing an elegant self-tolerance mechanism, receptor editing may have an additional biological benefit. V(D)J recombination is often far from random because differences exist within sets of variable and joining segments in their recombination signal sequences, promoter regions, and proximity to other cis-acting elements that affect the efficiency of recombination. As a result, overrepresentation of certain genes is common and perhaps inevitable. Secondary and higher order rearrangements may play an important role in promoting a higher degree of randomness in V and J gene usage than would occur if each locus underwent only a single recombination.

PERIPHERAL EDITING Receptor Revision Even mature B cells, it was shown recently, can undergo receptor editing during the immune response in germinal centers (190–193). The first indications of this were that RAG mRNA and protein were inducibly expressed in mature B cells under certain culture conditions or upon a germinal center immune reaction (190, 194). Histological analysis of spleens and lymph nodes of immunized mice indicated that the cell type expressing RAG in vivo was the centrocyte (190, 195). The centrocyte is a nondividing cell that has recently proliferated in the germinal center reaction and is intimately contacted and selected by the follicular dendritic cell network (196). Concerted treatment of splenic B cells with CD40 agonist and IL-4, agents that probably mimic T cell help and induce heavy chain class switching, could rapidly induce upregulation of RAG genes (193, 195, 197). Combined treatment with bacterial lipopolysaccharides and IL-4 had a similar effect (190, 195). Later studies suggested that IL-7 could substitute for IL-4 and was probably the critical cytokine in vivo because antibodies to IL-7R, though they permitted germinal center formation, prevented RAG expression in centrocytes (197).


RECEPTOR EDITING

37

RAG expression in mature B cells was strongly associated with B cell death both in vivo and in vitro and was often present in cells that were engulfed by macrophages (190, 195). In histological sections of germinal centers, RAG protein expression appeared to be localized to the cytoplasm rather than the nucleus (190, 194). RAG gene products induced in vivo and in vitro were nevertheless shown to cause double-stranded DNA breaks adjacent to recombination signal sequences (191) and new L-chain protein expression (192, 193, 198). Curiously, the cells undergoing these recombination reactions and receptor alterations did not manifest a loss of surface Ig expression (192), which would be predicted to be a common occurrence, suggesting rapid death of sIg- cells in these cultures. Perhaps most intriguingly, the cell fractions that expressed recombinase also expressed other markers characteristic of B cell precursors, including surrogate L-chain components (199), IL-7R (197), the surface marker GL-7 (194), and, in human germinal center cells, TdT (199). The reexpression of surrogate L-chain components is particularly interesting because of the likelihood that many editing events would initially silence L-chain production, perhaps requiring surrogate Lchain to temporarily pair with H-chain. Expression of surrogate L-chain components is also observed in cycling, centroblast cells (199). These unexpected similarities between bone marrow and germinal center cells, first pointed out by Han, Kelsoe, and colleagues on the basis of RAG expression (194), suggest that many elements of B cell development are recapitulated during the immune response. However, in the regulation of recombinase expression, there are critical differences as well. In the bone marrow microenvironment, immature B cells require only BCR ligation to undergo receptor editing, whereas different stimuli drive receptor editing in germinal center cells, and BCR stimulation under such conditions actively blocks the recombinase response (198, 199). In one study, in vivo RAG expression appeared to be present in a subset of germinal center centrocytes distinguishable by reduced CD45 levels and nonoptimal receptor gene usage for cognate immunogen (191). It is therefore unlikely that immune tolerance induces V(D)J recombination in germinal center cells. Instead, the data are most consistent with a role for receptor editing in the diversification of the receptors of antigenreactive cells. However, as we discuss below, this interpretation also presents problems. Very recently, data has been presented suggesting that receptor revision may occur concurrently with somatic hypermutation. Weigert and colleagues have identified a clone of anti-dsDNA reactive cells captured as hybridomas from MRL/lpr mice harboring targeted L-chain and H-chain genes encoding anti-DNA (F Brard, M Shannon, EL Park, S Litwin, M Weigert, in press). In this clone the L-chain Tg was heavily mutated, and all clone members shared a lethal stop mutation. The clone expressed a second L-chain on the other kappa allele that had acquired fewer mutations. Assuming that the new j-gene rearrangement appeared some time after the initiation of somatic mutation, these results imply that revision can occur in cells that undergo point hypermutation and that editing


38

NEMAZEE

may be specifically stimulated by the loss of BCR expression. In a second study, de Wildt et al (200) identified among 365 human primary IgG` clones a pair of cells expressing an identical hypermutated H chain gene and two different expressed L-chain genes, which were presumably rearranged after these mutations. Since the L-chain genes themselves showed evidence of somatic mutation, hypermutation and receptor editing can apparently occur concurrently in these cells. A second recent study argues that receptor revision can occur in mouse B-1 cells (201), which are not thought to take part in germinal center reactions. IgM`B220low peritoneal cells, particularly those from the autoimmune-prone NZB strain, express RAG mRNA and possess double-strand breaks at Jj RSSs. In analysis of mice carrying functional replacements of IgH and IgL variable genes, B-1 cells lost idiotypic determinants, indicative of receptor editing at the protein level. The function of such modification is unclear and, as in the case of germinal center B cells, has important implications for the generation of autoantibodies. Locus Specificity of Receptor Revision Several studies documented renewed recombination of L-chain genes in mature B cells (191–193, 197, 199, 201), and a single study provided possible evidence of rearrangements at the IgH locus (192). This latter result was obtained in a VDJ replacement mouse in which both nonphysiological D-to-VDJ joins and potentially physiological V-to-VDJ replacement reactions are possible. Thus, either this experimental system may represent the ideal model to reveal physiologically relevant receptor revision on the IgHlocus, or it may allow recombination events on the IgH-locus that rarely occur in normal cells. It will be important to determine if receptor revision can occur at the H-loci through V-to-VDJ replacements.

RECEPTOR EDITING IN T CELLS As discussed earlier, TCRa allelic exclusion essentially fails to occur because of ongoing rearrangements in cells prior to positive selection. This raises the question of whether an editing mechanism may play a role in thymocyte negative selection. Wang et al (202) established a targeted replacement mouse using a rearranged TCRa VJ gene from a pigeon cytochrome c (PCC)–specific T cell hybridoma inserted in the 58 region of the germ-line J locus. This gene was expressed in most double positive cells but was lost, presumably through nested rearrangements, in all but ;3% of peripheral T cells. Retention of gene expression was greatly improved when a TCRb Tg encoding the original hybridoma was coexpressed on a positively selecting background but was not improved on a nonselecting background. Most importantly, in the presence of PCC, negative selection was not associated with loss of thymic cellularity, in contrast to what was seen with conventional anti-PCC TCR transgenic mice, and nonautoreactive


RECEPTOR EDITING

39

cells populated the periphery. While these results failed to prove that autoreactivity promoted further rearrangements, the possibility is not excluded, and this model should allow this question to be further tested. It is equally possible that the autoreactive cells in this system were positively selected and subsequently lost by apoptosis, but the compartment was filled with competing cells that had lost expression of the targeted TCRa gene prior to antigen encounter. Perhaps more remarkable is the recent suggestion that receptor editing can occur during peripheral T cell tolerance (203). In this experiment, a TCRb-chain (Vb5) Tg mouse was constructed and tolerance to the endogenous mouse mammary tumor virus superantigen MTV-8 was assessed. In antigen-expressing mice most Vb5` cells were deleted, and the remainder were functionally inactivated. However, over time, a Vb51 population emerged. It was concluded that in a subset of the tolerant cells, extra-thymic V(D)J recombination occurred, altering and replacing their receptors because RAG expression and doublestranded breaks adjacent to TCR Db2 and Ja50 were detected in a subset of peripheral T cells. Furthermore, the appearance of Vb51 cells, but not tolerance, was dependent on the presence of B cells, suggesting that a B:T interaction, such as a germinal center reaction, might be required for this process. A potential inconsistency between the data and this model was that the RAG-expression was detected in Vb51 rather than Vb5` cells. These points will no doubt be addressed in the near future as experimental systems are designed with these questions in mind (202). Finally, it should be noted that the structure of the TCRb loci is consistent with receptor editing, as active genes using the 58-most D/J/Cb cluster can in theory be subsequently eliminated and replaced by V-to-DJ rearrangements to the 38 D/J/Cb cluster on the same chromosome (Figure 3). In addition, the spacers of their RSS elements are theoretically compatible with direct Vb-to-Jb rearrangement, but this has never been observed.

CONCLUSION The ability of immune cells to turn V(D)J recombinase on and off, or to redirect its activity to new gene loci, appears to regulate lymphocyte specificity in a novel way distinct from, and complementary to, the control of cellular growth and survival. This regulation of recombinase is largely controlled by signaling through the antigen receptor itself and constitutes a feedback mechanism that allows cells to ‘‘repair’’ receptor genes whose products are dysfunctional, or of inappropriate specificity, including those with too high or too low affinity for self-antigens. Receptor editing facilitates immune self-tolerance during B cell development and also permits developing thymocytes to screen multiple TCRs for potential positive selection. Mature B lymphocytes (and possibly also T cells) may re-express recombinase upon foreign antigenic stimulation, possibly allowing more rapid clonal evolution to high-affinity reactivity. The organization of antigen receptor genes appears to facilitate these processes, often at the cost of limiting potential

40

NEMAZEE

receptor gene diversity and lymphocyte monospecificity. The full extent to which these receptor selection processes normally occur, or benefit the organism, remains to be elucidated. ACKNOWLEDGMENTS This work was supported by grants from the National Institutes of Health and the Arthritis Foundation. The author thanks Anne Feeney, Norman Klinman, Martin Weigert, and members of the laboratory for their comments on the manuscript, and Kathy Offerding for secretarial assistance.


Note Added in Proof Since the submission of the manuscript, the cloning by Zachau’s group of the murine j-locus has made extraordinary progress. For information on this fastmoving area and associated recent publications, the reader is referred to their web site: http://www.med.uni-muenchen.de/biochemie/zachau/kappa,htm. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Radic MZ, Zouali M. 1996. Receptor editing, immune diversification, and selftolerance. Immunity 5:505–11 2. Hertz M, Nemazee D. 1998. Receptor editing and commitment in B lymphocytes. Curr. Opin. Immunol. 10:208–13 3. Nussenzweig MC. 1998. Immune receptor editing: revise and select. Cell 95:875–78 4. Ohmori H, Hikida M. 1998. Expression and function of recombination activating genes in mature B cells. Crit. Rev. Immunol. 18:221–35 5. Kofler R, Geley S, Kofler H, Helmberg A. 1992. Mouse variable-region gene families: complexity, polymorphism and use in non-autoimmune responses. Immunol. Rev. 128:5–21 6. Tonegawa S, Maxam AM, Tizard R, Bernard O, Gilbert W. 1978. Sequence of a mouse germ-line gene for a variable region of an immunoglobulin light chain. Proc. Natl. Acad. Sci. USA 75:1485–89

7. Cory S, Tyler BM, Adams JM. 1981. Sets of immunoglobulin V kappa genes homologous to ten cloned V kappa sequences: implications for the number of germline V kappa genes. J. Mol. Appl. Genet. 1:103–16 8. Kirschbaum T, Pourrajabi S, Zocher I, Schwendinger J, Heim V, Roschenthaler F, Kirschbaum V, Zachau HG. 1998. The 38 part of the immunoglobulin kappa locus of the mouse. Eur. J. Immunol. 28:1458–66 9. Max EE, Seidman JG, Leder P. 1979. Sequences of five potential recombination sites encoded close to an immunoglobulin kappa constant region gene. Proc. Natl. Acad. Sci. USA 76:3450–54 10. Seidman JG, Leder P. 1978. The arrangement and rearrangement of antibody genes. Nature 276:790–95 11. Seidman JG, Max EE, Leder P. 1979. A kappa-immunoglobulin gene is formed by site-specific recombination without

RECEPTOR EDITING

12.

13.

14.


15.

16.

17.

18.

19.

20.

21.

further somatic mutation. Nature 280: 370–75 Zachau HG. 1989. Immunoglobulin light chain genes of the j type in human in man and mouse. In Immunoglobulin Genes, ed. T Honjo, FW Alt, TH Rabbitts, pp. 91–110. San Diego: Harcourt Brace. 1st ed. Hieter PA, Korsmeyer SJ, Waldmann TA, Leder P. 1981. Human immunoglobulin kappa light-chain genes are deleted or rearranged in lambda-producing B cells. Nature 290:368–72 Zachau HG. 1995. The human immunoglobulin j genes. See Ref. 204, pp. 174– 91 Wood DL, Coleclough C. 1984. Different joining region J elements of the murine kappa immunoglobulin light chain locus are used at markedly different frequencies. Proc. Natl. Acad. Sci. USA 81:4756–60 Selsing E, Voss J, Storb U. 1984. Immunoglobulin gene ‘remnant’ DNA– implications for antibody gene recombination. Nucleic Acids Res. 12:4229–46 Lewis S, Rosenberg N, Alt F, Baltimore D. 1982. Continuing kappa-gene rearrangement in a cell line transformed by Abelson murine leukemia virus. Cell 30:807–16 Feddersen RM, Van Ness BG. 1985. Double recombination of a single immunoglobulin kappa-chain allele: implications for the mechanism of rearrangement. Proc. Natl. Acad. Sci. USA 82:4793–97 Shapiro MA, Weigert M. 1987. How immunoglobulin V kappa genes rearrange. J. Immunol. 139:3834–39 Feddersen RM, Martin DJ, Van Ness BG. 1990. Novel recombinations of the IG kappa-locus that result in allelic exclusion. J. Immunol. 145:745–50 Clarke S, McCray S. 1991. A shared kappa reciprocal fragment and a high frequency of secondary Jk5 rearrangements

22.

23.

24.

25.

26.

27.

28.

29.

30.

41

among influenza hemagglutinin specific B cell hybridomas. J. Immunol. 146:343– 49 Huber C, Klobeck HG, Zachau HG. 1992. Ongoing V kappa-J kappa recombination after formation of a productive V kappa-J kappa coding joint. Eur. J. Immunol. 22:1561–65 Prak EL, Trounstine M, Huszar D, Weigert M. 1994. Light chain editing in kappa-deficient animals: a potential mechanism of B-cell tolerance. J. Exp. Med. 180:1805–15 Harada K, Yamagishi H. 1991. Lack of feedback inhibition of V kappa gene rearrangement by productively rearranged alleles. J. Exp. Med. 173:409–15 Shimizu T, Iwasato T, Yamagishi H. 1991. Deletions of immunoglobulin C kappa region characterized by the circular excision products in mouse splenocytes. J. Exp. Med. 173:1065–72 Koop BF, Wilson RK, Wang K, Vernooij B, Zallwer D, Kuo CL, Seto D, Toda M, Hood L. 1992. Organization, structure, and function of 95 kb of DNA spanning the murine T-cell receptor C alpha/C delta region. Genomics 13:1209–30 Wilson RK, Koop BF, Chen C, Halloran N, Sciammis R, Hood L. 1992. Nucleotide sequence analysis of 95 kb near the 38 end of the murine T-cell receptor alpha/delta chain locus: strategy and methodology. Genomics 13:1198–1208 Concannon P, Lai E, Klein M, Siu S, Strauss E, Pickering L, Kung P, Gatti R, Hood L. 1986. Human T-cell receptor genes: organization, diversity, and polymorphism. Cold Spring Harbor Symp. Quant. Biol. 51(Pt. 2):785–95 Winoto A, Mjolsness S, Hood L. 1985. Genomic organization of the genes encoding mouse T-cell receptor alphachain. Nature 316:832–36 Takeshita S, Toda M, Yamagishi H. 1989. Excision products of the T cell receptor gene support a progressive rear-

42

31.

32.


33.

34.

35.

36.

37.

38.

NEMAZEE rangement model of the alpha/delta locus. EMBO J. 8:3261–70 Petrie HT, Livak F, Burtrum D, Mazel S. 1995. T cell receptor gene recombination patterns and mechanisms: cell death, rescue, and T cell production. J. Exp. Med. 182:121–27 Petrie HT, Livak F, Schatz DG, Strasser A, Crispe IN, Shortman K. 1993. Multiple rearrangements in T cell receptor alpha chain genes maximize the production of useful thymocytes. J. Exp. Med. 178:615–22 Livak F, Schatz DG. 1996. T-cell receptor alpha locus V(D)J recombination byproducts are abundant in thymocytes and mature T cells. Mol. Cell. Biol. 16:609– 18 Kong F, Chen CH, Cooper MD. 1998. Thymic function can be accurately monitored by the level of recent T cell emigrants in the circulation. Immunity 8:97– 104 Douek DC, McFarland RD, Keiser PH, Gage EA, Massey JM, Haynes BF, Polis MA, Haase AT, Feinberg MB, Sullivan JL, Jamieson BD, Zack JA, Picker LJ, Koup RA. 1998. Changes in thymic function with age and during the treatment of HIV infection. Nature 396:690– 95 Thompson SD, Pelkonen J, Hurwitz JL. 1990. First T cell receptor alpha gene rearrangements during T cell ontogeny skew to the 58 region of the J alpha locus. J. Immunol. 145:2347–52 Matsuda F, Ishii K, Bourvagnet P, Kuma K, Hayashida H, Miyata T, Honjo T. 1998. The complete nucleotide sequence of the human immunoglobulin heavy chain variable region locus. J. Exp. Med. 188:2151–62 Brodeur PH, Osman GE, Mackle JJ, Lalor TM. 1988. The organization of the mouse Igh-V locus. Dispersion, interspersion, and the evolution of VH gene family clusters. J. Exp. Med. 168:2261– 78

39. Reth MG, Jackson S, Alt FW. 1986. VHDJH formation and DJH replacement during pre-B differentiation: non-random usage of gene segments. EMBO J. 5:2131–38 40. Corbett SJ, Tomlinson IM, Sonnhammer ELL, Buck D, Winter G. 1997. Sequence of the human immunoglobulin diversity (D) segment locus: a systematic analysis provides no evidence for the use of DIR segments, inverted D segments, ‘‘minor’’ D segments or D-D recombination. J. Mol. Biol. 270:587–97 41. Kleinfield RW, Weigert MG. 1989. Analysis of VH gene replacement events in a B cell lymphoma. J. Immunol. 142:4475–82 42. Bernard O, Hozumi N, Tonegawa S. 1978. Sequences of mouse immunoglobulin light chain genes before and after somatic changes. Cell 15:1133–44 43. Brandle D, Muller C, Rulicke T, Hengartner H, Pircher H. 1992. Engagement of the T-cell receptor during positive selection in the thymus down-regulates RAG-1 expression. Proc. Natl. Acad. Sci. USA 89:9529–33 44. Carson S, Wu GE. 1989. A linkage map of the mouse immunoglobulin lambda light chain locus. Immunogenetics 29:173–79 45. Dildrop R, Gause A, Muller W, Rajewsky K. 1987. A new V gene expressed in lambda-2 light chains of the mouse. Eur. J. Immunol. 17:731–34 46. Sanchez P, Cazenave PA. 1987. A new variable region in mouse immunoglobulin lambda light chains. J. Exp. Med. 166:265–70 47. Selsing E, Daitch LE. 1995. Immunoglobulin k genes. See Ref. 204, pp. 194– 203 48. Eisen HN, Reilly EB. 1985. Lambda chains and genes in inbred mice. Annu. Rev. Immunol. 3:337–65 49. Kawasaki K, Minoshima S, Nakato E, Shibuya K, Shintani A, Schmeits JL, Wang J, Shimizu N. 1997. One-

RECEPTOR EDITING

50.

51.


52.

53.

54.

55.

56.

57.

58.

59.

megabase sequence analysis of the human immunoglobulin lambda gene locus. Genome Res. 7:250–61 Stiernholm NBJ, Berinstein NL. 1994. Immunoglobulin somatic variation; studies of receptor editing in a human B cell lymphoma. Leuk. Lymphoma 12:333–41 Siminovitch KA, Moore MW, Durdik J, Selsing E. 1987. The human kappa deleting element and the mouse recombining segment share DNA sequence homology. Nucleic Acids Res. 15:2699–705 Moore MW, Durdik J, Persiani DM, Selsing E. 1985. Deletions of kappa chain constant region genes in mouse lambda chain-producing B cells involve intrachromosomal DNA recombinations similar to V-J joining. Proc. Natl. Acad. Sci. USA 82:6211–15 Muller B, Stappert H, Reth M. 1990. A physical map and analysis of the murine C kappa-RS region show the presence of a conserved element. Eur. J. Immunol. 20:1409–11 Klobeck HG, Zachau HG. 1986. The human CK gene segment and the kappa deleting element are closely linked. Nucleic Acids Res. 14:4591–603 Durdik J, Moore MW, Selsing E. 1984. Novel kappa light-chain gene rearrangements in mouse lambda light chainproducing B lymphocytes. Nature 307: 749–52 Nadel B, Cazenave PA, Sanchez P. 1990. Murine lambda gene rearrangements: the stochastic model prevails over the ordered model. EMBO J. 9:435–40 Retter MW, Nemazee D. 1998. Receptor editing occurs frequently during normal B cell development. J. Exp. Med. 188:1231–38 Dunda O, Corcos D. 1997. Recombining sequence recombination in normal kappa-chain-expressing B cells. J. Immunol. 159:4362–66 Daitch LE, Moore MW, Persiani DM, Durdik JM, Selsing E. 1992. Transcrip-

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

43

tion and recombination of the murine RS element. J. Immunol. 149:832–40 Muller B, Reth M. 1988. Ordered activation of the Ig lambda locus in Abelson B cell lines. J. Exp. Med. 168:2131–37 Begley CG, Aplan PD, Davey MP, de Villartay JP, Cohen DI, Waldmann TA, Kirsch IR. 1989. Demonstration of delta rec-pseudo J alpha rearrangement with deletion of the delta locus in a human stem-cell leukemia. J. Exp. Med. 170:339–42 Hockett RD, de Villartay JP, Pollock K, Poplack DG, Cohen DI, Korsmeyer SJ. 1988. Human T-cell antigen receptor (TCR) delta-chain locus and elements responsible for its deletion are within the TCR alpha-chain locus. Proc. Natl. Acad. Sci. USA 85:9694–98 Verschuren MC, Wolvers-Tettero IL, Breit TM, Noordzij J, van Wering ER, van Dongen JJ. 1997. Preferential rearrangements of the T cell receptor-deltadeleting elements in human T cells. J. Immunol. 158:1208–16 Du Pasquier L, Schwager J. 1989. Evolutions of the immune system. Prog. Immunol. 7:1246–55 Ehlich A, Martin V, Muller W, Rajewsky K. 1994. Analysis of the B-cell progenitor compartment at the level of single cells. Curr. Biol. 4:573–83 Kitamura D, Kudo A, Schaal S, Muller W, Melchers F, Rajewsky K. 1992. A critical role of lambda 5 protein in B cell development. Cell 69:823–31 Kudo A, Melchers F. 1987. A second gene, VpreB in the lambda 5 locus of the mouse, which appears to be selectively expressed in pre-B lymphocytes. EMBO J. 6:2267–72 Karasuyama H, Rolink A, Melchers F. 1996. Surrogate light chain in B cell development. Adv. Immunol. 63:1–41 Reth M, Wienands J. 1997. Initiation and processing of signals from the B cell antigen receptor. Annu. Rev. Immunol. 15:453–79


44

NEMAZEE

70. Tsubata T, Tsubata R, Reth M. 1991. Cell surface expression of the short immunoglobulin mu chain (D mu protein) in murine pre-B cells is differently regulated from that of the intact mu chain. Eur. J. Immunol. 21:1359–63 71. Tsubata T, Tsubata R, Reth M. 1992. Crosslinking of the cell surface immunoglobulin (mu-surrogate light chains complex) on pre-B cells induces activation of V gene rearrangements at the immunoglobulin kappa locus. Int. Immunol. 4:637–41 72. Decker DJ, Boyle NE, Klinman NR. 1991. Predominance of nonproductive rearrangements of VH81X gene segments evidences a dependence of B cell clonal maturation on the structure of nascent H chains. J. Immunol. 147:1406– 11 73. Constantinescu A, Schlissel MS. 1997. Changes in locus-specific V(D)J recombinase activity induced by immunoglobulin gene products during B cell development. J. Exp. Med. 185:609–20 74. Born W, Rathbun G, Tucker P, Marrack P, Kappler J. 1986. Synchronized rearrangement of T-cell gamma and beta chain genes in fetal thymocyte development. Science 234:479–82 75. Snodgrass HR, Kisielow P, Kiefer M, Steinmetz M, von Boehmer H. 1985. Ontogeny of the T-cell antigen receptor within the thymus. Nature 313:592–95 76. Davis MM. 1985. Molecular genetics of the T cell-receptor beta chain. Annu. Rev. Immunol. 3:537–60 77. Honjo T, Matsuda F. 1995. Immunoglobulin heavy chain loci of mouse and human. See Ref. 204, pp. 145–71 78. Fehling HJ, Krotkova A, Saint-Ruf C, von Boehmer H. 1995. Crucial role of the pre-T-cell receptor alpha gene in development of alpha beta but not gamma delta T cells. Nature 375:795–98 79. von Boehmer H, Fehling HJ. 1997. Structure and function of the pre-T cell receptor. Annu. Rev. Immunol. 15:433–52

80. Mombaerts P, Clarke AR, Rudnicki MA, Iacomini J, Itohara S, Lafaille JJ, Wang L, Ichikawa Y, Jaenisch R, Hooper ML, et al. 1992. Mutations in T-cell antigen receptor genes alpha and beta block thymocyte development at different stages. Nature 360:225–31 81. Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa S, Papaioannou VE. 1992. RAG-1-deficient mice have no mature B and T lymphocytes. Cell 68:869–77 82. Shinkai Y, Rathbun G, Lam KP, Oltz EM, Stewart V, Mendelsohn M, Charron J, Datta M, Young F, Stall AM, et al. 1992. RAG-2-deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell 68:855–67 83. Uematsu Y, Ryser S, Dembic Z, Borgulya P, Krimpenfort P, Berns A, von Boehmer H, Steinmetz M. 1988. In transgenic mice the introduced functional T cell receptor beta gene prevents expression of endogenous beta genes. Cell 52:831–41 84. Shinkai Y, Koyasu S, Nakayama K, Murphy KM, Loh DY, Reinherz EL, Alt FW. 1993. Restoration of T cell development in RAG-2-deficient mice by functional TCR transgenes. Science 259:822–25 85. Marolleau JP, Fondell JD, Malissen M, Trucy J, Barbier E, Marcu KB, Cazenave PA, Primi D. 1988. The joining of germline V-alpha to J-alpha genes replaces the preexisting V-alpha-J-alpha complexes in a T-cell receptor-alpha, beta positive T-cell line. Cell 55:291–300 86. Casanova JL, Romero P, Widmann C, Kourilsky P, Maryanski JL. 1991. T cell receptor genes in a series of class I major histocompatibility complex-restricted cytotoxic T lymphocyte clones specific for a Plasmodium berghei nonapeptide: implications for T cell allelic exclusion and antigen-specific repertoire. J. Exp. Med. 174:1371–83 87. Jameson SC, Hogquist KA, Bevan MJ. 1995. Positive selection of thymocytes. Annu. Rev. Immunol. 13:93–126


RECEPTOR EDITING 88. Borgulya P, Kishi H, Uematsu Y, von Boehmer H. 1992. Exclusion and inclusion of alpha and beta T cell receptor alleles. Cell 69:529–37 89. Kouskoff V, Vonesch JL, Benoist C, Mathis D. 1995. The influence of positive selection on RAG expression in thymocytes. Eur. J. Immunol. 25:54–58 90. Turka LA, Schatz DG, Oettinger MA, Chun JJ, Gorka C, Lee K, McCormack WT, Thompson CB. 1991. Thymocyte expression of RAG-1 and RAG-2: termination by T cell receptor cross-linking. Science 253:778–81 91. Chien YH, Iwashima M, Kaplan KB, Elliott JF, Davis MM. 1987. A new Tcell receptor gene located within the alpha locus and expressed early in T-cell differentiation. Nature 327:677–82 92. Pardoll DM, Fowlkes BJ, Bluestone JA, Kruisbeek A, Maloy WL, Coligan JE, Schwartz RH. 1987. Differential expression of two distinct T-cell receptors during thymocyte development. Nature 326:79–81 93. Raulet DH, Garman RD, Saito H, Tonegawa S. 1985. Developmental regulation of T-cell receptor gene expression. Nature 314:103–7 94. Washburn T, Schweighoffer E, Gridley T, Chang D, Fowlkes BJ, Cado D, Robey E. 1997. Notch activity influences the alpha-beta versus gamma-delta T-cell lineage decision. Cell 88:833–43 95. Robey E, Fowlkes BJ. 1998. The alpha beta versus gamma-delta T-cell lineage choice. Curr. Opin. Immunol. 10:181–87 96. Li YS, Hayakawa K, Hardy RR. 1993. The regulated expression of B lineage associated genes during B cell differentiation in bone marrow and fetal liver. J. Exp. Med. 178:951–60 97. Wilson A, Held W, MacDonald HR. 1994. Two waves of recombinase gene expression in developing thymocytes. J. Exp. Med. 179:1355–60 98. Cebra JJ, Colberg JE, Dray S. 1966. Rabbit lymphoid cells differentiated with

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

45

respect to alpha-, gamma-, and muheavy polypeptide chains and to allotypic markers Aa1 and Aa2. J. Exp. Med. 123:547–58 Weiler E. 1965. Differential activity of allelic gamma-globulin genes in antibody-producing cells. Proc. Natl. Acad. Sci. USA 54:1765–72 Bernier GM, Cebra JJ. 1964. Polypeptide chains of human gamma-globulin cellular localization by fluorescent antibody. Science 144:1590–91 Pernis B, Chiappino G, Kelus AS, Gell PG. 1965. Cellular localization of immunoglobulins with different allotypic specificities in rabbit lymphoid tissues. J. Exp. Med. 122:853–76 Tsukamoto A, Weissman IL, Hunt SV. 1984. Allelic exclusion in rat kappa immunoglobulin chains: extent of Jk rearrangement in normal B lymphocytes. EMBO J. 3:975–81 Kitamura D, Rajewsky K. 1992. Targeted disruption of mu chain membrane exon causes loss of heavy-chain allelic exclusion. Nature 356:154–56 Coleclough C. 1983. Chance, necessity and antibody gene dynamics. Nature 303:23–26 Takeda S, Zou YR, Bluethmann H, Kitamura D, Muller U, Rajewsky K. 1993. Deletion of the immunoglobulin kappa chain intron enhancer abolishes kappa chain gene rearrangement in cis but not lambda chain gene rearrangement in trans. EMBO J. 12:2329–36 Malissen M, Trucy J, Jouvin-Marche E, Cazenave PA, Scollay R, Malissen B. 1992. Regulation of TCR alpha and beta gene allelic exclusion during T-cell development. Immunol. Today 13:315– 22 Padovan E, Casorati G, Dellabona P, Meyer S, Brockhaus M, Lanzavecchia A. 1993. Expression of two T cell receptor alpha chains: dual receptor T cells. Science 262:422–24 Sleckman BP, Khor B, Monroe R, Alt

46

109.

110.


111.

112.

113.

114.

115.

116.

117.

118.

NEMAZEE FW. 1998. Assembly of productive T cell receptor delta variable region genes exhibits allelic inclusion. J. Exp. Med. 188:1465–71 Davodeau F, Peyrat MA, Houde I, Hallet MM, De Libero G, Vie H, Bonneville M. 1993. Surface expression of two distinct functional antigen receptors on humangamma-delta T-cells. Science 260: 1800–2 Mostoslavsky R, Singh N, Kirillov A, Pelanda R, Cedar H, Chess A, Bergman Y. 1998. Kappa chain monoallelic demethylation and the establishment of allelic exclusion. Genes Dev. 12:1801–11 Alt FW, Enea V, Bothwell AL, Baltimore D. 1980. Activity of multiple light chain genes in murine myeloma cells producing a single, functional light chain. Cell 21:1–12 Coleclough C, Perry RP, Karjalainen K, Weigert M. 1981. Aberrant rearrangements contribute significantly to the allelic exclusion of immunoglobulin gene expression. Nature 290:372–78 Korsmeyer SJ, Hieter PA, Sharrow SO, Goldman CK, Leder P, Waldmann TA. 1982. Normal human B cells display ordered light chain gene rearrangements and deletions. J. Exp. Med. 156:975–85 Arakawa H, Shimizu T, Takeda S. 1996. Re-evaluation of the probabilities for productive arrangements on the kappa and lambda loci. Int. Immunol. 8:91–99 Wabl M, Steinberg C. 1982. A theory of allelic and isotypic exclusion for immunoglobulin genes. Proc. Natl. Acad. Sci. USA 79:6976–78 Sonoda E, Pewzner-Jung Y, Schwers S, Taki S, Jung S, Eilat D, Rajewsky K. 1997. B cell development under the condition of allelic inclusion. Immunity 6:225–33 Nemazee D. 1996. Antigen receptor ‘capacity’ and the sensitivity of selftolerance. Immunol. Today 17:25–29 Loffert D, Ehlich A, Muller W, Rajewsky K. 1996. Surrogate light chain expression

119.

120.

121.

122.

123.

124.

125.

126.

127.

is required to establish immunoglobulin heavy chain allelic exclusion during early B cell development. Immunity 4:133–44 Rusconi S, Kohler G. 1985. Transmission and expression of a specific pair of rearranged immunoglobulin mu and kappa genes in a transgenic mouse line. Nature 314:330–34 Sakaguchi N, Melchers F. 1986. Lambda 5, a new light-chain-related locus selectively expressed in pre-B lymphocytes. Nature 324:579–82 Nussenzweig MC, Shaw AC, Sinn E, Danner DB, Holmes KL, Morse HC, Leder P. 1987. Allelic exclusion in transgenic mice that express the membrane form of immunoglobulin mu. Science 236:816–19 Sakaguchi N, Berger CN, Melchers F. 1986. Isolation of a cDNA copy of an RNA species expressed in murine pre-B cells. EMBO J. 5:2139–47 Manz J, Denis K, Witte O, Brinster R, Storb U. 1988. Feedback inhibition of immunoglobulin gene rearrangement by membrane mu, but not by secreted mu heavy chains. J. Exp. Med. 168:1363–81 Kitamura D, Roes J, Kuhn R, Rajewsky K. 1991. A B cell-deficient mouse by targeted disruption of the membrane exon of the immunoglobulin mu chain gene. Nature 350:423–26 Nussenzweig MC, Shaw AC, Sinn E, Campos-Torres J, Leder P. 1988. Allelic exclusion in transgenic mice carrying mutant human IgM genes. J. Exp. Med. 167:1969–74 Papavasiliou F, Misulovin Z, Suh H, Nussenzweig MC. 1995. The role of Ig beta in precursor B cell transition and allelic exclusion. Science 268:408–11 Papavasiliou F, Jankovic M, Suh H, Nussenzweig MC. 1995. The cytoplasmic domains of immunoglobulin (Ig) alpha and Ig beta can independently induce the precursor B cell transition and allelic exclusion. J. Exp. Med. 182:1389–94


RECEPTOR EDITING 128. Weaver D, Costantini F, Imanishi-Kari T, Baltimore D. 1985. A transgenic immunoglobulin mu gene prevents rearrangement of endogenous genes. Cell 42:117– 27 129. Nussenzweig MC, Schmidt EV, Shaw AC, Sinn E, Campos-Torres J, MatheyPrevot B, Pattengale PK, Leder P. 1988. A human immunoglobulin gene reduces the incidence of lymphomas in c-Mycbearing transgenic mice. Nature 336:446–50 130. Keyna U, Beck-Engeser GB, Jongstra J, Applequist SE, Jack HM. 1995. Surrogate light chain-dependent selection of Ig heavy chain V regions. J. Immunol. 155:5536–42 131. Keyna U, Applequist SE, Jongstra J, Beck-Engeser GB, Jack HM. 1995. Ig mu heavy chains with VH81X variable regions do not associate with lambda 5. Ann. NY Acad. Sci. 764:39–42 132. Alt FW, Yancopoulos GD, Blackwell TK, Wood C, Thomas E, Boss M, Coffman R, Rosenberg N, Tonegawa S, Baltimore D. 1984. Ordered rearrangement of immunoglobulin heavy chain variable region segments. EMBO J. 3:1209–19 133. Hardy RR, Carmack CE, Shinton SA, Kemp JD, Hayakawa K. 1991. Resolution and characterization of pro-B and pre-pro-B cell stages in normal mouse bone marrow. J. Exp. Med. 173:1213–25 134. Gu H, Kitamura D, Rajewsky K. 1991. B cell development regulated by gene rearrangement: arrest of maturation by membrane-bound D mu protein and selection of DH element reading frames. Cell 65:47–54 135. Tornberg UC, Bergqvist I, Haury M, Holmberg D. 1998. Regulation of B lymphocyte development by the truncated immunoglobulin heavy chain protein Dmu. J. Exp. Med. 187:703–9 136. O’Shea CC, Thornell AP, Rosewell IR, Hayes B, Owen MJ. 1997. Exit of the pre-TCR from the ER/cis-Golgi is necessary for signaling differentiation, pro-

137.

138.

139.

140.

141.

142.

143.

144.

47

liferation, and allelic exclusion in immature thymocytes. Immunity 7:591– 99 Irving BA, Alt FW, Killeen N. 1998. Thymocyte development in the absence of pre-T cell receptor extracellular immunoglobulin domains. Science 280: 905–8 Chan VW, Meng F, Soriano P, DeFranco AL, Lowell CA. 1997. Characterization of the B lymphocyte populations in Lyndeficient mice and the role of Lyn in signal initiation and down-regulation. Immunity 7:69–81 Gollahon KA, Hagman J, Brinster RL, Storb U. 1988. Ig lambda-producing B cells do not show feedback inhibition of gene rearrangement. J. Immunol. 141:2771–80 Pauza ME, Rehmann JA, LeBien TW. 1993. Unusual patterns of immunoglobulin gene rearrangement and expression during human B cell ontogeny: human B cells can simultaneously express cell surface kappa and lambda light chains. J. Exp. Med. 178:139–49 Giachino C, Padovan E, Lanzavecchia A. 1995. Kappa(`)lambda(`) dual receptor B cells are present in the human peripheral repertoire. J. Exp. Med. 181:1245–50 Ritchie KA, Brinster RL, Storb U. 1984. Allelic exclusion and control of endogenous immunoglobulin gene rearrangement in kappa transgenic mice. Nature 312:517–20 Carmack CE, Camper SA, Mackle JJ, Gerhard WU, Weigert MG. 1991. Influence of a V kappa 8 L chain transgene on endogenous rearrangements and the immune response to the HA(Sb) determinant on influenza virus. J. Immunol. 147:2024–33 Pelanda R, Schaal S, Torres RM, Rajewsky K. 1996. A prematurely expressed Ig(kappa) transgene, but not V(kappa)J(kappa) gene segment targeted into the Ig(kappa) locus, can rescue B

48

145.

146.


147.

148.

149.

150.

151.

152.

153. 154.

155.

NEMAZEE cell development in lambda5-deficient mice. Immunity 5:229–39 Ma A, Fisher P, Dildrop R, Oltz E, Rathbun G, Achacoso P, Stall A, Alt FW. 1992. Surface IgM mediated regulation of RAG gene expression in E mu-N-myc B cell lines. EMBO J. 11:2727–34 Rolink A, Grawunder U, Haasner D, Strasser A, Melchers F. 1993. Immature surface Ig` B cells can continue to rearrange kappa and lambda L chain gene loci. J. Exp. Med. 178:1263–70 Levy S, Campbell MJ, Levy R. 1989. Functional immunoglobulin light chain genes are replaced by ongoing rearrangements of germline V kappa genes to downstream J kappa segment in a murine B cell line. J. Exp. Med. 170:1–13 Berinstein N, Levy S, Levy R. 1989. Activation of an excluded immunoglobulin allele in a human B lymphoma cell line. Science 244:337–39 Verkoczy LK, Stiernholm BJ, Berinstein NL. 1995. Up-regulation of recombination activating gene expression by signal transduction through the surface Ig receptor. J. Immunol. 154:5136–43 Prak EL, Weigert M. 1995. Light chain replacement: a new model for antibody gene rearrangement. J. Exp. Med. 182:541–48 Pelanda R, Schwers S, Sonoda E, Torres RM, Nemazee D, Rajewsky K. 1997. Receptor editing in a transgenic mouse model: site, efficiency, and role in B cell tolerance and antibody diversification. Immunity 7:765–75 Chen C, Prak EL, Weigert M. 1997. Editing disease-associated autoantibodies. Immunity 6:97–105 Nossal GJ. 1994. Negative selection of lymphocytes. Cell 76:229–39 Goodnow CC. 1996. Balancing immunity and tolerance: deleting and tuning lymphocyte repertoires. Proc. Natl. Acad. Sci. USA 93:2264–71 Nemazee D, Buerki K. 1989. Clonal deletion of autoreactive B lymphocytes

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

in bone marrow chimeras. Proc. Natl. Acad. Sci. USA 86:8039–43 Nemazee DA, Burki K. 1989. Clonal deletion of B lymphocytes in a transgenic mouse bearing anti-MHC class I antibody genes. Nature 337:562–66 Hartley SB, Crosbie J, Brink R, Kantor AB, Basten A, Goodnow CC. 1991. Elimination from peripheral lymphoid tissues of self-reactive B lymphocytes recognizing membrane-bound antigens. Nature 353:765–69 Tiegs SL, Russell DM, Nemazee D. 1993. Receptor editing in self-reactive bone marrow B cells. J. Exp. Med. 177:1009–20 Gay D, Saunders T, Camper S, Weigert M. 1993. Receptor editing: an approach by autoreactive B cells to escape tolerance. J. Exp. Med. 177:999–1008 Radic MZ, Erikson J, Litwin S, Weigert M. 1993. B lymphocytes may escape tolerance by revising their antigen receptors. J. Exp. Med. 177:1165–73 Radic MZ, Mascelli MA, Erikson J, Shan H, Weigert M. 1991. Ig H and L chain contributions to autoimmune specificities. J. Immunol. 146:176–82 Erikson J, Radic MZ, Camper SA, Hardy RR, Carmack C, Weigert M. 1991. Expression of anti-DNA immunoglobulin transgenes in non-autoimmune mice. Nature 349:331–34 Pewzner-Jung Y, Friedmann D, Sonoda E, Jung S, Rajewsky K, Eilat D. 1998. B cell deletion, anergy, and receptor editing in ‘‘knock in’’ mice targeted with a germline-encoded or somatically mutated anti-DNA heavy chain. J. Immunol. 161:4634–45 Taki S, Meiering M, Rajewsky K. 1993. Targeted insertion of a variable region gene into the immunoglobulin heavy chain locus. Science 262:1268–71 Chen C, Nagy Z, Prak EL, Weigert M. 1995. Immunoglobulin heavy chain gene replacement: a mechanism of receptor editing. Immunity 3:747–55


RECEPTOR EDITING 166. Kleinfield R, Hardy RR, Tarlinton D, Dangl J, Herzenberg LA, Weigert M. 1986. Recombination between an expressed immunoglobulin heavy-chain gene and a germline variable gene segment in a Ly 1` B-cell lymphoma. Nature 322:843–46 167. Cascalho M, Ma A, Lee S, Masat L, Wabl M. 1996. A quasi-monoclonal mouse. Science 272:1649–52 168. Bertrand FE, Golub R, Wu GE. 1998. VH gene replacement occurs in the spleen and bone marrow of non-autoimmune quasi-monoclonal mice. Eur. J. Immunol. 28:3362–70 169. Fanning L, Bertrand FE, Steinberg C, Wu GE. 1998. Molecular mechanisms involved in receptor editing at the Ig heavy chain locus. Int. Immunol. 10:241– 46 170. Nadel B, Tang A, Feeney AJ. 1998. V(H) replacement is unlikely to contribute significantly to receptor editing due to an ineffectual embedded recombination signal sequence. Mol. Immunol. 35:227–32 171. Lawton AR, Cooper MD. 1974. Modification of B lymphocyte differentiation by anti-immunoglobulins. Contemp. Top. Immunobiol. 3:193–225 172. Nossal GJ. 1983. Cellular mechanisms of immunologic tolerance. Annu. Rev. Immunol. 1:33–62 173. Sidman CL, Unanue ER. 1975. Receptor-mediated inactivation of early B lymphocytes. Nature 257:149–51 174. Raff MC, Owen JJ, Cooper MD, Lawton AR, Megson M, Gathings WE. 1975. Differences in susceptibility of mature and immature mouse B lymphocytes to anti-immunoglobulin-induced immunoglobulin suppression in vitro. Possible implications for B-cell tolerance to self. J. Exp. Med. 142:1052–64 175. Norvell A, Mandik L, Monroe JG. 1995. Engagement of the antigen-receptor on immature murine B-lymphocytes results in death by apoptosis. J. Immunol. 154:4404–13

49

176. Monroe JG. 1996. Tolerance sensitivity of immature-stage B-cells: Can developmentally regulated B-cell antigen receptor (BCR) signal transduction play a role? J. Immunol. 156:2657–60 177. Sandel PC, Monroe JG. 1999. Negative selection of immature B cells by receptor editing or deletion is determined by site of antigen encounter. Immunity 10:289– 99 178. Hartley SB, Cooke MP, Fulcher DA, Harris AW, Cory S, Basten A, Goodnow CC. 1993. Elimination of self-reactive Blymphocytes proceeds in two stages: arrested development and cell death. Cell 72:325–35 179. Hertz M, Nemazee D. 1997. BCR ligation induces receptor editing in IgMÌgD– bone marrow B cells in vitro. Immunity 6:429–36 180. Melamed D, Benschop RJ, Cambier JC, Nemazee D. 1998. Developmental regulation of B lymphocyte immune tolerance compartmentalizes clonal selection from receptor selection. Cell 92:173–82 181. Melamed D, Nemazee D. 1997. Selfantigen does not accelerate immature B cell apoptosis, but stimulates receptor editing as a consequence of developmental arrest. Proc. Natl. Acad. Sci. USA 94:9267–72 182. Lang J, Arnold B, Hammerling G, Harris AW, Korsmeyer S, Russell D, Strasser A, Nemazee D. 1997. Enforced Bcl-2 expression inhibits antigen-mediated clonal elimination of peripheral B-cells in an antigen dose-dependent manner and promotes receptor editing in autoreactive, immature B-cells. J. Exp. Med. 186:1513–22 183. Fang W, Weintraub BC, Dunlap B, Garside P, Pape KA, Jenkins MK, Goodnow CC, Mueller DL, Behrens TW. 1998. Self-reactive B lymphocytes overexpressing Bcl-xL escape negative selection and are tolerized by clonal anergy and receptor editing. Immunity 9:35–45 184. Spanopoulou E, Roman CA, Corcoran

50

185.

186.


187.

188.

189.

190.

191.

192.

193.

NEMAZEE LM, Schlissel MS, Silver DP, Nemazee D, Nussenzweig MC, Shinton SA, Hardy RR, Baltimore D. 1994. Functional immunoglobulin transgenes guide ordered B-cell differentiation in Rag-1deficient mice. Genes Dev. 8:1030–32 Xu H, Li H, Suri-Payer E, Hardy RR, Weigert M. 1998. Regulation of antiDNA B cells in recombination-activating gene-deficient mice. J. Exp. Med. 188:1247–54 Carsetti R, Kohler G, Lamers MC. 1995. Transitional B cells are the target of negative selection in the B cell compartment. J. Exp. Med. 181:2129–40 Allman DM, Ferguson SE, Cancro MP. 1992. Peripheral B cell maturation. I. Immature peripheral B cells in adults are heat-stable antigen hi and exhibit unique signaling characteristics. J. Immunol. 149:2533–40 Melamed D, Kench JA, Grabstein K, Rolink A, Nemazee D. 1997. A functional B cell receptor transgene allows efficient IL-7-independent maturation of B cell precursors. J. Immunol. 159:1233– 39 Lin WC, Desiderio S. 1995. V(D)J recombination and the cell cycle. Immunol. Today 16:279–89 Hikida M, Mori M, Takai T, Tomochika K, Hamatani K, Ohmori H. 1996. Reexpression of RAG-1 and RAG-2 genes in activated mature mouse B cells. Science 274:2092–94 Han S, Dillon SR, Zheng B, Shimoda M, Schlissel MS, Kelsoe G. 1997. V(D)J recombinase activity in a subset of germinal center B lymphocytes. Science 278:301–5 Papavasiliou F, Casellas R, Suh H, Qin XF, Besmer E, Pelanda R, Nemazee D, Rajewsky K, Nussenzweig MC. 1997. V(D)J recombination in mature B cells: a mechanism for altering antibody responses. Science 278:298–301 Hikida M, Ohmori H. 1998. Rearrangement of lambda light chain genes in

194.

195.

196.

197.

198.

199.

200.

201.

mature B cells in vitro and in vivo. Function of reexpressed recombinationactivating gene (RAG) products. J. Exp. Med. 187:795–99 Han S, Zheng B, Schatz DG, Spanopoulou E, Kelsoe G. 1996. Neoteny in lymphocytes: Rag1 and Rag2 expression in germinal center B cells. Science 274:2094–97 Hikida M, Mori M, Kawabata T, Takai T, Ohmori H. 1997. Characterization of B cells expressing recombination activating genes in germinal centers of immunized mouse lymph nodes. J. Immunol. 158:2509–12 Han S, Zheng B, Takahashi Y, Kelsoe G. 1997. Distinctive characteristics of germinal center B cells. Semin. Immunol. 9:255–60 Hikida M, Nakayama Y, Yamashita Y, Kumazawa Y, Nishikawa SI, Ohmori H. 1998. Expression of recombination activating genes in germinal center B cells: involvement of interleukin 7 (IL-7) and the IL-7 receptor. J. Exp. Med. 188:365– 72 Hertz M, Kouskoff V, Nakamura T, Nemazee D. 1998. V(D)J recombinase induction in splenic B lymphocytes is inhibited by antigen-receptor signalling. Nature 394:292–95 Meffre E, Papavasiliou F, Cohen P, de Bouteiller O, Bell D, Karasuyama, Schiff C, Banchereau J, Liu YJ, Nussenzweig MC. 1998. Antigen receptor engagement turns off the V(D)J recombination machinery in human tonsil B cells. J. Exp. Med. 188:765–72 de Wildt R, Hoet RMA, van Venrooij WJ, Tomlinson IM, Winter G. 1999. Analysis of heavy and light chain pairings indicates that receptor editing shapes the human antibody repertoire. J. Mol. Biol. 285:895–901 Qin XF, Schwers S, Yu W, Papavasiliou F, Suh HY, Nussenzweig A, Rajewsky K, Nussenzweig MC. 1999. Secondary

RECEPTOR EDITING


V(D)J recombination in B-1 cells. Nature 397:355–59 202. Wang F, Huang CY, Kanagawa O. 1998. Rapid deletion of rearranged T cell antigen receptor (TCR) Valpha-Jalpha segment by secondary rearrangement in the thymus: role of continuous rearrangement of TCR alpha chain gene and positive selection in the T cell repertoire

51

formation. Proc. Natl. Acad. Sci. USA 95:11834–39 203. McMahan CJ, Fink PJ. 1998. RAG reexpression and DNA recombination at T cell receptor loci in peripheral CD4` T cells. Immunity 9:637–47 204. Honjo T, Alt FW, eds. 1995. Immunoglobulin Genes. San Diego: Harcourt Brace. 2nd ed.



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


MOLECULAR BASIS OF CELIAC DISEASE Ludvig M. Sollid Institute of Immunology, Rikshospitalet, University of Oslo, N-0027 Oslo, Norway; e-mail: [email protected]


Key Words celiac disease, gluten, T cells, HLA, DQ2, peptides Abstract Celiac disease (CD) is an intestinal disorder with multifactorial etiology. HLA and non-HLA genes together with gluten and possibly additional environmental factors are involved in disease development. Evidence suggests that CD4` T cells are central in controlling an immune response to gluten that causes the immunopathology, but the actual mechanisms responsible for the tissue damage are as yet only partly characterized. CD provides a good model for HLA-associated diseases, and insight into the mechanism of this disease may well shed light on oral tolerance in humans. The primary HLA association in the majority of CD patients is with DQ2 and in the minority of patients with DQ8. Gluten-reactive T cells can be isolated from small intestinal biopsies of celiac patients but not of non-celiac controls. DQ2 or DQ8, but not other HLA molecules carried by patients, are the predominant restriction elements for these T cells. Lesion-derived T cells predominantly recognize deamidated gluten peptides. A number of distinct T cell epitopes within gluten exist. DQ2 and DQ8 bind the epitopes so that the glutamic acid residues created by deamidation are accommodated in pockets that have a preference for negatively charged side chains. Evidence indicates that deamidation in vivo is mediated by the enzyme tissue transglutaminase (tTG). Notably, tTG can also cross-link glutamine residues of peptides to lysine residues in other proteins including tTG itself. This may result in the formation of complexes of gluten-tTG. These complexes may permit gluten-reactive T cells to provide help to tTG-specific B cells by a mechanism of intramolecular help, thereby explaining the occurrence of gluten-dependent tTG autoantibodies that is a characteristic feature of active CD.

INTRODUCTION Celiac disease (CD), or gluten sensitive enteropathy, is a condition in which ingested wheat gluten or related proteins from rye and barley are not tolerated (1). CD, like type 1 diabetes, rheumatoid arthritis, and multiple sclerosis, has a chronic nature where particular HLA alleles are overrepresented among the patients (2). Commonly these disorders are multifactorial; HLA genes and other genes together with environmental factors are involved in disease development. The expression of CD is strictly dependent on dietary exposure to gluten and

0732–0582/00/0410–0053$14.00

53


54

SOLLID

similar cereal proteins (1). Patients go into complete remission when they are put on a gluten-free diet, and they relapse when gluten is reintroduced into the diet. CD is in this respect unique among the chronic inflammatory HLA- associated diseases in that a critical environmental factor has been identified. CD is primarily a disease of caucasians (1). It is most frequently recognized among Europeans, although there is an increasing awareness of this disorder in the United States. CD commonly presents in early childhood with classic symptoms including chronic diarrhea, abdominal distension, and failure to thrive (3). The general condition of these children is severely impaired. The disease may also present later in life with symptoms that tend to be more vague and include anemia, fatigue, weight loss, diarrhea, constipation, and neurological symptoms (4). CD patients on a gluten-containing diet have increased levels of serum antibodies to a variety of antigens, including gluten and the autoantigen tissue transglutaminase (tTG) (5, 6). The presence of antibodies to gluten and tTG is strictly dependent on dietary exposure to gluten. Testing of serum antibodies to gluten and tissue tTG is utilized to predict CD, and this provides a great aid in clinical practice (5, 7, 8). The final diagnosis of CD, nevertheless, rests on the demonstration of typical mucosal pathology by histological examination of small intestinal biopsies. The reported prevalence of disease with overt symptoms varies enormously in the populations of Europe and North America. Assessment of the prevalence by biopsy examination of individuals identified by antibody screening has however demonstrated surprisingly similar prevalence rates of about 1:200 to 1:400 throughout Europe and North America (9). Many of the patients identified in these studies have no symptoms or only mild symptoms that are often associated with decreased psychophysical well-being and anemia (4, 9). The clinical expression of CD is probably influenced by environmental factors. In Sweden an ‘‘epidemic’’ of CD in children under the age of two years produced a dramatic fourfold increase in incidence rates in the period 1985–1987 and a similar rapid decline in the incidence rates from 1995–1997 (10). These changes in incidence concur with changes in infant feeding practices and suggest that the amount and timing of the gluten introduction (perhaps in conjuction with the breast feeding duration) is important for precipitation of the disease in children (10). Whether the pattern of gluten feeding in infants affects only the age of onset of the disease or whether it ultimately changes the overall population prevalence is still an open question. Current treatment of CD is a lifelong exclusion of gluten from the diet. Poor diet compliance by patients and undiscovered disease are associated with complications including increased risk of anemia, infertility, osteoporosis, and intestinal lymphoma (4). Notably, untreated CD is associated with increased mortality. Research into the molecular basis of the disease has already lead to improved diagnosis, and it is hoped this research will lead to better treatment in the future.

MOLECULAR BASIS OF CELIAC DISEASE

55


THE CELIAC LESION The lesion in CD is localized in the proximal part of the small intestine. Villous atrophy, crypt cell hyperplasia, lymphocytic infiltration of the epithelium, and increased density of various leukocytes in the lamina propria characterize the classic textbook type of lesion (1). These alterations represent one end of a spectrum of mucosal pathology that Marsh (11) has classified into three stages: the infiltrative, the hyperplastic, and the destructive lesions. The infiltrative lesion is characterized by infiltration of small nonmitotic lymphocytes in the villous epithelium without any other sign of mucosal pathology. The hyperplastic lesion is similar to the infiltrate lesion but in addition has hypertrophic crypts whose epithelium may be infiltrated by lymphocytes. The destructive lesion is synonymous to the classic lesion described in textbooks. Oral challenge experiments with gluten have demonstrated that these stages are dynamically related (12). The existence of a spectrum of pathological stages in CD is interesting when considering the polygenic nature of CD. In the NOD mouse model of autoimmune diabetes, where at least 14 different loci are involved in the control of the disease, nearly all NOD mice develop insulitis, but many animals do not go on to develop diabetes (13). Notably fewer susceptibility genes are required to produce insulitis than diabetes (13). It is conceivable that in CD different susceptibility genes contribute at different stages to the development of the end-stage disease. The pathological alterations and the type of cellular infiltrates found in the classical, flat-destructive lesion are well characterized, and the major features are summarized in the following.

Enterocytes In CD there is an increased loss of epithelial cells and increased proliferation of epithelial cells in the crypts. Both these factors have been used to explain the villus atrophy found in CD (14, 15). It is not clear whether the two phenomena are causally linked, and if so, which of them is primary or secondary. The increased epithelial cell loss probably reflects increased apoptosis of enterocytes (16), whereas the increased enterocyte proliferation appears to be due to an increased production of keratinocyte growth factor (KGF) by stromal cells (17). Several molecules with immune function are known to have an altered expression in CD. There is an increased epithelial expression of HLA class II molecules with strong expression of DR and DP molecules, but with little or no expression of DQ molecules (18, 19). The expression of the polymeric Ig receptor is also upregulated (20). Notably, this enhanced expression of the polymeric Ig receptor is accompanied by increased transport of IgA and IgM into the gut lumen (21).

Intraepithelial Lymphocytes Three major lineages of intraepithelial lymphocytes (IELs) occur in the normal human small intestine; the most prominent is the TCRab` CD8`CD41 popu-


56

SOLLID

lation, while the TCRab` CD81CD4` population and the TCRcd` CD81CD41 population are also present. Both the TCRab` CD8`CD41 and the TCRcd` CD81CD41 populations are expanded in CD. In contrast to the TCRab` CD8` IELs that return to normal when gluten is removed from the diet, the TCRcd` IELs appear to remain at an elevated level (22). However, IELs of both the TCRab` CD8` and TCRcd` lineages express the Ki67 proliferation marker, suggesting intraepithelial proliferation of both populations in CD (23). Interestingly, the majority of TCRcd` IELs express the Vd1 TCR variable region (24, 25). Spies and co-workers have demonstrated that cd T cells expressing this variable region recognize MICA and MICB molecules (26)— molecules that are mainly expressed by intestinal epithelial cells (27). Activated human IELs are able to produce a number of cytokines including IFN-c, IL-2, IL-8 and TNF-a and are known to have a lytic potential (28). Furthermore, in CD, but not in giardiasis, the IELs stain positive for granzyme B and TiA (a marker characteristic for cytotoxic lymphocytes), indicating that some IELs in the celiac lesion may be activated cytotoxic T cells (29).

Lamina Propria Leukocytes A marked infiltration of TCRab` T cells appears in the lamina propria in the active lesion. These T cells are mostly CD4` and carry a memory phenotype (CD45RO`) (30). Notably, an increased percentage of these lamina propria T cells express the CD25 (IL2R a-chain) activation marker but lack the Ki67 marker associated with proliferation (23). Thus, gluten appears to induce a nonproliferative activation of CD4` lamina propria T cells. This fits well with the results of several studies reporting increased cytokine production by T cells in the lamina propria (31–33). There seems to be a particular increase in cells producing IFNc, whereas no increase appears in cells producing IL-4 or IL-10 (33, 34). mRNA for IFN-c has been found to be increased more than 1000-fold in untreated disease related to a small increase in the message for IL-2, IL-4, IL-6, and TNF-a (33). Furthermore, the IFN-c mRNA level of biopsies of treated patients has been demonstrated to reach that of untreated patients by in vitro stimulation with gluten (33). Altogether, these results are consistent with the conception that glutenreactive T cells in the lamina propria have a cytokine profile dominated by production of IFN-c. A characteristic of the CD lesion is an accumulation of IgA-, IgM-, and IgGproducing plasma cells (35). The specificities of the antibodies produced by these cells have been only partly characterized; however, in vitro culture of biopsies has demonstrated that antibodies to gliadin (36) and endomysium (i.e. tTG) (37) are produced. Just beneath the epithelium in the normal mucosa a high number of macrophage/dendritic-like cells stain positive for CD68 (38). It is conceivable that these cells are involved in sampling of luminal antigens. The expression of the HLA


57

class II, ICAM-1, and CD25 molecules is increased in these macrophage/ dendritic-like cells, suggesting that they are activated in the disease state (18, 23, 39).

The Extracellular Matrix


In the normal small intestine extracellular matrix formation (ECM) by stromal cells balances ECM degradation mediated by matrix metalloproteinases (MMPs). Increased ECM degradation has been suggested to play a role in the villous atrophy of CD. This is supported by the demonstration of a decreased ratio of cells expressing collagen I and tissue inhibitor of metalloproteinases (TIMP)-1 mRNA to those expressing matrix metalloproteinase (MMP)-1 and -3 mRNA in untreated CD (40). Expression of MMP-1 and MMP-3 mRNA is mainly localized to subepithelial fibroblasts and macrophages. It is likely that the increased expression of metalloproteinases is related to activation of mucosal T cells (see later).

THE GENETICS OF CELIAC DISEASE A high prevalence rate (10%) among first degree relatives of CD patients indicates a strong genetic influence on susceptibility to develop CD (41). Familial clustering can be expressed as the ratio of the prevalence in relatives of affected individuals over the prevalence within the population as a whole (42). The ratio ks based on the sibpair risk is the most commonly used. If this ratio is close to 1, then there is no evidence for genetic factors in susceptibility. In contrast, the ks value for CD is estimated to be 30–60 (42, 43), which is high compared with other multifactorial disorders like rheumatoid arthritis, type 1 diabetes, and multiple sclerosis. The strong genetic influence in CD is further supported by a high concordance rate of 70% in monozygotic twins (44). The sibship aggregation attributable to HLA (ks HLA) is estimated to be 2.3–5.5 (42, 43). Using these estimates and assuming a multiplicative model of disease predisposing genes, the overall importance of non-HLA genes has been calculated to be greater than that of HLA genes (42, 43). However, attempts to map predisposing genes by linkage analysis have, with the exception of the HLA, failed to reveal unambiguous candidate genes or chromosomal regions (45–48). This suggests that each of the yetunmapped predisposing CD genes has only a minor genetic influence. Indications for susceptibility regions at 5qter and 11qter are weak (47). As with other polygenic inflammatory diseases, little is known about the non-HLA susceptibility genes. Conceivably, however, the gene products of many of these genes have immune-related functions. In the case of CD, the HLA genes (see later) and the non-HLA genes shape the immune response to gluten so that immunopathology is produced in the small intestine. Relevant to this are the recent reports that the CTLA-4/CD28 gene region contains a CD susceptibility gene (49, 50), although this finding is not consistent in all populations (51).

58

SOLLID


HLA GENES IN CELIAC DISEASE CD was first found to be associated with the HLA class I molecule B8 (52, 53). Later stronger associations were found to the HLA class II molecules DR3 and DQ2 (54–56). The genes encoding DR3 and DQ2 are in strong linkage disequilibrium, and DR3 and DQ2 are both contained within the B8-DR3-DQ2 or the B18-DR3-DQ2 extended haplotypes. The B8-DR3-DQ2 and the B18-DR3-DQ2 haplotypes are both associated with CD (57, 58). This is significant, as these two haplotypes are conspicuously dissimilar in the regions outside the DR-DQ region. CD is also associated with DR7 (59, 60), but this association is seen almost only when DR7 occurs together with DR3 or DR5 (61, 62). This is unlike the susceptibility associated with DR3, which is seen irrespective of the accompanying DR allele. In studies to date, most CD patients have been shown to carry either the DR3-DQ2 haplotype or are DR5-DQ7/DR7-DQ2 heterozygous. Evidently, CD patients with these DR-DQ combinations share the genetic information conferring CD susceptibility (63) (see Figure 1, bottom part). The DQA1*0501 and DQB1*0201 alleles of the DR3-DQ2 haplotype (64, 65) are also found when

Figure 1 Patients with CD who are DR3 or DR5/DR7 heterozygous express the same HLA-DQ2 molecule, HLA-DQ(a1*0501, b1*02). The DQA1*0501 and DQB1*02 genes are located in cis (on the same chromosome) in DR3 individuals, whereas they are located in trans (on opposite chromosomes) in DR5/DR7 heterozygous individuals.



59

the DR5-DQ7 and DR7-DQ2 haplotypes are combined. The DR5-DQ7 haplotype carries the DQA1*0501 and DQB1*0301 alleles (66), and the DR7-DQ2 haplotype carries the DQA1*0201 and DQB1*0202 alleles (67, 68). Notably, the DQB1*0201 and DQB1*0202 alleles are identical except for the codon of residue 135 located in the membrane proximal domain of the DQb chain (69). Recombination (crossing over) seems to be an important mechanism for the generation of HLA haplotypes (70). Accumulating evidence suggests that the DR3-DQ2, DR7-DQ2, and the DR5-DQ7 haplotypes have a close evolutionary relationship. Based on microsatellite analysis, fragments of DNA flanking the DQA1 gene of the DR3-DQ2 haplotype have been identified on the DR5-DQ7 haplotype, and fragments of DNA flanking the DQB1 gene of the DR3-DQ2 haplotype have been identified on the DR7-DQ2 haplotype (71, 72). Thus, the genetic information in the DQ subregion of the DR3-DQ2 haplotype is reestablished in DR5-DQ7/ DR7-DQ2 heterozygotes, although the sequence information is split between two chromosomes. It can be argued that susceptibility for CD depends on an interaction between at least two genes on the DR3-DQ2 haplotype that are reunited in DR5-DQ7/ DR7-DQ2 heterozygous individuals. Theoretically this gene interaction could involve any HLA-linked genes in the DQ region. However, complete sequencing of an 86-kb genomic fragment spanning the DQ subregion of the DR3-DQ2 haplotype failed to identify genes other than the DQA1 and the DQB1 genes in this region (73). Furthermore, the DQA1 and DQB1 are very good candidates because their products interact by forming a class II heterodimer and because they are situated close to the putative recombination site. This evolutionary consideration together with the fact that most CD patients share a particular pair of DQA1 and DQB1 genes located either in cis or in trans are strong arguments that the DQA1*0501 and DQB1*0201 alleles jointly confer susceptibility to CD by coding for the DQ(a1*0501, b1*02) heterodimer (Figure 1, top part). In most populations studied, 90% or more of the CD patients carry the DQ(a1*0501, b1*02) heterodimer, compared to 20%–30% in healthy controls (63). The fraction of patients in different populations that encode this DQ heterodimer by genes in cis or in trans position depends on the haplotype frequencies of DR3-DQ2, DR5-DQ7, and DR7-DQ2 haplotypes in the given populations (74). In a few patients the DQ(a1*0501, b1*02) heterodimer may be found to be encoded in cis position by haplotypes other than DR3-DQ2 or in trans position by individuals being heterozygous for combinations other than DR5-DQ7/DR7DQ2 (63). There is no increase of the DQ(a1*0501, b1*0301) or DQ(a1*0201, b1*02) heterodimers alone in CD demonstrating that susceptibility is dependent on both the DQa and DQb chains in the DQ(a1*0501, b1*02) heterodimer. Many studies have reported a particular increased risk for CD among individuals who are DR3-DQ2 homozygous and DR3-DQ2/DR7-DQ2 heterozygous (for references, see 63). This could be explained by a gene dosage effect of the DQB1*02 allele possibly caused by an increased expression of the DQ(a1*0501, b1*0201) heterodimer in such individuals (75). A gene dosage effect of DQB1*02


60

SOLLID

could also provide an explanation of the high degree of HLA haplotype identity observed among affected siblings (74, 76). Depending on the populations studied, about 2%–10% of CD patients do not carry the DQ(a1*0501, b1*02) heterodimer. The great majority of these patients carry different subtypes of DR4. The genetic determinant responsible for the HLA association in these individuals is likely to be different from that of the DQ(a1*0501, b1*02)-expressing individuals. To unequivocally identify the responsible molecule encoded by the DR4 haplotype by a genetic approach is, however, difficult. Notably, there is a clear skewing in the representation of the DR4-DQ8 vs the DR4-DQ7 haplotype among these patients (77–79). This implies that DQ8, i.e. DQ(a1*0301, b1*0302), is most probably the molecule responsible for susceptibility. An opposing view is that the susceptibility is mediated by the DR53, i.e. DR(a*, b4*0101), molecule that is carried on most of the DR4, DR7, and DR9 haplotypes (80). The majority of DQ(a1*0501, b1*02)-negative patients would fit into this category. Importantly, however, this model does not account for the observed skewing of the DR4-DQ8 vs. the DR4-DQ7 haplotypes. Moreover, DQ(a1*0501, b1*02)-negative CD patients who carry the DRB1*0701DQB1*03032 haplotype exist (79), and this haplotype is reported to carry a non-expressed null allele at the DRB4 locus (81). Further studies including typing for the DRB4 null allele are needed to clarify the role of DR53 as a susceptibility molecule in CD. Genes located in the HLA gene complex other than DQ might also contribute to CD susceptibility. Associations to particular DP alleles have been reported in different populations, but many of these associations can be explained by linkage disequilibrium between the involved DP allele(s) and the DQA1*0501 and DQB1*02 alleles (for further discussion, see 63). Moreover, no independent associations to alleles at the TAP1 and TAP2 loci have been found (82–84). Several studies have consistently indicated that DQA1*0501/DQB1*02-positive individuals carrying the DR5/DR7 genotype have a higher risk to develop disease than do those of the DR3/DRX genotype (X ? DR7 and DR3) (84–86). Furthermore, it has been indicated that the risk of the DR3/DR7 genotype is higher than that of the DR3/DR3 genotype (84, 86), although this is not a consistent finding (75, 87). This has led to the suggestion that a gene on the DR7-DQ2 haplotype confers an additive effect to that of the DQA1*0501/DQB1*02 genes (86). To note, a locus with a protective allele of the DR3-DQ2 haplotype would produce the same effect. Studies of Irish CD patients have indicated an additional predisposing role of TNF genes, an association independent of DQ2 that has been demonstrated using a microsatellite polymorphism situated near the TNF genes (88). Moreover, a polymorphism of the TNF-a gene promoter has been demonstrated to be a component of the DR3-DQ2 haplotype (89). A Finnish study failed to reproduce the finding of a DQ2-independent association of the TNF microsattelites (90). These discrepant results may relate to population differences. Recently, an allele of a locus (D6S2223) that is located 2, 5 Mb telomeric to the HLA-F locus was found by Lie et al (91) to be less frequent among DR3-



61

DQ2 homozygous CD patients compared to DR3-DQ2 homozygous controls. The same allele of the D6S2223 locus was also found to be underrepresented among DR3-DQ2 homozygous type 1 diabetes patients, and it was transmitted less often than expected from DR3-DQ2 homozygous parents to diabetic siblings (92). These findings suggest that a gene(s) in the vicinity of D6S2223 is involved in the pathogenesis of both CD and type 1 diabetes. In addition, the MIC-A and MIC-B genes are interesting candidate susceptibility genes in CD, as the MIC molecules are ligands for TCRcd T cells. The MIC genes are located near the HLA-B locus, and the MIC-A*008 (5.1) allele is in strong positive linkage disequilibrium with HLA-B8 (93, 94). This allele is particularly interesting since it bears a frameshift and a premature stop codon in exon 5 (95) that might affect the expression of the molecule. Taken together, available data strongly suggest that susceptibility to develop CD is primarily associated to two conventional peptide-presenting DQ molecules: i.e. DQ(a1*0501, b1*02) (4DQ2) or to a lesser extent DQ(a1*03, b1*0302) (4DQ8). An issue still to be clarified is whether there are additional molecules encoded by unidentified genes in the HLA gene complex that also contribute to the genetic predisposition for CD. However, any effect of these additional genes is likely to be moderate. A key question for the understanding of the molecular basis for CD is therefore to define the functional role of the DQ2 and DQ8 molecules.

PEPTIDE BINDING MOTIF OF DISEASE-ASSOCIATED DQ MOLECULES Peptides binding to DQ2 have anchor residues in the relative positions P1, P4, P6, P7, and P9 (96–100). This is the same spacing as previously found for DR molecules, suggesting that DQ2 bound peptides adopt to a conformation similar to that of peptides bound to DR molecules. The peptide-binding motif of DQ2 illustrated in Figure 2 is quite different from other class II–binding motifs that have been identified (101). Notably, the preference for negatively charged residues for the three anchor positions in the middle seems to be unique for DQ2. The binding motif of DQ8 is different from that of DQ2, but DQ8 also displays a preference for binding negatively charged residues at several positions (i.e. P1, P4, and P9) (102, 103). Hence, both the DQ2 and DQ8 molecules share a preference for negatively charged residues at some of their anchor positions. The peptide-binding motif of DQ2, i.e. DQ(a1*0501, b1*02), is different from the motifs of the closely related DQ(a1*0501, b1*0301) and DQ(a1*0201, b1*02) molecules (96, 99, 104), which do not confer susceptibility to CD (see above). The binding motif of the DQ(a1*0501, b1*0301) molecule is clearly different from that of DQ(a1*0501, b1*02) with differences at the P4, P7 and P9 pockets (96), whereas the differences between DQ(a1*0501, b1*02) and


62

SOLLID

Figure 2 Schematic depiction of the peptide binding groove of HLA-DQ2 (i.e. HLADQ(a1*0501, b1*02) with the peptide-binding motif displayed with the one letter code for the amino acids. A bound peptide and a TCR recognizing the peptide/HLA complex are also indicated. This motif description is based on a compilation of results from references 96–100.

DQ(a1*0201, b1*02) are more subtle (96, 99). The molecules have similar binding motifs with the most apparent difference being an additional anchor residue at P3 for DQ(a1*0201, b1*02) (99, 105).

PREFERENTIAL PRESENTATION OF GLUTEN-DERIVED PEPTIDES BY DISEASE-ASSOCIATED HLA MOLECULES TO INTESTINAL T CELLS The DQ2 and DQ8 molecules could confer susceptibility to CD by presenting disease-related peptides in the target organ or alternatively by shaping the T cell repertoire during T cell development in the thymus. This issue has been addressed by studies of T cells derived from the celiac lesion. Stimulation of small intestinal biopsy specimens with a peptic/tryptic digest of gluten induces rapid activation (i.e. expression of CD25, the IL-2 receptor a-chain) of the T cells in the lamina propria of CD patients, but not of non-CD control subjects (106). Gluten-reactive T cells can be isolated and propagated from intestinal biopsies of CD patients but



63

not from non-CD controls (107–109). These T cells are CD4` and use the ab TCR. Importantly, T cells isolated from biopsy specimens of patients carrying the DR3-DQ2 haplotype typically recognize gluten fragments presented by the DQ2 molecule rather than the other HLA molecules carried by the patients (107). Both DR3-DQ2-positive and DR5-DQ7/DR7-DQ2-positive antigen-presenting cells (i.e. carrying the DQA1*0501 and DQB1*02 genes in cis or in trans position) are able to present the gluten antigen to these T cells (107, 110). Likewise, T cells isolated from small intestinal biopsies of DQ2-negative, DR4-DQ8-positive patients predominantly recognize gluten-derived peptides when presented by the DQ8 molecule (111). It is notable that no DR(a, b1*01)-restricted intestinal T cells specific for gluten have been reported supporting a role of DQ8 rather than the DR(a, b1*01) molecule in conferring susceptibility to CD. Taken together, these results allude to presentation of gluten peptides in the small intestine as the mechanism by which DQ2 and DQ8 confer susceptibility to CD. A thymic effect of the same DQ molecules on the TCR repertoire selection is, however, not excluded by these results. The DQ2 and DQ8 molecules are not preferential antigen-presenting molecules in the intestinal mucosa irrespective of antigen. T cells specific for astrovirus (a common gastroenteritis virus) are predominantly DR restricted (109), which suggests that the peculiar HLA restriction pattern of the gliadin-specific T cells of the intestine must be related to the antigen. Interestingly, gluten-specific T cells can also be found in the peripheral blood (112). These T cells are restricted either by DR, DP, or DQ molecules, and they do not therefore display the the predominant DQ2 or DQ8 restriction observed for gluten-specific T cells from the intestinal mucosa (112). One explanation for this could be that the majority of gluten-specific T cells of peripheral blood recognize epitopes different from those recognized by T cells of the small intestine. Studies of lamina propria T cells in situ have, as mentioned above, indicated that gluten reactive T cells have a cytokine profile dominated by IFN-c. This notion is sustained by the characterization of gut-derived DQ2 and DQ8-restricted gluten-specific T cell clones. These T cells uniformly secrete IFN-c at high concentrations, and some produce IL-4, IL-5, IL-6, IL-10, TNF-a, or TGF-b in addition (113).

T CELL RECOGNITION OF DEAMIDATED GLUTEN PEPTIDES Wheat gluten is a mixture of numerous proteins grouped into the gliadin and glutenin fractions. These proteins serve as a source of nitrogen and carbon for the growing seedling during germination. A vast sequence heterogeneity among gliadin and glutenin proteins probably reflects that these proteins have been subjected to few structural constraints during evolution. Generally, gluten proteins


64

SOLLID

contain a large percentage of proline and glutamine residues, while many other amino acids, including glutamic and aspartic acid, are unusually scarce. Feeding experiments have demonstrated that the gliadin fraction can precipitate CD (114), whereas the role of glutenins is still inconclusive. Proteins of the gliadin fraction can be subdivided according to their sequence into the a-, c-, and x-gliadins (115). A large number of different gliadins exist within each of these gliadin families. Estimates suggest that as many as 50 to 150 different a-gliadin genes may be present in a single wheat cultivar (116). For the work of identifying peptide fragments recognized by T cells, the complexity of this antigen presents a big challenge. Initially it was difficult to reconcile the DQ2 (and DQ8) binding motifs with presentation of gluten peptides because gluten proteins have an unusual scarcity of negatively charged residues. A clue to help explain this paradox came from the observation that the stimulatory capacity of gliadin preparations for gliadinspecific intestinal T cells was significantly enhanced following treatment at high temperatures and low pH (117). These conditions are known to cause nonspecific deamidation of glutamines to glutamic acid and may thus convert gliadin from a protein with very few peptides with the potential to bind to DQ2/DQ8 into one with many such. An important and general role for deamidation of gluten for T cell recognition was sustained by analysis of the response pattern of a panel of polyclonal, gliadin-specific T cell lines derived from biopsies (118). All the lines responded poorly to a gliadin antigen prepared under conditions of minimal deamidation (chymotrypsin-digestion), compared to the same antigen when further heat-treated in an acidic environment. The characterization of gluten epitopes recognized by intestinal T cells has extended the knowledge about the importance of deamidation for their T cell recognition. So far five unique epitopes of gluten that are recognized by gut T cells have been identified; three restricted by DQ2 (118, 119) (Table 1 and Figure 3) and two restricted by DQ8 (120, 121) (Table 1). The three DQ2-restricted peptides, one from c-gliadin and two from a-gliadins (DQ2-c-gliadin-I, DQ2-agliadin-I and DQ2-a-gliadin-II), fail to stimulate T cells in their native form but are potent antigens when a single glutamine residue is exchanged with glutamic

TABLE 1 Epitope DQ2-c-I-gliadin DQ2-a-I-gliadin DQ2-a-II-gliadin DQ8-a-I-gliadin DQ8-I-glutenin

Derived from protein

Presentation element

Recognized by patients

Reference

c-gliadin a-gliadin a-gliadin a-gliadin Glutenin

DQ2 DQ2 DQ2 DQ8 DQ8

Infrequently Frequently Frequently Frequently? Not known

(118) (119) (119) (120) (121)



65

Figure 3 A cartoon showing the amino acid sequence (one letter code) and binding of the three known gluten epitopes recognized by HLA-DQ2 restricted intestinal T cells of CD patients. All three epitopes contain a glutamic acid (E) residue that has been converted from glutamine (Q) by deamidation. The glutamic acid residues formed by deamidation improve the binding affinity and are critical for T cell recognition in all three epitopes. Notably, glutamic acid residues are accommodated in the P4 pocket for the DQ2-a-IIgliadin epitope (119), in the P6 pocket for the DQ2-a-I-gliadin epitope (119), and in the P7 pocket for the DQ2- c-I-gliadin epitope (118). In some a-gliadins, the DQ2-a-I-gliadin and DQ2-a-II-gliadin epitopes are part of the same fragment, and it is the very same glutamine that is modified by tTG in both epitopes.

acid in certain positions. The recognition of one of the DQ8-restricted peptides from a-gliadins (DQ8-a-gliadin-I) is augmented by introduction of negatively charged residues (122), whereas this is not seen for another DQ8-restricted peptide of glutenin (DQ8-glutenin-I) (121). These data demonstrate that most, but not all, gluten-specific intestinal T cells from CD patients recognize gluten proteins only after they have undergone deamidation. Moreover, the results with the glutenin epitope demonstrate that intestinal T cells can recognize gluten proteins other than gliadins (121). This raises the question of whether glutenins are also able to precipitate the disease.

66

SOLLID


DEAMIDATION IN VIVO IS LIKELY TO BE ENZYMATICALLY MEDIATED BY TISSUE TRANSGLUTAMINASE The deamidation of gliadin may take place in the acidic environment in the stomach (118). Alternatively, it can be mediated by the enzyme tissue transglutaminase (tTG) as demonstrated by Molberg et al (123) and later also by van de Wal et al (122). tTG is expressed in many different tissues and organs; in the small intestine it is expressed just beneath the epithelium in the gut wall (123). Notably the activity of tTG is elevated in the small intestinal mucosa of CD patients in both the active disease phase and in remission (123a). The enzyme is present both intracellularly and extracellulary, and in the extracellular environment tTG plays a role in extracellular matrix assembly, cell adhesion, and wound healing (124). The calcium-dependent transglutaminase activity of tTG catalyzes selective crosslinking or deamidation of protein-bound glutamine residues (125). Notably, tTG is the same protein that Dieterich et al found to be a major focus of the autoantibody response in CD (6). In contrast to the nonenzymatically mediated deamidation that results in a near random deamidation of the often numerous glutamine residues in gliadin peptides, tTG appears to carry out an ordered deamidation of some few specific glutamines (123). For all the three DQ2-restricted gliadin epitopes recognized by gut T cells and the DQ8-a-gliadin-I epitope, the residues critical for T cell recognition are all specifically targeted by tTG (119, 122, 123). Interestingly, the deamidation of glutamines that are not targeted by tTG (e.g. by acid treatment) can be deleterious for T cell recognition (105, 122). Additional evidence for a role of tTG comes from experiments where T cell lines have been established from biopsies challenged with a minimally deamidated gliadin antigen (chymotrypsin-digested) and then tested for recognition of this antigen or the same antigen treated with tTG (Ø Molberg, S McAdam, KEA Lundin, C Kristiansen, K Kett, EH Arentz-Hansen, LM Sollid, manuscript in preparation). In 14 out of 15 patients, the T cell lines responded better to the antigen that had been subjected to treatment with tTG. Similarly, T cell lines established from two DQ2` patients by stimulating biopsies with a chymotrypsin-digested recombinant a-gliadin were found to recognize synthetic peptides representing the DQ2a-gliadin-I and DQ2-a-gliadin-II epitopes, but not the corresponding nondeamidated peptides (Ø Molberg, S McAdam, KEA Lundin, C Kristiansen, EH Arentz-Hansen, K Kett, LM Sollid, manuscript in preparation). Taken together, these results indicate that deamidation in vivo is mediated by tTG. It is intriguing to hypothesize that tTG plays a central role in the selection of gliadin T cell epitopes. Credence to this idea comes from the observation that the intestinal T cell response to a-gliadin in adults is focused on a single deamidated glutamine (in the related DQ2-a-gliadin-I and DQ2-a-gliadin-II epitopes) that is targeted by tTG (119). Knowledge of the substrate recognition sites of tTG should allow further testing of this hypothesis. Unfortunately, the available information


67

on sequences targeted by tTG is not presently sufficient to establish the overall substrate specificity of the enzyme.


HIERARCHIES OF GLUTEN T CELL EPITOPES? The existence of multiple epitopes in gluten that are recognized by small intestinal T cells of CD patients raises several interesting questions: Are only some of the epitopes pathogenic and thereby relevant to explain the HLA association? Are responses toward some of the epitopes generated during the early phases of disease development, while the responses to others are a result of epitope spreading? Are different epitopes recognized by distinct groups of patients (e.g. children vs. adults)? Are some epitopes more relevant to disease as responses to them are found in the majority of the patients or because there is a higher precursor frequency of T cells in the lesion specific for these epitopes? The answers to most of these questions must await further investigations. At present we know that for the DQ2-a-gliadin-I and DQ2-a-gliadin-II epitopes, intestinal T cell reactivity is found in most if not all adult DQ2` patients (119), whereas for the DQ2-cgliadin-I epitope, intestinal T cell reactivity is found in only a minority of DQ2` patients (118). Less is known about the DQ8-restricted epitopes because few DQ8-positive patients have been tested so far. However, the DQ8-a-gliadin-I appears to be frequently recognized (120). What causes the variance in responsiveness to the different epitopes and whether this reflects qualitative or quantitative differences between the patients are presently unclear. Epitope spreading (126) may be a mechanism relevant to CD that could explain the existence of several gluten epitopes. In experimental autoimmune encephalomyelitis where epitope spreading occurs, along with spreading of new antigenic epitopes there is also a ‘‘spreading’’ of MHC class II molecules involved in epitope presentation (127, 128). The strict restriction of DQ2 and DQ8 as presentation elements for gluten-reactive T cells of the disease lesion clearly deviates from the picture found in experimental autoimmune encephalomyelitis and may suggest that other mechanisms are operating. Further studies are clearly needed to sort out this question. The mapping of epitopes of gluten proteins recognized by intestinal T cells is incomplete; the actual number of distinct epitopes is currently a matter of speculation. However, recent results from testing intestinal T cells of Norwegian adults against a panel of recombinant a-gliadins suggest that the number of epitopes might be more limited than initially thought (119). From the sequences represented in a panel of full-length recombinant a-gliadins, there seems to be only a single immunodominant fragment that contains the two related epitopes DQ2-agliadin-I and DQ2-a-gliadin-II. The disease relevance of epitopes defined using peripheral blood T cells must also be questioned because peripheral blood T cell gliadin epitopes do not appear to be limited in their presentation by DQ2 or DQ8 (112) nor to be enhanced by

68

SOLLID

treatment with tTG (123). Furthermore, a DQ2-restricted epitope of a-gliadin, which was defined by peripheral blood T cells of a CD patient (129) and which induces mucosal changes in peptide feeding experiments (130), fails to be recognized by gluten-reactive polyclonal intestinal T cell lines from six patients even after tTG treatment (119).


GLUTEN-SPECIFIC T CELLS MAY PROVIDE HELP FOR AUTOANTIBODY PRODUCTION The IgG and IgA serum antibodies to tTG (also termed anti-endomysial antibodies) are a hallmark of CD, and detection of serum IgA tTG-antibodies is utilized to predict the disease (7, 8). As the B cells producing the tTG antibodies have undergone an isotype switch, it is likely that these are T cell-dependent antibody responses. This poses a problem since the existence of T cells recognizing tTG is doubtful. tTG is expressed ubiquitously in the human body, and staining with sera from untreated CD patients indicates that the antigen is also expressed in fetal thymus (131). Most likely, T cells reactive with tTG are therefore deleted by negative selection in thymus, and if they should exist they would likely have induced serious systemic autoimmunity. Interestingly, gluten seems to drive the antibody production, as the presence of tTG antibodies is strictly dependent on dietary gluten exposure (8). This raises the possibility that gluten-reactive T cells provide help for tTG-specific B cells by a mechanism of intramolecular help (132) analogous to the hapten-carrier system (133). As mentioned earlier, an important physiological role of tTG is the catalysis of isopeptide bond formation between glutamine and lysine residues (125). Indeed, it is the substitution of water rather than lysine in this reaction that results in deamidation. In vitro treatment of gliadin fragments with tTG leads to some gliadin fragments becoming covalently attached to tTG by autocatalysis (6, 123). tTG-specific B cells may selectively bind and internalize gliadin-tTG complexes via specific surface immunoglobulins. The gliadin fragment may finally be processed and presented by DQ2 or DQ8 to the gliadin-specific T cells, thereby providing cognate help for B cell maturation, isotype switching, and antibody secretion. This model can explain why tTG antibody levels in CD are dependent on the presence of gliadin in the diet because its removal will also abolish the T cell help needed for antibody production. Autocatalysis by tTG should be more likely to occur when the concentration of other amine donors (lysine containing proteins/primary amines) is low. In fact, deamidation is also likely to happen when the amount of primary amines is low or absent (125). Formation of gliadin-tTG complexes and deamidation of gliadin may thus reflect an altered microenvironment in the gut mucosa. The unusual ability of gliadins to act as excellent amine acceptor substrates for tTG may result in a local depletion of lysine/polyamines, and the altered microenvironment may hence be established in situations where increased levels of gluten proteins get access to the subepithelial area.


69


PERTURBED ORAL TOLERANCE TO GLUTEN IN CELIAC DISEASE? Although the concept of oral tolerance is not as firmly established in humans as it is in rodents, it is clearly necessary that mechanisms that allow for tolerance to soluble food antigens exist in humans (134). In keeping with this thinking, oral tolerance to gluten in patients with CD either is not established properly or is broken. A deeper understanding of this issue should shed new light on the mechanism behind oral tolerance in humans. Given the preferential intestinal T cell response to deamidated gluten fragments in CD patients, it is conceivable that deamidation is central to the perturbation of the oral tolerance. Deamidation increases the binding affinity of gliadin peptides for DQ2 from poor but significant binders to epitopes with reasonable, but by no means exceptional, affinity (118, 119). The moderate binding affinity of these epitopes concurs with the finding that they do not carry optimal anchors in all the anchor positions. It is interesting that the modified glutamine residues for the three defined DQ2-restricted gliadin epitopes recognized by intestinal T cells occupy different pockets within DQ2 (Figure 3). This suggests that the altered affinity of the gliadin peptides for DQ2 is a critical factor involved in loss of tolerance rather than recognition of a single ‘‘pathogenic’’ motif that binds to DQ2 (119). Concurrent with the increase in affinity for DQ2 caused by deamidation of the gliadin peptide is a change in conformation of the gliadin/DQ2 complex. This is apparent by the failure of the T cells to recognize the unmodified peptides even at higher concentrations that should compensate for their lower affinity for DQ2. However, the simple modification of glutamine residues that act as major T cell receptor contact residues appears not to be sufficient to break tolerance as none of the modified glutamines are found in such positions (105, 119). Gliadin fragments containing two glutamine residues targeted by tTG may well be deamidated and cross-linked to other proteins that contain lysine. Conditions may exist in the gut, where T cell epitopes are both created and trapped locally by tTG, that prevent the epitopes from being presented by antigen-presenting cells that induce tolerance in the gut. Alternatively, it may prevent these epitopes from spreading systemically, a factor thought to be important in the establishment of oral tolerance (135). In this regard it is interesting that the motif targeted by tTG and shared in the DQ2-a-gliadin-I and DQ2-a-gliadin-II epitopes is repeated within many of the a-gliadins (119).

MECHANISMS INVOLVED IN FORMATION OF THE CELIAC LESION The evidence discussed above provides strong evidence that CD4` TCRab` T cells in the lamina propria are central for controlling the immune response to gluten that produces the immunopathology of CD. The knowledge of the events


70

SOLLID

downstream of T cell activation is, however, still incomplete. The characterization of mechanisms operating in the model of human fetal gut explant cultures, where activation of T cells induces villous atrophy and hyperplasia of the crypts, has provided interesting clues and indicated some major pathways (136). However, knowing how the immune system usually utilizes a multitude of effector mechanisms for fighting its opponents, it is reasonable to believe that multiple effector mechanisms may well be involved in the creation of the celiac lesion. Adding to the complexity, recent in vitro organ culture studies have indicated that gluten exerts additional immune relevant effects independent of T cell activation (137, 138). Some of these effects have rapid kinetics, and conceivably the direct effects of gluten may facilitate subsequent T cell responses. Cytokines produced by lamina propria CD4` T cells may be involved in the increased crypt cell proliferation and the increased loss of epithelial cells. IFN-c induces macrophages to produce TNF-a. TNF-a activates stromal cells to produce KGF, and KGF causes epithelial proliferation and crypt cell hyperplasia (17). IFN-c and TNF-a can jointly have a direct cytotoxic effect on intestinal epithelial cells (139). It is also conceivable that IELs and in particular cd T cells play a role in the epithelial cell destruction by recognizing MIC molecules induced by stress (26). Alterations of the extracellular matrix can also distort the epithelial arrangement, as the extracellular matrix provides the scaffold on which the epithelium lies. Enterocytes adhere to basement membrane through extracellular matrix receptors so that modification or loss of the basement membrane can result in enterocyte shedding. Evidence for increased extracellular matrix degeneration in CD exists, and this degeneration may be important for the mucosal transformation found in CD (40). The increased production of metalloproteinases by subepithelial fibroblasts and macrophages is likely to be directly or indirectly induced by cytokines that are released from activated T cells. Do the autoantibodies play a role in the pathogenesis of CD, or are they just an epiphenomenon? The significant increase in prevalence of CD among IgAdeficient individuals (1) speaks against a role of the antibodies. However, most CD patients also have elevated levels of serum IgG endomysial (i.e. tTG) antibodies (5), and little is known about the antibodies found locally in the mucosa of IgA-deficient CD patients. Interestingly, the endomysial (i.e. tTG) antibodies can, as suggested by Ma¨ki and coworkers (140), be involved in the disease development by blocking interactions between mesenchymal cells and epithelial cells during the migration of epithelial cells and fibroblasts from the crypts to the tips of the villi. tTG is necessary for activation of transforming growth factor-b (TGFb) (141). Indirect inhibition of TGF-b activation by anti-tTG antibodies can be envisaged to have broad effects as TGF-b is known to affect the differentiation of the intestinal epithelium (140), to stimulate extracellular matrix formation (142), and to regulate the function of many immune competent cells within the gut microenvironment (143). In addition, tTG has been demonstrated to be involved in attachment of fibroblasts to the extracellular matrix (144), suggesting


71

that the autoantibodies could also be involved in lesion formation by perturbing important contacts between fibroblasts and extracellular matrix components. The tTG antibodies may in addition modulate the deamidating activity of tTG in either an inhibiting or a promoting fashion (145). Further research is clearly needed to establish whether and how the tTG antibodies play a role in CD pathogenesis.


HLA ASSOCIATION WITH DISEASE: LESSONS TO BE LEARNED FROM CD Strong evidence suggests that the primary HLA association in CD is to the classical peptide presenting HLA molecules DQ2 and DQ8. These HLA molecules predispose to disease by presenting gluten peptides to CD4` T cells in the affected organ, although an effect mediated by shaping of the T cell repertoire in the thymus cannot yet be excluded. This has clear relevance for studies of other HLAassociated diseases where the identity of the HLA molecules involved are less well defined and where the triggering antigens have not been identified. The DQ2 and DQ8 molecules bind gluten peptides that after specific deamidation become good peptide ligands for DQ2 and DQ8. Exactly why no other class II molecules are able to present gluten peptides in the gut that result in disease is not yet fully understood. Likely related is that peptides that become deamidated in the gut mucosa are particularly effective in inducing a pathologic immune response, and that the DQ2 and DQ8 molecules are especially suited to bind deamidated peptides. The DQ(a1*0501, b1*0301) and DQ(a1*0201, b1*0202) molecules which are related to the predisposing DQ(a1*0501, b1*0201) molecule and which do not predispose to CD have different binding motifs, although the binding motif of DQ(a1*0201, b1*0202) is very similar. Interestingly, DQ(a1*0501, b1*0201) and DQ(a1*0201, b1*0202) expressing B lymphoblastoid cell lines exhibit abilities to present the DQ2-a-gliadin-II epitope that differ according to when the epitope is incorporated into a complex antigen that requires processing as compared with the peptide that is processing independent (105). This might suggest that factors involved in processing and peptide binding act differentially for loading of the gliadin peptides to two DQ molecules and that this is relevant for explaining the HLA association. Modification of self-proteins analogous to gliadin in CD would create epitopes recognized as nonself. This could be a more general mechanism for breaking of immunological tolerance and precipitation of autoimmune disease. Perhaps as many as 50% to 90% of the proteins in the human body are posttranslationally modified (P Roepstorff, personal communication), and the degree and type of modification are likely to be altered in an inflamed microenvironment. Epitopes harboring a posttranslational modification may go unreported, as the standard use of recombinant proteins and synthetic peptides for the characterization of T cell epitopes means that most in vivo modified epitopes would escape detection. This class of epitopes should not be overlooked, and it will be important to devise


72

SOLLID

strategies that will identify modified T cell epitopes of potential autoantigens so that their role in autoimmune disease can be clarified. Another important point illustrated from the studies of CD is how a foreign antigen drives autoantibody production. For most autoimmune diseases, autoantigens have been defined by use of the autoantibodies. It is often inferred that T cells must exist that are reactive with the autoantigen because the antibodies are of the IgG or IgA isotypes whose formation is dependent on T cell help. In some cases this assumption may turn out to be unjustified. It is in my opinion appropriate to intensify the search for unknown foreign agents that might be hosted by the human body (virus, bacteria, etc) and that are capable, after combining with a self-protein, of providing help for autoimmune responses similar to that found in CD. This review illustrates that the molecular basis of CD is complex. Given the multifactorial etiology of the disease with involvement of several genes and environmental factors, this is not unexpected. Despite the recent advances in understanding of critical steps in disease development, there is much still to be learned about the disease. Several predisposing genes are yet to be identified. Given the difficulty in defining susceptibility genes with modest effects, a combined functional and genetic approch will be required for their identification. The full understanding of multifactorial inflammatory diseases is surely a formidable challenge for scientists. Compared with the other diseases of this nature, however, CD stands out as a disease for which it should be easier to decipher both the actions of the predisposing gene products and how they interact with other gene products and environmental factors. In this situation it is justified to call for intensified research on CD as this can serve as an illuminator for the other multifactorial inflammatory diseases. ACKNOWLEDGMENTS I thank Knut E. A. Lundin, Stephen N. McAdam, Øyvind Molberg, and Frode Vartdal for critically reading the manuscript. I also thank members of the ‘‘celiac disease research group’’ at my Institute for stimulating discussions and Erik Thorsby for continous support. The author’s research is supported by grants from the Research Council of Norway and the European Commission (BMH4-CT98– 3087). Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Trier JS. 1991. Celiac sprue. N. Engl. J. Med. 325:1709–19 2. Thorsby E. 1997. Invited anniversary

review: HLA associated diseases. Hum. Immunol. 53:1–11 3. Schmitz J. 1992. Coelic disease in child-


4. 5.

6.


7.

8.

9.

10.

11.

12.

13.

hood. In Coeliac Disease, ed. MN Marsh, pp. 17–48. Oxford: Blackwell Ma¨ki M, Collin P. 1997. Coeliac disease. Lancet 349:1755–59 Ma¨ki M. 1995. The humoral immune system in coeliac disease. In Coeliac Disease, ed. PD Howdle, pp. 231–49. London: Baillie´re Tindall Dieterich W, Ehnis T, Bauer M, Donner P, Volta U, Riecken EO, Schuppan D. 1997. Identification of tissue transglutaminase as the autoantigen of celiac disease. Nat. Med. 3:797–801 Dieterich W, Laag E, Schopper H, Volta U, Ferguson A, Gillett H, Riecken EO, Schuppan D. 1998. Autoantibodies to tissue transglutaminase as predictors of celiac disease. Gastroenterology 115: 1317–21 Sulkanen S, Halttunen T, Laurila K, Kolho KL, Korponay-Szabo IR, Sarnesto A, Savilahti E, Collin P, Ma¨ki M. 1998. Tissue transglutaminase autoantibody enzyme-linked immunosorbent assay in detecting celiac disease. Gastroenterology 115:1322–28 Parnell N, Ciclitira PJ. 1999. Celiac disease. Curr. Opin. Gastroenterol. 15:120– 24 ˚ , Hernell O, Ivarsson A, Persson LA Ascher H, Cavell B, Danielsson L, Dannaeus A, Lindberg T, Lindquist B, Stenhammar L. 1999. The ‘‘epidemic’’ of coeliac disease in Swedish children. Acta Paediatr. In press Marsh MN. 1992. Mucosal pathology in gluten sensitivity. In Coeliac Disease, ed. MN Marsh, pp. 136–91. Oxford: Blackwell Leigh RJ, Marsh MN, Crowe P, Kelly C, Garner V, Gordon D. 1985. Studies of intestinal lymphoid tissue. IX. Dosedependent, gluten-induced lymphoid infiltration of coeliac jejunal epithelium. Scand. J. Gastroenterol. 20:715–19 Wicker LS, Todd JA, Peterson LB. 1995. Genetic control of autoimmune diabetes

14. 15.

16.

17.

18.

19.

20.

21.

22.

73

in the NOD mouse. Annu. Rev. Immunol. 13:179–200 Booth CC. 1970. Enterocyte in coeliac disease. Br. Med. J. 3:725–31 Walker-Smith J, MacDonald T. 1989. Insights provided by the study of the small intestine in the child and the foetus. Gut 30(Spec. No):11–16 Moss SF, Attia L, Scholes JV, Walters JR, Holt PR. 1996. Increased small intestinal apoptosis in coeliac disease. Gut 39:811–17 Bajaj-Elliott M, Poulsom R, Pender SL, Wathen NC, MacDonald TT. 1998. Interactions between stromal cell-derived keratinocyte growth factor and epithelial transforming growth factor in immunemediated crypt cell hyperplasia. J. Clin. Invest. 102:1473–80 Scott H, Sollid LM, Fausa O, Brandtzaeg P, Thorsby E. 1987. Expression of major histocompatibility complex class II subregion products by jejunal epithelium in patients with coeliac disease. Scand. J. Immunol. 26:563–71 Marley NJ, Macartney JC, Ciclitira PJ. 1987. HLA-DR, DP and DQ expression in the small intestine of patients with coeliac disease. Clin. Exp. Immunol. 70:386–93 Scott H, Brandtzaeg P, Solheim BG, Thorsby E. 1981. Relation between HLA-DR-like antigens and secretory component (SC) in jejunal epithelium of patients with coeliac disease or dermatitis herpetiformis. Clin. Exp. Immunol. 44:233–38 Colombel JF, Mascart-Lemone F, Nemeth J, Vaerman JP, Dive C, Rambaud JC. 1990. Jejunal immunoglobulin and antigliadin antibody secretion in adult coeliac disease. Gut 31:1345–49 Kutlu T, Brousse N, Rambaud C, Le Deist F, Schmitz J, Cerf-Bensussan N. 1993. Numbers of T cell receptor (TCR) ab` but not of TcR cd` intraepithelial lymphocytes correlate with the grade of

74

23.

24.


25.

26.

27.

28.

29.

30.

SOLLID villous atrophy in coeliac patients on a long term normal diet. Gut 34:208–14 Halstensen TS, Brandtzaeg P. 1993. Activated T lymphocytes in the celiac lesion: non-proliferative activation (CD25) of CD4` a/b cells in the lamina propria but proliferation (Ki-67) of a/b and c/d cells in the epithelium. Eur. J. Immunol. 23:505–10 Spencer J, Isaacson PG, Diss TC, MacDonald TT. 1989. Expression of disulfide-linked and non-disulfide-linked forms of the T cell receptor c/d heterodimer in human intestinal intraepithelial lymphocytes. Eur. J. Immunol. 19:1335– 38 Halstensen TS, Scott H, Brandtzaeg P. 1989. Intraepithelial T cells of the TcRcd` CD8- and Vd1/Jd1` phenotypes are increased in coeliac disease. Scand. J. Immunol. 30:665–72 Groh V, Steinle A, Bauer S, Spies T. 1998. Recognition of stress-induced MHC molecules by intestinal epithelial cd T cells. Science 279:1737–40 Groh V, Bahram S, Bauer S, Herman A, Beauchamp M, Spies T. 1996. Cell stress-regulated human major histocompatibility complex class I gene expressed in gastrointestinal epithelium. Proc. Natl. Acad. Sci. USA 93:12445–50 Lundqvist C, Melgar S, Yeung MM, Hammarstro¨m S, Hammarstro¨m ML. 1996. Intraepithelial lymphocytes in human gut have lytic potential and a cytokine profile that suggest T helper 1 and cytotoxic functions. J. Immunol. 157:1926–34 Oberhuber G, Vogelsang H, Stolte M, Muthenthaler S, Kummer AJ, Radaszkiewicz T. 1996. Evidence that intestinal intraepithelial lymphocytes are activated cytotoxic T cells in celiac disease but not in giardiasis. Am. J. Pathol. 148:1351–57 Halstensen TS, Farstad IN, Scott H, Fausa O, Brandtzaeg P. 1990. Intraepithelial TcRa/b` lymphocytes express CD45RO more often than the TcRc/d`

31.

32.

33.

34.

35.

36.

37.

38.

counterparts in coeliac disease. Immunology 71:460–66 Kontakou M, Sturgess RP, Przemioslo RT, Limb GA, Nelufer JM, Ciclitira PJ. 1994. Detection of interferon-c mRNA in the mucosa of patients with coeliac disease by in situ hybridisation. Gut 35:1037–41 Kontakou M, Przemioslo RT, Sturgess RP, Limb AG, Ciclitira PJ. 1995. Expression of tumour necrosis factor-a, interleukin-6, and interleukin-2 mRNA in the jejunum of patients with coeliac disease. Scand. J. Gastroenterol. 30:456–63 Nilsen EM, Jahnsen FL, Lundin KEA, Johansen FE, Fausa O, Sollid LM, Jahnsen J, Scott H, Brandtzaeg P. 1998. Gluten induces an intestinal cytokine response strongly dominated by interferon-c in patients with celiac disease. Gastroenterology 115:551–63 Beckett CG, Dell’Olio D, Kontakou M, Przemioslo RT, Rosen-Bronson S, Ciclitira PJ. 1996. Analysis of interleukin-4 and interleukin-10 and their association with the lymphocytic infiltrate in the small intestine of patients with coeliac disease. Gut 39:818–23 Baklien K, Fausa O, Thune PO, Gjone E. 1977. Immunoglobulins in jejunal mucosa and serum from patients with dermatitis herpetiformis. Scand. J. Gastroenterol. 12:161–68 Falchuk ZM, Strober W. 1974. Glutensensitive enteropathy: synthesis of antigliadin antibody in vitro. Gut 15:947–52 Picarelli A, Maiuri L, Frate A, Greco M, Auricchio S, Londei M. 1996. Production of antiendomysial antibodies after in-vitro gliadin challenge of small intestine biopsy samples from patients with coeliac disease. Lancet 348:1065–67 Nagashima R, Maeda K, Imai Y, Takahashi T. 1996. Lamina propria macrophages in the human gastrointestinal mucosa: their distribution, immunohistological phenotype, and function. J. Histochem. Cytochem. 44:721–31


MOLECULAR BASIS OF CELIAC DISEASE 39. Sturgess RP, Macartney JC, Makgoba MW, Hung CH, Haskard DO, Ciclitira PJ. 1990. Differential upregulation of intercellular adhesion molecule-1 in coeliac disease. Clin. Exp. Immunol. 82:489–92 40. Daum S, Bauer U, Foss HD, Schuppan D, Stein H, Riecken EO, Ullrich R. 1999. Increased expression of mRNA for matrix metalloproteinases-1 and -3 and tissue inhibitor of metalloproteinases-1 in intestinal biopsy specimens from patients with coeliac disease. Gut 44:17– 25 41. Ellis A. 1981. Coeliac disease: previous family studies. In The Genetics of Coeliac Disease, ed. RB McConnell, pp. 197–99. Lancaster: MTP 42. Risch N. 1987. Assessing the role of HLA-linked and unlinked determinants of disease. Am. J. Hum. Genet. 40:1–14 43. Petronzelli F, Bonamico M, Ferrante P, Grillo R, Mora B, Mariani P, Apollonio I, Gemme G, Mazzilli MC. 1997. Genetic contribution of the HLA region to the familial clustering of coeliac disease. Ann. Hum. Genet. 61:307–17 44. Polanco I, Biemond I, van Leeuwen A, Schreuder I, Meera Khan P, Guerrero J, D’Amaro J, Vazquez C, van Rood JJ, Penã AS. 1981. Gluten sensitive enteropathy in Spain: genetic and environmental factors. In The Genetics of Coeliac Disease, ed. RB McConnell, pp. 211–31. Lancaster: MTP 45. Zhong F, McCombs CC, Olson JM, Elston RC, Stevens FM, McCarthy CF, Michalski JP. 1996. An autosomal screen for genes that predispose to celiac disease in the western counties of Ireland. Nat. Genet. 14:329–33 46. Houlston RS, Tomlinson IP, Ford D, Seal S, Marossy AM, Ferguson A, Holmes GK, Hosie KB, Howdle PD, Jewell DP, Godkin A, Kerr GD, Kumar P, Logan RF, Love AH, Johnston S, Marsh MN, Mitton S, O’Donoghue D, Roberts A, Walker-Smith JA, Stratton MF. 1997.

47.

48.

49.

50.

51.

52.

53.

75

Linkage analysis of candidate regions for coeliac disease genes. Hum. Mol. Genet. 6:1335–39 Greco L, Corazza G, Babron MC, Clot F, Fulchignoni-Lataud MC, Percopo S, Zavattari P, Bouguerra F, Dib C, Tosi R, Troncone R, Ventura A, Mantavoni W, Magazz, Gatti R, Lazzari R, Giunta A, Perri F, Iacono G, Cardi E, De Virgiliis S, Cataldo F, De Angelis G, Musumeci S, Ferrari R, Balli F, Bardella MT, Volta U, Catassi C, Torre G, Eliaou JF, Serre JL, Clerget-Darpoux F. 1998. Genome search in celiac disease. Am. J. Hum. Genet. 62:669–75 Brett PM, Yiannakou JY, Morris MA, Bronson SR, Mathew C, Curtis D, Ciclitira PJ. 1998. A pedigree-based linkage study of coeliac disease: failure to replicate previous positive findings. Ann. Hum. Genet. 62:25–32 Djilali-Saiah I, Schmitz J, HarfouchHammoud E, Mougenot JF, Bach JF, Caillat-Zucman S. 1998. CTLA-4 gene polymorphism is associated with predisposition to coeliac disease. Gut 43:187– 89 Holopainen P, Arvas M, Sistonen P, Collin P, Ma¨ki M, Partanen J. 1999. CD28/ CTLA4 gene region on chromosome 2q33 confers genetic susceptibility to celiac disease. A linkage and familybased association study. Tissue Antigens 53:470–75 Clot F, Fulchignoni-Lataud MC, Percopo S, Bouguerra F, Babron MC, DjilaliSaiah I, Caillat-Zucman S, ClergetDarpoux F, Greco L, Serre JL. 1999. Linkage and association study of CTLA4 region in coeliac disease for Italian and Tunisian populations. Tissue Antigens. In press Falchuk ZM, Rogentine GN, Strober W. 1972. Predominance of histocompatibility antigen HL-A8 in patients with gluten-sensitive enteropathy. J. Clin. Invest. 51:1602–5 Stokes PL, Asquith P, Holmes GK,

76

54.

55.


56.

57.

58.

59.

60.

61.

SOLLID Mackintosh P, Cooke WT. 1972. Histocompatibility antigens associated with adult coeliac disease. Lancet 2:162–64 Keuning JJ, Pena AS, van Leeuwen A, van Hooff JP, van Rood JJ. 1976. HLADW3 associated with coeliac disease. Lancet 1:506–8 Solheim BG, Ek J, Thune PO, Baklien K, Bratlie A, Rankin B, Thoresen AB, Thorsby E. 1976. HLA antigens in dermatitis herpetiformis and coeliac disease. Tissue Antigens 7:57–59 Tosi R, Vismara D, Tanigaki N, Ferrara GB, Cicimarra F, Buffolano W, Follo D, Auricchio S. 1983. Evidence that celiac disease is primarily associated with a DC locus allelic specificity. Clin. Immunol. Immunopathol. 28:395–404 Alper CA, Fleischnick E, Awdeh Z, Katz AJ, Yunis EJ. 1987. Extended major histocompatibility complex haplotypes in patients with gluten-sensitive enteropathy. J. Clin. Invest. 79:251–56 Congia M, Frau F, Lampis R, Frau R, Mele R, Cucca F, Muntoni F, Porcu S, Boi F, Contu L, et al. 1992. A high frequency of the A30, B18, DR3, DRw52, DQw2 extended haplotype in Sardinian celiac disease patients: further evidence that disease susceptibility is conferred by DQ A1*0501, B1*0201. Tissue Antigens 39:78–83 DeMarchi M, Borelli I, Olivetti E, Richiardi P, Wright P, Ansaldi N, Barbera C, Santini B. 1979. Two HLA-D and DR alleles are associated with coeliac disease. Tissue Antigens 14:309–16 Betuel H, Gebuhrer L, Descos L, Percebois H, Minaire Y, Bertrand J. 1980. Adult celiac disease associated with HLA-DRw3 and -DRw7. Tissue Antigens 15:231–38 Mearin ML, Biemond I, Pena AS, Polanco I, Vazquez C, Schreuder GT, de Vries RR, van Rood JJ. 1983. HLA-DR phenotypes in Spanish coeliac children: their contribution to the understanding of

62.

63.

64.

65.

66.

67.

68.

69.

70.

the genetics of the disease. Gut 24:532– 37 Trabace S, Giunta A, Rosso M, Marzorati D, Cascino I, Tettamanti A, Mazzilli MC, Gandini E. 1984. HLA-ABC and DR antigens in celiac disease. A study in a pediatric Italian population. Vox Sang. 46:102–6 Sollid LM, Thorsby E. 1993. HLA susceptibility genes in celiac disease: genetic mapping and role in pathogenesis. Gastroenterology 105:910–22 Boss JM, Strominger JL. 1984. Cloning and sequence analysis of the human major histocompatibility complex gene DC-3b. Proc. Natl. Acad. Sci. USA 81:5199–5203 Schenning L, Larhammar D, Bill P, Wiman K, Jonsson AK, Rask L, Peterson PA. 1984. Both a and b chains of HLADC class II histocompatibility antigens display extensive polymorphism in their amino-terminal domains. EMBO J. 3:447–52 Schiffenbauer J, Didier DK, Klearman M, Rice K, Shuman S, Tieber VL, Kittlesen DJ, Schwartz BD. 1987. Complete sequence of the HLA DQa and DQb cDNA from a DR5/DQw3 cell line. J. Immunol. 139:228–33 Chang HC, Moriuchi T, Silver J. 1983. The heavy chain of human B-cell alloantigen HLA-DS has a variable N-terminal region and a constant immunoglobulinlike region. Nature 305:813–15 Karr RW, Gregersen PK, Obata F, Goldberg D, Maccari J, Alber C, Silver J. 1986. Analysis of DRb and DQb chain cDNA clones from a DR7 haplotype. J. Immunol. 137:2886–90 Hall MA, Lanchbury JS, Lee JS, Welsh KI, Ciclitira PJ. 1993. HLA-DQ2 second-domain polymorphisms may explain increased trans-associated risk in celiac disease and dermatitis herpetiformis. Hum. Immunol. 38:284–92 Carrington M. 1999. Recombination


71.

72.


73.

74.

75.

76.

77.

78.

79.

within the human MHC. Immunol. Rev. 167:245–56 Lin L, Jin L, Kimura A, Carrington M, Mignot E. 1997. DQ microsatellite association studies in three ethnic groups. Tissue Antigens 50:507–20 Lin L, Jin L, Lin X, Voros A, Underhill P, Mignot E. 1998. Microsatellite single nucleotide polymorphisms in the HLADQ region. Tissue Antigens 52:9–18 Ellis MC, Hetisimer AH, Ruddy DA, Hansen SL, Kronmal GS, McClelland E, Quintana L, Drayna DT, Aldrich MS, Mignot E. 1997. HLA class II haplotype and sequence analysis support a role for DQ in narcolepsy. Immunogenetics 46:410–17 Sollid LM, Thorsby E. 1990. The primary association of celiac disease to a given HLA-DQ a/b heterodimer explains the divergent HLA-DR associations observed in various Caucasian populations. Tissue Antigens 36:136–37 Ploski R, Ek J, Thorsby E, Sollid LM. 1993. On the HLA-DQ(a1*0501, b1*0201)-associated susceptibility in celiac disease: a possible gene dosage effect of DQB1*0201. Tissue Antigens 41:173–77 Greenberg DA, Hodge SE, Rotter JI. 1982. Evidence for recessive and against dominant inheritance at the HLA“linked” locus in coeliac disease. Am. J. Hum. Genet. 34:263–77 Spurkland A, Sollid LM, Polanco I, Vartdal F, Thorsby E. 1992. HLA-DR and -DQ genotypes of celiac disease patients serologically typed to be non-DR3 or non-DR5/7. Hum. Immunol. 35:188–92 Tighe MR, Hall MA, Ashkenazi A, Siegler E, Lanchbury JS, Ciclitira PJ. 1993. Celiac disease among Ashkenazi Jews from Israel. A study of the HLA class II alleles and their associations with disease susceptibility. Hum. Immunol. 38:270– 76 Polvi A, Arranz E, Fernandez-Arquero M, Collin P, Ma¨ki M, Sanz A, Calvo C,

80.

81.

82.

83.

84.

85.

86.

77

Maluenda C, Westman P, de la Concha EG, Partanen J. 1998. HLA-DQ2negative celiac disease in Finland and Spain. Hum. Immunol. 59:169–75 Clot F, Gianfrani C, Babron MC, Bouguerra F, Southwood S, Kagnoff MF, Troncone R, Percopo S, Eliaou JF, Clerget-Darpoux F, Sette A, Greco L. 1999. HLA-DR53 molecules are associated with susceptibility to celiac disease and selectively bind gliadin-derived peptides. Immunogenetics 49:800–7 O’Neill CM, Bunce M, Welsh KI. 1996. Detection of the DRB4 null gene, DRB4*0101102N, by PCR-SSP and its distinction from other DRB4 genes. Tissue Antigens 47:245–48 Colonna M, Bresnahan M, Bahram S, Strominger JL, Spies T. 1992. Allelic variants of the human putative peptide transporter involved in antigen processing. Proc. Natl. Acad. Sci. USA 89:3932– 36 Powis SH, Rosenberg WM, Hall M, Mockridge I, Tonks S, Ivinson A, Ciclitira PJ, Jewell DP, Lanchbury JS, Bell JI, et al. 1993. TAP1 and TAP2 polymorphism in coeliac disease. Immunogenetics 38:345–50 Meddeb-Garnaoui A, Zeliszewski D, Mougenot JF, Djilali-Saiah I, CaillatZucman S, Dormoy A, Gaudebout C, Tongio MM, Baudon JJ, Sterkers G. 1995. Reevaluation of the relative risk for susceptibility to celiac disease of HLA-DRB1, -DQA1, -DQB1, -DPB1, and -TAP2 alleles in a French population. Hum. Immunol. 43:190–99 Mazzilli MC, Ferrante P, Mariani P, Martone E, Petronzelli F, Triglione P, Bonamico M. 1992. A study of Italian pediatric celiac disease patients confirms that the primary HLA association is to the DQ(a1*0501, b1*0201) heterodimer. Hum. Immunol. 33:133–39 Fernandez-Arquero M, Figueredo MA, Maluenda C, de la Concha EG. 1995. HLA-linked genes acting as additive sus-

78

87.


88.

89.

90.

91.

92.

93.

94.

SOLLID ceptibility factors in celiac disease. Hum. Immunol. 42:295–300 Howell WM, Leung ST, Jones DB, Nakshabendi I, Hall MA, Lanchbury JS, Ciclitira PJ, Wright DH. 1995. HLADRB, -DQA, and -DQB polymorphism in celiac disease and enteropathyassociated T-cell lymphoma. Common features and additional risk factors for malignancy. Hum. Immunol. 43:29–37 McManus R, Wilson AG, Mansfield J, Weir DG, Duff GW, Kelleher D. 1996. TNF2, a polymorphism of the tumour necrosis-alpha gene promoter, is a component of the celiac disease major histocompatibility complex haplotype. Eur. J. Immunol. 26:2113–18 McManus R, Moloney M, Borton M, Finch A, Chuan YT, Lawlor E, Weir DG, Kelleher D. 1996. Association of celiac disease with microsatellite polymorphisms close to the tumor necrosis factor genes. Hum. Immunol. 45:24–31 Polvi A, Ma¨ki M, Collin P, Partanen J. 1998. TNF microsatellite alleles a2 and b3 are not primarily associated with celiac disease in the Finnish population. Tissue Antigens 51:553–55 Lie BA, Sollid LM, Ascher H, Ek J, Akselsen HE, Rønningen KS, Thorsby E, Undlien DE. 1999. A gene telomeric of the HLA class I region is involved in predisposition to type 1 diabetes and coeliac disease. Tissue Antigens 54:162–68 Lie BA, Todd JA, Pociot F, Nerup J, Akselsen HE, Joner G, Dahl-Jorgensen K, Rønningen KS, Thorsby E, Undlien DE. 1999. The predisposition to type 1 diabetes linked to the human leukocyte antigen complex includes at least one non-class II gene. Am. J. Hum. Genet. 64:793–800 Fodil N, Pellet P, Laloux L, Hauptmann G, Theodorou I, Bahram S. 1999. MICA haplotype diversity. Immunogenetics 49:557–60 Petersdorf EW, Schuler KB, Lanton GM, Spies T, Hansen JA. 1999. Population

95.

96.

97.

98.

99.

100.

101.

102.

study of alleleic diversity of the human MHC class I-related MIC-A gene. Immunogenetics 49:605–12 Mizuki N, Ota M, Kimura M, Ohno S, Ando H, Katsuyama Y, Yamazaki M, Watanabe K, Goto K, Nakamura S, Bahram S, Inoko H. 1997. Triplet repeat polymorphism in the transmembrane region of the MICA gene: a strong association of six GCT repetitions with Behcet disease. Proc. Natl. Acad. Sci. USA 94:1298–1303 Johansen BH, Vartdal F, Eriksen JA, Thorsby E, Sollid LM. 1996. Identification of a putative motif for binding of peptides to HLA-DQ2. Int. Immunol. 8:177–82 Vartdal F, Johansen BH, Friede T, Thorpe C, Stevanovic S, Eriksen JA, Sletten K, Thorsby E, Rammensee HG, Sollid LM. 1996. The peptide binding motif of the disease associated HLA-DQ(a1*0501, b1*0201) molecule. Eur. J. Immunol. 26:2764–72 van de Wal Y, Kooy YMC, Drijfhout JW, Amons R, Koning F. 1996. Peptide binding characteristics of the coeliac diseaseassociated DQ(a1*0501, b1*0201) molecule. Immunogenetics 44:246–53 van de Wal Y, Kooy YMC, Drijfhout JW, Amons R, Papadopoulos GK, Koning F. 1997. Unique peptide binding characteristics of the disease-associated DQ(a1*0501, b1*0201) vs the non-disease-associated DQ(a1*0201, b1*0202) molecule. Immunogenetics 46:484–92 Quarsten H, Paulsen G, Johansen BH, Thorpe CJ, Holm A, Buus S, Sollid LM. 1998. The P9 pocket of HLA-DQ2 (nonAspb57) has no particular preference for negatively charged anchor residues found in other type 1 diabetes-predisposing non-Aspb57 MHC class II molecules. Int. Immunol. 10:1229–36 Rammensee HG, Friede T, Stevanovic S. 1995. MHC ligands and peptide motifs: first listing. Immunogenetics 41:178–228 Godkin A, Friede T, Davenport M, Ste-


103.


104.

105.

106.

107.

108.

109.

vanovic S, Willis A, Jewell D, Hill A, Rammensee HG. 1997. Use of eluted peptide sequence data to identify the binding characteristics of peptides to the insulin-dependent diabetes susceptibility allele HLA-DQ8 (DQ 3.2). Int. Immunol. 9:905–11 Kwok WW, Domeier ML, Raymond FC, Byers P, Nepom GT. 1996. Allelespecific motifs characterize HLA-DQ interactions with a diabetes-associated peptide derived from glutamic acid decarboxylase. J. Immunol. 156:2171–77 Khalil-Daher I, Boisgerault F, Feugeas JP, Tieng V, Toubert A, Charron D. 1998. Naturally processed peptides from HLADQ7 (a1*0501-b1*0301): influence of both a and b chain polymorphism in the HLA-DQ peptide binding specificity. Eur. J. Immunol. 28:3840–49 Quarsten H, Molberg Ø, Fugger L, McAdam SN, Sollid LM. 1999. HLA binding and T cell recognition of a tissue transglutaminase modified gliadin epitope. Eur. J. Immunol. 29:2506–14 Halstensen TS, Scott H, Fausa O, Brandtzaeg P. 1993. Gluten stimulation of coeliac mucosa in vitro induces activation (CD25) of lamina propria CD4` T cells and macrophages but no crypt-cell hyperplasia. Scand. J. Immunol. 38:581– 90 Lundin KEA, Scott H, Hansen T, Paulsen G, Halstensen TS, Fausa O, Thorsby E, Sollid LM. 1993. Gliadin-specific, HLADQ(a1*0501,b1*0201) restricted T cells isolated from the small intestinal mucosa of celiac disease patients. J. Exp. Med. 178:187–96 Molberg Ø, Kett K, Scott H, Thorsby E, Sollid LM, Lundin KEA. 1997. Gliadin specific, HLA DQ2-restricted T cells are commonly found in small intestinal biopsies from coeliac disease patients, but not from controls. Scand. J. Immunol. 46:103–9 Molberg Ø, Lundin KEA, Nilsen EM, Scott H, Kett K, Brandtzaeg P, Thorsby

110.

111.

112.

113.

114.

115.

116.

117.

79

E, Sollid LM. 1998. HLA restriction patterns of gliadin- and astrovirus-specific CD4` T cells isolated in parallel from the small intestine of celiac disease patients. Tissue Antigens 52:407–15 Lundin KEA, Gjertsen HA, Scott H, Sollid LM, Thorsby E. 1994. Function of DQ2 and DQ8 as HLA susceptibility molecules in celiac disease. Hum. Immunol. 41:24–27 Lundin KEA, Scott H, Fausa O, Thorsby E, Sollid LM. 1994. T cells from the small intestinal mucosa of a DR4, DQ7/ DR4, DQ8 celiac disease patient preferentially recognize gliadin when presented by DQ8. Hum. Immunol. 41:285–91 Gjertsen HA, Sollid LM, Ek J, Thorsby E, Lundin KEA. 1994. T cells from the peripheral blood of coeliac disease patients recognize gluten antigens when presented by HLA-DR, -DQ, or -DP molecules. Scand. J. Immunol. 39:567– 74 Nilsen EM, Lundin KEA, Krajci P, Scott H, Sollid LM, Brandtzaeg P. 1995. Gluten specific, HLA-DQ restricted T cells from coeliac mucosa produce cytokines with Th1 or Th0 profile dominated by interferon gamma. Gut 37:766–76 van de Kamer JH, Weijers HA, Dicke WK. 1953. Coeliac disease. IV. An investigation into the injurious constituents of wheat in connection with their action on patients with coeliac disease. Acta Paediatr. 42:223–31 Shewry PR, Tatham AS, Kasarda DD. 1992. Cereal proteins and coeliac disease. In Coeliac Disease, ed. M Marsh, pp. 305–48. Oxford: Blackwell Anderson OD, Litts JC, Greene FC. 1997. The a-gliadin gene family. I. Characterization of ten new wheat a-gliadin genomic clones, evidence for limited sequence conservation of flanking DNA, and Southern analysis of the gene family. Theor. Appl. Genet. 95:50–58 Lundin KEA, Sollid LM, Noreń O,

80

118.


119.

120.

121.

122.

123.

SOLLID by gut-derived T cells. Nat. Med. 4:713– Anthonsen D, Molberg Ø, Thorsby E, 17 Sjo¨stro¨m H. 1997. Heterogenous reactivity patterns of HLA-DQ-restricted small 123a. Bruce SE, Bjarnason I, Peters TJ. 1985. Human jejunal transglutaminase: demintestinal T-cell clones from patients with onstration of activity, enzyme kinetics celiac disease. Gastroenterology 112: and substrate specificity with special 752–59 relation to gliadin and coeliac disease. Sjo¨stro¨m H, Lundin KEA, Molberg Ø, Clin. Sci. 68:573–79 Ko¨rner R, McAdam SN, Anthonsen D, Quarsten H, Noreń O, Roepstorff P, 124. Aeschlimann D, Paulsson M. 1994. Transglutaminases: protein cross-linking Thorsby E, Sollid LM. 1998. Identificaenzymes in tissues and body fluids. tion of a gliadin T-cell epitope in coeliac Thromb. Haemost. 71:402–15 disease: general importance of gliadin deamidation for intestinal T-cell recog- 125. Folk JE. 1983. Mechanism and basis for specificity of transglutaminase-catalyzed nition. Scand. J. Immunol. 48:111–15 e-(c-glutamyl) lysine bond formation. Arentz-Hansen EH, Ko¨rner R, Quarsten Adv. Enzymol. Relat. Areas Mol. Biol. H, Vader W, Kooy YMC, Lundin KEA, 54:1–56 Koning F, Roepstorff P, Sollid LM, 126. Lehmann PV, Sercarz EE, Forsthuber T, McAdam S. 1999. The intestinal T cell Dayan CM, Gammon G. 1993. Deterresponse to a-gliadin in adult celiac disminant spreading and the dynamics of ease is focused on a single deamidated the autoimmune T-cell repertoire. Immuglutamine targeted by tissue transglutanol. Today 14:203–8 minase. Submitted 127. Perry LL, Barzaga ME. 1987. Kinetics van de Wal Y, Kooy YM, van Veelen PA, and specificity of T and B cell responses Penã SA, Mearin LM, Molberg Ø, Lunin relapsing experimental allergic din KEA, Sollid LM, Mutis T, Benckencephalomyelitis. J. Immunol. 138: huijsen WE, Drijfhout JW, Koning F. 1434–41 1998. Small intestinal T cells of celiac 128. McCarron RM, Fallis RJ, McFarlin DE. disease patients recognize a natural pep1990. Alterations in T cell antigen specsin fragment of gliadin. Proc. Natl. Acad. ificity and class II restriction during the Sci. USA 95:10050–54 course of chronic relapsing experimental van de Wal Y, Kooy YMC, van Veelen P, allergic encephalomyelitis. J. NeuroimVader W, August SA, Drijfhout JW, Penã munol. 29:73–79 SA, Koning F. 1999. Glutenin is 129. Gjertsen HA, Lundin KEA, Sollid LM, involved in the gluten-driven mucosal T Eriksen JA, Thorsby E. 1994. T cells reccell response. Eur. J. Immunol. In press ognize a peptide derived from a-gliadin van de Wal Y, Kooy Y, van Veelen P, presented by the celiac disease-associPenã S, Mearin L, Papadopoulos G, Konated HLA-DQ(a1*0501, b1*0201) heting F. 1998. Selective deamidation by tiserodimer. Hum. Immunol. 39:243–52 sue transglutaminase strongly enhances 130. Sturgess R, Day P, Ellis HJ, Lundin gliadin-specific T cell reactivity. J. KEA, Gjertsen HA, Kontakou M, CicliImmunol. 161:1585–88 tira PJ. 1994. Wheat peptide challenge in Molberg Ø, McAdam SN, Ko¨rner R, coeliac disease. Lancet 343:758–61 Quarsten H, Kristiansen C, Madsen L, 131. Ha¨llstro¨m O. 1989. Comparison of IgAFugger L, Scott H, Roepstorff P, Lundin class reticulin and endomysium antibodKEA, Sjo¨stro¨m H, Sollid LM. 1998. Tisies in coeliac disease and dermatitis sue transglutaminase selectively modiherpetiformis. Gut 30:1225–32 fies gliadin peptides that are recognized 132. Sollid LM, Molberg Ø, McAdam S, Lun-


133.

134. 135.


136.

137.

138.

139.

din KEA. 1997. Autoantibodies in coeliac disease: tissue transglutaminase— guilt by association? Gut 41:851–52 Lake P, Mitchison NA. 1977. Regulatory mechanisms in the immune response to cell-surface antigens. Cold Spring Harbor Symp. Quant. Biol. 41:589–95 Garside P, Mowat AM, Khoruts A. 1999. Oral tolerance in disease. Gut 44:137–42 Gu¨tgemann I, Fahrer AM, Altman JD, Davis MM, Chien YH. 1998. Induction of rapid T cell activation and tolerance by systemic presentation of an orally administered antigen. Immunity 8:667– 73 Pender SL, Tickle SP, Docherty AJ, Howie D, Wathen NC, MacDonald TT. 1997. A major role for matrix metalloproteinases in T cell injury in the gut. J. Immunol. 158:1582–90 Maiuri L, Picarelli A, Boirivant M, Coletta S, Mazzilli MC, De Vincenzi M, Londei M, Auricchio S. 1996. Definition of the initial immunologic modifications upon in vitro gliadin challenge in the small intestine of celiac patients. Gastroenterology 110:1368–78 Maiuri L, Auricchio S, Coletta S, De Marco G, Picarelli A, Di Tola M, Quaratino S, Londei M. 1998. Blockage of Tcell costimulation inhibits T-cell action in celiac disease. Gastroenterology 115:564–72 Deem RL, Shanahan F, Targan SR. 1991.

140.

141.

142.

143.

144.

145.

81

Triggered human mucosal T cells release tumour necrosis factor- alpha and interferon-gamma which kill human colonic epithelial cells. Clin. Exp. Immunol. 83:79–84 Halttunen T, Ma¨ki M. 1999. Serum immunoglobulin A from patients with celiac disease inhibits human T84 intestinal crypt epithelial cell differentiation. Gastroenterology 116:566–72 Nunes I, Gleizes PE, Metz CN, Rifkin DB. 1997. Latent transforming growth factor-b binding protein domains involved in activation and transglutaminase-dependent cross-linking of latent transforming growth factor-b. J. Cell Biol. 136:1151–63 Bonewald LF. 1999. Regulation and regulatory activities of transforming growth factor ß. Crit. Rev. Eukaryo. Gene Expr. 9:33–44 Letterio JJ, Roberts AB. 1998. Regulation of immune responses by TGF-b. Annu. Rev. Immunol. 16:137–61 Verderio E, Nicholas B, Gross S, Griffin M. 1998. Regulated expression of tissue transglutaminase in Swiss 3T3 fibroblasts: effects on the processing of fibronectin, cell attachment, and cell death. Exp. Cell Res. 239:119–38 Sollid LM, Scott H. 1998. New tool to predict celiac disease on its way to the clinics. Gastroenterology 115:1584–86



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


POPULATION BIOLOGY OF LYMPHOCYTES: The Flight for Survival Antonio A. Freitas1 and Benedita Rocha2 1

Lymphocyte Population Biology Unit, URA CNRS 1961, Institut Pasteur, 25 Rue du Dr. Roux, 75015 Paris, France; e-mail: [email protected] 2 INSERM U345, Institut Necker, 156 Rue Vaugirard, 75015 Paris, France; e-mail: [email protected]


Key Words lymphocyte survival and life spans, lymphocyte competition, lymphocyte homeostasis, naive and memory lymphocytes Abstract In this essay we suggest that the primary goal of the cells of the immune system is to ensure their own growth and survival. In adults, in steady-state conditions, the number and distribution of lymphocyte populations is under homeostatic control. New lymphocytes that are continuously produced in primary and secondary lymphoid organs must compete with resident cells for survival. We discuss recent findings supporting lymphocyte survival as a continuous active process and implicating cognate receptor engagement as fundamental survival signals for both T and B lymphocytes. The conflict of survival interests between different cell types gives rise to a pattern of interactions that mimics the behavior of complex ecological systems. In their flight for survival and in response to competition, lymphocytes use different survival signals within different ecological niches during cell differentiation. This is the case for T and B lymphocytes and also for naive and memory/activated T and B cells. We discuss how niche differentiation allows the co-existence of different cell types and guarantees both repertoire diversity and efficient immune responses.

INTRODUCTION Life started when self-replicating molecules arose (1). The first unicellular organisms emerged when these molecules developed the capacity to control their immediate surroundings to ensure their own survival and replication. With time, single-cell individuals evolved to give rise to multicellular organisms. In these complex individuals, each cell is still imprinted with the same primordial program for self-replication and survival (2). A hierarchical organization is, however, established in which the survival and the rate of replication of the different cell types are restrained and their number controlled (3). The individual cells of the immune system follow the same program, but they can survive only within the limited constraints imposed by the host. In adult mice the total number of lymphocytes remains constant and shows a ‘‘return tendency, 0732–0582/00/0410–0083$14.00

83

84

FREITAS n ROCHA

due to a density dependent process to approach a stationary distribution of population densities’’ (4), usually referred to as homeostasis. T and B lymphocytes are, however, produced continuously in either the primary lymphoid organs or by peripheral cell division; it follows that each newly formed lymphocyte can persist only if another resident lymphocyte dies. In an immune system in which the total number of cells is limited, cell survival can no longer be a passive phenomenon; it is rather a continuous and active process (5) in which each lymphocyte must compete with other lymphocytes (6). It can be said that lymphocytes follow the Red Queen Hypothesis—‘‘it takes all the running you can do to keep in the same place’’ (5, 7).

LYMPHOCYTE SURVIVAL Annu. Rev. Immunol. 2000.18:83-111. Downloaded from arjournals.annualreviews.org by HINARI on 09/01/07. For personal use only.

Naive T Cells Maintenance of naive T cells in peripheral lymphoid organs requires continuous T cell receptor (TCR) engagement by major histocompatibility complex (MHC) molecules. This was established for both CD4` and CD8` lymphocytes using various experimental approaches. In the case of naive CD8` T cells, their capacity to survive was initially studied after adoptive transfer into mice expressing different MHC class I alleles. Monoclonal naive CD8` cells from TCR transgenic (TCR-Tg) mice, restricted to MHC class I H-2Db, survived without dividing when transferred to mice lacking H-2Kb but expressing normal levels of H-2Db (8). In mice expressing lower levels of H2Db, only a fraction of the transferred cells survived—this fraction was proportional to the level of H-2Db expression. In H-2Kb`Db1 host mice the TCR-Tg cells rapidly disappeared (8). The requirements for the peripheral survival of this naive CD8` T cell clone mimic, therefore, the requirements for positive selection of the same clone in the thymus (9); these requirements include the continuous ligation of the TCR by MHC and that the number of selected cells is related to the level of MHC class I expression (10). The existing experimental evidence indicates that the results obtained with this TCR-Tg clone apply to naive CD8` T cell populations in general. Naive monoclonal CD8` T cells, expressing a TCRTg specific for the gp33–41 peptide of LCMV and restricted to H-2Db (11), also persist after transfer into CD3e-/-H-2Kb1Db` hosts but these cells disappear by one week after injection into CD3e-/-H-2Kb`Db1 hosts (N Legrand, AA Freitas, unpublished). When wild-type thymic epithelium was grafted into MHC class I– deficient mice, mature CD8` T cell differentiation in the thymus was restored, but naive CD8` T cells produced in the thymus were unable to survive at the periphery (12). Naive CD4` T cells also require allele-specific MHC class II interactions to survive (13, 14). Monoclonal TCR-Tg CD4` T cells, restricted to I-Ad, survive after transfer into I-Ad` recipients but disappear after transfer into I-Ad1Ab` recip-


LYMPHOCYTE SURVIVAL

85

ients. Persistence of non-Tg CD4` polyclonal populations also depends on interactions with MHC class II. Using retroviral vectors, class II expression was transiently induced in the thymus of class II–deficient mice, allowing cohorts of CD4` T cells to be generated and exported into class II1 peripheral pools (15). Alternatively, MHC class II` fetal thymus grafts were implanted into class II– deficient hosts (16). In both cases, permanent CD4` survival was dependent on the expression of class II at the periphery, but the exported CD4` T cells took 100 days to disappear in the absence of class II. Although these results demonstrate that permanent mature CD4` T cell engraftment requires TCR/MHC class II interactions, in the absence of these interactions polyclonal CD4` populations appeared to survive much longer than CD8` T cells. This relatively prolonged survival could be due to the presence of activated cells; as most persistent survivor cells were of an activated phenotype (16), these cells (when compared to naive T cells) could be less dependent on MHC for survival (see below). It was also likely that within a heterogeneous polyclonal T cell population, some clones could interact with MHC class I- or class II–like molecules and thus would be able to survive longer in absence of bona fide TCR/class II interactions (17). Alternatively the prolonged CD4` survival could be due to the presence of MHC class II–bearing dendritic cells (DC) exported by the class II` thymus (18). When a class II` thymus depleted of all BM-derived class I` cells was grafted into class II1 mice, the CD4` T cells generated in the thymus did not survive at the periphery (18). When the transplanted class II–deficient hosts expressed a transgene of the same MHC class II molecule on the peripheral DC, the CD4` cells exported from the thymus survived in the periphery (18). In this case CD4` T cells were shown to be in close contact with transgenic class II` DC, suggesting that this direct interaction plays a key role in CD4` survival. When these hosts expressed at the periphery a class II transgene of a different haplotype from the thymus class II, the majority of the naive CD4` T cells generated and exported from the thymus did not survive (T Brocker, personal communication). These findings indicate that polyclonal naive CD4` T cell populations require TCR/MHCrestricted interactions for peripheral survival. TCR/MHC class II interactions are also required for the expansion of CD4` T cells in T cell–deficient hosts (19). Thymus positive selection requires the simultaneous recognition of MHC and peptide molecules presented by the MHC. The role of such peptide recognition in peripheral T cell survival was also investigated by comparing the fate of mature T cells in environments of restricted peptide complexity. In H-2Ma-deficient mice, the level of expression of MHC class II is not modified, but the loading of peptides into the MHC class II molecules is impaired. Most of the I-A molecules are occupied by the invariant chain peptide CLIP (20). This limited peptide display affects thymus selection, and the peripheral T cells from these mice have a restricted repertoire diversity (20). Polyclonal CD4` T cells and TCR-Tg cells restricted to I-Ab do not survive when transferred either to H-2Ma–deficient mice, to mice expressing a different MHC class II allotype, or to MHC class II–deficient hosts (14). The similar behavior of peripheral T cells in environments of restricted


86

FREITAS n ROCHA

peptide diversity or MHC deficiency establishes that peptide recognition is essential for T cell survival. The peptides involved are probably similar to those mediating thymus positive selection. Thus, CD4` T cells from H-2Ma–deficient mice that were selected by peptides expressed by the H-2Ma–deficient thymus survived when transferred into H-2Ma–deficient peripheral pools but not when transferred into MHC class II–deficient hosts (14). CD4` T cells selected by a single MHC class II/peptide complex survived without dividing in mice expressing the same complex at the periphery (21). These results demonstrate that, like thymus positive selection, peptide recognition plays a role in peripheral T cell survival. They also suggest that the peptides involved in thymus positive selection and peripheral survival may be similar. It thus appears that naive mature T cells, following emigration from the thymus, must continuously recognize MHC/peptide complexes in order to survive in the periphery. Like thymus positive selection, this TCR ligation event (a) is MHC restricted, (b) involves recognition of similar self-peptides, (c) maintains bcl-2 levels of expression, since when TCR/MHC interactions are discontinued bcl-2 levels decrease (13), and (d) does not induce extensive cell division because the vast majority of naive peripheral T cells appear to survive as resting T cells. CD441 TCR-Tg or wild-type CD8` cells do not divide (8, 22, 23). Naive CD4` T cells either do not divide or they cycle very slowly (21, 24). In spite of these similarities, peripheral survival and thymus positive selection signals probably do not overlap. Positive selection is mainly determined by interactions between T cells and MHC/peptide complexes expressed by the thymus epithelial cells, a cell type absent at the periphery. At the periphery, a close interaction between surviving CD4` T cells with class II` DC was identified (18). It is not known if survival of naive CD8` T cells also needs interactions with particular cell types. Besides, thymus positive selection and naive T cell survival also seem to differ in lymphokine requirements, as suggested by studies on H-Y antigen-specific TCR-Tg mice, deficient for the common c chain of the IL-2 receptor (cc-mice). The cc-deficient mice have severe defects of thymus T cell differentiation. Because productive TCR rearrangements are rare, immature precursors survive poorly, most dying before TCR gene rearrangement. These abnormalities of thymus differentiation can be partially overcome when the cc-mice are crossed into an ab-TCR-Tg background (25). Although producing CD8` mature T cells in the thymus, these mice still lack peripheral CD8` T cells. These results suggest that peripheral T cell survival is not solely dependent on signals mediated by the TCR; it may also depend on signals transmitted by lymphokine receptors. The lymphokine/s required for naive T cell survival were not yet identified.

Memory T Cells Memory T cells express at the cell surface a vast array of receptors (lymphokine receptors, costimulatory and adhesion molecules) that are absent or expressed at low levels in naive cells (8, 26–28). The expression of these receptors will likely


LYMPHOCYTE SURVIVAL

87

modify the interactions of memory cells with their environment in a yet unidentified way. Changes in overall receptor expression may contribute to modify the T cell avidity for antigen and/or MHC and to lower the thresholds of survival and activation (28–33). Moreover, some of these receptors may be able to transmit activation signals and induce cell division independently of TCR engagement. The requirements for the survival and division of memory T cells are not yet completely characterized. Current studies, however, indicate that they differ from those of naive T cells, suggesting that lymphocytes are ‘‘able to tune their threshold of activation in response to recurrent signals’’ (34). Memory CD8` T cells expressing the H-Y-specific TCR-Tg, restricted to MHC H-2Db class I, survive and expand in the absence of either H-Y male antigen or the MHC-restricting element in female H-2Kb`Db1, but they disappear in mice lacking both b2microglobulin and H-2Db (presumably expressing few MHC class I molecules) (8). The same anti-H-Y TCR-Tg cells stimulated in vitro by the cognate peptide were transferred to TAP-1` Rag-deficient or TAP-11 Rag-deficient mice (35). All memory cells survived for at least 70 days in mice expressing TAP-1, while most Tg cells disappear two weeks after transfer into TAP-1-deficient recipients. To investigate if the behavior of this T cell clone can be extrapolated to CD8` memory T cell populations in general, we studied the survival of CD8`CD44` polyclonal memory cells from normal B6 mice. Since the MHC-restriction elements of this polyclonal population cannot be identified we followed their fate after transfer into syngeneic hosts lacking simultaneously b2-microglobulin, H2Db and H-2Kb MHC class I molecules. We found that one week after transfer 80% to 90% of the transferred memory T cells disappeared (B Rocha, AA Freitas, F Lemonnier, unpublished). Although long-term survival has yet to be studied, these results suggest that the vast majority of the memory CD8`CD44` polyclonal T cell populations still require some type of TCR/MHC class I interactions to survive. The requirements for the survival in vivo of memory CD4` T cells have not yet been investigated, but probably they also differ from those of naive CD4` T cells. In MHC class II–deficient mice, the rare CD4` T cells that are able to persist at the periphery bear a very activated phenotype. Cells with a similar phenotype also persist in class II and b2-microglobulin–deficient mice, which appear to recognize MHC-like molecules (17). Although memory T cell survival may require TCR-MHC-peptide interactions, in contrast to naive cells this recognition does not appear to be MHC restricted (8). It is difficult to envisage how such nonrestricted interactions would trigger memory T cells. Nonspecific signaling through the CD4/CD8 coreceptors might be involved, as Lck was found to be targeted to the CD8 coreceptor in LCMVspecific memory CD8` T cells (36). Alternatively a lower threshold of T cell activation would allow signaling by allo-MHC. It must be noted, however, that so far memory T cell survival has been studied only after transfer into irradiated hosts. In these mice, memory T cells (but not naive cells) divide extensively (37). Part of this proliferation may be due to mitogenic factors liberated after irradia-


88

FREITAS n ROCHA

tion. The crossing of MHC-deficient mice into a CD3e-/- background will allow the study of memory T cell survival in intact hosts in the absence of possible artifacts induced by irradiation. Memory T cells also differ from naive cells in the state of activation associated with survival. While naive T cells persist mainly as long-lived resting cells, memory cell survival is accompanied by cell division (self-renewal). In normal mice a fraction of memory T cells are cycling and dividing in the absence of nominal antigen stimulation (8, 23, 32, 35, 37–42). The rate of memory T cell division can be enhanced by injection in vivo of lymphokines such as IL-2, IL-12, IL-15, and a-interferon (43). During immunization the increased production of growth factors also leads to the by-stander activation of memory T cells of unrelated specificity (31, 44). Memory T cell proliferation increases in conditions of T cell depletion. When mice are sublethally irradiated, residual activated T cells expand (37). T cell expansion is evident when T lymphocytes are transferred to T cell– deficient or irradiated hosts (38, 45). This expansion is at least partially responsible for the early increase in T cell numbers observed in treated AIDS patients (46) and in irradiated adults after BM transplantation (47). In these conditions peripheral mature T cells retain a considerable expansion potential that, in the absence of intentional immunization, allows a single CD4` T cell to generate up to 1015 daughter cells upon sequential transfers into successive T cell–depleted hosts (38). The finding that memory T cells divide in the absence of their cognate antigen raises several (as yet unanswered) questions. The first concerns the existence of long-lived resting memory T cells. It is possible that no memory T cell is actually in the Go phase of the cell cycle. Memory populations are likely not constituted by cells that cycle and cells that never divide. By studying the dilution of CFSE labeling in LCMV-specific memory T cells in steady-state conditions, all these antigen-specific memory T cells were shown to divide (K Murali-Khrishna, R Ahmed, in preparation). Results of studies of BrdU accumulation by memory CD8` T cells are also compatible with long cycle times and asynchronous division (23). If all memory cells are in some type of pre-activated state, this may favor their rapid and efficient engagement in secondary responses (H VeigaFernandes, U Walter, A McLean, B Rocha, submitted). The second question refers to the type of signals driving antigen-independent division. Survival and proliferation of memory cells may be driven by IL-15 (48) or by several other cytokines that are able to increase the rate of proliferation of memory T cells (43). It is probable, however, that some type of cognate TCR/MHC/peptide interactions may also be involved: CD4` cells selected by a single MHC/peptide complex do not divide after transfer to irradiated mice expressing the same complex, but they expand after transfer into normal mice (21). These results suggest that peptides responsible for thymus-positive selection and peripheral survival differ from those driving peripheral T cell division, the latter phenomenon involving some type of TCR ‘‘cross’’-reactivity.


LYMPHOCYTE SURVIVAL

89

Finally the most important question concerns the functional activity induced by non-specific T cell stimulation. What are the functional differences between these naturally activated T cells and T cells triggered in the course of an immune response? Several (non-exclusive) possibilities may be considered. Low-affinity signals may (a) reduce the frequency of the cells engaged in different effector functions, or (b) reduce the efficiency of each different effector function. For example, production of cytokines after non-specific stimulation may be so low that they can only promote autocrine growth (and maintain antigen-independent division of memory T cells); they would be unable to act on neighboring cells. (c) Different effector functions may require different thresholds of activation. Some type of cytokine secretion (required for memory T cell growth) may be induced by weak interactions, while T cell cytotoxicity depends on a strong T cell activation. In a lymphocyte receiving multiple signals through several cell surface receptors, what will determine the strength of T cell activation? How lymphocytes integrate TCR-mediated signals with signals transmitted by other cell surface molecules is largely unknown. These signals may act independently, antigen-specific engagement being absolutely required to induce some effector functions. Alternatively some type of integration may take place: In this latter case, a non-specific TCR engagement associated with other intense stimulation through other receptors may also be able to induce strong T cell activation. The answers to these questions have major implications for our understanding of autoimmunity. Mature memory T cells are continuously triggered by nonspecific TCR/MHC/peptide engagement. Strong activation through alternative receptors may explain the emergency of auto-aggression in conditions of lymphokine imbalance (49, 50) or of deficient control of T cell activation and growth (51, 52). Secondly, situations in which T cell repertoires are restricted are now frequent. In AIDS or after irradiation or chemotherapy, the restoration of immunecompetence may be dependent on our capacity to induce low-affinity T cells to mediate effector functions.

B Cell Survival Survival of naive B cells in the peripheral pools also appears to involve interactions between the B cell receptor (BCR) and yet unidentified ligand(s). This was first suggested by experiments in which a transgenic BCR could be ablated by an inducible Cre-LoxP recombination event. After BCR ablation, B cells reportedly rapidly disappeared from the peripheral pools (53). This study, however, did not directly correlate BCR signaling and peripheral B cell survival, as BCR ablation also leads to the arrest of new B cell production in the BM (54). In the absence of the newly formed BM migrants, a significant fraction of the peripheral B cells is rapidly lost (55). In mice with a deletion mutation of the Iga cytoplasmic tail, early B cell development in the BM exhibited only a small impairment, but the generation of the peripheral B cell pool was severely reduced (56). The question remained as to whether the mere presence of a signaling complex, e.g. IgM-


90

FREITAS n ROCHA

Iga-Igb or other, at the cell surface suffices to signal B cell survival, or if B cell survival requires ligand recognition. Other studies addressed directly the role of ligand-mediated recognition in peripheral B cell survival. B cells lacking the Vregion of the IgM receptor have a very short life-expectancy (57). The presence of the truncated membrane IgM transgene, lacking the V-region, provides constitutive signals that suffice to signal allelic exclusion and to promote pre-B cell development in the absence of the surrogate light chain k5 (58), but the transgene fails to support long-term survival of the Tg B cells (57). Differences in the antibody repertoires expressed by pre-B and peripheral B cells can also be invoked to suggest the involvement of ligand-mediated recognition in peripheral B cell positive selection and persistence (59–61). In contrast to T cells, in which TCR survival signaling seems to require the recognition of MHC class I or class II molecules, the nature of the ligand(s) that might be involved in B cell survival remains elusive. Analysis of VH-gene family expression has provided evidence for a very conserved pattern of VH-gene family usage that is independent of the VH-gene number, is strain and tissue specific, and is tightly regulated during B cell differentiation (60). After B cell generation in the BM, naive B cell survival is associated with a strong peripheral selection of B cells expressing particular VH-gene families (61, 62). These observations suggest that the recognition interactions related to B cell survival may not require the involvement of the full antigen-binding site and might be exclusively VHmediated. It is possible that this type of ligand recognition may not lead to full cell activation and that low avidity interactions suffice to promote cell survival. It was recently shown that the BCR is capable of differential signaling and that B cell responses may differ depending upon the properties of the antigen. Thus, while some B cell responses were better correlated with antigen-BCR affinity than with receptor occupancy, others were only weakly dependent on antigen affinity (63). Memory B cells not only show phenotypic changes and express a hypermutated BCR of a different isotype with an increased avidity for antigen, they also show a higher rate of cell division and a lower threshold of activation, and they occupy a different habitat within the secondary lymphoid organs. Studying memory responses to both thymus-dependent and independent antigens and using different cell transfer systems has revealed that long-term persistence of B cell memory requires continuous cell division in the presence of antigen (64–66). Recent observations in mice in which an inducible Cre-loxP-mediated gene inversion (67) is used to change the specificity of the BCR contradict these results. In these experiments, after antigenic stimulation and the generation of memory B cells, memory cells survived even after expressing a new BCR unable to bind to the original activating antigen (M Maruyama, K Lam, K Rajewsky, personal communication). These findings suggest that antigen may be required only at early stages of cell activation and selection of high-affinity B cells in the germinal centers. Once these B cells acquire a memory phenotype, they may no longer require antigen recognition for long-term survival.


LYMPHOCYTE SURVIVAL

91

In this context it is important to mention that the difference in the interactions required for the survival of B and T cells may explain the different rates of mutation of B and T cell receptors. Since T cell survival requires recognition of ubiquitous MHC molecules, somatic mutation of the TCR might always be disadvantageous as it may cause loss of MHC recognition and thus cell death. In contrast, somatic mutation may increase the BCR-ligand avidity and favor selection of high-affinity B cells which, compared to the initial population of nonmutated B cells, have a competitive survival advantage within the germinal center. In their flight for survival, B cells are still able to use mechanisms of receptor editing (68–70) and to change BCR specificity to escape cell death; they thus may gain a survival advantage over other rival cells. Lymphocyte survival is likely to involve multiple mechanisms. Different signals may engage different cell surface receptors using the same or differing survival pathways. Naive T cell survival depends on signals transmitted via lymphokine receptors (71), and it requires the expression of LKLF (72) or NFAT4 (73) transcription factors. The constitutive expression of the Epstein-Barr virus LMP2A protein in transgenic mice bypasses BCR signaling and allows the survival of receptor-less B cells in the peripheral pools (74). Downstream of the BCR signaling pathway, defects in CD45 (75), Btk (76), Syk (77), or NF-jB (78) or in the OBF-1 transcription factor (79) also affect peripheral B cell maintenance. Lymphocyte survival is also modified through the balance between different apoptotic and anti-apoptotic proteins (80–82). BCR signaling and increased levels of intracellular bcl-2 ensure lymphocyte survival through independent pathways (53, 57). The expression of bcl-2 may simply increase cell efficiency by lowering the threshold of resource requirements for survival.

HOMEOSTASIS AND COMPETITION Lymphocyte Production In adult mice the potential to produce new lymphocytes in the primary lymphoid organs or by peripheral cell division largely exceeds the number required to replenish the peripheral pools. Firstly, the number of immature T cell precursors in the thymus is not limiting. A normal-sized T cell pool can be generated in Rag2-deficient mice reconstituted with a mixture of 50% normal BM and 50% BM cells from a CD3e-/- deficient mice that are unable to generate T cells. The thymus production of mature T cells is also in excess as shown in mice in which thymus export is artificially reduced. Mice with reduced thymus export are Rag2-deficient hosts reconstituted with normal BM cells diluted in TCRa-deficient BM (that cannot generate mature T cells). In these mice, peripheral T cell numbers are maintained in spite of the lower production of mature T cells (A Almeida, AA Freitas, unpublished). Finally, T cells can also be produced by peripheral cell division. In a normal mouse,


92

FREITAS n ROCHA

peripheral division is restricted. When peripheral T cells are transferred into a T cell–deficient host, the cells are capable of considerable expansion (38). Each lymphocyte could generate a progeny of about 1015 cells (38). In theory, a single T cell could be sufficient to repopulate the peripheral T cell pool of a mouse. Secondly, an increase in lymphocyte production is not directly reflected in the size of the peripheral T cell pool. Mice grafted with increasing numbers of thymuses do not show a proportional increase in the total number of naive peripheral T cells (83–85). All these results indicate that the total number of peripheral T cells is not necessarily determined by T cell production but is limited at the periphery. Mature B cells are produced in much higher numbers than those required to fill the peripheral B cell pool. The number of B cell precursors is not limiting, as shown in mice in which the B cell precursor number is artificially reduced. The number of pre–B cells can be reduced in Rag2-deficient mice, which were lethally irradiated and reconstituted with mixtures of normal BM cells diluted among incompetent BM cells from B cell–deficient (lMT) donors. In these chimeras the number of BM pre–B cells is proportional to the fraction of normal BM cells injected. Mice containing less than 25% of the normal number of pre–B cells had reduced peripheral B cell numbers. A normal-sized peripheral B cell pool, however, was generated in mice containing only 30% of the normal number of pre– B cells. These results demonstrate that about one third of the normal number of BM B cell precursors suffices to maintain the peripheral B cell pool size (86). A similar conclusion was reached after parabiosis between one normal and one or two B cell–deficient mice. In these circumstances B cell production was restricted to the BM of the normal mouse since no chimerism was detectable in the BM of the different partners. In mice triads each individual mouse had physiological B cell numbers, i.e. the B cell production of one mouse was sufficient to populate the peripheral pools of three mice (86). These results demonstrate that peripheral B cell numbers are not determined by the rates of BM B cell production but are limited at the periphery.

Demonstration of Lymphocyte Competition In an immune system where new lymphocytes are continuously produced in excess but their total numbers are kept constant, newly generated cells have to compete with other newly produced or resident cells to survive. Competition can be defined as ‘‘an interaction between two populations, in which, for each, the birth rates are depressed or the death rates increased by the presence of the other population’’ (87). There are two main established criteria accepted as evidence of competition among populations: 1. The presence of competitors should modify the equilibrium size of a population, and 2. their presence should alter the dynamics, e.g. the life span, of a population (88). The question of whether competition arises between B and T cells was addressed by comparing the development and the fate of TCR or BCR-Tg and non-Tg populations in several different lines of


LYMPHOCYTE SURVIVAL

93

mouse BM chimeras (89, 90). It was found that (a) when injected alone, Tg and non-Tg cell populations show an identical behavior and generate peripheral pools of similar size and (b) when Tg and non-Tg cells are mixed in the same host, they initially accumulate at the same rate. However, after reaching steady-state numbers, a preferential selection of the non-Tg cells occurs at the periphery. These observations fulfil the first criterion for competition since they demonstrate that the presence of non-Tg populations modifies the number of the Tg cells (89, 90). In these experiments the life-expectancy of the Tg, T, or B cells varied according to the presence and the type of other competing cells (89–91). These latter findings fulfil the second main criterion as they prove that the presence of competitors alters the life span of a population. By considering each cell clone as a species, competition between Tg and nonTg cells represents an example of inter-specific competition. Competition may also occur between individuals of the same species (intra-specific competition) (87). Studies using monoclonal mice may allow determination of whether competition also occurs among cells of the same clone.

Types of Competition and Resources Competition may arise through different processes. In interference competition, populations may interact directly with each other, or one population can prevent a second population from occupying a habitat and from exploiting the resources in it. Thus, although interference competition may occur for a resource, it is ‘‘only loosely related to the resource level’’ (87). In exploitation competition, different populations have a common need for resources present in limited supply. In this case competition is directly related to level of resources available. We may define resource as any factor that can lead to increased cell survival or growth through at least some range of their availability (88). In a broad sense, a resource is any factor used by a cell and for that reason no longer available to other cells. Resources can be essential, complementary, substitutable, antagonistic, or even inhibitory (92). Ample evidence exists for the role of resources in lymphocyte competition. First, the kinetics of accumulation of Tg and non-Tg populations after reconstitution in different groups of mouse chimeras followed a density-dependent growth curve, which follows the Monod growth function, i.e. ‘‘it increases in a saturating manner with resource availability’’ (93). During the expansion phase of cell reconstitution, resources are abundant, and Tg and polyclonal populations accumulate at the same rate. When population growth reaches equilibrium, polyclonal populations become dominant, i.e. competition only occurs when resources are limiting (89, 90). Second, as discussed below, changes in resource level should alter the balance between populations. Administration of antigens to chimeras hosting different Tg cell populations favors dominance by the antigen-specific Tg cells (91, 94). Third, experiments demonstrating that the total numbers of peripheral T and B cells are not determined by rates of new cell production, but are

94

FREITAS n ROCHA

limited at the periphery, represent further evidence for the existence of resource competition. In the immune system, many molecules can function as resources, e.g. antigen, MHC molecules, ligands for costimulatory and adhesion molecules, mitogens, interleukins, chemokines, hormones, and other growth factors, etc. Resources can be external to the immune system or be produced by the lymphocytes themselves. By producing their own resources lymphocytes also contribute to generate their own ecological space.


Resource Availability Shapes Lymphocyte Populations Resource competition may also contribute to the regulation of the size of peripheral lymphocyte pools (94); if peripheral resources are plentiful, many more newly generated T and B cells are able to survive (86). Any manipulation of resources may modify lymphocyte populations. When resources are used by many cell types, changes in these common resources will modify overall cell numbers. Hormones and mitogens are examples of pleiotropic resources. The size of T and B cell pools is diminished in mice deficient in growth hormone and is increased in adrenalectomized, ovarectomized, castrated, pseudo-pregnant, and pregnant mice (6). In axenic mice, lymphocyte numbers are reduced and lymphocytes expressing activation or memory markers are rare or absent (B Rocha, unpublished). Normal-sized peripheral pools are rapidly restored as soon as the normal gut flora is reestablished or after its colonization by a single fungus or bacteria species. Activated lymphocytes and antigen-presenting cells (APC) are major producers of their own resources. Antigen stimulation increases resource availability by inducing APC and lymphocyte activation and the production of numerous cytokines. In this context, lymphocytes can increase in numbers as during the expansion phase of the immune response (95). Once antigen is eliminated, cytokine production decreases. Reduced resource levels are unable to maintain the same number of lymphocytes: Most will die during the contraction phase of the immune response (95). Antigen can also be considered as an example of a private resource, i.e., that used by a particular cell set. Changes in the levels of private resources modify the composition of lymphocyte populations, as when antigen injection to chimeras carrying mixtures of two different Tg B cells or of Tg and non-Tg B cells shifted the balance between populations to favor the antigen-specific Tg B cells (91, 94). In conditions of resource competition, one would also expect changes in morphology of a population in response to the presence of competitors: a process known as character displacement (96). The IgM secreted into the serum by a population of normal B cells exhibits different binding patterns according to the presence or absence of a population of Tg B cells (94). This implies that the presence of the Tg B cells leads to changes in the selection of the non-Tg B cell clones, a process that at a population level may be considered to mimic character displacement.

LYMPHOCYTE SURVIVAL

95

The Competition Exclusion/Diversity Paradox


According to the competition exclusion principle, competition should lead to the progressive dominance of a limited number of clones and consequently to the exclusion of most clones (87). In this context lymphocyte competition would be incompatible with maintenance of diversity. This traditional model of population dynamics assumes that all cells interact equally well with each other in small closed environments. The more recent metapopulation approach takes into account difference in the age and structure of the populations, their migration capacity, the heterogeneity and patchwork distribution of fragment habitats, and their change over time (4, 97). This approach explains the co-existence of multiple potentially competing species within larger areas of space. According to this theory, there are several mechanisms through which the immune system can solve the apparent competition-diversity paradox and select preferentially for diversity. 1. Continuous cell production. The continuous generation of naive cells ensures the permanent availability of new cell specificities. The contribution of new cell production to the maintenance of diversity can be evaluated by comparing lymphocyte repertoires in young and old individuals. With age the progressive decrease in new lymphocyte production both in the thymus and in the bonemarrow generally correlates with an increased frequency of pauciclonal repertoires. 2. Terminal differentiation. Control of cell proliferation by terminal differentiation prevents unlimited expansion and dominance (34). This is the case of B cell terminal differentiation into Ig-secreting plasma cells. 3. The redundancy of resources. The capacity to use multiple alternative factors for survival and proliferation may provide a critical advantage for cells competing for limiting resources. Indeed, B cells from a LPS-reactive mouse strain have a clear competitive advantage when compared to B cells from a non-LPS reactive strain (89). 4. Diversification in resource usage. The use of different resources by different populations of lymphocytes will allow the co-existence of different cell types. Cell differentiation is associated with the acquisition of receptors for different chemokines and growth factors or by regulation of homing receptors. These parameters contribute to create a heterogeneity of habitats that favors coexistence of diverse, potentially competing cell populations (4). Migration ensures the distribution of cells from source to the periphery and between the different peripheral environments (98): Cell survival is associated with the adaptation to new ecological niches.

The Ecological Niche The variety of resources and the heterogeneity of anatomical structures in the lymphoid organs permits that different lymphocyte types may find an ecological


96

FREITAS n ROCHA

niche to survive. The inability of the cell to migrate into the niche may compromise survival; exclusion of B cells from follicular niches in the secondary lymphoid organs has been claimed to result in cell death (99). Niches may also play a role in homeostasis and in the control of cell numbers by forcing the B cells to migrate into the primary follicles (100) where they must compete for trophic factors present in limited supply. Lymphocytes may help to define their appropriate niches. Thus, differentiation of epithelial cells in the thymus and the organization of the thymus cortex and medulla cellular networks depend on the presence of thymocytes (101); the maturation of the splenic follicular dendritic cell networks and the organization of B cell follicles depend on the expression of LTa and b and TNF by B lymphocytes (102–104). Germinal centers that develop in the B cell follicles during T cell–dependent antibody responses are one of the best-characterized ecological niches in immunology (105–107). Germinal centers are oligoclonal; the B cells that give rise to germinal centers are initially activated outside follicles; on the average three B cells colonize each follicle. Massive clonal expansion of the initial founder cells, driven by antigen held by follicular dendritic cells, prevents colonization of the germinal center by B cells specific for a second unrelated antigen. Expansion of the B cells is accompanied by BCR hypermutation. Competition among the proliferating B cells, based on their ability to interact with antigen held on FDC, will lead to the preferential survival of the cells capable of high-affinity recognition and to the death by apoptosis of other cells. Germinal centers play therefore a critical role in the maturation of immune responses by selecting cells with high avidity for antigen binding. They seem to have evolved to provide the appropriate niche for selecting B cells that attempt to gain competitive advantage for survival by somatic hypermutation. Indeed, somatic hypermutation is present in philogeny (108) well before the development of the affinity maturation of immune responses (109). Affinity maturation of the immune response, however, is only present in species capable of germinal center formation. When heterogeneous populations occupy heterogeneous habitats, the probability of extinction of a species is higher for the rare populations (97). The same rules may apply to the immune system. In mice, arrest of new B cell production leads to preferential extinction of B cells expressing rare V genes (61). The probability of extinction of a population is also related to the area of habitat occupied. Destruction of wide areas of habitat can be predicted to be more damaging for the survival of the most frequently found species with a wider distribution than for some rare species that commonly occupies restricted niches and may therefore escape catastrophic events. Large structural changes in secondary lymphoid organs (e.g. sub-lethal irradiation) affect predominantly dominant B cell clonal responses (110).

LYMPHOCYTE SURVIVAL

97

NICHE DIFFERENTIATION


Different Types of Lymphocytes Occupy Different Niches Different lymphocyte subsets likely use different resources. In adult mice the size of T and B lymphocyte populations is independently regulated. The number of mature B cells is similar in normal mice or in mice that lack T cells (111). In mice that lack B cells, the number of T cells is similar to that of normal mice (54). B and T cells also occupy different locations in the secondary lymphoid organs (112, 113). These observations suggest that B and T cells belong to different guilds, i.e. they exploit different resources. Among the T cell subsets the number of ab and cd T cells is also independently regulated (114): In the absence of ab T cells, the number of cd T cells does not increase significantly and vice versa. Within ab T lymphocytes, the number of CD4` and CD8` T cells is, however, coregulated (38). In the absence of either of the two CD4` or CD8` T cell subsets, cell loss can be compensated by the remaining cellular subset, and the total number of ab T cells remains similar to that of normal mice (115–117), suggesting that CD4` and CD8` T cells may partially share common resources.

Naive and Memory/Activated Cells Belong to Different Niches In response to the presence of competitors, a population may modify its survival requirements and adapt to a new ecological niche. During lymphocyte development, cells bearing the same clonal specificity but at different stages of differentiation (immature vs. mature; naive vs. memory; Th1 vs. Th2) either modify their use of resources or interact differently with the same resources. The variety and the patchy distribution of resources in different habitats will cause the adaptive lymphocyte environment to change according to time and tissue localization. Lymphocytes engage in niche differentiation as is the case for naive and activated/memory populations of cells. The peripheral pool of CD8` T cells was divided in two compartments (naive and memory) of equivalent size that are regulated by independent homeostatic mechanisms. Mice manipulated to contain only naive CD8` T cells, in spite of available space, have only half of the total number of CD8` T cells. Similarly, in mice engineered to contain only memory CD8` T cells, the total number of CD8` T cells is also reduced by half (37, 118). To reconstitute a normal-sized compartment, both naive and memory T cells are required. The existence of independent niches for naive and memory T cells ensures the coexistence of both cell types; this fact has major functional implications. Thus, the continuous thymus output of naive T cells cannot lead to the extinction of previously generated memory responses as exported naive thymus migrants do not replace resident memory T cells (119). Survival of naive T cells relies on the nature and number


98

FREITAS n ROCHA

of other competing naive T cells (90, 119). Conversely, accumulation of memory cells will not lead to the contraction of the naive T cell pool; successive antigen encounters will not necessarily result in pauciclonal repertoires, as the diverse naive repertoire will remain. Survival of resident memory T cells will rather be determined by the expansion of newly generated memory T cell populations. Upon the sequential immunization of mice with different types of virus (120), response to a second antigen leads to the reduction of the residual memory to a first antigen. It is not yet known if CD4` T cells follow the same strategy, but preliminary results suggest that this may be the case (A Almeida, AA Freitas, unpublished). In the case of B lymphocyte populations, the numbers of resting and activated IgM-secreting cells are also independently regulated (86, 121), as shown in mice engineered to contain different numbers of B cells. In mice with very low numbers, a significant fraction of the B cells shows an activated phenotype, and the number of IgM-secreting cells and serum IgM levels are as in normal mice (86, 121, 122). Mature resting B cells only accumulate once the activated pool is replenished (121). These results suggest that the immune system is organized to first ensure the maintenance of normal levels of natural antibodies (innate immunity), which constitutes an initial barrier of protection, while keeping a reserve of diversity in the resting B cell compartment that allows responses to new antigenic stimuli. The independent homeostatic regulation of resting and activated/ memory lymphocyte compartments implies the existence of a hierarchical organization of the immune system.

Lymphocyte Substitution From birth to adult life, lymphocyte pools increase in size and can accommodate newly generated lymphocytes. The situation differs once steady-state numbers are reached, and the ‘‘space’’ is full: Survival of newly generated lymphocytes will be determined by the rate of cell substitution. Substitution is conditioned by the rate of colonization (invasion) by new cells, but it is also determined by the rate of extinction of pre-established cells. To survive, a cell must migrate and find the appropriate niche. If the niche is empty, there is no obstacle to cell entering. If the niche is already colonized, the newly arriving cell is either able to out-compete and displace the established cell or it is prevented from entering the niche and dies. A population pre-established in the correct niche and with a low extinction rate benefits from the founder advantage ‘‘first come, first served.’’ It may resist replacement by a newly arrived cell. The most extreme case of substitution is succession, in which the vast majority of a resident population is replaced by a population of newly arriving cells. This situation probably occurs during ontogeny, as lymphocytes produced during the perinatal period are replaced by adult cells. Thus, T lymphocytes generated early in ontogeny perform poorly when compared to adult T cells (123). They proliferate poorly and secrete low levels of certain cytokines in response to in vitro


LYMPHOCYTE SURVIVAL

99

anti-TCR stimulation. In vivo proliferation in response to anti-TCR or antigenic stimulation is also impaired (124, 125), owing both to an intrinsic lymphocyte defect and to deficient antigen-presenting cell function (125). These characteristics of the perinatal environment are believed to be the basis for its increased susceptibility to tolerance induction (126). Affinity maturation may be absent, as suggested by the extensive cross-reactivity of neonatal T cell repertoires (127). Similarly, neonatal B cell repertoires differ from those of adult mice. In neonatal mice the DH proximal VH7183 gene family is used at the same frequency by 30% of the immature pre-B cells in the BM and 30% of the peripheral B cells (60), while in adult mice B cells expressing VH-genes of the VH7183 gene family are strongly counter-selected at the pre-B to B cell transition (62). Comparing the binding specificities of B cell hybridomas revealed that the vast majority of the antibodies were multireactive in neonatal mice, (128), while in adult mice most of the antibodies were monospecific (129). These changes of B cell repertoires correlate with overall modifications of B cell dynamics (55, 130), suggesting that B cell selection is permissive in neonatal mice during the phase of expansion of lymphocyte numbers, while it becomes strict in adult mice when resources become scarce and competition is established. The replacement of both T and B cell neonatal cross-reactive repertoires suggests that in ontogeny lymphocyte populations go through a process of ecological succession in which generalist cells are replaced by specialists. Analysis of the processes of succession also may reveal other types of interactions between populations (97). By modifying the environment, the presence of an early population may prepare the site for colonization by a later cell type, a process known as facilitation. Examples of these should be looked for in the immune system. How does substitution work in adult mice? Studies of the substitution of resident naive T cells by recent thymus migrants gave contradictory results. In irradiated TCR-Tg mice injected with normal BM cells, substitution of the naive Tg CD8` T cells by newly generated naive non-Tg cells was age-independent; e.g. thymus migrants and resident cells appeared to have the same survival probability at the periphery (119). Opposite results were obtained in mice transplanted with increased numbers of fetal thymus grafts. These grafts initially exported fetal T cells of donor origin but were subsequently colonized by host BM cells. The study of the substitution of T cells of graft origin by T cells of host origin suggested that recent thymus migrants survived better than resident naive T cells (85). Finally, the study of long-term colonization and persistence of recent thymus migrants labeled in vivo with CFSE seems to suggest the resistance of resident cells to replacement by the newly arriving cells (131). Similar conclusions were made from parabiosis experiments (90). The reasons for these conflicting results no doubt derive from the complexity of the different experimental systems required to study substitution. All experimental approaches involved stressful


100

FREITAS n ROCHA

conditions such as irradiation or surgery. Very diverse populations (Tg vs. polyclonal; fetal vs. adult) were compared. Finally, conclusions on substitution rates rely on the quantification of very low T cell numbers. The definite answer to this question requires results from new, as yet unavailable, experimental approaches. As naive and memory T cells belong to different niches (see section on Competition Exclusion/Diversity Paradox), naive recent thymus migrants do not dislodge resident memory T cells (119). Surprisingly, thymus migrants are able to dislodge resident tolerant T cells. Tolerant CD8` Tg cells can persist at high frequencies for prolonged periods of time in the absence of other CD8` T cells (132). Tolerant Tg CD8` T cells can be substituted by both thymus migrants (119) and memory CD8` T cells (B Rocha, unpublished). This replacement is such that a significant part of these cells disappear, but a few cells remain for prolonged time periods. The residual tolerant Tg T cells are unable to eliminate antigen but are not inert. They secrete IL-10 and c-IFN (133) and may also play a role in the control of immune responses by interference competition (134, 135). In the naive B cell pool, new B cells produced in the BM replace resident B cells continuously. Maintenance of the physiological size of the peripheral B cell pool requires a minimal continuous input of new cells because mouse chimeras with a 10-fold reduction of B cell production show diminished B cell numbers (86). Abrogation of the de novo B cell production leads to a rapid decrease in naive B cell numbers (55, 136). The excess of B cell production in normal physiological conditions suggests the existence of a high attrition rate at the periphery (55). Tolerant B cells from double-transgenic mice coexpressing hen egg lysozyme (HEL) and anti-HEL Ig-genes have a relatively short life span when compared to normal B cells (137). In contrast to naı¨ve B cells, naturally activated IgM-secreting B cells can resist replacement by newly arriving B cells. They can also prevent the entry of new emigrating B cells into the activated cell pool (121). This negative feedback regulation may be due to the secretion of inhibitory factors, e.g. Igs, IL-10, cIFN. Alternatively, the first established population may occupy a niche required for the selection and survival of incoming cells, preventing colonization (87). The ability of newly formed B cells to differentiate into IgM plasma cells is dependent on the nature and number of cells already present at the periphery. Early during development and expansion of the immune system, an initial pool of activated B cells is selected that may resist replacement. The same IgM-secreting cells are also found in axenic mice, suggesting their activation by self-antigens. Recent evidence suggests a possible role of self-antigen in the positive selection and activation of autoreactive B cells (138). Competition based on the BCR diversity and the antigenic environment heterogeneity eventually leads to the substitution of the initially selected population by new specificities formed in the BM. The question remains concerning the possible functional differences between these naturally activated B cells and the B cells triggered in the course of an immune response.

LYMPHOCYTE SURVIVAL

101


MUTUALISM Besides competition, populations of cells also interact differently. One of the most important types of interactions is cooperation, also called mutualism. Through cooperation the survival and growth rate of one population is increased in the presence of a second population. During evolution eukaryotic cells represent the first example of mutualism (139). Multicellular organisms, where no cell survives on its own, are yet another example of mutualism. Multiple examples of mutualism are found in the immune system: T and B lymphocytes evolved mutualistic interactions; CD4` T cells help proliferation and differentiation of antigen-specific B cells. Reciprocally, antigen presentation by B cells contributes to the expansion of antigen-specific T cells. CD4` T cells enhance cytolytic CD8` T cell responses. Similarly, T lymphocytes and antigenpresenting cells (APC) have evolved mutualistic interactions, essential for T cell survival and for triggering immune responses. Dendritic and other antigen-presenting cells may therefore play a keystone role in the establishment of lymphocyte communities. Predator-prey interactions may also shape lymphocyte populations. In chimeras reconstituted with mixed populations of BM cells from male and female donors, the immediate injection of TCR Tg CD8` T cells specific for the HY male antigen leads to the elimination of all cells of male origin, followed by an increase in the number of cells from female progeny (132). In fact, in this example, the fate of the two BM cell types mimics the dynamic behavior of two competing populations—apparent competition (87).

SELECTION OF REPERTOIRES: TOLERANCE AND AUTOIMMUNITY In an immune system where the total number of cells is limited, lymphocyte repertoires will be shaped by the differential ability of lymphocytes to survive. Lymphocyte survival relies on cell/ligand interactions, the availability of other resources, and the nature and number of competing rivals. In their continuous flight for survival, lymphocytes must acquire a selective advantage over their competitors. They must adapt to the immediate environment by modifying their survival thresholds (140) and or changing their requirements through differentiation. At different life stages, lymphocytes will require different signals to survive. These factors impose a continuous selective pressure throughout the lymphocyte life history. Initially, T and B cell repertoires are shaped in primary lymphoid organs at early stages of T and B cell differentiation (141, 142). In the thymus, selection is determined by the avidity of cognate receptor/ligand interactions; by the presence of growth and differentiation factors; by the size of the selecting niche (143);


102

FREITAS n ROCHA

by the number of competitors present at the different stages/compartments of lymphocyte differentiation (90, 144). Immature B cell precursors are particularly sensitive to deletion and receptor-editing mechanisms by low-affinity interactions (145), and the outcome of the ligand/B cell interaction may be determined by the degree of receptor engagement (146). Selection of T and B cells continues at the periphery (147, 148). After activation, lymphocytes may follow a pre-established hierarchical program and differentiate first to effector functions, then to memory, to anergy and/or terminal differentiation (149). The duration of antigen persistence may be fundamental to trigger sequentially these stochastic processes (29, 150) and thus to determine the class of immune responses. Lymphocyte selection and adaptation also require co-evolution involving several different cell lines. Different subclasses of lymphocytes and antigenpresenting cells represent patterns of co-evolved interactions. These patterns allow the emergence of the holistic properties of the immune system. The global properties of cellular communities by far exceed the simple properties of the individual cells. The multiple interactions between all differing cells suffice to explain the immune system’s development and functions. The context of homeostasis, survival, and competition excludes any manicheism as a realistic explanation of the global behavior of the immune system. In contrast to the original postulate of the clonal selection theory (151), lymphocyte selection through competition occurs even in absence of exogenous antigens. The idiotypic network of variable-to-variable region interactions is not essential (152). A closed autistic self-referential immune system (153) is impossible. The immune system’s functions cannot be directed by self/non-self (151) or sense/non-sense (154) discrimination or by the detection of danger (155). Lymphocyte activation, rather than being determined by the ability of the cells to discriminate between one or two signals (156), will likely be triggered by the capacity of the cell to integrate different exogenous stimuli. Indeed, both TCR and BCR are capable of differential signaling according to the properties of the antigen and the environmental conditions of stimulation (63, 132). In a complex system where a vast number of different new cell specificities are continuously generated in both the central and the peripheral compartments, the presence of self-reactive cells is unavoidable. Besides the described mechanisms of tolerance induction that work at a single cell level, the global properties of the system may also contribute to avoid autoaggression. One of these properties is niche segregation. When a lymphocyte encounters an activating ligand in the correct niche and in the presence of plenty of resources, it will respond and expand proportionally. This is the case for infectious agents presented to lymphocytes in the appropriate niche (in the secondary lymphoid organs) and in the presence of an excess of new resources provided by the primary inflammatory reaction. In contrast, encounters with a self-ligand frequently occur outside lymphoid organs, where deficient antigen presentation and limiting resources may be major factors to restrict clonal expansion and autoaggression.


LYMPHOCYTE SURVIVAL

103

The growth of the potentially harmful cells will be controlled by lymphocyte competition. In these circumstances, lymphocyte diversity associated with polyclonal stimulation may help prevent the emergence of large clone sizes, clonal dominance, and immune-pathology. Conversely, pauciclonal repertoires will determine higher fluctuations in clonal frequencies, which may increase the probability for the expansion and fixation of clones that in the absence of competition will become irreversibly dominant and induce pathology. This is supported by the inverse correlation between the world incidence of autoimmune diseases and the frequency of infectious diseases. The prevalence of autoimmune pathology is higher in developed countries with low population densities and very cold climates that do not favor spread of bacteria, parasites and viruses. Prevalence is lower in highly populated areas of third world countries (157). In this context polyclonal activation may have a therapeutic role in autoimmune disease, while immune suppression as a side effect may reinforce the already established selfreactive clonal drift. It may be more appropriate to refer to ‘‘horror monoclonicus’’ rather than ‘‘horror autotoxicus’’; in the survival/competition model the question is no longer ‘‘how large should the repertoire be to discriminate self and non-self?’’ (158, 159), but rather ‘‘what is the minimum size of the repertoire that protects against and avoids autoimmune pathology?’’ In their flight for survival lymphocytes may be caught in a Prisoner’s Dilemma game (160–162): They either cooperate, i.e. protect the host with strong immune responses, and continue to survive, or they defect, i.e. fail to protect or even destroy the host with uncontrolled responses, and so eventually die. ACKNOWLEDGMENTS We thank our colleagues F Agenes, A Almeida, S Guillaume, N Legrand, M Rosado, C Tanchot and H Veiga-Fernandes for continuous discussion. We thank Drs. T Brocker, P Ahmed and U Rajewsky for allowing us to quote their unpublished work. This work was supported by the Institute Pasteur, INSERM, CNRS, ANRS, ARC, DRET, Sidaction and MRT, France. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Miller SL, Orgel LE. 1974. The Origins of Life on Earth. Englewood Cliffs, NJ: Prentice-Hall 2. Dawkins R. 1989. The Selfish Gene. Oxford, UK: Oxford Univ. Press 3. Buss LW. 1987. The Evolution of Individuality. Princeton, NJ: Princeton Univ. Press

4. Hanski I. 1999. Metapopulation Ecology. Oxford, UK: Oxford Univ. Press 5. Freitas AA, Rocha B. 1997. Lymphocyte survival: a red queen hypothesis. Science 277:1950 6. Freitas AA, Rocha B. 1993. Lymphocyte lifespans: homeostasis, selection and competition. Immunol. Today. 14:25–29


104

FREITAS n ROCHA

7. Van Valen L. 1973. A new evolutionary law. Evol. Theory 1:1–30 8. Tanchot C, Lemonnier FA, Perarnau B, Freitas AA, Rocha B. 1997. Differential requirements for survival and proliferation of CD8 naive or memory T cells. Science 276:2057–62 9. Kisielow P, Miazek A. 1995. Positive selection of T cells: rescue from programmed cell death and differentiation require continual engagement of the T cell receptor. J. Exp. Med. 181:1975–84 10. van Santen HM, Woolsey A, Aston Rickardt PG, van Kaer L, Baas EJ, Berns A, Tonegawa S, Ploegh H. 1995. Increase in positive selection of CD8` T cells in TAP1-mutant mice by human beta2microglobulin transgene. J. Exp. Med. 181:787–92 11. Pircher H, Burki K, Lang R, Hengartner H, Zinkernagel RM. 1989. Tolerance induction in double specific T-cell receptor transgenic mice varies with antigen. Nature 342:559–61 12. Nesic D, Vukmanovic S. 1998. MHC class I is required for peripheral accumulation of CD8` thymic emigrants. J. Immunol. 160:3705–12 13. Kirberg J, Berns A, von Boehmer H. 1997. Peripheral T cell survival requires continual ligation of the T cell receptor to major histocompatibility complex– encoded molecules. J. Exp. Med. 186:1269–75 14. Viret C, Wong FS, Janeway CA. 1999. Designing and maintaining the mature TCR repertoire: the continuum of selfpeptide: self-MHC complex recognition. Immunity 10:559–68 15. Rooke R, Waltzinger C, Benoist C, Mathis D. 1997. Targeted complementation of MHC class II deficiency by intrathymic delivery of recombinant adenoviruses. Immunity 7:123–34 16. Takeda S, Rodewald H-R, Arakawa H, Bluethman H, Shimizu T. 1996. MHC class II molecules are not required for survival of newly generated CD4` T

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

cells, but affect their long-term life span. Immunity 5:217–28 Cardell S, Tangri S, Chan S, Kronenberg M, Benoist C, Mathis D. 1995. CD1restricted CD4` T cells in major histocompatibility complex class II-deficient mice. J. Exp. Med. 182:993–1004 Brocker T. 1997. Survival of mature CD4 T lymphocytes is dependent on major histocompatibility complex class IIexpressing dendritic cells. J. Exp. Med. 186:1223–32 Beutner U, MacDonald HR. 1998. TCRMHC class II interaction is required for peripheral expansion of CD4 cells in a T cell-deficient host. Int. Immunol. 10:305– 10 Martin WD, Hicks GG, Mendiratta SK, Leva HI, Ruley HE, Van Kaer L. 1996. H2-M mutant mice are defective in the peptide loading of class II molecules, antigen presentation, and T cell repertoire selection. Cell 84:543–50 Bender J, Mitchell T, Kapler J, Marrack P. 1999. CD4` T cell division in irradiated mice requires peptides distinct from those responsible for thymic selection. J. Exp. Med. 190:367–74 von Boehmer H, Hafen K. 1993. The life span of naive ab T cells in secondary lymphoid organs. J. Exp. Med. 177:891– 96 Tough DF, Sprent J. 1994. Turnover of naive- and memory-phenotype T cells. J. Exp. Med. 179:1127–35 Bruno L, Kirberg J, von Boehmer H. 1995. On the cellular basis of immunological T cell memory. Immunity 2:37– 43 DiSanto JP, Aifantis I, Rosmaraki E, Garcia C, Feinberg J, Fehling HJ, Fischer A, von Boehmer H, Rocha B. 1999. The common cytokine receptor gamma chain and the pre-T cell receptor provide independent but critically overlapping signals in early alpha/beta T cell development. J. Exp. Med. 189:563–73 Budd RC, Cerottini J-C, Horvath C, Bron

LYMPHOCYTE SURVIVAL

27.

28.


29.

30.

31.

32.

33.

34.

35.

C, Pedrazzini T, Howe RC, MacDonald HR. 1997. Distinction of virgin and memory T lymphocytes. J. Immunol. 138:3120–29 Curtsinger JM, Lins DC, Mescher MF. 1998. CD8` memory T cells (CD44high, Ly-6C`) are more sensitive than naive cells to (CD44low, Ly-6C1) to TCR/ CD8 signaling in response to antigen. J. Immunol. 160:3236–43 Kedl RM, Mescher MF. 1998. Qualitative differences between naive and memory T cells make a major contribution to the more rapid and efficient memory CD8` T cell response. J. Immunol. 161:674–83 Gett AV, Hodgkin PD. 1998. Cell division regulates the T cell cytokine repertoire, revealing a mechanism underlying immune class regulation. Proc. Natl. Acad. Sci. USA 95:9488–93 Ke Y, Ma H, Kapp JA. 1998. Antigen is required for the activation of effector activities, whereas interleukin 2 is required for the maintenance of memory in ovalbumin-specific, CD8` cytotoxic T lymphocytes. J. Exp. Med. 187:49–57 Nahill SR, Welsh RM. 1993. High frequency of cross-reactive cytotoxic T lymphocytes elicited during the virusinduced polyclonal cytotoxic T lymphocyte response. J. Exp. Med. 177:317–27 Tough DF, Borrow P, Sprent J. 1996. Induction of bystander T cell proliferation by viruses and Type I interferon in vivo. Science 272:1947–50 Zhang XH, Sun SQ, Hwang IK, Tough DF, Sprent J. 1998. Potent and selective stimulation of memory-phenotype CD8` T cells in vivo by IL-15. Immunity 8:591–99 Grossman Z. 1982. Recognition of self, balance of growth and competition: horizontal networks regulate immune responsiveness. Eur. J. Immunol. 12: 747–56 Markiewicz MA, Girao C, Opferman JT, Sun JL, Hu QH, Agulnik AA, Bishop

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

105

CE, Thompson CB, Ashton-Rickardt PG. 1998. Long-term T cell memory requires the surface expression of selfpeptide/major histocompatibility complex molecules. Proc. Natl. Acad. Sci. USA 95:3065–70 Bachmann MF, Gallimore A, Linkert S, Cerundolo V, Lanzavecchia A, Kopf M, Viola A. 1999. Developmental regulation of Lck targeting to the CD8 coreceptor controls signaling in naive and memory T cells. J. Exp. Med. 189:1521–30 Tanchot C, Rocha B. 1995. The peripheral T cell repertoire: independent homeostatic regulation of virgin and activated CD8` T cell pools. Eur. J. Immunol. 25:2127–36 Rocha B, Dautigny N, Pereira P. 1989. Peripheral T lymphocytes: expansion potential and homeostatic regulation of pool sizes and CD4/CD8 ratios ‘‘in vivo.’’ Eur. J. Immunol. 19:905–11 Hou S, Hyland L, Ryan KW, Portner A, Doherty PC. 1994. Virus-specific CD8` T-cell memory determined by the clonal burst size. Nature 369:652–54 Lau LL, Jamieson BD, Somasundaran T, Ahmed R. 1994. Cytotoxic T-cell memory without antigen. Nature 369:648–52 Mullbacher A. 1994. The long-term maintenance of cytotoxic T cell memory does not require persistence of antigen. J. Exp. Med. 179:317–21 Tripp RA, Lahti JM, Doherty PC. 1995. Laser light suicide of proliferating virusspecific CD8` T cells in an in vivo response. J. Immunol. 155:3719–21 Sprent J, Tough DF, Sun S. 1997. Factors controlling the turnover of T memory cells. Immunol. Rev. 156:79–85 Doherty PC, Topham DJ, Tripp RA. 1996. Establishment and persistence of virus-specific CD4` and CD8` T cell memory. Immunol. Rev. 150:23–44 Miller RA, Stutman O. 1984. T cell repopulation from functionally restricted splenic progenitors: 10,000-fold expansion documented by using limiting dilu-


106

FREITAS n ROCHA

tion analyses. J. Immunol. 133:2925–32 46. Hellerstein M, Hanley MB, Cesar D, Siler S, Papageorgopoulos C, Wieder E, Schmidt D, Hoh R, Neese R, Macallan D, Deeks S, McCune JM. 1999. Directly measured kinetics of circulating T lymphocytes in normal and HIV-1-infected humans. Nat. Med. 5:83–89 47. Mackall CL, Hakim FT, Gress RE. 1997. Restoration of T-cell homeostasis after Tcell depletion. Semin. Immunol. 9:339– 46 48. Lodolce JP, Boone DL, Chai S, Swain RE, Dassopoulos T, Trettin S, Ma A. 1998. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation. Immunity 9:669–76 49. Kramer S, Schimpl A, Hunig T. 1995. Immunopathology of interleukin (IL) 2deficient mice: thymus dependence and suppression by thymus-dependent cells with an intact IL-2 gene. J. Exp. Med. 182:1769–76 50. Kuhn R, Lohler J, Rennick D, Rajewsky K, Muller W. 1993. Interleukin-10deficient mice develop chronic enterocolitis. Cell 75:263–74 51. Waterhouse P, Penninger JM, Timms E, Wakeham A, Shahinian A, Lee KP, Thompson CB, Griesser H, Mak TW. 1995. Lymphoproliferative disorders with early lethality in mice deficient in Ctla-4. Science 270:985–88 52. Van Parijs L, Biuckians A, Ibragimov A, Alt FW, Willerford DM, Abbas AK. 1997. Functional responses and apoptosis of CD25 (IL-2R alpha)-deficient T cells expressing a transgenic antigen receptor. J. Immunol. 158:3738–45 53. Lam K-P, Kuhn R, Rajewsky K. 1997. In vivo ablation of surface immunoglobulin on mature B cells by inducible gene targeting results in rapid cell death. Cell 90:1073–83 54. Kitamura D, Roes J, Ku¨hn R, Rajewsky K. 1991. A B-cell deficient mouse by targeted disruption of the membrane exon

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

of the immunoglobulin mu chain gene. Nature 350:423–26 Freitas AA, Rocha B, Coutinho A. 1986. Lymphocyte population kinetics in the mouse. Immunol. Rev. 91:5–37 Torres RM, Flaswinkel H, Reth M, Rajewsky K. 1996. Aberrant B cell development and immune response in mice with a compromised BCR complex. Science 272:1804–8 Rosado MM, Freitas AA. 1998. The role of the B cell receptor V region in peripheral B cell survival. Eur. J. Immunol. 28:2685–93 Corcos D, Dunda O, Butor C, Cesbron JY, Lores P, Buchinni D, Jami J. 1995. Pre-B cell development in the absence of lambda5 in transgenic mice expressing a heavy-chain disease protein. Curr. Biol. 5:1140–48 Gu H, Tarlinton D, Muller W, Rajewsky K, Forster I. 1991. Most peripheral B cells are ligand selected. J. Exp. Med. 173:1357–71 Freitas AA, Lembezat MP, Coutinho A. 1989. Expression of antibody V-regions is genetically and developmentally controled and modulated by the B lymphocyte environment. Int. Immunol. 1:342– 54 Viale AC, Coutinho A, Heyman RA, Freitas AA. 1993. V region dependent selection of persistent resting peripheral B cells in normal mice. Int. Immunol. 5:599–605 Freitas AA, Andrade L, Lembezat MP, Coutinho A. 1990. Selection of VH gene repertoires: differentiating B cells of adult bone marrow mimic fetal development. Int. Immunol. 2:15–23 Kouskoff V, Famiglietti S, Lacaud G, Lang P, Rider JE, Kay BK, Cambier JC, Nemazee D. 1998. Antigens varying in affinity for the B cell receptor induce differential B lymphocyte responses. J. Exp. Med. 188:1453–64 Colle JH, Truffa-Bachi P, Freitas AA. 1988. Secondary antibody responses to

LYMPHOCYTE SURVIVAL

65.

66.

67.


68.

69.

70.

71.

72.

73.

74.

75.

thymus-independent antigens. Decline and life-span of memory. Eur. J. Immunol. 18:1307–14 Gray D, Skarvall H. 1988. B-cell memory is short-lived in the absence of antigen. Nature 336:70–73 Schittek B, Rajewsky K. 1990. Maintenance of B-cell memory by long-lived cells generated from proliferating precursors. Nature 346:749–51 Lam KP, Rajewsky K. 1998. Rapid elimination of mature autoreactive B cells demonstrated by Cre-induced change in B cell antigen receptor specificity in vivo. Proc. Natl. Acad. Sci. USA 95:13171–75 Gay D, Saunders T, Camper S, Weigert M. 1993. Receptor editing: an approach by autoreactive B cells to escape tolerance. J. Exp. Med. 177(4):999–1008 Tiegs SL, Russell DM, Nemazee D. 1993. Receptor editing in self-reactive bone marrow B cells. J. Exp. Med. 177:1009–20 Retter MW, Nemazee D. 1998. Receptor editing occurs frequently during normal B cell development. J. Exp. Med. 188:1231–38 DiSanto JP, Guy-Grand D, Fisher A, Tarakhovsky A. 1996. Critical role for the common cytokine receptor gamma chain in intrathymic and peripheral T cell selection. J. Exp. Med. 183:1111–18 Kuo CT, Veselits ML, Leiden JM. 1997. LKLF: a transcriptional regulator of single-positive T cell quiescence and survival. Science 277:1986–90 Oukka M, Ho I-C, de la Brousse FC, Hoey T, Grusby MJ, Glimcher LH. 1998. The transcription factor NFAT4 is involved in the generation and survival of T Cells. Immunity 9:295–304 Caldwell RG, Wilson JB, Anderson SJ, Longnecker R. 1998. Epstein-Barr virus LMP2A drives B cell development and survival in the absence of normal B cell receptor signals. Immunity 9:405–11 Byth KF, Conroy LA, Howlett S, Smith

76.

77.

78.

79.

80.

81.

82.

83. 84.

85.

107

AJ, May J, Alexander DR, Holmes N. 1996. CD45-null transgenic mice reveal a positive regulatory role for CD45 in early thymocyte development, in the selection of CD4`CD8` thymocytes, and B cell maturation. J. Exp. Med. 183:1707–18 Khan WN, Alt FW, Gerstein RM, Malynn BA, Larsson I, Rathbun G, Davidson L, Muller S, Kantor AB, Herzenberg LA. 1995. Defective B cell development and function in Btkdeficient mice. Immunity 3:283–99 Turner M, Gulbranson-Judge A, Quinn ME, Walters AE, MacLennan IC, Tybulewicz VL. 1997. Syk tyrosine kinase is required for the positive selection of immature B cells into the recirculating B cell pool. J. Exp. Med. 186:2013–21 Sha WC, Liou HC, Tuomanen EI, Baltimore D. 1995. Targeted disruption of the p50 subunit of NF-kappa B leads to multifocal defects in immune responses. Cell 80:321–30 Schubart DB, Rolink A, Kosco-Vilbois MH, Botteri F, Matthias P. 1996. B-cellspecific coactivator OBF-1/OCA-B/ Bob1 required for immune response and germinal centre formation. Nature 383:538–42 Adams JM, Cory S. 1998. The bcl-2 protein family: arbiters of cell survival. Science 281:1322–26 Chao DT, Korsmeyer SJ. 1998. Bcl-2 family: regulators of cell death. Annu. Rev. Immunol. 16:395–419 Minn AJ, Swain RE, Ma A, Thompson CB. 1998. Recent progress on the regulation of apoptosis by Bcl-2 family members. Adv. Immunol. 70:245–79 Metcalf D. 1965. Multiple thymus grafts in aged mice. Nature 208:87–88 Wallis V, Leuchars E, Collavo D, Davies AJ. 1976. ‘Hyperlymphoid’ mice. Adv. Exp. Med. Biol. 66:183–89 Berzins SP, Boyd RL, Miller JF. 1998. The role of the thymus and recent thymic migrants in the maintenance of the adult

108

86.

87.

88.


89.

90.

91.

92.

93.

94.

95.

96.

97.

FREITAS n ROCHA peripheral lymphocyte pool. J. Exp. Med. 187:1839–48 Agenes F, Rosado MM, Freitas AA. 1997. Independent homeostatic regulation of B cell compartments. Eur. J. Immunol. 27:1801–7 Begon M, Harper JL, Townsend CR. 1990. Ecology: Individuals, Populations and Communities. Oxford, UK: Blackwell’s Sci. Pianka ER. 1976. Competition and niche theory. In Theoretical Ecology. Principles and Applications, ed. RM May, pp. 114–41. Oxford, UK: Blackwell’s Sci. Freitas AA, Rosado M, Viale A-C, Grandien A. 1995. The role of cellular competition in B cell survival and selection of B cell repertoires. Eur. J. Immunol. 25:1729–38 Freitas AA, Agenes F, Coutinho GC. 1996. Cellular competition modulates the survival and selection of CD8 T cells. Eur. J. Immunol. 26:2640–49 Cyster JG, Goodnow CC. 1995. Antigeninduced exclusion from follicles and anergy are separate and complementary processes that influence peripheral B cell fate. Immunity 3:691–701 Tilman D. 1982. Resource competition and community structure. Monographs in Population Biology, Vol. 17. Princeton, NJ: Princeton Univ. Press Monod J. 1950. La technique de culture continue; theórie et applications. Ann. Inst. Pasteur. 79:390–410 McLean A, Rosado MM, Agenes F, Vasconcelos R, Freitas AA. 1997. Resource competition as a mechanism for B cell homeostasis. Proc. Natl. Acad. Sci. USA 94:5792–97 Ahmed R, Gray D. 1996. Immunological memory and protective immunity: understanding their relation. Science 272:54– 57 Schulter D. 1994. Experimental evidence that competition promotes divergence in adaptive radiation. Science 266:798–800 Tilman D, Kareiva P. 1997. The role of

98.

99.

100.

101.

102.

103.

104.

105.

106.

space in population dynamics and interspecific interactions. Monographs in Population Biology, Vol. 30. Princeton, NJ: Princeton Univ. Press Butcher EC, Picker LJ. 1996. Lymphocyte homing and homeostasis. Science 272:60–66 Cyster JG, Hartley SB, Goodnow CC. 1994. Competition for follicular niches excludes self-reactive cells from the recirculating B-cell repertoire. Nature 371:389–95 Forster R, Mattis AE, Kremmer E, Wolf E, Brem G, Lipp M. 1996. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell 87:1037–47 Penit C, Lucas B, Vasseur F, Rieker T, Boyd RL. 1996. Thymic medulla epithelial cells acquire specific markers by post-mitotic maturation. Dev. Immunol. 5:25–36 Gonzalez M, Mackay F, Browning JL, Kosco-Vilbois MH, Noelle RJ. 1998. The sequential role of lymphotoxin and B cells in the development of splenic follicles. J. Exp. Med. 187:997–1007 Fu YX, Huang G, Wang Y, Chaplin DD. 1998. B lymphocytes induce the formation of follicular dendritic cell clusters in a lymphotoxin alpha-dependent fashion. J. Exp. Med. 187:1009–18 Endres R, Alimzhanov MB, Plitz T, Futterer A, Kosco-Vilbois MH, Nedospasov SA, Rajewsky K, Pfeffer K. 1999. Mature follicular dendritic cell networks depend on expression of lymphotoxin beta receptor by radioresistant stromal cells and of lymphotoxin beta and tumor necrosis factor by B cells. J. Exp. Med. 189:159–68 Przylepa J, Himes C, Kelsoe G. 1998. Lymphocyte development and selection in germinal centers. Curr. Top. Microbiol. Immunol. 229:85–104 MacLennan IC. 1994. Germinal centers. Annu. Rev. Immunol. 12:117–39


LYMPHOCYTE SURVIVAL 107. Liu YJ, Arpin C. 1997. Germinal center development. Immunol. Rev. 156:111–26 108. Hinds-Frey KR, Nishikata H, Litman RT, Litman GW. 1993. Somatic variation precedes extensive diversification of germline sequences and combinatorial joining in the evolution of immunoglobulin heavy chain diversity. J. Exp. Med. 178:815–24 109. Litman GW, Stolen JS, Sarvas HO, Makela O. 1982. The range and fine specificity of the anti-hapten immune response: phylogenetic studies. J. Immunogenet. 9:465–74 110. Cazenave PA, Martinez C, Coutinho A. 1989. Differential L chain expression in the antibody responses to phosphorylcholine of adult bone marrow or peritoneum-derived B lymphocytes. J. Immunol. 142:8–11 111. Mombaerts P, Clarke AR, Rudnicki MA, Iacomini J, Itohara S, Lafaille J, Wang L, Ichikawa Y, Jaenisch R, Hooper M, Tonegawa S. 1992. Mutations in T-cell antigen receptor genes alpha and beta block thymocyte development at different stages. Nature 360:225–31 112. Parrott DV, De Sousa MA, East J. 1966. Thymus-dependent areas in the lymphoid organs of neonatally thymectomized mice. J. Exp. Med. 123:191–204 113. De Sousa MA, Ferguson A, Parrott DM. 1973. Ecotaxis of B cells in the mouse. Adv. Exp. Med. Biol. 29:55–62 114. Mombaerts P, Clarke AR, Hooper ML, Tonegawa S. 1991. Creation of a large genomic deletion at the T-cell antigen receptor beta-subunit locus in mouse embryonic stem cells by gene targeting. Proc. Natl. Acad. Sci. USA 88:3084–87 115. Rahemtulla A, Fung-Leung WP, Schillham MW, Kundig TM, Sambhara SR, Narendran A, Arabian A, Wakeham A, Paige CJ, Zinkernagel RM, Miller RG, Mak TW. 1991. Normal development and function of CD8` cells but markedly decreased helper cell activity in mice lacking CD4. Nature 353:180–84

109

116. Zijistra M, Bix M, Simister NE, Loring JM, Raulet DH, Jaenish R. 1990. b2Microglobulin deficient mice lack CD41 CD8` cytolytic T cells. Nature 344: 742–46 117. Cosgrove D, Gray D, Dierich A, Kaufman J, Lemeur M, Benoist C, Mathis D. 1991. Mice lacking MHC class II molecules. Cell 66:1051–66 118. Tanchot C, Rocha B. 1998. The organization of mature T-cell pools. Immunol. Today 19:575–79 119. Tanchot C, Rocha B. 1997. Peripheral selection of T cell repertoires: the role of continuous thymus output. J. Exp. Med. 186:1099–1106 120. Selin LK, Vergilis K, Welsh RM, Nahill SR. 1996. Reduction of otherwise remarkably stable virus-specific cytotoxic T lymphocyte memory by heterelogous viral infection. J. Exp. Med. 183:2489–99 121. Agenes F, Freitas AA. 1999. Transfer of small resting B cells into immunodeficient hosts results in the selection of a self-renewing activated B cell population. J. Exp. Med. 189:319–30 122. Kitamura D, Kudo A, Schaal S, Muller W, Melchers F, Rajewsky K. 1992. A critical role of lambda 5 protein in B cell development. Cell 69:823–31 123. Adkins B. 1999. T-cell function in newborn mice and humans. Immunol. Today 20:330–35 124. Rueff-Juy D, Liberman I, Drapier AM, Guillon JC, Leclerc C, Cazenave PA. 1991. Cellular basis of the resistance of newborn mice to the pathogenic effects of anti-CD3 treatment. Int. Immunol. 3:683–90 125. Le Bon A, Desaymard C, Papiernik M. 1995. Neonatal impaired response to viral superantigen encoded by MMTV(SW) and Mtv-7. Int. Immunol. 7:1897–1903 126. Billingham RE, Brent L, Medawar PB. 1953. Actively acquired tolerance of foreign cells. Nature 172:603


110

FREITAS n ROCHA

127. Gavin MA, Bevan MJ. 1995. Increased peptide promiscuity provides a rationale for the lack of N regions in the neonatal T cell repertoire. Immunity 3:793–800 128. Holmberg D, Forsgren S, Ivars F, Coutinho A. 1984. Reactions among IgM antibodies derived from normal, neonatal mice. Eur. J. Immunol. 14:435–41 129. Holmberg D, Freitas AA, Portnoi D, Jacquemart F, Avrameas S, Coutinho A. 1986. Antibody repertoires of normal BALB/c mice: B lymphocyte populations defined by state of activation. Immunol. Rev. 93:147–69 130. Thomas-Vaslin V, Freitas AA. 1989. Lymphocyte population kinetics during the development of the immune system. B cell persistance and life span can be determined by the host environment. Int. Immunol. 1:237–46 131. Graziano M, St-Pierre Y, Beauchemin C, Desrosiers M, Potworowski EF. 1998. The fate of thymocytes labeled in vivo with CFSE. Exp. Cell Res. 240:75–85 132. Rocha B, Grandien A, Freitas AA. 1995. Anergy and exhaustion are independent mechanisms of peripheral T cell tolerance. J. Exp. Med. 181:993–1003 133. Tanchot C, Guillaume S, Delon J, Bourgeois C, Franzke A, Sarukhan A, Trautmann A, Rocha B. 1998. Modifications of CD8` T cell function during in vivo memory or tolerance induction. Immunity 8:581–90 134. Qin S, Cobbold SP, Pope H, Elliott J, Kioussis D, Davies J, Waldmann H. 1993. ‘‘Infectious’’ transplantation tolerance. Science 259:974–77 135. Cobbold SP, Adams E, Marshall SE, Davies JD, Waldmann H. 1996. Mechanisms of peripheral tolerance and suppression induced by monoclonal antibodies to CD4 and CD8. Immunol. Rev. 149:5–33 136. Heyman RA, Borrelli E, Lesley J, Anderson D, Richman DD, Baird SM, Hyman R, Evans RM. 1989. Thymidine kinase obliteration: creation of transgenic mice

137.

138.

139. 140.

141.

142.

143.

144.

145.

146.

with controlled immune deficiency. Proc. Natl. Acad. Sci. USA 86:2698 Fulcher DA, Basten A. 1994. Reduced life span of anergic self-reactive B cells in a double-transgenic model. J. Exp. Med. 179:125–34 Hayakawa K, Asano M, Shinton SA, Gui M, Allman D, Stewart CL, Silver J, Hardy RR. 1999. Positive selection of natural autoreactive B cells. Science 285:113–16 Margulis L. 1981. Symbiosis in Cell Evolution. San Francisco: WH Freeman Grossman Z, Paul WE. 1992. Adaptive cellular interactions in the immune system: the tunable activation threshold and the significance of subthreshold responses. Proc. Natl. Acad. Sci. USA 89:10365–69 von Boehmer H. 1990. Developmental biology of T cells in T-cell receptor transgenic mice. Annu. Rev. Immunol. 8:531– 56 Goodnow CC, Cyster JG, Hartley SB, Bell SE, Cooke MP, Healy JI, Akkaraju S, Rathmell JC, Pogue SL, Shokat KP. 1995. Self-tolerance checkpoints in B lymphocyte development. Adv. Immunol. 59:279–368 Merkenschlager M, Benoist C, Mathis D. 1994. Evidence for a single-niche model of positive selection. Proc. Natl. Acad. Sci. USA 91:11694–98 Huesmann M, Scott B, Kisielow P, von Boehmer H. 1991. Kinetics and efficiency of positive selection in the thymus of normal and T cell receptor transgenic mice. Cell 66:533 Lang J, Jackson M, Teyton L, Brunmark A, Kane K, Nemazee D. 1996. B cells are exquisitely sensitive to central tolerance and receptor editing induced by ultralow affinity, membrane-bound antigen. J. Exp. Med. 184:1685–97 Fulcher DA, Lyons AB, Korn SL, Cook MC, Koleda C, Parish C, Fazekas de St. Groth B, Basten A. 1996. The fate of self-reactive B cells depends primarily

LYMPHOCYTE SURVIVAL

147.

148.

149.


150.

151.

152.

153.

on the degree of antigen receptor engagement and availability of T cell help. J. Exp. Med. 183:2313–28 Rocha B, von Boehmer H. 1991. Peripheral selection of the T cell repertoire. Science 251:1225–28 Russell DM, Dembic Z, Morahan G, Miller JF, Burki K, Nemazee D. 1991. Peripheral deletion of self-reactive B cells. Nature 354:308–11 Opferman JT, Ober BT, Ashton-Rickardt PG. 1999. Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 283:1742–45 Hodgkin PD, Lee JH, Lyons AB. 1996. B cell differentiation and isotype switching is related to division cycle number. J. Exp. Med. 184:277–81 Burnet M. 1959. The Clonal Selection Theory of Acquired Immunity. Nashville, TN: Vanderbilt Univ. Press Jerne NK. 1974. The immune system: a web of V-domains. Harvey Lect. 70:93– 110 Coutinho A, Forni L, Holmberg D, Ivars F, Vaz N. 1984. From an antigencentered, clonal perspective of immune responses to an organism-centered, network perspective of autonomous activity in a self-referential immune system. Immunol. Rev. 79:151–68

111

154. Vaz NM, Varela FJ. 1978. Self and nonsense: an organism-centered approach to immunology. Med. Hypotheses 4:231–67 155. Matzinger P. 1994. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12:991–1045 156. Bretscher P, Cohn M. 1970. A theory of self-nonself discrimination. Science 169:1042–49 157. Garchon HJ, Djabiri F, Viard JP, Gajdos P, Bach JF. 1994. Involvement of human muscle acetylcholine receptor alphasubunit gene (CHRNA) in susceptibility to myasthenia gravis. Proc. Natl. Acad. Sci. USA 91:4668–72 158. Perelson AS, Oster GF. 1979. Theoretical studies of clonal selection: minimal antibody repertoire size and reliability of self–non-self discrimination. J. Theor. Biol. 81:645–70 159. De Boer RJ, Perelson AS. 1993. How diverse should the immune system be? Proc. R. Soc. London Ser. B 252:171–75 160. Maynard-Smith J. 1982. Evolution and the Theory of Games. Cambridge, UK: Cambridge Univ. Press 161. Nowak MA, Sigmund K. 1998. Evolution of indirect reciprocity by image scoring. Nature 393:573–77 162. Turner PE, Chao L. 1999. Prisoner’s dilemma in an RNA virus. Nature 398:441–43



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


NONCLASSICAL MHC CLASS II MOLECULES Christopher Alfonso and Lars Karlsson The R.W. Johnson Pharmaceutical Research Institute, 3210 Merryfield Row, San Diego, California 92121; e-mail: [email protected]


Key Words antigen processing, HLA-DM, HLA-DO Abstract Major histocompatibility complex (MHC) class II molecules are cell surface proteins that present peptides to CD4` T cells. In addition to these wellcharacterized molecules, two other class II–like proteins are produced from the class II region of the MHC, HLA-DM (DM) and HLA-DO (DO) (called H2-M, or H2-DM and H2-O in the mouse). The function of DM is well established; it promotes peptide loading of class II molecules in the endosomal/lysosomal system by catalyzing the release of CLIP peptides (derived from the class II–associated invariant chain) in exchange for more stably binding peptides. While DM is present in all class II– expressing antigen presenting cells, DO is expressed mainly in B cells. In this cell type the majority of DM molecules are not present as free heterodimers but are instead associated with DO in tight heterotetrameric complexes. The association with DM is essential for the intracellular transport of DO, and the two molecules remain associated in the endosomal system. DO can clearly modify the peptide exchange activity of DM both in vitro and in vivo, but the physiological relevance of this interaction is still only partly understood.

INTRODUCTION The recognition of antigen-presenting major histocompatibility complex (MHC) class II molecules by CD4` T cells is a crucial component of the defense against pathogens. MHC class II expression is normally restricted to a subset of antigen presenting cells–dendritic cells, macrophages, B cells, and thymic epithelium–but a number of other cell types can be induced to express class II molecules after stimulation with cytokines, in particular interferon-c. The class II molecules at the cell surface are heterotrimeric complexes consisting of two transmembrane glycoprotein chains (a and b) that form a binding scaffold for the third component, a peptide of 11–20 amino acids. The peptides are derived from proteins of endogenous or exogenous origin that have been degraded in the endosomal or lysosomal system of the presenting cell (1). Class II molecules, like other transmembrane proteins, are translocated into the endoplasmic reticulum (ER) after synthesis, where they associate with a third protein, the invariant chain (Ii). This molecule is a type II transmembrane protein 0732–0582/00/0410–0113$14.00

113


114

ALFONSO n KARLSSON

that serves as a class II–specific chaperone that promotes the exit of class II-Ii complexes from the ER and prevents class II molecules from binding peptides and unfolded proteins in the ER and in the secretory pathway (2). A targeting motif in the cytoplasmic tail of Ii directs the complexes from the secretory pathway into the endosomal system, where the Ii is rapidly degraded and shed from the class II molecules. However, an Ii fragment, called CLIP, which occupies the peptide binding groove of the class II molecule, is in most cases not spontaneously released (3). The CLIP fragment serves as a substitute peptide that protects the class II binding pocket from collapsing both during intracellular transport and after Ii degradation in the endosomal system. Binding of antigenic peptides, generated from endocytosed proteins, requires an empty yet open binding site, and therefore CLIP has to be released while the open binding site needs to be stabilized to allow the binding of other peptides. DM has been well documented to mediate both of these functions, thus promoting the binding of antigenic peptides. After acquiring peptides, the class II molecules are transported to the cell surface via routes that are largely unknown. The general importance of DM and its mouse equivalent, H2-M, in the processing of antigens for CD4` T cells is now well established through a number of studies using mutant cell lines, isolated protein components, and genetically modified mice. Similar techniques have been used to study DO (and H2-O), but though it is clear that DO changes the function of DM, the physiological relevance of the DM-DO association is still being unraveled. In this review we present an overview of the current knowledge of the two nonclassical class II molecules, including their function and their influence on antigen presentation.

GENES AND TISSUE DISTRIBUTION The human MHC (also called HLA-D—human leukocyte antigen) is localized on the short arm of chromosome 6, while the equivalent mouse region (called H2) is localized on chromosome 17. The large interest in MHC-encoded molecules resulted in the class II genes (as well as the class I genes which are also located in this region) being among the earliest mammalian genes to be cloned and fully sequenced. Analysis by serology and of their ability to stimulate allogeneic T cells had revealed the existence of the three classical human class II molecules, DR, DQ, and DP, and the genes encoding these molecules had all been identified and characterized by the mid-1980s. In addition to these genes, however, two other genes encoding proteins similar to class II a and b chains, respectively, were found by cross-hybridization. These genes were called DZA (4, 5) (later renamed DNA and finally DOA) and DOB (6, 7), respectively, but no protein products were detected at the time. In the mouse, two b chain genes, Ab2 (7, 8) (homologous to DOB and later renamed Ob) and Eb2 (9, 10), were found, in addition to the genes encoding the two classical class II molecules, H2-A and H2-E. The number of class II genes exceeded what had been expected from

NONCLASSICAL MHC CLASS II MOLECULES

115

functional data, and it seemed likely that all or most of the genes in the class II region had been located. This assumption turned out to be wrong, and a number of genes encoding proteins involved in antigen processing for MHC class I molecules, as well as the genes encoding DM, have since been mapped to the class II region in both human and mouse (11). The majority of the human and mouse class II regions have now been sequenced and a number of other new genes have been identified (12), but since none of these have structural homology to class II genes they are not discussed here.


HLA-DM and H2-M The genes encoding DM and H2-M are quite dissimilar from other class II genes (including the DO and H2-O genes) (13, 14), and they were identified in a systematic search for genes in the class II region, rather than through crosshybridization (15). The overall amino acid identity between the DM chains and the corresponding a and b chains of the other human class II loci is only about 30%, while the DP, DQ, DR, and DO chains are about 60% identical when compared to each other (13, 14). Indeed, the b2 domain of DMb is almost as related to class I molecules as to class II molecules, suggesting that the DM/H2-M genes have evolved along different lines from other class II molecules. The DM a and b chain genes are located close to each other on the centromeric side of the TAP and Lmp genes (which are part of the peptide-generating machinery for class I molecules) (16). The human DM locus consists of one a chain gene, DMA, located on the centromeric side of a single b chain gene, DMB (13, 17). The mouse H2-M locus has a similar organization but contains two b chain genes, Mb1 and Mb2 (18, 19). The two b chain genes show a very high degree of identity– 96% in the coding regions, 93% if the entire genomic sequence is compared (19)–and they are likely to represent a recent gene duplication since other species including rats have only one Mb gene (20). Interestingly, the majority of amino acid differences are located in the b1 region (19, 21, 22), i.e. the region corresponding to the peptide-binding region of classical class II molecules, but it is unknown whether the presence of one or the other chain influences the efficiency or specificity of H2-M-mediated peptide exchange. It is somewhat unclear whether Mb1 and Mb2 are equally well expressed in vivo (21–23), but H2-M molecules consisting of either ab1 or ab2 heterodimers appear to be functional (21, 24, 25), although no systematic study comparing the two molecules has been published. Classical class II molecules display an extraordinary degree of polymorphism concentrated to the amino acids that contribute to the peptide binding site. In contrast, analysis of the DM and H2-M genes has revealed limited polymorphism where the minor differences that do occur appear to be spread throughout the coding regions (19, 21–23, 26, 27). A number of studies have investigated potential links between different DM haplotypes and autoimmune disease, but clear correlations have not yet been found (28–32). No functional comparisons have


116

ALFONSO n KARLSSON

been made between DM or H2-M molecules of different haplotypes, but it appears likely that the peptide exchange activity will not be seriously influenced by the polymorphisms. DM and H2-M appear to be expressed in all class II–expressing cells, although the level of transcription is generally lower than that of conventional class II molecules. Interferon-c, which increases the expression of class II genes, also increases the expression of the DM and H2-M genes (13, 14). The promoter regions of these genes share the characteristic motifs (referred to as S, X, X2, and Y boxes) (17, 19) that are present in other class II genes; the transcriptional activators CIITA and RFX-5, which control class II transcription, are also necessary for efficient transcription of the DM and H2-M genes (33, 34). However, while class II expression appears to be totally dependent on these transcription factors, a low basal level of H2-M expression seems to persist both in mice deficient for CIITA (35) and in mice deficient for RFX-5 (36), suggesting that differences in transcriptional control do exist. This notion is also reinforced by the isolation of mutant cell lines that appear to have normal expression of class II molecules, but where the DM genes are not transcribed normally due to an unknown genetic defect mapping to chromosome 6p (37). Genes encoding DM-like proteins have been characterized in a number of other species (20, 38, 39) beside mouse and human; although the gene products have not been analyzed, they are likely to be the functional equivalents of DM.

HLA-DO and H2-O In contrast to the DM genes, the genes encoding DO a and b are as similar to the classical class II genes as these are to each other, and as a consequence of this similarity both genes were found using low stringency cross-hybridization (4–7). Although they were initially proposed to form a pair (5), several pieces of evidence suggested that this was not the case. First, in contrast to other class II a and b chain genes that are located pairwise, close to each other, the two DO genes are located quite far from each other (200 kb), with a number of other genes located in between the two. Second, while the transcription of the a chain appeared to be similar to that of other class II genes (although the level of transcription is at least tenfold lower— 4), the transcription of the b chain was distinctly different (6). Thus, DOb cannot be induced by interferon-c in most cell types (although exceptions to this finding have been reported— 40), and its expression is independent of CIITA (41). The transcription from the DOB promoter is also distinctly weaker than the transcription from conventional class II genes, and one study even concluded that the gene was transcriptionally silent (42). The fact that DOa and DOb actually are expressed and do associate with each other was only shown after the identification of the equivalent mouse molecule, H2-O (43). The H2-O b chain, initially called Ab2 (44), was discovered due to its homology to other b chains, but the H2-O a chain was mapped and cloned based on its association with H2-Ob at the protein level (45). The H2-O



117

chains are both transcribed independently of CIITA, and H2-O is not upregulated by interferon-c (35, 46). In contrast, IL-4 appears to result in a moderate upregulation of H2-O protein expression in B cells (L Karlsson, unpublished data). H2-O and DO both have a more restricted tissue distribution than do other class II molecules, including DM (6, 43, 46–49). Thus, in peripheral tissues, H2O expression is largely restricted to B cells, while the vast majority of dendritic cells and macrophages do not express this molecule. The expression pattern of DO is less clear because this molecule has been reported to be expressed both by dendritic cells and by a melanoma cell line, as well as by B cells (49). The expression of DO by dendritic cells does not appear to be a general phenomenon, however (50; L Karlsson, unpublished), and it is possible that subsets of dendritic cells have different patterns of DO expression (A Vogt, H Kropshofer, personal communication). In the thymus, DO expression has been reported in both cortical and medullary epithelial cells (49), while immunohistochemistry showed H2-O expression to be most prominent on a subset of medullary epithelial cells (43, 47). Intriguingly, the thymic (but not the peripheral) H2-O staining was present also in mice lacking functional H2-Oa chains (51); (while this finding remains unexplained, analysis by in-situ hybridization has confirmed that mRNA for both H2-O chains are present in the thymic medulla (L Karlsson, unpublished data). The human DO chains have not been extensively analyzed for polymorphism, but the studies to date suggest that DO, like DM, is essentially nonpolymorphic (5, 7, 52, 53). This conclusion is also supported by sequence analysis of the mouse molecules where only minor differences could be found between different class II haplotypes (43, 54). The few allelic differences are spread throughout the open reading frames, suggesting that they are not likely to be functionally relevant.

H2-Eb2 The mouse class II region contains yet another class II b chain gene that may be classified as a nonclassical class II gene; Eb2. This gene, which does not appear to have any human homologue, is transcribed similarly to Ob (10) (i.e. it is expressed in B cells, but not macrophages, and is not inducible by interferon-c). No corresponding a chain gene has been identified, and it is unknown whether Eb2 is part of any functional protein.

INTRACELLULAR TRANSPORT AND SUBCELLULAR DISTRIBUTION The interaction between class II molecules and Ii, as well as the functional consequences of this interaction, have been very well characterized (reviewed in 2, 3, 55). Class II ab dimers rapidly associate with the Ii immediately after entry into the endoplasmic reticulum to form nonameric (abIi)3 complexes. The binding to Ii inhibits the class II molecules from associating with peptides in the endo-


118

ALFONSO n KARLSSON

plasmic reticulum and in the secretory pathway, and maybe more importantly, it protects from the binding of nascent unfolded protein chains to class II molecules in the ER (56, 57). Invariant chain binding to class II molecules is largely mediated through the CLIP region of Ii (amino acids 81–104), which is thought to occupy the peptide binding groove of the class II molecule (58–61). In some class II haplotypes, the affinity to the CLIP peptide sequence is low, however (62), and other regions of Ii are important for the association with class II molecules (61, 63–65). Once assembled, the nonameric class II-Ii complexes are rapidly transported out of the ER into the secretory pathway. Two extensively characterized di-leucine-like targeting motifs in the cytoplasmic tail of the Ii direct the class IIIi complexes into the endosomal/lysosomal system (66–69), where Ii is removed and peptide loading of the class II molecules occurs (see below), before the assembled peptide-class II complexes are transported to the cell surface. It is presently unclear where class II-Ii complexes enter the endosomal system. A number of morphological and biochemical studies have suggested a direct path from the trans-Golgi network (TGN) to late endosomal compartments (70–72), but class II-Ii complexes have also been detected at the plasma membrane, suggesting that a fraction of newly synthesized class II molecules reach late endosomes or lysosomes via the cell surface or early endosomes (73–76). In the absence of Ii expression, class II molecules are largely retained in the ER due either to aggregation and functional inactivation or to binding of unfolded proteins, and only a minority of the synthesized molecules reach the plasma membrane (77–79). However, some of the class II molecules that do appear on the cell surface can acquire peptides after internalization and recycling back to the plasma membrane (80– 82). The cytoplasmic tails of H2-Ab, and of DR a and b have been shown to be able to mediate internalization of mature Ii-free class II molecules from the cell surface into recycling endosomes where peptide loading occurs under certain conditions (81, 82). Most peptide loading does not involve recycling class II molecules, however, but requires newly synthesized Ii-associated class II proteins (1, 83). The endosomal or lysosomal compartments where newly synthesized class II molecules acquire peptides have been the subject of intensive study in recent years. Morphological and sub-cellular fractionation techniques have been used to characterize and isolate compartments that are enriched in their class II content and that are likely to be the location where peptide loading occurs. Initial studies using electron microscopy suggested that the majority of class II molecules were localized in multilamellar structures with lysosomal characteristics, called MHC class II compartments (MIIC) (70, 84). A number of subsequent studies have described compartments with both endosomal and lysosomal characteristics as being important for peptide loading (85–89), and it has become apparent that celltype dependent differences (72, 90) as well as haplotype-dependent differences (91) can explain many if not all of the discrepancies seen in different experimental systems. The term MIIC has with time been expanded to include class II–


119

containing multivesicular endosomal compartments as well as lysosomal compartments (72), but it remains uncertain whether class II–loading compartments are really specialized, or whether they are part of the normal endosomal/lysosomal system present in all cells. This question lies outside the scope of this review, however, and has been excellently reviewed elsewhere (1). After peptide acquisition, the mature class II–peptide complexes are transferred to the cell surface where they persist for extended periods of time, displaying the antigenic peptides to CD4` T cells.


Intracellular Transport of DM/H2-M The DMA and Ma genes encode mature proteins of about 26 kDa, but the addition of two carbohydrate residues increases the actual molecular weight to 33–35 kDa. The mature core proteins encoded by the DMB and Mb genes have a predicted size of 27 kDa, and while this is the actual size of Mb2, DMb (and probably Mb1) are modified by the addition of one carbohydrate moiety to have an actual size of 30–31 kDa (21, 92). Newly synthesized DM/H2-M chains assemble in the ER after synthesis to form heterodimers, but unlike newly synthesized conventional class II molecules that rapidly associate with the Ii, physical association between DM and Ii has been difficult to show (92). In contrast, complexes of H2M and Ii have been co-immunoprecipitated after solubilization in digitonin (a mild detergent that tends to preserve weak protein-protein interactions) or after chemical cross-linking, but the significance of this interaction is uncertain (93). Assembly and exit from the ER of either H2-M or DM clearly do not require association with the Ii, since expression of the two DM (or H2-M) chains into a number of different cell lines that do not express Ii results in efficient assembly and intracellular transport of apparently functional DM heterodimers (21, 93–95). While class II molecules transiently pass through the endosomal system before reaching the cell surface, mature DM and H2-M molecules are permanent residents of the endosomal/lysosomal system (21, 96). Early immuno-electron microscopy studies showed that the majority of DM was located in multivesicular or multilamellar late endosomal or lysosomal structures (MIICs) (96, 97), but subcellular fractionation experiments in combination with further electron microscopy have shown that DM is present throughout the endosomal system (72, 98). In addition, small amounts of DM can be detected at the plasma membrane (99; A Vogt, H Kropshofer, personal communication). Both DM and H2-M are directed to (and probably maintained in) the endosomal system by a tyrosinebased targeting motif (amino acids YTPL) located in the cytoplasmic tail of the b chain of both molecules (93, 94, 100). Transfection experiments in HeLa or NRK cells (neither of which express endogenous DM or Ii) have shown that removal or mutation of the essential tyrosine or leucine residues in the motif results in disrupted sorting of DM/H2-M and accumulation of the mutated


120

ALFONSO n KARLSSON

molecules at the cell surface. Somewhat surprisingly, considering the apparently low affinity between H2-M and Ii, co-expression with full length (but not targeting-deficient) Ii could rescue both the endosomal delivery and the peptide exchange function of the mutated H2-M, suggesting that the targeting motif in the H2-M tail is not essential for the localization of H2-M to endosomal/lysosomal structures in cells where Ii is present (93). Similar Ii-mediated rescue of DM function in cells transfected with targeting-deficient DM has also been reported, suggesting that DM does indeed associate with Ii to some extent (101), although this association may be weaker than the association between Ii and H2-M. How DM/H2-M reaches endosomal compartments has not been extensively studied, but transfection of a dominant negative form of clathrin that disrupts the formation of clathrin-coated vesicles suggests that DM and Ii-class II complexes may transit from the trans-Golgi network to the endosomal system in distinct vesicular compartments (102). Once in the endosomal system, DM and class II molecules must come in contact with each other in the same intracellular compartments, but it is not entirely clear whether the sorting signals in Ii and DM/ H2-M direct their respective cargoes to the same or to different compartments. One report has suggested that Ii–class II complexes are initially delivered to DMfree organelles where Ii is degraded before they are transported into DM-containing, peptide-loading compartments (103). Most available data seem to suggest, however, that Ii, class II, and DM are present in overlapping compartments (98, 104–106). The protease sensitivity of Ii complicates accurate detection in experimental systems, and the absence of detectable Ii does not necessarily mean that Ii was not responsible for the delivery of class II molecules to a particular compartment before its degradation. It is not known whether mature peptide-containing class II molecules (destined for delivery to the cell surface) are sorted away from DM and immature Iiassociated class II molecules in the endosomal system, or whether all the different molecules are deposited at the plasma membrane together, followed by rapid endocytosis of DM and possible residual Ii. Direct fusion of multivesicular class II–containing endosomes with the plasma membrane has been shown to occur (107), resulting in both the delivery of class II molecules to the plasma membrane and the secretion of the internal vesicles (called exosomes) into the surrounding medium (107, 108), but it is not clear if the majority of class II molecules are delivered to the plasma membrane by this mechanism. The vesicles have been reported to be enriched for class II molecules as well as for CD86 (B7.2) and several tetraspan molecules (CD37, CD53, CD 63, CD81 and CD82), while DM (and Lamp-1 and -2) were largely absent from the vesicles and instead concentrated on the surrounding endosomal membrane, suggesting that in this situation at least, DM and class II molecules may be sorted prior to delivery to the cell surface (109). Interestingly, several of the tetraspan molecules have been shown to associate with class II molecules as well as with DM and DO, but the significance of this finding is unclear (110).


121


Intracellular Transport of DO/H2-O The proteins encoded by the DO and H2-O genes are structurally similar to other class II proteins, consisting of a chains of about 25 kDa and b chains of about 28 kDa. Addition of 2 and 1 carbohydrate residues, respectively, results in an increase of the actual molecular weight to 32–34 kDa for the mature a chains and 33–34 kDa for the mature b chains (43, 48). The relatively large size of the b chains (compared to other class II b chains) is mainly a reflection of the unusually long cytoplasmic tail, present in both mouse and human molecules. After translocation into the ER, the a and b chains assemble to form heterodimers, but the dimerization was found to be relatively inefficient in transfected HeLa cells expressing transfected DO a and b alone, and exit from the ER of the assembled DO dimers could not be detected (48). The inefficient intracellular transport of DO is reminiscent of the transport of class II molecules expressed in the absence of Ii, but in the case of DO, co-expression of Ii resulted in only marginally improved transport out of the ER, and the association between the two molecules was substoichiometric. In contrast, co-expression of the two DM chains resulted in formation of heterotetrameric DMDO complexes that could effectively exit the ER (48). Analysis of B cell lines confirms that the DM-DO interaction is not a transfection artifact but occurs also in cells that express DM and DO endogenously (41, 48, 111, 112). A similar association between H2-O and H2-M is also seen in mouse B cells (48, 51). Both DO and H2-O do bind Ii to some extent in B cells (43, 48, 49), but this interaction is clearly less important than the interaction with DM and H2-M. Thus, H2-O dimers are formed in H2-Ma-deficient mice to some extent, but in the absence of functional H2-M, the H2-O molecules are not transported out of the ER and are instead rapidly degraded (48). In B cells the DMDO complexes are transported to endosomal/lysosomal compartments, and at steady state DO, like DM, is mainly localized in lysosomes (41, 48, 111). Most of the lysosomal DO molecules are associated with DM, and indeed it is uncertain whether free DO exists at all outside of the ER (48, 112). It is conceivable that DMDO complexes dissociate in lysosomes to generate free DM and DO, either spontaneously or in response to cellular activation or intracellular signaling events, but this has not been demonstrated so far. Free DM is present in DO-expressing cells, but the majority of DM seems to be associated with DO, both during intracellular transport and in lysosomes (48, 112). It is not known whether DMDO complexes have different targeting information than DM alone, and it is conceivable that DM and DMDO may be sorted to different compartments of the endosomal/lysosomal system, although no available data support this hypothesis. The long cytoplasmic tail of DOb contains two potential endosomal targeting motifs, and a recent report has shown that the DO tail may direct a transfected DRa-DOb heterodimer to lysosomal compartments, even though this chimeric molecule was also present at the cell surface (113). The targeting motifs are not conserved in the H2-Ob tail, however, and further study

122

ALFONSO n KARLSSON

will be necessary to determine whether the cytoplasmic tails of DO and H2-O are important for DO or DMDO localization and function.


FUNCTIONAL ASPECTS OF DM By the late eighties, it was becoming apparent that antigen processing and peptide aquisition by class II molecules were occurring in endosomal or lysosomal compartments of antigen presenting cells, but possible molecular requirements for this process were not known. A number of studies had shown that functional peptide– class II complexes could be formed by incubating antigenic peptides with class II molecules, either on cells or using purified class II molecules. Although this process was much slower than peptide loading in vivo, it was not immediately apparent that specific molecules would be necessary for mediating peptide loading. The first direct demonstration that peptide loading must be a regulated process came from the analysis of a panel of mutant cell lines that were defective in their ability to present protein antigens to CD4` T cells, but that had normal expression of class II proteins and normal or improved ability to present exogenously added peptides (114). These cell lines had been selected based on their failure to be recognized by 16.23, an HLA-DR3-specific monoclonal antibody that recognizes a subset of peptide-DR3 complexes (115). The antigen presentation defect was not limited to DR3-restricted antigens but was extended also to DP- and DQrestricted antigens. Biochemical analysis showed that the class II molecules from the mutant cells were less resistant to incubation in low concentrations of SDS than the molecules from wild-type control cells, leading to the suggestion that the class II molecules from the mutant cells might not contain any bound peptides. Subsequent studies in these and other mutant cell lines with similar phenotypes showed that the class II molecules were actually not empty, but instead contained a limited set of peptides, mainly derived from the Ii (116, 117). These peptides, called CLIP (class II associated invariant chain peptides), are intermediates in the degradation of Ii also in normal class II–expressing cells, but while they are normally released and replaced by antigenic peptides (118), CLIP release is blocked in the mutant cell lines. Genetic analysis showed that the defect in the mutant cell lines mapped to the class II region of the MHC (119, 120), and further work revealed that the deficiency could be rescued by transfection of functional HLA-DM genes (24, 121). How a class II–related molecule like DM could promote the loading of antigenic peptides onto conventional class II molecules was not immediately obvious and led to a period of intense study by a number of groups aimed at revealing the molecular mechanism of DM function.

DM Catalyzes Peptide Exchange In Vitro DM does have structural similarity to peptide-binding class II molecules; one initial suggestion was that DM, like other class II molecules would be able to



123

bind peptides and in particular CLIP peptides released from maturing class II molecules. The notion that DM might function as a ‘‘CLIP sink’’ does not appear to be correct, however, and later experiments have not been able to show any stable peptide association with DM (122, 123). This conclusion is further strengthened by the fact that the recently solved crystal structures of both DM and H2M show a very reduced binding pocket, which is too small to accommodate peptides (124, 125). However, experiments using purified DR and DM molecules, eiher isolated from B cell lines or produced as recombinant proteins, revealed that DM alone is able to facilitate the release of CLIP peptides from CLIP–class II complexes, while at the same time promoting the binding of antigenic peptides (122, 126–128). CLIP may be the most important substrate peptide in vivo, but the peptide releasing effect of DM in vitro is not limited to CLIP, and in fact other class II–associated peptides, as well as longer Ii-derived protein fragments, can also be exchanged for antigenic peptides by incubation with DM. In addition to increasing the release of peptides from class II molecules, DM also increases both the rate of peptide binding and the maximal level of peptide binding in vitro (122, 126, 127). This effect can partly be explained by the increased release of previously bound peptides. As more binding sites become available, more new peptides can be bound, but the increased binding does also appear to be due to a direct chaperone effect of DM resulting in the stabilization of empty class II molecules, thus allowing them to maintain the ability to bind peptides (123, 129). Peptidefree class II molecules appear to be very unstable, and prolonged incubation at acidic pH in the absence of suitable peptide or DM results in an irreversible loss of ability to bind peptides (123, 130–132). Further analysis has extended the initial peptide release experiments to suggest that the dissociation rates of all class II– peptide complexes are actually increased by DM. The rate of dissociation in the presence of DM has been shown to be directly proportional to the intrinsic rate of dissociation of the peptide from the class II molecule (i.e. the dissociation in the absence of DM) (133, 134), and although DM is not strictly an enzyme, DMmediated peptide release appears to follow general rules for enzyme kinetics (133, 135). The notion that DM functions catalytically also in vivo is supported both by the long half-life of DM (.24 hours) (92) and by the fact that the intracellular concentration of DM is lower than the concentration of DR. In B cell lysates the ratio between DR and DM has been estimated to be 20:1 (122), while the more relevant quantitation of DR and DM levels in endosomal/lysosomal loading compartments has suggested that DR is present in a four- to fivefold excess over DM (106, 123). It should be noted, however, that relatively minor differences in DM expression have been reported to influence the efficiency of peptide exchange in vivo, under certain conditions (136). It has not been experimentally determined how DM causes the release of class II–associated peptides, but a direct interaction between DM and the class II molecule appears to be required (122, 126, 127, 137). The strength of the interaction between DM and class II molecules is unclear; reliable measurements have not been published, but DM-DR complexes can be isolated from cell lysates after


124

ALFONSO n KARLSSON

solubilization using acidic conditions and gentle detergents (123, 129, 138). Many of these complexes do not contain bound peptides (123, 129), but it is presently not clear whether the stable interaction between DM and DR requires that the class II molecules are empty or whether the lack of peptides is a consequence of the extraction procedure. DR molecules containing a mutant form of DRa that has an additional glycosylation site in the membrane-proximal a2 domain, as well as several other mutations mapping to the same lateral face of DR, are unable to interact productively with DM both in vivo and in vitro, suggesting that the interacting DM and DR molecules are positioned parallel to each other in the membrane (126, 137; E Mellins, personal communication). The face of DR that contains these mutations corresponds to the end of the peptide-binding groove, which holds the N-terminus of the peptide, and this is likely to be the face that interacts with DM (124). The crystal structures have been determined for a number of class II molecules, including DR3-CLIP, but there are no obvious structural differences between the CLIP-containing molecule and other more stable class II–peptide complexes that could give more specific indications for how DM would bind to DR (59, 139, 140). However, both class II molecules and DM change conformation upon acidification, and these altered conformations may favor the association between the two molecules (141–144). The binding of the peptide to the class II–binding pocket is unusual in that it largely depends on a network of hydrogen bonds formed between the backbone of the bound peptide and the side chains of the amino acid residues lining the peptide binding pocket (59, 139). In addition to this network of hydrogen bonds, ‘‘good’’ class II–binding peptides are characterized by the presence of one or more anchor residues that fit in suitable pockets of the binding groove as well as by the absence of residues that are incompatible with binding (128, 145, 146). A likely explanation for how DM increases the rate of peptide exchange is that interaction between DM and a class II molecule stabilizes (or induces) a conformational change in the class II molecule that to some extent disrupts the hydrogen bond network holding the peptide in place (59, 124). Such structural changes would not have to be very substantial– a relatively minor dislocation of the two a-helices that line the peptide-binding pocket would be sufficient to ‘‘unzip’’ the hydrogen bonds locking the peptide in place (124). In this situation, the intrinsic affinity of the peptide for the open binding pocket may determine whether the peptide will be released or retained. It is not known how DM is released from the class II molecules when peptide binding has occurred, or indeed whether peptide binding is required for the release of DM. One likely possibility is that binding of a ‘‘good’’ peptide in the binding pocket of the class II molecule induces a conformational change that is sufficient to counteract the effect of DM, resulting in its release from the class II molecule. Data showing that DM appears to associate more stably with empty class II molecules than with peptide-containing class II molecules support this hypothesis (123, 129, 135). Alternatively, DM may be released from the class II molecule independently of the quality of the bound peptide. In this scenario ‘‘good’’ peptides would be trapped in the binding pocket more often than ‘‘bad’’ peptides,


125


simply because their dwell time in the open DM-stabilized binding pocket would be longer. Repeated cycles of association and dissociation between DM and the class II molecule would result in accumulation of ‘‘good’’ peptides also in this case, and the DM-class II interaction would be disrupted either by pH-induced conformational changes occurring during transfer from lysosomal compartments to the cell surface, or by the physical separation of the two molecules by differential sorting, either in the endosomal system or maybe even at the cell surface, depending on where class II and DM are actually sorted away from each other. This model for DM function is supported by data showing that class II molecules containing ‘‘good’’ stably bound peptides can also be found associated with DM (129, 138). The two models are not mutually exclusive, but whether one or the other (if any) is more correct will have to be determined experimentally. A schematic view of how DM functions is outlined in Figure 1 (see color insert).

DM/H2-M Function In Vivo The antigen presentation defects in human cells lacking DM are quite striking, and they are shared by all three types of classical class II molecules (DR, DP and DQ) expressed by these cells (114). However, when human cells transfected with mouse class II molecules were analyzed for their ability to present exogenous antigens, the defects were generally less prominent and appeared to be haplotypeas well as cell-type-dependent (147, 148). The relatively small differences seen between the presence or absence of DM indicated that certain class II haplotypes, or indeed all mouse class II molecules, were less dependent on DM activity than initially suggested. However, the generation of H2-M-deficient mice by three independent groups very convincingly showed that the dependence on DM (or H2-M) is not species-specific (149–151). The level of class II expression at the cell surface of APCs is normal in these mice, but the class II molecules (H2-Ab) have abnormal conformation, a finding explained by the fact that they are overwhelmingly associated with CLIP peptides. The fraction of class II molecules that bind CLIP is so high that an accurate estimate of the CLIP content is difficult to make, but analysis of H2-Ab-associated peptides suggests that .99% of the molecules are loaded with this peptide (A Rudensky, personal communication). In contrast to the mutant human cell lines, which can present many endogenously synthesized antigens (152), as well as antigens internalized by efficient receptormediated uptake (153), the H2-Ab molecules from the H2-M-deficient mice are essentially incapable of presenting antigens, whether expressed endogenously (95, 149, 154) or internalized either by fluid-phase (150, 151) or receptor-mediated endocytosis (C Alfonso, L Karlsson, unpublished data). In addition, the ability of APCs from mutant mice to present pre-processed exogenously added peptides is lower than the ability of APCs from wild-type littermates (95, 150, 151). In contrast, the capacity to raise responses to peptides in vivo is similar between mutant and wild-type mice (155). The almost total inability to present protein


126

ALFONSO n KARLSSON

antigens by H2-Ab is a haplotype-, rather than a species-dependent effect, and substantial fractions of H2-Ak and H2-Ek/b molecules expressed in the H2-Mdeficient animals (through breeding with H2-Ak or H2-Eak transgenic animals) are occupied with peptides other than CLIP, although their CLIP content is also increased compared to the H2-Ak and H2-Ek/b molecules from wild-type controls (156). Despite the lower CLIP content, the ability to form SDS-stable complexes is decreased for the H2-Ak and H2-Ek/b molecules from the H2-M-deficient mice, as is their ability to present exogenously added protein antigens to T cell hybridomas. In contrast, the ability to present exogenously added peptides is normal. The differences in antigen presentation and SDS stability of class II molecules between transfected human DM-deficient cell lines and antigen-presenting cells from H2-M-deficient mice are significant and are at least partly an effect of species mismatching, since DM interacts poorly with at least one class II allele, H2-Ad (25, 157, 158). Differences in Ii processing may also account for some of the discrepancies, since a fraction of CLIP peptides eluted from human DR molecules have extended N-termini compared with CLIP peptides eluted from mouse class II molecules (151, 159). This N-terminal extension, which has been reported to promote self-release of CLIP from DR molecules (160, 161), may be present also on CLIP bound by mouse class II molecules expressed in human cell lines, thus explaining their relative DM independence. The general importance of H2-M for peptide loading and peptide editing also in the absence of Ii is evident in mice lacking expression of both Ii and H2-M. The antigen presenting cells from these mice have low levels of class II (i.e. H2Ab) expression, comparable with the class II expression of APCs from Ii singledeficient mice (95, 162–164). The cells can present exogenously added peptides almost as well as the APCs lacking only Ii, but they have drastically decreased ability to present endogenously synthesized self-peptides, suggesting that H2-M is required not only for the removal of CLIP peptide, but also for the formation of stable peptide-class II complexes (95). In the absence of Ii, the peptide-binding pocket of the class II molecules is likely to be filled with various unstably bound endogenous peptides or proteins, and H2-M may be required to exchange these for better fitting endogenous peptides. Thus, H2-M appears to have a chaperonelike effect in vivo, consistent with results obtained with purified molecules in vitro. The antigen presentation defects in H2-M-deficient mice have expanded the understanding of antigen processing, but an equally important aspect is their contribution to the understanding of thymic selection. Thus, CD4` T cell selection in H2-M-deficient mice (expressing H2-Ab) is moderately decreased (by 50% to 70%) when compared to wild-type littermates (149–151). The T cells selected appear to have grossly normal repertoires of T cell receptor expression, but closer analysis has revealed that several different T cell receptors cannot be positively selected in the H2-M-deficient mice (154, 155, 165), and the receptors that are selected have more limited diversity than receptors from H2-M-expressing mice (166). A large part of this cell population (70%–80%) (165) appears to be made



127

up of cells that have not undergone normal negative selection (due to the absence of self-peptides other than CLIP in the thymic medulla, where negative selection occurs), and the cells react vigorously to APCs from MHC-matched wild-type littermates. Despite this reactivity, the CD4` T cells from the H2-M-deficient mice appear to have low affinity for wild-type H2-Ab molecules and are not capable of mediating skin transplant rejection or causing lethal graft-vs-host disease (167). Bone-marrow chimeras, where the bone-marrow-derived APCs are wild-type (and thus able to mediate negative selection in the thymus), have approximately 5% of normal CD4` T cells, and these cells are no longer reactive with wild-type APCs because the self-reactive T cells have been eliminated in the thymus (154, 155, 165). The remaining cells are polyclonal and can respond to antigen after immunization but are not as diverse as CD4` cells from wildtype littermates. It is presently not clear whether the positive selection occurring in the H2-M-deficient mice is mediated by CLIP-class II complexes or by the few other peptide-class II complexes that are likely to exist in the animals (154, 155, 165, 168). H2-M-deficient mice can raise antibody responses after immunization with protein antigens (155), but affinity maturation is not detectable (C Alfonso and L Karlsson, unpublished). Although the immune responses to pathogens have not been extensively studied, H2-M-deficient mice do survive infections with lymphatic choriomeningitis virus (LCMV) and vesicular stomatitis virus (VSV) (155), but in contrast to both wild-type mice and Ii-deficient mice, they cannot clear infection with Leishmania major (164). In mice expressing transgenic H2-Ak complexes, the number of selected CD4` T cells is not significantly changed compared to the mice expressing H2-Ab alone, but in contrast, the expression of H2-Ek,b results in wild-type levels of CD4` T cells (156). The selection of CD4` T cells by H2-Ek,b molecules is not normal, however, and the CD4` cells from these mice show a high degree of reactivity to H2-M-expressing littermates in mixed lymphocyte reactions (C Surh, personal communication). Surprisingly, despite the fact that the H2-Ek,b molecules are partly loaded with non-CLIP peptides, APCs from these mice are unable to stimulate allogeneic CD4` T cells. Mice lacking both H2-M and Ii have very low numbers of CD4` cells (10%–20% of normal, compared to 20%–30% in Ii-only–deficient mice), demonstrating that presentation of H2-M-dependent endogenous antigens is important for efficient T cell selection in the thymus (95, 162–164). Interestingly, B cell maturation has also been reported to be defective in mice lacking H2-M (or H2-M and Ii) (163), in a way similar to what has been described for mice lacking Ii (169) or H2-Aba (170) (but not H2-Abb) (169, 170).

FUNCTIONAL STUDIES OF DO The finding that a large proportion of DM is associated with DO in B cells was quite unexpected, particularly since DM alone appeared to be able to mediate peptide exchange both in vitro and in cell lines where DO is not expressed.

128

ALFONSO n KARLSSON


However, since the majority of DM is actually bound to DO in B cells, a function for DO as a modulator of DM activity was likely. At the molecular level, this interaction could lead to three different outcomes; increased DM function, decreased DM function, or a change in the specificity of which peptides are exchanged by DM. These three possibilities are not necessarily exclusive when molecular function and other factors such as intracellular location for peptide loading and mode of antigen uptake are considered, and data supporting all three possibilities have been presented (41, 51, 111, 112). Ultimately, however, the presence of DO is likely to result in improved presentation of antigens that are physiologically relevant to B cells, but since other antigen presenting cells are mostly not expressing DO, it is unlikely that expression of DO benefits antigen presentation in general. Some of the published data regarding the molecular functions of DO are conflicting, but more importantly, the physiologic function of DO is still only partly understood.

Analysis of DO Function In Vitro DM function can be reconstituted effectively in vitro using purified or semipurified preparations of DM; similar approaches have been used by several groups to study the function of DMDO complexes in vitro (41, 51, 111, 112). In a direct comparison, DMDO complexes are less capable of mediating peptide loading onto DR molecules than is DM alone, although the degree of inhibition appears to vary depending on the experimental system. DO by itself, in contrast, has no measurable effect on peptide exchange (51, 112). The inhibition of DM by association with DO is not absolute, and with time the number of peptide-loaded class II complexes actually reaches the number formed in the presence of DM alone (51, 112). More significantly, the influence of DO on DM-mediated peptide exchange function appears to be pH dependent, so that the inhibitory effect of DO is most noticeable at pH over 5.5, while the peptide loading efficiency at pH ,5.0 approaches the efficiency of DM in the absence of DO (51, 112; M van Ham, J Neefjes, personal communication). DMDO complexes, as well as DO alone, undergo a substantial reversible conformational change upon acidification, and this structural change might explain the increased exchange activity at the lower pH (51). DMDO complexes have been reported to associate with DR molecules in lysosomal lysates from B cell lines (110, 112), but it is unclear whether the increased peptide exchange activity at acidic pH is due to increased catalytic activity of the intact DMDO complex or to dissociation of the complexes into free active DM and DO. One report has used kinetic analysis to suggest that DMDO complexes bind more tightly to class II molecules than DM does, thus providing a more potent chaperone effect (112). A tighter interaction would potentially result in the binding of higher affinity peptides to the class II molecule in vivo, since the time that the class II molecule would be available for peptide exchange (and thus for sampling peptides) would be extended. The tighter binding is not well docu-


129


mented experimentally, however, and since other data show that peptide release from class II molecules, as well as peptide binding, is decreased in the presence of DO (at pH 5.5) (51), it remains to be shown that the time for peptide exchange is really increased in the presence of DO. When the rate of peptide release from DR molecules is analyzed for different peptide-DR complexes in the presence of DMDO complexes, the off-rate for any given peptide is proportional to the offrate in the presence of DM alone (or to the release in the absence of either DM or DMDO), suggesting that the presence of DO does not directly change the quality of released peptides, only the kinetics of peptide release (C Alfonso, L Karlsson, unpublished data). However, the repertoire of bound peptides is also influenced by the on-rate of available peptides (which is concentration-dependent) and by the rate of destruction of the molecules involved, so at any particular point in time the peptides bound by class II molecules may be different in the presence of DM or DMDO complexes (112).

DO/H2-O Function In Vivo Transfection of DO (or H2-O) in cell lines expressing class II molecules, Ii and DM, but not endogenous DO, has for the most part shown an inhibitory effect of DO on the function of DM. Thus, transfection of DO into class II–expressing T cell (41) or melanoma cell (111) lines has been reported to result in increased levels of CLIP–class II complexes both at the cell surface and intracellularly. The decreased release of CLIP in these cells was also accompanied by a decreased ability to present exogenous antigens to CD4` T cell clones, although the extent of this decrease was small (111). In contrast, another study has shown that the level of CLIP–class II complexes was decreased in a DO-transfected DR4 expressing melanoma cell line, compared with the DO negative parent line (112). The reason for these discrepancies is not clear, but the melanoma transfectants express different class II haplotypes (DR3 and DR4, respectively), and this may possibly explain the varying results. It should be noted, however, that the H2-Ab molecules from H2-O-deficient mice have normal (i.e. very low) CLIP content (51), suggesting that CLIP levels may not be a totally relevant marker for DO function. In contrast, several lines of evidence show that the repertoire of class II–associated peptides is different depending on whether the cells express DO or not. Thus, analysis of DR-associated peptides eluted from DO-transfected cell lines shows that these are partly different from the peptides eluted from the DOnegative parent cell lines (111, 112; M van Ham, J Neefjes, personal communication). Although the majority of the eluted peptides (around 90%) are identical in the DO-positive and DO-negative types of cells, some peptide species are present exclusively in one cell line or the other. Further evidence that the peptide repertoires are quantitatively different comes from analysis of antigen presentation by B cells derived from H2-O-deficient mice (51). Thus, while several protein antigens were presented equally well to T cell hybridomas independently of whether the B cells expressed H2-O, the presentation of other antigens was influ-

130

ALFONSO n KARLSSON


enced by the presence of this molecule. In general, presentation of antigens internalized by fluid phase uptake of the antigen was found to be as good or better when H2-O-deficient B cells were compared to B cells from wild-type controls. In contrast, presentation of antigens internalized by receptor-mediated antigen uptake (achieved through breeding with mice expressing transgenic membrane immunoglobulin receptors) was more effective when the antigen-presenting B cells expressed H2-O. Also in this case, however, the presentation of several analyzed epitopes was independent of whether H2-O was present or not. The differences in presentation are epitope- rather than antigen-dependent, since different epitopes derived from the same protein may display different dependence on H2-O (C Alfonso, L Karlsson, unpublished data). Thus, after receptormediated uptake, two out of three analyzed epitopes derived from ovalbumin are presented better by the wild-type B cells than by the H2-O-deficient B cells, while the third epitope is equally well presented by both types of cells.

Why Is DO/H2-O Expressed in B Cells, But Not in Most Other APCs? Both DO and H2-O have been reported to be expressed on thymic epithelium, suggesting that these molecules may influence the thymic selection of T cells. No functional data exist to support this suggestion, however, and since DO and H2O have been reported to be expressed in different parts of the thymus, it is not easy to predict a possible function. The fact that H2-Ob staining is intact also in mice lacking a functional H2-Oa gene also needs to be explained before a function can be attributed to H2-O in the thymus. DO expression has also been reported to occur in dendritic cells, and it is conceivable that expression of DO may be favorable under certain conditions or during certain maturation stages of these cells. However, most dendritic cells and macrophages clearly do not express DO, and thus, the physiological function of DO/H2-O is likely to be a modification to antigen presentation that is beneficial mainly to B cells, rather than to antigen presentation in general. The finding that the majority of DM is associated with DO in B cells, but not in other APCs, is one of the first defined molecular differences between different types of APCs and may explain some of the described differences in antigen presentation between B cells and dendritic cells (171, 172). The reason for the DO/H2-O expression in B cells is likely to be related to the specialized function of B cells; i.e. to present antigens internalized by the clonal Ig receptor in order to receive help from CD4` T cells. The unfolding and degradation of protein antigens to protein fragments and peptides suitable for binding to class II molecules require acidic pH as well as the action of endosomal or lysosomal proteases, but excessive proteolysis will also result in the destruction of potential peptide epitopes (reviewed in 1). Different proteins have different resistance to denaturation at acidic pH and to proteolysis, and thus antigenic epitopes are likely to be released at different locations of the endosomal/lysosomal system. The capture of particular class II–binding



131

epitopes may be optimal in different parts of the endosomal system, depending on the epitope, as well as on the haplotype of the class II molecule. Cathepsin S, the main enzyme degrading Ii in B cells and dendritic cells, is active in a broader pH range (including neutral pH) than most other lysosomal enzymes (173, 174), and it is therefore possible that Ii degradation precedes the bulk of antigen degradation (but not necessarily the bulk of epitope liberation). Since DM (in the absence of DO) is also active in the pH range between 6.5 and 4.5 (122, 126, 127), DM-mediated capture of antigenic epitopes by class II molecules is likely to occur throughout the endosomal/lysosomal system and may initially involve larger peptide fragments than the 11–20 amino acids long peptides that can be isolated from mature class II molecules. Most available in vitro data show that DO modifies the function of DMmediated peptide exchange by limiting the pH range in which effective peptide exchange can occur. The decreased activity of DM in the pH range between 6.5 and 5.5 suggests that the intracellular localization for peptide exchange may be skewed toward more acidic late endosomal and lysosomal compartments in cells expressing DO. Recent data support this conclusion by showing that expression of DO increases the sensitivity of antigen presentation to treatment with bafilomycin A, an inhibitor of the vacuolar ATPase (M van Ham, J Neefjes, personal communication). Treatment with this drug (which increases the endosomal pH) resulted in the abolition of antigen presentation by DO expressing cells, while presentation by the DO negative parent cell line was only partly decreased. In the case of B cells, relevant antigens are internalized after binding to membrane immunoglobulin (175, 176), and since the interaction with antibodies is well known to stabilize protein domains (177–179), the degradation patterns of free and antibody-bound antigens are often different. The pH in the endosomal system is not sufficient to dissociate most high-affinity antigen-antibody interactions, and release of antigenic epitopes from high-affinity antigen-antibody complexes is likely to require lysosomal conditions, due to the protease resistance of the antibodies themselves. Under these conditions, decreased DM-mediated peptide exchange in the earlier parts of the endosomal system due to the association with DO would limit presentation of antigens internalized by low-affinity receptors or by fluid phase uptake, without seriously influencing the presentation of antigens internalized by high-affinity immunoglobulin receptors. With age, H2O-deficient mice get increased levels of serum IgG1 (51), a finding that may reflect increased presentation of antigens internalized by low-affinity receptors or by fluid phase uptake. The presence of DO may serve to increase the specificity of the T cell–B cell interaction by promoting the presentation of peptides internalized by high-affinity antigen receptors, thus helping to focus the immune response and possibly avoiding the activation of irrelevant or harmful T and B cells. Antigen presentation in dendritic cells/macrophages compared to B cells is outlined in Figure 2 (see color insert).

132

ALFONSO n KARLSSON

CONCLUDING REMARKS


The general outline of antigen processing in the class II system in now well established, and it is quite apparent that DM plays an essential role in this process, not only by removing CLIP from maturing class II molecules, but also by editing the repertoire of bound peptides. The editing effect of DM can partly explain why certain peptides from a given protein are presented to T cells while others are not, but unfortunately it is still impossible to predict such peptides. In contrast to the clear and well-defined role of DM, the function of DO appears to be more subtle and is still being elucidated. Although our knowledge of antigen presentation and the components involved in this process is continually increasing, a complete picture of how antigens are processed will probably require a deeper understanding of how the endosomal/lysosomal system functions in general, and how this system is modified in different cell types. ACKNOWLEDGMENTS We thank our colleagues who have shared their preprints and unpublished data with us. We also thank A Brunmark for critically reading the manuscript and the rest of our colleagues at RWJPRI for their help and support. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Watts C. 1997. Capture and processing of exogenous antigens for presentation on MHC molecules. Annu. Rev. Immunol. 15:821–50 2. Wolf PR, Ploegh HL. 1995. How MHC class II molecules acquire peptide cargo: biosynthesis and trafficking through the endocytic pathway. Annu. Rev. Cell Dev. Biol. 11:267–306 3. Cresswell P. 1996. Invariant chain structure and MHC class II function. Cell 84:505–7 4. Trowsdale J, Kelly A. 1985. The human HLA class II alpha chain gene DZ alpha is distinct from genes in the DP, DQ and DR subregions. EMBO J. 4:2231–37 5. Inoko H, Ando A, Kimura M, Tsuji K. 1985. Isolation and characterization of

the cDNA clone and genomic clones of a new HLA class II antigen heavy chain, DO alpha. J. Immunol. 135:2156–59 6. Tonnelle C, DeMars R, Long EO. 1985. DO beta: a new beta chain gene in HLAD with a distinct regulation of expression. EMBO J. 4:2839–47 7. Servenius B, Rask L, Peterson PA. 1987. Class II genes of the human major histocompatibility complex. The DO beta gene is a divergent member of the class II beta gene family. J. Biol. Chem. 262:8759–66 8. Larhammar D, Hammerling U, Rask L, Peterson PA. 1985. Sequence of gene and cDNA encoding murine major histocompatibility complex class II gene A beta 2. J. Biol. Chem. 260:14111–19


NONCLASSICAL MHC CLASS II MOLECULES 9. Steinmetz M, Minard K, Horvath S, McNicholas J, Srelinger J, Wake C, Long E, Mach B, Hood L. 1982. A molecular map of the immune response region from the major histocompatibility complex of the mouse. Nature 300:35–42 10. Braunstein NS, Germain RN. 1986. The mouse E beta 2 gene: a class II MHC beta gene with limited intraspecies polymorphism and an unusual pattern of transcription. EMBO J. 5:2469–76 11. Trowsdale J. 1995. ‘‘Both man & bird & beast’’: comparative organization of MHC genes. Immunogenetics 41:1–17 12. Beck S, Trowsdale J. 1999. Sequence organisation of the class II region of the human MHC. Immunol. Rev. 167:201–10 13. Kelly AP, Monaco JJ, Cho SG, Trowsdale J. 1991. A new human HLA class II-related locus, DM. Nature 353: 571– 73 14. Cho SG, Attaya M, Monaco JJ. 1991. New class II-like genes in the murine MHC. Nature 353:573–76 15. Cho SG, Attaya M, Brown MG, Monaco JJ. 1991. A cluster of transcribed sequences between the Pb and Ob genes of the murine major histocompatibility complex. Proc. Natl. Acad. Sci. USA 88:5197–5201 16. Pamer E, Cresswell P. 1998. Mechanisms of MHC class I–restricted antigen processing. Annu. Rev. Immunol. 16:323–58 17. Radley E, Alderton RP, Kelly A, Trowsdale J, Beck S. 1994. Genomic organization of HLA-DMA and HLA-DMB. Comparison of the gene organization of all six class II families in the human major histocompatibility complex. J. Biol. Chem. 269:18834–38 18. Cho SG, Attaya M, Brown MG, Monaco JJ. 1991. A cluster of transcribed sequences between the Pb and Ob genes of the murine major histocompatibility complex. Proc. Natl. Acad. Sci. USA 88:5197–5201 19. Peleraux A, Karlsson L, Chambers J, Peterson PA. 1996. Genomic organiza-

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

133

tion of a mouse MHC class II region including the H2-M and Lmp2 loci. Immunogenetics 43:204–14 Hermel E, Monaco JJ. 1995. RT1.DMa and RT1.DMb: the rat homologues of H2-DMa and H2-DMb. Immunogenetics 42:446–47 Karlsson L, Peleraux A, Lindstedt R, Liljedahl M, Peterson PA. 1994. Reconstitution of an operational MHC class II compartment in nonantigen-presenting cells. Science 266:1569–73 Hermel E, Yuan J, Monaco JJ. 1995. Characterization of polymorphism within the H2-M MHC class II loci. Immunogenetics 42:136–42 Walter W, Loos M, Maeurer MJ. 1996. H2-M polymorphism in mice susceptible to collagen-induced arthritis involves the peptide binding groove. Immunogenetics 44:19–26 Morris P, Shaman J, Attaya M, Amaya M, Goodman S, Bergman C, Monaco JJ, Mellins E. 1994. An essential role for HLA-DM in antigen presentation by class II major histocompatibility molecules. Nature 368:551–54 Stebbins CC, Peterson ME, Suh WM, Sant AJ. 1996. DM-mediated release of a naturally occurring invariant chain degradation intermediate from MHC class II molecules. J. Immunol. 157:4892–98 Carrington M, Yeager M, Mann D. 1993. Characterization of HLA-DMB polymorphism. Immunogenetics 38:446–49 Sanderson F, Powis SH, Kelly AP, Trowsdale J. 1994. Limited polymorphism in HLA-DM does not involve the peptide binding groove. Immunogenetics 39:56–58 Djilali-Saiah I, Benini V, Schmitz J, Timsit J, Assan R, Boitard C, Bach J-F, Caillat-Zucman S. 1996. Absence of primary association between DM gene polymorphism and insulin-dependent diabetes mellitus or celiac disease. Hum. Immunol. 49:22–27 Kim T-G, Carrington M, Choi H-B, Kim

134

30.

31.


32.

33.

34.

35.

36.

37.

38.

ALFONSO n KARLSSON H-Y, Han H. 1996. Three HLA-DMB variants in Korean patients with autoimmune diseases. Hum. Immunol. 46:58–60 Kuwata S, Yanagisawa M, Nakagawa H, Saeki H, Etoh T, Miyamoto M, Juji T. 1996. HLA-DM gene polymorphisms in atopic dermatitis. J. Allergy Clin. Immunol. 98:S192–200 Ishihara M, Naruse T, Ohno S, Kawata H, Mizuki N, Yamagata N, Ishida T, Nose Y, Inoko H. 1996. Analysis of HLA-DM polymorphisms in sarcoidosis. Hum. Immunol. 49:144–46 Pinet V, Combe B, Avinens O, CaillatZucman S, Sany J, Clot J, Eliaou JF. 1997. Polymorphism of the HLA-DMA and DMB genes in rheumatoid arthritis. Arthritis Rheum. 40:854–58 Chang CH, Flavell RA. 1995. Class II transactivator regulates the expression of multiple genes involved in antigen presentation. J. Exp. Med. 181:765–67 Kern I, Steimle V, Siegrist CA, Mach B. 1995. The two novel MHC class II transactivators RFX5 and CIITA both control expression of HLA-DM genes. Int. Immunol. 7:1295–99 Chang CH, Guerder S, Hong SC, van Ewijk W, Flavell RA. 1996. Mice lacking the MHC class II transactivator (CIITA) show tissue-specific impairment of MHC class II expression. Immunity 4:167–78 Clausen BE, Waldburger JM, Schwenk F, Barras E, Mach B, Rajewsky K, Forster I, Reith W. 1998. Residual MHC class II expression on mature dendritic cells and activated B cells in RFX5-deficient mice. Immunity 8:143–55 Fling SP, Rak J, Muczynski KA, Arp B, Pious D. 1997. Novel mutants define genes required for the expression of human histocompatibility leukocyte antigen DM: evidence for loci on human chromosome 6p. J. Exp. Med. 186:1469– 80 Niimi M, Nakai Y, Aida Y. 1995. Nucleotide sequences and the molecular evolution of the DMA and DMB genes of

39.

40.

41.

42.

43.

44.

45.

46.

47.

the bovine major histocompatibility complex. Biochem. Biophys. Res. Commun. 217:522–28 Hermel E, Han M, Hague B, Kindt T, Monaco JJ. 1999. Isolation and mapping of the rabbit DM genes. Immunogenetics 49:295–302 Ponzoni M, Guarnaccia F, Corrias MV, Cornaglia-Ferraris P. 1993. Uncoordinate induction and differential regulation of HLA class-I and class-II expression by gamma-interferon in differentiating human neuroblastoma cells. Int. J. Cancer. 55:817–23 Denzin LK, Sant’Angelo DB, Hammond C, Surman MJ, Cresswell P. 1997. Negative regulation by HLA-DO of MHC class II-restricted antigen processing. Science 278:106–9 Voliva CF, Tsang S, Peterlin BM. 1993. Mapping cis-acting defects in promoters of transcriptionally silent DQA2, DQB2, and DOB genes. Proc. Natl. Acad. Sci. USA 90:3408–12 Karlsson L, Surh CD, Sprent J, Peterson PA. 1991. A novel class II MHC molecule with unusual tissue distribution. Nature 351:485–88 Larhammar D, Hammerling U, Denaro M, Lund T, Flavell RA, Rask L, Peterson PA. 1983. Structure of the murine immune response I-A beta locus: sequence of the I-A beta gene and an adjacent beta-chain second domain exon. Cell 34:179–88 Karlsson L, Peterson PA. 1992. The a chain gene of H-2O has an unexpected location in the major histocompatibility complex. J. Exp. Med. 176:477–83 Wake CT, Flavell RA. 1985. Multiple mechanisms regulate the expression of murine immune response genes. Cell 42:623–28 Surh CD, Gao EK, Kosaka H, Lo D, Ahn C, Murphy DB, Karlsson L, Peterson P, Sprent J. 1992. Two subsets of epithelial cells in the thymic medulla. J. Exp. Med. 176:495–505


NONCLASSICAL MHC CLASS II MOLECULES 48. Liljedahl M, Kuwana T, Fung-Leung WP, Jackson MR, Peterson PA, Karlsson L. 1996. HLA-DO is a lysosomal resident which requires association with HLA-DM for efficient intracellular transport. EMBO J. 15:4817–24 49. Douek DC, Altmann DM. 1997. HLADO is an intracellular class II molecule with distinctive thymic expression. Int. Immunol. 9:355–64 50. Vogt AB, Hammerling GJ, Kropshofer H. 1998. HLA-DM and HLA-DO: a chaperone and its modulator influencing the peptide repertoire seen by T cells. Immunologist 6:186–93 51. Liljedahl M, Winqvist O, Surh CD, Wong P, Ngo K, Teyton L, Peterson PA, Brunmark A, Rudensky AY, Fung-Leung WP, Karlsson L. 1998. Altered antigen presentation in mice lacking H2-O. Immunity 8:233–43 52. Jonsson AK, Rask L. 1989. Human class II DNA and DOB genes display low sequence variability. Immunogenetics 29:411–13 53. Naruse TK, Kawata H, Anzai T, Takashige N, Kagiya M, Nose Y, Nabeya N, Isshiki G, Tatsumi N, Inoko H. 1999. Limited polymorphism in the HLA-DOA gene. Tissue Antigens 53:359–65 54. Karlsson L, Peterson PA. 1992. The alpha chain gene of H-2O has an unexpected location in the major histocompatibility complex. J. Exp. Med. 176:477–83 55. Cresswell P. 1994. Assembly, transport, and function of MHC class II molecules. Annu. Rev. Immunol. 12:259–93 56. Busch R, Cloutier I, Sekaly RP, Hammerling GJ. 1996. Invariant chain protects class II histocompatibility antigens from binding intact polypeptides in the endoplasmic reticulum. EMBO J. 15: 418–28 57. Aichinger G, Karlsson L, Jackson MR, Vestberg M, Vaughan JH, Teyton L, Lechler RI, Peterson PA. 1997. Major histocompatibility complex class II–

58.

59.

60.

61.

62.

63.

64.

65.

66.

135

dependent unfolding, transport, and degradation of endogenous proteins. J. Biol. Chem. 272:29127–36 Romagnoli P, Germain RN. 1994. The CLIP region of invariant chain plays a critical role in regulating major histocompatibility complex class II folding, transport, and peptide occupancy. J. Exp. Med. 180:1107–13 Ghosh P, Amaya M, Mellins E, Wiley DC. 1995. The structure of an intermediate in class II MHC maturation: CLIP bound to HLA-DR3. Nature 378:457–62 Park SJ, Sadegh-Nasseri S, Wiley DC. 1995. Invariant chain made in Escherichia coli has an exposed N-terminal segment that blocks antigen binding to HLA-DR1 and a trimeric C- terminal segment that binds empty HLA-DR1. Proc. Natl. Acad. Sci. USA 92:11289–93 Jasanoff A, Song S, Dinner AR, Wagner G, Wiley DC. 1999. One of two unstructured domains of Ii becomes ordered in complexes with MHC class II molecules. Immunity 10:761–68 Sette A, Southwood S, Miller J, Appella E. 1995. Binding of major histocompatibility complex class II to the invariant chain-derived peptide, CLIP, is regulated by allelic polymorphism in class II. J. Exp. Med. 181:677–83 Siebenkotten IM, Carstens C, Koch N. 1998. Identification of a sequence that mediates promiscuous binding of invariant chain to MHC class II allotypes. J. Immunol. 160:3355–62 Wilson NA, Wolf P, Ploegh H, Ignatowicz L, Kappler J, Marrack P. 1998. Invariant chain can bind MHC class II at a site other than the peptide binding groove. J. Immunol. 161:4777–84 Thayer WP, Ignatowicz L, Weber DA, Jensen PE. 1999. Class II-associated invariant chain peptide-independent binding of invariant chain to class II MHC molecules. J. Immunol. 162:1502– 9 Bakke O, Dobberstein B. 1990. MHC

136

67.

68.


69.

70.

71.

72.

73.

74.

ALFONSO n KARLSSON class II-associated invariant chain contains a sorting signal for endosomal compartments. Cell 63:707–16 Lotteau V, Teyton L, Peleraux A, Nilsson T, Karlsson L, Schmid SL, Quaranta V, Peterson PA. 1990. Intracellular transport of class II MHC molecules directed by invariant chain. Nature 348:600–5 Odorizzi CG, Trowbridge IS, Xue L, Hopkins CR, Davis CD, Collawn JF. 1994. Sorting signals in the MHC class II invariant chain cytoplasmic tail and transmembrane region determine trafficking to an endocytic processing compartment. J. Cell Biol. 126:317–30 Pond L, Kuhn LA, Teyton L, Schutze MP, Tainer JA, Jackson MR, Peterson PA. 1995. A role for acidic residues in dileucine motif-based targeting to the endocytic pathway. J. Biol. Chem. 270:19989–97 Peters PJ, Neefjes JJ, Oorschot V, Ploegh HL, Geuze HJ. 1991. Segregation of MHC class II molecules from MHC class I molecules in the Golgi complex for transport to lysosomal compartments. Nature 349:669–76 Benaroch P, Yilla M, Raposo G, Ito K, Miwa K, Geuze HJ, Ploegh HL. 1995. How MHC class II molecules reach the endocytic pathway. EMBO J. 14:37–49 Kleijmeer MJ, Morkowski S, Griffith JM, Rudensky AY, Geuze HJ. 1997. Major histocompatibility complex class II compartments in human and mouse B lymphoblasts represent conventional endocytic compartments. J. Cell Biol. 139:639–49 Castellino F, Germain RN. 1995. Extensive trafficking of MHC class II– invariant chain complexes in the endocytic pathway and appearance of peptide-loaded class II in multiple compartments. Immunity 2:73–88 Roche PA, Teletski CL, Stang E, Bakke O, Long EO. 1993. Cell surface HLADR-invariant chain complexes are targeted to endosomes by rapid internal-

75.

76.

77.

78.

79.

80.

81.

82.

83.

ization. Proc. Natl. Acad. Sci. USA 90:8581–85 Wang K, Peterson PA, Karlsson L. 1997. Decreased endosomal delivery of major histocompatibility complex class IIinvariant chain complexes in dynamindeficient cells. J. Biol. Chem. 272: 17055–60 Pond L, Watts C. 1999. Functional early endosomes are required for maturation of major histocompatibility complex class II molecules in human B lymphoblastoid cells. J. Biol. Chem. 274:18049–54 Claesson-Welsh L, Peterson PA. 1985. Implications of the invariant gammachain on the intracellular transport of class II histocompatibility antigens. J. Immunol. 135:3551–57 Bikoff EK, Huang LY, Episkopou V, van Meerwijk J, Germain RN, Robertson EJ. 1993. Defective major histocompatibility complex class II assembly, transport, peptide acquisition, and CD4` T cell selection in mice lacking invariant chain expression. J. Exp. Med. 177:1699–1712 Viville S, Neefjes J, Lotteau V, Dierich A, Lemeur M, Ploegh H, Benoist C, Mathis D. 1993. Mice lacking the MHC class II-associated invariant chain. Cell 72:635–48 Pinet V, Malnati MS, Long EO. 1994. Two processing pathways for the MHC class II-restricted presentation of exogenous influenza virus antigen. J. Immunol. 152:4852–60 Pinet V, Vergelli M, Martin R, Bakke O, Long EO. 1995. Antigen presentation mediated by recycling of surface HLADR molecules. Nature 375:603–6 Zhong G, Romagnoli P, Germain RN. 1997. Related leucine-based cytoplasmic targeting signals in invariant chain and major histocompatibility complex class II molecules control endocytic presentation of distinct determinants in a single protein. J. Exp. Med. 185:429–38 Davidson HW, Reid PA, Lanzavecchia A, Watts C. 1991. Processed antigen


84.

85.


86.

87.

88.

89.

90.

91.

92.

binds to newly synthesized MHC class II molecules in antigen-specific B lymphocytes. Cell 67:105–16 Pieters J, Horstmann H, Bakke O, Griffiths G, Lipp J. 1991. Intracellular transport and localization of major histocompatibility complex class II molecules and associated invariant chain. J. Cell Biol. 115:1213–23 Amigorena S, Drake JR, Webster P, Mellman I. 1994. Transient accumulation of new class II MHC molecules in a novel endocytic compartment in B lymphocytes. Nature 369:113–20 Barnes KA, Mitchell RN. 1995. Detection of functional class II-associated antigen: role of a low density endosomal compartment in antigen processing. J. Exp. Med. 181:1715–27 Tulp A, Verwoerd D, Dobberstein B, Ploegh HL, Pieters J. 1994. Isolation and characterization of the intracellular MHC class II compartment. Nature 369:120– 26 Qiu Y, Xu XX, Wandinger-Ness A, Dalke DP, Pierce SK. 1994. Separation of subcellular compartments containing distinct functional forms of MHC class II. J. Cell Biol. 125:595–605 West MA, Lucocq JM, Watts C. 1994. Antigen processing and class II MHC peptide-loading compartments in human B-lymphoblastoid cells. Nature 369: 147–51 Pierre P, Turley SJ, Gatti E, Hull M, Meltzer J, Mirza A, Inaba K, Steinman RM, Mellman I. 1997. Developmental regulation of MHC class II transport in mouse dendritic cells. Nature 388:787– 92 Brachet V, Raposo G, Amigorena S, Mellman I. 1997. Ii chain controls the transport of major histocompatibility complex class II molecules to and from lysosomes. J. Cell Biol. 137:51–65 Denzin LK, Robbins NF, CarboyNewcomb C, Cresswell P. 1994. Assembly and intracellular transport of

93.

94.

95.

96.

97.

98.

99.

100.

137

HLA-DM and correction of the class II antigen processing defect in T2 cells. Immunity 1:595–606 Lindstedt R, Liljedahl M, Peleraux A, Peterson PA, Karlsson L. 1995. The MHC class II molecule H2-M is targeted to an endosomal compartment by a tyrosine-based targeting motif. Immunity 3:561–72 Marks MS, Roche PA, van Donselaar E, Woodruff L, Peters PJ, Bonifacino JS. 1995. A lysosomal targeting signal in the cytoplasmic tail of the beta chain directs HLA-DM to MHC class II compartments. J. Cell Biol. 131:351–69 Kovats S, Grubin CE, Eastman S, deRoos P, Dongre A, Van Kaer L, Rudensky AY. 1998. Invariant chain-independent function of H-2M in the formation of endogenous peptide-major histocompatibility complex class II complexes in vivo. J. Exp. Med. 187:245–51 Sanderson F, Kleijmeer MJ, Kelly A, Verwoerd D, Tulp A, Neefjes JJ, Geuze HJ, Trowsdale J. 1994. Accumulation of HLA-DM, a regulator of antigen presentation, in MHC class II compartments. Science 266:1566–69 Nijman HW, Kleijmeer MJ, Ossevoort MA, Oorschot VM, Vierboom MP, van de Keur M, Kenemans P, Kast WM, Geuze HJ, Melief CJ. 1995. Antigen capture and major histocompatibility class II compartments of freshly isolated and cultured human blood dendritic cells. J. Exp. Med. 182:163–74 Pierre P, Denzin LK, Hammond C, Drake JR, Amigorena S, Cresswell P, Mellman I. 1996. HLA-DM is localized to conventional and unconventional MHC class II-containing endocytic compartments. Immunity 4:229–39 Kim K-J, Jung H-H, Bae Y-S. 1996. Blood dendritic cells express detectable amounts of class II-related DM protein on the cell surfaces. Mol. Cells 6:684–91 Copier J, Kleijmeer MJ, Ponnambalam S, Oorschot V, Potter P, Trowsdale J,

138

101.

102.


103.

104.

105.

106.

107.

108.

ALFONSO n KARLSSON Kelly A. 1996. Targeting signal and subcellular compartments involved in the intracellular trafficking of HLA-DMB. J. Immunol. 157:1017–27 Copier J, Potter P, Sacks SH, Kelly AP. 1998. Multiple signals regulate the intracellular trafficking of HLA-DM in Blymphoblastoid cells. Immunology 93: 505–10 Liu SH, Marks MS, Brodsky FM. 1998. A dominant-negative clathrin mutant differentially affects trafficking of molecules with distinct sorting motifs in the class II major histocompatibility complex (MHC) pathway. J. Cell Biol. 140:1023–37 Ferrari G, Knight AM, Watts C, Pieters J. 1997. Distinct intracellular compartments involved in invariant chain degradation and antigenic peptide loading of major histocompatibility complex (MHC) class II molecules. J. Cell Biol. 139:1433–46 Kleijmeer MJ, Raposo G, Geuze HJ. 1996. Characterization of MHC class II compartments by immunoelectron microscopy. Methods Companion Methods Enzymol. 10:191–207 Robbins NF, Hammond C, Denzin LK, Pan M, Cresswell P. 1996. Trafficking of major histocompatibility complex class II molecules through intracellular compartments containing HLA-DM. Hum. Immunol. 45:13–23 Schafer PH, Green JM, Malapati S, Gu L, Pierce SK. 1996. HLA-DM is present in one-fifth the amount of HLA-DR in the class II peptide-loading compartment where it associates with leupeptininduced peptide (LIP)-HLA-DR complexes. J. Immunol. 157:5487–95 Raposo G, Nijman HW, Stoorvogel W, Liejendekker R, Harding CV, Melief CJ, Geuze HJ. 1996. B lymphocytes secrete antigen-presenting vesicles. J. Exp. Med. 183:1161–72 Zitvogel L, Regnault A, Lozier A, Wolfers J, Flament C, Tenza D, Ricciardi-

109.

110.

111.

112.

113.

114.

115.

Castagnoli P, Raposo G, Amigorena S. 1998. Eradication of established murine tumors using a novel cell-free vaccine: dendritic cell-derived exosomes. Nat. Med. 4:594–600 Escola JM, Kleijmeer MJ, Stoorvogel W, Griffith JM, Yoshie O, Geuze HJ. 1998. Selective enrichment of tetraspan proteins on the internal vesicles of multivesicular endosomes and on exosomes secreted by human B- lymphocytes. J. Biol. Chem. 273(32):121–27 Hammond C, Denzin LK, Pan M, Griffith JM, Geuze HJ, Cresswell P. 1998. The tetraspan protein CD82 is a resident of MHC class II compartments where it associates with HLA-DR, -DM, and -DO molecules. J. Immunol. 161:3282–91 van Ham SM, Tjin EPM, Lillemeier BF, Gruneberg U, van Meijgaarden KE, Pastoors L, Verwoerd D, Tulp A, Canas B, Rahman D, Ottenhoff TH, Pappin DJ, Trowsdale J, Neefjes J. 1997. HLA-DO is a negative modulator of HLA-DMmediated MHC class II peptide loading. Curr. Biol. 7:950–57 Kropshofer H, Vogt AB, Thery C, Armandola EA, Li BC, Moldenhauer G, Amigorena S, Hammerling GJ. 1998. A role for HLA-DO as a co-chaperone of HLA-DM in peptide loading of MHC class II molecules. EMBO J. 17:2971–81 Samaan A, Thibodeau J, Mahana W, Castellino F, Cazenave PA, Kindt TJ. 1999. Cellular distribution of a mixed MHC class II heterodimer between DRa and a chimeric DOb chain. Int. Immunol. 11:99–111 Mellins E, Smith L, Arp B, Cotner T, Celis E, Pious D. 1990. Defective processing and presentation of exogenous antigens in mutants with normal HLA class II genes. Nature 343:71–74 Pious D, Dixon L, Levine F, Cotner T, Johnson R. 1985. HLA class II regulation and structure. Analysis with HLA-DR3 and HLA-DP point mutants. J. Exp. Med. 162:1193–1207


NONCLASSICAL MHC CLASS II MOLECULES 116. Riberdy JM, Newcomb JR, Surman MJ, Barbosa JA, Cresswell P. 1992. HLA-DR molecules from an antigen-processing mutant cell line are associated with invariant chain peptides. Nature 360:474–77 117. Sette A, Ceman S, Kubo RT, Sakaguchi K, Appella E, Hunt DF, Davis TA, Michel H, Shabanowitz J, Rudersdorf R, et al. 1992. Invariant chain peptides in most HLA-DR molecules of an antigenprocessing mutant. Science 258:1801–4 118. Avva RR, Cresswell P. 1994. In vivo and in vitro formation and dissociation of HLA-DR complexes with invariant chain-derived peptides. Immunity 1:763– 74 119. Mellins E, Kempin S, Smith L, Monji T, Pious D. 1991. A gene required for class II-restricted antigen presentation maps to the major histocompatibility complex. J. Exp. Med. 174:1607–15 120. Riberdy JM, Cresswell P. 1992. The antigen-processing mutant T2 suggests a role for MHC-linked genes in class II antigen presentation. J. Immunol. 148:2586–90 121. Fling SP, Arp B, Pious D. 1994. HLADMA and -DMB genes are both required for MHC class II/peptide complex formation in antigen-presenting cells. Nature 368:554–58 122. Denzin LK, Cresswell P. 1995. HLA-DM induces CLIP dissociation from MHC class II ab dimers and facilitates peptide loading. Cell 82:155–65 123. Kropshofer H, Arndt SO, Moldenhauer G, Haammerling G, Vogt AB. 1997. HLA-DM acts as a molecular chaperone and rescues empty HLA-DR molecules at lysosomal pH. Immunity 6:293–302 124. Mosyak L, Zaller DM, Wiley DC. 1998. The structure of HLA-DM, the peptide exchange catalyst that loads antigen onto class II MHC molecules during antigen presentation. Immunity 9:377–83 125. Fremont DH, Crawford F, Marrack P, Hendrickson WA, Kappler J. 1998. Crys-

126.

127.

128.

129.

130.

131.

132.

133.

134.

139

tal structure of mouse H2-M. Immunity 9:385–93 Sloan VS, Cameron P, Porter G, Gammon M, Amaya M, Mellins E, Zaller DM. 1995. Mediation by HLA-DM of dissociation of peptides from HLA-DR. Nature 375:802–6 Sherman MA, Weber DA, Jensen PE. 1995. DM enhances peptide binding to class II MHC by release of invariant chain-derived peptide. Immunity 3:197– 205 van Ham SM, Gruneberg U, Malcherek G, Broker I, Melms A, Trowsdale J. 1996. Human histocompatibility leukocyte antigen (HLA)-DM edits peptides presented by HLA-DR according to their ligand binding motifs. J. Exp. Med. 184:2019–24 Denzin LK, Hammond C, Cresswell P. 1996. HLA-DM interactions with intermediates in HLA-DR maturation and a role for HLA-DM in stabilizing empty HLA-DR molecules. J. Exp. Med. 184:2153–65 Stern LJ, Wiley DC. 1992. The human class II MHC protein HLA-DR1 assembles as empty alpha beta heterodimers in the absence of antigenic peptide. Cell 68:465–77 Germain RN, Rinker A Jr. 1993. Peptide binding inhibits protein aggregation of invariant-chain free class II dimers and promotes surface expression of occupied molecules. Nature 363: 725–28 Sadegh-Nasseri S, Stern LJ, Wiley DC, Germain RN. 1994. MHC class II function preserved by low-affinity peptide interactions preceding stable binding. Nature 370:647–50 Weber DA, Evavold BD, Jensen PE. 1996. Enhanced dissociation of HLADR-bound peptides in the presence of HLA-DM. Science 274:618–20 Kropshofer H, Hammerling GJ, Vogt AB. 1997. How HLA-DM edits the MHC class II peptide repertoire: survival of the fittest? Immunol. Today 18:77–82


140

ALFONSO n KARLSSON

135. Vogt AB, Kropshofer H, Moldenhauer G, Haemmerling GJ. 1996. Kinetic analysis of peptide loading onto HLA-DR molecules mediated by HLA-DM. Proc. Natl. Acad. Sci. USA 93:9724–29 136. Ramachandra L, Kovats S, Eastman S, Rudensky AY. 1996. Variation in HLADM expression influences conversion of MHC class II alpha beta:class IIassociated invariant chain peptide complexes to mature peptide-bound class II alpha beta dimers in a normal B cell line. J. Immunol. 156:2196–204 137. Mellins E, Cameron P, Amaya M, Goodman S, Pious D, Smith L, Arp B. 1994. A mutant human histocompatibility leukocyte antigen DR molecule associated with invariant chain peptides. J. Exp. Med. 179:541–49 138. Sanderson F, Thomas C, Neefjes J, Trowsdale J. 1996. Association between HLA-DM and HLA-DR in vivo. Immunity 4:87–96 139. Stern LJ, Brown JH, Jardetzky TS, Gorga JC, Urban RG, Strominger JL, Wiley DC. 1994. Crystal structure of the human class II MHC protein HLA-DR1 complexed with an influenza virus peptide. Nature 368:215–21 140. Dessen A, Lawrence CM, Cupo S, Zaller DM, Wiley DC. 1997. X-ray crystal structure of HLA-DR4 (DRA*0101, DRB1*0401) complexed with a peptide from human collagen II. Immunity 7:473–81 141. Runnels HA, Moore JC, Jensen PE. 1996. A structural transition in class II major histocompatibility complex proteins at mildly acidic pH. J. Exp. Med. 183:127–36 142. Boniface JJ, Lyons DS, Wettstein DA, Allbritton NL, Davis MM. 1996. Evidence for a conformational change in a class II major histocompatibility complex molecule occurring in the same pH range where antigen binding is enhanced. J. Exp. Med. 183:119–26 143. Ullrich HJ, Doring K, Gruneberg U, Jah-

144.

145.

146.

147.

148.

149.

150.

151.

nig F, Trowsdale J, van Ham SM. 1997. Interaction between HLA-DM and HLADR involves regions that undergo conformational changes at lysosomal pH. Proc. Natl. Acad. Sci. USA 94:13163–68 Busch R, Reich Z, Zaller DM, Sloan V, Mellins ED. 1998. Secondary structure composition and pH-dependent conformational changes of soluble recombinant HLA-DM. J. Biol. Chem. 273:27557–64 Weber DA, Evavold BD, Jensen PE. 1996. Enhanced dissociation of HLADR-bound peptides in the presence of HLA- DM. Science 274:618–20 Raddrizzani L, Bono E, Vogt AB, Kropshofer H, Gallazzi F, Sturniolo T, Haemmerling GJ, Sinigaglia F, Hammer J. 1999. Identification of destabilizing residues in HLA class II-selected bacteriophage display libraries edited by HLA-DM. Eur. J. Immunol. 29:660–68 Brooks AG, Campbell PL, Reynolds P, Gautam AM, McCluskey J. 1994. Antigen presentation and assembly by mouse I-Ak class II molecules in human APC containing deleted or mutated HLA DM genes. J. Immunol. 153:5382–92 Stebbins CC, Loss G Jr, Elias CG, Chervonsky A, Sant AJ. 1995. The requirement for DM in class II-restricted antigen presentation and SDS-stable dimer formation is allele and species dependent. J. Exp. Med. 181:223–34 Fung-Leung W-P, Surh CD, Liljedahl M, Pang J, Leturcq D, Peterson PA, Webb SR, Karlsson L. 1996. Antigen presentation and T cell development in H2-M deficient mice. Science 271:1278–81 Martin WD, Hicks GG, Mendiratta SK, Leva HI, Ruley HE, Van Kaer L. 1996. H2-M mutant mice are defective in the peptide loading of class II molecules, antigen presentation, and T cell repertoire selection. Cell 84:543–50 Miyazaki T, Wolf P, Tourne S, Waltzinger C, Dierich A, Barois N, Ploegh H, Benoist C, Mathis D. 1996. Mice lacking H2-M complexes, enigmatic elements of


152.

153.


154.

155.

156.

157.

158.

159.

160.

the MHC class II peptide-loading pathway. Cell 84:531–41 Kovats S, Whiteley PE, Concannon P, Rudensky AY, Blum JS. 1997. Presentation of abundant endogenous class II DRrestricted antigens by DM-negative B cell lines. Eur. J. Immunol. 27:1014–21 Ma C, Blum JS. 1997. Receptor-mediated endocytosis of antigens overcomes the requirement for HLA-DM in class IIrestricted antigen presentation. J. Immunol. 158:1–4 Grubin CE, Kovats S, deRoos P, Rudensky AY. 1997. Deficient positive selection of CD4 T cells in mice displaying altered repertoires of MHC class IIbound self-peptides. Immunity 7:197– 208 Tourne S, Miyazaki T, Oxenius A, Klein L, Fehr T, Kyewski B, Benoist C, Mathis D. 1997. Selection of a broad repertoire of CD4` T cells in H-2Ma0/0 mice. Immunity 7:187–95 Wolf PR, Tourne S, Miyazaki T, Benoist C, Mathis D, Ploegh HL. 1998. The phenotype of H-2M-deficient mice is dependent on the MHC class II molecules expressed. Eur. J. Immunol. 28:2605–18 Katz JF, Stebbins C, Appella E, Sant AJ. 1996. Invariant chain and DM edit selfpeptide presentation by major histocompatibility complex (MHC) class II molecules. J. Exp. Med. 184:1747–53 Weenink S, Averdunk H, Boston T, Boswarva V, Guery JC, Adorini L, Mellins E, McCluskey J, Gautam AM. 1997. Impaired antigen presentation by murine I-Ad class II MHC molecules expressed in normal and HLA-DM-defective human B cell lines. Int. Immunol. 9:889– 96 Rammensee HG, Friede T, Stevanoviic S. 1995. MHC ligands and peptide motifs: first listing. Immunogenetics 41:178–228 Urban RG, Chicz RM, Strominger JL. 1994. Selective release of some invariant chain-derived peptides from HLA-DR1

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

141

molecules at endosomal pH. J. Exp. Med. 180:751–55 Kropshofer H, Vogt AB, Stern LJ, Hammerling GJ. 1995. Self-release of CLIP in peptide loading of HLA-DR molecules. Science 270:1357–59 Tourne S, Miyazaki T, Wolf P, Ploegh H, Benoist C, Mathis D. 1997. Functionality of major histocompatibility complex class II molecules in mice doubly deficient for invariant chain and H-2M complexes. Proc. Natl. Acad. Sci. USA 94:9255–60 Kenty G, Martin WD, Van Kaer L, Bikoff EK. 1998. MHC class II expression in double mutant mice lacking invariant chain and DM functions. J. Immunol. 160:606–14 Swier K, Brown DR, Bird JJ, Martin WD, Van Kaer L, Reiner SL. 1998. A critical, invariant chain-independent role for H2-M in antigen presentation. J. Immunol. 160:540–44 Surh CD, Lee D-S, Fung-Leung W-P, Karlsson L, Sprent J. 1997. Thymic selection by a single MHC/peptide ligand produces a semidiverse repertoire of CD4` T cells. Immunity 7:209–19 Sant’Angelo DB, Waterbury PG, Cohen BE, Martin WD, Van Kaer L, Hayday AC, Janeway CA Jr. 1997. The imprint of intrathymic self-peptides on the mature T cell receptor repertoire. Immunity 7:517–24 Lee DS, Ahn C, Ernst B, Sprent J, Surh CD. 1999. Thymic selection by a single MHC/peptide ligand: autoreactive T cells are low-affinity cells. Immunity 10:83–92 Barton GM, Rudensky AY. 1999. Requirement for diverse, low-abundance peptides in positive selection of T cells. Science 283:67–70 Shachar I, Flavell RA. 1996. Requirement for invariant chain in B cell maturation and function. Science 274:106–8 Rolink AG, Brocker T, Bluethmann H, Kosco-Vilbois MH, Andersson J, Mel-

142

171.


172.

173.

174.

ALFONSO n KARLSSON chers F. 1999. Mutations affecting either generation or survival of cells influence the pool size of mature B cells. Immunity 10:619–28 Nakagawa TY, Brissette WH, Lira PD, Griffiths RJ, Petrushova N, Stock J, McNeish JD, Eastman SE, Howard ED, Clarke SRM, Rosloniec EF, Elliott EA, Rudensky AY. 1999. Impaired invariant chain degradation and antigen presentation and diminished collagen-induced arthritis in cathepsin S null mice. Immunity 10:207–17 Gapin L, Bravo de Alba Y, Casrouge A, Cabaniols JP, Kourilsky P, Kanellopoulos J. 1998. Antigen presentation by dendritic cells focuses T cell responses against immunodominant peptides: studies in the hen egg-white lysozyme (HEL) model. J. Immunol. 160:1555–64 Chapman HA. 1998. Endosomal proteolysis and MHC class II function. Curr. Opin. Immunol. 10:93–102 Kirschke H, Wiederanders B. 1994. Cathepsin S and related lysosomal endopeptidases. Methods Enzymol. 244:500– 11

175. Rock KL, Benacerraf B, Abbas AK. 1984. Antigen presentation by haptenspecific B lymphocytes. I. Role of surface immunoglobulin receptors. J. Exp. Med. 160:1102–13 176. Lanzavecchia A. 1985. Antigen-specific interaction between T and B cells. Nature 314:537–39 177. Accolla RS, Cina R, Montesoro E, Celada F. 1981. Antibody-mediated activation of genetically defective Escherichia coli beta-galactosidases by monoclonal antibodies produced by somatic cell hybrids. Proc. Natl. Acad. Sci. USA 78:2478–82 178. Jemmerson R, Paterson Y. 1986. Mapping epitopes on a protein antigen by the proteolysis of antigen-antibody complexes. Science 232:1001–4 179. Simitsek PD, Campbell DG, Lanzavecchia A, Fairweather N, Watts C. 1995. Modulation of antigen processing by bound antibodies can boost or suppress class II major histocompatibility complex presentation of different T cell determinants. J. Exp. Med. 1815:1957– 63


ALFONSO

■

KARLSSON C-1

Figure 1 Schematic representation of DM function class II-CLIP complexes (top view; 1) are generated in the endosomal system. DM (2) binds to the class II-CLIP complexes, thus stabilizing a transient, more open, conformational state in the class II molecule (3). This conformational change results in the disruption of the hydrogen bond network that maintains CLIP in the peptide binding site and therefore the release of CLIP. The empty class II molecule is stabilized by DM (4). In the absence of peptides, DM may dissociate from the class II molecule, causing this to collapse and become functionally inactivated (5). This is not likely to be a common scenario in vivo, and instead the open DM-stabilized peptide binding site can bind other peptides generated by proteolysis in the endosomal system (6). Good stably binding peptides have anchor residues that fit in binding pockets on the class II molecule, and when DM is released from the class II molecule (7-8), these peptides become trapped in the binding site (8), while DM is free to catalyze another round of peptide exchange. Poorly stabilizing peptides (other than CLIP) trapped in the binding site of the class II molecule can be released in subsequent rounds of peptide exchange, thus resulting in a peptide editing effect of DM where stably binding peptides are favored.


C-2 ALFONSO

■

KARLSSON

Figure 2A A model of H2-O/DO-mediated differences in antigen processing between distinct APCs. Internalized extracellular antigens pass through the endocytic system becoming denatured (by acidic pH) and degraded (by proteolytic enzymes, including cathepsins (Cat)) as their migration proceeds through early endosomes (EE), late endosomes (EE), and lysosomes (Lys), where proteolysis continues. Antigens taken up by fluid phase internalization (blue) are subject to proteolysis generating peptides in early and late endosomes, as well as in lysosomes. Within endosomes, proteolytic enzymes degrade Ii generating class II-CLIP (DR-CLIP) complexes. In dendritic cells and macrophages (Figure 2A) antigens are internalized by fluid phase uptake, or by nonclonal receptors. In these cells, which do not express DO, DM activity is relatively high throughout the endosomal system (indicated by the bar to the left; yellow indicates low peptide exchange activity by DM, red indicates high activity), facilitating efficient CLIP release and loading of class II molecules wherever peptides become available. After peptide acquisition class II-peptide complexes (DR-pep) are transported to the cell surface.


ALFONSO

■

KARLSSON C-3

Figure 2B Internalization of exogenous antigens by B cells occurs by fluid phase and mIg receptor-mediated uptake (brown). As in other cell types, antigens internalized by fluid phase uptake are subject to proteolysis, generating peptides in early and late endosomes, as well as in lysosomes. Peptides generated in endosomes, however, are not favored for binding to class II molecules, since peptide exchange is poor at this pH when DM is complexed with DO (DMDO). mIg receptor-mediated uptake affords protection to antigens in the earlier parts of the endosomal pathway such that Ig-complexed antigens proceed to lysosomal compartments. Dissociation from the Ig molecule through proteolysis is favored in the acidic, highly proteolytic environment found in lysosomes. Here processing proceeds and, since downregulation by DO of DM-mediated peptide exchange is poor at lysosomal pH, the generated peptides can effectively bind class II molecules. Hence, peptides generated in lysosomal compartments may be favored for loading onto class II molecules, after which stable class II-peptide (DR-pep) complexes are transported to the B cell surface for presentation to CD4+ T cells.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


NEGATIVE REGULATION OF CYTOKINE SIGNALING PATHWAYS Hideo Yasukawa, Atsuo Sasaki, and Akihiko Yoshimura Institute of Life Science, Kurume University, Aikawamachi 2432–3, Kurume 839–0861, Japan, e-mail: [email protected]


Key Words JAK, STAT, CIS, JAB, SOCS, SSI, SH2 domain, tyrosine kinase, negative regulation, interferon gamma Abstract The Janus family of protein tyrosine kinases (JAKs) and STAT transcription factors regulate cellular processes involved in cell growth, differentiation, and transformation through their association with cytokine receptors. The CIS family of proteins (also referred to as the SOCS or SSI family) has been implicated in the regulation of signal transduction by a variety of cytokines. Most of them appear to be induced after stimulation with several different cytokines, and at least three of them (CIS1, CIS3/SOCS3, and JAB/SOCS1) negatively regulate cytokine signal transduction by various means: CIS1 inhibits STAT5 activation by binding to cytokine receptors that recruit STAT5, whereas JAB/SOCS-1 and CIS3/SOCS-3 directly bind to the kinase domain of JAKs, thereby inhibiting tyrosine-kinase activity. Therefore, these CIS family members seem to function in a classical negative feedback loop of cytokine signaling. Biochemical characterization as well as gene disruption studies indicate that JAB/SOCS1/SSI-1 is an important negative regulator of interferon c signaling. The mechanisms by which these inhibitors of cytokine signal transduction exert their effects have been extensively studied and will provide useful information for regulating tyrosine-kinase activity.

INTRODUCTION Growth, differentiation, and other functions of immune and hematopoietic cells are controlled by multiple cytokines, including interleukins (ILs) and colonystimulating factors (CSFs). Although lacking catalytic domains, members of the cytokine receptor superfamily mediate ligand-dependent activation of protein tyrosine phosphorylation through their association with and activation of members of the Janus kinase (JAK) family of protein tyrosine kinases. Cytokines induce homo- or heterodimerization and activation of their cognate receptors, resulting in the activation of associated JAK kinases (JAK1, JAK2, JAK3, and Tyk2). The activated JAKs phosphorylate the receptor cytoplasmic domains, thereby creating docking sites for SH2-containing signaling proteins. Among the substrates of tyrosine phosphorylation are members of the signal transducers and 0732–0582/00/0410–0143$14.00

143


144

YASUKAWA n SASAKI n YOSHIMURA

activators of the transcription family of proteins (STATs) (1). Although this pathway was initially found to be activated by interferons(s) (IFNs), it is now known that a large number of cytokines, growth factors, and hormonal factors activate JAK and/or STAT proteins. Various cytokines induce the tyrosine phosphorylation and activation of one or more of the seven STAT family members. Cytokines also activate Ras-MAP kinase pathways, phosphatidyl-inositol-3 (PI3) kinase, and phospholipase C-c (PLCc). The targeted disruption of genes for a number of Jaks and STATs in mice has revealed specific biological functions for many of them (for a review, see 2–4). For example, mice lacking JAK1 failed to manifest biologic responses to cytokines that bind to three distinct families of cytokine receptors: all class-II cytokine receptors, cytokine receptors that utilize the c(c) subunit for signaling, and the family of cytokine receptors that depend on the gp130 subunit for signaling. JAK2 gene disruption causes embryonic lethality due to the absence of definitive erythropoiesis. JAK3 knockout causes severe immunodeficiency. Mice in which the genes for STAT1, STAT4, and STAT6 are disrupted are viable but lack functions mediated by IFNs, interleukin 12 (IL-12), or IL-4, respectively, suggesting that these STATs perform very specific functions in immune responses. In contrast, the two highly related STAT5a and STAT5b are activated by various cytokines, including growth hormone (GH), prolactin (PRL), erythropoietin (EPO), interleukin-2 (IL-2), IL-3, granulocyte macrophage-colony stimulating factor (GMCSF), and thrombopoietin (TPO). The phenotypes of both STAT5a and STAT5b single-gene knockout mice and double knockout mice revealed an important role of STAT5s in PRL and GH signaling as well as in IL-2-dependent T cell functions (5–9). Over the years, the majority of investigations on cytokines have focused on the mechanisms by which they exert their actions. However, it is clear that the actions of cytokines are limited in both magnitude and duration, making it important to understand the mechanisms by which their actions are negatively controlled. Moreover, the downmodulation of one cytokine action by several factors, including another cytokine, are often found in important physiological circumstances. For example, Th1 and Th2 cytokines suppress each other (10). cAMP, calcium ionophore, and GM-CSF inhibit JAK1-STAT1 or STAT3 pathways (11), although the molecular basis of such modulation has not been well characterized. The deregulation of the negative feedback of the JAK-STAT pathway has been implicated in hematopoietic disorders, autoimmune and inflammatory diseases, interferon (IFN)-resistance, and cancer. For example, the constitutive activation of JAK and/or STATs has been implicated in transformation by v-src (12, 13). A constitutively active TEL-JAK fusion gene is implicated in T cell leukemia (14, 15). Mutations in the erythropoietin (EPO) receptor gene causing the deletion of the C-terminal region have been implicated in familial erythrocytosis, which is inherited as an autosomal dominant trait (16). In this disease, the disruption of the SHP-1 binding sites of the erythropoietin (EPO) receptor at the C-terminal

NEGATIVE REGULATION OF CYTOKINE SIGNALING

145


region has been shown to cause prolonged activation (phosphorylation) of JAK2 (17). In Table 1, the major regulatory mechanisms of JAK-STAT pathways are listed. The degradation of STATs and receptors should be involved in the regulation of signaling (18, 19). Naturally occurring dominant negative variants of STAT5 have also been reported (20, 21). Specific STAT1 and STAT3 inhibitors PIAS1 and PIAS3 have also been reported; however, their physiological meaning has not been clarified because these two genes are highly expressed without cytokine stimulation (22, 23). Tyrosine phosphatases containing the SH2 domain (SHP-1 and SHP-2) also regulate cytokine signaling, both positively and negatively. SHIP and Cbl have been implicated in the negative regulation of tyrosinekinase signaling and could also negatively regulate cytokine signaling (see below). However, recently, a new family of cytokine-inducible SH2 proteins (CISs) was identified and shown to be involved in the negative regulation of cytokine signals, especially JAKs and STATs.

SHP The activation of JAK tyrosine-kinase activity is positively regulated by the transphosphorylation of a critical tyrosine within the activation loop of the kinase domain (24–28). Considerable evidence suggests that one mechanism that terminates JAK activation involves the recruitment of a tyrosine phosphatase containing an SH2 domain (SHP-1) to receptor complexes resulting in the dephosphorylation of JAKs (28–30). The direct binding of JAK1 or JAK2 with SHP-1 has been demonstrated (31, 32). The potential importance of this mechanism is strongly suggested by the phenotype of motheaten (me/me) mice lacking SHP-1 that die from a disease with components of autoimmunity and inflammation (33). SHP-1 is a negative regulator of JAK kinases implicated in familial erythrocytosis (familial polycythemia) inherited as an autosomal dominant trait (16). In this disease, the disruption of the SHP-1 binding sites of the EPO receptor, as noted above, resulted in prolonged activation (phosphorylation) of JAK2 (17). SHP-1 is also implicated in familial hemophagocytic lymphohistiocytosis TABLE 1 Regulatory mechanism of the JAK-STAT pathway 1. Receptor-mediated endocytosis and degradation in lysosome 2. Dephosphorylation by tyrosine phosphatases, SHP and SHIP 3. Degradation by ubiquitin-proteasome systems, Cbl? SOCS box? 4. PIAS? 5. Dominant negative STATs 6. Expression of signaling molecules 7. CIS/SOCS/SSI family


146

(FHLH) (34). However, SHP-1 regulates a number of receptor and nonreceptor tyrosine kinases negatively, including c-kit and ZAP-70 (35, 36). Thus, it remains to be determined which kinases are specifically responsible for the phenotype. SHP-2, which is relatively ubiquitously expressed, has been implicated in positive regulation in cell growth and development. However, Kim et al found that SHP-2 negatively regulates gp130-mediated induction of acute-phase plasma protein genes in hepatic cells (37). You et al reported that SHP-2 also functions as a negative effector in IFN-induced growth-inhibitory and apoptotic pathways (38). They showed that treatment of mouse fibroblast cells lacking a functional SHP2 with IFN-a or IFN-c resulted in augmented suppression of cell viability compared to that of wild-type cells, probably by sustained activation of JAKs and STATs. Therefore, SHP-2 could function as a negative regulator of cytokine signaling in certain circumstances.


SHIP SH2-containing 5’-inositol phosphatase (SHIP) has been implicated in inhibitory signals not only from the immunotyrosine-based inhibition motif (ITIM) but also from some cytokine receptors. SHIP, which is widely expressed in hematopoietic cells, was first identified as a tyrosine phosphoprotein associated with Shc in response to numerous cytokines. Using SHIP–/– mast cells, Huber et al (1998) demonstrated that SHIP restricts the IgE-induced extracellular calcium entry by hydrolyzing phosphatidylinositol 3, 4, 5-trisphosphate [PI(3, 4, 5)P3] (39). SHIP–/– splenic B cells displayed prolonged Ca2` influx, increased proliferation in vitro, and enhanced mitogen-activated protein-kinase (MAPK) activation in response to B cell receptor–Fcc receptor IIB coligation (40). In addition, increased numbers of granulocyte-macrophage progenitors were observed in both the bone marrow and spleen of SHIP–/– mice, perhaps as a consequence of hyperresponsiveness to stimulation by the macrophage colony-stimulating factor, the granulocyte-macrophage colony-stimulating factor, interleukin-3, or the Steel factor as observed in vitro (41). Liu et al showed that engagement of the IL-3 and the GM-CSF receptors in SHIP –/– cells leads to increased and prolonged PI3 kinase-dependent PI(3, 4, 5)P3 accumulation and protein-kinase B (PKB) activation (42). These data indicate that SHIP negatively regulates the growth-factor and cytokine-mediated PKB activation as well as the cell survival of myeloid cells.

Cbl The adaptor molecule c-Cbl has also been demonstrated to be tyrosine phosphorylated in response to many stimuli, including cytokines (43, 44). However, genetic studies showed it to be a negative regulator of several tyrosine kinases. In C. elegans, the c-Cbl homologue Sli-1 acts as a negative regulator of the Ras homologue Let60 (45, 46). Moreover, several recent studies have demonstrated that mammalian c-Cbl has a negative regulatory function in intracellular signal



147

transduction, such as Syk tyrosine-kinase activation in mast cells (47), T cell receptor (TCR)-induced ERK2, and AP-1 activation (48). Cbl-b, a member of the Sli-1/c-Cbl protein family, has also been demonstrated to inhibit Vav-mediated cJun kinase activation (49). Indeed, c-Cbl gene knockout mice exhibited enhanced TCR signaling via Syk and ZAP-70 (50). The acceleration of ubiquitin/ proteasome-dependent degradation of receptor tyrosine kinases could be a mechanism of downregulation by c-Cbl (51, 52). Although the negative regulation of cytokine signals by c-Cbl has not been demonstrated, Ueno et al have reported that suppression of c-Cbl expression accelerated the EGF-dependent JAK/STAT pathway (54). Therefore, c-Cbl could also be a regulator of cytokine signaling. Previously, we cloned and characterized a novel adaptor molecule containing a Pleckstrin-homology (PH) domain and an Src homology-2 (SH2) domain, APS (55). APS was cloned using the yeast twohybrid system with the oncogenic c-kit kinase domain as bait. This protein forms a new subfamily of SH2 proteins with Lnk and SH2-B. APS is tyrosine phosphorylated in response to c-kit or B cell receptor stimulation; a single major tyrosine phosphorylation site, which is highly conserved in this family, was found at its C-terminus. The phosphorylation of the C-terminus creates a binding site for c-Cbl. Recently, an isoform of SH2-B (SH2-Bb) was shown to be a good substrate for JAK2; it was phosphorylated in response to IFNc and growth hormone (GH) (56). We also found that APS was highly phosphorylated in response to cytokine stimulation and suppressed erythropoietin (EPO)-induced STAT5 activation (57). Using the 293 cell-reconstitution systems, we showed that APS and c-Cbl synergistically inhibit cytokine-dependent STAT5 activation. Therefore, APS was suggested to function as a negative regulator of cytokine signaling in collaboration with c-Cbl.

CIS: Prototype of Cytokine-Induced SH2 Proteins The first member of this family was denoted as CIS (now called CIS1), for cytokine-inducible SH2 containing protein (58, 59) (Figure 1). CIS1 was cloned originally as an immediate early gene that was induced by IL-2, IL-3, and EPO. CIS1 was found to associate with the tyrosine phosphorylated EPO receptor and the IL-3 receptor b chain following stimulation with EPO or IL-3, respectively (58). Forced expression of CIS1 could partially suppress IL-3- or EPO-induced proliferation as well as STAT5 activation (58, 60). CIS1 binds specifically to the phosphorylated tyrosine residue of the EPO receptor, Y401 (61). This residue is one of the major STAT5 binding sites. Therefore, one possibility is that CIS1 might function by directly blocking phosphotyrosine motifs on receptors, preventing their coupling to stimulatory signaling pathways (Figure 2). Alternatively, given that CIS1 was observed to have a relatively short half-life, CIS1 might be a scavenger of tyrosine phosphorylated proteins, targeting them for degradation as well. It has been demonstrated that CIS1 itself is ubiquitinated (61). Thus, it is possible that CIS1 is involved in the ubiquitin/proteasome-dependent degradation


148


Figure 1 Schematic structure of the CIS/SOCS/SSI family and other CH domain/SOCS box-containing proteins. The numbers indicate amino acid length.

of the activated cytokine receptors by functioning as a ubiquitin-conjugating enzyme, an E3-like molecule. This is also supported by the recent finding that the C-terminal region of the CIS (CH domain or SOCS box) interacts with the Elongin B, C complex (62, 63) and probably with Rbx1, which activates cdc34like E2 molecules (64) (see below).



149

Figure 2 Signaling pathways activated in response to cytokines and action of three CIS family members. A schematic structure of the cytokine receptor dimer after ligand binding is shown. The shaded boxes in the cytoplasmic domain of the receptor represent boxes 1 and 2, which share homology among cytokine receptors. Two major signaling pathways, Ras-MAP kinase and JAK-STAT, are activated through cytokine receptors. CIS1 binds to the receptor and CIS3 and JAB bind to JAKs, thereby inhibiting signaling.

The CIS promoter contains two pairs of tandem TTCNNNGAA motifs that are capable of binding STAT5, and a CIS1 promoter reporter construct was induced by EPO in cells transfected with EPOR and STAT5 (60). The essential role of STAT5 in CIS expression was confirmed by the observation that CIS1 expression was not observed in the ovaries or thymus of STAT5a and b double knockout (KO) mice (8, 9). Therefore, CIS1 acts as a kind of negative feedback regulator of the JAK-STAT5 pathway. The negative effects of CIS1 on STAT5 activity were confirmed by the phenotypes observed in CIS1-transgenic mice (65). We created CIS1-transgenic mice under the control of b-actin promoter. These transgenic mice developed normally; however, their body size was smaller and their weight lower than those of the wild-type mice, suggesting a defect of growthhormone signaling. Female transgenic mice failed to lactate after parturition because of a failure of the terminal differentiation of the mammary gland, suggesting a defect of prolactin signaling. IL-2-dependent upregulation of the IL-2receptor a chain and proliferation were partially suppressed in splenocytes. These phenotypes remarkably resemble those found in STAT5a and/or STAT5b knockout mice (5–9). Indeed, STAT5 tyrosine phosphorylation was suppressed in the mammary gland and liver in CIS1 transgenic mice. Furthermore, IL-2-induced activation of STAT5 was markedly inhibited in T cells in transgenic mice, while leukemia inhibitory factor-induced STAT3 phosphorylation was not affected. We

150


also found that the number of cdT cells as well as that of natural killer (NK)-T cells was dramatically decreased in transgenic mice. Moreover, Th1/Th2 differentiation was also altered in CIS1-transgenic mice, although the mechanism is not clear at present. These data suggest that CIS1 functions as a specific negative regulator of STAT5 in vivo and plays an important regulatory role in T cell differentiation. Although there are no detailed data presented, CIS1-deficient mice have been shown to be phenotypically normal in all regards (mentioned in references 9 and 98). This strongly suggests a functional redundancy of CIS by related gene(s). Indeed, CIS turned out to be a member of a large family that is probably involved in regulation of cytokine signaling.


Identification of JAB/SOCS-1/SSI-1 Three groups, including ours, independently cloned the second CIS family member by entirely different experimental approaches. Starr et al cloned suppresser of cytokine signaling-1 (SOCS-1) as an inhibitor of IL-6-induced differentiation and growth arrest of the murine monocytic leukemia cell line M1 (66); Endo et al cloned the same protein as a JAK-binding protein (JAB) that could interact with the catalytic (JH1) domain of JAK2 (67); and Naka et al identified it by using an antibody that recognizes a common sequence of the SH2 domain of STATs and that was not a STAT protein but, rather, a STAT-induced STAT inhibitor-1 (SSI-1) (68) (Figure 2). When overexpressed, JAB/SOCS-1/SSI-1 can inhibit any signaling utilizing JAKs, such as STAT5 activation by EPO, STAT3 activation by leukemia inhibitory factor (LIF) or IL-6, and c-fos induction by IL2. However, since JAB did not inhibit either fibroblast growth-factor (FGF) induced c-fos activation or c-kit phosphorylation, although it binds to the FGF receptor and c-kit, JAB/SOCS-1/SSI-1 seems to be specific to JAK tyrosine kinases (69). Ohya et al also cloned the same gene (also called TIP-3) by twohybrid screening using a Tec kinase domain as bait (70). JAB/SOCS-1/SSI-1 was found to suppress Tec kinase activity. However, this inhibitory effect was marginal compared with that on JAKs. Recently, JAB/SOCS-1/SSI-1 has been shown to bind to the c-kit receptor tyrosine kinase via its SH2 domain (67, 71). Although JAB did not inhibit the catalytic activity of the c-kit tyrosine kinase, JAB inhibited c-kit mediated proliferation signals, probably by binding to the signaling proteins Grb-2 and the Rho-family guanine nucleotide-exchange factor Vav, via its SH3 domain (71). However, such effects seem to be observed only when JAB/SOCS-1 is extremely overexpressed at nonphysiological levels, and study of knockout mice suggests no connection between c-kit and JAB/SOCS-1 (see below). Losman et al reported that JAB/SOCS-1 potentially inhibits the activation of JAK1 kinase and STAT6 in response to IL-4 (72), suggesting that JAB/SOCS-1 is a negative regulator of IL-4 signaling. However, the most important physiological function of JAB/ SOCS-1 is the regulation of IFNc signaling (see below). Thus, as suggested by



151

Figure 3 Structure and model of kinase inhibition by JAB/SOCS-1 and CIS3/SOCS-3. (A) Schematic model of the functions of JAB and CIS3 domains. Essential amino acids in the kinase inhibitory region (KIR) and the extended SH2 subdomain are highlighted with bold circles. (B) The model of JH1 activation and inhibition by JAB/SOCS-1. Binding of JAB to the activation loop prevents the access of substrates and/or ATP to the catalytic pocket.

Venkataraman et al (73), JAB/SOCS-1, which is induced by IFNc, could play a key role in modulating IL-4 signaling, therefore in Th1/Th2 differentiation.

Involvement of JAB in Interferon-c Signaling IFNc is the most potent inducer of JAB/SOCS-1/SSI-1 in a wide variety of cell lines (74, 75). IFNb has no inducible effect. Previously, JAB/SOCS-1/SSI-1 was reported to be induced by IL-6 in mouse liver and M1 cells, suggesting that JAB/


152


SOCS-1 is a negative feedback regulator of IL-6/STAT3 (68). However, our experiments and those of others have suggested that IL-6 or LIF induces the expression of JAB much less efficiently than IFNc. JAB/SOCS-1 could be a negative feedback regulator of IFNc/STAT1 rather than IL-6/STAT3. Our recent study suggests that the dysregulated overexpression of JAB can confer IFN-resistance to cells (75). NIH-3T3 cells ectopically expressing JAB but not CIS3 lost responsiveness to the antiviral effect of IFNb and IFNc. M1 leukemic cells stably expressing JAB were also resistant to IFNc and IFNb-induced growth arrest. In both NIH-3T3 and M1 transformants expressing JAB, IFNc did not induce tyrosine phosphorylation or DNA-binding activity of STAT1. We also found that IFN-resistant clones derived from LoVo cells and Daudi cells expressed high levels of JAB without stimulation. In IFN-resistant LoVo and Daudi cells, IFN-induced STAT1 and JAK phosphorylation was partially reduced. Therefore, the overexpression of JAB could be, at least in part, a mechanism of IFNresistance with a dominant phenotype. It is clinically important to address the question of whether or not JAB is involved in the interferon resistance frequently found in patients with chronic myelogenous leukemia or those carrying the hepatitis type-C virus. Using M1 cells, we proposed that JAB/SOCS-1 induced by IFNc could be a mechanism of the antagonistic effect of IFNc against IL-6 or LIF (74). Pretreatment of M1 cells with IFNc increased the level of JAB/SOCS1 and prevented LIF- or IL-6-induced differentiation of M1 cells. Similarly, Venkataraman et al reported that IFNc stimulation results in a loss of IL-4-induced STAT6 tyrosine phosphorylation, nuclear translocation, and DNA binding (73). Using the fibrosarcoma cell-line U3A, which lacks Stat1, they demonstrated that the transcription-activation function of Stat1 is required for the IFNc-mediated repression. Treatment with IFNc, but not IL-4, specifically upregulates the expression of JAB/SOCS-1, and overexpression of JAB/SOCS-1 effectively blocks IL4-induced STAT6 phosphorylation and transcription. This suggests that the IFNc-mediated repression of several cytokine actions is at least in part mediated by JAB/SOCS-1.

Physiological Function of JAB/SOCS-1 Elucidated from Gene Disruption The most striking finding came from the phenotype of JAB/SOCS-1/SSI-1 KO mice (76–78). JAB/SOCS-1/SSI-1 -/- mice displayed growth retardation and died within three weeks after birth before weaning with fatty degeneration of the liver and monocytic infiltration of several organs. Lymphocytes in the thymus and spleen of the KO mice exhibited accelerated apoptosis with aging, and their number was 2–25% of that in wild-type mice at 10 days of age. Among various proand anti-apoptotic molecules examined, upregulation of Bax was found in the lymphocytes of the spleen and thymus of KO mice. In addition, a progressive loss occurred of maturing B lymphocytes in the bone marrow, spleen, and periph-


153


eral blood. Thus, JAB/SOCS-1/SSI-1 is required for the in vivo regulation of multiple cell types and is indispensable for normal postnatal growth and survival. Some of these phenotypes resemble those found in IFNc-transgenic mice (79, 80), and constitutive activation of STAT1 was found in JAB/SOCS-1/SSI-1–/– mice liver (J.C. Marine et al, unpublished data). Hematopoietic progenitor cells from JAB/SOCS-1/SSI-1–/– mice were hyperresponsive to inhibition by IFNc, the degree of inhibition varying markedly with the stimulating factor used (78). Importantly, much of the pathology of JAB/SOCS-1 –/– mice can be eliminated by injecting mice with antibodies to IFNc or by crossing the mice with IFNc knockout mice (81, 81a). Therefore, JAB/SOCS-1 is indeed a negative regulator of IFNc/STAT1 and probably plays an important role in the prevention of apoptosis induced by STAT1. Physiologically, it is also very interesting to examine the involvement of JAB/SOCS-1 in chronic inflammation.

Molecular Mechanism of Kinase Inhibition by JAB/SOCS-1 The activation of JAK tyrosine-kinase activity is positively regulated by the transphosphorylation of a critical tyrosine within the activation loop of the kinase domain. In the case of JAK2, that is the tyrosine residue Y1007. Recently, we demonstrated that JAB/SOCS-1 specifically binds to phosphorylated Y1007 (pY1007) in the activation loop of JAK2 (82). Extensive mutational analysis revealed the existence of three functional domains. Binding to the phosphorylated activation loop requires the JAB/SOCS-1 SH2 domain and an additional Nterminal 12 amino acids (extended SH2 subdomain) containing two residues (Ile68 and Leu75) that are conserved in JAB/SOCS-1 related proteins. An additional N-terminal 12 amino-acid region (kinase inhibitory region) of JAB/SOCS1 also contributes to high-affinity binding to the JAK2 tyrosine-kinase domain and is required for the inhibition of JAK2 signaling and kinase activity (Figure 3). The importance of the N-terminal region of JAB/SOCS-1 was also pointed out by Nicolson et al (83). These studies define a novel type of regulation of tyrosine kinases and may provide a basis for the design of specific tyrosine-kinase inhibitors (Figure 3). Without the kinase inhibitory region, JAB/SOCS-1 binds to the JAK2 kinase domain (JH1) with similar affinity to pY1007 phosphopeptide, suggesting that this region is a second binding site for JH1. We noticed that the kinase inhibitory region somewhat resembles the activation loop of JAKs. Therefore, we hypothesized that the kinase inhibitory region may mimic the activation loop by functioning as a pseudosubstrate. A similar structure and function relationship was found in CIS3/SOCS3, another JAK inhibitor (84). However, CIS3/SOCS3 did not inhibit JH1 kinase activity in vitro very efficiently (82, 83). Therefore, the true inhibitory mechanism by CIS3/SOCS-3 is still an open question. This hypothesis must be verified by kinetic analysis or more directly by co-crystal structure analysis using purified JAB/SOCS-1 and JH1.

154


Recently, other SH2 proteins, Grb10 and Grb14, have been shown to contain the second novel domain that interacts with the insulin receptor and insulin-like growth factor receptor in an activation loop-dependent manner (85, 86). These studies suggest that Grb10 and Grb14 may interact within or near the activation loop of the insulin receptor kinase domain, thereby physically blocking tyrosinekinase activity. Although there is no sequence similarity between the kinase inhibitory region of JAB and any domains of Grb10 or Grb14, it would be interesting to determine whether other tyrosine kinases possess their own specific inhibitory SH2 protein that interacts with the activation loop. It may also be possible to design a specific JAB-like tyrosine-kinase inhibitor acting against a given tyrosine kinase by modifying the SH2 domain to bind to the kinase activation loop as well as the kinase inhibitory region to the catalytic pocket of the kinase.


The CIS Family By database search, Masuhara et al reported five additional CIS family members (CIS2–6) (69) (Figure 1). Hilton’s group also identified one additional member (SOCS-7/CIS7; partial cDNA) (87). A partial cDNA of CIS5 was also cloned by Takenawa’s group as a binding protein for the SH3 domains of Nck, Ash (Grb2), PLCc (88). This gene was expressed mainly in brain. Others are expressed in hematopoietic and lymphoid tissues such as bone marrow, spleen, and thymus, and their expression is induced by a subset of cytokines in some hematopoietic cells (69). Thus, it is reasonable to call this family ‘‘the CIS family.’’ Figure 1 shows the structure of CIS family members (CISs). While the Nterminal regions share little sequence similarity, all seven contain an SH2 domain and a conserved C-terminal region that we designate as the CH (CIS-homology) domain, or others referred to as SOCS-box or SC-motif. The consensus motif in the CH domain is LXXPXXRX5SLQHLCRXXIX6–9IXXLPLPXXLXDYLXXYXY/F (X: any amino acid), although the entire length of the C-terminal region was varied. The SH2 domain exhibited 23–58% identity among CISs. Relatively high identity was seen between CIS1 and CIS2 (about 50%), CIS4 and CIS5 (about 58%), and JAB and CIS3 (about 40%). The -5 and -1 positions of the SH2 domain are fixed as L and G, respectively. Interestingly, the SH2 domains of STATs also contain the same amino acids at these positions (89). The Nterminal region is only 40–80 amino acids in length except for CIS4–7 which has a quite long N-terminal region (more than 350 amino acids). Among CIS family members, only CIS3/SOCS-3 and JAB/SOCS-1 can bind to JAK2-JH1 (69, 74, 84). Binding was confirmed by using two-hybrid analysis as well as co-precipitation assays in vitro and in vivo. The interaction of JAK2JH1 with JAB was stronger than that with CIS3/SOCS-3. By using a simple reporter-gene assay system, we found that JAB/SOCS-1 and CIS3/SOCS-3 almost completely blocked cytokine-dependent STAT3 and STAT5 reporter-gene activation (69). CIS2 and CIS4 had no effect or rather enhanced the reporter-gene activation. However, CIS3/SOCS-3 and JAB/SOCS-1 did not affect FGF- or



155

PDGF-induced c-fos promoter activation, suggesting that the effect of CIS3/ SOCS-3 and JAB is quite specific to JAK tyrosine kinases. CIS3/SOCS-3 is physiologically interesting, because it is expressed in a wide variety of tissues, including brain. Bjorbaek et al found that peripheral leptin administration to ob/ ob mice, but not db/db mice, rapidly induced CIS3/SOCS-3 mRNA in the hypothalamus, but had no effect on CIS1, JAB, or CIS2. In mammalian cell lines, CIS3/SOCS-3, but not CIS1 or CIS2, blocked leptin-induced signal transduction (90). The expression of CIS3/SOCS-3 mRNA in the arcuate and dorsomedial hypothalamic nuclei is increased in Ay/a mice, a model of leptin-resistant murine obesity. CIS3/SOCS-3 may be a leptin-inducible inhibitor of leptin signaling, and a potential mediator of leptin resistance in obesity. Auernhammer et al also found that injection of LIF or IL-1b increased CIS3/SOCS-3 mRNA in the pituitary, and, in AtT-20 cells, stable overexpression of CIS3/SOCS-3 inhibited basal and LIF-stimulated ACTH secretion (91). We also found that CIS3/SOCS-3 inhibited prolactin-induced STAT5 activation (92). CIS3/SOCS-3 was also implicated in the negative regulation of EPO signaling (84), IL-10 signaling (93, 94), and IL2 signaling (95). Recently, the promoter region of CIS3/SOCS-3 was characterized, and STAT3 binding sites were found in it (96). However, we recently reeported that CIS3/SOCS-3 is highly expressed in erythroid lineage cells during embryonic development (96a). And CIS3/SOCS3 deletion results in an embryonic lethality at 12–16 days associated with marked erythrocytosis. Moreover, the in vitro proliferative capacity of erythroid progenitors is greatly increased in response to EPO. These results demonstrate that CIS3SOCS3 is critical in negatively regulating fetal liver hematopoiesis, probably by negatively regulating EPO receptor/JAK2 signaling. Functions of CIS3/SOCS-3 in adult tissues remain to be clarified. CIS2/SOCS-2 binds to the insulin-like growth factor I (IGF-I)-receptor cytoplasmic kinase domain (97), but the functional meaning of this interaction is not clear.

Function of the CH-Domain/SOCS Box Database search also revealed that a similar CH domain is present in several proteins which contain ankyrin-like repeats, in the Ras-like GTPases and in proteins containing a WD domain (69). Hilton’s group called this domain ‘‘SOCS box’’ (66). They have identified 16 other proteins that contain this motif by using database search (87). These proteins fall into five classes based on the protein motifs found in the N-terminal of the CH domain. In addition to four new CIS proteins (SOCS-4 to SOCS-7) containing an SH2 domain, they describe four new families of proteins that contain either WD-40 repeats (WSB-1 and -2), SPRY domains (SSB-1 to -3), or ankyrin repeats (ASB-1 to -3) as well as G proteins. The CH domain may be involved in signal transduction because these domains in the N-terminal region are frequently found in signaling molecules, although

156


TABLE 2 List of CIS family members Other names CIS1 JAB CIS2 CIS3 CIS4 CIS5 CIS6 CIS7

CIS SOCS-1, SSI-1 SOCS-2, SSI-3 SOCS-3, SSI-3 SOCS-6 SOCS-7, NAP4 SOCS-5 SOCS-4

Amino Acids 257 211 197 225 535 .485 536 .521

Inductiona

Targetb

yes; STAT5 yes; STAT1 yes yes; STAT1,3 yes ? yes ?

cytokine receptors JAKs IGF-I receptor JAKs ? Grb2, PLCc ? ?

a


Induction of mRNA by various cytonkines in some hematopoietic cells. STATs possibly involved in induction are also indicated. b Apparent tyrosine phosphorylated proteins that bind to CISs. ? means not done or not identified.

this region is apparently not necessary for kinase inhibition by JAB or CIS3 (63, 82, 83, 84). Recently, Kamura et al reported that the Elongin BC complex interacts with the CH domain/SOCS box and prevents proteasome-dependent degradation of JAB/SOCS-1 (63). Elongin B is a ubiquitin-like protein, and Elongin C is a Skp1like protein that binds to a BC-box motif that is present in the Elongin A and CH domain/SOCS box. On the contrary, Zhang et al suggested that the SOCS box accelerates degradation of target proteins (including JAKs), from the analogous to the family of F-box-containing proteins, which mediate ubiquitin-proteasome degradation (62). The SOCS-box sequence has similarity to the a-domain of von Hippel-Lindau (VHL) tumor-suppressor gene product as well as the Skp2 F-box protein (98). The VHL protein is part of a complex that includes Elongin B/C, Cullin-2, and Rbx1, an evolutionarily conserved protein that contains a RING-H2 finger-like motif and interacts with Cullins (99). The yeast homologue of Rbx1 is a subunit and potent activator of the Cdc53-containing SCF-Cdc4 ubiquitin ligase (E2) required for the ubiquitination of the cyclin-dependent kinase inhibitor Sic1. Therefore, it is probable that JAB/SOCS-1 recruits the ubiquitin-ligation system into JAK2 by forming a complex with Elongin B/C, Rbx1, and Cullin-2. The CH domain/SOCS box-mediated degradation of JAK and cytokine receptors should be studied more intensively in the near future.

NOMENCLATURE Since CIS was the first member of this family to be identified and most of the family members are cytokine-inducible genes, we have referred to these proteins as the CIS family proteins. However, since SOCS is the most commonly used


157


one, we have also referred to SOCS by number in this review. Although SOCS, JAB, and SSI are all nomenclatures that have merit, the nomenclature should be agreed upon as soon as possible. In Table 2, the different names of each CIS member are listed. Three of the CIS family members, CIS1, JAB/SOCS-1, and CIS3/SOCS-3, have been extensively characterized. CIS1 is induced via STAT5 and then binds to the activated receptor, thereby inhibiting STAT5 activity by masking the STAT5 binding sites and/or promoting the degradation of the activated receptor. JAB/SOCS-1, and probably CIS3/SOCS-3, are induced by several cytokines (either STAT1 or STAT3 is probably involved in this induction) and then bind to JAK kinases directly, thereby inhibiting kinase activity. Although little is known about other CIS family members, a study of their physiological functions has been started. Understanding the mechanisms by which JAKs and STATs are inhibited by CISs/SOCSs is very important for developing specific pharmacological inhibitors of cytokine signaling as well as tyrosine kinases. ACKNOWLEDGMENTS We would like to express our appreciation to Ms. Miho Chikushi for preparing the manuscript. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Darnell JE Jr. 1997. STATs and gene regulation. Science 277:1630–35 2. Hoey T, Schindler U. 1998. STAT structure and function in signaling. Curr. Opin. Genet. Dev. 8:582–87 3. Liu KD, Gaffen SL, Goldsmith MA. 1998. JAK/STAT signaling by cytokine receptors. Curr. Opin. Immunol. 10:271– 78 4. Leonard WJ, O’Shea JJ. 1998. Jaks and STATs: biological implications. Annu. Rev. Immunol. 16:293–322 5. Udy GB, Towers RP, Snell RG, Wilkins RJ, Park SH, Ram PA, Waxman DJ, Davey HW. 1997. Requirement of STAT5b for sexual dimorphism of body growth rates and liver gene expression. Proc. Natl. Acad. Sci. USA 94:7239–44 6. Liu X, Robinson GW, Wagner KU, Garrett L, Wynshaw-Boris A, Hennighausen

L. 1997. Stat5a is mandatory for adult mammary gland development and lactogenesis. Genes Dev. 11:179–86 7. Nakajima H, Liu XW, Wynshaw-Boris A, Rosenthal LA, Imada K, Finbloom DS, Hennighausen L, Leonard WJ. 1997. An indirect effect of Stat5a in IL-2induced proliferation: a critical role for Stat5a in IL-2-mediated IL-2 receptor alpha chain induction. Immunity 7:691– 701 8. Teglund S, McKay C, Schuetz E, van Deursen JM, Stravopodis D, Wang D, Brown M, Bodner S, Grosveld G, Ihle JN. 1998. Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine responses. Cell 93:841–50 9. Moriggl R, Topham DJ, Teglund S, Sexl V, McKay C, Wang D, Hoffmeyer A, van

158

10.

11.


12.

13.

14.

15.

16.

17.

YASUKAWA n SASAKI n YOSHIMURA Deursen J, Sangster MY, Bunting KD, Grosveld GC, Ihle JN. 1999. Stat5 is required for IL-2-induced cell cycle progression of peripheral T cells. Immunity 10:249–59 Rook GA, Hernandez-Pando R, Lightman SL. 1994. Hormones, peripherally activated prohormones and regulation of the Th1/Th2 balance. Immunol. Today 15:301–3 Sengupta TK, Schmitt EM, Ivashkiv LB. 1996. Inhibition of cytokines and JAKSTAT activation by distinct signaling pathways. Proc. Natl. Acad. Sci. USA 93:9499–9504 Bromberg JF, Horvath CM, Besser D, Lathem WW, Darnell JE Jr. 1998. Stat3 activation is required for cellular transformation by v-src. Mol. Cell. Biol. 18:2553–58 Turkson J, Bowman T, Garcia R, Caldenhoven E, De Groot RP, Jove R. 1998. Stat3 activation by Src induces specific gene regulation and is required for cell transformation. Mol. Cell. Biol. 18: 2545–52 Lacronique V, Boureux A, Valle VD, Poirel H, Quang CT, Mauchauffe M, Berthou C, Lessard M, Berger R, Ghysdael J, Bernard OA. 1997. A TEL-JAK2 fusion protein with constitutive kinase activity in human leukemia. Science 278:1309–12 Peeters P, Raynaud SD, Cools J, Wlodarska I, Grosgeorge J, Philip P, Monpoux F, Van RL, Baens M, Van BH, Marynen P. 1997. Fusion of TEL, the ETS-variant gene 6 (ETV6), to the receptor-associated kinase JAK2 as a result of t (9;12) in a lymphoid and t (9;15;12) in a myeloid leukemia. Blood 90:2535–40 Kralovics R, Indrak K, Stopka T, Berman BW, Prchal JF, Prchal JT. 1997. Two new EPO receptor mutations: truncated EPO receptors are most frequently associated with primary familial and congenital polycythemias. Blood 90:2057–61 Klingmuller U, Lorenz U, Cantley LC,

18.

19.

20.

21.

22.

23.

24.

25.

26.

Neel BG, Lodish HF. 1995. Specific recruitment of SH-PTP1 to the erythropoietin receptor causes inactivation of JAK2 and termination of proliferative signals. Cell 80:729–38 Kim TK, Maniatis T. 1996. Regulation of interferon-gamma-activated STAT1 by the ubiquitin-proteasome pathway. Science 273:1717–19 Strous GJ, van Kerkhof P, Govers R, Ciechanover A, Schwartz AL. 1996. The ubiquitin conjugation system is required for ligand-induced endocytosis and degradation of the growth hormone receptor. EMBO J. 15:3806–12 Wang D, Stravopodis D, Teglund S, Kitazawa J, Ihle JN. 1996. Naturally occurring dominant negative variants of Stat5. Mol. Cell. Biol. 16:6141–48 Azam M, Lee C, Strehlow I, Schindler C. 1997. Functionally distinct isoforms of STAT5 are generated by protein processing. Immunity 6:691–701 Chung CD, Liao J, Liu B, Rao X, Jay P, Berta P, Shuai K. 1997. Specific inhibition of Stat3 signal transduction by PIAS3. Science 278:1803–5 Liu B, Liao J, Rao X, Kushner SA, Chung CD, Chang DD, Shuai K. 1998. Inhibition of Stat1-mediated gene activation by PIAS1. Proc. Natl. Acad. Sci. USA 95:10626–31 Gauzzi MC, Velazquez L, McKendry R, Mogensen KE, Fellous M, Pellegrini S. 1996. Interferon-alpha-dependent activation of Tyk2 requires phosphorylation of positive regulatory tyrosines by another kinase. J. Biol. Chem. 271: 20494–500 Feng J, Witthuhn BA, Matsuda T, Kohlhuber F, Kerr IM, Ihle JN. 1997. Activation of Jak2 catalytic activity requires phosphorylation of Y1007 in the kinase activation loop. Mol. Cell. Biol. 17: 2497–2501 Zhou YJ, Hanson EP, Chen YQ, Magnuson K, Chen M, Swann PG, Wange RL, Changelian PS, O’Shea JJ. 1997.


27.

28.


29.

30.

31.

32.

33.

34.

Distinct tyrosine phosphorylation sites in JAK3 kinase domain positively and negatively regulate its enzymatic activity. Proc. Natl. Acad. Sci. USA 94:13850–55 Liu KD, Gaffen SL, Goldsmith MA, Greene WC. 1997. Janus kinases in interleukin-2-mediated signaling: JAK1 and JAK3 are differentially regulated by tyrosine phosphorylation. Curr. Biol. 7:817–26 Weiss A, Schlessinger J. 1998. Switching signals on or off by receptor dimerization. Cell 94:277–80 Haque SJ, Harbor P, Tabrizi M, Yi T, Williams BR. 1998. Protein-tyrosine phosphatase Shp-1 is a negative regulator of IL-4- and IL-13-dependent signal transduction. J. Biol. Chem. 273:33893– 96 Migone TS, Cacalano NA, Taylor N, Yi T, Waldmann TA, Johnston JA. 1998. Recruitment of SH2-containing protein tyrosine phosphatase SHP-1 to the interleukin 2 receptor; loss of SHP-1 expression in human T-lymphotropic virus type I-transformed T cells. Proc. Natl. Acad. Sci. USA 95:3845–50 David M, Chen HE, Goelz S, Larner AC, Neel BG. 1995. Differential regulation of the alpha/beta interferon-stimulated Jak/ Stat pathway by the SH2 domaincontaining tyrosine phosphatase SHPTP1. Mol. Cell. Biol. 15:7050–58 Jiao H, Berrada K, Yang W, Tabrizi M, Platanias LC, Yi T. 1996. Direct association with and dephosphorylation of Jak2 kinase by the SH2-domain-containing protein tyrosine phosphatase SHP-1. Mol. Cell. Biol. 16:6985–92 Shultz LD, Schweitzer PA, Rajan TV, Yi T, Ihle JN, Matthews RJ, Thomas ML, Beier DR. 1993. Mutations at the murine motheaten locus are within the hematopoietic cell protein-tyrosine phosphatase (Hcph) gene. Cell 73:1445–54 Tabrizi M, Yang W, Jiao H, DeVries EM, Platanias LC, Arico M, Yi T. 1998.

35.

36.

37.

38.

39.

40.

41.

159

Reduced Tyk2/SHP-1 interaction and lack of SHP-1 mutation in a kindred of familial hemophagocytic lymphohistiocytosis. Leukemia 12:200–6 Kozlowski M, Larose L, Lee F, Le DM, Rottapel R, Siminovitch KA. 1998. SHP1 binds and negatively modulates the cKit receptor by interaction with tyrosine 569 in the c-Kit juxtamembrane domain. Mol.Cell. Biol. 18:2089–99 Plas DR, Johnson R, Pingel JT, Matthews RJ, Dalton M, Roy G, Chan AC, Thomas ML. 1996. Direct regulation of ZAP-70 by SHP-1 in T cell antigen receptor signaling. Science 272:1173–76 Kim H, Hawley TS, Hawley RG, Baumann H. 1998. Protein tyrosine phosphatase 2 (SHP-2) moderates signaling by gp130 but is not required for the induction of acute-phase plasma protein genes in hepatic cells. Mol. Cell. Biol. 18:1525–33 You M, Yu DH, Feng GS. 1999. Shp-2 tyrosine phosphatase functions as a negative regulator of the interferonstimulated Jak/STAT pathway. Mol. Cell. Biol. 19:2416–24 Huber M, Helgason CD, Damen JE, Liu L, Humphries RK, Krystal G. 1998. The src homology 2-containing inositol phosphatase (SHIP) is the gatekeeper of mast cell degranulation. Proc. Natl. Acad. Sci. USA 95:11330–35 Liu Q, Oliveira-Dos-Santos AJ, Mariathasan S, Bouchard D, Jones J, Sarao R, Kozieradzki I, Ohashi PS, Penninger JM, Dumont DJ. 1998. The inositol polyphosphate 5-phosphatase ship is a crucial negative regulator of B cell antigen receptor signaling. J. Exp. Med. 188:1333–42 Helgason CD, Damen JE, Rosten P, Grewal R, Sorensen P, Chappel SM, Borowski A, Jirik F, Krystal G, Humphries RK. 1998. Targeted disruption of SHIP leads to hemopoietic perturbations, lung pathology, and a shortened life span. Genes Dev. 12:1610–20


160


42. Liu Q, Sasaki T, Kozieradzki I, Wakeham A, Itie A, Dumont DJ, Penninger JM. 1999. SHIP is a negative regulator of growth factor receptor-mediated PKB/ Akt activation and myeloid cell survival. Genes Dev. 13:786–91 43. Barber DL, Mason JM, Fukazawa T, Reedquist KA, Druker BJ, Band H, D’Andrea AD. 1997. Erythropoietin and interleukin-3 activate tyrosine phosphorylation of CBL and association with CRK adaptor proteins. Blood 89:3166– 74 44. Odai H, Sasaki K, Iwamatsu A, Hanazono Y, Tanaka T, Mitani K, Yazaki Y, Hirai H. 1995. The proto-oncogene product c-Cbl becomes tyrosine phosphorylated by stimulation with GM-CSF or Epo and constitutively binds to the SH3 domain of Grb2/Ash in human hematopoietic cells. J. Biol. Chem. 270: 10800–5 45. Jongeward GD, Clandinin TR, Sternberg PW. 1995. sli-1, a negative regulator of let-23-mediated signaling in C. elegans. Genetics 139:1553–66 46. Yoon CH, Lee J, Jongeward GD, Sternberg PW. 1995. Similarity of sli-1, a regulator of vulval development in C. elegans, to the mammalian protooncogene c-cbl. Science 269:1102–5 47. Ota Y, Samelson LE. 1997. The product of the proto-oncogene c-cbl: a negative regulator of the Syk tyrosine kinase. Science 276:418–20 48. Rellahant BL, Graham LJ, Stoica B, DeBell KE, Bonvini E. 1997. Cblmediated regulation of T cell receptorinduced AP1 activation. Implications for activation via the Ras signaling pathway. J. Biol. Chem. 272:30806–11 49. Keane MM, Rivero-lezcano OM, Mitcell JA, Robbins KC, Lipkowitz S. 1995. Cloning and characterization of cbl-b: a SH3 binding protein with homology to the c-cbl proto-oncogene. Oncogene 10:2367–77 50. Murphy MA, Schnall RG, Venter DJ,

51.

52.

53.

54.

55.

56.

57.

Barnett L, Bertoncello I, Thien CB, Langdon WY, Bowtell DD. 1998. Tissue hyperplasia and enhanced T-cell signalling via ZAP-70 in c-Cbl-deficient mice. Mol. Cell. Biol. 18:4872–82 Bowtell DDL, Langdon WY. 1995. The protein product of the c-cbl oncogene rapidly complexes with the EGF receptor and is tyrosine phosphorylated following EGF stimulation. Oncogene 11:1561–67 Miyake S, Lupher ML Jr, Druker B, Band H. 1998. The tyrosine kinase regulator Cbl enhances the ubiquitination and degradation of the platelet-derived growth factor receptor alpha. Proc. Natl. Acad. Sci. USA 95:7927–32 Levkowitz G, Waterman H, Zamir E, Kam Z, Oved S, Langdon WY, Beguinot L, Geiger B, Yarden Y. 1998. c-Cbl/Sli1 regulates endocytic sorting and ubiquitination of the epidermal growth factor receptor. Genes Dev. 12:3663–74 Ueno H, Sasaki K, Miyagawa K, Honda H, Mitani K, Yazaki Y, Hirai H. 1997. Antisense repression of proto-oncogene c-Cbl enhances activation of the JAKSTAT pathway but not the ras pathway in epidermal growth factor receptor signaling. J. Biol. Chem. 272:8739–43 Yokouchi M, Suzuki R, Masuhara M, Komiya S, Inoue A, Yoshimura A. 1997. Cloning and characterization of APS, an adaptor molecule containing PH and SH2 domains that is tyrosine phosphorylated upon B-cell receptor stimulation. Oncogene 15:7–15 Rui L, Mathews LS, Hotta K, Gustafson TA, Carter-Su C. 1997. Identification of SH2-Bbeta as a substrate of the tyrosine kinase JAK2 involved in growth hormone signaling. Mol. Cell. Biol. 17: 6633–44 Wakioka T, Sasaki A, Mitsui K, Yokouchi M, Inoue A, Komiya S, Yoshimura A. 1999. APS, an adaptor protein containing PH and SH2 domains inhibits the JAK-STAT pathway in collaboration with c-Cbl. Leukemia 13:760–67


NEGATIVE REGULATION OF CYTOKINE SIGNALING 58. Yoshimura A, Ohkubo T, Kiguchi T, Jenkins NA, Gilbert DJ, Copeland NG, Hara T, Miyajima A. 1995. A novel cytokineinducible gene CIS encodes an SH2 containing protein that binds to tyrosinephosphorylated interleukin 3 and erythropoietin receptors. EMBO J. 14:2816– 26 59. Uchida K, Yoshimura A, Inazawa J, Yanagisawa K, Osada H, Masuda A, Saito T, Takahashi T, Miyajima A, Takahashi T. 1997. Molecular cloning of CISH, chromosome assignment to 3p21.3, and analysis of expression in fetal and adult tissues. Cytogenet. Cell. Genet. 78:209– 12 60. Matsumoto A, Masuhara M, Mitsui K, Yokouchi M, Ohtsubo M, Misawa H, Miyajima A, Yoshimura A. 1997. CIS, a cytokine inducible SH2 protein, is a target of the JAK-STAT5 pathway and modulates STAT5 activation. Blood 89:3148–54 61. Verdier F, Chretien S, Muller O, Varlet P, Yoshimura A, Gisselbrecht S, Lacombe C, Mayeux P. 1998. Proteasomes regulate erythropoietin receptor and signal ransducer and activator of transcription 5(STAT5) activation. Possible involvement of the ubiquitinated CIS protein. J. Biol. Chem. 273:28185–90 62. Zhang JG, Farley A, Nicholson SE, Willson TA, Zugaro LM, Simpson RJ, Moritz RL, Cary D, Richardson R, Hausmann G, Kile BJ, Kent SB, Alexander WS, Metcalf D, Hilton DJ, Nicola NA, Baca M. 1999. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. Proc. Natl. Acad. Sci. USA 96:2071–76 63. Kamura T, Sato S, Haque D, Liu L, Kaelin WG Jr, Conaway RC, Conaway JW. 1998. The Elongin BC complex interacts with the conserved SOCS-box motif present in members of the SOCS, ras,

64.

65.

66.

67.

68.

69.

70.

161

WD-40 repeat, and ankyrin repeat families. Genes Dev. 12:3872–81 Tyers M, Willems AR. 1999. One ring to rule a superfamily of E3 ubiquitin ligases. Science 284:601,603–4 Matsumoto A, Seki Y, Kubo M, Ohtsuka S, Suzuki A, Hayashi I, Tsuji K, Nakahata T, Okabe M, Yamada S, Yoshimura A. 1999. Suppression of STAT5 functions in liver, mammary glands, and T cells in cytokine-inducible SH2 protein1 (CIS1) transgenic mice Mol. Cell. Biol. 19:6396–6407 Starr R, Willson TA, Viney EM, Murray LJ, Rayner JR, Jenkins BJ, Gonda TJ, Alexander WS, Metcalf D, Nicola NA, Hilton DJ. 1997. A family of cytokineinducible inhibitors of signalling. Nature 387:917–21 Endo TA, Masuhara M, Yokouchi M, Suzuki R, Sakamoto H, Mitsui K, Matsumoto A, Tanimura S, Ohtsubo M, Misawa H, Miyazaki T, Leonor N, Taniguchi T, Fujita T, Kanakura Y, Komiya S, Yoshimura A. 1997. A new protein containing an SH2 domain that inhibits JAK kinases. Nature 387:921–24 Naka T, Narazaki M, Hirata M, Matsumoto T, Minamoto S, Aono A, Nishimoto N, Kajita T, Taga T, Yoshizaki K, Akira S, Kishimoto T. 1997. Structure and function of a new STAT-induced STAT inhibitor. Nature 387:924–29 Masuhara M, Sakamoto H, Matsumoto A, Suzuki R, Yasukawa H, Mitsui K, Wakioka T, Tanimura S, Sasaki A, Misawa H, Yokouchi M, Ohtsubo M, Yoshimura A. 1997. Cloning and characterization of novel CIS family genes. Biochem. Biophys. Res. Commun. 239:439–46 Ohya Ki, Kajigaya S, Yamashita Y, Miyazato A, Hatake K, Miura Y, Ikeda U, Shimada K, Ozawa K, Mano H. 1997. SOCS-1/JAB/SSI-1 can bind to and suppress Tec protein-tyrosine kinase. J. Biol. Chem. 272:27178–82


162


71. De Sepulveda P, Okkenhaug K, Rose JL, Hawley RG, Dubreuil P, Rottapel R. 1999. Socs1 binds to multiple signalling proteins and suppresses steel factordependent proliferation. EMBO J. 18:904–5 72. Losman JA, Chen XP, Hilton D, Rothman P. 1999. Cutting edge: SOCS-1 is a potent inhibitor of IL-4 signal transduction. J. Immunol. 162:3770–74 73. Venkataraman C, Leung S, Salvekar A, Mano H, Schindler U. 1999. Repression of IL-4-induced gene expression by IFNgamma requires Stat1 activation. J. Immunol. 162:4053–61 74. Suzuki R, Sakamoto H, Yasukawa H, Masuhara M, Wakioka T, Sasaki A, Yuge K, Komiya S, Inoue A, Yoshimura A. 1998. CIS3 and JAB have different regulatory roles in interleukin-6 mediated differentiation and STAT3 activation in M1 leukemia cells. Oncogene 17:2271– 78 75. Sakamoto H, Yasukawa H, Masuhara M, Tanimura S, Sasaki A, Yuge K, Ohtsubo M, Ohtsuka A, Fujita T, Ohta T, Furukawa Y, Iwase S, Yamada H, Yoshimura A. 1998. A Janus kinase inhibitor, JAB, is an interferon-gamma-inducible gene and confers resistance to interferons. Blood 92:1668–76 76. Naka T, Matsumoto T, Narazaki M, Fujimoto M, Morita Y, Ohsawa Y, Saito H, Nagasawa T, Uchiyama Y, Kishimoto T. 1998. Accelerated apoptosis of lymphocytes by augmented induction of Bax in SSI-1 (STAT-induced STAT inhibitor-1) deficient mice. Proc. Natl. Acad. Sci. USA 95:15577–82 77. Starr R, Metcalf D, Elefanty AG, Brysha M, Willson TA, Nicola NA, Hilton DJ, Alexander WS. 1998. Liver degeneration and lymphoid deficiencies in mice lacking suppressor of cytokine signaling-1. Proc. Natl. Acad. Sci. USA 95:14395–99 78. Metcalf D, Alexander WS, Elefanty AG, Nicola NA, Hilton DJ, Starr R, Mifsud S, Di Rago L. 1999. Aberrant hemato-

79.

80.

81.

81a.

82.

83.

poiesis in mice with inactivation of the gene encoding SOCS-1. Leukemia 13:926–34 Toyonaga T, Hino O, Sugai S, Wakasugi S, Abe K, Shichiri M, Yamamura K. 1994. Chronic active hepatitis in transgenic mice expressing interferon-gamma in the liver. Proc. Natl. Acad. Sci. USA 91:614–18 Young HA, Klinman DM, Reynolds DA, Grzegorzewski KJ, Nii A, Ward JM, Winkler-Pickett RT, Ortaldo JR, Kenny JJ, Komschlies KL. 1997. Bone marrow and thymus expression of interferongamma results in severe B-cell lineage reduction, T-cell lineage alterations, and hematopoietic progenitor deficiencies. Blood 89:583–95 Alexander WS, Starr R, Fenner JF, Scott CL, Handman E, Sprigg NS, Corbin JE, Cornish AL, Darwiche R, Owczarek CM, Kay TWH, Nicola NA, Hertzog PJ, Metcalf D, Hilton DJ. 1999. SOCS1 is a critical inhibitor of interferon gamma signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell 98:597–608 Marine J-C, Topham DJ, McKay C, Wang D, Parganas E, Stravopodis D, Yoshimura A, Ihle JN. 1999. SOCS1 deficiency causes a lymphocytedependent perinatal lethality. Cell 98:609–16 Yasukawa H, Misawa H, Sakamoto H, Masuhara M, Sasaki A, Wakioka T, Ohtsuka S, Imaizumi T, Matsuda T, Ihle JN, Yoshimura A. 1999. The JAK-binding protein JAB inhibits Janus tyrosine kinase activity through binding in the activation loop. EMBO J. 18:1309–20 Nicholson SE, Willson TA, Farley A, Starr R, Zhang JG, Baca M, Alexander WS, Metcalf D, Hilton DJ, Nicola NA. 1999. Mutational analyses of the SOCS proteins suggest a dual domain requirement but distinct mechanisms for inhibition of LIF and IL-6 signal transduction. EMBO J. 18;375–85


NEGATIVE REGULATION OF CYTOKINE SIGNALING 84. Sasaki A, Yasukawa H, Suzuki A, Kamizono S, Syoda T, Kinjyo I, Sasaki M, Johnston JA, Yoshimura A. 1999. Cytokine-inducible SH2 protein-3 (CIS3) inhibits Janus tyrosine kinase by binding through the N-terminal kinase inhibitory region as well as SH2 domain. Genes to Cells 4:339–51 85. He W, Rose DW, Olefsky JM, Gustafson TA. 1998. Grb10 interacts differentially with the insulin receptor, insulin-like growth factor I receptor, and epidermal growth factor receptor via the Grb10 Src homology 2 (SH2) domain and a second novel domain located between the pleckstrin homology and SH2 domains. J. Biol.Chem. 273:6860–67 86. Kasus-Jacobi A, Perdereau D, Auzan C, Clauser E, Van Obberghen E, MauvaisJarvis F, Girard J, Burnol A. 1998. Identification of the rat adapter Grb14 as an inhibitor of insulin actions. J. Biol. Chem. 273:26026–35 87. Hilton DJ, Richardson RT, Alexander WS, Viney EM, Willson TA, Sprigg NS, Starr R, Nicholson SE, Metcalf D, Nicola NA. 1998. Twenty proteins containing a C-terminal SOCS box form five structural classes. Proc. Natl. Acad. Sci. USA 95:114–19 88. Matuoka K, Miki H, Takahashi K, Takenawa T. 1997. A novel ligand for an SH3 domain of the adaptor protein Nck bears an SH2 domain and nuclear signaling motifs. Biochem. Biophys. Res. Commun. 239:488–92 89. Chen X, Vinkemeier U, Zhao Y, Jeruzalmi D, Darnell JE Jr, Kuriyan J. 1998. Crystal structure of a tyrosine phosphorylated STAT-1 dimer bound to DNA. Cell 93:827–39 90. Bjorbaek C, Elmquist JK, Frantz JD, Shoelson SE, Flier JS. 1998. Identification of SOCS-3 as a potential mediator of central leptin resistance. Mol. Cell 4:619–25 91. Auernhammer CJ, Chesnokova V, Bousquet C, Melmed S. 1998. Pituitary cor-

92.

93.

94.

95.

96.

96a.

163

ticotroph SOCS-3: novel intracellular regulation of leukemia-inhibitory factormediated proopiomelanocortin gene expression and adrenocorticotropin secretion. Mol. Endocrinol. 12:954–61 Helman D, Sandowski, Y, Cohen Y, Matsumoto A, Yoshimura A, Merchav S, Gertler A.1998 Cytokine-inducible SH2 protein (CIS3) and JAK2 binding protein (JAB) abolish prolactin receptormediated STAT5 signaling FEBS Lett. 441,287–91 Ito S, Ansari P, Sakatsume M, Dickensheets H, Vazquez N, Donnelly RP, Larner AC, Finbloom DS. 1999. Interleukin-10 inhibits expression of both interferon alpha- and interferon gamma- induced genes by suppressing tyrosine phosphorylation of STAT1. Blood 93:1456–63 Cassatella MA, Gasperini S, Bovolenta C, Calzetti F, Vollebregt M, Scapini P, Marchi M, Suzuki R, Suzuki A, Yoshimura A. 1999. IL-10 selectively enhances CIS3SOCS3 expression in human neutrophils. Evidence for an IL10-induced pathway that is independent of STAT protein activation. Blood. 94:2880–89 Cohney SJ, Sanden D, Cacalano NA, Yoshimura A, Mui A, Migone TS, Johnston JA. 1999. SOCS-3 is tyrosine phosphorylated in response to interleukin-2 and suppresses STAT5 phosphorylation and lymphocyte proliferation. Mol. Cell. Biol. 19:4980–88 Auernhammer CJ, Bousquet C, Melmed S. 1999. Autoregulation of pituitary corticotroph SOCS-3 expression: characterization of the murine SOCS-3 promoter. Proc. Natl. Acad. Sci. USA 96:6964–69 Marine J-C, McKay C, Wang D, Topham DJ, Parganas E, Nakajima H, Pendeville H, Yasukawa H, Sasaki A, Yoshimure A, Ihle JN. 1999. SOCS3 is essential in the regulation of fetal liver erythropoiesis. Cell 98: 617–27

164



97. Dey BR, Spence SL, Nissley P, Furlanetto RW. 1998. Interaction of human suppressor of cytokine signaling (SOCS)-2 with the insulin-like growth factor-I receptor. J. Biol. Chem. 273:24095–101 98. Stebbins CE, Kaelin WG Jr, Pavletich NP. 1999. Structure of the VHLElonginC-ElonginB complex: implica-

tions for VHL tumor suppressor function. Science 284:455–61 99. Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin WG Jr, Elledge SJ, Conaway RC, Harper JW, Conaway JW. 1999. Rbx1, a component of the VHL tumor suppressor complex and SCF ubiquitin ligase. Science 284:657–61


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


T CELL ACTIVATION AND THE CYTOSKELETON Oreste Acuto1 and Doreen Cantrell2 1 Molecular Immunology Unit, Department of Immunology, Pasteur Institute, 75724, Paris, cedex 15, France; e-mail: [email protected] 2 Lymphocyte Activation Laboratory, Imperial Cancer Research Funds, Lincolns Inn Field, London WC2A 3PX, United Kingdom; e-mail: [email protected]


Key Words T cell antigen receptor, CD28, protein tyrosine kinases, protein kinase C, Rho GTPases Abstract Ligation of the T cell antigen receptor (TCR) stimulates protein tyrosine kinases (PTKs), which regulate intracellular calcium and control the activity of protein kinase C (PKC) isozymes. PTKs activated by antigen receptors and costimulatory molecules also couple to phosphatidylinositol-3 kinase (PI3K) and control the activity of Ras- and Rho-family GTPases. T cell signal transduction is triggered physiologically by antigen in the context of antigen presenting cells (APC). The formation of stable and prolonged contacts between T cells and APCs is not neccessary to initiate T cell signaling but is required for effective T cell proliferation and differentiation. The stabilization of the T cell/ APC conjugate is regulated by intracellular signals induced by antigen receptors and costimulators. These coordinate the regulation of the actin and microtubule cytoskeleton and organize a specialized signaling zone that allows sustained TCR signaling.

INTRODUCTION The coordinated activation of T cells in response to foreign antigen ensures antigen-specific T cell clonal expansion and differentiation. Lymphocyte activation is controlled by signaling pathways initiated by antigen receptors and costimulatory molecules. The physiological ligand for T cell activation is a foreign peptide bound to major histocompatibility complex (MHC)-encoded molecules presented on the surface of professional APC such as dendritic cells (DC). The TCR comprises a/b subunits that recognise peptide/MHC and the signal transducing subunits e, c, d, and f (CD3-f complex). Biochemical signals initiated by this receptor determine the specificity of T cell activation, but events initiated by other membrane proteins such as the MHC receptors CD4/CD8, costimulators like CD2, CD28, and integrins modulate the intracellular signaling thresholds required to initiate a T cell immune reaction (1, 2). It is generally recognized that the integrity of the actin and microtubule cytoskeleton provides an essential structural support for T cell activation (3–6). The first part of this review describes 0732–0582/00/0410–0165$14.00

165

166

ACUTO n CANTRELL

the T cell signaling pathways identified biochemically in experiments with agonistic anti-TCR/CD3 antibodies. The second part will document the morphological changes that occur during T cell activation and cover recent progress towards understanding how symbiotic interactions between TCR signal transduction pathways and the cytoskeleton drive T cell activation.

THE INDUCTION OF TCR SIGNALING


The Regulation of Protein Tyrosine Kinases: the Importance of ZAP70/Lck Interaction Conventional biochemistry has established that the TCR initiates signaling by recruiting and activating protein tyrosine kinases (PTKs) of the Src, Syk, and Tec families. Moreover, genetic studies in mice and humans reveal that these PTKs are indispensible signaling molecules for T cell activation. An exhaustive description of the molecular events of this phase of the activation process has been reported (7–9). In this review we focus on the essential facts pertinent for understanding how TCR signaling initiates. Immunoreceptor tyrosine-based activation motifs (ITAMs) of the signal transducing subunits CD3 c, d, e and f chain are phosphorylated by Src PTKs, essentially Lck, thus making the ITAMs competent to associate with ZAP70 via tandem Src-homology 2 (SH2) domains characteristic of Syk family PTKs (7). ITAMbound ZAP70 is tyrosine phosphorylated and Lck binds to it via its own Srchomology 2 (SH2) domain (10, 11); ZAP70 catalytic activity is considerably increased as a consequence of Lck phosphorylation of the ZAP70 activation loop (12, 13). This initial process leads to the phosphorylation of ZAP70 substrates (e.g. SLP76, LAT) (14, 15) and may induce and/or amplify the activation of other PTKs (e.g., Src and Tec) downstream of ZAP70 (16). The LckSH2 domain binds Tyr319 in the linker region of ZAP70 (17) and predictibly, this tyrosine is essential for ZAP70 activation (18). Moreover, mutation of ZAP70 Tyr319 into a motif with three to fourfold increase in affinity for Lck (17) (and microcalorimetry data, F. Schaeffer, V. Di Bartolo, O. Acuto, unpublished data) causes a twenty to thirty fold increase in ZAP70 functional activity (17). Hence the SH2-mediated interaction of Lck with ZAP70 is of central importance; it promotes continued ITAMs phosphorylation with additional recuitment of ZAP70 and other lck substrates such as the Tec PTK Itk (19).

ITAM Phosphorylation Pre-Exists before T Cell Activation The stoichiometry of ITAM and ZAP-70 tyrosine phosphorylation is regulated by antigen and will be low if the TCR interaction with peptide/MHC is weak (20, 21). Phosphorylation is only led to completion, by strong TCR-ligand interactions or by the help of co-receptors CD4/CD8 (22). In quiescent T cells phosphorylation


CYTOSKELETON IN T CELL SIGNALING

167

of ITAMs by a pool of active Src PTKs is an ongoing reaction masked by the overwhelming repressive capacity of one or more tyrosine phosphatases (14, 15, 17, 23). CD45 is a good candidate (Figure 1). TCR ligation by antigen shifts this equilibrium with the driving force for amplification of the response almost certain to be exclusion of phosphatase molecules away from the TCR-ligand complex. This model was proposed by Davis & van der Merwe (24), who noted that the size of the TCR/MHC complex will ensure membrane approximation within 150 nm during T cell activation by antigen on the surface of an APC. Consequently, membrane molecules with large extracellular domains will be segregated out of the TCR/MHC contact whereas small-sized costimulatory receptors such as CD2 and CD28 would not (Figure 2). The tyrosine phosphatase CD45, which acts as a powerful negative regulator of TCR signaling (27), has a large extracellular domain that would be excluded on a size basis from proximity of a TCR/MHC complex. A sudden change in the local distribution of CD45 will promote a sudden increase in phosphorylation of ITAMs by proximal Src kinases. Note that CD45 also has a positive role in T cell signaling because this phosphatase is responsible for maintaining a pool of active Lck in quiescent cells (25, 26). T cell signal transduction should thus be viewed as a perturbation of an existing

Figure 1 A model of T cell activation based on a dynamic phosphorylation/dephosphorylation equilibrium. (Top) Steady-state level of ITAMs phosphorylation by Lck is of very low stoichiometry in an unstimulated T cell as it is counteracted by protein tyrosine phosphatases (PTPs). (Bottom) When the Ag/MHC: TCR interaction has a sufficient halflife, the balance shifts towards accumulation of phosphorylated ITAMs against a negative effect by PTPs. This effect may be reached by sudden local inactivation/physical exclusion of PTPs. Once ITAMs are sufficiently stably phosphorylated, ZAP70 binds, protects the ITAMs from dephosphorylation and all the downstream steps can proceed to completion (e.g. activation of ZAP-70 by Lck and substrates phosphorylation).


168

ACUTO n CANTRELL

Figure 2 A model explaining how CD28 can facilitate TCR engagement and signaling. Once ligated by B7, CD28 provides the T cell with an initial adhesion capable of approximating the T cell and APC membranes. Concomitantly, Fyn/Lck PTKs phosphorylate Vav and activate its GEF activity. Rac, the main Vav target then induces marked cortical actin changes that may favor an initial membrane re-organization leading to the formation of SMACs (e.g., receptor segretation and focalization of signaling protein complexes associated to coalescing membrane rafts). This process induced by CD28 may enhance membrane proximal signaling from the TCR.

phosphorylation/dephosphorylation equilibrium that will lead to amplification of ITAMs phosphorylation with subsequent propagation of the signal, provided that ZAP70 is activated by Lck (Figure 1).

How Are Coreceptors Involved in TCR Signal Transduction? A Model for Coreceptor Function Based on Lck/ZAP70 Association It was originally thought that the co-receptors CD8 and CD4, which are constitutively associated with Lck, are responsible for chaperoning this PTK to the ITAMs thus provoking their phosphorylation and ZAP70 binding. Two mechanisms have been proposed, one in which a simultaneous binding of TCR and coreceptor to the same peptide/MHC complex takes place (28) and a more recent sequential model in which first the TCR binds to peptide/MHC followed by the co-receptor (22, 29). However, these models do not explain how ITAMs can be phosphorylated in cells lacking CD4/CD8 co-receptors (e.g., most cd TCR- and some ab TCR-bearing T cells, early pre-T cells). Moreover, any model for coreceptor function has to explain how CD4 and CD8 recognize the very rare TCR/ peptide/MHC complexes and ignore all the other free TCRs and MHC molecules.



169

Three-dimensional structural information finds no evidence for a conformational differences between isolated MHC and TCR molecules or their complexes which would favor coreceptor interaction with triggered TCRs nor does the binding geometry of CD8 to class I MHC provide a basis for such a conformational change mechanism (30, 31). An alternative model (32) for coreceptor recruitment into activated TCR complexes is based on the evidence that the SH2 domain of Lck is important for co-receptor function (33) and on the SH2-mediated association of Lck with ZAP70 after TCR triggering (10, 11). In this model, Lck binding to ITAM-bound tyrosine phosphorylated ZAP70 is the chaperone that forces coreceptors to recognize the engaged TCR and the MHC molecule bound to it. The latter interaction will take place even with both strong and weak agonists (34). When the TCR binds to a weak agonist and the co-receptor is absent, phosphorylation of ITAMs and ZAP70 as well as Lck/ZAP70 association will be of very low stoichiometry. The presence of the co-receptor (whose affinity for MHC is in the three-digit lM range (1) will shift a short half-life association of Lck with ZAP70 into a more stable one, with consequent increase of phosphorylation reactions by a reciprocal positive feed-back loop between intracellular and extracellular interactions.

SIGNALING AFTER TYROSINE KINASES IN T CELLS Tyrosine kinases initiate a cascade of signaling pathways during T cell activation; the key events are the activation of Ras- and Rho-family GTPases signaling networks and regulation of inositol phospholipid metabolism. The latter response controls intracellular calcium and the activity of diverse serine/threonine kinases including members of the PKC family and phosphatidyl inositol-3 kinase (PI3K)controlled serine kinases.

The Regulation of Phospholipase C, Calcium and Protein Kinase C TCR coupling to phospholipase C gamma 1 (PLCc1) allows the TCR to regulate hydrolysis of phosphatidylinositol (4,5) biphosphate (PI(4,5)P2) and hence control the production of inositol 1,4,5-triphosphate and diacylglycerol. Initial models for TCR regulation of PLCc1 were simple and proposed that tyrosine phosphorylation of the enzyme resulted in its activation. Now, it is clear that antigen receptor activation of PLCc1 normally requires at least three classes of PTKs; Lck, ZAP70, and Tec kinases such as Rlk and Itk in T cells (Btk in B cells) (35–37). There is also a requirement for two adapter molecules that are both substrates for ZAP70: LAT (Linker for activation of T cells) and SLP76 (SH2-domain containing leukocyte protein of 76 kD) (38–41). LAT is an integral membrane protein with a short extracellular region and a long cytosolic tail with nine tyrosine residues conserved between mouse and human (42). LAT is selectively tyrosine phos-


170

ACUTO n CANTRELL

phorylated in response to triggering of antigen receptors and interacts with the PLCc1 SH-2 domain thereby recruiting this enzyme to the plasma membrane. The importance of SLP76 for PLCc1 activation is unequivocal but quite why is not yet established. One hypothesis is that PLCc1 tyrosine phosphorylation is mediated by a TEC family kinase and that interactions between these kinases and SLP76 are required to coordinate the TEC/PLCc1 interaction. LAT almost certainly acts as the scaffold for assembly of the PLCc1 signaling complex with SLP76 binding to LAT via the newly described Grb2-like adaptors Gads, GrpL, or Grf40 (43–45). The importance of PLCc1 regulation for lymphocyte activation stems from the fact that this pathway controls the production of inositol polyphosphates and diacylglycerols, which regulate increases in intracellular calcium and activate serine/threonine kinases of the protein kinase C family, respectively. Sustained elevation of intracellular calcium concentration ([Ca2`]i) is absolutely critical during the initial phases of T cell activation both for induction of cytokine gene expression and for controlling T cell cytolytic function (46, 47, 48). It must be emphasised that cytokine gene induction requires co-ordination of calcium signals with signals mediated by the Ras/Rho GTPases and by serine/threonine kinases of the PKC family (49). The potency of calcium and PKC signaling pathways for lymphocyte activation is underlined by the ability of pharmacological agents that elevate intracellular calcium levels and activate PKC (calcium ionophores and phorbol esters respectively) to mimic many aspects of antigen receptor triggering. It should also be emphasized that T cells express multiple calcium and diacylgycerol regulated molecules. For example, T cells express multiple functionally distinct PKC isoforms that can be classified into distinct groups: the classical PKCs (a, bI, bII and c), which are regulated by calcium, DAG and phospholipids; novel PKCs (d, e, g and h), which are regulated by DAG and phospholipids; the atypical PKCs (n and k), which lack calcium or diacylglycerol binding domains (50–52). Stunning confocal microscope images reveal the specific recruitment of PKCh to the plasma membrane at the contact zone formed between T cells and APCs (53). This spatial-temporal regulation of serine kinases during antigen receptor engagement has refocused attention on the relevance of PKCs for lymphocyte activation. The functional diversity of PKCs highlights the potential for PKCs to have a pleotropic function in lymphocytes.

p21ras Regulation in T Cells The guanine nucleotide binding protein p21ras rapidly accumulates in its active, GTP-bound form in antigen receptor activated T cells (49, 55). Ras function is essential for cytokine gene induction and has a critical role in thymocyte development (56, 57). The guanine nucleotide binding cycle of Ras is controlled by guanine nucleotide exchange proteins (GEFs), which promote formation of active GTP-bound Ras complexes, and GTPase activating proteins (GAPs), which stimulate the intrinsic GTPase activity of Ras resulting in hydrolysis of bound GTP



171

to GDP. The nucleotide exchange reaction switches Ras on whereas the hydrolysis of GTP turns it off (58). There has been much descriptive biochemistry about the regulation of Ras in lymphocytes (for review see 8, 59). The Ras GEF Sos, the mammalian homologue of the Drosophila ‘Son of Sevenless’ protein associates constitutively with the SH3 domains of an adapter molecule Grb2. In TCR-activated cells the Grb2 SH2 domain interacts with tyrosine-phosphorylated residues in the cytoplasmic tail of the adapter LAT thereby forming protein complexes that regulate the membrane localization and catalytic activity of Sos. One complication to this simple model is that another adapter molecule, SLP76, is also needed for TCR regulation of Ras although exactly why is not clear (41, 60). Once Ras is activated it couples to multiple biochemical effector signaling pathways including the Raf-1/MEK/ ERk1,2 kinases and signaling pathways controlled by the Rac/Rho GTPases (61, 62). In peripheral T cells Ras is important for pathways that activate transcription factors involved in cytokine gene induction. Surprisingly, despite the known importance of Ras, there is little solid information about the important targets for Ras action in the thymus.

Phosphatidylinositol-3 Kinase in T Cells TCR regulation of inositol phospholipids by the action of PLCc1 was one the earliest defined signaling pathway in lymphocytes. However, TCR triggering also stimulates inositol lipid turnover by controlling the activity of phosphatidyl inositol 3-kinase (PI3K) which phosphorylates PI(4,5)P2 on the D-3 position of the inositol ring to produce PI(3,4,5)P3 (63). There are multiple isoforms of PI3K and antigen receptors are thought to stimulate the activity of a PI3K complex that comprises a regulatory p85 and a catalytic p110 subunit. T cells express p110a and p110d and both these catalytic subunits appear to participate in lymphocyte signaling responses (63). Models for PI3K activation invoke p85 binding to adapters, which recruit the enzyme to the plasma membrane; constitutive membrane targeting of p110 catalytic subunits of PI3K creates a constitutively active enzyme that generates PI(3,4,5)P3 and PI(3,4)P2 when expressed in cells (64). The candidate adapters for recruiting PI3K to the membrane in TCR-activated cells are TRIM (65) or LAT (42). Genetic evidence for the importance of PI3K for B lymphocytes has been illustrated by the phenotype of mice lacking expression of the p85a regulatory adapter protein for PI3K that show profound defects in B cell function (66, 67). T lymphocytes lacking p85a appear normal but this is because they still contain functional PI3K. These data suggest that the regulation of PI3K in T cells is not mediated by p85a; whether p85b is important or whether there is a completely different adapter for PI3K regulation in T cells is unknown. The products of PI3K, PI(3,4,5)P3 and PI(3,4P)2, bind to the plextrin homology (PH) domains of proteins and either allosterically modify their activity or induce relocalization of the protein to defined areas of the plasma membrane where activation can occur (68). Recent studies have identified both tyrosine

172

ACUTO n CANTRELL


kinase and serine kinase regulation by PI3K in lymphocytes: the TEC family tyrosine kinases BTK and ITK (35, 69) and a serine/threonine kinase protein kinase B or Akt/PKB (70). These have PH domains that bind D-3 phosphoinositides and recruit these enzymes to the plasma membrane (71–73). TEC kinases function at the membrane to regulate PLCc1 activity. The other PI3K target, Akt/ PKB has several possible functions in T cells: It has been shown to play an important role in maintenance of cell survival in fibroblasts and epithelial cells (for review see 71–73). In T cells, PKB can stimulate the activity of E2F transcription factors, which are important components of the mechanisms that control the mammalian cell cycle (74). PKB also phosphorylates and inactivates GSK3 (71–73), an enzyme initially identified as a regulator of glycogen metabolism but which also has broader functions: GSK3 can also control the nuclear export of Nuclear Factor of Activated T cells (NFAT) transcription factors involved in cytokine gene induction (75).

PI3K is a Link to Rho Family GTPases. A third class of PH domain containing proteins regulated by the products of the PI3K cascade are GEFs for Rho GTPases such as Rac, Rho and Cdc42, (for review see 76). The best characterized Rho family GEF in lymphocytes is Vav1: The binding of PI(3,4,5)P3 to the PH domain of Vav-1 recruits it to the plasma membrane where tyrosine kinase mediated activation of the molecule occurs. Rho GTPases regulate lymphocyte survival, proliferation, and differentiation (55). They interact with multiple effectors and can initiate diverse signals. However, the first biological role ascribed to these GTPases was the dynamic organization of the actin cytoskeleton and the assembly of associated integrin structures. Experiments with constitutively active PI3K mutants have shown that the D-3 phosphoinositide products of PI3K are sufficient to induce Rac- and Rho-mediated cytokeletal responses (64). Cdc42, Rac1, and RhoA are responsible for distinct patterns of actin reorganization: RhoA regulates the formation of actin stress fibers and focal adhesions; Rac1 controls lamellipodia formation and focal complex assembly and subsequently can activate Rho-mediated cytoskeletal changes. Cdc42 first triggers induction of microspikes and filopodia, but consecutively can induce Rac- and Rho-mediated responses. T cells lacking the Rac GEF Vav-1 show defects in actin polymerization suggesting that the link between Rac and the maintenance of the actin cytoskeleton is maintained in lymphocytes (78, 79). An example of the ability of Vav-1 to control the actin cytoskeleton in T cells is shown in Figure 3 (see color insert), which illustrates how a T cell line overexpressing Vav-1 has abundant lamellipodia compared to control cells. There is also genetic and biochemical evidence for Cdc42 regulation of the actin cytoskeleton in lymphocytes. The genetic defect in Wiscott-Aldrich immunodeficiency syndrome (WAS) patients maps to an effector protein for Cdc42, termed WASP (80). The defects in WAS patients include cytoskeletal and cell activation abnormalities in lymphocytes. WASP-deficient


173

cells have a paucity of microvilli on the cell surface. WASP has been implicated in the control of in actin depolymerization (81), and recent data also indicate that Cdc42/WASP family complexes are strong stimulators of the actin nucleating activity of the Arp2/3 complex (82, 83). Moreover, Cdc42 can control T cell polarization toward APCs (84), a process that is critical for efficient T cell-APC contact and the directed release of cytokines.

SPATIO-TEMPORAL RESOLUTION OF GENERAL CYTOSKELETAL CHANGES PRECEDING AND ACCOMPANYING T CELL ACTIVATION


Morphological Modifications of T cell during the Encounter with APC The T cell signaling pathways described above were identified biochemically in experiments with agonistic anti-TCR/CD3 antibodies. Genetic studies have shown them to be important for T cell activation, but it should always be remembered that the physiological trigger for T cell activation is antigen presented to the T cell by an APC. Stimulation by agonistic anti-TCR/CD3 antibodies thus bypasses an important regulatory step in T cell activation; one that relies on the capacity of the T cell to establish an interaction with APCs, a process controlled by changes in cell cytoskeletal architecture. T cell activation is accompanied by a dynamic re-organization of cortical actin, exemplified by the increase in actin polymerization (e.g. increase in filamentous actin, F-actin) (85–87). Analysis at the single-cell level by time-lapse videomicroscopy has revealed the highly dynamic cytoskeletal and morphological modifications taking place in a T cell interacting with an APC (88–92). Trautmann and co-workers initially defined three main stages in the T cell/APC interaction: The first stage is the establishment of T cell/APC physical continuity, the second stage is characterized by T cell rounding, and it is followed by a third stage of T cell spreading (88) resulting in the formation of what has been termed by Dustin and colleagues as an immunological synapse (2). The initial area of contact between the T cell and the APC is formed within a few seconds to 1 minute (89, 91). Once established, the area of contact remains stable while moderate T cell membrane ruffling restricted to the contact zone can be observed (90–92). In the absence of antigen or at low Ag concentration, the T cell maintains its motility and continues to crawl on and around the APC and may leave in search of another partner (88, 89, 91). Initial contact and crawling is likely to depend on loose adhesion forces (1) generated by receptor couples such as CD28/B7, CD2/ CD58(CD48), or LFA1/ICAMs (93). Receptors on the APC (B7.1 and B7.2 may be the best candidates) may convey a ‘‘danger signal’’ (94) so that the T cell will pay particular attention and explore a professional APC (e.g., activated DC) in search of antigen, perhaps by receiving some intracellular signaling already at

174

ACUTO n CANTRELL

this stage (see below). These loose interactions favor fast serial scanning of different APCs by the T cell.


Signals That Control the Formation of the Contact Zone Pre-treatment of T cell with cytochalasin D, a drug that prevents actin polymerization, prevents the formation of T cell/APC contacts (91). Moreover, the integrity of the cytoskeleton needs to be maintained throughout signaling for T cell commitment toward differentiation and proliferation (95). Actin cytoskeletal changes are detectable in T cells making a contact with an APC during the short time before (or in absence of) antigen recognition (90, 91). The changes in the actin cytoskeleton that accompany formation of the first T cell/APC contacts cannot therefore be regulated by the TCR but must be induced by coreceptors, e.g CD28, CD2, or LFA1/ICAMs. Molecules like CD28 are good candidates for promoting the initial modifications because the relative abundance of the CD28 ligands CD80/CD86 on the APC (.104) is much higher than the levels of MHC/ peptide complexes that the TCR would likely encounter (100–200). Moreover, CD28 has a real kinetic advantage over the TCR when considering the affinity of the interaction of these two receptors with their respective ligand(s) on the APC (i.e CD80/CD86 versus peptide/MHC). CD28 association rate constant (Kon) for CD80 is at least two orders of magnitude above the Kon so far determined for most TCR/peptide/MHC interactions (1, 96). Although CD28/CD80 Koff is faster than those measured for TCR/peptide/MHC complexes, perhaps to leave sufficient flexibility at the cell cell interphase during the formation of the contact zone (1), the equilibrium dissociation constant (KD) of the CD28-CD80/CD86 complex is similar, if not lower. Similar data have been obtained with CD86 (S. Davis and A. van der Merwe, personal communication). These kinetic and mass data make it likely that CD28 will engage its ligand faster and at a higher stoichiometry than the TCR (Figure 2). CD28 ligation certainly induces an increase in F-actin, which concentrates at the contact site (87). The molecular basis for CD28 regulation of the actin cytoskeleton is probably explained by the connection between CD28 and PI3K. CD28 interaction with its ligands CD80 or CD86 results in tyrosine phosphorylation of the CD28 cytoplasmic domain and the recuitment of PI3K to the plasma membrane with a corresponding activation of PI3K activity that generates D3phosphoinositides (63). Membrane targeting of PI3K is sufficient to induce Racand Rho-mediated dynamic reorganization of the actin cytoskeleton (64). It is also striking that Vav-1 is rapidly tyrosine phosphorylated by the initial contact between CD28 and its physiological ligands CD80 or CD86 (97, 98). Tyrosine phosphorylation of Vav-1, which activates the catalytic function of this Rac GEF, is TCR autonomous in that it does not require TCR expression (98) but appears to depend on Src PTKs and in particular on Fyn (99). In this context, it should be emphasized that Vav-1 was originally identified as a substrate for antigen receptor activated tyrosine kinases but this induced phosphorylation was transient.



175

In contrast, CD28 induces a rapid and sustained tyrosine phosphorylation of Vav-1 in the absence of simultaneous ligation of the TCR (97). CD28 regulation of PI3K and Vav-1 would inevitably induce Rac/Rho-mediated actin cytoskeleton rearrangements of the type seen during the first stage of the T cell/APC contact. Recent studies have found that CD28 can influence proximal TCR-mediated signaling events such as the phosphorylation of the TCR f chain and ZAP70 (100). CD28/B7 ligation lowers the TCR activation threshold and potentiates T cell tyrosine phosphorylation pathways (101, 102). Ongoing work indicates that the promotion of TCR signaling during the initial stage of the T cell-APC encounter depends not only on CD28-mediated adhesion but also on CD28 signaling through Vav-1 (103), (F. Michel, L. Tuosto, G. Mangino and O. Acuto, unpublished data). These data force revision of previous models of T cell costimulation. It is unequivocal that signals generated by CD28 are essential for immune function, and it was proposed initially that CD28 initiates signaling to the nucleus in pathways parallel to the TCR. Now it seems that antigen-independent signals mediated by costimulatory molecules such as CD28 may promote signaling through the TCR complex by allowing the changes in the actin cytoskeleton that tighten and stabilize the T cell/APC conjugate.

Signals that Form a Stable T Cell/APC Contact Zone The second stage of the formation of the T cell/APC contact is antigen dependent and associated with a sudden widening of the contact zone characterized by a massive membrane protrusion from the T cell (in a pulsatile fashion) toward the APC (88, 89, 91). The signals that drive the second phase of the T cell/APC interaction are not known, but it occurs under conditions of poor or no [Ca2`]i elevation and requires the integrity of the actin cytoskeleton (89, 91). The only clue about how signal integration from the TCR and costimulators might control the actin cytoskeleton comes from reports that a unique protein complex consisting of tyrosine phosphorylated Vav-1 and tyrosine phosphorylated SLP76 assembles in TCR-triggered and CD28-costimulated T cells (98, 99, 104). SLP76 binds simultaneously to Vav-1 and p21-activated kinase (Pak) via the adaptor Nck (105), and the SLP76/Vav-1/Pak complex may facilitate physical proximity between activated Rho family GTPases and Pak. There is an enormous body of data from studies in fibroblasts that Pak co-ordinates the re-organization of the actin cytoskeleton. In particular, studies showing Pak regulation of myosin light chain kinase (106) indicate that the Vav-1/SLP76/Pak complex has the potential to dynamically control the type of changes in the actin cytoskeleton (107) essential for prolonging the T cell/APC contact. Cdc42 and/or Rac regulation of Pak activation in T cells has been described (60).

The Structure of the Stabilized T Cell/APC Contact Zone The final stabilization of the T cell/APC contact requires durable TCR engagement and is thought to be regulated by Ca2` (89), i.e. elevation of intracellular [Ca2`]i seems to initiate a signal that tells the T cell to stop and continue receiving


176

ACUTO n CANTRELL

signals from the APC (88, 108). As the T cell stops, it stabilizes contact with the APC. The stable contact zone between a T cell and the APC is not just a random distribution of receptors-counterreceptors. Rather, receptors and signaling molecules are organized within this area termed SMAC (for supramolecular activation cluster) or mature immunological synapse (6, 109). There is a specific pattern of receptor segregation containing in a concentric fashion two subregions, one internal where TCR, CD4, (together with Src PTKs Lck and Fyn and PKCh) are confined and the external one containing the integrin molecule LFA-1 with talin, which connects through vinculin to actin filaments (6, 109). The external structure may be analogous to focal complexes or adhesions (110). It has also been noted by Alcover and co-workers that the external zone of the SMAC contains ezrin (A. Roumier and A. Alcover personal communication), a cytoskeletal protein member of the ERM (ezrin, radixin, moesin) family, whose function is to connect membrane proteins to actin filaments (111). There will inevitably be intense interest in how the SMAC forms in such a precise and organized way. One clue comes from observations that the T cell/ APC contact zone is apparently enriched in plasma membrane glycosphingolipidenriched microdomains (GEMs, also referred to as lipid rafts) (102, 112, 113, 114). A number of molecules involved in antigen receptor signaling have been found concentrated within the GEMs including Src family PTKs (112, 115) and adaptor molecules such as LAT (116). Importantly, antigen receptors are transiently targeted to these lipid microdomains during an activation response, and it is proposed that this is where interactions with tyrosine kinases and adapter molecules occur (112, 117, 118). It has also been suggested that GEMs are the site of inositol lipid metabolism since both PLCc and PIP2 have been found concentrated within GEMs (114, 119). The signaling mechanisms that cordinate the reoganization of GEMs are not known but it has been described that GEMs interface with the actin cytoskeleton (120, 121), and in this regard it is noteworthy that CD28 signaling has been shown to induce GEM polarization toward the site of CD28 ligation (102). CD28 is uniquely coupled to the regulatory networks that control the activity of Rho family GTPases. The type of changes in the actin cytoskeleton controlled by Rho GTPases could control the polarization of GEMs into the SMAC. The GEMs would then spatially organize the key signaling molecules in the SMAC.

CONCLUDING REMARKS TCR Signal Transduction and the Actin Cytoskeleton? The complicated TCR-induced biochemical responses discussed in the first part of the review can be detected within seconds of T cell contact with APCs yet it takes 10–15 min to form a stable T cell/APC contact zone or SMAC (6, 109). Hence, the known TCR proximal signaling events are not dependent on stable T



177

cell/APC contacts or on the formation of a SMAC or immunological synapse. Rather, the signaling pathways initated by costimulators and antigen receptors organize the SMAC; the rapid changes in cell morphology, adhesion, and motility that form the SMAC are suggestive of an effective adaptive program to ensure that once the T cell identifies the appropriate ligand for its TCR clonotype on an APC a sustained lasting signal is established. Is the initiation of TCR signaling regulated by the actin cytoskelton? Yes in the sense that without dynamic changes in actin structures it is unlikely that the T cells will make sufficient effective contacts with APCs to enable TCR ligation with peptide MHC complexes. It should also be remembered that T cells do not randomly contact APCs but probably respond to chemokine gradients. The coordinated regulation of the actin cytoskeleton will be important for the motility of T lymphocytes and will be critical to allow T cells to respond to the chemokine gradients that will ensure the initial T cell contact with the APC (5). T cell activation and coordinated changes in the actin cytoskeleton are symbiotic processes. The first signals to control the actin cytoskeleton when a T cell contacts an APC are antigen independent and mediated by costimulators such as CD28; there are also likely contributions from other molecules such as CD2 or LFA-1 (122, 123, 124). Once the TCR signaling cascade is established changes in calcium, PI3K, and GTPases triggered by antigen receptors develop the actin dynamics that stabilize the T cell/APC conjugate and organize a signaling zone within the contact zone. This allows sustained T cell signaling to promote proliferation and differentiation.

ACKNOWLEDGMENTS We would like to thank past and present members of the Lymphocyte Activation Laboratory and of the Molecular Immunology Unit for their contributions that helped to forge our knowledge of T cell activation mechanisms. A special thanks to F. Michel and V. Di Bartolo for helping with confocal microscopy and the artwork pictures shown in this review; to A. Trautmann and A. van der Merwe for helpful discussions and suggestions and to Kim Rowan and Wendy Houssin for assistance in the preparation of this manuscript. We would also like to thank A. Alcover, A. Roumier, A. Shaw, M. Bachmann, A. van der Merwe and S. Davis for sharing results before publication. Work from our laboratories was made possible by support to D.C. from the Imperial Cancer Research Fund, the EEC and the Human Frontier Science Programme and to O.A. from the Institut Pasteur, Association pour la Recherche sur le Cancer, Centre National pour la Recherche Scientifique, and Human Frontier Science Programme. Visit the Annual Reviews home page at www.AnnualReviews.org.

178

ACUTO n CANTRELL


LITERATURE CITED 1. Davis SJ, Ikemizu S, Wild MK, van der Merwe PA. 1998. CD2 and the nature of protein interactions mediating cell-cell recognition. Immunol. Rev. 163:217–36 2. Shaw AS, Dustin ML. 1997. Making the T cell receptor go the distance: a topological view of T cell activation. Immunity 6:361–69 3. Cerottini JC, Brunner KT. 1972. Reversible inhibition of lymphocyte-mediated cytotoxicity by cytochalasin B. Nat. New Biol. 237:272–73 4. Kupfer A, Singer SJ. 1989. Cell biology of cytotoxic and helper T cell functions: immunofluorescence microscopic studies of single cells and cell couples. Annu. Rev. Immunol. 7:309–37 5. Serrador JM, Nieto M, Sanchez-Madrid F. 1999. Cytoskeletal rearrangement during migration and activation of T lymphocytes. Trends Cell Biol. 9:228–33 6. Monks CR, Freiberg BA, Kupfer H, Sciaky N, Kupfer A. 1998. Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 395:82–86 7. Chu DH, Morita CT, Weiss A. 1998. The Syk family of protein tyrosine kinases in T-cell activation and development. Immunol. Rev. 165:167–80 8. Clements JL, Boerth NJ, Lee JR, Koretzky GA. 1999. Integration of T cell receptor-dependent signaling pathways by adapter proteins. Annu. Rev. Immunol. 17:89–108 9. van Leeuwen JE, Samelson LE. 1999. T cell antigen-receptor signal transduction Curr. Opin. Immunol. 11:242–48 10. Duplay P, Thome M, Herve F, Acuto O. 1994. p56lck interacts via its src homology 2 domain with the ZAP-70 kinase. J. Exp. Med. 179:1163–72 11. Straus DB, Chan AC, Patai B, Weiss A. 1996. SH2 domain function is essential for the role of the Lck tyrosine kinase in

12.

13.

14.

15.

16.

17.

18.

19.

T cell receptor signal transduction. J. Biol. Chem. 271:9976–81 Iwashima M, Irving BA, van Oers NS, Chan AC, Weiss A. 1994. Sequential interactions of the TCR with two distinct cytoplasmic tyrosine kinases. Science 263:1136–39 Chan AC, Dalton M, Johnson R, Kong GH, Wang T, Thoma R, Kurosaki T. 1995. Activation of ZAP-70 kinase activity by phosphorylation of tyrosine 493 is required for lymphocyte antigen receptor function. EMBO J. 14:2499–2508 Qian D, Mollenauer MN, Weiss A. 1996. Dominant-negative zeta-associated protein 70 inhibits T cell antigen receptor signaling. J. Exp. Med. 183:611–20 Mege D, Di Bartolo V, Germain V, Tuosto L, Michel F, Acuto O. 1996. Mutation of tyrosines 492/493 in the kinase domain of ZAP-70 affects multiple T-cell receptor signaling pathways. J. Biol. Chem. 271:32644–52 Wong J, Straus D, Chan AC. 1998. Genetic evidence of a role for Lck in Tcell receptor function independent or downstream of ZAP-70/Syk protein tyrosine kinases. Mol. Cell. Biol. 18:2855– 66 Pelosi M, Di Bartolo V, Mounier V, Mege D, Pascussi JM, Dufour E, Blondel A, Acuto O. 1999. Tyrosine 319 in the interdomain B of ZAP-70 is a binding site for the Src homology 2 domain of Lck. J. Biol. Chem. 274:14229–37 Di Bartolo V, Mege D, Germain V, Pelosi M, Dufour E, Michel F, Magistrelli G, Isacchi A, Acuto O. 1999. Tyrosine 319, a newly identified phosphorylation site of ZAP-70, plays a critical role in T cell antigen receptor signaling. J. Biol. Chem. 274:6285–94 Heyeck SD, Wilcox HM, Bunnell SC, Berg LJ. 1997. Lck phosphorylates the activation loop tyrosine of the Itk kinase


20.

21.


22.

23.

24.

25.

26.

27.

28.

domain and activates Itk kinase activity. J. Biol. Chem. 272:25401–8 Sloan-Lancaster J, Shaw AS, Rothbard JB, Allen PM. 1994. Partial T cell signaling: altered phospho-zeta and lack of zap70 recruitment in APL-induced T cell anergy. Cell 79:913–22 Madrenas J, Wange RL, Wang JL, Isakov N, Samelson LE, Germain RN. 1995. Zeta phosphorylation without ZAP-70 activation induced by TCR antagonists or partial agonists. Science 267:515–18 Madrenas J, Chau LA, Smith J, Bluestone JA, Germain RN. 1997. The efficiency of CD4 recruitment to ligandengaged TCR controls the agonist/partial agonist properties of peptide-MHC molecule ligands. J. Exp. Med. 185:219–29 Scharenberg AM, Lin S, Cuenod B, Yamamura H, Kinet JP. 1995. Reconstitution of interactions between tyrosine kinases and the high affinity IgE receptor which are controlled by receptor clustering. EMBO J. 14:3385–94 Davis SJ, van der Merwe PA. 1996. The structure and ligand interactions of CD2: implications for T-cell function. Immunol. Today 17:177–87 Stone JD, Conroy LA, Byth KF, Hederer RA, Howlett S, Takemoto Y, Holmes N, Alexander DR. 1997. Aberrant TCRmediated signaling in CD45-null thymocytes involves dysfunctional regulation of Lck, Fyn, TCR-zeta, and ZAP-70. J. Immunol. 158:5773–82 Veillette A, Horak ID, Horak EM, Bookman MA, Bolen JB. 1988. Alterations of the lymphocyte-specific protein tyrosine kinase (p56lck) during T-cell activation. Mol. Cell. Biol. 8:4353–61 Weil R, Veillette A. 1994. Intramolecular and extramolecular mechanisms repress the catalytic function of p56lck in resting T-lymphocytes. J. Biol. Chem. 269: 22830–38 Janeway CA Jr. 1992. The T cell receptor as a multicomponent signalling machine: CD4/CD8 coreceptors and CD45 in T

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

179

cell activation. Annu. Rev. Immunol. 10:645–74 Hampl J, Chien YH, Davis MM. 1997. CD4 augments the response of a T cell to agonist but not to antagonist ligands. Immunity 7:379–85 Garcia KC, Teyton L, Wilson IA. 1999. Structural basis of T cell recognition. Annu. Rev. Immunol. 17:369–97 Zamoyska R. 1994. The CD8 coreceptor revisited: one chain good, two chains better. Immunity 1:243–46 Thome M, Duplay P, Guttinger M, Acuto O. 1995. Syk and ZAP-70 mediate recruitment of p56lck/CD4 to the activated T cell receptor/CD3/zeta complex. J. Exp. Med. 181:1997–2006 Xu H, Littman DR. 1993. A kinaseindependent function of Lck in potentiating antigen-specific T cell activation. Cell 74:633–43 Viola A, Salio M, Tuosto L, Linkert S, Acuto O, Lanzavecchia A. 1997. Quantitative contribution of CD4 and CD8 to T cell antigen receptor serial triggering. J. Exp. Med. 186:1775–79 Scharenberg AM, Kinet JP. 1998. PtdIns3, 4, 5-P3: a regulatory nexus between tyrosine kinases and sustained calcium signals. Cell 94:5 Debnath J, Chamorro M, Czar MJ, Schaeffer EM, Lenardo MJ, Varmus HE, Schwartzberg PL. 1999. rlk/ TXK encodes two forms of a novel cysteine string tyrosine kinase activated by Src family kinases. Mol. Cell. Biol. 19:1498– 1507 Schaeffer EM, Debnath J, Yap G, McVicar D, Liao XC, Littman DR, Sher A, Varmus HE, Lenardo MJ, Schwartzberg PL. 1999. Requirement for Tec kinases Rlk and Itk in T cell receptor signaling and immunity. Science 284:638–41 Finco TS, Kadlecek T, Zhang WG, Samelson LE, Weiss A. 1998. LAT is required for TCR-mediated activation of PLC gamma 1 and the Ras pathway. Immunity 9:617–26


180

ACUTO n CANTRELL

39. Zhang WG, Sommers CL, Burshtyn DN, Stebbins CC, DeJarnette JB, Trible RP, Grinberg A, Tsay HC, Jacobs HM, Kessler CM, Long EO, Love PE, Samelson LE. 1999. Essential role of LAT in T cell development. Immunity 10:323–32 40. Clements JL, Yang B, Rossbarta SE, Eliason SL, Hrstka RF, Williamson RA, Koretzky GA. 1998. Requirement for the leukocyte-specific adapter protein slp-76 for normal T-cell development. Science 281:416–19 41. Yablonski D, Kuhne MR, Kadlecek T, Weiss A. 1998. Uncoupling of nonreceptor tyrosine kinases from plcgamma-1 in an slp-76-deficient T-cell. Science 281:413–16 42. Zhang W, Sloan-Lancaster J, Kitchen J, Trible RP, Samelson LE. 1998. LAT: the ZAP-70 tyrosine kinase substrate that links T cell receptor to cellular activation. Cell 92:83–92 43. Asada H, Ishii N, Sasaki Y, Endo K, Kasai H, Tanaka N, Takeshita T, Tsuchiya S, Konno T, Sugamura K. 1999. Grf40, a novel Grb2 family member, is involved in T cell signaling through interaction with SLP-76 and LAT. J. Exp. Med. 189:1383–90 44. Liu SK, Fang N, Koretzky GA, McGlade CJ. 1999. The hematopoietic-specific adaptor protein gads functions in T-cell signaling via interactions with the SLP76 and LAT adaptors. Curr. Biol. 9:67– 75 45. Law CL, Ewings MK, Chaudhary PM, Solow SA, Yun TJ, Marshall AJ, Hood L, Clark EA. 1999. GrpL, a Grb2-related adaptor protein, interacts with SLP-76 to regulate nuclear factor of activated T cell activation. J. Exp. Med. 189:1243–53 46. Crabtree GR. 1999. Generic signals and specific outcomes: signaling through Ca2`, calcineurin, and NF-AT. Cell 96: 611–14 47. Golstein P, Smith ET. 1977. Mechanism of T-cell-mediated cytolysis: the lethal

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58. 59.

hit stage. Contemp. Top. Immunobiol. 7:273–300 Rao A, Luo C, Hogan PG. 1997. Transcription factors of the NFAT family: regulation and function. Annu. Rev. Immunol. 15:707–47 Izquierdo Pastor M, Reif K, Cantrell D. 1995. The regulation and function of p21ras during T-cell activation and growth. Immunol. Today 16:159–64 Mellor H, Parker PJ. 1998. The extended protein kinase C superfamily. Biochem. J. 332:281–92 Newton AC, Johnson JE. 1998. Protein kinase C: a paradigm for regulation of protein function by two membranetargeting modules. Biochim. Biophys. Acta 1376:155–72 Oancea E, Meyer T. 1998. Protein kinase C as a molecular machine for decoding calcium and diacylglycerol signals. Cell 95:307–18 Monks CPF, Kupfer H, Tamir I, Barlow A, Kupfer A. 1997. Selective modulation of protein-kinase C-theta during T-cell activation. Nature 385:83–86 Rozengurt E, Sinnett-Smith J, Zugaza JL. 1997. Protein kinase D: a novel target for diacylglycerol and phorbol esters. Biochem. Soc. Trans. 25:565–71 Henning SW, Cantrell DA. 1998. GTPases in antigen receptor signalling. Curr. Opin. Immunol. 10:322–29 Alberola-Ila J, Hogquist KA, Swan KA, Bevan MJ, Perlmutter RG. 1996. Positive and negative selection invoke distinct signalling pathways. J. Exp. Med. 184:9–18 Gartner F, Alt FW, Monroe R, Chu M, Sleckman BP, Davidson L, Swat W. 1999. Immature thymocytes employ distinct signaling pathways for allelic exclusion versus differentiation and expansion. Immunity 10:537–46 Downward J. 1996. Control of ras activation. Cancer Surv. 27:87–100 Cantrell D. 1996. T cell antigen receptor


60.

61.

62.


63.

64.

65.

66.

67.

68.

signal transduction pathways. Annu. Rev. Immunol. 14:259–74 Yablonski D, Kane LP, Qian D, Weiss A. 1998. A Nck-Pak1 signaling module is required for T-cell receptor-mediated activation of NFAT, but not of JNK. EMBO J. 17:5647–57 Genot E, Cleverley S, Henning S, Cantrell DA. 1996. Multiple p21ras effector pathways regulate nuclear factor of activated T cells. EMBO J. 15:3923–33 Genot E, Reif K, Beach S, Kramer I, Cantrell D. 1998. p21ras initiates Rac-1 but not phosphatidyl inositol 3 kinase/ PKB, mediated signaling pathways in T lymphocytes. Oncogene 17:1731–38 Ward SG, June CH, Olive D. 1996. PI 3kinase: a pivotal pathway in T cell activation. Immunol. Today 17:187–98 Reif K, Nobes CD, Thomas G, Hall A, Cantrell D. 1996. Phosphatidylinositol 3kinase signals activate a selective subset of Rac/Rho-dependent effector pathways. Curr. Biol. 6:1445–55 Bruyns E, Marie-Cardine A, Kirchgessner H, Sagolla K, Shevchenko A, Mann M, Autschbach F, Bensussan A, Meuer S, Schraven B. 1998. T cell receptor (TCR) interacting molecule (TRIM), a novel disulfide- linked dimer associated with the TCR-CD3-zeta complex, recruits intracellular signaling proteins to the plasma membrane. J. Exp. Med. 188:561–75 Fruman DA, Snapper SB, Yballe CM, Davidson L, Yu JY, Alt FW, Cantley LC. 1999. Impaired B cell development and proliferation in absence of phosphoinositide 3-kinase p85alpha. Science 283: 393–97 Suzuki H, Terauchi Y, Fujiwara M, Aizawa S, Yazaki Y, Kadowaki T, Koyasu S. 1999. Xid-like immunodeficiency in mice with disruption of the p85alpha subunit of phosphoinositide 3-kinase. Science 283:390–92 Rameh LE, Arvidsson A-K, Carraway KL, Couvillon AD, Rathbun G, Cromp-

69.

70.

71.

72.

73.

74.

75.

76.

77.

181

ton A, VanRenterghem B, Czech M-P, Ravichandran KS, Burakoff SJ, Wang D-S, Chen C-S, Cantley LC. 1997. A comparative analysis of the phosphoinositide binding specificity of pleckstrin homology domains. J. Biol. Chem. 272:22059–66 Bolland S, Pearse RN, Kurosaki T, Ravetch JV. 1998. SHIP modulates immune receptor responses by regulating membrane association of Btk. Immunity 8:509–16 Reif K, Burgering BMT, Cantrell DA. 1997. Phosphatidylinositol 3-kinase links the interleukin-2 receptor to protein kinase B and p70 S6 kinase. J. Biol. Chem. 272:14426–38 Alessi DR, Cohen P. 1998. Mechanism of activation and function of protein kinase B. Curr. Opin. Genet. Dev. 8:55– 62 Coffer PJ, Jin J, Woodgett JR. 1998. Protein kinase B (c-Akt): a multifunctional mediator of phosphatidylinositol 3kinase activation. Biochem. J. 335:1–13 Bellacosa A, Chan TO, Ahmed NN, Datta K, Malstrom S, Stokoe D, McCormick F, Feng J, Tsichlis P. 1998. Akt activation by growth factors is a multiple-step process: the role of the PH domain. Oncogene 17:313–25 Brennan P, Babbage JW, Burgering BM, Groner B, Reif K, Cantrell DA. 1997. Phosphatidylinositol 3-kinase couples the interleukin-2 receptor to the cell cycle regulator E2F. Immunity 7:679–89 Beals CR, Sheridan CM, Turck CW, Gardner P, Crabtree GR. 1997. Nuclear export of NF-ATc enhanced by glycogen synthase kinase-3. Science 275:1930–34 Cerione R, Zheng Y. 1996. The Dbl family of oncogenes. Curr. Opin. Cell Biol. 8:216–22 Nobes C, Hall A. 1995. Rho, Rac and Cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibres, lamellipodia and filopodia. Cell 81:53–62


182

ACUTO n CANTRELL

78. Holsinger LJ, Graef IA, Swat W, Chi T, Bautista DM, Davidson L, Lewis RS, Alt FW, Crabtree GR. 1998. Defects in actincap formation in Vav-deficient mice implicate an actin requirement for lymphocyte signal transduction. Curr. Biol. 8:563–72 79. Fischer K-D, Kong Y-Y, Nishina H, Tedford K, Marengere LEM, Kozieradzki I, Sasaki T, Starr M, Chan G, Gardener S, Nghiem MP, Bouchard D, Barbacid M, Bernstein A, Penninger JM. 1998. Vav is a regulator of cytoskeletal reorganization mediated by the T-cell receptor. Curr. Biol. 8:554–62 80. Kirchhausen T, Rosen FS. 1996. Disease mechanism: unravelling Wiskott-Aldrich syndrome. Curr. Biol. 6:676–78 81. Snapper SB, Rosen FS. 1999. The Wiskott-Aldrich Syndrome Protein (WASP): roles in signaling and cytoskeletal organization. Annu. Rev. Immunol. 17:905–29 82. Rohatgi R, Ma L, Miki H, Lopez M, Kirchhausen T, Takenawa T, Kirschner MW. 1999. The interaction between NWASP and the Arp2/3 complex links Cdc42- dependent signals to actin assembly. Cell 97:221–31 83. Machesky LM, Gould KL. 1999. The Arp2/3 complex: a multifunctional actin organizer. Curr. Opin. Cell Biol. 11:117– 21 84. Stowers L, Yelon D, Berg LJ, Chant J. 1995. Regulation of the polarization of T cells toward antigen-presenting cells by Ras-related GTPase CDC42. Proc. Natl. Acad. Sci. USA 92:5027–31 85. Parsey MV, Lewis GK. 1993. Actin polymerization and pseudopod reorganization accompany anti-CD3- induced growth arrest in Jurkat T cells. J. Immunol. 151:1881–93 86. Lowin-Kropf B, Shapiro VS, Weiss A. 1998. Cytoskeletal polarization of T cells is regulated by an immunoreceptor tyrosine-based activation motif-depen-

87.

88.

89.

90.

91.

92.

93.

94. 95.

dent mechanism. J. Cell Biol. 140:861– 71 Kaga S, Ragg S, Rogers KA, Ochi A. 1998. Stimulation of CD28 with B7–2 promotes focal adhesion-like cell contacts where Rho family small G proteins accumulate in T cells. J. Immunol. 160:24–27 Donnadieu E, Bismuth G, Trautmann A. 1994. Antigen recognition by helper T cells elicits a sequence of distinct changes of their shape and intracellular calcium. Curr. Biol. 4:584–95 Negulescu PA, Krasieva TB, Khan A, Kerschbaum HH, Cahalan MD. 1996. Polarity of T cell shape, motility, and sensitivity to antigen. Immunity 4:421– 30 Delon J, Bercovici N, Raposo G, Liblau R, Trautmann A. 1998. Antigendependent and -independent Ca2` responses triggered in T cells by dendritic cells compared with B cells. J. Exp. Med. 188:1473–84 Delon J, Bercovici N, Liblau R, Trautmann A. 1998. Imaging antigen recognition by naive CD4` T cells: compulsory cytoskeletal alterations for the triggering of an intracellular calcium response. Eur. J. Immunol. 28:716–29 Montes M, McIlroy D, Hosmalin A, Trautmann A. 1999. Calcium responses elicited in human T cells and dendritic cells by cell- cell interaction and soluble ligands. Int. Immunol. 11:561–68 Hauss P, Selz F, Cavazzana-Calvo M, Fischer A. 1995. Characteristics of antigen-independent and antigen-dependent interaction of dendritic cells with CD4` T cells. Eur. J. Immunol. 25:2285–94 Matzinger P. 1998. An innate sense of danger. Semin. Immunol. 10: 399–415 Valitutti S, Dessing M, Aktories K, Gallati H, Lanzavecchia A. 1995. Sustained signaling leading to T cell activation results from prolonged T cell receptor occupancy. Role of T cell actin cytoskeleton. J. Exp. Med. 181:577–84


CYTOSKELETON IN T CELL SIGNALING 96. Davis MM, Boniface JJ, Reich Z, Lyons D, Hampl J, Arden B, Chien Y-h. 1998. Ligand recognition by ab T cell receptors. Annu. Rev. Immunol. 16:523–44 97. Nunes JA, Collette Y, Truneh A, Olive D, Cantrell DA. 1994. The role of p21ras in CD28 signal transduction: Triggering of CD28 with antibodies, but not the ligand B7–1 activates p21ras. J. Exp. Med. 180:1067–76 98. Tuosto L, Michel F, Acuto O. 1996. p95vav associates with tyrosinephosphorylated SLP-76 in antigenstimulated T cells. J. Exp. Med. 184:1161–66 99. Michel F, Grimaud L, Tuosto L, Acuto O. 1998. Fyn and ZAP-70 are required for Vav phosphorylation in T cells stimulated by antigen-presenting cells. J. Biol. Chem. 273:31932–38 100. Tuosto L, Acuto O. 1998. CD28 affects the earliest signaling events generated by TCR engagement. Eur. J. Immunol. 28:2131–42 101. Viola A, Lanzavecchia A. 1996. T cell activation determined by T cell receptor number and tunable thresholds. Science 273:104–6 102. Viola A, Schroeder S, Sakakibara Y, Lanzavecchia A. 1999. T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science 283: 680–82 103. Bachmann MF, McKall-Faienza K, Schmits R, Bouchard D, Beach J, Speiser DE, Mak TW, Ohashi PS. 1997. Distinct roles for LFA-1 and CD28 during activation of naive T cells: adhesion versus costimulation. Immunity 7:549–57 104. Wu J, Motto DG, Koretzky GA, Weiss A. 1996. Vav and SLP-76 interact and functionally cooperate in IL-2 gene activation. Immunity 4:593–602 105. Wardenburg JB, Pappu R, Bu JY, Mayer B, Chernoff J, Straus D, Chan AC. 1998. Regulation of PAK activation and the T cell cytoskeleton by the linker protein SLP-76. Immunity 9:607–16

183

106. Sanders LC, Matsumura F, Bokoch GM, de Lanerolle P. 1999. Inhibition of myosin light chain kinase by p21-activated kinase. Science 283:2083–85 107. Rottner K, Hall A, Small JV. 1999. Interplay between Rac and Rho in the control of substrate contact dynamics. Curr. Biol. 9:640–48 108. Dustin ML, Bromley SK, Kan Z, Peterson DA, Unanue ER. 1997. Antigen receptor engagement delivers a stop signal to migrating T lymphocytes. Proc. Natl. Acad. Sci. USA 94:3909–13 109. Grakoui A, Bromley SK, Sumen C, Davis MM, SA, Allen PM, Dustin ML. 1999. The immunological synapse: a molecular mechanism machine controlling T cell activation. Science 285: In press 110. Hotchin NA, Hall A. 1995. The assembly of integrin adhesion complexes requires both extracellular matrix and intracellular rho/rac GTPases. J. Cell. Biol. 131:1857–65 111. Bretscher A. 1999. Regulation of cortical structure by the ezrin-radixin-moesin protein family. Curr. Opin. Cell Biol. 11:109–16 112. Montixi C, Langlet C, Bernard AM, Thimonier J, Dubois C, Wurbel MA, Chauvin JP, Pierres M, He HT. 1998. Engagement of T cell receptor triggers its recruitment to low-density detergentinsoluble membrane domains. EMBO J. 17:5334–48 113. Xavier R, Brennan T, Li QQ, McCormack C, Seed B. 1998. Membrane compartmentation is required for efficient T cell activation. Immunity 8:723–32 114. Simons K, Ikonen E. 1997. Functional rafts in cell membranes. Nature 387:569– 72 115. Kabouridis PS, Magee AL, Ley SC. 1997. S-acylation of LCK protein tyrosine kinase is essential for its signalling function in T lymphocytes. EMBO J. 16:4983–98 116. Zhang W, Trible RP, Samelson LE. 1998.

184

117.

118.

119.


120.

121.

ACUTO n CANTRELL LAT palmitoylation: its essential role in membrane microdomain targeting and tyrosine phosphorylation during T cell activation. Immunity 9:239–46 Stauffer TP, Meyer T. 1997. Compartmentalized IgE receptor-mediated signal transduction in living cells. J. Cell Biol. 139:1447–54 Xavier R, Seed B. 1999. Membrane compartmentation and the response to antigen. Curr. Opin. Immunol. 11:265–69 Jacobson K, Dietrich C. 1999. Looking at lipid rafts? Trends Cell Biol. 9:87–91 Moran M, Miceli MC. 1998. Engagement of GPI-linked CD48 contributes to TCR signals and cytoskeletal reorganization: a role for lipid rafts in T cell activation. Immunity 9:787–96 Harder T, Simons K. 1999. Clusters of glycolipid and glycosylphosphatidyli-

nositol-anchored proteins in lymphoid cells: accumulation of actin regulated by local tyrosine phosphorylation. Eur. J. Immunol. 29:556–62 122. Powell JD, Ragheb JA, KitagawaSakakida S, Schwartz RH. 1998. Molecular regulation of interleukin-2 expression by CD28 co-stimulation and anergy. Immunol. Rev. 165:287–300 123. Teh SJ, Killeen N, Tarakhovsky A, Littman DR, Teh HS. 1997. CD2 regulates the positive selection and function of antigen-specific CD4- CD8` T cells. Blood 89:1308–18 124. Dubey C, Croft M, Swain SL. 1995. Costimulatory requirements of naive CD4` T cells. ICAM-1 or B7–1 can costimulate naive CD4 T cell activation but both are required for optimum response. J. Immunol. 155:45–57


ACUTO

■

CANTRELL C-1

Figure 3 Overexpression of Vav-1 in Jurkat cells induces formation of lamellipodia. Jurkat cells were transfected with a vector expressing myc-tagged Vav-1. © (C) Differential interference contrast (DIC) microscopy image shows the formation of large lamellipodia only in the cell at the far right. (A) Anti-myc staining to reveal Vav-1 (green). (B) F-actin staining by phalloidin (red). It is clear that only the cell overexpressing Vav-1 has altered morphology. The white arrow indicates a cell not expressing transfected Vav-1. Note the strong increase in F-actin in the Vav-1 overexpressing cell and the coincident staining of Vav-1 and F-actin.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


THE SPECIFIC REGULATION OF IMMUNE RESPONSES BY CD8~ T CELLS RESTRICTED BY THE MHC CLASS IB MOLECULE, QA-1 Hong Jiang and Leonard Chess Department of Medicine and Pathology, Columbia University College of Physicians and Surgeons, New York, NY 10032


Key Words TCR, Qa-1, CD8, CD4, immunoregulation Abstract Over the last three decades considerable evidence has accumulated that CD8` T cells regulate peripheral immune responses, in part, by specifically controlling the outgrowth of antigen-triggered CD4` T cells. This regulatory function of CD8` T cells has been shown, in vivo, to control the emergence of autoreactive CD4` T cells as well as CD4` T cells reactive to conventional antigens, including alloantigens. In this review, we summarize the evidence that this immune suppression mediated by CD8` T cells is dependent, in part, on specific cognate interactions between MHC class I–restricted regulatory CD8` cells and antigen-activated CD4` T cells. Moreover, we review the evidence that regulatory CD8` T cells recognize antigen-activated CD4` T cells in a TCR specific manner restricted by the MHC class Ib molecule, Qa-1. The Qa-1 molecule may be uniquely qualified to serve this MHC restrictive function because, unlike conventional MHC molecules, it is preferentially and transiently expressed on activated and not resting CD4` T cells. This may assure that only recently antigen-activated CD4` T cells expressing Qa-1/TCR peptide complexes will induce regulatory CD8` T cells and subsequently become susceptible to regulation. Because Qa-1 also binds to self Qdm peptides that trigger NK (CD94/ NKG2) receptors on CD8` T cells, the machinery for homeostatic regulation of regulatory CD8` T cells can be envisioned. Finally, we propose a model by which these TCR specific, Qa-1-restricted regulatory CD8` T cells selectively downregulate antigen-activated T cells expressing TCRs of certain affinities. Ultimately these regulatory CD8` T cells control the peripheral TCR repertoire during the course of immune responses to both self and foreign antigens.

INTRODUCTION Immune responses are initiated when resting precursor CD4` T cells are triggered by MHC/peptide complexes in concert with costimulatory molecules on the surface of antigen presenting cells (APCs) (1–3). As a consequence of this triggering the CD4` T cells proliferate, begin to secrete cytokines (IL-2, IFN-c, IL-4, etc.) 0732–0582/00/0410–0185$14.00

185


186

JIANG n CHESS

and express important cell surface molecules including the IL-2 receptor (CD25), CTLA-4, and CD40 ligand. In order to regulate the immune response and to dampen the potential for autoimmunity, the immune system has also evolved several mechanisms to downregulate and control the outgrowth, differentiation, and function of peripheral antigen-activated CD4` T cells. One level of control resides at the initial clonal activation of the T cell receptor itself by MHC/peptide complexes. T cell signaling via this tri-molecular interaction is thought to be dependent on the affinity and duration of binding of the TCRs with MHC/peptide complexes (4, 5) and will subsequently influence many events during the evolution of both primary and secondary immune responses. In this regard, TCR triggering induces not only activation and differentiation events but, in addition, induces apoptotic pathways. These apoptotic signals are regulated, in part, by cell surface molecules, like FasL, expressed during antigen triggering (6). During the initial T cell/ MHC/peptide interaction, other receptor ligand interactions become pivotal in ultimately dictating the functional fate of T cells. For example, one of the earliest antigen activation–induced cell surface molecules expressed by T cells is CD40L (7). A critical consequence of the interaction of CD40L with CD40 expressed on APCs is the upregulation of other key costimulatory molecules, including CD80 and CD86 (8, 9). These molecules interact with T cell CD28 or CTLA-4 molecules to determine whether the outcome of antigen triggering will be either functional T cell activation or, alternatively, anergy and tolerance induction (3, 9–11). Thus, antigen triggering in the absence of CD80 or CD86 triggering is known to induce anergic T cells (3, 12, 13). Similarly, blockade of the CD4L/CD40 pathway can lead to tolerance induction (14). Another consequence of MHC/peptide triggering of CD4` T cells is the further differentiation of the CD4` T cells into the functional distinct TH1 and TH2 subsets (15–17). The elaboration of distinct cytokines by the TH subsets provides a second level of control intrinsic to the outgrowth and function of CD4` T cells (18). For example, IFN-c secreted by TH1 cells is known to downregulate the differentiation and function of TH2 cells, and conversely, IL-4 and IL-10 inhibit TH1 cell differentiation (17–19). In addition to these regulatory pathways, which are intrinsic to CD4 T cells, recent data strongly suggests that CD8` T cells also interact with CD4` T cells to regulate immune responses in a profound manner. For example, CD8` T cells have been shown to control the physiological outgrowth and function of autoreactive CD4` T cells in vivo (20–27). These regulatory interactions between CD4` and CD8` T cells are complex and involve both antigen specific and nonspecific mechanisms. In principle, one can envision three distinct but not mutually exclusive models by which CD8` T cells specifically regulate antigendriven CD4` T cells. First, CD8` T cells could be triggered by antigens to secrete lymphokines, which, in turn, control the CD4` T cells activated by the same antigens. Indeed, cytokines including IL-4 and IFN-c secreted by CD8` T cells have been shown in some systems to regulate CD4` T cell function (28). Because of the potential proximity of CD4` and CD8` T cells at the site of initial antigen


CD8` T CELL MEDIATED IMMUNOREGULATION

187

activation and because antigen-activated T cells may preferentially express the receptors for these cytokines, the effect of even these antigen nonspecific cytokines will be relatively antigen specific. Second, antigen triggering induces CD4` T cells to transiently express nonpolymorphic membrane activation molecules unrelated either to antigen or TCR that are recognized by regulatory CD8` T cells. The recognition of these activation molecules (sometimes referred to as ergotypic structures) may induce CD8` T cell differentiation into effector cells that have the potential to delete or inactivate antigen-activated CD4` T cells (29, 30). Third, T cell receptor related structures, including TCR-derived MHC/ peptide complexes expressed on antigen-activated CD4` cells may induce regulatory CD8` T cells (24, 25, 31–33). These CD8` T cells could then differentiate and recognize the TCR-derived peptide/MHC class I complexes expressed on the activated CD4` inducer cells. The effector phase of regulation mediated by these putative TCR peptide-recognizing CD8` T cells may involve either conventional cell-mediated cytotoxic (CTL) functions and/or the release of cytokines. This latter view of immune regulation involving recognition of TCR structures was initially suggested, in principle, by Jerne in his idiotype-driven network hypothesis (34). In addition, many of the suppressor cell interaction models proposed during the latter part of the 1970s and early 1980s involved complex interactions between T cell subsets (which now include the CD4` and CD8` subsets) as well as the recognition of TCR related structures by CD8` T cells (28, 35– 37). In this chapter we review recent evidence that CD8` T cells are generated during immune responses, in vivo and in vitro, which downregulate CD4` T cells in a TCR Vb–specific manner. We appreciate that this work resurrects two immunological hypotheses (i.e. TCR specific network regulation and control of immunity by suppressor T cells) that were more prevalent and/or popular two to three decades ago, at a time when the scientific revolution that has occurred in the molecular characterization of the immune system was in its infancy. We briefly review the history and experimental basis of these ideas. Moreover, we focus on and review some recent experiments that permit us to synthesize some of the old and new experiments and construct a general model of immune regulation by CD8` T cells.

HISTORICAL CONSIDERATIONS: THE INITIAL RISE AND FALL OF THE CD8` T CELL SUPPRESSOR PARADIGM The idea that immune responses must be tightly regulated was implicit in the early ideas of clonal selection as proposed by Ehrlich in 1900. He envisioned that during the ontogeny and outgrowth of the immunocompetent clones responsive to foreign antigens there had to be mechanisms to control the outgrowth of clones


188

JIANG n CHESS

reactive with self (38). In Ehrlich’s conception the failure to control the outgrowth of autoreactive cells would lead to a state of ‘‘horror autoxicus’’ or autoimmunity. Ehrlich’s ideas were amplified and refined in the 1950s by MacFarlane Burnet and Neils Jerne, with the elaboration of the Clonal Selection Hypothesis (34, 39, 40). This hypothesis states that (a) multiple clones of immunocompetent cells displaying unique antigen specific receptors exist prior to the introduction of foreign antigens; (b) following exposure to a particular foreign antigen, the antigen interacts with its unique receptor on the cell surface of an immunocompetent cell, and this specific receptor-antigen interaction triggers cells to divide and express their immunological programs; and (c) the majority of cells bearing receptors for self-antigens must be eliminated during differentiation or carefully regulated. These tenets of clonal selection have proven, over the years, to be essentially correct. A corollary of these notions is that autoimmune diseases arise from either the failure to eliminate or inactivate immunocompetent cells during their ontogeny and/or the failure of the immune system to control the outgrowth or function of self-reactive clones in the periphery. Thus, immunocompetent selfreactive cells circulating in the periphery either had to be incapable of responding to self-antigens (tolerant) or alternatively had to be specifically suppressed. The idea of suppressor T cells arose in the late 1960s and early 1970s at a time when it was already clear that immune responses were effected by lymphoid cells (41) and that an interesting subdivision of labor existed among populations of lymphocytes resident in distinct lymphoid organs (41–43). For example, it was discovered that thymic-derived T cells effected cell-mediated immune responses, whereas bone marrow–derived B cells produced antibodies (44–46). Moreover, B cell antibody production was found to be dependent on help from thymicderived T cells (47–50). In this setting, classic experiments at Yale University, in the Gershon laboratory, showed that adoptive transfer of T cells from animals made tolerant to antigen X could specifically suppress anti-X antibody production in recipient animals (51, 52). These experiments were confirmed in a variety of settings during the 1970s (53–55). It was also found that this suppressor activity of T cells could also downregulate delayed type hypersensitivity responses (DTH) and tumor immunity (56, 57). An important conceptual development in immunology in the 1970s that gave further credence to the notion of suppressor T cells was the idea that distinct T cell functions were mediated by phenotypically stable subsets of T cells expressing distinct cell surface molecules. Thus, the seminal experiments of Cantor and Boyse showed that the genetically well-defined Lyt allo-antisera could be used to define functionally distinct subsets of cells expressing either helper (inducer) function (Lyt1) or alternately cytotoxic functions (Lyt2) (58, 59). Moreover, the suppressor functions were found to be largely contained in the Lyt2 subset (60, 61). These findings were later extended to other species including man (62, 63), and with the advent of monoclonal antibodies the CD4 and CD8 surface molecules were identified and used to study T cell differentiation and immunoregulation in more precise detail (64, 65). The inducer Ly1 population was exclusively found in the CD4` population and the cytotoxic and suppressor



189

populations were largely contained within the CD8` population (66, 67). In addition to these functional distinctions, CD4` and CD8` T cell subsets also differed in a cognitive sense. Thus, CD4` T cells were found to be MHC class II restricted, whereas CD8` T cells were found to be MHC class I restricted (65, 68, 69). In both mice and man suppression was shown in vitro to require interactions between CD`8 cells and CD4` cells (69–72). These experiments led to the findings that a subset of activated CD4` T cells were required to induce CD8` suppressor cells. Moreover, the CD8` effector cells then could mediate suppression by suppressing the activity of the CD4` T cell inducing population. From these experiments emerged models in which extraordinarily complex cellular circuits were envisioned (37, 73–76). For example, a CD4` suppressor inducer T cell subset was characterized by the expression of two additional cell surface molecules identified by alloantisera, one termed I-J (77–80) and the other termed Qa-1 (72, 81). The precise genetic or molecular description of I-J has remained mysterious, in part because the cell surface structure recognized by I-J alloantisera was never identified and the original evidence that I-J was encoded in the MHC class II region was not correct (82, 83). In contrast, the Qa-1 alloantisera mapped to the 38 end of MHC class I region on chromosome 17 in a region known to encode other genes, like TL (84, 85). It is of interest that eventually Qa-1 was cloned and found to be a MHC class Ib molecule capable of presenting endogenous as well as exogenous peptides (86–88). However, the potential significance of Qa-1 expression on suppressor inducer cells was not delineated until the mid 1990’s (25, 33) (see sections below). In this regard, it is important to re-emphasize that the original suppressor cell circuits were initially conceived and/or deduced at a time when molecular immunology was in its infancy. For example, the nature of the TCR receptor was unknown as was the precise structure and function of MHC molecules in restricting T cell activity. In addition, the great majority of the cytokines that are now known to regulate immune functions and T cell differentiation were largely unknown. Clearly, understanding the precise role of these lymphokines in suppression would have significantly influenced the interpretation of data suggesting that an array of antigen specific and nonspecific suppressor factors were uniquely secreted by suppressor inducer and effector cells. To this day the precise biochemical definition of these putative antigen specific suppressive factors has not been resolved. An alternative hypothesis that will be discussed in depth below is the idea that the specificity of immune suppression is mediated not by the putative specific soluble factors but instead by cognate interactions between CD8` regulator cells and antigen-activated CD4` T cells. In this view, conventional ab TCRs expressed on the surface of suppressor CD8` T cells recognize activated CD4` T cells in an antigen-specific manner and are triggered to inactivate or delete the CD4` T cell. We in fact review below the evidence that the ‘‘uniqueness’’ of suppressor cells and their TCRs may simply be that they are endowed with the capacity to recognize self TCR peptides presented by MHC class I molecules expressed by antigen activated inducer clones. The consequences of TCR


190

JIANG n CHESS

triggering of ‘‘suppressor’’ CD8` T precursors could then be quite conventional and involve differentiation into specific cytotoxic lymphocytes (CTL), which lyse targets (inducer CD4` T cells) and/or secrete lymphokines that downregulate these inducer cells. Indeed, the cautionary notes concerning suppressor T cells and the doubts concerning the very existence of suppressor T cells that emanated in the middle to late 1980s were based on the lack of evidence that suppressor cells were a unique subset of T cells distinguished from helper and cytotoxic T cells (28, 89– 91). These cautionary notes emphasized that there was no evidence that suppressor T cells were endowed with unique sets of antigen receptors (distinct from the ab and cd TCRs), unique antigen specific soluble factors or unique I-J encoded suppressor structures. Indeed these arguments remain essentially valid. These concerns led to general skepticism for the role of T-T interactions in the downregulation of immune responses. On the other hand, evidence has continued to emerge during the last two decades that has substantiated many of the central ideas of immune suppression and particularly the idea that the immune system, not unlike all other complex physiologic and biochemical systems, is homeostatically regulated and often involves feedback suppressor mechanisms. In the next sections we will first briefly review the substantial in vivo data that unequivocally show that during immune responses regulatory suppressor CD8` T cells are induced that profoundly effect the outcome of autoimmunity and allograft responses by downregulating immunoreactive CD4` T cells.

THE RESURRECTION OF THE CD8` T SUPPRESSOR PARADIGM IN VIVO T Cell Vaccination Induces Regulator Cells, Including CD8` T Cells, Which Prevent Autoimmunity and Abrogate Immune Responses to Alloantigens One prediction of the model that suppressor T cells recognize TCRs expressed on antigen-activated CD4` T cells is the notion that CD4` T cell clones specific for antigen X may be used to induce regulatory CD8` T cell suppressors in vivo, which would specifically downregulate the immune responses to CD4` T cells expressing a TCR specific for antigen X. Early suggestions that this prediction is correct came from experiments that show that specific unresponsiveness to transplantation antigens could be induced by auto-immunization with syngeneic, antigen-specific T lymphoblasts (92, 93). Subsequent experiments provided evidence that these allospecific lymphoblasts serving as an immunogen can induce antiidiotypic immunity and tolerance induction. For example, immunization of F1 rats with alloreactive T-cell populations of parental strain origin, induces a hostmediated T-cell response that is specific for anti-MHC receptors on parental T cells. This protective immunity is rapid in onset and once induced, was shown to



191

provide highly effective, specific resistance to lethal GVH disease. Moreover, this resistance could be adaptively transferred to syngeneic recipients (94–96). Furthermore, Irun Cohen in the early 1980’s initiated a series of experiments that demonstrated that autoimmune T cell lines and clones can be used to vaccinate animals to prevent autoimmunity (97, 98). For example, the experimental autoimmune rodent model of multiple sclerosis, experimental allergic encephalomyelitis (EAE), can be induced either by injection of myelin antigens including myelin basic protein (MBP) or alternatively, by inoculation with ‘‘encephalogenic’’ MBP reactive T cell lines. However, if these encephalogenic T cells are attenuated by irradiation prior to injection, they do not cause disease but instead induce protection from subsequent induction of disease. As a consequence of these experiments, Irun Cohen, coined the term ‘‘T cell vaccination’’ (TCV) to denote the use of attenuated autoimmune T cells as vaccines to prevent autoimmune disease. Unlike conventional immunotherapy, T cell vaccination is selectively immunosuppressive and unlike other forms of immunotherapy TCV does not necessarily require precise knowledge of the autoantigen, which is often clinically difficult to define. Moreover, TCV has been widely used to ameliorate autoimmune diseases in a variety of animal models (99–102) and studies in some human autoimmune diseases have already been initiated (103–105). Although the precise mechanisms by which T cell vaccination abrogates autoimmunity are unknown, it has generally been thought that T cell vaccination augments the normal mechanisms employed by regulatory T cells to control the pathogenic potential and/or outgrowth of disease-causing T cells (106). In this regard, several lines of evidence have suggested that regulatory T cells mediate the protective effect of T cell vaccination by downregulating autoimmune T cells. For example, adoptive transfer of lymph node cells obtained from T cell vaccinated rats to naive rats, led to acquisition of EAE resistance by the recipient rats (20, 31). In vitro studies of these regulatory cells provided evidence for CD8` T cells that recognize and downregulate the vaccine T cells. Moreover, in the studies of multiple sclerosis in humans (104) the majority of the regulatory T cells isolated from CD4` T cell–vaccinated patients were CD8` T cells. These CD8` T cells were shown to inhibit the proliferation of vaccine T cell clones upon antigen stimulation and also specifically lysed the vaccine CD4` T cell clones in vitro (31, 104). Recently, experiments from our laboratory have shown that CD8` T cell depletion completely abolished TCV-induced protection of EAE induced by encephalogenic clones, providing direct evidence that CD8` T cells are required for TCV-induced protection from EAE in mice (107).

Direct Evidence that CD8` T Cells Downregulate Immune Response to Both Autoantigens and Foreign Antigens In Vivo Three sets of experiments in the early 1990’s provided further direct evidence that CD8` T cells regulate immune responses in vivo (22, 23, 108). The first experiment employed the EAE model in B10PL animals in which mice immu-


192

JIANG n CHESS

nized with an encephalitogenic peptide of MBP develop acute EAE and recover from the first episode of disease and do not develop relapsing EAE. Importantly, the first episode of EAE renders the animals resistant to the re-initiation of EAE by secondary immunization. In contrast, animals depleted of CD8` T cells in vivo, using anti-CD8 monoclonal antibody (Mab)–mediated clearance are no longer resistant to a second induction of disease with MBP (22). These data demonstrate that CD8` T cells play a major role in the resistance to a second induction of EAE. Furthermore, as noted above, TCV-induced protection of EAE is abrogated in mice depleted of CD8` T cells (107). In a second set of experiments designed to investigate the role of CD8` T cells in EAE, CD8-/- ‘‘knockout’’ mice were generated and bred with the EAEsusceptible PL/J H-2u strain. The onset of disease and susceptibility were similar in both the CD8-/- mice and wild-type mice. However, the mutant mice had more chronic EAE, reflected by a higher frequency of relapse (23). These experiments provide direct evidence that CD8` T cells play a role in both inducing resistance to autoimmune disease and in abrogating recurrent relapsing episodes of autoimmunity. These experiments are of clear clinical interest because it is known that the great majority of human autoimmune diseases manifest acute attacks of autoimmunity variably interspersed with periods of remission. In the third set of experiments, it was shown that CD8` T cells control the immune response of H-2d mice to the conventional antigen, hen egg lysozyme (HEL) (108). Thus, Vba haplotype mice that lack 10 TCR Vb gene segments respond efficiently to the HEL peptide 74–96. In contrast, normal Vbb H-2d mice do not respond to the HEL peptide 74–96. Importantly, the normal Vbb H-2d mice depleted of CD8` T cells in vivo (by anti-CD8 Mab treatment) respond vigorously to this antigen. These studies demonstrate CD8` T cells control and normally suppress the immune response to certain peptides derived from conventional antigens like HEL. These studies also suggest that the TCR Vb region of the genome encodes proteins involved in this CD8` T cell regulation (see below). These studies taken together with the experiments reviewed above suggest that CD8` T cells govern immune responses to both conventional and autoantigens.

EVIDENCE THAT CD8` T CELLS, IN VIVO, CONTROL THE OUTGROWTH OF CD4` T CELLS IN A Vb SPECIFIC MANNER Studies of TCR Vb Peptide Immunization A variety of studies have provided indirect evidence that the TCR or peptides derived from the TCR of autoimmune T cells are part of the target structure recognized by regulatory CD8` T cells. For example, regulatory CD8` T cells isolated from T cell–vaccinated animals preferentially recognize the vaccine T



193

cells activated by the specific antigen but not other T cells activated by different antigens in vitro (31, 104). Moreover, other studies have shown that pathogenic populations of immune cells are restricted in terms of TCR gene usage and the self peptide epitope recognized (109, 110). In addition, the classical TCV studies reviewed above have been extended in a significant manner to show that vaccination with TCR Vb peptides also efficiently prevents the development of EAE in rats and mice. Although TCR peptide immunization may function via more distinct mechanisms than ‘‘whole’’ T cell vaccination, it provides additional evidence that the TCR is involved in immune responses that control autoimmunity (111–113). The potential mechanism of this protection was investigated in experiments of Gaur et al who showed that vaccinating mice with a TCR Vb8.2 CDR2 peptide induces a state of TCR Vb specific unresponsiveness. Thus, the in vitro proliferative response to immunogens that usually induce an immune response mediated by Vb8.2 T cells are inhibited by prior TCR peptide vaccination. Moreover, the activation of T cells by anti-Vb8.2 crosslinking was also inhibited in a Vb specific manner in vaccinated mice. Importantly, depletion of CD8` cells before TCR peptide vaccination blocked such inhibition (24). Thus, the inhibition was dependent on CD8` T cells providing support for the idea that protection from EAE by TCR Vb peptide immunization is CD8` T cell dependent. Taken together, these TCR peptide immunization experiments provide substantial in vivo support for the prediction of some of the early notions of immune suppression. For example, one consequence of immunity is the induction of CD8` regulatory cells which are capable of specifically downregulating CD4` T cells through recognition of TCR related structures and suppressing immune responses to autoantigens and to alloantigens, in vivo. While the detailed mechanisms of this protection induced by vaccination with TCR Vb peptides is currently being actively investigated, preliminary clinical trials using relevant TCR Vb peptides have been initiated in patients with autoimmune disease, including multiple sclerosis (114–117).

Studies of the TCR Vb–Specific CD8` T Cell Downregulation of Antigen Activated CD4` T Cells In Vivo To develop direct evidence that CD8` T cells normally function to control the outgrowth of specific sets of CD4` T cells in vivo attention was turned to the study of superantigen-induced outgrowth of CD4` T cells (25). Superantigens bind to particular TCR Vb molecules and as a consequence activate a relatively large set of T cells expressing the particular TCR Vb to proliferate, differentiate and secrete cytokines. Administration of superantigen in vivo is followed by rapid changes in the populations of circulating T lymphocytes expressing the particular TCR Vbs. For example, after injection of the superantigen staphylococcus enterotoxin B (SEB), there is an initial deletion (12–24 h) followed by a proliferative expansion and a second phase of deletion (after day 4) of CD4` and CD8` T cells that express TCR Vb8 chains (118–120). The increase in Vb8` T


194

JIANG n CHESS

cells reaches a maximum on day 4 and by day 10 returns to background. However, the CD4`Vb8` but not the CD8`Vb8` T cell population becomes reduced to about 30%–40% below baseline and remains at this reduced level for at least 21 days. The mechanisms responsible for the initial depletion of CD4` Vb8` cells are thought to be predominately a consequence of apoptosis because it is known that triggering of the TCR alone by either superantigen or anti-TCR antibodies induces programmed cell death (121–123). On the other hand, the mechanism of the prolonged deletion of CD4` Vb8` T cells was of interest and we entertained the possibility that this reflected CD8` T cell regulation. To test this idea the regulation of CD4` Vb8` T cells was studied in SEB primed CD8` T cell–deficient mice. These experiments specifically showed that the downregulation of CD4`Vb8` T cells below baseline is not observed in mice depleted of CD8` cells or in CD8` T cell–deficient b2 M-/-mice. Further, the effects of CD8` T cell depletion were TCR Vb specific because CD8` T cell depletion had no effect on populations of CD4` T cells unaffected by SEB. Taken together, these experiments provided evidence that CD8` T cells, in vivo, control the outgrowth of CD4` T cells in a TCR Vb– specific manner (25). In experiments designed to determine possible mechanisms of the CD8` T cell regulation of CD4` Vb8` T cells, CD8` T cells derived from SEB primed mice, were found to be cytotoxic to CD4` T cell targets based on their TCR Vb expression. For example, Vb8` but not Vb81 targets were killed. Furthermore, this TCR Vb–specific cytolysis was b2-microglobulin (b2l)– dependent because target CD4`Vb8` T cells from b2 M-/-mice were not lysed by these CD8` CTL. Taken together, these data provided evidence that the Vb8– specific regulatory functions of CD8` T cells in vivo and in vitro is dependent on b2l-associated molecules. Moreover, in other experiments it was shown that Vb8 specific CD8` T cells are also induced by T cell vaccination with SEB– activated CD4` Vb8` T cells (33). Thus, the fact that CD8` Vb specific CTL can be induced in vivo by activated CD4` T cells strongly suggests that these CD8` Vb–specific CTL may contribute significantly to the protective effects of T cell vaccination. Further evidence for the in vivo protective role of CD8` Vb–specific T cells in T cell vaccination has come from preliminary experiments in an EAE model system in which a MBP-specific encephalogenic CD4`Vb8` T H1 clone, termed 1AE10, was used to initiate EAE (107). EAE is induced in virtually all mice within 10 days and TCV employing 1AE10 gives complete protection. This model permitted the direct analysis of the Vb specificity of TCV-induced protection as well as the effects of CD8` T cell depletion on this protection. For example, CD8` T cell–depleted animals were compared to control groups that either did not receive TCV or were not CD8` T cell depleted. As noted above, CD8` T cell depletion completely abolished TCV-induced protection in EAE animals showing that CD8` T cells are required for efficient protection from EAE in this model. Moreover, the TCR Vb specificity of this CD8` T cell–mediated protec-


195

tion was studied. Thus, MBP specific, TCR Vb8` and TCR Vb6` clones, were compared with respect to their capacity to serve as vaccine T cells and induce protection from EAE. These studies showed that Vb6` clones did not protect animals, whereas the Vb8` 1AE10–inducing clone and other MBP-specific TCR Vb8` clones protected animals from EAE. Therefore, TCV-induced in vivo protection from EAE is CD8` T cell dependent and Vb specific (107).


EVIDENCE THAT TCR Vb8–SPECIFIC REGULATORY CD8` T CELLS ARE RESTRICTED BY THE MHC CLASS IB MOLECULE, QA-1 Because the Vb8-specific regulatory functions of CD8` T cells were dependent on b2l-associated molecules it was expected that anti-sera to conventional H2 molecules would inhibit function. However, quite surprisingly, the M1/42.39 antibody, which is a rat anti mouse H-2 monoclonal antibody known to be specific for all H-2 haplotypes including H-2Kd and H-2Dd and H-2Ld MHC class I antigens (124), did not block the function of the Vb-specific CD8` CTL in BALB/ c mice despite the fact that the antibody efficiently blocked control of allospecific CTL. These experiments led us to the hypothesis that the Vb specific CTL were, perhaps, restricted by non-classical MHC class 1b molecules. Moreover, because of the experiments reviewed above showing that the Qa-1 molecule was prominently expressed on cells involved in T cell suppression, and because it was known that Qa-1 was preferentially expressed on activated lymphoblasts, we entertained the idea that Vb specific CD8` regulatory cells were Qa-1 restricted. Three distinct types of experiments subsequently showed that Qa-1 could function as restriction elements for CD8` TCR Vb–specific regulatory T cells. First, in a series of blocking studies using well characterized sera that initially defined the Qa-1a and Qa-1b alleles it was shown that the Qa-1b but not the Qa-1a determinant was expressed on the antigen-activated CD4` target cells derived from BALB/c mice. Moreover, the anti-Qa-1b but not the anti-Qa-1a antiserum efficiently blocked Vb8-specific CD8` CTL generated from the spleens of SEBprimed mice. The same anti-Qa-1b did not block the killing of target cells by CD8` anti-MHC class Ia allogeneic CTL. These conventional allospecific CTL, as expected, were blocked by the rat anti-mouse H-2 monoclonal antibody (25). In the second set of experiments a human B cell line was transfected with either Qa-1 alone or Qa-1 together with TCR Vb8 or Vb6 cDNAs. These transfectants were then used as targets for CD8` Vb-specific CTL induced by T cell vaccination with activated CD4` Vb8` T cells (33). These CTL efficiently killed only those target cells expressing both murine Qa-1b and Vb8. Transfectants expressing Vb8 or Qa-1b alone were not lysed and transfectants expressing Qa1b and Vb6 were also not killed. These results demonstrate that co-expression of murine Qa-1b and Vb8 cDNAs is necessary and sufficient for human B cells to


196

JIANG n CHESS

serve as targets for the murine CD8` anti-Vb8 CTL. It is important to emphasize that the target transfectants are B cells that do not express cell surface TCRs. Thus, Vb8 sequences do not need to be expressed on the cell surface as part of a mature TCRb polypeptide in order to be recognized by the Vb specific CD8` CTL. In this regard, it is well known that MHC class I molecules can efficiently present processed intracellular protein antigens. Moreover, it is also known that Qa-1 molecules can present foreign antigens (125) as well as self MHC class Ia peptides (126). In light of these considerations, these results are most consistent with the idea that CD8` anti-Vb8 CTL recognize TCR Vb peptide/Qa-1 complexes expressed on target T cells (33). The third line of evidence that Qa-1 molecules can serve as targets for CD8` TCR Vb specific regulatory cells stems from experiments employing CD8` T cell hybridoma clones derived from B10.PL mice that were vaccinated with an irradiated syngeneic MBP-specific, encephalogenic CD4` Vb8` T cell clone (33). For example, a CD8`Vb-reactive, Qa-1-restricted hybridoma clone (21– 5A9) was isolated that preferentially reacts with syngeneic target T cells expressing TCR Vb8, but not Vb6, in a Qa-1-restricted manner. In the course of these studies it was observed that antigen-activated but not resting CD4`Vb8` Qa-1 targets could trigger the hybridoma. Kinetic studies showed that activated CD4` T cells were most stimulatory on days 4–8 following activation and were nonstimulatory by day 12. One interpretation of this kinetic data is that it may reflect the kinetics of the activation-dependent expression of Qa-1/TCR peptide by CD4` T cells. In support of this, the level of Qa-1-expression increases with T cell activation and peaks around day 4 to 7 and is absent by day 10. Moreover, the responses induced by the activated CD4` T cell targets were inhibited in an allele specific manner by Qa-1 antisera. Taken together, these studies show at a clonal level that Qa-1-restricted, Vb-specific CD8` T cells are generated by T cell vaccination (33, 127). In summary, these experiments support the view that Qa-1 can function as a restriction element for regulatory CD8` TCR Vb–specific cells. These studies do not exclude the possibility that other MHC class Ib molecules, or even some MHC class Ia molecules, may also subserve this function. However, the inhibition of TCR Vb–specific regulatory CD8` cells by anti-Qa-1 but not MHC class Ia antibodies, suggest that Qa-1 is the predominant restriction element. On the other hand, the studies above also do not rule out the possibility that other types of regulatory cells including CD4` T cells may downregulate immune responses in a Vb-specific fashion and use entirely different MHC restriction elements (see below). Nevertheless, we will point out in subsequent sections, that there are a number of biologic features of the Qa-1 molecule that make it particularly suitable to function as a restricting element to present self-TCR peptides to regulatory CD8` T cells. For example, the preferential expression of Qa-1 on antigenactivated T cells would preclude resting T cells from being susceptible to downregulation.


197


EVIDENCE THAT TCR Vb–SPECIFIC REGULATORY CD8` T CELLS INDUCED BY TCV PREFERENTIALLY RECOGNIZE TH1 BUT NOT TH2 CD4` CELLS Further analysis of the Qa-1-restricted Vb-specific T hybridoma clone, 21–5A9, revealed that it preferentially recognized CD4`Vb8` TH1 clones but was not reactive with CD4`Vb8` TH2 clones (127). This was the case even for pairs of TH1 and TH2 clones expressing identical TCR Va and Vb chains. The mechanism that accounts for this differential recognition is not clear, but preliminary analysis suggests that it is not secondary to differential cytokine secretion. However, preliminary TCV studies suggest that the differential recognition of TH1 and TH2 cells by CD8` T cell hybridomas in vitro may represent a mechanism by which CD8` T cells regulate immune responses in vivo (107). Thus, mice were vaccinated with either the CD4`Vb8` T cell clones, 1AE10, (TH1) or the TCR identical CD4`Vb8` TH2 clone, 3AD2. The TH1 clone 1AE10 completely protected mice from subsequent induction of EAE, whereas the TH2 clone 3AD2 did not protect. Moreover, CD8` T cell depletion significantly abrogated this protection. This data is consistent with the model that Vb8` TH1 cells but not Vb8` TH2 cells are able to induce TCR Vb8-specific, Qa-1-restricted regulatory CD8` T cells that specifically downregulate TH1 T cell–induced disease in vivo. In this regard, it is of interest that TCV employing cDNA encoding Vb8 has been shown to protect animals from EAE. Although it is unknown if this form of TCV is also CD8` T cell–dependent, the protection observed is accompanied by a shift in the MBP-specific TH functional phenotype from TH1 to TH2 (128).

EVIDENCE THAT TCR Vb–SPECIFIC REGULATORY CD8` T CELLS EXIST IN HUMAN PERIPHERAL BLOOD A variety of studies in the human immune system provided evidence that recognition of the TCR is integral to the specificity of immune regulation between T cells subsets. For example, a series of experiments have shown that it is possible to generate human CD4` T cell clones that proliferate specifically to autologous CD4` clones that either show a particular antigen specificity or specific Vb expression (129–131). Moreover, human CD8` T cell clones raised against autologous allo-reactive CD4` T cell lines inhibit fresh autologous CD4` T cells from proliferating to the same alloantigen (132). These results are consistent with the idea that T-T cell interactions may involve all or part of the TCR as a target of recognition. Evidence that Vb-specific CD8` T cells exist in human peripheral blood came from studies showing that CD8` T cells raised to CD4` Vb2` T cell clones, in vitro, differentiated into CTL that specifically lyse independently isolated


198

JIANG n CHESS

autologous CD4` T cell targets expressing Vb2` but not Vb21 TCRs. Moreover, in reciprocal cloning experiments, CD8` T cells raised against autologous CD4` T cell clones with different TCR Vb usage were cytotoxic only to CD4`T cell targets expressing TCR Vbs expressed by the inducing clone. Taken together, these experiments indicate that CD8` T cells can kill autologous CD4` T cells based, at least in part, on the recognition of the TCR Vbs expressed. Interestingly, this TCR Vb–specific cytotoxicity was not blocked by the monoclonal antibody W6/32, which reacts with nonpolymorphic determinants present on HLA A/B/C Class Ia molecules. This data suggests that like in the mouse, MHC class Ib structures might serve as the restricting element (133). In contrast, Zhang et al showed that human CD8` T cells, isolated from T cell–vaccinated patients, which specifically recognize and kill potentially encephalogenic autologous MBPreactive CD4` T cell clones, could be blocked by the conventional HLA class Ia antibody, W6/32 (104). Taken together, these data suggest the possibility that under certain conditions MHC class Ia as well as MHC class Ib molecules may present TCR Vb peptides, at least in the human system.

REGULATORY CD4` T CELLS In this section we briefly review immunoregulatory functions effected by CD4` T cells. First, in a series of experiments, Kumar and Sercarz identified a CD4` T cell population in EAE-recovered mice, which specifically respond to a peptide derived from TCR Vb 8.2. These CD4` T cells downregulate 1–9Nac MBP proliferative responses, in vitro, and in vivo protect mice from EAE. For example, they isolated a CD4` T cell clone, specific for a peptide derived from framework region 3 of TCR Vb8.2 chain (B5). When adaptively transferred, this clone protects animals from subsequent induction of EAE. Moreover, immunizing animals with the B5 peptide protected animals from EAE (134–137). These CD4` T cells are analogous to the TCR Vb–specific CD8` T cells described above; however, as expected the CD4` regulatory T cells are restricted by MHC class II molecules instead of MHC class Ib molecules. Because activated murine T cells do not express MHC class II, they obviously cannot process and present MHC class II/ Vb peptide to regulatory CD4` T cells. Thus, the precise means by which the MHC class II restricted regulatory CD4` T cells are induced and function is unclear. Two distinct mechanisms were considered (138). First, it is possible that the TCR Vb peptides derived from antigen-activated CD4` T cells might be processed and presented by conventional antigen presenting dendritic cells or macrophages. This could occur, for example, if apoptotic CD4` T cells or if membrane fragments containing TCR Vb, are ingested by dendritic cells. These dendritic cells could then present TCR peptides complexed to either MHC class II or MHC class I and subsequently induce both Vb-specific CD4` or CD8` regulatory T cells. The CD4` regulatory T cells would not be able to directly downregulate antigen-activated CD4` T cells which do not express MHC class



199

II molecules. However, these cells could locally secrete modulatory cytokines, such as TGF-b or IL-10 at the site of CNS inflammation and as a consequence downregulate MBP-specific CD4` T cells. Alternatively, it is possible that dendritic cells that can present Vb peptides on either class I or class II MHC molecules might locally focus both CD4` regulatory cells with TCR Vb–specific CD8` regulatory T cells on the same cell surface in a CD40L-dependent manner analogous to mechanisms employed by CD4` T cells to help conventional cytotoxic CD8` CTL (139–141). This possibility is supported by experiments that show that the function of TCR Vb–specific regulatory CD4` T cells, in vivo, is dependent on CD8` T cells (134). The second set of studies demonstrate that under certain experimental conditions CD4` T regulatory cells can downregulate antigen specific CD4` T cells independent of other T cells. For example, it was shown that TCR transgenic mice (T/R`) specific for MBP rarely develop spontaneous EAE. However, when these T/R` mice are crossed with RAG-1-deficient mice to obtain mice (T/R1) that have T cells expressing the transgenic TCR almost all mice develop spontaneous EAE (142). Because both T/R` and T/R1 mice have large numbers of the potentially encephalitogenic CD4` anti-MBP T cells these results suggest that a very small number of nontransgenic lymphocytes, that are present in T/R` but absent in T/R1 mice can potently suppress the in situ activation of CD4` anti-MBP T cells mediating EAE. To identify the cellular regulatory mechanisms important in this suppression, T/R` mice were crossed into mice deficient in either B cells, CD8` T cells, NK1.1 CD4` (NKT) cells, c/d cells or a/b cells (143, 144). Only mice that were deficient in CD4` a/b T cells developed EAE. Moreover, T/R1 mice were protected from EAE by the early (prior to the onset of EAE) adoptive transfer of purified CD4` T cells from normal donors. These results support the view that under certain experimental conditions CD4` cells alone can suppress the initiation of autoimmunity. The specificity or mechanisms of this suppression is currently unknown. It is conceivable that the CD4` T regulatory cells are non-specifically activated to secrete cytokines known to downregulate EAE activity. In this regard, several groups have demonstrated that CD4` T cells secreting TGF-b IL-4 or IL-10 can down regulate a variety of autoimmune diseases including EAE, murine inflammatory bowel colitis models and NOD diabetes models (143, 145–149). It is of interest that both CD4` and CD8` regulatory T cell clones, which specifically recognize potentially encephalogenic autologous MBP-reactive CD4` T cell clones, have been observed in patients with multiple sclerosis treated by TCV with attenuated MPB-reactive clones (104, 150). In these experiments, although the majority of specific regulatory T cells were CD8`, approximately 10% were CD4`. Because activated human CD4` T cells can express MHC class II, it is conceivable that these regulatory CD4` T cells downregulate target CD4` T cells through recognition of TCR peptides presented by MHC class II molecules. If this were the case these regulatory CD4` T cells would be analogous to the Qa-1 restricted, TCR Vb–specific regulatory CD8` T cells described

200

JIANG n CHESS


above. Speculating further, it is of interest that some MHC class II molecules, including HLA-DQ are less polymorphic then the more prevalent HLA-DR molecules and could potentially serve to present a restricted set of TCR peptides. In this regard suppressor cells restricted by HLA-DQ class II molecules have been observed in patients with lepromatous leprosy who are selectively unable to mount cellular immunity to Mycobacterium leprae (151). Although it is unclear what factors govern whether CD4` or CD8` regulatory cells are induced, as noted above it is possible that dendritic cells that can present Vb peptides on either class I or class II MHC molecules may serve to induce either CD4` or CD8` TCR-specific regulatory cells. In this regard, it is known that dendritic cells can express both MHC class I and II molecules, including the less polymorphic HLA-DQ molecules (152, 153).

THE POTENTIAL FUNCTIONAL SIGNIFICANCE OF RESTRICTING T REGULATORY CELLS BY THE MHC CLASS IB MOLECULE, QA-1 There are a number of biologic features of the Qa-1 molecule that may make it particularly interesting with respect to its potential function as a restricting element to present self TCR peptides to CD8` T cells. First, although Qa-1 is expressed on a variety of hematopoietic tissues, it is only minimally or not expressed at all on resting T cells and thymocytes and its expression is significantly increased following peripheral T cell activation by antigen (154). For example, analysis of Qa-1 isolated from detergent lysates of surface labeled cells indicates that Qa-1 displays little charge heterogeneity on resting lymphocytes, but the level of expression and degree of charge heterogeneity are both increased on activated lymphocytes (155). Moreover, because the peripheral Qa-1 expression is short lived and persists for only a few days on activated CD4` T cells, this may exclude resting T cells from downregulation by Qa-1-restricted CD8` regulatory cells. In this way the relatively short-lived upregulation of Qa-1 on antigen-activated T cells is analogous to the transient expression of CD40L on antigen-activated T cells (7). The transient expression of both of these functional molecules, one for help (CD40L) and the other for suppression (Qa-1) on the surface of antigen-activated T cells would serve to confine the regulatory functions of these molecules to a time frame of specific antigen activation and may consequently prohibit promiscuous help or suppression. In addition, the rather limited expression of Qa-1-in the thymus may account for the fact that Qa-1restricted CD8` T thymocytes are not negatively selected in the thymus. The mechanism for the potential positive selection of Qa-1-restricted T cells in the thymus is currently unknown. It is conceivable that other MHC class Ia or class Ib molecules may serve to positively select those Qa-1-restricted T cells in the thymus.



201

Second, Qa-1 is known to be a MHC class Ib molecule of limited polymorphism with the potential to present self and foreign peptides to CD8` T cells (87, 88). The predominant self peptide presented by Qa-1 is termed Qdm (for Qa-1– determinant modifier), derived from the signal leader sequence of certain murine class Ia molecules (156–159). This peptide (AMAPRTLLL) binds with high affinity and accounts for the majority of the peptides associated with this molecule (159, 160). However, Qa-1 can also bind other self peptides including those derived from heat shock proteins (161) and preproinsulin leader sequences (162). In addition, data reviewed here suggests that CD8` T cells recognize TCR Vb motifs restricted by Qa-1. In addition, Qa-1 can bind exogenous peptides including bacterial derived peptides (160, 163–165). Many features of the Qa-1 binding cleft are also conserved in the rat RTBM.1 and in human HLA-E molecules, suggesting that these molecules may associate with structurally similar peptides (87). Thus, it is perhaps not surprising that HLA-E, like Qa-1, also binds to Qdm as well as Qdm-like leader sequence peptides derived from HLA class Ia molecules. The functional significance of the Qa-1/Qdm complex as well as similar HLAE/leader peptide complexes has recently begun to be illuminated by the finding that both Qa-1/Qdm complexes and HLA-E/leader peptide complexes interact with the CD94/NKG2 heterodimeric receptors expressed on NK cells (166). This set of receptors is comprised of an invariant C type lectin chain, CD94, that is disulfide-bonded with the NKG2 glycoproteins (167–169). The CD94/NKG2 receptor complexes are expressed on both human and murine NK cells as well as a subset of T cells (166, 170). The human CD94/NKG2 complex recognizes HLA-E/MHC class Ia leader sequence peptides (171), whereas the murine CD94/ NKG2 complex binds the homologous Qa-1/Qdm complex (170). As a consequence, tetramers of HLA-E or Qa-1 bind to a large fraction of human and murine NK cells, respectively (170–173). The CD94/NKG2 receptors is of interest in that they appear to gauge the level of MHC class I expression in targets, indirectly, by recognizing leader sequences of MHC molecules. Functionally these receptors are thought generally to limit NK killing of normal targets. The regulatory functions of CD94/NKG2 receptors are largely dependent on signaling through the NKG2 molecules. For example, the NKG2A/B molecules possess two immunoreceptor tyrosine-based inhibition motif (ITIM) sequences in their cytoplasmic domain, responsible for the inhibitory function of these receptors, whereas other NKG2 proteins lack ITIMs and transmit positive or activating signals (168, 174). For example, the human CD94/NKG2C binds to HLA-E, and noncovalently associates with DAP12, a membrane receptor containing an immunoreceptor tyrosine-based activating motif (ITAM). Efficient expression of CD94/ NKG2C on the cell surface requires the presence of DAP12, and charged residues in the transmembrane domains of DAP12 and NKG2C are necessary for this interaction (175). Phosphorylated DAP12 peptides bind ZAP-70 and Syk protein tyrosine kinases, suggesting that this activation pathway is similar to that of the T and B cell antigen receptors (176). Taken together, these various MHC class I

202

JIANG n CHESS

receptors expressed by NK cells are endowed with molecular mechanisms (ITIMs and DAP12-related ITAMS) not only to function as killer inhibitory receptors but perhaps more importantly as receptors maintaining homeostasis. (174, 177).


CD94/NKG2 RECEPTORS REGULATE THE FUNCTIONAL SIGNALING OF a/b AND c/d TCRS In this regard, the CD94/NKG2 receptors are also expressed on subsets of T cells expressing either a/b or c/d TCRs and therefore potentially, in concert with these TCRs, may regulate not only NK function but also antigen specific functions of more conventional T cells. For example, human c9/d2 T cells known to recognize phosphoantigens expressed by certain bacteria, mycobacterium and tumor cells have recently been shown to co-express CD94/NKG2 receptors, and signaling through the CD94/NKG2 receptor regulates the synthesis of cytokines (IFN-c and TNF-a) as well as antigen-induced cytotoxic functions (178). CD94/NKG2 receptors may also be co-expressed and regulate conventional MHC class I restricted a/b CD8` CTL specific for tumor or viral peptide/MHC complexes (179–182). Of interest, the co-expression of CD94/NKG2 receptors on a/b CD8` T cells is upregulated by cytokines including TGF-b and IL-10 (180, 181). The potential functional significance of the co-expression of a/b TCR with CD94/NKG2 receptors on CD8` CTL was suggested in experiments showing that ligation of CD94 on these cells can either activate or inhibit killing mediated by the a/b TCR (179, 183). These data support the idea that the co-expression of a/b TCR and CD94/ NKG2 receptors on conventional CD8` CTL homeostatically regulate CTL functions. It is of interest to speculate on how CD94/NKG2 receptors may functionally interact with the a/b TCRs expressed by Vb-specific, Qa-1-restricted regulatory CD8` T cells. First of all, recent data assaying Qa-1/Qdm teramers binding to T cell subsets suggest that a significant number of murine CD8` T cells bind to these tetramers and, thus, presumably express CD94/NKG2 receptors (170, 172). This suggests that clones of the regulatory CD8` T cells, in fact, co-express two distinct types of Qa-1 receptors. One type is an a/b TCR that recognizes Qa-1/ TCR peptide complexes and the other is the CD94/NKG2 receptor that recognizes Qa-1/Qdm. These two receptors, as noted above, could be differentially regulated on the cell surface of CD8` T cells and functionally regulate one another. Moreover, depending on whether CD94NKG2 receptors have ITIMs or alternatively bind DAP12 they may either enhance or downregulate CD8` T cell function. In addition, during CD4` T cell activation both Qa-1/Qdm and Qa-1/TCR Vb ligands may be differentially upregulated. For example, resting CD4` T cells may preferentially express Qa-1/Qdm, and with activation, other peptides (presumably TCR Vb peptides) may bind to Qa-1. This hypothesis is consistent with biochemical data that show that the Qa-1 molecule displays little charge


203


heterogeneity on resting lymphocytes but that the level of expression and degree of charge heterogeneity are both increased on activated lymphocytes (155). This increase in charge heterogeneity may reflect the presence of newly formed Qa-1/ TCR Vb complexes expressed during activation. The functional consequence of this model is that resting T cells will preferentially express Qa-1/Qdm, which following ligation of CD94/NKG2 may for the most part trigger inhibitory signals. At a point in time during T cell activation the prevalence of Qa-/TCR Vb expressed on target CD4` T cells may reach a threshold sufficient to trigger positive signals through the a/b TCR expressed on CD8` T cells. In this way the two types of Qa-1 receptors expressed on CD8` T cells may ultimately provide homeostatic control of the regulatory T cells.

GENERAL MODEL OF T CELL REGULATION BY CD8` T CELLS IN THE CONTEXT OF PERIPHERAL TCR REPERTOIRE SELECTION In this review we have summarized some of the pertinent evidence that CD8` T cells regulate ongoing peripheral immune responses, at least in part, by cognate interactions with antigen-activated CD4` T cells in a TCR-specific manner restricted by the MHC class Ib molecule, Qa-1 (see Figure 1, color insert). We have also reviewed the data that suggest that the Qa-1 molecule is uniquely qualified to serve this MHC restrictive function because unlike conventional MHC molecules it is preferentially expressed on activated and not resting CD4` T cells. In addition, Qa-1 expression in the thymus is either low or non-existent. Thus, precursor a/b TCR–expressing CD8` T cells restricted by Qa-1 are not negatively selected in the thymus. The conceptual origins of the current studies emerged from the experimental data and hypothesis that arose during the last three decades suggesting that suppressor CD8` T cells regulate peripheral normal immune responses and play a critical role in the regulation of autoimmune responses. In particular, key experiments defining the role of CD8` T cells in controlling autoimmunity stemmed from studies showing that mice that have recovered from the first episode of MBP peptide–induced EAE are resistant to re-induction of disease and that this resistance is CD8` T cell–dependent. More recent experiments have shown that this protection or resistance can be induced by TCV with activated MBP-reactive, CD4` T cells that induce Qa-1restricted TCR Vb–specific regulatory CD8` T cells. These data are compatible with the hypothesis that the regulatory CD8` T cells scan the surface of CD4` T cells and selectively delete or inactivate those T cells expressing certain TCR Vb peptides bound to Qa-1. This hypothesis would predict that the peripheral T cell repertoire of CD4` T cells reactive to MBP should be different in naı¨ve (EAE-susceptible) compared to EAE-recovered (EAE-resistant) animals. In this regard, elegant studies by Lehmann and Sercarz have shown that determinants of


204

JIANG n CHESS

MBP that are cryptic after primary immunization can become immunogenic in secondary responses and during the course of EAE (184–186). This diversification of the autoreactive T-cell repertoire has led to the concept of ‘‘determinant spreading’’ during the course of T-cell–driven autoimmune disease. Moreover, recent experiments in our laboratory have also investigated the peripheral TCR Vb repertoire of CD4` T cells to the 1–9 Nac MBP peptide in naı¨ve compared to EAE-recovered mice. These studies showed a significant reduction of certain Vb8.2-expressing MBP-reactive CD4` T cells in EAE recovered animals, accompanied by the emergence of other CD4` MBP-reactive T cells expressing other TCR Vbs, as well as the persistence of some MBP-reactive Vb8.2 clones. This change in the TCR repertoire was not observed in CD8` T cell–depleted animals. It is of interest that although EAE-recovered mice are resistant to disease, there is a persistence of some MBP-reactive CD4` T cells, including some TCR Vb8.2 clones. The persistent MBP-reactive CD4` clones may express distinct MBPspecific TCRs that escape TCR Vb-specific CD8` T cell recognition. For example, the TCRs on these persistent clones may be of too low affinity for MBP to trigger the mechanisms that normally permit downregulation. These observations lead us to propose the following hypothesis. Because regulatory CD8` T cells downregulate CD4` T cells based predominately on the recognition of Qa-1/TCR peptide complexes expressed on the surface of CD4` T cells, the cognitive distinction by which certain CD4` T cells are downregulated by CD8` T cells and others are not, is dictated by the ability of the CD4` T cells to process and present the relevant Qa-1/TCR peptide complexes on the cell surface. It is likely that the major determinant of whether or not a CD4` T cell will process and present Qa-1/TCR peptide will be the initial cognitive encounter between the particular TCR expressed by CD4` T cells and the MHC/ peptide complex presented on antigen-presenting cells. We envision that T cells expressing TCRs with either low affinity or high affinity interaction with MHC/ peptide will not signal Qa-1/TCR peptide presentation. Only T cells with ‘‘intermediate’’ affinity, above and under certain thresholds, will trigger Qa-1/TCR peptide presentation. The precise threshold for these CD4` T cells to process and present self TCR peptide coupled to Qa-1 molecules will not only be a function of the quality, intensity and duration of the initial tri-molecular complex interaction between antigen-specific CD4` T cells and antigen-presenting cells, but may also be influenced by signaling via co-stimulatory molecules including CD80, CD86 and CD40. Moreover, these same factors will determine whether the surface Qa-1 molecules are predominately comprised of Qa-1/Qdm or both Qa-1/Qdm and Qa-1/TCR peptide. The expression of Qa-1/TCR peptide will induce specific functional responses from regulatory CD8` T cells which, as discussed above, may be further regulated by Qa-1/Qdm expression. These considerations can be viewed as an ‘‘affinity model’’ of peripheral T cell regulation. If we consider this model with respect to self-reactive CD4` T cell clones, it is likely that high affinity clones will have been deleted in the thymus, so that the TCR affinity of clones for self determinants in the periphery



205

will either be low or intermediate. The T cells expressing TCRs with intermediate affinity to self are potentially pathogenic and capable of inducing auto-immune disease. These T cells will have been positively selected on self antigens in the thymus but escape negative thymic selection because they possess TCRs with intermediate affinity to self, sufficiently low enough not to induce apoptosis and thymic death. This type of T cells in the periphery may be triggered by self antigen to process and present Qa-1/TCR peptides and be subject to CD8` T cell– mediated downregulation. One biological consequence of this downregulation of self reactive T cells is to provide an additional mechanism to ensure peripheral self tolerance. In contrast, the T cells with very low affinity to self will have escaped negative selection in the thymus and also will not be subject to CD8` T cell–mediated downregulation in the periphery. We envision that this type of T cells may have high affinity to foreign antigens and thus be preserved as part of the peripheral T cell repertoire to foreign antigens. The MBP-reactive clones that persist in EAE-recovered mice or newly emerge in these mice may fall into this category of cells. These same principles can be readily extended to the CD8` T cell–mediated downregulation of the vast majority of T cells specific to foreign antigens. Thus, we propose that only T cells with intermediate affinity TCRs to MHC/foreign peptides will induce Qa-1/TCR peptide surface expression and be subject to CD8` T cell–mediated downregulation in periphery. The low affinity T cells may not receive sufficient TCR triggering to proliferate and/or differentiate and be diluted out by higher affinity T cell clones during the course of immune responses. We envision that T cells with high affinity TCRs to foreign antigens will also escape CD8` T cell–mediated downregulation. Perhaps the net effect of homeostatic mechanisms involving both Qa-1/Qdm and Qa-1/TCR peptide is to inhibit the downregulation of these high affinity T cells in periphery. Taken together, these considerations would predict that during immune responses T cells with intermediate affinity TCRs for either self peptides or foreign peptides would be downregulated by the Qa-1-restricted CD8` T cells. T cell clones with high affinity TCRs for foreign peptides would be most likely to survive this type of down-regulation in periphery. In this regard, it is known that during the evolution of an immune response and following repeated antigenic exposure there is a change in both the antigenspecific B cell immunoglobulin and the T cell receptor primary repertoire leading to a higher affinity repertoire. In the case of B cells, the special mechanisms of somatic hypermutation permit the creation of higher affinity receptors and the preferential selection of higher affinity receptors in secondary responses is a direct consequence of antigen triggering, although T-B contact–dependent help involving CD40 may also dictate which B cell clones will survive in germinal centers. In contrast to B cells, T cells do not undergo somatic hypermutation, and therefore antigen triggered peripheral T cells retain germline-encoded VDJ segments. Nevertheless, recent studies have shown that although the TCR repertoires between naı¨ve and primarily activated T cells are not significantly different there is a


206

JIANG n CHESS

significant change in the TCR Vb repertoire following secondary immunization (4, 187, 188). Molecular correlation of this change in TCR repertoire between primary and secondary responses has recently been evaluated in studies assaying MHC/peptide tetrameric complex binding to specific TCRs. These studies have shown that during the secondary response there is a selection against TCRs with the highest dissociation rates for MHC/peptide binding and for TCRs that bind MHC/peptide for a long duration. These parameters generally correlate with selection for higher affinity binding (4, 187, 188). We propose that superimposed on this direct antigen (MHC/peptide)–driven selection of the TCR repertoire, fine tuning of the TCR repertoire is made by regulatory Qa-1/TCR peptide–specific CD8` T cells. As emphasized above, these regulatory CD8` T cells are induced by and regulate only those activated CD4` T cells with intermediate affinity TCRs to foreign antigens that express surface Qa-1/TCR peptide–complexes. Thus, the peripheral TCR repertoire to foreign antigens changes during the course of immune responses, in the absence of somatic hypermutation mechanisms, by direct antigen (MHC/peptide)–driven TCR selection and additionally by regulatory CD8` T cells that selectively downregulate T cells expressing intermediate affinity TCRs. ACKNOWLEDGMENTS The authors wish to thank Drs. Ned Braunstein, Harvey Cantor, Lorraine Flaherty, Jim Forman, Mark Soloski, Alan Stall, and Robert Winchester for helpful discussions, support, and reagents. The research was siupported by NIH grants AI39630, AI30675 and National Multiple Sclerosis Society grant RG2938A to HJ, and NIH grants AI 39675, Ai40226A to LC. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Mueller DL, Jenkins MK, Schwartz RH. 1989. Clonal expansion versus functional clonal inactivation: a costimulatory signalling pathway determines the outcome of T cell antigen receptor occupancy. Annu. Rev. Immunol. 7:445 2. Janeway C Jr, Bottomly K. 1994. Signals and signs for lymphocyte responses. Cell 76:275 3. Lenschow DJ, Walunas TL, Bluestone JA. 1996. CD28/B7 system of T cell costimulation. Annu. Rev. Immunol. 14:233 4. Savage PA, Boniface JJ, Davis MM.

1999. A kinetic basis for T cell receptor repertoire selection during an immune response. Immunity 10:485 5. Davis MM, Boniface JJ, Reich Z, Lyons D, Hampl J, Arden B, Chien Y-h. 1998. Ligand recognition by ab T cell receptors. Annu. Rev. Immunol. 16:523 6. Lenardo M, Chan K-M, Hornung F, McFarland H, Siegel R, Wang J, Zheng L. 1999. Mature T lymphocyte apoptosis–immune regulation in a dynamic and unpredictable antigenic environment. Annu. Rev. Immunol. 17:221


CD8` T CELL MEDIATED IMMUNOREGULATION 7. Lederman S, Yellin MJ, Krichevsky A, Belko J, Lee JJ, Chess L. 1992. Identification of a novel surface protein on activated CD4` T cells that induces contact-dependent B cell differentiation (help). J. Exp. Med. 175:1091 8. Caux C, Massacrier C, Vanbervliet B, Dubois B, Van Kooten C, Durand I, Banchereau J. 1994. Activation of human dendritic cells through CD40 crosslinking. J. Exp. Med. 180:1263 9. Klaus SJ, Pinchuk LM, Ochs HD, Law CL, Fanslow WC, Armitage RJ, Clark EA. 1994. Costimulation through CD28 enhances T cell-dependent B cell activation via CD40-CD40L interaction. J. Immunol. 152:5643 10. Durie FH, Foy TM, Masters SR, Laman JD, Noelle RJ. 1994. The role of CD40 in the regulation of humoral and cellmediated immunity. Immunol. Today 15:406 11. Koulova L, Clark EA, Shu G, Dupont B. 1991. The CD28 ligand B7/BB1 provides costimulatory signal for alloactivation of CD4` T cells. J. Exp. Med. 173:759 12. Boussiotis VA, Freeman GJ, Gribben JG, Nadler LM. 1996. The role of B7–1/B7– 2:CD28/CLTA-4 pathways in the prevention of anergy, induction of productive immunity and down-regulation of the immune response. Immunol. Rev. 153:5 13. Jenkins MK. 1994. The ups and downs of T cell costimulation. Immunity 1:443 14. Durie FH, Foy TM, Masters SR, Laman JD, Noelle RJ. 1994. The role of CD40 in the regulation of humoral and cellmediated immunity. Immunol. Today 15:406 15. Coffman RL, Mosmann TR. 1991. CD4` T-cell subsets: regulation of differentiation and function. Res. Immunol. 142:7 16. Mosmann TR, Cherwinski H, Bond MW, Giedlin MA, Coffman RL. 1986. Two types of murine helper T cell clone. I.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

207

Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol. 136:2348 Mosmann TR, Coffman RL. 1989. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7:145 Fitch FW, McKisic MD, Lancki DW, Gajewski TF. 1993. Differential regulation of murine T lymphocyte subsets. Annu. Rev. Immunol. 11:29 Seder RA, Paul WE. 1994. Acquisition of lymphokine-producing phenotype by CD4` T cells. Annu. Rev. Immunol. 12:635 Lider O, Reshef T, Beraud E, Ben-Nun A, Cohen IR. 1988. Anti-idiotypic network induced by T cell vaccination against experimental autoimmune encephalomyelitis. Science 239:181 Mor F, Cohen IR. 1993. Experimental aspects of T cell vaccination. Clin. Exp. Rheumatol. 11:S55 Jiang H, Zhang SI, Pernis B. 1992. Role of CD8` T cells in murine experimental allergic encephalomyelitis. Science 256:1213 Koh D-R, Fung-Leung W-P, Ho A, Gray D, Acha-Orbea H, Mak T-W. 1992. Less mortality but more relapses in experimental allergic encephalomyelitis in CD8-/- mice. Science 256:1210 Gaur A, Ruberti G, Haspel R, Mayer JP, Fathman CG. 1993. Requirement for CD8` cells in T cell receptor peptideinduced clonal unresponsiveness. Science 259:91. Erratum. 1993. Science 260(5106):281 Jiang H, Ware R, Stall A, Flaherty L, Chess L, Pernis B. 1995. Murine CD8` T cells that specifically delete autologous CD4` T cells expressing V-b8 TCR: a role of the Qa-1 molecule. Immunity 2:185 Sun D, Whitaker JN, Wilson DB. 1999. Regulatory T cells in experimental allergic encephalomyelitis. I. Frequency and

208

27.

28.


29.

30.

31.

32.

33.

34. 35.

36.

JIANG n CHESS specificity analysis in normal and immune rats of a T cell subset that inhibits disease. Int. Immunol. 11:307 Sun D, Whitaker JN, Wilson DB. 1999. Regulatory T cells in experimental allergic encephalomyelitis. III. Comparison of disease resistance in Lewis and Fischer 344 rats. Eur. J. Immunol. 29: 1101 Bloom BR, Modlin RL, Salgame P. 1992. Stigma variations: observations on suppressor T cells and leprosy. Annu. Rev. Immunol. 10:453 Formby B, Shao T. 1993. T cell vaccination against autoimmune diabetes in nonobese diabetic mice. Ann. Clin. Lab. Sci. 23:137 Lohse AW, Spahn TW, Wolfel T, Herkel J, Cohen IR, Meyer-zum-Buschenfelde KH. 1993. Induction of the antiergotypic response. Int. Immunol. 5:533 Sun D, Qin Y, Chluba J, Epplen JT, Wekerle H. 1988. Suppression of experimentally induced autoimmune encephalomyelitis by cytolytic T-T cell interactions. Nature 332:843 Kumar V, Sercarz E. 1993. T cell regulatory circuitry: antigen-specific and TCR-idiopeptide-specific T cell interactions in EAE. Int. Rev. Immunol. 9:287 Jiang H, Kashleva H, Xu LX, Forman J, Flaherty L, Pernis B, Braunstein NS, Chess L. 1998. T cell vaccination induces T cell receptor V-beta-specific Qa-1-restricted regulatory CD8(`) T cells. Proc. Natl. Acad. Sci. USA 95:4533 Jerne NK. 1975. Immune system. Harvey Lecture—Web of V-domains. 1976:93 Weinberger JZ, Germain RN, Ju ST, Greene MI, Benacerraf B, Dorf ME. 1979. Hapten-specific T-cell responses to 4-hydroxy-3-nitrophenyl acetyl. II. Demonstration of idiotypic determinants on suppressor T cells. J. Exp. Med. 150:761 Dietz MH, Sy MS, Benacerraf B, Nisonoff A, Greene MI, Germain RN. 1981. Antigen- and receptor-driven regulatory mechanisms. VII. H-2-restricted anti-

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

idiotypic suppressor factor from efferent suppressor T cells. J. Exp. Med. 153:450 Green DR, Flood PM, Gershon RK. 1983. Immunoregulatory T-cell pathways. Annu. Rev. Immunol. 1:439 Ehrlich P. 1900. The Croonian Lecture: on immunity. Proc. R. Soc. London 66:424 Jerne NK. 1955. The natural selection theory of antibody formation. Proc. Natl. Acad. Soc. USA 41:849 Burnet FM. 1957. A modification of Jerne’s theory of antibody production using the concept of clonal selection. Aust. J. Sci. 20:67 Gowans JL, Gesner BM, McGregor DD. 1961. The immulogic activities of lymphocytes. In Ciba Found. Study Group, ed. G Wolstenhome, M O’Connor, p. 32. London: Churchill NacKean LD, Zak SJ, Varco RL, Godd RA. 1956. The role of the thymus in antibody production: an experimental study of immune response in the thymectomized rabbit. Transplant. Bull. 4:21 Glick B, Chang TS, Jaap RG. 1956. The bursa of Fabricius and antibody production. Poultry Sci. 35:224 Cooper MD, Raymond DA, Peterson RD, South MA, Good RA. 1966. The functions of the thymus system and the bursa system in the chicken. J. Exp. Med. 123:75 Dent PB, Good RA. 1965. Absence of antibody production in the bursa of Fabricius. Nature 207:491 Miller JF, Mitchell GF. 1967. The thymus and the precursors of antigen reactive cells. Nature 216:659 Claman HN, Chaperon EA, Triplett RF. 1966. Thymus-marrow cell combinations—synergism in antyibody formation. Proc. Soc. Exp. Biol. Med. 122:1167 Miller JF, Mitchell GF. 1968. Cell to cell interaction in the immune response. I. Hemolysin-forming cells in neonatally thymectomized mice reconstituted with


49.

50.


51.

52.

53.

54.

55.

56.

57.

58.

59.

thymus or thoracic duct lymphocytes. J. Exp. Med. 128:801 Mitchell GF, Miller JF. 1968. Cell to cell interaction in the immune response. II. The source of hemolysin-forming cells in irradiated mice given bone marrow and thymus or thoracic duct lymphocytes. J. Exp. Med. 128:821 Nossal GJ, Cunningham A, Mitchell GF, Miller JF. 1968. Cell to cell interaction in the immune response. 3. Chromosomal marker analysis of single antibodyforming cells in reconstituted, irradiated, or thymectomized mice. J. Exp. Med. 128:839 Gershon RK, Kondo K. 1970. Cell interactions in the induction of tolerance: the role of thymic lymphocytes. Immunology 18:723 Gershon RK, Kondo K. 1971. Infectious immunological tolerance. Immunology 21:903 Gershon RK. 1974. T cell control of antibody production. Contemp. Top. Immunobiol. 3:1 Gershon RK. 1975. A disquisition on suppressor T cells. Transplant. Rev. 26:170 Gershon RK. 1976. The role of suppression in immunoregulation. Adv. Exp. Med. Biol. 66:557 Gershon RK, Mokyr MB, Mitchell MS. 1974. Activation of suppressor T cells by tumour cells and specific antibody. Nature 250:594 Askenase PW, Hayden BJ, Gershon RK. 1975. Augmentation of delayed-type hypersensitivity by doses of cyclophosphamide which do not affect antibody responses. J. Exp. Med. 141:697 Cantor H, Boyse EA. 1975. Functional subclasses of T lymphocytes bearing different Ly antigens. II. Cooperation between subclasses of Ly` cells in the generation of killer activity. J. Exp. Med. 141:1390 Cantor H, Boyse EA. 1975. Functional subclasses of T-lymphocytes bearing dif-

60.

61.

62.

63.

64.

65.

66.

67.

209

ferent Ly antigens. I. The generation of functionally distinct T-cell subclasses is a differentiative process independent of antigen. J. Exp. Med. 141:1376 Jandinski J, Cantor H, Tadakuma T, Peavy DL, Pierce CW. 1976. Separation of helper T cells from suppressor T cells expressing different Ly components. I. Polyclonal activation: suppressor and helper activities are inherent properties of distinct T-cell subclasses. J. Exp. Med. 143:1382 Cantor H, Shen FW, Boyse EA. 1976. Separation of helper T cells from suppressor T cells expressing different Ly components. II. Activation by antigen: after immunization, antigen-specific suppressor and helper activities are mediated by distinct T-cell subclasses. J. Exp. Med. 143:1391 Chess L, Schlossman SF. 1977. Human lymphocyte subpopulations. Adv. Immunol. 25:213 Evans RL, Breard JM, Lazarus H, Schlossman SF, Chess L. 1977. Detection, isolation, and functional characterization of two human T-cell subclasses bearing unique differentiation antigens. J. Exp. Med. 145:221 Kung P, Goldstein G, Reinherz EL, Schlossman SF. 1979. Monoclonal antibodies defining distinctive human T cell surface antigens. Science 206:347 Reinherz EL, Schlossman SF. 1980. The differentiation and function of human T lymphocytes. Cell 19:821 Dialynas DP, Quan ZS, Wall KA, Pierres A, Quintans J, Loken MR, Pierres M, Fitch FW. 1983. Characterization of the murine T cell surface molecule, designated L3T4, identified by monoclonal antibody GK1.5: similarity of L3T4 to the human Leu-3/T4 molecule. J. Immunol. 131:2445 Swain SL, Dialynas DP, Fitch FW, English M. 1984. Monoclonal antibody to L3T4 blocks the function of T cells specific for class 2 major histocompati-

210

68.

69.


70.

71.

72.

73.

74.

75.

76.

77.

JIANG n CHESS bility complex antigens. J. Immunol. 132:1118 Rao A, Allard WJ, Hogan PG, Rosenson RS, Cantor H. 1983. Alloreactive T-cell clones. Ly phenotypes predict both function and specificity for major histocompatibility complex products. Immunogenetics 17:147 Thomas Y, Rogozinski L, Chess L. 1983. Relationship between human T cell functional heterogeneity and human T cell surface molecules. Immunol. Rev. 74:113 Cantor H, Hugenberger J, McVayBoudreau L, Eardley DD, Kemp J, Shen FW, Gershon RK. 1978. Immunoregulatory circuits among T-cell sets. Identification of a subpopulation of T-helper cells that induces feedback inhibition. J. Exp. Med. 148:871 Cantor H, McVay-Boudreau L, Hugenberger J, Naidorf K, Shen FW, Gershon RK. 1978. Immunoregulatory circuits among T-cell sets. II. Physiologic role of feedback inhibition in vivo: absence in NZB mice. J. Exp. Med. 147:1116 Eardley DD, Hugenberger J, McVayBoudreau L, Shen FW, Gershon RK, Cantor H. 1978. Immunoregulatory circuits among T-cell sets. I. T-helper cells induce other T-cell sets to exert feedback inhibition. J. Exp. Med. 147:1106 Dorf ME, Benacerraf B. 1984. Suppressor cells and immunoregulation. Annu. Rev. Immunol. 2:127 Germain RN, Benacerraf B. 1980. Helper and suppressor T cell factors. Springer Semin. Immunopathol. 3:93 Benacerraf B, Greene MI, Sy MS, Dorf ME. 1982. Suppressor T cell circuits. Ann. NY Acad. Sci. 392:300 Asherson GL, Colizzi V, Zembala M. 1986. An overview of T-suppressor cell circuits. Annu. Rev. Immunol. 4:37 Pierres M, Germain RN, Dorf ME, Benacerraf B. 1978. In vivo effects of antiIa alloantisera. I. Elimination of specific suppression by in vivo administration of

78.

79.

80.

81.

82.

83.

84.

85.

antisera specific for I-J controlled determinants. J. Exp. Med. 147:656 Eardley DD, Murphy DB, Kemp JD, Shen FW, Cantor H, Gershon RK. 1980. Ly-1 inducer and Ly-1,2 acceptor T cells in the feedback suppression circuit bear an I-J-subregion controlled determinant. Immunogenetics 11:549 Green DR, Eardley DD, Kimura A, Murphy DB, Yamauchi K, Gershon RK. 1981. Immunoregulatory circuits which modulate responsiveness to suppressor cell signals: characterization of an effector cell in the contrasuppressor circuit. Eur. J. Immunol. 11:973 Yamauchi K, Taniguchi M, Green D, Gershon RK. 1982. The use of a monoclonal i-j-specific antibody to distinguish cells in the feedback suppression circuit from those in the contrasuppressor circuit. Immunogenetics 16:551 Stanton TH, Calkins CE, Jandinski J, Schendel DJ, Stutman O, Cantor H, Boyse EA. 1978. The Qa-1 antigenic system. Relation of Qa-1 phenotypes to lymphocyte sets, mitogen responses, and immune functions. J. Exp. Med. 148:963 Kronenberg M, Steinmetz M, Kobori J, Kraig E, Kapp JA, Pierce CW, Sorensen CM, Suzuki G, Tada T, Hood L. 1983. RNA transcripts for I-J polypeptides are apparently not encoded between the I-A and I-E subregions of the murine major histocompatibility complex. Proc. Natl. Acad. Sci. USA 80:5704 Steinmetz M, Minard K, Horvath S, McNicholas J, Srelinger J, Wake C, Long E, Mach B, Hood L. 1982. A molecular map of the immune response region from the major histocompatibility complex of the mouse. Nature 300:35 Stanton TH, Boyse EA. 1979. A new serologically defined locus, Qa-1, in the Tla-region of the mouse. Immunogenetics 3:525 Kastner DL, Rich RR, Chu L. 1979. Qa1-associated antigens. II. Evidence for


86.

87.


88.

89. 90.

91.

92.

93.

94.

95.

functional differentiation from H-2K and H-2D antigens. J. Immunol. 123:1239 Lalanne JL, Transy C, Guerin S, Darche S, Meulien P, Kourilsky P. 1985. Expression of class I genes in the major histocompatibility complex: identification of eight distinct mRNAs in DBA/2 mouse liver. Cell 41:469 Connolly DJ, Cotterill LA, Hederer RA, Thorpe CJ, Travers PJ, McVey JH, Dyson J, Robinson PJ. 1993. A cDNA clone encoding the mouse Qa-1a histocompatibility antigen and proposed structure of the putative peptide binding site. J. Immunol. 151:6089 Shawar SM, Vyas JM, Rodgers JR, Rich RR. 1994. Antigen presentation by major histocompatibility complex class I-B molecules. Annu. Rev. Immunol. 12:839 Moller G. 1988. Do suppressor T cells exist? Scand. J. Immunol. 27:247 Janeway CA Jr. 1988. Do suppressor T cells exist? A reply [editorial]. Scand. J. Immunol. 27:621 Bloom BR, Salgame P, Diamond B. 1992. Revisiting and revising suppressor T cells. Immunol. Today 13:131 Anderson LC, Binz H, Wigzell H. 1976. Specific unresponsiveness to transplantation antigens induced by autoimmunisation with syngeneic, antigenspecific T lymphoblasts. Nature 264:778 Binz H, Wigzell H. 1976. Specific transplantation tolerance induced by autoimmunization against the individual’s own, naturally occurring idiotypic, antigenbinding receptors. J. Exp. Med. 144:1438 Bellgrau D, Wilson DB. 1978. Immunological studies of T-cell receptors. I. Specifically induced resistance to graftversus-host disease in rats mediated by host T-cell immunity to alloreactive parental T cells. J. Exp. Med. 148:103 Bellgrau D, Wilson DB. 1979. Immunological studies of T-cell receptors. II. Limited polymorphism of idiotypic determinants on T-cell receptors specific

96.

97.

98. 99.

100.

101.

102.

103.

104.

105.

106.

211

for major histocompatibility complex alloantigens. J. Exp. Med. 149:234 Kimura H, Pickard A, Wilson DB. 1984. Analysis of T cell populations that induce and mediate specific resistance to graft-versus-host disease in rats. J. Exp. Med. 160:652 Ben-Nun A, Wekerle H, Cohen IR. 1981. Vaccination against autoimmune encephalomyelitis with T-lymphocyte line cells reactive against myelin basic protein. Nature 292:60 Cohen IR, Weiner HL. 1988. T-cell vaccination. Immunol. Today 9:332 Holoshitz J, Naparstek Y, Ben-Nun A, Cohen IR. 1983. Lines of T lymphocytes induce or vaccinate against autoimmune arthritis. Science 219:56 Maron R, Zerubavel R, Friedman A, Cohen IR. 1983. T lymphocyte line specific for thyroglobulin produces or vaccinates against autoimmune thyroiditis in mice. J. Immunol. 131:2316 Kakimoto K, Katsuki M, Hirofuji T, Iwata H, Koga T. 1988. Isolation of T cell line capable of protecting mice against collegen-induced arthritis. J. Immunol. 140:78 Elias D, Reshef T, Birk OS, van der Zee R, Walker MD, Cohen IR. 1991. Vaccination against autoimmune mouse diabetes with a T-cell epitope of the human 65-kDa heat shock protein. Proc. Natl. Acad. Sci. USA 88:3088 van Laar JM, de Vries RR, Breedveld FC. 1993. T cell vaccination in humans: the experience in rheumatoid arthritis. Clin. Exp. Rheumatol. 11:S59 Zhang J, Medaer R, Stinissen P, Hafler D, Raus J. 1993. MHC-restricted depletion of human myelin basic proteinreactive T cells by T cell vaccination. Science 261:1451 Hafler DA, Cohen I, Benjamin DS, Weiner HL. 1992. T cell vaccination in multiple sclerosis: a preliminary report. Clin. Immunol. Immunopathol. 62:307 Mor F, Cohen IR. 1995. Vaccines to pre-

212

107.

108.

109.


110.

111.

112.

113.

114.

115.

JIANG n CHESS vent and treat autoimmune diseases. Int. Arch. Allergy Immunol. 108:345 Jiang H, Yu B, Braunstein N, Chess L. 1999. Protection of EAE by T cell vaccination is mediated by TCR Vb and TH type specific regualtory CD8` T cells. In preparation Nanda NK, Sercarz E. 1996. A truncated T cell receptor repertoire reveals underlying immunogenicity of an antigenic determinant. J. Exp. Med. 184:1037 Zamvil S, Nelson P, Trotter J, Mitchell D, Knobler R, Fritz R, Steinman L. 1985. T cell clones specific for myelin basic protein induce chronic relapsing paralysis and demyelination. Nature 317:355 Urban JL, Kumar V, Kono DH, Gomez C, Horvath SJ, Clayton J, Ando DG, Sercarz EE, Hood L. 1988. Restricted use of T cell receptor V genes in murine autoimmune encephalomyelitis raises possibilities for antibody therapy. Cell 54:577 Vandenbark AA, Hashim G, Offner H. 1989. Immunization with a synthetic Tcell receptor V-region peptide protects against experimental autoimmune encephalomyelitis. Nature 341:541 Offner H, Hashim GA, Vandenbark AA. 1991. T cell receptor peptide therapy triggers autoregulation of experimental encephalomyelitis. Science 251:430 Howell MD, Winters ST, Olee T, Powell HC, Carlo DJ, Brostoff SW. 1989. Vaccination against experimental allergic encephalomyelitis with T cell receptor peptides. Science 246:668. Erratum. 1990. Science 247(4947):1167 Vandenbark AA, Hashim GA, Offner H. 1996. T cell receptor peptides in treatment of autoimmune disease: rationale and potential. J. Neurosci. Res. 43:391 Gold DP, Smith RA, Golding AB, Morgan EE, Dafashy T, Nelson J, Smith L, Diveley J, Laxer JA, Richieri SP, Carlo DJ, Brostoff SW, Wilsonc DB. 1997. Results of a phase I clinical trial of a Tcell receptor vaccine in patients with multiple sclerosis. II. Comparative anal-

116.

117.

118.

119.

120.

121.

122.

ysis of TCR utilization in CSF T-cell populations before and after vaccination with a TCRV beta 6 CDR2 peptide. J. Neuroimmunol. 76:29 Bourdette DN, Chou YK, Whitham RH, Buckner J, Kwon HJ, Nepom GT, Buenafe A, Cooper SA, Allegretta M, Hashim GA, Offner H, Vandenbark AA. 1998. Immunity to T cell receptor peptides in multiple sclerosis. III. Preferential immunogenicity of complementaritydetermining region 2 peptides from disease-associated T cell receptor BV genes. J. Immunol. 161:1034 Vandenbark AA, Chou YK, Whitham R, Bourdette DN, Offner H. 1998. Effects of vaccination with T cell receptor peptides: epitope switching to a possible disease-protective determinant of myelin basic protein that is cross-reactive with a TCR BV peptide. Immunol. Cell Biol. 76:83 Gonzalo JA, Baixeras E, GonzalezGarcia A, George-Chandy A, VanRooijen N, Martinez C, Kroemer G. 1994. Differential in vivo effects of a superantigen and an antibody targeted to the same T cell receptor. Activationinduced cell death vs passive macrophage-dependent deletion. J. Immunol. 152:1597 Kawabe Y, Ochi A. 1991. Programmed cell death and extrathymic reduction of V-b8` CD4` T cells in mice tolerant to Staphylococcus aureus enterotoxin B [see comments]. Nature 349:245 Rellahan BL, Jones LA, Kruisbeek AM, Fry AM, Matis LA. 1990. In vivo induction of anergy in peripheral V-b 8` T cells by staphylococcal enterotoxin B. J. Exp. Med. 172:1091 Takahashi S, Maecker HT, Levy R. 1989. DNA fragmentation and cell death mediated by T cell antigen receptor/CD3 complex on a leukemia T cell line. Eur. J. Immunol. 19:1911 Lenardo MJ. 1991. Interleukin-2 pro-


123.

124.


125.

126.

127.

128.

129.

130.

131.

grams mouse alpha beta T lymphocytes for apoptosis. Nature 353:858 Boehme SA, Lenardo MJ. 1993. Propriocidal apoptosis of mature T lymphocytes occurs at S phase of the cell cycle. Eur. J. Immunol. 23:1552 Stallcup KC, Springer TA, Mescher MF. 1981. Characterization of an anti-H-2 monoclonal antibody and its use in largescale antigen purification. J. Immunol. 127:923 Bouwer HGA, Seaman MS, Forman J, Hinrichs DJ. 1997. MHC class Ibrestricted cells contribute to antilisterial immunity: evidence for Qa-1b as a key restricing element for Listeria-specific CTLs. J. Immunol. 159:2795 Aldrich CJ, DeCloux A, Woods AS, Cotter RJ, Soloski MJ, Forman J. 1994. Identification of a Tap-dependent leader peptide recognized by alloreactive T cells specific for a class Ib antigen. Cell 79:649 Braunstein NS, Jiang H, Kashleva H, Yu B, Tau G, Ruiz E, Winchester R, Chess L. 1999. Differential recognition of autoreactive, MBP-specific Th1 and Th2 clones by Qa-1-restricted CD8` T cells. Submitted Waisman A, Ruiz PJ, Hirschberg DL, Gelman A, Oksenberg JA, Brocke S, Mor F, Cohen IR, Steinman L. 1996. Suppressive vaccination with DNA encoding a variable region gene of the Tcell receptor prevents autoimmune encephalomyelitis and activates Th2 immunity. Nat. Med. 2:899 Lamb JR, Feldmann M. 1982. A human suppressor T cell clone which recognizes an autologous helper T cell clone. Nature 300:456 Naor D, Essery G, Tarcic N, Kahan M, Feldmann M. 1991. Regulatory interactions among autologous T cell clones. Human bifunctional T cell clones regulate the activity of an autologous T cell clone. Ann. NY Acad. Sci. 636:135 Naor D, Essery G, Tarcic N, Kahan M,

132.

133.

134.

135.

136.

137.

138.

139.

213

Feldmann M. 1991. Human bifunctional T cell clones regulate the activity of an autologous T cell clone. Ann. NY Acad. Sci. 636:135 Koide J, Engleman EG. 1990. Differences in surface phenotype and nechanism of action between alloantigen-specific CD8` cytotoxic and suppressor T cell clones. J. Immunol. 144:32 Ware R, Jiang H, Braunstein N, Kent J, Wiener E, Pernis B, Chess L. 1995. Human CD8` T lymphocyte clones specific for T cell receptor Vb families expressed on autologous CD4` T cells. Immunity 2:177 Kumar V, Sercarz EE. 1993. The involvement of T cell receptor peptidespecific regulatory CD4` T cells in recovery from antigen-induced autoimmune disease. J. Exp. Med. 178:909 Kumar V, Tabibiazar R, Geysen HM, Sercarz EE. 1995. Immunodominant framework region 3 peptide from TCR Vb8.2 chain controls murine experimental autoimmune encephalomyelitis. J. Immunol. 154:1941 Kumar V, Stellrecht K, Sercarz EE. 1996. Inactivation of T cell receptor peptidespecific CD4 regulatory T cells induces chronic experimental autoimmune encephalomyelitis (EAE). J. Exp. Med. 184:1609 Kumar V, Coulsell E, Ober B, Hubbard G, Sercarz E, Ward ES. 1997. Recombinant T cell receptor molecules can prevent and reverse experimental autoimmune encephalomyelitis: dose effects and involvement of both CD4 and CD8 T cells. J. Immunol. 159:5150 Kumar V, Sercarz E. 1996. Dysregulation of potentially pathogenic self reactivity is crucial for the manifestation of clinical autoimmunity. J Neurosci. Res. 45:334 Schoenberger SP, Toes REM, van der Voort EI, Offringa R, Melief CJM. 1998. T-cell help for cytotoxic T lymphocytes

214

140.

141.


142.

143.

144.

145.

146.

147.

148.

JIANG n CHESS is mediated by CD40-CD40L interactions. Nature 393:480 Ridge JP, Di Rosa F, Matzinger P. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4` Thelper and a T-killer cell. Nature 393:474 Bennett SR, Carbone FR, Karamalis F, Flavell RA, Miller JF, Heath WR. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393:478 Lafaille JJ, Nagashima K, Katsuki M, Tonegawa S. 1994. High incidence of spontaneous autoimmune encephalomyelitis in immunodeficient anti-myelin basic protein T cell receptor transgenic mice. Cell 78:399 Lafaille JJ. 1998. The role of helper T cell subsets in autoimmune diseases. Cytokine Growth Factor Rev. 9:139 Van de Keere F, Tonegawa S. 1998. CD4(`) T cells prevent spontaneous experimental autoimmune encephalomyelitis in anti-myelin basic protein T cell receptor transgenic mice. J. Exp. Med. 188:1875 Powrie F, Carlino J, Leach MW, Mauze S, Coffman RL. 1996. A critical role for transforming growth factor-beta but not interleukin 4 in the suppression of T helper type 1-mediated colitis by CD45RB (low) CD4` T cells. J. Exp. Med. 183:2669 Karpus WJ, Swanborg RH. 1989. CD4` suppressor cells differentially affect the production of IFN-gamma by effector cells of experimental autoimmune encephalomyelitis. J. Immunol. 143:3492 Karpus WJ, Swanborg RH. 1991. CD4` suppressor cells inhibit the function of effector cells of experimental autoimmune encephalomyelitis through a mechanism involving transforming growth factor-b. J. Immunol. 146:1163 Groux H, O’Garra A, Bigler M, Rouleau M, Antonenko S, de Vries JE, Roncarolo MG. 1997. A CD4` T-cell subset inhib-

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

its antigen-specific T-cell responses and prevents colitis. Nature 389:737 Chen Y, Kuchroo VK, Inobe J, Hafler DA, Weiner HL. 1994. Regulatory T cell clones induced by oral tolerance: suppression of autoimmune encephalomyelitis. Science 265:1237 Zhang J, Raus J. 1995. Clonal depletion of human myelin basic protein-reactive T-cells by T-cell vaccination. Ann. NY Acad. Sci. 756:323 Salgame P, Convit J, Bloom BR. 1991. Immunological suppression by human CD8` T cells is receptor dependent and HLA-DQ restricted. Proc. Natl. Acad. Sci. USA 88:2598 Young JW, Steinman RM. 1996. The hematopoietic development of dendritic cells: a distinct pathway for myeloid differentiation. Stem Cells 14:376 Thomas R, Davis LS, Lipsky PE. 1994. Rheumatoid synovium is enriched in mature antigen-presenting dendritic cells. J. Immunol. 152:2613 Tewarson S, Zaleska-Rutczynska Z, Figueroa F, Klein J. 1983. Polymorphism of Qa and Tla loci of the mouse. Tissue Antigens 22:204 Landolfi NF, Rich RR, Cook RG. 1985. The Qa-1 alloantigens. III. Biochemical analysis of the structure and extent of polymorphism of the Qa-1 allelic products. J. Immunol. 135:1264 Aldrich CJ, Rodgers JR, Rich RR. 1988. Regulation of Qa-1 expression and determinant modification by an H-2D- linked gene, Qdm. Immunogenetics 28:334 Aldrich CJ, Waltrip R, Hermel E, Attaya M, Lindahl KF, Monaco JJ, Forman J. 1992. T cell recognition of QA-1b antigens on cells lacking a functional Tap-2 transporter. J. Immunol. 149:3773 Lowen LC, Aldrich CJ, Forman J. 1993. Analysis of T cell receptors specific for recognition of class IB antigens. J. Immunol. 151:6155 Kurepa Z, Hasemann CA, Forman J. 1998. Qa-1b binds conserved class I


160.

161.


162.

163.

164.

165.

166.

167.

leader peptides derived from several mammalian species. J. Exp. Med. 188: 973 Cotterill LA, Stauss HJ, Millrain MM, Pappin DJ, Rahman D, Canas B, Chandler P, Stackpoole A, Simpson E, Robinson PJ, Dyson PJ. 1997. Qa-1 interaction and T cell recognition of the Qa-1 determinant modifier peptide. Eur. J. Immunol. 27:2123 Imani F, Soloski MJ. 1991. Heat shock proteins can regulate expression of the Tla region-encoded class Ib molecule Qa-1. Proc. Natl. Acad. Sci. USA 88: 10475 Chun T, Aldrich CJ, Baldeon ME, Kawczynski LV, Soloski MJ, Gaskins HR. 1998. Constitutive and regulated expression of the class IB molecule Qa-1 in pancreatic beta cells. Immunology 94:64 Bouwer HG, Bai A, Forman J, Gregory SH, Wing EJ, Barry RA, Hinrichs DJ. 1998. Listeria monocytogenes-infected hepatocytes are targets of major histocompatibility complex class Ib-restricted antilisterial cytotoxic T lymphocytes. Infect. Immun. 66:2814 Lo WF, Ong H, Metcalf ES, Soloski MJ. 1999. T cell responses to gram-negative intracellular bacterial pathogens: a role for CD8` T cells in immunity to Salmonella infection and the involvement of MHC class Ib molecules. J. Immunol. 162:5398 Seaman MS, Perarnau B, Lindahl KF, Lemonnier FA, Forman J. 1999. Response to Listeria monocytogenes in mice lacking MHC class Ia molecules. J. Immunol. 162:5429 Phillips JH, Chang CW, Mattson J, Gumperz JE, Parham P, Lanier LL. 1996. CD94 and a novel associated protein (94AP) form a NK cell receptor involved in the recognition of HLA-A, HLA-B, and HLA-C allotypes. Immunity 5:163 Chang CW, Rodriguez A, Carretero M, Lopez-Botet M, Phillips JH, Lanier LL. 1995. Molecular characterization of

168.

169.

170.

171.

172.

173.

174.

175.

215

human CD94: a type II membrane glycoprotein related to the C-type lectin superfamily. Eur. J. Immunol. 25:2433 Lazetic S, Chang C, Houchins JP, Lanier LL, Phillips JH. 1996. Human natural killer cell receptors involved in MHC class I recognition are disulfide-linked heterodimers of CD94 and NKG2 subunits. J. Immunol. 157:4741 Plougastel B, Trowsdale J. 1997. Cloning of NKG2-F, a new member of the NKG2 family of human natural killer cell receptor genes. Eur. J. Immunol. 27:2835 Vance RE, Kraft JR, Altman JD, Jensen PE, Raulet DH. 1998. Mouse CD94/ NKG2A is a natural killer cell receptor for the nonclassical major histocompatibility complex (MHC) class I molecule Qa-1(b). J. Exp. Med. 188:1841 Braud VM, Allan DS, O’Callaghan CA, Soderstrom K, D’Andrea A, Ogg GS, Lazetic S, Young N, Bell JI, Phillips JH, Lanier LL, McMichael AJ. 1998. HLAE binds to natural killer cell receptors CD94/NKG2A, B and C. Nature 391:795 Salcedo M, Bousso P, Ljunggren HG, Kourilsky P, Abastado JP. 1998. The Qa1b molecule binds to a large subpopulation of murine NK cells. Eur. J. Immunol. 28:4356 Sivakumar PV, Gunturi A, Salcedo M, Schatzle JD, Lai WC, Kurepa Z, Pitcher L, Seaman MS, Lemonnier FA, Bennett M, Forman J, Kumar V. 1999. Cutting edge: expression of functional CD94/ NKG2A inhibitory receptors on fetal NK1.1`Ly-49- cells: a possible mechanism of tolerance during NK cell development. J. Immunol. 162:6976 Lanier LL, Corliss BC, Wu J, Leong C, Phillips JH. 1998. Immunoreceptor DAP12 bearing a tyrosine-based activation motif is involved in activating NK cells. Nature 391:703 Lanier LL, Corliss B, Wu J, Phillips JH. 1998. Association of DAP12 with activating CD94/NKG2C NK cell receptors. Immunity 8:693


216

JIANG n CHESS

176. Lanier LL. 1998. NK cell receptors. Annu. Rev. Immunol. 16:359 177. Lopez-Botet M, Bell T. 1999. Natural killer cell activation and inhibition by receptors for MHC class I. Curr. Opin. Immunol. 11:301 178. Poccia F, Cipriani B, Vendetti S, Colizzi V, Poquet Y, Battistini L, Lopez-Botet M, Fournie JJ, Gougeon ML. 1997. CD94/ NKG2 inhibitory receptor complex modulates both anti-viral and anti-tumoral responses of polyclonal phosphoantigenreactive V gamma 9V delta 2 T lymphocytes. J. Immunol. 159:6009 179. Noppen C, Schaefer C, Zajac P, Schutz A, Kocher T, Kloth J, Heberer M, Colonna M, De Libero G, Spagnoli GC. 1998. C-type lectin-like receptors in peptide-specific HLA class I-restricted cytotoxic T lymphocytes: differential expression and modulation of effector functions in clones sharing identical TCR structure and epitope specificity. Eur. J. Immunol. 28:1134 180. Bertone S, Schiavetti F, Bellomo R, Vitale C, Ponte M, Moretta L, Mingari MC. 1999. Transforming growth factorbeta-induced expression of CD94/ NKG2A inhibitory receptors in human T lymphocytes. Eur. J. Immunol. 29:23 181. Galiani MD, Aguado E, Tarazona R, Romero P, Molina I, Santamaria M, Solana R, Pena J. 1999. Expression of killer inhibitory receptors on cytotoxic cells from HIV-1-infected individuals. Clin. Exp. Immunol. 115:472

182. Speiser DE, Valmori D, Rimoldi D, Pittet MJ, Lienard D, Cerundolo V, MacDonald HR, Cerottini JC, Romero P. 1999. CD28-negative cytolytic effector T cells frequently express NK receptors and are present at variable proportions in circulating lymphocytes from healthy donors and melanoma patients. Eur. J. Immunol. 29:1990 183. Bellon T, Heredia AB, Llano M, Minguela A, Rodriguez A, Lopez-Botet M, Aparicio P. 1999. Triggering of effector functions on a CD8` T cell clone upon the aggregation of an activatory CD94/ kp39 heterodimer. J. Immunol. 162:3996 184. Sercarz EE, Lehmann PV, Ametani A, Benichou G, Miller A, Moudgil K. 1993. Dominance and crypticity of T cell antigenic determinants. Annu. Rev. Immunol. 11:729 185. Lehmann PV, Forsthuber T, Miller A, Sercarz EE. 1992. Spreading of T-cell autoimmunity to cryptic determinants of an autoantigen. Nature 358:155 186. Lehmann PV, Sercarz EE, Forsthuber T, Dayan CM, Gammon G. 1993. Determinant spreading and the dynamics of the autoimmune T-cell repertoire. Immunol. Today 14:203 187. Busch DH, Pamer EG. 1999. T cell affinity maturation by selective expansion during infection. J. Exp. Med. 189:701 188. Busch DH, Pilip I, Pamer EG. 1998. Evolution of a complex T cell receptor repertoire during primary and recall bacterial infection. J. Exp. Med. 188:61


JIANG

■

CHESS C-1

Figure 1 Model of cognate interactions in the induction and function of regulatory CD8+ T cells. CD8+ T cells regulate peripheral immune responses by cognate interactions with antigen activated CD4+ T cells in a TCR specific, Qa-1 restricted manner. Antigen activated CD4+ T cells begin to express surface Qa-1 molecules associated with TCR Vβ peptides. These Qa-1/Vβ complexes expressed on the surface of CD4+ T cells trigger the differentiation of regulatory CD8+ T cells, which subsequently downregulate activated CD4+ T cells in a Qa-1 restricted, Vβ specific manner.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


THE BIOLOGY OF CHEMOKINES AND THEIR RECEPTORS Devora Rossi1 and Albert Zlotnik2 1

Pharmingen Inc., 10975 Torreyana Road, San Diego, California 92121–1111; e-mail: [email protected] 2 DNAX Research Institute, 901 California Avenue, Palo Alto, California, 94304; e-mail: [email protected]


Key Words chemokines, inflammation, trafficking, metastasis, angiogenesis Abstract During the last five years, the development of bioinformatics and EST databases has been primarily responsible for the identification of many new chemokines and chemokine receptors. The chemokine field has also received considerable attention since chemokine receptors were found to act as co-receptors for HIV infection (1). In addition, chemokines, along with adhesion molecules, are crucial during inflammatory responses for a timely recruitment of specific leukocyte subpopulations to sites of tissue damage. However, chemokines and their receptors are also important in dendritic cell maturation (2), B (3), and T (4) cell development, Th1 and Th2 responses, infections, angiogenesis, and tumor growth as well as metastasis (5). Furthermore, an increase in the number of chemokine/receptor transgenic and knock-out mice has helped to define the functions of chemokines in vivo. In this review we discuss some of the chemokines’ biological effects in vivo and in vitro, described in the last few years, and the implications of these findings when considering chemokine receptors as therapeutic targets.

INTRODUCTION In the last few years, we have witnessed the development of EST (expressed sequence tag) databases and the widespread application of bioinformatics for new gene discovery. As a result, the pace of novel gene discovery has accelerated dramatically. One of the best examples of the impact of these technologies has been in the number of members of the chemokine superfamily identified in this time. The chemokines are ideal molecules to be discovered through bioinformatics because they are small secreted molecules that exhibit very specific cysteine motifs in their amino acid sequence. Most chemokines have four characteristic cysteines, and depending on the motif displayed by the first two cysteines, they have been classified into CXC or alpha, CC or beta, C or gamma, and CX3C or delta chemokine classes. The only exception to the four Cys rule is lymphotactin (6), which has only two Cys residues. Two disulfide bonds are formed between 0732–0582/00/0410–0217$14.00

217

ROSSI n ZLOTNIK


218

the first and third Cys and between the second and fourth. Thus, lymphotactin manages to retain a functional structure with only one disulfide bond. In addition, the CXC or alpha subfamily has been divided into two groups depending on the presence of the ELR motif preceding the first cysteine: the ELR-CXC chemokines and the non-ELR-CXC chemokines. Chemokine receptors are G-protein coupled, seven-transmembrane receptors. Based on the chemokine class they bind, the receptors have been named CXCR1, 2, 3, 4, and 5 (bind CXC chemokines); CCR1 through CCR9 (bind CC chemokines); XCR1 (binds the C chemokine, Lptn); and CX3CR1 (binds the CX3C chemokine, fractalkine or neurotactin) (Table 1). Along with the accelerated rate of discovery of chemokines has come the realization that these molecules not only control hemopoietic cell migration, but also are involved in a number of other physiological and pathological processes. Figure 1 summarizes our current understanding of the involvement of chemokine biology in other areas. Also, the characterization of new chemokines has uncovered new roles of these molecules, for example, in lymphoid cell development. The mouse and human genome/EST projects have generated large sequence databases, which resulted in the identification of most of the members of the chemokine superfamily. When viewed in this light, we can discern certain trends.

TABLE 1 Summary of the known chemokine receptors and some of their known human ligands Chemokine receptors

Human chemokine ligands

CXCR1 CXCR2 CXCR3 CXCR4 CXCR5

IL-8, GCP-2 IL-8, GCP-2, Gro a, Gro b, Gro c, ENA-78, PBP MIG, IP-10, I-TAC SDF-1/PBSF BLC/BCA-1

CCR1 CCR2 CCR3 CCR4 CCR5 CCR6 CCR7 CCR8 CCR9

MIP-1a, MIP-1b, RANTES, HCC-1, 2, 3, and 4 MCP-1, MCP-2, MCP-3, MCP-4 eotaxin-1, eotaxin-2, MCP-3 TARC, MDC, MIP-1a, RANTES MIP-1a, MIP-1b, RANTES MIP-3a/LARC MIP-3b/ELC, 6Ckine/LC I-309 TECK

XCR1 CX3CR1

Lymphotactin Fractalkine/neurotactin


BIOLOGY OF CHEMOKINES AND RECEPTORS

219

Figure 1 Biological functions of chemokines and chemokine receptors.

For example, among the first ones discovered were those chemokines produced by many cellular sources. Conversely, the last ones found exhibit restricted tissue and cellular specificity. This specificity suggests organ-specific functions, as in the case of TECK described below. This suggests that other chemokines that remain to be discovered will likely exhibit limited tissue distribution and will not be identified until their cellular source is analyzed by EST production. This chapter does not aim to be a comprehensive review on chemokines; rather, we highlight those areas of chemokine biology with new developments or therapeutic potential.

ROLE OF CHEMOKINES IN ANGIOGENESIS/ ANGIOSTASIS AND METASTASIS Angiogenesis is a biological process through which blood vessels are generated. Although it is a strictly controlled and transient event during wound healing, angiogenesis is also associated with several chronic inflammatory diseases, such as psoriasis, rheumatoid arthritis (pannus formation), and idiopathic pulmonary fibrosis (7) as well as with tumor growth and metastasis (8), where neovascularization appears to be aberrantly upregulated. Solid tumor growth requires the presence of neovascularization to guarantee an adequate supply of oxygen and nutrients (9). Strieter (10) and others have proposed that an imbalance between angiogenic and angiostatic factors may be responsible for tumor growth and the development or progress of some chronic inflammatory diseases. It is well established that ELR-CXC chemokines are potent angiogenic factors, able to stimulate endothelial cell chemotaxis, while most non-ELR-CXC che-


220

ROSSI n ZLOTNIK

mokines are strong angiostatic factors, which inhibit the endothelial cell chemotaxis induced by ELR-CXC chemokines (10). In fact, CXC chemokines act as angiogenic or angiostatic factors depending solely on the presence of the ELR motif (10). ELR-CXC chemokines bind to CXCR2 and few to CXCR1, while non-ELR-CXC chemokines bind to CXCR3, CXCR4, and CXCR5 (Table 1). Several examples of diseases in which the balance between angiogenic and angiostatic chemokines is altered appear in the literature. For example, Keane et al (7) demonstrated that lung tissues from IPF (idiopathic pulmonary fibrosis) patients constitutively express more IL-8 and less IP-10 than those from healthy individuals. This suggests that a net angiogenic balance of produced factors may be established in IPF patients, which causes fibroplasia and deposition of extracellular matrix, leading to progressive fibrosis and loss of lung function. Similar observations (high ELR-CXC chemokines vs. low non-ELR-CXC chemokines) have been described for other inflammatory diseases such as chronic pancreatitis, inflammatory bowel disease (11), and psoriasis (12). In the latter case, high levels of IL-8 and low levels of thrombospondin-1 (an angiogenesis inhibitor) produced by psoriatic keratinocytes directly correlated with their angiogenic capacity, and moreover, correlated with the overexpression of its receptor, CXCR2, but not CXCR1 (13). This suggests a tight relationship between ELRCXC chemokine expression and its receptor, CXCR2, in angiogenesis. Interestingly, CXCR2 shares a high degree of homology with a G-proteincoupled receptor (ORF74) encoded by the Kaposi’s sarcoma–associated herpesvirus-8 (KSHV/HHV8) (14). Bais et al (14) have shown that conditioned media from cells stably transfected with KSHV-GPCR stimulate the growth of human umbilical vein endothelial cells (HUVEC), and they have also suggested that VEGF (vascular endothelial growth factor) might mediate this response. These results are in accordance with earlier studies done by Arvanitakis and collaborators (15), in which they show that this constitutively activated (agonistindependent) KSHV-GPCR signals through the phosphoinositideinositoltrisphosphate-protein kinase C pathway. Thus, PKC (protein kinase C) stimulates the transcription of genes with promoters containing AP-1 sites, such as angiogenic factors (e.g. IL-8 and VEGF) (16, 17). In summary, this KSHVGPCR with high CXCR2 sequence identity is promoting endothelial cell growth and angiogenesis. Therefore, it is possible that CXCR2 mediates angiogenesis in a similar fashion: in a controlled manner during normal wound repair and an abnormal upregulated mode during tumorigenesis.

METASTASIS Tumor development and progression depend heavily on the presence of angiogenic factors. This group of factors includes chemokines as well as growth factors such as VEGF, FGF (fibroblast growth factor), EGF (epidermal growth factor), PDGF (platelet-derived endothelial growth factor), angiogenin, HGF/SF



221

(hepatocyte growth factor/scatter factor), and many others not discussed in this review but having equal importance in tumor growth (for a review, see 18). Some of these angiogenic factors, i.e. the chemokines, display a pattern of expression that appears to be imbalanced during tumor development. Luan et al (19) have described an abnormally increased expression of MGSA/GRO (melanoma growth stimulatory activity/growth-related proteins) and decreased levels of IP-10 in melanoma lesions. Similar results were reported by Arenberg et al (8) in non–small cell lung cancer (NSCLC). Interestingly, two subtypes of NSCLC, adenocarcinoma and squamous cell carcinoma, behave very differently, and this behavior correlates with a different chemokine expression pattern. Squamous cell carcinomas show higher levels of IP-10, have lesser metastatic potential, greater patient survival, and less vascularization than adenocarcinomas, which are associated with a worse prognosis. These data support the model established by Strieter et al (10), in which a shift in the balance of ELR vs. non-ELR-CXC chemokine expression determines whether a tumor grows and metastasizes or regresses. Besides the role of chemokines and their receptors in angiogenesis, they also seem to be involved in the process of tumor cell migration, invasion, and metastasis. It is known that certain tumors exhibit certain patterns of metastasis (or invasion) to certain organs; in other words, tumor cells do not migrate randomly. One explanation for this phenomenon is that this specific migration of tumor cells may be determined by the chemokine receptors they express and by the chemokines expressed in the target organs. There is some evidence supporting this hypothesis. Youngs et al (20) have reported that different breast carcinoma cell lines respond differentially to distinct chemokines. Indeed, some of them were unresponsive to the chemokines tested, indicating that tumor cells are not uniform in their ability to migrate in response to chemokines. Kleeff et al (21) have shown that MIP-3a and its receptor CCR6 were expressed by all pancreatic cancer cell lines tested. However, it may be too simplistic to envision that chemokines alone determine the site/s where metastasis develop. Other molecules, such as adhesion molecules (integrin, cadherins), proteases, angiogenic factors, etc, may also be involved in the metastatic fate of tumor cells. Although the involvement of these other molecules in metastasis is beyond the scope of this review (for a review, see 22), an example for this process is given. High levels of the b2 integrin CD11b/CD18 have been detected on an adherent T lymphoma cell line that had high metastatic potential in the kidney. This cell line was very responsive to the chemokines RANTES and MCP-1, produced by a normal kidney cell line. In contrast, the nonadherent parental cell line did not migrate in response to these chemokines, even though both cell lines were able to bind to them. Interestingly, the nonadherent cell line expressed lower levels of CD11b/CD18, which explains its inability to migrate and metastasize in response to RANTES and MCP-1 (23). This suggests that while locally produced chemokines (in this case, by normal kidney cells) may guide these metastatic tumor cells preferentially to the kidney, adhesion molecules expressed by these tumor cells also contribute to their final homing site.

222

ROSSI n ZLOTNIK

In conclusion, these results indicate that not all tumor cells respond equally to chemokines, and they also suggest that chemokines and chemokine receptors work in a coordinated fashion with adhesion molecules to determine the ability of tumor cells to invade and colonize other tissues.


ANGIOSTASIS Angiostasis is another potential antitumor therapeutic effect of chemokines. The expression of most non-ELR-CXC chemokines (IP-10, MIG, I-TAC, PF4) (24– 27) and the receptor they bind (CXCR3) (28) are induced by IFN-c. The angiostatic activity of these molecules seems to be associated with binding to CXCR3. Mouse 6Ckine/SLC (29, 30), a CC chemokine that can bind mouse CXCR3, is also angiostatic (28). However, the mechanism of the angiostatic phenomenon is not known. Luster et al (31) have shown that only IP-10 and PF-4 binding inhibited endothelial cell proliferation in a calcium flux–independent fashion and required the presence of a specific proteoglycan, HSPG, on the surface of the endothelial cells. Whether the HSPG or the CXCR3 is the receptor responsible for chemokine-induced angiostasis is not clear and will require further study.

How Do Cytokines Regulate the Angiostatic/Angiogenic Balance? Several cytokines affect chemokine expression (Figure 2). For example, IL-12 and IL-18, cytokines that induce IFN-c production, synergistically induce tumor regression by inhibiting angiogenesis (32). Moreover, IFN-c is known to be angiostatic, not only because it induces the expression of non-ELR-CXC chemokines, but also because it suppresses the expression of angiogenic CXC chemokines (33, 34). In addition, IFNs are known to inhibit wound healing (35, 36), probably due to their inhibitory effects on endothelial cells (37). Another important cytokine that can affect angiogenesis is IL-10. This cytokine downregulates the expression of MHC II on antigen-presenting cells, which results in the inhibition of proinflammatory cytokine production, such as IFN-c by Th1 cells (38–40). Hence, IL-10 may induce angiogenesis by suppressing IFNc production by T cells and, indirectly, the expression of angiostatic CXC chemokines (41). However, there is evidence that IL-10 may suppress tumor growth by inhibiting angiogenesis. Huang et al (42) have shown that IL-10 blocks macrophage-derived angiogenic factors (e.g. VEGF) and, therefore, tumor growth and metastasis. In summary, an understanding of how the cytokine network regulates this delicate balance of angiostatic and angiogenic chemokines is a promising approach for the control of chronic inflammatory diseases and neoplasias.



223

Figure 2 Biological processes and molecules involved in the regulation of angiogenesis.

CHEMOKINES AS THERAPEUTIC AGENTS Chemokines are likely to be primarily responsible for the cell infiltration observed in many disease states. The presence of proinflammatory chemokines may not be beneficial in a chronic inflammatory disease, but it is desirable in diseases where the immune response needs to be promoted, (e.g. cancer). In theory, any chemokine capable of inducing the migration of T, NK cells, dendritic cells, and/or macrophages could promote the regression or even eradication of a tumor mass by boosting the immune response against the tumor. Therefore, several chemokines are now being studied for their potential use as adjuvants in antitumor immune responses. Recent findings have demonstrated that IP-10, MIG, and Lptn have antitumor activity, as well as MCP-1, MCP-3, TCA-3 and others (see below, Antitumor Activity).

224

ROSSI n ZLOTNIK

Conversely, a chronic deleterious immune response can be reduced by interfering with the proinflammatory actions of chemokines. Based on this idea, three types of antagonists have received attention: (a) small molecule inhibitors of chemokine receptors, (b) modified chemokines or N-terminal peptides, and (c) neutralizing monoclonal antibodies against chemokines or their receptors (see below, Chemokines in Infectious and Inflammatory Diseases).


AntiTumor Activity: Chemokines as Agonists In a mouse syngeneic tumor model, Lptn (lymphotactin) has shown antitumor activity only when combined with IL-2 (43). The most likely explanation for this effect is that Lptn induces T and NK cell infiltration (44) to the tumor site, while IL-2 expands the T cell clones upon TCR activation, enhancing a specific immune response. Similar results were obtained by Emtage and collaborators combining Lptn with IL-2 or IL-12 delivered using adenoviral vectors (45). These methods attempt to increase the probability of encounter between T lymphocytes and malignant cells to induce a specific antitumor response. Another approach using chemokines in cancer immunotherapy is DC (dendritic cell)–based vaccines, which favor DC antigen presentation and induce Agspecific CTL responses. For example, Cao et al (46) transduced DC with an adenovirus expressing Lptn and then pulsed the cells with tumor peptides. The modified DC were more effective than uninfected pulsed peptide-DC in reducing metastasis in the 3LL tumor model. These results illustrate the potential use of chemokines as antitumor agents. In addition to Lptn, other chemokines display antitumor activities. IP-10 and MIG are potent chemoattractants for Th1 cells and induce tumor infiltration with CD8` lymphocytes. Several laboratories have demonstrated that IP-10 and MIG are responsible, at least in part, for the antitumor effect of IL-12, mediated through IFN-c (47, 48). Either expression or inoculation of these chemokines or cytokines intratumorally results in tumor regression, associated with extensive vascular damage and necrosis (49–51). Evidently, both the inhibition of angiogenesis and the increased T cell recruitment (52, 31) are a result of the concerted actions of IP-10 and MIG through IL-12, IFN-c, IL-18 (32), and possibly other factors yet to be determined. Other examples of antitumor activity by chemokine gene transfer are MCP-3 (53), MIP-1a (54), MCP-1 (55), and TCA-3 (56). In most cases, augmented leukocyte infiltration occurred in the perivascular areas, peritumoral tissue, or within the tumor that led to tumor necrosis. Various approaches have been used to deliver chemokines in vivo. Some examples, chemokine-transfected tumors (54) or dendritic cells (46), protein injection, and adenoviral vectors, have been mentioned here. Although gene transfer technology is still in its early stages, it has large therapeutic potential, particularly for chemokine delivery. We conclude that inducing chemokine expression in tumors


225

and inhibiting it in chronic inflammatory diseases may restore the normal cytokine and chemokine balance and reverse the disease process.


Chemokines in Infectious and Inflammatory Diseases Chronic inflammatory diseases are characterized by the presence of cell infiltrates, and chemokines are likely to be involved in this phenomenon. The most common way to resolve, or at least ameliorate, this type of disease is to decrease the inflammatory response. Traditionally, this has been accomplished by administering corticosteroids, cyclosporin A, and similar drugs. But due to the success of small molecule inhibitors of GPCRs (G protein–coupled receptors) in the treatment of various diseases, and the fact that chemokine receptors are GPCRs expressed on hematopoietic cells, the pharmaceutical and biotechnology industries have demonstrated particular interest in the production of small-molecule inhibitors of chemokine receptors. In view of the critical role of CCR5 and CXCR4 in HIV infection (1), the search for and development of antagonists against these two chemokine receptors have been the focus of interest of many laboratories. A successful example of antagonists is NSC 651016, a distamycin analog that inhibits HIV-1 replication by downregulating CCR5 and CXCR4 expression (57). Unfortunately, some HIV strains (HIV-2) appear to also use other chemokine receptors as co-receptors for cell entry, such as CCR2b, CCR3, GPR15/BOB (58), US28 (59), and others. Still, the most relevant co-receptor in HIV infection seems to be CCR5, since individuals with defective CCR5 alleles exhibit resistance to HIV-1 infection (60–62). However, the discovery and development of pharmacological antagonists is a long and expensive process. For this reason, other avenues are being explored. For example, modified chemokines and N-terminal peptides can be engineered to allow them to retain binding specificity and affinity to a receptor while blocking it. Several examples have been described in the literature. Plater-Zyberk et al (63) have shown reduced incidence of arthritis in DBA/1 mice treated with the modified chemokine MetRANTES. A similar antagonist, (AOP)-RANTES (amino oxypentane), inhibits HIV-1 infectivity in macrophages and lymphocytes (64). However, the success of modified chemokines or N-terminal peptides as antagonists depends mostly on their capacity to fully occupy the chemokine receptor/ s at nanomolar concentrations, competing with the natural ligand/s binding and thus blocking signaling. One of the advantages of using a modified ligand is that most of the receptors used by that ligand can be blocked or partially blocked by a single antagonist (65). Finally, a third type of antagonist is monoclonal antibodies against chemokines or their receptors. Eosinophils and Th2 lymphocytes are main players in allergic responses. Both cell types express CCR3 (66, 67) and respond through this receptor to several chemokines (RANTES, eotaxin, MCP-4, MCP-3, etc), which makes it a good target for development of antagonists. Heath et al (68) have made a

ROSSI n ZLOTNIK

226

monoclonal antibody against CCR3 that blocked chemotaxis and calcium flux induced by all CCR3 ligands. Furthermore, monoclonal antibodies have been proven to work successfully as antagonists in vivo, for example, the anti-TNF-a antibody for the treatment of rheumatoid arthritis (69, 70). This effect may be due, at least in part, to the inhibition of chemokine production brought about by the neutralization of excess TNF-a in the joints of these patients. In sum, all these types of antagonists are promising candidates to block inflammation and/or HIV infection.


CHEMOKINES IN ORGANOGENESIS AND LYMPHOCYTE TRAFFICKING The proinflammatory activities of chemokines during immune responses have been extensively described in the literature (71). However, in recent years we have learned that chemokines are also involved in noninflammatory functions such as the regulation of lymphocyte trafficking, T and B cell development, and particularly cell compartmentalization within lymphoid tissues (T cell vs. B cell areas). The redundancy of chemokine expression plus the promiscuity of ligandreceptor binding have made it very difficult to understand how the cell migration process works in vivo. Nevertheless, recent findings suggest that chemokine/ receptor expression changes in cells during development and between organs (Figure 3), as do the chemotactic responses.

THYMUS Various chemokines are expressed in the thymus at significant levels: MDC, TARC, 6Ckine/SLC/TCA-4, TECK, SDF-1, and MIP-3b/ELC. Some thymusexpressed chemokines attract mature lymphocytes. For example, MDC (72) attracts activated T cells among other cells (73); TARC attracts peripheral blood T cells (74); TCA-4/6Ckine attracts mature T cells (75). Therefore, this group of chemokines may coordinate the trafficking of mature T cells into the thymus. However, chemokines such as TECK, MIP-3b, and SDF-1 are also able to attract immature T cells and thymocyte subsets with varying efficacy (76). Thus, the specific localization of thymocyte subsets within the thymus (cortex or medulla) may possibly be explained by the differential chemotactic response of these subsets. Mouse TECK is expressed by thymic dendritic cells and induces the migration of thymocytes, but not mature peripheral T cells (4). Recently, GPR9–6/CCR9, a thymus-expressed GPCR, has been reported to be the receptor for TECK (77) and was specifically detected in immature and mature thymocytes. Based on these



227

Figure 3 Chemokine expression in lymphoid organs.

data, TECK could attract T cell progenitors from bone marrow to fetal thymus and retain them until T cell development is completed. Probably, once T cells are ready to move to the periphery, they may lose GPR9–6 expression and be able to leave the thymus microenvironment. It is possible that this receptor may be downregulated in peripheral T lymphocytes, since Vicari et al (4) did not detect any chemotactic response with these cells, nor did Wilkinson et al (78). In contrast, there is evidence that TECK may not be responsible for the migration of T cell progenitors to the thymus, since Wilkinson et al (78) have shown that antibodies to TECK did not prevent thymus recolonization by T cell precursors. GPR9–6-deficient mice may help resolve this paradox. SDF-1 and MIP-3b are also chemoattractants for thymocytes. Kim et al (76) have done an extensive characterization of the chemotactic responses of thymocyte subsets to these two chemokines. Thymocytes can be divided into four subpopulations based on their expression of CD4 and CD8 that comprise the immature double negative (DN) CD41 CD81, and double positive (DP) CD4` CD8` subsets and the mature CD4` and CD8` single positive cells (SP). Interestingly, immature DN and DP thymocytes are more responsive to SDF-1 and are located in the thymic cortex, while mature CD4` SP and CD8` SP are more responsive to MIP-3b and are located in the medulla, from where mature T cells migrate to the peripheral blood. These results suggest that chemokines may control the compartmentalization seen within lymphoid organs and coordinate, along with other molecules, T cell development.

228

ROSSI n ZLOTNIK

It is very likely that chemokines will be found to regulate intrathymic T cell migration in a very precise manner. As developing T cells achieve specific stages of development (79), they may change their chemokine receptor expression and become responsive to other chemokines that will induce their migration to new intrathymic locations where their subsequent development may proceed. These observations, however, raise the question of whether chemokines merely control intrathymic cell migration or also have direct differentiation effects on the developing thymocytes.


BONE MARROW SDF-1 is a chemoattractant for T cells, B cells, and megakaryocytes (80) that selectively binds to CXCR4. SDF-1 is a growth factor for progenitor B cells, and chemotactic for human pre- and pro-B cell lines, as well as mature B cells (81, 82). Bone marrow (BM) stromal cells produce SDF-1, which attracts B cell progenitors and places them in contact with the stromal cells. These cells release growth and differentiation factors that are necessary for B cell maturation (81). Also, SDF-1 induces bone marrow colonization by hematopoietic precursor cells (CD34` cells) during embryogenesis (83). These data suggest that SDF-1 and CXCR4 expression are essential for B cell development and maturation in the BM. This has been confirmed by analysis of SDF-1- and CXCR4-deficient mice, which show defects in the hematopoietic, nervous, and vascular systems (84– 86). Ma et al (87) and Kawabata et al (88) have reported the phenotype of the CXCR4-deficient mouse. In these mice, B cell precursors are decreased in fetal liver and BM and abnormally increased in blood. Granulocytes are also decreased in BM and elevated, but immature, in blood. Evidently, B cell and granulocytic precursors are released prematurely into the periphery. Although T cells express CXCR4, their development does not seem to be impaired in the CXCR4-deficient mice. Interestingly, these mice show normal T and B cell localization in secondary lymphoid organs. In summary, CXCR4 may regulate B lymphopoiesis and myelopoiesis by retaining the precursors within the fetal liver and BM microenvironment. MIP-3b is also expressed by bone marrow stromal cells, but only upon LPS (lipopolysaccharide) stimulation. It specifically attracts macrophage progenitors (76) and may increase the number of macrophage progenitors in the BM during inflammatory responses. In this way, more macrophages are generated and exported to the blood during inflammation. Other chemokines expressed in the BM are HCC-1 (89), MCP-2 (90), and MIP-1a (91). However, the B cell–specific chemoattractant BCA-1/BLC is not expressed in BM (92). The role that each chemokine plays in the BM microenvironment is still under investigation. Nevertheless, some chemokines have shown inhibitory effects on the proliferation of hematopoietic progenitors. Some


229

examples are MPIF-1, MPIF-2 (93), MIP-1a (94), PF4, and IL-8 (95). Some of these antiproliferative chemokines may be useful therapeutically by preventing cellular damage during anticancer therapies such as chemotherapy or radiation.

SECONDARY LYMPHOID ORGANS


B Cells The B cell homing chemokine BLC/BCA-1 (96, 92) is expressed in follicles in the spleen, Peyer’s patches, and lymph nodes, specifically in FDC (follicular dendritic cells). Its receptor, BLR-1/CXCR5 (97) is mainly expressed on B cells, Burkitt’s lymphoma cells, and to a lesser extent, in T cells (3) and monocytes (98). Forster et al (97) demonstrated that BLR-1/CXCR5 is required for B cell migration into splenic and Peyer’s patch follicles and for germinal center formation, but not for migration into lymph node follicles. This correlates with a report by Gunn et al (96) who did not detect BLC expression in lymph node follicles. Evidently, other B cell chemoattractants may be responsible for B cell migration into lymph node follicles. One candidate is MIP-3b, which is expressed in lymph nodes and induces B, T, and dendritic cell chemotaxis (76, 99, 100). Interestingly, MIP-3b attracts these cell populations to lymph nodes that are involved in the antigen presentation process, assisting in the initiation of the immune response. Another probable candidate responsible for the B cell migration into lymph node follicles is MIP-3a, which also attracts B cells (101). Another B cell chemoattractant expressed in secondary lymphoid organs is SDF-1. Bleul et al (82) showed that naive and memory B cells are responsive to SDF-1; however, germinal center B cells are not. These cells are responsive to BCA-1, which retains them within the germinal center to undergo somatic hypermutation and affinity maturation. These data correlate very well with chemokine receptor expression, since only naive and resting B cells express CXCR5 (BCA1/BLC binds to CXCR5). Once B cells have been activated, the cells surrounding GC (germinal center/s) express SDF-1 (SDF-1 binds to CXCR4), which probably attracts them out of the GC. Although B cells express CXCR4 at all stages, GC B cells are not responsive to SDF-1a. This could be due to the downregulated surface expression of CXCR4 on GC B cells being much slower than on naive or memory B cells, when incubated with SDF-1. In addition, when GC cells are differentiated in vitro into memory B cells, they increase L-selectin, ICAM-1, LFA-3, and B7.1/CD80 expression as well as responsiveness to SDF-1a (82). This indicates that other molecules, such as adhesion molecules, can also be responsible for changes in chemokine responsiveness of cells.

T Cells SLC/6Ckine/Exodus-2 (30, 29, 102) is expressed in HEV (high endothelial venules) and in T cell areas in the spleen, Peyer’s patches, and lymph nodes, and the


230

ROSSI n ZLOTNIK

marginal zone of follicles. Indeed, Willimann et al (103) demonstrated that the cells expressing SLC/6Ckine in human lymph nodes were interdigitating dendritic cells. Although SLC/6Ckine is expressed in other nonlymphoid organs, Gunn et al showed that this is due to its expression on lymphatic endothelium (104). Probably, the role of SLC/6Ckine on lymphatic endothelium is to recruit lymphocytes from tissues to draining lymphatics. SLC/6Ckine attracts naive T cells (104), dendritic cells (100), and less efficiently, B cells (104). In addition, it induces firm adhesion of naive T cells via b2 integrin binding to ICAM-1 (104). Moreover, Nakano et al (105) have described a T cell–homing defect in plt mice ( paucity of lymph node T cell) that is related to a gene localized on mouse chromosome 4, where SLC/6Ckine gene maps. Recently, Gunn et al (106) have reported that these mice do not express SLC/6Ckine mRNA, though the sequence of SLC/6Ckine introns and exons is normal. In addition, DCs do not accumulate in the T cell zones of lymph nodes and spleen in the plt mice, which correlates with the DC-chemoattractant ability of SLC/6Ckine described by Kellerman et al (100). These data indicate that SLC/ 6Ckine mediates the homing of naive T lymphocytes and dendritic cells to secondary lymphoid organs. Interestingly, DC-CK1/PARC also attracts naive T lymphocytes. It is expressed by dendritic cells of lymph node–germinal centers and in T cell areas of secondary lymphoid organs (107, 108). One possibility is that SLC/6Ckine directs the migration of naive T cells to lymph nodes through HEV, and then DC-CK1/PARC guides them to the germinal centers. DC-CK-1/ PARC is expressed by a specific subpopulation of human DCs (107) located in close contact with T cells in germinal centers (in human tonsils, spleen, and lymph nodes). However, this human DC subset appears to be involved in maintaining the activation state of GC memory T cells and promoting T-B cell interaction (109) rather than activating naive T cells, since no naive T cells could be detected in the GCs (109). In fact, it is more likely that MDC, TARC, or MIP-3a perform this function. MDC is expressed by monocyte-derived-DC and lymph node-DCs (Langerhans cells) (72, 110) and attracts activated T lymphocytes. TARC is also produced by monocyte-derived DCs and attracts activated T cells (111). In fact, it may be the most DC-specific chemokine. MIP-3a/LARC is expressed in lymph nodes (112, 113) and attracts B cells, activated T cells (114), and CD34`-derived DCs (115). In summary, most of the chemokines previously mentioned attract T, B, and dendritic cells, cells that are involved in the antigen presentation process. This suggests that the seeming redundancy of chemokine expression in lymphoid organs may serve the purpose of attracting different cell subsets into different microenvironments and, consequently, may regulate the antigen presentation process. In addition, this redundancy can also be explained by the fact that chemokines can work sequentially, guiding each cell subpopulation to a specific site (116). Further investigation will be needed to clarify the individual role of each of these chemokines in lymphocyte trafficking and antigen presentation.


231


HOMEOSTATIC VERSUS INFLAMMATORY CHEMOKINES When discussing the function of chemokines, it is possible to make a distinction between those whose expression suggests more of a homeostatic function than a regulatory function in inflammation. Many chemokines show an expression pattern that strongly suggests a role in inflammation. They are typically induced in either monocytes or macrophages or in epithelial, endothelial, or fibroblastic cells by proinflammatory cytokines (IL-1, TNF-a, or IFN-c) or stimuli (LPS). This is in fact the most common perceived role for chemokines, a proinflammatory function, frequently associated with a Th1 cytokine expression profile (IFN-c, IL-2, IL-12) and thus with a Th1 cell infiltrate at the inflammation site. However, not all chemokines fit this pattern. For example, other chemokines [e.g. C10 (117), DC-CK1/AMAC-1/PARC (118)] are specifically induced by Th2 cytokines (IL4, IL-10, IL-13) in monocytes or other cells. We recently reported that a new member of the HCC subfamily of chemokines, HCC-4 (119), is induced by IL10 in monocytes (IL-10 is generally considered an anti-inflammatory cytokine). The latter observations raise the possibility that the production of certain chemokines will be associated with Th1 responses (proinflammatory) while others will be associated with Th2 responses. Recently, it has been reported that Th1 cells produce more chemokines than Th2 (120). It is therefore very likely that there will be a segregation of Th1 vs Th2 chemokines depending on the nature of the developing immune responses. This is an area that deserves further investigation. While the identification of chemokines associated with Th1 or Th2 responses is a developing story, the association of chemokine receptor expression with the Th1 or the Th2 phenotypes is well established (reviewed by Lanzavecchia et al in this volume). Several chemokine receptors are associated with the Th1 phenotype (including CXCR3 and CCR5), while others are associated with the Th2 phenotype (CCR3, CCR4, and CCR8) (for a review, see 121). However, some receptors are associated more than others. For example, CCR8 expression is far more abundant at the mRNA level in activated Th2 cells than CCR3. It is likely that CCR3 may actually define a subset within the Th2 population of cells. The expression of chemokine receptors in T helper cells depends on the state of activation of these cells. For example, CXCR3 is present in resting Th1, while CCR8 is only present in activated Th2. Finally, the issue arises of whether the expression of these receptors is only for the purposes of directing the migration of developing Th1 or Th2 cells to the appropriate sites where these responses are occurring, or whether the corresponding ligands may actually have effects on the differentiation of Th1 or Th2 cells. In support of the latter possibility, it has been reported that MIP-1a (a CCR5 ligand) promotes differentiation toward a Th1 phenotype (122). In contrast, there are several chemokines whose expression strongly suggests a homeostatic role. Typically, this is due to the fact that they are expressed in

232

ROSSI n ZLOTNIK


normal organs or tissues in the apparent absence of inflammatory stimuli, and/or they may be produced by cells that do not typically participate as active inducers of inflammation. For example, the chemokine HCC-1 is present in large concentrations in serum and is constitutively expressed in several organs (89). Several members of this family (HCC-2 and HCC-4—also called LEC) show strong constitutive expression in various organs, suggesting a homeostatic role. In this category, we could also include many other chemokines that tend to be tissue specific, including some described earlier, such as TECK (abundantly expressed in the thymus), which is likely to have an organ-specific function (for example, involvement in T cell development). Still, other chemokines may represent a ‘‘mixed’’ category, with both homeostatic and inflammatory functions. An example of the latter is fractalkine (123), which is highly expressed in the brain and may have a homeostatic function there but is also induced by TNF-a in endothelial cells and may thus participate in inflammatory reactions.

CONCLUSION The chemokines and their receptors have received increasing attention in the last few years. Besides their role in HIV pathogenesis, it is now clear that chemokines participate intimately in many pathological conditions like inflammation and autoimmunity. They also play a very important role in normal homeostasis, including lymphoid development and migration. Some chemokines have potential therapeutic applications, mainly in cancer through their ability to attract subpopulations of lymphoid cells and also through their angiostatic effects. The nature of their receptors (seven-transmembrane G-protein-coupled receptors) also makes them compelling candidates as therapeutic targets in many areas where chemokines are involved. In addition, chemokines are likely key regulators of immune responses. In particular, the emerging picture is that of a discrete network of cells interacting through their specific production of chemokines or expressing a highly specific pattern of chemokine receptors. The emerging view of this specificity has replaced the older view of a redundant system and in turn suggests that chemokine receptors, in particular, have the potential of evolving into specific markers to identify subsets of cells involved in a given immune response. Little effort has so far been directed toward the functional characterization of subsets of immune cells defined by chemokine receptor expression. This situation will probably change when reagents become available. There is an important need for reagents for flow cytometry as well as a need for neutralizing monoclonal antibodies in order to advance this field. We should expect advances (and surprises) to come from the chemokine field. It is one of the first molecular families on which we have witnessed the impact of bioinformatics and genomics. It is therefore also valuable to learn from these


233

TABLE 2 Proposed new nomenclature for human chemokines Systemic name

Human ligand

Mouse ligand (if only the mouse ligand is known)


CXCL CXCL1 CXCL2 CXCL3 CXCL4 CXCL5 CXCL6 CXCL7 CXCL8 CXCL9 CXCL10 CXCL11 CXCL12 CXCL13 CXCL14 CXCL15

Groa/MGSA-a Grob/MGSA-b Groc PF4 ENA-78 GCP-2 NAP-2 IL-8 Mig IP-10 I-TAC SDF-1/PBSF BLC/BCA-1 BRAK/Bolekine Lungkine

XCL XCL1 XCL2

lymphotactin/SCM-1a/ATAC SCM-1b

CX3CL CX3CL1

Fractalkine/neurotactin

CCL CCL1 CCL2 CCL3 CCL4 CCL5 (CCL6) CCL7 CCL8

I-309 MCP-1 MIP-1a MIP-1b RANTES C10/MRP-1 MCP-3 MCP-2

234

ROSSI n ZLOTNIK

TABLE 2

(continued) Proposed new nomenclature for human chemokines


Systemic name (CCL9/10) CCL11 (CCL12) CCL13 CCL14 CCL15 CCL16 CCL17 CCL18 CCL19 CCL20 CCL21 CCL22 CCL23 CCL24 CCL25 CCL26 CCL27

Human ligand

Mouse ligand (if only the mouse ligand is known) MRP-2/CCF18/MIP-1c

Eotaxin MCP-5 MCP-4 HCC-1/HCC-3 HCC-2/leukotactin HCC-4/LEC TARC DC-CK1/PARC/AMAC-1 MIP-3b/ELC/exodus-3 MIP-3a/LARC/exodus-1 6Ckine/SLC/exodus-2 MDC/STCP-1/ABCD-1 MPIF-1 MPIF-2/Eotaxin-2 TECK SCYA26/Eotaxin-3 (MCC)/ALP/CTACK/ESkine

developments so that in the future we will be able to apply some of the lessons learned from the chemokines to other molecular families. ACKNOWLEDGMENTS We thank Elizabeth Oldham and Lewis Lanier for their comments and suggestions, and Dovie Wyler for reviewing the manuscript. DNAX Research Institute is supported by Schering-Plough Corporation. Note Added in Proof: Proposed Nomenclature At the 1998 Gordon Conference on Chemokines, a Committee headed by Dr. Osamu Yoshie (Kinki University, Japan) was formed to study options for a standardized new chemokine ligand nomenclature. It became clear that with the proliferation of new chemokines, there was a problem with their nomenclature. Often, various groups published the same molecule with different names, leading to confusion in the literature. Dr. Yoshie presented his recommendation at the 1999 Keystone Symposium on Chemokines; it was approved by the audience



235

composed of chemokine researchers from most leading laboratories worldwide. Table 2 here shows the recommendation for a new systematic nomenclature for human chemokines. It is based on the fact that the genes for chemokines have already received a standardized designation. For example, the CC gene family is called SCYA (small cytokine-A family) followed by a number. The proposal uses the same numbering system already in use for chemokine genes, but replaces the SCY abbreviation with CXCL for the CXC family, CCL for the CC family, XC for the lymphotactin, and CX3CL for fractalkine. In the case of the CC chemokines, there are some spaces for ligands that have been described in the mouse, but not yet in the human. However, it offers the advantage that the number representing a particular gene will be the same as its ligand, which is represented by the letter ‘‘L’’ after the family (CXC, CC, XC, CX3C). Thus, this nomenclature proposal is analogous to the current one in use for receptors and offers the added advantage that each ligand will be immediately recognizable as belonging to a particular subclass. But the main advantage is that it eliminates nomenclature ambiguities when referring to each chemokine. It is recommended that future articles use this nomenclature. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Feng Y, Broder CC, Kennedy PE, Berger EA. 1996. HIV-1 entry cofactor: functional cDNA cloning of a seventransmembrane, G protein-coupled receptor. Science 272:872–77 2. Sozzani S, Allavena P, D’Amico G, Luini W, Bianchi G, Kataura M, Imai T, Yoshie O, Bonecchi R, Mantovani A. 1998. Differential regulation of chemokine receptors during dendritic cell maturation: a model for their trafficking properties. J. Immunol. 161:1083–86 3. Forster R, Emrich T, Kremmer E, Lipp M. 1994. Expression of the G-protein– coupled receptor BLR1 defines mature, recirculating B cells and a subset of Thelper memory cells. Blood 84:830–40 4. Vicari AP, Figueroa DJ, Hedrick JA, Foster JS, Singh KP, Menon S, Copeland NG, Gilbert DJ, Jenkins NA, Bacon KB, Zlotnik A. 1997. TECK: a novel CC chemokine specifically expressed by thymic

dendritic cells and potentially involved in T cell development. Immunity 7:291– 301 5. Strieter RM, Polverini PJ, Arenberg DA, Walz A, Opdenakker G, Van Damme J, Kunkel SL. 1995. Role of C-X-C chemokines as regulators of angiogenesis in lung cancer. J. Leukocyte Biol. 57:752– 62 6. Kelner GS, Kennedy J, Bacon KB, Kleyensteuber S, Largaespada DA, Jenkins NA, Copeland NG, Bazan JF, Moore KW, Schall TJ, Zlotnik A. 1994. Lymphotactin: a cytokine that represents a new class of chemokine. Science 266:1395–99 7. Keane MP, Arenberg DA, Lynch JP, Whyte RI, Iannettoni MD, Burdick MD, Wilke CA, Morris SB, Glass MC, DiGiovine B, Kunkel SL, Strieter RM. 1997. The CXC chemokines, IL-8 and IP-10, regulate angiogenic activity in

236

8.

9.


10.

11.

12.

13.

14.

15.

ROSSI n ZLOTNIK idiopathic pulmonary fibrosis. J. Immunol. 159:1437–43 Arenberg DA, Polverini PJ, Kunkel SL, Shanafelt A, Hesselgesser J, Horuk R, Strieter RM. 1997. The role of CXC chemokines in the regulation of angiogenesis in non-small cell lung cancer. J. Leukocyte Biol. 62:554–62 Folkman J, Cotran R. 1976. Relation of vascular proliferation to tumor growth. Int. Rev. Exp. Pathol. 16:207–48 Strieter RM, Polverini PJ, Kunkel SL, Arenberg DA, Burdick MD, Kasper J, Dzuiba J, Van Damme, J, Walz A, Marriott D, Kunkel S. 1995. The functional role of the ELR motif in CXC chemokine-mediated angiogenesis. J. Biol. Chem. 270:27348–57 Walz A, Schmutz P, Mueller C, Schnyder-Candrian S. 1997. Regulation and function of the CXC chemokine ENA-78 in monocytes and its role in disease. J. Leukocyte Biol. 62:604–11 Nickoloff BJ, Mitra RS, Varani J, Dixit VM, Polverini PJ. 1994. Aberrant production of interleukin-8 and thrombospondin-1 by psoriatic keratinocytes mediates angiogenesis. Am. J. Pathol. 144:820–28 Kulke R, Bornscheuer E, Schluter C, Bartels J, Rowert J, Sticherling M, Christophers E. 1998. The CXC receptor 2 is overexpressed in psoriatic epidermis. J. Invest. Dermatol. 110:90–94 Bais C, Santomasso B, Coso O, Arvanitakis L, Raaka EG, Gutkind JS, Asch AS, Cesarman E, Gershengorn MC, Mesri EA, Gerhengorn MC. 1998. G-proteincoupled receptor of Kaposi’s sarcomaassociated herpesvirus is a viral oncogene and angiogenesis activator. Nature 391:86–89. Erratum. 1998. Nature 392(6672):210 Arvanitakis L, Geras-Raaka E, Varma A, Gershengorn M, Cesarman E. 1997. Human herpesvirus KSHV encodes a constitutively active G-protein-coupled

16.

17.

18.

19.

20.

21.

22.

23.

24.

receptor linked to cell proliferation. Nature 385:347–50 Roebuck K, Carpenter LR, Lakshminarayanan V, Page SM, Moy JN, Thomas LL. 1999. Stimulus-specific regulation of chemokine expression involves differential activation of the redox-responsive transcription factors AP-1 and NFkappaB. J. Leukocyte Biol. 65:291–98 Kolch W, Martiny-Baron G, Kieser A, Marme D. 1995. Regulation of the expression of the VEGF/VPS and its receptors: role in tumor angiogenesis. Breast Cancer Res. Treat. 36:139–55 Leek RD, Harris AL, Lewis CE. 1994. Cytokine networks in solid human tumors: regulation of angiogenesis. J. Leukocyte Biol. 56:423–35 Luan J, Shattuck-Brandt R, Haghnegahdar H, Owen JD, Strieter R, Burdick M, Nirodi C, Beauchamp D, Johnson KN, Richmond A. 1997. Mechanism and biological significance of constitutive expression of MGSA/GRO chemokines in malignant melanoma tumor progression. J. Leukocyte Biol. 62:588–97 Youngs SJ, Ali SA, Taub DD, Rees RC. 1997. Chemokines induce migrational responses in human breast carcinoma cell lines. Int. J. Cancer 71:257–66 Kleeff J, Kusama T, Rossi DL, Ishiwata T, Maruyama H, Friess H, Buchler MW, Zlotnik A, Korc M. 1999. Detection and localization of Mip-3alpha/LARC/Exodus, a macrophage proinflammatory chemokine, and its CCR6 receptor in human pancreatic cancer. Int. J. Cancer 81:650– 57 Meyer T, Hart IR. 1998. Mechanisms of tumour metastasis. Eur. J. Cancer 34: 214–21 Wang JM, Deng X, Gong W, Su S. 1998. Chemokines and their role in tumor growth and metastasis. J. Immunol. Methods 220:1–17 Vanguri P, Farber JM. 1990. Identification of CRG-2. An interferon-inducible


25.

26.


27.

28.

29.

30.

31.

mRNA predicted to encode a murine monokine. J. Biol. Chem. 265:15049–57 Farber JM. 1993. HuMig: a new human member of the chemokine family of cytokines. Biochem. Biophys. Res. Commun. 192:223–30 Cole KE, Strick CA, Paradis TJ, Ogborne KT, Loetscher M, Gladue RP, Lin W, Boyd J, Moser B, Wood DE, Sahagan BG, Neote K. 1998. Interferon-inducible T cell alpha chemoattractant (I-TAC): a novel non- ELR CXC chemokine with potent activity on activated T cells through selective high affinity binding to CXCR3. J. Exp. Med. 187:2009–21 Gasperini S, Marchi M, Calzetti F, Laudanna C, Vicentini L, Olsen H, Murphy M, Liao F, Farber J, Cassatella MA. 1999. Gene expression and production of the monokine induced by IFN-gamma (MIG), IFN-inducible T cell alpha chemoattractant (I-TAC), and IFN- gammainducible protein-10 (IP-10) chemokines by human neutrophils. J. Immunol. 162:4928–37 Soto H, Wang W, Strieter RM, Copeland NG, Gilbert DJ, Jenkins NA, Hedrick J, Zlotnik A. 1998. The CC chemokine 6Ckine binds the CXC chemokine receptor CXCR3. Proc. Natl. Acad. Sci. USA 95:8205–10 Hedrick J, Zlotnik A. 1997. Identification and characterization of a novel beta chemokine containing six conserved cysteines. J. Immunol. 159:1589–93 Nagira M, Imai T, Yoshida R, Takagi S, Iwasaki M, Baba M, Tabira Y, Akagi J, Nomiyama H, Yoshie O. 1998. A lymphocyte-specific CC chemokine, secondary lymphoid tissue chemokine (SLC), is a highly efficient chemoattractant for B cells and activated T cells. Eur. J. Immunol. 28:1516–23 Luster AD, Greenberg SM, Leder P. 1995. The IP-10 chemokine binds to a specific cell surface heparan sulfate site shared with platelet factor 4 and inhibits

32.

33.

34.

35.

36.

37.

38.

39.

237

endothelial cell proliferation. J. Exp. Med. 182:219–31 Coughlin CM, Salhany KE, Wysocka M, Aruga E, Kurzawa H, Chang AE, Hunter CA, Fox JC, Trinchieri G, Lee WMF. 1998. Interleukin-12 and interleukin-18 synergistically induce murine tumor regression which involves inhibition of angiogenesis. J. Clin. Invest. 101:1441– 52 Gusella GL, Musso T, Bosco MC, Espinoza-Delgado I, Matsushima K, Varesio L. 1993. IL-2 up-regulates but IFNgamma suppresses IL-8 expression in human monocytes. J. Immunol. 151: 2725–32 Schnyder-Candrian S, Strieter RM, Kunkel SL, Walz A. 1995. Interferon-alpha and interferon-gamma down-regulate the production of interleukin-8 and ENA-78 in human monocytes. J. Leukocyte Biol. 57:929–35 Stout AJ, Gresser I, Thompson WD. 1993. Inhibition of wound healing in mice by local interferon alpha/beta injection. Int. J. Exp. Pathol. 74:79–85 Pober JS, Cotran RS. 1990. Cytokines and endothelial cell biology. Physiol. Rev. 70:427–51 Maier J, Morelli D, Lazzerini D, Menard S, Colnaghi MI, Balsari A. 1999. Inhibition of fibronectin-activated migration of microvascular endothelial cells by interleukin-1alpha, tumour necrosis factor alpha and interferon gamma. Cytokine 11:134–39 Howard M, O’Garra A, Ishida H, de Waal Malefyt R, de Vries J. 1992. Biological properties of interleukin 10. J. Clin. Immunol. 12:239–47 de Waal Malefyt R, Haanen J, Spits H, Roncarolo MG, te Velde A, Figdor C, Johnson K, Kastelein R, Yssel H, de Vries JE. 1991. Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigenspecific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via downregula-

238

40.

41.


42.

43.

44.

45.

46.

47.

ROSSI n ZLOTNIK tion of class II major histocompatibility complex expression. J. Exp. Med. 174:915–24 Fiorentino DF, Zlotnik A, Vieira P, Mosmann TR, Howard M, Moore KW, O’Garra A. 1991. IL-10 acts on the antigen-presenting cell to inhibit cytokine production by Th1 cells. J. Immunol. 146:3444–51 Tebo JM, Kim HS, Gao J, Armstrong DA, Hamilton TA. 1998. Interleukin-10 suppresses IP-10 gene transcription by inhibiting the production of class I interferon. Blood 92:4742–49 Huang S, Xie K, Bucana CD, Ullrich SE, Bar-Eli M. 1996. Interleukin 10 suppresses tumor growth and metastasis of human melanoma cells: potential inhibition of angiogenesis. Clin. Cancer Res. 2:1969–79 Dilloo D, Bacon K, Holden W, Zhong WY, Burdach S, Zlotnik A, Brenner M. 1996. Combined chemokine and cytokine gene transfer enhances antitumor immunity. Nat. Med. 2:1090–95 Hedrick JA, Saylor V, Figueroa D, Mizoue L, Xu Y, Menon S, Abrams J, Handel T, Zlotnik A. 1997. Lymphotactin is produced by NK cells and attracts both NK cells and T cells in vivo. J. Immunol. 158:1533–40 Emtage PC, Wan Y, Hitt M, Graham FL, Muller WJ, Zlotnik A, Gauldie J. 1999. Adenoviral vectors expressing lymphotactin and interleukin 2 or lymphotactin and interleukin 12 synergize to facilitate tumor regression in murine breast cancer models. Hum. Gene Ther. 10:697–709 Cao XT, Zhang WP, He L, Xie ZF, Ma SH, Tao Q, Yu YZ, Hamada H, Wang JL. 1998. Lymphotactin gene-modified bone marrow dendritic cells act as more potent adjuvants for peptide delivery to induce specific antitumor immunity. J. Immunol. 161:6238–44 Kanegane C, Sgadari C, Kanegane H, Teruya-Feldstein J, Yao L, Gupta G, Farber JM, Liao F, Liu L, Tosato G. 1998.

48.

49.

50.

51.

52.

53.

54.

Contribution of the CXC chemokines IP10 and Mig to the antitumor effects of IL-12. J. Leukocyte Biol. 64:384–92 Tannenbaum CS, Tubbs R, Armstrong D, Finke JH, Bukowski RM, Hamilton TA. 1998. The CXC chemokines IP-10 and Mig are necessary for IL-12-mediated regression of the mouse RENCA tumor. J. Immunol. 161:927–32 Sgadari C, Angiolillo AL, Cherney BW, Pike SE, Farber JM, Koniaris LG, Vanguri P, Burd PR, Sheikh N, Gupta G, Teruya-Feldstein J, Tosato G. 1996. Interferon-inducible protein-10 identified as a mediator of tumor necrosis in vivo. Proc. Natl. Acad. Sci. USA 93:13791–96 Sgadari C, Farber JM, Angiolillo AL, Liao F, Teruya-Feldstein J, Burd PR, Yao L, Gupta G, Kanegane C, Tosato G. 1997. Mig, the monokine induced by interferon-gamma, promotes tumor necrosis in vivo. Blood 89:2635–43 Angiolillo AL, Sgadari C, Taub D, Liao F, Farber JM, Maheshwari S, Kleinman HK, Reaman GH, Tosato G. 1995. Human interferon-inducible protein 10 is a potent inhibitor of angiogenesis in vivo. J. Exp. Med. 182:155–62 Luster AD, Leder P. 1993. IP-10, a -CX-C- chemokine, elicits a potent thymusdependent antitumor response in vivo. J. Exp. Med. 178:1057–65 Fioretti F, Fradelizi D, Stoppacciaro A, Ramponi S, Ruco L, Minty A, Sozzani S, Garlanda C, Vecchi A, Mantovani A. 1998. Reduced tumorigenicity and augmented leukocyte infiltration after monocyte chemotactic protein-3 (MCP-3) gene transfer: perivascular accumulation of dendritic cells in peritumoral tissue and neutrophil recruitment within the tumor. J. Immunol. 161:342–46 Nakashima E, Oya A, Kubota Y, Kanada N, Matsushita R, Takeda K, Ichimura F, Kuno K, Mukaida N, Hirose K, Nakanishi I, Ujiie T, Matsushima K. 1996. A candidate for cancer gene therapy: MIP1 alpha gene transfer to an adenocarci-


55.


56.

57.

58.

59.

60.

61.

noma cell line reduced tumorigenicity and induced protective immunity in immunocompetent mice. Pharm. Res. 13:1896–1901 Hoshino Y, Hatake K, Kasahara T, Takahashi Y, Ikeda M, Tomizuka H, Ohtsuki T, Uwai M, Mukaida N, Matsushima K, et al. 1995. Monocyte chemoattractant protein-1 stimulates tumor necrosis and recruitment of macrophages into tumors in tumor-bearing nude mice: increased granulocyte and macrophage progenitors in murine bone marrow. Exp. Hematol. 23:1035–39 Laning J, Kawasaki H, Tanaka E, Luo Y, Dorf ME. 1994. Inhibition of in vivo tumor growth by the beta chemokine, TCA3. J. Immunol. 153:4625–35 Howard OM, Korte T, Tarasova NI, Grimm M, Turpin JA, Rice WG, Michejda CJ, Blumenthal R, Oppenheim JJ. 1998. Small molecule inhibitor of HIV-1 cell fusion blocks chemokine receptor- mediated function. J. Leukocyte Biol. 64:6–13 Morner A, Bjorndal A, Albert J, Kewalramani VN, Littman DR, Inoue R, Thorstensson R, Fenyo EM, Bjorling E. 1999. Primary human immunodeficiency virus type 2 (HIV-2) isolates, like HIV- 1 isolates, frequently use CCR5 but show promiscuity in coreceptor usage. J. Virol. 73:2343–49 Pleskoff O, Treboute C, Brelot A, Heveker N, Seman M, Alizon M. 1997. Identification of a chemokine receptor encoded by human cytomegalovirus as a cofactor for HIV-1 entry. Science 276:1874–78 Liu R, Paxton WA, Choe S, Ceradini D, Martin SR, Horuk R, MacDonald ME, Stuhlmann H, Koup RA, Landau NR. 1996. Homozygous defect in HIV-1 coreceptor accounts for resistance of some multiply-exposed individuals to HIV-1 infection. Cell 86:367–77 Samson M, Libert F, Doranz BJ, Rucker J, Liesnard C, Farber CM, Saragosti S,

62.

63.

64.

65.

66.

67.

68.

239

Lapoumeroulie C, Cognaux J, Forceille C, Muyldermans G, Verhofstede C, Burtonboy G, Georges M, Imai T, Rana S, Yi Y, Smyth RJ, Collman RG, Doms RW, Vassart G, Parmentier M. 1996. Resistance to HIV-1 infection in caucasian individuals bearing mutant alleles of the CCR-5 chemokine receptor gene. Nature 382:722–25 Quillent C, Oberlin E, Braun J, Rousset D, Gonzalez-Canali G, Metais P, Montagnier L, Virelizier JL, ArenzanaSeisdedos F, Beretta A. 1998. HIV-1-resistance phenotype conferred by combination of two separate inherited mutations of CCR5 gene. Lancet 351:14–18 Plater-Zyberk C, Hoogewerf AJ, Proudfoot AE, Power CA, Wells TN. 1997. Effect of a CC chemokine receptor antagonist on collagen induced arthritis in DBA/1 mice. Immunol. Lett. 57:117–20 Simmons G, Clapham PR, Picard L, Offord RE, Rosenkilde MM, Schwartz TW, Buser R, Wells TNC, Proudfoot AE. 1997. Potent inhibition of HIV-1 infectivity in macrophages and lymphocytes by a novel CCR5 antagonist. Science 276:276–79 Elsner J, Petering H, Hochstetter R, Kimmig D, Wells TN, Kapp A, Proudfoot AE. 1997. The CC chemokine antagonist Met-RANTES inhibits eosinophil effector functions through the chemokine receptors CCR1 and CCR3. Eur. J. Immunol. 27:2892–98 Daugherty BL, Siciliano SJ, DeMartino JA, Malkowitz L, Sirotina A, Springer MS. 1996. Cloning, expression, and characterization of the human eosinophil eotaxin receptor. J. Exp. Med. 183:2349– 54 Sallusto F, Mackay CR, Lanzavecchia A. 1997. Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Science 277:2005–7 Heath H, Qin S, Rao P, Wu L, LaRosa G, Kassam N, Ponath PD, Mackay CR.

240

69.

70.


71.

72.

73.

74.

75.

ROSSI n ZLOTNIK 1997. Chemokine receptor usage by human eosinophils. The importance of CCR3 demonstrated using an antagonistic monoclonal antibody. J. Clin. Invest. 99:178–84 Elliott MJ, Maini RN, Feldmann M, Long-Fox A, Charles P, Katsikis P, Brennan FM, Walker J, Bijl H, Ghrayeb J. 1993. Treatment of rheumatoid arthritis with chimeric monoclonal antibodies to tumor necrosis factor alpha. Arthritis Rheum. 36:1681–90 Feldmann M, Elliott MJ, Brennan FM, Maini RN. 1994. Use of anti-tumor necrosis factor antibodies in rheumatoid arthritis. J. Interferon Res. 14:299–300 Schluger NW, Rom WN. 1997. Early responses to infection: chemokines as mediators of inflammation. Curr. Opin. Immunol. 9:504–8 Godiska R, Chantry D, Raport CJ, Sozzani S, Allavena P, Leviten D, Mantovani A, Gray PW. 1997. Human macrophagederived chemokine (MDC), a novel chemoattractant for monocytes, monocyte-derived dendritic cells, and natural killer cells. J. Exp. Med. 185:1595–1604 Chang M, McNinch J, Elias C, Manthey C, Grosshans D, Meng T, Boone T, Andrew D. 1997. Molecular cloning and functional characterization of a novel CC chemokine, stimulated T cell chemotactic protein (STCP-1) that specifically acts on activated T lymphocytes. J. Biol. Chem. 272:25229–37 Imai T, Yoshida T, Baba M, Nishimura M, Kakizaki M, Yoshie O. 1996. Molecular cloning of a novel T cell–directed CC chemokine expressed in thymus by signal sequence trap using Epstein-Barr virus vector. J. Biol. Chem. 271:21514– 21 Tanabe S, Lu Z, Luo Y, Quackenbush EJ, Berman MA, Collins-Racie LA, Mi S, Reilly C, Lo D, Jacobs KA, Dorf ME. 1997. Identification of a new mouse betachemokine, thymus-derived chemotactic agent 4, with activity on T lymphocytes

76.

77.

78.

79.

80.

81.

82.

83.

and mesangial cells. J. Immunol. 159:5671–79 Kim CH, Pelus LM, White JR, Broxmeyer HE. 1998. Differential chemotactic behavior of developing T cells in response to thymic chemokines. Blood 91:434–43 Zaballos A, Gutierrez J, Varona R, Ardavin C, Marquez G. 1999. Cutting edge: identification of the orphan chemokine receptor GPR-9–6 as CCR9, the receptor for the chemokine TECK. J. Immunol. 162:671–75 Wilkinson B, Owen JJT, Jenkinson EJ. 1999. Factors regulating stem cell recruitment to the fetal thymus. J. Immunol. 162:873–81 Godfrey D, Zlotnik A. 1993. Control points in early T cell development. Immunol. Today 14:457–53 Hamada T, Mohle R, Hesselgesser J, Hoxie J, Nachman RL, Moore MA, Rafii S. 1998. Transendothelial migration of megakaryocytes in response to stromal cell-derived factor 1 (SDF-1) enhances platelet formation. J. Exp. Med. 188:39– 48 D’Apuzzo M, Rolink A, Loetscher M, Hoxie JA, Clark-Lewis I, Melchers F, Baggiolini M, Moser B. 1997. The chemokine SDF-1, stromal cell-derived factor 1, attracts early stage B cell precursors via the chemokine receptor CXCR4. Eur. J. Immunol. 27:1788–93 Bleul CC, Schultze JL, Springer TA. 1998. B lymphocyte chemotaxis regulated in association with microanatomic localization, differentiation state, and B cell receptor engagement. J. Exp. Med. 187:53–62 Aiuti A, Webb IJ, Bleul C, Springer T, Gutierrez-Ramos JC. 1997. The chemokine SDF-1 is a chemoattractant for human CD34` hematopoietic progenitor cells and provides a new mechanism to explain the mobilization of CD34` progenitors to peripheral blood. J. Exp. Med. 185:11–20


BIOLOGY OF CHEMOKINES AND RECEPTORS 84. Nagasawa T, Hirota S, Tachibana K, Takakura N, Nishikawa S, Kitamura Y, Yoshida N, Kikutani H, Kishimoto T. 1996. Defects of B-cell lymphopoiesis and bone-marrow myelopoiesis in mice lacking the CXC chemokine PBSF/SDF1. Nature 382:635–38 85. Ma Q, Jones D, Borghesani PR, Segal RA, Nagasawa T, Kishimoto T, Bronson RT, Springer TA. 1998. Impaired Blymphopoiesis, myelopoiesis, and derailed cerebellar neuron migration in CXCR4- and SDF-1-deficient mice. Proc. Natl. Acad. Sci. USA 95:9448–53 86. Tachibana K, Hirota S, Iizasa H, Yoshida H, Kawabata K, Kataoka Y, Kitamura Y, Matsushima K, Yoshida N, Nishikawa S, Kishimoto T, Nagasawa T. 1998. The chemokine receptor CXCR4 is essential for vascularization of the gastrointestinal tract. Nature 393:591–94 87. Ma Q, Jones D, Springer TA. 1999. The chemokine receptor CXCR4 is required for the retention of B lineage and granulocytic precursors within the bone marrow microenvironment. Immunity 10:463–71 88. Kawabata K, Ujikawa M, Egawa T, Kawamoto H, Tachibana K, Iizasa H, Katsura Y, Kishimoto T, Nagasawa T. 1999. A cell-autonomous requirement for CXCR4 in long-term lymphoid and myeloid reconstitution. Proc. Natl. Acad. Sci. USA 96:5663–67 89. Schulz-Knappe P, Magert HJ, Dewald B, Meyer M, Cetin Y, Kubbies M, Tomeczkowski J, Kirchhoff K, Raida M, Adermann K. 1996. HCC-1, a novel chemokine from human plasma. J. Exp. Med. 183:295–99 90. Van Coillie E, Froyen G, Nomiyama H, Miura R, Fiten P, Van Aelst I, Van Damme J, Opdenakker G. 1997. Human monocyte chemotactic protein-2: cDNA cloning and regulated expression of mRNA in mesenchymal cells. Biochem. Biophys. Res. Commun. 231:726–30 91. Jukita T, Nomiyama H, Ohmoto Y,

92.

93.

94.

95.

96.

97.

241

Kukita A, Shuto T, Hotokebuchi T, Sugioka Y, Miura R, Iijima T. 1997. Macrophage inflammatory protein-1 alpha (LD78) expressed in human bone marrow: its role in regulation of hematopoiesis and osteoclast recruitment. Lab. Invest. 76:399–406 Legler DF, Loetscher M, Roos RS, Clark-Lewis I, Baggiolini M, Moser B. 1998. B cell-attracting chemokine 1, a human CXC chemokine expressed in lymphoid tissues, selectively attracts B lymphocytes via BLR1/CXCR5. J. Exp. Med. 187:655–60 Patel VP, Kreider BL, Li Y, Li H, Leung K, Salcedo T, Nardelli B, Pippalla V, Gentz S, Thotakura R, Parmelee D, Gentz R, Garotta G. 1997. Molecular and functional characterization of two novel human C-C chemokines as inhibitors of two distinct classes of myeloid progenitors. J. Exp. Med. 185:1163–72 Maze R, Sherry B, Kwon BS, Cerami A, Broxmeyer HE. 1992. Myelosuppressive effects in vivo of purified recombinant murine macrophage inflammatory protein-1 alpha. J. Immunol. 149:1004–9 Broxmeyer HE, Sherry B, Cooper S, Lu L, Maze R, Beckmann MP, Cerami A, Ralph P. 1993. Comparative analysis of the human macrophage inflammatory protein family of cytokines (chemokines) on proliferation of human myeloid progenitor cells. Interacting effects involving suppression, synergistic suppression, and blocking of suppression. J. Immunol. 150:3448–58 Gunn M, Ngo VN, Ansel KM, Ekland EH, Cyster JG, Williams LT. 1998. A Bcell-homing chemokine made in lymphoid follicles activates Burkitt’s lymphoma receptor-1. Nature 391:799– 803 Forster R, Mattis AE, Kremmer E, Wolf E, Brem G, Lipp M. 1996. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs

242

98.

99.


100.

101.

102.

103.

104.

ROSSI n ZLOTNIK and specific anatomic compartments of the spleen. Cell 87:1037–47 Kaiser E, Forster R, Wolf I, Ebensperger C, Kuehl WM, Lipp M. 1993. The G protein-coupled receptor BLR1 is involved in murine B cell differentiation and is also expressed in neuronal tissues. Eur. J. Immunol. 23:2532–39 Yoshida R, Imai T, Hieshima K, Kusuda J, Baba M, Kitaura M, Nishimura M, Kakizaki M, Nomiyama H, Yoshie O. 1997. Molecular cloning of a novel human CC chemokine EBI1-ligand chemokine that is a specific functional ligand for EBI1, CCR7. J. Biol. Chem. 272:13803–9 Kellermann SA, Hudak S, Oldham ER, Liu YJ, McEvoy L. 1999. The CC chemokine receptor-7 ligands 6Ckine and macrophage inflammatory protein-3 beta are potent chemoattractants for in vitroand in vivo-derived dendritic cells. J. Immunol. 162:3859–64 Tanaka Y, Imai T, Baba M, Ishikawa I, Uehira M, Nomiyama H, Yoshie O. 1999. Selective expression of liver and activation-regulated chemokine (LARC) in intestinal epithelium in mice and humans. Eur. J. Immunol. 29:633–42 Hromas R, Kim CH, Klemsz M, Krathwohl M, Fife K, Cooper S, SchnizleinBick C, Broxmeyer HE. 1997. Isolation and characterization of Exodus-2, a novel C-C chemokine with a unique 37-amino acid carboxyl-terminal extension. J. Immunol. 159:2554–58 Willimann K, Legler DF, Loetscher M, Roos RS, Delgado MB, Clark-Lewis I, Baggiolini M, Moser B. 1998. The chemokine SLC is expressed in T cell areas of lymph nodes and mucosal lymphoid tissues and attracts activated T cells via CCR7. Eur. J. Immunol. 28:2025–34 Gunn M, Tangemann K, Tam C, Cyster JG, Rosen SD, Williams LT. 1998. A chemokine expressed in lymphoid high endothelial venules promotes the adhesion and chemotaxis of naive T lympho-

105.

106.

107.

108.

109.

110.

111.

cytes. Proc. Natl. Acad. Sci. USA 95:258–63 Nakano H, Mori S, Yonekawa H, Nariuchi H, Matsuzawa A, Kakiuchi T. 1998. A novel mutant gene involved in Tlymphocyte-specific homing into peripheral lymphoid organs on mouse chromosome 4. Blood 91:2886–95 Gunn MD, Kyuwa S, Tam C, Kakiuchi T, Matsuzawa A, Williams LT, Nakano H. 1999. Mice lacking expression of secondary lymphoid organ chemokine have defects in lymphocyte homing and dendritic cell localization. J. Exp. Med. 189:451–60 Adema GJ, Hartgers F, Verstraten R, de Vries E, Marland G, Menon S, Foster J, Xu Y, Nooyen P, McClanahan T, Bacon KB, Figdor CG. 1997. A dendritic-cellderived C-C chemokine that preferentially attracts naive T cells. Nature 387:713–17 Hieshima K, Imai T, Baba M, Shoudai K, Ishizuka K, Nakagawa T, Tsuruta J, Takeya M, Sakaki Y, Takatsuki T, Miura R, Opdenakker G, Van Damme J, Yoshie O, Nomiyama H. 1997. A novel human CC chemokine PARC that is most homologous to macrophage-inflammatory protein-1 alpha/LD78 alpha and chemotactic for T lymphocytes, but not for monocytes. J. Immunol. 159:1140–49 Grouard G, Durand I, Filgueira L, Banchereau J, Liu YJ. 1996. Dendritic cells capable of stimulating T cells in germinal centres. Nature 384:364–67 Tang HL, Cyster JG. 1999. Chemokine Up-regulation and activated T cell attraction by maturing dendritic cells. Science 284:819–22 Imai T, Nagira M, Takagi S, Kakizaki M, Nishimura M, Wang J, Gray PW, Matsushima K, Yoshie O. 1999. Selective recruitment of CCR4-bearing Th2 cells toward antigen- presenting cells by the CC chemokines thymus and activationregulated chemokine and macrophage-


112.


113.

114.

115.

116.

117.

derived chemokine. Int. Immunol. 11:81– 88 Rossi DL, Vicari AP, Franz-Bacon K, McClanahan TK, Zlotnik A. 1997. Identification through bioinformatics of two new macrophage proinflammatory human chemokines: MIP-3alpha and MIP-3beta. J. Immunol. 158:1033–36 Hieshima K, Imai T, Opdenakker G, Van Damme J, Kusuda J, Tei H, Sakaki Y, Takatsuki K, Miura R, Yoshie O, Nomiyama H. 1997. Molecular cloning of a novel human CC chemokine liver and activation- regulated chemokine (LARC) expressed in liver. Chemotactic activity for lymphocytes and gene localization on chromosome 2. J. Biol. Chem. 272:5846– 53 Baba M, Imai T, Nishimura M, Kakizaki M, Takagi S, Hieshima K, Nomiyama H, Yoshie O. 1997. Identification of CCR6, the specific receptor for a novel lymphocyte-directed CC chemokine LARC. J. Biol. Chem. 272:14893–98 Greaves DR, Wang W, Dairaghi DJ, Dieu MC, Saint-Vis B, Franz-Bacon K, Rossi D, Caux C, McClanahan T, Gordon S, Zlotnik A, Schall TJ. 1997. CCR6, a CC chemokine receptor that interacts with macrophage inflammatory protein 3alpha and is highly expressed in human dendritic cells. J. Exp. Med. 186:837–44 Foxman EF, Campbell JJ, Butcher EC. 1997. Multistep navigation and the combinatorial control of leukocyte chemotaxis. J. Cell Biol. 139:1349–60 Orlofsky A, Lin EY, Prystowsky MB.

118.

119.

120.

121.

122.

123.

243

1994. Selective induction of the beta chemokine C10 by IL-4 in mouse macrophages. J. Immunol. 152:5084–91 Kodelja V, Muller C, Politz O, Hakij N, Orfanos CE, Goerdt S. 1998. Alternative macrophage activation-associated CCchemokine-1, a novel structural homologue of macrophage inflammatory protein-1 alpha with a Th2-associated expression pattern. J. Immunol. 160: 1411–18 Hedrick JA, Helms A, Vicari A, Zlotnik A. 1998. Characterization of a novel CC chemokine, HCC-4, whose expression is increased by interleukin-10. Blood 91:4242–47 Bradley LM, Asensio VC, Schioetz LK, Harbertson J, Krahl T, Patstone G, Woolf N, Campbell IL, Sarvetnick N. 1999. Islet-specific Th1, but not Th2, cells secrete multiple chemokines and promote rapid induction of autoimmune diabetes. J. Immunol. 162:2511–20 O’Garra A, McEvoy LM, Zlotnik A. 1998. T-cell subsets: chemokine receptors guide the way. Curr. Biol. 8:R646– 49 Karpus WJ, Kennedy KJ. 1997. MIP1alpha and MCP-1 differentially regulate acute and relapsing autoimmune encephalomyelitis as well as Th1/Th2 lymphocyte differentiation. J. Leukocyte Biol. 62:681–67 Bazan JF, Bacon KB, Hardiman G, Wang W, Soo K, Rossi D, Greaves DR, Zlotnik A, Schall TJ. 1997. A new class of membrane-bound chemokine with a CX3C motif. Nature 385:640–44



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


DENDRITIC CELLS IN CANCER IMMUNOTHERAPY Lawrence Fong and Edgar G. Engleman Departments of Pathology and Medicine, Stanford University School of Medicine, Palo Alto, California 94304; e-mail: [email protected]


Key Words clinical trial, tumor immunology, T cell, vaccine Abstract The potential to harness the potency and specificity of the immune system underlies the growing interest in cancer immunotherapy. One such approach uses bone marrow–derived dendritic cells, phenotypically distinct and extremely potent antigen-presenting cells, to present tumor-associated antigens and thereby generate tumor-specific immunity. Support for this strategy comes from animal studies that have demonstrated that dendritic cells, when loaded ex vivo with tumor antigens and administered to tumor-bearing hosts, can elicit T cell–mediated tumor destruction. These observations have led to clinical trials designed to investigate the immunologic and clinical effects of antigen-loaded dendritic cells administered as a therapeutic vaccine to patients with cancer. In the design and conduct of such trials, important considerations include antigen selection, methods for introducing the antigen into MHC class I and II processing pathways, methods for isolating and activating dendritic cells, and route of administration. Although current dendritic cell–based vaccination methods are cumbersome, promising results from clinical trials in patients with malignant lymphoma, melanoma, and prostate cancer suggest that immunotherapeutic strategies that take advantage of the antigen presenting properties of dendritic cells may ultimately prove both efficacious and widely applicable to human tumors.

INTRODUCTION To date, human tumor immunotherapy has met with only limited success. Among the reasons for this have been the limited availability of tumor-associated antigens and the inability to deliver such antigens in a manner that renders them immunogenic in patients with cancer. However, recent insights into the role of dendritic cells (DC) as the pivotal antigen-presenting cells (APC) that initiate immune responses may provide the basis for generating more effective antitumor immune responses. These rare bone marrow–derived cells are uniquely capable of sensitizing naive T cells to protein antigens. The role of DC in initiating or ‘‘priming’’ immune responses to viral and bacterial antigens in vivo is well established (1– 0732–0582/00/0410–0245$14.00

245

246

FONG n ENGLEMAN

3). By harnessing the capacity of DC to present tumor antigens to T cells, DC may serve as the centerpiece of an immunotherapeutic approach to cancer. The ability of the immune system to recognize and attack tumors has been demonstrated unequivocally. While both humoral and cellular effector mechanisms can contribute to tumor lysis, the latter are felt to be responsible for tumor regression in the majority of cases. CD8` cytotoxic T lymphocytes (CTL), in particular, have been demonstrated to recognize and kill cancer cells in various tumor models (4 –6). The ability of DC to prime T cells capable of recognizing and killing tumor cells in an antigen-specific fashion has also been demonstrated in various animal models (7–10). Moreover, DC-based immunization can lead to immunologic memory with protection against subsequent tumor challenges (10).


CONSIDERATIONS IN DC IMMUNOTHERAPY Defining DC Mature DC have a distinct morphology characterized by the presence of numerous membrane processes that can extend for up to hundreds of micrometers. These processes can take the form of dendrites, pseudopods, or veils (Figure 1). Additional morphologic features of DC include high concentrations of intracellular structures related to antigen processing such as endosomes, lysosomes, and the Birbeck granules of Langerhans cells (LC) of the epidermis. DC are also characterized by the presence on their surface of large amounts of class II MHC antigens and the absence of lineage markers including CD14 (monocyte), CD3 (T cell), CD19,20,24 (B cell), CD56 (natural killer cell), and CD66b (granulocyte) (Table 1) (11). Not surprisingly, in light of their antigenpresenting functions, DC also express various adhesion and costimulatory molecules. Examples of the former include CD11a (LFA-1), CD11c, CD50 (ICAM-2), CD54 (ICAM-1), CD58 (LFA-3), and CD102 (ICAM-3), although all of these markers can be found on monocytes and macrophages (12). Costimulatory molecules such as CD80 (B7.1) and CD86 (B7.2), and molecules regulating costimulation such as CD40 are also expressed on mature myeloid DC (13, 14). DC phenotype varies with different stages of maturation and activation. Human DC precursors circulating in the blood initially can express CD2, 4, 13, 16, 32, and 33, but they gradually lose their expression of these antigens with maturation (15). In contrast, adhesion molecules, costimulatory molecules, and MHC antigens increase with maturation. CD80 and 86 are upregulated with activation, particularly with CD40 ligation. CD86 tends to appear earlier in maturation, while CD80, which is almost unmeasurable in blood precursors, appears later (16). Several antibodies have been described that preferentially but not exclusively stain mature DC. Antibodies reactive against human DC include anti-CD83 and CMRF-44. Antibodies to CD83 stain mature activated DC, but not DC precursors, and also cross-react with activated B cells (17). CMRF-44 can stain circulating


DENDRITIC CELLS IN IMMUNOTHERAPY

247

Figure 1 DC morphology. A. Dendritic cells purified from peripheral blood demonstrate characteristic dendrites after several days of culture in vitro. B. Dendritic cells within the skin are evident on silver stain.

blood DC as well as activated DC, but it also cross-reacts with macrophage and monocytes (18). Antibodies directed at mouse DC include 33D1, N418 (antiCD11c), and DEC-205, although these antibodies also stain monocytes to varying degrees (19–21). DC possess various functional properties that distinguish them from other APC in vitro and in vivo. The ability to present antigens to naive T cells represents one of the most crucial functions of DC. While DC priming can be assessed in vitro, these assays can be difficult to perform routinely (22, 23). As a result, many

FONG n ENGLEMAN

248

TABLE 1 DC surface phenotype


DC Langerhans Activated MonocyteActivated precursors cells DC derived DC Monocytes B cells MHC Class I Class II

` `

` ``

` ```

` ``

` `

`` ``

Lymphoid CD2 CD3 CD4 CD19 CD56

`/1 1 ` 1 1

1 1 1 1 1

1 1 1 1 1

1 1 1 1 1

`/1 1 1 1 1

`/1 1 1 ` 1

1 1 1 1 `

` 1 `/1 1 `

` ` ` 1 `

1 ` ` 1 `

`/1 `/1 `/1 ` `/1

1 1 1 `/1 1

Costimulatory molecules CD80 1 CD86 ` CD40 `/1

` ` `

`` ``` `

` ` `

1 ` `

`` `` ``

Antigen receptors Mannose receptor Dec-205 CD16, CD32 CD35 CD36

1 1 ` ` `

` ` ` ` `

` ` 1 `/1

` ` ` ` `

` `/1 ` ` `

1 `/1 `/1 ` 1

Adhesion molecules CD11a (LFA-1) CD58 (LFA-3) CD54 (ICAM-1)

` ` `

` ` `

` ` ``

` ` `

` ` `

` ` ``

DC markers CD83 CMRF-44

1 1

`/1 1

` `

`/1 `

1 `

` `

Myeloid CD1a CD1b CD1c CD14 CD33

groups rely on the ability of DC to induce an allogeneic mixed leukocyte reaction (MLR) (24). In this respect, DC are 10 to 100 times more potent than monocytes or B cells at stimulating allogeneic T cells to proliferate. DC can also stimulate



249

autologous T cells, potentially presenting either self or exogenous antigens (including those potentially derived from foreign serum such as BSA) (25, 26). Allogeneic and autologous MLR assays, however, do not assess antigen uptake and processing. While macrophages are more efficient at phagocytosis of larger particles, such as cellular debris and bacteria, freshly isolated DC are active at pinocytosis of soluble antigen. Like macrophage, DC possess nonspecific antigen uptake receptors though at lower levels. Some DC express FccR (CD16, CD32) and complement receptors (CD11b, CD11c, CD35). CD11c may also act as a receptor for LPS as DC lack the classical LPS receptor, CD14, and yet respond to this stimulus. DC also can take up antigen through mannose receptors, potentially through the macrophage mannose receptor or through the receptor recognized by the DEC205 antibody (27). With DC activation and migration from the tissues, antigen uptake activity and the associated antigen receptors are downregulated, resulting in a switch in APC function from antigen uptake to antigen presentation (28). DC are capable of processing antigen via classical pathways: endogenous antigens via the proteasome into the MHC class I compartment, and exogenous antigens via endocytic lysosomes into the MHC class II compartment (29). DC also possess alternative pathways of antigen processing and can route exogenous antigen into the MHC class I pathway through a mechanism known as cross-priming (30). DC may also utilize molecular chaperones such as heat shock proteins (e.g. hsp96) to deliver antigens via the class I pathway (31). Sensitizing the immune system to specific antigens is certainly the most pertinent function for DC, and this has been examined both in vitro and in vivo. Early studies in mice demonstrated that DC exposed to infectious influenza virus or influenza nucleoprotein peptide, in vitro, induce a primary proliferative and antiviral CTL response (3, 32). Similarly, DC induced specific CTL for sendai, herpes simplex, and Moloney leukemia viruses (2). We have demonstrated that human DC, but not monocytes or B cells, can sensitize naive T cells to soluble protein antigens, enabling the generation of antigen-specific CD4` helper and CD8` CTL lines, in vitro (22, 23). In vivo priming of CD4` helper and CD8` cytotoxic T cells with antigenloaded DC has been demonstrated in several animal models. DC can also prime immune responses to antigens in vivo in humans (33, 34). Moreover, administration of DC loaded with tumor-associated peptides or proteins can lead not only to specific proliferative responses and CTL, but also to tumor protection in animal models (7–10). These observations provide a compelling rationale for pursuing DC-based immunotherapy for cancer.

DC Ontogeny and Function Although there is general agreement that DC are derived from hematopoietic stem cells, recent studies indicate that DC can arise from at least two distinct lineages. The precise relationship between these lineages remains controversial, partly


250

FONG n ENGLEMAN

because no specific markers for DC precursors have yet been identified. However, the heterogeneity of DC lineage and the possibility that the resulting DC subpopulations differ in their functions have significant implications when considering DC-based vaccination strategies. In humans, the more classical ‘‘myeloid’’ DC are derived from a committed DC precursor or granulocyte/monocyte precursor. DC from this lineage appear in the blood as typical lymphoid cells that express high levels of MHC class II antigens and lack most lineage-specific markers. Myeloid DC can also be derived from several cell types previously thought to be terminally committed. For example, monocytes and granulocyte precursors can differentiate into DC when exposed, in vitro, to appropriate combinations of cytokines including GM-CSF, TNF-a with or without IL-4 (35, 36). More recently, Shortman and colleagues described a population of DC derived from lymphoid progenitors in mice (37, 38). This cell type appears to arise from CD4`CD8` lymphoid precursors and can be induced to differentiate, in vitro, without GM-CSF (39, 40). Knockout mouse models have also implicated a lymphoid DC lineage. For example, Rel B knockout mice, in addition to their defects in NK and B cell development, have impaired production of mature DC, although LC develop normally in these mice (41, 42). Ikaros knockout mice lack T, B, and NK cells as well as DC, again implicating a role for a lymphoid lineage in the development of DC (43). While these ‘‘lymphoid’’ DC, characterized by their expression of CD8a homodimers, are felt to play a role in negative selection in the thymus, they clearly can prime naive T cells in the periphery. However, CD8a` DC appear to bias CD4` T cell priming to a Th1 response, while CD8a1 DC appear to bias toward a Th2 response (44, 45). No human CD8` DC have been described. DC derived from CD10` lymphoid precursors have been described in humans, although their role in the immune system has not been resolved (46– 48). More recently, reports of a distinct human DC subpopulation expressing high levels of CD123 (IL-3 receptor) and CD4 and lacking the CD11c myeloid DC marker have appeared (48, 49). Identified in blood and tonsil, these CD123` DC precursors require IL-3 for survival and an activation signal, such as CD40L, for maturation. They appear to bias CD4` T cell priming to a Th2 response, in contrast to myeloid CD11c` DC, which preferentially induce a Th1 response (50). These CD123hi DC also appear to be a major source of type I interferons and may, therefore, possess effector immune function as well (51).

Dendritic Cell Life Cycle In order to effectively exploit DC function, the different stages of DC maturation must be understood. DC precursors migrate from bone marrow and circulate in the blood to specific sites in the body where they mature and act as sentinels for the immune system (Figure 2) (52). This trafficking to the tissues is directed by expression of the chemokine receptors CCR1, CCR5, and CCR6 as well as adhe-



251

Figure 2 DC life cycle. DC migrate into peripheral tissues through the vascular endothelium, drawn to such sites by chemokines (for example, MIP1a, MIP-3a, and RANTES). Following activation by one or more stimuli [e.g. LPS, IFN-a or c, RANTES, bacteria, viruses, certain unmethylated immunostimulatory oligonucleotides, also known as CpG, MIP-3b. or SLC (secondary lymphoid tissue chemokine)], DC emigrate from the tissues via the afferent lymphatics to lymph nodes, where they interact with T cells.

sion molecules such as the CD62P-ligand (53–55). Tissue-resident DC, including LC in skin, hepatic DC in the portal triads, mucosal DC, and lung DC, sample the environment, processing and presenting antigens in the context of MHC class I and II molecules. Sources of antigen can include viral proteins, bacterial proteins, or apoptotic bodies (1, 56, 57).


252

FONG n ENGLEMAN

In addition to encountering antigen, DC are thought to require an antigenindependent danger signal in order to become activated. These signals can include LPS, interferon (IFN) a and c, IL-1b, and potentially direct signals from viruses and bacteria. With activation, DC downregulate their antigen uptake and processing functions and shift to antigen presentation and upregulate their expression of MHC, costimulatory, and adhesion molecules as described earlier. With maturation during in vitro culture, MHC class I and II densities on the cell surface increases 10–100-fold, while MHC synthesis is reduced. Class II peptide-MHC complex turnover falls dramatically, as reflected by the change in surface halflife of these complexes from a few hours on immature DC to 2 days on mature DC, providing a stable source of processed antigen when danger signals are received (58). Once activated, DC leave the tissues and migrate via the afferent lymphatics to the T cell–rich paracortex of the draining lymph nodes, drawn by chemokines MIP-3b and SLC through upregulation of their receptor CCR7 (59, 60). Within the secondary lymphoid organs, activated DC elaborate chemokines such as DCCK1 and MDC that attract naive and memory T cells for priming (61, 62). Activated DC also secrete cytokines including IL-7 and IL-12, which contribute to their unique functions. IL-7 can induce CD4 and CD8 T cell proliferation and B cell differentiation (63, 64). IL-12 biases T helper responses toward a Th1 pattern (65, 66). In contrast to DC, monocytes secrete significant levels of IL-4, -5, and -10 and do not secrete IL-7. This combination of cytokines will typically bias T helper responses to Th2 or antibody generation. This stimulatory milieu produced by activated DC, combined with the presentation of epitopes in MHC class I and class II and the expression of costimulatory molecules, contributes to the generation of potent antigen-specific immune responses (67). Within the secondary lymphoid tissues, activated CD4` T cells can further activate DC through CD40/ 40L interactions, and this also provides a survival signal to the DC (68–70). Furthermore, CD40 ligation allows DC to prime CTL, even in the absence of immediate CD4` T cell help. Preliminary evidence also exists for the ability of DC to present antigen to natural killer cells (NK), perhaps through CD1 (71). This may serve as another mechanism through which DC can elicit an effector immune response.

Sources of DC The earliest studied DC were derived from mouse spleen (72). Their isolation relied upon a transient exposure of the cells to tissue culture plates, which allowed for the depletion of adherent monocytes. Lymphocytes were then depleted by passage over a bovine serum albumin (BSA) gradient to isolate low-density DC (73). Related techniques to obtain DC from murine tissues, including lymph nodes and skin, were subsequently developed, but these techniques yield low numbers of DC, making experiments difficult to perform. More recent approaches for obtaining DC have involved their generation in vitro from precursor cells cultured



253

in the presence of different cytokines. One method, in which murine bone marrow cells are depleted of differentiated precursors and grown in medium containing GM-CSF and IL-4, has become widely used because of its greater DC yield (74). Human DC can also be enriched as circulating precursors from the blood by density-based purification techniques (75). After a period of in vitro culture and maturation, DC precursors become larger and less dense. Gradient solutions lacking potentially immunogenic proteins such as BSA have been employed including Percollt (Pharmacia Biotech), Nycodenzt (Nycomed Pharma, Oslo, Norway), and metrizamide. These different fluids are osmotically active to varying degrees and have additional stimulatory properties as well. The use of density-based isolation is, however, limited by the low frequency of DC precursors in blood, representing around 1% of peripheral blood mononuclear cells (PBMC) (22, 75). As a result, a leukapheresis must be performed to generate sufficient numbers (on average 5 x 106) of DC for therapeutic vaccination in humans. As with the animal models, more recent approaches for generating human DC from bone marrow precursors utilize CD34` cells cultured in the presence of exogenous GM-CSF, usually in combination with IL-4 and/or TNF-a (76–79). Stem cell factor (SCF) and/or Flt3-ligand (FL) are often added to increase DC yields by inducing the proliferation of DC progenitors (80, 81). Sources for human CD34` precursors include bone marrow, cord blood, and G-CSF-mobilized peripheral blood. Many groups have generated DC-like cells by culturing CD14` monocyteenriched PBMC in vitro. When cultured for 1–2 weeks with media supplemented with GM-CSF and IL-4, monocytes give rise to large numbers of cells that are morphologically and phenotypically similar to the ‘‘classical’’ density purified DC (35, 77). These cytokine-generated DC require additional maturation in vitro with TNF-a or monocyte-conditioned media in order to fully stimulate in an allogeneic MLR or prime antigen-specific T cell responses in vitro and in vivo. Moreover, without this additional maturation step, the DC phenotype can revert to that of a monocyte. Nonetheless, these ‘‘monocyte-derived DC’’ are capable of inducing strong antitumor responses in vivo, in mice (7, 8). Recently, clinical trials using these cells in patients with several malignancies have demonstrated induction of specific immune responses as well as tumor regression (34, 82). One advantage of this approach includes easy access to CD14` monocytes. 50 mL of whole blood is sufficient to generate over 10 x 106 DC for vaccine production. However, this approach requires at least one week of in vitro culture with exogenous cytokines, which increases the risk of bacterial contamination as well as cost. An alternative approach is to expand DC in vivo. Administration of FL to mice in vivo results in preferential mobilization or release of DC precursors from the bone marrow to the periphery and into lymphoid organs (81). Recently, in a clinical protocol that utilizes FL from Immunex Corp., we have observed that administration of this growth factor can increase the number of circulating blood DC 10–30-fold (Figure 3). These DC can be harvested with a leukapheresis pro-

FONG n ENGLEMAN

254

B.

CD3, CD14, CD19, CD56

A.

1.8%

11.6%


MHC Class II Figure 3 FL administration in vivo expands circulating blood DC. A. DC in circulating blood as identified by expression of MHC class II and lack of lineage markers represent approximately 1% of the PBMC pre-FL administration. B. Following FL administration in vivo, the fraction of DC is increased approximately 10-fold as a percentage of PBMC.

cedure for ex vivo manipulation and used to prime humans to antigens in vivo (manuscript in preparation). Such an approach would avoid the need for prolonged in vitro culture or repeated leukaphereses. In generating DC for use as a cellular vaccine, regardless of source, the infused cells should possess a stable as well as an activated phenotype. In addition to expressing the requisite MHC and costimulatory molecules to prime T cells, the cells should express appropriate adhesion molecules and chemokine receptors to attract the DC to secondary lymphoid organs for priming. Otherwise, ineffective priming may occur, particularly if the DC are administered systemically rather than locally into the relevant draining lymph nodes. Intralymphatic or intranodal injection of DC may be used to deliver DC directly to secondary lymphoid organs, although these routes of administration have not yet been shown conclusively to produce more effective vaccination (83).

DC and Cancer The ability of DC to generate antitumor immune responses in vivo has been documented in many animal models. Most of these experiments have involved in vitro isolation of DC, followed by loading of the DC with tumor antigen and injection of the antigen-bearing DC into syngeneic animals as a cancer vaccine. DC loaded with tumor lysates, tumor antigen-derived peptides, synthetic MHC class I–restricted peptides, and whole protein have all been demonstrated to generate tumor-specific immune responses and antitumor activity (9, 10, 84). Furthermore, antigen-loaded DC can be used therapeutically to induce regression of


255

preexisting tumors (7, 85). These observations have established the rationale for evaluating tumor antigen–bearing DC as therapeutic vaccines in humans.

Candidate Target Antigens DC based immunization requires that the cells present one or more tumor antigens to the host’s T cells. Fortunately, a number of antigens associated with various malignancies have been described (Table 2). Some groups have focused upon

TABLE 2 Potential tumor antigens for immunotherapy


Antigen Class Tumor specific antigens

Antigen Immunoglobulin idiotype TCR Mutant p21/ras Mutant p53 p210/bcr-abl fusion product

Developmental antigens

Viral antigens

Tissue-specific selfantigens

Overexpressed selfantigens

MART-1/Melan A MAGE-1, MAGE-3 GAGE family Telomerase Human papilloma virus Epstein Barr virus

Tyrosinase gp100 Prostatic acid phosphatase Prostate-specific antigen Prostate-specific membrane antigen Thyroglobulin a-fetoprotein Her-2/neu Carcinoembryonic antigen Muc-1

Malignancy B-Cell non-Hodgkin’s lymphoma, multiple myeloma T cell non-Hodgkin’s lymphoma Pancreatic, colon, lung cancer Colorectal, lung, bladder, head and neck cancer Chronic myelogenous leukemia, acute lymphoblastic leukemia Melanoma Melanoma, colorectal, lung, gastric cancers Melanoma Various Cervical, penile cancer Burkitt’s lymphoma, nasopharyngeal carcinoma, post-transplant lymphoproliferative disorders Melanoma Melanoma Prostate cancer Prostate cancer Prostate cancer Thyroid cancer Liver cancer Breast and lung cancers Colorectal, lung, breast cancer Colorectal, pancreatic, ovarian, lung cancer


256

FONG n ENGLEMAN

tumor-derived protein extracts or RNA as the source for antigens. When these approaches are used, the vaccine contains multiple antigens, increasing the probability of inducing immunity to more than one tumor-associated antigen. Although the target proteins are initially undefined, with this approach, the immunogen can be identified later (6). On the other hand, there is increased potential for the induction of a destructive autoimmune response to antigens expressed on normal tissues. Truly tumor-specific antigens offer theoretical advantages as immunotherapy targets. The immune response induced by such antigens presumably would be limited to tumor cells bearing epitopes of the antigen, limiting the risk of collateral damage to normal tissues (86, 87). If such proteins arise in the tumor following thymic development, antigen-reactive T cells could exist in the repertoire as they may have avoided thymic deletion. Nevertheless, the immunogenicity of these antigens may be limited by the number of epitopes contained within the protein (e.g., p53 or ras possessing a single amino acid mutation). Moreover, tumorspecific antigens may vary between individuals with the same tumor type. Viral tumor antigens can represent desirable immune targets given the inherent immunogenicity of many viral proteins (88). A number of antigens expressed on normal tissues as well as tumors have also been examined as immunotherapy targets. Despite the relatively small number of such antigens identified to date, many have been shown to stimulate T cell responses in vitro when presented by DC (89, 90). This is surprising since most T cells capable of recognizing self-proteins are presumably deleted from the repertoire during T cell differentiation in the thymus. The extent to which this occurs may vary depending on the extent of protein expression during the period of T cell development. However, preexisting T cells reactive to self-antigens have been demonstrated in vivo in cancer patients. Moreover, tumor-specific immune responses have been induced with various immunization strategies in vivo. Tumor-reactive T cells, therefore, apparently exist in the T cell repertoire of cancer patients but are rendered tolerant or anergic (91). Alternatively, the antigenreactive T cells may recognize and respond to subdominant or cryptic epitopes that are not ordinarily presented at levels sufficient to induce immunity (92, 93). While animal models have been used frequently to examine the immunogenicity of tumor-associated antigens, significant caveats exist in extrapolating animal data to humans. Many of the animal tumor antigen models focus upon proteins that are foreign to the animal (e.g. carcinoembryonic antigen, or Her-2/ neu in mice). Transgenic tumor models (e.g. MMTV/neu) could be used to address this issue although differences in T cell receptor repertoire remain as a confounding factor. Other tumor models (e.g. fibrosarcoma) are based upon tumors that are inherently immunogenic. These model systems contrast with cancer in human patients, where many of the tumor antigens are self-antigens, and tumors other than melanoma and renal cell carcinoma are rarely immunogenic. Another consideration in selecting candidate target antigens is tissue or tumor specificity. Most of the tumor-associated proteins being explored in clinical trials



257

(e.g. CEA and Her2/neu) are expressed by some normal tissues (albeit at low density) as well as by neoplastic cells. There exists a theoretical risk that DCbased presentation of these individual antigens, or crude mixtures of tumorderived peptides or nucleic acids, may induce autoimmune disease affecting vital organs. Antigens uniquely expressed in non-essential organs (e.g. prostate) may induce an acceptable form of tissue-specific autoimmunity, especially if these organs have been removed as part of treatment (e.g. prostatectomy). Antigen-loss variants leading to tumor escape have been observed with immunotherapy (94, 95). Ideally target antigens should represent antigens universal to the tumor type and required for their survival and growth. Vaccine formulations may also require mixtures of commonly expressed tumor antigens to target heterogeneous expression of antigens within the tumor. Recently, approaches have been devised to increase the immunogenicity of tumor-associated antigens by modifying their structure or using xenogeneic analogues. For example, immunization with xenogeneic prostate acid phosphatase (PAP), e.g. immunizing rats with human PAP, generated autoimmunity to the prostate gland (96). In contrast, immunizing with the self-protein, rat PAP, did not. Other approaches have focused upon modifying peptides to increase HLA binding or T cell receptor interactions and have shown that these modified peptides can serve as agonist peptides (97, 98). In one recent melanoma trial, patients immunized with a modified melanoma-derived peptide developed clinical responses, while those immunized with the native peptide did not (99). The ability to develop more immunogenic antigens will improve as MHC-peptide–T cell receptor interactions are better understood.

Approaches to Antigen Delivery To date, several approaches have been used to arm DC with target antigen for use in clinical trials. Tumor-associated peptides either in cell lysates or acid eluted fractions from MHC class I molecules on tumor cells have been used to load MHC molecules at the surface of the cell. Specific MHC-restricted peptides derived from tumor-associated antigens such as MART-1, MAGE-1, CEA, and prostate-specific membrane antigen (PSMA) have also been used (34, 100, 101). This approach relies on predictive algorithms to identify peptides with high binding affinity to the HLA molecule, most commonly the HLA*A0201 alleles. Those peptides with demonstrated binding to HLA*A0201 are then studied for immunogenicity based upon their ability to generate CTL, in vitro. Although this approach simplifies production of target antigens to a limited number of peptides, selection of patients is restricted to those with this specific HLA haplotype (30– 45% of the population). Moreover, additional biologically important epitopes may be missed, as binding affinity for a particular HLA allele does not translate directly into immunogenicity. Using MHC class I–restricted peptides also ignores the role of MHC class II–restricted T helper cells in initiating and sustaining an immune response.


258

FONG n ENGLEMAN

An alternative approach is to utilize full-length recombinant proteins as antigens. By introducing whole protein into DC, subsequent immune responses would potentially be restricted by multiple alleles rather than single alleles as they are with peptides. The antigen-processing machinery of DC would be relied upon to stimulate responses to the important epitopes within the antigen. Under normal circumstances, addition of intact soluble proteins to DC would be expected to result in entry of the proteins into MHC class II processing pathway but not the MHC class I processing pathway. Such processing would allow for presentation of antigenic epitopes to CD4` T cells, which are important sources of proinflammatory cytokines such as IL-2 and IFN-c, but would not allow for efficient presentation of antigenic epitopes to CD8` CTL, which are believed to be critical effectors of antitumor immune responses. Although DC may also ‘‘leak’’ exogenous antigen into the MHC class I pathway, which can lead to stimulation of CD8` CTL, this occurs inefficiently. Nonetheless, since most solid tumors do not express MHC class II determinants, the generation of tumor-reactive CD8` T cells may be required for an effective antitumor response, and this would necessitate that antigens be delivered to DC in a manner that assures their access to the class I MHC processing pathway. A number of approaches are being developed to accomplish this goal, including gene transfer methods that result in antigen processing in the MHC class I pathway of DC and presentation to CD8` T cells. Conjugating certain ‘‘transporter’’ peptides onto full-length proteins allows these proteins to translocate across cell membranes and into the MHC class I pathway. One such peptide derived from HIV tat can increase the efficiency of MHC class I–restricted antigen presentation by more than 100-fold (102). Peptides derived from bacterial toxins may also have similar capabilities (103). DNA transfection of DC, with or without liposomal encapsulation, has been tried with varying success (104). RNA derived from tumors or encoding tumor antigens can also be transfected into DC leading to the generation of class I MHC restricted responses (90). Recombinant viral vectors such as adenovirus, vaccinia, or fowlpox expressing the antigen have been used to deliver transgene products into the class I MHC pathway of DC (105, 106). This approach offers the additional advantage of exposing the DC to immune-activating components of the virus, which leads to a more potent APC. Recombinant retroviruses have also been used, although this approach has so far been effective only in DC precursors because proliferation is required and differentiated DC do not proliferate (107). However, retroviral constructs capable of infecting resting, nondividing cells have recently been described (108, 109).

DC Clinical Trials Human trials using DC loaded with tumor antigens are underway at several institutions. In the first reported DC trial, the effect of autologous DC pulsed ex vivo with tumor-specific antigen was investigated in patients with malignant B cell lymphoma for whom conventional chemotherapy had failed. Like other B lym-



259

phocytes, the neoplastic cells in these patients express surface immunoglobulin receptors, and because B cell lymphomas are monoclonal, all the cells of a given tumor express identical surface immunoglobulin. Moreover, this immunoglobulin is potentially immunogenic by virtue of its unique idiotypic determinants, which are formed by the combination of the variable regions of immunoglobulin heavy and light chains (110–112). To prepare idiotype proteins for this clinical study, patients underwent tumor biopsies; the immunoglobulin (idiotype) produced by each tumor was ‘‘rescued’’ by somatic cell fusion techniques and purified from hybridoma supernatants (113). Peripheral blood leukocyte preparations containing DC precursors were obtained from patients by leukapheresis, and monocytedepleted mononuclear cells were incubated for 36 h in the presence of either tumor-specific idiotype protein or a control protein, keyhole limpet hemocyanin (KLH). This incubation technique not only enabled the DC precursors to take up and process the exogenous proteins, but also induced their maturation into potent APC. Following this culture period, the idiotype and KLH-pulsed DC were separated from contaminating lymphocytes on the basis of their differences in buoyant density. The DC were then extensively washed to remove free protein, resuspended in sterile saline, and administered to the patients by intravenous infusion. This procedure was repeated three times at monthly intervals with a booster immunization given four to six months later. Two weeks after each DC infusion, the patients received subcutaneous injections of idiotype, which in the absence of DC or a chemical adjuvant would not be expected to induce an immune response (114). Throughout the trial the patients were followed for the development of an immune response to the idiotype and their tumor burden was monitored. A report of the results obtained in our initial four patients has been published (33). All of these treated patients, as well as six not described in our published report, tolerated their infusions well, and none have experienced clinically significant toxicity at any point during the study. In addition, the majority of the patients developed T cell–mediated anti-idiotype and anti-KLH proliferative responses that were not observed prior to treatment initiation. The anti-idiotype responses were specific for autologous tumor immunoglobulin when compared to irrelevant, isotype-matched immunoglobulins. In addition to these proliferative responses, T cells from one patient were expanded for several weeks in vitro in the presence of idiotype protein and shown to lyse autologous tumor hybridoma cells but not an isotype-matched, unrelated hybridoma. In contrast to earlier idiotype vaccination studies that utilized chemical adjuvants rather than DC (111, 114), a humoral response to idiotype was not detected in our patients, consistent with the interpretation that the idiotype-pulsed DC induced a Th1 rather than Th2 response. Most importantly, two of the patients experienced complete tumor regression, including one who entered the trial with bulky disease and remained in complete remission for more than three years. A third patient experienced a partial response while three have had stable disease states and three have experienced disease progression. Recently, a new cohort of patients has been vacci-


260

FONG n ENGLEMAN

nated while in remission, and their follow-up is ongoing. In the DC trial for B cell lymphoma, we do not know whether the idiotype pulsed DC used induced CD8` CTL. B cell lymphoma is unusual among malignancies in that the neoplastic cells express both MHC class I and II antigens and would, therefore, be susceptible to recognition and attack by either CD4` or CD8` T cells. A Phase I DC trial in prostate cancer has since been reported by Murphy et al (101). DC were derived from peripheral blood–derived monocytes and cultured in GM-CSF and IL-4. Subsets of the 51 patients were treated with unloaded DC or DC pulsed with HLA*A2–restricted peptides derived from prostate-specific membrane antigen (PSMA). Non-HLA*A2 patients were included in the trial, and no consistent T cell responses were detected. Seven patients were scored as partial responders. Of these patients, 33 were subsequently enrolled in a Phase II study utilizing two PSMA peptides loaded onto autologous DC. Nine patients were scored as partial responders, including four who were partial responders in the original Phase I trial (115). Patients were not, however, controlled for concurrent treatment modifications (including hormone withdrawal), so follow-up studies will be very important. Two vaccine trials for melanoma have also been reported investigating DC pulsed with a panel of melanoma-derived, HLA-restricted peptides. Both trials utilized DC-like cells derived from monocytes by culture in GM-CSF and IL-4. Nestle et al limited their clinical trial to patients expressing the HLA*A1 or A2 alleles and reported that 5 of 15 patients developed clinical responses, including two complete remissions which have been durable (34; personal communication). Induction of delayed type hypersensitivity (with skin testing) to the antigen was seen with this vaccination approach. Lotze et al also reported the results of their clinical trial in melanoma with one complete response in their cohort of six patients (82). Our group has recently initiated several new DC trials. One trial, in multiple myeloma patients who have undergone peripheral blood stem cell rescue following ablative chemotherapy, employs the myeloma-specific immunoglobulin idiotype as the tumor antigen. Each patient’s idiotype is purified from serum. Following high dose therapy and bone marrow recovery, these patients are treated with idiotype pulsed DC using a protocol similar to that used in our lymphoma trials. This trial is currently ongoing, although in our initial cohort of 12 patients idiotype-specific T cell responses were detected only in those patients who achieved a complete response with high dose chemotherapy and had no detectable circulating myeloma protein (116). We have also initiated a trial in prostate cancer employing prostatic acid phosphatase (PAP) as the tumor antigen. Utilizing DC purified from peripheral blood by density gradient centrifugation, this protocol seeks to immunize patients with metastatic prostate cancer with PAP-pulsed DC. Based upon our preclinical data, we are utilizing a xenogeneic PAP, derived from rodents, rather than human PAP as the immunogen. Preliminary data demonstrate the ability of xenogeneic antigen-loaded DC to generate cross-reactive immunity to the self-antigen as well


261

as clinical responses in a subset of patients (83). Many other clinical trials have been initiated at various institutions (Table 3).


Future Directions Although the results to date of the various DC trials are provocative, the procedures used for isolating and ‘‘arming’’ DC are tedious and not yet broadly applicable to clinical practice. Indeed, our observation in the original lymphoma trial that a small number of APC could induce an immune response to a weak antigen in tumor-bearing patients, most of whom were leukopenic as a consequence of prior chemotherapy, was surprising. Not all of the treated patients, however , have experienced tumor regression, suggesting that the DC dose, route of administration, choice of antigen, or method of antigen loading may have been suboptimal. As discussed, many groups are already pursuing techniques for in vitro generation of DC from CD34` precursors or CD14` monocytes. These approaches, while expensive, can generate large numbers of cells for use in clinical trials. Administration of growth factors that increase numbers of circulating DC in patients in vivo offers another possibility. In a limited pilot trial using GM-CSF for DC mobilization, we found that yields of monocytes were significantly increased, making DC purification difficult (E Engleman, unpublished results). In contrast to GM-CSF, FL, which appears to preferentially expand DC precursors in humans in vivo, shows significant promise. It remains to be seen whether FL is useful alone as a therapeutic agent, or whether the DC mobilized with FL must be harvested, armed in vitro with antigen, and infused to induce antitumor immunity. Treatment of mice bearing certain immunogenic tumors with FL has been shown to result in tumor regression (117). These responses presumably occur, at least in part, through DC priming of effector T cells, although NK cells have also been shown to be activated by FL. However; whether such an approach will circumvent the inhibitory milieu within cancer patients will require further investigation. Studies to investigate this possibility are currently underway. In addition to questions regarding the best source of DC for use in clinical trials, the choice of tumor antigen with which to ‘‘arm’’ the DC is almost certainly going to have a profound influence on clinical outcome. At present, the choices are limited because only a few antigens have been identified that are either tumoror tissue-restricted in their expression. Only a subset of these antigens are immunogenic. On the other hand, the number of candidate antigens has increased with time and is expected to increase further as a consequence of intensive gene identification efforts. Ultimately, combinations of antigens should be used to reduce the risk of generating antigen-loss variants that evade the immune response. The increasing use of microarray technology in assessing tumors may someday allow for individualized antigen combinations. Antigen delivery also remains to be optimized. Use of protein, whether tumor derived or with recombinant DNA methods, can be cumbersome and potentially limiting, especially at concentrations that may be necessary for MHC class I

262

TABLE 3 Dendritic cell clinical trials


Malignancy

Antigen

Antigen construct

DC type

Institution

Investigator

Breast, colorectal, Tumor derived pancreas, lung, melanoma Breast, colorectal, lung Mutant Ras derived

Autologous tumor cell lysate

GM-CSF ` IL-4 cultured monocytes

Uni. of Michigan

Mule et al.(84)

Ras peptides

Vanderbilt

Carbone et al.1

Breast, colorectal, lung Mutant p53

p53 DNA

Vanderbilt

Carbone et al.1

Breast, colorectal

CEA

USC

Weber et al.1

Lung, colorectal

CEA

Stanford

Fong, Engleman et al.

Lung

CEA

Duke

Lyerly, Gilboa et al.1

Lung

CEA

HLA*A2 restricted peptide CAP-1 HLA*A2 restricted peptide CAP-1 (610M)3 HLA*A2 restricted peptide CAP-1 CEA RNA

GM-CSF ` IL-4 cultured monocytes GM-CSF ` IL-4 cultured monocytes CF-CSF ` IL-4 cultured monocytes FL mobilized, density purified DC

Duke

Lyerly, Gilboa et al.1

Melanoma

gp100, Mart-1, & Tyrosinase gp100, Mart-1, & Tyrosinase

HLA*A2 restricted peptides HLA*A2 restricted peptides

Uni. of Zurich

Nestle et al.(34)

Univ. Pittsburgh

Lotze et al.(82)

gp100 (210M)2, MART-1, & Tyrosinase

HLA*A2 restricted peptides

NCI

Marincola et al.1

Melanoma

Melanoma

GM-CSF ` IL-4 cultured monocytes GM-CSF ` IL-4 cultured monocytes GM-CSF ` IL-4 cultured monocytes GM-CSF ` IL-4 cultured monocytes or CD34` cells GM-CSF ` IL-4 cultured monocytes

TABLE 3 Continued Malignancy Melanoma

Non-Hodgkin’s lymphoma Prostate Prostate

gp100, Mart-1, MAGE-3 & Tyrosinase gp100 (210M),2 & Tyrosinase Immunoglobulin idiotype Immunoglobulin idiotype Immunoglobulin idiotype Xenogeneic PAP PSMA

Prostate

PAP

Renal

Tumor derived

Melanoma Multiple myeloma Multiple myeloma

1

263


Antigen

Antigen construct

DC type

Institution

Investigator

HLA*A2 restricted peptides

GM-CSF ` TNF-a cultured CD34` cells

Baylor

Bancerau et al.1

HLA*A2 restricted peptides Purified protein

GM-CSF ` IL-4 cultured monocytes Dendreon density purified DC Dendreon density purified DC Density purified DC

USC

Weber et al.1

Stanford

Levy, Engleman et al.

Dendreon (multicenter) Stanford

Valone et al.1

Purified protein Recombinant protein Recombinant protein HLA*A2 restricted peptides Recombinant fusion protein Autologous tumor cell lysate

Density purified DC GM-CSF ` IL-4 cultured monocytes 5 adjuvant GM-CSF Dendreon density purified DC GM-CSF ` IL-4 cultured monocytes

Stanford Northwest Biotherapeutics (multicenter) Dendreon (multicenter) UCLA

Personal communication. 2gp100 derived peptide with modification at position 210. 3CEA derived peptide with modification at position 610.

Levy, Engleman et al.(33) Fong, Engleman et al. Murphy et al.(129)

Valone et al.(130) Figlin, Belldegrun, et al.1


264

FONG n ENGLEMAN

delivery. Use of specific peptide conjugates or fusion constructs (e.g. with HIV tat) may increase the efficiency of presenting epitopes from these soluble proteins. RNA, DNA, and viral vectors are more easily produced and may offer a viable alternative, although issues of transfection efficiency and DC viability remain unresolved. The potential benefits of administering cytokines or other DC activators in combination with DC vaccination remain relatively unexplored. Clearly, DC can elaborate their own cytokines for antigen priming as previously discussed. However, supplementing culture media used for generating DC in vitro with additional cytokines may enhance the APC functions of these DC. TNF-a or CD40 ligand are known to activate DC in vitro and could increase the potency of DC-based vaccines (70, 118, 119). The addition of IL-12 may aid in antigen priming and generating Th1 responses in vitro or in vivo (120). A singular study to date has examined the role of IL-2 in augmenting the effector phase of the immune response to peptide vaccination in melanoma patients. This trial is particularly important in that patients who were immunized and given high doses of IL-2 developed immune responses, while those patients who received the vaccine alone did not (99). Synergy between DC vaccination and IL-2 has already been demonstrated in an animal model (121). No consensus exists on the optimal approach to assessment of immune responses in patients undergoing immunotherapy, let alone DC vaccination. Delayed-type hypersensitivity testing with an antigen challenge injected into the skin has been used to assess gross immune reactivity. Cytokine production by both CD4` and CD8` T cells can be detected by cytokine ELISA, ELISPOT, or intracellular cytokine staining. Other techniques for assessing CD4` T cell responses include measurement of antigen-specific proliferation. CD8 T cell responses have also been assessed with CTL assays and more recently tetramer staining. However, these assays assess the functions of circulating T cells but do not necessarily reflect immune responses in lymphoid organs. Improved immunologic assays that better correlate with clinical outcome will be required before these assays can serve as surrogates for evaluation of tumor status. Before closing, it is worth considering the following question. Why does administration of small numbers of antigen-pulsed DC induce antigen-specific T cell responses and tumor regression in patients in whom both the antigen and DC are already plentiful? The explanation that we favor is that most of the circulating DC and DC residing in or near malignant tissues may be defective in their capacity to take up, process, and present tumor-associated antigens because they fail to become activated and differentiate into mature APC. This view is supported by the observation that DC from tumor-bearing animals can be defective in their capacity to stimulate T cells (86, 122). Moreover, tumors secrete various mediators that have recently been found to inhibit DC differentiation and/or maturation in vitro. Examples include VEGF, IL-6, and IL-10 (91, 123, 124). Tumors may even inhibit migration of DC to tumor-bearing tissues through their secretion of IL-10 or downregulation of vascular adhesion molecules (125, 126). Indeed, the


265

presence of mature DC in lung, cervical, colon, gastric, nasopharyngeal, thyroid, and prostate cancers is associated with a favorable prognosis (127, 128). By isolating and arming DC with antigen ex vivo, we may be allowing the cells to mature away from an inhibitory milieu. If defective maturation of DC is a common occurrence in malignancy, then identification of agents that induce their maturation, in vivo, may represent an elegant solution to this problem. In the absence of such agents, administration of DC derived from circulating DC precursors, which have been induced to mature in vitro and armed with appropriate tumor antigens, may prove useful in the treatment of a variety of tumors. Visit the Annual Reviews home page at www.AnnualReviews.org.


LITERATURE CITED 1. Hengel H, Lindner M, Wagner H, Heeg K. 1987. Frequency of herpes simplex virus-specific murine cytotoxic T lymphocyte precursors in mitogen- and antigen-driven primary in vitro T cell responses. J. Immunol. 139:4196–4202 2. Kast WM, Boog CJ, Roep BO, Voordouw AC, Melief CJ. 1988. Failure or success in the restoration of virus-specific cytotoxic T lymphocyte response defects by dendritic cells. J. Immunol. 140:3186–93 3. Nonacs R, Humborg C, Tam JP, Steinman RM. 1992. Mechanisms of mouse spleen dendritic cell function in the generation of influenza-specific, cytolytic T lymphocytes. J. Exp. Med. 176:519–29 4. Feltkamp MC, Smits HL, Vierboom MP, Minnaar RP, de Jongh BM, Drijfhout JW, ter Schegget J, Melief CJ, Kast WM. 1993. Vaccination with cytotoxic T lymphocyte epitope-containing peptide protects against a tumor induced by human papillomavirus type 16- transformed cells. Eur. J. Immunol. 23:2242–49 5. Celluzzi CM, Mayordomo JI, Storkus WJ, Lotze MT, Falo LD Jr. 1996. Peptide-pulsed dendritic cells induce antigen-specific CTL-mediated protective tumor immunity. J. Exp. Med. 183:283– 87

6. Boon T, Cerottini JC, Van den Eynde B, van der Bruggen P, Van Pel A. 1994. Tumor antigens recognized by T lymphocytes. Annu. Rev. Immunol. 12:337– 65 7. Zitvogel L, Mayordomo JI, Tjandrawan T, DeLeo AB, Clarke MR, Lotze MT, Storkus WJ. 1996. Therapy of murine tumors with tumor peptide-pulsed dendritic cells: dependence on T cells, B7 costimulation, and T helper cell 1-associated cytokines. J. Exp. Med. 183:87–97 8. Paglia P, Chiodoni C, Rodolfo M, Colombo MP. 1996. Murine dendritic cells loaded in vitro with soluble protein prime cytotoxic T lymphocytes against tumor antigen in vivo. J. Exp. Med. 183:317–22 9. Flamand V, Sornasse T, Thielemans K, Demanet C, Bakkus M, Bazin H, Tielemans F, Leo O, Urbain J, Moser M. 1994. Murine dendritic cells pulsed in vitro with tumor antigen induce tumor resistance in vivo. Eur. J. Immunol. 24:605– 10 10. Mayordomo JI, Zorina T, Storkus WJ, Zitvogel L, Celluzzi C, Falo LD, Melief CJ, Ildstad ST, Kast WM, Deleo AB, et al. 1995. Bone marrow-derived dendritic cells pulsed with synthetic tumour peptides elicit protective and therapeutic

266

11.

12.

13.


14.

15.

16.

17.

18.

FONG n ENGLEMAN antitumour immunity. Nat. Med. 1:1297– 1302 Hart DN. 1997. Dendritic cells: unique leukocyte populations which control the primary immune response. Blood 90: 3245–87 Hart DN, Prickett TC. 1993. Adhesion molecules in tonsil DC-T cell interactions. Adv. Exp. Med. Biol. 329:65–69 Banchereau J, Bazan F, Blanchard D, Briere F, Galizzi JP, van Kooten C, Liu YJ, Rousset F, Saeland S. 1994. The CD40 antigen and its ligand. Annu. Rev. Immunol. 12:881–922 Fagnoni FF, Takamizawa M, Godfrey WR, Rivas A, Azuma M, Okumura K, Engleman EG. 1995. Role of B70/B7–2 in CD4` T-cell immune responses induced by dendritic cells. Immunology 85:467–74 Takamizawa M, Rivas A, Fagnoni F, Benike C, Kosek J, Hyakawa H, Engleman EG. 1997. Dendritic cells that process and present nominal antigens to naive T lymphocytes are derived from CD2` precursors. J. Immunol. 158: 2134–42 McLellan AD, Starling GC, Williams LA, Hock BD, Hart DN. 1995. Activation of human peripheral blood dendritic cells induces the CD86 co-stimulatory molecule [published erratum appears in Eur. J. Immunol. 1995 Dec;25(12):3525]. Eur. J. Immunol. 25:2064–68 Zhou LJ, Schwarting R, Smith HM, Tedder TF. 1992. A novel cell-surface molecule expressed by human interdigitating reticulum cells, Langerhans cells, and activated lymphocytes is a new member of the Ig superfamily. J. Immunol. 149:735–42 Hock BD, Starling GC, Daniel PB, Hart DN. 1994. Characterization of CMRF44, a novel monoclonal antibody to an activation antigen expressed by the allostimulatory cells within peripheral blood, including dendritic cells. Immunology 83:573–81

19. Nussenzweig MC, Steinman RM, Witmer MD, Gutchinov B. 1982. A monoclonal antibody specific for mouse dendritic cells. Proc. Natl. Acad. Sci. USA 79:161–65 20. Metlay JP, Witmer-Pack MD, Agger R, Crowley MT, Lawless D, Steinman RM. 1990. The distinct leukocyte integrins of mouse spleen dendritic cells as identified with new hamster monoclonal antibodies. J. Exp. Med. 171:1753–71 21. Kraal G, Breel M, Janse M, Bruin G. 1986. Langerhans’ cells, veiled cells, and interdigitating cells in the mouse recognized by a monoclonal antibody. J. Exp. Med. 163:981–97 22. Mehta-Damani A, Markowicz S, Engleman EG. 1995. Generation of antigenspecific CD4` T cell lines from naive precursors. Eur. J. Immunol. 25:1206–11 23. Mehta-Damani A, Markowicz S, Engleman EG. 1994. Generation of antigen-specific CD8` CTLs from naive precursors. J. Immunol. 153:996–1003 24. Steinman RM, Witmer MD. 1978. Lymphoid dendritic cells are potent stimulators of the primary mixed leukocyte reaction in mice. Proc. Natl. Acad. Sci. USA 75:5132–36 25. Inaba K, Steinman RM. 1986. Accessory cell-T lymphocyte interactions. Antigendependent and -independent clustering. J. Exp. Med. 163:247–61 26. Inaba K, Romani N, Steinman RM. 1989. An antigen-independent contact mechanism as an early step in T cell-proliferative responses to dendritic cells. J. Exp. Med. 170:527–42 27. Reis e Sousa C, Stahl PD, Austyn JM. 1993. Phagocytosis of antigens by Langerhans cells in vitro. J. Exp. Med. 178:509–19 28. Hart DN, McKenzie JL. 1988. Isolation and characterization of human tonsil dendritic cells. J. Exp. Med. 168:157–70 29. Lanzavecchia A. 1996. Mechanisms of antigen uptake for presentation. Curr. Opin. Immunol. 8:348–54


DENDRITIC CELLS IN IMMUNOTHERAPY 30. Norbury CC, Chambers BJ, Prescott AR, Ljunggren HG, Watts C. 1997. Constitutive macropinocytosis allows TAPdependent major histocompatibility complex class I presentation of exogenous soluble antigen by bone marrowderived dendritic cells. Eur. J. Immunol. 27:280–88 31. Arnold-Schild D, Hanau D, Spehner D, Schmid C, Rammensee HG, de la Salle H, Schild H. 1999. Cutting edge: receptor-mediated endocytosis of heat shock proteins by professional antigen-presenting cells. J. Immunol. 162:3757–60 32. Macatonia SE, Taylor PM, Knight SC, Askonas BA. 1989. Primary stimulation by dendritic cells induces antiviral proliferative and cytotoxic T cell responses in vitro. J. Exp. Med. 169:1255–64 33. Hsu FJ, Benike C, Fagnoni F, Liles TM, Czerwinski D, Taidi B, Engleman EG, Levy. R 1996. Vaccination of patients with B-cell lymphoma using autologous antigen-pulsed dendritic cells. Nat. Med. 2:52–58 34. Nestle FO, Alijagic S, Gilliet M, Sun Y, Grabbe S, Dummer R, Burg G, Schadendorf D. 1998. Vaccination of melanoma patients with peptide- or tumor lysate-pulsed dendritic cells. Nat. Med. 4:328–32 35. Palucka KA, Taquet N, Sanchez-Chapuis F, Gluckman JC. 1998. Dendritic cells as the terminal stage of monocyte differentiation. J. Immunol. 160:4587–95 36. Oehler L, Majdic O, Pickl WF, Stockl J, Riedl E, Drach J, Rappersberger K, Geissler K, Knapp W. 1998. Neutrophil granulocyte-committed cells can be driven to acquire dendritic cell characteristics. J. Exp. Med. 187:1019–28 37. Vremec D, Zorbas M, Scollay R, Saunders DJ, Ardavin CF, Wu L, Shortman K. 1992. The surface phenotype of dendritic cells purified from mouse thymus and spleen: investigation of the CD8 expression by a subpopulation of dendritic cells. J. Exp. Med. 176:47–58

267

38. Ardavin C, Wu L, Li CL, Shortman K. 1993. Thymic dendritic cells and T cells develop simultaneously in the thymus from a common precursor population. Nature 362:761–63 39. Wu L, Li CL, Shortman K. 1996. Thymic dendritic cell precursors: relationship to the T lymphocyte lineage and phenotype of the dendritic cell progeny. J. Exp. Med. 184:903–11 40. Suss G, Shortman K. 1996. A subclass of dendritic cells kills CD4 T cells via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183:1789–96 41. DiMolfetto L, Reilly C, Wei Q, Lo D. 1997. Dendritic-like cells from relB mutant mice. Adv. Exp. Med. Biol. 417:47–54 42. Burkly L, Hession C, Ogata L, Reilly C, Marconi LA, Olson D, Tizard R, Cate R, Lo D. 1995. Expression of relB is required for the development of thymic medulla and dendritic cells. Nature 373:531–36 43. Wu L, Nichogiannopoulou A, Shortman K, Georgopoulos K. 1997. Cell-autonomous defects in dendritic cell populations of Ikaros mutant mice point to a developmental relationship with the lymphoid lineage. Immunity 7:483–92 44. Pulendran B, Smith JL, Caspary G, Brasel K, Pettit D, Maraskovsky E, Maliszewski CR. 1999. Distinct dendritic cell subsets differentially regulate the class of immune response in vivo. Proc. Natl. Acad. Sci. USA 96:1036–41 45. Maldonado-Lopez R, De Smedt T, Michel P, Godfroid J, Pajak B, Heirman C, Thielemans K, Leo O, Urbain J, Moser M. 1999. CD8a` and CD8a1 subclasses of dendritic cells direct the development of distinct T helper cells in vivo. J. Exp. Med. 189:587–92 46. Galy A, Travis M, Cen D, Chen B. 1995. Human T, B, natural killer, and dendritic cells arise from a common bone marrow progenitor cell subset. Immunity 3:459– 73


268

FONG n ENGLEMAN

47. Marquez C, Trigueros C, Franco JM, Ramiro AR, Carrasco YR, Lopez-Botet M, Toribio ML. 1998. Identification of a common developmental pathway for thymic natural killer cells and dendritic cells. Blood 91:2760–71 48. Grouard G, Rissoan MC, Filgueira L, Durand I, Banchereau J, Liu YJ. 1997. The enigmatic plasmacytoid T cells develop into dendritic cells with interleukin (IL)-3 and CD40-ligand. J. Exp. Med. 185:1101–11 49. Olweus J, BitMansour A, Warnke R, Thompson PA, Carballido J, Picker LJ, Lund-Johansen F. 1997. Dendritic cell ontogeny: a human dendritic cell lineage of myeloid origin. Proc. Natl. Acad. Sci. USA. 94:12551–56 50. Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, Liu YJ. 1999. Reciprocal control of T helper cell and dendritic cell differentiation. Science 283:1183–86 51. Siegal FP, Kadowaki N, Shodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, Antonenko S, Liu YJ. 1999. The nature of the principal type 1 interferon-producing cells in human blood. Science 284:1835–37 52. Austyn JM, Larsen CP. 1990. Migration patterns of dendritic leukocytes. Implications for transplantation. Transplantation 49:1–7 53. Robert C, Fuhlbrigge RC, Kieffer JD, Ayehunie S, Hynes RO, Cheng G, Grabbe S, von Andrian UH, Kupper TS. 1999. Interaction of dendritic cells with skin endothelium: A new perspective on immunosurveillance. J. Exp. Med. 189:627–36 54. Greaves DR, Wang W, Dairaghi DJ, Dieu MC, Saint-Vis B, Franz-Bacon K, Rossi D, Caux C, McClanahan T, Gordon S, Zlotnik A, Schall TJ. 1997. CCR6, a CC chemokine receptor that interacts with macrophage inflammatory protein 3alpha and is highly expressed in human dendritic cells. J. Exp. Med. 186:837–44

55. Sozzani S, Allavena P, D’Amico G, Luini W, Bianchi G, Kataura M, Imai T, Yoshie O, Bonecchi R, Mantovani A. 1998. Cutting edge: differential regulation of chemokine receptors during dendritic cell maturation: a model for their trafficking properties. J. Immunol. 161:1083–86 56. Albert ML, Sauter B, Bhardwaj N. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class Irestricted CTLs. Nature 392:86–89 57. Svensson M, Stockinger B, Wick MJ. 1997. Bone marrow-derived dendritic cells can process bacteria for MHC-I and MHC-II presentation to T cells. J. Immunol. 158:4229–36 58. Pure E, Inaba K, Crowley MT, Tardelli L, Witmer-Pack MD, Ruberti G, Fathman G, Steinman RM. 1990. Antigen processing by epidermal Langerhans cells correlates with the level of biosynthesis of major histocompatibility complex class II molecules and expression of invariant chain. J. Exp. Med. 172:1459– 69 59. Dieu MC, Vanbervliet B, Vicari A, Bridon JM, Oldham E, Ait-Yahia S, Briere F, Zlotnik A, Lebecque S, Caux C. 1998. Selective recruitment of immature and mature dendritic cells by distinct chemokines expressed in different anatomic sites. J. Exp. Med. 188:373–86 60. Chan VW, Kothakota S, Rohan MC, Panganiban-Lustan L, Gardner JP, Wachowicz MS, Winter JA, Williams LT. 1999. Secondary lymphoid-tissue chemokine (SLC) is chemotactic for mature dendritic cells. Blood 93:3610–16 61. Adema GJ, Hartgers F, Verstraten R, de Vries E, Marland G, Menon S, Foster J, Xu Y, Nooyen P, McClanahan T, Bacon KB, Figdor CG. 1997. A dendritic-cellderived C-C chemokine that preferentially attracts naive T cells. Nature 387:713–17 62. Tang HL, Cyster JG. 1999. Chemokine up-regulation and activated T cell attrac-


63.

64.


65.

66.

67.

68.

69.

70.

71.

tion by maturing dendritic cells. Science 284:819–22 Zhou LJ, Tedder TF. 1995. A distinct pattern of cytokine gene expression by human CD83` blood dendritic cells. Blood 86:3295–3301 Armitage RJ, Namen AE, Sassenfeld HM, Grabstein KH. 1990. Regulation of human T cell proliferation by IL-7. J. Immunol. 144:938–41 Heufler C, Koch F, Stanzl U, Topar G, Wysocka M, Trinchieri G, Enk A, Steinman RM, Romani N, Schuler G. 1996. Interleukin-12 is produced by dendritic cells and mediates T helper 1 development as well as interferon-gamma production by T helper 1 cells. Eur. J. Immunol. 26:659–68 Yawalkar N, Brand CU, Braathen LR. 1996. IL-12 gene expression in human skin-derived CD1a` dendritic lymph cells. Arch. Dermatol. Res. 288:79–84 Steinman RM. 1991. The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9:271–96 Ridge JP, Di Rosa F, Matzinger P. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4` Thelper and a T-killer cell. Nature 393:474–78 Bennett SR, Carbone FR, Karamalis F, Flavell RA, Miller JF, Heath WR. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393:478–80 Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 393:480–83 Fernandez NC, Lozier A, Flament C, Ricciardi-Castagnoli P, Bellet D, Suter M, Perricaudet M, Tursz T, Maraskovsky E, Zitvogel L. 1999. Dendritic cells directly trigger NK cell functions: crosstalk relevant in innate anti-tumor immune responses in vivo. Nat. Med. 5:405–11

269

72. Steinman RM, Adams JC, Cohn ZA. 1975. Identification of a novel cell type in peripheral lymphoid organs of mice. IV. Identification and distribution in mouse spleen. J. Exp. Med. 141:804–20 73. Van Voorhis WC, Hair LS, Steinman RM, Kaplan G. 1982. Human dendritic cells. Enrichment and characterization from peripheral blood. J. Exp. Med. 155:1172–87 74. Inaba K, Inaba M, Romani N, Aya H, Deguchi M, Ikehara S, Muramatsu S, Steinman RM. 1992. Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 176:1693– 1702 75. Markowicz S, Engleman EG. 1990. Granulocyte-macrophage colony-stimulating factor promotes differentiation and survival of human peripheral blood dendritic cells in vitro. J. Clin. Invest. 85:955–61 76. Caux C, Dezutter-Dambuyant C, Schmitt D, Banchereau J. 1992. GM-CSF and TNF-alpha cooperate in the generation of dendritic Langerhans cells. Nature 360:258–61 77. Romani N, Gruner S, Brang D, Kampgen E, Lenz A, Trockenbacher B, Konwalinka G, Fritsch PO, Steinman RM, Schuler G. 1994. Proliferating dendritic cell progenitors in human blood. J. Exp. Med. 180:83–93 78. Bernhard H, Disis ML, Heimfeld S, Hand S, Gralow JR, Cheever MA. 1995. Generation of immunostimulatory dendritic cells from human CD34` hematopoietic progenitor cells of the bone marrow and peripheral blood. Cancer Res. 55:1099–1104 79. Sallusto F, Lanzavecchia A. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4

270

80.

81.


82.

83.

84.

85.

86.

87.

FONG n ENGLEMAN and downregulated by tumor necrosis factor alpha. J. Exp. Med. 179:1109–18 Rosenzwajg M, Canque B, Gluckman JC. 1996. Human dendritic cell differentiation pathway from CD34` hematopoietic precursor cells. Blood 87:535– 44 Maraskovsky E, Brasel K, Teepe M, Roux ER, Lyman SD, Shortman K, McKenna HJ. 1996. Dramatic increase in the numbers of functionally mature dendritic cells in Flt3 ligand-treated mice: multiple dendritic cell subpopulations identified. J. Exp. Med. 184:1953–62 Lotze MT, Hellerstedt B, Stolinski L, Tueting T, Wilson C, Kinzler D, Vu H, Rubin JT, Storkus W, Tahara H, Elder E, Whiteside T. 1997. The role of interleukin-2, interleukin-12, and dendritic cells in cancer therapy. Cancer J. Sci. Am. 3:S109 Fong L, Benike C, Brockstedt C, Strang G, Hao ZM, Smits B, Ruegg CL, Engleman EG. 1999. Immunization with dendritic cells pulsed with xenogeneic prostatic acid phosphatase administered via different routes induces cellular immune responses in prostate cancer patients. Proc. Am. Assoc. Cancer Res. 40:85 Fields RC, Shimizu K, Mule JJ. 1998. Murine dendritic cells pulsed with whole tumor lysates mediate potent antitumor immune responses in vitro and in vivo. Proc. Natl. Acad. Sci. USA 95:9482–87 Gong J, Chen D, Kashiwaba M, Kufe D. 1997. Induction of antitumor activity by immunization with fusions of dendritic and carcinoma cells. Nat. Med. 3:558–61 Gabrilovich DI, Nadaf S, Corak J, Berzofsky JA, Carbone DP. 1996. Dendritic cells in antitumor immune responses. II. Dendritic cells grown from bone marrow precursors, but not mature DC from tumor-bearing mice, are effective antigen carriers in the therapy of established tumors. Cell. Immunol. 170:111–19 Mayordomo JI, Loftus DJ, Sakamoto H,

88.

89.

90.

91.

92.

93.

De Cesare CM, Appasamy PM, Lotze MT, Storkus WJ, Appella E, DeLeo AB. 1996. Therapy of murine tumors with p53 wild-type and mutant sequence peptide-based vaccines. J. Exp. Med. 183:1357–65 Murray RJ, Kurilla MG, Brooks JM, Thomas WA, Rowe M, Kieff E, Rickinson AB. 1992. Identification of target antigens for the human cytotoxic T cell response to Epstein-Barr virus (EBV): implications for the immune control of EBV-positive malignancies. J. Exp. Med. 176:157–68 Kawashima I, Hudson SJ, Tsai V, Southwood S, Takesako K, Appella E, Sette A, Celis E. 1998. The multi-epitope approach for immunotherapy for cancer: identification of several CTL epitopes from various tumor-associated antigens expressed on solid epithelial tumors. Hum. Immunol. 59:1–14 Nair SK, Boczkowski D, Morse M, Cumming RI, Lyerly HK, Gilboa E. 1998. Induction of primary carcinoembryonic antigen (CEA)-specific cytotoxic T lymphocytes in vitro using human dendritic cells transfected with RNA. Nat. Biotechnol. 16:364–69 Lee PP, Yee C, Savage PA, Fong L, Brockstedt D, Weber JS, Johnson D, Swetter S, Thompson J, Greenberg PD, Roederer M, Davis MM. 1999. Characterization of circulating T cells specific for tumor-associated antigens in melanoma patients. Nat. Med. 5:677–85 Feltkamp MC, Vreugdenhil GR, Vierboom MP, Ras E, van der Burg SH, ter Schegget J, Melief CJ, Kast WM. 1995. Cytotoxic T lymphocytes raised against a subdominant epitope offered as a synthetic peptide eradicate human papillomavirus type 16-induced tumors. Eur. J. Immunol. 25:2638–42 Tsai V, Southwood S, Sidney J, Sakaguchi K, Kawakami Y, Appella E, Sette A, Celis E. 1997. Identification of subdominant CTL epitopes of the GP100 mela-


94.


95.

96.

97.

98.

99.

noma-associated tumor antigen by primary in vitro immunization with peptide-pulsed dendritic cells. J. Immunol. 158:1796–1802 Maryanski JL, Marchand M, Uyttenhove C, Boon T. 1983. Immunogenic variants obtained by mutagenesis of mouse mastocytoma P815. VI. Occasional escape from host rejection due to antigen-loss secondary variants. Int. J. Cancer. 31:119–23 Seung S, Urban JL, Schreiber H. 1993. A tumor escape variant that has lost one major histocompatibility complex class I restriction element induces specific CD8` T cells to an antigen that no longer serves as a target. J. Exp. Med. 178:933–40 Fong L, Ruegg CL, Brockstedt D, Engleman EG, Laus R. 1997. Induction of tissue-specific autoimmune prostatitis with prostatic acid phosphatase immunization: implications for immunotherapy of prostate cancer. J. Immunol. 159:3113–17 Overwijk WW, Tsung A, Irvine KR, Parkhurst MR, Goletz TJ, Tsung K, Carroll MW, Liu C, Moss B, Rosenberg SA, Restifo NP. 1998. gp100/pmel 17 is a murine tumor rejection antigen: induction of ‘‘Self’’- reactive, tumoricidal T cells using high-affinity, altered peptide ligand. J. Exp. Med. 188:277–86 Zaremba S, Barzaga E, Zhu M, Soares N, Tsang KY, Schlom J. 1997. Identification of an enhancer agonist cytotoxic T lymphocyte peptide from human carcinoembryonic antigen. Cancer Res. 57:4570–77 Rosenberg SA, Yang JC, Schwartzentruber DJ, Hwu P, Marincola FM, Topalian SL, Restifo NP, Dudley ME, Schwarz SL, Spiess PJ, Wunderlich JR, Parkhurst MR, Kawakami Y, Seipp CA, Einhorn JH, White DE. 1998. Immunologic and therapeutic evaluation of a synthetic peptide vaccine for the treatment

100.

101.

102.

103.

104.

105.

106.

107.

271

of patients with metastatic melanoma. Nat. Med. 4:321–27 Alters SE, Gadea JR, Sorich M, O’Donoghue G, Talib S, Philip R. 1998. Dendritic cells pulsed with CEA peptide induce CEA-specific CTL with restricted TCR repertoire. J. Immunother. 21:17–26 Tjoa BA, Erickson SJ, Bowes VA, Ragde H, Kenny GM, Cobb OE, Ireton RC, Troychak MJ, Boynton AL, Murphy GP. 1997. Follow-up evaluation of prostate cancer patients infused with autologous dendritic cells pulsed with PSMA peptides. Prostate 32:272–78 Kim DT, Mitchell DJ, Brockstedt DG, Fong L, Nolan GP, Fathman CG, Engleman EG, Rothbard JB. 1997. Introduction of soluble proteins into the MHC class I pathway by conjugation to an HIV tat peptide. J. Immunol. 159:1666–68 Goletz TJ, Klimpel KR, Leppla SH, Keith JM, Berzofsky JA. 1997. Delivery of antigens to the MHC class I pathway using bacterial toxins. Hum. Immunol. 54:129–36 Tuting T, Wilson CC, Martin DM, Kasamon YL, Rowles J, Ma DI, Slingluff CL Jr, Wagner SN, van der Bruggen P, Baar J, Lotze MT, Storkus WJ. 1998. Autologous human monocyte-derived dendritic cells genetically modified to express melanoma antigens elicit primary cytotoxic T cell responses in vitro: enhancement by cotransfection of genes encoding the Th1- biasing cytokines IL-12 and IFN-a. J. Immunol. 160:1139–47 Dietz AB, Vuk-Pavlovic S. 1998. High efficiency adenovirus-mediated gene transfer to human dendritic cells. Blood 91:392–98 Kim CJ, Cormier J, Roden M, Gritz L, Mazzara GP, Fetsch P, Hijazi Y, Lee KH, Rosenberg SA, Marincola FM. 1998. Use of recombinant poxviruses to stimulate anti-melanoma T cell reactivity. Ann. Surg. Oncol. 5: 64–76 Reeves ME, Royal RF, Lam JS, Rosenberg SA, Hwu P. 1996. Retroviral trans-

272

108.

109.


110.

111.

112.

113.

114.

115.

FONG n ENGLEMAN duction of human dendritic cells with a tumor- associated antigen gene. Cancer Res. 56:5672–77 Kafri T, Blomer U, Peterson DA, Gage FH, Verma IM. 1997. Sustained expression of genes delivered directly into liver and muscle by lentiviral vectors. Nat. Genet. 17:314–17 Johnston JC, Gasmi M, Lim LE, Elder JH, Yee JK, Jolly DJ, Campbell KP, Davidson BL, Sauter SL. 1999. Minimum requirements for efficient transduction of dividing and nondividing cells by feline immunodeficiency virus vectors. J. Virol. 73:4991–5000 Lynch, RG, Rohrer JW, Odermatt B, Gebel HM, Autry JR, Hoover RG. 1979. Immunoregulation of murine myeloma cell growth and differentiation: a monoclonal model of B cell differentiation. Immunol. Rev. 48:45–80 Hsu FJ, Kwak L, Campbell M, Liles T, Czerwinski D, Hart S, Syrengelas A, Miller R, Levy R. 1993. Clinical trials of idiotype-specific vaccine in B-cell lymphomas. Ann. NY Acad. Sci. 690:385–87 Maloney DG, Kaminski MS, Burowski D, Haimovich J, Levy R. 1985. Monoclonal anti-idiotype antibodies against the murine B cell lymphoma 38C13: characterization and use as probes for the biology of the tumor in vivo and in vitro. Hybridoma 4:191–209 Carroll WL, Lowder JN, Streifer R, Warnke R, Levy S, Levy R. 1986. Idiotype variant cell populations in patients with B cell lymphoma. J. Exp. Med. 164:1566–80 Kwak LW, Campbell MJ, Czerwinski DK, Hart S, Miller RA, Levy R. 1992. Induction of immune responses in patients with B-cell lymphoma against the surface-immunoglobulin idiotype expressed by their tumors. N. Engl. J. Med. 327:1209–15 Tjoa BA, Simmons SJ, Bowes VA, Ragde H, Rogers M, Elgamal A, Kenny GM, Cobb OE, Ireton RC, Troychak MJ,

116.

117.

118.

119.

120.

121.

122.

123.

Salgaller ML, Boynton AL, Murphy GP. 1998. Evaluation of phase I/II clinical trials in prostate cancer with dendritic cells and PSMA peptides. Prostate 36:39–44 Reichardt VL, Okada CY, Liso A, Benike CJ, Stockerl-Goldstein KE, Engleman EG, Blume KG, Levy R. 1999. Idiotype vaccination using dendritic cells after autologous peripheral blood stem cell transplantation for multiple myeloma–a feasibility study. Blood 93:2411–19 Lynch DH, Andreasen A, Maraskovsky E, Whitmore J, Miller RE, Schuh JC. 1997. Flt3 ligand induces tumor regression and antitumor immune responses in vivo. Nat. Med. 3:625–31 Chen B, Shi Y, Smith JD, Choi D, Geiger JD, Mule JJ. 1998. The role of tumor necrosis factor alpha in modulating the quantity of peripheral blood-derived, cytokine-driven human dendritic cells and its role in enhancing the quality of dendritic cell function in presenting soluble antigens to CD4` T cells in vitro. Blood 91:4652–61 Morse MA, Lyerly HK, Gilboa E, Thomas E, Nair SK. 1998. Optimization of the sequence of antigen loading and CD40-ligand-induced maturation of dendritic cells. Cancer Res. 58:2965–68 Gabrilovich DI, Cunningham HT, Carbone DP. 1996. IL-12 and mutant P53 peptide-pulsed dendritic cells for the specific immunotherapy of cancer. J. Immunother. Emphasis Tumor Immunol. 19:414–18 Shimizu K, Fields RC, Giedlin M, Mule JJ. 1999. Systemic administration of interleukin 2 enhances the therapeutic efficacy of dendritic cell-based tumor vaccines. Proc. Natl. Acad. Sci. USA 96:2268–73 Gabrilovich DI, Ciernik IF, Carbone DP. 1996. Dendritic cells in antitumor immune responses. I. Defective antigen presentation in tumor-bearing hosts. Cell Immunol. 170:101–10 Gabrilovich DI, Chen HL, Girgis KR,


124.


125.

126.

127.

Cunningham HT, Meny GM, Nadaf S, Kavanaugh D, Carbone DP. 1996. Production of vascular endothelial growth factor by human tumors inhibits the functional maturation of dendritic cells [published erratum appears in Nat. Med. 1996 Nov;2(11):1267]. Nat. Med. 2:1096–103 Steinbrink K, Wolfl M, Jonuleit H, Knop J, Enk AH. 1997. Induction of tolerance by IL-10-treated dendritic cells. J. Immunol. 159:4772–80 Qin Z, Noffz G, Mohaupt M, Blankenstein T. 1997. Interleukin-10 prevents dendritic cell accumulation and vaccination with granulocyte-macrophage colony-stimulating factor gene-modified tumor cells. J. Immunol. 159:770–76 Piali L, Fichtel A, Terpe HJ, Imhof BA, Gisler RH. 1995. Endothelial vascular cell adhesion molecule 1 expression is suppressed by melanoma and carcinoma. J. Exp. Med. 181:811–16 Nestle FO, Nickoloff BJ. 1994. Role of dendritic cells in benign and malignant

273

lymphocytic infiltrates of the skin. Dermatol. Clin. 12:271–82 128. Tsujitani S, Kakeji Y, Watanabe A, Kohnoe S, Maehara Y, Sugimachi K. 1990. Infiltration of dendritic cells in relation to tumor invasion and lymph node metastasis in human gastric cancer. Cancer 66:2012–16 129. Salgaller ML, Lodge PA, Tjoa BA, McLean JG, Ragde H, Kenny RJ, Rogers M, Boynton AL, Murphy GP. 1998. Monitoring of prostate-specific membrane antigen- (PSMA) specific immune responses and prostate markers in a phase II clinical trial with patients infused with dendritic cells pulsed with PSMA-derived peptides. Proc. Am. Assoc. Cancer Res. 39:173 130. Valone FH, Small EJ, Whisenant S, Peshwa MV, Strang G, Laus R, van Schooten WCA. 1998. Antigen-pulsed dendritic cell therapy for hormone refractory prostate cancer: A phase I/II trial. Proc. Am. Assoc. Cancer Res. 39:173



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


CD8~ T CELL EFFECTOR MECHANISMS IN RESISTANCE TO INFECTION John T. Harty, Amy R. Tvinnereim, and Douglas W. White Department of Microbiology and Interdisciplinary Graduate Program in Immunology, University of Iowa, Iowa City, Iowa, e-mail: [email protected], [email protected], [email protected]


Key Words CD8` T cells, cytolysis, cytokines, pathogens, infection Abstract Based on T cell subset depletion studies and the analysis of gene knockout mice, it is evident that CD8` T cells contribute to resistance against intracellular infections with certain viral, protozoan, and bacterial pathogens. Although they are known primarily for their capacity to kill infected cells, CD8` T cells elaborate a variety of effector mechanisms with the potential to defend against infection. Microbes use multiple strategies to cause infection, and the nature of the pathogenhost interaction may determine which CD8` T cell effector mechanisms are required for immunity. In this review, we summarize our current understanding of the effector functions used by CD8` T cells in resistance to pathogens. Analyses of mice deficient in perforin and/or Fas demonstrate that cytolysis is critical for immunity against some, but not all, infections and also reveal the contribution of cytolysis to the pathogenesis of disease. The role of CD8` T cell–derived cytokines in resistance to infection has been analyzed by systemic treatment with neutralizing antibodies and cytokine gene knockout mice. These studies are complicated by the fact that few, if any, cytokines are uniquely produced by CD8` T cells. Thus, the requirement for CD8` T cell– derived cytokines in resistance against most pathogens remains to be defined. Finally, recent studies of human CD8` T cells reveal the potential for novel effector mechanisms in resistance to infection.

STIMULATION OF PATHOGEN-SPECIFIC CD8` T CELL RESPONSES CD8` T lymphocytes are important mediators of adaptive immunity against certain viral, protozoan, and bacterial pathogens. During the initial encounter with a microbe, CD8` T cells bearing TCR specific for pathogen-derived antigens are selected to undergo clonal expansion. The specificity of this selection is driven by the interaction of the TCR with short (8–10 amino acid) pathogen-derived peptides presented by MHC class I molecules on the surface of professional 0732–0582/00/0410–0275$12.00

275


276

HARTY n TVINNEREIM n WHITE

antigen-presenting cells (APC) (1). APC such as dendritic cells (DC), by virtue of their expression of specific costimulatory molecules, are capable of stimulating naive T cells to proliferate and differentiate in response to antigen (2). Priming of CD8` T cells occurs in response to antigens that gain access to the cytosol of APC for processing by the endogenous MHC class I presentation pathway (3). Thus, CD8` T cells are activated in response to cytosolic infections with viruses, intracytoplasmic bacteria, and protozoa. CD8` T cells also respond to antigens from noncytosolic sources such as phagosomal-bacterial and -protozoan pathogens; these antigens are processed by exogenous MHC class I presentation pathways that may be limited to professional APC (4). Recent evidence demonstrates that DC are also able to ‘‘cross present’’ MHC class I–restricted antigens obtained from apoptotic, infected cells (5). New technologies for detection of antigen-specific CD8` T cells, including staining with tetrameric MHC class I–peptide complexes, intracellular cytokine staining, and ELISPOT analysis, have revised our understanding of the magnitude of CD8` T cell responses to infection (6). As a result of clonal expansion, pathogen-specific CD8` T cells rapidly increase from virtually undetectable in the naive host to levels that are readily detectable (1–2% of splenic CD8` T cells in primary responses to certain bacterial infections) (7) or even dominate the repertoire, reaching 50% or more of splenic CD8` T cells in the primary response to certain viral infections (8, 9). These expanded populations of ‘‘effector’’ CD8` T cells contribute to clearance of the pathogen and then decline in numbers to a memory level that may be maintained at 5–10% of the initial clonal burst size (8). Memory CD8` T cells may be present for the life of the host and are able to mount rapid, heightened responses to reinfection with the specific pathogen (10). Experiments to understand the signals that maintain the memory pool are ongoing, with evidence for and against such mediators as persistent antigen (11, 12), specific cytokines (13), or MHC molecules (14).

CD8` T CELL EFFECTOR MECHANISMS Concurrent with the initiation of proliferation is the instigation of a program of gene expression that arms the CD8` T cell with a veritable arsenal of effector mechanisms to combat the infection. The capacity to mobilize these effector mechanisms, which include cytolysis of infected cells and the production of cytokines, chemokines, or microbicidal molecules, develops over time after the initial stimulation of naive CD8` T cells and manifests upon subsequent encounter of the effector CD8` T cells with infected cells. Since MHC class I molecules are expressed on essentially all nucleated cells, effector CD8` T cells can respond against intracellular pathogens in most tissues and facilitate clearance of the infection. In the case of memory CD8` T cells, mobilization of effector mechanisms is very rapid (15, 16), occurs in response to reduced antigen levels, perhaps due to the increased expression of adhesion molecules on memory T cells (10), and


CD8` T CELL EFFECTOR MECHANISMS

277

is less dependent on the costimulatory signals required for priming antigenspecific naive T cells (17). Activated CD8` T cells are able to induce cytolysis of infected cells by two distinct molecular pathways (18): the granule exocytosis pathway, dependent on the pore-forming molecule perforin, or by the upregulation of FasL (CD95L), which can initiate programmed cell death by aggregation of Fas (CD95) on target cells. Both of these pathways, activated in response to signals from the TCR, stimulate the caspase cascade in the target cell, leading to apoptotic death (19). Efficient lysis by the granule exocytosis pathway requires the coordinated delivery of perforin and granule enzymes, such as granzymes A and B, into the target cell (20, 21). CD8` T cells also elaborate cytokines, including IFN-c and TNF, as well as chemokines that function to recruit and/or activate the microbicidal activities of effector cells such as macrophages and neutrophils (22). Cytokines may also directly interfere with pathogen attachment or pathogen gene expression, or they may restrict intracellular replication. As with cytolytic effector mechanisms, expression of cytokine molecules by CD8` T cells is tightly regulated through TCR-dependent signals.

UNDERSTANDING THE ROLE OF SPECIFIC CD8` T CELL EFFECTOR MECHANISMS It is important to note that many of the effector mechanisms employed by CD8` T cells are also employed by other cells of the immune system. Natural killer (NK) cells and at least some MHC class II–restricted CD4` T cells express perforin and/or FasL and are capable of cytolysis. TH1-type CD4` T cells and NK cells produce IFN-c as do cd T cells. TNF is produced by many cell types, including CD8` T cells, and has a variety of effects on cells expressing one or both of the TNF receptors, ranging from activation to death (23). In addition to their potential as CD8` T cell effector mechanisms most, if not all, of these molecules play regulatory roles ranging from T cell homeostasis to enhancement of antigen presentation. Given the complexity of the CD8` T cell arsenal, the non-exclusive expression of these molecules and their pleiotrophic effects on the immune system, understanding which mechanisms are important in CD8` T cell– mediated resistance to infection is a substantial challenge. In the last several years, gene knockout mice, lacking specific CD8` T cell effector molecules, have become important tools in addressing this significant biological question. In this review, we summarize the results from selected studies of experimental infections of mice with viral, protozoan, and bacterial pathogens as well as from recent experiments in human systems describing potentially novel CD8` T cell effector mechanisms in resistance to infection.

278



CD8` T CELLS IN RESISTANCE TO VIRAL INFECTIONS Viruses are obligate intracellular pathogens that depend on host cell molecules to complete their replication cycles. They range in complexity from less than ten to several hundred potential open reading frames, and they infect an enormous range of eukaryotic cells and a variety of tissues. The results of infection include destruction of the host cell (lytic), replication in the absence of host cell damage (nonlytic), or an interaction where viral gene expression is limited or absent (latent) but reactivation can occur. Some viruses exhibit exquisite host specificity, while other viruses are capable of infecting any mammalian cell. Thus, viruses exhibit both shared and unique interactions with their host and, as a group, present a complex challenge to the immune response. By virtue of their dependence on host protein synthesis machinery, all viral proteins are potentially accessible to the endogenous MHC class I–presentation pathway and to the attention of CD8` T cells. Despite the obvious potential for CD8` T cells to provide immunity against viruses, these cells are not always the dominant effectors of antiviral immunity. However, a role for CD8` T cells in resistance to at least some viruses is clearly documented in animal models and is further supported by the finding that several viruses express molecules that specifically interfere with MHC class I antigen presentation (24) or CD8` T cell– induced cell death (25). Given the limited genome size of even the largest viruses, it would be remarkable if such molecules were maintained in the absence of an important biological function.

ROLE OF CYTOLYSIS IN RESISTANCE TO LYTIC AND NONLYTIC VIRAL INFECTION A major attribute of CD8` T cells is their ability to kill target cells expressing specific MHC class I–peptide complexes. Cytolysis was recognized as potentially significant in resistance to viruses because lysis of the infected cell should eliminate the factory for viral production and contribute to resolution of the infection. Experimental support for this notion was convincingly presented upon analysis of perforin gene knockout (PKO) mice, which lack the granule exocytosis pathway of cytolysis (26–28). This early work, which was elegantly summarized in a prior volume of the Annual Review of Immunology (29), revealed that perforin, but not Fas-dependent cytolysis, was critical in clearance of lymphocytic choriomeningitis virus (LCMV), a nonlytic virus that is eliminated by a CD8` T cell–dependent mechanism (26, 27). Interestingly, another study demonstrated that neither perforin nor Fas-mediated cytolysis was required for resistance to vaccinia virus, vesicular stomatitis virus, or Semliki Forest virus–lytic viral infections where antibodies, CD4` T cells, and cytokines also contribute significantly



279

to viral clearance (30). These results led to the hypothesis that CD8` T cell– mediated cytolysis may be required for immunity to nonlytic viruses but is not a necessary effector mechanism in resistance to those viruses that lyse their host cell (29). The requirement for perforin-dependent cytolysis in viral clearance has subsequently been analyzed in several murine models. Control of primary influenza virus infection, a lytic infection of lung epithelial cells, involves CD8` T cells (31). Immunity to reinfection, in both mice and humans, requires the participation of antibodies. Using CD4` T cell depletion to prevent a vigorous antibody response, and a series of bone marrow and splenocyte chimeric mice to control perforin expression in CD8` T cells or Fas expression on lung epithelia, Topham et al demonstrated that perforin-deficient CD8` T cells clear influenza virus from Fas-positive epithelia but fail to clear virus from Fas-minus epithelia (32). In addition, wild-type CD8` T cells were able to clear influenza infection from Fasminus epithelia. Thus, either perforin- or Fas-mediated cytolysis was required and sufficient to eliminate primary influenza infection from the lung, indicating that CD8` T cell control of lytic infection with influenza virus is dependent on cytolysis. CD8` T cells are critical in recovery of mice from infection with ectromelia virus, a murine-specific orthopoxvirus that causes lytic infection. Wild-type B6 mice survive footpad infection with doses up to 106 PFU of ectromelia virus. In contrast, B6-PKO mice and B6-PKO/granzyme A double deficient mice succumb to infection with as little as 10 PFU (33). Mortality was attributed to the failure of the B6-PKO mice to control infection, since levels of virus were substantially elevated at 8 days post infection compared to wild-type mice. Consistent with previous studies (30), the absence of perforin had little impact on resistance to cowpox virus, which exhibits less virulence for mice than does ectromelia virus. These results demonstrate that perforin-dependent cytolysis is critical in resistance to lytic infection with ectromelia virus, a highly virulent natural mouse pathogen. One of the proposed functions of perforin is to facilitate the delivery of granzymes into the target cell (34). Granzyme B (GrmB) activates the caspase cascade in target cells and induces apoptosis associated with rapid DNA fragmentation (20). GrmB-deficient CD8` T cells are unable to induce rapid DNA fragmentation and apoptosis of target cells. However, GrmB-deficient CD8` T cells will eventually induce DNA fragmentation. Recently, data have been presented that Granzyme A (GrmA) also induces target cell DNA fragmentation in a caspaseindependent fashion, but with delayed kinetics compared to GrmB (35, 36). Consistent with this information are studies indicating a severe block in cytolysis in GrmA x B double-deficient mice, similar to that observed with perforin deficiency (35). These results indicate that perforin, delivered by CD8` T cells, is not sufficient for cell lysis and that perforin-induced membrane alterations may be repaired in the absence of granzymes. GrmA-deficient mice, in contrast to perforin-deficient mice, do not exhibit defects in clearance of LCMV (21). However, they do exhibit a defect in clearance of ectromelia virus (37). This result may


280


involve the expression of protease inhibitors (serpins) by the ectromelia virus, which could inhibit GrmB-mediated activation of the caspase cascade (25). Thus, GrmA may be required for destruction of cells infected with ectromelia virus or other pathogens that are able to block the GrmB-induced caspase cascade (35, 36). Taken together, these results do not support the hypothesis that CD8` T cell– mediated cytolysis is an unnecessary effector mechanism for resistance to lytic viruses. This hypothesis was originally based on analysis of three lytic viruses, vaccinia virus, vesicular stomatitus virus, and Semliki Forest virus, which are not highly virulent, natural pathogens of mice. Consistent with these results, perforin was not required for resistance to cowpox virus infection of mice. In contrast, perforin-dependent cytolysis was clearly required for resistance to lytic infection with the highly virulent mouse poxvirus, ectromelia virus. In addition, either perforin or Fas-mediated cytolysis was required for clearance of lytic infection of lung epithelia with influenza virus, in the absence of CD4` T cells. Thus, either perforin- or Fas-dependent cytolysis may be required to resist highly virulent infection with host-adapted viruses or in situations where CD8` T cells are the sole line of defense. The converse hypothesis, that CD8` T cell–mediated cytolysis is required for clearance of nonlytic viruses, requires further analysis. However, recent data in the LCMV system, described below, suggest that noncytolytic mechanisms may also be effective against this nonlytic virus (38). Finally, it should be noted that the designation of viruses as lytic or nonlytic is generally based on in vitro analysis in cell culture. For instance, the fact that LCMV is lytic for some cell types is the basis for the plaque assay to quantify the virus (39). Thus, lytic or nonlytic designations from in vitro analyses may not reflect the true nature of viral infection in vivo.

PERFORIN-DEPENDENT CYTOLYSIS: PROTECTION, IMMUNOPATHOLOGY, AND HOMEOSTASIS Intracranial infection of adult B6 mice with LCMV results in lethal CNS disease mediated by CD8` T cells (40). In contrast, B6-PKO mice do not develop severe CNS disease after intracranial infection (26). Thus, perforin-dependent cytolysis plays an important role in the immunopathogenesis of LCMV-induced CNS disease. Interestingly, a fraction of perforin-deficient mice succumb to peripheral infection with LCMV (26, 27), indicating that the failure to clear this normally nonlethal infection can be fatal. Infection of resistant B6 mice with Theiler’s virus (TMEV), a murine picorna virus, results in an acute encephalitis that is cleared by a CD8` T cell-dependent mechanism in approximately 10 days post infection (41). TMEV is lytic for murine fibroblasts, but viral replication is restricted in other cell types such as



281

macrophages and may be restricted in the CNS as well (42). Intracranial infection of B6-PKO mice with 2 x 104 PFU of TMEV resulted in death within 20 days (43). Mortality was associated with a substantial viral load in the CNS of B6PKO mice at 11 days post infection, a time when virus is cleared from infected B6 mice. Thus, perforin appears to be essential for control of acute CNS infection by TMEV. Studies by another group revealed that intracranial infection of B6PKO mice with 2 x 106 PFU of TMEV resulted in a persistent infection under conditions where the infection was eliminated in B6 mice (44). Interestingly, although these B6-PKO animals exhibit some demyelination, they did not exhibit overt clinical symptoms. The extent of demyelination in these animals and the precise relationship between demylination and overt disease remain to be determined. These data, in conjunction with the finding that b2M-deficient mice develop persistent CNS infection and demyelination without overt CNS disease (45), suggest that perforin-dependent, CD8` T cell-mediated cytolysis contributes to both the clearance of TMEV and the clinical symptoms of CNS disease. The reason for the differences in mortality after infections with the same strain of virus is unknown. Interestingly, survival correlated with the highest challenge dose of virus. This result may be due to small differences in the virus strains, CD8` T cell exhaustion (46), or high-dose intracranial TMEV infection, which may induce a peripheral infection that inhibits migration of CD8` T cell into the CNS, as observed with LCMV (47). Intracranial infection of B6 mice with the neurotropic JHM virus, a member of the mouse hepatitis virus group of coronaviruses, results in demyelination that is at least partially dependent on virus-specific CD8` T cells (48). JHMV is rapidly cytolytic in vitro, but evidence for in vivo cytopathology is lacking. Perforin-deficient mice, infected intracranially with doses of JHMV that are lethal for wild-type mice, exhibit a moderate reduction in morbidity and extended survival compared to wild-type mice, although all perforin-deficient mice eventually succumb to infection (49). CNS viral loads were elevated, and clearance of virus from the CNS was delayed in the perforin-deficient mice. These results suggest that perforin-dependent cytolysis contributes to the efficiency of virus clearance but also contributes to clinical disease. CD8` T cells play a role in clearance of cardiopathogenic strains of the picornavirus, coxsackievirus B3, and also in the pathology of viral myocarditis (50). In contrast to wild-type B6 mice, which die from myocarditis within 10 days post infection, all perforin-deficient mice survive the infection (51). Inflammatory heart lesions were present in infected perforin-deficient mice, but at a reduced level compared to wild-type mice. Perforin-deficiency had no impact on viral replication in heart tissue or clearance of the virus from the host. Thus, perforindependent cytolysis contributes significantly to the immunopathology of myocarditis but not to the clearance of the inducing virus. CD8` T cell-derived perforin may also play a role in T cell homeostasis. B6 mice infected with LCMV clone 13 develop chronic infections associated with deletion of virus-specific CD8` T cells (52). Perforin-deficient mice develop


282


chronic infections with LCMV clone 13 that are indistinguishable, based on the titers of virus in tissue, from those seen in B6 mice (53). However, deletion of LCMV clone 13–specific CD8` T cells in perforin-deficient mice is less efficient than in B6 mice, and a substantial fraction of the B6-PKO mice die within 2–4 weeks after LCMV clone 13 infection. This mortality is reversed by depletion of CD8` T cells. Similar results were observed in a model of high dose LCMV infection where deletion of CD8` T cells was less efficient in B6-PKO mice (46). Together, the results suggest that perforin plays a regulatory role in CD8` T cell homeostasis that is apparently independent of antigen levels, at least in LCMV clone 13 infections. Absence of this regulatory mechanism leads to death, perhaps related to the chronic activation of CD8` T cells. While the mechanism of action of perforin in this situation is undefined, the authors speculate that CD8` T cells may develop susceptibility to perforin-mediated damage after a certain upper limit of cell division (53). However, it is equally possible that other constituents of the cytotoxic granules are important in T cell homeostasis during chronic viral infection and that perforin is required for their function. Resolution of this interesting issue will require further investigation. These studies reveal the capacity of perforin-dependent cytolytic mechanisms to exacerbate disease, including cardiac and CNS pathology, after viral infection. Thus, perforin-dependent cytolysis is a two-edged sword, involved in viral clearance but also capable of mediating pathology. In addition, intriguing evidence suggests that the same granule exocytosis pathway may participate in CD8` T cell homeostasis during chronic viral infection.

RESISTANCE TO VIRUSES IN THE ABSENCE OF PERFORIN As mentioned above, perforin-dependent cytolysis was not required for resistance to lytic infections with vaccinia virus, Semliki Forest virus, vesicular stomatitis virus, or cowpox virus (30, 33). CD8` T cells are also required for clearance of acute lung (lytic) infection with murine gamma herpesvirus 68 (MHV-68) and to control latently infected B cells in this natural mouse infection model, with similarities to Epstein-Barr virus infections (54). Intranasal infection of B6 and B6PKO mice with MHV-68 resulted in similar kinetics of viral clearance in the lung, indicating that perforin-mediated cytolysis is not required to clear the acute lytic infection (55). In addition, levels of virus in the spleen up to 21 days post infection were not significantly different between B6 and B6-PKO mice, indicating that establishment of the latent infection was perforin independent. These results, together with the results with coxsackievirus B3 (51), indicate that perforindependent cytolysis is not required for resistance to all murine-adapted viruses. Similarly, IFN-c-deficient mice cleared lung infections and developed latent infection with MHV-68 in a manner that was only slightly, if at all, deficient compared to wild-type mice (56). In contrast, while IFN-c receptor–deficient mice



283

clear lung infections with MHV-68, establishment of the chronic infection is delayed and the mice develop an unusual spleen pathology that is reversed upon depletion of CD8` T cells (57). This discrepancy in the nature of infection in IFN-c and IFN-c receptor–deficient mice is potentially interesting and may suggest the existence of alternate ligands for the IFN-c receptor. Consistent with this notion, herpes simplex virus infection is more virulent in IFN-c receptor–deficient mice than in IFN-c-deficient mice (58). The long-term control of latency or the sufficiency of Fas-mediated cytolysis for control of lytic or latent infections with MHV-68 was not addressed. This issue is of interest because Fas-mediated cytolysis clears influenza virus from lung tissue (32). CD8` T cells contribute to clearance of primary rotavirus infection in B cell– deficient mice as demonstrated by development of chronic infections after CD8` T cell depletion (59, 60). In CD8` T cell–depleted mice, or in b2M-deficient mice, clearance of rotavirus is slightly delayed compared to wild-type mice (60). Perforin- and Fas-deficient mice clear rotavirus infection with the same kinetics as wild-type mice, and infection of these mice is similarly exacerbated by CD8` T cell depletion (61). Perforin-deficient mice also clear rotavirus infection when treated with neutralizing anti-mouse IFN-c, and perforin-deficient CD8` T cells clear virus from chronically infected Rag-2 deficient mice with kinetics similar to those of wild-type CD8` T cells. While these results clearly demonstrate a contribution of CD8` T cells in resistance to viral infection in the absence of perforin, Fas, or IFN-c, they are also potentially consistent with the relevance of either perforin- or Fas-dependent lysis as described in the influenza system (32). However, it is also clear that CD8` T cells are not the major mediators of rotavirus immunity in wild-type mice, and thus, it may not be surprising that no single CD8` T cell effector mechanism is required for clearance. Despite the evidence that CD8` T cells participate in control of these viruses, perforin-dependent cytolysis is not required and does not contribute in a detectable fashion to control of a murine-adapted virus, MHV-68, as well as coxsackievirus, rotavirus, vaccinia virus, Semliki Forest virus, vesicular stomatitis virus, and cowpox virus. In the majority of these studies, the role of Fas-dependent cytolysis in the absence of perforin remains to be examined. Each of these viruses is capable of lytic infection, but they represent a broad spectrum in the virulence they bring to mouse infection, and evidence exits that CD8` T cell–independent mechanisms also participate in clearance of infection. Further studies are required to determine which, if any, of the known CD8` T cell effector mechanisms are involved in clearance of these pathogens.

CD8` T CELL CYTOKINES IN RESISTANCE TO VIRAL INFECTION In addition to their capacity to induce cytolysis of infected cells, effector and memory CD8` T cells produce cytokines such as IFN-c and TNF after stimulation through their antigen receptors. A role for CD8` T cell–derived IFN-c in


284


resistance to LCMV infection was suggested by the observation that viral clearance was delayed by several days in IFN-c receptor deficient mice compared to controls (62). The IFN-c receptor–deficient mice mounted an LCMV-specific CD8` T cell response that was similar in magnitude to that observed in wildtype mice but which reached peak levels with delayed kinetics, associated with the delayed clearance of the infection. However, in contrast to perforin-deficient mice that fail to clear LCMV and exhibit altered homeostasis of the CD8` T cell compartment (26, 27, 53), IFN-c receptor–deficient mice eventually clear the infection, and homeostasis of the CD8` T cell compartment appears normal. Similarly, IFN-c-deficient mice eliminate acute LCMV infection within 12 days and develop significant CD8` T cell responses (63). Although these studies do not formally demonstrate that CD8` T cell derived IFN-c participates in resolution of LCMV infection they clearly indicate that IFN-c is not a required CD8` T cell effector mechanism for clearance of this virus. Infection of newborn mice with LCMV causes a persistent, lifelong infection (40). Transfer of LCMV-specific memory CD8` T cells, in the form of immune splenocytes, into persistently infected wild-type mice results in clearance of the infection from most tissues by 21 days post transfer. Although clearance is absolutely dependent on CD8` T cells, effective immunotherapy requires both CD8` and CD4` T cells (52). Interestingly, LCMV-immune splenocytes from IFN-cdeficient mice fail to clear the persistent infection from wild-type mice (63). This result suggests that IFN-c from CD4` or CD8` T cells is critical for clearance of the persistent infection, perhaps through its immune regulatory functions. However, recent work demonstrated that CD8` T cells, but not CD4` T cells, from IFN-c-deficient mice survive after transfer into Listeria monocytogenes–infected SCID mice (64). Since clearance of chronic LCMV infection requires both CD4` and CD8` T cells, a role for IFN-c in survival of CD4` T cells could also explain the failure to clear chronic infection. This issue could be resolved by mixing experiments with wild-type and IFN-c-deficient T cell populations. As described above, clearance of JHMV from the CNS is at least partially dependent on CD8` T cell–derived perforin (49). IFN-c-deficient mice also exhibit delayed clearance of JHMV, more severe paralysis, and increased mortality in comparison to wild-type mice (65). IFN-c-deficient mice develop virusspecific CD8` T cells responses at levels similar to those of wild-type mice. Interestingly, although the IFN-c-deficient mice clear virus from most CNS cell types such as neurons, microglia, and astrocytes, they fail to clear virus from oligodendrocytes. This finding may reflect the fact that oligodendrocytes apparently lack MHC class I molecules but do express the IFN-c receptor (66). Viral hepatitis, caused by hepatitis B virus (HBV), is thought to occur through the destruction of infected hepatocytes by MHC class I–restricted CD8` T cells (67). Interestingly, essentially all hepatocytes are infected, yet viral clearance and clinical recovery occur without destruction of the entire liver. Using a transgenic model, in which murine hepatocytes express the entire HBV genome, evidence was provided for a nonlytic mechanism by which CD8` T cells are able to



285

extinguish viral gene expression (68). Transfer of MHC class I–restricted antiHBV CD8` T cell clones into HBV transgenic mice eliminated expression of the viral genome with minimal hepatocyte death. HBV-specific CD8` T cell clones from perforin-deficient mice, which caused no discernible hepatocyte death, also mediated the effect. The ability of wild-type CD8` T cells to inhibit viral gene expression was abrogated by neutralization of both IFN-c and TNF and was partially blocked when either cytokine was neutralized. Although these results demonstrate the participation of CD8` T cells, IFN-c, and TNF in the shutdown of viral gene expression, they do not conclusively demonstrate that CD8` T cells are the source of the relevant cytokines, both of which are made by multiple cell types. This issue is relevant to understanding the specificity and mechanism of viral clearance because infection of the same HBV transgenic mice with LCMV, adenovirus, and cytomegalovirus also results in elimination of HBV gene expression (69, 70). Thus, HBV clearance can be independent of CD8` T cell recognition of HBV-infected cells. This model has now been extended to primates infected with HBV, which clear the virus without substantial liver damage and prior to extensive T cell infiltration (71) and persistent LCMV infection, initiated by infection of newborn mice (38). The latter mice harbor LCMV in hepatocytes and Kupffer cells of the liver. As described above, transfer of LCMV-immune CD8` T cells into persistently infected mice causes clearance of infection unless the transferred cells do not express IFN-c (63). CD8` T cell–mediated clearance of LCMV from persistently infected mice occurs in the absence of extensive liver damage. Interestingly, viral RNA and antigen were rapidly cleared from hepatocytes but only slowly cleared from nonparenchymal Kupffer cells. Administration of IL-12 or infection with adenovirus also resulted in clearance of LCMV infection of hepatocytes but not nonparenchymal cells. These data suggest that cytolysis is essential for complete clearance of virus, specifically from those cells that derive from the regenerating macrophage pool. In contrast, viral clearance from hepatocytes appeared to result from a nonlytic mechanism. The authors speculate that such nonlytic mechanisms may be operative to clear infections with extensive involvement of essential tissues such as the liver (38). CD8` T cells are thought to be important in control of HIV infection of humans, although the effector mechanisms involved in viral clearance remain largely undefined (72). However, it was recognized in the 1980s that CD8` T cells from HIV-infected individuals secreted a factor or factors that inhibit HIV replication in vitro, in the absence of cytolysis, and it was suggested that these factors may contribute to control of infection in vivo (73). Shortly before the demonstration by multiple groups that b-chemokine receptors are coreceptors for HIV (74), a study showed that the b-chemokines RANTES, MIP-1a, and MIP1b could be purified from the supernatants of transformed CD8` T cells and inhibit HIV replication in a T cell line in vitro (75). In this study, neutralization of these b-chemokines with an antibody cocktail completely abrogated the ability of CD8` T cell supernatants, from transformed cells and 3 of 4 HIV-infected donors, to inhibit HIV replication. From these data and other similar reports, it


286


was concluded that the nonlytic inhibition of HIV replication in vitro was a consequence of interference with HIV binding to the b-chemokine coreceptors. However, addition of antibodies that neutralize these same b-chemokines did not abrogate the ability of CD8` T cells to inhibit HIV replication in macrophages (76) or CD4` T cell/dendritic cell cocultures (77). These results suggest that novel CD8` T cell–derived anti-HIV factors remain to be identified. The discrepancy in the results of these studies may relate to the virus strain and the target cell types analyzed; further studies will be needed to resolve this issue. Together, these studies reveal the potential of CD8` T cell–derived soluble factors (cytokines and chemokines) to influence viral infection in the absence of cytolysis. Signaling through cytokine receptors for IFN-c and TNF results in altered gene expression in the responding cell (78, 79). Since viruses utilize host molecules to transcribe and translate their gene products, it is possible that cytokines can directly inhibit viral infection by influencing host factors required for viral gene expression. Such nonlytic mechanisms to eliminate viruses could potentially lessen the degree of immunopathology in response to viral infection of essential tissues such as the liver or CNS. Despite the strong evidence that bchemokines, and perhaps other unidentified CD8` T cell factors, limit HIV infection of cells in vitro, the challenge lies in determining the role of this mechanism in resistance to HIV infection in vivo and the reason for the ultimate failure of the host to control this infection. Finally, it should be pointed out that none of the studies reviewed here conclusively demonstrated that CD8` T cells are the relevant in vivo source of these cytokines, all of which are produced by other cell types during infection.

CD8` T CELLS IN RESISTANCE TO INTRACELLULAR PROTOZOA Mouse models of infection with intracellular protozoan parasites have been used to analyze the pathogenesis of these organisms and the immune response to infection. These pathogens undergo intricate developmental processes, often involving multiple host species, and the immune response to parasite infection is complex and multifactorial. Experiments with T cell subset depletion and gene-deficient mice reveal that CD8` T cells contribute to immunity against a number of protozoan parasites including Toxoplasma gondii (80), Plasmodium berghei (81–83), Trypanosoma cruzi (84, 85), and Encephalitozoon cuniculi (86). Information on CD8` T cell effector mechanisms is currently limited to these pathogens.

INFECTIONS WITH INTRACELLULAR PROTOZOAN PATHOGENS Each of these parasites has a distinct life cycle that needs to be considered in the context of CD8` T cell effector functions. T. gondii, the causative agent of toxo-



287

plasmosis, is an apicomplexan protozoan found in nearly all animals and most birds. Transmission of oocysts, which develop in the cat gut, occurs by consumption of contaminated undercooked meat, in utero by hematogenous spread, or via tissue transplant. The oocysts contain sporozoites that invade host cells and differentiate into tachyzoites, which reproduce rapidly in the parasitophorus vacuole, eventually rupturing the infected cell to release infectious tachyzoites. A normal immune response leads to encystment within brain and muscle tissue; the parasites within such cysts reproduce very slowly, if at all, and persist for years. Loss of immune function, as seen in immunocompromised patients, allows reactivation and/or spread of the disease. Sporozoites of Plasmodium species, the causative agents of malaria in humans and animal models, are transmitted to the host bloodstream via mosquito bite. The sporozoites enter hepatocytes where they mature within a parasitophorus vesicle. Within approximately 2 weeks, the sporozoites mature into tissue schizonts that rupture the infected cell, producing 10,000 to 40,000 merozoites. Merozoites initiate the red blood cell stage of infection and are the source for mosquito transmission. Since hepatocytes express MHC class I molecules, it is assumed that they are the targets of Plasmodium-specific CD8` T cells. T. cruzi, the causative agent of Chagas’ disease, is a sarcomastigophoran protozoan capable of infecting most mammalian species. The parasite, which is naturally transmitted by the bite of an insect, can infect cells such as macrophages and cardiac myocytes. Parasites enter cells in a parasitophorus vacuole but then escape to the cytoplasm where they differentiate into amastigotes and proliferate. The infected cells then burst, liberating trypomastigotes that invade other cells. Major sites of infection include the CNS, reticuloendothelial system, and heart. Disease occurs as an acute febrile phase with intense inflammatory changes and as a chronic phase lasting many years where there is gradual tissue destruction. E. cuniculi is one of the causative agents of microspiridosis in AIDS patients. These organisms, which belong to the phylum Microspora, are spread horizontally among mammals via contaminated excreta. The parasites initiate infection of immunocompromised hosts by entry into monocytes or other cells where they proceed through a complex replication and developmental cycle in the parasitophorus vacuole and eventually lyse the infected cell. In mouse models, immunocompetent mice resist infection by a T cell-dependent mechanism.

CD8` T CELL–MEDIATED CYTOLYSIS IN RESISTANCE TO PROTOZOA Intracellular protozoan pathogens depend on host cells to complete their complex life cycles. Thus, cytolysis is a potentially relevant effector function in resistance to these infections. CD8` T cells play an important role in resistance to primary and secondary T. gondii infection (80). The requirement for perforin-dependent


288


cytolysis in resistance to acute, chronic, and secondary infections with T. gondii has been addressed (87). Infection of perforin-deficient mice with a virulent strain of T. gondii revealed no differences in acute infection but did result in accelerated mortality, compared to wild-type mice, beginning 75 days post infection. The perforin-deficient mice produced normal levels of IFN-c during the chronic infection with T. gondii, indicating that death in these animals was not due to a lack of IFN-c. In addition, this study found that perforin-deficient mice vaccinated with an attenuated mutant of T. gondii were completely resistant to normally lethal challenge with a virulent strain. As is the case with wild-type mice (88), secondary resistance of perforin-deficient mice was abrogated by treatment with neutralizing anti-IFN-c antibodies. Thus, perforin-dependent cytolysis is not required to control acute infection or provide secondary resistance after vaccination but is required to control chronic infections with T. gondii. Experiments to assess the fate of T. gondii–specific CD8` T cells during the course of infection will be of interest in light of the evidence that perforin deficiency alters CD8` T cell homeostasis in chronic viral infection (53). Immunization with irradiated sporozoites induces resistance to P. berghei infection of mice that is dependent on CD8` T cells (82). Immunization of wildtype, perforin-, Fas-, FasL-, or perforin/FasL-deficient mice resulted in similar immunity to challenge with P. berghei (89). Therefore, neither perforin nor Fasdependent cytolysis of infected host cells by CD8` T cells is required to control P. berghei infection. These results are consistent with the potential for nonlytic CD8` T cell effector mechanisms to impact infection of the liver. Although CD8` T cells contribute to resistance against T. cruzi infection (84, 85), the perforin- and GrmB-mediated cytolytic pathways are not required (86). Perforin- and GrmB-deficient mice infected with T. cruzi had parasitemia levels and survival rates comparable to those of wild-type mice. Additionally, the perforin- and GrmB-deficient mice were protected from subsequent challenge with T. cruzi, while challenge with T. cruzi was lethal in b2M- and TAP-1-deficient mice. These results suggest that control of T. cruzi by CD8` T cells is dependent on other effector mechanisms such as the Fas pathway or the production of cytokines. CD8` T cells participate in control of infection with E. cuniculi, as demonstrated by lethal infection in CD8-deficient but not CD4-deficient mice (90). Perforin-deficient mice infected with E. cuniculi succumbed to infection at the same time as CD8-deficient mice. These results suggest that perforin-mediated cytotoxicity is a required effector mechanism for CD8` T cell control of murine infection with E. cuniculi. Although the number of examples is small, they suffice to demonstrate that the contribution of perforin-dependent cytolysis is not uniform in CD8` T cell– mediated control of protozoan infection. The contribution of CD8` T cell– mediated, Fas-dependent cytolysis in the absence of perforin remains to be determined for infections with T. cruzi and T. gondii.


289


CD8` T CELL–DERIVED CYTOKINES IN RESISTANCE TO PROTOZOA The role of IFN-c in CD8` T cell–mediated protection from T. gondii infection has been analyzed by in vivo depletion studies (88). Mice that received CD8` T cells from T. gondii immune mice had prolonged survival compared to mice that received CD8` T cells from naive mice. Treatment of recipient mice with neutralizing antibody against IFN-c, but not control IgG, abrogated the CD8` T cell– mediated immunity. However, this study did not demonstrate that CD8` T cell–derived IFN-c was necessary, as the required IFN-c may be produced by other cell types in this experimental system. This issue can best be addressed by analysis of the protective capacity of CD8` T cells from IFN-c deficient mice. IFN-c is also an important mediator of CD8` T cell–dependent immunity in animals infected with malaria parasites. Naive recipient mice were protected from P. berghei challenge by adoptive transfer of a CD8` T cell clone specific for the circumsporozoite protein. Treating the recipient mice with anti-IFN-c abolished the protection mediated by CD8` T cells (81). Again, these results do not conclusively demonstrate that the required IFN-c is derived from CD8` T cells. Thus, the role of CD8` T cell–derived IFN-c as an effector mechanism in resistance to T. gondii and P. berghei remains to be determined. This issue has not been investigated for infections with T. cruzi and E. cuniculi. Furthermore, the role of other cytokines produced by activated CD8` T cells has not been investigated in any of these infections. Therefore, additional CD8` T cell effector mechanisms may play a role in the control of these important parasitic infections.

CD8` T CELLS IN RESISTANCE TO INTRACELLULAR BACTERIA All intracellular bacteria enter eukaryotic cells in a membrane-bound structure. Organisms such as Mycobacteria, Salmonella, and Chlamydia survive within a membrane-bound structure, whereas Listeria and Shigella escape from the vesicle into the cytosol of the infected cell. As with most pathogens, the immune response to bacterial infection is complex, and CD8` T cells may or may not be major effectors of immunity (91). In this section we discuss the evidence that CD8` T cells contribute to resistance against Listeria, Chlamydia, and Mycobacteria infections. Listeria monocytogenes (LM) infection of mice evokes strong CD8` T cell responses against a number of antigens presented by MHC class Ia and class Ib molecules (92, 93). CD8` T cells are highly effective mediators of immunity against LM as demonstrated by experiments in mice lacking CD8` T cells (94– 99). In addition, CD8` T cells from immunized mice transfer antilisterial resistance to naive animals (100–104) and mediate clearance of chronic listeriosis in


290


SCID mice (105). Interestingly, recent data suggest that MHC class Ib–restricted CD8` T cells are capable of mediating antilisterial immunity (106). The immune response to Chlamydia has been analyzed in a variety of mouse models including genital, respiratory. and intravenous infection with serovars derived from human and mouse infection. Antibodies, CD4` and CD8` T cells, and cytokines all contribute to resolution of Chlamydia infection. Chlamydiaspecific CD8` T cell lines that recognize peptide antigens presented by MHC class Ia molecules can be stimulated from infected mice (107–109). A role for CD8` T cells in resistance to C. trachomatis infection is supported by experiments demonstrating that transient depletion of CD8` T cells prior to respiratory challenge exacerbates lung infection and that b2M-deficient mice exhibit exacerbated lung infection and mortality (110). Similarly, CD8-deficient mice and mice that lack CD8` T cells, due to double deficiency of b2M and TAP1, exhibit exacerbated lung infection with C. pneumoniae (110). Although most evidence suggests that CD4` T cells are the major effectors of immunity against Mycobacterium tuberculosis (MTB), Mycobacteria-specific CD8` T cell lines have been generated from mice immunized with heat-killed MTB and M. bovis BCG (111). Other experiments demonstrated that immunization with BCG, expressing defined recombinant antigens, leads to a CD8` T cell response to these antigens (112–114). In addition, MHC class Ia-restricted CD8` T cells specific for secreted antigens of MTB can be isolated from tuberculosis patients (115). Presentation of MTB antigens by nonclassical human MHC class Ib molecules such as CD1 has also been described (116, 117). Adoptive transfer experiments have variably demonstrated a protective role for CD8` T cells against MTB (118–20). In addition, a CD8` T cell clone specific for M. leprae Hsp65 protected naive, irradiated mice from subsequent challenge with MTB (121). The specificity of protection in vivo was not assessed and may result from Hsp65-specific CD8` T cell recognition of uninfected, stressed eukaryotic cells (122, 123). Despite this caveat, this approach provides evidence that a single antigen-specific CD8` T cell response may provide immunity to MTB. Finally, the observation that b2M- (124) and TAP1- (125) deficient mice (both of which lack functional CD8` T cells) are deficient in control of MTB is consistent with the notion that CD8` T cells not only respond to MTB, but also play a requisite role in resistance to infection. The relevance of the CD1 antigen presentation pathway to immunity against MTB infection remains to be determined (116, 117).

BACTERIAL INFECTIONS AND ANTIGEN PRESENTATION TO CD8` T CELLS To understand the role of CD8` T cell effector mechanisms in resistance to these organisms requires some discussion of their interaction with the infected host.



291

LM is a gram-positive, facultative intracellular bacterium that causes disseminated infection in immunocompromised human hosts. Following injection into mice, LM is phagocytosed by macrophages and neutrophils in the spleen and liver. In these cells, most bacteria are destroyed, but some escape from the phagosome into the cytoplasm, a process dependent on the listeriolysin O (LLO) molecule (126). LLO-deficient LM strains are avirulent and fail to prime CD8` T cell responses (127, 128). Once in the cytoplasm, LM multiplies and polymerizes host-derived F-actin, thereby initiating movement and direct spread to neighboring host cells (129–132). In this fashion, LM can spread from macrophages to hepatocytes without leaving the intracellular space. ActA mutant strains of LM fail to polymerize host actin, are deficient in cell-to-cell spread, and are severely attenuated in wild-type mice (133, 134). In the new cell, virulent LM escape from the double-membrane-bound vacuole–a process dependent on two bacterial phospholipases (135, 136)–and then reinitiate the infectious cycle. In the host cell cytoplasm, virulent LM secrete a number of proteins that thereby intersect with the endogenous MHC class I antigen presentation pathway (137). Indeed, several MHC class Ia–restricted epitopes have been identified from secreted LM proteins, including the LLO and p60 molecules (138, 139). In addition, the infected host can efficiently mount CD8` T cell responses against nonsecreted LM antigens (140, 141), which are likely to be processed by an exogenous MHC class I presentation pathway, as well as responses to MHC class Ib–restricted LM antigens (142–144). Chlamydia are species of gram-negative, obligate intracellular bacteria that are a major cause of genital and eye infections (C. trachomatis) as well as a variety of respiratory syndromes (C. pneumoniae) in humans. They are spread by direct contact or via the respiratory route as infectious elementary bodies that are taken up by mammalian cells and that subsequently differentiate into reticulate bodies capable of division. The noninfectious reticulate bodies replicate with a doubling time of 2–3 h and, after several divisions, differentiate into elementary bodies that are released by disintegration of infected cells, a process that occurs approximately 48 h after infection. Chlamydia are capable of infecting numerous eukaryotic cell types, including macrophages and mucosal epithelial cells. The route of presentation of Chlamydia antigens and the potential for MHC class Ib–restricted CD8` T cell responses to this pathogen remain to be elucidated. MTB is a slow-growing, acid-fast organism with a doubling time of approximately 48 h in vivo. MTB causes a chronic infection that typically begins in the lungs but can disseminate to a wide range of tissues, perhaps by passage through M cells of the respiratory tract (145). Bacteria can be isolated from the lungs, kidneys, liver and spleen of mice that were infected months previously. Phagocytosis of MTB by macrophages takes place via Fc, mannose, and complement receptors (146). In contrast to LM, intracellular MTB remain localized to the phagosome, where they survive and cause chronic infections by interfering with fusion between phagosomes and lysosomes (147). In most humans, an effective

292


granulomatous response prevents dissemination of the organism from the lung. Active tuberculosis is predominantly observed in immunocompromised individuals and often represents reactivation of an infection that has been latent for years. As is the case for Chlamydia, the precise route by which MTB antigens are presented to CD8` T cells is not well defined. Given the vacuolar location of both these organisms, it is likely that an exogenous route of MHC class I presentation is involved.


CD8` T CELL–MEDIATED CYTOLYSIS IN RESISTANCE TO BACTERIA The role of cytolysis in CD8` T cell–mediated resistance to LM infection was initially addressed using perforin-deficient mice (148). Survival following primary infection with virulent LM did not differ between perforin-deficient and wild-type mice. However, while bacterial clearance from the livers of perforindeficient mice occurred with the same kinetics as control mice, clearance of bacteria from the spleen was delayed. These results have been confirmed and extended to perforin/Fas double-deficient mice, which exhibit delayed clearance of LM from both spleen and liver (149). The results suggest a role for perforin in CD8` T cell–mediated resistance to primary LM infection. Interestingly, the degree of infection, as measured by LM numbers in the spleen and liver, was actually lower in perforin-deficient mice than wild-type mice during early listeriosis (148). The basis for this finding is not understood but may result from immune stimulation by undetected opportunistic pathogens (150). The increased resistance of perforin-deficient mice suggested the possibility that CD8` T cell priming might be decreased as a consequence of reduced antigen load and may account for the deficit in immunity to LM. In contrast to this notion, priming of LM antigen-specific CD8` T cells in perforin-deficient mice is not reduced, and may be enhanced, in comparison to wild-type animals (151). Increased priming of CD8` T cells in perforin-deficient mice has also been reported in chronic LCMV infection (53). In contrast, secondary immunity to LM was impaired in both spleen and livers of perforin-deficient (148) and perforin/Fas double-deficient (149) mice. Whereas CD8-depleted splenocytes from immunized wild-type or perforin-deficient mice transferred equivalent low levels of immunity to naive recipients, CD4-depleted splenocytes from immunized perforin-deficient mice mediated 10-fold (in the liver) and 100-fold (in the spleen) less immunity compared to wild-type cells (148). These data suggest a role for perforin in CD8` T cell immunity to LM that is more pronounced in the spleen than the liver, but the data also suggest the existence of perforin-independent mechanisms of CD8` T cell–mediated resistance to LM.



293

Subsequently, the role of perforin-dependent cytolysis was examined in experiments with LM antigen-specific CD8` T cells lines derived from H-2b and H2d MHC perforin-deficient mice (151, 152). Although perforin-deficient CD8` T cells mediated significant reduction in bacterial numbers in the spleen and livers of infected mice, the degree of immunity in the spleen was consistently less than that provided by antigen-matched wild-type CD8` T cells. In contrast, immunity provided by perforin-deficient and wild-type CD8` T cells was similar in the liver. While these findings confirmed a role for perforin in CD8` T cell–mediated immunity to LM, they emphasized the potential importance of perforin-independent mechanisms and strengthen the notion of tissue-specific CD8` T cell effector mechanisms. The relevance of perforin-independent mechanisms of immunity to LM is underscored by the finding that both wild-type and perforin-deficient CD8` T cells protect naive recipients from otherwise lethal challenges with virulent LM. The exquisite antigen specificity of this immunity was shown when perforindeficient CD8` T cells, specific for a single antigen, conferred survival on mice infected with a normally lethal dose of recombinant LM expressing that antigen, but these cells did not protect mice infected with the equally virulent parental LM strain (151). Thus, while perforin plays a role in CD8` T cell–mediated resistance to LM, especially in the spleen, perforin-independent mechanisms mediated by CD8` T cells are sufficient to confer survival from lethal infection. The importance of Fas-dependent cytolysis in CD8` T cell control of LM was addressed by determining the ability of LM antigen-specific CD8` T cells, from wild-type or perforin-deficient mice, to provide antilisterial immunity to Fasdeficient (B6.MRLlpr/lpr) hosts. LM antigen-specific CD8` T cells, whether they express perforin or not, were capable of mediating similar levels of immunity to LM in Fas-deficient and wild-type mice (152). These results demonstrate CD8` T cell–mediated immunity against LM in the complete absence of the major pathways of cytolysis. CD8` T cells can recognize and lyse Chlamydia infected cells in vitro (108, 109). Lysis of cells containing the replicative reticulate body is an attractive mechanism to interfere with infection since this form is noninfectious. However, genital infections of perforin-deficient mice with C. trachomatis (153) or respiratory infections with C. pneumoniae (154) reveal no delay in clearance of infection compared to wild-type mice. Genital infection of perforin/FasL double-deficient mice with C. trachomatis revealed a slight delay in clearance compared to wildtype mice, but this was attributed to the mixed background of these double knockout animals (153). Thus, a role for CD8` T cell–mediated cytolysis has not been demonstrated in resistance to Chlamydia infection. The role of cytolysis by CD8` T cells in resistance to infection with MTB has been explored in perforin-, GrmB-, and Fas-deficient mice (155, 156). These studies did not reveal a role for cytolysis in resistance against primary, short-term infection with MTB. However, further experiments are required to determine whether perforin plays a role in control of reactivation or long-term MTB infections.

294



CD8` T CELL CYTOKINES IN RESISTANCE TO BACTERIA Activated CD8` T cells produce IFN-c after antigen stimulation. This cytokine, which is also produced by NK and CD4` T cells, is known to be important in the innate immune response against LM (157). IFN-c activates listericidal activity in macrophages (158) and increases the expression of a number of immunologically relevant molecules including those involved in antigen presentation (159). Experiments using Ab-mediated cytokine neutralization did not provide a consensus on the requirement for IFN-c in secondary immunity to LM (160–162). IFN-c-deficient mice provided an experimental system in which differences in Ab reagents or the effectiveness of in vivo depletion of IFN-c was not an issue. Naive IFN-c (163) and IFN-c receptor (164, 165) deficient mice are extremely susceptible to primary infection with virulent LM, consistent with the important role of IFN-c during the innate immune response. However, IFN-c-deficient mice survive high dose infection with an ActA mutant strain of LM, which localizes in the cytoplasm but fails to spread to neighboring cells (133, 134). Interestingly, immunization of IFN-c-deficient or wild-type mice with ActA mutant LM generated equivalent resistance to high-dose challenge with virulent LM (163). Secondary immunity to LM in IFN-c- deficient mice was shown to be antigen specific and dependent on CD8` T cells. Thus, the adaptive immune response could overcome the absence of a cytokine that was critical for resistance of a naive animal. Furthermore, using adoptive transfer experiments, it was found that CD8` T cells from IFN-c-deficient mice were as effective as CD8` T cells derived from wild-type mice in eliminating virulent LM from naive recipients. These results have been corroborated by experiments in which CD8`, but not CD4`, T cells from IFN-c- deficient mice cleared LM from chronically infected SCID mice (64). Although these data do not rule out a contribution of CD8` T cell–derived IFN-c in resistance to LM in normal animals, they do convincingly demonstrate that IFN-c is not required for the development or the expression of antilisterial immunity mediated by CD8` T cells. Antigen-stimulated CD8` T cells produce TNF, a cytokine that also plays an important role in the innate immune response to LM infection (166–168). Recently, it has been shown that TNF- (169) and TNFRI- (170, 171) deficient mice are extremely susceptible to primary LM infection. TNF derived from CD8` T cells could participate in antilisterial immunity by at least three independent mechanisms. CD8` T cell–derived TNF might contribute to the maximal activation of listericidal capacities in macrophages (158). Besides activation of proinflammatory genes, signaling through TNFRI can activate the caspase cascade and induce apoptosis (172), suggesting that CD8` T cells might deliver a TNFdependent signal resulting in the death of the infected target cell. This would release the LM into the extracellular space where they are susceptible to phagocytosis by listericidal macrophages and neutrophils. Finally, TNF has been dem-



295

onstrated to upregulate adhesion molecule expression on endothelial cells, which could aid in the recruitment of accessory phagocytes to the site of infection (173– 175). Treatment of LM-immune mice with anti-TNF Abs inhibited resistance against a secondary challenge with LM (176). In addition, anti-TNF Abs eliminated antilisterial immunity provided by LM antigen-specific, perforin-deficient CD8` T cells (152). Although the results show that TNF participates in secondary resistance to LM, they do not demonstrate a role for TNF derived from CD8` T cells versus other cell types. While TNF-deficient mice are extremely susceptible to primary infection with virulent LM, those that survive infection or those that are immunized with attenuated LM, develop high-level secondary resistance (L Alexopoulou, G Kollias, personal communication). The reason(s) for the differences in secondary immunity in TNF-deficient mice compared to mice treated with antiTNF Ab’s is under investigation. To address the role of CD8` T cell–derived TNF in antilisterial resistance, we generated LM antigen-specific CD8` T cells from MHC-matched B6, TNF-deficient and TNFRI-deficient mice (DW White, G Kollias, JT Harty, unpublished observations). CD8` T cells derived from LMimmune TNF- and TNFRI-deficient animals were cytolytic, produced IFN-c in an antigen-specific fashion, and mediated substantial antilisterial immunity when transferred to wild-type mice. However, TNF-deficient CD8` T cells fail to provide immunity when transferred into TNF-deficient hosts. Thus, CD8` T cell– derived TNF was not required for antilisterial immunity. However, TNF produced by other cells may be required for the optimal expression of CD8` T cell–mediated immunity against LM. Macrophage inhibitory protein-1 alpha (MIP-1a) is a CC chemokine produced by multiple cell types including CD8` T cells. However, the precise role(s) and relevant sources of MIP-1a in response to infections in vivo are unknown. Recent work suggests that CD8` T cell–derived MIP-1a may be required for immunity to LM (177). CD8` T cells derived from LM-infected MIP-1a-deficient mice were capable of LM-specific cytolysis in vitro but failed to transfer antilisterial immunity to naive mice. MIP-1a expression by other cells types was not critical as MIP-1a-deficient recipients and wild-type recipients were equally protected by LM immune wild-type cells. While these experiments did not analyze the priming of the CD8` T cell response in MIP-1a-deficient mice, the failure of which could also explain their results, they suggest that MIP-1a and chemokine production in general may represent another mechanism for CD8` T cell– mediated antilisterial immunity. The finding that no single effector mechanism studied to date can account for all of the antilisterial resistance mediated by CD8` T cells has led to the hypothesis that multiple mechanisms, each capable of functioning independently, may carry out antilisterial immunity (22). Limited studies along these lines, however, have yet to uncover a combination of known CD8` T cell effector functions that are convincingly required for antilisterial immunity. For example, depletion of IFN-c using high doses of mAb did not interfere with antilisterial immunity medi-


296


ated by perforin-deficient CD8` T cells (152). Recent studies of LM-specific CD8` T cells from mice deficient in both perforin and IFN-c support this contention as they provide high level immunity to LM infection (V Badovinac and JT Harty, in preparation). Similarly, perforin-deficient CD8` T cells functioned equally well when transferred into wild-type and Fas-deficient host mice but failed to provide immunity in the presence of anti-TNF antibody (152). Analysis of perforin/TNF double-deficient mice, recently generated in our laboratory, will address the potential redundancy of these effector mechanisms. Systematic elimination of known CD8` T cell effector mechanisms through combination of specific gene knockout mice should eventually identify critical combinations of effector functions or reveal the existence of novel effector mechanisms in resistance to LM infection. A role for IFN-c in CD8` T cell immunity to Chlamydia infection was demonstrated by Ab-mediated neutralization of systemic IFN-c, which inhibited the ability of a C. trachomatis–specific CD8` T cell line to provide immunity in transfer assays (108). Importantly, this group next demonstrated that a Chlamydiaspecific CD8` T cell line, derived from IFN-c-deficient mice, failed to provide immunity under conditions where the wild-type CD8` T cell line mediated significant protection (178). This study provides evidence that CD8` T cell–derived IFN-c can contribute to resolution of Chlamydia infection. The mechanism for this protection is unclear but could involve direct effects on the metabolic activities of intracellular Chlamydia (179–181) or activation of the microbicidal activities of host macrophages (182). While TNF and IFN-c are both critical molecules in resistance to MTB (183, 184), their role as CD8` T cell effector mechanisms in this infection has not been described.

MICROBICIDAL MOLECULES OF CD8` T CELLS Intriguing recent data suggest that CD8` T cells may elaborate molecules with direct microbicidal activity (185, 186). Purified recombinant granulysin, a protein normally found in the granules of activated human CD8` T cells and NK cells, mediates direct microbicidal activity against a range of pathogens (186). Furthermore, granulysin is capable of inhibiting the growth of intracellular MTB in the presence, but not the absence, of perforin. Consistent with these studies, T cell-mediated cytolysis of MTB-infected cells by the granule exocytosis pathway, but not the Fas pathway, resulted in decreased survival of the microbe (185). Besides suggesting an additional role for perforin, these results indicate that CD8` T cells may do more than simply interrupt the life cycle of intracellular pathogens by destroying the eukaryotic host cell. Studies to assess the relevance of granulysin in the defense against microbial pathogens in vivo would be greatly simplified by the identification of a mouse homologue for this molecule.


297


CONCLUDING REMARKS Infections of gene knockout mice reveal that CD8` T cell–mediated cytolysis, either perforin- or Fas-dependent, is critical for resistance against some nonlytic and at least some lytic viruses. In contrast, cytolysis is not involved in the control of other viruses, although it may contribute to the pathological consequences of infection. Conclusive evidence that CD8` T cell–derived cytokines are critical in resistance to viruses is lacking, although increasing evidence supports the potential of such nonlytic mechanisms to contribute in the resolution of viral infection, specifically in essential organs such as the liver and CNS. Given the relatively small number of examples, it is difficult to generalize about CD8` T cell effector mechanisms in resistance to protozoan pathogens. Despite the evidence that CD8` T cells are important in immunity to these infections, perforin-dependent cytolysis does not appear to contribute to control of primary or secondary infections with P. berghei, T. cruzi, or T. gondii, although it does contribute to control of chronic T. gondii infection and is clearly important in resistance of mice to E. cuniculi. In contrast, clear evidence indicates that CD8` T cell–mediated immunity to T. gondii and P. berghei depends on IFN-c. However, formal evidence that CD8` T cells are the source of the required cytokine has not been provided. Further analysis is required to understand the contribution of specific CD8` T cell effector mechanisms in resistance to these complex intracellular pathogens. Perforin-dependent cytolysis contributes to adaptive immunity against L. monocytogenes infection, but its action appears to be most potent in the spleen and dispensable in the liver. Furthermore, perforin is not required for CD8` T cell–mediated survival from lethal challenge. These results are consistent with the existence of tissue-specific effector mechanisms, specifically nonlytic mechanisms to eliminate pathogens in the liver, as suggested from viral studies. In contrast, no evidence for perforin-dependent immunity against Chlamydia or M. tuberculosis has been presented, although additional analyses of chronic MTB infections need to be performed. CD8` T cells lacking IFN-c or TNF also provide high-level immunity against L. monocytogenes infection. Current studies to evaluate combinations of these effector functions are underway and should distinguish between effector function redundancy and the existence of novel mechanisms such as the recently described granulysin pathway of direct microbicidal activity. In contrast, CD8` T cell– derived IFN-c is an important mediator of resistance to Chlamydia infection, and this issue remains to be addressed in MTB infection. Although limited in scope, the data suggest the possibility that the subcellular location of the bacterial pathogen may impact the relevance of specific CD8` T cell effector mechanisms. In conclusion, CD8` T cells contribute to resistance against intracellular infections with viral, protozoan, and bacterial pathogens. Each of these pathogens has developed a unique interaction with the host and the specifics of this interaction

298


may impact which CD8` T cell effector mechanisms are required for immunity. Thus, pathogen complexity likely provides the impetus for maintaining an extensive arsenal of CD8` T cell effector mechanisms in resistance to infection. ACKNOWLEDGMENTS The authors wish to thank Richard Roller and Michael Starnbach for helpful discussion and Stanley Perlman for critical review of the manuscript. Work from the author’s laboratory was supported by NIH grants AI36864 and AI42767, The Roy J. Carver Charitable Trust, and the Arthritis Foundation. Visit the Annual Reviews home page at www.AnnualReviews.org.


LITERATURE CITED 1. Germain RN. 1994. MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation. Cell 76:287–99 2. Banchereau J, Steinman RM. 1998. Dendritic cells and the control of immunity Nature 392:245–52 3. Pamer E, Cresswell P. 1998. Mechanisms of MHC class I-restricted antigen processing. Annu. Rev. Immunol. 16:323–58 4. Jondal M, Schirmbeck R, Reimann J. 1996. MHC class I-restricted CTL responses to exogenous antigens. Immunity 5:295–302 5. Albert ML, Sauter B, Bhardwaj N. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class Irestricted CTLs. Nature 392:86–89 6. Doherty PC. 1998. The numbers game for virus-specific CD8` T cells. Science 280:227 7. Busch DH, Pilip IM, Vijh S, Pamer EG. 1998. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 8:353–62 8. Murali-Krishna K, Altman JD, Suresh M, Sourdive DJ, Zajac AJ, Miller JD, Slansky J, Ahmed R. 1998. Counting antigen-specific CD8 T cells: a reevalu-

9.

10.

11.

12.

13.

14.

15.

ation of bystander activation during viral infection. Immunity 8:177–87 Butz EA, Bevan MJ. 1998. Massive expansion of antigen-specific CD8` T cells during an acute virus infection. Immunity 8:167–75 Ahmed R, Gray D. 1996. Immunological memory and protective immunity: understanding their relation. Science 272:54– 60 Oehen S, Waldner H, Kundig TM, Hengartner H, Zinkernagel RM. 1992. Antivirally protective cytotoxic T cell memory to lymphocytic choriomeningitis virus is governed by persisting antigen. J. Exp. Med. 176:1273–1 Lau LL, Jamieson BD, Somasundaram T, Ahmed R. 1994. Cytotoxic T-cell memory without antigen. Nature 369:648–2 Tough DF, Sprent J. 1996. Viruses and T cell turnover: evidence for bystander proliferation. Immunol. Rev. 150:129–2 Tanchot C, Lemonnier FA, Perarnau B, Freitas AA, Rocha B. 1997. Differential requirements for survival and proliferation of CD8 naive or memory T cells. Science 276:2057–62 Selin LK, Welsh RM. 1997. Cytolytically active memory CTL present in lymphocytic choriomeningitis virus-immune


16.

17.

18.


19.

20.

21.

22.

23.

24. 25.

26.

27.

mice after clearance of virus infection. J. Immunol. 158:5366–73 Lalvani A, Brookes R, Hambleton S, Britton WJ, Hill AV, McMichael AJ. 1997. Rapid effector function in CD8` memory T cells. J. Exp. Med. 186:859– 65 Iezzi G, Karjalainen K, Lanzavecchia A. 1998. The duration of antigenic stimulation determines the fate of naive and effector T cells. Immunity 8:89–95 Berke G. 1995. The CTL’s kiss of death. Cell 81:9–12 Shresta S, Pham CT, Thomas DA, Graubert TA, Ley TJ. 1998. How do cytotoxic lymphocytes kill their targets? Curr. Opin. Immunol. 10:581–87 Heusel JW, Wesselschmidt RL, Shresta S, Russell JH, Ley TJ. 1994. Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 76:977–87 Ebnet K, Hausmann M, Lehmann-Grube F, Mullbacher A, Kopf M, Lamers M, Simon M.M. 1995. Granzyme A-deficient mice retain potent cell-mediated cytotoxicity. EMBO J. 14:4230–39 Harty JT, Bevan MJ. 1999. Responses of CD8` T cells to intracellular bacteria. Curr. Opin. Immunol. 11:89–93 Ashkenazi A, Dixit VM. 1998. Death receptors: signalling and modulation. Science 281:1305–8 Ploegh HL. 1998. Viral strategies of immune evasion. Science 280:248–53 Tewari M, Telford WG, Miller RA, Dixit VM. 1995. CrmA, a poxvirus-encoded serpin, inhibits cytotoxic T-lymphocytemediated apoptosis. J. Biol. Chem. 270:22705–8 Kagi D, Ledermann B, Burki K, Seiler P, Odermatt B, Olsen KJ, Podack ER, Zinkernagel RM, Hengartner H. 1994. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforin-deficient mice. Nature 369:31–37 Walsh CM, Matloubian M, Liu CC, Ueda

28.

29.

30.

31.

32.

33.

34.

35.

299

R, Kurahara CG, Christensen JL, Huang MT, Young JD, Ahmed R, Clark WR. 1994. Immune function in mice lacking the perforin gene. Proc. Natl. Acad. Sci. USA 91:10854–58 Lowin B, Beermann F, Schmidt A, Tschopp J. 1994. A null mutation in the perforin gene impairs cytolytic T lymphocyte- and natural killer cell-mediated cytotoxicity. Proc. Natl. Acad. Sci. USA 91:11571–75 Kagi D, Ledermann B, Burki K, Zinkernagel RM, Hengartner H. 1996. Molecular mechanisms of lymphocytemediated cytotoxicity and their role in imunological protection and pathogenesis. Annu. Rev. Immunol. 14:207–32 Kagi D, Seiler P, Pavlovic J, Ledermann B, Burki K, Zinkernagel RM, Hengartner H. 1995. The roles of perforin- and Fasdependent cytotoxicity in protection against cytopathic and noncytopathic viruses. Eur. J. Immunol. 25:3256–62 Doherty PC. 1996. Cytotoxic T cell effector and memory function in viral Immunity Curr. Top. Microbiol. Immunol. 206:1–14 Topham DJ, Tripp RA, Doherty PC. 1997. CD8` T cells clear influenza virus by perforin or Fas-dependent processes. J. Immunol. 159:5197–5200 Mullbacher A, Hla RT, Museteanu C, Simon MM. 1999. Perforin is essential for control of ectromelia virus but not related poxviruses in mice. J. Virol. 73:1665–67 Froelich CJ, Orth K, Turbov J, Seth P, Gottlieb R, Babior B, Shah GM, Bleackley RC, Dixit VM, Hanna W. 1996. New paradigm for lymphocyte granule-mediated cytotoxicity—target cells bind and internalize granzyme B, but an endosomolytic agent is necessary for cytosolic delivery and subsequent apoptosis. J. Biol. Chem. 271:29073–79 Shresta S, Graubert TA, Thomas DA, Raptis SZ, Ley TJ. 1999. Granzyme A initiates an alternative pathway for

300

36.

37.


38.

39.

40.

41.

42.

43.

44.

HARTY n TVINNEREIM n WHITE granule-mediated apoptosis. Immunity 10:595–605 Beresford PJ, Xia ZN, Greenberg AH, Lieberman J. 1999. Granzyme A loading induces rapid cytolysis and a novel form of DNA damage independently of caspase activation. Immunity 10:585–94 Mullbacher A, Ebnet K, Blanden RV, Hla RT, Stehle T, Museteanu C, Simon MM. 1996. Granzyme A is critical for recovery of mice from infection with the natural cytopathic viral pathogen, ectromelia. Proc. Natl. Acad. Sci. USA 93:5783–87 Guidotti LG, Borrow P, Brown A, McClary H, Koch R, Chisari FV. 1999. Noncytopathic clearance of lymphocytic choriomeningitis virus from the hepatocyte. J. Exp. Med. 189:1555–64 Ahmed R, Salmi A, Butler LD, Chiller JM, Oldstone MB. 1984. Selection of genetic variants of lymphocytic choriomeningitis virus in spleens of persistently infected mice. Role in suppression of cytotoxic T lymphocyte response and viral persistence. J. Exp. Med. 160:521– 40 Buchmeier MJ, Welsh RM, Dutko FJ, Oldstone MB. 1980. The virology and immunobiology of lymphocytic choriomeningitis virus infection. Adv. Immunol. 30:275–331 Dethlefs S, Brahic M, Larsson-Sciard EL. 1997. An early, abundant cytotoxic Tlymphocyte response against Theiler’s virus is critical for preventing viral persistence. J. Virol. 71:8875–78 Cash E, Chamorro M, Brahic M. 1985. Theiler’s virus RNA and protein synthesis in the central nervous sytem of demyelinating mice. Virology 144:290–94 Rossi CP, McAllister A, Tanguy M, Kagi D, Brahic M. 1998. Theiler’s virus infection of perforin-deficient mice. J. Virol. 72:4515–19 Murray PD, McGavern DB, Lin X, Njenga MK, Leibowitz J, Pease LR, Rodriguez M. 1998. Perforin-dependent

45.

46.

47.

48.

49.

50.

51.

52.

neurologic injury in a viral model of multiple sclerosis. J. Neurosci. 18:7306–14 Rodriguez M, Dunkel AJ, Thiermann RL, Leibowitz J, Zijlstra M, Jaenisch R. 1993. Abrogation of resistance to Theiler’s virus-induced demyelination in H2b mice deficient in b2-microglobulin. J. Immunol. 151:255–66 Gallimore A, Glithero A, Godkin A, Tissot AC, Pluckthun A, Elliott T, Hengartner H, Zinkernagel R. 1998. Induction and exhaustion of lymphocytic choriomeningitis virus-specific cytotoxic T lymphocytes visualized using soluble tetrameric major histocompatibility complex class I peptide complexes. J. Exp. Med. 187:1383–93 Marker O, Thomsen AR, Volkert M, Hansen BL, Clemmensen IH. 1985. High-dose survival in the lymphocytic choriomeningitis virus infection is accompanied by suppressed DTH but unaffeced T cell cytotoxicity. Scand. J. Immunol. 21:81–91 Williamson JSP, Stohlman S. 1990. Effective clearance of mouse hepatitis virus from the central nervouse sysem requires both CD4` and CD8` T cells. J. Virol. 64:4589–92 Lin MT, Stohlman SA, Hinton DR. 1997. Mouse hepatitis virus is cleared from the central nervous systems of mice lacking perforin-mediated cytolysis. J. Virol. 71:383–91 Henke A, Huber SA, Stelzner A, Whitton JL. 1995. The role of CD8` T lymphocytes in coxsackie virus B3-induced myocarditis. J. Virol. 69:6720–28 Gebhard JR, Perry CM, Harkins S, Lane T, Mena I, Asensio VC, Campbell I, Whitton JL. 1998. Coxsackievirus B3induced myocarditis: perforin exacerbates disease, but plays no detectable role in virus clearance. Am. J. Pathol. 153:417–28 Matloubian M, Concepcion RJ, Ahmed R. 1994. CD4` T cells are required to sustain CD8` cytotoxic T-cell responses


53.

54.


55.

56.

57.

58.

59.

60.

61.

during chronic viral infection. J. Virol. 68:8056–63 Matloubian M, Suresh M, Glass A, Galvan M, Chow K, Whitmire JK, Walsh CM, Clark W.R., Ahmed, R. 1999. A role for perforin in downregulating T-cell responses during chronic viral infection. J. Virol. 73:2527–36 Ehtisham S, Sunil CN, Nash AA. 1993. Pathogenesis of murine gammaherpesvirus infection in mice deficient in CD4 and CD8 T cells. J. Virol. 67:5247–52 Usherwood EJ, Brooks JW, Sarawar SR, Cardin RD, Young WD, Allen DJ, Doherty PC, Nash AA. 1997. Immunological control of murine gammaherpesvirus infection is independent of perforin. J. Gen. Virol. 78:2025–30 Sarawar SR, Cardin RD, Brooks JW, Mehrpooya M, Hamilton EA, Mo XY, Doherty PC. 1997. Gamma interferon is not essential for recovery from acute infection with murine gammaherpesvirus 68. J. Virol. 71:3916–21 Dutia BM, Clarke CJ, Allen DJ, Nash AA. 1997. Pathological changes in the spleens of gamma interferon receptordeficient mice infected with murine gammaherpesvirus: a role for CD8 T cells. J. Virol. 71:4278–83 Cantin E, Tanamachi B, Openshaw H, Mann J, Clarke K. 1999. Gamma interferon (IFN-gamma) receptor null-mutant mice are more susceptible to herpes simplex virus type 1 infection than IFNgamma ligand null-mutant mice. J. Virol. 73:5196–5200 McNeal MM, Siobhan Barone K, Rae MN, Ward RL. 1995. Effector functions of antibody and CD8` cells in resolution of rotavirus infection and protection against reinfection in mice. Virology 214:387–97 Franco MA, Greenberg HB. 1995. Role of B cells and cytotoxic T lymphocytes in clearance of and immunity to rotavirus infection in mice. J. Virol. 69:7800–6 Franco MA, Tin C, Rott LS, VanCott JL,

62.

63.

64.

65.

66.

67.

68.

69.

70.

301

McGhee JR, Greenberg HB. 1997. Evidence for CD8` T-cell immunity to murine rotavirus in the absence of perforin, fas, and gamma interferon. J. Virol. 71:479–86 Lohman BL, Welsh RM. 1998. Apoptotic regulation of T cells and absence of immune deficiency in virus-infected gamma interferon receptor knockout mice. J. Virol. 72:7815–21 Tishon A, Lewicki H, Rall G, Von HM, Oldstone MB. 1995. An essential role for type 1 interferon-gamma in terminating persistent viral infection. Virology 212: 244–50 Bhardwaj V, Kanagawa O, Swanson PE, Unanue ER. 1998. Chronic Listeria infection in SCID mice: requirements for the carrier state and the dual role of T cells in transferring protection or suppression. J. Immunol. 160:376–84 Parra B, Hinton DR, Marten NW, Bergmann CC, Lin MT, Yang CS, Stohlman SA. 1999. IFN-gamma is required for viral clearance from central nervous system oligodendroglia. J. Immunol. 162: 1641–47 Torres C, Arangurz I, Rubio N. 1995. Expression of interferon-gamma receptors on murine oligodendrocytes and its regulation by cytokines and mitogens. Immunology. 86:250–55 Chisari FV. 1997. Cytotoxic T cells and viral hepatitis. J. Clin. Invest. 99:1472– 77 Guidotti LG, Ishikawa T, Hobbs MV, Matzke B, Schreiber R, Chisari FV. 1996. Intracellular inactivation of the hepatitis B virus by cytotoxic T lymphocytes. Immunity 4:25–36 Guidotti LG, Borrow P, Hobbs MV, Matzke B, Gresser I, Oldstone MB, Chisari FV. 1996. Viral cross talk: intracellular inactivation of the hepatitis B virus during an unrelated viral infection of the liver. Proc. Natl. Acad. Sci., USA 93:4589–94 Cavanaugh VJ, Guidotti LG, Chisari FV.

302

71.

72.


73.

74. 75.

76.

77.

78.

79.

80.

HARTY n TVINNEREIM n WHITE 1998. Inhibition of hepatitis B virus replication during adenovirus and cytomegalovirus infections in transgenic mice. J. Virol. 72:2630–37 Guidotti LG, Rochford R, Chung J, Shapiro M, Purcell R, Chisari FV. 1999. Viral clearance without destruction of infected cells during acute HBV infection. Science 284:825–29 McMichael AJ, Phillips RE. 1997. Escape of human immunodeficiency virus from immune control. Annu. Rev. Immunol. 15:271–96 Walker CM, Moody DJ, Stites DP, Levy JA. 1986. CD8` lymphocytes can control HIV infection in vitro by suppressing virus replication. Science 234:1563–66 Bates, P. 1996. Chemokine receptors and HIV-1: an attractive pair? Cell 86:1–3 Cocchi F, DeVico AL, Garzino-Demo A, Arya SK, Gallo RC, Lusso P. 1995. Identification of RANTES, MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive factors produced by CD8` T cells. Science 270:1811–15 Barker E, Bossart KN, Levy JA. 1998. Primary CD8` cells from HIV-infected individuals can suppress productive infection of macrophages independent of beta-chemokines. Proc. Natl. Acad. Sci. USA 95:1725–29 Rubbert A, Weissman D, Combadiere C, Pettrone KA, Daucher JA, Murphy PM, Fauci AS. 1997. Multifactorial nature of noncytolytic CD8` T cell-mediated suppression of HIV replication: betachemokine-dependent and -independent effects. AIDS Res. Hum. Retrovirus 13:63–9 Bach EA, Aguet M, Schreiber RD. 1997. The IFN-gamma receptor: a paradigm for cytokine receptor signaling. Annu. Rev. Immunol. 15:563–91 Ashkenazi A, Dixit VM. 1998. Death receptors: signaling and modulation. Science 281:1305–8 Suzuki Y, Remington JS. 1988. Dual regulation of resistance against Toxo-

81.

82.

83.

84.

85.

86.

87.

88.

89.

plasma gondii infection by Lyt-2` and Lyt-1`,L3T4` T cells in mice. J. Immunol. 140:3943 Weiss WR, Berzofsky JA, Houghten RA, Sedegah M, Hollindale M, Hoffman SL. 1992. A T cell clone directed at the circumsporozoite protein which protects mice against Plasmodium yoelii and Plasmodium berghei. J. Immunol. 149: 2103–9 Weiss WR, Sedegah M, Beaudoin RL, Miller LH, Good MF. 1988. CD8` T cells (cytotoxic/suppressors) are required for protection in mice immunized with malaria sporozoites. Proc. Natl. Acad. Sci. USA 85:573–76 Romero P, Maryanski JL, Corradin G, Nussenzweig RS, Nussenzweig V, Zavala F. 1989. Cloned cytotoxic T cells recognize an eptiope in the circumsporozoite protein and protect against malaria. Nature 341:323–26 Tarleton RL, Koller BH, Latour A, Postan M. 1992. Susceptibility of b2-microglobulin-deficient mice to Trypanosoma cruzi infection. Nature 356:338–40 Tarleton RL. 1990. Depletion of CD8` T cells increases susceptibility and reverses vaccine-induced immunity in mice infected with Trypanosoma cruzi. J. Immunol. 144:717–24 Kumar S, Tarleton RL. 1998. The relative contribution of antibody production and CD8` T cell function to immune control of Trypanosoma cruzi. Parasite Immunol. 20:207–16 Denkers EY, Yap G, Scharton-Kersten T, Charest H, Butcher BA, Caspar P, Heiny S, Sher A. 1997. Perforin-mediated cytolysis plays a limited role in host resistance to Toxoplasma gondii. J. Immunol. 159:1903–8 Suzuki Y, Remington JS. 1990. The effect of anti-IFN-c antibody on the protective effect of LYt-2` immune T cells against Toxoplasmosis in mice. J. Immunol. 144:1954–56 Renggli J, Hahne M, Matile H, Betschart


90.

91.


92.

93.

94.

95.

96.

97.

B, Tschopp J, Corradin G. 1997. Elimination of P. berghei liver stages is independent of Fas (CD95/apo-I) or perforin-mediated cytotoxicity. Parasite Immunol. 19:145–48 Khan IA, Schwartzman JD, Kasper LH, Moretto M. 1999. CD8` CTLs are essential for protective immunity against Encephalitozoon cuniculi infection. J. Immunol. 162:6086–91 White DW, Wilson RL, Harty JT. 1996. CD8` T cells in intracellular bacterial infections of mice. Res. Immunol. 147:519–24 Pamer EG, Sijts AJ, Villanueva MS, Busch DH, Vijh S. 1997. MHC class I antigen processing of Listeria monocytogenes proteins: implications for dominant and subdominant CTL responses. Immunol. Rev. 158:129–36 Lenz LL, Bevan MJ. 1997. CTL responses to H2-M3-restricted Listeria epitopes. Immunol. Rev. 158:115–21 Mielke ME, Ehlers S, Hahn H. 1988. Tcell subsets in delayed-type hypersensitivity, protection, and granuloma formation in primary and secondary Listeria infection in mice: superior role of Lyt-2` cells in acquired Immunity Infect. Immun. 56:1920–25 Czuprynski CJ, Brown JF. 1990. Effects of purified anti-Lyt-2 mAb treatment on murine listeriosis: comparative roles of Lyt-2` and L3T4` cells in resistance to primary and secondary infection, delayed-type hypersensitivity and adoptive transfer of resistance. Immunology 71:107–12 Mielke ME, Niedobitek G, Stein H, Hahn H. 1989. Acquired resistance to Listeria monocytogenes is mediated by Lyt-2` T cells independently of the influx of monocytes into granulomatous lesions. J. Exp. Med. 170:589–94 Roberts AD, Ordway DJ, Orme IM. 1993. Listeria monocytogenes infection in beta 2 microglobulin-deficient mice. Infect. Immun. 61:1113–16

303

98. Ladel CH, Flesch IE, Arnoldi J, Kaufmann SH. 1994. Studies with MHC-deficient knock-out mice reveal impact of both MHC I- and MHC II-dependent T cell responses on Listeria monocytogenes infection. J. Immunol. 153:3116– 22 99. Fu YX, Roark CE, Kelly K, Drevets D, Campbell P, O’Brien R, Born W. 1994. Immune protection and control of inflammatory tissue necrosis by gamma delta T cells. J. Immunol. 153:3101–15 100. Czuprynski CJ, Brown JF. 1987. Dual regulation of anti-bacterial resistance and inflammatory neutrophil and macrophage accumulation by L3T4` and Lyt 2` Listeria-immune T cells. Immunology 60:287–93 101. Bishop DK, Hinrichs DJ. 1987. Adoptive transfer of immunity to Listeria monocytogenes. The influence of in vitro stimulation on lymphocyte subset requirements. J. Immunol. 139:2005–9 102. Harty JT, Bevan MJ. 1992. CD8` T cells specific for a single nonamer epitope of Listeria monocytogenes are protective in vivo. J. Exp. Med. 175:1531–38 103. Lukacs K, Kurlander R. 1989. Lyt-2` T cell-mediated protection against listeriosis. Protection correlates with phagocyte depletion but not with IFN-gamma production. J. Immunol. 142:2879–86 104. Harty JT, Bevan MJ. 1996. CD8 T-cell recognition of macrophages and hepatocytes results in immunity to Listeria monocytogenes. Infect. Immun. 64:3632– 40 105. Bhardwaj V, Kanagawa O, Swanson PE, Unanue ER. 1998. Chronic Listeria infection in SCID mice: requirements for the carrier state and the dual role of T cells in transferring protection or suppression. J. Immunol. 160:376–84 106. Seaman MS, Perarnau B, Lindahl KF, Lemonneir FA, Forman J. 1999. Response to L. monocytogenes in mice lacking MHC class Ia molecules. J. Immunol. 162:5429–36


304


107. Igietseme JU, Magee DM, Williams DM, Rank RG. 1994. Role for CD8` T cells in antichlamydial immunity defined by Chlamydia-specific T-lymphocyte clones. Infect. Immun. 62:5195–97 108. Starnbach MN, Bevan MJ, Lampe MF. 1994. Protective cytotoxic T lymphocytes are induced during murine infection with Chlamydia trachomatis. J. Immunol. 153:5183–89 109. Starnbach MN, Bevan MJ, Lampe MF. 1995. Murine cytotoxic T lymphocytes induced following Chlamydia trachomatis intraperitoneal or genital tract infection respond to cells infected with multiple serovars. Infect. Immun. 63: 3527–30 110. Magee DM, Williams DM, Smith JG, Bleicker CA, Grubbs BG, Schachter J, Rank RG. 1995. Role of CD8 T cells in primary Chlamydia infection. Infect. Immun. 63:516–21 111. De Libero G, Flesch I, Kaufmann SH. 1988. Mycobacteria-reactive Lyt-2` T cell lines. Eur. J. Immunol. 18:59–66 112. Winter N, Lagranderie M, Gangloff S, Leclerc C, Gheorghiu M, Gicquel B. 1995. Recombinant BCG strains expressing the SIVmac251nef gene induce proliferative and CTL responses against nef synthetic peptides in mice. Vaccine 13:471–78 113. Stover CK, de la Cruz VF, Fuerst TR, Burlein JE, Benson LA, Bennet LT, Bansal GP, Young JF, Lee MH, Hatfull GF, Snapper SB, Barletta RG, Jacobs WR Jr, Bloom BR. 1991. New use of BCG for recombinant vaccines. Nature 351:456– 60 114. Aldovini A, Young RA. 1991. Humoral and cell-mediated immune responses to live recombinant BCG-HIV vaccines. Nature 351:479–82 115. Lalvani A, Brookes R, Wildinson RJ, Malin AS, Pathan AA, Andersen P, Dockrell H, Pasvol G, Hill AV. 1998. Human cytolytic and interferon gammasecreting CD8` T lymphocytes specific

116.

117.

118.

119.

120.

121.

122.

123.

for Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 95:270–75 Sieling PA, Chatterjee D, Porcelli SA, Prigozy TI, Mazzaccaro RJ, Soriano T, Bloom BR, Brenner MB, Kronenberg M, Brennan PJ, Modlin RL. 1995. CD1restricted T cell recognition of microbial lipoglycan antigens. Science 269:227–30 Mazzaccaro RJ, Stenger S, Rock KL, Porcelli SA, Brenner MB, Modlin RL, Bloom BR. 1998. Cytotoxic T lymphocytes in resistance to tuberculosis. Adv. Exp. Med. Biol. 452:85–101 Orme IM. 1987. The kinetics of emergence and loss of mediator T lymphocytes acquired in response to infection with Mycobacterium tuberculosis. J. Immunol. 138:293–98 Hubbard RD, Flory CM, Collins FM. 1991. Memory T cell-mediated resistance to Mycobacterium tuberculosis infection in innately susceptible and resistant mice. Infect. Immun. 59:2012– 16 Kawamura I, Tsukada H, Yoshikawa H, Fujita M, Nomoto K, Mitsuyama M. 1992. IFN-gamma-producing ability as a possible marker for the protective T cells against Mycobacterium bovis BCG in mice. J. Immunol. 148:2887–93 Silva CL, Silva MF, Pietro RCLR, Lowrie DB. 1994. Protection against tuberculosis by passive transfer with T-cell clones recognizing mycobacterial heatshock protein 65. Immunology. 83:341– 46 Steinhoff U, Schoel B, Kaufmann SH. 1990. Lysis of interferon-gamma activated Schwann cell by cross-reactive CD8` alpha/beta T cells with specificity for the mycobacterial 65 kd heat shock protein. Internat. Immunol. 2:279–84 Zugel U, Schoel B, Yamamoto S, Hengel H, Morein B, Kaufmann SH. 1995. Crossrecognition by CD8 T cell receptor alpha beta cytotoxic T lymphocytes of peptides in the self and the mycobacterial


124.

125.


126.

127.

128.

129.

130.

131.

132.

hsp60 which share intermediate sequence homology. Eur. J. Immunol. 25:451–58 Flynn JL, Goldstein MM, Triebold KJ, Koller B, Bloom BR. 1992. Major histocompatibility complex class Irestricted T cells are required for resistance to Mycobacterium tuberculosis infection. Proc. Natl. Acad. Sci. USA. 89:12013–17 Behar SM, Dascher CC, Grusby MJ, Wang C-R, Brenner MB. 1999. Susceptibility of mice deficient in CD1D or TAP1 to infection with Mycobacterium tuberculosis. J. Exp. Med. 189:1973–80 Portnoy DA, Chakraborty T, Goebel W, Cossart P. 1992. Molecular determinants of Listeria monocytogenes pathogenesis. Infect. Immun. 60:1263–67 Barry RA, Bouwer HG, Portnoy DA, Hinrichs DJ. 1992. Pathogenicity and immunogenicity of Listeria monocytogenes small-plaque mutants defective for intracellular growth and cell-to-cell spread. Infect. Immun. 60:1625–32 Brunt LM, Portnoy DA, Unanue ER. 1990. Presentation of Listeria monocytogenes to CD8` T cells requires secretion of hemolysin and intracellular bacterial growth. J. Immunol. 145:3540– 46 Tilney LG, Portnoy DA. 1989. Actin filaments and the growth, movement, and spread of the intracellular bacterial parasite, Listeria monocytogenes. J. Cell Biol. 109:1597–1608 Tilney LG, Connelly PS, Portnoy DA. 1990. Actin filament nucleation by the bacterial pathogen, Listeria monocytogenes. J. Cell Biol. 111:2979–88 Welch MD, Iwamatsu A, Mitchison TJ. 1997. Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature 385: 265–69 Welch MD, Rosenblatt J, Skoble J, Portnoy DA, Mitchison TJ. 1998. Interaction of Human Arp2/3 Complex and the Listeria monocytogenes ActA protein in

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

305

actin filament nucleation. Science 281:105–8 Kocks C, Gouin E, Tabouret M, Berche P, Ohayon H, Cossart P. 1992. L. monocytogenes-induced actin assembly requires the actA gene product, a surface protein. Cell 68:521–31 Brundage RA, Smith GA, Camilli A, Theriot JA, Portnoy DA. 1993. Expression and phosphorylation of the Listeria monocytogenes ActA protein in mammalian cells. Proc. Natl. Acad. Sci. USA 90:11890–94 Smith GA, Portnoy DA. 1993. The role of two phospholipases in the pathogenicity of Listeria monocytogenes. Infect. Agent. Dis. 2:183–85 Smith GA, Marquis H, Jones S, Johnston NC, Portnoy DA, Goldfine H. 1995. The two distinct phospholipases C of Listeria monocytogenes have overlapping roles in escape from a vacuole and cell-to-cell spread. Infect. Immun. 63:4231–37 Sijts AJ, Villanueva MS, Pamer EG. 1996. CTL epitope generation is tightly linked to cellular proteolysis of a Listeria monocytogenes antigen. J. Immunol. 156:1497–1503 Pamer EG, Harty JT, Bevan MJ. 1991. Precise prediction of a dominant class I MHC-restricted epitope of Listeria monocytogenes. Nature 353:852–55 Pamer EG. 1994. Direct sequence identification and kinetic analysis of an MHC class I-restricted Listeria monocytogenes CTL epitope. J. Immunol. 152:686–94 Schafer R, Portnoy DA, Brassell SA, Paterson Y. 1992. Induction of a cellular immune response to a foreign antigen by a recombinant Listeria monocytogenes vaccine. J. Immunol. 149:53–59 Shen H, Miller JF, Fan X, Kolwyck D, Ahmed R, Harty JT. 1998. Compartmentalization of bacterial antigens: differential effects on priming of CD8 T cells and protective Immunity Cell 92:535–45 Gulden PH, Fischer PR, Sherman NE, Wang W, Engelhard VH, Shabanowitz J,

306

143.

144.


145.

146.

147.

148.

149.

150.

151.

HARTY n TVINNEREIM n WHITE Hunt DF, Pamer EG. 1996. A Listeria monocytogenes pentapeptide is presented to cytolytic T lymphocytes by the H2-M3 MHC class Ib molecule. Immunity 5:73– 79 Lenz LL, Dere B, Bevan MJ. 1996. Identification of an H2-M3-restricted Listeria epitope: implications for antigen presentation by M3. Immunity 5:63–72 Princiotta MF, Lenz LL, Bevan MJ, Staerz UD. 1998. H2-M3 restricted presentation of a Listeria-derived leader peptide. J. Exp. Med. 187:1711–19 Teitelbaum R, Schubert W, Gunther L, Kress Y, Macaluso F, Pollard JW, McMurray DN, Bloom BR. 1999. The M cell as a portal of entry to the lung for the bacterial pathogen Mycobacterium tuberculosis. Immunity 10:641–50 Schlesinger LS. 1996. Entry of Mycobacterium tuberculosis into mononuclear phagocytes. Curr. Top. Microbiol. Immunol. 215:71–96 Russell DG. 1995. Of microbes and macrophages: entry, survival and persistence. Curr. Opin. Immunol. 7:479–84 Kagi D, Ledermann B, Burki K, Hengartner H, Zinkernagel RM. 1994. CD8` T cell-mediated protection against an intracellular bacterium by perforindependent cytotoxicity. Eur. J. Immunol. 24:3068–72 Jensen ER, Glass AA, Clark WR, Wing EJ, Miller JF, Gregory SH. 1998. Fas (CD95)-Dependent cell-mediated immunity to Listeria monocytogenes. Infect. Immun. 66:4143–50 Czuprynski CJ, Brown JF. 1985. Phagocytes from flora-defined and germfree athymic nude mice do not demonstrate enhanced antibacterial activity. Infect. Immun. 50:425–30 White DW, MacNeil A, Busch DH, Philip IM, Pamer EG, Harty JT. 1999. Perforin-deficient CD8` T cells: in vivo priming and antigen-specific immunity against Listeria monocytogenes. J. Immunol. 162:980–88

152. White DW, Harty JT. 1998. Perforindeficient CD8` T cells provide immunity to Listeria monocytogenes by a mechanism that is independent of CD95 and IFN-gamma but requires TNF-alpha. J. Immunol. 160:898–905 153. Perry LL, Feilzer K, Hughes S, Caldwell HD. 1999. Clearance of Chlamydia trachomatis from the murine genital mucosa does not require perforin-mediated cytolysis or Fas-mediated apoptosis. Infect. Immun. 67:1379–85 154. Rottenberg ME, Gigliotti RA, Gigliotti D, Svanholm C, Bandholtz L, Wigzell H. 1999. Role of innate and adaptive immunity in the outcome of primary infection with Chlamydia pneumoniae, as analyzed in genetically modified mice. J. Immunol. 162:2829–36 155. Laochumroonvorapong P, Wang J, Liu CC, Ye WG, Moreira AL, Elkon KB, Freedman VH, Kaplan G. 1997. Perforin, a cytotoxic molecule which mediates cell necrosis, is not required for the early control of mycobacterial infection in mice. Infect. Immun. 65:127–32 156. Cooper AM, Dsouza C, Frank AA, Orme IM. 1997. The course of Mycobacterium tuberculosis infection in the lungs of mice lacking expression of either perforin- or granzyme-mediated cytolytic mechanisms. Infect. Immun. 65:1317–20 157. Bancroft GJ, Schreiber RD, Unanue ER. 1991. Natural immunity: a T-cellindependent pathway of macrophage activation, defined in the SCID mouse. Immunol. Rev. 124:5–24 158. Portnoy DA, Schreiber RD, Connelly P, Tilney LG. 1989. Gamma interferon limits access of Listeria monocytogenes to the macrophage cytoplasm. J. Exp. Med. 170:2141–46 159. Fruh K, Yang Y. 1999. Antigen presentation by MHC class I and its regulation by interferon-c. Curr. Opin. Immunol. 11:76–81 160. Tripp CS, Kanagawa O, Unanue ER. 1995. Secondary response to Listeria


161.

162.


163.

164.

165.

166.

167.

168.

169.

infection requires IFN-gamma but is partially independent of IL-12. J. Immunol. 155:3427–32 Harty JT, Schreiber RD, Bevan MJ. 1992. CD8 T cells can protect against an intracellular bacterium in an interferon gamma-independent fashion. Proc. Natl. Acad. Sci. USA 89:11612–16 Samsom JN, Langermans JA, Savelkoul HF, van Furth R. 1995. Tumour necrosis factor, but not interferon-gamma, is essential for acquired resistance to Listeria monocytogenes during a secondary infection in mice. Immunology 86:256– 62 Harty JT, Bevan MJ. 1995. Specific immunity to Listeria monocytogenes in the absence of IFN gamma. Immunity 3:109–17 Huang S, Hendriks W, Althage A, Hemmi S, Bluethmann H, Kamijo R, Vilcek J, Zinkernagel RM, Aguet M. 1993. Immune response in mice that lack the interferon-gamma receptor. Science 259:1742–45 Dai WJ, Bartens W, Kohler G, Hufnagel M, Kopf M, Brombacher F. 1997. Impaired macrophage listericidal and cytokine activities are responsible for the rapid death of Listeria monocytogenesinfected IFN-gamma receptor-deficient mice. J. Immunol. 158:5297–5304 Bancroft GJ, Sheehan KC, Schreiber RD, Unanue ER. 1989. Tumor necrosis factor is involved in the T cell-independent pathway of macrophage activation in SCID mice. J. Immunol. 143:127–30 Havell EA. 1987. Production of tumor necrosis factor during murine listeriosis. J. Immunol. 139:4225–31 Nakane A, Minagawa T, Kato K. 1988. Endogenous tumor necrosis factor (cachectin) is essential to host resistance against Listeria monocytogenes infection. Infect. Immun. 56:2563–69 Pasparakis M, Alexopoulou L, Episkopou V, Kollias G. 1996. Immune and inflammatory responses in TNF-alpha-

170.

171.

172.

173.

174.

175.

176.

307

deficient mice–a critical requirement for TNF-alpha in the formation of primary B cell follicles, follicular dendritic cell networks and germinal centers, and in the maturation of the humoral immune response. J. Exp. Med. 184:1397–1411 Pfeffer K, Matsuyama T, Kundig TM, Wakeham A, Kishihara K, Shahinian A, Wiegmann K, Ohashi PS, Kronke M, Mak TW. 1993. Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73:457–67 Endres R, Luz A, Schulze H, Neubauer H, Futterer A, Holland SM, Wagner H, Pfeffer K. 1997. Listeriosis in p47 (phox-/-) and TRp55-/- mice: protection despite absence of ROI and susceptibility despite presence of RNI. Immunity 7:419–32 Ashkenazi A, Dixit VM. 1998. Death receptors: signalling and modulation. Science 281:1305–8 Henninger DD, Panes J, Eppihimer M, Russell J, Gerritsen M, Anderson DC, Granger DN. 1997. Cytokine-induced VCAM-1 and ICAM-1 expression in different organs of the mouse. J. Immunol. 158:1825–32 Kondo S, Sauder DN. 1997. Tumor necrosis factor (TNF) receptor type 1 (p55) is a main mediator for TNF-alphainduced skin inflammation. Eur. J. Immunol. 27:1713–18 Lukacs NW, Chensue SW, Strieter RM, Warmington K, Kunkel SL. 1994. Inflammatory granuloma formation is mediated by TNF-alpha-inducible intercellular adhesion molecule-1. J. Immunol. 152:5883–89 Samsom JN, Langermans JA, Savelkoul HF, van Furth R. 1995. Tumour necrosis factor, but not interferon-gamma, is essential for acquired resistance to Listeria monocytogenes during a secondary infection in mice. Immunology. 86:256– 62


308


177. Cook DN, Smithies O, Strieter RM, Frelinger JA, Serody JS. 1999. CD8` T cells are a biologically relevant source of macrophage inflammatory protein-1 alpha in vivo. J. Immunol. 162:5423–28 178. Lampe MF, Wilson CB, Bevan MJ, Starnbach MN. 1998. Gamma interferon production by cytotoxic T lymphocytes is required for resolution of Chlamydia trachomatis infection. Infect. Immun. 66:5457–61 179. Perry LL, Su H, Feilzer K, Messer R, Hughes S, Whitmire W, Caldwell HD. 1999. Differential sensitivity of distinct Chlamydia trachomatis isolates to IFNgamma-mediated inhibition. J. Immunol. 162:3541–48 180. Kane CD, Byrne GI. 1998. Differential effects of gamma interferon on Chlamydia trachomatis growth in polarized and nonpolarized human epithelial cells in culture. Infect. Immun. 66:2349–51 181. Kane CD, Vena RM, Ouellette SP, Byrne GI. 1999. Intracellular tryptophan pool sizes may account for differences in gamma interferon-mediated inhibition and persistence of chlamydial growth in polarized and nonpolarized cells. Infect. Immun. 67:1666–71

182. Bancroft GJ, Schreiber RD, Bosma GC, Bosma MJ, Unanue ER. 1987. A T cellindependent mechanism of macrophage activation by interferon-gamma. J. Immunol. 139:1104–7 183. Flynn JL, Chan J, Triebold KJ, Dalton DK, Stewart TA, Bloom BR. 1993. An essential role for interferon gamma in resistance to Mycobacterium tuberculosis infection. J. Exp. Med. 178:2249–54 184. Flynn JL, Goldstein MM, Chan J, Triebold KJ, Pfeffer K, Lowenstein CJ, Schreibe RR, Mak TW, Bloom BR. 1995. Tumor necrosis factor-alpha is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2:561–72 185. Stenger S, Mazzaccaro RJ, Uyemura K, Cho S, Barnes PF, Rosat J-P, Sette A, Brenner MB, Porcelli SA, Bloom BR, Modlin RL. 1997. Differentail effects of cytolytic T cell subsets on intracellular infection. Science 276:1684 186. Stenger S, Hanson DA, Teitelbaum R, Dewan P, Niazi KR, Froelich CJ, Ganz T, Thoma-Uszynski S, Melian A, Bogdan C, Porcelli SA, Bloom BR, Krensky AM, Modlin RL. 1998. An antimicrobial activity of cytolytic T cells mediated by granulysin. Science 282:121–25


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:309–345

GLUCOCORTICOIDS IN T CELL DEVELOPMENT AND FUNCTION* Jonathan D. Ashwell, Frank W. M. Lu, and Melanie S. Vacchio Laboratory of Immune Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892; e-mail: [email protected]


Key Words thymocyte selection, T cell activation, immunosuppression, apoptosis Abstract Glucocorticoids are small lipophilic compounds that mediate their many biological effects by binding an intracellular receptor (GR) that, in turn, translocates to the nucleus and directly or indirectly regulates gene transcription. Perhaps the most recognized biologic effect of glucocorticoids on peripheral T cells is immunosuppression, which is due to inhibition of expression of a wide variety of activationinduced gene products. Glucocorticoids have also been implicated in Th lineage development (favoring the generation of Th2 cells) and, by virtue of their downregulation of fasL expression, the inhibition of activation-induced T cell apoptosis. Glucocorticoids are also potent inducers of apoptosis, and even glucocorticoid concentrations achieved during a stress response can cause the death of CD4`CD8` thymocytes. Perhaps surprisingly, thymic epithelial cells produce glucocorticoids, and based upon in vitro and in vivo studies of T cell development it has been proposed that these locally produced glucocorticoids participate in antigen-specific thymocyte development by inhibiting activation-induced gene transcription and thus increasing the TCR signaling thresholds required to promote positive and negative selection. It is anticipated that studies in animals with tissue-specific GR-deficiency will further elucide how glucocorticoids affect T cell development and function.

INTRODUCTION The development and function of cells that comprise the immune system are subject to regulation by many intrinsic and extrinsic factors. Immunologists are extremely familiar with many of the ligands and receptors that induce cells to become activated, to migrate, adhere, and express effector functions. Cytokines, typically glycoproteins synthesized and/or secreted de novo in response to an immune or inflammatory stimulus, are the paradigm for soluble molecules that *The US government has the right to retain a nonexclusive, royalty-free license in and to any copyright covering this paper.

309


310

ASHWELL n LU n VACCHIO

can act at a distance and modulate the nature and intensity of the immune response. Another class of molecules that can affect immune cells belongs to the neuroendocrine system. Despite a vast literature on the effects of these soluble products on lymphocytes, the characterization of the molecular mechanisms by which they work, and an appreciation of their fundamental importance in the immune response, lags far behind that for cytokines. Nevertheless, countless observations in humans and animal models and in vitro strongly suggest that neuroendocrine influences can and do participate in shaping the immune response. Perhaps the best-studied mediators in this category are steroids, small lipophilic molecules that participate in an enormous number of normal and pathologic processes. Steroids bind intracellular DNA-binding factors that, in turn, regulate gene transcription in virtually all cell types. This review concentrates on one particular type of steroid: glucocorticoids. The case that this hormone has substantial and important physiologic and pharmacologic effects on the immune response is solid, and accumulating evidence has suggested an unexpected role for glucocorticoids in regulating thymocyte development and selection.

OVERVIEW OF GLUCOCORTICOIDS AND GLUCOCORTICOID RECEPTORS Steroid Hormones The term steroids refers to a group of small lipophilic compounds derived from a common precursor, cholesterol. The four major types of steroids: progestins, androgens, estrogens, and corticoids, differ in the number of carbon atoms they contain, the receptors they bind, and the biological activities they possess. One can further divide the corticoids into two groups: mineralocorticoids, which regulate ion transport and thus fluid and electrolyte balance, and glucocorticoids, which have many activities, including resistance to stress, regulation of intermediary metabolism, and immunosuppressive and anti-inflammatory effects. The conversion of cholesterol to the various steroids is performed by an array of dehydrogenases and cytochrome P450 enzymes, membrane-bound and hemecontaining monooxygenases that catalyze dehydroxylation-oxidation reactions (Figure 1). The first and rate-limiting step in steroid biosynthesis is the cleavage of the side chain of cholesterol by P450scc to generate the first steroid, pregnenolone. P450scc expression seems to be limited to steroidogenic tissues such as the adrenals, placenta, gonads, brain, and thymus (1–5). Pregnenolone is hydroxylated at position 17 by P450c17, resulting in two possible parallel pathways of corticoid synthesis. Although rodents express P450c17 in the gonads, in adult animals it is not detectable in the adrenal cortex (6), so the major circulating glucocorticoid in mice (corticosterone) differs slightly from the preponderant circulating glucocorticoid in most species, including human (cortisol). Progesterone (or its 17-OH form) is hydroxylated in the endoplasmic reticulum by P450c21 to


311

Figure 1 Simplified scheme detailing some of the major steps in steroid biosynthesis. The major enzymes shown are P450scc (CYP11A, ‘‘scc’’ denotes ‘‘side chain cleavage’’), P450c11 (CYP11B1, or 11b-hydroxylase), P450c17 (CYP17, or 17 a-hydroxylase/lyase), P450c21 (CYP21, 21-hydroxylase), P450aromatase (CYP19, or aromatase). The structures of cholesterol, corticosterone, and cortisol are shown.


312


yield 11-deoxycorticosterone (or 11-deoxycortisol), which has little glucocorticoid activity, and is then converted in mitochondria to the active glucocorticoid corticosterone (or cortisol) by P450c11. A small amount of corticosterone undergoes a series of intermediate steps that result in conversion to the mineralocorticoid aldosterone. Because steroids are lipophilic they are transported in the blood in a reversible complex with protein. The major high-capacity low-affinity carrier is albumin; the major low-capacity high-affinity protein is transcortin, also known as corticosteroid-binding globulin (CBG). Only a small fraction (on the order of 1– 10%) of glucocorticoids is free and available to mediate biological functions, so the effective glucocorticoid concentration in plasma is considerably lower than the total that is typically measured. The synthetic glucocorticoid dexamethasone (Dex) is particularly potent in vivo because of its high affinity for the GR and relatively low level of binding to plasma proteins (7, 8). Secretion of glucocorticoids by the adrenals is under control of the hypothalamo-pituitary axis. Adrenocorticotrophic hormone (ACTH) produced by the anterior pituitary causes an immediate increase in the secretion of glucocorticoids as well as an increase in the production of steroid biosynthetic enzymes (9). Glucocorticoid levels are maintained, in part, by a feedback loop with the hypothalamus and anterior pituitary in which low systemic glucocorticoid levels increase, and high levels suppress, ACTH secretion (8). Under normal conditions, secretion of ACTH occurs with a circadian pattern, in humans peaking prior to waking and reaching a nadir in the evening. Perhaps of special note to immunologists, in nocturnal animals such as rodents this pattern is reversed, with the peak in ACTH secretion occurring in the late afternoon (8). The resulting changes in glucocorticoid secretion typically yield fluctuations in plasma concentration, with a range of threefold to as much as tenfold over the course of the day (10). The major stimulus to increased ACTH, and thus glucocorticoid, secretion is ‘‘stress,’’ a term that covers a wide range of physiologic (e.g. exercise, emotional disturbance, etc) and pathologic (e.g. trauma, hemorrhage, fever, etc) situations (11).

Glucocorticoid Receptors Nonprotein-bound corticosteroids passively diffuse across the plasma membrane into the cell where they encounter the glucocorticoid receptor (GR). The GR is a member of a large superfamily that includes receptors for other steroid hormones, thyroid hormone, vitamin D3, retinoic acid, and a number of orphan receptors, including Nur77. Receptors of this superfamily have many conserved structural elements, including a COOH-terminal ligand-binding domain (which also contains residues required for dimerization and hormone-dependent gene transactivation), a nearby hinge region containing nuclear localization signals, a central zinc-finger-containing DNA-binding domain, and an NH2-terminal variable region important for ligand-independent gene transactivation (12). Although many of these receptors are nuclear-resident, the GR exists in the cytosol in a


GLUCOCORTICOIDS AND T CELLS

313

complex with heat shock proteins such as hsp90, hsp70, and the immunophilins hsp56 and CyP-40 (13), and it translocates to the nucleus when it is occupied by ligand (14). Once in the nucleus, the GR binds as a homodimer to specific DNA sequences (glucocorticoid responsive elements, or GREs), where it acts to enhance or inhibit transcription of corresponding genes. The classic consensus GRE consists of two conserved 6-nucleotide halves separated by three nonconserved bases, and it typically acts as an enhancer element (15). However, ‘‘negative’’ GREs that suppress gene transcription have also been described (15, 16), and some GREs contain sites for other transcription factors embedded in the GRbinding region, yielding a ‘‘composite’’ response element that may have enhancing or repressing activity (17). Thus, the GR is a ligand-regulated transcription factor. The members of this receptor superfamily interact with a cohort of molecules to mediate their function as transcriptional regulators, and the GR is no exception. For example, GRs are thought to stabilize the formation of a preinitiation complex that contains components of the basal transcriptional machinery (18). Furthermore, the ligand-binding domain contains a region (AF-2) that binds a number of proteins important for GR (and other members of this receptor superfamily) function. One group of co-activators includes steroid receptor coactivator-1 (SRC-1), activator of thyroid hormone and retinoid receptors (ACTR), and transcriptional intermediary factor 2 (TIF2)/glucocorticoid receptor interacting protein 1 (GRIP1) gene products (18). A second co-activator group includes proteins such as CREB binding protein (CBP) and its homolog p300, and p300/ CBP-associated factor (P/CAF). One of the more exciting developments in this area in recent years is the realization that SRC-1, ACTR, p300/CBP, and P/CAF all have intrinsic histone acetytransferase (HAT) activity (19–23). Acetylation of core histones alters nucleosomal packing to allow increased access of transacting factors and components of the basal transcriptional machinery to the local DNA (24). Thus, chromatin remodeling in response to recruitment of HATs by the liganded GR is at least one mechanism by which glucocorticoids enhance gene transcription. The GR is constitutively phosphorylated on serines and threonines. All of the potential phosphorylation sites (eight in the mouse) are N-terminal of the DNAbinding domain (25, 26). Many of these sites are ‘‘proline-directed’’ consensus sequences, favored by cyclin-dependent kinases (CDKs) and MAP kinases, and in fact in vitro cyclin/CDK complexes and ERK2 have been found to phosphorylate different serine/threonine residues in the GR (27). Baseline GR phosphorylation is cell-cycle-dependent, being highest in G2/M and lowest in S phase. GR phosphorylation is enhanced after ligation with hormone in as short a time as 5– 10 min, the degree of increase correlating inversely with the basal phosphorylation level (i.e. greatest in S phase) (reviewed in 26, 28). The biologic consequences of GR phosphorylation are controversial. Initial studies with GRs with mutated phosphorylation sites found little effect on nuclear localization or transactivation (29–31). A more recent study, however, reported that while transiently expressed GR phosphorylation-defective mutants transactivated a reporter driven by the


314


MMTV promoter, they were less good (by two- to fourfold) at transactivating a simple GRE-containing reporter (32). Moreover, receptors containing multiple mutated phosphorylation sites were, unlike wild-type GRs, incapable of downregulating transcription of their own gene, and the half-life of GR protein was substantially increased. Thus, although at this time it seems that the phosphorylation status of the GR does not have an obvious major impact on function, this posttranslational modification may have subtle but perhaps biologically significant effects on receptor expression and transactivation of some target genes. Humans have two GR isoforms: a (the classic GR) and b (33). These receptors share the first 727 amino acids (encoded by the first 8 exons of the GR gene and containing the transactivating and DNA-binding domains), but due to alternate mRNA splicing the receptors contain carboxy-terminal residues encoded by either exon 9a or 9b, respectively. Thus, in GRb the last 50 amino acids of GRa are replaced with a unique 15 amino acid sequence that renders the molecule incapable of binding glucocorticoids and therefore transcriptionally inactive. Unlike GRa, GRb constitutively resides in the nucleus (34). Indeed, GRb can bind a GRE consensus sequence, and overexpression of GRb can inhibit gene transactivation mediated by GRa in a dominant negative fashion. Screens for mRNA expression found that GRa and GRb are widely expressed in the same tissues (33, 34), leading to the speculation that these GR isoforms might interact, perhaps via heterodimerization, to regulate the transcriptional effects of glucocorticoids. This possibility was supported by the finding that peripheral blood mononuclear cells from patients with glucocorticoid-resistant asthma had decreased binding of the GR to DNA, accompanied by elevated numbers of cells in which GRb was detected by immunohistochemistry (35). Binding analyses indicated that the decrease in GR-DNA binding was due to lowered GR affinity for the GRE. Moreover, overexpression of GRb in HepG2 cells (hepatocytes) had the same effect on binding of the endogenous GRa to DNA. GR DNA binding activity increased in peripheral blood mononuclear cells cultured in medium but not those cultured in IL-2 and IL-4, and culture of cells from normal donors with IL-2 and IL-4 resulted in increased GRb expression. Together, these data suggest that GRb expression can be regulated by cytokines and that there may in fact be situations in vivo in which the interplay between GRa and GRb has significant biological consequences.

Glucocorticoid Signaling Although the activity of the GR is often thought of simply in terms of direct gene transactivation, considerable cross-talk also occurs between the GR and other transcription factors that can modify each of their biologic activities. The first such interaction described was between the GR and AP-1 (17, 36–38). Typically, these factors antagonize each other’s transcription enhancing activity, although AP-1 consisting of c-Jun dimers can also enhance GRE-mediated transactivation (reviewed in 39, 40). Most, although not all, studies that examined the issue



315

detected a direct physical interaction between the GR and AP-1, suggesting at least one direct molecular basis for cross-talk, although simple sequestration of these factors into inactive complexes does not seem to be likely to explain the phenomenon. Other nuclear factors that have been found to bind the GR and modify its activity are NF-jB (41–44), the cAMP response element binding protein CREB (45), and the signaling and transcription factors STAT3 and STAT5 (46, 47). Another potential mechanism is competition for co-activators (‘‘squelching’’). Both AP-1 and the GR are co-activated by CBP/p300, and in fact overexpression of CBP or p300 reverses the antagonism between AP-1 and the GR (46). Similarly, overexpression of CPB or SRC-1 reverses the transcriptional antagonism between the GR and NF-jB (47). These results support the notion that in some circumstances cross-talk between the GR and other transcription factors is due to competition for limiting co-activators of transcription. In the case of NF-jB, yet another mechanism for cross-talk with the GR has been proposed: Glucocorticoids increase the transcription and synthesis of IjB and thus may inhibit NF-jB by promoting its retention in the cytosol (48, 49). Although the possibility that glucocorticoids inhibit NF-jB by upregulating IjB is attractive, its biological relevance is uncertain at best, as there is a steadily increasing number of examples in which inhibition of NF-jB occurs in the absence of IjB upregulation (50–53). Regardless of the particular mechanism involved, the extensive degree of cross-talk between the GR and other transcription factors provides a rich framework for mutual regulation (positive or negative) between glucocorticoids and other signaling pathways.

GLUCOCORTICOIDS AND APOPTOSIS At least since the end of the nineteenth century it has been known that adrenal insufficiency in humans (54) and adrenalectomy of animals result in thymic hypertrophy that cannot be reversed by the adrenal medullary product epinephrine (55–58), and that stress and drug-induced involution of the thymus is prevented by adrenalectomy (58). These observations were followed by the findings that administration of ACTH to mice caused a marked reduction in thymus and lymph node mass (59), and that a purified corticosteroid caused the regression of a lymphosarcoma (60)1. It is now appreciated that lymphoid cells, especially 1 In an oddly touching footnote to this article, the editor explains that although the manuscript was received in June, 1942, it was withheld from publication, at the authors’ request, until April, 1944. Among the reasons the authors gave for this voluntary delay are two that would now be considered remarkable: 1) ‘‘Publication of the results would necessarily create hopes for a prompt enlargement of the scope of the investigation to certain forms of malignancy in clinical medicine for which answers could not be given’’, and 2) ‘‘The amount of 11-dehydro-17-hydroxycorticosterone available was so limited that extension of the investigation even to larger experimental animals was impossible. Also confirmation of the work by other laboratories could not be undertaken until a satisfactory source of the material became available.’’


316


CD4`CD8` (double positive, or DP) thymocytes, are among the few cell types that undergo apoptosis in response to corticosteroids [among the rare reports of other cells that have been observed to undergo glucocorticoid-induced apoptosis are osteoclasts and osteocytes (61, 62), dendritic cells (63), and some neuronal subsets (61)]. Despite the enormous strides made in our understanding of regulated cell death, the mechanism(s) by which glucocorticoids cause apoptosis is still largely unknown. What is known is that glucocorticoid-induced thymocyte apoptosis is mediated via the ‘‘mitochondrial’’ pathway: inhibitable by Bcl-2 and Bcl-xL and requiring Apaf-1 and caspase-9 (64–69). Since the GR is a transcriptional regulator, the simplest model is that glucocorticoids induce the expression of one or more gene products that directly or indirectly cause cell death. Data with thymocytes support this, since glucocorticoid-induced apoptosis requires ATP (70) and is prevented by inhibitors of protein synthesis (71, 72), although reportedly the latter is not true of splenic T cells (73). An alternative possibility is that the lethal activity of glucocorticoids could be indirect, due to interference with other transcription factors required for cell survival (transcriptional repression). Such a possibility was suggested by the finding that expression of a transcriptionally inactive form of the GR that can still interact with transcription factors such as AP-1 was capable of signaling for apoptosis when expressed in GR-negative Jurkat T cells (74). This observation was countered, however, by a perhaps more physiological study using mice in which the wild-type GR was replaced with a point mutant that cannot dimerize and, therefore, cannot directly transactivate gene transcription, although it can interact with other transcription factors (75). Although grossly normal, thymocytes from these animals were refractory to corticosteroid-induced apoptosis. Therefore, while receptor cross-talk may account for some biological responses to corticosteroids, at this time it seems likely that glucocorticoid-induced thymocyte apoptosis requires GR-mediated gene transactivation. Many attempts have been made to isolate steroid-induced genes that mediate cell death (76–80). Unfortunately, to date there are no convincing data that any of the candidates play such a role. However, a growing number of gene products have been implicated in blocking glucocorticoid-induced apoptosis. Among these are the now classic inhibitors of mitochondrial-dependent cell death such as Bcl2 and Bcl-xL (64, 66, 81) as well as IAPs (inhibitors of apoptosis), which are thought to work at least in part by directly binding and inhibiting some caspases, including caspase-3, -7, and -9 (82–85). In addition to these, a number of gene products are not ‘‘pure’’ apoptosis inhibitors but nonetheless prevent glucocorticoid-induced apoptosis. One example is Notch, a transmembrane receptor that has been implicated in regulating the CD4` vs. CD8` and perhaps the TCR ab vs. cd differentiation decisions in the thymus (86, 87), and which has a developmental pattern of expression reminiscent of Bcl-2: highly expressed in early thymocyte progenitors (CD41CD81), absent in DP thymocytes, and expressed at intermediate levels in single positive thymocytes (88). Overexpression of a



317

constitutively active intracellular portion of Notch in DP thymocytes rendered them relatively resistant to, and in T cell hybridomas completely prevented, glucocorticoid-induced apoptosis (89). Another example is RAP46 (the murine homolog is BAG-1), which binds the GR and inhibits its function (90). Inhibitors of the proteasome, a complex intracytoplasmic protease complex that degrades ubiquitinated substrates, prevent thymocyte apoptosis induced by p53-dependent and -independent stimuli such as corticosteroids (91). Also, as discussed in detail below, activation via the TCR (or other transmembrane molecules that can transduce activating signals) potently inhibits glucocorticoid-induced cell death (5, 92, 93). Reportedly the interaction of B7 with CD28 and/or CTLA-4 can rescue thymocytes even in the absence of TCR occupancy (94). Although these disparate molecules and signaling pathways do not at this time move us appreciably closer to the goal of defining the molecular mechanisms by which glucocorticoids kill, they do emphasize the multiple levels at which this process can be controlled. CD4`CD8` thymocytes are exquisitely sensitive to glucocorticoid-mediated cell death, and even the physiologic concentrations achieved during a stress response can be sufficient to cause their apoptosis (95–97). Resting peripheral T cells, however, are comparatively resistant to glucocorticoid-induced death (72). The reason for the difference appears to be the expression of Bcl-2, which is present in CD4`CD81 and CD41CD8` thymocytes and peripheral T cells but not in CD4`CD8` thymocytes (98). Although normally relatively resistant to glucocorticoid-induced apoptosis, TCRhi thymocytes and mature T cells derived from Bcl-2-deficient ES cells are just as sensitive as CD4`CD8` thymocytes (99). It is noteworthy in this regard that, after initially developing normally, the lymphoid organs of Bcl-2-deficient animals undergo massive apoptotic involution at approximately 4 weeks of age (99, 100), the same time at which circulating corticosteroids achieve adult levels (101, 102). It should be possible to determine if there is a cause-and-effect relationship between these two events by characterizing lymphoid development in adrenalectomized Bcl-2-deficient mice. Activation of peripheral T cells (which does not cause a decrease in Bcl-2) makes them more sensitive to glucocorticoids (103, 104), perhaps at least in part due to upregulation of GR levels (105). Glucocorticoid-induced apoptosis, unlike that induced by irradiation or genotoxic reagents, does not require p53 (106, 107).

GLUCOCORTICOIDS AND PERIPHERAL T CELLS Immunosuppression Immunosuppression is arguably the most widely appreciated effect of exogenously administered corticosteroids. Cortisone was first administered to patients suffering from rheumatoid arthritis in 1948, and soon thereafter glucocorticoids became a staple in the treatment of a myriad of autoimmune and inflammatory conditions (108, 109). Early evidence that glucocorticoids interfere with the


318


immune response was provided by the observation that cortisone or ACTH prolongs the survival of allogeneic skin grafts (transplantation immunity) (110, 111). Once it became possible to study lymphocytes in vitro, it was found that corticosteroids inhibit proliferative responses to a variety of mitogenic stimuli, largely by inhibiting the secretion of T-cell growth factor (i.e., the growth-promoting activity of T cell–derived lymphokines) (112). Corticosteroids are now known to inhibit the production of a large number of cytokines, including IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-8, IL-10, IL-13, GMCSF, TNFa, and c-interferon (113–116). In those instances in which the mechanism has been studied, inhibition was found to be largely due to interference with gene expression. In the case of IL-2, for example, it was initially reported that glucocorticoids inhibited a reporter construct containing the entire IL-2 enhancer but not a reporter construct containing a triplicated NF-AT binding site; this effect was mapped to two proximal regions in the IL-2 promoter that bind AP-1-like proteins (117). At the same time an AP-1 site in the IL-2 enhancer was reported to confer sensitivity to glucocorticoid inhibition, but only when present in conjunction with an NF-AT binding site (118). Another study found that Dex prevented binding of AP-1 and NF-AT to their corresponding binding sites in the IL-2 gene promoter and blocked activation of an AP-1 but not an NF-AT reporter (119). A direct effect of glucocorticoids on the generation of these and other important transcription factors seems unlikely. Initial studies using electrophoretic DNA mobility shift assays (gel shifts) to quantitate transcription factors reported that the levels of activation-induced transcription factors that bind the IL-2 enhancer, including NF-AT, AP-1, AP-3, Oct-1, and NF-jB, were not decreased by glucocorticoids (117, 120). The consensus from these studies seems to be that glucocorticoids generally do not prevent activation-induced generation of transcription factors; even in cases in which AP-1 binding to DNA was inhibited, there was no decrease in the actual amount of c-Fos and c-Jun protein in the nucleus (121). Rather, glucocorticoids appear to interfere with the binding and/ or function of critical transcription factors (notably AP-1 in the case of the IL-2 gene), probably by direct protein-protein interactions between these factors and the liganded GR. The differences between reports on the effect (or lack thereof) of glucocorticoids on gel shift assays may be accounted for at least in part by the methods used to prepare nuclear extracts, some of which do not recover the GR (117). Direct inhibition of gene transcription by the GR is another possible mechanism of interference, and in fact a functional ‘‘negative’’ GRE has been identified in the promoter region of the IL-1b gene (16). The literature is replete with reports of the effects of glucocorticoids on a host of transcription-dependent and -independent events that might be expected to affect the immune response. Glucocorticoids decrease the stability and half-life of mRNA encoding IL-1, IL-2, IL-6, IL-8, TNFa, and GM-CSF (122). Paradoxically, glucocorticoids also upregulate the expression of receptors for some of these same cytokines, including IL-1, IL-6, and GM-CSF, as well as IFN-c



319

(reviewed in 113). The effect on IL-1 receptors is not straightforward: Dex augments mRNA levels in human peripheral mononuclear cells for both type I and type II IL-1 receptors and causes the shedding of the latter as a soluble protein (123). The IL-1 type II receptor binds IL-1 but does not signal, and thus it is thought to act as an inactive ‘‘decoy’’ for IL-1. In the same vein, glucocorticoids were found to increase expression of intracellular IL-1 receptor antagonist (IL1ra) (124). Therefore, upregulation of IL-1 signaling antagonists may be yet another mechanism by which glucocorticoids mediate their anti-inflammatory effects. The effect of corticosteroids on IL-2 receptors is also not clear-cut. In fact, for T cells glucocorticoids have been reported to (a) increase IL-2Ra mRNA and protein, perhaps in synergy with IL-2 (125–127), and (b) decrease IL-2Ra and IL-2Rb mRNA and protein levels (128, 129). In those cases in which IL2Ra was decreased, it was likely a secondary effect due to inhibition of IL-2 production, since exogenous IL-2 prevented the downregulation. Whatever their effect on IL-2R expression, corticosteroids appear to directly decrease IL-2dependent, but not IL-4- or IL-9-dependent, T cell proliferation, perhaps by inhibiting proximal signaling via the IL-2R (130, 131). Among the other reported effects of glucocorticoids are downregulation of cell surface adhesion molecules like ICAM-1 and E-selection (43, 132), inhibition of CD40 ligand upregulation on activated CD4` T cells (133), interference with transcription of the CTL serine protease granzyme B (134), and even the direct inhibition of early TCR signaling events (135). While the literature is confusing and inconclusive in many respects, what these studies make clear is that the immunosuppressive activity of glucocorticoids is a reflection of its activity on multiple molecular targets. While the inhibition of cytokine production is widely accepted to be biologically significant, the degree to which the other reported events (inhibition/induction of receptors and receptor antagonists, inhibition of receptor signaling, downregulation of adhesion molecules, etc) contribute to immunosuppression in vivo is not well established.

Th Lineage Commitment Mature helper T cells can be divided into subsets that differ in the spectrum of cytokines they secrete (136). Th1 cells produce IL-2, IFN-c, and TNFb and contribute largely to T cell–mediated responses such as delayed-type hypersensitivity. Th2 cells produce IL-4, IL-5, IL-6, IL-10, and IL-13 and participate in humoral and allergic responses. The means by which helper T cells are induced to differentiate down either of these pathways has been an area of intense interest, and clearly the cytokines the subsets secrete, and perhaps the nature of the antigenpresenting cell they interact with during differentiation (137), have an enormous influence. There is also a body of data indicating that glucocorticoids participate in guiding the differentiation of helper T cells. For example, in early studies mice were implanted with sustained-release pellets containing Dex and then immunized


320


with OVA (138). Upon antigen restimulation in vitro, OVA-specific T cells from animals that had received Dex produced substantially lower amounts of IL-2 but higher amounts of IL-4. A similar skewing of lymphokine production away from IL-2 and toward IL-4 was observed when T cells from OVA-primed mice were rechallenged in vitro with OVA in the presence of low concentrations of glucocorticoids. It was deduced that the effect of Dex was directly on the T cells, because glucocorticoid treatment of a cloned T cell line and a T cell hybridoma also favored the production of IL-4. Consistent with this, Dex either had no effect (139) or it enhanced anti-TCR-induced proliferation of the murine D10 Th2 cell line, presumably by increasing its secretion of IL-4 (140). Other studies have shown that the cytokine profile of CD4` rat T cells that had initially been activated in the presence of Dex was skewed toward Th2 cytokines when subsequently activated in the absence of glucocorticoids (141). Interestingly, it has also been possible to generate antigen-specific Th2 cells from unprimed mice by repeated simulation in vitro with antigen in the presence of IL-2 and anti-IL-10 antibodies in serum-free medium, but only in the presence of glucocorticoids (142). In humans, acute glucocorticoid treatment of patients with nonsteroiddependent asthma results in depressed levels of IgG and IgA and increased levels of IgE (143). This activity may or may not reflect a direct effect on the Th subsets, however, because glucocorticoids can synergize with IL-4 to increase IgE synthesis by enhancing isotype switching, even in the absence of T cells (144–146). In fact, and in contrast to the observations with murine cells mentioned above, many reports indicate that glucocorticoids inhibit the production of Th2 cytokines by human T cells, leading to the suggestion that murine and human T cells may be fundamentally different in their susceptibility to glucocorticoids (116, 147, 148). This seems unlikely, however, and although glucocorticoids do inhibit both Th1 and Th2 cytokines by activated human T cells, the effect on Th1 cytokines may be more pronounced (149). Furthermore, IL-12 production (which favors Th1 development) by LPS-stimulated PBL was inhibited by glucocorticoids, while IL-10 production (favoring Th2 development) was relatively resistant (150). Therefore, if Th differentiation is regulated by the balance between Th1 and Th2 cytokines rather than the absolute amount produced, glucocorticoids might be expected to support the generation of Th2 cells in humans. Furthermore, although activation of rat T cells in the presence of glucocorticoids resulted in the generation of Th2 cells, addition of Dex during the subsequent activation of these Th2 cells inhibited production of IL-4, just as it inhibits IL-4 production by human T cells (151). Collectively, the data are consistent with the hypothesis that while glucocorticoids inhibit the acute production of both Th1 and Th2 cytokines, their presence during initial activation may nevertheless promote the differentiation to the Th2 phenotype. The finding that Dex inhibits IL-4 production but that exogenous IL-4 potentiates the ability of Dex to enhance the generation of Th2 cells (141, 151) suggests that IL-4 and glucocorticoids may promote differentiation of Th2 cells by different mechanisms.


321


Antagonism of Activation-Induced Apoptosis Activation-induced T cell apoptosis was first described with T cell hybridomas (152–154) and subsequently with pre-activated mature T cells and T cell clones (155–157). In the course of studies aimed at comparing activation- and glucocorticoid-induced apoptosis of T cell hybridomas, it was unexpectedly found that these lethal stimuli, when administered simultaneously, no longer caused cell death (92). This phenomenon was termed ‘‘mutual antagonism,’’ and was presumed to result from transcriptional interference between the GR and activationinduced transcription factors such as AP-1, although the relevant genes were not known (39). It is now clear that activation-induced apoptosis in these cells is caused by the upregulation of Fas ligand (FasL) expression and its subsequent interaction with Fas (158–161). Further studies found that glucocorticoids prevent activation-induced FasL upregulation but do not prevent signaling via Fas itself, demonstrating that interference with FasL expression by the GR is responsible for one arm of the mutual antagonism (161). One candidate for mediating this effect is the product of a newly described glucocorticoid-induced gene, GILZ (glucocorticoid-induced leucine zipper) (162). Stable expression of GILZ in T cell hybridomas was found to block activation-induced FasL upregulation and subsequent apoptosis. Our increasing appreciation that many instances of apoptosis rely on the expression of FasL has fueled interest in understanding the molecular regulation of this molecule. At least three transcription factors have been directly implicated in its upregulation by activation: NF-AT (163, 164), Egr2 and Egr-3 (165, 166), and NF-jB (167, 168). Of these, only the Egr family members are synthesized de novo, and the finding that inhibition of protein synthesis prevents activation-induced upregulation of fasL mRNA is consistent with the notion that these are critical mediators of activation-induced FasL upregulation (166). Furthermore, Egr-2 and -3 are themselves upregulated by NF-AT (165, 169, 170), and thus the contribution of NF-AT to FasL expression may be accounted for, at least in part, by indirect effects via the Egr transcription factors. In preliminary experiments, we have found that glucocorticoids inhibit fasL mRNA upregulation when cDNA encoding Egr-3 is introduced into HeLa cells (P Mittelstadt and JD Ashwell, unpublished data). It appears, then, that the interplay between the GR, perhaps its induced genes, and other transcriptional regulators may affect fasL transcription in an interesting but complex fashion. Another molecule that may participate in the antagonism of activation-induced apoptosis is GITR (glucocorticoid-induced TNFR family related), a transmembrane molecule homologous to the TNF/NFGR family members 4–1BB, CD27, and OX-40 that is expressed in thymocytes and peripheral T cells (171–173). GITR was initially identified as a murine glucocorticoid-induced gene that was also upregulated relatively late after activation, and its overexpression in T cell hybridomas inhibited CD3-mediated but not anti-Fas- or dexamethasone-induced apoptosis (171). However, although co-expression of the human homolog (hGITR or AITR) and its ligand, hGITRL/TL6, in Jurkat T cells also prevented anti-CD3-


322


induced cell death (presumably by interacting with TRAF2 to upregulate NF-jB activity), hGITR/AITR is upregulated by activation but not by glucocorticoids (172, 173). Thus, the possibility that GITR family members are glucocorticoidinduced genes that prevent activation-mediated apoptosis is uncertain. Finally, Nur77 is a nuclear orphan receptor whose induction is required for activationinduced death of T cell hybridomas and whose overexpression in thymocytes has been implicated in FasL upregulation, although as yet no Nur77 regulatory elements have been identified in the fasL promoter (174–176). Glucocorticoids inhibit the Nur77-dependent transcription of reporter constructs driven by either of the two known Nur77 DNA-binding response elements, one of which (NurRE) binds Nur77 homodimers and exhibits increased activity in T cells upon activation (177). Moreover, titration of Nur77 and the GR in transient transfection studies revealed that they each antagonize the transcriptional activity of each other (178). These results raise the possibility that antagonism between the GR and Nur77 may account, at least in part, for the observed antagonism between TCR- and glucocorticoid signaling in the induction of apoptosis. Inhibition of activation-induced T cell death is not confined to transformed T cell hybridomas. Encephalitogenic myelin basic protein–specific T cell lines undergo apoptosis when restimulated with antigen and IL-2 in vitro; cell death is prevented by simultaneous exposure to glucocorticoids (179). Glucocorticoids also prevent activation-induced death of human CD4` T cells acutely infected with HIV and then restimulated with anti-TCR antibodies, and prevent the accelerated apoptosis of cultured CD4` and CD8` T cells from people infected with HIV (180). Since the susceptibility of T cells from HIV` individuals to spontaneous and activation-induced apoptosis is largely conferred by increased Fas expression and sensitivity to FasL-mediated apoptosis (181, 182), it is likely that the protective effect of glucocorticoids is due to their suppression of FasL expression.

GLUCOCORTICOIDS IN THYMOCYTE DEVELOPMENT Thymocytes undergo an ordered series of phenotypic transitions as they differentiate from early precursors to mature T cells. Immature CD41CD81 (double negative) thymocytes that successfully rearrange the TCR b gene locus express a pre-TCR that consists of CD3-c,-d, and -e, a f homodimer, and TCR b heterodimerized with the nonpolymorphic pre-TCR a chain (183). Subsequent to preTCR expression, thymocytes begin to divide rapidly, acquire CD4 and CD8, and undergo TCR a gene rearrangement. For those cells that generate a functional a chain, the pre-TCR is replaced with low levels of the mature ab TCR (184). Due to the mostly random nature of the process by which TCR a and b chains are generated, each thymocyte will produce a unique ab TCR, and from this point forward thymocyte development is tightly linked to the specificity of the TCR for self-peptides bound to MHC-encoded molecules. The prevailing view at this



323

time is that TCRloCD4`CD8` thymocytes having TCRs with subthreshold avidity for self-antigen/MHC survive for approximately 3.5 days (185) and then undergo a default death pathway that has been called ‘‘death by neglect.’’ Thymocytes bearing TCRs that recognize self-antigen/MHC with high avidity undergo activation-induced apoptosis (negative selection), a major mechanism for promoting self-tolerance. The molecular mechanisms underlying this form of apoptosis are poorly defined, but antigen-specific negative selection appears to be largely independent of the Fas and TNF receptor signaling pathways (186– 188). Finally, thymocytes bearing TCRs with intermediate avidity for self antigen/ MHC are rescued from the default death pathway, differentiate into TCRhiCD4`CD81 or TCRhiCD41CD8`cells, and migrate to the periphery (positive selection). The outcome of these selection processes largely determines the mature T cell antigen-specific repertoire. The molecular means by which signaling via the TCR causes both rescue from default cell death and induction of cell death is an area of great interest. Presumably, low levels of TCR occupancy result in a change in expression or function of molecules involved in enhancing survival, while higher levels of occupancy alter the levels or activities of molecules that regulate apoptosis. Although the signaling pathways that lead to positive selection (which involves calcineurin and the Ras/MAP kinase signaling pathways) (189–191) and negative selection (which does not) may differ, the effector molecules regulated by these pathways have not been identified, making it difficult to determine how thymocytes make the critical decision of whether to live or die. However, considerable evidence suggests that glucocorticoids participate in this process and in doing so shape the peripheral T cell antigen-specific repertoire of adult animals. The initial observation that activation rescues T cell hybridomas from glucocorticoid-induced apoptosis (and vice versa) led to the speculation that similar interactions occurring in vivo could account for the relationship between TCR avidity for self-antigen/ MHC and cell fate (92, 93, 192). In this ‘‘mutual antagonism’’ model, DP thymocytes with subthreshold avidity for self-antigen/MHC undergo death (by neglect) at least in part because of glucocorticoids. Encounter of self-antigen/ MHC by thymocytes with intermediate avidity TCRs results in signaling that would otherwise lead to apoptosis, but due to antagonism by glucocorticoids the cells survive (positive selection). Finally, signals leading to apoptosis in thymocytes bearing TCRs with high avidity for self-antigen/MHC are too strong to be overcome by ambient corticosteroids, and these cells are deleted (negative selection). There are a few inferences from this model that are not necessarily obvious. First, in this scenario TCR occupancy at any level does not in itself deliver a ‘‘positive’’ signal. That is, low-to-moderate levels of TCR-mediated activation do not initiate signaling pathways that intrinsically promote survival while higher levels of activation initiate apoptotic signals. Rather, biologically relevant levels of TCR signaling result in apoptosis. Intermediate levels of TCR signaling can be viewed as positive only in that they counter apoptotic signals delivered via the


324


GR (and vice versa)—in the absence of glucocorticoids, it would be expected that these same intermediate levels of TCR signals would result in activation-induced apoptosis. This is in many ways analogous to TCR-mediated activation of mature T cells. TCR occupancy alone causes anergy and apoptosis; only in the presence of costimulation, typically via CD28, does TCR occupancy result in proliferation and survival (193). Second, in addition to apoptotic signals, intermediate levels of TCR occupancy must result in signals that lead to differentiation, as evidenced by increases in expression of molecules such as the TCR, CD5, and CD69, and extinguishing of expression of CD4 or CD8. Differentiation is of course thwarted in cells also induced to die but will proceed if death is prevented by glucocorticoids. If glucocorticoids play a significant role in thymocyte differentiation, they must necessarily be present in late fetal and neonatal life, a time during which a tremendous amount of thymocyte development occurs. However, the availability of circulating glucocorticoids is not uniform throughout development. In utero the placenta forms a partial barrier to transfer of maternal glucocorticoids (194), and in humans placental enzymes convert biologically active cortisol to the less active cortisone (195). Moreover, the expression of some steroidogenic enzymes is developmentally regulated, and as a result the levels of circulating glucocorticoids are low in fetal and neonatal life, not reaching adult levels in rodents until approximately 4 weeks of age (101, 102). These considerations prompted us to ask if the thymus itself produces glucocorticoids, and if so whether it does so during neonatal life. It was found that thymic epithelium, but not thymocytes, macrophages, or dendritic cells, produces pregnenolone and deoxycorticosterone (see Figure 1) when cultured in vitro, and that steroid production increased approximately twofold in response to ACTH (5). Immunohistochemical studies revealed that P450scc and P450c11 (Figure 1) are expressed mainly in cortical epithelium and that there is a subpopulation of large and intensely staining cells in a subcapsular and cortical periphery similar to that of thymic nurse cells. Subsequent studies have confirmed that thymic epithelial cells express steroidogenic enzymes and produce all of the steroids from pregnolone to corticosterone (196, 197). The ontogeny of thymic steroid production also appears to differ from that of the adrenal; cultured thymic epithelium from fetal and neonatal mice produced approximately twice as much as pregnenolone epithelium from four-week-old animals (5). A thymic epithelial cell line has been also reported to produce glucocorticoids, as evidenced by its ability induce apoptosis of a CD3`CD4`CD8` radiation leukemia virus–transformed thymocyte clone (197). This activity was inhibited by drugs that block either steroid production (aminoglutethimide) or GR occupancy (RU-38486). As with T cell hybridomas and normal thymocytes, death of the thymocyte clone was antagonized by activation via the TCR. Interestingly, the epithelial cell line appeared to produce glucocorticoids only when in contact with the CD3` thymocyte cell line, an activity that was not induced by contact with a CD31CD41CD81 thymocyte clone. Irradiated normal thymus cells were also reported to produce progesterone (3bHSD activity) but not corticosterone


325

(P450c11 activity) (196). These data raise the intriguing possibility that thymocyte-epithelial cell interactions may be required for the latter to produce glucocorticoids.


In Vitro Models The mutual antagonism model of thymocyte selection predicts that diminishing glucocorticoid levels or responsiveness should affect antigen-specific thymocyte selection by causing the activation-induced death of cells that would ordinarily be positively selected. That is, levels of TCR-mediated signaling that normally result in positive selection should now result in negative selection. This prediction has been tested in a number of different experimental model systems. Since the thymus produces its own steroids, it is possible to ask how thymocyte development proceeds in fetal thymic organ culture (FTOC) in which glucocorticoid production is prevented. Blockade of corticosteroid production with metyrapone, or responsiveness with RU-486, did in fact make thymocytes much more sensitive to apoptosis induced by anti-TCR antibodies or a low avidity ligand (5). Importantly, the effect of metyrapone was largely reversed by exogenous corticosterone, ruling out a pharmacologic effect or toxicity independent of the blockade of glucocorticoid production. The most direct test of the prediction was performed with mice expressing a transgenic ab TCR specific for the male H-Y antigen presented by the H-2Db class I molecule. Thymocytes from H-2bmale, but not female, mice bearing this TCR are negatively selected (198). Importantly, thymocytes from female mice undergo positive selection and mature into clonotypebearing CD41CD8` cells when expressed in an H-2b, but not an H-2d, animal, indicating that this transgenic TCR recognizes some unknown antigen plus an H2b-encoded molecule with low avidity (199). To determine if inhibition of glucocorticoid production could ‘‘turn positive into negative selection,’’ FTOC was performed with thymuses from RAG-2-deficient (to prevent expression of endogenous TCRs) H-2b female mice that expressed the anti-H-Y/Db TCR (200). Inhibition of corticosteroid production with metyrapone resulted in a substantial increase in DP apoptosis at 24 hr and a marked decrease in thymocyte recovery at 3 days that was largely prevented by the addition of physiologic levels of corticosterone. Notably, metyrapone did not induce apoptosis or affect thymocyte recovery when TCR ab transgenic littermates of the H-2d haplotype were analyzed, demonstrating that TCR occupancy was required for thymocyte death to occur.

In Vivo Models A number of different methods have been used to ask how glucocorticoids affect thymocyte development in vivo. One approach was to adrenalectomize pregnant rat dams, which resulted in early maturation of thymic stroma in the fetuses (201). Although this was interpreted as the direct result of glucocorticoid insufficiency, the possibility that this effect might be due to increased ACTH secretion by the


326


adrenalectomized dams and secondary upregulation of corticosteroid production by the thymic epithelium of the fetuses was not addressed. Another approach has been to generate transgenic mice that express antisense transcripts to the 3’ untranslated region of the GR. One set of animals (termed TKO mice) was created in which expression of the antisense transcripts is driven by the lck proximal promoter—a T lineage–specific promoter that is most active in immature thymocytes and relatively inactive in peripheral T cells (202). As expected, the transgene was expressed only in the thymus, and DP thymocytes homozygous for the antisense transgene had approximately a twofold reduction in GR mRNA and protein, with a corresponding reduction in sensitivity to corticosteroid-induced upregulation of a GRE-luciferase reporter construct and apoptosis (203). Thymuses from mice homozygous for the antisense transgene were as much as 90% smaller than controls due to a decrease in the number of DP thymocytes and a secondary decrease in CD4`CD81 and CD41CD8` thymocytes; heterozygous mice had an intermediate phenotype. A substantial increase occurred in the amount of spontaneous apoptosis of TKO DP thymocytes, and analysis of thymocyte ontogeny revealed that differences in thymocyte cellularity between wildtype and TKO mice were first apparent between d15 and d16 of gestation. As ab TCRs are not yet expressed at this stage of fetal development, it was speculated that glucocorticoids may be required to antagonize apoptotic signals delivered via the pre-TCR as well as the mature ab TCR. Furthermore, the DP cells that did develop in TKO mice were deleted by anti-TCR antibodies at approximately 100fold lower concentration than was required for wild-type cells, indicating that, like deprivation of glucocorticoids in FTOC, glucocorticoid hyporesponsiveness in vivo rendered DP thymocytes exquisitely sensitive to activation-induced apoptosis. A different antisense GR mouse, using a similar construct driven by a neurofilament promoter, has been analyzed as well (204, 205). Despite the presumably tissue-specific promoter, GR levels were found to be reduced two- to threefold in all tissues analyzed, including the anterior pituitary gland, hippocampus, liver, thymus, and spleen, with resultant increases in circulating ACTH and corticosterone, and hyporesponsiveness of the lymphoid cells to glucocorticoidmediated inhibition of mitogenesis. Unlike the TKO mice, there was no reduction in CD4`CD8` thymocytes, and if anything a small increase in total thymocyte number, with a failure of the thymus to regress after puberty. An increase in the ratio of CD4` to CD8` cells in the spleen was also noted. There are several possible resolutions for the substantial differences between the two GR antisense animal models. First, the antisense GR was restricted to thymocytes in the TKO mice but was expressed widely in the other strain, and it is possible if not likely that systemic consequences of glucocorticoid hyporesponsiveness can indirectly affect thymocyte and peripheral T cell development. Second, because of the tissue distribution of the antisense transgenes, circulating (and perhaps thymus-derived) corticosterone levels were elevated due to increased secretion of ACTH only in the second model, and it may be that this compensated for the decreased thymocyte sensitivity to glucocorticoids. A GR ‘‘knock-in’’ mouse has also been


327


reported in which the wild-type GR was replaced with a point mutant that cannot dimerize and, therefore, cannot directly regulate gene transcription. Although a detailed analysis of thymocyte phenotype was not performed, DP thymocytes in these mice were resistant to glucocorticoid-induced apoptosis, and the gross thymus phenotype (cellularity and CD4 and CD8 staining) was normal (75). Once again, there are multiple possible interpretations for this result, and although it argues that direct gene transactivation is necessary for glucocorticoids to induce thymocyte apoptosis, it does not address the possibility that corticosteroids regulate thymocyte development by more indirect mechanisms (e.g., by interfering with or enhancing the activities of other transcription factors). Careful analysis of thymocyte phenotype and function in these animals and, ultimately, in conditional GR knockout mice will be required to fully explore the effect of eliminating glucocorticoid responsiveness in immature thymocytes.

Proposed Mechanism for the Effect of Glucocorticoids on Antigen-Specific Thymocyte Development Both in vitro and in vivo methods of suppressing the response of thymocytes to corticosteroids have been used to address the mechanistic basis for possible effects of glucocorticoids on thymocyte selection. Metyrapone was used in FTOC to prevent local glucocorticoid production by thymuses from normal C57BL/6 (H2b) mice (206). Just as with the female H-2b mice bearing the anti-H-Y/Db ab TCR transgenes, this caused a decrease in DP thymocyte recovery (in this case approximately 50%) that was reversed by the addition of physiologic levels of free corticosterone (1 nM). If this decrease was due to deletion of cells with lowto-moderate avidity TCRs for self-antigen/MHC, it is much greater than expected given that only 3–4% of thymocytes undergo positive selection (207). To determine if TCR occupancy is indeed required for thymocyte loss in the absence of glucocorticoids, the availability of TCR ligands was varied by comparing the effect of metyrapone on thymuses from MHC-congenic C57BL mice. MHCcongenic animals differ in the number of different MHC-encoded molecules they express and therefore in the quantity and variety of potential TCR ligands. The following haplotypes (and the number of different MHC-encoded molecules they express) were tested: H-2d (five), H-2k (four), H-2b (three), and b2-microglobulindeficient H-2b (b2M -/-; one). A strict correlation was found between the ‘‘complexity’’ of the MHC and the loss of thymocytes caused by metyrapone, with H-2d mice (five MHC-encoded molecules) being the most affected and b2M -/mice (one MHC-encoded molecule) the least. The role of apoptosis in the decrease in thymocyte recovery was confirmed with a modified TUNEL assay after 24 hr of FTOC: metyrapone caused an increase in apoptotic cells that was directly related to the number of MHC molecules expressed (H-2d . H-2b . b2M -/-). Strikingly, when antigen-presentation was prevented by blocking the sole MHCencoded molecule in b2M -/- mice (I-Ab) with a monoclonal antibody, metyrapone actually caused an increase in cell recovery. Since no ligands are available


328


for TCR recognition under these conditions, all thymocyte death must be ‘‘by neglect,’’ and thus the enhanced thymocyte recovery caused by metyrapone provides direct evidence for the notion that glucocorticoids in fact participate in this default death pathway. These results demonstrate that TCR occupancy is required for the loss of local glucocorticoid production to result in DP thymocyte death, and they furthermore indicate that a large fraction (.50%) of the preselection TCR repertoire must recognize self-antigen/MHC with biologically significant avidity, consistent with recent data from a number of experimental systems (208–210). Moreover, since these self-reactive cells are ‘‘revealed’’ by the removal of glucocorticoids, the data suggest that glucocorticoids in fact prevent thymocytes bearing TCRs with relatively low (but still biologically significant) avidity for self-antigen/MHC from entering the positive selection ‘‘window.’’ To test this hypothesis, the effect of metyrapone on the levels of CD5, a transmembrane molecule whose upregulation on DP thymocytes is an early and sensitive measure of TCR/ligand interactions (209, 211), was assessed in FTOC of MHC-congenic H-2b/b2M -/- (one MHC-encoded molecule) and H-2a mice (5 MHC-encoded molecules). H-2a DP thymocytes exhibited approximately a 40% increase in CD5 levels when cultured with metyrapone. In contrast, metyrapone caused no appreciable change in CD5 expression on b2M -/- DP thymocytes. To address the possibility that CD4 and/ or CD8 interactions with MHC might have a primary role in this effect, wildtype H-2b mice were compared to H-2b mice deficient for TCRa, whose thymocytes progress to the DP stage of development but fail to express a mature ab TCR. Metyrapone upregulated CD5 expression on wild-type but not TCRadeficient DP thymocytes. Similar results were observed in the TKO mouse model in which GR levels are decreased in thymocytes: DP thymocytes from wild-type H-2k mice expressed lower levels of CD5 than DP thymocytes from MHCmatched TKO animals. These results indicate that there is a direct relationship between TCR occupancy, glucocorticoids, and cellular activation. Based upon these data, we suggest that for immature thymocytes, just as for mature peripheral T cells, glucocorticoids are ‘‘immunosuppressive.’’ That is, by a variety of mostly transcriptional mechanisms, glucocorticoids blunt the biological consequences of TCR signaling. In the case of peripheral T cells, this means inhibition of effector function; for DP thymocytes the consequence is inhibition of differentiation (positive selection) or apoptosis (negative selection). As shown in Figure 2, the simplest conventional model of thymocyte selection holds that the vast majority of thymocytes die by neglect, with only a small number bearing TCRs with sufficient avidity for self-antigen/MHC to undergo positive or negative selection. The proposed model holds that, in fact, most thymocytes are not ‘‘neglected,’’ but rather there is a bias of preselection TCRs toward those with an avidity for self-antigen/MHC sufficient to signal. Just as for mature T cells, the biological consequences of TCR signaling in DP thymocytes will be inhibited by glucocorticoids. Reductions in glucocorticoid signaling will increase the thymocyte response at any given degree of TCR occupancy, the result being



329

that cells bearing TCRs with avidities in the ‘‘normal’’ positive selection window will undergo activation-induced apoptosis, and some cells bearing TCRs with avidities that are normally inadequate for selection will now fall in the positive selection window. This model, therefore, predicts that alterations in glucocorticoid effects will change the average TCR avidity for selecting self-antigen/MHC and thus the peripheral T cell antigen–specific repertoire. The ability of glucocorticoids to regulate the T cell repertoire by setting the avidity window for selection is the most important implication of the mutual antagonism model. Two distinct experimental approaches employing the TKO GR antisense mice have been taken to test this hypothesis. The first was based upon the notion that the introduction of this transgene into an autoimmune mouse strain should decrease disease because of the decrease in the average TCR avidity for self-antigen/MHC. To do this, the GR antisense transgene was introduced into MRL-lpr/lpr mice (212), a model of spontaneous autoimmune disease that, due to a defect in the fas gene and the accompanying decrease in T cell activation– induced apoptosis, exhibits progressive lymphadenopathy because of the accumulation of T cells with the TCR`Thy-1`CD41CD81B220` cell surface phenotype. Although the number of thymocytes was decreased by the introduction of the antisense transgene, lymph node T cell numbers between the wild-type and lpr mice and the lpr.TKO were within 20% of each other by 7 weeks of age. Notably, although the proliferative response to mitogens, anti-TCR antibodies, and alloantigen was identical in these two animals, there was a marked and transgene-dose-dependent decrease in lymphadenopathy, anti-double stranded DNA antibodies, glomerulonephritis, and mortality in the TKO animals. Some TCR Vbs promote positive selection in mice of certain MHC haplotypes. This is not superantigen-mediated but is a consequence of recognition of unknown selfantigens presented by MHC-encoded molecules (213–215). As a result, overexpression of an affected Vb is seen in the CD4` or CD8` T cell subset, but not both. To determine if the beneficial effect of the transgene was indeed due to a change in the TCR repertoire, TCR Vb use was quantitated. In each of the five cases examined, T cells bearing Vbs that are normally overexpressed in one of the T cell subsets in H-2k mice (the MRL-lpr/lpr haplotype) were reduced in fractional representation toward the levels found in non-H-2k mice. This result is analogous to the FTOC results obtained with metyrapone and the anti-H-Y/Db ab TCR transgenic mice: A decrease in response to glucocorticoids leads to the deletion of cells that would otherwise be positively selected. Another approach to address the nature of the TCR repertoire of the TKO mice has been to look for ‘‘holes in the repertoire,’’ that is, antigens to which these mice cannot respond because the T cells that can respond to them are not present or are present at low numbers. Reasoning that the most easily demonstrated examples of nonresponsiveness, if they in fact exist, would be with small peptide antigens that generate a homogenous response, H-2k mice that do or do not express the antisense GR transgene were immunized with the 81–104 C-terminal fragment of pigeon cytochrome c in complete Freund’s adjuvant (F Lu and JD

330



Ashwell, manuscript in preparation). This peptide fragment is presented by I-Ek MHC-encoded molecules, and normally the large majority of responding T cells in these animals express TCRs containing Va11 and Vb3 (216). T cells were isolated from the draining lymph nodes of immunized mice and assayed for their response to restimulation. The proliferative response to anti-CD3 and to PPD (present in complete Freund’s adjuvant) was virtually identical between the two groups. However, while the wild-type mice responded well to pigeon cytochrome c fragment 81–104, the TKO mice proliferated very poorly. Moreover, the fre-

Figure 2 See next page for legend.


331

quency of Va11 and Vb3` T cells that responded to this antigen was markedly reduced in the TKO mice. It appears, therefore, that a responder animal has been converted to a nonresponder (or ‘‘hyporesponder’’) by a reduction in the expression of the GR during thymocyte development (recall that GR levels are normal in the peripheral T cells of TKO mice). Taken together, these data strongly support the prediction made by the mutual antagonism model that alterations in thymocyte glucocorticoid production or responsiveness will alter the antigen-specific T cell repertoire.


Implications One of the more intriguing possible predictions of the mutual antagonism model is that increases in local glucocorticoid production and/or responsiveness might prevent thymocytes bearing TCRs with relatively high avidity for self-antigen/ MHC from undergoing negative selection. These would represent untolerized cells that would provide a pool of potentially autoreactive T cells once in the periphery. At this time little evidence supports or militates against this possibility, but it is interesting to note that serum corticosterone levels are higher in autoimmune-prone mice than in normal animals (217) and that thymocytes from autoimmune obese chickens are resistant to glucocorticoid-induced apoptosis (218). It may be noteworthy in this regard that patients with the autoimmune disease multiple sclerosis also have significantly higher baseline plasma cortisol levels than do matched controls (219). Another interesting issue is the teleological reason for the involvement of corticosteroids in thymocyte selection. One possibility is that this evolved to deal with the problem posed by the elevation of systemic glucocorticoids that occurs during stress, including that incurred by acute infection. Given their potent immunosuppressive activity, it might be expected that Figure 2 Proposed relationship between TCR avidity for self-antigen/MHC and thymocyte fate. (A) The conventional model in which the majority of thymocytes die by neglect (subthreshold biologically relevant avidity). Small numbers of cells with intermediate avidity are positively selected (`) and with high avidity are negatively (1) selected. (B) The mutual antagonism model. In a mouse with a full complement of MHCencoded molecules, the large majority of thymocytes engage self-antigen/MHC with sufficient avidity to generate activating signals. In the presence of glucocorticoids (upper panel) most of these cells are inadequately activated to undergo positive selection and ultimately die in the thymus. As glucocorticoid levels are reduced (lower panel) their inhibition of activation is diminished, resulting in greater effective signal transduction at the same degree of receptor occupancy. Thus, occupancy of receptors that normally lead to only partial (inadequate) activation now lead to greater activation and either positive or negative selection. Note that this schematic does not address the question of whether fewer cells with very low avidity TCRs will die when glucocorticoid levels are reduced. Although there are data showing that in short-term FTOC thymocyte survival is enhanced by removing glucocorticoids, other factors that affect the viability of CD4`CD8` thymocytes may keep this population relatively constant in vivo.

332


weakly or suboptimally activated T cells would be rendered useless under these conditions. The requirement that only thymocytes bearing TCRs with sufficient avidity to function in the presence of glucocorticoids (as evidenced by activationinduced positive selection) would help ensure that mature T cells would be capable of activation and expression of effector functions in the face of elevated glucocorticoid levels. It might be possible to explore such a possibility by challenging TKO mice with pathogens and assessing the robustness and efficacy of the T-dependent immune response.


CONCLUDING REMARKS The anti-inflammatory and immunosuppressive effects of glucocorticoids are widely appreciated. What is less well known is that glucocorticoids influence T cells in a number of other, more subtle, ways. As detailed in this chapter, an extensive literature indicates that glucocorticoids are involved in Th lineage commitment, survival of activated T cells, and thymocyte development. Although much remains to be learned about molecular mechanisms, it is possible if not likely that all of these effects are mediated in large part by glucocorticoid-induced suppression of gene transcription. Fuller characterization and appreciation of the normal functions of glucocorticoids in vivo has been hampered by the multitude of effects glucocorticoids have on virtually all tissues, as well as the lack of a good GR-null model, since these animals die perinatally (220). It is to be expected, however, that the recent generation of viable mice with transcriptionally inactive GRs and, ultimately, mice with tissue-specific GR-deficiency will provide animal models that will make it possible to explore in more detail and with greater assurance the physiologic role for glucocorticoids in T cell development and the generation of the immune response. ACKNOWLEDGMENTS We are grateful to Barbara Osborne, Allan Weissman, and Georg Wick for critical review of this manuscript. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Mintami F, Shimizu T, Ueno R, Ishimura Y, Izumi S, Komatsu N, Watanabe K. 1982. Cytochrome P450–11b and P450scc in adrenal cortex: zonal distribution and intramitochondrial localiza-

tion by the horseradish peroxidaselabeled antibody method. J. Histochem. Cytochem. 30:1066–74 2. Geuze HJ, Slot JW, Yanagibashi K, McCracken JA, Schwartz AL, Hall PF.


3.

4.

5.


6.

7.

8.

9.

10.

11.

12.

13.

1987. Immunogold cytochemistry of cytochromes P-450 in porcine adrenal cortex. Histochemistry 86:551–57 Payne AH. 1990. Hormonal regulation of cytochrome P450 enzymes, cholesterol side-chain cleavage and 17a-hydroxylase/C17–20 lyase in Leydig cells. Biol. Reprod. 42:399–404 LeGoascogne C, Robel P, Gouezou M, Sananes N, Baulieu EE, Waterman M. 1987. Neurosteroids: cytochrome P450scc in rat brain. Science 237:1212–15 Vacchio MS, Papadopoulos V, Ashwell JD. 1994. Steroid production in the thymus: implications for thymocyte selection. J. Exp. Med. 179:1835–46 Keeney DS, Jenkins CM, Waterman MR. 1995. Developmentally regulated expression of adrenal 17 alpha-hydroxylase cytochrome P450 in the mouse embryo. Endocrinology 136:4872–79 Meikle AW, Tyler FH. 1977. Potency and duration of action of glucocorticoids. Effects of hydrocortisone, prednisone and dexamethasone on human pituitaryadrenal function. Am. J. Med. 63:200–7 Vinson GP, Whitehouse B, Hinson J. 1992. The Adrenal Cortex. Englewood Cliffs, NJ: Prentice-Hall Simpson ER, Waterman MR. 1988. Regulation of the synthesis of steroidogenic enzymes in adrenal cortical cells by ACTH. Annu. Rev. Physiol. 50:427– 40 Vinson GP, Whitehouse BJ, Hinson JP. 1998. Structure and function of the adrenal cortex. In Clinical Endocrinology, ed. A Grossman, p. 395. London: Blackwell Sci. Jones MT, Gillham B. 1988. Factors involved in the regulation of adrenocorticotropic hormone/b-lipotropic hormone. Physiol. Rev. 68:743–818 Beato M, Herrlich P, Schutz G. 1995. Steroid hormone receptors: many actors in search of a plot. Cell 83:851–57 Owens-Grillo JK, Hoffmann K, Hutchison KA, Yem AW, Deibel MR Jr, Hand-

14.

15.

16.

17.

18.

19.

20.

21.

22.

333

schumacher RE, Pratt WB. 1995. The cyclosporin A-binding immunophilin CyP-40 and the FK506-binding immunophilin hsp56 bind to a common site on hsp90 and exist in independent cytosolic heterocomplexes with the untransformed glucocorticoid receptor. J. Biol. Chem. 270:20479–84 Abraham N, Miceli MC, Parnes JR, Veillette A, Abbot S. 1991. Enhancement of T-cell responsiveness by the lymphocytespecific tyrosine protein kinase p56lck. Nature 350:62–66 Beato M. 1991. Transcriptional control by nuclear receptors. FASEB J. 5:2044– 51 Zhang G, Zhang L, Duff GW. 1997. A negative regulatory region containing a glucocorticosteroid response element (nGRE) in the human interleukin-1b gene. DNA Cell Biol. 16:145–52 Diamond MI, Miner JN, Yoshinaga SK, Yamamoto KR. 1990. Transcription factor interactions: selectors of positive or negative regulation from a single DNA element. Science 249:1266–72 Shibata H, Spencer TE, Onate SA, Jenster G, Tsai SY, Tsai MJ, O’Malley BW. 1997. Role of co-activators and corepressors in the mechanism of steroid/ thyroid receptor action. Recent Prog. Horm. Res. 52:141–64 Bannister AJ, Kouzarides T. 1996. The CBP co-activator is a histone acetyltransferase. Nature 384:641–43 Chen H, Lin RJ, Schiltz RL, Chakravarti D, Nash A, Nagy L, Privalsky ML, Nakatani Y, Evans RM. 1997. Nuclear receptor coactivator ACTR is a novel histone acetyltransferase and forms a multimeric activation complex with P/CAF and CBP/p300. Cell 90:569–80 Ogryzko VV, Schiltz RL, Russanova V, Howard BH, Nakatani Y. 1996. The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87:953–59 Spencer TE, Jenster G, Burcin MM, Allis

334

23.

24.


25.

26.

27.

28.

29.

30.

31.

ASHWELL n LU n VACCHIO CD, Zhou J, Mizzen CA, McKenna NJ, Onate SA, Tsai SY, Tsai MJ, O’Malley BW. 1997. Steroid receptor coactivator1 is a histone acetyltransferase. Nature 389:194–98 Yang XJ, Ogryzko VV, Nishikawa J, Howard BH, Nakatani Y. 1996. A p300/ CBP-associated factor that competes with the adenoviral oncoprotein E1A. Nature 382:319–24 Wade PA, Pruss D, Wolffe AP. 1997. Histone acetylation: chromatin in action. Trends Biochem. Sci. 22:128–32 Bodwell JE, Orti E, Coull JM, Pappin DJ, Smith LI, Swift F. 1991. Identification of phosphorylated sites in the mouse glucocorticoid receptor. J. Biol. Chem. 266:7549–55 Bodwell JE, Webster JC, Jewell CM, Cidlowski JA, Hu JM, Munck A. 1998. Glucocorticoid receptor phosphorylation: overview, function and cell cycledependence. J. Steroid Biochem. Mol. Biol. 65:91–99 Krstic MD, Rogatsky I, Yamamoto KR, Garabedian MJ. 1997. Mitogen-activated and cyclin-dependent protein kinases selectively and differentially modulate transcriptional enhancement by the glucocorticoid receptor. Mol. Cell. Biol. 17:3947–54 Weigel NL. 1996. Steroid hormone receptors and their regulation by phosphorylation. Biochem. J. 319:657–67 Mason SA, Housley PR. 1993. Sitedirected mutagenesis of the phosphorylation sites in the mouse glucocorticoid receptor. J. Biol. Chem. 268:21501–4 Almlof T, Wright AP, Gustafsson JA. 1995. Role of acidic and phosphorylated residues in gene activation by the glucocorticoid receptor. J. Biol. Chem. 270:17535–40 Jewell CM, Webster JC, Burnstein KL, Sar M, Bodwell JE, Cidlowski JA. 1995. Immunocytochemical analysis of hormone mediated nuclear translocation of wild type and mutant glucocorticoid

32.

33.

34.

35.

36.

37.

38.

39.

40.

receptors. J. Steroid Biochem. Mol. Biol. 55:135–46 Webster JC, Jewell CM, Bodwell JE, Munck A, Sar M, Cidlowski JA. 1997. Mouse glucocorticoid receptor phosphorylation status influences multiple functions of the receptor protein. J. Biol. Chem. 272:9287–93 Bamberger CM, Bamberger AM, de Castro M, Chrousos GP. 1995. Glucocorticoid receptor b, a potential endogenous inhibitor of glucocorticoid action in humans. J. Clin. Invest. 95:2435–41 Oakley RH, Sar M, Cidlowski JA. 1996. The human glucocorticoid receptor b isoform. Expression, biochemical properties, and putative function. J. Biol. Chem. 271:9550–59 Leung DYM, Hamid Q, Vottero A, Szefler SJ, Surs W, Minshall E, Chrousos GP, Klemm DJ. 1997. Association of glucocorticoid insensitivity with increased expression of glucocorticoid receptor b. J. Exp. Med. 186:1567–74 Schu¨le R, Rangarajan P, Kliewer S, Ransone LJ, Bolado J, Yang N, Verma IM, Evans RM. 1990. Functional antagonism between oncoprotein c-Jun and the glucocorticoid receptor. Cell 62:1217–26 Yang-Yen HF, Chambard JC, Sun YL, Smeal T, Schmidt TJ, Drouin J, Karin M. 1990. Transcriptional interference between c-Jun and the glucocorticoid receptor: mutual inhibition of DNA binding due to direct protein-protein interaction. Cell 62:1205–15 Jonat C, Rahmsdorf HJ, Park K-K, Cato ACB, Gebel S, Ponta H, Herrlich P. 1990. Antitumor promotion and antiinflammation: down-modulation of AP-1 (Fos/Jun) activity by glucocorticoid hormone. Cell 62:1189–204 Zacharchuk CM, Ashwell JD. 1992. Fruitful outcomes of intracellular crosstalk. Curr. Biol. 2:246–48 Gottlicher M, Heck S, Herrlich P. 1998. Transcriptional cross-talk, the second


41.

42.


43.

44.

45.

46.

47.

48.

mode of steroid hormone receptor action. J. Mol. Med. 76:480–89 Stein B, Baldwin ASJ, Ballard DW, Greene WC, Angel P, Herrlich P. 1993. Cross-coupling of the NF-jB p65 and Fos/Jun transcription factors produces potentiated biological function. EMBO J. 12:3879–91 Ray A, Prefontaine KE. 1994. Physical association and functional antagonism between the p65 subunit of transcription factor NF-jB and the glucocorticoid receptor. Proc. Natl. Acad. Sci. USA 91:752–56 Caldenhoven E, Liden J, Wissink S, Van de Stolpe A, Raaijmakers J, Koenderman L, Okret S, Gustafsson JA, Van der Saag PT. 1995. Negative cross-talk between RelA and the glucocorticoid receptor: a possible mechanism for the antiinflammatory action of glucocorticoids. Mol. Endocrinol. 9:401–12 Scheinman RI, Gualberto A, Jewell CM, Cidlowski JA, Baldwin AS Jr. 1995. Characterization of mechanisms involved in transrepression of NF-jB by activated glucocorticoid receptors. Mol. Cell. Biol. 15:943–53 Imai E, Miner JN, Mitchell JA, Yamamoto KR, Granner DK. 1993. Glucocorticoid receptor-cAMP response element-binding protein interaction and the response of the phosphoenolpyruvate carboxykinase gene to glucocorticoids. J. Biol. Chem. 268:5353–56 Zhang Z, Jones S, Hagod JS, Fuentes NL, Fuller GM. 1997. STAT3 acts as a coactivator of glucocorticoid receptor singaling. J. Biol. Chem. 272:30607–10 Stocklin E, Wissler M, Gouilleux F, Groner B. 1996. Functional interactions between Stat5 and the glucocorticoid receptor. Nature 383:726–28 Auphan N, Didonata JA, Rosette C, Hemberg A, Karin M. 1995. Immunosuppression by glucocorticoids: inhibition of NF-jB activity through induction of IjB synthesis. Science 270:286–90

335

49. Scheinman RI, Cogswell PC, Lofquist AK, Baldwin AS Jr. 1995. Role of transcriptional activation of IjBa in mediation of immunosuppression by glucocorticoids. Science 270:283–86 50. Brostjan C, Anrather J, Csizmadia V, Stroka D, Soares M, Bach FH, Winkler H. 1996. Glucocorticoid-mediated repression of NFjB activity in endothelial cells does not involve induction of IjBa synthesis. J. Biol. Chem. 271: 19612–16 51. Heck S, Bender K, Kullmann M, Gottlicher M, Herrlich P, Cato AC. 1997. IjBa-independent downregulation of NF-jB activity by glucocorticoid receptor. EMBO J. 16:4698–707 52. De Bosscher K, Schmitz ML, Vanden Berghe W, Plaisance S, Fiers W, Haegeman G. 1997. Glucocorticoid-mediated repression of nuclear factor-jB-dependent transcription involves direct interference with transactivation. Proc. Natl. Acad. Sci. USA 94:13504–9 53. Wissink S, van Heerde EC, vand der Burg B, van der Saag PT. 1998. A dual mechanism mediates repression of NFjB activity by glucocorticoids. Mol. Endocrinol. 12:355–63 54. Star P. 1895. An unusual case of Addison’s disease; sudden death; remarks. Lancet 1:284 55. Boinet E. 1899. Recherches expe´rimetales sur les fonctions des capsules surreńales. C.R. Seances Soc. Biol. Fil. 51:671– 74 56. Jaffe HL. 1924. The influence of the supradrenal gland on the thymus. I. Regeneration of the thymus following double supradrenalectomy in the rat. J. Exp. Med. 40:325–41 57. Jaffe HL. 1924. The influence of the suprarenal gland on the thymus. III. Stimulation of the growth of the thymus gland following double supradrenalectomy in young rats. J. Exp. Med. 40:753– 59 58. Selye H. 1936. Thymus and adrenals in

336

59.

60.


61.

62.

63.

64.

65.

66.

ASHWELL n LU n VACCHIO the response of the organism to injuries and intoxication. Br. J. Exp. Pathol. 17:234–48 Dougherty TF, White A. 1943. Effect of pituitary adrenotropic hormone on lymphoid tissue. Proc. Soc. Exp. Biol. Med. 53:132–33 Heilman FR, Kendall EC. 1944. The influence of 11-dehydro-17-hydroxycorticosterone (compound E) on the growth of a malignant tumor in the mouse. Endocrinology 34:416–20 Mitchell IJ, Cooper AJ, Griffiths MR, Barber DJ. 1998. Phencyclidine and corticosteroids induce apoptosis of a subpopulation of striatal neurons: a neural substrate for psychosis? Neuroscience 84:489–501 Weinstein RS, Jilka RL, Parfitt AM, Manolagas SC. 1998. Inhibition of osteoblastogenesis and promotion of apoptosis of osteoblasts and osteocytes by glucocorticoids. Potential mechanisms of their deleterious effects on bone. J. Clin. Invest. 102:274–82 Brokaw JJ, White GW, Baluk P, Anderson GP, Umemoto EY, McDonald DM. 1998. Glucocorticoid-induced apoptosis of dendritic cells in the rat tracheal mucosa. Am. J. Respir. Cell Mol. Biol. 19:598–605 Sentman CL, Shutter JR, Hockenbery D, Kanagawa O, Korsmeyer SJ. 1991. bcl2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 67:879–88 Siegel RM, Katsumata M, Miyashita T, Louie DC, Greene MI, Reed JC. 1992. Inhibition of thymocyte apoptosis and negative antigenic selection in bcl-2 transgenic mice. Proc. Natl. Acad. Sci. USA 89:7003–7 Grillot DA, Merino R, Nunez G. 1995. Bcl-XL displays restricted distribution during T cell development and inhibits multiple forms of apoptosis but not clonal deletion in transgenic mice. J. Exp. Med. 182:1973–83

67. Yoshida H, Kong YY, Yoshida R, Elia AJ, Hakem A, Hakem R, Penninger JM, Mak TW. 1998. Apaf1 is required for mitochondrial pathways of apoptosis and brain development. Cell 94:739–50 68. Hakem R, Hakem A, Duncan G. S, Henderson JT, Woo M, Soengas MS, Elia A, de la Pompa JL, Kagi D, Khoo W, Potter J, Yoshida R, Kaufman SA, Lowe SW, Penninger JM, Mak TW. 1998. Differential requirement for caspase 9 in apoptotic pathways in vivo. Cell 94:339–52 69. Kuida K, Haydar TF, Kuan C. Y, Gu Y, Taya C, Karasuyama H, Su MS, Rakic P, Flavell RA. 1998. Reduced apoptosis and cytochrome c-mediated caspase activation in mice lacking caspase 9. Cell 94:325–37 70. Stefanelli C, Bonavita F, Stanic{\hacek} I, Farruggia G, Falcieri E, Robuffo I, Pignatti C, Muscari C, Rossoni C, Guarnieri C, Caldarera CM. 1997. ATP depletion inhibits glucocorticoid-induced thymocyte apoptosis. Biochem. J. 322:909–17 71. Thomas N, Edwards JL, Bell PA. 1983. Studies of the mechanism of glucocorticoid-induced pyknosis in isolated rat thymocytes. J. Steroid Biochem. 18:519–24 72. Cohen JJ, Duke RC. 1984. Glucocorticoid activation of a calcium-dependent endonuclease in thymocyte nuclei leads to cell death. J. Immunol. 132:38–42 73. Perandones CE, Illera VA, Peckham D, Stunz LL, Ashman RF. 1993. Regulation of apoptosis in vitro in mature murine spleen T cells. J. Immunol. 151: 3521–29 74. Helmberg A, Auphan N, Caelles C, Karin M. 1995. Glucocorticoid-induced apoptosis of human leukemic cells is caused by the repressive function of the glucocorticoid receptor. EMBO J. 14:452–60 75. Reichardt HM, Kaestner K. H, Tuckermann J, Kretz O, Wessely O, Bock R, Gass P, Schmid W, Herrlich P, Angel P, Schutz G. 1998. DNA binding of the glu-


76.

77.


78.

79.

80.

81.

82.

83.

84.

cocorticoid receptor is not essential for survival. Cell 93:531–41 Harrigan MT, Baughman G, Campbell NF, Bourgeois S. 1989. Isolation and characterization of glucocorticoid- and cyclic AMP-induced genes in T lymphocytes. Mol. Cell. Biol. 9:3438–46 Baughman G, Harrigan MT, Campbell NF, Nurrish SJ, Bourgeois S. 1991. Genes newly identified as regulated by glucocorticoids in murine thymocytes. Mol. Endocrinol. 5:637–44 Baughman G, Lesley J, Trotter J, Hyman R, Bourgeois S. 1992. Tcl-30, a new T cell-specific gene expressed in immature glucocorticoid-sensitive thymocytes. J. Immunol. 149:1488–96 Owens GP, Hahn WE, Cohen JJ. 1991. Identification of mRNAs associated with programmed cell death in immature thymocytes. Mol. Cell. Biol. 11:4177–88 Goldstone SD, Lavin MF. 1991. Isolation of a cDNA clone, encoding a human bgalactoside binding protein, overexpressed during glucocorticoid-induced cell death. Biochem. Biophys. Res. Commun. 178:746–50 Memon SA, Petrak D, Moreno MB, Zacharchuk CM. 1995. A simple assay for examining the effect of transiently expressed genes on programmed cell death. J. Immunol. Methods 180:15–24 Deveraux QL, Takahashi R, Salvesen GS, Reed JC . 1997. X-linked IAP is a direct inhibitor of cell-death proteases. Nature 388:300–4 Duckett CS, Li F, Wang Y, Tomaselli KJ, Thompson CB, Armstrong RC. 1998. Human IAP-like protein regulates programmed cell death downstream of BclxL and cytochrome c. Mol. Cell. Biol. 18:608–15 Deveraux QL, Roy N, Stennicke HR, Van Arsdale T, Zhou Q, Srinivasula SM, Alnemri ES, Salvesen GS, Reed JC. 1998. IAPs block apoptotic events induced by caspase-8 and cytochrome c

85.

86.

87.

88.

89.

90.

91.

92.

93.

337

by direct inhibition of distinct caspases. EMBO J. 17:2215–23 Roy N, Deveraux QL, Takahashi R, Salvesen GS, Reed JC. 1997. The c-IAP-1 and c-IAP-2 proteins are direct inhibitors of specific caspases. EMBO J. 16:6914– 25 Robey E, Chang D, Itano A, Cado D, Alexander H, Lans D, Weinmaster G, Salmon P. 1996. An activated form of Notch influences the choice between CD4 and CD8 T cell lineages. Cell 87:483–92 Washburn T, Schweighoffer E, Gridley T, Chang D, Fowlkes BJ, Cado D, Robey E. 1997. Notch activity influences the ab versus cd T cell lineage decision. Cell 88:833–43 Hasserjian RP, Aster JC, Davi F, Weinberg DS, Sklar J. 1996. Modulated expression of notch1 during thymocyte development. Blood 88:970–76 Deftos ML, He YW, Ojala EW, Bevan MJ. 1998. Correlating notch signaling with thymocyte maturation. Immunity 9:777–86 Kullmann M, Schneikert J, Moll J, Heck S, Zeiner M, Gehring U, Cato AC. 1998. RAP46 is a negative regulator of glucocorticoid receptor action and hormoneinduced apoptosis. J. Biol. Chem. 273:14620–25 Grimm LM, Goldberg AL, Poirier GG, Schwartz LM, Osborne BA. 1996. Proteasomes play an essential role in thymocyte apoptosis. EMBO J. 15:3835–44 Zacharchuk CM, Mercep M, Chakraborti P, Simons SS Jr, Ashwell JD. 1990. Programmed T lymphocyte death: cell activation- and steroid-induced pathways are mutually antagonistic. J. Immunol. 145:4037–45 Iwata M, Hanaoka S, Sato K. 1991. Rescue of thymocytes and T cell hybridomas from glucocorticoid-induced apoptosis by stimulation via the T cell receptor/ CD3 complex: a possible in vitro model

338

94.

95.


96.

97.

98.

99.

100.

101.

ASHWELL n LU n VACCHIO for positive selection of the T cell repertoire. Eur. J. Immunol. 21:643–48 Wagner DH Jr, Hagman J, Linsley PS, Hodsdon W, Freed JH, Newell MK. 1996. Rescue of thymocytes from glucocorticoid-induced cell death mediated by CD28/CTLA-4 costimulatory interactions with B7–1/B7–2. J. Exp. Med. 184:1631–38 Gruber J, Sgonc R, Hu YH, Beug H, Wick G. 1994. Thymocyte apoptosis induced by elevated endogenous corticosterone levels. Eur. J. Immunol. 24:1115–21 Ayala A, Herdon CD, Lehman DL, DeMaso CM, Ayala CA, Chaudry IH. 1995. The induction of accelerated thymic programmed cell death during polymicrobial sepsis: control by corticosteroids but not tumor necrosis factor. Shock 3:259–67 Tarcic N, Ovadia H, Weiss DW, Weidenfeld J. 1998. Restraint stress-induced thymic involution and cell apoptosis are dependent on endogenous glucocorticoids. J. Neuroimmunol. 82:40–46 Hockenbery DM, Zutter M, Hickey W, Nahm M, Korsmeyer SJ. 1991. BCL2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc. Natl. Acad. Sci. USA 88:6961–65 Nakayama K, Nakayama K, Negishi I, Kuida K, Shinkai Y, Louie MC, Fields LE, Lucas PJ, Stewart V, Alt FW, Loh DY. 1993. Disappearance of the lymphoid system in Bcl-2 homozygous mutant chimeric mice. Science 261: 1584–88 Veis DJ, Sorenson CM, Shutter JR, Korsmeyer SJ. 1993. Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair. Cell 75:229–40 Henning SJ. 1978. Plasma concentrations of total and free corticosterone during development in the rat. Am. J. Physiol. 235:E451–56

102. Savu L, Zouaghi H, Nunez EA. 1985. Serum inflammatory responses of transcortin binding activities and of total and free corticosterone and progesterone levels in developing rats: a kinetic approach. Int. J. Tissue React. 7:443–48 103. Lanza L, Scudeletti M, Puppo F, Bosco O, Peirano L, Filaci G, Fecarotta E, Vidali G, Indiveri F. 1996. Prednisone increases apoptosis in in vitro activated human peripheral blood T lymphocytes. Clin. Exp. Immunol. 103:482–90 104. Horigome A, Hirano T, Oka K, Takeuchi H, Sakurai E, Kozaki K, Matsuno N, Nagao T, Kozaki M. 1997. Glucocorticoids and cyclosporine induce apoptosis in mitogen-activated human peripheral mononuclear cells. Immunopharmacology 37:87–94 105. Crabtree GR, Munck A, Smith KA. 1980. Glucocorticoids and lymphocytes. I. Increased glucocorticoid receptor levels in antigen-stimulated lymphocytes. J. Immunol. 124:2430–35 106. Clarke A. R, Purdie CA, Harrison DJ, Morris RG, Bird CC, Hooper ML, Wyllie AH. 1993. Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature 362:849–52 107. Lowe SW, Schmitt E. M, Smith SW, Osborne BA, Jacks T. 1993. p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature 362:847– 49 108. Hench PS, Kendall EC, Slocumb CH, Polley HF. 1950. Effects of cortisone acetate and pituitary ACTH on rheumatoid arthritis, rheumatic fever and certain other conditions. Arch. Int. Med. 85:545– 66 109. Plotz CM. 1975. Corticosteroids as immunosuppressive drugs: a brief perspective. Transplant. Proc. 7:5–6 110. Billingham RE, Krohn BA, Medawar PB. 1951. Effect of cortisone on survival of skin homografts in rabbits. Br. Med. J. 1:1157–63 111. Medawar PB, Sparrow EM. 1956. The


112.

113.


114.

115.

116.

117.

118.

119.

effects of adrenocortical hormones, adrenocorticotrophic hormone and pregnancy on skin transplantation immunity in mice. J. Endocrinol. 14:240–56 Crabtree GR, Gillis S, Smith KA, Munck A. 1980. Mechanisms of glucocorticoidinduced immunosuppression: inhibitory effects on expression of Fc receptors and production of T-cell growth factor. J Steroid Biochem 12:445–49 Almawi WY, Beyhum HN, Rahme AA, Rieder MJ. 1996. Regulation of cytokine and cytokine receptor expression by glucocorticoids. J. Leukoc. Biol. 60:563–72 Kunicka JE, Talle MA, Denhardt GH, Brown M, Prince LA, Goldstein G. 1993. Immunosuppression by glucocorticoids: inhibition of production of multiple lymphokines by in vivo administration of dexamethasone. Cell. Immunol. 149:39– 49 Fushimi T, Shimura S, Suzuki S, Saitoh H, Okayama H, Shirato K. 1998. Suppression of gene expression and production of interleukin 13 by dexamethasone in human peripheral blood mononuclear cells. Tohoku J. Exp. Med. 185:157–60 Rolfe FG, Hughes JM, Armour CL, Sewell WA. 1992. Inhibition of interleukin-5 gene expression by dexamethasone. Immunology 77:494–99 Northrop JP, Crabtree GR, Mattila PS. 1992. Negative regulation of interleukin 2 transcription by the glucocorticoid receptor. J. Exp. Med. 175:1235–45 Vacca A, Felli MP, Farina AR, Martinotti S, Maroder M, Screpanti I, Meco D, Petrangeli E, Frati L, Gulino A. 1992. Glucocorticoid receptor-mediated suppression of the interleukin 2 gene expression through impairment of the cooperativity between nuclear factor of activated T cells and AP-1 enhancer elements. J. Exp. Med. 175:637–46 Paliogianni F, Raptis A, Ahuja SS, Najjar SM, Boumpas DT. 1993. Negative transcriptional regulation of human interleukin 2 (IL-2) gene by glucocorticoids

120.

121.

122.

123.

124.

125.

126.

339

through interference with nuclear transcription factors AP-1 and NF-ATJ. Clin. Invest. 91:1481–89 Granelli-Piperno A, Nolan P, Inaba K, Steinman RM. 1990. The effect of immunosuppressive agents on the induction of nuclear factors that bind to sites on the interleukin 2 promoter. J. Exp. Med. 172:1869–72 Adcock IM, Brown C. R, Gelder CM, Shirasaki H, Peters MJ, Barnes PJ. 1995. Effects of glucocorticoids on transcription factor activation in human peripheral blood mononuclear cells. Am. J. Physiol. 268:C331–38 Tobler A, Meier R, Seitz M, Dewald B, Baggiolini M, Fey MF. 1992. Glucocorticoids downregulate gene expression of GM-CSF, NAP-1/IL-8, and IL-6, but not of M-CSF in human fibroblasts. Blood 79:45–51 Re F, Muzio M, De Rossi M, Polentarutti N, Giri J. G, Mantovani A, Colotta F. 1994. The type II ‘‘receptor’’ as a decoy target for interleukin 1 in polymorphonuclear leukocytes: characterization of induction by dexamethasone and ligand binding properties of the released decoy receptor. J. Exp. Med. 179:739–43 Levine SJ, Benfield T, Shelhamer JH. 1996. Corticosteroids induce intracellular interleukin-1 receptor antagonist type I expression by a human airway epithelial cell line. Am. J. Respir. Cell Mol. Biol. 15:245–51 Lamas M, Sanz E, Martin-Parras L, Espel E, Sperisen P, Collins M, Silva AG. 1993. Glucocorticoid hormones upregulate interleukin 2 receptor a gene expression. Cell. Immunol. 151:437–50 Fernandez-Ruiz E, Rebollo A, Nieto MA, Sanz E, Somoza C, Ramirez F, Lopez-Rivas A, Silva A. 1989. IL-2 protects T cell hybrids from the cytolytic effect of glucocorticoids. Synergistic effect of IL-2 and dexamethasone in the induction of high-affinity IL-2 receptors. J. Immunol. 143:4146–51


340


127. Wiegers GJ, Labeur MS, Stec IE, Klinkert WE, Holsboer F, Reul JM. 1995. Glucocorticoids accelerate anti-T cell receptor-induced T cell growth. J. Immunol. 155:1893–902 128. Boumpas DT, Anastassiou ED, Older SA, Tsokos GC, Nelson DL, Balow JE. 1991. Dexamethasone inhibits human interleukin 2 but not interleukin 2 receptor gene expression in vitro at the level of nuclear transcription. J. Clin. Invest. 87:1739–47 129. Batuman OA, Ferrero AP, Diaz A, Berger B, Pomerantz RJ. 1994. Glucocorticoidmediated inhibition of interleukin-2 receptor-a and -b subunit expression by human T cells. Immunopharmacology 27:43–55 130. Paliogianni F, Ahuja SS, Balow JP, Balow JE, Boumpas DT. 1993. Novel mechanism for inhibition of human T cells by glucocorticoids. Glucocorticoids inhibit signal transduction through IL-2 receptor. J. Immunol. 151:4081–89 131. Louahed J, Renauld JC, Demoulin JB, Baughman G, Bourgeois S, Sugamura K, Van Snick J. 1996. Differential activity of dexamethasone on IL-2-, IL-4-, or IL9-induced proliferation of murine factordependent T cell lines. J. Immunol. 156:3704–10 132. Cronstein BN, Kimmel SC, Levin RI, Martiniuk F, Weissmann G. 1992. A mechanism for the antiinflammatory effects of corticosteroids: the glucocorticoid receptor regulates leukocyte adhesion to endothelial cells and expression of endothelial-leukocyte adhesion molecule 1 and intercellular adhesion molecule 1. Proc. Natl. Acad. Sci. USA 89:9991–95 133. Bischof F, Melms A. 1998. Glucocorticoids inhibit CD40 ligand expression of peripheral CD4` lymphocytes. Cell. Immunol. 187:38–44 134. Wargnier A, Lafaurie C, Legros-Maida S, Bourge JF, Sigaux F, Sasportes M, Paul P. 1998. Down-regulation of human

135.

136.

137.

138.

139.

140.

141.

142.

143.

granzyme B expression by glucocorticoids. Dexamethasone inhibits binding to the ikaros and ap-1 regulatory elements of the granzyme b promoter. J. Biol. Chem. 273:35326–31 Baus E, Andris F, Dubois PM, Urbain J, Leo O. 1996. Dexamethasone inhibits the early steps of antigen receptor signaling in activated T lymphocytes. J. Immunol. 156:4555–61 Mosmann TR, Sad S. 1996. The expanding universe of T-cell subsets: Th1, Th2 and more. Immunol. Today 17:138–46 Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, Liu YJ. 1999. Reciprocal control of T helper cell and dendritic cell differentiation. Science 283:1183–86 Daynes RA, Araneo BA. 1989. Contrasting effects of glucocorticoids on the capacity of T cells to produce the growth factors interleukin 2 and interleukin 4. Eur. J. Immunol. 19:2319–25 Sierra-Honigmann MR, Murphy PA. 1992. T cell receptor-independent immunosuppression induced by dexamethasone in murine T helper cells. J. Clin. Invest. 89:556–60 Bertoglio JH, Leroux E. 1988. Differential effects of glucocorticoids on the proliferation of a murine helper and a cytolytic T cell clone in response to IL2 and IL-4. J. Immunol. 141:1191–96 Ramirez F, Fowell DJ, Puklavec M, Simmonds S, Mason D. 1996. Glucocorticoids promote a Th2 cytokine response by CD4` T cells in vitro. J. Immunol. 156:2406–12 Dozmorov IM, Miller RA. 1998. Generation of antigen-specific Th2 cells from unprimed mice in vitro: effects of dexamethasone and anti-IL-10 antibody. J. Immunol. 160:2700–5 Posey WC, Nelson HS, Branch B, Pearlman DS. 1978. The effects of acute corticosteroid therapy for asthma on serum immunoglobulin levels. J. Allergy Clin. Immunol. 62:340–48


GLUCOCORTICOIDS AND T CELLS 144. Jabara HH, Ahern DJ, Vercelli D, Geha RS. 1991. Hydrocortisone and IL-4 induce IgE isotype switching in human B cells. J. Immunol. 147:1557–60 145. Wu CY, Sarfati M, Heusser C, Fournier S, Rubio-Trujillo M, Peleman R, Delespesse G. 1991. Glucocorticoids increase the synthesis of immunoglobulin E by interleukin 4-stimulated human lymphocytes. J. Clin. Invest. 87:870–77 146. Bohle B, Willheim M, Baier K, Stadler B, Spitzauer S, Scheiner O, Kraft D, Ebner C. 1995. Hydrocortisone enhances total IgE levels—but not the synthesis of allergen-specific IgE—in a monocytedependent manner. Clin. Exp. Immunol. 101:474–79 147. Wu CY, Fargeas C, Nakajima T, Delespesse G. 1991. Glucocorticoids suppress the production of interleukin 4 by human lymphocytes. Eur. J. Immunol. 21:2645– 47 148. Wang Y, Campbell HD, Young IG. 1993. Sex hormones and dexamethasone modulate interleukin-5 gene expression in T lymphocytes. J. Steroid Biochem. Mol. Biol. 44:203–10 149. Franchimont D, Louis E, Dewe W, Martens H, Vrindts-Gevaert Y, De Groote D, Belaiche J, Geenen V. 1998. Effects of dexamethasone on the profile of cytokine secretion in human whole blood cell cultures. Regul. Pept. 73:59–65 150. Visser J, van Boxel-Dezaire A, Methorst D, Brunt T, de Kloet ER, Nagelkerken L. 1998. Differential regulation of interleukin-10 (IL-10) and IL-12 by glucocorticoids in vitro. Blood 91:4255–64 151. Ramirez F. 1998. Glucocorticoids induce a Th2 response in vitro. Dev. Immunol. 6:233–43 152. Ashwell JD, Cunningham RE, Noguchi PD, Hernandez D. 1987. Cell growth cycle block of T cell hybridomas upon activation with antigen. J. Exp. Med. 165:173–94 153. Mercep M, Bluestone JA, Noguchi PD, Ashwell JD. 1988. Inhibition of trans-

154.

155.

156.

157.

158.

159.

160.

161.

341

formed T cell growth in vitro by monoclonal antibodies directed against distinct activating molecules. J. Immunol. 140:324–35 Ucker DS, Ashwell JD, Nickas G. 1989. Activation driven T cell death. I. Requirements for de novo transcription and translation and association with genome fragmentation. J. Immunol. 143:3461–69 Russell JH, White CL, Loh DY, Meleedy-Rey P. 1991. Receptor-stimulated death pathway is opened by antigen in mature T cells. Proc. Natl. Acad. Sci. USA 88:2151–55 Ucker DS, Meyers J, Obermiller PS. 1992. Activation-driven T cell death. II. Quantitative differences alone distinguish stimuli triggering nontransformed T cell proliferation or death. J. Immunol. 149:1583–92 Wesselborg S, Janssen O, Kabelitz D. 1993. Induction of activation-driven death (apoptosis) in activated but not resting peripheral blood T cells. J. Immunol. 150:4338–45 Dhein J, Walczak H, Baümler C, Debatin K.-M, Krammer PH. 1995. Autocrine Tcell suicide mediated by APO-1/(Fas/ CD95). Nature 373:438–41 Brunner T, Mogil RJ, LaFace D, Yoo NJ, Mahboubi A, Echeverri F, Martin SJ, Force WR, Lynch DH, Ware CF, Green DR. 1995. Cell-autonomous Fas CD95 / Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas. Nature 373:441–44 Ju S-T, Panka DJ, Cui H, Ettinger R, ElKhatib M, Sherr DH, Stanger BZ, Marshak-Rothstein A. 1995. Fas(CD95)/ FasL interactions required for programmed cell death after T-cell activation. Nature 373:444–48 Yang Y, Mercep M, Ware CF, Ashwell JD. 1995. Fas and activation-induced Fas ligand mediate apoptosis of T cell hybridomas: inhibition of Fas ligand

342

162.

163.


164.

165.

166.

167.

168.

169.

170.

ASHWELL n LU n VACCHIO expression by retinoic acid and glucocorticoids. J. Exp. Med. 181:1673–82 D’Adamio F, Zollo O, Moraca R, Ayroldi E, Bruscoli S, Bartoli A, Cannarile L, Migliorati G, Riccardi C. 1997. A new dexamethasone-induced gene of the leucine zipper family protects T lymphocytes from TCR/CD3-activated cell death. Immunity 7:803–12 Latinis KM, Norian LA, Eliason SL, Koretzky GA. 1997. Two NFAT transcription factor binding sites participate in the regulation of CD95 (Fas) ligand expression in activated human T cells. J. Biol. Chem. 272:31427–34 Holtz-Heppelmann CJ, Algeciras A, Badley AD, Paya CV. 1998. Transcriptional regulation of the human FasL promoter-enhancer region. J. Biol. Chem. 273:4416–23 Mittelstadt PR, Ashwell JD. 1999. Role of Egr-2 in up-regulation of Fas ligand in normal T cells and aberrant doublenegative lpr and gld T cells. J. Biol. Chem. 274:3222–27 Mittelstadt PR, Ashwell JD. 1998. The cyclosporin A-sensitive transcription factor Egr-3 regulates Fas ligand expression. Mol. Cell. Biol. 18:3744–51 Matsui K, Fine A, Zhu B, Marshak-Rothstein A, Ju ST. 1998. Identification of two NF-jB sites in mouse CD95 ligand (Fas ligand) promoter: functional analysis in T cell hybridoma. J. Immunol. 161:3469–73 Kasibhatla S, Genestier L, Green DR. 1999. Regulation of fas-ligand expression during activation-induced cell death in T lymphocytes via NF-jB. J. Biol. Chem. 274:987–92 Mages HW, Baag R, Steiner B, Kroczek RA. 1998. Utilization of an NF-ATp binding promoter element for EGR3 expression in T cells but not fibroblasts provides a molecular model for the lymphoid cell-specific effect of cyclosporin A. Mol. Cell. Biol. 18:7157–65 Rengarajan J, Mittelstadt PR, Mages H-

171.

172.

173.

174.

175.

176.

177.

W, Gerth A, Kroczek R, Ashwell JD, Glimcher LH. 1999. Sequential involvement of NFAT and Egr transcription factors in Th1-specific FasL regulation. Submitted. Nocentini G, Giunchi L, Ronchetti S, Krausz LT, Bartoli A, Moraca R, Migliorati G, Riccardi C. 1997. A new member of the tumor necrosis factor/nerve growth factor receptor family inhibits T cell receptor-induced apoptosis. Proc. Natl. Acad. Sci. USA 94:6216–21 Gurney AL, Marsters SA, Huang RM, Pitti RM, Mark DT, Baldwin DT, Gray AM, Dowd AD, Brush AD, Heldens AD, Schow AD, Goddard AD, Wood WI, Baker KP, Godowski PJ, Ashkenazi A. 1999. Identification of a new member of the tumor necrosis factor family and its receptor, a human ortholog of mouse GITR. Curr. Biol. 9:215–18 Kwon B, Yu KY, Ni J, Yu GL, Jang IK, Kim YJ, Xing L, Liu D, Wang SX, Kwon BS. 1999. Identification of a novel activation-inducible protein of the tumor necrosis factor receptor superfamily and its ligand. J. Biol. Chem. 274:6056–61 Liu ZG, Smith SW, McLaughlin KA, Schwartz LM, Osborne BA. 1994. Apoptotic signals delivered through the T-cell receptor of a T-cell hybrid require the immediate-early gene nur77. Nature 367:281–84 Woronicz JD, Calnan B, Ngo V, Winoto A. 1994. Requirement for the orphan steroid receptor Nur77 in apoptosis of T-cell hybridomas. Nature 367:277–81 Weih F, Ryseck RP, Chen L, Bravo R. 1996. Apoptosis of nur77/N10-transgenic thymocytes involves the Fas/Fas ligand pathway. Proc. Natl. Acad. Sci. USA 93:5533–38 Philips A, Maira M, Mullick A, Chamberland M, Lesage S, Hugo P, Drouin J. 1997. Antagonism between Nur77 and glucocorticoid receptor for control of transcription. Mol. Cell. Biol. 17:5952– 59


GLUCOCORTICOIDS AND T CELLS 178. Philips A, Lesage S, Gingras R, Maira MH, Gauthier Y, Hugo P, Drouin J. 1997. Novel dimeric Nur77 signaling mechanism in endocrine and lymphoid cells. Mol. Cell. Biol. 17:5946–51 179. Pelfrey CM, Tranquill LR, Boehme SA, McFarland HF, Lenardo MJ. 1995. Two mechanisms of antigen-specific apoptosis of myelin basic protein (MBP)specific T lymphocytes derived from multiple sclerosis patients and normal individuals. J. Immunol. 1548:6191–202 180. Lu W, Salerno-Goncalves R, Yuan J, Sylvie D, Han DS, Andrieu JM. 1995. Glucocorticoids rescue CD4` T lymphocytes from activation-induced apoptosis triggered by HIV-1: implications for pathogenesis and therapy. AIDS 9:35–42 181. Baümler CB, Bo¨hler T, Herr I, Benner A, Krammer PH, Debatin K-M. 1996. Activation of the CD95 (APO-1/Fas) system in T cells from human immunodeficiency virus type-1-infected children. Blood 88:1741–46 182. Yang Y, Liu Z-H, Ware CF, Ashwell JD. 1997. A cysteine protease inhibitor prevents activation-induced T cell apoptosis and death of peripheral blood cells from HIV-infected individuals by inhibiting upregulation of Fas ligand. Blood 89:550–57 183. van Oers NS, von Boehmer H, Weiss A. 1995. The pre-T cell receptor (TCR) complex is functionally coupled to the TCR-f subunit. J. Exp. Med. 182:1585– 90 184. Kisielow P, von Boehmer H. 1995. Development and selection of T cells: facts and puzzles. Adv. Immunol. 58:87– 209 185. Huesmann M, Scott B, Kisielow P, von Boehmer H. 1991. Kinetics and efficacy of positive selection in the thymus of normal and T cell receptor transgenic mice. Cell 66:533–40 186. Singer GG, Abbas AK. 1994. The fas antigen is involved in peripheral but not thymic deletion of T lymphocytes in T

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

343

cell receptor transgenic mice. Immunity 1:365–71 Page DM, Roberts EM, Peschon JJ, Hedrick SM. 1998. TNF receptor-deficient mice reveal striking differences between several models of thymocyte negative selection. J. Immunol. 160:120–33 Gravestein LA, Borst J. 1998. Tumor necrosis factor receptor family members in the immune system. Semin. Immunol. 10:423–34 Alberola-Ila J, Hogquist KA, Swan KA, Bevan MJ, Perlmutter RM. 1996. Positive and negative selection invoke distinct signaling pathways. J. Exp. Med. 184:9–18 O’Shea CC, Crompton T, Rosewell IR, Hayday AC, Owen MJ. 1996. Raf regulates positive selection. Eur. J. Immunol. 26:2350–55 Urdahl KB, Pardoll DM, Jenkins MK. 1994. Cyclosporin A inhibits positive selection and delays negative selection in ab TCR transgenic mice. J. Immunol. 152:2853–59 Zacharchuk CM, Mercep, M, Ashwell, JD. 1991. Thymocyte activation and death: a mechanism for molding the T cell repertoire. In Antigen and CloneSpecific Immunoregulation, ed. RL Edelson, p. 52. New York: Ann. NY Acad. Sci. Noel PJ, Boise LH, Green JM, Thompson CB. 1996. CD28 costimulation prevents cell death during primary T cell activation. J. Immunol. 157:636–42 Ballard PL. 1979. Glucocorticoids and differentiation. In Glucocorticoid Hormone Action, ed. JD Baxter, GG Rousseau, p. 493. Berlin: Springer-Verlag Murphy BEP, Diez d’Aux RC. 1972. Steroid levels in the human fetus: cortisol and cortisone. J. Clin. Endocrinol. Metab. 35:678–83 Lechner O, Wiegers GJ, Oliveira-dosSantos AJ, Dietrich H, Recheis H, Waterman M, Boyd R, Wick G. 1999.

344

197.

198.


199.

200.

201.

202.

203.

204.

205.

ASHWELL n LU n VACCHIO Glucocorticoid production in the murine thymus. Eur. J. Immunol. In press Pazirandeh A, Xue Y, Rafter I, Sjovall J, Jondal M, Okret S. 1999. Paracrine glucocorticoid activity produced by mouse thymic epithelial cells. FASEB J. 13: 893–901 Kisielow P, Teh HS, Blu¨thmann H, von Boehmer H. 1988. Positive selection of antigen-specific T cells in thymus by restricting MHC molecules. Nature 335:730–33 Teh HS, Kisielow P, Scott B, Kishi H, Uematsu Y, Blu¨thmann H, von Boehmer H. 1988. Thymic major histocompatibility complex antigens and the ab T-cell receptor determine the CD4/CD8 phenotype of T cells. Nature 335:229–33 Vacchio MS, Ashwell JD. 1997. Thymus-derived glucocorticoids regulate antigen-specific positive selection. J. Exp. Med. 185:2033–38 Sacedon R, Varas A, Jimenez E, Munoz JJ, Vicente A, Zapata AG. 1999. Accelerated maturation of the thymic stroma in the progeny of adrenalectomized pregnant rats. Neuroimmunomodulation 6:23–30 Garvin AM, Abraham KM, Forbush KA, Farr AG, Davison BL, Perlmutter RM. 1990. Disruption of thymocyte development and lymphomagenesis induced by SV40 T-antigen. Int. Immunol. 2:173–80 King LB, Vacchio MS, Dixon K, Hunziker R, Margulies DH, Ashwell JD. 1995. A targeted glucocorticoid receptor antisense transgene increases thymocyte apoptosis and alters thymocyte development. Immunity 3:647–56 Pepin MC, Pothier F, Barden N. 1992. Impaired type II glucocorticoid-receptor function in mice bearing antisense RNA transgene. Nature 355:725–28 Morale MC, Batticane N, Gallo F, Barden N, Marchetti B. 1995. Disruption of hypothalamic-pituitary-adrenocortical system in transgenic mice expressing type II glucocorticoid receptor antisense

206.

207.

208.

209.

210.

211.

212.

213.

214.

ribonucleic acid permanently impairs T cell function: effects on T cell trafficking and T cell responsiveness during postnatal development. Endocrinology 136: 3949–60 Vacchio MS, Lee JYM, Ashwell JD. 1999. Thymus-derived glucocorticoids set the thresholds for thymocyte selection by inhibiting TCR-mediated thymocyte activation. J. Immunol. 163:1327–33 Shortman K, Egerton M, Spangrude GJ, Scollay R. 1990. The generation and fate of thymocytes. Semin. Immunol. 2:3–12 Zerrahn J, Held W, Raulet DH. 1997. The MHC reactivity of the T cell repertoire prior to positive and negative selection. Cell 88:627–36 Merkenschlager M, Graf D, Lovatt M, Bommhardt U, Zamoyska R, Fisher AG. 1997. How many thymocytes audition for selection? J. Exp. Med. 186:1149–58 Vidovic D, Boulanger N, Kuye O, Toral J, Ito K, Guenot J, Bluethmann H, Nagy ZA. 1997. The helper T-cell repertoire of mice expressing class II major histocompatibility complex b chains in the absence of a chains. Immunogenetics 45:325–35 Dutz JP, Ong CJ, Marth J, Teh HS. 1995. Distinct differentiative stages of CD4`CD8` thymocyte development defined by the lack of coreceptor binding in positive selection. J. Immunol. 154: 2588–99 Tolosa E, King LB, Ashwell JD. 1998. Thymocyte glucocorticoid resistance alters positive selection and inhibits autoimmunity and lymphoproliferative disease in MRL-lpr/lpr mice. Immunity 8:67–76 Mountz JD, Zhou T, Eldridge J, Berry K, Blu¨thmann H. 1990. Transgenic rearranged T cell receptor gene inhibits lymphadenopathy and accumulation of CD41CD81B220` T cells in lpr/lpr mice. J. Exp. Med. 172:1805–17 Perkins DL, Listman JA, Marshak-Rothstein A, Kozlow W, Kelley VR, Finn PW,


GLUCOCORTICOIDS AND T CELLS Rimm IJ. 1996. Restriction of the TCR repertoire inhibits the development of memory T cells and prevents autoimmunity in lpr mice. J. Immunol. 156:4961–68 215. Peng SL, Fatenejad S, Craft J. 1996. Induction of nonpathologic, humoral autoimmunity in lupus-prone mice by a class II-restricted, transgenic ab T cell. Separation of autoantigen-specific and -nonspecific help. J. Immunol. 157: 5225–30 216. Davis MM, McHeyzer-Williams M, Chien YH. 1995. T-cell receptor Vregion usage and antigen specificity. The cytochrome c model system. Ann. NY Acad. Sci. 756:1–11 217. Lechner O, Hu Y, Jafarian-Tehrani M, Dietrich H, Schwarz S, Herold M, Haour F, Wick G. 1996. Disturbed immunoendocrine communication via the hypoth-

345

alamo-pituitary-adrenal axis in murine lupus. Brain Behav. Immun. 10:337–50 218. Wick G, Sgonc R, Lechner O. 1998. Neuroendocrine-immune disturbances in animal models with spontaneous autoimmune diseases. Ann. NY Acad. Sci. 840:591–98 219. Michelson D, Stone L, Galliven E, Magiakou MA, Chrousos GP, Sternberg EM, Gold PW. 1994. Multiple sclerosis is associated with alterations in hypothalamic-pituitary-adrenal axis function. J. Clin. Endocrinol. Metab. 79:848–83 220. Cole TJ, Blendy JA, Monaghan AP, Krieglstein K, Schmid W, Aguzzi A, Fantuzzi G, Hummler E, Unsicker K, Schu¨tz G. 1995. Targeted disruption of the glucocorticoid receptor gene blocks adrenergic chromaffin cell development and severely retards lung maturation. Genes Dev. 9:1608–21

Note Added in Proof We referred to an article as “in preparation” on page 330. It has been accepted, and this is the current citation: Lu, F. W. M., Yasutomo, K., McHeyzer-Williams, L. J., McHeyzer-Williams, M. G., Goodman, G., Germain, R. N., and Ashwell, J. D. 2000. Thymocyte resistance to glucocorticoids leads to antigen-specific unresponsiveness due to “holes” in the T cell repertoire. Immunity. In press.



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


MOLECULAR GENETICS OF ALLERGIC DISEASES Santa Jeremy Ono Schepens Eye Research Institute and Brigham & Women’s Hospital and Committee on Immunology, Harvard University, Boston, Massachusetts, 02115; e-mail: [email protected]


Key Words atopy, genome analysis, linkage, inflammation Abstract Allergic diseases affect approximately one third of the general population. This class of disease, characterized by elevated serum IgE levels and hypersensitivity to normally innocuous antigen, can manifest in practically any mucosal tissue or as a systemic response. A few examples of serious allergic diseases include asthma, dermatitis, bee sting allergy, food allergy, conjunctivitis, and severe systemic anaphylaxis. Taken together, allergic diseases constitute one of the major problems of modern day medicine. A considerable portion of the healthcare budget is expended in the treatment of allergic disease, and morbidity rates of inner city asthmatics are rising steadily. Due to the enormity of the problem, there has been a worldwide effort to identify factors that contribute to the etiology of allergic diseases. Epidemiologic studies of multigeneration families and large numbers of twins clearly indicate a strong genetic component to atopic diseases. At least two independently segregating diseasesusceptibility genes are thought to come together with environmental factors to result in allergic inflammation in a particular tissue. On the basis of the strong genetic studies, multiple groups have attempted to identify disease-susceptibility genes via either a candidate gene approach or by genome-wide scans. Both of these approaches have implicated multiple regions in the human and mouse genomes, which are currently being evaluated as harboring putative atopy genes.

INTRODUCTION Atopy defines a general predisposition to develop allergic reactions to otherwise innocuous substances. These substances, allergens, although usually small airborne glycoproteins, can also be low molecular weight substances such as antibiotics, metalloproteins, or perfumes. Genetically predisposed individuals develop hypersensitivity to these substances during early childhood and adolescence. The initial encounter with an allergen is termed the ‘‘sensitization’’ phase. At this point, polypeptides are broken down into amino acid stretches of approximately two dozen amino acids, which are then bound and presented by class II major histocompatibility complex (MHC) molecules to T helper cells. Allergenspecific Th2 cells then communicate with allergen-reactive, surface immunoglobulin-positive B cells to promote their growth and differentiation. This cell-cell 0732–0582/00/0410–0347$14.00

347


348

ONO

communication results in the synthesis of allergen-reactive IgE molecules in sensitized individuals, thus priming them for subsequent encounters with the same allergen. Even prior to the next phase of the allergic process (the provocation or challenge phase), atopic individuals present an altered immunologic repertoire, compared to normal individuals. Atopic individuals have serum IgE levels that are up to one thousand-fold higher than that of a normal individual. Thus, completely separately from the allergen driven variations in specific IgE levels, atopic individuals are genetically predisposed to overproduce total IgE. During the ‘‘challenge’’ phase, IgE receptor-positive cells become involved in the disease process because allergen-reactive IgE molecules are bound to the surface of these cells. Mast cells, basophils, and eosinophils play a major role in this phase of the disease. Exposure to allergen during this period results in receptor cross-linking of membrane-bound IgE molecules on these cells. Receptor crosslinking results in activation of signal transduction from the IgE receptor that rapidly leads to release of both preformed and newly synthesized mediators. A few examples of these mediators include cytokines, histamines, leukotrienes, and tryptases. Each of these substances has been implicated in different aspects of the acute phase reaction (that which occurs within the first seconds to minutes) as well as in recruiting inflammatory cells that will drive the chronic phase of allergic diseases (which occurs usually several hours after initial encounter with allergen). This very cursory description of the phases of an allergic reaction already illustrates the complexity of the cell-cell communications and the array of molecules that might be altered in the atopic individual. Indeed, it is this very complexity that has made the identification of bona fide atopy genes a challenging process.

Vertical Transmission of Atopic Disease Many studies have been performed over the past 25 years to assess the familial nature of atopy. In practically every case, overwhelming evidence indicates that atopy or genetic predisposition for allergic disease is heritable. The initial studies of medical students and asthmatic children in London simply determined the prevalence of allergic disease in first degree relatives of affected individuals (1, 2). Invariably, the prevalence of allergic disease in first degree relatives of affected individuals was significantly higher than in relatives of unaffected individuals. Additional studies involving asthma, hay fever, atopic dermatitis, and allergic conjunctivitis, assessed over three generations and involving thousands of families, clearly indicate that each of these diseases are significantly overrepresented in relatives of atopic individuals. As these studies were performed on different continents, the impact of genetics on expression of allergic disease appears to be separate from any geographical component. Undoubtedly, the strongest evidence for a genetic component to allergic disease stems from the analysis of monozygotic and dizygotic twins. Analysis of


MOLECULAR GENETICS OF ALLERGIC DISEASES

349

several hundred pairs of twins clearly shows that quantifiable traits associated with allergic diseases such as total serum IgE levels, skin test results, and methacholine sensitivity in the lungs show intrapair correlation coefficients twofold higher for monozygotic than dizygotic twins (3). These data strongly suggest a genetic component to the expression of these traits. Additional strong evidence for vertical transmission of asthma, allergic rhinitis, and atopic dermatitis comes from a prospective study of several hundred newborns observed over a five-year period (4). While 51% of children with a family history of atopy developed allergic disease within the first 5 years of life, only 19% of children with no family history of atopy developed any symptoms of allergic disease within this span. Interestingly, as we have confirmed in more recent studies of allergic conjunctivitis, the tissue-specificity of the allergic reaction also appears to be inherited from one’s parents. Therefore, children of asthmatic parents are more likely to develop asthma than children from parents with atopic dermatitis and vice versa. Soon after the early studies of families and twins, several investigators tested the heritable nature of atopy in animal models. Using either guinea pigs or mice, sensitized to develop anaphylactic responses in the lung in response to ovalbumin, these studies clearly showed that the nature of both the inherited class II MHC gene and other independently segregating genes control the allergic response in the lung in these animal models (5). Several investigators have tried to determine a mode of inheritance of atopy in different human and animal populations. While certain susceptibility genes (such as class II MHC genes) appear to be inherited as autosomal dominant characters, it is very difficult to assign a mode of inheritance for atopy, as it is most certainly multifactorial and influenced by interaction with environmental factors (6).

Approach Toward Identification of Atopy Genes Once a clear genetic component for allergic diseases was established by segregation studies in humans and in animals, investigators immediately began research programs to try to identify regions of the genome that might harbor these susceptibility/protection genes. The methods used have evolved with time as more sophisticated approaches toward genome-wide analysis have emerged. In addition, as investigators have identified progressively larger numbers of cell surface and secreted molecules involved in the immune response, these have been included in the list of potential atopy genes, to be analyzed via a ‘‘candidate’’ approach. The candidate gene approach rests on the assumption that testing selected genes for association with the atopic phenotype (where the gene product encoded by that gene is involved in the immune response) is likely to identify an atopy gene more rapidly than a random, genome-wide approach. For example, since allergic responses require both the cytokine IL-4 and the process of immuno-


350

ONO

globulin class switching, both the IL-4 gene and components of a class-switch recombinase (perhaps specific for IgE) would be potential atopy genes evaluated under the candidate gene approach. Other genes, by their very biological function related to the allergic process (eotaxin 1, RANTES, VLA-4, and the mast-cell tryptases), are examples of potential atopy genes. Undoubtedly, as profiles of gene expression in inflamed tissues become carefully characterized, many more genes will be included in the candidate gene approach. Initial studies of ‘‘linkage’’ were carried out with studies of polymorphic surface markers, as gene probes were either absent or difficult to use. Thus, analyses of association with particular class II MHC types could be accomplished by serological analysis. When gene probes became available, these studies were complemented by restriction fragment length polymorphism analysis, and nucleotide sequence analysis. Somewhat later, with the advent of polymerase chain reaction technology, locus-specific oligonucleotide hybridization analysis dramatically improved the speed of genotyping individuals at specific loci. Most recently, the identification of broadly dispersed microsatellite repeats throughout the genome has allowed a detailed analysis of linkage in practically any part of the human genome. Once the human genome project has been completed, such linkage studies will become even more facile. Just as our knowledge of the human genome and our ability to rapidly genotype individuals throughout the genome facilitates the candidate gene approach, it has also ushered in the possibility of genome-wide screens for atopy genes. Indeed, multiple groups have already accomplished initial, low-density genome-wide screens for atopy or asthma genes (Table 1). These studies are likely to continue and to expand as our knowledge of the human genome and our analysis of both genotypic and phenotypic data become more sophisticated. Finally, two trends in genome analysis will likely greatly accelerate the speed of such linkage analyses: TABLE 1 Human genetic loci implicated in atopic disease Locus Chromosome 2 Chromosome 5q23-33 Chromosome 5p15 Chromosome 6p21 Chromosome 7p11.1-q11.2 Chromosome 8 Chromosome 11q13 Chromosome 12 Chromosome 13q21.3 Chromosome 14q11.2-13 Chromosome 17p11.1-11.2

Candidate Genes ? IL-4 cytokine cluster ? HLA D region genes, hsp70 ? ? FceRI-b near D12S97 ? ? ?


351

the push by the National Institutes of Health to streamline high throughput analyses of the genome and of genotyping, and the development of centers of proteomics/genomics at several institutions worldwide, should certainly expedite studies of this type. Improved programs for the analysis of complex traits and multiple segregating genes should also allow investigators to make sense of what is now often difficult to interpret.


Human Chromosome 6 (The Major Histocompatibility Complex) Serious studies on the segregation of atopy within multigeneration families began in the early 1980s. This coincided with a flood of interest in the products of the major histocompatibility complex. During this time, monoclonal antibodies were developed that could determine what polymorphic MHC product was produced in a given individual, and the first molecular probes for RFLP analyses were also being isolated. For these reasons, the genes of the major histocompatibility complex were the first ‘‘candidate’’ genes analyzed for association with atopy. Since the early 1980s when studies of MHC association of atopy began, dozens of studies have investigated this issue with respect to multiple atopic diseases and multiple-defined allergens. Although there is not complete agreement that class II MHC genes determine immune responsiveness to all allergens, they remain the strongest candidates for atopy genes within the human genome. The majority of studies into a potential linkage between a particular MHC allele and an allergic response confirm a strong role for MHC gene products in determining predisposition for the disease. One relatively early study analyzed approximately 100 individuals who were either positive or negative for a family history of asthma or atopy. Genomic DNA isolated from these individuals was genotyped at the DQ, DR, and DP genes by PCR-RFLP. Strikingly, this study indicated that the HLA-DR4 and DR7 alleles are associated with susceptibility for disease (7). Since these individuals were simply defined as atopic or nonatopic, the strong positive association suggests that the class II alleles confer susceptibility not via traditional peptide-binding mechanisms, but by a more general mechanism (perhaps during thymic selection). The investigators also provided some evidence that certain HLA-DQ alleles might confer protection against asthma. More recently, genome-wide searches for asthma-susceptibility loci in ethnically diverse populations, as well as in a United Kingdom population of asthmatics, have provided additional evidence for linkage between human chromosome 6p2l.3–23 and the asthma-associated phenotypes, bronchial hyperresponsiveness, and reversible airflow obstruction (8, 9). Two very recent studies of the HLA-D region provide an example of the power of careful studies of specific immune response to defined allergens (10, 11). In both cases, more than 135 individuals were tested to determine whether linkage could be demonstrated between specific markers on human chromosome 6p2l and high titers of Derpl-specific IgE antibodies. In one case, Caucasian families in


352

ONO

Baltimore were recruited, and in the second case a sample of the Venezuelan population was studied. In the Baltimore study, particular alleles of the D6S1281 and the DQCAR loci showed significant excess transmission of very specific IgE responses to Derp-p polypeptides. Other alleles at D6S1281 and D6S291 showed decreased transmission of specific IgE responses. While these results confirmed the Baltimore group’s previous findings that chromosome 6p2l controls responsiveness to Derp-p-specific immune responses, it provided clearer resolution of the complexity of the response. The Venezuelan study analyzed both class I and class II gene allele frequencies in the study population. Once again, linkage analysis was performed in relation to specific IgE levels to Derp p and Derp f. In agreement with the Baltimore study, a strong association was shown between atopic asthma in this population and HLA-Derpl*1101, DQA1*0501, DQB1*0301. Once again, in agreement with previous reports, other class II and class I alleles appear to have a protective role in the development of asthma in this population. There have also been reports of studies that do not confirm MHC association of atopic disease. In a very early report, a University of Maryland group studied 20 families during an 8-year period; the disease phenotype was bronchial hyperresponsiveness. Using highly polymorphic markers mapping into human chromosomes 11q and very near HLA-DR, these investigators were failed to show any association between these loci and atopy and bronchial hyperresponsiveness (12). One explanation for this result has been that these individuals were asthmatic due to a diverse array of allergens, and under such a situation one might not expect to find the tight linkage that has been observed in the carefully controlled Baltimore and Venezuelan studies. However, subsequent studies in Germany, New Zealand, and Spain also failed to show any evidence for significant association between any HLA class II locus and immune responsiveness to house dust mite allergens (13–15). Since these studies were analyzing immune responsiveness to the same allergens studied by the Baltimore group, it is more difficult to explain the discrepancies based upon peptide-presentation. In such a case, it is more likely that discrepancies are due either to differences in the phenotyping of the diseases in different locations and/or to geographical or racial differences in the mechanism of genetic predisposition. Despite these discrepancies, it seems clear that in many cases HLA class II MHC gene products do control immune responsiveness to particular allergens.

Human Chromosome 11 (FceRI) The human genetic locus that has received by far the most recent attention as a potential atopy locus is chromosome 11Q13. This linkage, although still controversial, was initially discovered during the analysis of seven families using a DNA polymorphism defined by the p lambda MS.51 probe, where inheritance appears to be dominant and assigned to chromosome 11q (16). The association between 11q and the IgE response immediately became controversial since the group of



353

Lee was unable to confirm linkage between D11F97 on chromosome 11q and either atopy or bronchial hyperresponsiveness to methacholine (17). Since this early phase of studies of 11q, the field has been split as to whether or not a real linkage exists between 11q and atopy. Investigators in support of a role for 11q in atopy have made significant progress in their analysis of the association. For example, since epidemiologic data has suggested that the risk of atopy is higher in children of atopic mothers than fathers, the Cookson group tested to see whether there was evidence of maternal inheritance of markers on chromosome 11q in their study population. Their analysis strongly suggested that the transmission of the ‘‘atopy’’ gene on chromosome 11 only occurs through the maternal line (18). These investigators interpreted the data to suggest that either genomic imprinting or maternal modification of the immune response affects the expression of the allergic phenotype. Subsequent follow-up studies have also honed in on a potential candidate gene on chromosome 11q (19). An analysis of 155 sibling-pairs indicated that the beta subunit of the high-affinity receptor for IgE (FORIP) maps into the susceptibility locus on 11q. To facilitate a careful molecular analysis of this region, Cookson and his collaborators next generated a 2.8 Mb YAC contig in the 11q12-q13-locus (20). Seven genes were mapped in the interval and were covered by at least 8 YAC clones. As indicated previously, the beta subunit of the high-affinity IgE receptor was found within this region as was a second marker, CD20. Attention naturally was focused primarily on the high-affinity IgE receptor, due to its clear role in the allergic process. Therefore, there was great excitement when the Cookson/Hopkin group identified a common variant of FceRI-P with an Ile181Leu substitution within the 4th-transmembrane domain. This variant showed significant association in the initial study with positive IgE responses in a random patient population (21). Moreover, the variant was maternally inherited and showed a strong association with both asthma and rhinitis. Since this was consistent with the initial mapping into 11q, the data strongly suggested that the atopy gene might be FceRI-b itself. Soon after the identification of variant forms of the high-affinity receptor, the group of Morimoto in Osaka provided important support for the 11q-association (22). Two hundred seventy atopic asthmatic individuals were analyzed by linkage analysis. The results from this Japanese population provided additional evidence for genetic linkage between atopy and human chromosome 11q13, and they supported the phenomenon of maternal genomic imprinting at that locus. During this time, the Hopkin/Cookson group constructed a more detailed genetic map of chromosome 11q using 15 markers spanning 27 cm. The atopy locus was mapped to a 7 cm interval between D11S480 and D11S451. Importantly, this interval was shown to contain FceRI-b (23). Using this refined map, the group then set out to establish the prevalence of particular FceRI-b polymorphisms in the general population (24). Greater than 1000 members of 230 twogeneration families were analyzed in this massive study. The study provided


354

ONO

strong evidence that maternal inheritance of FceRI-b Leul8l/Leul83 is indeed a genetic risk factor for atopy and bronchial hyperesponsiveness. Additional support for the FceRI-b association came from a large study performed by the group of Walters and coworkers (25). In the123 affected sibling pairs analyzed, there was clear evidence of significant linkage between a highly polymorphic microsatellite marker in the fifth intron of FceRI-b and asthma. Surprisingly, significant linkage was also observed in siblings sharing bronchial hyperreactivity, but not atopy. Moreover, in contrast to the previous data from the Cookson/Hopkin group, atopy in the absence of bronchial hyperreactivity did not show linkage to FceRI-b. The situation on human chromosome 11q became even more complex when the Morton group performed nonparametric linkage and association analyses using the NOPAR program (26). Parametric analyses were also performed with the complex inheritance with diathesis and severity (COMDS) program. Interestingly, on 11q, allele 168 at the D11S527 locus was associated with bronchial hyperreactivity, but not with log IgE levels. At the D11S534 locus, allele 235 was associated with log IgE, but not with bronchial hyperreactivity. In addition, both of the markers, D11S527 and D11S534, are quite distant from the FceRI-b gene. Therefore, while these experiments provided support for an atopy gene on 11q, they raised the possibility that perhaps two genes, linked to FceRI-b, but distinct from it, may control serum IgE levels and bronchial hyperesponsiveness. Additional support for the existence of an atopy gene on 11q13 came from a careful study of atopic dermatitis families in Germany (27). Using multiple informative markers on 11q, two markers (D11, F903 and FCER1B) showed evidence for strong linkage with both atopic dermatitis and atopy. Lod-score analysis mapped the FCER1B gene close to B11S903 and iss supportive of Cookson’s data implicating the IgE high-affinity receptor in an oligogenic mode of inheritance of atopy and atopic dermatitis. Similar studies by the Cookson group (28) also suggested a strong association of FceRI-b Rsal polymorphisms and atopic dermatitis. Very recent studies of polymorphisms of the beta chain of FceRI-b in South African black and white asthmatic and non-asthmatic individuals also provide support for a role for chromosome 11q in atopy (29). Using amplification refractory mutation system polymerase chain reaction ARMS-PCR, the prevalence of three mutations in the beta chain of the high-affinity IgE receptor (I181L, V183L, and E237G) in black and white asthmatic and control subjects in South Africa was documented. The extensive phenotyping of these amino acids suggests that I181L may dispose to atopy in the white population, but not in the black population. Moreover, a high prevalence of E2351 in blacks was interpreted as being a possible explanation for the increased severity of asthma in blacks and the higher mortality rate for black asthmatics than for whites. Most recently, the US Collaborative Study on the Genetics of Asthma (CSGA) also showed evidence for linkage to chromosome 11q13 in African-American families (30). In 58 African-American families, 328 individuals were evaluated



355

at both 5q31233 and 11q13. Linkage to 11q13 was observed when specific IgE responses to dermatophagoides farinae, cat, and Bermuda grass pollen allergens were analyzed. These data strongly support that an atopy gene resides at 11q13 and that this gene product is generally relevant for allergic responses to a multiplicity of allergens. Despite the strong support from many groups for the idea that an atopy gene exists on 11q, this review would be incomplete without a discussion of the many reports that do not confirm the existence of an atopy gene on 11q. Although there is astonishing detail in the data of Cookson and Hopkin, and the recent confirmatory report from the US CSGA provides compelling evidence for an atopy gene on 11q, the large number of nonconfirming studies also need to be evaluated. As indicated earlier in this review, the group of Lee, also in the United Kingdom, almost immediately challenged the initial report for linkage between 1lq and atopy. The same group rapidly followed up their initial work with an analysis of nine two-generation and, in some instances, three-generation families with asthma and atopy (31). Initially, restriction fragment length polymorphism analysis, using the PYGM and INT2 probes, was used to evaluate association between 11q12q13.2 and asthma/atopy. This study did not confirm linkage between 11q and atopy as defined by skin testing and specific IgE levels. Another study of 12 Australian pedigrees (32) also showed no evidence of association between atopy and l1q. Additional studies on subpopulations of these families also failed to show any evidence of linkage with this locus. Furthermore, analysis of affected sibpairs in this population showed no evidence for linkage between atopy and 11q. Kawakami and coworkers then tried to replicate the 11q13 linkage in a Japanese population (33). In this initial study, there was no evidence for any association between atopy and 11q13 in Japanese families. Similar studies by Pignatti and coworkers (33a), evaluating 45 Italian families with 213 subjects, failed to show significant maternal allele sharing and the association between ILE181LEU at FceRI-b and atopy. Similarly, Kofler and coworkers (34) failed to observe the ILE181LEU polymorphism in 40 unrelated atopic patients. This result suggested that the polymorphism occurs at a lower frequency than reported and may only be associated (35) with atopy in restricted populations. Holgate’s group also evaluated the potential contribution of 11q to atopy (36). Although their analysis of FceRI-b did not confirm the nucleotide changes in exon 6, they were able to identify an amino acid substitution in exon 7 that was strongly linked to asthma and atopy. Although their broad general analysis was unable to prove conclusively an association with 11q, the association observed with the new amino acid substitution was consistent with the previous reports of Cookson/Hopkin. A very recent sibling-pair analysis of 306 Japanese individuals also failed to find evidence for linkage of asthma or atopy with the IgE high-affinity receptor on 11q13. While these data are in direct conflict with those of Shirakawa and colleagues (21, 22), they suggested that certain ‘‘atopy’’ genes operate without geographical constraints (such as those on 5q3lq33), while others (such as that

356

ONO

on 11q) may only be relevant for subpopulations of a particular racial group (37). Such an interpretation of 11q association may also explain the lack of association between atopy and the RsaI polymorphism within intron 2 of the FceRI-b in a Spanish population (38). This analysis of the Rsal polymorphism (originally identified in one of the Japanese studies) showed that its frequency in this sample of Japanese patients did not differ from nonatopic control individuals. Taken together, although the confirmatory reports of 11q and atopy require that investigators continue to analyze the association carefully, the abundant nonconfirming reports also require that investigators seek to determine why the association holds in certain populations and not others. As the database of genotyping/ phenotyping information grows in different racial and geographic areas, the picture should clarify.


Human Chromosome 5q (The IL-4 Cytokine Cluster) The genomic location that has attracted the next greatest interest with respect to predisposition for allergic diseases is human chromosome 5q3l-q33. Two independent groups (those of Marsh and Bleecker—30, 40) originally detected evidence for association with this locus. As is the case with chromosome 11q, several follow-up studies have either confirmed or failed to confirm this initial association. Much like the situation with 11q, strong support for the existence of an atopy gene on 5q23–31 comes from a large study supported by the US CSGA (39). In this analysis of a founder population of Hutterites, markers in 5q23–31 did show possible linkage in both primary and replication samples with asthma phenotypes. Interestingly, none of the 12 markers associated with asthma or associated phenotypes were located in 11q. The group of Bleecker/Meyers extended their initial analysis of data from 92 families, showing recessive inheritance of high IgE levels and asthma to a twolocus segregation and linkage analysis. These studies provided strong evidence that a second locus, unlinked to 5q, contributes to these asthma phenotypes (40). Once again, the potential atopy gene was mapped to 5q3l-q33, with the D5S436 marker being the most informative marker when the two-locus model was used. When the Bleecker/Meyers group extended these studies further to a large siblingpair study of 303 children and grandchildren of 84 probands, the results showed linkage of bronchial hyperresponsiveness with several markers on chromosome 5q, including D5S436 (41). Taken together with the previous studies of this group and of Marsh, the data suggest that elevated serum IgE levels are coinherited with bronchial hyperresponsiveness and are controlled by a gene or genes on chromosome 5q3l-q33. Confirmatory results were obtained by the groups of Holgate/ Morton, using nonparametric linkage and association analyses (26). Using this rather conservative evaluation method, the data supported the view that chromosome 5q contains genes relevant to asthma and atopy. Levitt and coworkers (42) have taken the initial mapping of a human atopy gene to 5q3l-q33 into the murine system to facilitate the identification of the



357

relevant gene at this location. Since the homologs of candidate genes on 5q3lq33 are located on mouse chromosomes 18, 11, and 13, Levitt and coworkers assessed the linkage of bronchial hyperresponsiveness with markers on each chromosome. They found that bronchial hyperresponsiveness in mice was strongly linked to chromosome 13. Since the gene for interleukin-9 (IL-9) is located in this linked region, the investigators asked whether IL-9 might indeed be an atopy gene. Phenotypic analysis of IL-9 expression in bronchial hyperresponsive and control mice showed that IL-9 expression was markedly reduced in hyperresponsive mice. These data provide compelling evidence that IL-9 may very well be an atopy gene for human asthma. This must, certainly, be confirmed by careful analyses of IL-9 expression and linkage in the human population. Since the associations with 11q showed clear evidence of geographical or racial heterogeneity, it was important to see whether association with 5q might hold in other racial groups. An analysis of 306 Japanese individuals (37) via the siblingpair approach confirmed a linkage between asthma and gene markers in or near the IL-4, IL-9, and D5S393 markers on 5q3l-q33. Interestingly, the same analyses on 11q failed to show linkage with asthma or other atopy phenotypes. The data showed an association between 5q and childhood asthma in this population. Thus, this major Japanese study provides confirmatory evidence that an atopy gene resides on 5q3hqp3. Finally, the very recent study of the US CSGA provides even more detailed analysis of chromosome 5q linkage to particular IgE responses to defined allergens (30). In this study of African-American families (once again confirming the cross-racial relevance of 5q), linkage to chromosome 5q3l-q33 was observed in specific IgE responses to American cockroach and dog allergens. However, the atopy genes on 5q3l-q33 and 11q13 (which was also observed in IgE responses to other allergens) showed contrasting effects on atopy. Once again, as was stated in the analysis of 11q associations, these data underscore the importance of highly specific phenotypic information on allergic responses within each studied population, so that meaningful conclusions can be drawn from each study. Such detailed analysis can be performed retrospectively on databases that already exist, or it can be folded into future studies. One of the strongest studies indicating no linkage between chromosome 5q31 and asthma was performed on a Finnish founder population of 157 nuclear families (43). Sixteen polymorphic markers, including the IL-4 and IL-9 genes, were evaluated using both the sibling pair and cousin pair analyses. No evidence of genetic linkage between any of the markers on 5q and either serum IgE levels or asthma was found. Complex haplotype association studies were also carried out, and once again no evidence for linkage disequilibrium or association were observed. These data provide relatively compelling evidence that 5q31 does not contribute to inheritance of asthma or high IgE levels in this Finnish population. The Holgate group also investigated linkage to 5q in two distinct sample populations from the United Kingdom (36). Although allelic associations were identified in both regions, there was no significant evidence for linkage. In this case,


358

ONO

these investigators did observe linkage between an amino acid substitution in exon 7 and asthma and atopy. One other major study that has failed to confirm linkage of atopy to 5q31 involved a study of 45 sibling pairs in Australia. Using the polymorphic microsatellite marker D5S399, no evidence of linkage was found in this study (44). An even more extensive study starting with a set of 2415 adults, with a case/control sample of 181, also failed to observe any strong allelic association between a series of markers on 5q31–33 and either bronchial hyperresponsiveness or total serum IgE levels (45). There was some weak association between the markers D4S404, IRF-1, and D5S210, but the significance of those results is unclear. In this latter study, association analysis was performed using the nonparametric ChiZ Mann-Whitney test. Following the initial reports of association between 5q and serum IgE levels, two groups (Rosenwasser and Ono–45, 46) investigated whether nucleotide polymorphisms within IL-4 promoter sequences might affect transcriptional activity from the promoter and might explain association with markers of atopy. The studies of Rosenwasser focused on upstream (e.g. –590) polymorphisms. In our own studies (46), we screened a panel of B lymphoblastoid cell lines for polymorphisms within the proximal 200 nucleotides upstream of the transcription initiation site. A few polymorphic nucleotides were observed in this region, and some of these were found to affect transcriptional activity when heterologous reporter constructs were transfected into the Jurkat cell line. Subsequent studies of transcription factor interaction with these modified cis-elements indicated that these polymorphic nucleotides could affect the affinity of transcription factors–cis-element interaction, which in certain cases was consistent with promoter activity. The Rosenwasser group with their upstream polymorphic nucleotides also performed similar studies. Follow-up experiments of both the upstream and proximal promoter nucleotide substitutions have provided little evidence that these have a significant impact on either asthma or atopy (47, 48). However, these initial studies were useful in that they did demonstrate that such nucleotide polymorphisms could affect promoter activities, and they have spurred a comprehensive analysis of polymorphisms in several promoters of candidate atopy genes (e.g. RANTES and IL-10; SJ Ono, unpublished data; Table 2).

IL-4 Receptor a A fourth locus that has attracted a great deal of attention with respect to atopy is the gene encoding the interleukin-4 receptor a chain. The analysis of this region followed the discovery that the cytokines IL-4 and IL-13 exert their biological activities by binding to distinct heterodimer receptors, having the IL-4 receptor a chain as a shared component (49). Since IL-4 and IL-13 both induce IgE synthesis in B cells and promote Th2 differentiation, several investigators hypothesized that augmented signal transduction from their receptors might account for the atopic condition. Alternatively, signal transduction molecules linked to these


359

Table 2 Candidate atopy genes Gene


Class II MHC genes FceRI-b T cell receptor alpha locus RANTES IL-4 receptor alpha chain Mast cell chymase TAP1 b-adrenergic receptor

receptors, e.g. Stat 6, could also account for the elevated IgE synthesis and Th2biased profiles in atopic individuals. ‘The discovery of the IL-4 receptor a chain spurred genetic studies, in many laboratories, on the potential association of this locus with atopy. Preliminary results obtained from these laboratories indicated that the IL-4 receptor a chain was a viable atopy gene. Chatila and coworkers performed the pioneering studies on allelic forms of this receptor that might be associated with atopy (50). In this initial study, several dozen DNA samples obtained from allergic individuals were analyzed by single-strand confirmation polymorphism analysis and DNA sequencing to detect mutations in the IL-4 receptor a chain gene. Subjects with atopy were defined on the basis of elevated serum IgE levels and positive RAST tests. Importantly, signal transduction from the allelic IL-4 receptors was examined by multiple assays. A particular allelic form of IL-4 receptor a having a guanine-to-adenine substitution at nucleotide 1902 results in glutamine-toarginine chains at position 576. R576 is located in the cytoplasmic domain of the IL-4 receptor adjacent to tyrosine 575, which is the target of association with signal transduction molecules. R576 was strongly associated with atopy in patients with hyper-IgE syndrome and severe atopic dermatitis. Among a group of 50 prospectively recruited individuals, R576 was strongly associated with atopy (M.001). The relative risk for atopy in individuals with R576 was 9.3. Subsequent studies by Shirakawa, Izuhara, and Hopkin within the Japanese population also identified a polymorphism of the IL-4 receptor a chain associated with predisposition for atopic asthma (51). These investigators identified two variants of IL-4 receptor a . Four different cell lines were established with four types of IL-4 receptor a by stable transfection. Transfectants carrying either Val or Ile at position 50 or Gln or Arg at position 551 were then tested for responsiveness to IL-4. Only the Ile-to-Val substitution resulted in enhanced signal transduction from the IL-4 receptor, as assessed by Stat 6 activation, proliferation, and transcription from the Ie promoter. Importantly, Ile 50 was associated with atopic asthma in this Japanese population, and CD23 expression and IgE synthesis by IL-4 were augmented in peripheral blood mononuclear cells bearing the Ile 50 variant of IL-4 receptor a.

360

ONO

Therefore, two comprehensive and carefully controlled studies have provided evidence that mutations with the IL-4 receptor a chain can contribute to the atopic condition.


The BCL-6 Transcriptional Repressor A fifth gene that has been implicated as a potential atopy gene encodes the transcriptional repressor BCL-6. Initially identified by Dalla-Favera as the locus of frequent translocations; in non-Hodgkin’s lymphoma, BCL-6 is now known to encode a zinc-finger transcriptional repressor expressed in both B cells and CD4` T cells in germinal centers. Interestingly, the BCL-6 polypeptide binds sequencespecifically to DNA recognition motifs resembling those found by STAT transcription factors. Thus, the tissue-specific expression patterns of this factor as well as its DNA-binding specificity suggested that a normal function of the protein may be to regulate cytokine responses mediated by STAT molecules. Two groups generated BCL-6 knockout mice to investigate the normal function of BCL-6 (52). BCL-6-deficient mice had normal B cell and T cell development but exhibited a profound defect in T cell–dependent humoral responses. Germinal centers were lacking in BCL-6-deficient mice, and affinity maturation of antibody molecules was blocked. Dramatically, BCL-6-knockout mice developed multiple organ inflammation characterized by eosinophilia and an elevation of IgE bearing B cells. These studies, therefore, uncovered critical roles of BCL-6 in germinal center formation and in the regulation of the Th2 response. Since a Th2 hyperimmune response is a hallmark of atopic disease, mutations in BCL-6 or its promoter could very well contribute to predisposition to allergic disease. Since the BCL-6 promoter is a target for somatic hypermutation, a detailed screen for mutations in BCL-6 that might be associated with atopy is warranted. More research is also needed to properly understand the mechanism by which BCL-6 may contribute to a hyperimmune Th2 response. Although BCL-6 and STAT 6 interact with a highly homologous DNA recognition motif, the analysis of double-mutant (STAT 6-deficient/BCL 6-deficient mice) clearly indicates that Th2 inflammatory disease observed in BCL-6-deficient mice occurs via a STAT 6–independent mechanism.

Control of Th Development in Mice One other locus of considerable interest, Tpm l, has been discovered by Murphy and colleagues during their studies of genetic control of IL-12 responsiveness. The locus was discovered in their investigations of the molecular basis of developmental commitment in the Th lineage (54). They determined that signaling from the IL-4 receptor could reverse early Thl differentiation, but that IL-12 cannot reverse early Th2 differentiation. Th2 commitment results from a loss of responsiveness to IL-12 signaling. The IL-12 signaling defect then results in the failure of phosphorylation of Jak-Stat molecules.


361

Pioneering studies on the effect of genetic background on Thl/Th2 development showed clearly that the genetic background influences IL-12 responsiveness. In addition, IL-12 responsiveness was shown to have important implications for disease progression. Genetic mapping studies determined that mouse chromosome 11 or human chromosome 5q31 controls IL-12 responsiveness and the development of atopic, infectious, and autoirnmune disorders (55). The genetic locus, termed T cell phenotype modifier 1 (Tpm 1), was shown to control differential maintenance of IL-12 responsiveness. The analysis of a series of recombinants from first-generation backcrosses of B10.D2 and BALB/c has narrowed the genetic boundaries of the locus (35). In this process, IRF-1 and the IL-13/IL4 gene cluster were shown to be unlikely candidates for Tpm 1. Continuing efforts by this and other laboratories to pinpoint the relevant gene within Tpm 1 will likely uncover a gene playing a major role in predisposition for allergic diseases.


Other Chromosomal Regions Since several genome-wide screens have now been completed in the analysis of atopy, high IgE levels, and asthma, several other regions of the genome have attracted some attention as sites for other atopy genes. However, the data on these locations are too preliminary and, in general, the associations too weak for a detailed analysis at this time. The other regions that deserve a closer look based upon these studies include: chromosome 21 (10, 56), chromosome 8 (32, 56), chromosome 12 (8, 57, 58), and chromosome 17 (8). Other candidate genes that are attracting some attention include the b-adrenergic receptor (33, 59–61), the TAP1 gene (62), the T cell receptor b-chain locus (63), the RANTES gene (64), and the mast cell kinase gene (65, 66). Each of these candidate genes requires further scrutiny as a potential atopy gene.

CONCLUSIONS An enormous amount of effort throughout the world has been invested in the identification of atopy genes. This effort is justified in view of the large number of individuals affected with allergic diseases. Over the past 10 years, reasonable evidence has mounted that three genetic loci on chromosomes 5, 6, and 11 are likely to harbor atopy genes. However, each of these cases has involved considerable controversy because many groups have been unable to confirm these associations. In the most recent linkage analyses, it is somewhat comforting that when analyses are performed with highly specific data on IgE responses in each individual, evidence supporting these initial associations can often be found. The takehome lesson from this phase of investigation should be that definitive conclusions can only come after the multifactorial influences of race, geography, and multilocus contributions are taken into account. With the growing field of bioinformatics, and the power of analysis programs, the future remains bright that this

362

ONO

line of investigation will provide the clues for the genetic basis of allergic diseases. It is hoped, once the confounding problems for these sorts of analyses are corrected, the information obtained will be used to design novel therapies and to genotype individuals prior to clinical manifestation of disease.


ACKNOWLEDGMENTS National Institutes of Health grants R01 GM49661 and R01 EY1901 supported the author’s investigations in this area. The author is also grateful to the Lucille P. Markey Charitable Foundation for support of investigations in his laboratory at both Johns Hopkins and Harvard for the past eight years. The author is also grateful to his colleagues at Johns Hopkins and within the Committee on Immunology at Harvard for encouragement and stimulating discussions. Finally, the author would like to dedicate this article to: Drs. Abraham Fuks, Jack L. Strominger, Tom Maniatis, Lawrence M. Lichtenstein, Gilbert Jay, Tasuku Honjo, and J. Wayne Streilein, for everything. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Taylor B, Broom BC. 1981. Atopy in medical students. Ann. Allergy 47:197– 99 2. Sibbald B, Horn ME, Brain EA, Gregg 1. 1980. Genetic factors in childhood asthma. Thorax 35:671–74 3. Hopp RJ, Bewtra AK, Watt GD, Nair NM, Townley RG. 1984. Genetic analysis of allergic disease in twins. J. Allergy Clin. Immunol. 73:265–70 4. Luoma R, Koivikko A, Viander M. 1983. Development of asthma, allergic rhinitis and atopic dermatitis by the age of five years. A prospective study of 543 newborns. Allergy 38:339–46 5. Clausen JT, Sparck JV, Werdelin O, Larsen SO, Lundberg L. 1985. Immunogenetic studies of respiratory anaphylaxis in selectively bred guinea pigs. Inheritance of atopy, linkage to MFICcomplex. Allergy 40:15–22 6. Cookson WO ,Hopkin JM. 1988. Dominant inheritance of atopic immunoglobulin E responsiveness. Lancet 1:86–88

7. Aron Y, Desmazes-Dufeu N, Matran R, Polla BS, Dusser D, Lockhart A, Swierczewski E. 1996. Evidence of a strong, positive association between atopy and the HLA class II alleles DR4 and DR7. Clin. Exp. Allergy 26:821–28 8. Marsh DG, Ober C, Bleecker E. 1997. A genome-wide search for asthma susceptibility loci in ethnically diverse populations. Collaborative Study on the Genetics of Asthma (CSGA). Nat. Genet. 15:389–92 9. Daniels SE, Bhattacharrya S, James A, Leaves NI, Young A, Hill MR, Faux JA, Ryan GF, le Souef PN, Lathrop GM. 1996. A genome-wide search for quantitative trait loci underlying asthma. Nature 383:247–50 10. Hizawa N, Collins G, Rafnar T, Huang SK, Duffy DL, Weber JL, Freidhoff LR, Ehrlich E, Marsh DG, Beaty TH . 1998. Linkage analysis of Dermatophagoides pteronyssinus-specific IgE responsive-


11.


12.

13.

14.

15.

16.

17.

ness with polymorphic markers on chromosome 6p2l (HLA-D region) in Caucasian families by the transmission/ disequilibrium test. Collaborative Study on the Genetics of Asthma (CSGA). J. Allergy Clin. Immunol. 102:4418 Lara-Marquez ML, Yunis JJ, Layrisse Z, Ortega F, Carvallo-Gil E, Montagnani S, Makhatadze NJ, Pocino M, Granja C, Yunis E. 1999. Immunogenetics of atopic asthma: association of DRB1*1101 DQA1*0501 DQB1*0301 haplotype with Dermatophagoides spp.-sensitive asthma in a sample of the Venezuelan population. Clin. Exp. Allergy 29:60–71 Amelung PJ, Panhuysen CI, Postma DS, Levitt RC, Koeter GH, Francomano CA, Bleecker ER, Meyers DA. 1992. Atopy and bronchial hyperresponsiveness: exclusion of linkage to markers on chromosomes 11q and 6p. Clin. Exp. Allergy 22:1077–84 Stephan V, Schmid V, Frischer T, Sparholt S, Forster J, Wahn V, Kuehr J. 1996. Mite allergy, clinical atopy, and restriction by HLA class II immune response genes. Pediatr. Allergy Immunol. 7:28–34 Holloway JW, Doull I, Begishvili B, Beasley R, Holgate ST, Howell WM. 1996. Lack of evidence of a significant association between HLA-DR, DQ and DP genotypes and atopy in families with HDM allergy. Clin. Exp. Allergy 26: 1142–49 Torres-Galvan MJ, Quiralte J, Blanco C, Castillo R, Carrillo T, Perez-Aciego P, Sanchez-Garcia F. 1999. Linkage of house dust mite allergy with the HLA region. Ann. Allergy Asthma Immunol. 82:198–203 Cookson WO, Sharp PA, Faux JA, Hopkin JM. 1989. Linkage between immunoglobulin E responses underlying asthma and rhinitis and chromosome 11q. Lancet 1:1292–95 Lympany P, Welsh K, MacCochrane G, Kemeny DM, Lee TH. 1992. Genetic analysis using DNA polymorphism of

18.

19.

20.

21.

22.

23.

24.

363

the linkage between chromosome 11q13 and atopy and bronchial hyperresponsiveness to methacholine. J. Allergy Clin. Immunol. 89:619–28 Cookson WO, Young RP, Sandford AJ, Moffatt MF, Shirakawa T, Sharp PA, Faux JA, Julier C, Nakumuura Y, Nakumura Y . 1992. Maternal inheritance of atopic IgE responsiveness On chromosome 1lq [published erratum. appears in Lancet 1992 Oct 31; 340(8827):1110] [see comments]. Lancet 340:381–84 Sandford AJ, Shirakawa T, Moffatt MF, Daniels SE, Ra C, Faux JA, Young RP, Nakamura Y, Lathrop GM, Cookson WO. 1993. Localisation of atopy and beta subunit of high-affinity IgE receptor (Fc epsilon RI) on chromosome 11q [see comments]. Lancet 341:332–34 Stafford AN, Rider SH, Hopkin JM, Cookson WO, Monaco AP. 1994. A 2.8 Mb YAC contig in 11q12-q13 localizes candidate genes for atopy: Fc epsilon RI beta and CD20. Hum. Mol. Genet. 3:779–85 Shirakawa T, Li A, Dubowitz M, Dekker JW, Shaw AE, Faux JA, Ra C, Cookson WO, Hopkin JM. 1994. Association between atopy and variants of the beta subunit of the high- affinity immunoglobulin E receptor [see comments]. Nat. Genet. 7:125–29 Shirakawa T, Hashimoto T, Furuyama J, Takeshita T, Morimoto K. 1994. Linkage between severe atopy and chromosome 11q13 in Japanese families. Clin. Genet. 46:228–32 Sandford AJ, Moffatt MF, Daniels SE, Nakamura Y, Lathrop GM, Hopkin JM, Cookson WO. 1995. A genetic map of chromosome 11q, including the atopy locus. Eur. J. Hum. Genet. 3:188–94 Hill MR, James AL, Faux JA, Ryan G, Hopkin JM, le Souef P, Musk AW, Cookson WO. 1995. Fc epsilon RI-beta polymorphism and risk of atopy in a general population sample [see comments] [published erratum appears in Br. Med. J.

364

25.

26.


27.

28.

29.

30.

31.

ONO 1995. Nov 4;311(7014):11961.] Br. Med. J. 311:776–79 van Herwerden L, Harrap SB, Wong ZY, Abramson MJ, Kutin JJ, Forbes AB, Raven J, Lanigan A, Walters EH. 1995. Linkage of high-affinity IgE receptor gene with bronchial hyperreactivity, even in absence of atopy [see comments]. Lancet 346:1262–65 Doull IJ, Lawrence S, Watson M, Begishvili T, Beasley RW, Lampe F, Holgate T, Morton NE. 1996. Allelic association of gene markers on chromosomes 5q and 11q with atopy and bronchial hyperresponsiveness. Am. J. Respir. Crit.Care Med. 153:1280–84 Folster-Holst R, Moises HW, Yang L, Fritsch W, Weissenbach J, Christophers E. 1998. Linkage between atopy and the IgE high-affinity receptor gene at 11q13 in atopic dermatitis families. Hum. Genet. 102:236–39 Cox HE, Moffatt MF, Faux JA, Walley AJ, Coleman R, Trembath RC, Cookson WO, Harper JI. 1998. Association of atopic dermatitis to the beta subunit of the high affinity immunoglobulin E receptor. Br. J. Dermatol. 138:182–87 Green SL, Gaillard MC, Song E, Dewar JB, Halkas A. 1998. Polymorphisms of the beta chain of the high-affinity immunoglobulin E receptor (Fc epsilon RIbeta) in South African black and white asthmatic and nonasthmatic individuals. Am. J. Respir. Crit. Care Med. 158:1487– 92 Hizawa N, Freidhoff LR, Ehrlich E, Chiu YF, Duffy DL, Schou C, Dunston GM, Beaty TH, Marsh DG, Barnes KC . 1998. Genetic influences of chromosomes 5q3l-q33 and 11q13 on specific IgE responsiveness to common inhaled allergens among African American families. Collaborative Study on the Genetics of Asthma (CSGA). J. Allergy Clin. Immunol. 102:449–53 Lympany P, Welsh KI, Cochrane GM, Kemeny DM, Lee TH. 1992. Genetic

32.

33.

33a.

34.

35.

36.

37.

38.

analysis of the linkage between chromosome 11q and atopy. Clin. Exp. Allergy 22:108–192 Brereton HM, Ruffin RE, Thompson PJ, Turner DR. 1994. Familial atopy in Australian pedigrees: adventitious linkage to chromosome 8 is not confirmed nor is there evidence of linkage to the high affinity IgE receptor. Clin. Exp. Allergy 24:868–77 Hizawa N, Ohe M, Kawakami 1996. [Genetic analysis of bronchial asthma in Japanese population–Fc epsilon RI beta gene and beta 2 adrenergic receptor gene]. Nippon Rinsho. 54:539–43 Martinati LC, Trabetti E, Casartelli A, Boner AL, Pignatti PF. 1996. Affected sib-pair and mutation analyses of the high affinity IgE receptor beta chain locus in Italian families with atopic asthmatic children. Am J. Respir. Crit. Care Med. 153(5):1682–5 Kofler H, Aichberger S, Ott G, Casari A, Kofler R. 1996. Lack of association between atopy and the Ile181Leu variant of the beta-subunit of the high-affinity immunoglobulin E receptor. Int. Arch .Allergy Immunol. 111:44–47 Guler ML, Gorham JD, Dietrich WF, Murphy TL, Steen RG, Parvin CA, Fenoglio D, Grupe A, Peltz G, Murphy KM. 1999. Tpml, a locus controlling IL-12 responsiveness, acts by a cellautonomous mechanism. J. Immunol. 162:1339–47 Thomas NS, Wilkinson J, Holgate ST. 1997. The candidate region approach to the genetics of asthma and allergy. Am. J. Respir. Crit. Care Med. 156:S144–51 Noguchi E, Shibasaki M, Arinami T, Takeda K, Maki T, Miyamoto T, Kawashima T, Kobayashi K, Hamaguchi H. 1997. Evidence for linkage between asthma/atopy in childhood and chromosome 5q3l-q33 in a Japanese population. Am. J. Respir. Crit. Care Med. 156:1390– 93 Castro J, Telleria JJ, Blanco-Quiros A,


39.


40.

41.

42.

43.

44.

Linares P, Andion R. 1998. Lack of association between atopy and Rsal polymorphism within intron 2 of the Fc(epsilon)RIbeta gene in a Spanish population sample. Allergy 53:1083–86 Ober C, Cox 54, Abney M, Di Rienzo A, Lander ES, Changyaleket B, Gidley H, Kurtz B, Lee J, Nance IV. 1998. Genome-wide search for asthma susceptibility loci in a founder population.Collaborative Study on the Genetics of Asthma. Hum. Mol. Genet. 7:1393–98 Xu J, Levitt RC, Panhuysen CI, Postma DS, Taylor EW, Amelung PJ, Holroyd KJ, Bleecker ER, Meyers DA. 1995. Evidence for two unlinked loci regulating total serum IgE levels. Am. J. Hum. Genet. 57:425–30 Postma DS, Bleecker ER, Amelung PJ, Holroyd KJ, Xu J, Panhuysen Cl, Meyers DA, Levitt RC. 1995. Genetic susceptibility to asthma–bronchial hyperresponsiveness coinherited with a major gene for atopy. N. Engl. J. Med. 333:894–900 Holroyd KJ, Martinati LC, Trabetti E, Scherpbier T, Eleff SM, Boner AL, Pignatti PF, Kiser MB, Dragwa CR, Hubbard F, Sullivan CD, Grasso L, Messler CJ, Huang M, Hu Y, Nicolaides NC, Buetow KH, Levitt RC. 1998. Asthma and bronchial hyperresponsiveness linked to the XY long arm pseudoautosmal region. Genomics 52(2):233–535 Laitinen T, Kauppi P, Ignatius J, Ruotsalainen T, Daly MJ, Kaariainen H, Kruglyak L, Laitinen H, de la Chapelle A, Lander ES . 1997. Genetic control of serum IgE levels and asthma: linkage and linkage disequilibrium studies in an isolated population. Hum. Mol. Genet. 6:2069–76 Kamitani A, Wong ZY, Dickson P, van Herwerden L, Raven J, Forbes AB, Abramson MJ, Walters EH, Harrap SB. 1997. Absence of genetic linkage of chromosome 5q31 with asthma and atopy in the general population. Thorax 52:8167

365

45. Rosenwasser LJ. Promoter polymorphism in the candidate genes, IL-4, IL-9, TGF-beta1, for atopy and asthma. Int Arch Allergy Immunol 1999 Feb–Apr; 118(2-4):268–70 46. Deleted in proof 47. Noguchi E, Shibasaki M, Arinami T, Takeda K, Yokouchi Y, Kawashima T, Yanagi H, Matsui A, Hamaguchi H. 1998. Association of asthma and the interleukin-4 promoter gene in Japanese. Chin. Exp. Allergy 28:449–53 48. Walley AJ,Cookson WO. 1996. Investigation of an interleukin-4 promoter polymorphism for associations with asthma and atopy. J. Med. Genet. 33:689–92 49. Izuhara K, Shirakawa T. 1999. Signal transduction via the interleukin-4 receptor and its correlation with atopy. Int. J. Mol. Med. 3:3–10 50. Hershey GK, Friedrich MF, Esswein LA, Thomas ML, Chatila TA. 1997. The association of atopy with a gain-of-function mutation in the alpha subunit of the interleukin-4 receptor. N. Engl. J. Med. 337:1720–25 51. Adra CN, Mao XQ, Kawada H, Gao PS, Korzycka B, Donate JL, Shaldon SR, Coull P, Dubowitz M, Enomoto T, Ozawa A, Syed SA, Horiuchi T, Khaeraja R, Khan R, Lin SR, Flinter F, Beales P, Hagihara A, Inoko H, Shirakawa T, Hopkin JM. 1999. Chromosome 11q13 and atopic asthma. Clin. Genet. 55(6):431–37 52. Dent AL, Shaffer AL, Yu X, Allman D, Staudt LM. 1997. Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 276:589–92 53. Dent AL, Hu-Li J, Paul WE, Staudt LM. 1998. T helper type 2 inflammatory disease in the absence of interleukin 4 and transcription factor STAT6. Proc. Natl. Acad. Sci. USA 95:13823–28 54. Szabo SJ, Jacobson NG, DigheAS, Gubler U, Murphy KM. 1995. Developmental commitment to the Th2 lineage by

366

55.


56.

57.

58.

59.

60.

ONO extinction of IL-12 signaling. Immunity 2:665–75 Gorham JD, Guler ML, Murphy KM. 1997. Genetic control of interleukin 12 responsiveness: implications for disease pathogenesis. J. Mol. Med. 75:502–11 Hizawa N, Freidhoff LIZ, Chiu YF, Ehrlich E, Luehr CA, Anderson JL, Duffy DL, Dunston GM, Weber JL, Huang SK. 1998. Genetic regulation of Dermatophagoides pteronyssinus-specific IgE responsiveness: a genome-wide multipoint linkage analysis in families recruited through 2 asthmatic sibs. Collaborative Study on the Genetics of Asthma (CSGA). J. Allergy Clin. Immunol. 102:436–42 Wilkinson J, Thomas NS, Morton N, Holgate ST. 1999. Candidate gene and mutational analysis in asthma and atopy. Int. Arch. Allergy Immunol. 118:265–67 Wilkinson J, Grimley S, Collins A, Thomas NS, Holgate ST, Morton N. 1998. Linkage of asthma to markers on chromosome 12 in a sample of 240 families using quantitative phenotype scores. Genomics 53:251–59 Dewar JC, Wheatley AP, Venn A, Morrison JF, Britton J, Hall IP. 1998. Beta2adrenoceptor polymorphisms are in linkage disequilibrium, but are not associated with asthma in an adult population. Clin. Exp. Allergy 28:442–48 Dewar JC, Wilkinson J, Wheatley A, Thomas NS, Doull I, Morton N, Lio P, Harvey JF, Liggett SB, Holgate ST . 1997. The glutamine 27 beta2-adrenoceptor polymorphism is associated with elevated IgE levels in asthmatic families. J. Allergy Clin. Immunol. 100:261–65

61. Ohe M, Munakata M. Hizawa N, Itoh A, Doi I, Yamaguchi E, Homma Y, Kawakami Y. 1995. Beta 2 adrenergic receptor gene restriction fragment length polymorphism and bronchial asthma. Thorax 50:353–59 62. Ismail A, Bousaffara R, Kaziz J, Zili J, el Kamel A, Tahar Sfar M, Remadi S, Chouchane L. 1997. Polymorphism in transporter antigen peptides gene (TAP1) associated with atopy in Tunisians. J. Allergy Clin. Immunol. 99:216–23 63. Noguchi E, Shibasaki M, Arinami T, Takeda K, Kobayashi K, Matsui A, Hamaguchi H. 1998. Evidence for linkage between the development of asthma in childhood and the T-cell receptor beta chain gene in Japanese. Genomics 47:121–24 64. Nickel RG, Saitta FP, Freidhoff LIZ, Yu XY, Ehrlich E, Barnes KC, Beaty T, Huang SK. 1999. Positional candidate gene approach and functional genomics strategy in atopy gene discovery. Int. Arch. Allergy Immunol. 118:282–84 65. Kawashima T, Noguchi E, Arinami T, Kobayashi K, Otsuka F, Hamaguchi H. 1998. No evidence for an association between a variant of the mast cell chymase gene and atopic dermatitis based on case-control and haplotype- relative-risk analyses. Hum. Hered. 48:271–74 66. Mao XQ, Shirakawa T, Yoshikawa T, Yoshikawa K, Kawai M, Sasaki S, Enomoto T, Hashimoto T, Furuyama J, Hopkin JM . 1996. Association between genetic variants of mast-cell chymase and eczema [see comments] [published erratum appears in Lancet 1997 Jan 4;349(9044):64]. Lancet 348:581–83


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


IMMUNOLOGY AT THE MATERNAL-FETAL INTERFACE: Lessons for T Cell Tolerance and Suppression A. L. Mellor and D. H. Munn Program in Molecular Immunology, Institute of Molecular Medicine and Genetics, Departments of Medicine and Pediatrics, Medical College of Georgia, 1120, 15th St., Augusta, Georgia 30912; e-mail: [email protected].


Key Words pregnancy, placenta, macrophages, inflammation, T cells Abstract Mammalian reproduction poses an immunological paradox because fetal alloantigens encoded by genes inherited from the father should provoke responses by maternal T cells leading to fetal loss. Current understanding of T cell immunobiology and the critical role of inflammatory processes during pregnancy is reviewed and discussed. Lessons derived from studies on the regulation of T cell responsiveness during mammalian gestation are considered in the wider context of T cell tolerance toward some microbial infections and tumors, avoidance of autoimmunity, and tissue allograft rejection.

INTRODUCTION—THE BIOLOGY OF PREGNANCY Viviparity is a unique characteristic of mammals. Gestational outcomes avoiding fetal defects or loss, maternal infection, or morbidity are contingent upon an intimate association between mother and developing fetus that nurtures the fetus without provoking maternal immune responses. The process of nurturing new individuals in this way necessitates exquisite integration and coordination of several complex biological processes, including metabolic, endocrine, vascular, and immune functions. Almost certainly, ancestral mammals evolved fundamental mechanism(s) to allow successful viviparity. Rodents and primates possess hemochorial placentas, in which maternal and fetal tissues are not segregated by basement membranes. Broad similarities in the structure, organization, and development of the maternal-fetal interface are shared by many mammalian species, but the detailed anatomy and cell biology of the maternal-fetal interface is surprisingly diverse. Most likely, this reflects the need to accommodate variable fetal masses, numbers of fetus per mother, and different gestational periods as ancestral mammals evolved. Thus, evolutionary embellishments and diversity between mammalian species may have obscured fundamental processes necessary for the 0732–0582/00/0410–0367$14.00

367


368

MELLOR n MUNN

success of viviparity. It is not our intention to review the evolutionary aspects of viviparity per se. However, we point out that possible evolutionary links between viviparity in mammalian vertebrates and processes that allow commensal invertebrate life forms to recognize and regulate cellular interactions are the subject of recent debate (1, 2). In our view, some adaptations were crucial to development of the placental method of nurturing the young. However, we do not hold that one critical process allowed mammals to emerge through an evolutionary bottleneck. Most likely, several processes, which evolved for other biological purposes were exploited and, in time, were adapted to allow mammals to evolve their unique method of reproduction. Not surprisingly, attempts to elucidate biological mechanisms that promote successful outcomes of pregnancy have excited great interest and proved to be a tremendous challenge; they have also generated a large and complex literature. We do not address non-immunological aspects of pregnancy in this review except to illuminate our discussion of selected immunobiological aspects of pregnancy. From an immunological perspective, placental nurturing of the developing fetus poses some perplexing questions, the answers to which will help us to understand other immunological phenomena. Put simply, the questions that exercise most immunological interest are how the mother provides protection from microbial infections without mounting a lethal immune response against fetal tissues expressing paternally inherited alloantigens. In this review we focus on the specific questions of why and how the maternal T cell repertoire tolerates the fetus throughout gestation. This issue has provoked enormous interest, debate, and controversy, which testifies to the importance of the issue as well as to the complexity and ignorance that surround the apparent paradox of maternal tolerance of the fetal allograft. Although this issue has been reviewed extensively, the field is still active and merits further discussion in view of recent developments that have appeared in the literature. Unraveling the secrets that resolve the paradox of fetal tolerance promises to illuminate other important immunological phenomena in which tissues or cells displaying new or foreign antigens are tolerated despite the potential to eliminate them via host immune responses.

THE IMMUNOBIOLOGY OF PREGNANCY Excellent reviews on the immunobiology of pregnancy are available, which address issues such as trophoblast cell biology and anatomy and the role of fetal antigen presentation, inflammation, uterine macrophages and NK cells, cytokine production, immunoregulation, tolerance and the innate immune system during pregnancy (3–9). The reader is referred to these reviews for detailed discussion of these and other aspects of pregnancy. While our goal is to review new developments relating to the effect of pregnancy on maternal T cell immunobiology we will refer to the excellent summations of the field to be found in these reviews to illustrate the background to our discussion. Space and time limit our ability to


T CELL TOLERANCE DURING PREGNANCY

369

review this large and developing field. While we have tried to do justice to the field as it now stands, inevitably we will have omitted some elements that merit attention. Undoubtedly, these issues will surface in due course if they shed new light on the complex and perplexing relationship between the maternal immune system and the developing fetus. As declared above, our goal is to focus on a specific issue: regulation of maternal T cell responsiveness during gestation. To this end, we address the impact of pregnancy on maternal T cell responsiveness and, in particular, examine the critical role of inflammation and the innate immune system in this unique immunological situation. While it is necessary to emphasize results obtained from experimental animal (mostly murine) models of pregnancy, we are aware of the potential pitfalls of this selective approach. Where possible we refer to evidence that lessons learned from animal models are relevant to human pregnancy, and we identify controversies that arise from potential distinctions between mice and humans. While our selective approach ignores other important aspects of reproductive immunology, we justify this on the grounds that T cells are critical mediators of immunity that lead to clearance of microbial infections, autoimmune diseases, tissue transplant rejection, and, in some experimental systems, to tumor regression. Lessons learned from studying how maternal T cell responsiveness is controlled during pregnancy are likely to illuminate the role of T cell immunoregulation in these important immunological phenomena. Throughout this review we consider the current status of the paradox of the fetal allograft, which has been the dominant hypothesis driving research and debate in the reproductive immunology field for the last four decades (10). However, we also synthesize older with more contemporary ideas to elaborate a coherent rationale that explains fetal survival and from which potential lessons for understanding other immunological phenomena involving regulation of T cell responsiveness also emerge.

THE FETUS AS A TISSUE ALLOGRAFT Genetic Polymorphism and Tissue Alloantigens From a genetic perspective mother and fetus are never identical in outbred populations because the fetus inherits a different set of polymorphic genes from each parent. Only in exceptional circumstances, such as experimental matings between pure inbred animals, are mother and female fetus genetically identical (syngeneic). Strictly, male offspring of syngeneic matings and their mothers are not completely identical due to Y-chromosome genes encoding tissue antigens such as the male-specific transplantation antigen H-Y (11). In most mating combinations, multiple tissue antigens differ between fetus and mother and are potential tissue alloantigens recognizable by T cells. Many genetic polymorphisms are immunologically inert because they do not change protein-coding sequences. Not all polymorphisms that change protein-coding sequences are detected by the


370

MELLOR n MUNN

immune system because structural polymorphisms must change peptides bound to major histocompatibility complex (MHC) molecules to generate tissue alloantigens recognizable by T cells. By recognizing peptides bound to MHC molecules, T cells can detect extremely low amounts of peptides generated by intracellular proteolytic degradation of proteins (12). Polymorphisms that alter the rate of gene transcription can have immunological effects if the density of peptide/MHC complexes is altered as a consequence. In mice, a large number of tissue alloantigens recognized by T cells, called minor histocompatibility (miH) antigens, have been characterized structurally (11–13). While these are likely to be important fetal alloantigens, polymorphic MHC genes pose the most significant immunological barrier to fetal survival because maternal and paternal MHC mismatches are very frequent in outbred populations and they excite potent T cell responses in tissue graft recipients. Recent reviews of the role of MHC and miH antigens in tissue allograft rejections provide detailed assessments of the contribution of antigenic differences, antigen processing pathways, and T cell subsets to graft rejection (14, 15).

Medawar’s Paradox and Potential Solutions According to the laws of tissue transplantation fetal alloantigens encoded by polymorphic genes inherited from the father ought to provoke maternal immune responses leading to fetal rejection soon after blastocyst implantation in the uterine wall (10, 16). No other tissue, when surgically transplanted between genetically different individuals, enjoys the impunity from lethal host (maternal) immune responses that characterize the maternal-fetal relationship. This unique relationship resembles a parasitic condition, albeit a temporary one, in which the fetus is nurtured and given immunological protection from microbial infections and potentially lethal maternal immune responses. Originally, Medawar proposed three broad hypotheses to explain the paradox of maternal immunological tolerance to her fetus: (a) physical separation of mother and fetus; (b) antigenic immaturity of the fetus; and (c) immunologic inertness of the mother (10). The ‘Holy Grail’ of reproductive immunology has been to elucidate the fundamental processes that explain fetal survival in all mammalian species. Consequently, investigators have sought experimental evidence arguing definitively for or against each of Medawar’s three hypotheses. The field has been reviewed comprehensively in the context of these three guiding hypotheses. The key issue of whether a single mechanism explains Medawar’s paradox has not been answered decisively, and it seems likely that no single mechanism completely resolves the paradox. An overview of the current literature reveals no broad agreement on which fundamental processes resolve Medawar’s paradox for all mammalian species. Nevertheless, compared to surgically transplanted tissue allografts, ample evidence suggests that the fetus enjoys special privileges and employs specific procedures to evade or minimize the risk of being rejected by the maternal immune system during gestation.



371

During the discussion that follows we use the paradigm of the fetus as a tissue allograft tolerated by the mother, contrasting the immunological situation of a fetus with that of a tissue allograft surgically transplanted onto a nontolerant (allogeneic) recipient. Although some commentators now question the value of this comparison (1, 5), it is a useful starting point to consider the immunogenetics and immunobiology of the maternal-fetal relationship. To mimic the immunogenetics of pregnancy, the hypothetical donor graft should express two haploid genotypes (paternal x maternal, F1) and the recipient should match one donor haplotype (the maternal genotype). To aid our discussion, we provide a diagram comparing processes in the afferent (antigen presentation leading to T cell activation) and efferent (T cell differentiation, migration, and effector functions) phases of a T cell response that lead to T cell–mediated rejection of tissue allografts (Figure 1). We consider whether these processes occur or are moderated during gestation to explain fetal allograft survival. This diagram is based on current knowledge of processes that provoke T cell activation leading to tissue allograft destruction and rejection after transplantation (14, 15).

Fetal Loss Syndromes and Immune Dysregulation Spontaneous Fetal Loss in Mice After genetic considerations, the most important reason for viewing the fetus as a tissue allograft is experimental evidence that maternal immune responses can cause spontaneous fetal loss in some mating combinations. However, it has long been established that provoking a systemic maternal T cell response against paternal MHC alloanantigens during murine pregnancy does not induce fetal loss. These results were interpreted as evidence that the fetus was impervious to attack, even when the afferent arm of the T cell response was bypassed experimentally, and that local mechanisms provided protection from effector T cells in such circumstances. Even though the fetus is a de facto tissue allograft, spontaneous fetal loss due to maternal immunity is considered to be a rare event. Circumstantial evidence suggests that some instances of spontaneous fetal loss are due to immunological complications during pregnancy. For many years, the most compelling animal model of immunologically mediated spontaneous fetal loss has been the murine [CBA/J x DBA/2] mating combination in which between 20% and 50% of fetus is resorbed by gestational day 13 (17). The critical importance of this precise parental genetic combination, and the ability to suppress fetal loss by preimmunizing females with peritoneal macrophages or by prior mating to a BALB/ cJ male, point to an immunological component that contributes to this fetal loss syndrome. NK cells and macrophages have been implicated as cellular mediators of the syndrome and excessive nitric oxide (NO) and TNF-a release by decidual mononuclear cells as effector mechanisms that become dysregulated in this strain combination (18, 19). An indirect role for GM-CSF and, significantly, maternal CD8` T cells in preventing NK cell-mediated fetal loss in this system has also been suggested (20). Recent studies reinforce the connection between TNF-a


372 Figure 1 T cell immunobiology of tissue allograft rejection compared to the situation of the fetus. Tissue alloantigens (Ag) are delivered to draining lymph nodes (LN) by donor or host APCs (step 1) where they encounter naı¨ve T cells (Tn). This leads to T cell activation and differentiation (step 2) into cytotoxic (Tc) and helper (Th1 and Th2) T cells. Effector T cells (Tc and Th1) recirculate to donor graft tissues or help B cells to produce antibodies (Ig), which contribute to the destruction of cells of donor origin (step 3). Evidence that steps 1–3 are subject to regulation during the maternal-fetal relationship are considered in detail in the text. Two additional steps are also considered; step 4, Th2 dependent suppression of Th1 and Tc effector T cells, and step 5, direct access of Tn to fetal tissues avoiding the need to circulate through LN.



373

release by decidual macrophages, which precedes fetal loss (21). These findings send clear messages that cytokine imbalances associated with inappropriate activation of macrophages and NK cells of the innate immune system are detrimental to fetal survival in this experimental model. Indeed, a recent review proposes to categorize cytokines and growth/differentiation factors as either deleterious (TNFa, IFN-c, IL-2) or beneficial (TGF-b, LIF, CSF-1, GM-CSF, IL-1, IL-3, IL-4, IL6, IL-10 and IFN-s) to fetal survival, implying a link between regulated activation of cells that produce the second set of factors and fetal survival in normal pregnancies (22). As yet, it is unclear how dysregulated activation of NK cells and decidual macrophages relate to the distinctions made between fetal and tissue allografts by maternal T cells. Moreover, how relevant these observations are to successful pregnancy remains to be seen, originating as they do from a single murine mating combination. Indeed, IFN-c (23) and nitric oxide (NO) (24) production by maternal monocytes and uterine NK cells (25) that invade the decidua shortly after implantation are features of normal pregnancies in mice. One problem is that it is not obvious whether cytokine imbalances are caused by inappropriate T cell responses or vice versa. Chaouat et al showed that spontaneous fetal loss in the CBAxDBA/2 model was corrected by administering IL-10 or the unusual interferon variant IFN-s, which is expressed in ruminant placentas (26). Collectively, these results point to a critical role for the local inflammatory milieu at the maternal-fetal interface, which might shape the context in which maternal T cells encounter fetal alloantigens. We discuss this issue in more detail later. Connections between the activation status of the innate immune system and human pregnancy outcomes have been reviewed recently (9). An increasing number of reports document spontaneous fetal loss syndromes in matings involving gene-deficient (knockout) mice. In some cases, these syndromes arise from defective placental vascularization or fetal development rather than immune dysfunction and are not discussed here. Mice deficient in the production of myeloid growth/differentiation factors, granulocyte-macrophage colony stimulatory factor (GM-CSF) and macrophage colony stimulatory factor (CSF-1 or M-CSF), both exhibit poor reproductive performance. In GM-CSF gene-deficient mice, the effects manifest as poor placental development, which increases the rate of fetal loss and compromises fetal growth (27). CSF-1 gene– deficient mice exhibit low pregnancy rates and small litter sizes that might indicate immune dysfunction leading to fetal rejection. However, the major effect of CSF1 deficiency in spontaneously mutant osteopetrotic (op/op) mice is on the frequency and rate of ovulation (28). This highlights a major difficulty in using gene-deficient mice to assess requirements for successful pregnancy since total loss of a cytokine or growth factor may result in multiple effects, with cumulative impacts on reproductive performance, without necessarily compromising immunological protection of the fetus. Another technical problem is that many investigators assess reproductive performance in syngeneic or at least MHC-matched mating combinations because nearly all gene-deficient (knockout) mice are generated on the 129/Sj (H-2b haplotype) background, which are then intercrossed

374

MELLOR n MUNN


with H-2b haplotype C57BL/6 mice to establish the lineage. Given the importance of assessing the potential immunological stress due to parental genetic disparity, it is imperative to examine the effects of gene deficiencies in allogeneic as well as syngeneic pregnancies. Linked to this, it may also be necessary to determine whether the effects of gene deficiencies manifest only when inherited as maternal and/or paternal traits rather than as traits inherited from both parents. Induced Fetal Loss and Tryptophan Metabolism Increased rates of fetal loss are induced by injecting pregnant mice with IL-2, which promotes Th1-type T helper cell responses at the expense of Th2-type responses (29). This reinforces the view that a delicate immunological equilibrium is established during gestation that can be subverted by inappropriate activation of monocytes or lymphocytes of the innate or adaptive immune systems (30). However, definitive proof that maternal T cells specific for paternally inherited alloantigens could be induced to participate in immune responses leading to fetal rejection was not obtained until recently. Experimental evidence supporting this view was obtained from studies in which pregnant mice carrying syngeneic or allogeneic fetus were treated with a pharmacologic inhibitor of an enzyme called indoleamine 2,3 dioxygenase (IDO) (31), which is expressed by cells in the maternal decidua (32) and which catabolizes tryptophan (33). This treatment induced uniform loss of allogeneic fetuses, which was complete by gestational day 9.5 when pregnant mice were exposed to IDO inhibitor at gestational day 4.5, the time of blastocyst implantation. Further, the same treatment had no effect on development to term of syngeneic fetus or allogeneic fetus carried by immunodeficient RAG-1 gene-deficient mothers, which have no lymphocytes. The rationale for these experiments came from studies on immunosuppressive human macrophages that prevent T cell activation in vitro by depriving T cells of tryptophan (34). These findings demonstrate that allogeneic fetus are potentially vulnerable to maternal T cell–dependent processes that could provoke fetal loss, and cells expressing IDO and degrading tryptophan provide protection from maternal T cells. Interestingly, golden hamsters placed on high tryptophan diets exhibited high rates of fetal loss in an earlier study (35). This study was designed to test whether increased synthesis of serotonin analogues derived metabolically from dietary tryptophan would induce rapid fetal loss due to non-immunological, pharmacologic effects on the placenta. However, the outcome obtained can also be interpreted in terms of a link between tryptophan metabolism and immunological protection of allogeneic fetus during gestation. More studies are needed to determine whether these findings can be generalized to other mouse mating combinations and to other mammalian species, including humans. However, the observation that serum tryptophan levels decrease progressively from the first trimester of human pregnancy provides circumstantial evidence that this link may be relevant to human pregnancy (36). Recurrent Spontaneous Abortion in Humans Spontaneous human fetal loss is a significant clinical problem. Some commentators estimate that early fetal loss



375

after implantation may occur in as many as 30% of human pregnancies since many may not be diagnosed (6). Moreover, approximately 60–70% of spontaneous abortions in women are unexplained by fetal defects or infections. Studies on recurrent spontaneous abortion syndromes are dominated by suggestions of immunologic causation or at least observations that can be interpreted as such. This evidence includes genetic (epidemiological) analyses, anatomical, physiological, and cell biological analyses, and evidence for cytokine dysregulation linked to inappropriate activation of the innate and adaptive immune systems during human pregnancy. Wegmann and colleagues suggested that Th2-type cytokines and T helper cell responses correlated with successful pregnancy outcomes (37, 38). Recent reports provide additional support for this hypothesis (6, 39). Although beyond the scope of the present review, these issues are discussed in depth by others (6, 30, 40, 41). Summarizing the current state of the field, there are compelling reasons for assuming that some, perhaps many, of human pregnancy failures are causedby immune dysfunction that upsets the delicate balance between maternal tolerance and fetal nurturing. However, it is very difficult to discriminate whether abnormalities observed during difficult pregnancies are causes or effects of immune dysfunction. Given the largely and necessarily descriptive nature of the evidence available from studies on human pregnancy, it is difficult to assess definitively whether maternal T cells are participants, facilitators, or merely bystanders in processes that lead to postimplantation loss of human fetuses . Immunological dysfunction has also been suggested as a potential cause of the hypertensive disorder pre-eclampsia, which manifests late in human pregnancy (42–44). However, maternal age and parity, male factors (45), and defective endothelial and vascular tissue development (46) have all been considered as factors that predispose to pre-eclamptic pregnancy, although it is difficult to know which factors predispose toward the syndrome and which are effects.A recent radical suggestion is that pre-eclampsia is far more frequent when partners cohabit for only a brief period before conception compared with partners who have a long history of cohabitation (44, 47). While this observation needs to be confirmed, it raises intriguing questions about mechanisms that would decrease the likelihood of problematic pregnancy during a lengthy period of cohabitation. Although speculative, one answer might be that exposure to sperm may precondition the female, perhaps even her immune system, if genetic material from sperm induces immunosuppression or tolerance over time.

THE IMMUNOBIOLOGY OF FETAL SURVIVAL AND GRAFT REJECTION In this section we compare and contrast fetal implantation and growth in the uterus during gestation with the immunobiology of T cell–mediated allograft rejection (Figure 1). Using this approach we address Medawar’s paradox of the tissue

376

MELLOR n MUNN

allograft and describe recent experimental data that relate to regulation of maternal T cell responsiveness during pregnancy. Our discussion focuses on studies of murine pregnancy, which are amenable to genetic manipulation and intense immunological scrutiny. We also discuss how immunological knowledge gained from studying pregnancy might apply more generally to understanding T cell regulation in other immunological phenomena.


T Cell–Mediated Tissue Allograft Rejection—A Brief Synopsis Tissue allografts transplanted onto recipients that are not T cell deficient or immunosuppressed are rejected, after a delay of several days to a few weeks (Figure 1). During this period, donor alloantigens are delivered by specialized antigen presenting cells (APCs) of myeloid origin to draining lymph nodes (step 1, afferent phase). There, APCs encounter and activate naı¨ve T cells that recognize donor alloantigens, which proliferate and differentiate into helper and cytotoxic effector T cells (step 2). Effector T cells recirculate to the grafted donor tissue where they kill cells displaying donor alloantigens, which eventually destroys the graft (step 3, efferent phase). T cells also coordinate B cell activation and production of complement-fixing antibodies that bind to donor cells. Recent reviews are available for readers wishing to know more about the details of these processes (15, 48–50). Interactions between the adaptive and innate immune systems at the cellular and molecular level regulate the T cell response to tissue allografts. In particular, myeloid cells play critical roles in regulating each stage of this entire process. Dendritic cells (DCs) are specialized to accumulate antigens at the site of tissue engraftment, to migrate to draining lymph nodes, and to present antigens under conditions that provoke T cell activation. Although immunologists have largely focussed on elucidating mechanisms that generate effective T cell responses, there has also been interest in addressing how T cell responses are regulated to prevent loss of control leading to autoimmunity, which would inflict collateral damage on healthy (host) cells and tissues. The concept of immunosuppression grew out of the realization that there must be mechanisms that limit immune responses spatially and temporally to minimize the risk of autoimmunity. We discuss the critical role of inflammation and immunoregulatory processes during pregnancy later. However, the unqualified term inflammation is, in our opinion, inadequate to describe the complex processes that occur in tissue microenvironments undergoing immune attack. We prefer the concept of ‘‘aggressive’’ and ‘‘suppressive’’ inflammatory processes, which co-exist temporally and even spacially, and counterbalance one another. We envisage that the outcomes of immune responses targeted to particular tissue microenvironments are critically dependent on the precise nature of the cellular processes, molecular cues, and biochemical status of microenvironments as perceived by naı¨ve or effector T cells. By these criteria it is more important to know whether the collective effect of all these processes


377

tends to promote or suppress T cell responses in vivo than to measure individual components that may be overruled by other processes in the complex milieu of inflamed tissues. Thus, the overall balance of this milieu dictates the outcome and direction of T cell responses.


Preparing For Engraftment—Pre-Implantation Processes Unlike recipients of surgically transplanted tissue allografts, females experience many physiological changes before blastocyst implantation takes place. Three events, which are necessary preludes to implantation and pregnancy, trigger these changes: ovulation, copulation, and fertilization. These events directly or indirectly induce dramatic changes in metabolism and the hormone and cytokine balance in the reproductive tract, which is thereby prepared for subsequent blastocyst implantation. Many changes in uterine physiology that occur following these events resemble classical inflammation at the mucosal surfaces of the female reproductive tract, and it is quite likely that these changes impact locally on the maternal immune system well before the blastocyst implants in the uterus. Consequently, the outcome of interactions between T cells and myeloid cells are likely to differ markedly during pregnancy, compared to outcomes in nonpregnant hosts. These distinctions are likely to be most acute at mucosal surfaces lining the female reproductive tract and in lymph nodes draining the maternal-fetal unit. Thus, the uterus may be preconditioned to accept the blastocyst (the incoming ‘‘tissue allograft’’).This situation contrasts with that of a surgically inserted tissue allograft because there is no prior conditioning of the graft bed or draining lymph nodes in graft recipients. Hormonal changes precede implantation and persist throughout gestation. Steroid hormones induced by ovulation are potent modulators of myeloid APC and lymphocyte functions and may influence the nature and potency of immune responses during pregnancy (51, 52). For example, progesterone, the hallmark of pregnancy, suppresses T cell effector functions by modulating potassium channels and calcium signaling, which affects gene expression (53). Male factors, the consequence of exposure to sperm or semen, may also trigger events that have beneficial effects favoring successful pregnancy outcomes (45). Sperm or components of semen could elicit physiological changes, for example via prostaglandins, that promote immunosuppression at the mucosal surfaces of the female reproductive tract (5, 54). In addition, Tremellen et al demonstrated that transforming growth factor (TGF-b1), a component of seminal fluid, stimulates GM-CSF production and recruitment of inflammatory cell infiltrates into the uterus (55).

Nurture Without Immunity—Postimplantation Fetal Development After implantation, the key immunological questions are whether maternal T cells are aware of fetal alloantigens that would provoke immune responses if presented on a paternal tissue graft, and, if they are, why does this not lead to fetal loss

378

MELLOR n MUNN


during gestation? The frequency of fetus-specific maternal T cells is a significant fraction (;10–30%) of the peripheral T cell repertoire in most parental combinations in outbred populations because of MHC mismatch. This explains why tissue allografts provoke potent anti-graft T cell responses and why maternal acceptance of the same alloantigenic differences is such an intriguing issue to immunologists. Maternal T Cell Awareness of Fetal Alloantigens During allogeneic pregnancy the maternal immune system contains naı¨ve T cells that are not tolerized to fetal alloantigens and yet do not attack the fetus. An explanation for maternal acceptance of the fetus during gestation is that either maternal T cells are not exposed to fetal alloantigens or exposure occurs but results in T cell tolerance rather than immunity. Several studies involving female T cell receptor (TCR) transgenic mice mated with allogeneic males provide strong evidence that maternal T cells are aware of fetal alloantigens during pregnancy. Using this approach, Tafuri et al demonstrated that maternal H-2Kb-specific CD8` T cells were functionally tolerized by fetal H-2Kb alloantigen, but that this state lasted only until shortly after parturition (56). Using another H-2Kb-specific system, we have reported similar phenotypic changes during allogeneic pregnancy (57). In a third study, Jiang & Vacchio concluded that maternal male-(H-Y)-antigen-specific CD8` T cells were tolerized by Fas-dependent deletion mediated by fetal trophoblast cells, and by nondeletional processes that rendered residual maternal T cells unresponsive to antigenic challenge in vitro (58). The conclusion that maternal T cells are aware of fetal alloantigens, at least in murine pregnancy, has several important implications that relate to Medawar’s three postulates to explain fetal survival despite potential maternal immunity. First, no cell-impermeable barrier between mother and fetus prevents exposure of fetal alloantigens to maternal T cells throughout gestation. Second, antigenic immaturity of the fetus does not explain why maternal T cells fail to mount a response to the fetus. And third, fetal survival depends on tolerogenic mechanisms that block maternal T cell responses. For obvious practical and ethical reasons it is not possible to conduct similar experiments in humans; however the recently developed ability to focus on fetal-specific maternal T cells in peripheral blood using highly specific MHC/peptide tetramer reagents may aid experimental analyses of the effect of pregnancy on these cells (59). Fetal Cell Traffic into Maternal Circulation The highly organized trophoblast, endothelial and mesenchymal cells that form the outer trophoblast, segregate fetal and maternal blood circulatory systems (5, 60). As well as physical barriers, molecular (24) and biochemical (31) barriers may also prevent the passage or functioning of maternal T cells that attempt to access fetal tissues. However,there is general agreement that the barrier to cellular and subcellular traffic between mother and fetus is not completely impermeable during human or rodent pregnancy, and the maternal or host immune system is alerted to the presence of


379


foreign alloantigens during pregnancy and after tissue engraftment (5, 60, 61). Thus, while the anatomical organization of the maternal-fetal interface may have evolved at least in part to minimize immunological contacts, this is an imperfect strategy necessitating additional mechanisms to ensure fetal survival during gestation. The experiments cited above demonstrating maternal T cell awareness of unprocessed fetal H-2Kb alloantigen (56, 57) or male H-Y peptide bound to unprocessed H-2Db (58) imply that fetal cells expressing MHC alloantigens alert maternal T cells to the presence of fetal MHC alloantigens. (A less likely explanation which cannot be entirely ruled out is that MHC alloantigens shed from fetal cells attach to maternal APCs.) Thus, the conclusion from these studies is that either maternal T cells encounter fetal cells as they circulate through the maternal-fetal unit (Figure 1, step 5) or fetal cells migrate to local draining lymph nodes where they present fetal alloantigens to maternal T cells (step 2).

Antigenic Immaturity of the Fetus As pointed out by Medawar over four decades ago, a simple explanation for lack of maternal immune responses during pregnancy is a lack of MHC expression on fetal tissues during gestation (10). This assumes that the maternal-fetal interface is the only point of contact between maternal lymphocytes and fetal alloantigens. Murine trophoblast cells express MHC class I genes and alloantigens at high levels from early times in gestation when MHC class I expression is barely detectable on fetal tissues (62, 63). As pointed out by other commentators, it is highly unlikely that maternal T cells circulating through the murine maternal-fetal interface do not encounter cells of fetal origin (60). Thus, this pattern of MHC class I expression at the maternalfetal interface is incompatible with the hypothesis that fetal tissues at the interface are antigenically immature to minimize the risk of provoking maternal T cell responses. Nevertheless, it is important to stress that there is general agreement that MHC class II alloantigens are not expressed by cells of fetal origin at the maternal-fetal interface in humans or rodents (64–66). This removes a critical factor that elicits potent CD4` T cell responses, which are significant components of effector and helper T cell responses that target tissue allografts leading to rejection (Figure 1, step 3) (15). In this respect, the argument that antigenic immaturity contributes to fetal survival has merit. It remains to be seen whether fetal loss ensues if allogeneic MHC class II expression is forced on fetal trophoblast cells, for example in transgenic mice. Several studies report attempts to express MHC class II molecules in trophoblast cells following DNA transfection of cultured trophoblast cell lines (67) or to force MHC class II expression in murine trophoblast (68– 70). It is not clear whether these attempts succeeded in transgenic mice because it was not reported whether surface MHC class II expression was observed in trophoblast cells in mice carrying a transgene with a strong CMV promoter. However, under normal circumstances MHC class II expression on murine trophoblast is strongly repressed at the transcriptional level (71).

380

MELLOR n MUNN


The observation that human trophoblast cells express the nonclassical and relatively nonpolymorphic MHC class I molecules HLA-G and, possibly, HLAE has inspired much research and speculation about their immunological significance, particularly with respect to their possible effect on maternal T cell responses (66, 72). Recently, debate has focussed on their possible roles in regulating NK cell activity, rather than T cell responses (73, 74). This debate is difficult to evaluate experimentally, largely because analogous MHC-like molecules and their receptors have not yet been identified in rodents. This last point suggests either that immunological roles for HLA-G/E may have evolved after rodent and primate ancestors diverged, or that rodent and human NK cell receptors are no longer homologous. Perhaps these possible roles, if they can be demonstrated experimentally, are examples of processes that evolved due to the need to protect more fetal mass for longer periods in larger mammalian species. Fetal Microchimerism In the field of transplantation research, energetic efforts are currently directed at the therapeutic potential of induced microchimerism for prolonging survival of tissue grafts (75, 76). In pregnancy, maternal T cell tolerance must be induced and maintained either by fetal trophoblast cells or by fetal cells that enter the maternal circulation and establish a reservoir of fetal cells in maternal tissues that is either stable or continuously replenished. At present, it is not clear which route is most critical. However, recent studies have generated compelling evidence that fetal microchimerism affects the maternal immune system. Studies on pregnant women show that fetal cells appear in maternal circulation at an early stage in gestation and that genetic microchimerism persists for many years after parturition (77–80). An earlier study on pregnant mice by Bonney & Matzinger demonstrated that male fetal cells containing Y-chromosomal DNA access maternal circulation in about 20% of immunocompetent mice and in a higher proportion (;40%) of immunocompromised mice (60). These observations prompted the authors to conclude that fetal microchimerism could not explain maternal T cell tolerance and that maternal T cell immunity eradicated all traces of fetal cells in maternal circulation in most cases. However, this conclusion would seem incompatible with data showing that maternal splenic T cells not only are aware of fetal alloantigens but are tolerized to them in all pregnant mice (56–58). These results can be reconciled by assuming that fetal cells migrating into maternal tissues excite a T cell response that results in T cell tolerance and destruction of the fetal cells.This scenario might also help to explain why maternal tolerance disappears shortly after parturition, in the study by Tafuri et al, since a continuous supply of fetal cells may be necessary to maintain T cell tolerance (56); T cell tolerance did, however, persist after parturition, as reported in the recent study by Jiang & Vacchio (58). More speculatively, it has also been suggested that circulating fetal cells may help induce or maintain thymic involution that persists during human and animal pregnancy, and that this may contribute to maternal T cell tolerance of the fetus (81, 82) (although it is difficult to eliminate neuroendocrine effects and nonspe-


381


cific effects of steroid hormones released during pregnancy as factors contributing to thymic involution). Nevertheless, modification of thymic output through microchimerism is an active field of current research as an potential therapy to improve the success rate of tissue allografts (83). At present, however, there is no consensus on whether thymic microchimerism is beneficial or detrimental to the outcome of tissue transplantation (84). Thus, there is no compelling evidence that modifications to maternal thymus output during pregnancy makes significant contributions to maternal T cell tolerance of the fetus. Maternal T Cell Tolerance The mechanisms that induce maternal T cell tolerance are not understood in detail, although there is some progress from recent studies on pregnant TCR transgenic mice, as reviewed above. Tolerance induction is likely to be linked to the unique cocktail of hormones and cytokines that are elaborated before and after blastocyst implantation. The nature of inflammatory changes that occur during pregnancy and their potential impact on maternal T cell responsiveness are discussed in detail in the next section. These, or other factors, might modify the afferent (Figure 1, steps 1, 2) or effector (steps 3, 4) arms of the immune response elaborated against tissue allografts. As alluded to above, pregnancy correlates with enhanced Th2-type responses (step 4), which may help to ameliorate potentially lethal Th1 and cytotoxic (Tc) T cell responses (6, 38, 85). Thus, encounters between APCs displaying fetal alloantigens may induce T cell deletion or anergy or enhance CD4` Th2-type responses. Jiang & Vacchio recently reported that maternal H-Y-specific CD8` T cells were either deleted, via Fas-dependent processes, or anergized during pregnancy (58). However, this result is inconsistent with reports going back many years that pregnancy does not suppress systemic T cell immunity to paternal tissue allografts or cells expressing paternal MHC alloantigens. Two potential explanations for this discrepancy are the weak nature of the single fetal alloantigen under scrutiny (11) or the absence of T cells that do not express the H-Y-specific TCR clonotype on CD8` T cells (58). It should also be emphasized that tumor cells used by Tafuri et al to detect maternal T cell tolerance in their experimental system (56) are less immunogenic than tissue allografts and therefore much easier to tolerate. To reconcile earlier experiments showing lack of systemic T cell tolerance with those of Tafuri et al, it is necessary to postulate that T cell tolerance can be reversed in vivo without compromising fetal development. This hints at a role for other protective mechanisms operating locally at the maternal-fetal interface that can prevent activated effector T cells from attacking fetal cells (Figure 1, step 3). Antigenic immaturity of fetal trophoblast cells or local immunosuppression at the maternal-fetal interface could contribute to continuing survival of the fetus under such circumstances. As discussed in the next section, a final tolerogenic phenomenon that might contribute to the onset of transient maternal T cell tolerance during gestation is linked suppression by maternal T cells that become anergized in early encounters with cells displaying fetal alloantigens. This is analogous to experimentally induced long-term T cell tolerance (also termed infectious toler-

382

MELLOR n MUNN

ance) induced in vivo when naı¨ve T cells encounter APCs in the presence of nondeleting monoclonal antibodies that block T cell activation (86).

THE ROLE AND NATURE OF INFLAMMATION DURING PREGNANCY


The Role of Inflammation in Adaptive Immune Responses The immune system can be conceptually divided into the adaptive system—T and B cells, which respond to specific antigens—and the innate system, which responds to a wide range of signals of infection and injury in an antigenindependent fashion. The adaptive system confers exquisite specificity and memory, while the innate system provides most of the actual inflammatory and effector functions needed for the organism to fight infection. A decade ago, Janeway pointed out that although the adaptive immune system is a marvel of flexibility and precision, immunization with foreign antigen nonetheless requires a crude, proinflammatory adjuvant in order to elicit an adaptive response. Janeway thus proposed that costimulatory signals from the innate immune system, indicating that infection was present, were necessary in order to give ‘‘permission’’ to the adaptive immune system to activate. By extension, tissues that were not infected or inflamed would not support lymphocyte activation and would thus operationally be considered ‘‘self’’ regardless of the antigens they displayed (87). Janeway’s hypothesis has been extended to include any signals that indicate microenvironmental tissue injury, not just infection (88), and Matzinger has proposed that any condition that the immune system recognizes as connoting danger to the host could be permissive for lymphocyte activation (89). Fearon has pointed out that, far from being a primitive first line of defense, the innate immune system actually acquires critically important information regarding the nature of the threat to the host via its array of specialized receptors to detect microbial products, tissue damage, complement activation, clotting factors, etc (90). Thus, it makes sense that the highly specialized recognition and memory functions of the adaptive system should be regulated and directed by the older and wiser innate immune system (91). A growing body of evidence now supports the view that the adaptive immune response is regulated by the innate system. While the perspectives cited above differ in their emphasis, they have in common the concept that lymphocyte responses are not simply driven by encounter with antigen, but rather are strongly influenced by the context in which this encounter occurs. As a general principle in these models, inflammation favors activation. However, in the case of pregnancy this presents a difficulty, since the decidua would appear to be a highly reactive tissue, both at implantation and throughout gestation, with many signs of inflammation (reviewed in 3). Why does this not favor activation of alloreactive maternal T cells and rejection of the fetus?


383


Not All Inflammation Is the Same One clue may come from the particular nature of the inflammatory response provoked by the fetus. It has been shown using models of infectious disease that the nature of the initial inflammatory response strongly influences the subsequent adaptive response. Infection of susceptible mouse strains with the parasite Leishmania major induces inflammation and a strong immune reaction, but the result is almost entirely a Th2 and humoral response, which is not protective against this intracellular pathogen. In contrast, when the initial inflammatory milieu is altered by infection or immunization with Leishmania in the presence of IL-12, the result is primarily a Th1 response, which is protective (92, 93). Thus, the nature of the initial innate response is critical in determining the nature of the subsequent adaptive response. In pregnancy, the maternal-fetal interface is an immunologically active site, with the production of many immunoregulatory cytokines. Several recent reviews have drawn together the considerable evidence that this cytokine milieu predominantly favors development of Th2 responses. As discussed above, in a number of settings, Th2 cytokines have tended to promote successful pregnancy, whereas Th1 cytokines incline toward fetal loss (6, 38). While this model has been somewhat more clean cut in mice than in humans, the available evidence supports at least a circumstantial link in humans as well (85). Moreover, the concept that different types of inflammation lead to different T cell responses need not be confined to the Th1/Th2 paradigm. In the immune system, there are numerous examples of one inflammatory stimulus triggering a counterregulatory response, which limits the damage done by the initial agent– e.g., pretreatment with TNF protects against subsequent lethal challenge with TNF or IL-1 (94), and pretreatment with IL-12 protects against rechallenge with IL-12 (95). In this regard, IFNc is particularly interesting. Data from IFNc and IFNc-receptor knockout animals reveal that, while they are defective in their ability to clear microbial pathogens, as would be expected from the loss of a proinflammatory cytokine, they are also unexpectedly resistant to tolerance induction (96–98) and susceptible to a variety of autoimmune disorders (99–101). This suggests that IFNc is also important in recruiting counterregulatory systems that limit inappropriate or excessive T cell activation. Conceptually, such opposing regulatory systems are often thought of as ‘‘proinflammatory’’ and ‘‘anti-inflammatory’’ responses. In reality, however, from the point of view of the innate immune system, both reflect proactive recruitment of effector cells and cytokine networks, and hence both can be legitimately considered forms of inflammation.It is only from the point of view of the T cell that certain types of inflammation appear immunosuppressive. We propose that in pregnancy the innate immune system is activated, local and systemic inflammation occurs, and the adaptive immune system is fully aware of paternal alloantigens, but—unlike the case of an organ allograft—the nature of the inflammatory

384

MELLOR n MUNN

response is such that this awareness of antigenic difference is not allowed to lead to fetal rejection.


What Is Different About the Inflammatory Milieu in Pregnancy? As outlined in Figure 1, there are several potential points at which the mammalian fetus could differ from a simple paternal tissue allograft. The first (labeled Step 1 in the Figure) is in antigen presentation. Although studies in mice indicate that the maternal immune system becomes aware of fetal antigens during gestation, it is not clear by what route these antigens are presented. As discussed above, fetal tissues such as trophoblasts suppress expression of MHC class II and have reduced or absent MHC class I. Those alloantigens, which are expressed on trophoblast, may be presented in an incomplete or immunosuppressive fashion (e.g., without costimulatory signals). Likewise, local maternal APCs such as decidual macrophages may be ineffective or tolerogenic presenters of fetal antigens. However, since there is strong evidence that the maternal immune system also encounters fetal antigens outside of the specialized setting of the placenta (see above), additional mechanisms must exist to prevent maternal sensitization and fetal rejection. We have alluded to the apparent Th2 bias in pregnancy which may help protect against rejection. That bias might be introduced locally at the maternal-fetal interface, e.g., by factors such as IL-4, progesterone, or PGE2 (37, 52, 102). Alternatively, the bias may be introduced regionally in the lymph nodes draining the uterus. Recent studies suggest that lymph nodes which drain sites of mucosal tolerance are instrumental in tolerance induction and are functionally different from lymph nodes draining nontolerogenic sites (103). The mechanism by which this difference is introduced has not been established, but the regional lymph nodes draining the uterus are potential sites of tolerance induction or Th2 bias. Finally, there could in theory be a systemic Th2 bias introduced by circulating cytokines or pregnancy-related hormones. This last possibility is the most problematic, since it must be reconciled with the fact that pregnant females mount successful Th1-type responses to infectious agents and skin grafts. Even if potentially cytotoxic T cells are generated by exposure to fetal antigens, the uterine microenvironment may suppress development of effector function (step 3 in Figure 1). PGE2 and cytokines such as IL-10 and TGFb can suppress T cell activation and proliferation and are found in placenta. For example, in the CBA x DBA/2 murine model of recurrent fetal loss, there is a local defect in placental IL-10 production, and systemic administration of IL-10 restores the ability of abortion-prone females to carry to term (26). In addition to soluble suppressive factors, there may also be structural features of the maternal-fetal interface that confer some degree of immune privilege. Hunt et al have described fas-ligand (fas-L) expressed in uterus and placenta, and have demonstrated increased leukocyte infiltrate into the maternal-fetal interface of fasL deficient (gld) mice (104). Although this infiltrate was composed of granulo-



385

cytes and was not antigen-driven (matings were syngeneic), it nonetheless raises the possibility that Fas-L could modulate local T cell activation in the placenta. Likewise, IDO has been described at the maternal-fetal interface in human placenta (32), where it may serve an immunosuppressive role (31). Finally, the most recent and controversial suggestion concerning maternal tolerance of the fetus asks whether there may be a systemic perturbation of specific T cell responses to fetal antigens. There is growing agreement that naive T cells that do not receive their initial activation under the normal (i.e., proinflammatory) conditions may be rendered unresponsive or anergic (reviewed in 105). Under certain circumstances, tolerized or anergized T cells may spread their unresponsiveness to other T cells recognizing the same antigen or to additional clones recognizing different, linked antigens (106, 107; and reviewed in 86). The mechanisms of these phenomena are not fully understood, but they suggest that there may exist antigen-specific tolerogenic mechanisms independent of the general bias toward Th2 in pregnancy. Evidence that such tolerogenic mechanisms might operate during pregnancy has been discussed above. Taken together, these findings suggest that the T cell repertoire capable of responding to fetal antigens is made aware of the presence of these antigens during pregnancy and rendered unresponsive to them in an antigen-specific manner. Whether this condition arises by direct exposure of all potentially responsive cells to antigen under tolerizing conditions or by tolerization of a subset of T cells with subsequent spread to the remainder of the repertoire remains to be determined.

CONCLUSIONS We propose that the key difference between the fetal allograft and a solid-organ transplant lies not in the ability of the adaptive immune system to see and respond to fetal alloantigens, but rather in the way in which the innate immune system treats the presence of the fetus. The innate system is alerted and responds actively to the fetal ‘‘invasion,’’ but the type of inflammation jointly created by fetally derived cells and the maternal innate immune system is not a milieu in which rejecting T cell responses are produced. However, far from being hidden from the maternal adaptive immune system, fetal alloantigens are actively involved in establishing a condition of antigen-specific tolerance during pregnancy.

ACKNOWLEDGMENTS These studies were supported by grants AI44219. AI42247, HL60137, and AI44759 from the National Institutes of Health, the Departments of Medicine and Pediatrics, Medical College of Georgia and generous support from the Trustees of the Carlos and Marguerite Mason Trust.

386

MELLOR n MUNN



LITERATURE CITED 1. King A, Loke YW. 1999. The influence of the maternal uterine immune response on placentation in human subjects. Proc. Nutr. Soc. 58:69–73 2. Rinkevich B. 1998. Immunology of human implantation: from the invertebrates’ point of view. Hum. Reprod. 13:455–59 3. Hunt JS, Robertson SA. 1996. Uterine macrophages and environmental programming for pregnancy success. Am. J. Reprod. Immunol. 32:1–25 4. Redline RW. 1997. The structural basis of maternal-fetal immune interactions in the human placenta. Curr. Top. Microbiol. Immunol. 222:25–44 5. Olding LB, Papadogiannakis N, Barbieri B, Murgita RA. 1997. Suppressive cellular and molecular activities in maternofetal immune interactions; suppressor cell activity, prostaglandins, and alphafetoproteins. Curr. Top. Microbiol. Immunol. 222:159–87 6. Raghupathy R. 1997. Th1-type immunity is incompatible with successful pregnancy. Immunol. Today 18:478–82 7. Torry DS, McIntyre JA, Faulk WP. 1997. Immunobiology of the trophoblast: mechanisms by which placental tissues evade maternal recognition and rejection. Curr. Top. Microbiol. Immunol. 222:127–40 8. Yokoyama WM. 1997. The mother-child union: the case of missing-self and protection of the fetus. Proc. Natl. Acad. Sci. USA 94:5998–6000 9. Sacks G, Sargent I, Redman C. 1999. An innate view of human pregnancy. Immunol. Today 20:114–18 10. Medawar PB. 1953. Some immunological and endocrinological problems raised

11.

12.

13.

14.

15.

16.

17.

18.

19.

by evolution of viviparity in vertebrates. Symp. Soc. Exp. Biol. 7:320–28 Simpson E, Scott D, Chandler P. 1997. The male-specific histocompatibility antigen, H-Y: a history of transplantation, immune response genes, sex determination and expression cloning. Annu. Rev. Immunol. 15:39–61 Shastri N, Serwold T, Paz P. 1998. Reading within the lines: naturally processed peptides displayed by MHC class I molecules. Curr. Opin. Immunol. 10:137–44 Simpson E. 1998. Minor transplantation antigens: animal models for human hostversus- graft, graft-versus-host, and graft-versus-leukemia reactions. Transplantation 65:611–16 Auchincloss H, Sultan H. 1996. Antigen processing and presentation in transplantation. Curr. Opin. Immunol. 8: 681–87 Gould DS, Auchincloss H Jr. 1999. Direct and indirect recognition: the role of MHC antigens in graft rejection. Immunol. Today 20:77–82 Billingham RE. 1964. Transplantation immunity and the maternal-fetal relation. New Engl. J. Med. 270:667–72 Bobe P, Kiger N. 1989. Immunogenetic studies of spontaneous abortion in mice. III. Non-H-2 antigens and gestation. J. Immunogenet. 16:223–31 Haddad EK, Duclos AJ, Baines MG. 1995. Early embryo loss is associated with local production of nitric oxide by decidual mononuclear cells. J. Exp. Med. 182:1143–51 Haddad EK, Duclos AJ, Lapp WS, Baines MG. 1997. Early embryo loss is associated with the prior expression of macrophage activation markers in the decidua. J. Immunol. 158:4886–92


T CELL TOLERANCE DURING PREGNANCY 20. Clark DA, Chaouat G, Mogil R, Wegmann TG. 1994. Prevention of spontaneous abortion in DBA/2-mated CBA/J mice by GM-CSF involves CD8` T celldependent suppression of natural effector cell cytotoxicity against trophoblast target cells. Cell. Immunol. 154:143–52 21. Lea RG, McIntyre S, Baird JD, Clark DA. 1998. Tumor necrosis factor-alpha mRNA-positive cells in spontaneous resorption in rodents. Am. J. Reprod. Immunol. 39:50–57 22. Martal J, Chene N, Camous S, Huynh L, Lantier F, Hermier P, L’Haridon R, Charpigny G, Charlier M, Chaouat G. 1997. Recent developments and potentialities for reducing embryo mortality in ruminants: the role of IFN-tau and other cytokines in early pregnancy. Reprod. Fertil. Dev. 9:355–80 23. Platt JS, Hunt JS. 1998. Interferongamma gene expression in cycling and pregnant mouse uterus: temporal aspects and cellular localization. J. Leuko. Biol. 64:393–400 24. Hunt JS, Miller M, Vassmer D, Croy BA. 1997. Expression of the inducible nitric oxide synthase gene in mouse uterine leukocytes and potential relationships with uterine function during pregnancy. Biol. Repro. 57:827–36 25. Head JR. 1996. Uterine natural killer cells during pregnancy in rodents. Nat. Immun. 15:7–21 26. Chaouat G, Assal Meliani A, Martal J, Raghupathy R, Elliot J, Mosmann T, Wegmann TG. 1995. IL-10 prevents naturally occurring fetal loss in the CBA x DBA/2 mating combination, and local defect in IL-10 production in this abortion-prone combination is corrected by in vivo injection of IFN-tau. J. Immunol. 154:4261–68 27. Robertson SA, Roberts CT, Farr KL, Dunn AR, Seamark RF. 1999. Fertility impairment in granulocyte-macrophage colony-stimulating factor-deficient mice. Biol. Repro. 60:251–61

387

28. Cohen PE, Zhu L, Pollard JW. 1997. Absence of colony stimulating factor-1 in osteopetrotic (csfmop/csfmop) mice disrupts estrous cycles and ovulation. Biol. Repro. 56:110–18 29. Shiraishi H, Hayakawa S, Satoh K. 1996. Murine experimental abortion by IL-2 administration is caused by activation of cytotoxic T lymphocytes and placental apoptosis. J. Clin. Lab. Immunol. 48:93– 108 30. Chaouat G, Tranchot Diallo J, Volumenie JL, Menu E, Gras G, Delage G, Mognetti B. 1997. Immune suppression and Th1/ Th2 balance in pregnancy revisited: a (very) personal tribute to Tom Wegmann. Am. J. Reprod. Immunol. 37:427–34 31. Munn DH, Zhou M, Attwood JT, Bondarev I, Conway SJ, Marshall B, Brown C, Mellor AL. 1998. Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 281:1191–93 32. Kamimura S, Eguchi K, Yonezawa M, Sekiba K. 1991. Localization and developmental change of indoleamine 2,3dioxygenase activity in the human placenta. Acta. Med. Okayama 45:135– 39 33. Taylor MW, Feng G. 1991. Relationship between interferon-c, indoleamine 2,3dioxygenase, and tryptophan catabolism. FASEB J. 5:2516–22 34. Munn DH, Shafizadeh E, Attwood JT, Bondarev I, Pashine A, Mellor AL. 1999. Inhibition of T cell proliferation by macrophage tryptophan catabolism. J. Exp. Med. 189:1363–72 35. Meier AH, Wilson JM. 1983. Tryptophan feeding adversely influences pregnancy. Life Sci. 32:1193–96 36. Schrocksnadel H, Baier-Bitterlich G, Dapunt O, Wachter H, Fuchs D. 1996. Decreased plasma tryptophan in pregnancy. Obstet. Gynecol. 88:47–50 37. Lin H, Mosmann TR, Guilbert L, Tuntipopipat S, Wegmann TG. 1993. Synthesis of T helper 2-type cytokines at the

388

38.

39.


40.

41.

42.

43.

44.

45.

46.

MELLOR n MUNN maternal-fetal interface. J. Immunol. 151:4562–73 Wegmann TG, Lin H, Guilbert L, Mosmann TR. 1993. Bidirectional cytokine interactions in the maternal-fetal relationship: Is successful pregnancy a TH2 phenomenon? Immunol. Today 14:353– 56 Piccinni MP, Beloni L, Livi C, Maggi E, Scarselli G, Romagnani S. 1998. Defective production of both leukemia inhibitory factor and type 2 T- helper cytokines by decidual T cells in unexplained recurrent abortions. Nat. Med. 4:1020–24 Vince GS, Johnson PM. 1995. Maternofetal immunobiology in normal pregnancy and its possible failure in recurrent spontaneous abortion? Hum. Reprod. 10(Suppl.)2:107–13 Shaarawy M, Nagui AR. 1997. Enhanced expression of cytokines may play a fundamental role in the mechanisms of immunologically mediated recurrent spontaneous abortion. Am. J. Obst. Gynec. 76:205–11 Sacks GP, Studena K, Sargent K, Redman CW. 1998. Normal pregnancy and preeclampsia both produce inflammatory changes in peripheral blood leukocytes akin to those of sepsis. Am. J. Obst. Gynec. 179:80–86 Redman CW, Sacks GP, Sargent IL. 1999. Preeclampsia: an excessive maternal inflammatory response to pregnancy. Am. J. Obst. Gynec. 180:499–506 Dekker GA, Sibai BM. 1999. The immunology of preeclampsia. Semin. in Perinatol. 23:24–33 Hamamah S, Fignon A, Lansac J. 1997. The effect of male factors in repeated spontaneous abortion: lesson from invitro fertilization and intracytoplasmic sperm injection. Hum. Reprod. Update 3:393–400 Labarrere CA, Faulk WP. 1995. Intercellular adhesion molecule-1 (ICAM-1) and HLA-DR antigens are expressed on endovascular cytotrophoblasts in abnor-

47.

48.

49.

50.

51.

52.

53.

54.

55.

mal pregnancies. Am. J. Reprod. Immunol. 33:47–53 Robillard P, Dekker GA. 1999. Revisiting the epidemiological standard of preeclampsia: primigravidity or primipaternity? Eur. J. Obst. Gynec. 84:37–41 Banchereau J, Steinman RM. 1998. Dendritic cells and the control of immunity. Nature 392:245–52 Van-Parijs L, Abbas AK. 1998. Homeostasis and self-tolerance in the immune system: turning lymphocytes off. Science 280:243–48 Goerdt S, Orfanos CE. 1999. Other functions, other genes: alternative activation of antigen- presenting cells. Immunity 10:137–42 Prabhala RH, Wira CR. 1995. Sex hormone and IL-6 regulation of antigen presentation in the female reproductive tract mucosal tissues. J. Immunol. 155:5566– 73 Piccinni MP, Giudizi MG, Biagiotti R, Beloni L, Giannarini L, Sampognaro S, Parronchi P, Manetti R, Annunziato F, Livi C, et al. 1995. Progesterone favors the development of human T helper cells producing Th2-type cytokines and promotes both IL-4 production and membrane CD30 expression in established Th1 cell clones. J. Immunol. 155:128–33 Ehring GR, Kerschbaum HH, Eder C, Neben AL, Fanger CM, Khoury RM, Negulescu PA, Cahalan MD. 1998. A nongenomic mechanism for progesterone-mediated immunosuppression: inhibition of K` channels, Ca2` signaling, and gene expression in T lymphocytes. J. Exp. Med. 188:1593–602 Miao D, Skibinski G, James K. 1996. The effects of human seminal plasma and PGE2 on mitogen induced proliferation and cytokine production of human splenic lymphocytes and peripheral blood mononuclear cells. Am. J. Reprod. Immunol. 30:97–114 Tremellen KP, Seamark RF, Robertson SA. 1998. Seminal transforming growth


56.

57.


58.

59.

60.

61.

62.

63.

64.

65.

factor beta1 stimulates granulocytemacrophage colony-stimulating factor production and inflammatory cell recruitment in the murine uterus. Biol. Repro. 58:1217–25 Tafuri A, Alferink J, Moller P, Hammerling GJ, Arnold B. 1995. T cell awareness of paternal alloantigens during pregnancy. Science 270:630–33 Zhou M, Mellor AL. 1998. Expanded cohorts of maternal CD8` T cells specific for paternal MHC class I accumulate during pregnancy. Am. J. Reprod. Immunol. 40:47–62 Jiang SP, Vacchio MS. 1998. Multiple mechanisms of peripheral T cell tolerance to the fetal ‘‘allograft’’. J. Immunol. 160:3086–90 Altman JD, Moss PAH, Goulder PJR, Barouch DH, McHeyzer-Wiliams MG, Bell JI, McMichael AJ. 1996. Phenotypic analysis of antigen-specific T lymphocytes. Science 274:94–96 Bonney EA, Matzinger P. 1997. The maternal immune system’s interaction with circulating fetal cells. J. Immunol. 158:40–47 Papadogiannakis N. 1997. Traffic of leukocytes through the maternofetal placental interface and its possible consequences. Curr. Top. Microbiol. Immunol. 222:141–57 Chatterjee-Hasrouni S, Lala PK. 1982. Localization of paternal H-2K antigens on murine trophoblast cells in vivo. J. Exp. Med. 155:1679–89 Philpott KL, Rastan S, Brown S, Mellor AL. 1988. Expression of H-2 class I genes in murine extra-embryonic tissues. Immunology 64:479–85 Rodriguez AM, Mallet V, Lenfant F, Arnaud J, Girr M, Urlinger S, Bensussan A, Le Bouteiller P. 1997. Interferongamma rescues HLA class Ia cell surface expression in term villous trophoblast cells by inducing synthesis of TAP proteins. Eur. J. Immunol. 27:45–54 Mattsson R. 1998. The non-expression of

66.

67.

68.

69.

70.

71.

72.

73.

74.

389

MHC class II in trophoblast cells. Am. J. Reprod. Immunol. 40:383–84 Ober C. 1998. HLA and pregnancy: the paradox of the fetal allograft. Am. J. Hum. Genet. 62:1–5 Arvola M, Gustafsson E, Brunsberg U, Mattsson R. 1999. Human choriocarcinoma-derived JEG-3 cells transfected with murine MHC class II Aq expression vectors present antigen to Aq-restricted murine T-cell hybridoma. Am. J. Reprod. Immunol. 42:17–30 Brahic M, Ozden S. 1992. Simultaneous detection of cellular RNA and proteins. In In Situ Hybridization, ed. D Rickwood, BD Hames. Oxford: Oxford Univ. Press Brunsberg U, Gustafsson K, Jansson L, Michaelsson E, Ahrlund-Richter L, Pettersson S, Mattsson R, Holmdahl R. 1994. Expression of a transgenic class II Ab gene confers susceptibility to collagen-induced arthritis. Eur. J. Immunol. 24:1698–1702 Gustafsson E, Arvola M, Brunsberg U, Mattsson A, Mattsson R. 1997. Lack of detectable major histocompatibility complex class II a beta- chain messenger ribonucleic acid in placentas of interferon-gamma- and 5-azacytidine-treated mice. Biol. Repro. 57:715–22 Morris AC, Riley JL, Fleming WH, Boss JM. 1998. MHC class II gene silencing in trophoblast cells is caused by inhibition of CIITA expression. Am. J. Reprod. Immunol. 40:385–94 Hiby SE, King A, Sharkey A, Loke YW. 1999. Molecular studies of trophoblast HLA-G: polymorphism, isoforms, imprinting and expression in preimplantation embryo. Tissue Antigens 53:1–13 Rouas-Freiss N, Goncalves RM, Menier C, Dausset J, Carosella ED. 1997. Direct evidence to support the role of HLA-G in protecting the fetus from maternal uterine natural killer cytolysis. Proc. Natl. Acad. Sci. USA 94:11520–25 Rajagopalan S, Long EO. 1999. A

390

75.

76.


77.

78.

79.

80.

81.

82.

83.

MELLOR n MUNN human histocompatibility leukocyte antigen (HLA)-G-specific receptor expressed on all natural killer cells. J. Exp. Med. 189:1093–1100 Fuchimoto Y, Yamada K, Shimizu A, Yasumoto A, Sawada T, Huang CH, Sachs DH. 1999. Relationship between chimerism and tolerance in a kidney transplantation model. J. Immunol. 162:5704–11 Simon AR, Warrens AN, Sykes M. 1999. Efficacy of adhesive interactions in pigto-human xenotransplantation. Immunol. Today 20:323–30 Knight M, Redman CW, Linton EA, Sargent IL. 1998. Shedding of syncytiotrophoblast microvilli into the maternal circulation in pre-eclamptic pregnancies. Br. J. Obstet. Gynaecol. 105:632–40 Lo YM, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, Wainscoat JS. 1997. Presence of fetal DNA in maternal plasma and serum. Lancet 350:485–87 Lo YM, Hjelm NM, Fidler C, Sargent IL, Murphy MF, Chamberlain PF, Poon PM, Redman W, Wainscoat JS. 1998. Prenatal diagnosis of fetal RhD status by molecular analysis of maternal plasma. New Engl. J. Med. 339:1734–38 Evans PC, Lambert N, Maloney S, Furst DE, Moore JM, Nelson JL. 1999. Longterm fetal microchimerism in peripheral blood mononuclear cell subsets in healthy women and women with scleroderma. Blood 93:2033–37 Clarke AG, Kendall MD. 1994. The thymus in pregnancy: the interplay of neural, endocrine and immune influences. Immunol. Today 15:545–51 Savion S, Toder V. 1995. Pregnancyassociated effect on mouse thymocytes in vitro. Cell. Immunol. 162:282–87 Remuzzi G. 1998. Cellular basis of longterm organ transplant acceptance: pivotal role of intrathymic clonal deletion and thymic dependence of bone marrow

84.

85.

86.

87.

88.

89.

90.

91. 92.

93.

microchimerism-associated tolerance. Am. J. Kidney Dis. 31:197–212 Ota H, Gotoh M, Ohzato H, He L, Tanigawa T, Nagano H, Dono K, Takeda Y, Okuyama M, Shimizu J, Umeshita K, Nakamori S, Sakon M, Nishisho I, Monden M. 1999. Microchimerism in thymus is associated with up-regulated T helper type 1 cytokine transcription during cardiac allograft rejection in rats. Transplantation 67:165–73 Hill JA, Polgar K, Anderson DJ. 1995. T-helper 1-type immunity to trophoblast in women with recurrent spontaneous abortion. JAMA 273:1933–36 Waldmann H, Cobbold S. 1998. How do monoclonal antibodies induce tolerance? A role for infectious tolerance? Annu. Rev. Immunol. 16:619–44 Janeway CA. 1992. The immune system evolved to discriminate infectious nonself from noninfectious self. Immunol. Today 13:11–16 Ibrahim MAA, Chain BM, Katz DR. 1995. The injured cell: the role of the dendritic cell system as a sentinel receptor pathway. Immunol. Today 16:181–86 Matzinger P. 1994. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12:991–1045 Fearon DT, Locksley RM. 1996. The instructive role of innate immunity in the acquired immune response. Science 272:50–54 Fearon DT. 1997. Seeking wisdom in innate immunity. Nature 388:323–24 Nabors GS, Afonso LC, Farrell JP, Scott P. 1995. Switch from a type 2 to a type 1 T helper cell response and cure of established Leishmania major infection in mice is induced by combined therapy with interleukin 12 and Pentostam. Proc. Natl. Acad. Sci. USA 92:3142–46 Afonso LCC, Scharton TM, Vieira LQ, Wysocka M, Trinchieri G, Scott P. 1994. The adjuvant effect of interleukin-12 in a vaccine against leishmania major. Science 263:235–37


T CELL TOLERANCE DURING PREGNANCY 94. Wallach D, Holtmann H, Engelmann H, Nophar Y. 1988. Sensitization and desensitization to lethal effects of tumor necrosis factor and IL-1. J. Immunol. 140:2994–99 95. Leonard JP, Sherman ML, Fisher GL, Buchanan LJ, Larsen G, Atkins MB, Sosman JA, Dutcher JP, Vogelzang NJ, Ryan JL. 1997. Effects of single-dose interleukin-12 exposure on interleukin12- associated toxicity and interferongamma production. Blood 90:2541–48 96. Konieczny BT, Dai Z, Elwood ET, Saleem S, Linsley PS, Baddoura FK, Larsen CP, Pearson TC, Lakkis FG. 1998. IFN-gamma is critical for longterm allograft survival induced by blocking the CD28 and CD40 ligand T cell costimulation pathways. J. Immunol. 160:2059–64 97. Cauley LS, Cauley KA, Shub F, Huston G, Swain SL. 1997. Transferable anergy: superantigen treatment induces CD4` T cell tolerance that is reversible and requires CD4-CD8- cells and interferon gamma. J. Exp. Med. 186:71–81 98. Kweon MN, Fujihashi K, VanCott JL, Higuchi K, Yamamoto M, McGhee JR, Kiyono H. 1998. Lack of orally induced systemic unresponsiveness in IFNgamma knockout mice. J. Immunol. 160:1687–93 99. Krakowski M, Owens T. 1996. Interferon-c confers resistance to experimental allergic encephalomyelitis. Eur. J. Immunol. 26:1641–46 100. Willenborg DO, Fordham S, Bernard CCA, Cowden WB, Ramshaw IA. 1996. IFN-c plays a critical down-regulatory role in the induction and effector phase of myelin oligodendrocyte glycoprotein-

101.

102.

103.

104.

105.

106.

107.

391

induced autoimmune encephalomyelitis. J. Immunol. 157:3223–27 Manoury-Schwartz B, Chiocchia G, Bessis N, Abehsira-Amar O, Batteux F, Muller S, Huang S, Boissier MC, Fournier C. 1997. High susceptibility to collageninduced arthritis in mice lacking IFNgamma receptors. J. Immunol. 158: 5501–6 Abe N, Katamura K, Shintaku N, Fukui T, Kiyomasu T, Iio J, Ueno H, Tai G, Mayumi M, Furusho K. 1997. Prostaglandin E2 and IL-4 provide naive CD4` T cells with distinct inhibitory signals for the priming of IFN-gamma production. Cell. Immunol. 181:86–92 Wolvers DA, Coenen-de Roo CJ, Mebius RE, van der Cammen MJ, Tirion F, Miltenburg AM, Kraal G. 1999. Intranasally induced immunological tolerance is determined by characteristics of the draining lymph nodes: studies with OVA and human cartilage gp-39. J. Immunol. 162:1994–98 Hunt JS, Vassmer D, Ferguson TA, Miller L. 1997. Fas ligand is positioned in mouse uterus and placenta to prevent trafficking of activated leukocytes between the mother and the conceptus. J. Immunol. 158:4122–28 Schwartz RH. 1996. Models of T cell anergy: Is there a common molecular mechanism? J. Exp. Med. 184:1–8 Davies JD, Leong LY, Mellor A, Cobbold SP, Waldmann H. 1996. T cell suppression in transplantation tolerance through linked recognition. J. Immunol. 156: 3602–7 Wise MP, Bemelman F, Cobbold SP, Waldmann H. 1998. Linked suppression of skin graft rejection can operate through indirect recognition. J. Immunol. 161:5813–16



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


REGULATION OF B LYMPHOCYTE RESPONSES TO FOREIGN AND SELF-ANTIGENS BY THE CD19/CD21 COMPLEX Douglas T. Fearon1 and Michael C. Carroll2


1 Wellcome Trust Immunology Unit, Department of Medicine, School of Clinical Medicine, University of Cambridge, Cambridge, United Kingdom; e-mail: [email protected]; 2 Center for Blood Research, Department of Pathology, Harvard Medical School, Boston, Massachusetts 02115; e-mail: [email protected]

Key Words signal transduction, tolerance, costimulation, innate immunity, complement Abstract The membrane protein complex CD19/CD21 couples the innate immune recognition of microbial antigens by the complement system to the activation of B cells. CD21 binds the C3d fragment of activated C3 that becomes covalently attached to targets of complement activation, and CD19 co-stimulates signaling through the antigen receptor, membrane immunoglobulin. CD21 is also expressed by follicular dendritic cells and mediates the long-term retention of antigen that is required for the maintenance of memory B cells. Understanding of the biology of this receptor complex has been enriched by analyses of genetically modified mice; these analyses have uncovered roles not only in positive responses to foreign antigens, but also in the development of tolerance to self-antigens. Studies of signal transduction have begun to determine the basis for the coreceptor activities of CD19. The integration of innate and adaptive immune recognition at this molecular site on the B cell guides the appropriate selection of antigen by adaptive immunity and emphasizes the importance of this coreceptor complex.

BACKGROUND: INNATE IMMUNITY AND THE CD19/ CD21 COMPLEX In 1968 Nussenzweig found that a subset of lymphocytes in the lymph node, but not in the thymus, bound immune complexes that had fixed complement (1). Although perhaps not fully appreciated at that time, this observation may have been the first to suggest that the role of the innate immune system of complement might be more than that of serving as an effector of humoral adaptive immune response. Because of the presence of a complement receptor on lymphocytes, complement might also have a role in the afferent limb of adaptive immunity. Four years later this suggestion was shown to be correct when Pepys found that 0732–0582/00/0410–0393$14.00

393


394

FEARON n CARROLL

depleting mice of C3 impaired their IgG response to a T-dependent antigen (2). Within the next five years, activated C3 was shown to bind covalently to complement-activators such as immune complexes (3). The bound C3 mediated uptake of the immune complexes by cells bearing C3 receptors, including follicular dendritic cells (FDCs) (4) and B lymphocytes, the latter having receptors, CR1 (CD35) and CD21 (CD21), for two of the bound C3 fragments, C3b and C3d, respectively (5, 6). C3 was also shown to be required for the normal generation of memory B cells (7). Thus, a rough outline of the molecular basis of Pepys’ observation was apparent early on, with antigen activating complement becoming coated with fragments of C3 and binding to B cells and FDCs, thereby enhancing the response to T-dependent antigens. Yet, this function of complement received relatively little attention from the immunological community, perhaps because innate immune systems seemed only peripherally relevant to the adaptive immune response, which presented the pressing problem of understanding lymphocyte function and diversity. By the late 1980s, the possible relevance of complement to activation of B lymphocytes was re-examined. Both complement receptors of B lymphocytes had been cloned (8–10). Coligating CD21 to membrane immunoglobulin, as would occur with antigen to which C3d had covalently bound, was found to enhance signaling through the antigen receptor by one to two orders of magnitude (11). On a more general level, the immunological community was reminded by Janeway of the strategic importance of innate immune systems to adaptive immunity by his remarking the ‘‘immunologists’ dirty little secret,’’ which was the need to mix antigens with adjuvants to induce robust immune responses (12). By the mid1990s, we had come to accept as an immunological principle that innate immunity must influence the selection of the targets and the modes of adaptive immune responses (13–15). This principle is based on an essential difference between the genes encoding the receptors of innate immunity, which are encoded in the germline, and those of adaptive immunity, the antigen receptors of B and T lymphocytes, which are products of somatically rearranging genes. The receptors of innate immunity reflect evolutionary selection for the recognition of structures that are associated with infectious microorganisms, but these receptors have evolved to detect only those structures that are highly conserved and cannot bear mutation without affecting microbial viability, such as lipopolysaccharide (LPS). On the other hand, adaptive immunity of vertebrate organisms has sacrificed its connection to evolutionary selection for microbial recognition to achieve, by the mechanism of somatic recombination of B and T cell receptor genes, an extraordinary diversity of antigen-binding proteins that are not constrained by prior selection. Therefore, a complementarity exists between the two branches of immunity in which their respective strengths and weaknesses are precisely balanced. Innate immunity detects relatively few structures that imply the presence of infectious organisms and transmits this information to adaptive immunity by promoting the activation


REGULATION OF B LYMPHOCYTE RESPONSES

395

and differentiation of those lymphocyte clones that express highly specific receptors capable of binding antigens associated with the infectious process. The CD19/CD21 complex of B lymphocytes is reviewed here as a molecular example of this complementarity. The complex, which was initially sought as a possible explanation for the paradox of the costimulatory function of CD21 on B cells, despite its short cytoplasmic domain, represents a discrete molecular site at which innate immunity could direct an adaptive response (16). The extracellular region of CD21 is comprised entirely of an ancient structural motif, the short consensus repeat, that is found in complement proteins of invertebrates (17), while that of CD19, a membrane protein present on all B cells, contains two Ig-like domains (18, 19), which form the structural motif characteristic of adaptive immunity. In the complex, CD21 serves as a ligand binding subunit that links the recognition of antigen by complement to the signaling function of CD19. This activity of CD19 is required for the formation and/or function of germinal centers (20, 21), which are the site of somatic hypermutation and selection for B cells having high affinity Ig, and also the site of the generation of memory B cells and long-lived plasma cells. Therefore, the association of CD21 with CD19 gives the complement system input into the most essential aspects of the humoral immune response to T-dependent antigens. Accordingly, an analysis of the CD19/CD21 complex should promote an understanding of the humoral immune response as well as exploration of one pathway of innate immune instruction of adaptive immunity.

REGULATION OF EXPRESSION OF CD19 AND CD21 The real, and apparent, composition of complexes of membrane proteins that contain CD19 is variable and may depend on the stage of B cell development and on technical aspects, such as the type of solubilizing detergent that is used. The complex was first described as being comprised of CD19, CD21 (16), and CD81 (TAPA-1) (22), a member of the tetraspan proteins, but subsequent studies using different detergents have found two other tetraspan proteins, CD9 and CD82 (23). The roles of the tetraspan members in this complex are not clear, although genetic deletion of CD81 in mice impairs the membrane expression of CD19 (24– 26), perhaps suggesting a role in trafficking of the complex to the plasma membrane. Unlike CD21, which is not an essential component and is actually lacking in pro- and pre-B cells, and present only at low levels in transitional B cells (27), CD81 is expressed at all stages of B cell development. It may be, therefore, an essential component of the complex, as suggested by the effects of its absence on CD19 expression, although this has not been specifically examined. CD81 and CD19 associate through their extracellular regions. As CD81 has been the subject of a recent excellent review (28), it is not discussed further. Although present on the B cell in a 1:1 complex (16) that requires the transmembrane and extracellular regions of both proteins (29, 30), CD19 and CD21

FEARON n CARROLL


396

can be independently expressed, in keeping with their differing transcriptional regulation. Even when both are present on the B cell, CD19 is present in molar excess to CD21, so that free CD19 is available. This finding may be relevant to the occurrence of ligands for CD19 that do not involve the complement system. Transcription of CD19 involves several transcription factors, the best characterized being the B cell–specific transcription factor, BSAP (also termed Pax5) (31). This paired-box transcription factor is expressed throughout B cell development until the plasma cell stage, in the testis and the mid-brain. Mice lacking BSAP have altered posterior mid-brain development, as well as complete arrest of B cell development at an early precursor stage leading to loss of small pre-B, B, and plasma cells (32). The promoter region of the Cd19 gene has a high-affinity BSAP binding site (31) that is occupied in a B cell line, but not in a non-B cell line. This site confers B cell specificity to a reporter gene in transient transfection experiments. Additional studies of the Cd19 promoter have suggested that other cis elements, designated PyG and GC boxes, are also necessary (33), but the trans-acting factors interacting with these sites have not been defined. The expression of CD21 is regulated independently of CD19, as both human and murine CD21 are expressed on cell types other than the B lymphocyte. For human CD21, these include follicular dendritic cells (FDCs), early thymocytes, a subset of mature T lymphocytes, and epithelial cells (34), the latter being an important observation that may account for the ability of the Epstein-Barr virus to infect epithelial cells in addition to B lymphocytes. Murine CD21 expression appears to be limited to B cells and FDCs. CD21 in humans and the mouse also differs from CD19 in its absence from pro-, pre- and immature B cells, appearing first in the IgMhiIgDlo transitional B cell (27). Studies of the molecular elements that regulate CD21 transcription have found that the 5’ promoter sequences lack tissue specificity. Rather, tissue-specific expression of both human and murine CD21 is determined by an intronic silencer that restricts expression to the cell lines and tissues that normally have CD21 (35, 36). The human silencer differs from the murine element in requiring stable integration into chromatin to mediate its effects, raising the possibility that nuclear matrix or chromatin interactions are mediated by this region of the Cd21 gene.

TABLE 1

Roles of CD21 in regulating antigen-dependent responses of B cells

CD21 expression

B cell activation (18 response)

Formation of memory B cells (GC reaction)

Long-term memory & affinity maturation

Negative selection

Co-receptor FDC

```` `

```` `

? ```

? ?


397


LIGATION OF THE CD19/CD21 COMPLEX Early attempts at demonstrating functions for CD19 or CD21 on the B lymphocyte relied on these membrane proteins being ligated independently of the antigen receptor. These studies led to conclusions that CD19 suppressed activation of the B cell (37), and that C3d or antibodies to CD21 enhanced B cell proliferative responses (38, 39). It is now apparent that these studies may have been misleading and that ligating either membrane protein should yield relatively comparable responses since they reside, in part, as a complex on the B lymphocyte. On human B cells, CD21 is complexed either with CD19 (16) and CD81 (21) or with CR1 (CD35) (41), and probably few, if any, reside alone. CD19 is in molar excess of CD21 at all stages of B cell development, so that a proportion of it will be uncomplexed. Cross-linking either CD19 or CD21 to limited numbers of mIgM on B cell enhances by 10- to 1000-fold the cellular responses, elevation of intracellular Ca2` (11, 41) activation of mitogen-activated proteins (MAP) kinases (42, 43), and proliferation (44). This strongly suggests that the physiological function of the CD19/CD21 complex is to promote signaling through the antigen receptor. That is, the CD19/CD21 complex serves as a coreceptor for membrane Ig, with costimulation requiring that the complex be coligated to the antigen receptor, rather than as an independent signaling complex. There may be some effects, however, that the complex can mediate that do not require ligation of the antigen receptor; this is discussed in a later section (Figure 1). Several experiments have also suggested that the antigen receptor can recruit CD19 function constitutively, that is, without coligating the coreceptor. This possibility is supported by the findings that capping of membrane IgM causes the cocapping of CD19 (45) and that small proportions of membrane IgM and CD19 coimmune precipitate via an association that requires a specific region of the cytoplasmic domain of CD19 (46). This same region is required for the tyrosine phosphorylation of CD19 when membrane IgM alone is ligated, which probably indicates that this reaction requires CD19 to be close to membrane Ig. However, some functional studies are not consistent with a major contribution of CD19 to antigen receptor signaling in the absence of coligation. Responses to type 2 Tindependent (TI-2) antigens in vivo are normal in Cd19-/- mice (20). In vitro proliferative responses to ligating membrane IgM alone are also normal (20), or 50% of normal (21), which contrasts with a 100-fold enhancement when membrane IgM and CD19 are coligated (44). Therefore, most evidence indicates that the full costimulatory function of the CD19/CD21 complex requires that the complex be cross-linked to the antigen receptor. If the basic requirement for costimulation is coligation of the CD19/CD21 complex with the antigen receptor, it follows that the ligands for the CD19/CD21 complex should be physically associated with antigen. In the instance of CD21, all preceding work on the complement system had indicated that C3, through its


398

FEARON n CARROLL

Figure 1 Coligation of the CD19/CD21 complex to membrane IgM by antigen coated with C3d recruits the signal transducing functions of CD19 that regulate essential functions of B cells.

ability to bind covalently to complement activators, was uniquely adapted for this purpose. The bound C3b would be proteolytically processed to iC3b and C3dg by factor I and a cofactor, either the plasma protein, factor H, or the other complement receptor of B cells, CR1 (CD35). The association of CR1 with CD21 on human B cells would then facilitate the transfer of processed antigen to CD21, which could then associate with CD19 to achieve signaling potential. A similar outcome would be achieved in the mouse, but by a different mechanism. Murine CR1 is an alternatively spliced extension of CD21 in which the C3b/C4b binding site is added N-terminal to the C3d-binding site (47, 48); that is, murine CR1 binds both C3b and C3d, and both forms of the murine complement receptor are associated with CD19 (49). An important corollary of the proposal that the CD19/CD21 complex functions as a coreceptor of the antigen receptor is that the complement system augments the adaptive immune response in an antigen-specific manner; that is, complement ‘‘tags’’ antigen with C3d. This property distinguishes complement from some other innate immune systems that promote the recognition of any antigen within the microenvironment of an infectious process. An example of this less specific



399

means for promoting adaptive immune recognition of antigens is the LPS receptor on dendritic cells that induces their maturation to competent antigen-presenting cells. Two considerations have suggested that other ligands for the CD19/CD21 complex, in addition to complement, exist. First, the two Ig-like domains in the extracellular region of CD19 raised the possibility that this membrane protein evolved with its own independent means for ligation. Presumably, CD21 then was selected for its capacity to associate with CD19, thereby integrating the complement system with the humoral adaptive immune response. Second, and more convincing, was the finding that the phenotypes of Cd211/1 mice and Cd191/1, although generally similar, differ in that of the Cd191/1mice is more pronounced. Cd191/1 mice are more severely immunodeficient than are Cd211/1mice. The serum immunoglobulin levels of the latter are normal (50, 51), whereas Cd191/1 mice have IgM levels that are reduced by 75% and IgG1 levels by 90% (20, 21). The IgM levels may be largely dependent on B-1 B cells, which are diminished in genetic and acquired deficiencies in CD19 (20, 21, 52), but less consistently so in Cd211/1 mice (50, 51). The serum IgG1 level probably reflects the responses of B-2 cells to environmental T-dependent antigens, and its reduction could indicate impaired T-dependent B cell responses in the Cd191/1 mice (20, 21). In accord with this suggestion, the memory B cell response to Tdependent antigens in the presence of adjuvant is normal or modestly impaired in Cd211/1 mice, whereas in Cd191/1 mice this response is absent despite the use of alum or complete Freund’s adjuvant (CFA). As is discussed in a later section, these contrasting responses may be caused by the more severe impairment in the formation of germinal centers in Cd191/1 mice than in Cd211/1 mice. Therefore, an alternative means for recruiting the function of CD19 that does not involve complement or CD21 seemed likely to exist. One report presented the glycosphingolipid globotriaosyl ceramide, CD77, as a candidate CD19 ligand (53). Nevertheless, CD77 does not seem to offer a means by which CD19 can be coligated to mIg, as it neither directly associates with antigen, nor is present on the cells that are specialized in presenting antigen to B cells, the FDCs. A recent unpublished study, however, has found biologically plausible candidates for CD19 ligands (54). A fusion protein between the membrane proximal Ig-like domain of CD19 and the Fc region of human IgG1 specifically bound to murine IgM and IgG3, but not to other Ig isotypes. The CD19-Fc fusion protein also bound to heparan sulfate, presumably as a proteoglycan, on murine stromal cell lines. The binding to IgM and IgG3 involved a common site in the extracellular domain of CD19, whereas heparan sulfate, although binding to the same Ig-like domain, interacted with a distinct site, as judged by competition experiments with IgM. The presumptive biological relevance of these findings is that IgM and IgG3, when bound to antigen, could coligate membrane Ig and CD19. Also possibly relevant is the origin of these two isotypes, as both are produced by B-1 cells. This subset of B cells makes polyreactive antibodies in a T-independent manner, possibly in response to self-antigens (55), that cross-react with microbial antigens


400

FEARON n CARROLL

having repetitive epitopes. IgM and IgG3 also are the major isotypes made in response to TI-2 antigens, a reaction that is relatively CD19 independent. Therefore, the recruitment of CD19, which is required for the T-dependent formation of the germinal center, may occur with antibody isotypes that are produced before the T-dependent phase of the B cell response to antigen, and, because of the unusual biology of B-1 cells, before any cellular response to a microbial antigen. This attractive synthesis, which avoids the circularity of requiring CD19 function in order to elicit CD19 function, may fall apart, however, when one considers that the development and/or maintenance of B-1 cells is itself CD19 dependent. Nevertheless, the manner by which CD19 function is recruited by B-1 cells is unknown and may not require IgM; for example, mice unable to secrete IgM have elevated numbers of B-1 cells (56, 57). The binding of CD19 by heparan sulfate on stromal cells, assuming that this characteristic of in vitro cultured stromal cells is shared by FDCs, would be another means by which antigen, when attached to an FDC through interaction with Fc or complement receptors, could coligate mIg and CD19. Finally, the possible role of CD21 in maintaining B-1 cells indicates that complement, which may be activated by pathways that do not require immunoglobulin, may have a role in eliciting CD19 function. The ordering of these possible means for recruiting CD19 function will require complex genetic experiments involving mutated forms of CD19, and mice that lack serum IgM and IgG3.

BIOLOGICAL FUNCTIONS OF THE CD19/CD21 COMPLEX Development The ability of cross-linked CD19 to block IL-7-induced downregulation of RAG expression by in vitro cultured B cell progenitors suggested that CD19 might have a role in early B cell development (58). Nevertheless, mice in which the Cd19 gene has been interrupted have relatively normal development and numbers of B-2 cells. Such mice do exhibit diminished numbers of B-1 cells. Conversely, Cd191/1 mice expressing graded numbers of human CD19 on their B cells in addition to endogenous CD19 have increasing numbers of B-1 cells and fewer B-2 cells (59). Treatment of mice with antibody to CD19 gradually leads to a reduction in B-1 cells, which is caused by their decreased replication rather than accelerated death (52). One of two Cd211/1 mouse strains had low numbers of B-1 cells (50). Therefore, the function of the CD19/CD21 complex is required for the normal development and/or maintenance of the B-1, but not the B-2, subset of B cells. However, as is subsequently discussed, the response of B-2 cells to Tdependent antigens is CD19 dependent. This apparent paradox of a requirement for CD19 in the development of B-1 cells but not B-2 cells, and in T cell- and antigen-dependent phase of B-2 cells,



401

and possibly B-1 cells, may have been partially resolved by a recent study of Hayakawa et al (55). The B-1 population of B cells produces ‘‘natural’’ autoantibodies, mainly of the IgM and IgG3 classes. Many of these cross-react with microbial antigens and are thought to provide immediate resistance to infection in the non-immunized host, making B-1 cells an innate immune functional adaptation of acquired immunity. The origin of these of B cells has been incompletely understood. Either they arise early during ontogeny and are maintained by a process of self-renewal, or they represent a form of antigen-activated B cells. Hayakawa et al have developed evidence in favor of the latter possibility by finding that the normal development of B-1 cells specific for the Thy-1 antigen requires the presence of this self-antigen. Therefore, B-1 cells are antigen experienced. It is interesting that the other major population of antigen-experienced B cells, the memory B-2 cell, like B-1 cells, are CD19 dependent (60), perhaps indicating that this membrane protein has a special role in the maintenance of antigen-selected B cells. The parallel, however, is imperfect in that B-1 cells maintain themselves by slowly replicating, whereas memory B-2 cells have been reported to be maintained without cellular division (61). Other aspects of B-1 cells that may be related to their requirement for CD19 are a dependency on Vav (62), Btk (63), and CD81 (24–26) for their presence. Vav is a cytosolic guanine nucleotide exchange factor for the Rho/Rac/ Cdc42 family of GTPases. Tyrosine phosphorylated CD19 binds and activates Vav (64, 65), perhaps accounting for the loss of B-1 cells when Vav is genetically deleted. Btk, or Bruton’s tyrosine kinase, is mutated in xid mice, which have a deficiency in B-1 cells, and CD19 has been reported to be involved in the activation of Btk (66). The decrease in B-1 cells in Cd811/1 mice may be caused by a secondary decrease in the surface expression of CD19.

Antigen-Dependent Responses of B Lymphocytes: CD19 The role of the CD19/CD21 complex in the activation of B cells by TI-1, TI-2, and T-dependent antigens has been examined. In one study, [3H]-thymidine incorporation induced by treating B cells with the TI-1 antigen LPS was reduced by 50% in Cd191/1 B cells as compared with wild-type B cells (21), but in another study there was no difference (20). Similarly, Cd191/1 B cells showed a 50% reduction in proliferation when activated by antibody to IgM in one study but did not differ from normal B cells in another study. Also, there was no impairment (20) or even an improvement (67) in the TI-2 response in Cd191/1 mice, which leads to the conclusion that T-independent responses of B cells are marginally, if at all, dependent on CD19. This finding is somewhat surprising because of the capacity of CD19, when coligated with membrane Ig, to amplify signaling in vitro, and the constitutive, albeit weak, association of CD19 with membrane IgM. Furthermore, the dependence of B-1 cells on CD19 seems inconsistent with this membrane protein having little role in T-independent responses, especially when we consider the recent demonstration that B-1 cells respond to a self-antigen,


402

FEARON n CARROLL

presumably in a T-independent manner. Some of these apparent discrepancies may ultimately be related to the context in which the B cell encounters type-2 Tindependent antigens, such as the presence or absence of a ligand for CD19. Responses to T-dependent antigens are markedly impaired in Cd191/1 mice in all studies that have been reported (20, 21, 60). The impairment is manifest as a decrease in the total IgG antibody titer elicited to the antigen, impaired affinity maturation of the antigen-specific antibody, absence of memory B cells, and no generation of long-lived plasma cells. The effect of CD19 deficiency on the formation of germinal centers depends on the source, and possibly the abundance, of the antigen. Following immunization with nonreplicating, T-dependent antigens, even with a potent adjuvant, Cd191/1 mice have a severe reduction in the number of germinal centers that are formed (20, 21). Since the germinal center reaction is the site for somatic hypermutation of V(D)J genes, memory cell differentiation, and development of plasma cells that home to the viability-promoting environment of the bone marrow, the absence of germinal centers may account for the impaired humoral response of Cd191/1. However, with a replicating viral antigen, a somewhat different outcome is observed (60). The Cd191/1 mice form germinal centers in response to infection with vesicular stomatitis virus but still do not generate memory B cells or long-lived plasma cells. Although memory B cells harboring somatically mutated V(D)J genes have been observed in genetically manipulated mice not capable of forming germinal centers (68), the Cd191/ mouse may be the only example in which this characteristic anatomical structure is formed but does not generate its normal output. These results may be related to two distinct functions for CD19 when serving as a coreceptor of the antigen receptor. CD19 lowers the threshold for the number of membrane Ig that must be ligated to cause the B cell to enter cell cycle, which was the first coreceptor function defined for CD19 (44). CD19 also may permit the antigen receptor to engage signaling pathways that would not otherwise be effectively activated, even when optimal numbers of membrane Ig had been ligated. An example of this second function of CD19 is the activation of certain MAP kinases only when CD19 is cross-linked to membrane Ig (43). Thus, when antigen is limiting, CD19 is required both to lower the threshold for membrane Ig signaling in germinal centers and for differentiation of memory B cells and plasma cells. When antigen is not limiting, as with a replicating virus, CD19 may not be required to amplify membrane Ig signaling in the germinal center, but its ability to alter the quality of mIg signaling remains essential. This suggestion is not yet proven.

Antigen-Dependent Responses of B Lymphocytes: CD21 Early observations made with mice having transient, acquired deficiencies of C3 (2) and guinea pigs with inherited deficiencies of C2 or C3 (69), and more recent studies of mice in which the C3 and C4 genes, respectively, had been interrupted by homologous recombination (70), have consistently indicated a role for com-



403

plement in the humoral response to T-dependent antigens. The complexity of the complement cascade, however, precluded the simple conclusion that CD21 was involved in this immunological response. Instead, the complement receptor itself had to become the subject of in vivo studies, the first of which was conducted by Heyman and her colleagues who used a monoclonal antibody blocking procedure to downregulate either CD35 alone or CD21 and CD35 jointly (71). Recall that in the mouse (47, 48), unlike the situation in humans (72, 73), both CD21 and CD35 are encoded at the Cd21 locus. For this discussion, the two receptors are referred to jointly as CD21/CD35. An antibody that blocked the binding of C3d by CD21/CD35 suppressed the response to horse red blood cells greater than 99%. The CD35-specific and joint CD21/CD35-specific antibodies had similar, although less marked, effects (71), suggesting that CD21 may have a role in humoral responses in the mouse and may mediate the effects of C3. A later study by Gustavsson et al using a similar approach demonstrated that CD4` helper T cells were generated despite the absence of a B cell response following treatment with the CD21 antibody (74). The mechanism of immunosuppression by the antiCD21/CD35 antibodies, however, may have been secondary to perturbation of B cell function by the cross-linked complement receptors. This alternative explanation was excluded by the finding that a soluble form of CD21, in which the C3d-binding site was fused to the N terminus of the heavy chain of an irrelevant murine IgG1, also suppressed T-dependent, but not T-independent, antibody responses in the mouse (75). The CD21-IgG1 fusion protein presumably competed with cellular CD21 for C3d-antigen complexes. However, like the Heyman report, these studies did not determine whether the important site for CD21 function was the B cell in the CD19/CD21 complex or on FDCs. The direct analysis of the function of CD21 awaited the application of gene targeting technology and development of Cd211/1 mice (50, 51). As already noted, the immunological phenotype of such mice is similar to, but less severe than, that of Cd191/1 mice. Cd211/1 mice have impaired primary and secondary humoral responses to T-dependent antigens, and a variable reduction in the number and size of germinal centers (50, 51, 76). This hallmark of the T-dependent B cell response is more dependent on CD19 than CD21 because in the latter, but not the former, the abnormality in germinal center formation and short-term secondary response can be overcome by the use of an adjuvant (77). This difference may be related to the occurrence of alternative means for coligating CD19 to the antigen receptor of the B cell. It should also be pointed out, as discussed below, that CD21/CD35 on FDCs has a role independent of CD19 in memory responses of B cells. Comparison of mice deficient in CD21/CD35 with those lacking C3 or C4 indicated a similar level of impairment both in primary and secondary humoral responses (50, 78). These findings suggest that the only biologically relevant ligand for CD21/CD35 is an activated product of C3, that is, C3d(g). This is an important point because in humans, CD21 has been reported to bind not only activated products of C3 but also the low affinity IgE receptor, CD23 (79). How-


404

FEARON n CARROLL

ever, in the mouse these interactions have not been reported, and the phenotype of Cd231/1mice does not resemble that of Cd211/1 or C31/1 mice. To clarify the relative importance of CD21 expression on B cells and FDCs, Croix et al (80) used the technique of complementing RAG-2 blastocysts with Cd211/1 embryonic stem (ES) cells. This provided a model in which the B cells, which are wholly derived from the ES cells, lacked CD21/CD35, whereas FDCs expressed complement receptors. Although the Cd211/1 B cells in the RAG-2 chimeras expressed normal levels of CD19, the secondary response of these mice to a T-dependent antigen was diminished. By contrast, Cd21`/`/RAG-2 chimeras displayed a normal secondary response following immunization. Furthermore, greater than 50% of the splenic B cell follicles had germinal centers in the chimeric mice having Cd21`/`B cells, but few germinal centers were observed in the mice with Cd211/1 B cells. Thus, the blastocyst complementation studies demonstrated that FDC expression of CD21/CD35 was not sufficient in the face of B cell deficiency in this receptor for development of memory B cells and their short-term maintenance. Long-term maintenance, however, is dependent on the expression of CD21/CD35 on FDCs (see below). An alternative approach to address the relative roles of complement receptor expression on B cells and FDCs used radiation chimeras in which lethally irradiated Cd211/1 mice were reconstituted with Cd21`/` bone marrow. Earlier studies had determined that reconstitution of irradiated mice with adult bone marrow did not lead to the appearance of bone marrow–derived FDCs (81), so that this approach could be used to repair the B cell defect in Cd211/1recipients without altering FDC expression. Analysis of the short-term secondary antibody response of the chimeric mice demonstrated a similar initial IgG titer as found with control normal mice (50). Therefore, CD21 expression on B cells was necessary and sufficient for a normal short-term secondary response, although it was noted that the IgG titer dropped abnormally rapidly in the reconstituted chimeric mice. Using a similar approach, Fang et al have recently reported that similar reconstitution of Cd211/1 mice with normal bone marrow only partially restored the humoral response to T-dependent antigens (76). These observations are discussed further as they suggest that the long-term memory or recall response might be impaired in the absence of CD21/CD35 expression on FDC. A possible mechanism explaining the impaired secondary response in mice either homozygous for CD21 deficiency or in chimeric mice lacking B cell expression of CD21 is that their B cells fail to receive a sufficient activation signal. As is discussed later, biochemical studies have revealed that coligation of CD21 or CD19 to the antigen receptor reduces the level of antigen receptor ligation required for B cell activation by approximately 100-fold. To examine this phenomenon in vivo, an adoptive transfer approach was used in which Cd21`/` or Cd211/1 antigen-specific B cells were adoptively transferred into immunized normal mice at the optimal B cell response period, that is, on day 8 following antigen injection (82). This approach (83) permits an analysis of the importance of CD21 in the presence of a constant level of T cell help and antigen retention



405

on FDC. Moreover, the assay is highly sensitive, as the transferred B cells must compete with endogenous cells for T help and antigen. B cells expressing the Ig transgene specific for hen egg lysozyme (HEL) were used as a source of B cells (84). The relative binding affinity of the hybridoma (Hy-HEL-10) from which the transgenic mice were constructed for HEL is approximately three orders of magnitude greater than for duck lysozyme (DEL), 4 x 1010 M11 versus 2 x 107 M11 (85). In transgenic mice immunized with DEL and given splenocytes from either Cd21`/` or Cd211/1 HEL transgenic mice on day 7 with additional HEL, expression of CD21 on B cells was required for maintenance or survival within the splenic white pulp region. By contrast, there was no significant difference between the two groups of transgenic B cells in this response in non-immunized chimeric mice. Thus, the presence of CD21 on B cells is required for their localization in the splenic white pulp in the presence of a limiting amount of antigen. This abnormality could be overcome by administering HEL, which was bound 1000-fold better by the membrane Ig of the transgenic B cells. Therefore, as predicted by in vitro studies of the coreceptor function of the CD19/CD21 complex, B cell activation for localization within the splenic follicles is dependent on costimulation of antigen receptor signaling. Preliminary results from studies with Cd211/1, HEL-transgenic/lpr mice suggest that the mechanism of elimination of the Cd211/1 B cells in this model requires Fas (M Zhang, personal communication). It may be that regulation of B cells receiving subthreshold stimulation by antigen by Fas represents a checkpoint to prevent bystander activation of self-reactive B cells. A similar proposal was made to explain elimination of anergic B cells that fail to respond to antigen in the presence of antigen-specific T cells (86). The finding that germinal centers, although fewer in number and smaller in area in Cd211/1 mice, did occur suggested that despite limited activation within the lymphoid compartment, some B cells do survive and initiate a germinal reaction in the absence of CD21. Thus, it was predicted that if CD21/CD35 expression were only necessary for ‘‘tuning-up’’ the membrane Ig signal, then Cd211/1, HEL transgenic B cells that received an optimal signal from the HEL antigen, for which they have a high affinity, would successfully compete with endogenous B cells for space within the germinal center (82). Unexpectedly, however, few such Cd21/1B cells survive within germinal centers of HEL-immune recipients. These results are consistent with the biochemical findings discussed below that coligating the CD19/CD21 complex to membrane Ig not only lowers the threshold for B cell stimulation but also induces a different activation pattern than does crosslinking of the antigen receptor alone. The interaction of antigen that has been coupled to C3d with the CD19/CD21 complex interaction appears to provide a selective advantage for germinal center survival and entry into the memory pool. Given the importance of the germinal center reaction to production of highaffinity memory B cells, the process of clonal selection must be rigorous. For example, interference with CD40 signaling results in a rapid loss of antigen-


406

FEARON n CARROLL

specific B cells (87). Similarly, flooding of the GC with excess antigen, as might occur in the mutation of a B cell to self, results in rapid elimination or migration out of the splenic white pulp (88–90). A possible mechanism to explain the latter observations is that injection of excess antigen bypasses coupling of C3d to antigen. This hypothesis supports the notion that the innate immune system has been conscripted to identify or mark those antigens to which an immune response would be biologically advantageous. Antigen-specific B cells undergo affinity maturation in response to Tdependent antigens (91), a process that most likely represents the preferential selection of high-affinity clones of B cells that have undergone somatic hypermutation within germinal centers. Accordingly, it might be predicted that Cd211/1 B cells that escape elimination within the follicles and survive the germinal center checkpoint would bind antigen at a higher affinity than Cd21`/`, normal B cells because they do not have the benefit of complement receptor– dependent amplified signaling. This prediction has experimental support. Using a high dose of antigen complexed with alum, Chen et al overcame the impaired germinal center response in the Cd211/1 mice and found that the overall affinity of the antibody response 4 months after immunization was higher in Cd211/1 than in the Cd21`/` controls (77). Genetic analysis revealed an increased frequency of coding region mutations in the variable region of antigen receptors expressed by Cd211/1 as compared with normal germinal center B cells. Moreover, a broader range of VH gene usage was observed in Cd211/1 B cells. Perhaps in the absence of an intact coreceptor that cannot be recruited by complement-bearing antigen, clonal selection is more intense and only cells bearing higher affinity receptors were selected. The maintenance of memory B cells requires the presence of antigen (92), and FDCs are generally considered to be the site of antigen retention for long-term B cell memory and continued affinity maturation of Ig genes (4). The mechanism for antigen retention on FDCs, however, is not well established. The two principle receptors that have been thought to bind antigen are FcgammaRs and CD21/ CD35. Of the three types of FcgammaRs, i.e. RI, RII, and RIII, RII is expressed on FDC (93). Its role in this function is not apparent, however, because FcgammaRII1/1 mice have normal or even enhanced secondary antibody responses to T-dependent antigens (94), perhaps because this receptor on B cells is also a negative regulator of signaling by membrane Ig. Early studies of complement receptors on FDCs suggesting a role in binding antigen found that transient reduction in C3 caused a significant decrease in the binding by these cells of aggregated IgG (4, 95). Nonetheless, chimeric mice bearing Cd21`/` B cells and Cd211/1 FDCs do make substantial secondary IgG responses, although these were elicited relatively soon after primary immunization, and they do not persist relative to secondary responses in normal mice. Recent experiments using an adoptive transfer approach have further dissected the importance of FDC expression of CD21/CD35 in longer-term memory, that is, 3 to 4 months after primary responses. Hapten- and carrier-primed B and T



407

cells were isolated from immunized normal mice and transferred, with antigen, into either irradiated Cd211/1 (FDC-CD211) or Cd21`/` (FDC-CD21`) recipients (96). Mice were challenged with antigen at either 3 weeks or 4 months after transfer, and the memory response was evaluated. Secondary immunization 3 weeks after transfer revealed no significant differences between the two mice in either the titer or affinity of antibody. By contrast, antigen challenge of the mice at 4 months showed a striking difference in both measurements, as antibody titer and affinity were both reduced approximately 10-fold in the FDC-CD211 mice compared with the FDC-CD21` chimeras. Interestingly, the antibody titer and affinity in the FDC-CD211 mice was similar to that found after the 3-week injection, suggesting that in the absence of CD21 on FDCs, the transferred memory B cells failed to continue a maturation process. A threshold of antigen and/or C3d on FDCs may be required for the maintenance of memory B cells that have exited the germinal center, and perhaps the continued selection for higher affinity B cell clones. Consistent with the conclusion that CD21/CD35 on FDCs has a major role in retaining C3d-bearing antigen is the finding that the deposits of C3d that are present on FDCs within splenic follicles of normal mice are absent in Cd211/1 mice.

Tolerance: CD19 A third aspect of the response of the B cell to antigen that can be influenced by CD19 and CD21 is the induction of B cell tolerance. In a model system in which mice express a transgene encoding human CD19, which leads to a two- to threefold increase in CD19 expression by B cells, elevated serum levels of rheumatoid factor and autoantibodies to DNA occur (59). One might have ascribed this increase to the augmented numbers of B-1 cells in these mice, except for the additional experiments performed in triple transgenic mice expressing membrane IgM specific for HEL, HEL, and human CD19 (97). In contrast to B cell tolerance and low levels of serum IgM anti-HEL that characterize mice having only murine CD19, mice expressing the additional human CD19 had serum levels of anti-HEL that approached those of nontolerant mice lacking the HEL transgene. This apparent loss of tolerance became more evident as the mice aged. Moreover, immunization of the triple transgenic mice, even at a young age and before the apparent loss of tolerance, caused an immune response leading to high levels of serum anti-HEL. Interestingly, administering only adjuvant increased serum anti-HEL. An aspect of this study that would be particularly important to examine would be the state of HEL-specific T cells in the immunized, triple transgenic mice. Although one would have anticipated that they would have been tolerant to HEL, the response to immunization suggests that they may have been activated to provide help for the B cell response to this T-dependent antigen. These experiments strongly suggest that CD19 can overcome the changes in signaling by membrane Ig induced by exposure of developing B cells to selfantigen. There may be several interpretations for these observations. A trivial,


408

FEARON n CARROLL

although unlikely, explanation may be that human CD19, which the transgene encoded in these studies, is not representative of murine CD19. For example, it may associate more firmly with membrane Ig than does murine CD19, or it may bind CD19 ligands more effectively. A nontrivial explanation would be that CD19 does not amplify signaling by membrane IgM during tolerance induction in immature B cells, but does rescue mature B cells from antigen-induced loss when T cell help is absent. This proposed function might be related to the role of CD19 in B-1 cells, which, as noted earlier, are likely to be selected positively by selfantigen rather than being tolerized. In addition, mice expressing human CD19 have increased numbers of B-1 cells (59). This possibility, however, does not account for the absence of increased B-1 cells in the triple transgenic mice. Finally, as suggested by the authors, B-2 cells having excessive CD19 may simply be hyperreactive, so they proliferate and differentiate in response to levels of stimulation through the antigen receptor that would be insufficient in a normal B cell. Their hyperreactivity may also cause them to respond to ‘‘inflammation,’’ accounting for induction of anti-HEL responses of triple transgenic mice given only adjuvant. An implication of this finding may be that the regulation of CD19 levels on B cells may be one of many genetic factors that contribute to the development of autoimmunity.

Tolerance: CD21 A role for complement in promoting adaptive immune responses to foreign antigens raises the possibility that it is involved in regulating self-reactive B cells. Studies with transgenic mice have led to a model in which the ‘‘strength’’ of the signal emanating from the antigen receptors on immature B cells determines whether they will be deleted or anergized or will continue to develop (98). For example, mutations in genes expressed in B cells that increase the intensity of membrane Ig signaling result in clonal deletion within the bone marrow by a selfantigen that usually causes anergy (99). The form of self-antigen also is important as membrane antigens can result in clonal deletion, whereas soluble forms of the same antigens only induce anergy (84, 100, 101). Given the importance of CD19/ CD21 complex as a coreceptor in lowering the threshold for antigen activation and the role of CD21/CD35 on FDCs in antigen retention, it is reasonable to consider the possibility that complement may modulate this process. Comparison of Cd21`/` with Cd211/1 mice that express transgenes for HEL and its corresponding membrane Ig revealed impaired induction of tolerance in the absence of CD21/CD35 (102). Thus, splenic B cells harvested from Cd211/1 double tg mice and cultured overnight with HEL responded by expressing B-7.2 (CD86) and releasing intracellular Ca2`, and both responses were diminished in the Cd21`/` mice. Consistent with this obvservation was the finding that C4null double tg chimeric mice also broke tolerance in this model. The combined results support a role for complement and CD21/CD35 in the regulation of tolerance against soluble self-antigens, which may be mediated through the



409

CD19/CD21 complex, as suggested by the findings of autoimmunity in mice overexpressing CD19. The precise stage of B cell development at which complement regulates B cell tolerance may be the transition from immature to mature B cells, as this is the point at which CD21/CD35 expression by this cell type occurs. The identification of a role for CD21/CD35 in negative selection in the transgenic model provides a basis for speculation that complement is involved in maintenance of tolerance to a subset of natural self-antigens. Related support for this function comes from the long-standing observation that individuals deficient in complement proteins C1q or C4 frequently develop systemic lupus erythematosus (SLE) (103), and that mice deficient in C1q develop autoantibodies and glomerulonephritis (104). Further support for this concept comes from the finding that the absence of CD21/CD35 exacerbates the autoimmunity occurring in lpr mice lacking Fas (102). Moreover, these mice develop severe glomerulonephritis, which is characteristic of SLE, while Cd21`/` lpr mice develop only mild disease, as has been reported for lpr on the C57BL/6 background. Therefore, some complement deficiencies may prediospose to autoimmunity because of impaired generation of C3d leading to altered B cell tolerance that is dependent on CD21/ CD35, and perhaps on the CD19/CD21 complex. The finding that complement is involved in B cell tolerance may seem at odds with the hypothesis that a function of innate immunity, of which the complement system is a part, is to direct the adaptive immune response to microbial antigens (14, 15). Thus, the primary role of complement may be to enhance the humoral response to infectious, non-self-antigens, yet another role is to identify noninfectious self-antigens for the induction of tolerance. A distinction between the two functions is the stage of maturation of the B cell at which they are manifest, with the former affecting mature B cells and the latter influencing immature B cells. Thus, immature cells are ‘instructed’’ to become tolerant or nonresponsive to specific self-antigens. It would be of interest if the relevant self-antigens are those that have conserved epitopes among vertebrates and microorganisms. In any event, the cell biological mechanisms of tolerance and immune enhancement may be similar and involve the costimulation of membrane Ig signaling by the CD19/CD21 complex.

SIGNAL TRANSDUCTION The current level of understanding of how CD19 costimulates the B cell, like that of its biological role in costimulation, is intriguing but incomplete. CD19 has been shown to costimulate several intracellular pathways that could regulate transcriptional responses of the B cell to antigen, but there has been little work in identifying these genes.

410

FEARON n CARROLL


Tyrosine Phosphorylation The major means by which CD19 is thought to couple extracellular ligation to intracellular signaling pathways is by the tyrosine phosphorylation of its large, 242 amino acid cytoplasmic domain, which is required for its signaling function (105). This phosphorylation may occur following ligation of either membrane Ig or the CD19/CD21 complex alone (106, 107), but it is at least 10- to 20-fold greater when these are coligated (65). This may indicate that the tyrosine kinase(s) acting on CD19 are most effectively activated by membrane IgM, and CD19 must be brought into proximity to the antigen receptor for optimal phosphorylation to occur. This conclusion may not be inconsistent with potential roles for the two members of the src family of tyrosine kinases, Lyn and Fyn, that have been reported to associate with CD19 (108, 109), as they may be involved in the relatively modest phosphorylation of CD19 when the membrane protein is ligated alone. However, Lyn is now known to have predominantly inhibitory functions in the B cell, which it mediates by phosphorylating the immunoreceptor tyrosinebased inhibitory motifs (ITIMs) of CD22 and FcgammaRII (110–112). In short, the tyrosine kinase that phosphorylates CD19 is not known, but its activity toward CD19 is best observed when membrane Ig and the CD19/CD21 complex are coligated. There are nine tyrosines in the cytoplasmic domain of CD19, but only three have been formally shown by mutational analysis to be phosphorylated: tyrosines391, -482 and -513 (in the human CD19 sequence) (42, 65, 106). The former binds Vav through its SH2 domain, and the latter two cooperate to bind phosphatidylinositol 3-kinase (PIP3-kinase) through the SH2 domains of this enzyme’s p85 subunit. Phosphotyrosyl peptides with sequences corresponding to Y405 and Y445 of CD19, and having the sequence, YEND/E, have been shown to bind a fusion protein containing the SH2 domain of Fyn (109), but CD19 has not been mutated at these positions to determine their role in signaling by CD19.

Calcium Responses Before CD19 was discovered to be a coreceptor of membrane Ig, cross-linking it alone with monoclonal antibodies was found to elevate intracellular Ca2`(37). However, CD19 was also reported to be associated with membrane IgM, and its Ca2`-inducing function to be ‘‘desensitized’’ by prior ligation of the antigen receptor (113), raising the possibility that this function of CD19 was merely caused by indirect ligation of membrane IgM. This possibility was excluded by the finding that the coligation of CD19 with membrane Ig synergistically increased intracellular Ca2`, since the costimulated response was greater than the additive response achieved by simultaneously and individually ligating the two receptors (41). Furthermore, whereas cross-linking membrane IgM alone caused the tyrosine phosphorylation of phospholipase C (PLC)-gamma, CD19 did not, and coligating membrane IgM and CD19 did not increase the tyrosine phosphor-



411

ylation of PLC-gamma, despite markedly increasing the generation of inositol1,4,5-trisphosphate (IP3) and elevating intracellular Ca2`. These findings suggested that CD19 promotes the interaction between PLC-gamma and its substrate, phosphatidylinositol 4,5-bisphosphate (PIP2), rather than enhancing the activation of the enzyme, as does, for example, SLP-76 in the T cell (114). CD19 could accomplish this either by facilitating the localization of PLC-gamma to the plasma membrane, or by promoting the synthesis of PIP2. The former has not been excluded, and evidence has been obtained for the latter possibility by analysis of the interaction of CD19 with Vav. Vav is a guanine nucleotide exchange factor for the Rho/Rac/Cdc42 family of small GTPases. Rho has been shown to regulate a phosphatidylinositol 4phosphate 5-kinase (PIP5-kinase) that synthesizes PIP2 from its precursor, phosphatidylinositol 4-phosphate. The activation of a PLC rapidly depletes cells of PIP2, and its availability becomes rate-limiting in the generation of IP3. IP3 is responsible for the release of intracellular Ca2` and, ultimately, the activity of the Ca2` release-activated calcium current that maintains long-term elevations of intracellular Ca2`. Therefore, the finding that ligating CD19 alone modestly activated and that coligating CD19 and membrane IgM strongly activated PIP5kinase in primary murine B cells suggested that the CD19 may contribute to the B cell Ca2` response by regulating the availability of PIP2 (65). Vav contributed to this function of CD19 because the prolonged Ca2` elevation caused by coligating CD19 and membrane IgM was suppressed in B cells from Vav1/1 mice. That the direct interaction between CD19 and Vav mediated the effect was confirmed by finding that mutating tyrosine-391 abolished the costimulatory function of CD19 in the Ca2` response of a B cell line. Thus, CD19 regulates activation of PIP5-kinase and cellular levels of PIP2 by recruiting the function of Vav. This activity of CD19 may also influence cytoskeletal reactions of costimulated B cells that are dependent on PIP2 and cellular responses that are related to phosphatidylinositol 3,4,5-phosphate, which is synthesized from PIP2. Recently, an additional mechanism for the enhancement of Ca2` response by CD19 has been proposed that involves PIP3-kinase and Btk (66, 115). This tyrosine kinase is mutated in xid mice and in humans with X-linked agamma globulinemia. The human phenotype is severe, with a marked arrest in B cell development, principally at the pre- to the immature B cell stage, whereas the murine developmental phenotype is less severe, although such mice do not respond to TI-2 antigens. Btk function is required for optimal stimulation by membrane IgM of Ca2` responses, perhaps through phosphorylation and activation of PLC-gamma. This function of Btk is dependent on its pleckstrin homology domain that interacts with PIP3 and causes the enzyme to be localized to the plasma membrane where it can more effectively interact with PLC-gamma. Therefore, it was postulated that the ability of CD19 to activate PIP3-kinase would cause CD19 to activate Btk. In a transfected plasmacytoma cell line and primary murine B cells, the presence of CD19 was shown to enhance the activation of PI3-kinase and elevation of intracellular Ca2` induced by ligating membrane IgM

412

FEARON n CARROLL

alone. This finding would be consistent with the association of small numbers of CD19 with the antigen receptor. Of more interest, however, is the finding of moderately diminished activation of Btk by membrane IgM in B cells from Cd191/ mice, and in a cell line expressing CD19 mutated at the tyrosines that bind PI3-kinase. Therefore, further work on the possible link between CD19 and Btk is merited that takes into account the markedly different phenotypes of xid mice and Cd191/1 mice. Among these differences are the absence of TI-2 responses and abnormal B-2 B cell development in xid mice, neither of which has been observed in Cd191/1 mice.


Activation of MAP Kinases The other major pathways that are costimulated by membrane IgM and CD19 are those leading to the MAP kinases, ERK2, JNK1, and p38. In primary murine B cells, coligating CD19 to membrane IgM caused a 5- to 10-fold augmentation of the activation of all three MAP kinases, relative to the effects of ligating optimal amounts of membrane IgM alone (43). At suboptimal levels of antigen receptor stimulation, the increment in MAP kinase activation was in the range of 100-fold. Ligating CD19 alone had no or modest effects, mainly on p38, which is no greater than twofold over that of unstimulated cells. Enhanced activation of ERK2 by coligated membrane Ig and CD19 on a human B cell line also occurs (42). In other transformed B cell lines, CD19 alone activated ERK2, and this function was related to the developmental stage of the cell lines (116). Interestingly, activation of nuclear factor-kappaB by CD19 was observed in an immature B cell line, the only example of this transcription factor being downstream of CD19. The pathways leading from CD19 to the MAP kinases are poorly defined. In one study (117) the ability of CD19 to costimulate ERK2 activation in the Burkitt lymphoma line, Daudi, was not dependent on elevations in Ca2`. In this same cell line, ERK2 activation was shown to be dependent on tyrosine-391 (42), but in murine B cells lacking Vav, ERK2 costimulation by CD19 was found to be intact (43). In this latter study, Vav was shown to participate in the activation of JNK1, as might have been anticipated based on its role in stimulating Rac, which is proximal to JNK in other cellular systems. In an interesting analysis of the costimulation of enzymes that lead to activation ERK2, only enhanced activation of MAPK kinase 1 (MEK1), and not of Ras or Raf, was observed (117). This raises the possibility that MEK1 is the point at which the membrane Ig and CD19 signals converge on the way to ERK2. Enhanced activation of the Ras pathway may also be related to the tyrosine phosphorylation and membrane translocation of a Shc complex caused by costimulation by CD19 and the antigen receptor (118).

Genes Induced by the CD19/CD21 Complex An important aim of signaling studies is to identify those genes that are uniquely induced by the costimulatory function of the CD19/CD21 complex. These would account for the essential role of CD19 in the differentiation of memory B cells



413

and long-lived plasma cells. Although this aim has not been achieved, two studies have examined the role of the CD19/CD21 complex in augmenting the antigenpresenting function of B cells, and in promoting their resistance to apoptosis. In the first, ligating CR1/CD21 or CD19 alone on murine B cells upregulated the expression of B7–1 and B7–2 (120). Remarkably, costimulation with mIgM did not further enhance this response, and cross-linking mIgM alone only increased the expression of B7–2. These findings might suggest that for this function the CD19/CD21 complex acts independently of the antigen receptor, which certainly would be possible. The need for the germinal center reaction to select for mutated B cells having high-affinity membrane Ig, however, would seem to require the costimulated expression of B-7, which would link effective antigen presentation to germinal center T cells with uptake of antigen by the selecting B cells. Thus, the findings of this study are surprising. Perhaps the means by which the CD19/ CD21 complex was cross-linked, which was with primary rat monoclonal antibodies followed by a goat ant-rat antibody, caused the coligation of the small numbers of mIgM because the goat antibody weakly cross-reacted membrane IgM of the murine B cells. In any event, that the CD19/CD21 complex has a role in inducing expression of the gene encoding B7–1 and B7–2 is an important finding when considering functions of CD19 that contribute the germinal center response. The induction of the anti-apoptotic protein, bcl-2, by coligating CD19/CD21 with membrane IgM, but not by membrane Ig alone (120), is also potentially relevant to the role of this complex in the B cell response to T-dependent antigens. For example, it may be related to the role of CD19 in the development or in the maintenance of memory B cells that exit the germinal center, as bcl-2 is expressed in these cells (121). However, bcl-2 expression is low in the germinal center (122, 123), perhaps to permit the negative selection of autoreactive, hypermutated antibody V regions at this site, so that this finding may not be related to functions of CD19 at this site.

EFFECTS OF INHIBITORY RECEPTORS: CD22, FCGAMMARIIB As discussed earlier in this review, the participation of CD19 in B cell activation is governed mainly by its cross-linking to the antigen receptor, so that inhibitors of signaling by the antigen receptor may also suppress costimulation by the CD19/ CD21 complex. Of the several inhibitory receptors on the B cell, only one, CD22, has been examined for this function. CD22 is tyrosine phosphorylated by Lyn following ligation of membrane Ig. Of the six tyrosines in the cytoplasmic domain of CD22, three have the characteristic ITIM sequence that is found in many inhibitory membrane receptors of the immune system and, when phosphorylated on tyrosine, that mediates the binding of SH2 domain-containing phosphatases. These sequences in CD22 bind and activate the phosphotyrosine phosphatase,


414

FEARON n CARROLL

SHP-1 by which CD22 downregulates signaling through membrane Ig (124, 125). Sequestration of CD22 from membrane Ig amplifies by 100-fold the proliferative response of B cells in vitro to ligating the antigen receptor. Moreover, B cells from Cd221/1 mice are hyperresponsive to membrane Ig signaling (126, 127), and develop autoantibodies (128). These inhibitory functions of CD22 are dominant over the costimulatory function of the CD19/CD21 complex. The synergistic activation of the MAP kinases induced by coligating the complex to membrane IgM are suppressed in a dose-related manner by the additional juxtaposition of CD22 (43). The basis for this function of CD22 is its inhibitory effect on proximal steps of signaling by membrane Ig because it suppresses the tyrosine phosphorylation of many cellular proteins in addition to that of CD19 and presumably inhibits the activation of the protein tyrosine kinase responsible for phosphorylation of CD19. The FcgammaRIIB1 has a single ITIM motif and, when phosphorylated by cross-linking to membrane Ig, binds the inositol polyphosphate 5’-phosphatase, SHIP, rather than SHP-1 (129). This phosphatase regulates the influx of Ca2` by mechanisms that are not entirely clear, although its role in regulating Btk function may be relevant (130). SHIP converts PIP3 to phosphatidylinositol 3,4-bisphosphate, abolishing its ability to recruit Btk to the plasma membrane. As Btk is involved in the activation of PLC-gamma (131), this may be one means by which SHIP regulates the Ca2` response. Cross-linking FcgammaRIIB1 to membrane Ig suppresses the tyrosine phosphorylation of CD19 and the association of PI3kinase with the membrane protein (132). It was proposed that this terminates IP3 production, perhaps because the production of PIP3 is diminished and Btk is no longer activated. These studies have not been extended to an analysis of the effects of FccRIIB1 on co-stimulation by the CD19/CD21 complex that is cross-linked to the antigen receptor.

CONCLUSION The CD19/CD21 complex, paradoxically, guards against self-reactivity during the development of B cells while, in mature B cells, it guides the selection of appropriate foreign antigens for the humoral adaptive immune response. The development and maintenance of memory B cells and B-1 cells are dependent on the CD19/CD21 complex. CD19 has an obligatory signal-transducing role, whereas CD21 may serve only as a ligand-binding subunit that links the B cell to the innate immune system of complement. CD19 also may be directly stimulated by other ligands that are associated with antigen, such as IgM, which probably accounts for its coreceptor function in the absence of CD21. CD21 also has CD19-independent functions in relation to its expression by FDCs, which enables C3d-bearing antigens to persist for long times in the host, thereby providing continual stimulation of antigen-specific B cells. The costimulatory functions of the CD19/CD21 complex also are involved in the conditioning of B cells to selfantigens, so that under- or overexpression of either component leads to altered


415

tolerance. The signaling function of CD19 involves the regulation of several intracellular pathways, including Ca2`, phosphoinositol metabolism, and MAP kinase activation, but the genes that CD19 uniquely regulates to permit the maturation of germinal center B cells into memory B cells have not been identified. The latter remains as perhaps the outstanding challenge to be met in our understanding of this coreceptor complex. Visit the Annual Reviews home page at www.AnnualReviews.org.


LITERATURE CITED 1. Lay WH, Nussenzweig V. 1968. Receptors for complement of leukocytes. J. Exp. Med. 128:991–1009 2. Pepys MB. 1972. Role of complement in induction of the allergic response. Nat. New. Biol. 237:157–59 3. Law SK, Levine RP. 1977. Interaction between the third complement protein and cell surface macromolecules. Proc. Natl. Acad. Sci. USA 74:2701–5 4. Klaus GG, Humphrey JH, Kunkl A, Dongworth DW. 1980. The follicular dendritic cell: its role in antigen presentation in the generation of immunological memory. Immunol. Rev. 53:13–28 5. Eden A, Miller GW, Nussenzweig V. 1973. Human lymphocytes bear membrane receptors for C3b and C3d. J. Clin. Invest. 52:3239–42 6. Ross GD, Polley MJ, Rabellino EM, Grey HM. 1973. Two different complement receptors on human lymphocytes. One specific for C3b and one specific for C3b inactivator-cleaved C3b. J. Exp. Med. 138:798–811 7. Klaus GG, Humphrey JH. 1977. The generation of memory cells. I. The role of C3 in the generation of B memory cells. Immunology 33:31–40 8. Klickstein LB, Bartow TJ, Miletic V, Rabson LD, Smith JA, Fearon DT. 1988. Identification of distinct C3b and C4b recognition sites in the human C3b/C4b

9.

10.

11.

12.

13.

14.

15.

receptor (CR1, CD35) by deletion mutagenesis. J. Exp. Med. 168:1699–717 Moore MD, Cooper NR, Tack BF, Nemerow GR. 1987. Molecular cloning of the cDNA encoding the Epstein-Barr virus/ C3d receptor (complement receptor type 2) of human B lymphocytes. Proc. Natl. Acad. Sci. USA 84:9194–98 Weis JJ, Toothaker LE, Smith JA, Weis JH, Fearon DT. 1988. Structure of the human B lymphocyte receptor for C3d and the Epstein-Barr virus and relatedness to other members of the family of C3/C4 binding proteins. J. Exp. Med. 167:1047–66 Carter RH, Spycher MO, Ng YC, Hoffman R, Fearon DT. 1988. Synergistic interaction between complement receptor type 2 and membrane IgM on B lymphocytes. J. Immunol. 141:457–63 Janeway CA Jr. 1989. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harbor Symp. Quant. Biol. 54:1–13 Matzinger P. 1994. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12:991–1045 Fearon DT, Locksley RM. 1996. The instructive role of innate immunity in the acquired immune response. Science 272:50–53 Medzhitov R, Janeway CA Jr. 1997. Innate immunity:the virtues of a non-

416

16.


17.

18.

19.

20.

21.

22.

23.

FEARON n CARROLL clonal system of recognition. Cell 91:295–98 Matsumoto AK, Kopicky-Burd J, Carter RH, Tuveson DA, Tedder TF, Fearon DT. 1991. Intersection of the complement and immune systems: a signal transduction complex of the B lymphocytecontaining complement receptor type 2 and CD19. J. Exp. Med. 173:55–64 Smith LC, Chang L, Britten RJ, Davidson EH. 1996. Sea urchin genes expressed in activated coelomocytes are identified by expressed sequence tags. Complement homologues and other putative immune response genes suggest immune system homology within the deuterostomes. J. Immunol. 156:593– 602 Stamenkovic I, Seed B. 1988. CD19, the earliest differentiation antigen of the B cell lineage, bears three extracellular immunoglobulin-like domains and an Epstein-Barr virus-related cytoplasmic tail. J. Exp. Med. 168:1205–10 Tedder TF, Isaacs CM. 1989. Isolation of cDNAs encoding the CD19 antigen of human and mouse B lymphocytes. A new member of the immunoglobulin superfamily. J. Immunol. 143:712–17 Rickert RC, Rajewsky K, Roes J. 1995. Impairment of T-cell-dependent B-cell responses and B-1 cell development in CD19-deficient mice. Nature 376:352– 55 Engel P, Zhou LJ, Ord DC, Sato S, Koller B, Tedder TF. 1995. Abnormal B lymphocyte development, activation, and differentiation in mice that lack or overexpress the CD19 signal transduction molecule. Immunity 3:39–50 Bradbury LE, Kansas GS, Levy S, Evans RL, Tedder TF. 1992. The CD19/CD21 signal transducing complex of human B lymphocytes includes the target of antiproliferative antibody-1 and Leu-13 molecules. J. Immunol. 149:2841–50 Horvath G, Serru V, Clay D, Billard M, Boucheix C, Rubinstein E. 1998. CD19

24.

25.

26.

27.

28.

29.

30.

31.

32.

is linked to the integrin-associated tetraspans CD9, CD81, and CD82. J. Biol. Chem. 273:30,537–43 Maecker HT, Levy S. 1997. Normal lymphocyte development but delayed humoral immune response in CD81-null mice. J. Exp. Med. 185:1505–10 Tsitsikov EN, Gutierrez-Ramos JC, Geha RS. 1997. Impaired CD19 expression and signaling, enhanced antibody response to type II T independent antigen and reduction of B-1 cells in CD81deficient mice. Proc. Natl. Acad. Sci. USA 94:10844–49 Miyazaki T, Muller U, Campbell KS. 1997. Normal development but differentially altered proliferative responses of lymphocytes in mice lacking CD81. EMBO J. 16:4217–25 Tedder TF, Clement LT, Cooper MD. 1984. Expression of C3d receptors during human B cell differentiation: immunofluorescence analysis with the HB-5 monoclonal antibody. J. Immunol. 133:678–83 Levy S, Todd SC, Maecker HT. 1998. CD81 (TAPA-1): a molecule involved in signal transduction and cell adhesion in the immune system. Annu. Rev. Immunol. 16:89–109 Matsumoto AK, Martin DR, Carter RH, Klickstein LB, Ahearn JM, Fearon DT. 1993. The CR2/CD19 complex. Functional dissection of the CD21/CD19/ TAPA-1/Leu-13 complex of B lymphocytes. J. Exp. Med. 178:1407–17 Bradbury LE, Goldmacher VS, Tedder TF. 1993. The CD19 signal transduction complex of B lymphocytes. Deletion of the CD19 cytoplasmic domain alters signal transduction but not complex formation with TAPA-1 and Leu 13. J. Immunol. 151:2915–27 Kozmik Z, Wang S, Dorfler P, Adams B, Busslinger M. 1992. The promoter of the CD19 gene is a target for the B-cellspecific transcription factor BSAP. Mol. Cell. Biol. 12:2662–72 Urbanek P, Wang ZQ, Fetka I, Wagner


33.

34.


35.

36.

37.

38.

39.

40.

EF, Busslinger M. 1994. Complete block of early B cell differentiation and altered patterning of the posterior midbrain in mice lacking Pax5/BSAP. Cell 79:901– 12 Riva A, Wilson GL, Kehrl JH. 1997. In vivo footprinting and mutational analysis of the proximal CD19 promoter reveal important roles for an SP1/Egr-1 binding site and a novel site termed the PyG box. J. Immunol. 159:1284–92 Birkenbach M, Tong X, Bradbury LE, Tedder TF, Kieff E. 1992. Characterization of an Epstein-Barr virus receptor on human epithelial cells. J. Exp. Med. 176:1405–14 Hu H, Martin BK, Weis JJ, Weis JH. 1997. Expression of the murine CD21 gene is regulated by promoter and intronic sequences. J. Immunol. 158: 4758–68 Makar KW, Pham CT, Dehoff MH, O’Connor SM, Jacobi SM, Holers VM. 1998. An intronic silencer regulates B lymphocyte cell- and stage-specific expression of the human complement receptor type 2 (CR2, CD21) gene. J. Immunol. 160:1268–78 Pezzutto A, Dorken B, Rabinovitch PS, Ledbetter JA, Moldenhauer G, Clark EA. 1987. CD19 monoclonal antibody HD37 inhibits anti-immunoglobulin-induced B cell activation and proliferation. J. Immunol. 138:2793–99 Nemerow GR, McNaughton ME, Cooper NR. 1985. Binding of monoclonal antibody to the Epstein Barr virus (EBV)/ CR2 receptor induces activation and differentiation of human B lymphocytes. J. Immunol. 135:3068–73 Carter RH, Fearon DT. 1989. Polymeric C3dg primes human B lymphocytes for proliferation induced by anti-IgM. J. Immunol. 143:1755–60 Tuveson DA, Ahearn JM, Matsumoto AK, Fearon DT. 1991. Molecular interactions of complement receptors on B lymphocytes:a CR1/CR2 complex dis-

41.

42.

43.

44.

45.

46.

47.

48.

49.

417

tinct from the CR2/CD19 complex. J. Exp. Med. 173:1083–89 Carter RH, Tuveson DA, Park DJ, Rhee SG, Fearon DT. 1991. The CD19 complex of B lymphocytes. Activation of phospholipase C by a protein tyrosine kinase-dependent pathway that can be enhanced by the membrane IgM complex. J. Immunol. 147:3663–71 Li X, Sandoval D, Freeberg L, Carter RH. 1997. Role of CD19 tyrosine 391 in synergistic activation of B lymphocytes by coligation of CD19 and membrane Ig. J. Immunol. 158:5649–57 Tooze RM, Doody GM, Fearon DT. 1997. Counterregulation by the coreceptors CD19 and CD22 of MAP kinase activation by membrane immunoglobulin. Immunity 7:59–67 Carter RH, Fearon DT. 1992. CD19: lowering the threshold for antigen receptor stimulation of B lymphocytes. Science 256:105–7 Pesando JM, Bouchard LS, McMaster BE. 1989. CD19 is functionally and physically associated with surface immunoglobulin. J. Exp. Med. 170:2159–64 Carter RH, Doody GM, Bolen JB, Fearon DT. 1997. Membrane IgM-induced tyrosine phosphorylation of CD19 requires a CD19 domain that mediates association with components of the B cell antigen receptor complex. J. Immunol. 158: 3062–69 Kurtz CB, O’Toole E, Christensen SM, Weis JH. 1990. The murine complement receptor gene family. IV. Alternative splicing of Cr2 gene transcripts predicts two distinct gene products that share homologous domains with both human CR2 and CR1. J. Immunol. 144:3581–91 Molina H, Kinoshita T, Webster CB, Holers VM. 1994. Analysis of C3b/C3d binding sites and factor I cofactor regions within mouse complement receptors 1 and 2. J. Immunol. 153:789–95 Krop I, Shaffer AL, Fearon DT, Schlissel

418

50.


51.

52.

53.

54. 55.

56.

57.

FEARON n CARROLL MS. 1996. The signaling activity of murine CD19 is regulated during cell development. J. Immunol. 157:48–56 Ahearn JM, Fischer MB, Croix D, Goerg S, Ma M, Xia J, Zhou X, Howard RG, Rothstein TL, Carroll MC. 1996. Disruption of the Cr2 locus results in a reduction in B-1a cells and in an impaired B cell response to T-dependent antigen. Immunity 4:251–62 Molina H, Holers VM, Li B, Fung Y, Mariathasan S, Goellner J, StraussSchoenberger J, Karr RW, Chaplin DD. 1996. Markedly impaired humoral immune response in mice deficient in complement receptors 1 and 2. Proc. Natl. Acad. Sci. USA 93:3357–61 Krop I, de Fougerolles AR, Hardy RR, Allison M, Schlissel MS, Fearon DT. 1996. Self-renewal of B-1 lymphocytes is dependent on CD19. Eur. J. Immunol. 26:238–42 Maloney MD, Lingwood CA. 1994. CD19 has a potential CD77 (globotriaosyl ceramide)-binding site with sequence similarity to verotoxin B-subunits: implications of molecular mimicry for B cell adhesion and enterohemorrhagic Escherichia coli pathogenesis. J. Exp. Med. 180:191–201 de Fougerolles AR, Fearon DT. 1999. Unpublished observations Hayakawa K, Asano M, Shinton SA, Gui M, Allman D, Stewart CL, Silver J, Hardy RR. 1999. Positive selection of natural autoreactive B cells. Science 285:113–16 Ehrenstein MR, O’Keefe TL, Davies SL, Neuberger MS. 1998. Targeted gene disruption reveals a role for natural secretory IgM in the maturation of the primary immune response. Proc. Natl. Acad. Sci. USA 95:10089–93 Boes M, Esau C, Fischer MB, Schmidt T, Carroll M, Chen J. 1998. Enhanced B1 cell development, but impaired IgG antibody responses in mice deficient in secreted IgM. J. Immunol. 160:4776–87

58. Billips LG, Nunez CA, Bertrand FE 3rd, Stankovic AK, Gartland GL, Burrows PD, Cooper MD. 1995. Immunoglobulin recombinase gene activity is modulated reciprocally by interleukin 7 and CD19 in B cell progenitors. J. Exp. Med. 182:973–82 59. Sato S, Ono N, Steeber DA, Pisetsky DS, Tedder TF. 1996. CD19 regulates B lymphocyte signaling thresholds critical for the development of B-1 lineage cells and autoimmunity. J. Immunol. 157:4371–78 60. Fehr T, Rickert RC, Odermatt B, Roes J, Rajewsky K, Hengartner H, Zinkernagel RM. 1998. Antiviral protection and germinal center formation, but impaired B cell memory in the absence of CD19. J. Exp. Med. 188:145–55 61. Schittek B, Rajewsky K. 1990. Maintenance of B-cell memory by long-lived cells generated from proliferating precursors. Nature 346:749–51 62. Zhang R, Alt FW, Davidson L, Orkin SH, Swat W. 1995. Defective signalling through the T- and B-cell antigen receptors in lymphoid cells lacking the vav proto-oncogene. Nature 374:470–73 63. Hayakawa K, Hardy RR, Herzenberg LA. 1986. Peritoneal Ly-1 B cells: genetic control, autoantibody production, increased lambda light chain expression. Eur. J. Immunol. 450–56 64. Weng WK, Jarvis L, LeBien TW. 1994. Signaling through CD19 activates Vav/ mitogen-activated protein kinase pathway and induces formation of a CD19/ Vav/phosphatidylinositol 3-kinase complex in human B cell precursors. J. Biol. Chem. 269:32514–21 65. O’Rourke LM, Tooze R, Turner M, Sandoval DM, Carter RH, Tybulewicz VL, Fearon DT. 1998. CD19 as a membraneanchored adaptor protein of B lymphocytes: costimulation of lipid and protein kinases by recruitment of Vav. Immunity 8:635–45 66. Buhl AM, Cambier JC. 1999. Phosphorylation of CD19 Y484 and Y515,


67.


68.

69.

70.

71.

72.

73.

74.

and linked activation of phosphatidylinositol 3-kinase, are required for B cell antigen receptor-mediated activation of Bruton’s tyrosine kinase. J. Immunol. 162:4438–46 Sato S, Steeber DA, Tedder TF. 1995. The CD19 signal transduction molecule is a response regulator of B-lymphocyte differentiation. Proc. Natl. Acad. Sci. USA 92:11558–62 Matsumoto M, Lo SF, Carruthers CJ, Min J, Mariathasan S, Huang G, Plas DR, Martin SM, Geha RS, Nahm MH, Chaplin DD. 1996. Affinity maturation without germinal centres in lymphotoxin-alpha-deficient mice. Nature 382: 462–66 Bitter-Suermann D, Burger R. 1990. C3 deficiencies. Curr. Top. Microbiol. Immunol. 153:223–33 Carroll MC. 1998. The role of complement and complement receptors in induction and regulation of immunity. Annu. Rev. Immunol. 16:545–68 Heyman B, Wiersma EJ, Kinoshita T. 1990. In vivo inhibition of the antibody response by a complement receptorspecific monoclonal antibody. J. Exp. Med. 172:665–68 Carroll MC, Alicot EM, Katzman PJ, Klickstein LB, Smith JA, Fearon DT. 1988. Organization of the genes encoding complement receptors type 1 and 2, decay-accelerating factor, and C4-binding protein in the RCA locus on human chromosome 1. J. Exp. Med. 167:1271– 80 Rey-Campos J, Rubinstein P, Rodriguezde-Cordoba S. 1988. A physical map of the human regulator of complement activation gene cluster linking the complement genes CR1, CR2, DAF, and C4BP. J. Exp. Med. 167:664–69 Gustavsson S, Kinoshita T, Heyman B. 1995. Antibodies to murine complement receptor 1 and 2 can inhibit the antibody response in vivo without inhibiting T

75.

76.

77.

78.

79.

80.

81.

82.

83.

419

helper cell induction. J. Immunol. 154:6524–28 Hebell T, Ahearn JM, Fearon DT. 1991. Suppression of the immune response by a soluble complement receptor of B lymphocytes. Science 254:102–5 Fang Y, Xu C, Fu Y, Holers VM, Molina H. 1998. Expression of complement receptors 1 and 2 on follicular dendritic cells is necessary for the generation of a strong antigen-specific IgG response. J. Immunol. 160:5273–79 Chen Z, Koralov S, Gendelman M, Carroll M, Kelsoe G. 1999. In situ studies of the primary immune response to (4-hydroxy-3-nitrophenyl) acetyl. VI. Regulation of antibody affinity and persistence by CD21/CD35. Manuscript submitted Fischer M, Ma M, Goerg S, Zhou X, Xia J, Finco O, Han S, Kelsoe G, Howard R, Rothstein T, Kremmer E, Rosen F. 1996. Regulation of the B cell response to Tdependent antigens by classical pathway complement. J. Immunol. 157:549–56 Aubry J-P, Pochon S, Grabar P, Jansen K, Bonnefoy J-Y. 1992. CD21 is a ligand for CD23 and regulates IgE production. Nature 358:505–7 Croix D, Ahearn J, Rosengard A, Han S, Kelsoe G, Ma M, Carroll M. 1996. Antibody response to a T-dependent antigen requires B cell expression of complement receptors. J. Exp. Med. 183: 1857–64 Humphrey JH, Grennan D, Sundaram V. 1984. The origin of follicular dendritic cells in the mouse and the mechanism of trapping of immune complexes on them. Eur. J. Immunol. 14:859–64 Fischer MB, Goerg S, Shen LM, Prodeus AP, Goodnow CC, Kelsoe G, Carroll MC. 1998. Dependence of germinal center B cells on expression of CD21/CD35 for survival. Science 280:582–85 Shokat K, Goodnow CC. 1995. Antigeninduced B-cell death and elimination

420

84.

85.


86.

87.

88.

89.

90.

91.

92.

FEARON n CARROLL during germinal centre immune response. Nature 375:334–38 Goodnow CC, Crosbie J, Adelstein S, Lavoie TB, Smith-Gill SJ, Brink RA, Pritchard-Briscoe H, Wotherspoon JS, Loblay RH, Raphael K, Trent RJ, Basten A. 1988. Altered immunoglobulin expression and functional silencing of selfreactive B lymphocytes in transgenic mice. Nature 334:676–82 Lavoie T, Drohan W, Smith-Gill S. 1992. Experimental analysis by site-directed mutagenesis of somatic mutation effects on affinity and fine specificity in antibodies specific for lysozyme. J. Immunol. 148:503–13 Rathmell JC, Cooke MP, Ho WY, Grein J, Townsend SE, Davis MM, Goodnow CC. 1995. CD95 (Fas)-dependent elimination of self-reactive B cells upon interaction with CD4` T cells. Nature 376:181–84 Han S, Hathcock K, Zheng B, Kepler TB, Hodes R, Kelsoe G. 1995. Cellular interaction in germinal centers. Roles of CD40 ligand and B7–2 in established germinal centers. J. Immunol. 155:556– 67 Han S, Zheng B, Dal Porto J, Kelsoe G. 1995. In situ studies of the primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. IV. Affinity-dependent, antigen-driven B cell apoptosis in germinal centers as a mechanism for maintaining self-tolerance. J. Exp. Med. 182:1635–44 Pulendran B, Kannourakis G, Nouri S, Smith KGC, Nossal GJV. 1995. Soluble antigen can cause enhanced apoptosis of germinal center B cells. Nature 375:331– 33 Shokat K, Goodnow CC. 1995. Antigeninduced B-cell death and elimination during germinal centre immune response. Nature 375:334–38 Rajewsky K. 1996. Clonal selection and learning in the antibody system. Nature 381:751–57 Gray D, Skarvall H. 1988. B cell memory

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

is short-lived in the absence of antigen. Nature 336:70–73 Bolland S, Ravetch J. 1999. Inhibitory pathways triggered by ITIM-containing receptors. Adv. Immunol. 72:149–77 Takai T, Ono M, Hikida M, Ohmori H, Ravetch J. 1996. Augmented humoral and anaphylactic responses in Fc gamma RII-deficient mice. Nature 379:346–49 Papamichail M, Gutierrez C, Embling P, Johnson P, Holborow EJ, Pepys MB. 1975. Complement dependence of localization of aggregated IgG in germinal centers. Scand. J. Immunol. 4:343–47 Fischer MB, Ma M, Prodeus AP, Carroll MC. 1999. Maintenance of memory and affinity maturation of B cells is dependent on expression of CD21/CD35 on follicular dendritic cells. Submitted for publication Inaoki M, Sato S, Weintraub BC, Goodnow CC, Tedder TF. 1997. CD19regulated signaling thresholds control peripheral tolerance and autoantibody production in B lymphocytes. J. Exp. Med. 186:1923–31 Goodnow CC. 1996. Balancing immunity and tolerance: Deleting and tuning lymphocyte repertoires. Proc. Natl. Acad. Sci. USA 93:2264–71 Cyster JG, Goodnow C C. 1995. Protein tyrosine phosphatase 1C negatively regulates antigen receptor signaling in B lymphocytes and determines the thresholds for negative selection. Immunity 2:13–24 Goodnow CC, Crosbie J, Jorgensen H, Brink RA, Basten A. 1989. Induction of self-tolerance in mature peripheral B lymphocytes. Nature 342:385–91 Nemazee DA, Burki K. 1989. Clonal deletion of B lymphocytes in a transgenic mouse bearing anti-MHC class I antibody genes. Nature 337:562–66 Prodeus AP, Goerg S, Shen LM, Pozdnykova OO, Chu L, Alicot E, Goodnow CC, Carroll MC. 1998. A critical role for


103.

104.


105.

106.

107.

108.

109.

110.

111.

complement in maintenance of self-tolerance. Immunity 9:721–31 Walport MJ, Morgan BP. 1991. Complement deficiency and disease. Immunol. Today 12:301–6 Botto M, Dell’Agnola C, Bygrave AE, Thompson EM, Cook HT, Petry F, Loos M, Pandolfi PP, Walport MJ. 1998. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nat. Genet. 19:56–59 Sato S, Miller AS, Howard MC, Tedder TF. 1997. Regulation of B lymphocyte development and activation by the CD19/CD21/CD81/Leu 13 complex requires the cytoplasmic domain of CD19. J. Immunol. 159:3278–87 Tuveson DA, Carter RH, Soltoff SP, Fearon DT. 1993. CD19 of B cells as a surrogate kinase insert region to bind phosphatidylinositol 3-kinase. Science 260:986–89 Chalupny NJ, Kanner SB, Schieven GL, Wee SF, Gilliland LK, Aruffo A, Ledbetter JA. 1993. Tyrosine phosphorylation of CD19 in pre-B and mature B cells. EMBO J. 12:2691–96 van Noesel CJ, Lankester AC, van Schijndel GM, van Lier RA. 1993. The CR2/CD19 complex on human B cells contains the src-family kinase Lyn. Int. Immunol. 5:699–705 Chalupny NJ, Aruffo A, Esselstyn JM, Chan PY, Bajorath J, Blake J, Gilliland LK, Ledbetter JA, Tepper MA. 1995. Specific binding of Fyn and phosphatidylinositol 3-kinase to the B cell surface glycoprotein CD19 through their src homology 2 domains. Eur. J. Immunol. 25:2978–84 Smith KGC, Tarlinton DM, Doody GM, Hibbs ML, Fearon DT. 1998. Inhibition of the B cell by CD22: a requirement for Lyn. J. Exp. Med. 187:807–11 Nishizumi H, Horikawa K, MlinaricRascan I, Yamamoto T. 1998. A doubleedged kinase Lyn: a positive and negative regulator for antigen receptor-

112.

113.

114.

115.

116.

117.

118.

119.

120.

421

mediated signals. J. Exp. Med. 187: 1343–48 Chan VW, Lowell CA, DeFranco AL. 1998. Defective negative regulation of antigen receptor signaling in Lyn-deficient B lymphocytes. Curr. Biol. 8:545– 53 Rijkers GT, Griffioen AW, Zegers BJ, Cambier JC. 1990. Ligation of membrane immunoglobulin leads to inactivation of the signal-transducing ability of membrane immunoglobulin, CD19, CD21, and B-cell gp95. Proc. Natl. Acad. Sci. USA 87:8766–70 Yablonski D, Kuhne MR, Kadlecek T, Weiss A. 1998. Uncoupling of nonreceptor tyrosine kinases from PLC-gamma1 in an SLP-76-deficient T cell. Science 281:413–16 Buhl AM, Pleiman CM, Rickert RC, Cambier JC. 1997. Qualitative regulation of B cell antigen receptor signaling by CD19: selective requirement for PI3kinase activation, inositol-1,4,5-trisphosphate production and Ca2` mobilization. J. Exp. Med. 186:1897–1910 Weng WK, Shah N, O’Brien D, Van Ness B, LeBien TW. 1997. Differential induction of DNA-binding activities following CD19 cross-linking in human B lineage cells. J. Immunol. 159:5502–8 Li X, Carter RH. 1998. Convergence of CD19 and B cell antigen receptor signals at MEK1 in the ERK2 activation cascade. J. Immunol. 161:5901–8 Lankester AC, van Schijndel GM, Rood PM, Verhoeven AJ, van Lier RA. 1994. B cell antigen receptor cross-linking induces tyrosine phosphorylation and membrane translocation of a multimeric Shc complex that is augmented by CD19 co-ligation. Eur. J. Immunol. 24:2818–25 Kozono Y, Abe R, Kozono H, Kelly RG, Azuma T, Holers VM. 1998. Cross-linking CD21/CD35 or CD19 increases both B7–1 and B7–2 expression on murine splenic B cells. J. Immunol. 160:1565– 72 Roberts T, Snow EC. 1999. Recruitment

422

121.


122.

123.

124.

125.

126.

FEARON n CARROLL of the CD19/CD21 coreceptor to B cell antigen receptor is required for antigenmediated expression of Bcl-2 by resting and cycling hen egg lysozyme transgenic B cells. J. Immunol. 162:4377–80 Bovia F, Nabili-Tehrani AC, WernerFavre C, Barnet M, Kindler V, Zubler RH. 1998. Quiescent memory B cells in human peripheral blood co-express bcl-2 and bcl-x(L) anti-apoptotic proteins at high levels. Eur. J. Immunol. 28:4418–23 Liu YJ, Mason DY, Johnson GD, Abbot S, Gregory CD, Hardie DL, Gordon J, MacLennan IC. 1991. Germinal center cells express bcl-2 protein after activation by signals which prevent their entry into apoptosis. Eur. J. Immunol. 21: 1905–10 Martinez-Valdez H, Guret C, de Bouteiller O, Fugier I, Banchereau J, Liu YJ. 1996. Human germinal center B cells express the apoptosis-inducing genes Fas, c-myc, P53, and Bax but not the survival gene bcl-2. J. Exp. Med. 183: 971–77 Doody GM, Justement LB, Delibrias CC, Matthews RJ, Lin J, Thomas ML, Fearon DT. 1995. A role in B cell activation for CD22 and the protein tyrosine phosphatase SHP. Science 269:242–44 Campbell MA, Klinman NR. 1995. Phosphotyrosine-dependent association between CD22 and protein tyrosine phosphatase 1C. Eur. J. Immunol. 25: 1573–79 O’Keefe TL, Williams GT, Davies SL, Neuberger MS. 1996. Hyperresponsive B

127.

128.

129.

130.

131.

132.

cells in CD22-deficient mice. Science 274:798–801 Sato S, Miller AS, Inaoki M, Bock CB, Jansen PJ, Tang ML, Tedder TF. 1996. CD22 is both a positive and negative regulator of B lymphocyte antigen receptor signal transduction: altered signaling in CD22-deficient mice. Immunity 5:551– 62 O’Keefe TL, Williams GT, Batista FD, Neuberger MS. 1999. Deficiency in CD22, a B cell-specific inhibitory receptor, is sufficient to predispose to development of high affinity autoantibodies. J. Exp. Med. 189:1307–13 Ono M, Bolland S, Tempst P, Ravetch JV. 1996. Role of the inositol phosphatase SHIP in negative regulation of the immune system by the receptor Fc(gamma)RIIB. Nature 383:263–66 Bolland S, Pearse RN, Kurosaki T, Ravetch JV. 1998. SHIP modulates immune receptor responses by regulating membrane association of Btk. Immunity 8:509–16 Takata M, Kurosaki T. 1996. A role for Bruton’s tyrosine kinase in B cell antigen receptor-mediated activation of phospholipase C-gamma 2. J. Exp. Med. 184:31– 40 Hippen KL, Buhl AM, D’Ambrosio D, Nakamura K, Persin C, Cambier JC. 1997. Fc gammaRIIB1 inhibition of BCR-mediated phosphoinositide hydrolysis and Ca2` mobilization is integrated by CD19 dephosphorylation. Immunity 7:49–58


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:423–449

REGULATORY T CELLS IN AUTOIMMMUNITY* Ethan M. Shevach Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892; e-mail: [email protected]


Key Words suppressor T cells, immune regulation, organ-specific autoimmunity, gastritis Abstract Clonal deletion of autoreactive T cells in the thymus is not the sole mechanism for the induction of tolerance to self-antigens since partial depletion of peripheral CD4` T cells from neonatal and adult animals results in the development of organ-specific autoimmunity. Reconstitution of these immunodeficient animals with populations of regulatory CD4`T cells prevents the development of autoimmunity. The lineage of regulatory CD4` T cells is generated in the thymus and can be distinguished from effector cells by the expression of unique membrane antigens. The target antigens for these suppressor populations and their mechanisms of action remain poorly defined. Depletion of regulatory T cells may be useful in the induction of immunity to weak antigens, such as tumor-specific antigens. Conversely, enhancement of regulatory T cell function may be a useful adjunct to the therapy of autoimmune diseases and for prevention of allograft rejection.

INTRODUCTION Shortly after the discovery in the late 1960s that T lymphocytes functioned as helper cells for B lymphocytes, RK Gershon (1) proposed that T cells could also act as regulatory cells that could suppress the immune response. The prevailing view in the 1970s–1980s was that suppressor T cells mediated their biologic effects by producing soluble factors that were responsible for their biologic activity. The early volumes of this series contained detailed reviews describing the immunologic properties of these factors. Many of these soluble suppressor molecules were claimed to be antigen-specific and contained I-J determinants encoded by the MHC and/or determinants encoded by Ig genes. Molecular studies in the mid-1980s, however, clearly demonstrated that the genes encoding the T cell receptor were not identical to those encoding Ig and failed to identify an I-J gene within the MHC. As succinctly pointed out by Green & Webb (2), the ‘‘S’’ word became the nearest thing to a dirty word in cellular immunology, and its use was considered synonymous with over-interpretation of scanty data and mystical phenomenology. *The US government has the right to retain a nonexclusive, royalty-free license in and to any copyright covering this paper.

423

424

SHEVACH


Tolerance still remains a fundamental concept of modern immunology, but clonal deletion of autoreactive T cells in the thymus cannot be the sole mechanism for the induction of tolerance. Many of the in vivo experiments performed over the past 30 years contain compelling data that supports the existence of suppressor T cells. The discovery of Th1/Th2 cells in the late 1980s prompted most workers in the field to abandon the concept that suppressor T cells were a specialized population; suppression was merely the result of the activity of counter-regulatory cytokines. The purpose of this review is to resurrect the concept that the suppressor T cell is a member of a separate lineage of cells that mediates its downregulatory functions by a variety of effector mechanisms. I focus on the role of suppressor T cells in the prevention of organ-specific autoimmunity and begin with an overview of the animal models initially described 30 years ago, which formed the foundation for more recent studies that have facilitated the characterization of suppressor T cells.

DEPLETION OF REGULATORY T CELLS FROM NEONATAL ANIMALS During the course of studies investigating the role of the thymus in mediating tumor immunity, Nishizuka & Sakakura (3) observed that female mice that were thymectomized (Tx) early in life became infertile secondary to the development of oophoritis. Ovarian atrophy was observed when the thymus was removed from neonatal animals on day 3 of life and not on day 1 or day 7. Thymus grafting at 7 days of age prevented the development of disease, while grafting at 40 days of age had no effect. Later studies demonstrated that grafting of an intact thymus was not required, as disease could be prevented by injection of a suspension of thymocytes from 7-day-old or adult mice or of spleen or lymph node cell suspensions from adult mice. In contrast, grafting of allogeneic thymus, syngeneic newborn spleen, injection of bone marrow cells, spleen cells from 7-day-old intact donors, or spleen cells from adult d3Tx donors failed to protect. The conclusions drawn from these studies were that thymus-derived lymphocytes were responsible for the suppression of disease and that the cells that prevented the disease in adult spleen were derived from the thymus. Although these cells developed in the thymus of the neonate, they were not peripheralized to spleen or lymph nodes during the first 3 days of life since spleen cells of newborn mice were not effective in preventing disease and d3Tx completely extinguished this lineage. These investigators considered the possibility early on that a suppressor factor might be extractable from the thymus, which would be effective in vivo, but crude extracts from the thymus did not substitute for the intact thymus! Thus, they did not embark on the treacherous journey of the ‘‘mainstream’’ immunologists who had simultaneously discovered suppressor T cells.


REGULATORY T CELL FUNCTION

425

The syndrome of post-d3Tx-induced organ-specific autoimmunity was characterized in depth over the next 10 years (4, 5). In addition to autoimmune oophoritis and orchitis in A strain mice, thyroiditis was most prominently observed in C3H mice, while autoimmune gastritis was most often seen in BALB/c mice. The gastritis was characterized by a loss of chief and parietal cells and by varying degrees of lymphoid infiltration along thickened gastric muscularis mucosa. Antiparietal cell antibodies were present in the sera of mice with gastritis and megaloblastic anemia subsequently developed; the gastritis in BALB/c mice resembles autoimmune pernicious anemia in humans. Other organs involved included the coagulating glands, prostate, and epididymus. More than one organ could be involved in a single mouse, but the disease process was completely organ-specific, with no evidence of the development of systemic autoimmune disease, antinuclear antibodies, immune complex nephritis, or manifestations of graft-versus-host disease (GVHD). Only small numbers (15 2 104) of regulatory T cells were needed to prevent disease from developing in d3Tx mice, while much larger numbers of cells (107) were used to demonstrate suppression of antibody formation (1). Autoimmune oophoritis or gastritis was transferred successfully into newborn mice or adult nu/nu mice with splenic T cells (Thy-1`, Lyt-1`, Lyt-231) obtained from d3Tx mice with disease (6, 7); both organ-specific lesions and circulating autoantibodies developed in the recipients. The effector cell was sensitive to antithymocyte serum, but resistant to cyclophosphamide treatment or in vitro X-irradiation. Later studies (8, 9) demonstrated that the effector and the suppressor T cell populations were CD4`CD8-. Although the reconstitution experiments strongly implied that disease was prevented by suppressor T cells, a major advance in this field was the demonstration (10) that removal of suppressor cells T cells from an immune system of an otherwise normal animal resulted in disease, and that reconstitution of the recipient with suppressor T cells then reestablished self-tolerance and prevented autoimmunity. When T cells from normal adult mice were separated based on the relative expression of the Lyt-1 (CD5) antigen, transfer of the Lyt-1low cells to nu/nu mice resulted in the development of autoimmune diseases in several organs. In contrast, transfer of the Lyt-1high T cells failed to result in autoimmunity and cotransfer of Thy-1`, Lyt-1high cells with Lyt-1low cells completely prevented the development of disease. Potentially pathogenic CD4` self-reactive T cells were present in the periphery of normal mice, and their activation/expansion was controlled by CD4` cell suppressor cells. It is ironic that this clear demonstration of the presence of suppressor T cells in normal mice was published at about the same time that the entire field of suppression was in a state of chaos, since molecular cloning of the TCR a- and b-chain genes ruled out the presence of the Ig determinants, which were purported to be associated with many of the soluble T cell factors. Very little progress was made in the further characterization of the suppressor T cell population involved in this model of organ-specific autoimmunity over the next 13–14 years. One problem was that the Lyt-1 antigen was actually expressed on all T cells and that the suppressor cells were contained in the large subset (80–


426

SHEVACH

90%) of cells susceptible to lysis by anti-Lyt-1 and complement. The next seminal advance in the field was again made by Sakaguchi and associates (11, 12), who further defined the suppressor cell in this model as a minor (10%) subset of CD4` T cells which coexpressed the CD25 (IL-2R a-chain) antigen. When CD4`CD251 T cells from normal BALB/c mice were transferred to 6-week-old syngeneic nu/nu recipients, the mice developed inflammatory lesions in various organs that resembled the pattern of autoimmune disease seen post-d3Tx; some recipients also developed pathologic changes consistent with systemic autoimmunity, including polyclonal B cell activation and hypergammaglobulinemia, proteinuria secondary to immune complex glomerulonephritis, and a wasting disease that clinically resembled GVHD. Most importantly, disease could be prevented by cotransfer of CD25` T cells within 10 days of transfer of the CD25- cells. In order to demonstrate that a deficiency of the same CD25` population of regulatory T cells was also responsible for the autoimmunity seen post-d3Tx, d3Tx mice were inoculated on day 10 of life with either whole spleen cells or spleen cells from which CD25` T cells had been depleted (13). Autoimmune gastritis was abolished only in those recipients that were reconstituted with CD25` T cells. A logical extension of this finding would be that CD25` T cells would not be detectable in the peripheral lymphoid tissues of unmanipulated 3day-old mice. While Asano et al (12) could not detect CD4` CD25` in the spleen of 3-day-old mice, this result is difficult to interpret, as CD3` T cells cannot be detected in the spleen until after day 3 of life. In contrast, Suri-Payer et al (14) could readily identify CD4` T cells in lymph nodes of mice as young as 2 days of age, and 10% of these CD4` T cells expressed CD25; it is possible that the CD25` cells home to the lymph node in the neonate rather than the spleen. The presence of CD25` T cells in 2-day-old animals raised the question of whether they actually were responsible for preventing autoimmunity. It remains possible that they are nonfunctional or even that they may represent effector cells that have been induced to express CD25 following recognition of their autoantigen. Alternatively, the small number of CD25` cells present on day 2 of life may not expand in the periphery after d3Tx, and their number may be too low to mediate immunoregulation. Interestingly, d3Tx resulted in an immediate increase in the percentage of CD4`CD25` cells, which reached a plateau of approximately 30% of total CD4` T cells 10 days after Tx. These cells are likely effector cells that have been very rapidly activated after d3Tx. A number of other protocols for the induction of the lymphopenic state have been described that resulted in the development of a spectrum of autoimmune diseases which closely resembles that seen post-d3Tx; presumably, all of these manipulations result in selective depletion and/or delayed development of the suppressor CD4`CD25` populations. When newborn mice were treated with cyclosporine A (CsA) during the first 7 days of life or when adult mice were treated for 2 weeks, transplantation of their thymuses to nu/nu recipients resulted in the development of organ-specific autoimmunity (15, 16). Cotransplantation of normal thymus with CsA thymus or injection of spleen cells from normal adult



427

mice prevented autoimmune disease. d7Tx of CsA-treated newborn mice markedly enhanced the development of disease; thymocytes from hydocortisonetreated mice contained functionally active regulatory T cell populations (17). Administration of CsA to euthymic newborn, but not adult, mice caused organspecific autoimmune diseases. Although CsA can selectively deplete regulatory T cells from the thymus, autoimmune disease will only develop when regulatory T cells are absent from or have not yet migrated to the periphery. High-dose (42.5 Gy) fractionated (2.5 Gy, 17 times) total lymphoid irradiation (TLI) of mice also resulted in the development of organ-specific autoimmunity (18). Radiation-induced tissue damage was not the primary cause of the autoimmune disease because irradiation of the target organs alone failed to elicit autoimmunity and shielding the organs from irradiation was unable to prevent it. Inoculation of spleen, thymocyte, or bone marrow cell suspensions within 2 weeks of TLI prevented the development of autoimmunity. In contrast to the experiments of Penhale et al (19) in the rat model (vide infra), TLI alone was more effective than Tx` TLI in inducing autoimmunity. In the mouse model, the T cells regenerating from the irradiated thymus may be selectively enriched in autoreactive effectors, which may mediate the autoimmune responses synergistically with those derived from the peripheral T cell pool. Neonatal infection with mouse T lymphotropic virus (MTLV) resulted in depletion of CD4` T cells from the thymus and periphery and the subsequent development of organ-specific autoimmunity (20). Reconstitution of the infected animals 3 weeks after neonatal virus infection with syngeneic adult CD4`CD25` T cells from noninfected mice prevented the autoimmune disease. MTLV infection of the newborn appears to selectively impair suppressor T cell function. All of the models described above for induction of autoimmunity likely result from depletion of regulatory T cells from the thymus and/or the peripheral lymphoid tissue. One of the most intriguing models for induction of autoimmunity described by Sakaguchi and associates (21) appears to involve a delayed maturation of the regulatory population. Transgenic mice that expressed any rearranged TCR a-chain transgene, but not TCR b-chain gene, under control of the IgH chain enhancer developed T cell–mediated autoimmune disease in multiple organs in a single mouse. The disease could be transferred by T cells that expressed endogenous TCR a- and b-chains. It is not known in this model how transgene expression resulted in the induction of autoreactive effector T cells. Although it is theoretically possible that expression of the transgene leads to the production of more self-reactive cells, which then overwhelmed the regulatory cells, it is more likely that transgene expression selectively inhibited or delayed the development of the suppressor population. I favor this second possibility as we (A Thornton, EM Shevach, unpublished observations) have been able to isolate CD4`CD25` T cells from adult TCR a-chain alone transgenic mice, which are potent inhibitors of T cell activation in vitro. Thus, expression of the transgene may create a window of opportunity (a transient 3dTx effect) for the self-reactive


428

SHEVACH

T cells that emerge from the thymus to initiate autoimmune damage in the absence of regulatory T cells in the periphery. For many years one of the ‘‘standard’’ protocols for the selective depletion of suppressor T cell function in vivo involved the treatment of animals with cyclophosphamide. Studies by Barrett et al (22) offer some support for this procedure. Autoimmune gastritis was induced in adult BALB/c mice by Tx at 6–8 weeks of age followed by a single dose of cyclophosphamide (300 mg/kg). This treatment transiently reduced the number of splenic T and B cells 25-fold, but by day 8 after treatment, the number of splenic T and B cells had returned to normal adult levels. Cyclophosphamide without Tx did not have any effect. The autoimmune gastritis that developed was identical to that seen post-3dTx, and it could be transferred from gastritic mice to nu/nu recipients. A more careful analysis of the susceptibility of the CD4`CD25` population to depletion by cyclophosphamide is clearly warrented. The autoimmune gastritis that develops post-d3Tx is one of the few animal models where the target antigen has been clearly defined. Circulating autoantibodies in post-d3Tx gastritis and in pernicious anemia in humans specifically react with the 95-kDa a-subunit and the 60–90 kDa b-subunit of the membrane bound proton pump, the H/K ATPase (23). To define the target antigen for CD4` T cells in d3Tx induced gastritis, Alderuccio et al (24) produced transgenic mice that expressed the b-subunit of the H/K ATPase under control of an MHC class II promoter with resultant widespread tissue expression of the b-chain. Transgenic expression prevented the production of autoantibodies and the development of gastritis following d3Tx. The incidence of oophoritis in the transgenic mice was the same as that seen in control mice. Since the transgene was expressed in the thymus, it was likely that tolerance induction occurred in the thymus. Indeed, thymocytes from normal adult BALB/c mice, but not the b-subunit transgenics, transferred autoimmune gastritis to nu/nu recipients. mRNA for the a-subunit gene was easily detectable in normal thymus (25) and transgenic expression of the a-subunit in the thymus failed to prevent the development of autoimmune gastritis. It was postulated that high-affinity T cells specific for the b-subunit could escape to the periphery and be capable of initiating autoimmune gastritis in the absence of regulatory T cells. In contrast, high-affinity T cells specific for the asubunit would be deleted in the thymus, whereas the low-affinity cells escape to the periphery. These low-affinity cells would not be activated following d3Tx but might be recruited by determinant spreading following the inflammatory response initiated by the activation of b-subunit reactive T cells by Tx; responses to the a-subunit may represent secondary events. Suri-Payer et al (26) demonstrated that H/K ATPase-reactive CD4` T cells could be identified in d3Tx mice with gastritis. The frequency of ATPase-reactive T cells was highest in the gastric lymph node since proliferative responses of lymphocytes isolated from more distal sites gave undetectable responses. More detailed analysis (14) of the fine specificity of the CD4` H/K ATPase-reactive cells demonstrated that freshly explanted gastric lymph node cells from the d3Tx



429

mice were highly reactive to the a-subunit, but only minimally reactive to the bsubunit of the H/K ATPase. Two T cell clones were isolated that recognized distinct peptides of the a-subunit. Although one clone secreted a Th1 pattern of cytokines while the other demonstrated a Th2 cytokine profile, both clones were equally potent in inducing gastritis with distinct profiles of cellular infiltration in nu/nu recipients. The capacity of either of the cell lines to induce disease was abrogated by cotransfer of CD4`CD25` T cells from normal BALB/c mice. It is unlikely that the target epitopes recognized by these two clones were immunodominant for the pathogenesis of gastritis because proliferative responses to these peptides were not observed in freshly explanted gastric lymph node cells from animals with gastritis. The preferential reactivity of the CD4` T cells in proliferation assays with the a-subunit should be contrasted with the studies in the H/K ATPase transgenic mice, where reactivity to the b-subunit played a key role in induction of disease. It remains possible the a-chain dominant proliferative responses were observed at a time when determinant spreading had already occurred or that more complex mechanisms are operative in the protection of the b-, but not a-, subunit transgenic mice from disease post-d3Tx. One of the most intriguing questions that remains to be addressed in the immunopathogenesis of autoimmunity post-d3Tx is why only certain organs are targeted for involvement by the autoimmune process. When transgenic mice, which expressed the b-subunit of the H/K ATPase in the pancreatic islets under the regulation of the rat insulin promoter, were subject to d3Tx, they developed both gastritis and insulitis (27). The peri-insulitis, however, did not progress to invasion of the islets or to diabetes. Insulitis was only observed in animals that developed gastritis. It was concluded from this study that tissue-specific factors must play a fundamental role in the development of organ-specific autoimmunity. One possibility is that the high turnover of gastric parietal cells (23-day half-life) generated a higher level of antigen presentation in the stomach than in the pancreas. Antigen-specific T cell activation would first occur in the stomach, and then the activated T cells would infiltrate the pancreas. The lack of islet cell destruction may be due to the inability of the infiltrating cells to be restimulated, lack of appropriate cytokines, or the inability to recruit CD8` T cells.

DEPLETION OF REGULATORY T CELLS FROM ADULT ANIMALS The theory that regulatory T cells control antibody production was rapidly applied to the study of autoimmunity by Penhale and colleagues (19, 28, 29), who hypothesized that antibody mediated autoimmune diseases might develop because of a failure of regulatory T cells to control autoantibody production. In a series of innovative experiments, procedures were devised to deplete the regulatory T cells and leave the helper T cell population responsible for autoantibody production


430

SHEVACH

intact. The disease model selected for study was autoimmune thyroiditis because circulating antibody to thyroglobulin was believed to play an important pathogenic role. Spontaneous thyroiditis developed in 60% of Wistar rats following the selective depletion of T cells by Tx and irradiation. Tx was performed between 3 and 5 weeks of age, and the rats were given 4–5 repeated doses of 200 rad at 14-day intervals. Circulating IgG antibody to thyroglobulin also developed, and it was assumed that the major effector cell in this model was the B cell that produced the autoantibody; no other manifestations of autoimmunity were seen. No evidence of thyroiditis was seen in rats that received only local irradiation to the thyroid region, indicating that irradiation itself did not induce pathologic changes. The conclusion drawn from these studies was that in the unmanipulated animal, B cells autoreactive with thyroid antigens were prohibited from differentiating into autoantibody producing cells by an active controlling T cell mechanism. Although not specifically tested at the time, it was assumed that the suppressor T cell population was mediating its functions by acting directly on the B cell and not by regulating other T cells. The active role of T cells in preventing the development of autoimmunity in this model was confirmed by reconstituting the Txirradiated mice with lymphoid cells from normal donors (29). Lymph node, spleen cells, or thymocytes would abrogate disease when administered intravenously shortly after the final dose of irradiation. A limited characterization of the suppressor cells with the reagents available at that the time suggested that they were T cells; they also appeared to have been activated in vivo, as they were found in the fractions containing large cells when separated on a Ficoll gradient. Taken together, these experiments demonstrated that normally autoreactive helper and suppressor cells may coexist and that certain autoimmune responses are held in check by the equilibrium favoring suppressor activity. Although Penhale and colleagues (30) also demonstrated that autoimmune diabetes would develop in a strain of rats that was normally not susceptible to this disease by the Tx-irradiation protocol, this potentially powerful experimental model for the characterization of regulatory T cells in autoimmunity was largely ignored by other investigators for the next 15 years. Powrie & Mason (31) were the first to use a combination of cell surface markers and functional differences between T cell subsets to define regulatory and effector T cell populations. In the rat, CD4` T cells could be divided into two subsets based on their differential expression of the CD45RB isoform. CD45RBhigh T cells mediated GVHD and produced IL-2 and IFN-c, but little IL4, upon polyclonal activation in vitro. In contrast, the CD45RBlow cells provided the majority of help for secondary antibody production both in vivo and in vitro and produced significant amounts of IL-4, but less IL-2 and IFN-c. Most importantly, when athymic rats were reconstituted with small numbers of CD45RBhigh T cells, they developed a severe wasting disease characterized by extensive mononuclear cell infiltration in the lungs, liver, thyroid, stomach, and pancreas 6–10 weeks later. No pathology developed in animals that received unseparated CD4`



431

T cells or CD45RBlow cells. It seemed likely from these studies that the CD45RBlow subset controlled the capacity of the CD45RBhigh subset to mediate the wasting disease. The suppressive effects of the CD45RBlow subset were directly demonstrated by Fowell & Mason (32) in the Tx-irradiation model described by Penhale’s group (19). Transfer of 5 2 106 CD45RBlow CD4` T cells completely inhibited the development of diabetes and insulitis. CD45RBlow T cells from long-term Tx donors could protect as efficiently as cells from normal donors, demonstrating that the regulatory T cell is long-lived in the periphery. When mouse T cells were separated on the basis of the differential expression of CD45RB isoforms and injected into immunodeficient SCID recipients, severe autoimmune manifestations were also observed (33–35). In contrast to the rat model in which widely dispersed pathologic lesions were observed, mice that received the CD45RBhigh subset only developed severe intestinal pathology. Mice that received CD45RBlow, unseparated CD4` cells, or a mixture of CD45RBhigh and CD45RBlow cells never developed intestinal lesions. The pathologic changes in the colon included extensive mononuclear cell infiltrates, ulceration, and pronounced epithelial cell hyperplasia. Although the early studies in the rat model suggested that the CD45RBhigh population was enriched in IFN-c producers and may represent Th1 cells, while the CD45RBlow populations contained IL-4 producers consistent with Th2 cell function, subsequent studies in the mouse model were inconsistent with this functional separation. Indeed, it appeared that, in the normal mouse, the CD45RBlow population was enriched in memory T cells, which were primed to be either IL-4 or IFN-c producers, while the CD45RBhigh subset contained the majority of naive T cells that had not yet differentiated to produce IFN-c or IL-4. Not surprisingly, when disease was induced in SCID mice by transfer of the CD45RBhigh cells, the majority of the CD4` T cells in the sick animals were found to be CD45RBlow. The role of cytokines in disease pathogenesis and the potential contribution of cytokines in protection from disease has been addressed by Powrie and colleagues (36, 37). The presence of colitis in SCID recipients of CD45RBhigh cells correlated with elevations of IFN-c, but not IL-4 or IL-10, mRNA in the involved tissues. CD4` T cells purified from colonic lesions produced 10 times more IFN-c than did CD4` T cells from normal BALB/c colons. Treatment of animals with antiIFN-c on days 1 and 14 after T cell transfer completely protected animals from disease, and protection was not dependent on the continuous presence of the antibody. Anti-IFN-c may have permanently altered a step in effector cell differentiation, possibly by inhibiting IL-12R expression (38). In contrast, weekly antiTNF treatment was required for suppression of disease; when the anti-TNF treatment was stopped at 8 weeks, severe disease was observed in mice sacrificed at 12 weeks. Mice treated with IL-10 were highly protected, but IL-10 also did not cause a long lasting modulation of the immune response, as mice sacrificed 4 weeks after the last treatment with IL-10 developed severe colitis. IL-4 treatment did not affect the induction of colitis, and none of the treatments was associated with induction of Th2 responses.


432

SHEVACH

The approach used (37) to analyze the contribution of cytokines to the protective effects of the CD45RBlow cells involved treatment of SCID mice that had been reconstituted with a mixture of CD45RB45high and CD45RBlow cells with anti-TGFb or anti-IL-10R (F Powrie, personnal communication) at the time of reconstitution, and then weekly for 6 weeks. Protection was abrogated in mice treated with either anti-TGFb or anti-IL-10R when the mice were killed on week 7. There was no evidence that IL-4 played any role in mediating protection since CD45RBlow T cells from IL-4-deficient mice were able to inhibit colitis and wasting disease. The regulatory T cell population differs from classical Th2 cells, which either would produce IL-4 or are dependent on IL-4 for their development and differentiation. A similar approach was used in the rat Tx-irrradiation model of autoimmune thyroiditis (39). Both CD45RBlow peripheral T cells as well as thymocytes will protect against disease when given immediately following the course of irradiation. When rats were reconsituted with either 106 CD4` thymocytes or 5 2 106 CD45RBlow peripheral T cells and then treated with either anti-IL-4 or anti-TGFb for 1 month after reconstitution, protection was abrogated. The potential contribution of IL-10 was not evaluated. Surprisingly, while abrogating suppression, blockade of IL-4 or TGFb did not alter the basic disease process. Collectively, these studies suggest that the regulatory T cells are an unusual T cell subset, perhaps related to Th2 or what has been termed Tr1 (40) cells. It should be noted that in both of these models, treatment is continued for many weeks, and the effects of the anti-cytokine reagents may be mediated on cells that are induced to produce these suppressor cytokines and not on the regulatory cells themselves. It is even possible that these cytokines are primarily required for differentiation of the regulatory cell populations in vivo and not for their suppressor effector functions. The requirement for IL-10 in mediating protection in the inflammatory bowel disease model may be indicative of the normal physiologic role of this cytokine in preventing inflammation in the gut since the induction of IL-12 production by bacteria may constantly stimulate Th1 responses to gut antigens. Although these recent studies suggest that certain classes of regulatory T cells mediate their protective effects by producing classic suppressor cytokines, it should also be emphasized that neither of these studies included control groups in which the animals reconstituted with only the CD45RBhigh subset, which develop disease, were also treated with anticytokine antibodies. It is quite conceivable that both the magnitude and pathology of the disease state in such animals would also have been altered, rendering a comparison with animals in which suppressor function had been abrogated problematic.

ANTIGENIC SPECIFICITY OF REGULATORY T CELLS Any approach to an analysis of the mechanism whereby regulatory T cells inhibit organ-specific autoimmunity requires an understanding of their target antigen. Very little progress has been made in this area because suppressor T cells in these



433

models have not been propagated long term in vitro or cloned. On the other hand, there exists a body of controversial data which suggests that the suppressor populations recognize a target antigen that is specific for the organ under attack. Spleen cells from normal adult male mice were much more effective suppressors of autoimmune orchitis post-3dTx than were spleen cells from female mice or male mice that had undergone a neonatal orchiectomy (41). Spleen cells from female mice were as effective as spleen cells from male mice in preventing gastritis. Studies in prostatitis suggested that the differences were relative rather than absolute, as protection could be seen with 4 2 106 spleen cells from normal males, but not females, while 4 2 107 spleen cells from females were protective. Smith et al (8) demonstrated that Con A–stimulated spleen cells from d3Tx male mice were less efficient at transferring oophoritis than were spleen cells from female mice (42). This result suggested that for T cells to initiate oophoritis they may need to be primed by endogenous ovarian antigen. In contrast to the earlier studies (41), normal male, normal female, and spleen cells from females that were oophorectomized at birth were found to suppress d3Tx oophoritis with comparable efficiency. Although these investigators used a wide range of spleen cells (5–20 2 106) per recipent, the magnitude of the suppressive effect at the lower doses used was only modest (50–60% suppression with 5 2 106 cells), and differences between male and female suppressor populations may have been missed. Taguchi et al (43) used a rather novel approach to study the tissue specificity of the regulatory T cells. The suppressor population that inhibited autoimmune prostatitis again appeared to be organ-specific since 4 2 106 spleen cells from males, but not females or orchiectomized males, inhibited disease post-d3Tx. All of these populations were equally effective in inhibiting the lacrimal gland adenitis, which also developed post-d3Tx. d3Tx mice, orchiectomized at birth, never developed prostatitis. Prostatitis developed following treatment with dihydrotestosterone, indicating that expression of the prostate antigen responsible for evoking autoimmunity could be induced by the appropriate hormonal stimulus. Most importantly, mice orchiectomized at birth, thymectomized as adults, and then treated with dihydrotestosterone developed a mature prostate. Spleen cells from these mice when inoculated into d3Tx male mice prevented the development of post-d3Tx autoimmune prostatitis. This result demonstrates that prostate-specific regulatory T cells can be activated by the induction of the organ-specific antigen. Activation of these regulatory cells takes place in the periphery since the cells could be isolated from adult Tx hormone-treated mice. It is difficult to reconcile these disparate findings on the presence of tissue-specific suppressors. All the experiments were performed in different disease models involving autoimmune damage to reproductive organs; it is possible that some tissue-specific antigens are not gender-specific, and this may account for the differences between the results with orchitis and oophoritis. Future studies should employ purified populations of CD4`CD25` cells rather than unseparated spleen cells and should include very careful dose-response studies.


434

SHEVACH

The term subtractive was initially used by McCullagh (44) to categorize protocols in which the control of the development of tolerance is investigated by preventing the developing immune system from gaining access to an antigen or to the developing organ until self/non-self discrimination had been established. When the fetal rat thyroid gland was destroyed by exposure to radioactive iodine and syngeneic thyroid tissue was implanted into the athyroid rats as adults, autoimmune thyroiditis developed in the grafted tissue. The autoimmune attack was organ-specific, as other endocrine organs did not exhibit signs of inflammation. The development of thyroiditis in rats that had been exposed to radioactive iodine in utero could be prevented by parabiosis to normal syngeneic partners (44). Parabiosis was only protective if instituted at the time of thyroid grafting, but not 1–2 weeks after graft implantation. Normal rats possess migratory regulatory cells capable of blocking the antithyroid immune response of rats in which thyroid development had been disrupted before the development of immunocompetence. These studies on ‘‘subtraction’’ of regulatory T cells by removal of the target organ during fetal life are quite consistent with the results described earlier on the failure to detect regulatory T cells in animals in which the reproductive organs were extirpated in the neonatal period. A critical time period of exposure is therefore needed to allow for the development of organ-specific suppressor T cells. Seddon & Mason (45) used the approach developed by McCullagh (44) to demonstrate that peripheral CD45RBlow T cells from rats rendered athyroid were unable to prevent thyroid-specific autoimmunity induced by the adult Txirradiation protocol. The loss of regulatory cells was specific for the extirpated organ, as T cells from the athyroid rats could prevent the development of diabetes. In contrast, CD4`CD8- thymocytes from the same athyroid donors were as effective as those from normal rats at preventing thyroiditis. Regulatory T cells were generated normally in the thymus in the absence of their target organ, but the target organ was needed for survival and/or expansion.

IN VITRO MODELS OF SUPPRESSOR T CELL FUNCTION There is little doubt that the immunoregulatory T cells described in this review have potent inhibitory functions in vivo. As the model systems described frequently require weeks to months of assessment of disease activity, it has been difficult to determine the mechanism of action, antigen specificity, or cellular targets of the suppressor T cell populations. A number of recent studies (46, 47) have demonstrated that CD4`CD25` T cells are potent inhibitors of polyclonal T cell activation in vitro. Purified CD4`CD25` were completely nonresponsive to stimulation by TCR-derived signals. In one study (46), the CD25` cells remained unresponsive even when costimulatory signals were provided by addition of anti-CD28, while in another study (47), a modest response of the



435

CD4`CD25` T cells was seen in the presence of anti-CD28. Most importantly, the CD4`CD25` cells could adoptively suppress the responses of CD4`CD251 cells in coculture studies (Figure 1). Suppression appeared to be mediated by a cell contact–dependent mechanism because suppression was not seen when the suppressors were separated from the responders by a semipermeable membrane, and supernatants from activated suppressors could not mediate suppression. Neutralizing antisuppressor cytokine antibodies (anti-IL-4, -10, TGF-b) alone, or in any combination, also failed to reverse the suppression. CD4`CD25` T cells from IL-4-/- or IL-10-/- mice were as effective suppressors in vitro as CD4`CD25` T cells from wild-type mice. This result should be compared with the effectiveness of these reagents in reversing the suppressive effects of CD45RBlow cells in vivo (37, 39). Suppression in vitro required that the suppressor population be activated via the TCR. The antigen-specific proliferative response of TCR transgenic T cells was not suppressed by the CD4`CD25` T cells, whereas the responses of the same cell mixture to anti-CD3 stimulation were completely suppressed. No evidence was obtained that the suppression mediated by the CD4`CD25` population was mediated by killing of the responder population by Fas/Fas-L– dependent mechanisms. The CD4`CD25` T cells would respond to the combination of anti-CD3 ` IL-2, and suppression in coculture studies could be overcome by the addition of IL-2 or by the enhancement of endogenous IL-2 production by anti-CD28. Nevertheless, the CD4`CD25` population did not mediate suppression by binding and/or consuming IL-2 (functioning as an ‘‘IL2 sink’’), as the suppressor cells completely inhibited IL-2 gene transcription and IL-2 production in the responder T cell population. No significant differences in the VaVb T cell repertoire between the CD25` and CD25- populations could be defined, indicating that the CD25` population is not mono- or oligoclonal (47). The CD25` T cells did not express NK1.1, and they appear to be distinct from Figure 1 CD4`CD25` T cells are powerful suppressors of polyclonal T cell activation in vitro. CD4`CD25- T cells (5 x 104) were stimulated with soluble antiCD3, T-depleted spleen cells, and different numbers of CD4`CD25` suppressor cells. Almost complete suppression was observed when 12.5 x 104 CD4`CD25` cells were added (suppressor/responder 4 1/4).


436

SHEVACH

NK T cells. When compared with CD4`CD251 T cells, the CD25` cells were similar in their expression of CD5, had a slightly higher proportion of CD62Llow cells, and had a modest increase in the number of CD69` cells (5% –15%). They had an unusual pattern of expression of CD45RB; they lacked the CD45RBhigh population and were composed primarily CD45RBint and CD45RBlow cells. Although these results suggested that the CD25` population might be heterogeneous and composed of a mixture of suppressor T cells and conventional memory T cells, we (47a) have been unable to identify a subpopulation of CD25` T cells that failed to manifest potent suppressor activity in vitro. We have, therefore, concluded that the CD4`CD25` population represents a relatively homogeneous population of suppressor T cells. While CD25` T cells from normal mice could not suppress antigen-specific responses, CD4`CD25` T cells isolated from TCR transgenic mice on a conventional background could readily suppress the proliferative responses of TCR transgenic CD4`CD25- T cells from the same mice. When TCR transgenic mice were bred to SCID mice and mice were selected that only expressed the transgenic TCR, the number of CD4`CD25` T cells was decreased by .90% (13). Expression of the endogenous TCR a-chain was therefore required to generate the CD4`CD25` population in TCR transgenic mice on a conventional background. One must therefore conclude that the true specificity of the CD25` population is defined by the endogeous receptor and not by the transgenic TCR. It is only fortuitous that ‘‘antigen-specific’’ suppression can be mediated by the transgene expressing CD4`CD25` T cells. Once activated by antigen, suppressor effector function was completely antigen nonspecific. This was best illustrated when CD4`CD25` T cells were propagated in culture by stimulation with anti-CD3 in the presence of IL-2. After 7 days of culture, the activated CD25` cells remain anergic to restimulation with anti-CD3 in the absence of IL-2; moreover, they had enhanced suppressor activity when cultured with fresh CD4`CD25- T cells and anti-CD3. More importantly, the activated CD25` population suppressed antigenspecific responses of a wide variety of CD25- T cell populations derived from different TCR transgenic mice. There was no requirement for MHC restriction or antigen recognition for this suppression to occur (47a). Even though these studies on the in vitro functions of the CD4`CD25` cells reveal a potent suppressor activity, both the target cell and the mechanism by which suppression is mediated remain elusive. Some observations strongly suggested that the suppressor population acted on the APC. The CD25` cells readily suppressed the APC-dependent response to soluble anti-CD3, but not the response to the relatively APC-independent stimulus, plate-bound anti-CD3. Furthermore, some studies suggested that the suppressors may inhibit the generation or delivery of costimulatory signals, as suppression could be overcome either by the addition of exogenous IL-2 or by the generation of endogenous IL-2 by costimulation with anti-CD28. So far we have been unable to overcome suppression by addition of large numbers of fully activated APC, nor have we observed inhibition of expression of costimulatory molecules when APC were cultured in the presence of the



437

suppressor cells. Suppressor T cells functioned normally when fixed, activated APC were used in the cultures. Could the responder T cell be the target for suppression? This certainly remains an attractive possibility, but none of the studies by our group or others revealed a possible mechanism of such an activity. Although Fas/Fas ligand interactions do not play a role in suppression, it is possible that other known or unknown members of the TNF/TNFR family may be mediating the inhibition of cell growth in this system. Of course, it is still possible that we are dealing with a relatively labile, unknown, soluble suppressor factor. What is the relationship between the CD4`CD25` suppressor T cell population that functions in vivo and in vitro in a suppressor cytokine-independent manner and the CD45RBlow T cells that mediate suppression of autoimmunity in vivo by secreting suppressor cytokines? It has recently been demonstrated (48) that the CD45RBlow population can be subdivided based on expression of the CD38 antigen; approximately 50% of the CD45RBlow cells express CD38. The CD38` population is about 50% CD25` and behaves in a similar fashion in vitro, i.e. it is anergic and suppressive. The CD381 population contained antigenprimed cells that were capable of mounting vigorous secondary responses in vitro and in vivo. Previous studies had shown that CD45RBlow T cells were poor responders to stimulation via the TCR. Although these studies were interpreted as indicating that memory cells were more difficult to activate than naive cells, perhaps secondary to CD4-mediated downregulatory signals, it is much more likely that the poor responses of the CD45RBlow cells to stimulation with antiCD3 were secondary to active suppression mediated by the CD38` cells in that population. Both CD25`CD38` and CD25`CD381 T cells are anergic and suppressive (47a), so it remains possible that all of the suppressive activity of the CD38` population is mediated by the CD25` cells. Both the CD38` and CD381 subpopulations of the RBlow pool were capable of inhibiting colitis in vivo. The suppressive capacity of CD25-CD38` cells has not yet been evaluated in vivo or in vitro. It remains possible that multiple functional lineages of suppressor cells may be identified based on differential expression of these markers. It is puzzling that the in vitro suppressor activity of the CD38` cells was suppressor cytokineindependent, while their in vivo activity was cytokine-mediated.

REGULATORY T CELLS IN OTHER AUTOIMMUNE DISEASES Although this review has focused on autoimmune diseases that develop in animals from which regulatory T cells have been depleted, a considerable body of experimental data has been accumulated over the past 10 years which suggests that regulatory T cells are involved in almost all experimental animals models of autoimmunity. In the NOD mouse model of diabetes, a number of observations are compatible with a requirement for diminished regulatory T cell function for


438

SHEVACH

transfer of disease. Diabetes could only be transferred from sick mice to normal syngeneic recipients if the recipients were fewer than 5 weeks of age (female) or 3 weeks of age (male). Furthermore, cyclophosphamide was required for induction of diabetes in young male as well as female NOD mice. Most importantly, CD4` T cells from nondiabetic NOD mice could prevent the transfer of diabetes from overtly diabetic mice into sublethally irradiated NOD recipients (49, 50). The protective cell population was not present in the spleen until 3 weeks of age and reached its highest activity at 8 weeks of age; suppressor cells were present in the thymus of neonates, which may explain why thymectomy at weaning accelerates disease. Diabetes could also be efficiently transferred to non-irradiated adult NOD recipients if they were Tx and CD4` T cell depleted. Depletion of CD4` T cells alone was not sufficient for disease transfer, and it was likely that Tx was needed to limit re-expansion of the CD4` regulatory T cells (51). Islet-infiltrating T cells from young nondiabetic mice could transfer diabetes to NOD/SCID mice, but cotransfer of CD4`CD45RBlow splenic T cells from the same mouse delayed the onset of disease (52). CD4`CD45RBlow T cells from overtly diabetic mice transferred disease and produced IFN-c, while the protective CD4`CD45RBlow cells produced a Th2 or Th0 cytokine profile. There is very little data as to the nature of the target antigen recognized by the regulatory T cells in NOD mice. Both CD4` (53, 54) and CD8` (55) T cell clones with suppressive activities have been isolated, but the antigens recognized by these clones have been very poorly characterized. In some cases the clones appeared to be autoreactive with self-MHC class II molecules (54), while in other cases both reactive to components in fetal calf serum and specifically reactive with islet cells (53). These clones produced a variety of cytokines including several uncharacterized suppressor factors, while other clones mediated their suppressive effects by producing TGFb (54). Although NK-like T cells have not been shown to play a regulatory role in the other models of autoimmunity described here, Gombert et al (56) demonstrated that NOD mice had a deficit in NK T cells, which was first seen at 3 weeks of age and persisted until 8 weeks of age. Both NK T cells in the thymus and the spleen lacked the ability to produce IL-4. It appears that NK T cells emerge from the thymus later in life in NOD mice, and this may explain why thymectomy at 3 weeks of age aggravates disease. NK T cells contained within the population of double negative thymocytes were capable of preventing the spontaneous onset of diabetes when transferred into prediabetic recipients (57). The protective effect of the transferred NK T cells was mediated by IL-4 and IL-10 because neutralization of IL-4 and IL-10 during the first week after NK T cell transfer inhibited protection by these cells. These studies have been interpreted as demonstrating that a deficiency of NK T cells in NOD mice contributes to the pathogenesis of IDDM by permitting the development of pathogenic Th1 effector cells. Preliminary studies suggest that a similar defect may be seen in human diabetes, as diabetic siblings had lower frequencies of CD4-CD8- NK T cells that expressed the Va24JaQ TCR (58). NK T cell clones from the diabetic twins secreted only



439

IFN-c upon stimulation, while clones from the at risk nonprogressor siblings and normals secreted both IL-4 and IFN-c. As in the mouse, the loss of the capacity of NK T cells to produce IL-4 appears to be the major factor regulating disease susceptibility. Clearly, more studies are needed in both animal models and human to validate this hypothesis. Furthermore, it is not clear how the target antigens recognized by the NK T cells play a role in the pathogenesis of diabetes. While one might have predicted that mice which expressed a transgenic TCR specific for an autoantigen would rapidly develop autoimmune disease, a number of studies have demonstrated that only a small percentage of mice which express a TCR specific for an autoantigen develop disease. Transgenic mice that expressed the a-and b-chains of an MBP-specific TCR exhibited only a low incidence of spontaneous EAE, which developed after 1 year of age. When T cells from these mice were stimulated in vitro they were not anergic; immunization of the transgenics with MBP peptide also resulted in the rapid development of EAE. When crossed to RAG-1 –/– mice, all the mice rapidly developed EAE spontaneously. It is likely that the resistance of mice that expressed the transgenic TCR on a conventional background was mediated by the low proportion of T cells expressing TCR encoded by the endogenous a- and b- TCR genes, which are not present in the RAG –/– mice (59, 60). When CD4` T cells from nontransgenic mice were transferred into the 3-week-old TCR transgenic RAG –/– mice, EAE onset was delayed, severity diminished, and the animals recovered from the disease. Recipients older than 45 days were less susceptible to protection. Crosses of TCR transgenic RAG-1 –/– mice with mice deficient in B cells, CD8` T cell, NK T cells, c/d T cells, or a/b T cells indicated that a/b`, CD4` T cells were the only cell population capable of mediating protection. Susceptibility of the mice to EAE correlated inversely with the diversity of their T cell repertoire. The nature of the target antigen recognized by the regulatory T cells in this model and the relationship of the regulatory T cells to the other types of regulatory cells described in this review remain to be determined. A similar protective effect of T cells that expressed receptors encoded by endogenous a- and b-chain genes was observed in mice that expressed a transgenic TCR specific for an pancreatic islet cell antigen (61).

REGULATORY T CELLS ARE GENERATED IN THE THYMUS In both the 3dTx model of induction of autoimmunity and the Tx-irradiation models, regulatory CD4` thymocytes can protect in reconstitution studies. In fact, CD4` thymocytes were considerably more protective in the rat Tx-irradiation diabetes model than were peripheral CD4` T cells (62). While fewer than 5 2 106 peripheral CD4` cells were not protective, significant protection was seen with 0.6 2 106 CD4` thymocytes. This simple finding raises a number of important questions about the potential antigens (s) that might be recognized by these


440

SHEVACH

cells. It does not appear that regulatory T cells are generated by the antigen-driven expansion of a limited number of thymocyte precursors in the periphery. It remains possible that tissue-specific antigens may be expressed in thymus and that the regulatory T cells may be generated by recognition of such antigens on thymic epithelium, but the studies in the athyroid rat model (45) demonstrate that maintenance/expansion of the regulatory T cells also required the presence of the relevant autoantigen in the periphery as well as in the thymus. Papiernik et al (63) were the first to demonstrate that CD4`CD25` T cells originate in the thymus and are induced to express CD25 at the CD4` single positive stage. The CD4`CD25` single positive thymocytes were not derived from CD25` double positive cells. Approximately 5% of CD4` thymocytes express CD25, while less than 0.3% of CD8` thymocytes express CD25 (64). The phenotype of the CD4`CD25` thymocyte resembles that of the peripheral CD4`CD25` T cells in that there is an enhanced expression of membrane activation markers. The capacity of CD25` T cells to migrate from the thymus to the periphery was studied after intrathymic injection of FITC. The percentage of CD25` cells within migrants and resident T cells was identical, suggesting that CD25` cells in the periphery can originate in the thymus. CD4`CD25` thymocytes have the capacity for peripheral expansion, as shown by the transfer of CD4` cells to nu/nu recipients. Most importantly, CD4`CD25` T cells were absent from the periphery and from the CD4` single positive thymocyte pool of IL-2–deficient mice. This finding indicates that IL-2 itself is important for the generation of these regulatory cells in the thymus and possibly for their maintenance/expansion in the periphery. It is still unclear how the expression of CD25 relates to this requirement for IL-2 stimulation. The thymocyte population also functionally resembled the CD4`CD25` population found in the periphery since the CD4`CD25` thymocytes were both anergic and suppressive. Nu/nu mice reconstituted with CD25-depleted adult thymocytes developed a wider spectrum of disease than that seen after d3Tx, including the development of signs of systemic autoimmunity. Although studies of thymocyte differentiation over the past 10–15 years have offered major insights to the processes of positive and negative selection, it is still difficult to develop a model for the generation of regulatory T cells, particularly since their antigenic specificity is so ill defined. A number of animal models are now available that permit one to take a reductionist approach to the creation of models that might explain regulatory T cell differentiation in the thymus. The simplest view, based on the selective expression of the CD25 antigen at the CD4` single positive stage, is that regulatory T cells are generated during the process of negative selection. If the process of negative selection was inhibited, regulatory T cells would not be induced and autoimmunity might result. When the Keratin 14 promoter was used to re-express a class II MHC antigen in class II negative mice, the transgenic molecule was expressed only on thymic cortical epithelium; thymic medullary epithelium and bone marrow–derived cells were MHC class II



441

negative (65). Such mice lacked MHC class II expression on the critical cell types, which have been postulated to mediate negative selection. When CD4` T cells from these mice were cultured in vitro, they exhibited an enhanced autoreactivity, as measured by an elevated proliferative response in the syngeneic MLR and the generation of CD4` MHC class II–restricted cytolytic cells. These mice showed no evidence of autoimmunity since they did not express MHC class II antigens on their peripheral APC; however, when CD4` T cells from the mice were transferred into lethally irradiated syngeneic C57BL/6 mice, they induced acute graftversus-host disease with bone marrow failure (66). Furthermore, these autoreactive CD4` T cells caused hypergammaglobulinemia and the production of autoantibodies when transferred into unirradiated C57BL/6 hosts. It cannot be determined from these studies if the autoreactivity resulted from the failure of deletion of high-affinity autoreactive T cells, from the failure to generate immunoregulatory cells, or a combination of both. It would be of interest to examine these mice as well as H2-M–deficient mice, which also exhibit a defect in the negative selection process (67), for the presence of CD25` anergic/suppressive T cells. Could regulatory T cells be generated during the process of positive selection on thymic epithelium? Modigliani et al (68) have characterized a complex model in which an embryonic thymic rudiment from a day 10 fetus was transplanted onto a nu/nu recipient. This type of thymic epithelial (TE) graft is devoid of hematopoietic cells and was colonized by the host’s hematopoietic cells with restoration of the T cell compartment. When TE was grafted to an allogeneic recipient, tolerance was induced to a variety of peripheral tissues of donor type that expressed nonthymic antigens (skin and heart). Transfer of high numbers of T cells from tolerant animals to athymic nu/nu recipients resulted in maintenance of the tolerant state in the adoptive host; the regulatory T cell that transferred tolerance was shown to be a CD4` T cell. The mechanisms whereby such regulatory T cells would suppress effector cells that recognized tissue-specific antigens have not yet been determined. This model system supports the view that the generation of regulatory T cells might occur at stages of the T cell differentiation process other than negative selection. It is also possible that autoimmunity may result from defects in both positive and negative selection. Ridgway and Fathman (69, 70) have demonstrated that immunization of NOD mice with self-peptides resulted in an immune response to the self-peptide, with resultant autoproliferation of peripheral lymphocytes. NOD mice retained the capacity to respond normally to foreign peptides while demonstrating an abnormal response to self-peptides. The induction of autoreactivity was only seen in NOD mice, and homozygous expression of I-Ag7 was required. It was proposed that the poor peptide binding properties of I-Ag7 molecules were responsible for defective positive selection. Inefficient positive selection would require an increased affinity of the T cells to attain the requisite avidity needed for selection. These high-affinity T cells would then enter the periphery, and in collaboration with other genes they would mediate autoimmunity once an

442

SHEVACH

inflammatory event broke self-tolerance. This model does not address the altered generation of regulatory T cells, but it is likely that the process of differentiation of regulatory T cells would also be abnormal and suboptimal.


CONCLUDING REMARKS Taken together, the potential contribution of disordered thymic selection to the generation of regulatory T cell function is a confusing, but important, area for future study. It is clear that manipulations of thymic architecture will likely result in the generation of both altered autoimmune effector cell function and in defects in regulatory T cell function. The evidence presented here strongly suggests that the regulatory CD4` cell lineage is unique and is a normal product of T cell differentiation in the thymus. I would like to propose a model to explain the generation of regulatory T cells under normal physiologic conditions. This model is heavily biased by the concept that the process of positive selection in the thymus is much less stringent that the process of negative selection; furthermore, selfantigens, which are expressed in or transported from the periphery to the thymus, play a very minor role in positive selection, but a critical role in negative selection. The experimental evidence to support this view is derived from experiments with MBP-deficient mice (71, 72). Such mice can generate a vigorous high-avidity T cell response to MBP, while the wild-type control strain is poorly responsive to this protein. First, it is clear that endogenous MBP is not required for positive selection. Second, the major function of self-antigens in the thymus is to induce clonal deletion of high-avidity T cells and to permit only the export from the thymus of low-avidity MBP-reactive T cells. How then are regulatory T cells generated in the thymus? As is demonstrable in most immunologic systems, ‘‘all or none’’ phenomena are rare. I would propose that the process of negative selection may have outcomes other than simple clonal deletion or passage through to the periphery. Studies (73) with mature peripheral T cells and T cell clones have demonstrated that engagement of the TCR by socalled ‘‘altered-peptide ligands’’ may result in a permanent change in the effector functions of the clones. If the fit between the self-peptide/MHC complex and the TCR during negative selection is not of sufficient avidity to result in clonal deletion, yet not weak enough to allow the autoreactive T cells to pass through to the periphery, the outcome may fall into what we would term‘‘altered negative selection’’ (Figure 2). This process would result in a permanent change in the capacity of the self-reactive T cells to signal through the TCR. Cells surviving this process would leave the thymus in an incapacitated state: 1. They would not be able to differentiate into Th1 cells capable of producing organ-specific autoimmunity, 2. they might resemble the CD4`CD25` cells and be completely anergic, 3. they might alternatively resemble the CD45RBlow populations and produce suppressor cytokines, or 4. they may be unable to express membrane molecules which are critical for their differentiation into pathogenic Th1 effectors, such as the CD40



443

Figure 2 The ‘‘altered negative selection model’’ for the generation of suppressor T cells. Suppressor T cells with a variety of functional phenotypes would be generated during the process of negative selection in the thymus because the fit between their TCR and the selfpeptide/MHC complex is not optimal and results in a permanent change in the signaling capacity of their TCR. The precise mechanism by which they inhibit the activation of lowaffinity anti-self effector cells on the surface of dendritic cells in the target organ remains to be defined.

ligand (38). A number of different functional phenotypes might be created that would account for the diversity of suppressor effector mechanisms described in this review.


444

SHEVACH

What is the antigenic specificity of the regulatory T cells? As none of the studies we have described are really compatible with a model where the effectors and suppressors compete for the same peptide determinant of an antigen, I would favor the possibility that the suppressor populations are either specific for other peptide determinants on the target autoantigen or are specific for determinants on other peptides derived from proteins of the target organ. Once the pathologic process has been initiated by autoreactive effector cells, the organ-specific suppressors would home to the target organ, be activated by their target peptide, and mediate suppression. They might even be activated by the same APC that activates the effector cell, so a process of ‘‘linked-suppression’’ (74) would result. Although it is easiest to invoke a suppressor effector function that is mediated by a cocktail of suppressor cytokines (IL-4, -10, -13, and TGF-b), all the data on the CD4`CD25` population suggest that a novel cell contact–dependent mechanism of effector T cell inactivation also plays a role. Do regulatory T cells exist in humans and do they play a role in the pathogenesis of autoimmunity? Thus far, almost no information is available on the existence of subpopulations of CD4` T cells in humans that have any of the characteristics of the CD4`CD25` or CD4`CD45RBlow cells that have been described in the mouse. We therefore focus on the possible implications of the findings in rodents to the pathogenesis and treatment of disease in humans. The studies by Sakaguchi et al (75) have emphasized that one potential drawback of a unique lineage of regulatory T cells is that they may be susceptible to a variety of environmental insults or genetic abnormalities. For example, while immunosuppressive therapy may be used to reduce or eliminate activated T cells, such treatment may actually induce autoimmune disease by also depleting regulatory T cells. Depletion of suppressor cells by different agents (drugs, radiation, infection, etc) might lead to development of autoimmune disease in the same organ system in a susceptible individual. The particular disease that would develop would be determined by the genetic background of the host. Although the focus of this review has been on the role of regulatory T cells in controlling autoimmunity, it should be emphasized that regulatory T cells must have a much broader role in controlling immune responses. Do they also inhibit immune responses to foreign antigens? Our bias is that the majority of foreign antigen-specific T cells have much higher affinities than those autoreactive T cells which recognize organ-specific antigens; hence foreign antigen-reactive populations will be much less susceptible to inhibition. Needless to say, some members of the pool of foreign antigen–reactive T cells may have lower affinity receptors, and those cells may be susceptible to down-regulation by regulatory T cells. The most important class of antigens the response to which will be regulated by suppressor T cells will be the class of autoantigens involved in tumor immunity. Temporary depletion of regulatory T cell function might facilitate vaccination protocols to both tumor-specific antigens and weakly immunogenic pathogens. Lastly, an extension of this concept would involve attempts to enhance regulatory T cell function in patients with autoimmunity or in recipients of allografts.


445


Although most of the animal studies suggest that the regulatory T cells will be less effective as inhibitors of primed or activated T cells, they may still be capable of inhibiting relapses of chronic diseases by blocking epitope spreading and sensitization of new effectors. In a timely, but scathing, editorial in 1988, G. Moller (76) summarized his reasons for questioning the existence of suppressor T cells. In 2000, there is little doubt that suppressor T cells play a critical role in regulating autoimmune disease in a large number of animal models. Although markers are now available that allow enrichment of suppressor T cell populations, they are still imperfect because all are also expressed on other cell types. More importantly, most of the properties of suppressor T cells that were poorly characterized in 1988—antigen-specificity, MHC restriction, frequency of reactive cells, mechanism of action, and cellular targets of suppression—remain so today. All of these topics should be the subject of fruitful investigations in the future. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Gershon RK, Kondo K. 1970. Cell interactions in the induction of tolerance: the role of thymic lymphocytes. Immunology 18:723–35 2. Green DR, Webb DR. 1993. Saying the ‘‘S’’ word in public. Immunol. Today 14:523–25 3. Nishizuka Y, Sakakura T. 1969. Thymus and reproduction: sex-linked dysgenesia of the gonad after neonatal thymectomy in mice. Science 166:753–55 4. Kojima A, Tanaka-Kojima Y, Sakakura T, Nishizuka Y. 1976. Prevention of postthymectomy autoimmune thyroiditis in mice. Lab. Invest. 34:601–5 5. Kojima A, Taguchi O, Nishizuka Y. 1980. Experimental production of possible autoimmune gastritis followed by macrocytic anemia in athymic nude mice. Lab. Invest. 42:387–95 6. Sakaguchi S, Takahashi T, Nishizuka Y. 1982. Study on cellular events in postthymectomy autoimmune oophoritis in mice. I. Requirement of Lyt-1 effector cells for oocyte damage after adoptive transfer. J. Exp. Med. 156:1565–76

7. Sakaguchi S, Takahashi T, Nishizuka Y. 1982. Study on cellular events in postthymectomy autoimmune oophoritis in mice. II. Requirement of Lyt-1 cells in normal female mice for the prevention of oophoritis. J. Exp. Med. 156:1577–86 8. Smith H, Sakamoto Y, Kasai K, Tung KSK. 1991. Effector and regulatory cells in autoimmune oophoritis elicited by neonatal thymectomy. J. Immunol. 147: 2928–33 9. Smith H, Lou Y-H, Lacy P, Tung KSK. 1992. Tolerance mechanism in experimental ovarian and gastric autoimmune diseases. J. Immunol. 149:2212–18 10. Sakaguchi S, Fukuma K, Kuribayashi K, Masuda T. 1985. Organ-specific autoimmune diseases induced in mice by elimination of T cell subset. I. Evidence for the active participation of T cells in natural self-tolerance; deficit of a T cell subset as a possible cause of autoimmune disease. J. Exp. Med. 161:72–87 11. Sakaguchi S, Sakaguchi N, Asano M, Itoh M, Toda M. 1995. Immunologic self-tolerance maintained by activated T

446

12.

13.


14.

15.

16.

17.

18.

19.

20.

SHEVACH cells expressing IL-2 receptor a-chains (CD25). J. Immunol. 155:1151–64 Asano M, Toda M, Sakaguchi N, Sakaguchi S. 1996. Autoimmune disease as a consquence of developmental abnormality of a T cell subpopulation. J. Exp. Med. 184:387–96 Suri-Payer E, Amar AZ, Thornton AM, Shevach EM. 1998. CD4`CD25` T cells inhibit both the induction and effector function of autoreactive T cells and represent a unique lineage of immunoregulatory cells. J. Immunol. 160:1212– 18 Suri-Payer E, Amar AZ, McHugh R, Natarajan K, Margulies DH, Shevach EM. 1999. Post-thymectomy autoimmune gastritis: fine specificity and pathogenicity of anti-H/K ATPase-reactive T cells. Eur. J. Immunol. 29:669–77 Sakaguchi S, Sakaguchi N. 1988. Thymus and autoimmunity. Transplantation of the thymus from cyclosporin A-treated mice causes organ-specific autoimmune disease in athymic nude mice. J. Exp. Med. 167:1479–85 Sakaguchi S, Sakaguchi N. 1989. Organspecific autoimmune disease induced in mice by elimination of T cell subsets. J. Immunol. 142:471–80 Sakaguchi S, Sakaguchi N. 1990. Thymus and autoimmunity: capacity of the normal thymus to produce pathogenic self-reactive T cells and conditions required for their induction of autoimmune disease. J. Exp. Med. 172:537–45 Sakaguchi N, Miyai K, Sakaguchi S. 1994. Ionizing radiation and autoimmunity. Induction of autoimmune disease in mice by high dose fractionated total lymphoid irradiation and its prevention by inoculating normal T cells. J. Immunol. 152:2586–95 Penhale WJ, Farmer A, McKenna RP, Irvine WJ. 1973. Spontaneous thyroiditis in thymectomized and irradiated wistar rats. Clin. exp. Med. 15:225–36 Morse SS, Sakaguchi N, Sakaguchi S.

21.

22.

23.

24.

25.

26.

27.

28.

1999. Virus and autoimmunity: induction of autoimmune disease in mice by mouse T lymphotropic virus (MTLV) destroying CD4` T cells. J. Immunol. 162: 5309–16 Sakaguchi S, Ermak TH, Toda M, Berg LJ, Ho W, Fazekas de St. Groth B, Peterson PA, Sakaguchi N, Davis MM. 1994. Induction of autoimmune disease in mice by germline alteration of the T cell receptor gene expression. J. Immunol. 152: 1471–84 Barrett SP, Toh BH, Alderuccio F, van Driel I, Gleeson PA. 1995. Organspecific autoimmunity induced by adult thymectomy and cyclophosphamideinduced lymphopenia. Eur. J. Immunol. 25:238–44 Gleeson PA, Toh B-H, van Driel I. 1996. Organ-specific autoimmunity induced by lymphopenia. Immunol. Rev. 149:97–126 Alderuccio F, Toh B-H, Tan S-S, Gleeson PA, van Driel IR. 1993. An autoimmune disease with multiple molecular targets abrogated by the transgenic expression of a single autoantigen in the thymus. J. Exp. Med. 178:419–26 Alderuccio F, Gleeson PA, Berzins SP, Martin M, van Driel IR, Toh B-H. 1997. Expression of the gastric H/K-ATP-asubunit in the thymus may explain the dominant role of the b-subunit in the pathogenesis of autoimmune gastritis. Autoimmunity 25:167–75 Suri-Payer E, Kehn PJ, Cheever AW, Shevach EM. 1996. Pathogenesis of post-thymectomy autoimmune gastritis. J. Immunol. 157:1799–1805 Barrett SP, van Driel IR, Tan S-S, Toh BH, Gleeson PA. 1995. Expression of a gastric autoantigen in pancreatic islets results in non-destructive insulitis after neonatal thymectomy. Eur. J. Immunol. 25:2686–94 Penhale WJ, Farmer A, Irvine WJ. 1975. Thyroiditis in T cell-depleted rats: influence of strain, radiation dose, adjuvants


29.

30.


31.

32.

33.

34.

35.

36.

and antilymphocyte serum. Clin. exp. Immunol. 21:362–75 Penhale WJ, Irvine WJ, Inglis JR, Farmer A. 1976. Thyroiditis in T cell-depleted rats: suppression of the autoallergic response by reconstitution with normal lymphoid cells. Clin. Exp. Immunol. 25:6–16 Penhale WJ, Stumbles PA, Huxtable CR, Sutherland RJ, Pethick DW. 1990. Induction of diabetes in PVG/c strain rats by manipulation of the immune system. Autoimmunity 7:169–79 Powrie F, Mason D. 1990. OX-22high CD4` T cells induce wasting disease with multiple organ pathology: prevention by the OX-22low subset. J. Exp. Med. 172:1701–8 Fowell D, Mason D. 1993. Evidence that the T cell repertoire of normal rats contains cells with the potential to cause diabetes. Characterization of the CD4` T cell subset that inhibits this autoimmune potential. J. Exp. Med. 177:627–36 Powrie F, Correa-Oliveira R, Mauze S, Coffman RL. 1994. Regulatory interactions between CD45RBhigh and CD45RBlow CD4` T cells are important for the balance between protective and pathogenic cell-mediated immunity. J. Exp. Med. 179:589–600 Powrie F, Leach MW, Mauze S, Caddle LB, Coffman RL. 1993. Phenotypically distinct subsets of CD4` T cells induce or protect from chronic intestinal inflammation in C.B-17 scid mice. Int. Immunol. 5:1461–71 Morrissey PJ, Charrier K, Braddy S, Liggitt D, Watson JD. 1993. CD4` T cells that express high levels of CD45RB induce wasting disease when transferred into congenic severe combined immunodeficient mice. Disease development is prevented by cotransfer of purified CD4` T cells. J. Exp. Med. 178:237–44 Powrie F, Leach MW, Mauze S, Menon S, Caddle LB, Coffman RL. 1994. Inhibition of Th1 responses prevents inflam-

37.

38.

39.

40.

41.

42.

43.

44.

45.

447

matory bowel disease in scid mice reconstituted with CD45RBhi CD4` T cells. Immunity 1:553–62 Powrie F, Carlino J, Leach MW, Mauze S, Coffman RL. 1996. A critical role for transforming growth factor-b but not interleukin 4 in the suppression of T helper type 1-mediated colitis by CD45RBlow CD4` T cells. J. Exp. Med. 183:2669–74 Chang JT, Shevach EM, Segal BM. 1999. Regulation of interleukin (IL)-12 receptor subunit expression by endogneous IL-12: a critical step in the differentiation of pathogenic autoreactive T cells. J. Exp. Med. 189:969–78 Seddon B, Mason D. 1999. Regulatory T cells in the control of autoimmunity: the essential role of transforming growth factor b and interleukin 4 in the prevention of autoimmune thyroiditis in rats by peripheral CD4` CD45RC- cells and CD4`CD8- thymocytes. J. Exp. Med. 189:279–88 Groux H, O’Garra A, Bigler M, Rouleau M, Antonenko S, de Vries JE, Roncarolo MG. 1997. A CD4` T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature 389:737–41 Taguchi O, Nishizuka Y. 1981. Experimental autoimmune orchitis after neonatal thymectomy in the mouse. Clin. exp. Immunol. 46:425–34 Taguchi O, Nishizuka Y. 1987. Self tolerance and localized autoimmunity. Mouse models of autoimmune disease that suggest tissue-specific suppressor T cells are involved in self tolerance. J. Exp. Med. 165:146–56 Taguchi O, Kontani K, Ikeda H, Kezuka T, Takeuchi M, Takahashi T. 1994. Tissue-specific suppressor T cells involved in self-tolerance are activated extrathymically by self-antigens. Immunology 82:365–69 McCullagh P. 1996. The significance of immune suppression in normal self tolerance. Immunol. Rev. 149:127–54 Seddon B, Mason D. 1999. Peripheral

448

46.

47.


47a.

48.

49.

50.

51.

52.

53.

SHEVACH autoantigen induces regulatory T cells that prevent autoimmunity. J. Exp. Med. 189:877–81 Thornton AM, Shevach EM. 1998. CD4` immunoregulatory T cells suppress polyclonal T cell activation in vitro by inhibiting interleukin 2 production. J. Exp. Med. 188:287–96 Takahashi T, Kuniyasu Y, Toda M, Sakaguchi N, Itoh M, Iwata M, Shimizu J, Sakaguchi S. 1998. Immunologic selftolerance maintained by CD25` CD4` naturally anergic and suppressive T cells: induction of autoimmune disease by breaking their anergic/suppressive state. Int. Immunol. 10:1969–80 Thornton AM, Shevach EM. 1999. Suppressor effector function of CD4` CD25` immunoregulatory T cells is antigen non-specific. J. Immunol. In press Read S, Mauze S, Asseman C, Bean A, Coffman R, Powrie F. 1998. CD38` CD45RBlow CD4` T cells: a population of T cells with immune regulatory activities in vitro. Eur. J. Immunol. 28:3435– 47 Boitard C, Yasunami R, Dardenne M, Bach JF. 1989. T cell-mediated inhibition of the transfer of autoimmune diabetes in NOD mice. J. Exp. Med. 169:1669–80 Hutchings PR, Cooke A. 1990. The transfer of autoimmune diabetes in NOD mice can be inhibited or accelerated by distinct cell populations present in normal splenocytes taken from young males. J. Autoimmun. 3:175–85 Sempe P, Richard M-F, Bach J-F, Boitard C. 1994. Evidence of CD4` regulatory T cells in the non-obese diabetic male mouse. Diabetologia 37:337–43 Shimada A, Charlton B, Rohane P, Taylor-Edwards C, Fathman CG. 1996. Immune regulation in type 1 diabetes. J. Autoimmun. 9:263–69 Akhtar I, Gold JP, Pan L-Y, Ferrara JLM, Yang X-D, Kim JI, Tan K-N. 1995. CD4` b islet cell-reactive T cell clones

54.

55.

56.

57.

58.

59.

60.

that suppress autoimmune diabetes in nonobese diabetic mice. J. Exp. Med. 182:87–97 Han H-S, Jun H-S, Utsugi T, Yoon J-W. 1996. A new type of CD4` suppressor T cell completely prevents spontaneous autoimmune diabetes and recurrent diabetes in syngeneic islet-transplanted NOD mice. J. Autoimmun. 9:331–39 Pankewycz O, Strom TB, Rubin-Kelley VE. 1991. Islet-infiltrating T cell clones from non-obese diabetic mice that promote or prevent accelerated onset diabetes. Eur. J. Immunol. 21:873–79 Gombert J-M, Herbelin A, TancredeBohin E, Dy M, Carnaud C, Bach J-F. 1996. Early quantitative and functional deficiency of NK1`-like thymocytes in the NOD mouse. Eur. J. Immunol. 26:2989–98 Hammond KJL, Poulton LD, Palmisano LJ, Silveira PA, Godfrey DI, Baxter AG. 1998. a/b-T cell receptor (TCR)` CD41 CD81 (NKT) thymocytes prevent insulin-dependent diabetes mellitus in nonobese diabetic (NOD)/Lt mice by the influence of interleukin (IL)-4 and/or IL10. J. Exp. Med. 187:1047–56 Wilson SB, Kent SC, Patton KT, Orban T, Jackson RA, Exley M, Porcelli S, Schatz DA, Atkinson MA, Balk SP, Strominger JL, Hafler DA. 1998. Extreme Th1 bias of invariant Va24JaQ T cells in type 1 diabetes. Nature 391:177–81 Olivares-Villagomez D, Wang Y, Lafaille JJ. 1998. Regulatory CD4` T cells expressing endogenous T cell receptor chains protect myelin basic protein-specific transgenic mice from spontaneous autoimmune encephalomyelitis. J. Exp. Med. 188:1883–94 Van de Keere F, Tonegawa S. 1998. CD4` cells prevent spontaneous experimental autoimmune encephalomyelitis in anti-myelin basic protein T cell recep-


61.

62.


63.

64.

65.

66.

67.

tor transgenic mice. J. Exp. Med. 10:1875–82 Luhder F, Katz J, Benoist C, Mathis D. 1998. Major histocompatibility complex class II molecules can protect from diabetes by positively selecting T cells with additional specificities. J. Exp. Med. 187:379–87 Saoudi A, Seddon B, Fowell D, Mason D. 1996. The thymus contains a high frequency of cells that prevent autoimmune diabetes on transfer into prediabetic recipients. J. Exp. Med. 184:2393–98 Papiernik M, Leite de Moraes M, Pontoux C, Vasseur F, Penit C. 1998. Regulatory CD4 T cells: expression of IL-2Ra chain, resistance to clonal deletion and IL-2 dependency. Int. Immunol. 10:371– 78 Itoh M, Takahashi T, Sakaguchi N, Kuniyasu Y, Shimizu J, Otsuka F, Sakaguchi S. 1999. Thymus and autoimmunity: production of CD25` CD4` naturally anergic and suppressive T cells as a key function of the thymus in maintaining immunologic self-tolerance. J. Immunol. 162:5317–26 Laufer TM, DeKoning J, Markowitz JS, Lo D, Glimcher LH. 1996. Unopposed positive selection and autoreactivity in mice expressing class II MHC only on thymic cortex. Nature 383:81–85 Laufer TM, Fan L, Glimcher LH. 1999. Self-reactive T cells selected on thymic cortical epithelium are polyclonal and are pathogenic in vivo. J. Immunol. 162: 5078–84 Grubin CE, Kovats S, deRoos P, Rudensky AY. 1997. Deficient positive selection of CD4 T cells in mice displaying altered repertoires of MHC class II-

68.

69.

70.

71.

72.

73.

74.

75.

76.

449

bound self-peptides. Immunity 7:197– 208 Modigliani Y, Bandeira A, Coutinho A. 1996. A model for developmentally acquired thymus-dependent tolerance to central and peripheral antigens. Immunol. Rev. 149:155–74 Ridgway WM, Fasso M, Lanctot A, Garvey C, Fathman CG. 1996. Breaking self-tolerance in nonobese diabetic mice. J. Exp. Med. 183:1657–62 Ridgway WM, Ito H, Fasso M, Yu C, Fathman CG. 1998. Analysis of the role of variations of major histocompatibility complex class II expression on nonobese diabetic (NOD) peripheral T cell response. J. Exp. Med. 188:2267–75 Targoni OS, Lehmann PV. 1998. Endogenous myelin basic protein inactivates the high avidity T cell repertoire. J. Exp. Med. 187:2055–63 Harrington CJ, Paez A, Hunkapiller T, Mannikko V, Brabb T, Ahearn M, Beeson C, Goverman J. 1998. Differential tolerance is induced in T cells recognizing distinct epitopes of myelin basic protein. Immunity 8:571–80 Sloan-Lancaster J, Allen PM. 1996. Altered peptide ligand-induced partial T cell activation: molecular mechanisms and role in T cell biology. Annu. Rev. Immunol. 14:1–27 Frasca G, Carmichael R, Lechler R, Lombardi G. 1997. Anergic T cells effect linked suppression. Eur. J. Immunol. 27:3191–97 Sakaguchi S, Toda M, Asano M, Itoh M, Morse SS, Sakaguchi N. 1996. T cellmediated maintenance of natural selftolerance: its breakdown as a possible cause of various autoimmune diseases. J. Autoimmun. 9:211–20 Moller G. 1988. Do suppressor T cells exist? Scand. J. Immunol. 27:247–50



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


SIGNALING AND TRANSCRIPTION IN T HELPER DEVELOPMENT Kenneth M. Murphy, Wenjun Ouyang, J. David Farrar, Jianfei Yang, Sheila Ranganath, Helene Asnagli, Maryam Afkarian, and Theresa L. Murphy Department of Pathology, and Howard Hughes Medical Institute, Washington University School of Medicine, St. Louis, Missouri, 63110; e-mail: [email protected]


Key Words Th subsets, cytokines, signaling, transcription, gene expression Abstract The recognition of polarized T cell subsets defined by cytokine production was followed by a search to define the factors controlling this phenomenon. Suitable in vitro systems allowed the development of cytokine ‘‘recipes’’ that induced rapid polarization of naı¨ve T cells into Th1 or Th2 populations. The next phase of work over the past several years has begun to define the intracellular processes set into motion during Th1/Th2 development, particularly by the strongly polarizing cytokines IL-12 and IL-4. Although somewhat incomplete, what has emerged is a richly detailed tapestry of signaling and transcription, controlling an important T cell developmental switch. In addition several new mediators of control have emerged, including IL-18, the intriguing Th2-selective T1/ST2 product, and heterogeneity in dendritic cells capable of directing cytokine-independent Th development.

OVERVIEW The developmental regulation of T helper responses during infection is critically important to the form and effectiveness of acquired immunity. Over the past five years, tremendous progress has been made in understanding the signaling pathways and transcriptional mechanisms that guide the differentiation of T helper phenotypes from naı¨ve precursors. While additional patterns of cytokines are beginning to emerge, studies of classical Th1 and Th2 subtypes, as originally defined by Mosmann & Coffman (1), have benefited from robust in vitro developmental systems. These systems have allowed application of molecular analysis during the developmental process of naı¨ve to committed T cells, which has revealed both satisfying regulatory paradigms and some challenging observations. This review focuses on the progress primarily in signaling and transcriptional regulation in Th1/Th2 development since the last general Annual Review chapters on this topic (2, 3). 0732–0582/00/0410–0451$14.00

451

452

MURPHY ET AL


REGULATION OF IFN-c PRODUCTION IN TH1 DEVELOPMENT Among various cytokines produced by Th1 cells, IFN-c perhaps most defines this subset in terms of its selective expression and functional properties. Studies of IFN-c regulation predating recognition of the Th1/Th2 subsets mapped specific DNase I-hypersensitive sites and attributed T cell–specific expression to an 8.6Kb genomic region (4, 5) that directed tissue-specific expression in transgenic mice (6). In a 500-bp IFN-c1inducible promoter region (7, 8), two elements were identified, a distal consensus GATA motif, and a proximal region homologous to NFIL-2A of the IL-2 promoter (9). GATA-3 was proposed to augment expression via the distal element (9), but at that time it was not known that it is selectively expressed in Th2, not Th1 cells. Flavell and colleagues analyzed the proximal and distal elements in transgenic reporter mice (10). Interestingly, both proximal and distal elements directed reporter expression in CD4` T cells, but in CD8` T cells, only the distal element appeared active. The cAMP response element binding protein-activation transcription factor-1 (CREB-ATF1) (11) interacted with these regulatory elements, although whether this factor mediated the Th1-selective expression of IFN-c was not certain (12). The NFAT and NF-jB transcription factors are thought to regulate IFN-c gene expression (13–16). Young & colleagues showed that c-Rel interacts with ciselements within the IFN-c promoter (14). However, NFAT and NF-jB transcription factors can bind to similar sequences in vitro, requiring careful functional analysis. NFAT sites in the IFN-c promoter were identified (15), termed P1 and P2, and they showed interaction with NFAT factors using specific supershift analysis. Further, the P2 site was shown to exert enhancer activity and to mediate cyclosporin sensitivity in reporter assays (15). Young and colleagues examined P2 (C3–3P) (16), finding that NF-jB proteins were also capable of functional interactions via this element, suggesting possible coordinate actions of NFAT and NF-jB in overlapping cis elements. More recently, two strong NFAT binding sites were identified using DNase footprint analysis (13), arguing that IFN-c inducibility might be directed by signals either from NFAT or NF-jB factors. Agarwal & Rao (17) extended earlier descriptions of hypersensitive regions within the IFN-c gene, describing acquisition of specific DNase-hypersensitive sites during the development of naive T cells toward IFN-c-producing Th1 cells, although the factors interacting at these sites were not characterized. The MAP kinase pathway also appears involved in IFN-c production in CD4` T cells (18–20). First, inhibitors of p38 MAPK and dominant-negative p38 selectively impaired Th1 responses (18). Mice with a targeted disruption of Mkk3 (21), an upstream activator of p38, have diminished Th1 responses, primarily from decreased IL-12 production, but they show an impaired IL-12-induced Th1 development, implying a role in T cells as well. A potential role for the c-Jun NH2-terminal kinase (JNK) pathway was implied by the enhanced Th2 bias


T HELPER DEVELOPMENT

453

observed in JNK1-deficient T cells (20). JNK1-deficient T cells showed enhanced IL-4 production, but similar levels of IFN-c production when T cells were activated under conditions in vitro that did not drive either toward Th1 or Th2 phenotypes, which implies that JNK1 may not be necessary for IFN-c production. However, in JNK2-deficient T cells, a more substantial reduction in IFN-c production was observed (19). One mechanism for this decreased IFN-c production was lower levels of IL-12Rb2 expression in JNK2-deficient T cells, rather than perhaps a direct requirement for JNK2 expression for IFN-c promoter activation. Members of the IRF-1 family have also been suggested to participate in Th1 development and IFN-c production (22, 23). Two studies examined IRF-1deficient mice, each describing defects in IL-12 production, but with somewhat differing interpretations of T cell effects. IRF-1-deficient CD4` T cells developed normal Th1 responses when IL-12 was provided (23). However, using IRF-1deficient TCR-transgenic T cells, another study (22) found significantly diminished IFN-c production under IL-12-driven in vitro conditions. Thus, IRF-1 may act not only in IL-12 transactivation, but also directly in Th1 development. Recently, these results were extended (24) by demonstrating that IL-12 signaling in T cells induces IRF-1 expression, apparently through activation of Stat4. Thus, IRF-1 expression may be downstream of IL-12, but direct actions in the IFN-c promoter are not clear. Several transcription factors have been found to be selectively expressed in Th1 cells. An IL-12-induced Ets family transcription factor, ERM, is selectively induced by IL-12 via Stat4 activation in Th1 cells (25). While Ets-1 is important for the development of natural killer (NK) cells in mice, another important physiological source of IFN-c production in vivo (26), forced expression of ERM, did not satisfy the stringent requirement of restoring IFN-c production in Stat4deficient T cells (25). However, recently another Th1-specific transcription factor T-bet, a novel member of the T brachiury family of transcription factors, was identified as Th1 specific and could strongly activate IFN-c production when expressed in T cells developing under neutral conditions, and even when expressed in Th2 cells (SJ Szabo, JI Kim, L Glimcher, unpublished observations). The class II MHC transactivator (CIITA) is selectively expressed in Th1 cells and represses IL-4 but apparently does not directly transactivate IFN-c (27). Finally, the homeodomain transcription factor Hlx was found to be Th1 specific, with transgenic overexpression enhancing Th1-type responses (RA Flavell, unpublished observations) (Figure 1). A potential negative regulatory element in the IFN-c promoter was found and shown to bind to the nuclear factor Yin-Yang 1 (YY1) (28). Subsequently, two more YY1 binding sites were identified (29), one overlapping an AP1 site, suggesting repression by YY1 could occur either through cooperation with an unidentified repressor protein or by competition with AP1 for positive transactivation. Disruption of the YY1 cis-element within the IFN-c promoter failed to significantly enhance or diminish reporter activity (13), suggesting this factor either is not active in the cells studied or may play a positive rather than

MURPHY ET AL


454

Figure 1 Pathways and transcription in IFN-c production by Th1 cells. Th1-specific components (asterisks) include the receptors for IL-12 and IL-18 and recently identified transcription factors. A number of putative cis-acting elements are shown, but none are known to interact with Th1-specific transcription factors. Stat4 activation is restricted to Th1 cells, but alone does not transactivate the IFN-c gene.



455

negative role in IFN-c production. Certain other factors inhibit Th1 development but may not directly inhibit IFN-c gene transcription. GATA-3 inhibits Th1 development, independent of its positive actions for Th2 development (30), perhaps through inhibition of IL-12Rb2 expression, which could also occur in IL-4deficient and Stat6-deficient T cells. Further, Glimcher and colleagues reported that the Th2-specific factor c-maf (31) also attenuates Th1 differentiation by both IL-4-dependent and -independent mechanisms (32). DNA methylation is potentially important for IFN-c regulation. Differential methylation at sites in the IFN-c promoter was found between Th1 and Th2 cells (33), but this study did not distinguish whether this difference caused or resulted from differential gene expression. In analyzing human B cell lines, Pang et al (34) described hypomethylation in the IFN-c promoter in B cell lines which both expressed and did not express IFN-c. In CD8` T cells, a correlation was shown between methylation of the IFN-c promoter and IFN-c production (35). Demethylation of sites -205 , -191, and -53 correlated well with IFN-c expression between naive and effector CD8 T cells.

ROLE OF STAT4 IN TH1 DEVELOPMENT The Jak/STAT signaling pathway has been recently reviewed (36, 37). Th1 development involves signaling through Stat4, initially cloned by homology to other STAT family members (38) but later recognized to mediate IL-12 signaling (39). Stat4 activation was correlated with the capacity to promote IFN-c production, based on the loss of Stat4 activation by IL-12 in Th2 cells (40), but proof of its role in Th1 development came from analysis of targeted Stat4 deficiency in mice (41, 42). Not all IFN-c production appears to depend on Stat4 (43, 44). In Stat6/ Stat4 double deficient mice, IFN-c production was observed, indicating production by a Stat4-independent mechanism (43). Further, CD8` T cells exhibit Stat4independent TCR-induced IFN-c production (44). However, the specific transcriptional targets of Stat4 involved in mediating the process of Th1 development were not evident from these early studies. Hoey and colleagues (45) suggested that activated Stat4 acts directly on the IFN-c gene to augment transcription via cooperative binding to nonconsensus low-affinity STAT sites within both the promoter and first intron. However, fully differentiated Th1 cells are capable of production of IFN-c at high levels in the absence of Stat4 activation (25, 46). Also, Sinigaglia and colleagues showed that Stat4 activation by IFN-a in human T cells was not sufficient for the direct activation of the IFN-c gene (47), suggesting additional requirements. Conceivably, the STAT sites in the IFN-c gene may not be important for either Th1 development or acute IFN-c production. Other factors that are induced or repressed by Stat4 may act to affect IFN-c expression. Signaling through IL-12 receptor also activates Stat1 and Stat3 in addition to Stat4, and yet neither Stat1 nor Stat3 induces Th1 development. Thus, this private


456

MURPHY ET AL

action of Stat4 is not solely a result of IL-12R signaling specificity but could result from the recognition by Stat4 of unique target sequences in the genome, or through selective interactions with coactivators in a context-dependent or -independent manner. Several STAT factors have been shown to interact with the general transcriptional coactivators p300/CBP (48–50). A direct interaction of Stat4 with p300/CBP has not been reported but seems plausible. The crystal structure of Stat1 lacking the amino-terminal domain (51) suggested additional potential interaction sites for STAT proteins with other transcription factors. A coiled-coil domain consisting predominantly of hydrophilic surfaces was suggested to provide a convenient interface for potential STAT-specific protein interactions (51). Interestingly, the coiled-coil domain of Stat5b was used in a two-hybrid screen to identify interactions with the protein Nmi (52), a protein of previous unknown function. Nmi interaction was generalizable to all of the STATs except Stat2 and enhanced coassociation of the coactivators p300/CBP, thereby providing a general coactivation mechanism. Higher order interactions of Stat4 were observed in binding adjacent but nonconsensus and low-affinity STAT sites in the IFN-c first intron (45). The capacity for tetramer interaction with adjacent nonconsensus STAT sites provides a mechanism that could allow selective target activation via Stat4 differentially from other STAT factors, such as Stat1, that bind similar sequences as dimers. Recently, Kuriyan and colleagues (53) provided a structural basis for such a mechanism. The Stat4 amino-terminus crystallized as a dimer, with extensive interactions along one face that could mediate association of two Stat4 dimers into a tetramer. However, the interaction of the amino-terminus to form STAT tetramers is not unique to Stat4. The amino-terminus of Stat1 was also shown to mediate interactions between two dimers and to augment interactions with adjacent STAT targets in DNA (54). The Stat5 amino-terminus may also mediate tetramer formation important for Stat5 transactivation via adjacent cis-acting elements within the IL-2Ra gene, in both murine (55) and human (56) systems. In summary, although Stat4 contains a number of properties shared with other STAT factors, the basis for the private physiological actions of Stat4 is not yet entirely clear.

ROLE OF TYPE I INTERFERONS IN TH1 RESPONSES Although over the past several years much of the control of Th1 development in humans and mice has been attributed to IL-12, a number of early reports also attributed control of Th1 development in humans to type I interferons, particularly IFN-a. As early as 1992, human T cells activated with IFN-a and c were seen to favor the development of Th1 rather than Th2 cytokine profiles (57). These results were independently confirmed (58) again in human CD4` T cells in an anti-CD3 rather than antigen-specific system. Very recently, the direct action of IFN-a on human T cell Th1 development was confirmed by Sinigaglia and colleagues (59). Here, priming of human T cells in the presence of IFN-a markedly



457

enhanced IFN-c production, even in the absence of IL-12. This excludes the possibility that the action of Type-I interferon may be indirect through bystander induction of IL-12. Further, IFN-a was able to induce Th1 development even in the presence of IL-4. This highlights the potential for IFN-a to act through receptors that are retained during Th2 development, when IL-12 receptors are extinguished. These actions of IFN-a in human T cells contrast with the inability of IFN-a in the murine system to induce Th1 development (60). Very recently these results were confirmed (47) by directly comparing human and murine systems in the same study. Again, in the human, but not the murine system, IFN-a/b was clearly capable of driving Th1 development, bypassing the need for IL-12 signaling. The basis of selective IFN-a induction of Th1 development in human T cells may be explained by differences in IFN-a signaling between the murine and human IFN-a receptors. Stat4 was first suspected to be important in Th1 development by its identification in murine T cells as being uniquely activated by IL12 (39); this was confirmed by O’Shea and colleagues for the human system, where IL-12 signaling also activated Stat4 (61). Interestingly, whereas in the mouse system IL-12 activated Stat1, Stat3, and Stat4 (39), in human T cells IL12 activated only Stat4, but not Stat1 and Stat2 (61). It, therefore, came as a surprise when O’Shea reported activation of Stat4 by IFN-a in the human system (62). This observation initially was interpreted to indicate that Stat4 may not be involved in Th1 development, since it was not universally accepted that IFN-a drove Th1 development, a bias predominantly emerging from the murine system. However, in direct comparisons of human and mouse T cell development, both IL-12 and IFN-a activated Stat4 and induced Th1 development in human T cells, but only IL-12 exerts these effects in murine T cells (47). Thus, while in the murine system, the role of Stat4 and Th1 development has been confirmed by knockout experiments (41, 42), expectations would now define Stat4 as important for Th1 development in human CD4 T cells, and also as able to undergo activation by IFN-a and IL-12. No Stat4 mutations in humans have been reported to confirm its role in Th1 development. Whereas IL-12 receptor expression appears to be modulated, the constitutive expression of IFN-a receptors on human T cells potentially provides a constant capacity for response to interferons for promoting Th1-type responses to various pathogens under many circumstances.

Species-Specific Differences in IFN-a Receptor Signaling While the critical role of IFN-a signaling for protection against viral pathogens has been confirmed by a targeted mutation of the interferon receptor in mice (63), most of the structural studies of the IFN-a signaling pathway have been carried out in the human system (64). Both human and mouse IFN-a receptors are composed of two subunits, the IFNAR1 (64, 65) and IFNAR2 (66–68). The human IFNAR1 and IFNAR2 subunits are significantly dissimilar in primary sequence compared to their murine homologues (69, 70), which explains the sequence


458

MURPHY ET AL

specificity of interferon actions on human versus mouse cells. IFN-a1induced ligation of the IFNAR1 and IFNAR2 chains activates Jak1 (71) and Tyk2 (72– 74). Activation of STAT proteins involves the recruitment of Stat2 to the IFNR1 subunit (74, 75). Stat2 bound to the R1 receptor then serves as a docking site for the recruitment of Stat1 to the receptor complex (76). Tyrosine 690 of human Stat2 undergoes phosphorylation and serves as a binding site for the recruitment of Stat1, causing a hierarchy of dependence for Stat1 activation on Stat2 for activation by IFN-a signaling (77). Stat3 appears to undergo direct recruitment to a separate region of the IFNAR1 receptor, which is highly conserved between the human and mouse receptors (78, 79). To date, no publications have described the basis for differential Stat4 activation by human and mouse interferons. Several possibilities exist, including differences in the primary sequences of the human and mouse IFNAR1 and IFNAR2 subunits and differences in docking involved in STAT recruitment. It is not yet known whether Stat4 is directly recruited to the cytoplasmic domains of either the IFNAR1 or IFNAR2 receptor, or whether Stat4 is recruited to Stat2, as is Stat1, or potentially to Stat3. Finally, since Stat3 (79) and CrkL (80) can each act as adaptor molecules coupled to downstream signaling pathways, potentially, species-specific differences in adaptor molecules could mediate Stat4 activation in human, but not mouse.

ROLE OF IL-1 FAMILY CYTOKINES AND SIGNALING TH1/TH2 DEVELOPMENT While IL-12 and IL-4 have been considered dominant factors for inducing Th1 or Th2 development following primary T cell activation, other factors are also recognized to contribute. Differential actions of IL-1 on T cell subsets have long been recognized (81, 82). Also, O’Garra and colleagues recognized the costimulatory role played by IL-1 for promotion of strong Th1 phenotypes in BALB/c, but not B.10, background T cells (83). This recognition foreshadowed identification of additional IL-1 family members that can significantly contribute to the biological regulation of Th1 and Th2 cells. First, Okamura & colleagues identified a novel factor, initially called interferon-c-inducing factor (IGIF), on the basis of its ability to augment IFN-c production from Th1 clones and natural killer (NK) cells (84, 85). Subsequent analyses (86) recognized that IGIF was a new member of the IL-1 cytokine family, suggesting the term IL-1c, although IGIF was eventually renamed IL-18. Subsequent studies showed close linkage between the mechanism of IL-18 and IL-1 production, signaling, and mechanisms of action. Two reports showed that deficiency in the enzyme IL-1b converting enzyme (ICE) exhibited defects in IGIF (IL-18) production (87, 88). The importance of ICE-dependent processing of IL-18 in vivo was subsequently confirmed by comparison of antigen-dependent



459

IFN-c production in wild-type and ICE-deficient mice (89). However, regulation of IL-18 and IL-1b may exhibit subtle differences in terms of gene expression, synthesis, and processing between human and murine cells, implying additional methods of regulation to be determined by future studies (90). After its initial discovery, IL-18 (IGIF) was shown to be a selective activator of IFN-c in Th1 but not Th2 cells. First, a powerful synergy between IL-12 and IL-18 for IFN-c production was found (91). Moreover, the effects of IL-18 as a costimulator for IFN-c were found to be selectively targeted toward Th1 cells, but not Th2 cells (92). An important role for IL-18 in vivo was shown by targeted disruption of the IL-18 gene (93). IL-18-deficient mice were found to have a significant reduction in LPS-induced IFN-c production following priming with Propionibacterium acnes. Further, in crossing IL-18-deficient mice to IL-12-deficient mice, the partial deficiency in IFN-c production found in either strain was exaggerated to generate a much more severe defect in IFN-c production in the double cytokinedeficient mice. While IL-18 was clearly important for the strength of IFN-c production in responses in vivo, its mechanism of action in T cell development remained unclear. This issue was clarified by the demonstration that IL-18 does not induce development of Th1 cells but acts subsequent to IL-12-induced Th1 development, to augment levels of IFN-c produced by differentiated Th1 cells (94). Finally, additional effector functions of IL-18 have recently been reported by Young et al (95), who showed IL-18 as a potent activator of IL-13 production in natural killer (NK) cells and T cells. Molecular mechanisms of IL-1 receptor signaling have been defined recently in great detail. First, Cao & Gao identified a protein kinase IRAK as a component of the IL-1R signaling pathway (96). IRAK was shown to be recruited to the IL1R complex through interactions with the IL-1R accessory protein (IL-1RAcP) (97). Further, IL-1RAcP is essential for IL-1-induced IRAK activation (98). A second kinase, IRAK-2, and a death domain–containing adaptor molecule, MyD88, were identified as proximal components of IL-1 signaling (99). MyD88 is recruited to the IL-1R complex following IL-1 stimulation (100), perhaps recruiting downstream signaling molecules such as TRAF6, which may function as a signal transducer for NF-jB activation (101). While a precise understanding of the role of IL-1 in regulating Th1/Th2 responses is not yet clear, some connection is implied by two recent findings: IL-1R-deficient mice showed enhanced Th2 responses in vivo (102), and so implied an IL-4-independent autocrine pathway for Th2 proliferation in which IL-1 may serve as a mitogen (103). Subsequently, IL-1 and IL-18 signaling pathways were shown to share similar components. First, IL-18 was demonstrated to activate NF-jB in Th1 cells (94, 104) and to activate IRAK (94). In addition, a potential for IL-18 induction of a MAP kinase pathway has been reported (105). An orphan member of the IL-1R family, IL-1Rrp, was initially identified (106), but its function or relationship to IL-1 cytokine action was not initially clear. Through biochemical purification of authentic IL-18 receptors (107), the previous IL-1Rrp was found to be identical


460

MURPHY ET AL

to one chain of the human IL-18R complex. However, IL-1Rrp alone did not comprise a functional IL-18R, implying the need for a second component (108). Sims and colleagues identified a second protein, termed AcPL (109), which, together with the previously identified IL-1Rrp, can mediate IL-18 activation of NF-jB and JNK. Thus, a functional IL-18R complex can be composed of IL-1Rrp and AcPL. IL-18 also activates TRAF6 similarly to IL-1 (110), and MyD88-deficient mice were shown to lose not only IL-1-induced function, but also IL-18-mediated functions (111). IRAK-deficient mice were recently shown to have defective IL18-mediated natural killer (NK) and Th1-type responses to pathogens in vivo (112). The role for IL-1Rrp in authentic IL-18 signaling was shown by targeted deletion of this protein in vivo (113). IL-1Rrp-deficient mice lacked activation of NF-jB and c-Jun N-terminal kinase (JNK) in response to IL-18 treatment of Th1 cells. While TCR signaling was considered the exclusive stimulus for IFN-c production by CD4` T cells, recently O’Garra and colleagues uncovered a second pathway for induction (94). The cytokines IL-12 and IL-18 together induced full IFN-c production by Th1 cells, independent of stimulus through the TCR. The synergistic actions of IL-12 and IL-18 on the IFN-c promoter were proposed to be mediated by a Stat4 binding site and an adjacent AP-1 binding site (114). IL12 was argued to contribute via a Stat4 binding element, and IL-18 by activation of AP-1 (114). However, the elements proposed to bind Stat4 are poorly conserved between human and mouse promoters (45), and the functional data supporting the roles of these elements require additional study. Further, IL-18 activated the IRAK/NF-jB pathway (94), but this study found no evidence for IL-18 activation of AP-1 in CD4` T cells. Notably, in the latter study, using primary human CD4` T cells (114), IL-18 alone could induce IFNc, whereas in the former study (94), both IL-12 and IL-18 were needed for IFNc production. A subsequent study (46) confirmed the findings of O’Garra and further showed there was clear distinction in the pathways and transcription factors activated by TCR and IL-12/IL-18 stimulation. This study suggested that the IFN-c promoter may be composed of distinct sets of cis-acting elements, a composite promoter, that undergo transactivation by two signaling pathways using two distinct sets of factors (46). The basis for synergy between IL-12 and IL-18 is not yet clear. IL-12 treatment of T cells can increase expression of the IL-18R (115, 116), suggesting one potential basis for synergy. For example, Liew and colleagues (117) reported that IL18 receptors become selectively expressed on Th1 but not Th2 cells, and they suggested that, reciprocally, IL-18 may augment IL-12R expression. However, resting Th1 cells express functional IL-12 responses without IL-18 treatment, and conversely express functional IL-18 responses without IL-12 treatment (46). Rather, it appears that there are two modes of activation for IFN-c in Th1 cells, TCR and IL-12/IL-18 signaling, which use distinct pathways and perhaps different transcription factors (46). For example, activation of IFN-c by IL-18 was



461

recently found to involve activation of NF-jB acting at discrete cis-acting elements in the IFN-c regulatory region (118). The definition of the specific transcription factors and cis-acting elements responding to this and the TCR pathway are subjects for additional work. While IL-18 and IL-18R augment Th1 responses, another IL-1R family member was discovered to be selectively expressed on Th2 cells. First identified as a serum- and oncoprotein-induced gene, designated T1 (119), the T1/ST2 gene was later established as a member of the IL-1R family (120) through demonstration of its structural and functional properties. First, comparisons between T1/ST2 and other IL-1R homologues, such as Toll and 18-Wheeler, revealed clusters of sequence similarity (120). In addition, chimeric receptors expressing the putative signaling domain of T1/ST2 indicated its capacity to signal similarly to IL-1, for example, by its ability to activate the NF-jB signaling pathway. Subsequently, selective expression of the T1/ST2 gene on Th2 but not Th1-type helper T cells was demonstrated (121). In addition, a strong correlation between expression of T1/ST2 and IL-4 production suggested T1/ST2 as a stable marker for Th2 cells in vivo. Subsequently, the Th2 selective expression of T1/ST2 was confirmed by Kamradt and colleagues, who showed that antibodies against T1/ST2 were able to attenuate the eosinophilic inflammation in airways and suppress Th2 cytokines in vivo following adoptive transfer of Th2 cells (122). In addition, T1/ST2 was highly upregulated on CD4` T cells following infection by Schistosoma mansoni, and its surface expression correlated with production of type 2 cytokines ex vivo (123). To date, a single report has suggested the cloning of a putative ligand for the T1/ST2 receptor (124), but no additional studies have confirmed this result.

IL-12 GENE REGULATION While the mechanism for Stat4 directing Th1 development is not resolved, it is clear that IL-12 is a critical mediator of Th1 development in both murine (125– 127) and human systems (128–130), and therefore IL-12 production is of some importance. IL-12, composed of p40 and p35 subunits, is largely induced in macrophages and dendritic cells by transcriptional induction of the p40 subunit (131). The IL-12 p40 subunit of IL-12 was initially recognized to be controlled through proximal cis-acting elements interacting with NF-jB family members (132). While multiple pathways converge on NF-jB activation, IL-12 production optimally requires priming of monocytes by IFN-c, which may involve other ciselements, including sites for IRF-1, NF-IL6, and Ets-2 (133, 134). Also, the IFN consensus sequence binding protein (ICSBP) participates in Th1 development (135), at least partly through regulating IL-12 production (136–138). Further, IFN-c may directly induce expression of the IL-12 p35 gene (133). The genomic organization and chromosomal assignments for both the p35 and p40 subunits have been characterized, with sequence scanning suggesting some additional putative regulatory elements (139, 140).


462

MURPHY ET AL

Additional analysis of the human p40 promoter suggested a requirement for an Ets-2 binding site located approximately 100 bp upstream from the NF-jB site (141), although this element was not required in the murine system (132, 142). While the basis for this difference has not been clarified, it may involve the presence of an Ets-related factor binding site immediately adjacent to the NF-jB site in the murine system, which is not conserved in the human sequence (132). Smale and colleagues further analyzed the proximal promoter region to demonstrate the presence of C/EBP protein interactions immediately downstream of the NF-jB target site in the p40 promoter (142). Additionally, the Ets transcription factors may synergistically interact with NF-jB for p40 promoter activation (143). Bacterial LPS can activate NF-jB for p40 induction, but another important pathway involves CD40-CD40 ligand (CD40L) interaction (144–148), which may be of particular importance for IL-12 production in dendritic cells. Inhibitors of IL-12 induction have also been reported, including retinoids (149), acetyl salicylic acid (150), and 1,25 dihydroxyvitamin D3 (151). Interestingly, each of these inhibitors operated through inhibition of NF-jB-mediated induction via the proximal NF-jB binding site within the p40 promoter (132). A defect in IL-12 production observed in Mkk3-deficient mice, mediated through lack of p40 promoter activity, may involve NF-jB activation also (21). The immune suppression occurring following measles infection may also involve suppression of IL-12, through a mechanism related to cross-linking of CD46 on monocytes (152), although the biochemical mechanism of this inhibition is unresolved. Monocytes and dendritic cells previously activated with endotoxin also become unable to produce IL-12 upon further stimulation, termed endotoxin tolerance (153). The mechanism of this repression is thought to be transcriptional, independent of IL-10, and not rescued by IFN-c or GMC-SF treatment (153). Finally, IL-10 is an important physiologic inhibitor of IL-12 production. In fact, this action was likely a key to its discovery, as its initial moniker, cytokine synthesis inhibitory factor (CSIF), implied (154). Although the inhibitory action of IL-10 is well known, its mechanism has been elusive. Involvement of IL-12 transcription at the level of the p40 promoter was proposed (155), but mapping the responsive cis-elements in the p40 promoter has not been successful. Interestingly, very recently Hamilton and colleagues have identified a potential mechanism that may relate to IL-12 p40 inhibition (156). These authors showed that IL-10 mediates mRNA destabilization of several cytokines via actions at clustered AU-rich elements present within the 38UTR (156). While this study analyzed TNF-a specifically, similar elements are present within the 38UTR of the IL-12 p40 mRNA. Since its initial identification as a Th1-inducing factor, the production of IL12 has been considered generally beneficial for antipathogen immune responses. However, in a model of LCMV infection, Biron and co-workers showed that lowdose, but not high-dose administration of IL-12 was beneficial (157). In addition, for MCMV infection, low-dose administration of IL-12 is also of benefit (158). Further, in MCMV infection, these coworkers showed that early responses were

463


characterized by induction of IFN-a/b, TNF, IL-12, and IFN-c, each of which contributed to both T cell–dependent and –independent antiviral functions (159). These and subsequent studies (160, 161) suggest that IL-12 is but one road to promoting T cell IFN-c responses during viral infections, with IFN-a/b production induced by certain viruses representing the other. For example, in response to influenza virus infection, IL-12 production contributes to early NK cell IFN-c production, but not to later T cell IFN-c production (162) (Table 1).


IL-4 Signaling and Th2 Development Seder & Paul (163) have reviewed the extensive data documenting the role for IL-4 in Th2 development from naive T cell precursors. Recently Huang & Paul (164) reported the distinct IL-4 independence for Th2 cytokine production by fully differentiated T helper cells. IL-4 receptor signaling, reviewed in (165), involves the IL-4-specific a chain and common gamma chain (cc), and activation of Janus tyrosine kinases, Jak1 and Jak3. Phosphorylation of specific tyrosine residues in a chain cytoplasmic domain recruits downstream signaling molecules (166) with distinct regions controlling growth responses and IL-4-specific gene induction (167). Mitogenic effects involve activation of 4PS (IRS-1) (167), while IL-4-specific gene induction involves Stat6 (168, 169). TABLE 1 Targeted and natural mutations influencing Th1 development Targeted murine mutations Gene

Phenotype

References

IL-12

f but not absent IFN-c production

(125, 126)

IL-18

f IFN-c production

(93)

IL-12Rb1

f IL-12 induced Th1

(127)

IL-18Rrp

f but not absent IFN-c production

(113)

Stat1

f IL-12R, intact Th1 development

(308)

Stat4

f IFN-c production

(41, 42)

IRAK

f IL-18 induced IFN-c

(112)

MyD88

f IL-18 induced IFN-c production

(111)

JNK1

Mild F in IL-4, no D in IFN-c

(20)

JNK2

f IL-12Rb2 and IFN-c

(19)

MKK3

f IL-12 production

(21)

IRF-1

f IL-12 production

(22, 23)

Human deficiencies IL-12 Impaired mycobacterial immunity

(130)

IL-12Rb1

(128, 129)

Impaired resistance to mycobacteria and Salmonella


464

MURPHY ET AL

Discrete tyrosine residues are responsible for the recruitment of Stat6 to the receptor complex (167), but additional levels of regulation appear involved in tuning receptor activation. For example, an IL-4Ra allelic variant among humans, consisting of glutamine 576 to arginine substitution (Q576R) within the region of Stat6 recruitment, has been observed to be strongly associated with allergic inflammatory disorders (170). However, direct transfer of this mutation and analysis in vitro using cell lines failed to produce a detectable change in IL-4 sensitivity (171), suggesting the observed hypersensitive induction of CD23 in human patients who have this allele may actually reside elsewhere in the IL-4 signaling pathway. In addition to the earlier identified Stat6-docking tyrosine residues, tyrosines Y497 and Y713 appeared to mediate protection from apoptosis, apparently independently of 4PS (IRS-2), Shc, or SHIP (172). Further, residues adjacent to the conserved tyrosine residues that serve as IRS-2 and Stat6 recruitment sites play a role in the efficiency of IL-4 signaling (173). Finally, a direct role for Stat6 in Th2 development was confirmed through targeting of the Stat6 locus in murine T cells (174–176). Following priming with IL-4, Th2 populations become progressively more stable to reversal by Th1-inducing cytokines (177), and IL-4 production becomes independent of extrinsic IL-4 (164). Loss of IL-12R expression by Th2 cells (40, 179) may explain early Th2 stability. In contrast, the stability of the Th1 phenotype is less clear, but recently IL-4 signaling impairment was also reported for Th1 cells (180), although in other systems, diminished but not absent IL-4 signaling was seen (40). Huang & Paul (180) reported striking decreases in Stat6 phosphorylation in Th1 cells relative to Th2 cells, with reduced phosphorylation of Jak3 and IRS-2. Stat6 and Jak3 proteins were present in both Th1 and Th2 cells, with decreased IRS-2 protein described in Th1 cells.

SOCS PROTEINS AND NEGATIVE REGULATION OF CYTOKINE SIGNALING Diminished IL-4 signaling in Th1 cells is not understood but may involve SOCS proteins (181). SOCS proteins were identified first as novel cytokine-inducible SH2 containing (CIS) proteins (182). Later, SOCS-1, SOCS-2, and SOCS-3 (183), JAK-binding protein (JAB) (identical to SOCS-1) (184), and SSI-1 (identical to SOCS-1) (185) were cloned. Soon thereafter, human homologues to these family members were identified (186, 187). Hilton and colleagues recognized a common motif in these three proteins called the SOCS box (183). A number of additional proteins containing this C-terminal SOCS box can be divided among five structural classes (188): those having SH2-domains, WD-40 and ankyrin repeats, GTPase domains, or the SSB family of proteins containing the SPRY domain (189).



465

SOCS proteins are now recognized to inhibit signaling in several pathways, including SOCS-2 in insulin-like growth factor I signaling (190), SOCS-3 in growth hormone (191) and leptin signaling (192), and CIS3 and SOCS-1 in prolactin signaling (193). Cytokine signaling related to Th1/Th2 development was recognized by the finding that SOCS-1 could inhibit the signaling of both IFNa/b and IFN-c signaling (194). The actions of SOCS proteins require the SH2 domain and a region immediately N-terminal to the SH2 domain (195, 196), consistent with previous findings that the JAB (SOCS-1) SH2 domain alone did not fully inhibit Jak2 (184). Yoshimura and colleagues identified two regions involved in inhibition of JAK kinases (197), including an extended SH2 subdomain that bridges the SH2 to a kinase inhibitory region. Interaction between these SOCS and JAK proteins may be facilitated by the tyrosine phosphorylation on the JAK kinase. Subsequently, this interaction may target JAK kinases to a degradation pathway (198–201). CIS undergoes ubiquitinization (198), leading to the proteasome degradation pathway (200). Indeed, the SOCS-box binds Elongins B and C (200), suggesting a direct coupling to proteasomal degradation. Another study, however, suggested that rather than promoting degradation, this interaction may increase expression of SOCS proteins by inhibiting their degradation (199). The recently determined crystal structure of Elongin C–Elongin B complex (201) is consistent with its suggested interaction with SOCS-box containing family members. Initially, several cytokine pathways were found to induce expression of SOCS1, SOCS-2, and SOCS-3 (183). Induction of SOCS-1 appeared to be Stat3dependent (185), suggesting involvement by various cytokines. For CIS, induction appears to be mediated via activation of Stat5 (202, 203). Further, the CIS promoter possesses distinct Stat5-responsive elements (202). Interestingly, there is feedback regulation in LIF signaling by specific SOCS-3 induction (204). LIF signaling activates Stat3, which binds specific sites in the SOCS-3 promoter. Further, the stable transfection of SOCS-3 inhibited LIF signaling, implying completion of the feedback inhibition loop. Specificity may also reside in which SOCS can inhibit which cytokine pathways. For example, SOCS-1 and SOCS-3, but not SOCS-2, inhibit IFN-c-induced activation of antiviral and antiproliferative responses (205). SOCS-1 also inhibits IL-4 signaling in B cells and fibroblasts (181, 206), and perhaps in human monocytes (207). More recently, a connection of SOCS protein actions to the effector and phenotype function of T cells was suggested by the phenotypes of SOCS-1-deficient mice (208–210). The involvement of SOCS proteins in a widespread group of signaling pathways was implied by the initial finding of lethality caused by liver degeneration occurring in SOCS-1-deficient mice. In thymocytes of SOCS-1deficient animals, the transient activation of Stat6 induced by IL-4 was converted to a significantly prolonged duration of Stat6 activation, implying a normal role of SOCS-1 in attenuation of IL-4 signaling (209). Finally, the phenotype of T cells from SOCS-1-deficient mice exhibit a hyperresponsiveness to IFN-c imply-

466

MURPHY ET AL


ing an additional role of SOCS-1 in attenuation of IFN signaling in vivo (178, 210). Currently, the in vivo patterns of SOCS activation by cytokines in T cells are relatively uncharacterized but promise to provide a basis for differential regulation of IFN and IL-4 signaling. Implication of SOCS-3 in inhibition of IL-2 responses has been shown in human T cell line and PBL (211). These actions could potentially explain observations related to differences between Th1 and Th2 cells reported for both IFN-c signaling (212, 213) and for various levels of IL-4 signaling reported for Th1 cells (180, 214). Finally, while the basis for the diminished IL-4 signaling in Th1 cells is not fully understood, it may involve SOCS-1 (181). Rothman and colleagues showed the potential for SOCS-1 to inhibit Jak1 and Stat6 activation in response to IL-4 but did not explicitly demonstrate a role in the differential signaling between Th1 and Th2 cells (181).

MOLECULAR REGULATION OF IL-4 AND TH2 CYTOKINE GENES The molecular regulation of the IL-4 gene and the role of NFAT in cytokine regulation were recently reviewed (215, 216). Initial targeting studies of NFAT1 implied that this factor exerted a negative influence on cytokine expression and immune responses (217, 218), showing modest effects on the bias in T helper responses. Kiani et al (219) showed that NFAT1 is involved in termination of the late phase of IL-4 gene transcription, thereby inhibiting Th2 responses. NFAT1deficient mice were more susceptible to infection with Leishmania major and to prolonged levels of IL-4 gene transcripts (219). Further, increased levels of IL-5 and IL-13 were reported in sensitized NFAT1 knockout mice after allergenchallenge in vivo (220). Similar results were confirmed by Serfling and colleagues (221, 222). RAG-2 blastocyst complementation showed that NFAT2-deficient T lymphocytes have impaired activation and decreased production of IL-4 (223, 224). NFAT4 is preferentially expressed in immature double positive thymocytes and is required for development of mature single positive CD4` and CD8` thymocytes and peripheral T cells (225). NFAT4-deficient peripheral T cells are hyperactive, perhaps reflecting an increased sensitivity to TCR-mediated signaling (225). While the single deficiencies of NFAT1 and NFAT4 exhibited modest disturbances in lymphocyte function, dual deficiency of these two transcription factors generated a profound lymphoproliferative disorder suggestive of a lower threshold of TCR signaling and increased resistance to apoptosis (226). Additionally, remarkable increases in serum IgG1 and IgE and dramatic selective increases in Th2 cytokines resulted (226). The mechanism of NFAT-dependent gene induction was recently shown to involve p300/CBP coactivators (227, 228), similar to a number of other transcription factors.



467

The Th2-specific transcription factor c-maf (31) promotes IL-4 promoter activity. Transgenic expression of c-maf did not dramatically enhance IL-4 production in mature Th1 cells, but did diminish IFN-c production (32). Further, targeting the c-maf locus in vivo led to an abnormality in lens development (229) and a selective defect in IL-4 production, but not other Th2 cytokines (230). Recently, Bcl-6 was found to influence Th2 development (231, 232). Targeted disruption of the Bcl-6 locus in mice led to increased Th2 cytokine production, to heart and lung vasculitis, and to the absence of germinal centers (231). Bcl-6 binds sites recognized by Stat6 and represses IL-4-induced transcription at such sites. Interestingly, the inflammatory disorders and elevated Th2 cytokines occurred even when the Bcl-6 disruption was crossed onto IL-4-deficient, or Stat6-deficient backgrounds (232). Thus, although Stat6 appears necessary for Th2 differentiation, Bcl-6 may repress Th2 responses in vivo by a pathway that is IL-4- and Stat6-independent (231). This may suggest repression of targets downstream of IL-4 or Stat6 in the hierarchy of Th2 transcription factors, such as GATA-3. Flavell and coworkers described NFAT-dependent transcriptional activity (233) and AP-1 activity (JunB) (234) to be relatively higher in Th2 compared to Th1 cells. AP-1 binds OCT sites near the IL-4 promoter elements P1 (235) or P0 (236). Structural analysis showed that NFAT proteins possess an extended interface surface that can mediate interactions with AP-1 when bound to adjacent sites on DNA (237), perhaps explaining cooperativity observed in transactivation via these proximal IL-4 promoter elements. Similarly, an interaction between c-maf and JunB was recently reported, further providing a basis for Th2-selective synergistic interactions of factors on the IL-4 promoter (238) (Figure 2). GATA-3 plays roles in several distinct developmental pathways. Targeted disruption of GATA-3 causes lethal abnormalities of the central nervous system and in hematopoiesis (239). Using RAG-2 blastocyst complementation, Leiden and colleagues showed that GATA-3-deficient thymocytes arrest at the immature double negative stage of development (240). In addition, these cells fail to reach the earliest CD44`CD251 stage of development (241). Two groups (242, 243) identified GATA-3 as selectively expressed in Th2 cells. A role for GATA-3 was shown for the Th2-specific cytokine IL-5, and these authors argued that while GATA-3 directly transactivates the IL-5 promoter, it had significantly less activity in the IL-4 promoter (243, 244). The IL-5 promoter region between –70 and –59 was suggested as a target for GATA-3 (245), findings recently confirmed (246). Consistent with this idea, several regions with GATA-3-dependent enhancer activity were identified within the IL-4/IL-13 locus (247), suggesting that GATA-3 could act to enhance expression of the IL-4/IL-13 locus via interactions at more distant sites. In addition to increasing expression of Th2-selective cytokines, GATA-3 inhibits Th1 development (30), in part mediated by inhibition of IL-12Rb2 expression (30, 248). Interestingly, a correlation was reported linking GATA-3 expression with increased atopic asthma in humans (249), suggesting generality of these


468

MURPHY ET AL

Figure 2 Pathways and transcription in IL-4 production by Th2 cells. Th2-specific components (asterisks) include the transcription factors GATA-3, c-Maf, and JunB. Acute transcriptional activation by NF-AT, c-Maf, JunB, and perhaps other factors may require prior chromatin remodeling of the IL-4 locus or enhancer activity provided by GATA-3. Th2-specific DNase-hypersensitive sites are indicated with asterisks.

findings into humans. Finally, Agarwal & Rao showed that structural changes within the IL-13/IL-4 locus can be detected during Th2 development, with specific DNase hypersensitive sites emerging in a Stat6-dependent manner (17). The


469

nuclear factors responsible for the DNase hypersensitive regions are currently not identified, but several of the sites are in close proximity to earlier identified GATA-3 sites (247). A cis-element capable of Stat6 binding downstream of the IL-4 structural gene was proposed to act as a subset-specific silencing element in Th1 cells (214). Further, a Stat6 binding site in the proximal IL-4 promoter has been identified by several groups (250–252), although its precise role in differentiation or acute transcriptional activation are not completely clear.


IL-13 AND TH2 DEVELOPMENT IL-13, initially cloned as an inducible cytokine (253), was recognized to be closely linked to IL-4, to share the common IL-4Ra subunit, and to induce many of the same target genes and inflammatory responses (254). More recently, IL-13 was implicated as a mediator of allergic asthma (255, 256). The role of IL-13 in peripheral T cell development is less clear. IL-13-deficient T cells show defective Th2 development in vitro, both in response to IL-4 and the combination of IL-4 and IL-13 treatment (257). The reason for the inability of IL-13 to restore the normal phenotype in CD4` T cells could be a failure of naive T cells to express IL-13R, but the failure of Th2 development in response to IL-4 was surprising (257). The IL-13-specific receptor (IL-13Ra1) is expressed on B cells, specifically, with highest expression on IgD` CD381 populations, representing naive B cells (258). Additionally, although this receptor showed some mRNA expression in T cells, no surface expression was detected on T cells using a specific monoclonal antibody to the IL-13Ra1 subunit (258). Th2 responses promote resolution of gastrointestinal nematode infection by effector mechanisms that involve worm expulsion. Indeed, IL-4 treatment promotes the cure of Nippostrongylus brasiliensis infection (259), and yet IL-4-deficient mice could restrict egg production and expel adult N. brasiliensis as do wild-type animals. These in vivo results challenge the concept that IL-4 was required for the initiation of Th2 development in vivo (260). Although the initial IL-4 source responsible for priming Th2 responses in vivo has been controversial (261), the ability of mice with targeted mutations in the IL-4Ra subunit to generate an IL-4 response in vivo suggested the existence of an IL-4-independent pathway of Th2 development (262). In an analysis of N. brasiliensis resistance, Finkelman and colleagues showed that the presence of an intact pathway involving IL-13, IL-4Ra chain, and Stat6 is required for full competence to expel N. brasiliensis (263). Further, these results were confirmed recently in a direct comparison of response to N. brasiliensis of IL-4Ra-deficient, and IL-4-deficient mice (264), indicating a role of IL-13 in the generation of Th2 responses controlling expulsion of parasites. Finally, additional evidence of partially redundant roles of IL-4 and IL-13 in vivo in generating Th2 responses was provided by analysis of doubly-deficient IL-4 and IL-13 knockout mice (257). In summary, while IL-4 appears to be a dominant factor in the initiation of T cell skewing toward the Th2 phenotype, IL-13 clearly

470

MURPHY ET AL

plays roles in effector responses, which are mediated through a Stat6-dependent signaling pathway and partially depend on a common receptor shared with IL-4.


CONTROL OF TH1/TH2 DEVELOPMENT BY APCS While the cytokines are considered dominant in controlling T cell polarization, additional mechanisms have been uncovered (265, 266). Liu and colleagues showed that distinct dendritic cell populations exerted divergent polarizing effects on CD4` T cells in vitro (265). While a DC1 population, capable of producing IL-12, preferentially induced Th1 differentiation as expected, unexpectedly the CD4` CD3– CD11c–-derived DC2 dendritic cells promoted Th2 development, even when IL-4 was neutralized. This IL-4-independent priming for Th2 development challenges the long-held notion that IL-4 is required for this pathway. Maliszewski and colleagues extended these studies in vivo, obtaining similar results (266). Distinct dendritic cell subsets were found to promote not only distinct polarizing influences on T cell–derived cytokine patterns, but also on the characteristic pattern of isotype switching in B cells. Intriguingly, the precursor of the DC2 cells was recently shown to be the major source of type I interferons present in human blood in response to viral challenge (267). Thus, the precursor of DC2 cells is an effector cell type in the immune system important for antiviral and antitumor responses, yet generates a dendritic cell type, DC2, which subsequently can promote Th2 responses, not thought to be most appropriate for antiviral immune responses. Clearly dendritic cell control of Th1/Th2 development will be an area of active research (Table 2).

TC1/TC2 AND TH1/TH2: PARALLOGS OR ORTHOLOGS? Polarization in T cells initially centered on CD4` cells, but CD8` T cells were found capable of acquiring an IL-4-producing phenotype (268), reviewed in (269, 270). Confirming this work, others subsequently used allogeneic or TCR transgenic systems to propose a model of CD8` T cell polarization similar to that proposed earlier proposed for CD4` T cells (271, 272). Mosmann and colleagues proposed the nomenclature Tc1/Tc2 for CD8 counterparts of Th1/Th2 subsets (271). Despite general similarities, Tc2 cells generally produce significantly less IL-4 than Th2 cells and can continue to produce IFN-c, whereas Th2 cells lose this capacity. Toward understanding these differences between Tc1/Tc2 cells and Th1/Th2 cells, several recent findings are worth mentioning. Two studies have shown a requirement for CD4 in Th2 development (273, 274). First, interaction CD4 and MHC class II/peptide complex were found to be necessary for normal Th2 development (273, 274). One study found that the cytoplasmic domain of CD4 was not required for Th2 responses against N. brasiliensis (273). A second reported that only IL-4-induced Th2 development could

471

T HELPER DEVELOPMENT TABLE 2 Targeted and natural mutations influencing Th2 development Targeted murine mutations


Gene

Phenotype of deficiency

Reference

IL-4

f Th2 development

(309)

IL-4Ra

Marked f Th2 development

(262)

Stat6

f Th2 development

(174–176)

NFAT1

Moderate F Th2 cytokines

(217, 218)

NFAT2

f IL-4 (RAG complementation)

(223, 224)

NFAT4/NFAT2 DKO

Startling F in Th2 cytokines

(225)

JNK1

Mild F in IL-4, no D in IFN-c

(20)

GATA-3

Lethal. Loss of DN thymocytes (RAG complementation)

(239) (240, 241)

c-Maf

f IL-4

(229, 230)

Bcl-6

F Th2 cytokines

(231)

CD4

f Th2 development

(273, 274)

CIITA

F IL-4 in Th1 cells

(27)

Human deficiency IL-4R variant

Associated with atopy

(170)

occur in the absence of this domain (274). Nonetheless there are compelling reasons to consider CD4 signaling as a potential modifier of Th2 development. The cytoplasmic domain of CD4 associates with the tyrosine kinase p56lck (Lck) (275, 276). Expression of a dominant-negative Lck was found to inhibit Th2 development (277), suggesting a link to the CD4 requirement described earlier. Although Lck overexpression correlated with Stat3 and Stat5 activation (278), no direct role for Lck in Stat4 or Stat6 activation has been reported. However, Lck activation has been linked to the pathway of activation of the Itk kinase (279), possibly providing a basis for cytokine gene regulation. Targeted disruption of the Itk gene causes defective TCR-induced proliferation (280), subsequently shown to result from decreased IL-2 production in part due to an uncoupling of TCR signaling from capacitative calcium entry (281). Differences between Th1 and Th2 cells in calcium signaling have been recognized for some time (282) and confirmed by recent analyses of Ca2` signaling pathways during Th1 and Th2 development (283). Upon activation, Th1 cells exhibited multiple calcium transients, whereas Th2 cells exhibited an initial spike followed by a very low level of sustained calcium flux (283). No comparison of calcium fluxes in polarized CD8` T cells has been reported. However, two reports are notable in this regard (284, 285). Ohashi and colleagues described the capacity of naive CD8` T cells to generate a sustained calcium flux in response to full


472

MURPHY ET AL

TCR agonists, with attenuation of signaling occurring with partial agonists and antagonists (284). In addition, two phases of calcium signaling were demonstrated to occur for regulation of CTL activity (285). For perforin/granule exocytosis, a large calcium transient and sustained release from intracellular stores was required, whereas for FasL induction and killing, only the sustained influx of extracellular calcium was required. These distinct components of calcium signaling (transient spikes versus sustained plateaus) may relate selectively to distinct downstream effector molecules. Crabtree and colleagues (286) reported that transient Ca2` spikes are not sufficient to maintain NFAT-dependent transcription, but that sustained calcium plateaus were required for this activity. Further, large transient calcium influxes selectively activate NF-jB and JNK, whereas even low, but sustained calcium plateaus are sufficient for activation of the NFAT transcription factor (287). Finally, activated CD8` T cells express significantly higher levels of the enzyme protein kinase C (PKC) c, a calcium-dependent isoform of PKC, suggesting that perhaps CD4` and CD8` T cells may differ intrinsically in their regulation of downstream signaling (288). Conceivably differential capacities between Th1, Th2, Tc1, and Tc2 cells in regulation of calcium transients and plateaus could produce differences in NFAT-, NF-jB-, and JNK-dependent transcription of cytokines genes such as IL-4. Beyond differences in IL-4 production, CD4` and CD8` T cells may also have differences in regulation of other cytokines. Recently, differences between CD4` and CD8` T cells for their requirement for Stat4 in the regulation of IFNc production were reported (44). Differences potentially could also reside in the component of the MAPK kinase pathway discussed above, although this has not been directly addressed. For example, a dominant-negative Ras inhibits Th2 development (289), but its role in Tc2 development has not been addressed. Also, no direct comparison of the initial methylation status of the IFN-c locus between CD4 and CD8 T cells has been reported.

INSTRUCTION, SELECTION, AND THE POSSIBLY RANDOM NATURE OF COMMITMENT The robust effects of cytokines for inducing Th1 or Th2 development from naive T cells were clearly demonstrated using mono-specific T cells primed in vitro in bulk cultures. While highly reproducible and having strong polarizing effects, the studies from bulk culture systems do not necessarily distinguish between instructive or selective models of T cell differentiation. A purely instructive model proposes that all naive T cells are fully capable for differentiation down either pathway, and that this commitment decision is based upon the instructive signals (i.e., cytokines), experienced during the process of primary T cell activation. In contrast, a selective model proposes that the population of naive T cells appears to polarize because of the selective outgrowth of precommitted precursors, whose



473

differentiation is restricted to one or another pathway. In yet a third model, which could be called ‘‘stochastic without instruction,’’ a multipotent population of naive precursors would undergo primary activation to generate an initial repertoire of all possible cytokine production profiles, and this early random commitment would be fixed by epigenetic mechanisms. Subsequent polarization in this model would result from the selective outgrowth of certain cell types, influenced by either the stochastic expression of growth factor receptors or by the stochastic expression of cytokines acting as growth factors. While initially, an instructive model was all but presumed, based on the precedent from numerous developmental paradigms (for an interesting instructional parallel see 290), some recent results have attracted attention to the ‘‘instruction versus selection’’ debate within Th1/Th2 biology. To begin addressing the basis of Th commitment, several studies compared the stability versus reversibility of early polarized cultures (177, 291). However, related studies based on single-cell cloning have not been reported. Flavell and colleagues have addressed this issue using a lineage ablation approach (292–294). By expressing herpes simplex virus 1 thymidine kinase under the control of the genomic IL-4 promoter, activation of early naive T cells in the presence of ganciclovir was shown to delete the precursors to both IL-4- and IFN-c-producing cells. This result was interpreted to mean that precursors to both the Th1 and Th2 lineage proceed through a stage in which the IL-4 promoter is expressed at some level, indicating a common precursor. Caveats include the potential for integration-dependent variations in transgene expression and that the sensitivity to deletion may not fully reflect endogenous IL-4 expression. Indeed, the recent observations (see below) on monoallelic versus biallelic cytokine expression, and the epigenetic regulation of the endogenous IL-4 locus may argue for caution in taking hard-and-fast interpretations of these results. While early populations were shown to contain some potential for reversibility, Th2 cells rapidly lost IL-12R expression and became resistant to subsequent IL12 reversal (295). In contrast, early Th1 populations retained a greater degree of reversibility (295–297). These results indicated that the reversal seen in early Th1 populations toward IL-4-producing cells was at least in part based upon the recruitment of uncommitted precursors that exist in a heterogeneous early Th1 population (297). Very recently, our concepts of early Th development and cytokine expression have changed based upon several novel and surprising results. First, Burakoff and coworker observed, both in T cell clones and naive developing T cells, that monoallelic expression of IL-2 genes occurred in a substantial fraction of T cells (298). By carefully examining the quantity, characteristics, and cloning efficiency of wild-type, IL-2 heterozygous, and IL-2-null T cells, the suspicion was raised that not all alleles of the IL-2 locus are available to all T cells during activation. To directly address the issue, expression of polymorphic IL-2 alleles in (C57BL/ 6 x m. Spretus) F1 mice was analyzed, and the clear finding of monoallelic expression was documented. Moreover, the timing of replication of the IL-2


474

MURPHY ET AL

alleles was analyzed by fluorescence in situ hybridization (FISH), revealing asynchronous replication in most activated T cells during S phase. This finding suggests a possible mechanism for marking one allele for expression or silencing, potentially related to replication-dependent marking of DNA, perhaps as described for PCNA and CAF-1-coupled inheritance of chromatin in yeast (299). In a similar approach, Bix & Locksley (300) demonstrated similar phenomena occurring for the IL-4 locus in Th2 cells. Subsequent studies have extended these analyses through the use of mice harboring a reporter gene targeted to both the IL-2 (301) or the IL-4 locus (302). By allowing analysis of individual alleles directly in populations of developing T cells, these later studies provided for a deeper understanding of the control and timing of the emergence of mono- and biallelic expression patterns (301). In contrast to the findings of Burakoff et al (298), Gu and colleagues found that biallelic expression of IL-2 was the predominant phenotype upon T cell activation (301). Although it is formally possible that the GFP cDNA and polyadenylation sequences used in this system disturb the pattern of monoallelic expression within the IL-2 locus, there is no evidence to suggest this. A second study using a similar approach but analyzing IL-4 expression, or a replacement of human CD2 knocked into the IL-4 locus, also found high frequency of biallelic expression (302). Interestingly, here, exploiting a mono-specific TCR population, the level of T cell activation was found to influence the distribution of monoallelic versus biallelic IL-4 locus expression when observing protein expression directed from distinct alleles (302). This study favored a model in which activation of the individual alleles is regulated by a stochastic process, and probability of activation is a function of strength of signal delivered by the TCR. As the regulation of cytokine genes comes under greater scrutiny in the process of Th1 and Th2 development from naive T cells, the general process of chromatin remodeling is being considered. Reiner and colleagues analyzed the requirement for cell division in the initiation of cytokine expression during differentiation of T helper 1 and T helper 2 cells from naive T cells (303). Interestingly, although T cells were immediately capable of making IL-2, production of IFN-c and IL-4 correlated with the number of cell divisions following activation. In particular, IFN-c appeared to require one to two cell divisions, and IL-4 production was delayed until four cell divisions. Similar findings have been reported by Gett & Hodgkin for a more extensive list of T cell–derived cytokines (304). While the mechanisms underlying these phenomena are not established, Reiner and colleagues (303) argued for a role of cell cycle and a division counter mechanism in the commitment process. Alternately, Gett & Hodgkin described similar requirements for cell division in the regulation of various cytokines, which are not differentially expressed among Th1 and Th2 cells, such as IL-3 (304). Two recent studies directly analyzed the IL-4/IL-13 locus for evidence of chromatin remodeling during Th1 and Th2 differentiation (17, 305). Agarwal & Rao demonstrated the acquisition of several specific DNase I–hypersensitive sites in the IL-4 and IL-13 locus (17). Acquisition of these sites in chromatin was dependent upon IL-4 signaling and Stat6 activation, suggesting a link to the process of



475

Th2 development. Similarly, Arai and colleagues constructed a map of DNase I– hypersensitive sites in polarized T cells, with the similar finding of specific sites being present exclusively in Th2, and not Th1 or naive T cells (305). The factors mediating the acquisition of these hypersensitive sites during early Th2 development have not been identified. It is possible that transcription factors differentially expressed between Th1 and Th2 cells could participate in altering chromatin structure or alter the accessibility of these loci to acutely activate transcription factors acting on the promoters of these cytokine genes. The role of methylation in IFN-c gene expression has been recently reviewed by Ye & Young (306). More recently, methylation of CpG residues in DNA may also play a role in the regulation of the IL-4 gene (17, 35, 303, 307). While a correlation between gene expression and methylation status has been long recognized, these studies extend this correlation to cytokine genes expressed by Th1 or Th2 cells. At present, whether methylation is the causative agent, or a reflection of differential expression established through other means, is unresolved. In summary, a stochastic component of cytokine gene expression has been established by the findings of monoallelic expression, but it is still undetermined whether this implies a random/stochastic model of commitment, or whether the data are still consistent with a cytokine-based instructional model. At least this controversy, if not many other unresolved issues, ensures a future for this field in the new millennium. ACKNOWLEDGMENTS Thanks to Robert D. Schreiber for helpful conversations and advice and to the many colleagues who provided us with preprints prior to publication. Work cited from our own laboratory was supported by grants from the National Institutes of Health, the Juvenile Diabetes Foundation, and the Howard Hughes Medical Institute. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Mosmann TR, Cherwinski H, Bond MW, Giedlin MA, Coffman RL. 1986. Two types of murine helper T cell clone. I. Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol. 136:2348 2. Romagnani S. 1994. Lymphokine production by human T cells in disease states. Annu. Rev. Immunol. 12:227 3. Seder RA, Paul WE. 1994. Acquisition of lymphokine-producing phenotype by

CD4` T cells. Annu. Rev. Immunol. 12:635 4. Hardy KJ, Peterlin BM, Atchison RE, Stobo JD. 1985. Regulation of expression of the human interferon gamma gene. PNAS 82:8173 5. Hardy KJ, Manger B, Newton M, Stobo JD. 1987. Molecular events involved in regulating human interferon-gamma gene expression during T cell activation. J. Immunol. 138:2353


476

MURPHY ET AL

6. Young HA, Komschlies KL, Ciccarone V, Beckwith M, Rosenberg M, Jenkins NA, Copeland NG, Durum SK. 1989. Expression of human IFN-gamma genomic DNA in transgenic mice. J. Immunol. 143:2389 7. Ciccarone VC, Chrivia J, Hardy KJ, Young HA. 1990. Identification of enhancer-like elements in human IFNgamma genomic DNA. J. Immunol. 144:725 8. Chrivia JC, Wdrychowicz T, Young HA, Hardy KJ. 1990. A model of human cytokine regulation based on transfection of gamma interferon gene fragments directly into isolated peripheral blood T lymphocytes. J. Exp. Med. 172:661 9. Penix L, Weaver WM, Pang Y, Young HA, Wilson CB. 1993. Two essential regulatory elements in the human interferon gamma promoter confer activation specific expression in T cells. J. Exp. Med. 178:1483 10. Aune TM, Penix LA, Rincon MR, Flavell RA. 1997. Differential transcription directed by discrete gamma interferon promoter elements in naive and memory (effector) CD4 T cells and CD8 T cells. Mol. Cell. Biol. 17:199 11. Penix LA, Sweetser MT, Weaver WM, Hoeffler JP, Kerppola TK, Wilson CB. 1996. The proximal regulatory element of the interferon-gamma promoter mediates selective expression in T cells. J. Biol. Chem. 271:31964 12. Zhang F, Wang DZ, Boothby M, Penix L, Flavell RA, Aune TM. 1998. Regulation of the activity of IFN-gamma promoter elements during Th cell differentiation. J. Immunol. 161:6105 13. Sweetser MT, Hoey T, Sun YL, Weaver WM, Price GA, Wilson CB. 1998. The roles of nuclear factor of activated T cells and ying-yang 1 in activation-induced expression of the interferon-gamma promoter in T cells. J. Biol. Chem. 273:34775 14. Sica A, Tan TH, Rice N, Kretzschmar M,

15.

16.

17.

18.

19.

20.

21.

22.

23.

Ghosh P, Young HA. 1992. The c-rel protooncogene product c-Rel but not NFkappa B binds to the intronic region of the human interferon-gamma gene at a site related to an interferon-stimulable response element. PNAS 89:1740 Campbell PM, Pimm J, Ramassar V, Halloran PF. 1996. Identification of a calcium-inducible, cyclosporine sensitive element in the IFN-gamma promoter that is a potential NFAT binding site. Transplantation 61:933 Sica A, Dorman L, Viggiano V, Cippitelli M, Ghosh P, Rice N, Young HA. 1997. Interaction of NF-kappaB and NFAT with the interferon-gamma promoter. J. Biol. Chem. 272:30412 Agarwal S, Rao A. 1998. Modulation of chromatin structure regulates cytokine gene expression during T cell differentiation. Immunity 9:765 Rincon M, Enslen H, Raingeaud J, Recht M, Zapton T, Su MS, Penix LA, Davis RJ, Flavell RA. 1998. Interferon-gamma expression by Th1 effector T cells mediated by the p38 MAP kinase signaling pathway. EMBO J. 17:2817 Yang DD, Conze D, Whitmarsh AJ, Barrett T, Davis RJ, Rincon M, Flavell RA. 1998. Differentiation of CD4` T cells to Th1 cells requires MAP kinase JNK2. Immunity 9:575 Dong C, Yang DD, Wysk M, Whitmarsh AJ, Davis RJ, Flavell RA. 1998. Defective T cell differentiation in the absence of Jnk1. Science 282:2092 Lu HT, Yang DD, Wysk M, Gatti E, Mellman I, Davis RJ, Flavell RA. 1999. Defective IL-12 production in mitogenactivated protein (MAP) kinase kinase 3 (Mkk3)-deficient mice. EMBO J. 18: 1845 Taki S, Sato T, Ogasawara K, Fukuda T, Sato M, Hida S, Suzuki G, Mitsuyama M, Shin EH, Kojima S, Taniguchi T, Asano Y. 1997. Multistage regulation of Th1-type immune responses by the transcription factor IRF-1. Immunity 6:673 Lohoff M, Ferrick D, Mittrucker HW,


24.


25.

26.

27.

28.

29.

30.

31.

Duncan GS, Bischof S, Rollinghoff M, Mak TW. 1997. Interferon regulatory factor-1 is required for a T helper 1 immune response in vivo. Immunity 6:681 Coccia EM, Passini N, Battistini A, Pini C, Sinigaglia F, Rogge L. 1999. Interleukin-12 induces expression of interferon regulatory factor-1 via signal transducer and activator of transcription-4 in human T helper type 1 cells. J. Biol. Chem. 274:6698 Ouyang W, Jacobson NG, Bhattacharya D, Gorham JD, Fenoglio D, Sha WC, Murphy TL, Murphy KM. 1999. The Ets transcription factor ERM is Th1-specific and induced by IL-12 through a Stat4dependent pathway. PNAS 96:3888 Barton K, Muthusamy N, Fischer C, Ting CN, Walunas TL, Lanier LL, L Leiden JM. 1998. The Ets-1 transcription factor is required for the development of natural killer cells in mice. Immunity 9:555 Gourley T, Roys S, Lukacs NW, Kunkel SL, Flavell RA, Chang CH. 1999. A novel role for the major histocompatibility complex class II transactivator CIITA in the repression of IL-4 production. Immunity 10:377 Ye J, Ghosh P, Cippitelli M, Subleski J, Hardy KJ, Ortaldo JR, Young HA. 1994. Characterization of a silencer regulatory element in the human interferon-gamma promoter. J. Biol. Chem. 269:25728 Ye J, Cippitelli M, Dorman L, Ortaldo JR, Young HA. 1996. The nuclear factor YY1 suppresses the human gamma interferon promoter through two mechanisms: inhibition of AP1 binding and activation of a silencer element. Mol. Cell. Biol. 16:4744 Ouyang W, Ranganath SH, Weindel K, Bhattacharya D, Murphy TL, Sha WC, Murphy KM. 1998. Inhibition of Th1 development mediated by GATA-3 through an IL-4-independent mechanism. Immunity 9:745 Ho IC, Hodge MR, Rooney JW,

32.

33.

34.

35.

36.

37.

38.

39.

477

Glimcher LH. 1996. The proto-oncogene c-maf is responsible for tissue-specific expression of interleukin-4. Cell 85:973 Ho IC, Lo D, Glimcher LH. 1998. c-maf promotes T helper cell type 2 (Th2) and attenuates Th1 differentiation by both interleukin 4-dependent and -independent mechanisms. J. Exp. Med. 188:1859 Young HA, Ghosh P, Ye J, Lederer J, Lichtman A, Gerard JR Jr, Penix L, Wilson CB, Melvin AJ, McGurn ME. 1994. Differentiation of the T helper phenotypes by analysis of the methylation state of the IFN-gamma gene. J. Immunol. 153:3603 Pang Y, Norihisa Y, Benjamin D, Kantor RR, Young HA. 1992. Interferon-gamma gene expression in human B-cell lines: induction by interleukin-2, protein kinase C activators, and possible effect of hypomethylation on gene regulation. Blood 80:724 Fitzpatrick DR, Shirley KM, McDonald LE, Bielefeldt-Ohmann H, Kay GF, Kelso A. 1998. Distinct methylation of the interferon gamma (IFN-gamma) and interleukin 3 (IL-3) genes in newly activated primary CD8` T lymphocytes: regional IFN-gamma promoter demethylation and mRNA expression are heritable in CD44(high)CD8` T cells. J. Exp. Med. 188:103 Leonard WJ, O’Shea JJ. 1998. Jaks and STATs: biological implications. Annu. Rev. Immunol. 16:293 Hoey T, Grusby MJ. 1999. STATs as mediators of cytokine-induced responses. Adv. Immunol. 71:145 Zhong Z, Wen Z, Darnell JE Jr. 1994. Stat3 and Stat4: members of the family of signal transducers and activators of transcription. PNAS 91:4806 Jacobson NG, Szabo SJ, Weber-Nordt RM, Zhong Z, Schreiber RD, Darnell JE Jr, Murphy KM. 1995. Interleukin 12 signaling in T helper type 1 (Th1) cells involves tyrosine phosphorylation of sig-

478

40.

41.


42.

43.

44.

45.

46.

47.

48.

MURPHY ET AL nal transducer and activator of transcription (Stat)3 and Stat4. J. Exp. Med. 181:1755 Szabo SJ, Jacobson NG, Dighe AS, Gubler U, Murphy KM. 1995. Developmental commitment to the Th2 lineage by extinction of IL-12 signaling. Immunity 2:665 Thierfelder WE, van Deursen JM, Yamamoto K, Tripp RA, Sarawar SR, Carson RT, Sangster MY, Vignali DA, Doherty PC, Grosveld GC, Ihle JN. 1996. Requirement for Stat4 in interleukin-12mediated responses of natural killer and T cells. Nature 382:171 Kaplan MH, Sun YL, Hoey T, Grusby MJ. 1996. Impaired IL-12 responses and enhanced development of Th2 cells in Stat4-deficient mice. Nature 382:174 Kaplan MH, Wurster AL, Grusby MJ. 1998. A signal transducer and activator of transcription (Stat)4-independent pathway for the development of T helper type 1 cells. J. Exp. Med. 188:1191 Carter LL, Murphy KM. 1999. Lineagespecific requirement for signal transducer and activator of transcription (Stat)4 in interferon gamma production from CD4` versus CD8` T cells. J. Exp. Med. 189:1355 Xu X, Sun YL, Hoey T. 1996. Cooperative DNA binding and sequence-selective recognition conferred by the STAT amino-terminal domain. Science 273:794 Yang J, Murphy TL, Ouyang W, Murphy KM. 1999. Induction of interferongamma production in Th1 CD4` T cells: evidence for two distinct pathways for promoter activation. Eur. J. Immunol. 29:548 Rogge L, D’Ambrosio D, Biffi M, Penna G, Minetti LJ, Presky DH, Adorini L, Sinigaglia F. 1998. The role of Stat4 in species-specific regulation of Th cell development by type I IFNs. J. Immunol. 161:6567 Bhattacharya S, Eckner R, Grossman S, Oldread E, Arany Z, D’Andrea A, Liv-

49.

50.

51.

52.

53.

54.

55.

56.

57.

ingston DM. 1996. Cooperation of Stat2 and p300/CBP in signalling induced by interferon-alpha. Nature 383:344 Korzus E, Torchia J, Rose DW, Xu L, Kurokawa R, McInerney EM, Mullen TM, Glass CK, Rosenfeld MG. 1998. Transcription factor-specific requirements for coactivators and their acetyltransferase functions. Science 279:703 Zhang JJ, Vinkemeier U, Gu W, Chakravarti D, Horvath CM, Darnell JE Jr. 1996. Two contact regions between Stat1 and CBP/p300 in interferon gamma signaling. PNAS 93:15092 Chen X, Vinkemeier U, Zhao Y, Jeruzalmi D, Darnell JE Jr, Kuriyan J. 1998. Crystal structure of a tyrosine phosphorylated STAT-1 dimer bound to DNA. Cell 93:827 Zhu M, John S, Berg M, Leonard WJ. 1999. Functional association of Nmi with Stat5 and Stat1 in IL-2- and IFN-gammamediated signaling. Cell 96:121 Vinkemeier U, Moarefi I, Darnell JE Jr, Kuriyan J. 1998. Structure of the aminoterminal protein interaction domain of STAT-4. Science 279:1048 Vinkemeier U, Cohen SL, Moarefi I, Chait BT, Kuriyan J, Darnell JE Jr. 1996. DNA binding of in vitro activated Stat1 alpha, Stat1 beta and truncated Stat1: interaction between NH2-terminal domains stabilizes binding of two dimers to tandem DNA sites. EMBO J. 15:5616 Meyer WK, Reichenbach P, Schindler U, Soldaini E, Nabholz M. 1997. Interaction of STAT5 dimers on two low affinity binding sites mediates interleukin 2 (IL2) stimulation of IL-2 receptor alpha gene transcription. J. Biol. Chem. 272:31821 John S, Vinkemeier U, Soldaini E, Darnell JE JR, Leonard WJ. 1999. The significance of tetramerization in promoter recruitment by Stat5. Mol. Cell. Biol. 19:1910 Parronchi P, De Carli M, Manetti R, Simonelli C, Sampognaro S, Piccinni


58.


59.

60.

61.

62.

63.

64.

65.

MP, Macchia D, Maggi E, Del Prete G, Romagnani S. 1992. IL-4 and IFN (alpha and gamma) exert opposite regulatory effects on the development of cytolytic potential by Th1 or Th2 human T cell clones. J. Immunol. 149:2977 Brinkmann V, Geiger T, Alkan S, Heusser CH. 1993. Interferon alpha increases the frequency of interferon gammaproducing human CD4` T cells. J. Exp. Med. 178:1655 Rogge L, Barberis-Maino L, Biffi M, Passini N, Presky DH, Gubler U, Sinigaglia F. 1997. Selective expression of an interleukin-12 receptor component by human T helper 1 cells. J. Exp. Med. 185:825 Wenner CA, Guler ML, Macatonia SE, O’Garra A, Murphy KM. 1996. Roles of IFN-gamma and IFN-alpha in IL-12induced T helper cell-1 development. J. Immunol. 156:1442 Bacon CM, Petricoin EF, Ortaldo JR, Rees RC, Larner AC, Johnston JA, O’Shea JJ. 1995. Interleukin 12 induces tyrosine phosphorylation and activation of STAT4 in human lymphocytes. PNAS 92:7307 Cho SS, Bacon CM, Sudarshan C, Rees RC, Finbloom D, Pine R, O’Shea JJ. 1996. Activation of STAT4 by IL-12 and IFN-alpha: evidence for the involvement of ligand-induced tyrosine and serine phosphorylation. J. Immunol. 157:4781 Steinhoff U, Muller U, Schertler A, Hengartner H, Aguet KM, Zinkernagel RM. 1995. Antiviral protection by vesicular stomatitis virus-specific antibodies in alpha/beta interferon receptor-deficient mice. J. Virol. 69:2153 Uze G, Lutfalla G, Gresser I. 1990. Genetic transfer of a functional human interferon alpha receptor into mouse cells: cloning and expression of its cDNA. Cell 60:225 Cook JR, Cleary CM, Mariano TM, Izotova L, Pestka S. 1996. Differential responsiveness of a splice variant of the

66.

67.

68.

69.

70.

71.

72.

73.

479

human type I interferon receptor to interferons. J. Biol. Chem. 271:13448 Novick D, Cohen B, Rubinstein M. 1994. The human interferon alpha/beta receptor: characterization and molecular cloning. Cell 77:391 Soh J, Mariano TM, Lim JK, Izotova L, Mirochnitchenko O, Schwartz B, Langer JA, Pestka S. 1994. Expression of a functional human type I interferon receptor in hamster cells: application of functional yeast artificial chromosome (YAC) screening. J. Biol. Chem. 269:18102 Colamonici OR, Domanski P. 1993. Identification of a novel subunit of the type I interferon receptor localized to human chromosome 21. J. Biol. Chem. 268:10895 Kim SH, Cohen B, Novick D, Rubinstein M. 1997. Mammalian type I interferon receptors consist of two subunits: IFNaR1 and IFNaR2. Gene 196:279 Owczarek CM, Hwang SY, Holland KA, Gulluyan LM, Tavaria M, Weaver B, Reich NC, Kola I, Hertzog PJ. 1997. Cloning and characterization of soluble and transmembrane isoforms of a novel component of the murine type I interferon receptor, IFNAR 2. J. Biol. Chem. 272:23865 Domanski P, Fish E, Nadeau OW, Witte M, Platanias LC, Yan H, Krolewski J, Pitha P, Colamonici OR. 1997. A region of the beta subunit of the interferon alpha receptor different from box 1 interacts with Jak1 and is sufficient to activate the Jak-Stat pathway and induce an antiviral state. J. Biol. Chem. 272:26388 Colamonici O, Yan H, Domanski P, Handa R, Smalley D, Mullersman J, Witte M, Krishnan K, Krolewski J. 1994. Direct binding to and tyrosine phosphorylation of the alpha subunit of the type I interferon receptor by p135tyk2 tyrosine kinase. Mol. Cell. Biol. 14:8133 Yan H, Krishnan K, Lim JT, Contillo LG, Krolewski JJ. 1996. Molecular characterization of an alpha interferon receptor

480

74.


75.

76.

77.

78.

79.

80.

81.

82.

MURPHY ET AL 1 subunit (IFNaR1) domain required for TYK2 binding and signal transduction. Mol. Cell. Biol. 16:2074 Abramovich C, Shulman LM, Ratovitski E, Harroch S, Tovey M, Eid P, Revel M. 1994. Differential tyrosine phosphorylation of the IFNAR chain of the type I interferon receptor and of an associated surface protein in response to IFN-alpha and IFN-beta. EMBO J. 13:5871 Yan H, Krishnan K, Greenlund AC, Gupta S, Lim JT, Schreiber RD, Schindler CW, Krolewski JJ. 1996. Phosphorylated interferon-alpha receptor 1 subunit (IFNaR1) acts as a docking site for the latent form of the 113 kDa STAT2 protein. EMBO J. 15:1064 Leung S, Qureshi SA, Kerr IM, Darnell JEJ, Stark GR. 1995. Role of STAT2 in the alpha interferon signaling pathway. Mol. Cell. Biol. 15:1312 Qureshi SA, Leung S, Kerr IM, Stark GR, Darnell JE Jr. 1996. Function of Stat2 protein in transcriptional activation by alpha interferon. Mol. Cell. Biol. 16:288 Yang CH, Shi W, Basu L, Murti A, Constantinescu SN, Blatt L, Croze E, Mullersman JE, Pfeffer LM. 1996. Direct association of STAT3 with the IFNAR-1 chain of the human type I interferon receptor. J. Biol. Chem. 271:8057 Pfeffer LM, Mullersman JE, Pfeffer SR, Murti A, Shi W, Yang CH. 1997. STAT3 as an adapter to couple phosphatidylinositol 3-kinase to the IFNAR1 chain of the type I interferon receptor. Science 276:1418 Ahmad S, Alsayed YM, Druker BJ, Platanias LC. 1997. The type I interferon receptor mediates tyrosine phosphorylation of the CrkL adaptor protein. J. Biol. Chem. 272:29991 Lichtman AH, Chin J, Schmidt JA, Abbas AK. 1988. Role of interleukin 1 in the activation of T lymphocytes. PNAS 85:9699 Taylor-Robinson AW, Phillips RS. 1994.

83.

84.

85.

86.

87.

88.

89.

Expression of the IL-1 receptor discriminates Th2 from Th1 cloned CD4` T cells specific for Plasmodium chabaudi. Immunology 81:216 Shibuya K, Robinson D, Zonin F, Hartley SB, Macatonia SE, Somoza C, Hunter CA, Murphy KM, O’Garra A. 1998. IL1 alpha and TNF-alpha are required for IL-12-induced development of Th1 cells producing high levels of IFN-gamma in BALB/c but not C57BL/6 mice. J. Immunol. 160:1708 Okamura H, Tsutsi H, Komatsu T, Yutsudo M, Hakura A, Tanimoto T, Torigoe K, Okura T, Nukada Y, Hattori K. 1995. Cloning of a new cytokine that induces IFN-gamma production by T cells. Nature 378:88 Okamura H, Nagata K, Komatsu T, Tanimoto T, Nukata Y, Tanabe F, Akita K, Torigoe K, Okura T, Fukuda S. 1995. A novel costimulatory factor for gamma interferon induction found in the livers of mice causes endotoxic shock. Infect. Immun. 63:3966 Bazan JF, Timans JC, Kastelein RA. 1996. A newly defined interleukin-1? Nature 379:591 Gu Y, Kuida K, Tsutsui H, Ku G, Hsiao K, Fleming MA, Hayashi N, Higashino K, Okamura H, Nakanishi K, Kurimoto M, Tanimoto T, Flavell RA, Sato V, Harding MW, Livingston DJ, Su MS. 1997. Activation of interferon-gamma inducing factor mediated by interleukin-1beta converting enzyme. Science 275:206 Ghayur T, Banerjee S, Hugunin M, Butler D, Herzog L, Carter A, Quintal L, Sekut L, Talanian R, Paskind M, Wong W, Kamen R, Tracey D, Allen H. 1997. Caspase-1 processes IFN-gamma-inducing factor and regulates LPS-induced IFN-gamma production. Nature 386:619 Fantuzzi G, Puren AJ, Harding MW, Livingston DJ, Dinarello CA. 1998. Interleukin-18 regulation of interferon gamma production and cell proliferation as shown in interleukin-1beta-converting


90.


91.

92.

93.

94.

95.

96.

97.

enzyme (caspase-1)-deficient mice. Blood 91:2118 Puren AJ, Fantuzzi G, Dinarello CA. 1999. Gene expression, synthesis, and secretion of interleukin 18 and interleukin 1beta are differentially regulated in human blood mononuclear cells and mouse spleen cells. PNAS 96:2256 Micallef MJ, Ohtsuki T, Kohno K, Tanabe F, Ushio S, Namba M, Tanimoto T, Torigoe K, Fujii M, Ikeda M, Fukuda S, Kurimoto M. 1996. Interferon-gammainducing factor enhances T helper 1 cytokine production by stimulated human T cells: synergism with interleukin-12 for interferon-gamma production. Eur. J. Immunol. 26:1647 Kohno K, Kataoka J, Ohtsuki T, Suemoto Y, Okamoto I, Usui M, Ikeda M, Kurimoto M. 1997. IFN-gamma-inducing factor (IGIF) is a costimulatory factor on the activation of Th1 but not Th2 cells and exerts its effect independently of IL12. J. Immunol. 158:1541 Takeda K, Tsutsui H, Yoshimoto T, Adachi O, Yoshida N, Kishimoto T, Okamura H, Nakanishi K, Akira S. 1998. Defective NK cell activity and Th1 response in IL-18-deficient mice. Immunity 8:383 Robinson D, Shibuya K, Mui A, Zonin F, Murphy E, Sana T, Hartley SB, Menon S, Kastelein R, Bazan F, O’Garra A. 1997. IGIF does not drive Th1 development but synergizes with IL-12 for interferon-gamma production and activates IRAK and NFkappaB. Immunity 7:571 Hoshino T, Wiltrout RH, Young HA. 1999. IL-18 is a potent coinducer of IL13 in NK and T cells: a new potential role for IL-18 in modulating the immune response. J. Immunol. 162:5070 Cao Z, Henzel WJ, Gao X. 1996. IRAK: a kinase associated with the interleukin1 receptor. Science 271:1128 Huang J, Gao X, Li S, Cao Z. 1997. Recruitment of IRAK to the interleukin 1 receptor complex requires interleukin 1

98.

99.

100.

101.

102.

103.

104.

105.

106.

481

receptor accessory protein. PNAS 94:12829 Wesche H, Korherr C, Kracht M, Falk W, Resch K, Martin MU. 1997. The interleukin-1 receptor accessory protein (IL1RAcP) is essential for IL-1-induced activation of interleukin-1 receptor-associated kinase (IRAK) and stress-activated protein kinases (SAP kinases). J. Biol. Chem. 272:7727 Muzio M, Ni J, Feng P, Dixit VM. 1997. IRAK (Pelle) family member IRAK-2 and MyD88 as proximal mediators of IL1 signaling. Science 278:1612 Wesche H, Henzel WJ, Shillinglaw W, Li S, Cao Z. 1997. MyD88: an adapter that recruits IRAK to the IL-1 receptor complex. Immunity 7:837 Cao Z, Xiong J, Takeuchi M, Kurama T, Goeddel DV. 1996. TRAF6 is a signal transducer for interleukin-1. Nature 383:443 Satoskar AR, Okano M, Connaughton S, Raisanen S, David JR, Labow M. 1998. Enhanced Th2-like responses in IL-1 type 1 receptor-deficient mice. Eur. J. Immunol. 28:2066 Huber M, Beuscher HU, Rohwer P, Kurrle R, Rollinghoff M, Lohoff M. 1998. Costimulation via TCR and IL-1 receptor reveals a novel IL-1alpha-mediated autocrine pathway of Th2 cell proliferation. J. Immunol. 160:4242 Matsumoto S, Tsuji-Takayama K, Aizawa Y, Koide K, Takeuchi M, Ohta T, Kurimoto M. 1997. Interleukin-18 activates NF-kappaB in murine T helper type 1 cells. Biochem. Biophys. Res. Commun. 234:454 Tsuji-Takayama K, Matsumoto S, Koide K, Takeuchi M, Ikeda M, Ohta T, Kurimoto M. 1997. Interleukin-18 induces activation and association of p56(lck) and MAPK in a murine TH1 clone. Biochem. Biophys. Res. Commun. 237:126 Parnet P, Garka KE, Bonnert TP, Dower SK, Sims JE. 1996. IL-1Rrp is a novel

482

107.


108.

109.

110.

111.

112.

113.

MURPHY ET AL receptor-like molecule similar to the type I interleukin-1 receptor and its homologues T1/ST2 and IL-1R AcP. J. Biol. Chem. 271:3967 Torigoe K, Ushio S, Okura T, Kobayashi S, Taniai M, Kunikata T, Murakami T, Sanou O, Kojima H, Fujii M, Ohta T, Ikeda M, Ikegami H, Kurimoto M. 1997. Purification and characterization of the human interleukin-18 receptor. J. Biol. Chem. 272:25737 Thomassen E, Bird TA, Renshaw BR, Kennedy MK, Sims JE. 1998. Binding of interleukin-18 to the interleukin-1 receptor homologous receptor IL-1Rrp1 leads to activation of signaling pathways similar to those used by interleukin-1. J. Interferon Cytokine Res. 18:1077 Born TL, Thomassen E, Bird TA, Sims JE. 1998. Cloning of a novel receptor subunit, AcPL, required for interleukin18 signaling. J. Biol. Chem. 273:29445 Kojima H, Takeuchi M, Ohta T, Nishida Y, Arai N, Ikeda M, Ikegami H, Kurimoto M. 1998. Interleukin-18 activates the IRAK-TRAF6 pathway in mouse EL4 cells. Biochem. Biophys. Res. Commun. 244:183 Adachi O, Kawai T, Takeda K, Matsumoto M, Tsutsui H, Sakagami M, Nakanishi K, Akira S. 1998. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18-mediated function. Immunity 9:143 Kanakaraj P, Ngo K, Wu Y, Angulo A, Ghazal P, Harris CA, Siekierka JJ, Peterson PA, Fung-Leung Wp. 1999. Defective interleukin (IL)-18-mediated natural killer and T helper cell type 1 responses in IL-1 receptor-associated kinase (IRAK)-deficient mice. J. Exp. Med. 189:1129 Hoshino K, Tsutsui H, Kawai T, Takeda K, Nakanishi K, Takeda Y, Akira S. 1999. Cutting edge: generation of IL-18 receptor-deficient mice: evidence for IL1 receptor-related protein as an essential

114.

115.

116.

117.

118.

119.

120.

121.

IL-18 binding receptor. J. Immunol. 162:5041 Barbulescu K, Becker C, Schlaak JF, Schmitt E, Meyer Z, Buschenfelde KH, Neurath MF. 1998. IL-12 and IL-18 differentially regulate the transcriptional activity of the human IFN-gamma promoter in primary CD4` T lymphocytes. J. Immunol. 160:3642 Ahn HJ, Maruo S, Tomura M, Mu J, Hamaoka T, Nakanishi K, Clark S, Kurimoto M, Okamura H, Fujiwara H. 1997. A mechanism underlying synergy between IL-12 and IFN-gamma-inducing factor in enhanced production of IFNgamma. J. Immunol. 159:2125 Yoshimoto T, Takeda K, Tanaka T, Ohkusu K, Kashiwamura S, Okamura H, Akira S, Nakanishi K. 1998. IL-12 upregulates IL-18 receptor expression on T cells, Th1 cells, and B cells: synergism with IL-18 for IFN-gamma production. J. Immunol. 161:3400 Xu D, Chan WL, Leung BP, Hunter D, Schulz K, Carter RW, McInnes IB, Robinson JH, Liew FY. 1998. Selective expression and functions of interleukin 18 receptor on T helper (Th) type 1 but not Th2 cells. J. Exp. Med. 188:1485 Kojima H, Aizawa Y, Yanai Y, Nagaoka K, Takeuchi M, Ohta T, Ikegami H, Ikeda M, Kurimoto M. 1999. An essential role for NF-kappa B in IL-18-induced IFNgamma expression in KG-1 cells. J. Immunol. 162:5063 Klemenz R, Hoffmann S, Werenskiold AK. 1989. Serum- and oncoproteinmediated induction of a gene with sequence similarity to the gene encoding carcinoembryonic antigen. PNAS 86:5708 Mitcham JL, Parnet P, Bonnert TP, Garka KE, Gerhart MJ, Slack JL, Gayle MA, Dower SK, Sims JE. 1996. T1/ST2 signaling establishes it as a member of an expanding interleukin-1 receptor family. J. Biol. Chem. 271:5777 Xu D, Chan WL, Leung BP, Huang F,


122.


123.

124.

125.

126.

127.

128.

Wheeler R, Piedrafita D, Robinson JH, Liew FY. 1998. Selective expression of a stable cell surface molecule on type 2 but not type 1 helper T cells. J. Exp. Med. 187:787 Lohning M, Stroehmann A, Coyle AJ, Grogan JL, Lin S, Gutierrez-Ramos JC, Levinson D, Radbruch A, Kamradt T. 1998. T1/ST2 is preferentially expressed on murine Th2 cells, independent of interleukin 4, interleukin 5, and interleukin 10, and important for Th2 effector function. PNAS 95:6930 Lohning M, Grogan JL, Coyle AJ, Yazdanbakhsh M, Meisel C, GutierrezRamos JC, Radbruch A, Kamradt T. 1999. T1/ST2 expression is enhanced on CD4` T cells from schistosome egginduced granulomas: analysis of Th cell cytokine coexpression ex vivo. J. Immunol. 162:3882 Gayle MA, Slack JL, Bonnert TP, Renshaw BR, Sonoda G, Taguchi, Testa T Jr, Dower SK, Sims JE. 1996. Cloning of a putative ligand for the T1/ST2 receptor. J. Biol. Chem. 271:5784 Magram J, Connaughton SE, Warrier RR, Carvajal DM, Wu CY, Ferrante J, Stewart C, Sarmiento U, Faherty DA, Gately MK. 1996. IL-12-deficient mice are defective in IFN gamma production and type 1 cytokine responses. Immunity 4:471 Piccotti R, Li K, Chan SY, Ferrante J, Magram J, Eichwald EJ, Bishop DK. 1998. Alloantigen-reactive Th1 development in IL-12-deficient mice. J. Immunol. 160:1132 Wu C, Ferrante J, Gately MK, Magram J. 1997. Characterization of IL-12 receptor beta1 chain (IL-12Rbeta1)-deficient mice: IL-12Rbeta1 is an essential component of the functional mouse IL-12 receptor. J. Immunol. 159:1658 de Jong R, Altare F, Haagen IA, Elferink DG, Boer T, van Breda V, Kabel PJ, Draaisma JM, van Dissel JT, Kroon FP, Casanova JL, Ottenhoff TH. 1998.

129.

130.

131.

132.

133.

134.

135.

483

Severe mycobacterial and Salmonella infections in interleukin-12 receptordeficient patients. Science 280:1435 Altare F, Durandy A, Lammas D, Emile JF, Lamhamedi S, Le Deist F, Drysdale P, Jouanguy E, Doffinger R, Bernaudin F, Jeppsson O, Gollob JA, Meinl E, Segal AW, Fischer A, Kumararatne D, Casanova JL. 1998. Impairment of mycobacterial immunity in human interleukin-12 receptor deficiency. Science 280:1432 Altare F, Lammas D, Revy P, Jouanguy E, Doffinger R, Lamhamedi S, Drysdale P, Scheel-Toellner D, Girdlestone J, Darbyshire P, Wadhwa M, Dockrell H, Salmon M, Fischer A, Durandy A, Casanova JL, Kumararatne DS. 1998. Inherited interleukin 12 deficiency in a child with bacille Calmette-Guerin and Salmonella enteritidis disseminated infection. J. Clin. Invest. 102:2035 Trinchieri G. 1995. Interleukin-12: a proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity. Annu. Rev. Immunol. 13:251 Murphy TL, Cleveland MG, Kulesza P, Magram J, Murphy KM. 1995. Regulation of interleukin 12 p40 expression through an NF-kappa B half-site. Mol. Cell. Biol. 15:5258 Ma X, Gri G, Trinchieri G. 1996. A novel ets-2-related nuclear factor is involved in transcriptional activation of the human interleukin-12 p40 gene promoter in response to interferon-gamma and LPS stimulation of monocytic cells. Ann. NY Acad. Sci. 795:357 Ma X, Chow JM, Gri G, Carra G, Gerosa F, Wolf SF, Dzialo R, Trinchieri G. 1996. The interleukin 12 p40 gene promoter is primed by interferon gamma in monocytic cells. J. Exp. Med. 183:147 Holtschke T, Lohler J, Kanno Y, Fehr T, Giese N, Rosenbauer F, Lou J, Knobeloch KP, Gabriele L, Waring JF, Bachmann MF, Zinkernagel RM, Morse HC, Ozato K, Horak I. 1996. Immunodefi-

484

136.


137.

138.

139.

140.

141.

142.

MURPHY ET AL ciency and chronic myelogenous leukemia-like syndrome in mice with a targeted mutation of the ICSBP gene. Cell 87:307 Wu CY, Maeda H, Contursi C, Ozato K, Seder RA. 1999. Differential requirement of IFN consensus sequence binding protein for the production of IL-12 and induction of Th1-type cells in response to IFN-gamma. J. Immunol. 162:807 Giese NA, Gabriele L, Doherty TM, Klinman DM, Tadesse-Heath L, Contursi C, Epstein SL, Morse HC. 1997. Interferon (IFN) consensus sequence-binding protein, a transcription factor of the IFN regulatory factor family, regulates immune responses in vivo through control of interleukin 12 expression. J. Exp. Med. 186:1535 Scharton-Kersten T, Contursi C, Masumi A, Sher A, Ozato K. 1997. Interferon consensus sequence binding protein-deficient mice display impaired resistance to intracellular infection due to a primary defect in interleukin 12 p40 induction. J. Exp. Med. 186:1523 Yoshimoto T, Kojima K, Funakoshi T, Endo Y, Fujita T, Nariuchi H. 1996. Molecular cloning and characterization of murine IL-12 genes. J. Immunol. 156:1082 Tone Y, Thompson SA, Babik JM, Nolan KF, Tone M, Raven C, Waldmann H. 1996. Structure and chromosomal location of the mouse interleukin-12 p35 and p40 subunit genes. Eur. J. Immunol. 26:1222 Ma X, Neurath M, Gri G, Trinchieri G. 1997. Identification and characterization of a novel Ets-2-related nuclear complex implicated in the activation of the human interleukin-12 p40 gene promoter. J. Biol. Chem. 272:10389 Plevy SE, Gemberling JH, Hsu S, Dorner AJ, Smale ST. 1997. Multiple control elements mediate activation of the murine and human interleukin 12 p40 promoters: evidence of functional syn-

143.

144.

145.

146.

147.

148.

149.

ergy between C/EBP and Rel proteins. Mol. Cell. Biol. 17:4572 Gri G, Savio D, Trinchieri G, Ma X. 1998. Synergistic regulation of the human interleukin-12 p40 promoter by NFkappaB and Ets transcription factors in Epstein-Barr virus-transformed B cells and macrophages. J. Biol. Chem. 273:6431 Yoshimoto T, Nagase H, Ishida T, Inoue J, Nariuchi H. 1997. Induction of interleukin-12 p40 transcript by CD40 ligation via activation of nuclear factor-kappaB. Eur. J. Immunol. 27:3461 Shu U, Kiniwa M, Wu CY, Maliszewski C, Vezzio N, Hakimi J, Gately M, Delespesse G. 1995. Activated T cells induce interleukin-12 production by monocytes via CD40-CD40 ligand interaction. Eur. J. Immunol. 25:1125 Armant M, Armitage R, Boiani N, Delespesse G, Sarfati M. 1996. Functional CD40 ligand expression on T lymphocytes in the absence of T cell receptor engagement: involvement in interleukin2-induced interleukin-12 and interferongamma production. Eur. J. Immunol. 26:1430 Avice MN, Demeure CE, Delespesse G, Rubio M, Armant M, Sarfati M. 1998. IL-15 promotes IL-12 production by human monocytes via T cell-dependent contact and may contribute to IL-12mediated IFN-gamma secretion by CD4` T cells in the absence of TCR ligation. J. Immunol. 161:3408 Avice MN, Sarfati M, Triebel F, Delespesse G, Demeure CE. 1999. Lymphocyte activation gene-3, a MHC class II ligand expressed on activated T cells, stimulates TNF-alpha and IL-12 production by monocytes and dendritic cells. J. Immunol. 162:2748 Na SY, Kang BY, Chung SW, Han SJ, Ma X, Trinchieri G, Im SY, Lee JW, Kim TS. 1999. Retinoids inhibit interleukin-12 production in macrophages through physical associations of retinoid X recep-


150.

151.


152.

153.

154.

155.

156.

157.

158.

tor and NFkappaB. J. Biol. Chem. 274:7674 Mazzeo D, Panina-Bordignon P, Recalde H, Sinigaglia F, D’Ambrosio D. 1998. Decreased IL-12 production and Th1 cell development by acetyl salicylic acidmediated inhibition of NF-kappaB. Eur. J. Immunol. 28:3205 D’Ambrosio D, Cippitelli M, Cocciolo MG, Mazzeo D, Di Lucia P, Lang R, Sinigaglia F, Panina-Bordignon P. 1998. Inhibition of IL-12 production by 1,25dihydroxyvitamin D3. Involvement of NF-kappaB downregulation in transcriptional repression of the p40 gene. J. Clin. Invest. 101:252 Karp CL, Wysocka M, Wahl LM, Ahearn JM, Cuomo PJ, Sherry B, Trinchieri G, Griffin DE. 1996. Mechanism of suppression of cell-mediated immunity by measles virus. Science 273:228 Karp CL, Wysocka M, Ma X, Marovich M, Factor RE, Nutman T, Armant M, Wahl L, Cuomo P, Trinchieri G. 1998. Potent suppression of IL-12 production from monocytes and dendritic cells during endotoxin tolerance. Eur. J. Immunol. 28:3128 Fiorentino DF, Bond MW, Mosmann TR. 1989. Two types of mouse T helper cell. IV. Th2 clones secrete a factor that inhibits cytokine production by Th1 clones. J. Exp. Med. 170:2081 Aste-Amezaga M, Ma X, Sartori A, Trinchieri G. 1998. Molecular mechanisms of the induction of IL-12 and its inhibition by IL-10. J. Immunol. 160:5936 Kishore R, Tebo JM, Kolosov M, Hamilton TA. 1999. Cutting edge: clustered AU-rich elements are the target of IL-10mediated mRNA destabilization in mouse macrophages. J. Immunol. 162:2457 Orange JS, Wolf SF, Biron CA. 1994. Effects of IL-12 on the response and susceptibility to experimental viral infections. J. Immunol. 152:1253 Orange JS, Wang B, Terhorst C, Biron

159.

160.

161.

162.

163.

164.

165.

166.

167.

485

CA. 1995. Requirement for natural killer cell-produced interferon gamma in defense against murine cytomegalovirus infection and enhancement of this defense pathway by interleukin 12 administration. J. Exp. Med. 182:1045 Orange JS, Biron CA. 1996. Characterization of early IL-12, IFN-alphabeta, and TNF effects on antiviral state and NK cell responses during murine cytomegalovirus infection. J. Immunol. 156:4746 Cousens LP, Orange JS, Su HC, Biron CA. 1997. Interferon-alpha/beta inhibition of interleukin 12 and interferongamma production in vitro and endogenously during viral infection. PNAS 94:634 Cousens LP, Peterson R, Hsu S, Dorner A, Altman JD, Ahmed R, Biron CA. 1999. Two roads diverged: interferon alpha/beta- and interleukin 12-mediated pathways in promoting T cell interferon gamma responses during viral infection. J. Exp. Med. 189:1315 Monteiro JM, Harvey C, Trinchieri G. 1998. Role of interleukin-12 in primary influenza virus infection. J. Virol. 72:4825 Seder RA, Paul WE. 1994. Acquisition of lymphokine-producing phenotype by CD4` T cells. Annu. Rev. Immunol. 12:635 Huang H, Hu-Li J, Chen H, Ben-Sasson SZ, Paul WE. 1997. IL-4 and IL-13 production in differentiated T helper type 2 cells is not IL-4 dependent. J. Immunol. 159:3731 Nelms K, Keegan AD, Zamorano J, Ryan JJ, Paul WE. 1999. The IL-4 receptor: signaling mechanisms and biologic functions. Annu. Rev. Immunol. 17:701 Wang HY, Paul WE, Keegan AD. 1996. IL-4 function can be transferred to the IL-2 receptor by tyrosine containing sequences found in the IL-4 receptor alpha chain. Immunity 4:113 Ryan JJ, McReynolds LJ, Keegan A,

486

168.


169.

170.

171.

172.

173.

174.

175.

MURPHY ET AL Wang LH, Garfein E, Rothman P, Nelms K, Paul WE. 1996. Growth and gene expression are predominantly controlled by distinct regions of the human IL-4 receptor. Immunity 4:123 Hou J, Schindler U, Henzel WJ, Ho TC, Brasseur M, McKnight SL. 1994. An interleukin-4-induced transcription factor: IL-4 Stat. Science 265:1701 Quelle FW, Shimoda K, Thierfelder W, Fischer C, Kim A, Ruben SM, Cleveland JL, Pierce JH, Keegan AD, Nelms K. 1995. Cloning of murine Stat6 and human Stat6, Stat proteins that are tyrosine phosphorylated in responses to IL-4 and IL-3 but are not required for mitogenesis. Mol. Cell. Biol. 15:3336 Hershey GK, Friedrich MF, Esswein LA, Thomas ML, Chatila TA. 1997. The association of atopy with a gain-of-function mutation in the alpha subunit of the interleukin-4 receptor [see comments]. N. Engl. J. Med. 337:1720 Wang HY, Shelburne CP, Zamorano J, Kelly AE, Ryan JJ, Keegan AD. 1999. Cutting edge: effects of an allergy-associated mutation in the human IL-4R alpha (Q576R) on human IL-4-induced signal transduction. J. Immunol. 162: 4385 Zamorano J, Keegan AD. 1998. Regulation of apoptosis by tyrosine-containing domains of IL-4R alpha: Y497 and Y713, but not the STAT6-docking tyrosines, signal protection from apoptosis. J. Immunol. 161:859 Ryan JJ, McReynolds LJ, Huang H, Nelms K, Paul WE. 1998. Characterization of a mobile Stat6 activation motif in the human IL-4 receptor. J. Immunol. 161:1811 Kaplan MH, Schindler U, Smiley ST, Grusby MJ. 1996. Stat6 is required for mediating responses to IL-4 and for development of Th2 cells. Immunity 4:313 Takeda K, Tanaka T, Shi W, Matsumoto M, Minami M, Kashiwamura S, Nakan-

176.

177.

178.

179.

180.

181.

182.

183.

ishi K, Yoshida N, Kishimoto T, Akira S. 1996. Essential role of Stat6 in IL-4 signalling. Nature 380:627 Shimoda K, van Deursen J, Sangster MY, Sarawar SR, Carson RT, Tripp RA, Chu C, Quelle FW, Nosaka T, Vignali DA, Doherty PC, Grosveld G, Paul WE, Ihle JN. 1996. Lack of IL-4-induced Th2 response and IgE class switching in mice with disrupted Stat6 gene. Nature 380:630 Murphy E, Shibuya K, Hosken N, Openshaw P, Maino V, Davis K, Murphy K, O’Garra A. 1996. Reversibility of T helper 1 and 2 populations is lost after long-term stimulation. J. Exp. Med. 183:901 Alexander WS, Starr R, Fenner JE, et al. 1999. SOCS1 is a critical inhibitor of interferon c signaling and prevents the potentially fatal neonatal actions of this cytokine. Cell 98:597 Murphy KM, Murphy TL, Szabo SJ, Jacobson NG, Guler ML, Gorham JD, Gubler U. 1997. Regulation of IL-12 receptor expression in early T-helper responses implies two phases of Th1 differentiation: capacitance and development. Chem. Immunol. 68:54 Huang H, Paul WE. 1998. Impaired interleukin 4 signaling in T helper type 1 cells. J. Exp. Med. 187:1305 Losman JA, Chen XP, Hilton D, Rothman P. 1999. Cutting edge: SOCS-1 is a potent inhibitor of IL-4 signal transduction. J. Immunol. 162:3770 Yoshimura A, Ohkubo T, Kiguchi T, Jenkins NA, Gilbert DJ, Copeland NG, Hara T, Miyajima A. 1995. A novel cytokineinducible gene CIS encodes an SH2containing protein that binds to tyrosine-phosphorylated interleukin 3 and erythropoietin receptors. EMBO J. 14:2816 Starr R, Willson TA, Viney EM, Murray LJ, Rayner JR, Jenkins BJ, Gonda TJ, Alexander WS, Metcalf D, Nicola NA, Hilton DJ. 1997. A family of cytokine-


184.

185.


186.

187.

188.

189.

190.

inducible inhibitors of signalling. Nature 387:917 Endo TA, Masuhara M, Yokouchi M, Suzuki R, Sakamoto H, Mitsui K, Matsumoto A, Tanimura S, Ohtsubo M, Misawa H, Miyazaki T, Leonor N, Taniguchi T, Fujita T, Kanakura Y, Komiya S, Yoshimura A. 1997. A new protein containing an SH2 domain that inhibits JAK kinases. Nature 387:921 Naka T, Narazaki M, Hirata M, Matsumoto T, Minamoto S, Aono A, Nishimoto N, Kajita T, Taga T, Yoshizaki K, Akira S, Kishimoto T. 1997. Structure and function of a new STAT-induced STAT inhibitor. Nature 387:924 Minamoto S, Ikegame K, Ueno K, Narazaki M, Naka T, Yamamoto H, Matsumoto T, Saito H, Hosoe S, Kishimoto T. 1997. Cloning and functional analysis of new members of STAT induced STAT inhibitor (SSI) family: SSI-2 and SSI-3. Biochem. Biophys. Res. Commun. 237:79 Masuhara M, Sakamoto H, Matsumoto A, Suzuki R, Yasukawa H, Mitsui K, Wakioka T, Tanimura S, Sasaki A, Misawa H, Yokouchi M, Ohtsubo M, Yoshimura A. 1997. Cloning and characterization of novel CIS family genes. Biochem. Biophys. Res. Commun. 239:439 Hilton DJ, Richardson RT, Alexander WS, Viney EM, Willson TA, Sprigg NS, Starr R, Nicholson SE, Metcalf D, Nicola NA. 1998. Twenty proteins containing a C-terminal SOCS box form five structural classes. PNAS 95:114 Vasiliauskas D, Hancock S, Stern CD. 1999.SWiP-1: novel SOCS box containing WD-protein regulated by signalling centres and by Shh during development. Mech. Dev. 82:79 Dey BR, Spence SL, Nissley P, Furlanetto RW. 1998. Interaction of human suppressor of cytokine signaling (SOCS)-2 with the insulin-like growth factor-I receptor. J. Biol. Chem. 273:24095

487

191. Adams TE, Hansen JA, Starr R, Nicola NA, Hilton DJ, Billestrup N. 1998. Growth hormone preferentially induces the rapid, transient expression of SOCS3, a novel inhibitor of cytokine receptor signaling. J. Biol. Chem. 273:1285 192. Bjorbaek C, Elmquist JK, Frantz JD, Shoelson SE, Flier JS. 1998. Identification of SOCS-3 as a potential mediator of central leptin resistance. Mol. Cell 1:619 193. Helman D, Sandowski Y, Cohen Y, Matsumoto A, Yoshimura A, Merchav S, Gertler A. 1998. Cytokine-inducible SH2 protein (CIS3) and JAK2 binding protein (JAB) abolish prolactin receptormediated STAT5 signaling. FEBS Lett. 441:287 194. Sakamoto H, Yasukawa H, Masuhara M, Tanimura S, Sasaki A, Yuge K, Ohtsubo M, Ohtsuka A, Fujita T, Ohta T, Furukawa Y, Iwase S, Yamada H, Yoshimura A. 1998. A Janus kinase inhibitor, JAB, is an interferon-gamma-inducible gene and confers resistance to interferons. Blood 92:1668 195. Nicholson SE, Willson TA, Farley A, Starr R, Zhang JG, Baca M, Alexander WS, Metcalf D, Hilton DJ, Nicola NA. 1999. Mutational analyses of the SOCS proteins suggest a dual domain requirement but distinct mechanisms for inhibition of LIF and IL-6 signal transduction. EMBO J. 18:375 196. Narazaki M, Fujimoto M, Matsumoto T, Morita Y, Saito H, Kajita T, Yoshizaki K, Naka T, Kishimoto T. 1998. Three distinct domains of SSI-1/SOCS-1/JAB protein are required for its suppression of interleukin 6 signaling. PNAS 95:13130 197. Yasukawa H, Misawa H, Sakamoto H, Masuhara M, Sasaki A, Wakioka T, Ohtsuka S, Imaizumi T, Matsuda T, Ihle JN, Yoshimura A. 1999. The JAK-binding protein JAB inhibits Janus tyrosine kinase activity through binding in the activation loop. EMBO J. 18:1309 198. Verdier F, Chretien S, Muller O, Varlet P,

488

199.


200.

201.

202.

203.

204.

MURPHY ET AL Yoshimura A, Gisselbrecht S, Lacombe C, Mayeux P. 1998. Proteasomes regulate erythropoietin receptor and signal transducer and activator of transcription 5 (STAT5) activation. Possible involvement of the ubiquitinated Cis protein. J. Biol. Chem. 273:28185 Kamura T, Sato S, Haque D, Liu L, Kaelin WGJ, Conaway RC, Conaway JW. 1998. The Elongin BC complex interacts with the conserved SOCS-box motif present in members of the SOCS, ras, WD-40 repeat, and ankyrin repeat families. Genes Dev. 12:3872 Zhang JG, Farley A, Nicholson SE, Willson TA, Zugaro LM, Simpson RJ, Moritz RL, Cary D, Richardson R, Hausmann G, Kile BJ, Kent SB, Alexander WS , Metcalf D, Hilton DJ, Nicola NA, Baca M. 1999. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. PNAS 96:2071 Stebbins CE, Kaelin WGJ, Pavletich NP. 1999. Structure of the VHL-ElonginCElonginB complex: implications for VHL tumor suppressor function. Science 284:455 Verdier F, Rabionet R, Gouilleux F, Beisenherz-Huss C, Varlet P, Muller O, Mayeux P, Lacombe C, Gisselbrecht S, Chretien S. 1998. A sequence of the CIS gene promoter interacts preferentially with two associated STAT5A dimers: a distinct biochemical difference between STAT5A and STAT5B. Mol. Cell. Biol. 18:5852 Matsumoto A, Masuhara M, Mitsui K, Yokouchi M, Ohtsubo M, Misawa H, Miyajima A, Yoshimura A. 1997. CIS, a cytokine inducible SH2 protein, is a target of the JAK-STAT5 pathway and modulates STAT5 activation. Blood 89:3148 Auernhammer CJ, Bousquet C, Melmed S, 1999. Autoregulation of pituitary corticotroph SOCS-3 expression: character-

205.

206.

207.

208.

209.

210.

211.

212.

ization of the murine SOCS-3 promoter. PNAS 96:6964 Song MM, Shuai K. 1998. The suppressor of cytokine signaling (SOCS) 1 and SOCS3 but not SOCS2 proteins inhibit interferon-mediated antiviral and antiproliferative activities. J. Biol. Chem. 273:35056 Venkataraman C, Leung S, Salvekar A, Mano H, Schindler U. 1999. Repression of IL-4-induced gene expression by IFNgamma requires Stat1 activation. J. Immunol. 162:4053 Dickensheets HL, Donnelly RP. 1999. Inhibition of IL-4-inducible gene expression in human monocytes by type I and type II interferons. J. Leuk. Biol. 65:307 Starr R, Metcalf D, Elefanty AG, Brysha M, Willson TA, Nicola NA, Hilton DJ, Alexander WS. 1998. Liver degeneration and lymphoid deficiencies in mice lacking suppressor of cytokine signaling-1. PNAS 95:14395 Naka T, Matsumoto T, Narazaki M, Fujimoto M, Morita Y, Ohsawa Y, Saito H, Nagasawa T, Uchiyama Y, Kishimoto T. 1998. Accelerated apoptosis of lymphocytes by augmented induction of Bax in SSI-1 (STAT-induced STAT inhibitor-1) deficient mice. PNAS 95:15577 Metcalf D, Alexander WS, Elefanty AG, Nicola NA, Hilton DJ, Starr R, Mifsud S, Di Rago L. 1999. Aberrant hematopoiesis in mice with inactivation of the gene encoding SOCS-1. Leukemia 13:926 Cohney SJ, Sanden D, Cacalano NA, Yoshimura A, Mui A, Migone TS, Johnston JA. 1999. SOCS-3 is tyrosine phosphorylated in response to interleukin-2 and suppresses STAT5 phosphorylation and lymphocyte proliferation. Mol. Cell. Biol. 19, 4980–88. 7–99 Bach EA, Szabo SJ, Dighe AS, Ashkenazi A, Aguet M, Murphy KM, Schreiber RD. 1995. Ligand-induced autoregula-


213.

214.


215.

216.

217.

218.

219.

220.

221.

tion of IFN-gamma receptor beta chain expression in T helper cell subsets. Science 270:1215 Pernis A, Gupta S, Gollob KJ, Garfein E, Coffman RL, Schindler C, Rothman P. 1995. Lack of interferon gamma receptor beta chain and the prevention of interferon gamma signaling in TH1 cells [see comments]. Science 269:245 Kubo M, Ransom J, Webb D, Hashimoto Y, Tada T, Nakayama T. 1997. T-cell subset-specific expression of the IL-4 gene is regulated by a silencer element and STAT6. EMBO J. 16:4007 Szabo SJ, Glimcher LH, Ho IC. 1997. Genes that regulate interleukin-4 expression in T cells. Curr. Opinion Immunol. 9:776 Rao A, Luo C, Hogan PG. 1997. Transcription factors of the NFAT family: regulation and function. Annu. Rev. Immunol. 15:707 Xanthoudakis S, Viola JP, Shaw KT, Luo C, Wallace JD, Bozza PT, Luk DC, Curran T, Rao A. 1996. An enhanced immune response in mice lacking the transcription factor NFAT1 Science 272:892 Hodge MR, Ranger AM, Charles dlB, Hoey T, Grusby MJ, Glimcher LH. 1996. Hyperproliferation and dysregulation of IL-4 expression in NF-ATp-deficient mice. Immunity 4:397 Kiani A, Viola JP, Lichtman AH, Rao A. 1997. Down-regulation of IL-4 gene transcription and control of Th2 cell differentiation by a mechanism involving NFAT1. Immunity 7:849 Viola JP, Kiani A, Bozza PT, Rao A. 1998. Regulation of allergic inflammation and eosinophil recruitment in mice lacking the transcription factor NFAT1: role of interleukin-4 (IL-4) and IL-5. Blood 91:2223 Heyer J, Kneitz B, Schuh K, Jankevics E, Siebelt F, Schimpl A, Serfling E. 1997. Inefficient termination of antigen

222.

223.

224.

225.

226.

227.

228.

229.

489

responses in NF-ATp-deficient mice. Immunobiology 198:162 Schuh K, Kneitz B, Heyer J, Bommhardt U, Jankevics E, Berberich-Siebelt F, Pfeffer K, Muller-Hermelink HK, Schimpl A, Serfling E. 1998. Retarded thymic involution and massive germinal center formation in NF-ATp-deficient mice. Eur. J. Immunol. 28:2456 Ranger AM, Hodge MR, Gravallese EM, Oukka M, Davidson L, Alt FW, de la Brousse FC, Hoey T, Grusby M, Glimcher LH. 1998. Delayed lymphoid repopulation with defects in IL-4-driven responses produced by inactivation of NF-ATc. Immunity i:125 Yoshida H, Nishina H, Takimoto H, Marengere LE, Wakeham AC, Bouchard D, Kong YY, Ohteki T, Shahinian A, Bachmann M, Ohashi PS, Penninger JM, Crabtree GR, Mak TW. 1998. The transcription factor NF-ATc1 regulates lymphocyte proliferation and Th2 cytokine production. Immunity 8:115 Oukka M, Ho IC, de la Brousse FC, Hoey T, Grusby MJ, Glimcher LH. 1998. The transcription factor NFAT4 is involved in the generation and survival of T cells. Immunity 9:295 Ranger AM, Oukka M, Rengarajan J, Glimcher LH. 1998. Inhibitory function of two NFAT family members in lymphoid homeostasis and Th2 development. Immunity 9:627 Garcia-Rodriguez C, Rao A. 1998. Nuclear factor of activated T cells (NFAT)-dependent transactivation regulated by the coactivators p300/CREBbinding protein (CBP). J. Exp. Med. 187:2031 Chuvpilo S, Zimmer M, Kerstan A, Glockner J, Avots A, Escher C, Fischer C, Inashkina I, Jankevics E, BerberichSiebelt F, Schmitt E, Serfling E. 1999. Alternative polyadenylation events contribute to the induction of NF-ATc in effector T cells. Immunity 10:261 Kim JI, Li T, Ho IC, Grusby MJ,

490

230.

231.


232.

233.

234.

235.

236.

237.

MURPHY ET AL Glimcher LH. 1999. Requirement for the c-Maf transcription factor in crystallin gene regulation and lens development. PNAS 96:3781 Kim JI, Ho IC, Grusby MJ, Glimcher LH. 1999. The transcription factor c-Maf controls the production of interleukin-4 but not other Th2 cytokines. Immunity 10:745–51 Dent AL, Shaffer AL, Yu X, Allman D, Staudt LM. 1997. Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 276:589 Dent AL, Hu-Li J, Paul WE, Staudt LM. 1998. T helper type 2 inflammatory disease in the absence of interleukin 4 and transcription factor STAT6. PNAS 95:13823 Rincon M, Flavell RA. 1997. Transcription mediated by NFAT is highly inducible in effector CD4` T helper 2 (Th2) cells but not in Th1 cells. Mol. Cell. Biol. 17:1522 Rincon M, Derijard B, Chow CW, Davis RJ, Flavell RA. 1997. Reprogramming the signalling requirement for AP-1 (activator protein-1) activation during differentiation of precursor CD4` T-cells into effector Th1 and Th2 cells. Genes Function 1:51 Szabo SJ, Gold JS, Murphy TL, Murphy KM. 1993. Identification of cis-acting regulatory elements controlling interleukin-4 gene expression in T cells: roles for NF-Y and NF-ATc. Mol. Cell. Biol. 13:4793 Li-Weber M, Salgame P, Hu C, Davydov LV, Laur O, Klevenz S, Krammer PH. 1998. Th2-specific protein/DNA interactions at the proximal nuclear factor-AT site contribute to the functional activity of the human IL-4 promoter. J. Immunol. 161:1380 Chen L, Glover JN, Hogan PG, Rao A, Harrison SC. 1998. Structure of the DNA-binding domains from NFAT, Fos

238.

239.

240.

241.

242.

243.

244.

245.

246.

and Jun bound specifically to DNA. Nature 392:42 Li B, Tournier C, Davis RJ, Flavell RA. 1999. Regulation of IL-4 expression by the transcription factor JunB during T helper cell differentiation. EMBO J. 18:420 Pandolfi PP, Roth ME, Karis A, Leonard MW, Dzierzak E, Grosveld FG, Engel JD, Lindenbaum MH. 1995. Targeted disruption of the GATA3 gene causes severe abnormalities in the nervous system and in fetal liver haematopoiesis. Nat. Genet. 11:40 Ting CN, Olson MC, Barton KP, Leiden JM. 1996. Transcription factor GATA-3 is required for development of the T-cell lineage. Nature 384:474 Hendriks RW, Nawijn MC, Engel JD, van Doorninck H, Grosveld F, Karis A. 1999. Expression of the transcription factor GATA-3 is required for the development of the earliest T cell progenitors and correlates with stages of cellular proliferation in the thymus. Eur. J. Immunol. 29(6):1912–18 Zheng W, Flavell RA. 1997. The transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4 T cells. Cell 89:587 Zhang DH, Cohn L, Ray P, Bottomly K, Ray A. 1997. Transcription factor GATA-3 is differentially expressed in murine Th1 and Th2 cells and controls Th2-specific expression of the interleukin-5 gene. J. Biol. Chem. 272:21597 Zhang DH, Yang L, Ray A. 1998. Differential responsiveness of the IL-5 and IL-4 genes to transcription factor GATA3. J. Immunol. 161:3817 Siegel MD, Zhang DH, Ray P, Ray A. 1995. Activation of the interleukin-5 promoter by cAMP in murine EL-4 cells requires the GATA-3 and CLE0 elements. J. Biol. Chem. 270:24548 Lee HJ, O’Garra A, Arai K, Arai N. 1998. Characterization of cis-regulatory elements and nuclear factors conferring


247.

248.


249.

250.

251.

252.

253.

254.

255.

Th2-specific expression of the IL-5 gene: a role for a GATA-binding protein. J. Immunol. 160:2343 Ranganath S, Ouyang W, Bhattarcharya D, Sha WC, Grupe A, Peltz G, Murphy KM. 1998. GATA-3-dependent enhancer activity in IL-4 gene regulation. J. Immunol. 161:3822 Ferber IA, Lee HJ, Zonin F, Heath V, Mui A, Arai N, O’Garra A. 1999. GATA-3 significantly downregulates IFN-gamma production from developing Th1 cells in addition to inducing IL-4 and IL-5 levels. Clin. Immunol. 91:134 Nakamura Y, Ghaffar O, Olivenstein R, Taha RA, Soussi-Gounni A, Zhang DH, Ray A, Hamid Q. 1999. Gene expression of the GATA-3 transcription factor is increased in atopic asthma. J. Allergy Clin. Immunol. 103:215 Curiel RE, Lahesmaa R, Subleski J, Cippitelli M, Kirken RA, Young HA, Ghosh P. 1997. Identification of a Stat-6-responsive element in the promoter of the human interleukin-4 gene. Eur. J. Immunol. 27:1982 Georas SN, Cumberland JE, Burke TF, Chen R, Schindler U, Casolaro V. 1998. Stat6 inhibits human interleukin-4 promoter activity in T cells. Blood 92:4529 Lederer JA, Perez VL, DesRoches L, Kim SM, Abbas AK, Lichtman AH. 1996. Cytokine transcriptional events during helper T cell subset differentiation. J. Exp. Med. 184:397 Minty A, Chalon P, Derocq JM, Dumont X, Guillemot JC, Kaghad M, Labit C, Leplatois P, Liauzun P, Miloux B. 1993. Interleukin-13 is a new human lymphokine regulating inflammatory and immune responses. Nature 362:248 Chomarat P, Banchereau J. 1998. Interleukin-4 and interleukin-13: their similarities and discrepancies. Int. Rev. Immunol. 17:1 Wills-Karp M, Luyimbazi J, Xu X, Schofield B, Neben TY, Karp CL, Donaldson

256.

257.

258.

259.

260.

261.

262.

263.

264.

491

DD. 1998. Interleukin-13: central mediator of allergic asthma. Science 282:2258 Wills-Karp M. 1999. Immunologic basis of antigen-induced airway hyperresponsiveness. Annu. Rev. Immunol. 17:255– 81 McKenzie GJ, Emson CL, Bell SE, Anderson S, Fallon P, Zurawski G, Murray R, Grencis R, McKenzie AN. 1998. Impaired development of Th2 cells in IL13-deficient mice. Immunity 9:423 Graber P, Gretener D, Herren S, Aubry JP, Elson G, Poudrier J, Lecoanet-Henchoz S, Alouani S, Losberger C, Bonnefoy JY, Kosco-Vilbois MH, Gauchat JF. 1998. The distribution of IL-13 receptor alpha1 expression on B cells, T cells and monocytes and its regulation by IL13 and IL-4. Eur. J. Immunol. 28:4286 Urban JFJ, Maliszewski CR, Madden KB, Katona IM, Finkelman FD. 1995. IL-4 treatment can cure established gastrointestinal nematode infections in immunocompetent and immunodeficient mice. J. Immunol. 154:4675 Lawrence RA, Gray CA, Osborne J, Maizels RM. 1996. Nippostrongylus brasiliensis: cytokine responses and nematode expulsion in normal and IL-4-deficient mice. Exp. Parasitol. 84:65 Murphy KM. 1998. T lymphocyte differentiation in the periphery. Curr. Opin. Immunol. 10:226 Noben-Trauth N, Shultz LD, Brombacher F, Urban JFJ, Gu H, Paul WE. 1997. An interleukin 4 (IL-4)-independent pathway for CD4` T cell IL-4 production is revealed in IL-4 receptordeficient mice. PNAS 94:10838 Urban JFJ, Noben-Trauth N, Donaldson DD, Madden KB, Morris SC, Collins M, Finkelman FD. 1998. IL-13, IL-4Ralpha, and Stat6 are required for the expulsion of the gastrointestinal nematode parasite Nippostrongylus brasiliensis. Immunity 8:255 Barner M, Mohrs M, Brombacher F, Kopf M. 1998. Differences between IL-

492

265.

266.


267.

268.

269.

270.

271.

272.

273.

274.

MURPHY ET AL 4R alpha-deficient and IL-4-deficient mice reveal a role for IL-13 in the regulation of Th2 responses. Curr. Biol. 8:669 Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal M, Liu YJ. 1999. Reciprocal control of T helper cell and dendritic cell differentiation [see comments]. Science 283:1183 Pulendran B, Smith JL, Caspary G, Brasel K, Pettit D, Maraskovsky E, Maliszewski CR. 1999. Distinct dendritic cell subsets differentially regulate the class of immune response in vivo. PNAS 96:1036 Siegal FP, Kadowaki N, Shodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, Antonenko S, Liu YJ. 1999. The nature of the principal type 1 interferon-producing cells in human blood. Science 284:1835 Seder RA, Boulay JL, Finkelman F, Barbier S, Ben-Sasson SZ, Le Gros G, Paul WE. 1992. CD8` T cells can be primed in vitro to produce IL-4. J. Immunol. 148:1652 Le Gros G, Erard F. 1994. Non-cytotoxic, IL-4, IL-5, IL-10 producing CD8` T cells: their activation and effector functions. Curr. Opin. Immunol. 6:453 Mosmann TR, Li L, Sad S. 1997. Functions of CD8 T-cell subsets secreting different cytokine patterns. Semin. Immunol. 9:87 Sad S, Marcotte R, Mosmann TR. 1995. Cytokine-induced differentiation of precursor mouse CD8` T cells into cytotoxic CD8` T cells secreting Th1 or Th2 cytokines. Immunity 2:271 Croft M, Carter L, Swain SL, Dutton RW. 1994. Generation of polarized antigen-specific CD8 effector populations: reciprocal action of interleukin (IL)-4 and IL-12 in promoting type 2 versus type 1 cytokine profiles. J. Exp. Med. 180:1715 Fowell DJ, Magram J, Turck CW, Killeen N, Locksley RM. 1997. Impaired Th2 subset development in the absence of CD4. Immunity 6:559 Brown DR, Moskowitz NH, Killeen N,

275.

276.

277.

278.

279.

280.

281.

282.

283.

284.

Reiner SL. 1997. A role for CD4 in peripheral T cell differentiation. J. Exp. Med. 186:101 Veillette A, Bookman MA, Horak EM, Bolen JB. 1988. The CD4 and CD8 T cell surface antigens are associated with the internal membrane tyrosine-protein kinase p56lck. Cell 55:301 Turner JM, Brodsky MH, Irving BA, Levin SD, Perlmutter RM, Littman DR. 1990. Interaction of the unique N-terminal region of tyrosine kinase p56lck with cytoplasmic domains of CD4 and CD8 is mediated by cysteine motifs. Cell 60:755 Yamashita M, Hashimoto K, Kimura M, Kubo M, Tada T, Nakayama T. 1998. Requirement for p56(lck) tyrosine kinase activation in Th subset differentiation. Int. Immunol. 10:577 Yu CL, Jove R, Burakoff SJ. 1997. Constitutive activation of the Janus kinaseSTAT pathway in T lymphoma overexpressing the Lck protein tyrosine kinase. J. Immunol. 159:5206 Heyeck SD, Wilcox HM, Bunnell SC, Berg LJ. 1997. Lck phosphorylates the activation loop tyrosine of the Itk kinase domain and activates Itk kinase activity. J. Biol. Chem. 272:25401 Liao XC, Littman DR. 1995. Altered T cell receptor signaling and disrupted T cell development in mice lacking Itk. Immunity 3:757 Liu KQ, Bunnell SC, Gurniak CB, Berg LJ. 1998. T cell receptor-initiated calcium release is uncoupled from capacitative calcium entry in Itk-deficient T cells. J. Exp. Med. 187:1721 Gajewski TF, Schell SR, Fitch FW. 1990. Evidence implicating utilization of different T cell receptor-associated signaling pathways by TH1 and TH2 clones. J. Immunol. 144:4110 Sloan-Lancaster J, Steinberg TH, Allen PM. 1997. Selective loss of the calcium ion signaling pathway in T cells maturing toward a T helper 2 phenotype. J. Immunol. 159:1160 Bachmann MF, Mariathasan S, Bouchard


285.

286.


287.

288.

289.

290.

291.

292.

D, Speiser DE, Ohashi PS. 1997. Four types of Ca2` signals in naive CD8` cytotoxic T cells after stimulation with T cell agonists, partial agonists and antagonists. Eur. J. Immunol. 27:3414 Esser MT, Haverstick DM, Fuller CL, Gullo CA, Braciale VL. 1998. Ca2` signaling modulates cytolytic T lymphocyte effector functions. J. Exp. Med. 187:1057 Timmerman LA, Clipstone NA, Ho SN, Northrop JP, Crabtree GR. 1996. Rapid shuttling of NF-AT in discrimination of Ca2` signals and immunosuppression. Nature 383:837 Dolmetsch RE, Lewis RS, Goodnow CC, Healy JI. 1997. Differential activation of transcription factors induced by Ca2` response amplitude and duration [see comments] [published erratum appears in Nature 1997 Jul 17; 388(6639):308]. Nature 386:855 Harris W, Gollapudi S, Gupta S. 1995. Anti-CD3-induced changes in protein kinase C isozymes expression in human CD4` and CD8` T lymphocytes. J. Clin. Immunol. 15:232 Yamashita M, Kimura M, Kubo M, Shimizu C, Tada T, Perlmutter RM, Nakayama T. 1999. T cell antigen receptor-mediated activation of the Ras/ mitogen-activated protein kinase pathway controls interleukin 4 receptor function and type-2 helper T cell differentiation. PNAS 96:1024 Dasen JS, O’Connell SM, Flynn SE, Treier M, Gleiberman AS, Szeto DP, Hooshmand F, Aggarwal AK, Rosenfeld MG, Rosenfeld M. 1999. Reciprocal interactions of Pit1 and GATA2 mediate signaling gradient-induced determination of pituitary cell types. Cell 97:587 Coffman RL, Mocci S, O’Garra A. 1999. The stability and reversibility of Th1 and Th2 populations. Curr. Top. Microbiol. Immunol. 238:1 Minasi LE, Kamogawa Y, Carding S, Bottomly K, Flavell RA. 1993. The selective ablation of interleukin 2-pro-

293.

294.

295.

296.

297.

298.

299.

300.

301.

302.

493

ducing cells isolated from transgenic mice. J. Exp. Med. 177:1451 Kamogawa Y, Minasi LA, Carding SR, Bottomly K, Flavell RA. 1993. The relationship of IL-4- and IFN gamma-producing T cells studied by lineage ablation of IL-4-producing cells. Cell 75:985 Nakamura T, Kamogawa Y, Bottomly K, Flavell RA. 1997. Polarization of IL-4and IFN-gamma-producing CD4` T cells following activation of naive CD4` T cells. J. Immunol. 158:1085 Szabo SJ, Dighe AS, Gubler U, Murphy KM. 1997. Regulation of the interleukin (IL)-12R beta 2 subunit expression in developing T helper 1 (Th1) and Th2 cells. J. Exp. Med. 185:817 Mocci S, Coffman RL. 1995. Induction of a Th2 population from a polarized Leishmania-specific Th1 population by in vitro culture with IL-4. J. Immunol. 154:3779 Mocci S, Coffman RL. 1997. The mechanism of in vitro T helper cell type 1 to T helper cell type 2 switching in highly polarized Leishmania major-specific T cell populations. J. Immunol. 158:1559 Hollander GA, Zuklys S, Morel C, Mizoguchi E, Mobisson K, Simpson S, Terhorst C, Wishart W, Golan DE, Bhan AK, Burakoff SJ. 1998. Monoallelic expression of the interleukin-2 locus [see comments]. Science 279:2118 Shibahara K, Stillman B. 1999. Replication-dependent marking of DNA by PCNA facilitates CAF-1-coupled inheritance of chromatin. Cell 96:575 Bix M, Locksley RM. 1998. Independent and epigenetic regulation of the interleukin-4 alleles in CD4` T cells. Science 281:1352 Naramura M, Hu RJ, Gu H. 1998. Mice with a fluorescent marker for interleukin 2 gene activation. Immunity 9:209 Riviere I, Sunshine MJ, Littman DR. 1998. Regulation of IL-4 expression by

494

303.

304.

305.


306.

MURPHY ET AL activation of individual alleles. Immunity 9:217 Bird JJ, Brown DR, Mullen AC, Moskowitz NH, Mahowald MA, Sider JR, Gajewski TF, Wang CR, Reiner SL. 1998. Helper T cell differentiation is controlled by the cell cycle. Immunity 9:229 Gett AV, Hodgkin PD. 1998. Cell division regulates the T cell cytokine repertoire, revealing a mechanism underlying immune class regulation. PNAS 95:9488 Takemoto N, Koyano-Nakagawa N, Yokota T, Arai N, Miyatake S, Arai K. 1998. Th2-specific DNase I-hypersensitive sites in the murine IL-13 and IL-4 intergenic region. Int. Immunol. 10:1981 Ye J, Young HA. 1997. Negative regulation of cytokine gene transcription. FASEB J. 11:825

307. Fitzpatrick DR, Shirley KM, Kelso A. 1999. Cutting edge: stable epigenetic inheritance of regional IFN-gamma promoter demethylation in CD44highCD8` T lymphocytes. J. Immunol. 162:5053 308. Meraz MA, White JM, Sheehan KC, Bach EA, Rodig SJ, Dighe AS, Kaplan DH, Riley JK, Greenlund AC, Campbell D, Carver-Moore K, DuBois RN, Clark R, Aguet M, Schreiber RD. 1996. Targeted disruption of the Stat1 gene in mice reveals unexpected physiologic specificity in the JAK-STAT signaling pathway. Cell 84:431 309. Kopf M, Le Gros G, Bachmann M, Lamers MC, Bluethmann H, Kohler G. 1993. Disruption of the murine IL-4 gene blocks Th2 cytokine responses. Nature 362:245


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


THE RAG PROTEINS AND V(D)J RECOMBINATION: Complexes, Ends, and Transposition Sebastian D. Fugmann1, Alfred Ian Lee, Penny E. Shockett, Isabelle J. Villey, and David G. Schatz


Howard Hughes Medical Institute, Section of Immunobiology, Yale University School of Medicine, New Haven, Connecticut 06520–8011; e-mail: [email protected]

Key Words antigen receptor, site-specific recombination, transposition, RAG1, RAG2 Abstract V(D)J recombination proceeds through a series of protein:DNA complexes mediated in part by the RAG1 and RAG2 proteins. These proteins are responsible for sequence-specific DNA recognition and DNA cleavage, and they appear to perform multiple postcleavage roles in the reaction as well. Here we review the interaction of the RAG proteins with DNA, the chemistry of the cleavage reaction, and the higher order complexes in which these events take place. We also discuss postcleavage functions of the RAG proteins, including recent evidence indicating that they initiate the process of coding end processing by nicking hairpin DNA termini. Finally, we discuss the evolutionary and functional implications of the finding that RAG1 and RAG2 constitute a transposase, and we consider RAG protein biochemistry in the context of several bacterial transposition systems. This suggests a model of the RAG protein active site in which two divalent metal ions serve alternating and opposite roles as activators of attacking hydroxyl groups and stabilizers of oxyanion leaving groups.

INTRODUCTION The genes encoding immunoglobulin and T cell receptor proteins are unique in being split into multiple gene segments in the germline that are then made contiguous by recombination in somatic tissues. The assembly process, known as V(D)J recombination, is named for the V (variable), D (diversity), and J (joining) gene segments that are the targets of the reaction. For assembly of an antigen receptor gene, one V, one J, and in some cases one D gene segment are joined by recombination to create an exon that encodes the antigen binding portion of 1

The first four authors are listed alphabetically and contributed equally to this work.

0732–0582/00/0410–0495$14.00

495


496

FUGMANN ET AL

the receptor chain. Following transcription, the V(D)J exon is spliced to one or more exons encoding the constant region to produce the mature mRNA and subsequently the receptor polypeptide. In many species, including primates and rodents, V(D)J recombination is responsible for generating a great deal of the diversity found in these receptors. Such diversity arises from two sources. The first, combinatorial diversity, is a consequence of the fact that there are typically many different V, D, and J gene segments and that each different V(D)J combination yields a different receptor specificity. The second, junctional diversity, arises from imprecise joining of the V, D, and J gene segments. The sites of recombination are specified by recombination signal sequences (RSSs) that immediately flank each gene segment. Each RSS consists of a highly conserved 7-bp sequence (the heptamer; consensus 58-CACAGTG) and an ATrich 9-bp sequence (the nonamer; consensus 58-ACAAAAACC) that are separated by a poorly conserved spacer whose length is either 12 5 1 or 23 5 1 bp. Spacer length therefore defines two types of RSSs, termed the 12-RSS and the 23-RSS. Efficient recombination occurs only between a 12-RSS and a 23-RSS, a restriction known as the 12/23 rule. V(D)J recombination of the endogenous antigen receptor gene loci is a complex and highly regulated process. Events can span more than a megabase and can occur in lineage and developmental stage-specific patterns. Factors such as nucleosomal positioning and higher order chromatin structure are likely to play critical roles in regulating the reaction (for reviews, see 1, 2). Relatively little is understood about how these forces regulate V(D)J recombination, and they are not discussed further here. Instead, we consider the reaction at a much simplified level, that of artificial DNA substrates containing one or two RSSs, typically analyzed in cell-free reaction systems. At this level, the reaction can be considered to occur in two phases (Figure 1). In the first, the two RSSs are recognized by the recombination machinery, brought into close juxtaposition (synapsis), and the DNA is cleaved precisely between the RSSs and their flanking coding elements. This generates four free ends: two blunt, 58-phosphorylated signal ends and two covalently sealed, hairpin coding ends. In the second phase of the reaction, the coding ends are processed, often with the loss and addition of a small number of nucleotides, and joined to form a coding joint (CJ), while the signal ends are joined, typically precisely, to form a signal joint. It is also possible, though less frequent, for a signal end to become joined to a coding end. If the signal end is joined to the coding end to which it was previously connected, an open/shut (O/S) joint results, whereas if it is joined to the opposite coding end, a hybrid joint results. The first phase of V(D)J recombination can be performed in its entirety (3, 4) by the proteins encoded by the recombination activating genes RAG1 and RAG2 (5, 6). The high mobility group proteins –1 and –2 (HMG1/2) also make an important contribution to this phase of the reaction (7, 8). The RAG1 and RAG2 proteins are coexpressed only in cells of the B and T lymphocyte lineages and


RAG PROTEINS AND V(D)J RECOMBINATION

497

Figure 1 Schematic model of the protein-DNA complexes in V(D)J recombination. See text for details. The 12-RSS and 23-RSS are represented as black and white triangles, respectively, coding segments as rectangles and proteins as shaded ovals. Several aspects of the reaction are not depicted, including nicking adjacent to RSSs (which may occur before or after synapsis), asymmetric opening of the hairpin coding ends to generate Pnucleotides, and nucleotide addition by TdT. Coding end processing likely occurs in the context of the CSC, but the existence of a complex containing just the coding ends (brackets) cannot be ruled out. Figure adapted from (24) and relies also on data from (7, 31, 65).


498

FUGMANN ET AL

are each essential for cleavage activity (9–11). Mutagenesis studies have defined minimal ‘‘core’’ regions of each RAG protein required for catalytic activity (12– 15). While it is clear that the ‘‘nonessential’’ regions of the RAG proteins make important contributions to their activity (16–19), these roles are not well understood. One striking feature of the RAG genes is that they are located immediately adjacent to one another in the genomes of all jawed vertebrate species examined to date, and in most species they lack introns in their coding regions (20). The second phase of V(D)J recombination is not well understood but appears to require RAG1, RAG2, and a group of ubiquitous DNA repair proteins. The enzyme terminal deoxynucleotidyl transferase (TdT) is not essential, but when present it contributes substantially to receptor diversity by adding nontemplated (N) nucleotides to coding junctions (21, 22). The essential DNA repair proteins include the three components of the DNA-dependent protein kinase (DNAPK), Ku70, Ku80, and the kinase catalytic subunit DNAPKcs, and the XRCC4 and DNA ligase IV proteins (23 and references therein). A deficiency in any of these factors results in defective V(D)J recombination, hypersensitivity to ionizing radiation, and an early block in lymphocyte development. A full consideration of these proteins is beyond the scope of this review. Recent biochemical studies have demonstrated that V(D)J recombination probably proceeds through a series of well-defined protein:DNA complexes (Figure 1). The RAG proteins first recognize the 12-RSS and 23-RSS to form stable complexes referred to as the 12-SC and 23-SC (nomenclature adapted from 24). Synapsis of these complexes leads to formation of the paired complex (PC), within which DNA cleavage is completed. After cleavage, the four ends are held in a cleaved signal complex (CSC), which is likely the complex within which much or all of CJ formation occurs. This leaves behind the quite stable signal end complex (SEC), which is the precursor for signal joint formation. In this review, we focus on the properties and activities of the RAG proteins, particularly the protein:DNA complexes in which they participate. We also discuss the recent finding that RAG1 and RAG2 constitute a transposase (25, 26) and compare the properties of this transposase-become-recombinase with several well-studied bacterial transposases.

RECOGNITION OF THE RSS RAG-RSS Interaction The first step in V(D)J recombination generates a stable complex of the RAG proteins and the RSS (12-SC or 23-SC; Figure 1), and both the heptamer and nonamer make important contributions to complex formation. An in vivo onehybrid system and a surface plasmon resonance analysis revealed an essential role of the nonamer for detectable binding (27, 28). This is mediated by an interaction with the NBD (‘‘nonamer-binding-domain’’, aa 390–460) of RAG1. RAG1 alone



499

binds the 12- and 23-RSSs with approximately equal affinities, and the heptamer makes a small but clearly detectable contribution to this interaction (27, 29). RAG2 by itself displays no detectable binding activity, but the 12-SC and 23-SC, containing RAG1 and RAG2, are much more stable and sequence-specific than complexes containing RAG1 alone (30–32). Overall, these results suggest that the RAG-RSS interaction may involve two steps: a ‘‘recruitment’’ step in which the nonamer acts as an anchoring motif; and a ‘‘stabilization’’ step, in which the heptamer, in the presence of RAG2, now supports a larger part of the interactions, closer to the coding flank and the site of cleavage. In all cases, formation of the 12-SC or 23-SC requires a proper spacer length and the presence of a divalent metal cation (31, 33). Interference and Footprinting Studies Interactions of the RAG proteins with the RSS have been studied at the nucleotide level by methylation and ethylation interference and footprinting assays. Interaction of RAG1 with the RSS is clearly evident (30, 34, 35), and ethylation interference studies (34) revealed that this interaction is contributed by nonspecific interactions with the DNA backbone, especially in the spacer region, as well as by base-specific interactions, restricted to the nonamer. Strong overall nonamer occupancy is also observed within the RAG1-RAG2-RSS complex, and in this situation the protection extends through the spacer to the spacer-proximal side of the heptamer (Figure 2, see color insert). In contrast with what was observed with RAG1 alone, interactions outside the nonamer are then more base-specific, consistent with the observation that RAG2 and the heptamer favor a more stable complex. For the 23-RSS, the interaction with the nonamer is evident in a RAG1-RAG2 complex, but the protein:DNA contacts do not propagate as far toward the heptamer as with the 12-RSS. This fits well with the observation that the 23-SC is less stable than the 12-SC (31) and implies that additional factor(s) are required to stabilize the 23SC (see below). Strikingly, in any situation, the interactions are biased toward one face of the helix throughout the RSS (34) (Figure 2). This could explain why cleavage requires appropriate phasing (an integral number of helical turns) between the heptamer and the nonamer. Indeed, changing the length of the spacer by more than 51 bp dramatically reduces cleavage and recombination (36–40). UV Cross-Linking Footprinting studies have failed to reveal any interaction of the RAG proteins with nucleotides near the coding flank/heptamer border. Since this is the site at which cleavage occurs, there must be protein:DNA interactions in this region, but they are likely to be transient or weak. Indeed, several studies have suggested that RAG1 interacts with the coding flank (31, 41, 42). To detect such weak interactions, we designed a UV cross-linking assay that involved substituting 5-iodouridine in place of thymidine at several positions of the coding flank and the heptamer in RSS oligonucleotides (43). RAG1 was shown to be cross-linked at positions flanking the site of cleavage (Figure 2): the first position of the coding flank on the bottom strand (C-1B), the second position of the heptamer on the bottom strand (H2B), and the third position of the heptamer on the


500

FUGMANN ET AL

top strand (H3T). In the presence of RAG1, RAG2 could be cross-linked at the C-1B and H3T positions, although much more weakly than RAG1 (Figure 2). Differences in intensity of cross-linking may simply reflect the selectivity of the assay. Such biases include the facts that the 5-position of pyrimidines lies in the major groove of the DNA helix, and that iodo-groups preferentially cross-link to aromatic residues (44). Cross-linking of RAG1 and RAG2 to the RSS was also observed using aryl-azide-modified bases (45). However, these results are more difficult to interpret because the azido group allows cross-linking to more distant residues. Another study using iodo-groups confirmed the RAG1-heptamer interaction but did not reveal cross-linking to RAG2 (32). Interestingly, the C-1T position (first position of the coding flank on the top strand) was also strongly cross-linked to RAG1 in our assay, but only if the DNA of the coding flank was unpaired (which may mimic the DNA melting that is thought to occur during cleavage). It is reasonable that RAG1 would establish additional DNA contacts near the site of cleavage to stabilize the unpaired state. Severe distortion of the DNA at the coding flank/heptamer border favors binding (34), and unpairing of the coding flank has been shown to enhance cleavage (36, 39). Taken together, these studies suggest that single-stranded coding flank/ heptamer borders are recognized as well as, if not better than, double-stranded borders, which implies that the RAG proteins are able to bind and cleave singlestranded DNA. Consistent with this, cleavage of single-stranded RSSs occurs (36, 39), and our unpublished data show that RAG1 and perhaps RAG2 can be crosslinked by UV light to a single-stranded RSS containing an iodo group at position C-1B (IJVilley and DG Schatz, unpublished). Together, the results demonstrate that the RAG proteins contact and surround the site of cleavage. Stoichiometry and Configuration of Subunits The composition of the RAGRSS complex has been investigated recently, revealing that core RAG1 exists as a homodimer in solution and retains its dimeric form upon binding to the RSS (29, 45, 46), as was suggested by a previous study (33). One study found that RAG2 forms multimers, particularly dimers, in solution, and that the 12-SC consists of a tetramer of two molecules of each RAG protein bound to a 12-RSS (46). This conflicts with other data which suggest that only a monomer of RAG2 is found in the RAG1-RAG2-RSS complex (45). The configuration of the different subunits within the complex remains hypothetical. In a first model, one RAG1 molecule contacts both the nonamer and the heptamer/coding flank. Alternatively, two molecules of RAG1 might interact with one RSS, with one contacting the nonamer and the other the heptamer/coding flank (29, 45).

Role of HMG1/2 How are the RAG proteins able to recognize the 12-RSS and 23-RSS given the extra helical turn found in the latter? Parallels with other recombination systems suggested the involvement of a DNA-bending accessory factor, and it has been



501

shown that HMG1/2, very abundant and ubiquitous DNA-binding and bending proteins, enhance binding and cleavage by the RAG proteins (7, 8, 47, 48). Their effect, however, is different with the two types of RSSs. Formation of the 12-SC with purified RAG proteins is relatively efficient and only slightly improved by the addition of HMG1/2. In contrast, formation of the 23-SC is stimulated over tenfold by HMG1/2 and is then formed as efficiently as the 12-SC (7). The interaction of HMG1/2 with RAG1 (see below) can only account for a portion of the effect on 23-SC formation (29). How might HMG proteins stimulate formation of the 23-SC? One possibility is that they act to bend the DNA in the spacer region, thereby bringing the nonamer and heptamer elements closer together and allowing the RAG proteins to more easily contact both elements simultaneously (7). However, a recent study has shown that RAG1 and RAG2 by themselves induce a bend in the RSS, whose magnitude (about 608) is not increased by the addition of HMG1/2 (49). It was therefore proposed that HMG1/2 are incorporated into the complex to stabilize it through the stabilization of the bend. In this model, the RAG proteins would bind and bend the DNA, and the HMG1/2 protein would behave like a clamp to ensure a durable and favorable bending of the RSS. HMG1/2 contain two HMG boxes and an acidic C-terminal tail, with the HMG boxes involved in both protein:DNA and protein:protein interactions (50). Interestingly, the protein:protein interactions invariably involve the DNA-bindingdomain of the partner protein as well. In keeping with this, it is the NBD of RAG1 that interacts with HMG1, and both HMG boxes are required for the binding (49). Based on sequence and functional similarities with the Hin recombinase, the NBD of RAG1 has been postulated to consist of a homeobox-like domain containing a GGRPR motif and three a-helices (27, 28). HMG1/2 display no sequence-specific DNA-binding properties. Instead, they recognize unusual structures, such as DNA bends, and themselves induce sharp bends (808) in the helix upon binding. It is now clear that HMG1/2 interact with the DNA binding domains of many transcription factors (51). Although these factors are able to bind their target sequence by themselves, the addition of HMG proteins typically greatly strengthens their binding. It is interesting that steroid hormone receptors, and not other hormone receptors, are stimulated by HMG1/2 (52). The DNA binding domain of nonsteroid nuclear receptors is composed of three a-helices that contact both major and minor grooves in the target sequence, whereas for the steroid hormone receptors, only two a-helices are involved, with contacts restricted to the major groove. Because the HMG1/2 proteins bind DNA in the minor groove, it is plausible that they provide an additional interaction surface for the steroid hormone receptors, but they would compete with the third helix of the nonsteroid hormone receptors (52). RAG1 interacts extensively with both the major and minor grooves of the nonamer (34, 35). In contrast, in the RAG1-RAG2-RSS complex, the interaction with the heptamer/spacer border appears to involve the major groove only. It is tempting to think that HMG1/2

502

FUGMANN ET AL

may stabilize this complex by providing minor groove DNA contacts in vicinity of the heptamer.

SYNAPSIS, CLEAVAGE, AND THE 12/23 RULE Chemistry of the Cleavage Reaction


The cleavage reaction occurs in two steps and introduces a DNA double-strand break between the coding gene element and the flanking RSS. In the first step (nicking), a single-strand break is introduced on the top strand between the gene element and the first nucleotide of the heptamer. In the subsequent hairpin formation step, the bottom strand is cleaved, creating a hairpin structure at the coding end and a blunt, 58-phosphorylated signal end. Nicking The phosphate ester between the last nucleotide of the coding element and the first nucleotide of the heptamer on the top strand is hydrolyzed, creating a nick characterized by a 38-OH group at the end of the coding element and a 58phosphorylated end at the heptamer (Figure 3, see color insert) (3). The exact reaction mechanism has not been characterized yet, but the analysis of nicking reactions catalyzed by other site-specific recombinases suggests two possibilities. The first is a one-step reaction in which the recombinase proteins catalyze the direct hydrolysis of the phosphate ester. The second is a two-step reaction in which a serine or tyrosine residue of the recombinase protein acts as a nucleophile in a transesterification reaction, creating a covalent protein-DNA linkage, as occurs with bacteriophage lambda integrase, resolvases, and invertases (reviewed in 53). Subsequent hydrolysis of this ester generates the nicked product. The comparison of the absolute configuration of the central phosphor atom before and after the reaction would distinguish between the two different mechanisms. One-step hydrolysis is a single SN2 reaction leading to inversion of the configuration of the phosphor atom, whereas the two-step mechanism consists of two SN2 reactions leading to retention of the configuration. For HIV-1 integrase it was shown, by substituting one of the nonbridging oxygen atoms of the phosphate with sulfur (creating a chiral phosphorothioate), that the configuration of the phosphor atom was inverted after the nicking reaction (54). But it is important to note that, for the analyzed reaction products, the 38-OH group at the end of the top strand was used as the nucleophile instead of water. Therefore, strictly speaking, the direct hydrolysis mechanism still remains a conclusion drawn from indirect evidence. The use of a phosphorothioate ester to determine the mechanism of RAGmediated nicking was not successful because the ester was not cleaved by the RAG proteins (33). The failure, thus far, to isolate a covalently linked RAG-DNA complex suggests that the RAG proteins catalyze direct hydrolysis analogous to


503


HIV-1 integrase. It is conceivable, however, that the covalently linked intermediate is very unstable and thus quickly hydrolyzed. Hairpin Formation Hairpin formation by the RAG proteins occurs by direct transesterification (55). The top strand 38-OH group acts as a nucleophile attacking the central phosphor atom of the phosphate on the lower strand in a SN2 reaction (Figure 3), with the lower strand of the heptamer serving as the leaving group. Hairpin formation also occurs starting from a pre-nicked substrate, indicating that the nicked product is a real intermediate of the cleavage reaction (3, 56). Attack of the bottom strand by the 38-OH of the top strand could not occur without a significant bend in one or both strands of the DNA. Some coding flank sequences inhibit cleavage at an isolated RSS in Mn2` in vitro, selectively at the hairpin formation step (36, 39). Unpairing of the first two positions of the coding flank (C-2 and C-1) usually restores hairpin formation on these substrates, suggesting that flexibility of the DNA at the site of cleavage is mandatory. Because the RAG proteins have been shown to contact the site of cleavage, one can imagine that they are directly involved in its physical modification. In other recombination systems, unpairing of the flanks facilitates the strand transfer reaction (57, 58).

The Precleavage Synaptic Complex (Paired Complex) In general, V(D)J recombination occurs effectively only on 12/23 RSS pairs (59). Given that gene segments in the antigen receptor loci are separated by up to a megabase, synapsis of the two RSS to generate the PC is critical for reaction fidelity and coordinate cleavage. RAG1 and RAG2 form multimeric complexes in vivo (60–62) and in vitro (45, 46), and it is likely that the two RSSs in the PC are held together by the RAG proteins (24, 46). Recently the PC has been detected and isolated in vitro by blocking the nuclease activity of the RAG proteins using Ca2` as the divalent metal ion (24). This complex was only detected in the presence of HMG1. The catalytic activity of the RAG proteins in this ‘‘captured’’ PC was restored by adding an excess of Mg2`, suggesting that the isolated complex had a structure similar to that of the PC formed in the presence of Mg2` ions. Alignment of RSSs and Stoichiometry In the PC, RSSs may be aligned in a parallel, antiparallel, or some intermediate orientation. Thus far, this issue has not been investigated using imaging techniques such as electron microscopy or atomic force microscopy. However, the influence of the distance between two RSSs on the efficiency of the reaction has been studied in vitro (37) and in vivo (63). The steric constraint imposed by a parallel alignment of the RSS is stronger for inversional than deletional substrates. Indeed, with inversional substrates, cleavage and recombination were more sensitive to shorter inter-RSS distances than with dele-

504

FUGMANN ET AL


tional substrates (37, 63), suggesting a parallel alignment (or something close to it) of the RSSs in the PC. Except for the presence of two RSSs, the stoichiometry of the PC is unknown. As described above, the complex assembled on a single RSS contains two molecules of RAG1 and either one (45) or two (46) molecules of RAG2. From symmetry considerations, it seems likely that the minimum protein content of the PC is two molecules of each RAG protein, although this requires each molecule of RAG1 to be able to interact with a different RSS. Another reasonable possibility is that the PC contains four molecules each of RAG1 and RAG2. A similar situation exists for the PC of the Mu transposon, which contains two MuA homodimers, and only one subunit of each MuA dimer uses its active site during the transposition reaction (64). HMG1/2 is likely to be a component of the PC, since it is a stable component of at least two postcleavage complexes, the SEC (65) and the strand transfer complex that arises from RAG-mediated transposition (25). In addition, HMG2 is stably incorporated into an RSS-RAG1 complex (29). Cleavage Within the Paired Complex In vitro analyses of the nicking reaction in Mg2` (the physiological divalent cation) show that nicking occurs readily on substrates containing a single RSS (31, 66), or on substrates containing two RSSs separated by short distances that prevent synapsis within the same molecule (66). In these latter substrates, the fraction of doubly-nicked RSSs is equal to the product of the fractions of singly-nicked RSSs. However, as the distance between the two RSSs increases, the fraction of doubly-nicked RSSs also increases. Thus, nicking at two RSSs is a largely uncoupled phenomenon, although synapsis of 12- and 23-RSSs substantially stimulates nicking at both RSSs. By contrast, in vitro studies on hairpin formation show that products in which both RSSs are fully cleaved appear before products in which only one RSS is cleaved (37), suggesting that single-end cleavage products result from offpathway reactions. Moreover, although mutations in one RSS do not significantly affect nicking at the partner RSS, such mutations can impair hairpin formation at both RSSs (24, 37, 66). Hence, it appears that hairpin formation at the two coding flank/heptamer junctions is both physically and temporally coupled. This conclusion has recently gained strong support from in vitro cleavage experiments in Mg2` using RSS oligonucleotide substrates in which the sequence of the coding flank was varied (66a). Certain coding flank sequences can dramatically inhibit recombination by wild-type RAG proteins in vivo (66b) and cleavage by purified RAG proteins in vitro (66a). These coding flanks greatly slow the nicking step of the reaction but have no effect on hairpin formation if a prenicked substrate is used. Interestingly, if nicking at one RSS is inhibited hairpin formation (but not nicking) is inhibited at the partner RSS (66a). This indicates that nicking of both RSSs is required for hairpin formation at either one and that the proteins engaging one RSS can sense the status (nicked or unnicked) of the partner RSS. A question that remains to be answered concerns whether RAG1 and RAG2 in the PC cleave the RSS to which they are bound (cis cleavage) or the opposite


505


RSS (trans cleavage). At first glance, the fact that nicking occurs in the absence of a proper PC and the results of cleavage experiments in Mn2` (see below) seem to suggest that RAG proteins can perform both nicking and hairpin formation in cis. However, the arrangement of the RAG proteins and stoichiometry of RSSs within ‘‘single-RSS’’ complexes are unknown. Such complexes may represent half of the true synaptic complex, or they may represent the full synaptic complex with one RSS lacking. An alternative, albeit less likely, scenario is that they may be improper synaptic complexes generated between RSSs on different DNA molecules. The 12/23 Rule The specific requirements for enforcement of the 12/23 rule, the precise step at which enforcement occurs, and the molecular mechanism that mediates it have all been subjects of considerable debate. Initial in vitro studies demonstrated the importance of the divalent cation in conferring 12/23-coordinated cleavage. In the presence of Mn2`, both RAGcontaining lymphoid extracts (37, 56) and purified core RAG proteins (3) catalyze nicking and hairpin formation on substrates containing either 12/23 pairs of RSSs or isolated 12- or 23-RSSs; in substrates containing 12/23 pairs, single-RSS cleavages predominate. In Mn2`, therefore, cleavage events at two RSSs are uncoupled. By contrast, in the presence of Mg2`, lymphoid extracts show a 25- to 50-fold preference for double cleavage of 12/23 substrates over cleavage of 12/12 or 23/ 23 substrates (37), in good agreement with in vivo data (38). The traditional explanation for this ion dependency is that Mn2` allows RAG1 and RAG2 to cleave at isolated RSSs, perhaps by altering the geometry of the active site in relation to the substrate. Indeed, Mn2` skews both the temperature and pH activity profiles of the RAG proteins (33), suggesting that the architecture of the RAGRSS complex in Mn2` is distorted. Interestingly, Tn10 cleavage in the presence of low levels of Mn2` yields an accumulation of single-end events (67), although cleavage is still believed to occur within the context of a PC. Purified core RAG proteins in the presence of Mg2` show only a three- to fivefold preference for 12/23 double cleavage versus 12/12 double cleavage (4); single cleavage events are still seen at low frequency and mainly on the 12-RSS. Addition of HMG1 or HMG2 has a moderate stimulatory effect on 23-RSS single cleavage and a large stimulatory effect on 12/23 double cleavage (7, 8), but this addition does not result in strict conformity to the 12/23 rule because substantial cleavage of 12/12 and 23/23 substrates still occurs. Full restoration of the 12/23 rule can be achieved by adding crude cellular extracts (8). Thus, it appears that both Mg2` as the divalent cation and additional nonlymphoid factors including HMG are involved in full enforcement of 12/23-coordinated cleavage. Recent studies have suggested that purified core RAG and HMG proteins may be the crucial protein factors required to fully recreate the 12/23 rule. Using short double-stranded oligonucleotides containing single RSSs, Hiom & Gellert showed that RAG1, RAG2, and HMG1 in Mg2` cleave only 12/23 pairs of oligonucleotides and not 12/12 or 23/23 pairs (24). Kim & Oettinger constructed an


506

FUGMANN ET AL

oligonucleotide substrate with a single-stranded region between the RSSs; using purified core RAG proteins, HMG1, and nonspecific double-stranded competitor DNA in the presence of Mg2`, strict 12/23-coordinated cleavage is seen, with suppression of 12/12 or 23/23 events, suggesting that the effects of cell extracts might be at least partly due to nonspecific DNA (48). West & Lieber constructed a double-stranded oligonucleotide substrate with a nick between the RSSs; while purified core RAG proteins in Mg2` have no cleavage activity on this substrate, strict 12/23-regulated double cleavage is seen upon addition of HMG1 (47). The precise step at which the 12/23 rule is imposed remains to be determined. Using EMSA, Hiom & Gellert showed that in the presence of Ca2`, purified core RAG proteins with HMG1 synapse 12/23 oligonucleotide pairs more efficiently than either 12/12 or 23/23 pairs, although both 12/12 and 23/23 oligonucleotides show a significant background level of synapsis (24). In contrast, West & Lieber showed that preincubation of their labeled 12/23 substrate with unlabeled 12/23, 12/12, or 23/23 competitor substrate reduces double hairpin formation in the labeled substrate 10- to 20-fold in all cases (47). This suggests that RAG and HMG1 proteins can synapse 12/23, 12/12, and 23/23 substrates, all with similar efficiencies, and that enforcement of the 12/23 rule occurs at the level of double hairpin formation. Both studies raise important issues. It is possible that, in the study by Hiom & Gellert, the use of Ca2` as the divalent cation alters the events that take place under normal physiological conditions. Also, assaying RAG-mediated synapsis by using isolated RSSs on two separate DNA molecules may be a less sensitive means of study, as lack of connectivity between the two RSSs automatically decreases the probability of synapsis. On the other hand, the nicked substrate used in West & Lieber’s study may allow for artifactual synapsis. The RSSs in their substrate are separated by less than 70 bp, and the tight tethering of two RSSs combined with increased flexibility between the RSSs may allow for noncanonical synapsis to occur at increased efficiencies. Finally, it is possible that the 12/ 23 rule is imposed at both synapsis ($threefold) and double hairpin formation ($tenfold), generating the $30-fold preference for 12/23 substrates seen in vivo (38).

POSTCLEAVAGE SYNAPTIC COMPLEXES Before they are joined, hairpin coding ends must be nicked open and nucleotides may be inserted or deleted. Nucleotide insertion usually results either from the action of TdT or from asymmetric opening of the hairpin to generate short stretches of palindromic (P) nucleotides (reviewed in 59). While N nucleotide addition occurs predominantly in postnatal animals, P nucleotides are found in V(D)J coding junctions at most loci in fetal and adult animals. The mechanism of nucleotide deletion from coding ends is unknown.


507


The Cleaved Signal Complex (CSC) Immediately after cleavage, the RAG proteins are bound to the signal ends and coding ends in a four-end complex, the CSC (Figure 1). One model of coding end processing predicts that coding ends are held in this complex long enough for modifying enzymes, ligase, and necessary cofactors to be recruited and to act, and hence that CJ formation occurs in the CSC. Most of the available data is consistent with this model. Alternatively, it is possible that the two signal ends and two coding ends are associated in the CSC immediately after cleavage, but that the complex falls apart rapidly and coding ends are processed or joined independent of the SEC (Figure 1). Although the stability of association of coding ends in the CSC has rightly been questioned, the available experimental data argue against this model. The earliest hint that coding ends and signal ends must exist at least transiently within a common protein:DNA complex came from studies that identified hybrid joints and O/S joints with high efficiency in transfected cells (68, 69). The ability of coding ends to be joined to signal ends implies that all four ends must exist within a common postcleavage complex. Additionally, the occurrence of a similar spectrum of coding end processing events in hybrid, O/S, and coding junctions supports the idea that coding ends and signal ends exist within a common complex during coding end processing (68–70). Recently, direct physical evidence has been provided for the CSC (24). RAGmediated cleavage in trans of RSS oligonucleotide substrates (one labeled with biotin and the other with 32P) allowed streptavidin capture of 32P-labeled signal ends and coding ends. Coding ends were captured less efficiently than signal ends, consistent with the failure to detect stable coding end/coding end or coding end/ signal end complexes in an earlier study (65). Definitive proof of the four-end complex awaits experiments demonstrating that both RSSs have been cleaved in the captured complexes.

Hairpin Opening and the Processing of Coding Ends Signal ends were the first V(D)J recombination-specific DNA intermediates to be detected in normal lymphoid precursors, and they are considerably more abundant than coding ends (71–75). A pre–B cell line with inducible RAG expression was used to show that coding ends are present at 10- to 100-fold lower levels than signal ends at the Ig Jj locus (76). After RAG protein induction, CJ formation correlates with the appearance of signal ends, suggesting that while signal ends persist, coding ends are processed and joined rapidly after cleavage. The RAG proteins appear to be necessary for coupling the cleavage and joining stages of V(D)J recombination. In vitro, coding ends generated by RAG-mediated cleavage cannot be joined if they are first deproteinized (77, 78). There are several possible postcleavage roles for the RAG proteins. First, they might help to maintain the structure or nuclear location of the postcleavage complex. Second, they

508

FUGMANN ET AL


may be required for the recruitment of DNA processing or repair proteins. Finally, as discussed below, they may play a catalytic role in coding end processing or joining. Ubiquitous Hairpin Opening Activities Activities have been detected that can resolve and join transiently transfected synthetic hairpin DNA substrates in a variety of cell lines (79, 80, and discussed in 81). These observations indicated that ubiquitous hairpin opening activities exist in vivo and raised the possibility that hairpin opening during V(D)J recombination is performed by a ubiquitous factor. Recent work has suggested that the RAD50/Mre11/NBS DNA repair complex could potentially serve in this capacity (82, 83). In vitro, the human Mre11/ RAD50/NBS complex possesses activities that can nick synthetic DNA hairpins, remove 38 overhangs, and mediate 38 to 58 resection of DNA termini (83). This latter activity promotes homology-mediated ligation of DNA ends and could conceivably resect coding ends to promote joining stimulated by short homologies during V(D)J recombination (82). As yet, no direct involvement of RAD50/ Mre11/NBS in V(D)J recombination has been demonstrated. Hairpin Opening by the RAG Proteins Recent studies have suggested that hairpin opening during V(D)J recombination might be performed by the RAG proteins (84, 85). In vitro, the RAG proteins nick synthetic hairpins a few nucleotides 58 of the tip, and the presence of both RAG1 and RAG2 is required. The hairpin structure is not required for this nuclease activity because a homologous DNA duplex substrate is nicked at the same position. With synthetic hairpins, HMG2 focuses the nicking activity to the vicinity of the hairpin tip, presumably by interacting with the RAG proteins and promoting their interaction with the altered DNA structure at the tip (85). HMG1/2 may perform a similar function in vivo. Like the disregulated cleavage of isolated RSS oligonucleotides, nicking of synthetic hairpins occurs in Mn2` and not Mg2` (84, 85). Strikingly, hairpin nicking can occur in Mg2`, either in trans on oligonucleotide substrates or in cis with plasmid substrates, but only in the context of 12/ 23 regulated cleavage. Using oligonucleotide substrates, coding end hairpins are opened at the tip (84). After coupled cleavage of plasmid substrates, however, coding end hairpins are opened predominantly 1 to 2 nt 58 of the tip, generating palindromic extensions similar in length to the short P regions found in CJs in vivo (Figure 4) (85). Hairpin nicking in Mg2` has also been observed using a large DNA fragment containing an RSS at one end and a hairpin at the other (PE Shockett, DG Schatz, unpublished). Since nicking of synthetic hairpin substrates (lacking a signal end) does not occur in Mg2`, it is possible that transient synapsis occurs between the signal end bound by the RAG proteins and the coding end, to some extent mimicking the postcleavage complex. Additionally, synthetic hairpin nicking in Mg2` appears to be stimulated by short signal end oligonucleotides



509

Figure 4 Summary of sites of hairpin nicking by the RAG proteins within the postcleavage complex (A) and types of coding ends detected in vivo (B).

(PE Shockett, DG Schatz, unpublished). Thus, in Mg2`, organization of the RAG protein active site for hairpin nicking probably requires signal end binding. These findings suggest that in vitro, the RAG proteins nick coding end hairpins within the CSC, generating full-length coding ends, P nucleotides, and conceivably deletions at coding ends. Similarities between mechanistic aspects of V(D)J recombination and bacterial transposition (see below) further support the hypothesis that the RAG proteins initiate both RSS cleavage and subsequent hairpin opening (86, 87). These similarities were further extended by the recent finding that both the RAG proteins and Tn10 transposase are sequence-nonspecific 38 flap endonucleases (87a). Both 38 extensions and 38 flaps can be removed by endonucleolytic cleavage at or near the double strand/single strand junction. In addition to raising the possibility of an additional role for RAG1 and RAG2 in creating junctional diversity, the results suggest that sequence-specific nicking by the RAG proteins may proceed through a flap-like structure at the RSS-coding flank border (87a).

Questions Raised by RAG-Mediated Hairpin Opening One unresolved question is why oligonucleotide and plasmid substrates yielded different patterns of coding end hairpin opening (Figure 4) in the two studies cited above (84, 85). This difference may result from a combination of factors including the sequence, structure, and single-stranded character of the hairpin coding end, the substrate orientation, and the preparation of RAG proteins (88 and references therein). Preliminary data suggest that the differences observed in the two studies can be


510

FUGMANN ET AL

accounted for by differences in both the substrates and the protein preparations used (85; P Cortes, personal communication). A second question arises from the fact that while coding end hairpins are opened at, or 58 of, the tip in vitro, coding ends in vivo exhibit deletions and extensions consistent with 38 overhangs and nicking of the hairpin 38 of the tip (Figure 4) (74–76). Whether the ends detected in vivo represent products of the initial coding end hairpin processing event and a true intermediate in V(D)J recombination remains to be shown. A possible precursor-product relationship between some of the deleted ends detected in vivo and deletions found within actual CJs has been pointed out (75). In vivo, both deletion and P nucleotide formation are influenced by the coding end sequence (88, 89 and references therein). It has been proposed that double-stranded hairpin nicking at AT-rich sequences serves as a mechanism of nucleotide loss from coding ends (88). The low recovery of coding ends with nucleotides deleted on both strands in vitro after RAG-mediated cleavage might be a function of the coding end sequences used or the need for a more stable and prolonged hairpin:RAG association for double-stranded nicking of the hairpin. Additionally, the studies of RAGmediated hairpin opening in vitro have been performed with truncated RAG proteins, and it is possible that the full-length proteins will behave differently. We have observed synthetic hairpin nicking in the presence of full-length RAG2, but full-length RAG1 has not yet been examined (PE Shockett, DG Schatz, unpublished). Finally, in lymphocytes of DNAPKcs- and Ku80-deficient mice, hairpin coding ends accumulate despite RAG protein expression (90–92). This presents an interesting paradox: In vivo, DNAPKcs and Ku are necessary for RAG-mediated hairpin opening, while in vitro, the RAG proteins can nick coding end hairpins without DNAPKcs or Ku. One resolution is suggested by the observation that coding ends are retained inefficiently in the CSC in vitro (24, 65). It is plausible that in vivo other factors are required to retain coding ends in the postcleavage complex long enough to be nicked by the RAG proteins. Ku and DNAPKcs are good candidates for factors that would act to stabilize coding ends in the CSC. Ku has been proposed to regulate remodeling or disassembly of the postcleavage complex, thereby facilitating coding end processing (91). Furthermore, the Ku proteins have been shown to facilitate the ligation of DNA ends similar to those generated during V(D)J recombination, suggesting that they might bridge two coding ends to promote joining (93). DNAPKcs has been shown to bind to DNA hairpins (94), and furthermore, in the presence of Ku but in the absence of DNA ends, it can be activated in vitro by the putative nuclear matrix protein C1D (95). By virtue of this interaction, DNAPKcs could tether the postcleavage complex to the nuclear matrix, thereby stabilizing coding end association with the complex and enhancing RAG-mediated hairpin nicking. Another possibility is that DNAPKcs may regulate the RAG proteins or other proteins in the postcleavage complex by phosphorylation. DNAPKcs activity might be required in vivo to


511

move DNA end-binding proteins out of the way so that processing enzymes, including the RAG proteins, have access to coding ends (91, 96). Coding end blocking proteins might include Ku and DNAPKcs themselves, or PARP, which binds and is activated by DNA ends and which may stimulate DNAPKcs by ADPribosylation (96–98). Finally, it is possible that a complex of Ku and possibly DNAPKcs must move along the coding end hairpin altering its conformation before it can be nicked by the RAG proteins. It has been proposed that Ku might unwind the hairpin coding end to allow nicking (99).


The Signal End Complex In vitro, signal ends remain associated with each other in the signal end complex (SEC) (Figure 1), a nuclease resistant protein:DNA complex containing the RAG proteins and HMG1/2 (24, 65). This complex has not been directly demonstrated in vivo; however, several observations suggest that it exists. First, signal ends are usually not subject to deletion and N-nucleotide insertion as are coding ends, and this could be explained if RAG proteins protect signal ends from the processing activities that act on coding ends (65). Second, in normal lymphoid precursors, signal ends are easily detected (71–73) and persist before being joined (76). This persistence causes no obvious increase in p53 in normal animals (100). Thus, signal ends might not trigger conventional DNA damage responses in the cell because of their association with the RAG proteins. Finally, the formation of signal joints correlates with the downregulation of RAG gene expression, consistent with a need to remove RAG proteins from the SEC before joining (76). Recent experiments demonstrated that removal of short terminal portions of RAG1 and RAG2, previously thought to be ‘‘nonessential’’ for signal joint formation, is associated with an accumulation of signal ends in transient recombination assays, consistent with decreased efficiency of signal joint formation (19). Requirements for chaperone-mediated disassembly of Mu transposase protein:DNA complexes have prompted speculations that signal joint formation requires the binding of chaperones for disassembly of the SEC and subsequent recruitment of repair proteins (19, 65). It has been suggested that these RAG terminal domains may interact with such a factor (19). Additionally, it has been proposed that the requirement for SEC disassembly before signal joint formation directs processing and joining activities to coding ends first, thereby minimizing hybrid joint formation (99). While it is accepted that Ku proteins are required for signal joint formation, the role of DNAPKcs is more controversial (92, 101, 102). After RAG-mediated cleavage in cell extracts, associations between signal ends and either Ku or DNAPKcs have been detected (65). The finding that signal joints on transfected substrates and at TCR loci can exhibit N-nucleotide addition and deletion suggests that in some circumstances signal ends encounter some of the same processing enzymes as coding ends (103, 104, and references therein).

512

FUGMANN ET AL

RAG-MEDIATED TRANSPOSITION


From the beginning, it had been postulated that V(D)J recombination must have evolved from an ancient transposition or site-specific recombination system; both the compact genomic structure of the RAG locus and the organization of the RSSs were compatible with this notion (6, 105). Subsequent elucidation of the reaction pathway (3), the observation that alcohols could serve as nucleophiles in the nicking reaction, and the demonstration that hairpin formation proceeded via a direct SN2 transesterification mechanism (4) all provided strong correlations with transposition. RAG-mediated hybrid joint formation in vitro appeared to be analogous to the disintegration reactions catalyzed in vitro by retroviral integrase (70, 106). Finally, the demonstration that purified RAG proteins in vitro could catalyze transpositional insertion confirmed the link between V(D)J recombination and transposition (25, 26).

Basic Features of RAG-Mediated Transposition RAG-mediated transposition was demonstrated in two different studies using either plasmid (25) or oligonucleotide (26) substrates containing RSSs. These studies showed that RAG proteins catalyze efficient insertion of signal end substrates into a target vector, generating a strand transfer product. The target vector can be the same molecule as the signal end substrate, resulting in an intramolecular transposition product (25); or the target vector can be a different molecule from the signal end substrate, resulting in an intermolecular transposition product. As with all transposition reactions, RAG-mediated coupled strand transfer of two signal ends results in a target site duplication at the site of insertion. The duplication is generally 5-bp long for RAG-catalyzed transposition, although duplications arising from intramolecular reactions show more variation in length. The sites at which insertions occur vary in sequence, and GC-rich sequences are somewhat preferred; a GC-rich hotspot was observed in one instance (25). RAGmediated transposition is strongly dependent on the presence of HMG1/2 (25, 26) and typically requires both a 12- and a 23-RSS, although a low level of strand transfer products can be seen using only a 12-RSS substrate (26). The reaction can occur using either Mg2` or Ca2` as the divalent cation.

Comparison to Other Transposition Systems Many aspects of RAG behavior are shared by composite transposons, large elements whose internal sequences encode drug resistance and accessory factors, and whose ends are functionally dissociable insertion sequences (IS) that encode transposase (for reviews of different classes of transposable elements, see 107– 109 and references therein). In particular, many parallels are seen with the composite transposon Tn10/IS10, the first transposon in which formation of an obligatory hairpin intermediate was conclusively demonstrated (87).



513

Divalent Cation As a general rule, it appears that Mg2` is the physiological divalent cation for proper functioning of transposases (107–110). In the absence of a divalent cation, RAG1 and RAG2 fail to stably bind isolated RSSs (31); by contrast, MuA can bind its cognate RSSs but is incapable of performing synapsis (107), whereas Tn10 can perform both binding and synapsis (108, 111). Without a divalent cation, transposases have no catalytic activity presumably because, as discussed below, divalent cations are essential components of the active site. In the presence of Ca2`, transposase binding and synapsis are quite stable. However, enzymatic activities are impaired, with Tn10 showing no activity (67) and RAG proteins and MuA capable of only strand transfer (24, 57). In the presence of Mn2`, transposases behave strangely. In various Mn2`-based assays, both RAG and Tn10 show (a) uncoordinated single-end cleavage (47, 56, 67), (b) inability to complete cleavage, resulting in accumulation of nicked intermediates (36, 39, 112), and (c) relaxed sequence specificities (33, 67). Importantly, Mn2`-induced relaxation of sequence specificity is also seen in other enzymes that catalyze phosphoryl transfer reactions, such as polymerases (113) and restriction endonucleases (114). Signal Sequences The organization of the RSS separates the primary site of RAG protein binding (the nonamer) from the site of cleavage (the heptamer). The dissociation of these two sites appears to be a general property among transposons (108, 115, 116). It has been suggested that the RSS resembles the sites at the ends of Tc family transposons (117). The RSS heptamer motif is almost identical to the Tc end sequence; the two sequences differ primarily in the first nucleotide (C for the heptamer; T for Tc transposons). However, the cleavage products generated by the two reactions differ (115). The RSS nonamer motif resembles an AT-rich sequence in the Tc1 transposase binding site immediately downstream of the heptamer-like motif; however, this sequence is not well conserved among the Tc family (115, 117). Thus, the functional significance of the parallels between RSSs and Tc end sequences are uncertain. Synapsis Transposons such as Mu or Tn7 have strict requirements for synapsis (107, 109). In Mu, under physiological conditions, topological requirements restrict PC formation to supercoiled substrates in which the two end sequences have specific orientations along the DNA (107). By contrast, Tn10 shows considerable versatility in synapsis; in vitro and in vivo, transposition can occur using many different arrangements of IS10 ends on the same or on different molecules (108). The behavior of the RAG proteins in synapsis appears to be similar to that of Tn10 transposase. Protein-DNA interactions within PCs are quite strong. Mu and Tn10 cleavage from supercoiled donor substrates retains supercoils within the transposon elements; similarly, insertion of excised Tn10 into a supercoiled target molecule

514

FUGMANN ET AL

retains the supercoils in the target (107, 108). Thus we may expect protein:DNA interactions within the RAG PC to be similarly strong (see below).


The 12/23 Rule The 12/23 rule provides a means by which the two ends of the primordial RAG transposon can be distinguished. This asymmetry in end usage is echoed in varying degrees in other transposition systems, particularly in more complicated systems such as Mu and Tn7 (107, 109). Perhaps the most striking parallel is seen in the apparent ‘‘9–21’’ rule of P-element transposition (116). The two ends of P-elements contain a 31-bp inverted repeat and a 10-bp transposase binding site, separated by a spacer whose length is either 9 bp (38 end) or 21 bp (58 end). Cleavage occurs within the 31-bp inverted repeat and typically requires both 38 and 58 ends, although in vitro, uncoupled cleavage can be observed at low levels on substrates containing a single 38 end. Cleavage In addition to restrictions governing PC formation, Mu and Tn7 transposons have another control feature to regulate initiation of cleavage. In the absence of target DNA, MuA performs the initial nicking reaction somewhat slowly; addition of target DNA greatly stimulates the nicking rate (107, 108). Tn7 is even more rigid in that cleavage is not performed at all unless a target DNA sequence is present (109). In contrast, both RAG and Tn10 transposases readily cleave synapsed ends in the absence of target DNA (108). This dissociation between cleavage and target interaction may have been an important evolutionary factor in allowing the RAG transposase to function in recombination. As described previously, RAG1 and RAG2 appear to catalyze coupled cleavage at two RSSs almost simultaneously. However, in many other systems (e.g., Tn7 and Tn10), the cleavage reactions at the two ends are temporally separable (108, 109). The cleavage reactions catalyzed by RAG, Tn5, and Tn10 transposases all result in the formation of hairpin intermediates (87, 118). In Tn5 and Tn10 cleavage, however, the hairpin resides on the end of the transposon, whereas in V(D)J cleavage, the hairpin resides on the end of the flanking sequence. This seemingly minor distinction may have important functional and evolutionary ramifications (see below). It has been suggested that excision of certain elements in plants (Ac/ Ds, Tam3, and Slide), Drosophila (hobo), and fungi (Ascot-1) may resemble RAG cleavage in this regard (119). Interestingly, the ends of Ascot-1 bear the sequence CAGTG that is found at the ends of Tc family transposons (117, 119) and that represents the last five bases of the heptamer. Postcleavage Complex As discussed above, current evidence indicates that after RAG-mediated cleavage has occurred, the CSC retains all four cleaved ends in a single postcleavage complex (24). Available evidence on Tn10 cleavage, however, suggests that the Tn10 transpososome releases each flanking sequence as it is cleaved, resulting in a two-end postcleavage complex (111). This release may


515

be necessary to ensure that efficient hairpin opening occurs during Tn10 transposition (see below).


Target Immunity Transposons such as Mu or Tn7 typically do not transpose into the same donor molecule from which they originate, a phenomenon known as target immunity (107, 109, 120). The mechanistic basis for this resides in the fact that under physiological conditions, strand transfer requires additional proteins (MuB; TnsC) that bind to target molecules other than the transposon donor. By contrast, RAG and Tn10 transposons do not appear to be subject to target immunity; in both systems, intramolecular transposition occurs at relatively high frequencies (25, 108). The rationale for this discrepancy is that, while intramolecular transposition would be deleterious to a transposon such as Mu or Tn7, such intramolecular events within a composite transposon are advantageous for an insertion sequence whose primary aim is to disseminate as widely as possible (108). Thermodynamics of Transposition In transposition reactions, the stability of each complex along the reaction pathway increases as the reaction progresses (107, 108). Consistent with this notion, RAG complexes with cleaved signal end– signal end molecules are extremely stable (65), and RAG proteins appear to remain tightly associated with strand transfer complexes resulting from intramolecular transposition (25).

THE RAG1-RAG2 ACTIVE SITE One of the most interesting and still open questions is the structure and functionality of the active site of the RAG complex. No structure has been reported for the core regions of the RAG proteins thus far, and hence this section attempts to address this issue based primarily on analogies to other site-specific recombinases, together with evidence from in vitro experiments.

How Many Active Sites Participate in Cleavage? Comparison of active site stoichiometries in other transposition systems suggests that in V(D)J cleavage, a single active site may catalyze both nicking and hairpin formation at a given coding end–signal end junction (55). Both Tn5 and Tn10 transposition generate double-strand break intermediates that are formed via a nicking-hairpin approach, and both transposases appear to use one active site for cleavage at each transposon end (86, 108, 118). By contrast, Tn7 transposition generates double-strand break intermediates that are formed by nicking each of the two complementary strands; two separate proteins are required to catalyze the two single-strand cleavages, apparently to accommodate the two substrate strands with opposite polarities (109, 121).

516

FUGMANN ET AL

Neither RAG1 nor RAG2 alone exhibits any catalytic activity (3, 25, 26, 70, 84, 85), suggesting either that both proteins contribute amino acid side chains to the active site, or that interaction between them leads to a structural change in one of them, bringing the catalytic site into its active conformation. In keeping with the latter model, transposases often have cofactors that bind to and activate the catalytic subunit, e.g. MuA-MuB and TnsA/B-TnsC (reviewed in 107, 109).


A Two-Divalent Metal Ion Model for the Active Site HIV-1/ASV integrase, MuA, Tn5, and Tn10 transposases all have catalytic activities similar to those of the RAG complex. Initial hydrolysis of a phosphate ester bond on one strand releases a 38-OH group that is subsequently used in a direct transesterification reaction, creating either a strand transfer product or a DNA double-strand break. Therefore it is possible that all these proteins contain a similar active site. Although at the amino acid level the homology between the core regions of ASV integrase, HIV-1 integrase, MuA, and Tn5 is only 9–15% (118), X-ray crystal structure analyses have revealed that these four proteins, together with RuvC and RNaseH, belong to a polynucleotidyl-transferase family defined by structural similarities (118, reviewed in 122, 123). The active site of all four proteins contains a triplet of acidic residues, known as the DDE motif (124), that is shared among many other transposases. These residues chelate divalent metal ions and are important for all chemical steps of transposition and integration. Given the mechanistic parallels between transposition and V(D)J recombination, and the strong influence that divalent metal ions have on the activity and specificity of the RAG proteins, it is appealing to think that the active site of the RAG proteins contains one or more divalent metal ions that are directly involved in the catalytic process. Indeed, recent experiments (SD Fugmann, IJ Villey, LM Ptaszek, DG Schatz, unpublished) have identified two aspartic acid residues within RAG1 (murine amino acids D600 and D708) that function specifically in catalysis. Mutation of either of these residues abolishes all cleavage and strand transfer activities, but results in properly folded proteins that are able to bind to the RSS and form synaptic complexes with pairs of RSSs. Additional data indicate that D708 directly coordinates a divalent metal ion. The results suggest that these two amino acids and at least one divalent metal ion are critical, catalytic components of the RAG active site. A model describing how two divalent metal ions catalyze phosphoryl-transfer reactions was originally proposed for the DNA polymerase 38-58-exonuclease domain (126, 127), and that model is now thought to be a general principle in many phosphoryl-transfer enzymes (128). Applied to the RAG proteins, the model offers a simple explanation for all their activities reported so far: nicking, hairpin formation, hybrid joint and O/S joint formation, hairpin opening, and strand transfer. The general principle is that the two metal ions stabilize the trigonal bipyramidal transition state of the SN2 reactions and either activate hydroxyl nucleo-



517

philes or act as a Lewis acid stabilizing oxyanion leaving groups for the SN2 reactions. The two metal ions switch between these two functions after each step of the reaction, i.e. the metal ion that activates the nucleophile in the first step stabilizes the leaving group in the next step and vice versa (Figure 3). The nucleophile and the target phosphor atom are well defined for the two steps of the V(D)J cleavage reaction. First, water is activated as a nucleophile by a Mg2` ion (designated number 1 in Figure 3), whereas Mg2` ion number 2 stabilizes the 38-OH leaving group on the top strand of the coding end. Second, the hairpin is formed by the attack of the previous 38-OH leaving group, now activated by Mg2` number 2 as a nucleophile, at the phosphate ester on the lower strand. Mg2` number 1 enhances the leaving group properties of the 38-OH group on the resulting signal end. The final outcome of the recombination event has not been established at this point. It will be determined by the nucleophile and the target phosphate ester in the third reaction step. If the signal end 38-OH is used as the nucleophile, either hybrid joint (or O/S joint) formation, a reversal of the second step or transposition occurs, depending on whether the hairpin coding end or a different DNA molecule, respectively, is attacked (25, 26, 70). However, if a water molecule is activated to perform the nucleophilic attack, hairpin opening can occur. The model therefore predicts that hairpin opening and transposition (and also hybrid joint and O/S joint formation) are mutually exclusive. Once hairpin hydrolysis occurs (which starts the reaction down the pathway leading to CJ formation), the signal ends lose their ability to perform strand transfer reactions.

EVOLUTION AND CONTROL OF RAG-MEDIATED TRANSPOSITION According to the proposed two-divalent metal ion model (Figure 3), the fate of the SEC is determined in the CSC shortly after the cleavage. Either coding ends and signal ends get rejoined to form hybrid joints or O/S joints, or the hairpins are opened, or the hairpins fall out of the complex before they get opened and the 38-OH groups at the signal ends are active for transposition. If the hairpins get opened, the signal end 38-OH groups are inactive for strand transfer and the signal ends can be ligated to form a signal joint. Since inversional recombination and subsequent genome integrity require the formation of signal joints, regulatory mechanisms should exist to control which pathway is chosen.

Does RAG-Mediated Transposition Occur In Vivo? In general, transposition activity is maintained at a low level, since highly active transposons would be detrimental to the host cell due to the mutagenic effect of genome rearrangements. It has been proposed that chromosomal translocations involving antigen receptor gene loci arise from incomplete RAG-mediated trans-


518

FUGMANN ET AL

position events (26), but no experimental evidence for these events has been presented so far. To date, no RAG-mediated transposition has been observed in vivo, although SECs are not rapidly joined or degraded after the cleavage reaction (see above). So, why don’t we see transposition? First, transposition events might be too rare to be detected with the current methods. Second, the transposase activity might be suppressed in vivo either by the RAG proteins themselves, or by other factors. Thus far, no active transposon has been isolated from vertebrates. The putative vertebrate transposase genes isolated to date contain multiple, inactivating mutations (129). In one case, clever reversion of these mutations resulted in an active transposase (130). The RAG proteins might also have acquired mutations affecting their ability to perform transposition in vivo, but these would have had to spare all of their activities necessary for V(D)J recombination. Perhaps these mutations affected target capture or strand transfer, although both activities are robust in vitro. As mentioned above, RAG-mediated hairpin opening might be detrimental to transposition, and an alteration of the active site might have caused a switch in the preferred nucleophile for the third transesterification (Figure 3) from the 38-OH group on the signal end to a water molecule. Additionally, since the in vitro experiments were performed using the core regions of RAG1 and RAG2 (25, 26), it is possible that the amino-terminus of RAG1 and/or the carboxy-terminus of RAG2, missing in the core proteins, might inhibit transposition in vivo, by facilitating the disassembly of the SEC (19). Additional factors that are present in the nucleus of the cell but not in the in vitro reactions might reduce or abolish the capacity of the RAG proteins to catalyze a complete transposition reaction. These factors could cause conformational changes in the SEC either by performing modifications (e.g. phosphorylation) or by interacting directly with the RAG proteins or the DNA in the SEC. Such changes could destabilize or inactivate this complex and therefore inhibit transposition. The DNA double-strand break repair factors XRCC4, DNA Ligase IV, Ku70, Ku80 are candidates because they are important for the formation of signal joints, a process that is mutually exclusive of the strand transfer reaction. Additionally, the Ku70/Ku80 heterodimer and DNAPKcs could facilitate disassembly of the SEC by competing with the RAG complex for the binding to the signal ends. This mechanism might at least prevent reassembly of SEC in the event that the RAG proteins dissociated from the signal ends.

Evolutionary Implications Previously, the RAG transposon was proposed to consist of the RAG genes flanked by a single RSS on each side (25, 131). To explain the separation of the RAG genes and the RSSs, we would suggest a modification of this idea in which the RAG transposon, like Tn10/IS10 and Tn5/IS50, was a composite transposon, consisting of a central sequence flanked by a pair of insertion sequences (IS), each flanked by a pair of RSSs (Figure 5). The RAG genes may have resided in


519

Figure 5 Schematic model of the ancient RAG transposon and the generation of the split antigen receptor genes. See text for details. The 12-RSS and 23-RSS are represented as black and white triangles, respectively. The RAG genes are drawn as boxes with arrows indicating the direction of transcription; the RAG-proteins in the SECs are shown as shaded ovals.

520

FUGMANN ET AL

the central region or, as is the case for the composite bacterial transposons, inside one of the insertion sequences. After this complete transposon integrated into the germline, one or both of the IS elements was excised and integrated elsewhere in the genome. If the RAG genes resided in the central portion of the composite transposon, such an event would also have rendered them incapable of further transposition. At some point during evolution, one of the IS elements integrated into an exon of a receptor gene. The functional receptor gene could then only be produced if the RAG proteins excised the IS and the chromosomal break was repaired. All of the current split antigen receptor gene loci would thereafter have arisen by repeated gene duplications. The putative microorganism from which the RAG transposon originated has not been identified, and it is possible that it has not survived subsequent evolutionary selection processes.


ACKNOWLEDGMENTS This review is dedicated to the memory of our friend and colleague, Eugenia Spanopoulou. We would like to thank L Ptaszek for help in preparation of Figure 2, and M Lieber, P Cortes, and M Bianchi for sharing unpublished data. This work was supported by a Howard Hughes Medical Institute predoctoral fellowship to AIL and by NIH grant AI32524 to DGS. SDF, IJV, and DGS are employees of the Howard Hughes Medical Institute. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Sleckman BP, Gorman JR, Alt FW. 1996. Accessibility control of antigen-receptor variable-region gene assembly—role of cis-acting elements. Annu. Rev. Immunol. 14:459–81 2. Schlissel MS, Stanhope-Baker P. 1997. Accessibility and the developmental regulation of V(D)J recombination. Sem. Immunol. 9:161–70 3. McBlane JF, van Gent DC, Ramsden DA, Romeo C, Cuomo CA, Gellert M, Oettinger MA. 1995. Cleavage at a V(D)J recombination signal requires only RAG1 and RAG2 proteins and occurs in two steps. Cell 83:387–95 4. van Gent DC, Ramsden DA, Gellert M. 1996. The RAG1 and RAG2 proteins establish the 12/23 rule in V(D)J recombination. Cell 85:107–13

5. Schatz DG, Oettinger MA, Baltimore D. 1989. The V(D)J recombination activating gene (RAG-1). Cell 59:1035–48 6. Oettinger MA, Schatz DG, Gorka C, Baltimore D. 1990. RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination. Science 248:1517–23 7. van Gent DC, Hiom K, Paull TT, Gellert M. 1997. Stimulation of V(D)J cleavage by high mobility group proteins. EMBO J. 16:2665–70 8. Sawchuk DJ, Weis-Garcia F, Malik S, Besmer E, Bustin M, Nussenzweig MC, Cortes P. 1997. V(D)J recombination: modulation of RAG1 and RAG2 cleavage activity on 12/23 substrates by whole cell extract and DNA bending proteins. J. Exp. Med. 185:2025–32


RAG PROTEINS AND V(D)J RECOMBINATION 9. Schatz DG, Oettinger MA, Schlissel MS. 1992. V(D)J recombination: molecular biology and regulation. Annu. Rev. Immunol. 10:359–83 10. Schatz DG. 1997. V(D)J recombination moves in vitro. Semin. Immunol. 9:149– 59 11. Gellert M. 1996. Recent advances in understanding V(D)J recombination. Adv. Immunol. 64:39–64 12. Sadofsky MJ, Hesse JE, McBlane JF, Gellert M. 1993. Expression and V(D)J recombination activity of mutated RAG1 proteins. Nucleic Acids Res. 22:5644– 50 13. Sadofsky MJ, Hesse JE, Gellert M. 1994. Definition of a core region of RAG-2 that is functional in V(D)J recombination. Nucleic Acids Res. 22:1805–9 14. Cuomo CA, Oettinger MA. 1994. Analysis of regions of RAG-2 important for V(D)J recombination. Nucleic Acids Res. 22:1810–14 15. Silver DP, Spanopoulou E, Mulligan RC, Baltimore D. 1993. Dispensable sequence motifs in the RAG-1 and RAG2 genes for plasmid-V(D)J recombination. Proc. Natl. Acad. Sci. USA 90:6100–4 16. McMahan CJ, Difilippantonio MJ, Rao N, Spanopoulou ES, Schatz DG. 1997. A basic motif in the N-terminal region of RAG1 enhances recombination activity. Mol. Cell. Biol. 17:4544–52 17. Roman CAJ, Cherry SR, Baltimore D. 1997. Complementation of V(D)J recombination deficiency in RAG-1(1/1) B cells reveals a requirement for novel elements in the N-terminus of RAG-1. Immunity 7:13–24 18. Kirch SA, Rathbun GA, Oettinger MA. 1998. Dual role of RAG2 in V(D)J recombination–catalysis and regulation of ordered Ig gene assembly. EMBO J. 17:4881–86 19. Steen SB, Han JO, Mundy C, Oettinger MA, Roth DB. 1999. Roles of the ‘‘dispensable’’ portions of RAG-1 and RAG-

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

521

2 in V(D)J recombination. Mol. Cell. Biol. 19:3010–17 Litman GW, Anderson MK, Rast JP. 1999. Evolution of antigen binding receptors. Annu. Rev. Immunol. 17:109– 47 Komori T, Okada A, Stewart V, Alt FW. 1993. Lack of N regions in antigen receptor variable region genes of TdT-deficient lymphocytes. Science 261:1171–75 Gilfillan S, Dierich A, Lemeur M, Benoist C, Mathis D. 1993. Mice lacking TdT: mature animals with an immature lymphocyte repertoire. Science 261:1175–78 Critchlow SE, Jackson SP. 1998. DNA end-joining: from yeast to man. Trends Biochem. Sci. 23:394–98 Hiom K, Gellert M. 1998. Assembly of a 12/23 paired signal complex: a critical control point in V(D)J recombination. Mol. Cell. 1:1011–19 Agrawal A, Eastman QM, Schatz DG. 1998. Transposition mediated by RAG1 and RAG2 and its implications for the evolution of the immune system. Nature 394:744–51 Hiom K, Melek M, Gellert M. 1998. DNA transposition by the RAG1 and RAG2 proteins: a possible source of oncogenic translocations. Cell 94:463– 70 Difilippantonio MJ, McMahan CJ, Eastman QM, Spanopoulou E, Schatz DG. 1996. RAG1 mediates signal sequence recognition and recruitment of RAG2 in V(D)J recombination. Cell 87:253–62 Spanopoulou E, Zaitseva F, Wang F-H, Santagata S, Baltimore D, Panayotou G. 1996. The homeodomain of Rag-1 reveals the parallel mechanisms of bacterial and V(D)J recombination. Cell 87:263–76 Rodgers KK, Villey IJ, Ptaszek L, Corbett E, Schatz DG, Coleman JE. 1999. A dimer of the lymphoid protein RAG1 recognizes the recombination signal sequence and the complex stably incor-

522

30.

31.

32.


33.

34.

35.

36.

37.

38.

39.

40.

FUGMANN ET AL porates the high mobility group protein HMG2. Nucleic Acids Res. 27:2938–46 Akamatsu Y, Oettinger MA. 1998. Distinct roles of RAG1 and RAG2 in binding the V(D)J recombination signal sequences. Mol. Cell. Biol. 18:4670–78 Hiom K, Gellert M. 1997. A stable RAG1-RAG2-DNA complex that is active in V(D)J cleavage. Cell 88:65–72 Mo X, Bailin T, Sadofsky MJ. 1999. RAG1 and RAG2 cooperate in specific binding to the recombination signal sequence in vitro. J. Biol. Chem. 274:7025–31 Santagata S, Aidinis V, Spanopoulou E. 1998. The effect of Me2` cofactors at the initial stages of V(D)J recombination. J. Biol. Chem. 273:16325–31 Swanson PC, Desiderio S. 1998. V(D)J recombination signal recognition–distinct, overlapping DNA-protein contacts in complexes containing RAG1 with and without RAG2. Immunity 9:115–25 Nagawa F, Ishiguro K, Tsuboi A, Yoshida T, Ishikawa A, Takemori T, Otsuka A, Sakano H. 1998. Footprint analysis of the RAG protein recombination signal sequence complex for V(D)J type recombination. Mol. Cell. Biol.:655–63 Cuomo CA, Mundy CL, Oettinger MA. 1996. DNA sequence and structure requirements for cleavage of V(D)J recombination signal sequences. Mol. Cell. Biol. 16:5683–90 Eastman QM, Leu TMJ, Schatz DG. 1996. Initiation of V(D)J recombination in vitro obeying the 12/23 rule. Nature 380:85–88 Hesse JE, Lieber MR, Mizuuchi K, Gellert M. 1989. V(D)J recombination: a functional definition of the joining signals. Genes Dev. 3:1053–61 Ramsden DA, McBlane JF, van Gent DC, Gellert M. 1996. Distinct DNA sequence and structure requirements for the two steps of V(D)J recombination signal cleavage. EMBO J. 15:3197–3206 Steen SB, Gomelsky L, Speidel SL, Roth

41.

42.

43.

44.

45.

46.

47.

48.

49.

DB. 1997. Initiation of V(D)J recombination in vivo: role of recombination signal sequences in formation of single and paired double-strand breaks. EMBO J. 16:2656–64 Roman CAJ, Baltimore D. 1996. Genetic evidence that the RAG1 protein directly participates in V(D)J recombination through substrate recognition. Proc. Natl. Acad. Sci. USA 93:2333–38 Sadofsky MJ, Hesse JE, van Gent DC, Gellert M. 1995. RAG-1 mutations that affect the target specificity of V(D)J recombination—a possible direct role of RAG-1 in site recognition. Genes Dev. 9:2193–99 Eastman QM, Villey IJ, Schatz DG. 1999. Detection of RAG protein-V(D)J recombination signal interactions near the site of DNA cleavage by UV crosslinking. Mol. Cell. Biol. 19:3788–97 Meisenheimer K, Meisenheimer P, Willis M, Koch T. 1996. High yield photocrosslinking of a 5-iodocytidine substituted RNA to its associated protein. Nucleic Acids Res. 24:981–82 Swanson PC, Desiderio S. 1999. RAG-2 promotes heptamer occupancy by RAG1 in the assembly of a V(D)J initiation complex. Mol. Cell. Biol. 19:3674–83 Bailin T, Mo X, Sadofsky MJ. 1999. A RAG1 and RAG2 tetramer complex is active in cleavage in V(D)J recombination. Mol. Cell. Biol. 19:4664–71 West RB, Lieber MR. 1998. The RAGHMG1 complex enforces the 12/23 rule of V(D)J recombination specifically at the double-hairpin formation step. Mol. Cell. Biol. 18:6408–15 Kim DR, Oettinger MA. 1998. Functional analysis of coordinated cleavage in V(D)J recombination. Mol. Cell. Biol. 18:4679–88 Aidinis V, Bonaldi T, Beltrame M, Santagata S, Bianchi ME, Spanopoulou E. 1999. The RAG1 homeodomain recruits HMG1,2 to facilitate RSS binding and to enhance the intrinsic DNA bending


50.

51.


52.

53.

54.

55.

56.

57.

58.

59.

activity of RAG1/2. Mol. Cell. Biol. 19:6532–42 Bustin M, Reeves R. 1996. High-mobility-group chromosomal proteins: architectural components that facilitate chromatin function. Prog. Nucleic Acid Res. Mol. Biol. 54:35–100 Bianchi ME, Beltrame M. 1998. Flexing DNA: HMG-box proteins and their partners. Am. J. Hum. Genet. 63:1573–77 Boonyaratanakornkit V, Melvin V, Prendergast P, Altmann M, Ronfani L, Bianchi ME, Taraseviciene L, Nordeen SK, Allegretto EA, Edwards DP. 1998. Highmobility group chromatin proteins 1 and 2 functionally interact with steroid hormone receptors to enhance their DNA binding in vitro and transcriptional activity in mammalian cells. Mol. Cell. Biol. 18:4471–87 Craig NL. 1988. The mechanism of conservative site-specific recombination. Annu. Rev. Genet. 22:77–105 Engelman A, Mizuuchi K, Craigie R. 1991. HIV-1 DNA integration: mechanism of viral DNA cleavage and DNA strand transfer. Cell 67:1211–21 van Gent DC, Mizuuchi K, Gellert M. 1996. Similarities between initiation of V(D)J recombination and retroviral integration. Science 271:1592–94 van Gent DC, McBlane JF, Ramsden DA, Sadofsky MJ, Hesse JE, Gellert M. 1995. Initiation of V(D)J recombination in a cell-free system. Cell 81:925–34 Savilahti H, Rice PA, Mizuuchi K. 1995. The phage mu transpososome core–DNA requirements for assembly and function. EMBO J. 14:4893–4903 Zhu XD, Sadowski PD. 1998. Selection of novel, specific single-stranded DNA sequences by Flp, a duplex-specific DNA binding protein. Nucleic Acids Res. 26:1329–36 Lewis SM. 1994. The mechanism of V(D)J joining: lessons from molecular, immunological, and comparative analyses. Adv. Immunol. 56:27–150

523

60. Spanopoulou E, Cortes P, Shih C, Huang CM, Silver DP, Svec P, Baltimore D. 1995. Localization, interaction, and RNA binding properties of the V(D)J recombination-activating proteins RAG1 and RAG2. Immunity 3:715–26 61. Leu TMJ, Schatz DG. 1995. Rag-1 and rag-2 are components of a high-molecular-weight complex, and association of rag-2 with this complex is rag-1 dependent. Mol. Cell. Biol. 15:5657–70 62. Grawunder U, Schatz DG, Leu TMJ, Rolink A, Melchers F. 1996. The half-life of RAG-1 protein in precursor B cells is increased in the absence of RAG-2 expression. J. Exp. Med. 183:1731–37 63. Sheehan KM, Lieber MR. 1993. V(D)J recombination: signal and coding joint resolution are uncoupled and depend on parallel synapsis of the sites. Mol. Cell. Biol. 13:1363–70 64. Namgoong SY, Harshey RM. 1998. The same two monomers within a MuA tetramer provide the DDE domains for the strand cleavage and strand transfer steps of transposition. EMBO J. 17:3775–85 65. Agrawal A, Schatz DG. 1997. RAG1 and RAG2 form a stable post-cleavage synaptic complex with DNA containing signal ends in V(D)J recombination. Cell 89:43–53 66. Eastman QM, Schatz DG. 1997. Nicking is asynchronous and stimulated by synapsis in 12/23 rule-regulated V(D)J cleavage. Nucleic Acids Res. 25:4370–78 66a. Yu K, Lieber MR. 1999. Mechanistic basis for coding end sequence effects in the initiation of V(D)J recombination. Mol. Cell. Biol. 19:8094–8102 66b. Gerstein RM, Lieber MR. 1993. Coding end sequence can markedly affect the initiation of V(D)J recombination. Genes Dev. 7:1459–69 67. Junop MA, Haniford DB. 1996. Multiple roles for divalent metal ions in DNA transposition: distinct stages of Tn10 transposition have different Mg2` requirements. EMBO J. 15:2547–55 68. Lewis SM, Hesse JE, Mizuuchi K, Gel-

524

69.

70.

71.


72.

73.

74.

75.

76.

77.

78.

79.

FUGMANN ET AL lert M. 1988. Novel strand exchanges in V(D)J recombination. Cell 55:1099–107 Lewis SM, Hesse JE. 1991. Cutting and closing without recombination in V(D)J joining. EMBO J. 10:3631–39 Melek M, Gellert M, van Gent DC. 1998. Rejoining of DNA by the RAG1 and RAG2 proteins. Science 280:301–3 Roth DB, Nakajima PB, Menetski JP, Bosma MJ, Gellert M. 1992. V(D)J recombination in mouse thymocytes: double-stranded breaks near T-cell receptor delta rearrangement signals. Cell. 69:41–53 Schlissel M, Constantinescu A, Morrow T, Baxter M, Peng A. 1993. Doublestrand signal sequence breaks in V(D)J recombination are blunt, 5’-phosphorylated, RAG-dependent, and cell cycle regulated. Genes Dev. 7:2520–32 Liva`k F, Schatz DG. 1996. T-cell receptor alpha locus V(D)J recombination byproducts are abundant in thymocytes and mature T cells. Mol. Cell. Biol. 16:609– 18 Liva`k F, Schatz DG. 1997. Identification of V(D)J recombination coding end intermediates in normal thymocytes. J. Mol. Biol. 267:1–9 Schlissel MS. 1998. Structure of nonhairpin coding-end DNA breaks in cells undergoing V(D)J recombination. Mol. Cell. Biol. 18:2029–37 Ramsden DA, Gellert M. 1995. Formation and resolution of double-strand break intermediates in V(D)J rearrangement. Genes Dev. 9:2409–20 Ramsden DA, Paull TT, Gellert M. 1997. Cell-free V(D)J recombination. Nature 388:488–91 Leu TMJ, Eastman QM, Schatz DG. 1997. Coding joint formation in a cell free V(D)J recombination system. Immunity 7:303–14 Lewis SM. 1994. P nucleotide insertions and the resolution of hairpin DNA structures in mammalian cells. Proc. Natl. Acad. Sci. USA 91:1332–36

80. Staunton JE, Weaver DT. 1994. Scid cells efficiently integrate hairpin and linear DNA substrates. Mol. Cell. Biol. 14:3876–83 81. Lewis SM. 1999. Palindromy is eliminated through a structure-specific recombination process in rodent cells. Nucleic Acids Res. 27:2521–28 82. Paull TT, Gellert M. 1998. The 3’ to 5’ exonuclease activity of Mre11 facilitates repair of DNA double-strand breaks. Mol. Cell. 1:969–79 83. Paull TT, Gellert M. 1999. Nbs1 potentiates ATP-driven DNA unwinding and endonuclease cleavage by the Mre11/ Rad50 complex. Genes Dev. 13:1276–88 84. Besmer E, Mansilla-Soto J, Cassard S, Sawchuk DJ, Brown G, Sadofsky M, Lewis SM, Nussenzweig MC, Cortes P. 1998. Hairpin coding end opening is mediated by RAG1 and RAG2 proteins. Mol. Cell. 2:817–28 85. Shockett PE, Schatz DG. 1999. DNA hairpin opening mediated by the RAG1 and RAG2 proteins. Mol. Cell. Biol. 9:4159–66 86. Bolland S, Kleckner N. 1996. The three chemical steps of Tn10/IS10 transposition involve repeated utilization of a single active site. Cell 84:223–33 87. Kennedy A, Guhathakurta A, Kleckner N, Haniford D. 1998. Tn10 transposition via a DNA hairpin intermediate. Cell 95:125–34 87a. Santagata S, Besmer E, Villa A, Bozzi F, Allingham JS, Sobacchi C, Haniford DB, Vezzoni P, Nussenzweig MC, Pan Z-Q, Cortes P. 1999. The RAG1/RAG2 complex constitutes a 3’ flap endonuclease: implications for junctional diversity in V(D)J and transpositional recombination. Mol. Cell. Biol. In press 88. Nadel B, Feeney AJ. 1997. Nucleotide deletion and P addition in V(D)J recombination—a determinant role of the coding-end sequence. Mol. Cell. Biol. 17:3768–78 89. Nadel B, Feeney AJ. 1995. Influence of


90.

91.


92.

93.

94.

95.

96.

97.

98.

coding-end sequence on coding-end processing in V(D)J recombination. J. Immunol. 155:4322–29 Roth DB, Menetski JP, Nakajima PB, Bosma MJ, Gellert M. 1992. V(D)J recombination: broken DNA molecules with covalently sealed (hairpin) coding ends in scid mouse thymocytes. Cell 70:983–91 Zhu CM, Bogue MA, Lim DS, Hasty P, Roth DB. 1996. Ku86-deficient mice exhibit severe combined immunodeficiency and defective processing of V(D)J recombination intermediates. Cell 86: 379–89 Gao YJ, Chaudhuri J, Zhu CM, Davidson L, Weaver DT, Alt FW. 1998. A targeted DNA-PKcs-null mutation reveals DNAPK-independent functions for ku in V(D)J recombination. Immunity 9:367– 76 Ramsden DA, Gellert M. 1998. Ku protein stimulates DNA end joining by mammalian DNA ligases: a direct role for Ku in repair of DNA double-strand breaks. EMBO J. 17:609–14 Smider V, Rathmell WK, Brown G, Lewis S, Chu G. 1998. Failure of hairpinended and nicked DNA to activate DNAdependent protein kinase—implications for V(D)J recombination. Mol. Cell. Biol. 18:6853–58 Yavuzer U, Smith GCM, Bliss T, Werner D, Jackson SP. 1998. DNA end-independent activation of DNA-PK mediated via association with the DNA-binding protein C1d. Genes Dev. 12:2188–99 Calsou P, Frit P, Humbert O, Muller C, Chen DJ, Salles B. 1999. The DNAdependent protein kinase catalytic activity regulates DNA end processing by means of Ku entry into DNA. J. Biol. Chem. 274:7848–56 Satoh MS, Lindahl T. 1992. Role of poly(ADP-ribose) formation in DNA repair. Nature 356:356–58 Ruscetti T, Lehnert BE, Halbrook J, Le Trong H, Hoekstra MF, Chen DJ, Peter-

99.

100.

101.

102.

103.

104.

105.

106.

525

son SR. 1998. Stimulation of the DNAdependent protein kinase by poly(ADP-ribose) polymerase. J. Biol. Chem. 273:14461–67 Smider V, Chu G. 1997. The end-joining reaction in V(D)J recombination. Semin. Immunol. 9:189–97 Guidos CJ, Williams CJ, Ildiko G, Knowles G, Huang MTF, Danska JS. 1996. V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid lymphocyte precursors. Genes Dev. 10:2038–54 Bogue MA, Jhappan C, Roth DB. 1998. Analysis of variable (diversity) joining recombination in DNA dependent protein kinase (DNA-PK)-deficient mice reveals DNA-PK-independent pathways for both signal and coding joint formation. Proc. Natl. Acad. Sci. USA 95:15559–64 Taccioli GE, Amatucci AG, Beamish HJ, Gell D, Xiang XH, Arzayus MIT, Priestley A, Jackson SP, Rothstein AM, Jeggo PA, Herrera VLM. 1998. Targeted disruption of the catalytic subunit of the DNA-PK gene in mice confers severe combined immunodeficiency and radiosensitivity. Immunity 9:355–66 Lieber MR, Hesse JE, Mizuuchi K, Gellert M. 1988. Lymphoid V(D)J recombination: nucleotide insertion at signal joints as well as coding joints. Proc. Natl. Acad. Sci. USA 85:8588–92 Candeias S, Muegge K, Durum SK. 1996. Junctional diversity in signal joints from T cell receptor beta and delta loci via terminal deoxynucleotidyl transferase and exonucleolytic activity. J. Exp. Med. 184:1919–26 Sakano H, Hu¨ppi K, Heinrich G, Tonegawa S. 1979. Sequences at the somatic recombination sites of immunoglobulin light-chain genes. Nature 280:288–94 Chow SA, Vincent KA, Ellison V, Brown PO. 1992. Reversal of integration and DNA splicing mediated by integrase of

526

107.

108.

109.


110.

111.

112.

113.

114.

115.

116.

117.

FUGMANN ET AL human immunodeficiency virus. Science 255:723–26 Mizuuchi K. 1992. Transpositional recombination: mechanistic insights from studies of Mu and other elements. Annu. Rev. Biochem. 61:1011–51 Kleckner N, Chalmers RM, Kwon D, Sakai J, Bolland S. 1996. Tn10 and IS10 transposition and chromosome rearrangements: mechanism and regulation in vivo and in vitro. Curr. Top. Microbiol. Immunol. 204:49–82 Craig NL. 1996. Transposon Tn7. Curr. Top. Microbiol. Immunol. 204:27–48 Asante-Appiah E, Skalka AM. 1997. Molecular mechanisms in retrovirus DNA integration. Antiviral Res. 36:139– 56 Sakai J, Chalmers RH, Kleckner N. 1995. Identification and characterization of a pre-cleavage synaptic complex that is an early intermediate in Tn10 transposition. EMBO J. 14:4374–83 Bolland S, Kleckner N. 1995. The two single-strand cleavages at each end of Tn10 occur in a specific order during transposition. Proc. Natl. Acad. Sci. USA 92:7814–18 Beckman RA. 1985. On the fidelity of DNA replication: manganese mutagenesis in vitro. Biochemistry 24:5810–17 Hsu M, Berg P. 1978. Altering the specificity of restriction endonuclease: effect of replacing Mg2` with Mn2`. Biochemistry 17:131–38 Plasterk RH. 1996. The Tc1/mariner transposon family. Curr. Top. Microbiol. Immunol. 204:125–43 Beall EL, Rio DC. 1997. Drosophila Pelement transposase is a novel site-specific endonuclease. Genes Dev. 11:2137– 51 Dreyfus DH. 1992. Evidence suggesting an evolutionary relationship between transposable elements and immune system recombination sequences. Mol. Immunol. 29:807–19

118. Davies DR, Braam LM, Reznikoff WS, Rayment I. 1999. The three-dimensional structure of a Tn5 transposase-related ˚ resolution. J. protein determined to 2.9-A Biol. Chem. 274:11904–13 119. Colot V, Haedens V, Rossignol JL. 1998. Extensive, nonrandom diversity of excision footprints generated by ds-like transposon Ascot-1 suggests new parallels with V(D)J recombination. Mol. Cell. Biol. 18:4337–46 120. Stellwagen AE, Craig NL. 1997. Avoiding self–two Tn7-encoded proteins mediate target immunity in Tn7 transposition. EMBO J. 16:6823–34 121. Gary PA, Biery MC, Bainton RJ, Craig NL. 1996. Multiple DNA processing reactions underlie Tn7 transposition. J. Mol. Biol. 257:301–16 122. Yang W, Steitz TA. 1995. Recombining the structures of HIV integrase, RuvC and RNaseH. Structure 3:131–34 123. Rice P, Craigie R, Davies DR. 1996. Retroviral integrases and their cousins. Curr. Opin. Struct. Biol. 6:76–83 124. Kulkosky J, Jones KS, Katz RA, Mack JP, Skalka AM. 1992. Residues critical for retroviral integrative recombination in a region that is highly conserved among retroviral/retrotransposon integrases and bacterial insertion sequence transposases. Mol. Cell. Biol. 12:2331– 38 125. Polard P, Chandler M. 1995. Bacterial transposases and retroviral integrases. Mol. Microbiol. 15:13–23 126. Freemont PS, Friedman JM, Beese LS, Sanderson MR, Steitz TA. 1988. Cocrystal structure of an editing complex of Klenow fragment with DNA. Proc. Natl. Acad. Sci. USA 85:8924–28 127. Beese LS, Steitz TA. 1991. Structural basis for the 3’-5’ exonuclease activity of Escherichia coli DNA polymerase I: a two metal ion mechanism. EMBO J. 10:25–33



128. Steitz TA, Steitz JA. 1993. A general two-metal-ion mechanism for catalytic RNA. Proc. Natl. Acad. Sci. USA 90:6498–502 129. Lohe AR, Moriyama EN, Lidholm DA, Hartl DL. 1995. Horizontal transmission, vertical inactivation, and stochastic loss of mariner-like transposable elements. Mol. Biol. Evol. 12:62–72

527

130. Ivics Z, Hackett PB, Plasterk RH, Izsvak Z. 1997. Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91:501–10 131. Thompson CB. 1995. New insights into V(D)J recombination and its role in the evolution of the immune system. Immunity 3:531–39

Figure 2 Map of the sites of contact of the RAG proteins on a 12-RSS. A. Linear model of the 12RSS and coding flank DNA. B. Three-dimensional depiction as standard B-form DNA. Colors: yellow, the heptamer and nonamer; green, positions protected from cleavage in footprinting assays; light purple (phosphates); and orange (bases), sites of ethylation/methylation interference; dark purple and blue, sites of UV cross-linking to RAG1 only, or to both RAG1 and RAG2, respectively. Data derived from (30, 34, 35, 43).


FUGMANN ET AL C-1


C-2 FUGMANN ET AL

Figure 3 Schematic model of the two divalent metal ion active site model. The figure is focused on the phosphate ester bridges between the coding flank on left side and the RSS on the right side. The three steps, nicking (1), hairpin formation (2), and hairpin opening, transposition, hybrid joint and O/S joint formation (3) are proposed to be catalyzed by a single active site containing two Mg2+ ions. The Mg2+ ion that activates the nucleophile is shown in red; the ion that stabilizes the leaving group is shown in blue; and the nucleophile is shown in green. Solid arrows represent nucleophilic attacks, and dashed arrows indicate the activation of nucleophiles and the stabilization of leaving groups. Phosphate groups are drawn as a circled "P". See text for additional details.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


THE ROLE OF THE THYMUS IN IMMUNE RECONSTITUTION IN AGING, BONE MARROW TRANSPLANTATION, AND HIV-1 INFECTION Barton F. Haynes, M. Louise Markert, Gregory D. Sempowski, Dhavalkumar D. Patel, and Laura P. Hale


Departments of Medicine (BFH, DDP, GDS), Pediatrics (MLM) and Pathology (LPH) and the Human Vaccine Institute and Duke Center for AIDS Research, Duke University Medical Center, Durham, North Carolina 27710; e-mail: [email protected]

Key Words AIDS, autoimmunity, cytokines, DiGeorge syndrome, T cells Abstract The human thymus is a complex chimeric organ comprised of central (thymic epithelial space) and peripheral (perivascular space) components that functions well into adult life to produce naive T lymphocytes. Recent advances in identifying thymic emigrants and development of safe methods to study thymic function in vivo in adults have provided new opportunities to understand the role that the human thymus plays in immune reconstitution in aging, in bone marrow transplantation, and in HIV-1 infection. The emerging concept is that there are age-dependent contributions of thymic emigrants and proliferation of postthymic T cells to maintain the peripheral T cell pool and to contribute to T cell regeneration, with the thymus contributing more at younger ages and peripheral T cell expansion contributing more in older subjects. New studies have revealed a dynamic interplay between postnatal thymus output and peripheral T cell pool proliferation, which play important roles in determining the nature of immune reconstitution in congenital immunodeficiency diseases, in bone marrow transplantation, and in HIV-1 infection. In this paper, we review recent data on human postnatal thymus function that, taken together, support the notion that the human thymus is functional well into the sixth decade and plays a role throughout life to optimize human immune system function.

INTRODUCTION The thymus is essential for the initial establishment of the peripheral T cell pool in animals and humans (1–3). In humans, children born without a thymus (complete DiGeorge syndrome) lack functional peripheral T cells (3). The normal human thymus develops early on in fetal development, and is colonized by stem cells at 7 to 8 weeks of gestational age (4). By the time a human baby is born, 0732–0582/00/0410–0529$14.00

529


530

HAYNES ET AL

the peripheral T cell repertoire is established to the point that thymectomy (either for myasthenia gravis or in the context of other thoracic surgeries) does not cause immediate immune deficiency. The supposition is that thymectornized subjects, regardless of the age at thymectomy, will be immunologically healthy unless a need arises for the T cell arm of the immune system to be regenerated. In recent years, the most common clinical settings in which the T cell pool needs to be regenerated are cancer chemotherapy, bone marrow transplantation, and the acquired immunodeficiency syndrome due to human immunodeficiency virus type I (HIV-1) infection. While bone marrow transplantation has been observed to be clinically successful in cancer and congenital immune deficiency diseases, and varying degrees of immune reconstitution have been observed in AIDS patients on highly active antiretroviral therapy (HAART), until recently no direct techniques were available for study of human postnatal thymus function. This has been due to lack of definitive phenotypic markers of human recent thymic emigrants and to lack of accessibility of the thymus for routine biopsy and analysis. Recent advances have been made in understanding the structure of the human thymus, in identifying recent thymic emigrants, and in defining the half-life and proliferative capacity of various human T cell subsets. This paper presents a synthesis of our own data with data of others that support the view that the human thymus functions throughout life to modulate human immune function. Moreover, we discuss the prospects for reconstitution of the T cell arm of the immune system of children and adults when the need arises.

NEW TECHNIQUES FOR STUDY OF POSTNATAL THYMIC FUNCTION Kong et al demonstrated in chickens that recent thymic emigrants could be identified by a monoclonal antibody, chT1, and showed that peripheral chT1` T cell levels fell two weeks after thymectomy (5, 6). Unfortunately, the human homologue of chT1 is not expressed in the same manner in humans, and no single human marker of recent thymic emigrants with the characteristics of chT1 has been found. Nonetheless, a variety of markers are now used in studies of animal and human T cells that, in specific circumstances, are able to identify newly produced T cells (Table 1).

Phenotypic Markers of Recent Thymic Emigrants In mice and rats, markers have been found that identify thymic emigrants, including RT6hi`, CD45RChi` T cells in rat (7), and intermediate expression of heat stable antigen (HSA) in mice (8, 9). In humans, differential expression of CD45 isoforms has been used to identify the naive versus the memory population of CD4` and CD8` T cells (10–14). Concomitant use of high expression of CD62L

531

FUNCTION OF THE POSTNATAL HUMAN THYMUS TABLE 1 Surface markers and assays for measurement of recent thymic emigrants


Species

Marker

Comments

Reference

Chicken

chT1 monoclonal antibody

chT1` cells decrease 12 days after thymectomy

5, 6

Chicken

T cell receptor excision circles (TRECs)

TREC half-life of 2 weeks after thymectomy

6

Rat

Thy-1lo, CD45RChi`, RT6hi`

RTE markers in rats

7

These mark ;50% of RTE in LN and spleen

8, 9

int`

hi`

Mouse

HSA

Mouse

BrdUlo`

BrdU in drinking water labels only appproximately 1/2 of RTE

9

Human

CD4 or CD8, CD45RA`, CD62L`

Mark RTE but also marks a subset of long-lived naive (not RTE) and a small subset of memory T cells

10–15

Human

Signal joint (sj) TRECs

CD4RA` cells exclusivly contain TRECs

25

Human

?CD95 (fas) expression on CD4` CD45RA`, and CD4` CD45RO` T cells

CD95` T cells fall in both CD4` and CD8` subsets with aging

23

, Qa-2

RTE, recent thymic emigrants; BrdU, Bromodeoxyuridine; HSA, heat stable antigen; TREC, T cell receptor excision circles; int, intermediate expression; high, high expression

with CD45RA has provided improved identification of human naive T cells (15). In the setting of establishment of the peripheral T cell pool (14), or in studies of regenerating immune systems such as occurs after bone marrow transplantation (BMT), use of the CD45RA`, CD62L` markers identifies naive CD4` and CD8` T cells (13). However, CD45RO` T cells can revert back to CD45RA` cells and, in some circumstances, can remain functionally memory T cells (16– 19). Moreover, the long-lived population of T cells are primarily naive phenotype CD45RA` T cells, while the majority of CD45RO` memory T cells have relatively short life spans (20–22). Thus, in the adult CD45RA` T cells may be naive, but they need not be recent thymus emigrants. What is needed in humans for the simplified analysis of thymocyte dynamics in vivo is a marker of human recent thymic emigrants with the characteristics of chT1 marker in chicken (5, 6). Recently Aspinall et al have shown that CD4`, CD45RA`, CD95(fas)` T cells as well as CD8` CD45RA` CD95` T cells decline with age in humans, although the use of this combination of T cell subset markers for recent thymic emigrants has yet to be validated (23).

532

HAYNES ET AL


Signal Joint T Cell Receptor Excision Circles During T cell receptor rearrangement in newly developed T cells, the excised DNA persists episomally as T cell receptor (TCR) excision circles (TRECs) (24). Two TCRD locus circles are produced per cell from TCRA gene rearrangements in new TCRab` T cells, and then are diluted out of naive T cell populations as T cells proliferate and expand in the presence of antigen in the periphery (Figure 1) (24, 25). Kong et al have shown in the chicken (6) and Douek et al in human (25) that signal joint (sj) TREC` T cells are either recent thymic emigrants or are thymic emigrants of unknown time that have not divided to the point of diluting out sjTRECs. Thus, the level of sjTRECS in peripheral blood (PB) T cells is an excellent measure of thymic function in postnatal humans during periods of reconstitution of the peripheral immune system, and a reasonable measure of thymic function in general (25). Importantly, Douek et al found sjTRECs predominated in CD45RA` but not in CD45RO` peripheral blood T cells, confirming the recent emigration or lack of extensive postthymic division of the CD45RA` T cell subset (25). Moreover, they found that thymectomized subjects had approximately a log fewer sjTRECs than did nonthymectomized agedmatched controls (25). These latter data suggested that the thymus might function to some degree throughout life to maintain sjTREC` (naive) T cells at the required levels via production of new sjTREC` thymocytes. In addition, these data also demonstrated the existence of a long-lived population of CD45RA`, sjTREC` T cells that, even 10` years after thymectomy, had divided few enough times to remain sjTREC` (25). Thus, by combining phenotypic analysis with sjTREC determinations, investigators have now begun to probe human thymic function in adults and children and to study the role of the thymus versus peripheral expansion of the postthymic T cell pool in reconstitution of T cells in a variety of diseases and clinical settings.

THE HUMAN THYMUS DURING AGING Understanding the changes the human thymus undergoes in aging, in HIV-1 infection, and in autoimmune diseases such as myasthenia gravis has required an appreciation of the chimeric nature of the human thymus (26–31). Studies over the last 15 years have revealed that the thymus is comprised of two key components, the thymic epithelial space in which thymopoiesis occurs, and the nonepithelial perivascular space, which usually contains no thymopoiesis (26–31). Although previous studies suggested that the thymus begins to atrophy only after puberty, in 1986, Steinmann et al made the important observation that the thymopoietic thymic epithelial space actually begins to atrophy at the age of one year, and shrinks in volume by ;3%/yr through middle age, then shrinks by ,1%/yr throughout the rest of life (26, 28). Extrapolation of these data suggests that total loss of all thymopoietic tissue would occur by approximately age 105 (32). When these two


FUNCTION OF THE POSTNATAL HUMAN THYMUS

533

Figure 1 Generation of signal joint (sj) and coding joint (cj) TRECs during TCRA rearrangement. Figure shows a simplified representation of the TCRD locus flanked by portions of the TCRA locus. Rearrangement of the TCRA gene forms a single TREC containing a unique sj sequence. The cjTREC is formed after TCRA V to TCRA J occurs. (Adapted and used with permission from Ref. 25).

human thymic components are evaluated separately, it can be seen that between ages 1 and 50` years of age, the perivascular space increases and the thymic epithelial space decreases to approximately 10% of its original volume (33, 34) (Figures 2 and 3; for Figure 2 see color insert). Flores and colleagues in our group have studied the thymic perivascular space in 87 normal and 31 myasthenia gravis thymuses from subjects newborn to 78 years of age (33). In this study, we found that the thymus perivascular space


534

HAYNES ET AL

Figure 3 Changes in the thymic epithelial space and thymic perivascular space of the normal human thymus with aging. Figure shows means of % perivascular space (PVS) or thymic epithelial space (TES) of groups of thymuses by age quintile (33, 34). The thymopoietic TES falls over time to ;20% of the total space by 50` years, while the PVS reciprocally rises to ;80% of the total space (data summarized from Ref. 33, 34).

contained primarily mature CD1a-, CD4` or CD8` T lymphocytes that expressed CD45RO, including clusters of T cells that expressed TIA-l` cytotoxic granules (33). Taken together with our observation in HIV-1` thymuses of expansion of thymic perivascular space lymphocytes that were predominantly CD3`, CD8`, and TIA-l` (30) (see below), these data suggested the hypothesis that in adults, prime components of the perivascular space were peripheral T cells that were not directly involved in thymopoiesis (30, 31, 33). It has long been recognized that thymuses with well-defined cortex and medulla suggestive of ongoing thymopoiesis are frequently found at autopsy of middle-aged subjects. Smith & Ossa-Gomez found thymocytes in both the thymic cortex and medulla in three subjects in their 70s (35). Bertho et al found histologic evidence of active thymopoiesis to age 49 (the oldest thymus studied) (36). While not able to provide data on mature thymocyte production, these studies strongly suggested human thymus function throughout life. The gradual decline in thymic epithelial space from age 1 to 50` year also predicted that human thymus function should be present in most normal subjects (26, 28, 33, 34). Douek et al used sjTREC analysis of peripheral blood CD4` and CD8` T cells to show a gradual fall in peripheral blood CD4` and CD8` T cell sjTREC levels during postnatal life (25). In comparing both normal and thymectornized subjects during aging, they found measurable sjTREC levels up to age 70.


535


Douek et al (25), Jameson et al (37), and our group (34) have studied sjTREC levels in human thymocytes and found constant sjTREC levels per 105 thymocytes during postnatal life through age 50 years. When whole thymus tissue DNA was analyzed for sjTREC levels and expressed as sjTRECs per mg of thymus tissue, we found that sjTREC levels fell throughout life from 1 to 68 years but were readily detectable through age 60 (34). Using the linker-mediated polymerase chain reaction (LM-PCR) assay, we showed ongoing TCRB gene rearrangements up to age 46 (the oldest patient studied). Similarly, Poulin et al have recently demonstrated TCRB TRECs in adult thymus tissue (38). Thus, sjTREC levels are relatively constant per 105 thymocytes with age, while per unit of whole thymus tissue weight, sjTRECs declined during adult life. That is to say, thymus function is constant per thymocyte unit, but the total number of thymocyte units of the thymic engine falls with aging. Taken together, these data have definitively shown that the human thymus is indeed functional in adults at least through approximately age 60.

Morphology of the Aging Human Thymus Whereas the human thymic epithelial space thymocyte component begins to involute in year 1 of life, the thymic perivascular space lymphocyte component involutes at a different pace, initially increasing in size from year 1 to young adulthood and then decreasing in volume thereafter (26, 28, 33, 34, 36) (Figure 3). In contrast to the mouse thymus, in which the total size shrinks with age, the human thymus does not lose its volume during aging, but rather over time the thymus content is gradually shifted from thymic epithelial to perivascular space (28). The perivascular lymphoid and other hematopoietic cell components decrease after young adulthood and are taken over by adipocyte proliferation and differentiation as the thymus ages (26, 28, 33, 34) (Figure 2C).

Pathophysiology of Thymic Atrophy Theories of why the thymus atrophies have included lack or block of TCR rearrangements (39), loss of self-peptides on thymic epithelial MHC molecules (40), aging of thymic stroma with loss of trophic cytokines (41–43), and aging of the stem cell population (44, 45). We have recently studied steady-state cytokine mRNA expression by RNase protection assays in normal aging thymuses, and we identified the changes that occurred in cytokine gene expression during aging (Figure 4) (34). We found a group of thymus cytokines whose expression fell during aging (IL-2, IL-9, IL-10, IL-13, and IL14) as well as a group whose expression levels did not significantly change with aging [IL-15, granulocyte colony stimulating factor (G-CSF), and IL-7]. Unexpectedly, we found a series of cytokines whose expression dramatically increased in the thymus with age. These were leukemia inhibitor factor (LIF), oncostatin M (OSM), IL-6, stem cell factor (SCF), and macrophage-colony stimulating factor (M-CSF) (34). That IL7 mRNA levels remained constant in the face of progressive atrophy suggested to us that overproduction of thymic cytokines might be involved in an active


536

HAYNES ET AL

Figure 4 Summary of cytokine changes that occur in normal human thymus during aging. Figure schematically represents 3 groups of cytokine mRNAs expressed during thymus aging as determined by quantitative RNase protection assays (34). One group of cytokine mRNAs fell with aging (IL-2, IL-9, IL- 10, IL-13, and IL-14), one group did not change (lL-7, IL-15, and G-CSF), and a group was overexpressed (LIF, OSM, SCF, IL-6, and M-CSF) (34).

process of cytokine induction of thymic atrophy. Administration of these overexpressed cytokines individually to Balb/c mice demonstrated that each cytokine except M-CSF induced significant thymic atrophy, with LIF and OSM inducing the most severe thymocyte loss (34). It is interesting that transgenic mice that overexpress LIF, OSM, and SCF in thymus have thymic abnormalities (46–49). LIF and OSM transgenic mice have decreased double positive thymocytes and numerous thymic B cell follicles (46, 47), and injection by others of mice with LIF has also induced thymic atrophy (48). Thus, LIF, OSM, and SCF all have been reported by others to have regulatory effects on thymopoiesis. Active suppression of thymopoiesis has been shown by administration of testosterone or by reversal of thymic atrophy by testosterone inhibitors such as administration of luteinizing hormone releasing hormone or surgical castration (50). Corticosteroids can actively suppress thymopoiesis with selective loss of cortical thymocytes (52). Pituitary ACTH production drives adrenal corticosteroid production. Interestingly, LIF-deficient animals have decreased ACTH; thus, pituitary ACTH production is regulated by LIF in an as-yet-unknown way (53). Taken together, these data have raised the notion that as with systemic hormone production, locally produced thymus cytokines may actively suppress thymopoiesis.



537

One theory of human thymus aging is the wear and tear model, with thymic atrophy due to loss of thymic microenvironment function (32, 41–43). Although the human thymus is capable of dramatic regeneration after involution caused by stress, steroids, infections, pregnancy, and cancer chemotherapy, it does not regenerate in response to thymic aging. Thus, a key question is why the thymus does not counter the forces that drive thymic atrophy during aging. George & Ritter have proposed that evolutionarily the thymus has not been needed beyond age 40 until recently, when life spans have increased from age 40 in the mid-1800s to nearly age 80 today (32). They postulated that when humans did not travel and lived short lives, they interacted with all of the pathogens in the environment at early ages (32). Thymic involution then was beneficial in that it saved energy for the species to use for reproduction. Another postulated advantage to the organism of thymic atrophy is to protect the host from autoimmune diseases (54). Thus, it is plausible that the combined evolutionary advantages of not maintaining an unnecessary thymus after establishing a sufficiently broad T cell repertoire to deal with the myriad of infectious agents, as well as requiring protection against generation of new autoreactive T cell clones, together confers an evolutionary advantage favoring thymic atrophy.

The Aging Human Thymus in Autoimmune Disease: Myasthenia Gravis Myasthenia gravis is an autoimmune disease in which autoantibodies are generated that react with acetylcholine receptors (AchR) and block neuromuscular transmission, resulting in muscle weakness (27). Because 10% of myasthenia gravis patients have a thymoma (epithelial tumor of the thymus that is routinely surgically removed), it was realized early on that thymectomy in myasthenia gravis was associated with improvement in muscle strength. While B cells producing anti-AchR antibody are found in myasthenic thymuses, many patients improve clinically after thymectomy, yet have no change in anti-AchR titers. Thus, the role the thymus plays in myasthenia gravis and how thymectomy improves symptoms in myasthenia gravis are unknown. Thymuses from myasthenia gravis patients have morphologies that are exaggerated versions of changes found in normal aging thymuses (27, 31). Sixty percent of myasthenia gravis thymuses are termed hyperplastic with B cell germinal centers seen throughout the perivascular space (27). The term thymic hyperplasia is actually a misnomer, since the true thymic epithelial space itself is generally hypoplastic in myasthenia gravis with decreased thymopoiesis (Figure 5, see color insert) (27, 31). When the myasthenia gravis thymic epithelial space during aging is compared with the that of normal aging thymus, a more rapid atrophy of the true epithelial space of the thymus is observed in myasthenia gravis thymuses than in normal thymuses (Figure 6) (33, 34). Study of the cytokines in aged myasthenia gravis thymuses showed similar overproduction of the same thymopoiesis-suppressing cytokines as was found in normal atrophic thymuses (34).


538

HAYNES ET AL

Figure 6 Atrophy of the thymopoietic thymic epithelial space in normal subjects and myasthenia gravis (MG) during aging. Figure shows that after approximately age 25, MG thymuses atrophy faster than normals (33, 34). Standard error of the mean for normals ages 26–49 was 54, and for age .50 was 53.6. For MG ages 26–49 standard error was 54, and for age .50 was 57. p,.05 for MG versus normal for % thymic epithelial space in these two age groups (data summarized from Ref. 33).

Comparison of T Cell Regeneration Via Thymopoiesis Versus Peripheral T Cell Expansion Table 2 summarizes a number of features of the T cell repertoires generated from thymopoiesis versus peripheral T cell expansion. The T cell repertoire generated from thymopoiesis has a full TCR repertoire and thus is capable of mediating T cell responses to neoantigens. In contrast, peripheral expansion of existing T cell pools may lead to T cell repertoires limited to those of existing memory T cells, with limited capacity to respond to new antigens (55–57). Thymopoiesis is more efficient in restoring peripheral T cell levels than is peripheral expansion of T cells. Mackall & Gress have demonstrated that in the presence of a functioning thymus, peripheral T cell expansion may be suppressed; this suppression is reversed by eliminating thymic-derived progeny (13). In the absence of active thymopoiesis, peripheral expansion of T cells is enhanced (13), although peripheral expansion may not restore a full peripheral T cell repertoire to normal (14, 55–57). The phenotype of peripheral blood T cells can be helpful for evaluating the mode of T cell reconstitution. T cell reconstitution via thymopoiesis is associated with the presence of CD45RA` T cells, while T cell regeneration via peripheral expansion is associated with CD45RO` T cells (14, 55–57). Because PB T cells


539

TABLE 2 Attributes of T cell regeneration via thymopoiesis versus peripheral T cell expansion Thymopoiesis 1. Responses to new antigens 2. Full polyclonal TCR repertoire from ‘‘recapitulation of T cell ontogeny’’


3. Efficient in restoring T cell pool to normal levels 4. In presence of thymus, peripheral T cell expansion is downregulated 5. New T cells from thymic reconstitution have CD45RA`, CD62L` phenotype 6. Expansion driven by MHC-peptide signals of positive selection 7. More often present in younger subjects

Peripheral T Cell Expansion 1. Poor responses to new antigens 2. Frequently oligoclonal TCR repertoire. TCR diversity limited to starting population of peripheral T cells 3. Not efficient in restoring normal T cell pool levels 4. In absence of thymus, peripheral T cell expansion has CD45RO` phenotype 5. New progeny of peripheral T cell expansion have CD45RO` phenotype 6. Expansion driven by peripheral MHCpeptide stimulation of dividing memory T cells 7. More often present in older subjects

Summarized from References 55–57 and 107.

are rapidly dividing during peripheral T cell expansion, regenerating T cells can express CD45RO even if the expanding T cells were CD45RA` prior to activation (14). Because no peripheral T cell expansion occurs in mouse models in the absence of peripheral MHC antigen expression, it has been realized that maintenance of the peripheral pool of T cells is driven by peripheral T cell TCR ligation of self-antigen peptides complexed to MHC expressed on peripheral antigen presenting cells (reviewed in 58). In an elegant series of experiments, Mackall & Gress have clearly defined the age-related dependence of two T cell regeneration pathways in patients receiving cancer chemotherapy. In younger patients (,15 yrs of age), T cell regeneration was faster and involved CD4` CD45RA` cells (14, 55–57). In older patients (.18 yrs of age), CD4` T cell level was delayed, and CD4` CD45RA` T cell levels were low after chemotherapy (14). There was an inverse correlation between age and CD4` T cell levels 6 months after chemotherapy. Thus, thymopoiesis is more likely operative in younger subjects, and peripheral expansion of T cells is more likely operative in older subjects to regenerate and maintain the pool of peripheral T cells (14).

The Interplay Between the Thymus and the Peripheral T Cell Pools Recent data in mice regarding how the naive and memory CD8` T cell pools are maintained have shaped the way we think about human CD4` and CD8` T cell pools. In particular, from the previous section it has become clear that in most


540

HAYNES ET AL

postnatal humans, some combination of thymopoiesis and peripheral T cell expansion—the relative amounts dependent on host age—combine to make up and maintain the peripheral T cell pool. But how does the peripheral pool know how many cells to have in it, and how is the ratio of naive to memory T cells maintained? There is emerging evidence in mice that the naive and memory CD8` T cell pools are independently regulated, with each pool of T cells having its own niche (Table 3). For CD8` T cells, both naive and memory T cells require MHC class I expression on peripheral antigen-presenting cells to survive (58). However, naive T cells survive by not dividing, while memory T cells survive by proliferating upon interaction with MHC (58). There is some evidence in animals that the rate of thymic emigrant production is fixed and is not responsive to events in the periphery (59, 60). If so, the importance of separate niches for naive and memory T cells is paramount for the maximal assurance that there will be naive T cells present for the longest period of time, particularly if thymic export dwindles (58, 61). Thus, when thymus function is high and naive T cell emigrants are produced in large amounts, turnover is high in the peripheral naive T cell pool, in order to make ‘‘room’’ for new thymic emigrants. However, the immune system must also retain its memory pool to be versatile (58, 61). As thymus output gradually dwindles with age, the half-life of naive T cells grows longer, thus ensuring the organism of a supply of naive T cells even when the thymus has stopped production (58). These data explain the persistence of very low but measurable levels of sjTREC` T cells in thymectomized humans many years after thymectomy (25). These concepts also may explain why peripheral T cell expansion alone is less efficient in filling up the peripheral T cell pool. Memory T cells fill the memory T cell niche but not the naive T cell niche. While direct data are TABLE 3 Attributes of naive versus memory CD8` T cell pools Naive

Memory

1. Require dendritic cell and MHC class I expression to survive

1. Require MHC class I expression to survive

2. Survive in host tissues without dividing

2. Survive in host tissues by dividing

3. Long-lived

3. Short-lived

4. Has naive CD8` T cell niche independent of other T cell subsets

4. Has memory CD8` T cell niche independent of other T cell subsets

5. Thymic output of naive cells is independent of the peripheral naive T cell pool

5. Memory T cell proliferation increases in response to T cell depletion but initially only fills the memory T cell pool niche

6. Naive CD8` T cells have short life spans when thymopoiesis is active and long life spans when thymopoiesis is minimal Summarized from Reference 58.


541

lacking for a similar peripheral niche for CD4` naive T cells, CD4` T cells do require MHC class II antigen to survive in the periphery (58), and in particular, naive CD4` T cells cannot survive without peripheral dendritic cells (62, 63).


RECONSTRUCTING THE HUMAN THYMUS: THYMIC TRANSPLANTATION IN COMPLETE DIGEORGE SYNDROME A key issue for exploring the regenerative capacity of the human immune system has been to determine the regenerative capacity of the postnatal human thymus. Studies of growth of fetal human thymus tissue in the SCID/Hu Thy/Liv model clearly demonstrated the superiority of fetal versus postnatal thymus tissue as a transplant graft (64). However, we found human postnatal thymus would grow in SCID mice, and when mouse NK cells were eliminated in SCID mice with anti-asialo-GM1 antibody, human postnatal thymus grafts formed mouse-human chimeric thymuses with ‘‘leaky’’ SCID mouse T cell precursors (65). Given the ready availability of human postnatal thymus tissue versus the scarcity of fetal thymus for transplantation, a team at Duke led by Louise Markert has studied postnatal human thymus transplantation to reconstitute the T cell arm of immunity in complete DiGeorge syndrome, a congenital malformation of the third and fourth pharyngeal pouches in which the thymus is absent (3, 66–68). Some DiGeorge syndrome patients have low but present T cell levels and function (partial DiGeorge syndrome), while others are completely devoid of T cells or function (the latter are termed complete DiGeorge syndrome) (3). While most attempts to treat complete DiGeorge syndrome have been inconclusive, rare patients have clearly been reconstituted by adoptive transfer of mature T cells (69–71) or by fetal thymus transplantation (72–74). In studies of allogeneic postnatal thymus transplants, we have used deoxyguanosine-treated thymus explants to kill immunocompetent mature donor thymocytes to minimize graft-versus-host disease (3, 66, 67). T cell function developed in 4 of 5 patients treated with allogeneic partially HLA-matched thymus tissue transplanted into the thigh (3, 60). Three patients died from causes unrelated to the transplantation, while two patients have had complete restoration of T cell function and establishment of a normal T cell repertoire (3). Both of the latter patients have had thymus graft biopsies after restoration of T cell function that showed normal thymus with large areas of normal thymopoiesis (Figure 7, see color insert) and rises in sjTREC levels in peripheral T cells (3). Sequencing of T cell receptor genes and reactivity with anti-TCR Vb monoclonal antibodies showed normal diversification of the T cell repertoire after thymus transplantation (3, 66). Thus, in humans, postnatal thymus (0–3 months of age) is capable of regenerating a new T cell repertoire in vivo in the setting of complete DiGeorge syndrome (3, 66).

542

HAYNES ET AL


That allogeneic thymus tissue is successful in this setting is important for two reasons. First, the lack of any host T cell response to alloantigens is key to thymus transplant survival and function, since similar thymus transplants in AIDS patients are uniformly rejected by functional host T cells (75). Second, both successful thymus grafts were mismatched at several HLA class I and class II loci, yet patient T cells recognized antigen in the context of host HLA (L Markert, unpublished observations). Thus, as has been described with bone marrow transplantation in SCID syndrome, it is likely that peripheral tolerance to host HLA antigen is present in engrafted thymus transplant patients (76). To date, none of these patients has developed autoimmune phenomena as a consequence of breaks in peripheral or central tolerance.

ROLE OF THE THYMUS IN IMMUNE RECONSTITUTION IN BONE MARROW TRANSPLANTATION Bone Marrow Transplantation in Children Studies of modes of reconstitution in children in the setting of BMT have been hampered both by lack of assays and measurements for evaluation of thymopoiesis versus peripheral T cell expansion and by the need to use chemotherapy to prepare for BMT. Table 4 is a list of factors known to damage the thymus or prevent immune reconstitution in adults as well as children (77–79). Early attempts to look at the thymus of children who had undergone BMT revealed no thymopoiesis after BMT with the majority of mature activated lymphocytes present in the thymic perivascular space (77). Of four children studied (ages 7–12), 3 had clinical graft-versus-host disease and one had disseminated candidiasis (77). In a later series, 4 of 36 patients studied at autopsy had histopathologic evidence of thymopoiesis after BMT (79). Of the 36 studied, 16 were aged 16 or less, and 2 of these 16 pediatric thymuses had active thymopoiesis (79). The other two patients in this series with active thymopoiesis were 18 and 21 years of age (79).

TABLE 4 Clinical events and diseases that decrease thymic function and prevent immune reconstitution Thymic atrophy due to aging Cancer chemotherapy Thymus irradiation Graft-versus-host disease Cyclosporine therapy Serious systemic illness or physical conditions that lead to systemic stress responses



543

These studies were important in that they showed that thymopoiesis was involved in T cell regeneration in some patients treated with T cell depleted BMT. Miniero et al reported rebound thymic hyperplasia following high dose chemotherapy and allogeneic BMT in a 3-year-old child, with normal thymic histology on thymic biopsy (80). Mackall et al have demonstrated that thymic rebound likely indicates ongoing thymopoiesis (14). Thymic rebound occurred in 6 of 7 children under age 16 treated with cancer chemotherapy. In their series of 15 subjects aged 1–2 years, the CD4` CD45RA` T cell were at highest levels 6 months after chemotherapy in those with thymic rebound shadows on computerized tomography (14). In 1987, Hong et al transplanted allogeneic BM mismatched at one MHC class I allele into a 2-year-old patient with SCID (81). No chemotherapy conditioning was given, and the donor marrow was treated with an anti-T cell monoclonal antibody plus complement prior to BMT. The BM graft reconstituted the child, and biopsy of the thymus showed normal morphology (81). In 1990, Incefy et al reported a correlation of engraftment in BMT for SCID with the appearance of thymulin levels in plasma following transplant, suggesting ongoing thymopoiesis (82). Recently, Buckley et al have summarized the long-term results over 16 years of BMT in 89 infants with SCID (83). Of the 89 infants, 72 were still alive 3 months to 16.5 years after transplantation, including 12 of 12 who received HLA identical BMT and 60 of the 77 (78%) given T cell–depleted, haploidenticalrelated BMT (83). Patel et al have studied pre- and post-cryopreserved peripheral blood samples from this patient cohort and demonstrated that of the children studied that successfully engrafted, all had dramatic rises in PB T cell sjTREC levels (84). Thus, taken together, current data show in children with SCID that engraft with stem cell BMT, there is a high likelihood of contribution to T cell regeneration from the thymus.

Bone Marrow Transplantation in Adults Mackall et al’s data in chemotherapy-treated patients predicted that only younger adult BMT-treated patients would have significant T cell regeneration via thymopoiesis, while most adults will have T cell regeneration via passenger donor mature postthymic T cells (14). Miller & Stutman demonstrated by limiting dilution analysis that mature T cells can expand 10,000-fold to repopulate T celldepleted mice (85). LeBlond et al studied 33 patients with allogeneic (17 patients) or autologous (10 patients) BMT and showed high levels of regenerating CD4` CD45` CD7 memory T cells after BMT, suggesting peripheral T cell regeneration (86). ln adult mice, Mackall et al showed the complexity of BMT reconstitution in adult animals and identified contributions to T cell regeneration from BM-derived thymic dependent progeny, from peripheral T cell progenitor-derived thymic independent progeny, and from host-derived T cells (13).


544

HAYNES ET AL

In a preliminary set of experiments, we have tested pre- and postperipheral blood samples of partially matched allogeneic cord blood transplants in adults who reconstituted, and we found only 3 of 16 who had minimal sjTREC rises posttransplant—consistent with peripheral T cell expansion in most adults studied (87). Whether lack of detection of contribution of the thymus in adults is due to chemotherapy or graft-versus-host disease damage to a thymus already comprised due to age remains to be determined. These results notwithstanding, it is expected that BMT in at least younger adults will result in some combination of thymopoiesis and peripheral T cell expansion contributing to T cell regeneration. Storek et al noted that BMT in adults predominately resulted in CD4` CD45RO` T cell with low levels of CD4` CD45RA` T cells 1 year after BMT (88). They concluded that adult BMT recipients were highly unlikely to recapitulate T cell development and rather were more likely to have been reconstituted with expanded peripheral T cells (88). The origin of the late arriving CD4` CD45RA` ‘‘naive’’ phenotype peripheral blood T cells in adult BMT has been debated; it can come from a trickle of naive T cells from an atrophic but still functioning adult thymus or it could be CD45RA` ‘‘revertants’’ of CD45RO` T cells that had stopped dividing after the initial round of T cell regeneration post-BMT (89). Only functional studies of these late CD45RA` T cells for memory T cell responses as well as sjTREC analysis will directly answer this question. Finally, an important experiment was performed by Heitgen et al who gave a BMT to a 15-year-old cancer patient who had been completely thymectomized (90). The patient was prepared for BMT with total body irradiation and cyclosporine A. Interestingly, all CD4` T cells that regenerated were CD45RO`, but the level of CD8` CD45RA` T cells was normal (90). Recent studies have shown that CD8`, CD45RA`, CD62L1 T cells are memory CTL effector cells and are not antigen naive (91, 92). Thus, BMT in children and young adults has shown the dependence of the thymus for naive T cell generation, and it has also shown that the young thymus in the setting of BMT can reconstitute the peripheral T cell pool. That the adult thymus does function suggests that when factors such as irradiation, intense chemotherapy, and graft-versus-host disease are not present to suppress thymopoiesis, then the adult postnatal thymus will also be able to contribute to regeneration of the T cell pool in adults after BMT.

ROLE OF THE THYMUS IN IMMUNE RECONSTITUTION IN HIV INFECTION AND AIDS Morphology of the Thymus in Early and Late HIV-1 Infection To understand the effect of HIV-1 infection on thymus, it is important to compare the thymus in early and late stages of HIV-1 infection with age-matched controls (Figures 8A and 8B) (30, 31). In early HIV infection, the lymphoid infiltrate of



545

Figure 8 Schematic representations of morphology of the human thymus in early (Panel A) and late (Panel B) stages of HIV-1 infection (adapted and used with permission from Ref. 31).

the thymic perivascular space is increased compared to normal thymus, with perivascular space germinal centers (reviewed in 30, 31, 93, 94). These changes are very similar to those seen in the hyperplastic thymus in myasthenia gravis (27). MECA-l` high endothelial venules (HEV) that are typical for peripheral lymph node are induced in the thymic perivascular space of both myasthenia gravis and HIV-1 infected thymuses (27, 31) (Figure 8A). In late HIV infection, the HIV-1 infected thymus is prematurely atrophic with changes that are similar to, but more exaggerated than, normal atrophic thymus (30, 31, 95). As shown


546

HAYNES ET AL

schematically in Figure 8B, the thymus of a 30-year-old subject with late stage AIDS can appear similar to the normal atrophic thymus of a 60-year-old subject with the exception that in HIV-1 infection, Hassall’s bodies (swirls of medullary epithelial cells) are more frequently calcified (indicating thymic epithelial cell death), and thymic epithelial islands devoid of lymphocytes are frequently surrounded by an intense lymphoid infiltrate of CD3`, CD8` T cells, monocytes, and B cells (30, 31, 93). However, even in patients with profound lymphopenia in late stage AIDS, small islands of active thymopoiesis can be found (30) (Figure 9, see color insert). In the perivascular space, CD8` T cells are present that express the cytotoxic T lymphocyte (CTL) granule antigen TIA-1, suggesting mature CTL effector cell function similar to the CD8` T cell infiltrate seen in HIV-1` peripheral lymph node (30). In early and late HIV-1 infection, HIV-1 expressing cells are found both in the thymic perivascular space and within the true thymic epithelial space (30) (Figure 9). Thus, a hallmark of HIV infection of the thymus is the induction of premature thymic atrophy. The morphologic changes induced in the thymus by HIV-1, namely increased perivascular space infiltrates of CD3` CD8` T cells, calcified Hassall’s bodies, and condensation of thymic epithelium with large areas devoid of thymopoiesis, are either identical to or are exaggerated versions of, changes seen in older normal or myasthenia gravis thymuses.

T Cell Dynamics and Thymus Function in HIV-1 Infection In simian immunodeficiency virus (SIV) infection, there is an early stage of suppressed thymopoiesis, reflected in an early drop in CD4` T cell count in primary infection, and an increased rate of thymocyte apoptosis reflected by a decrease in thymocyte numbers (96). Then, independent of viral load, a rebound in thymic function occurs with increased thymocyte production in SIV-infected monkeys beginning ;2 months after infection (96). Rebound of thymus function occurred before any evidence of peripheral lymphopenia, suggesting that the rebound is in response to thymic injury rather than in response to signals from T cell pool depletion (96). In HIV infection, this rebound phenomenon has been best documented using computerized tomogram scans in 20–40-year-old HIV-infected patients (97). McCune et al found abundant thymus tissue in 50% of HIV-1infected subjects, and the level of thymus tissue correlated with CD4`, CD45RA` naive T cell levels (97). To directly address the issue of thymic contribution to rises in peripheral T cell levels, Douek et al used the sjTREC assay to demonstrate that HAART induces rises in peripheral CD4` T cell counts and sjTREC levels in younger HIV-1-infected patients (25). They also demonstrated that sjTREC production was suppressed in the thymus of untreated HIV-infected patients with high viral loads, demonstrating a direct suppressive effect of HIV viremia on thymic sjTREC` cell production (26). HIV-1 may directly kill thymocytes (98–100), may kill dendritic cells required for normal thymocyte development (101), may damage thymic epithelial cells required for normal



547

thymopoiesis (98), or may inhibit thymocyte signaling. Up to 30% of recent thymic emigrants in the mouse model have recently divided just prior to emigration (9), providing a potential vulnerable point in the late stages of human T cell development for HIV-1 to infect activated newly generated naive T cells (102). Thus, in most patients who are . 60 years of age at the time of HIV infection, the thymus is functional to a degree based on age. If we assume that human thymic output is maximal at birth and comparable to the level of thymic export in mice (;2% of total thymocytes exported per day) (103), and that total thymocytes at birth are $ 50 2 109 (104), then thymic output per day in normal newborn subjects will be $ 109 new naive T cells. sJTRECs in age-matched HIV1` patients versus HIV-11 subjects were reduced by ;80% (25) (Table 5). Thus, although thymic output has not been directly measured in humans as yet, for the purposes of discussion we will assume 80% reduction in thymic output in HIV1 infection (Table 5). If we assume thymic output is proportional to the percentage of the thymus comprised of true thymic epithelial space, then thymic output can be expected to fall over time with age from ;109 naive T cells per day to ;1.8 2 108 naive T cells per day for . 50-year-old subjects (Table 5). For HIV-1infected patients with 80% reduction in thymic function, thymic output would range from ;2 2 108 naive T cells/day at birth to ;3.6 2 107 naive T cells per day in $ 50-year-old subjects. Figures 10 and 11 show schematic illustrations of the thymic and peripheral T cell pools in normal and HIV-infected subjects ;20 years of age. Two studies of the effect of combination antiretroviral therapy on peripheral T cell dynamics suggested very high levels of CD4` T cell turnover per day with correspondingly extraordinarily high levels of new T cells generated to replace these losses (105, 106). As Mackall Gress have clearly shown there are only two ways TCRab` T cells can be made in humans, via thymopoiesis and via peripheral expansion of preexisting mature postthymic peripheral T cells (reviewed in 107). Recent studies have shown that a major component of early rises of CD4` T cells in peripheral blood during highly active antiretroviral therapy (HAART) is due to decrease in the activation state of sequestered T cells in lymphoid tissues and redistribution of T cells to the peripheral circulation from tissues rather than from newly produced T cells from the thymus (108–110). This phenomenon was most clearly seen in the study of HAART-treated HIV-1` patients previously thymectomized for myasthenia gravis (30). In thymectornized subjects, the initial rises of CD4` T cells were comprised of rises in existing (redistributed) CD4` CD45RA` CD62L` (naive phenotype) T cells as well as CD4` CD45RO` (memory phenotype) T cells (Figure 12) (30). Moreover, sjTREC levels were very low at the time of CD4`, CD45RA`, CD62L` T cell rises, proving that these CD4` CD45RA`, CD62L` T cells were redistributed revertants from CD45RO` T cells since the patient in Figure 12 had been totally thymectomized 8 years prior to onset of HAART (30). After the initial CD4` T cell rise over the first 3 months, CD4` CD45RA` CD62L` T cell levels stabilized and all sub-

548


TABLE 5 Estimated thymic output of recent thymic emigrants in normal adults and HIV-infected patients Age quintile

0-1 year

2-10 year

11-25 year

26-49 year

$50 year

Percent thymic epithelial space#

93%

88%

63%

45%

18%

Estimated daily thymocyte output in normal subjects before HIV-1 infection

$109

8.8 2 108

6.3 2 108

4.5 2 108

1.8 2 108

Estimated daily thymocyte output in HIV-1 infected patients (80% reduction of normal output based on measured peripheral blood sjTREC levels in HIV-1` patients*)

2 2 108

1.8 2 108

1.3 2 108

9.0 2 107

3.6 2 107

*sjTREC levels in normal T cells peripheral blood CD4` and CD8` T cells and HIV-1` CD4` and CD8` T cells from Reference 25. #Thymic epithelial space percentages from References 33, 34.


549


Figure 10 Schematic representation of estimates of thymic and peripheral T cell pools in a normal 20-year-old human.

Figure 11 Schematic representation of estimates of thymic and peripheral T cell pools in an HIV- I infected 20year-old patient with AIDS.


550

HAYNES ET AL

Figure 12 Immune reconstitution after HAART in a thymectomized HIV-1` patient, aged 30. Figure shows rises in CD4` CD45RA`, CD62L` and CD4`, CD45RO` T cells were essentially sjTREC- per 105 cells, indicating both populations were derived from redistribution and/or expansion of existing peripheral T cells (adapted and used with permission from Reference 30).

sequent rises in CD4` T cell levels were sjTREC-, CD4`, CD45RO` T cells due to peripheral T cell expansion (Figure 12) (30). Using a novel nonradioactive deuterium-labeled glucose technique for assessing in vivo cellular turnover in humans, Hellerstein et al demonstrated that total T cell half-life in HIV-1 infection is shortened from ;82 days (87 days, CD4` and 77 days, CD8`) in uninfected controls to 23 days in HIV-1` patients (24 days, CD4` and 22 days, CD8`) (111). In HIV-1` patients, CD8` T cell production increased approximately fourfold over normals, while CD4` T cell production was unchanged (111). Interestingly, the T cell half-life decreased after 12 weeks of HAART, (14 days for both CD4` and CD8` T cells), with absolute production rates of CD4` and CD8` T cells increased (111). Taken together with the recent studies of others, these data demonstrated that the CD4` T cell lymphopenia in HIV-1 infection is due both to a shortened CD4` T cell survival time and to a failure to increase production of circulating T cells (reviewed in 112, 113). Given the new data that the human thymus is functioning to variable degrees in most adults who become infected with HIV-1, and that a potential production site of new CD4` naive T cells is the thymus, it is reasonable to speculate that at least a portion of the gradual loss of naive CD4` T cells over time in HIV infection in some patients could be due in part to loss of thymus production of new CD4` and CD8` T cells, with shrinkage in the CD4` and CD8` naive T



551

cell niches, and replacement of some of the volume of the peripheral T cell pool with CD8 . CD4 memory T cells (Figure 11). The total peripheral T cell pool is estimated to be reduced by ;50% in lymphopenic patients with AIDS (112). From the studies of Mackall & Gress, we know that peripheral expansion of postthymic T cells is neither efficient in reconstitution of the peripheral T cell pool, nor regenerates a complete T cell repertoire (55–57, 107). Thus, untreated adult HIV-1` patients frequently have oligoclonal T cell repertoires, which with prolonged HAART can become more polyclonal (presumably with new naive T cells from a functional thymus) (114, 115). Zhang et al have demonstrated normal thymus function in many HIV` patients, and HAART-induced rises in TCRA TRECs in a subset of HIV-infected patients with low pretreatment TRECs (116). They suggest that in spite of documented thymus function in adults, the quantity of new T cells post-HAART is not sufficient to totally explain the post-HAART increase in peripheral T cells (116). Walker et al used CD4` T cells transfected with the neomycin phosphotransferase gene and adoptively transferred into HIV` twins to demonstrate that the CD4` T cell pool in adults is maintained primarily by the proliferation of mature CD4` T cells rather than by thymopoiesis (117). Thus, it has been conclusively shown that the postnatal thymus is functional to some degree in many adults with HIV infection. However, the key question remains as to how much of a quantitative contribution the thymus makes to the peripheral T cell pool in HIV infection, both before and after HAART. That HAART doesn’t usually restore the T cell level to normal might be explained by the fundings of Rocha and colleagues (58) that naive and memory T cells may occupy separate niches in the periphery. Thus, only when thymus function is fully restored by prolonged HAART (and thymus naive T cell export is restored), can we expect that normal T cell levels will be restored. In a protocol studying thymic function of HIV-1` patients on long-term HAART, we are now indeed starting to see normal peripheral blood CD4` T cell levels return in select patients whose PBT cells are sjTREC` (ML Markert, A Alverez-McLeod, GD Sempowski, LP Hale, JA Bartlett, BF Haynes, unpublished). Gobbi et al have shown that human herpesvirus-6 (HHV-6) infection induces thymocyte depletion in SCID/Hu Thy/Liv mice (118). However, in an analysis of 8 HIV-infected human thymuses, no expression of HHV-6 was found (A Gobbi, BF Haynes, L Hale, N Abbey, B Herndier, JMcCune, unpublished observations). Thus, at present, data are inconclusive regarding the contribution of HHV-6 to thymus damage in HIV-1 infection. Finally, it is interesting to compare aspects of immune reconstitution in thymectomized HIV` subjects after HAART therapy (30) with that of the thymectomized patient given BMT (90). After BMT of the thymectomized patient, CD4` CD45RA` PB T cells were absent, while after HAART treatment of the thymectomized HIV-1` patients, PB CD4` CD45RA` T cells were present. Importantly, prior to BMT the thymectomized patient received conditioning irradiation, and after BMT received cyclosporine A. Thus, any existing CD4` naive T cells were likely reduced prior to BMT, resulting in no peripheral CD4`, CD45RA`

552

HAYNES ET AL

T cells to proliferate after BMT. In contrast, our HAART-treated thymectomized patient had no irradiation, and the CD4`, CD45RA` T cells that increased after HAART were sjTREC- and therefore were postthymic peripheral T cell pool in origin; initially they rose in the PB due to redistribution from tissue sites.


PROSPECTS FOR RECONSTITUTION OF THE POSTNATAL HUMAN T CELL POOL The prospects are excellent for achieving immune reconstitution of the postnatal T cell pool in a variety of settings. In particular, development of new strategies such as CTLA-4-Ig administration (119) to prevent graft-versus-host disease in the setting of BMT should facilitate regeneration via thymopoiesis. In adults with no thymus function, the ability to use strategies such as CTLA-4-Ig or CD40L (120) administration may facilitate allogeneic thymus transplantation in AIDS, allowing the formation of chimeric thymic grafts similar to those in DiGeorge syndrome (3, 66). The extraordinary proliferative and regenerative capacity of the peripheral T cell pool complements and augments thymus output and, though suboptimal, is a fail-safe mechanism for T cell regeneration if the thymus is profoundly atrophic or irreversibly damaged. One key issue noted in this review is that to date no firm feedback loops have been defined between events in the peripheral T cell pool feeding back on the thymus to increase thymic output (59, 60). Rather, numerous events/processes/ drugs affect the thymus to reduce thymopoiesis (Table 4), and after removal of the event/process/drug, the thymus regenerates (‘‘rebounds’’) (14). After initial near-total thymectomy for corrective cardiovascular surgery as newborns, children at ages 1–2 who are reoperated have substantial thymic regrowth (R Ungerleider, personal communication). Thus, it is key to understand the cytokines and processes that drive thymus regeneration in order to be able to drive thymopoiesis. An important concern regarding stimulating thymopoiesis in adults is that stimulation of adult thymopoiesis will trigger defective negative selection processes in the thymus and precipitate autoimmune diseases (54). Mathis has clearly shown in a mouse model that inflammatory synovitis can be caused by a postnatal (acquired) loss of (central) T cell tolerance in the thymus (121). In this regard, it is of great interest that reports are now emerging of the onset of fulminant autoimmune thyroid disease (Graves disease) in the setting of T cell reconstitution after HAART (122; ML Markert, A Alverez-McLeod, GD Sempowski, LP Hale, JA Bartlett, BF Haynes, unpublished). Current strategies targeted toward expanding the peripheral T cell pool in HIV-1 such as IL-2 therapy (123) to date have not been associated with the development of autoimmune syndromes. The drawback of only expanding the peripheral T cell pool in HAART-treated AIDS patients is that the pool may be limited and may not reach normal levels long-


553

term after treatment is stopped. Thus, for TCR repertoire versatility, developing strategies to regenerate both the naive and memory T cell pools is desirable in clinical conditions requiring T cell reconstitution. Further attempts to stimulate new thymopoiesis will require close attention to the risk of development of autoimmunity. Nonetheless, the past two years have brought exciting new insights toward understanding human postnatal thymus biology. These insights hold great promise for leading to a more complete understanding of the pathogenesis of T cell– mediated autoimmune syndromes and for safely reconstituting the T cell arm of the immune system in congenital and acquired immunodeficiency syndromes.


ACKNOWLEDGMENTS The authors thank Richard Koup, David Ho, Max Cooper, Ashley Haase, and Mike McCune for preprints of papers, Ashley Haase and Kristine GebhardMitchell for performing in situ hybridization for Figure 9B, Daniel Douek and Richard Koup for performing sjTREC assays in Figure 12, and Kim McClammy for expert secretarial assistance. In particular, we are grateful to Dr. Rebecca Buckley for discussions, mentoring, leadership, and inspiration for many years. Supported by NIH grants CA28936, A138550, MOI-RR-00035, AI47604, A138387, CA09058, and A107217. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Miller JFAP. 1962. Immunological significance of the thymus of the adult mouse. Nature 195:1318–19 2. Miller JFAP. 1965. Effect of thymectomy in adult mice on immunological responsiveness. Nature 208:1337–38 3. Markert LM, Boeck A, Hale LP, Kloster AL, McLaughlin TM, Batchvarova MN, Douek DC, Koup RA, Kostyu DD, Ward FE, Rice HE, Mahaffey SM, Buckley RH, Haynes BF. 1999. Thymus transplantation in complete DiGeorge syndrome. New Engl. J. Med. 341:1180–89 4. Haynes BF, Heinly CS. 1995. Early human T cell development: analysis of the human thymus at the time of initial entry of hematopoietic stem cells into the fetal thymic microenvironment. J. Exp. Med. 181:1445–58

5. Kong F-K, Chen CH, Cooper MD. 1998. Thymic function can be accurately monitored by the level of recent T cell emigrants in the circulation. Immunity 8:97– 104 6. Kong F-K, Chen CH, Six A, Hockett RD, Cooper MD. 1999. T cell receptor gene deletion circles identify recent thymic emigrants in the peripheral T cell pool. Proc. Natl Acad. Sci. USA 96:1536–40 7. Hosseinzadeh H, Goldschneider I. 1993. Recent thymic emigrants in the rat express a unique antigen phenotype and undergo post-thymic maturation in peripheral lymphoid tissue. J. Immunol. 150:1670–79 8. Scollay R, Wilson A, Shortman K. 1984. Thymus cell migration: analysis of thymus emigrants with markers which dis-

554

9.

10.


11.

12.

13.

14.

15.

16.

17.

HAYNES ET AL tinguish medullary thymocytes from peripheral T cells. J. Immunol. 132:1089–94 Penit C, Vasseur F. 1997. Expansion of mature thymocyte subsets before emigration to the periphery. J. Immunol. 59:4848–56 Pilarski LM, Gillitzer R, Zola H, Shortman K, Scollay R. 1989. Definition of the thymic generative lineage by selective expression of high molecular weight isoforms of CD45 (T200). Eur. J. Immunol. 19:589–97 Fujii Y, Okumura M, Inada K, Nakahara K, Matsuda H. 1992. CD45 isoform expression during T cell development in the thymus. Eur. J. Immunol. 22:1843– 50 Erkeller-Yuksel FM, Deneys V, Yuksel B, Hannet 1, Hulstaert F, Hamilton C, Mackinnon H, Stokes LT, Munhyeshuli V, Vanlangendonck F. 1992. Age related changes in human blood lymphocyte subpopulations. J. Pediatr. 120:216–22 Mackall CL, Granger L, Sheard MA, Cepeda R, Gress RE. 1993. T-cell regeneration after bone marrow transplantation: differential CD45 isoform expression on thymic-derived versus thymic-depleted progeny. Blood 82:2585– 94 Mackall CL, Fleisher TA, Brown MR, Andrich MP, Chen CC, Feuerstein IM, Horowitz ME, Magrath IT, Shad AT, Steinberg SM, Wexler LH, Gress RE. 1995. Age, thymopoiesis, and CD4` Tlymphocyte regeneration after intensive chemotherapy. N. Engl. J. Med. 332: 143–49 Picker LJ, Treer JR, Ferguson-Darnell B, Collin PA, Buck D, Terstappen LWMM. 1993. Control of lymphocyte recirculation in man. J. Immunol. 150:1105–21 Bell EB, Sparshott SM. 1991. Interconversion of CD45R subsets of CD4 T cells in vivo. Nature 348:163–66 Hargreaves M, Bell EB. 1997. Identical expression of CD45R isoforms by

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

CD45RC` ‘‘revertant’’ memory and CD45RC` naive CD4 T cells. Immunology 91:323–30 Sparshott SM, Bell EB Sarawar SR. 1991. CD45R CD4 T cell subset reconstituted nude rats: subset-dependent survival of recipients and bi-directional isoforms switching. Eur. J. Immunol. 21:993–1000 Richards D, Chapman MD, Sasama J, Lee TH, Kemeny DM. 1997. Immune memory in CD4` CD45R` T cells. Immunology 91:331–39 Michie CA, McLean A, Alcock C, Beverley PCL. 1992. Lifespan of human lymphocyte subsets defined by CD45 isoforms. Nature 360:264–65 Tough DF, Sprent J. 1995. Life span of naive and memory T cells. Stem Cells 13:242–49 Weng N-P, Levine BL, June CH, Hodes RJ. 1995. Human naive and memory T lymphocytes differ in telomeric length and replicative potential. Proc. Natl. Acad. Sci. USA 92:1109–94 Aspinall R, Carroll J, Jiang S. 1998. Agerelated changes in the absolute number of CD95 positive cells in T cell subsets in the blood. Exp. Gerontol. 33:581–91 Livak E, Schatz D. 1996. T-cell receptor a locus V(D) recombination by products are abundant in thymocyte and mature T cells. Mol. Cell. Biol. 16:609–18 Douek DC, McFarland RD, Keiser PH, Gage EA, Massey JM, Haynes BF, Polis MA, Haase AT, Feinberg MB, Sullivan JL, Jamieson BD, Zack JA, Picker LJ, Koup RA. 1998. Changes in thymic function with age and during the treatment of HIV infection. Nature 396:690– 95 Steinmann GG, Klaus B, Muller-Hermelin HK. 1985. The involution of the aging human thymic epithelium is independent of puberty. A morphometric study. Scand. J. Immunol. 22:563–75 Bofill M, Janossy G, Willcox N, Chilosi M. 1985. Microenvironments in the nor-


28.

29.


30.

31.

32.

33.

34.

35.

36.

37.

mal thymus and the thymus in myasthenia gravis. Am. J. Pathol. 119:462–73 Steinmann GG. 1986. Changes in the human thymus during aging. Curr. Topics Pathol. 75:43–88 Tamaoki N, Habu S, Kameya T. 1997. Thymic lymphoid follicles in autoimmune diseases. 11. Histological, histochemical, and electron microscopic studies. Keio J. Med. 20:57–68 Haynes BF, Hale LP, Weinhold KJ, Patel DD, Liao H-X, Bressler PB, Jones DM, Demarest JF, Gebhard-Mitchell K, Haase AT, Bartlett JA. 1999. Analysis of the adult thymus in reconstitution of T lymphocytes in HIV-1 infection. J. Clin. Invest. 103:453–60 Haynes BF, Hale LP. 1998. The human thymus: a chimeric organ comprised of central and peripheral lymphoid components. Immunol. Res. 3:175–92 George AJT, Ritter MA. 1996. Thymic involution with ageing: obsolescence or good housekeeping? Immunol. Today 17:267–71 Flores KG, Li J, Sempowski GD, Haynes BF, Hale LP. 1999. Analysis of the human thymic perivascular space during aging. J. Clin. Invest. 104:1031–39 Sempowski GD, Hale LP, Sundy JS, Koup RA, Douek DC, Patel DP, Haynes, BF. 2000. Expression of thymic cytokines and human thymic atrophy. J. Immunol. 164: In press Smith SM, Ossa-Gomez LJ. 1981. A quantitative histologic comparison of the thymus in 100 healthy and diseased adults. Am. J. Clin. Pathol. 76:657–65 Bertho J-M, Demarquay C, Moulian N, Meeren A, Berrih-Aknin S, Gourmelon P. 1997. Phenotypic and immunohistological analysis of the human adult thymus: evidence for an active thymus during adult life. Cell. Immunol. 179:30– 40 Jamieson BD, Douek DC, Killian S, Hultin LE, Scripture-Adams DD, Giorgi JV, Marell D, Koup RA, Zack JA. 1999.

38.

39.

40.

41.

42.

43.

44.

45.

46.

555

Generation of functional thymocytes in the human adult. Immunity 10:569–75 Poulin J-F, Viswanathan MN, Komanduri KV, Wieder E, Ringuette N, Harris JM, Jenkins M, McCune JM, Sekaly RP. 1999. Direct evidence for thymic function in adult humans. J. Exp. Med. 190:479–86 Aspinall R. 1997. Age-associated thymic atrophy in the mouse is due to a deficiency affecting rearrangement of the TCR during intrathymic T cell development. J. Immunol. 158:3037–45 Hartwig M, Steinmann G. 1994. On a causal mechanism of chronic thymic involution in man. Mech. Ageing Dev. 75:151–56 Hirokawa K, Sato K, Makinodan T. 1982. Influence of age of thymic grafts on the differentiation of T cells in nude mice. Clin. Immunol. Immunopathol. 24:251–62 Leiner H, Greinert U, Scheiwe W, Bathmann R, Muller-Hermelink HK. 1984. Repopulation of lymph nodes and spleen in thymus chimeras after lethal irradiation and bone marrow transplantation: dependence on the age of the thymus. Immunobiology 167:345–58 Utsuyama M, Kasai M, Kurashima C, Hirokawa K. 1991. Age influence on the thymic capacity to promote differentiation of T cells: induction of different composition of T cell subsets by aging thymus. Mech. Ageing Dev. 58:267–77 Tyan ML. 1977. Age-related decrease in mouse T cell progenitors. J. Immunol. 118:846–51 Kadish JL, Basch RS. 1976. Hematopoietic thymocyte precursors. 1. Assay and kinetics of the appearance of progeny. J. Exp. Med. 143:1082 Shen MM, Skoda RC, Cardiff RD, Campos-Torres J, Leder P, Ornitz DM. 1994. Expression of LIF in transgenic mice results in altered thymic epithelium and apparent interconversion of thymic and

556

47.

48.


49.

50.

51.

52.

53.

54.

55.

HAYNES ET AL lymph node morphologies. EMBO J. 13:1375 Malik N, Haugen HS, Modrell B, Shoyab M, Clegg CH. 1995. Developmental abnormalities in mice transgenic for bovine oncostatin M. Mol. Cell. Biol. 15:2349 Metcalf D, Nicola NA, Gearing DP. 1990. Effects of injected leukemia inhibitory factor on hematopoietic and other tissues in mice. Blood 76:50–56 Kapur R, Everett ET, Uffman J, McAndrews-Hill M, Cooper R, Ryder J, Vik T, Williams DA. 1997. Overexpression of human stem cell factor impairs melanocyte, mast cell, and thymocyte development: a role for receptor tyrosine kinase-mediated mitogen activated protein kinease activation in cell differentiation. Blood 90:3018–26 Kendall MD, Fitzpatrick FT, Greenstein BD, Khoylou F, Safieh B, Hamblin A. 1990. Reversal of ageing changes in the thymus of rats by chemical or surgical castration. Cell Tissue Res. 261:555–64 Greenstein BD, Fitzpatrick FT, Adcock IM, Kendall MD, Wheeler MJ. 1986. Reappearance of the thymus in old rate after orchidectomy: inhibition of regeneration by testosterone. J. Endocrinol. 110:417–22 Wylie AH. 1980. Glucocorticoidinduced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature 284:555–56 Akita S, Malkin J, Melmed S. 1996. Disrupted murine leukemia inhibitory factor (LIF) gene attenuates adrenocorticotropic hormone (ACTH) secretion. Endocrinology 137:3140–43 Aronson M. 1991. Hypothesis: involution of the thymus with aging—programmed and beneficial.Thymus 18:7–13 Mackall CL, Fleisher TA, Brown MR, Andrich MP, Chen CC, Feuerstein IM, Magrath IT, Wexler LH, Dimitrov DS, Gress RE. 1997. Distinctions between CD8` and CD4` T-cell regenerative

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

pathways result in prolonged T-cell subset imbalance after intensive chemotherapy. Blood 89:3700–7 Mackall CL, Bare CV, Granger LA, Sharrow SO, Titus JA, Gress RE. 1996. Thymic-independent T cell regeneration ocurs via antigen-driven expansion of peripheral T cells resulting in a repertoire that is limited in diversity and prone to skewing. J. Immunol. 156:4609–16 Mackall CL, Hakim FT, Gress RE. 1997. T-cell regeneration: All repertoires are not created equal. Immunol. Today 18:245–51 Tanchot C, Rocha B. 1998. The organization of mature T-cell pools. Immunol. Today 19:575–79 Gabor MJ, Scollay R, Godfrey DI. 1997. Thymic T cell export is not influenced by the peripheral T cell pool. Eur. J. Immunol. 27:2986–93 Ramanathan S, Norwich K, Poussier P. 1998. Antigen activation rescues recent thymic emigrants from programmed cell death in the BB rat. J. Immunol. 160:5757–64 Tanchot C, Rocha B. 1995. The peripheral T cell repertoire: independent homeostatic regulation of virgin and activated CD8` T cell pools. Eur. J. Immunol. 25:2127–36 DeKoning J, DiMolfetto L, Reilly C, Wei Q, Havran WL, Lo D. 1997. Thymic cortical epithelium is sufficient for the development of mature T cells in re1B deficient mice. J. Immunol. 158:2558–66 Laufer TM, DeKoning J, Markowitz JS, Lo D, Glimcher LH. 1996. Unopposed positive selection and autoreactivity in mice expressing class 11 MHC only on thymic cortex. Nature 383:81–85 McCune JM, Namikawa R, Kaneshima H, Shultz LD, Lieberman M, Weissman IL. 1988. The SCID-hu mouse: murine model for the analysis of human hematolymphoid differentiation and function. Science 241:1632–39 Barry TS, Jones DM, Richter CB,


66.


67.

68.

69.

70.

71.

72.

73.

74.

Haynes BF. 1991. Successful engraftment of human postnatal thymus in severe combined immune deficient (SCID) mice: differential engraftment of thymic components with irradiation versus anti-asialo GM-1 immunosuppressive regimens. J. Exp. Med. 173:167–80 Markert ML, Kostyu DD, McLaughlin TM, Watson TJ, Ward FE, Buckley RH et al. 1997. Successful formation of a chimeric human thymus allograft following transplantation of cultured postnatal human thymus. J. Immunol. 158:998– 1005 Hong R, Moore AL. 1996. Organ culture for thymus transplantation. Transplantation 61:444–48 Markert ML, Watson TJ, Kaplan 1, Hale LP, Haynes BF. 1997. The human thymic microenvironment during organ culture. Clin. Immunol. Immunopathol. 82:26 Goldsobel AB, Haas A, Stiehm ER. 1987. Bone marrow transplantation in DiGeorge syndrome. J. Pediatr. 111:40– 44 Borzy NIS, Ridgway D, Noya, FJ, Shearer WT. 1989. Successful bone marrow transplantation with split lymphoid chimerism in DiGeorge syndrome. J. Clin. Immunol. 9:386–92 Bowers DC, Lederman HM, Sicherer SH, Winkeistein JA, Chen AR. 1998. Immune reconstitution of complete DiGeorge anomaly by transplantation of unmobilized blood mononuclear cells. Lancet 352: 1983–84 August CS, Berkel AL, Levey RH, Rosen FS. 1970. Establishment of immunological competence in a child with congenital thymic aplasia by a graft of fetal thymus. Lancet 1:1080–83 Cleveland WW, Fogel BJ, Brown WT, Kay HEM. 1968. Foetal thymic transplant in a case of DiGeorge’s syndrome. Lancet 2:1211–14 Jose DG, Barnes G, Rossiter EJ, Myers NA, Fitzgerald MG. 1974. Reconstitution of cellular immune function in a

75.

76.

77.

78.

79.

80.

81.

557

child with thymic aplasia by foetal thymus grafting. Aust. NZ J. Med. 4:267–73 Markert LM, Hick CB, Bartlett JA, Harmon JL, Hale LP, Greenberg ML, Ferrari G, Ottinger J, Boeck A, Kloster AL, McLaughlin TM, Bleich KB, Ungerleider RM, Lyerly HK, Wilkerson WE, Rousseau FS, Heath-Chiozzi ME, Leonard JM, Haase AT, Shaw GM, Bucy RP, Bolognesi DP, Weinhold KJ. 2000. Effect of HAART and thymic transplantation on immunoreconstitution in HIV infection. Aids Res. Hum. Retrovirus. In press Bacchetta R, Parkman R, McMahon M, Weinberg K, Bigler M, de Vries JE, Roncarolo M-G. 1995. Dysfunctional cytokine production by host-reactive T cell clones isolated from a chimeric severe combined immunodeficiency patient transplanted with haploidentical bone marrow. Blood 85:1944–52 Thomas JA, Sloane JP, Imrie SF, Ritter MA, Schuurman H-J, Huber J. 1986. Immunohistology of the thymus in bone marrow transplant recipients. Am. J. Pathol. 122:531–40 Muller-Hermelink HK, Gulden M, Bathmann R. 1984. Restitution of the thymus in lethally irradiated mice after transplantation of syngeneic or allogeneic bone marrow. Immunobiology 167:462–82 Muller-Hermelink HK, Sale GE, Borisch B, Storb R. 1987. Pathology of the thymus after allogeneic bone marrow transplantation in man. Am. J. Pathol. 129:242–56 Miniero R, Busca A, Leonardo E, Mossetti C, Machado D, Vassallo E, Madon E. 1993. Rebound thymic hyperplasia following high dose chemotherapy and allogeneic BMT. Bone Marrow Transplant. 11:67–70 Hong R, Horowitz S, Moen R, Trigg M, Sondel P, Billing R, Borcherding W, Johnson JP, Kanoff M, Bias WB, Winkelstein J. 1987. Thymus and B cell reconstitution in severe combined immu-

558

82.


83.

84.

85.

86.

87.

88.

HAYNES ET AL nodeficiency after transplantation of monoclonal antibody depleted parental mismatched bone marrow. Bone Marrow Transplant. 1:405–9 Incefy GS, Flomenberg N, Heller G, Kernan NA, Brochstein J, Kirkpatric D, Kapor N, Groshen S, O’Reilly RJ. 1990. Evidence that appearance of thymulin in plasma follows lymphoid chimerism and precedes development of immunity in patients with lethal combined immunodeficiency transplanted with T celldepleted haploidentical marrow. Transplantation 50:55–61 Buckley RH, Schiff SE, Schiff RI, Markert ML, Williams LW, Roberts JL, Myers LA, Ward FE. 1999. Hematopoietic stem-cell transplantation for the treatment of severe combined immunodeficiency. N. Engl. J. Med. 340:508–61 Patel DD, Gooding M, Parrott, R, Curtis, K, Haynes BF, Buckley RH. 2000. Contribution of the thymus to T cell regeneration following treatment of severe combined immune deficiency diseases with T cell depleted bone marrow transplantation. Submitted Miller RA, Stutman 0. 1984. T cell repopulation from functionally restricted splenic progenitors: 10,000-fold expansion documented by using limiting dilution analyses. J. Immunol. 133:2925–32 Leblond V, Othman TB, Blanc C, Theodorou 1, Choquet S, Sutton L, Debre P, Autran B. 1997. Expansion of CD4` CD71 T cells, a memory subset with preferential interleukin-4 production, after bone marrow transplantation. Transplantation 64:1453–59 Chao NJ, Chen BJ, Cui X, Kurtzberg J, Haynes BF, Patel DD. 2000. Measurement of thymic output using T cell receptor excision circles (TRECs) Following allogeneic hematopoietic stem cell transplantation: adult vs. pediatric patients. Submitted Storek J, Witherspoon RP, Storb R. 1995. T cell reconstitution after bone marrow

89.

90.

91.

92.

93.

94.

95.

transplantation into adult patients does not resemble T cell development in early life. Bone Marrow Transplant. 16:413– 25 Michie CA, McLean AR. 1993. Bone marrow transplantation and lymphocyte subset kinetics. Bone Marrow Transplant. 12:547 Heitger A, Neu Nk, Kern H, Panzer-Grumayer ER, Greinix H, Nachbaur D, Niederwieser D, Fink FM. 1997. Essential role of the thymus to reconstitute naive (CD45RA`) T-helper cells after human allogeneic bone marrow transplantation. Blood 90:850–57 Hamann D, Baars PA, Rep MHG, Hooibrink B, Kerkhof-Garde SR, Klein M, van Lier RAW. 1997. Phenotypic and functional separation of memory and effector human CD8` T cells. J. Exp. Med. 186:1407–18 Roederer M, Dubs JG, Anderson MT, Raju PA, Herzenberg LA, Herzenberg LA. CD8 naive T cell counts decrease progressively in HIV-infected adults. J. Clin. Invest. 95:2061–66 Burke AP, Anderson D, Benson W, Turnicky R, Mannan P, Liang Y-H, Smialek J, Virmani R. 1995. Localization of human immunodeficinecy virus 1 RNA in thymic tissues from asymptomatic drug addicts. Arch. Pathol. Lab. Med. 119:3641 Prevot S, Audouin J, Andre-Bougaran J, Griffais R, Tourneau AL, Fournier JG, Diebold J. 1992. Thymic pseudoturnorous enlargement due to follicular hyperplasia in a human immunodeficiency virus seropositive patient. Am. J. Clin. Pathol. 97:420–25 Schuurman HJ, Krone WJ, Broekhuizen R, van Baarlen J, van Veen P, Golstein AL, Huber J, Goudsmit J. 1989. The thymus in acquired immunodeficiency syndrome: comparison with other types of immunodeficiency diseases, and presence of components of human immuno-


96.

97.


98.

99.

100.

101.

102.

deficiency virus type 1. Am. J. Pathol. 134:1329–38 Wykrzykowska JJ, Rosenzweig M, Veazey RS, Simon MA, Halvorsen K, Desrosiers RC, Johnson RP, Lackner AA. 1998. Early regeneration of thymic progenitors in rhesus macaques infected with simian immunodeficiency virus. J. Exp. Med. 187:1767–78 McCune JM, Loftus R, Schmidt DK, Carroll P, Webster D, Swor-Yim LB, Francis IR, Gross BH, Grant RM. 1998. High prevalence of thymic tissue in adults with human immunodeficiency virus-1 infection. J. Clin. Invest. 101:2301–08 Stanley SK, McCune JM, Kaneshima H, Justement JS, Sullivan M, Boone E., Baseler M, Adelsberg J, Bonyhadi M, Orenstein J, Fox CH, Fauci AS. 1993. Human immunodeficiency virus infection of the human thymus and disruption of the thymic microenvironment in the SCID-hu mouse. J. Exp. Med. 178:1151– 63 Bonyhadi ML, Rabin L, Salimi S, Brown DA, Kosek J, McCune JM, Kaneshima H. 1993. HIV induces thymus depletion in vivo. Nature 363:728–32 Kourtis AP, Ibegbu C, Nahmias AJ, Lee FK, Clark WS, Sawyer MK, Nesheirn S. 1996. Early progression of disease in HIV-infected infants with thymus dysfunction. N. Engl. J. Med. 335:1431–36 Valentin H, Nugeyre M-T, Vuillier F, Bournsell L, Schmid M, Barre-Sinoussi F, Pereira RA. 1994. Two subpopulations of human triple-negative thymic cells are susceptible to infection by human immunodeficiency virus type 1 in vitro. J. Virol. 68:3041–50 Schnittman SM, Denning SM, Greenhouse JJ, Justement JS, Baseler M, Kurtzberg J, Haynes BF, Fauci AS. 1990. Evidence for susceptibility of intrathymic T-cell precursors and their progeny carrying T-cell antigen receptor phenotypes TCRap` and TCRy6` to

103.

104.

105.

106.

107.

108.

109.

110.

559

human immunodeficiency virus infection: a mechanism for CD4` (T4) lymphocyte depletion. Proc. Natl. Acad. Sci. USA 87:7727–73 Shortman K, Egerton M, Spangrude GJ, Scollay R. 1990. The generation and fate of thymocytes. Semin. Immunol. 2:3–12 Westermnn J, Pabst R. 1992. Distribution of lymphocyte subsets and natural killer cells in the human. Clin. lnvest. 70:539– 544 Wei X, Ghosh SK, Taylor ME, Johnson VA, Emini EA, Deutsch P, Lifson JD, Bonhoeffer S, Nowak MA, Hahn BH. 1995. Viral dynamics of human immunodeficiency virus type 1 infection. Nature 373:117–22 Pakker NG, Notermans DW, de Boer RJ, Roos MT, de Wolf F, Hill A, Leonard JM, Danner SA, Miedema F, Schellekens PT. 1998. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: a composite of redistribution and proliferation. Nat. Med. 4:208–14 Mackall CL, Hakim FT, Gress RE. 1997. Restoration of T-cell homestasis after Tcell depletion. Semin. Immunol. 9:339– 46 Pakker NG, Notermans DW, DeBoer RJ, Roos MTL, de Wolf F, Hill A, Leonard JM, Danner SA, Miedema F, Schellekens PTA. 1998. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: A composite of redistribution and proliferation. Nat. Med. 4:208–14 Wolthers KC, Wisman GBA, Otto SA, de Roda Husman A-M, Schaft N, de Wolf F, Goudsmit J, Coutinho RA, van der Zee AGJ, Meyaard L, Miedema F. 1996. T cell telomere length in HIV-1 infection: No evidence for increased CD4` T cell turnover. Science 254:1543–47 Bucy RP, Hockett DR, Derdeyn CA, Saag MS, Squires K, Sillers M, Mitsuyasu RT, Kilby JM. 1999. Initial increase in blood CD4` lymphocytes after HIV antiretroviral therapy reflects redistribu-

560

111.

112.


113. 114.

115.

116.

117.

HAYNES ET AL tion from lymphoid tissues. J. Clin. Invest. 103:1391–98 Hellerstein M, Hanley MB, Cesar D, Siler S, Papageorgopoulos C, Wieder Schmidt D, Hoh R, Neese R, Macallan D, Deeks S, McCune JM. 1999. Directly measured kinetics of circulating T lymphocytes in normal and HIV- I infected humans. Nat. Med. 5:83–89 Haase AT. 1999. Population biology of HIV- I infection: Viral and CD4` T cell demographics and dynamics in lymphatic tissues. Annu. Rev. Immunol. 17:625–56 Horton R. 1999. HIV infection: Where have all the T cells gone? Lancet 354:5 Gorochov G, Neumann AU, Kereveur A, Parizot C, Li T, Katlama C, Karmochkine M, Raguin G, Autran B, Debre P. 1998. Perturbation of CD4` and CD8` T cell repertories during progression to AIDS and regulation of the CD4` repertoire during antiviral therapy. Nat. Med. 4:215–21 Autran B, Carcelain G, Li TS, Blanc C, Mathez D, Tubiana R, Katlama C, Leibowitch DJ. 1997. Positive effects of combined antiretroviral therapy on CD4` T cell homestasis and function in advanced HIV disease. Science 277:112– 16 Zhang L, Lewin S, Markowitz M, Lin HH, Skulsky E, Karanicolas R, He Y, Jin X, Tuttleton S, Vesanen M, Spiegel H, Kost R, Wolinsky S, Borkowsky W, Palumbo P, Kostrikis L, Ho DD. 2000. Measuring recent thymic emigrants in blood of normal persons and HIV- I -infected patients before and after effective therapy. J. Exp. Med. 190:725–732 Walker RE, Carter CS, Muui L, Natarajan V, Herpin BR, Leitman SF, Klein HG, Mullen CA, Metcalf JA, Baseler M, Falloon J, Davey RT Jr., Kovacs JA, Polis MA, Masur H, Blaese RM, Lane HC.

118.

119.

120.

121.

122.

123.

1998. Peripheral expansion of pre-existing mature T cells is an important means of CD4` T-cell regeneration HIVinfected adults. Nat. Med. 4:850–56 Gobbi A, Stoddart CA, MaInati MS, Locatelli G, Santoro F, Abbey NW, Bare C, Linquist-Stepps V, Moreno MB, Herndier BG, Lusso P, McCune JM. 1999. Human herpervirus 6 (HHV-6) causes severe thymocytes depletion in SCID-hu Thy/Liv Mice. J. Exp. Med. 189:1953–60 Guinan EC, Boussiotis VA, Neuberg D, Brennan LL, Hirano N, Nadler LM, Gribben JG. 1999. Transplantation of anergic histoincompatible bone marrow allografts. New Engl. J. Med. 340:1704– 14 Kirk AD, Burkly LC, Batty DS, Baumgartner RE, Berning JD, Buchanan K, Fechner JH Jr, Germond RL, Kampen RL, Patterson NB, Swanson SJ, Tadaki DK, Tenhoor CN, White L, Knechtle SJ, Harlan DM. 1999. Treatment with humanized monoclonal antibody against CD 154 prevents acute renal allograft rejection in nonhuman primates. Nat. Med. 5:686–93 Kouskoff V, Korganow AS, Duchatelle V, Degott C, Benoist C, Mathis D. 1996. Organ-specific disease provoked by systemic autoirnmunity. Cell 87:811–22 Gilquin J, Viard J-P, Jubault V, Sert C, Kazatchkine MD. 1998. Delayed occurance of Graves’ disease after immune restoration with HAART. Lancet 352: 1907–8 Chun T-W, Engel D, Mizell SB, Hallahan CW, Fischette M, Park S, Davey RF Jr, Dybul M, Kovacs JA, Metcalf JA, Mican JM, Berrey MM, Corey L, Lane HC, Fauci AS. Effect of interleukin-2 on the pool of latently infected, resting CD4` T cell in HIV- I infected patients receiving highly active anti-retroviral therapy. Nat. Med. 5:651–56


HAYNES ET AL C-1

Figure 2 The human thymus during normal aging. All panels are hematoxylin and eosin stained sections of human thymus, counter-stained with an antibody against keratin (brown areas). Top panel is thymus of a 2-month-old male with cortex (C) and medulla (M) of the true thymic epithelial space and narrow perivascular spaces (P) indicated. Dark central area is a Hassall's body. Panel B is thymus of a 36-year-old female with well-defined cortex (C) and medulla (M) identified. The perivascular space (P) is enlarged with both lymphoid cells (blue P areas) and adipose tissue (white P areas). Panel C shows thymus of a 60-yearold male with very small areas of normal cortex (C) and medulla (M), and perivascular space (P) adipose tissue (white P area) predominating, and a small amount of lymphoid tissue in the perivascular space around thymic epithelium (blue P area) ( all panels x 10).


C-2 HAYNES ET AL

Figure 5 The thymus in myasthenia gravis (MG). Panel A shows reactivity of a hematoxylinstained section of MG thymus with a monoclonal antibody against the CD1 antigen (brown) that is expressed by cortical thymocytes. Arrows point to areas of thymopoiesis in the cortex of the thymic epithelial space that are CD1+. G points out germinal centers in the perivascular space (P). Other white P areas point out perivascular space with adipocytes. Panel B shows the same areas in Panel A stained with an anti-keratin antibody (brown) (arrows) showing the thymic epithelial space and areas of compression of keratin+ thymic epithelial cells by expanded germinal centers (G) within the perivascular space (P) (x 10).


HAYNES ET AL C-3

Figure 7 Immunohistologic analysis of a chimeric thymus allograft placed 3 months earlier in the thigh of a patient with complete DiGeorge syndrome. Panel A is a hematoxylinstained section of the lymphoid graft (arrows) surrounded by muscle (M). Panel B is stained in immunohistochemistry using an anti-CD3 mab. Arrow shows positive cells. Panels C-F are stained using indirect immunofluorescence with mabs against CD1 (C), CD4 (D) donor thymus MHC class I (E), and recipient MHC class 11 (F) mabs. Graft had areas of normal CD4+ cortex thymocytes (C) that were CD4+ (D) and CD8+ (not shown), and had thymic microenvironment expansion of donor thymus MHC molecules (arrows) (E). Panel F shows host MHC class II+ macrophages outside and migrating inside (arrows) graft. (All panels x 400).


C-4 HAYNES ET AL

Figure 9 The thymus in HIV infection. Panel A shows keratin positive empty thymic epithelium (E) (brown) surrounded by an intense lymphoid infiltrate in the perivascular space (P) in late stage AIDS (x 10). Panel B shows in situ hybridization for HIV- I RNA expression and shows HIV-1 expressing cells in the perivascular space (P) and in the thymic epithelial space (brown) (E) (arrows) (x 40). Panel C shows the thymus epithelial space (E) stained red with anti-keratin antibody and perivascular space (P) CD8+ lymphocytes stained brown (arrows) (x 10). Panel D shows TIA-l+ DonhamCTL granules in perivascular space (P) lymphocytes (arrows). E is epithelial space. (x 20). Panels E and F show a thymic epithelial tuft in an end stage AIDS thymus with residual thymopoiesis. Panel E shows keratin + thymic epithelium (arrows) and panel F shows CD1+ developing thymocytes (brown) (arrows) (x 40).


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


ACCESSING COMPLEXITY: The Dynamics of Virus-Specific T Cell Responses Peter C. Doherty and Jan P. Christensen Department of Immunology, St Jude Children’s Research Hospital, Memphis, Tennessee 38105; e-mail: [email protected]


Key Words tetramers, cell-mediated immunity, inflammation, pneumonia, memory, recall response Abstract The cellular dynamics of the immune system are complex and difficult to measure. Access to this problematic area has been greatly enhanced by the recent development of tetrameric complexes of MHC class I glycoprotein ` peptide (tetramers) for the direct staining of freshly isolated, antigen-specific CD8` T cells. Analysis to date with both naturally acquired and experimentally induced infections has established that the numbers of virus-specific CD8` T cells present during both the acute and memory phases of the host response are more than tenfold in excess of previously suspected values. The levels are such that the virus-specific CD8` set is readily detected in the human peripheral blood lymphocyte compartment, particularly during persistent infections. Experimentally, it is now possible to measure the extent of cycling for tetramer `CD8` T cells during the acute and memory phases of the host response to viruses. Dissection of the phenotypic, functional, and molecular diversity of CD8` T cell populations has been greatly facilitated. It is hoped it will also soon be possible to analyze CD4` T cell populations in this way. Though these are early days and there is an enormous amount to be done, our perceptions of the shape of virus-specific cell-mediated immunity are changing rapidly.

INTRODUCTION Vertebrates interact with the infinite variety of external challenge via two great complex systems, the immune system and the central nervous system (CNS). Apparently-inclusive terms such as recognition, triggering, response, and memory are used by both immunologists and neurobiologists. The meanings, however, are very different in the two fields. The CNS reacts to and recalls events by establishing pathways of electrical conduction and interconnectivity. The immune system, on the other hand, responds and remembers by processes that involve proliferation of the specifically committed cellular elements (the lymphocytes) that bear receptors for the inducing stimulus (antigen). Whereas the CNS is ana0732–0582/00/0410–0561$14.00

561


562

DOHERTY n CHRISTENSEN

tomically stable and occupies the same, defined space from maturity, the size of the immune system can vary dramatically and is difficult to quantify. The present discussion looks at the issue of measurement. Perhaps the greatest challenge facing immunologists is to understand how lymphocyte numbers are regulated, the nature of the underlying homeostatic control mechanisms (1–4). The questions posed are very different from those that arise when we consider the volume and shape of a stable, solid organ such as the brain or liver. Although the size of the lymph nodes and spleen and the cellularity of the blood phase in the ‘‘resting’’ immune system may be relatively predictable, it is much more difficult to measure the extent of the lymphocyte ‘‘diaspora’’ to other solid tissues. Also, we have little understanding of lymphocyte turnover rates in vivo, the relationship between cell cycling and cell death, and the kinetics of entry and exit from various body compartments (3, 5, 6). No other organ system in the body has this duality of a solid-tissue and an anatomically independent dispersed phase, consisting of cellular elements that vary greatly in receptor specificity, life span, and differentiation state. One place to start with the immune /homeostasis problem is to focus the analysis on the types of selective pressures that have shaped the mammalian host response. All the available information suggests that the immune system as we know it has evolved to protect an interdependent, multiorgan complex (the vertebrate) from the consequences of colonization by simpler, and more specialized, life forms. The viruses are obligate intracellular parasites and are thus the most intimate of all such pathogens. In a sense, when we analyze the nature of the cellular events that result from the confrontation with infectious viruses, we are studying the basic physiology of vertebrate immunity. At least at the level of the CD4` and CD8` T cell responses, the problem can be approached as one of population biology at the single cell level, interfacing the virus-infected cells on the one hand with the antigen-specific lymphocytes on the other. This is where immunobiology is starting to become accessible to the mathematical biologists, offering the possibility of predictive models of lymphocyte dynamics (7–9) that may help us to deconstruct immune homeostasis. Much of what follows addresses the insights and questions that have been raised by a recent technological breakthrough, the development of tetrameric complexes of MHC class I glycoprotein`peptide (tetramers) that allow the direct staining of antigen-specific CD8` T cells (10). An alternative strategy that uses a dimeric, peptide-loaded complex of MHC class I-Ig gives similar results (11). The term tetramer is used in a generic sense throughout this discussion to describe both approaches. Comparable protocols are also being developed for the CD4` set (12), but little has yet been published. For the first time, the tetramers (Tet) are giving us real numbers. However, nobody is really sure what these numbers mean, particularly when the new data implies levels of clonal expansion and persistence that are in great excess of anything predicted previously (10, 13–16). What is clear is that we have started down a path that is rapidly changing our understanding of cell-mediated immunity (CMI).

DYNAMICS OF T CELL RESPONSES

563

THE VIRUS SYSTEMS


Before discussing the available data generated with the tetramers, it is important to say a little about the various virus models that have been analyzed so far. Although the situation has changed dramatically since the emergence of the human immunodeficiency viruses (HIV) as a major threat, the numbers of experimental immunologists working with viruses remain small. If a discussion of cellular dynamics in virus infections is to be useful, it is first necessary to understand at least some of the characteristics of these systems that are complex in their own right. Those that have been looked at to date are broadly representative of the diversity of viral pathogens. Some useful, general conclusions about T cell responsiveness can be drawn when we take into account the nature of these different experimental and natural infections.

Influenza Virus and Sendai Virus: Localized Infections Without Persistence Most of the experiments that have focused on quantifying CMI to viruses that cause localized infections of limited duration have used two negative-strand RNA viruses, the influenza A viruses, and the murine parainfluenza type 1 virus, Sendai virus (6, 14, 17, 18). Both are respiratory pathogens that, because they require an anatomically restricted trypsin-like enzyme to cleave the surface hemagglutin (influenza) or fusion (Sendai) proteins, cause productive infections limited largely to the superficial epithelial cells of the respiratory tract (19–21). These viruses are also thought to cause defective infection (no progeny virus) of the prominent dendritic cell layer in the upper part of the lung (22, 23). There is good evidence that dendritic cells are the major stimulator population in the regional lymph nodes and spleen (24, 25) although, at least in the mouse, it is difficult to prove that they have indeed migrated from the virus-infected respiratory tract (26). Both viruses are thought to be completely eliminated from normal, previously unexposed mice after 10–12 days, and there is no evidence for the persistence of viral genome following respiratory challenge (27). Some of the more virulent influenza A viruses will also replicate to a more limited extent in other tissues, such as the liver, brain, and lymphoid tissue (28, 29), although most mouse experiments have used virus strains that do not have these characteristics. Many influenza A viruses are maintained in nature as infections of the alimentary tract of birds, and they also infect a variety of other mammals such as horses, pigs, and seals (30). The great variety of surface hemagglutinin (H) and neuraminidase (N) glycoprotein types that are circulating as a consequence of this broad species distribution provides a potential source of pandemic strains that will not be recognized by pre-existing antibody (30). The Hong Kong (HK) H3N2 influenza A viruses that have been circulating in humans for the past 40 years probably arose originally as a consequence of the segmented

564


influenza genome (8 genes) repackaging in a respiratory epithelial cell infected simultaneously with a duck virus and a human virus (30). As is the case for many viruses (31), much of the influenza virus–specific CD8` T cell response is directed against peptides derived from conserved internal components such as the nucleoprotein (NP). Even before Townsend et al used the influenza model to establish for the first time that CD8` T cells are specific for NP peptide`MHC class I glycoprotein (32), some of the basic rules of the recall response were being worked out using priming and challenge experiments with the serologically distinct H1N1 and H3N2 influenza A viruses (33, 34). This continues to be an extremely useful experimental system (14, 29), which mimics to some extent what happens in the natural epidemiology of influenza.


Persistent CNS Infections with Lytic RNA Viruses Variants of both Theiler’s murine encephalomyelitis virus (TMV) and murine hepatitis virus (MHV) have been selected for the capacity to persist at a low level in the murine CNS (35, 36). The chronic pathology caused by TMV (enterovirus) and MHV (coronavirus) is considered to mimic some aspects of human multiple sclerosis (MS). Intracerebral challenge of mice with Sendai virus has also been associated with evidence of CNS persistence (37). In each case, the continued presence of virus is restricted to the brain and spinal cord, reflecting the particular limitations of CMI effector mechanisms in neural tissue, another complex issue not discussed in detail here. A somewhat analogous situation in humans is the disease subacute sclerosing panencephalitis, where the failure to clear a defective measles virus from the brain is associated with massive B cell/plasma cell infiltration, local, virus-specific antibody production, coma, and death (38, 39).

Lymphocytic Choriomeningitis Virus (LCMV): Systemic Infection With or Without Persistence No virus model has taught us more about CD8` T CMI than LCMV (21, 40– 45). The CD8` T cell response is extremely potent and readily measured by the CTL assays that dominated this field until 1996. However, it is important to understand that this RNA virus is very unusual. LCMV induces massive, generalized infections, but causes so little damage that mice infected in utero remain persistent, asymptomatic carriers for life (46). Classically, adult mice that have not previously been exposed to ‘‘neurotropic’’ LCMV die from CD8` T cellmediated immunopathology following intracerebral challenge or develop severe hepatitis following extraneural exposure to a ‘‘viscerotropic’’ strain. The consequence of peripheral exposure to a moderate dose of LCMV is generally rapid virus growth followed by complete elimination (47), though the latter view is challenged (48). This nonlytic infection is one of the few showing an absolute requirement (49) for perforin to mediate CD8` T cell effector function in vivo.


565


Because LCMV induces little damage in the absence of CMI, it is possible to achieve enormous virus titers in vivo without any obvious deleterious effect on the host. A typical, lytic virus would be lethal under these conditions. The consequence can be ‘‘high dose immune paralysis’’ (50) or, as it has been called more recently, ‘‘immune exhaustion’’ of the CD8` T cell response (51), with subsequent persistence of high-level LCMV infection. Although the immune exhaustion idea has now been around for some time, it is not clear how applicable it is to any other virus. The case has been looked at most thoroughly for HIV (52– 54), but HIV is much more lytic and it is likely that infected individuals will die rapidly from virus-induced pathology in the absence of CD8` T cell–mediated control (21, 55). Thus, though LCMV has taught us a great deal about both CMI and tolerance, the results from this model should be considered at one end of a spectrum and far from typical of what happens in the great majority of virus infections.

Persistent Retrovirus Infections The need to deal with HIV has led to the development of quantitative approaches for analyzing both virus load and the CD8` T cell response (56–58). However, the more detailed dissection of pathogenesis mechanisms in vivo has been severely constrained by the lack of a comparable mouse pathogen. Infection of rhesus macaques with the related simian immunodeficiency virus (SIV) lentivirus has been exploited with considerable effect (15, 55), but the experiments are limited by economic considerations and the lack of the sophisticated genetic technology provided by the availability of mice with selectively disrupted host response genes. Even so, the studies with HIV and SIV are teaching us a great deal about the limits of immune control, and they have forced the development of better protocols for studying human (and nonhuman primate) immunity. The analysis of people infected with HIV provided the first account of the direct staining of virus-specific CD8` T cells with tetramers that has so changed our thinking about the magnitude of CD8` T cell responses (10). Similar studies (11, 59–61)have also been done with people infected with human T cell leukemia virus type 1 (HTLV-1), an oncogenic retrovirus that can cause both transformation of CD4` T cells and the neurological disease, tropical spastic paraparesis (TSPE).

The Gammaherpesviruses (cHV): Large DNA Viruses that Establish Latency in Lymphoid Tissue The basic pathogenesis of the cHVs is that they cause initial, lytic infections of the oral/respiratory mucosa, then establish persistent latency in the B lymphocyte compartment (62–64). These are very evolved ‘‘companion’’ pathogens of mammals that seem, at least in part, to have achieved a profile of stable parasitism by appropriating (or mimicking) various host response genes (65–67). The two viruses that have received the most attention from the CMI aspect (64) are the


566


prototypic human type 1 cHV, Epstein Barr virus (EBV), and a type 2 cHV (cHV68) that was originally isolated from voles but also causes substantial infections in laboratory mice (63). Much less is known about the Kaposi’s sarcoma–related (68, 69) type 2 cHV human herpesvirus 8 (HHV-8), though recent experiments have shown that it does indeed induce a CD8` T cell response (70). The consequences of reactivation from latency following in vitro culture are somewhat different for EBV and cHV-68. While EBV causes B cell transformation (71), an effect that has been used experimentally over the years to develop lymphoblastoid cell lines from the peripheral blood lymphocyte (PBL) pool, coculturing B cells that are latently infected with cHV-68 on susceptible cell monolayer leads to reactivation to lytic phase and cell death (72). Even so, latent EBV and cHV-68 both revert to cause productive infection in the in vivo situation, with evidence of excretion from the oropharynx (EBV) or virus replication in the lung (cHV-68) being most apparent under conditions of immunosuppression (73– 75). These viruses also cause massive, long-term stimulation of the CD8` T cell compartment (76). The prolonged, selective increase of activated CD8` T cell numbers in the PBL (and accompanying lymph node enlargement) associated with primary EBV infection of adolescents (77) is called infectious mononucleosis (IM). An apparently similar profile is seen following the resolution of the acute phase of cHV-68 infection (78). The numbers of CD4` T cells and B cells are also substantially increased, but the effect on the CD8` population is proportionally greater with the consequence that this subset dominates the blood profile (4). Even so, it is not clear that the CD8` T cell expansion caused by cHV-68 can indeed be considered equivalent to the IM syndrome associated with EBV. The expanded CD8` set in mice infected with cHV-68 is dominated by a population expressing the Vb4 T cell receptor (TCR) chain in association with a spectrum of TCRa chains (78, 79). These CD8`Vb4` lymphocytes have no known peptide specificity and are detected at high frequency in mice expressing a variety of MHC haplotypes (78, 80). In fact, though the expansion of this Vb4`CD8` set depends on the presence of both B cells and CD40 ligand (CD40L)-mediated CD4` T help (81, 82), there may be no requirement for the Vb4`CD8` T cells themselves to interact with either MHC class I or class II glycoprotein (83). No comparable response has been described for EBV, where many of the CD8` T cells in the blood are clearly specific for viral peptides presented conventionally by MHC class I glycoproteins (16).

A BRIEF HISTORY OF ANTIGEN-SPECIFIC T CELL MEASUREMENT Our understanding of CMI has developed progressively over the 30 years or so since the realization that there are distinct categories of T and B lymphocytes (84). The tetramer staining approach is the latest, and most powerful, of a series


567

of advances determined by the emergence of new technologies. Though substantial insight has also been gained from adoptive transfer experiments using lymphocytes from mice transgenic for particular TCR ab heterodimers (85), these studies are not considered in detail here. The present discussion focuses on the quantitation of viral immunity in unmanipulated hosts with a normal T cell repertoire. The following summarizes briefly the history of attempts to measure T cell responses, particularly for the CD8` subset.


In Vivo Bioassays Early efforts at determining the potency of antigen-specific T cell populations utilized the adoptive transfer of serially diluted lymphocytes, followed by assay for some measure of in vivo effector function. The readouts included increase in lymph node size and weight (for alloreactivity), rapidity of allograft rejection, foot pad swelling, the induction of inflammatory pathology, and the reduction in titers for an infectious agent growing in the liver or spleen, particularly Listeria monocytogenes, ectromelia virus, or LCMV(86–89). None of these protocols was capable of determining effector T cell numbers. An in vivo–in vitro approach that did allow direct measurement (but had limited applicability) was to do colony counts for xenoreactive T cells following proliferation on the chorioallantoic membrane (CAM) of the chick embryo (90), a protocol that was pursued briefly in the 1960s by FM Burnet. Burnet had developed the chick CAM technology in the 1930s for quantitating the growth of a spectrum of mammalian viruses, following on from his earlier work with the bacteriophages. These are among the last experiments published by Burnet, who won the Nobel Prize with PB Medawar in 1960 for the theory of immunological tolerance (91). The hundredth anniversary of Burnet’s birth was celebrated in 1999. When we talk about lymphocyte numbers, the discussion automatically assumes the fact of clonal selection, the other major immunology concept that was formalized in a conceptually satisfying way by Burnet. It is interesting to speculate that Burnet’s earlier focus on the selection of virus variants and quantitative virology (91, 92) led him to perceive immediately the utility of the clonal selection idea, which seems to have been first proposed by David Talmage (92– 94).

The Cytotoxic T Lymphocyte (CTL) Assay The exploitation of the in vitro 51Cr release assay for measuring alloreactivity by Cerottini & Brunner (95), and the extension of this approach to the virus models (96), opened out the field of CD8` T CMI. The discovery of MHC class I restriction in 1973/1974 was a direct consequence of the use of this technique, allowing us to develop an appropriate conceptual framework much more rapidly than those who had similar findings with the CD4` subset (97). Various applications of the

568


CTL approach continued to dominate the analysis of the CD8` T cell response until the mid-1990s. Though the CTL assay provides numbers, it cannot be regarded as truly quantitative. Comparisons were made using ‘‘lytic unit’’ calculations (98, 99), but these provided no real understanding of effector T cell prevalence. Attempts were also made to develop single cell binding assays (99, 100) with a cytotoxicity readout, but these experiments were difficult and (as we know in hindsight) probably confounded by the spectrum of adhesion molecules that are upregulated on the activated T cell.


Limiting Dilution Analysis (LDA) The combination of 6–7 day microculture (in 96 well plates) under LDA conditions, followed by assaying individual wells for CTL activity, was the standard protocol (101, 102) for determining virus-specific CD8` T cell numbers from the late 1970s to the mid-1990s (3). The experiments were tedious to perform and notoriously subject to variability for a range of technical reasons. The approach was also extremely cumbersome. If, for example, we wished to determine the TCR or activation (e.g., CD62L, CD44, CD69) phenotype (103) for virus-specific CD8` T cells, it was necessary to first sort the lymphocytes in the flow cytometer, then stimulate them under LDA conditions (104, 105). It was also very clear that LDA greatly underestimated the numbers of effector CD8` T cells recovered, for example, from a site of virus-induced pathology (106, 107). The reason for this is obvious. A single virus-specific CTL precursor (CTLp) must undergo 10–15 cycles of replication before achieving a colony size that can be read out by assaying lytic activity (108). Highly activated T cells are driven to apoptosis when stimulated continuously in this way. Even so, the kinetic data describing the establishment and persistence of CD8` T cell memory, which was so painfully developed from LDA experiments over many years, looks to be reasonably valid (3, 45) though the numbers may be far too low (13). There is also a continuing debate about whether the profile for functional CD8` T cell memory is more accurately reflected by LDA, or by staining with tetramers (36). More experiments need to be done before we can reach a useful consensus on this question. The LDA protocol has also been used to determine CD4` T cell frequencies, the readout being IL-2 production rather than cytotoxicity (109, 110). This approach has now been replaced in most laboratories (111–113) by the single cell ELISA Spot (ELISPOT) assays (114). At this stage, however, much less is known about the quantitative aspects of virus-specific CD4` T cell responses. The area merits more attention, as CD4` T cell–mediated effector function is clearly of primary importance in some types of virus infections (115–117).

Direct Staining of Virus-Specific CD8` T Cells The direct staining of virus-specific CD8` T cells with tetrameric complexes of MHC class I glycoprotein`peptide is in the process of revolutionizing both our



569

understanding and the way that we approach the analysis of CD8` T CMI. The experimental utility of the tetramers is, however, limited by the sensitivity of flow cytometry. Populations that comprise less than 1.0% of a particular lymphocyte subset are clearly at the limit of detection. The rate of data acquisition can obviously be improved by first enriching for the cells in question. One possibility is to use magnetic separation protocols to select, for example, the tetramer (Tet)` set prior to FACS analysis. Otherwise, when the prevalence falls to less than 0.5%, it may be appropriate to use an LDA approach with ELISPOT as the readout. The ELISPOT may give frequencies that are equivalent to those determined with the tetramers, or values that are two- to fourfold lower, depending on the experimental system and the particular laboratory (13, 118). Relatively few studies have been done to date with CD8`Tet` cells that were first sorted, then analyzed by ELISPOT. The level of confidence that the tetramer protocol indeed provides valid numbers is (apart from the usual specificity controls) considerably strengthened by the fact that stimulation for 6 h with high doses of cognate peptide in the presence of Brefeldin A (to stop protein secretion), followed by fixation and staining for intracellular interferon c (IFN-c), gives essentially identical results (13, 14, 17, 75, 119). Much of the continuing challenge is to develop a clearer understanding of the functional and molecular diversity of these lymphocyte populations (120). A recent molecular analysis indicates, for example, that the CD8`Tet` memory population in mice that have recovered from LCMV infection is derived from distinct subset generated during the course of the acute host response (121). Most of the experiments so far have, however, focused on measuring the dynamics of virus-specific immune responses, the subject of the present review.

ANALYTICAL POSSIBILITIES OPENED BY THE TETRAMER TECHNOLOGY The capacity to analyze unfixed, viable antigen-specific CD8` T cells directly by multiparameter flow cytometry enables a spectrum of applications that were simply not possible with earlier functional approaches. The CTL assay measures a single property, without providing any information on the number of effectors within that particular lymphocyte population. Use of the CTL/51Cr –release-based LDA has effectively been limited to academic research groups with substantial budgets for expendable supplies (medium, fetal calf serum, 96 well plates) and access to a multi-channel c-counter. Prior to 1996, determining something as straightforward as the level of CD44 expression on a CTLp required that the lymphocytes first be stained and sorted, then stimulated for 6 days under LDA conditions. The availability of the tetramers means that any group that has access to a FACScan (Becton Dickinson), or equivalent instrument, can now determine the relative prevalence of antigen-specific CD8`CD44hi and CD8`CD44lo simply by staining and two-color FACS analysis. A few examples of how tetramer staining has greatly facilitated the dissection CMI are listed below.

570



Monitoring Human Disease The clinical situation obviously becomes much more accessible when functional assays can be replaced with morphometric analysis of relatively small numbers of lymphocytes. Most clinical immunology, pathology, and infectious disease programs have access to a flow cytometer and are accustomed to making such determinations. So far, much of the tetramer analyses with humans has been done for situations where antigen is known to persist. Prevalence rates of 1–10% for a particular Tet` set in the PBL have been shown for people persistently infected with EBV, HIV, and HTLV-1 (10, 16, 54, 55, 57–61). Tumor-specific CD8` T cells can be found at frequencies of 2–5% in at least a proportion of melanoma patients (122). Though influenza A viruses do not persist, all human populations are subject to regular challenge with cross-reactive viruses. Again, Tet` influenzaspecific T cells are readily shown to be circulating in the PBL compartment (10). Once the appropriate Tet reagents become more widely available, their utility for clinical studies will obviously be determined by cost and by the particular spectrum of HLA types present in the individuals being studied. Apart from looking at patients, a likely application will be to monitor trials with both viral (HIV) and tumor vaccines. It is reasonable to expect that the tetramers will soon be regarded as an essential reagent within (at least) the clinical research community, leading to a ‘‘democratization’’ of quantitative cellular immunology. One consequence will undoubtedly be to bring new, active minds into this area of investigation.

Antigen-Specific CD8` T Cells in Local Anatomical Niches and Sites of Inflammatory Pathology Respiratory challenge with the influenza viruses, Sendai virus, and cHV-68 causes severe lung pathology. In each case, the inflammatory exudate recovered by bronchoalveolar lavage (BAL) contains potent CTL effectors (eCTL). Previous LDA analysis suggested that viral peptide-specific CTLps were present at frequencies less than 1:50, a finding that seemed at odds with the extremely strong CTL activity in this site of high virus growth. Depending on the model system, it is now possible to account for up to 80% of the CD8` T cells in these virus-induced inflammatory exudates by probing with the appropriate peptide (IFN-c assay) and/or tetramer (14, 18, 76). We were obviously unable to expand most of the eCTL by LDA, probably because such highly activated lymphocytes are readily driven to apoptosis (3, 107). Similar situations have now been described for CD8` T cells recovered from the CNS of mice infected with MHV and Theiler’s virus (123, 124). Enrichment (three times greater than PBL) of HTLV-1–specific CD8` T cells has also been described for cerebrospinal fluid from a TSPE patient (11). Virus-specific T cells comprised .20% of the CD8` population in seminal fluid obtained from an SIV-


571

infected macaque (125). Again, these activated T cells would probably not have been detected by LDA.


Analyzing Lymphocyte Proliferation In Vivo Prior to the availability of tetramers, measurement of T cell proliferation rates in vivo was basically limited to adoptive transfer experiments using, in particular, donor lymphocytes from mice transgenic for a particular TCR ab pair. We tried to determine the extent of cycling by giving influenza virus–infected mice high levels of the thymidine analogue, bromodeoxyuridine (BrdU), then using the fact that BrdUhi cells die subsequent to exposure to bright light to ‘‘suicide’’ the responding lymphocytes (126). The residual CTLp frequencies were then measured by LDA. The key conclusions from this ‘‘suicide’’ approach have been confirmed in later experiments with the tetramers, but the technique was too cumbersome for general use and the excessive amounts of BrdU proved toxic in some experiments. The proliferation analysis can now be done quite simply by feeding mice much lower levels of BrdU throughout the period of antigen challenge, then using threecolor FACS analysis to determine the prevalence of the CD8`Tet`BrdUhi set (Figure 1). Such experiments (3, 5) have shown us that there is massive CD8` T cell proliferation in both primary and secondary responses, with the duration of the proliferative phase and the likely antigenic exposure correlating quite closely. However, it is important to recognize the inherent limitations of this analysis. If, for example, we have a starting CD8`Tet` memory population at a prevalence of 1.0% in the spleen and then give antigen, the entry of only 10% (0.1% of total) of these cells into the cell cycle could, assuming one division every 12 h and the survival of all progeny, lead to the conclusion that the great majority of ‘‘resting’’ memory T cells were indeed capable of dividing further. An alternative protocol is to measure the extinction of labeling following secondary stimulation of Tet`BrdU` memory cells that were generated during the acute, antigen-driven phase of the primary response. The particular difficulty with this analysis is that we have little real data on either the extent of cell death or localization to sites other than those being monitored. Both death and dispersion are likely to be confounding variables for any in vivo study of this type. The analysis of rate effects for antigen-specific T cells is one area likely to benefit from the involvement of the mathematical biologists.

Looking at ‘‘Bystander’’ Activation The term bystander has a red flag quality for some cellular immunologists, so it is important to establish what is actually meant. The definition of bystander used here is that memory T cells of one specificity are induced to express some phenotypic or functional change characteristic of activation during the course of an apparently unrelated infectious process (107, 127). The focus of the contro-


572


Figure 1 Proliferation status of influenza A virus specific (CD8`NPP`) memory T cells (A) recovered from the spleen at 8 days after challenge with cross-reactive (B) or unrelated viruses (C,D). The C57BL/6J (B6, H-2b) mice were primed by intraperitoneal (i.p.) injection of a large dose of the PR8 (H1N1) influenza A virus, rested for 1 month, then anesthetized and infected intranasally (i.n.) with the HKx31 (H3N2) virus that shares the NP366–374 peptide. These secondarily stimulated mice were held for a further month, then fed BrdU in the drinking water. One group (A) was not challenged again, while others were infected i.p. with LCMV (D) or i.n. with the B/HK influenza B virus (C). A control group of singly-primed, PR8-immune mice was infected i.n. at the same time with the HKx31 virus (B). The profiles were generated by three-color analysis using a FACScan and Lysis 2 software. The percentage values in the upper right quadrants are for the BrdUhi set (R3 region in ‘‘A’’).

versy is whether such effects are simply a consequence of local exposure to cytokine (IFN- L, IL-15) or depend also on low avidity/affinity TCR-mediated recognition (127–130). The magnitude of the clonal expansion that occurs in at least some virus-specific CD8` T cell response indeed suggests the possibility


573


of a measure of cross-reactivity. Other evidence indicates that the sporadic production of cytokines like IL-15 may be important for the maintenance of T cell memory (131). Whatever the explanation for the ‘‘bystander’’ effect, the phenomenon can obviously be analyzed much more effectively using the tetramer approach. The example in Figure 1 shows BrdU staining profiles for CD8` memory T cells specific for the influenza A virus epitope H-2Db`NP366–374 (NPP). Some of these doubly primed (H3N2rH1N1) mice were infected again with LCMV or with an unrelated influenza B virus. The response to LCMV, for example, is associated with evidence of cycling for .30% of the CD8`NPP` population (Figure 1D). However, it is also possible that proliferating, influenza A virus–specific CD8` memory T cells are localizing selectively to ‘‘activated’’ lymphoid tissue. Can we find any evidence of cross-reactivity with LCMV for this population of ‘‘blasted’’ influenza A–peptide specific, CD8`NPP` T cells? The experiments are in progress.

QUANTITATION OF THE PRIMARY VIRUS-SPECIFIC CD8` T CELL RESPONSE Tetramer analysis of the primary CD8` T cell response has, for obvious reasons, only been possible to date with the experimental animal models. As might be expected, there are common features and differences for the various viruses. Also, the same questions have not been asked in every study.

Clonal Expansion in Lymphoid Tissue and the Role of Antigen Naı¨ve CD8` Tet` lymphocytes are present at too low a frequency to be detected by flow cytometric analysis of either the thymus or the secondary lymphoid tissue. Following primary infection, the CD8`Tet` set is first detected in the regional lymph nodes and/or spleen within 5–7 days. The numbers of virus-specific CD8` T cells generally remain relatively low (,3.0% of the CD8` population) for the influenza A viruses (primary, Figure 2), MHV and TMV, all of which cause substantially localized infections. Higher frequencies are found for the systemic infections caused by cHV-68 and SIV, with the prevalence of the CD8`Tet` set in the PBL compartment of mice given cHV-68 and macaques infected with SIV broadly reflecting the values determined for lymphoid tissue (80, 132). The latter observation should bring particular satisfaction to immunologists studying humans, who are generally dependent on the PBL as the only readily accessible source of antigen-specific T cells. The cumulated response to three different cHV-68 peptides accounted for 7– 10% of the splenic CD8` population at 10–15 days after primary challenge (76). The frequency of CD8` T cells specific for the NP324–332 peptide of Sendai virus


574


Figure 2 The persistence of CD8`NPP` T cells in the MLN and spleen following primary or secondary (H1N1-primed) challenge with the HKx31 (H3N2) influenza A virus (see legend to Figure 1). This data is given in more complete form in Reference 6, and is supported by more extensive findings from later experiments that are not yet published.

may also reach levels of 10% or so in lymphoid tissue (18), though this virus does not normally replicate outside the respiratory tract. The Sendai NP324–332 peptide seems to be immunodominant to an unusual extent, engaging an extremely diverse spectrum of TCR ab pairs (133). Evidence of much greater clonal expansion in the antigen-specific CD8` set has been described following primary challenge with LCMV (13), supporting the long-held view that the host response to this virus is heavily skewed toward CD8` T CMI. The prevalence of CD8`Tet` LCMV-specific populations in spleen can be .50%, reflecting the high level of antigen challenge with this essentially non-lytic virus. Antigen availability does seem to be the main parameter determining the duration of the proliferative phase in the primary response phase. This has been analyzed by BrdU staining for the influenza-specific CD8`NPP` set (5). The findings indicate that every CD8`NPP` T cell divides a number of times during the 8 days following the initial exposure to virus. Evidence of continued cycling can also be detected for a few days after virus can no longer be isolated from the lung. There are indications that antigen presenting cells (24) may persist for a


575

few days after the clearance of infectious virus from a particular location, though this question has only been analyzed in any detail for Sendai virus (25).


Accumulation in Sites of Virus-Induced Pathology Analysis with the tetramers has demonstrated very clearly that antigen-specific CD8` T cells concentrate at foci of maximum antigen production. The prevalence of CD8` T cells specific for an immunodominant epitope can range from 15% (influenza NPP) to 70% (Kb ` Sendai NP324–332) of the CD8` set obtained by BAL of the pneumonic lung (14, 18). Similarly, more than 80% of the inflammatory CD8` T cells recovered from the respiratory tract of mice challenged intranasally (i.n.) with cHV-68 can be shown to be specific for peptides derived from proteins expressed during the lytic phase of virus replication (75, 76). The values for CD8` T cells isolated from the CNS of mice infected with TMV or MHV (123, 124) can be as high as 60% or 26%, respectively, for the most prominent peptide, with the total MHV-specific component comprising 50% of the inflammatory CD8` T cells when other epitopes are included in the analysis. In general, the frequencies for any particular peptide may be five- to tenfold higher than those found in the responding lymphoid tissue (13, 14, 18). With TMV, however, it has not been possible to detect virus-specific CD8` T cells in the cervical lymph nodes or spleen using tetramers that give evidence of massive localization to the CNS (124). On the other hand, the extraordinarily high values (.50% CD8`Tet`) for the spleens of mice infected with LCMV (13) can be considered to reflect that the lymphoid tissue is a site of virus-induced pathology in this infectious process. At least for the pneumonia models, the recruitment to (and retention of) virusspecific CD8` T cells to the infected lung has generally been considered to reflect localization via the blood subsequent to clonal expansion in the regional lymph nodes and spleen (3). The lymphocyte populations obtained by BAL also contain memory T cells specific for other antigens (107), though the tetramer studies indicate that these lymphocytes are relatively less prevalent in a given inflammatory exudate than was originally thought to be the case (14). The LDA undoubtedly reads out passively recruited ‘‘resting’’ T cells more effectively (36) than the highly activated eCTL that are controlling the infectious process. The magnitude of the nonspecific T cell component may also reflect the molecular characteristics of the inducing virus, being greater for the influenza A viruses than for cHV-68 or Sendai virus. These inflammatory processes resolve rapidly once antigen is eliminated from the target organ (Figure 3). A proportion of the virus-specific CD8` set in the lungs of mice with influenza pneumonia may return to the circulation and be removed in the liver (17). Other, apoptotic T cells are probably engulfed by macrophages and eliminated via retrograde mucus flow and coughing. Otherwise, few apoptotic CD8` T cells are detected in the BAL population at any stage of influenza virus–induced pathology (17). Also, though the total BAL counts


576


Figure 3 Characteristics of the non-fatal pneumonia caused by the HKx31 (H3N2) influenza A virus. The B6 mice were either immunologically naive (primary) or had been infected i.p. with the PR8 (H1N1) virus more than 1 month previously (secondary). The inflammatory cell populations were recovered by BAL, and the numbers of CD8` and CD8`NPP` T cells were calculated from the total cell counts and the percent determined by two-color FACS analysis. The virus titers (k) were determined for lung homogenates. This figure is reproduced from Reference 6. An earlier publication (14) shows comparable results expressed as percentage (rather than total) CD8`NPP` T cells, with the secondary challenge data being for mice primed with the PR8 virus 7 months previously. Any concern about the durability of CD8` T cell memory to a virus that does not persist should be removed by this study.

decline in the recovery phase, the percentrage of CD8`NPP` cells in the residual lymphocyte population remains fairly constant. The loss of lymphocytes from the site of pathology is not, therefore, obviously selective for virus-specfic T cells.


577


It is also worth noting that the total magnitude of any virus-induced inflammatory processes will generally be such that, even though a CD8`Tet` population may be greatly enriched in a site of pathology, the majority of the virus-specific CD8` T cells will still be located in the lymphoid tissue (5, 14, 75, 76). At the peak of the primary influenza-specific response, for example, the CD8`NPP` set comprises approximately 2% of the of the CD8` population in the spleen (3 x 105 in 6 x 106) compared with 15% of the inflammatory CD8` T cells recovered by BAL (1.0 x 105 in 6 x 105). Though there is no formal proof that the CD8`NPP` T cells in the lymphoid tissue have not circulated through the virus infected lung, it seems likely that most of the virus-specific CD8` T cells detected in the lymph nodes and spleen during the course of these local infections may never have encountered high doses of antigen. At least a proportion of this ‘‘less driven’’ lymphocyte population may be considered to give rise to the memory T cell pool.

MEASUREMENT OF VIRUS-SPECIFIC CD8` T CELL MEMORY Earlier LDA studies with mice established very clearly that the massive clonal expansion characteristic of the primary virus-specific CD8` T cell response leads invariably to the development of lifelong memory (3). The overall impression was that the size of the memory T cell pool reflects the magnitude of the initial, antigen-driven clonal burst (134). The virus-specific CD8` CTLp were generally found to persist at levels of 1:103–1:104 spleen or lymph node cells (3). The remarkable stability of these memory CD8`CTLp frequencies over time, both within and between the various virus models, raised the problematic question of how such set points might be determined in the context of lymphocyte homeostasis. The tetramer analysis seems to be confirming these themes in a broad sense but has provided us with much higher memory T cell numbers and is opening a variety of new questions (5, 13). The use of LDA subsequent to FACS sorting also indicated that the expression of various cell-surface markers might be used to probe the activation status of memory T cells. Though all CD8` memory T cells are consistently CD44hi, the profile of CD62L expression changes with time. While the majority (Sendai) or all (influenza) of the virus-specific CTLp set switches to an activated CD62Llo phenotype during the acute phase of the primary response, progressive reversion to the naive CD62Lhi form is apparent within 6 (Sendai) to 12 (influenza) months (105). Again, the much simpler flow cytometric analysis with tetramers is confirming these profiles.

‘‘Memory’’ in Persistent Infections It is debatable whether the virus-specific CD8`Tet` lymphocyte populations that are maintained in the long term subsequent to primary infection with a persistent virus should be described as memory T cells. The consequences of continuous,


578


or sporadic, restimulation with even a low dose of an inducing peptide are likely to be very different from the effects of short-term exposure to a single (though substantial) pulse of antigen. This nomenclature problem is not resolved and, for the purpose of the present discussion, both situations are described as T cell memory. A consistent feature following experimental infection with viruses that are (SIV, cHV-68) or are not (influenza, Sendai, some LCMV models) maintained in the lymphoid compartment is that CD8`Tet` T cell numbers are maximal during the primary response, then decline to a much lower level thereafter (5, 13, 14, 18). Once CD8` T cell memory is established, the findings with SIV and cHV68 (15, 55, 76) are in accord with the long-term profiles for HIV, HTLV-1, and EBV (10, 16, 54, 55, 57–61). The general pattern is that these persistent infectious processes keep virus-specific T cell numbers at high, though not overwhelming (generally ,10%) levels. The rapid decrease in antigen load following effective antiviral therapy in HIV patients is associated with a progressive decline in the numbers of virus-specific CD8` T cells in the PBL compartment (58, 135). The best experimental evidence to date of a correlation between the intensity of continuing antigen stimulation and the size of the ‘‘memory’’ T cell pool is available for the cHV-68 model (75). The lack of IFN-c–producing CD4` effectors in MHC class II1/1 mice (117) persistently infected with cHV-68 leads to defective immune control and the reactivation of latent virus to lytic phase (74). This in turn results in the development of a progressive wasting disease and eventual (100–120 days) death, which is thought (though not proven) to be a consequence of virus-induced cytopathology. The prevalence of cHV-68–specific CD8`Tet` lymphocytes in the PBL and lymphoid compartments of these MHC class II1/1 mice is consistently higher than for the MHC class II`/` controls, reflecting the difference in virus load. Even at the terminal stages of the disease process in CD4-deficient mice, the cHV-68–specific CD8` T cells are still able to secrete IFN-c following stimulation with high doses of peptide (75). The evidence of anergy found for a proportion of the CD8`Tet` set in melanoma patients (122, 136) and in mice with persistent LCMV infection has not been replicated for the cHV-68 model. Again, this may be a dose effect, with the levels of stimulation resulting from the reactivation of cHV-68 to lytic phase being minimal when compared with the antigen concentrations provided by a large tumor mass or persistent, but nonlytic, LCMV infection. Further dissection of the balance between functional memory, anergy, and immune exhaustion will continue to be a major focus in studies of persistent infectious processes (particularly HIV) and tumor immunity.

Memory to Viruses that Do Not Persist With the exception of Sendai virus and LCMV (13, 18), the percentage of CD8`Tet` cells detected in the PBL and lymphoid tissue following primary infection with viruses that do not persist, or where there is no possibility of natural



579

rechallenge (human influenza infection), is too low to add much to the information base developed from the LDA approach (14, 123, 124). Secondary stimulation with the influenza virus NPP epitope, however, causes massive expansion of the CD8`NPP` set (14). The high frequencies of CD8`NPP` T cells detected in the spleens of H1N1-primed mice following an H3 N2 challenge (Figure 2) are remarkable when it is realized that the H3N2 virus causes little (if any) productive infection outside the respiratory tract. It is reasonable to assume that antigenpresenting dendritic cells (25) travel via afferent lymph from the infected lung to the regional mediastinal lymph nodes (MLN). Do they also distribute via the blood to the spleen? Such massively increased frequencies for the CD8`NPP` population (Figure 2) were not predicted from comparable experiments analyzed previously by LDA. Also, the LDA profiles showed no indication of the steady, progressive decline in prevalence with time that is apparent for the CD8`NPP` population (6), an observation which calls into question earlier thinking about stable ‘‘set sizes’’ for memory T cells (3). Analysis using the BrdU–incorporation protocol (Figure 1) also indicates that these memory CD8`NPP` lymphocytes are very heterogeneous, with some turning over at a high rate while others show little evidence of further cell division (Figure 4). Our perceptions of CD8` T cell memory are in the process of evolving, but a number of very obvious experiments need be done. A key requirement is to develop a clearer understanding of the relationship between tetramer staining and function for CD8` memory T cells.

SECONDARY CD8` T CELL RESPONSES AND PROTECTION The massive increase in virus-specific CD8` T cell prevalence (6, 14) detected in lymphoid tissue (Figure 2) following secondary challenge with the same influenza virus epitope (NPP) is even more obvious in the site of pathology (Figure 3). The remarkable feature of these experiments was that the large numbers of virus-specific CD8`NPP` T cells (ultimately 70% of the inflammatory CD8` population) that were generated only shortened the duration of the infectious process in the lung by a matter of 2–3 days (Figure 3). This confirmed earlier studies showing that the recall of CD8` T cell memory could provide only partial resistance to reinfection (137, 138). The prevalence of the primed CD8`NPP` set in the MLN and spleen of the H1N1-primed mice was ,1.0 % (of CD8` T cells) immediately prior to the H3N2 challenge. The lag in virus control apparently resulted from the fact that the memory T cells needed first to be restimulated in the lymphoid tissue before they could traffic to the virus-infected lung. Current experiments are asking whether increasing the number of influenza–specific memory T cells will lead to a more rapid response, but this analysis is still incomplete.


580


Figure 4 The extent of continuing lymphocyte turnover for CD8`NPP` memory T cell populations in the MLN and spleen. The PR8-primed B6 mice were fed BrdU in the drinking water for 8 days after secondary challenge with the HKx31 influenza A virus (see legend to Figure 1). The BrdU staining profiles for acutely stimulated (day 8) and memory (day 73) splenic T cells are shown in A, while the persistence of the BrdUhi set is illustrated in B. The numbers of virus-specific CD8` T cells decline over this interval (Figure 2), but those that remain in the MLN (in particular) seem to be cycling little, if at all. The apparent stability of this population provides further evidence for the lack of antigen persistence in the mouse influenza model. The figure is reproduced from Reference 6. Later experiments (J Christensen & J Riberdy, unpublished data) have shown that at least a proportion of these BrdUhi memory T cells can be driven back into cell cycle subsequent to further virus challenge.

Other studies with cHV-68 do suggest that the prevalence of virus-specific CD8` T cells may indeed determine the rapidity and effectiveness of the recall response. Mice were primed (139) singly or doubly with recombinant vaccinia (vacc) or influenza A (flu) viruses expressing a lytic phase epitope (p56). The respiratory phase of cHV-68 infection was more rapidly controlled following prior exposure to one or another of these viruses. However, the protective effect was much greater for mice that were given vacc-p56, rested, then infected with flup56 and rested again. The massively enhanced numbers of p56-specific CD8` T cells in these flu-p56r vacc-p56–primed mice provided almost total protection against the development of productive cHV-68 infection in the lung. Somewhat surprisingly, however, the extent of cHV-68 latency in the spleen was equivalent for doubly-immunized and naı¨ve mice within 3 weeks of respiratory cHV-68 challenge. Also, there was little effect on the magnitude of the


581

IM-like disease that develops subsequent to the control of the lytic infectious process in the lung. Perhaps the situation might be improved if the vaccine also incorporated a peptide expressed during latency, but no major latent phase epitope has yet been identified for cHV-68. It remains open whether primed (but ‘‘resting’’) CD8` T cell memory can indeed provide useful protection against any virus that has the potential to maintain persistent, or latent, infections of lymphocytes in the face of an effective CD8` T cell response.


QUANTITATION OF VIRUS-SPECIFIC CD4` T CELL RESPONSES AND MEMORY The tetramer approach (12) has yet to be utilized for the in vivo analysis of virusspecific CD4` T cell responses. Apart from the technical difficulties associated with making the necessary reagents, there is concern that the prevalence of lymphocytes specific for any one viral peptide may be too low to allow detection in freshly isolated lymphocyte populations recovered during the course of an initial encounter with antigen. The difficulty could presumably be overcome using secondary challenge protocols that focus the response to a particular epitope, but this would not tell us what happens in a normal primary response. The general impressions from IFN-c ELISPOT and LDA experiments that read out IL-2 production are consistent (110–113). Though substantial clonal expansion occurs in the virus-specific CD4` set prior to clearance of the inducing antigen, both the overall magnitude and the duration of the proliferative phase are much more limited than for the responding CD8` population (13, 14, 76). This should translate into lower numbers of CD4` T cells and could explain why these lymphocytes tend to be three- to fourfold less prevalent than the CD8` set in sites of virus-induced pathology. Surprisingly little attention has been given to the recall of virus-specific CD4` T cell memory following secondary challenge although, again, there is evidence of lymphocyte proliferation (29, 112). The longterm consequences of such restimulation for the size of CD4` T cell memory pool have not been analyzed. At least in the influenza model, the virus-specific CD4` T cell response seems to proceed reasonably normally in the absence of either B cells or antibody (Ig1/1 lMT mice), while the resultant memory populations are maintained in the long-term after recovery (140). A difference for the B cell–deficient lMT mice (which have very small spleens) is that the memory CD4` T cells seem to localize principally to the lymph nodes. The long-term concentration of the virus-specific CD4` set in the spleens of normal, Ig`/` mice may be due to the lack of the lymph node homing receptor (CD62L) and the consequent inability to enter lymph nodes via the high endothelial venules (141). Perhaps CD8` memory T cells are more likely to migrate through solid tissues and enter the nodes via the afferent lymphatics.


582


Sendai-specific CD4` memory T cells progressively reacquire CD62L expression with time (110), although this switch back to a ‘‘naive’’ CD62Lhi phenotype is much less apparent (113) for the cHV-68 model (Figure 5). The difference presumably reflects that Sendai virus is completely eliminated, while cHV-68 persists. Even so, the reactivation of latent cHV-68 to express lytic phase epitopes is probably a relatively low-frequency event (74). The numbers of memory CD4` T cells specific for cHV-68 also tend to be maintained at a fairly constant level (Figure 5), while the prevalence of the Sendai-specific CD4` set falls, then increases again in old mice (110). These late profiles for Sendai virus suggest enrichment subsequent to the progressive loss of thymic function and the diminution of the naive CD4` T cell compartment. A similar concentration effect has not been noted for virus-specific CD8` T cell memory. The overall impression is that there are substantial differences in both the acute and memory phases of virus-specific CD4` and CD8` T cell responses. Until, however, it is possible to stain antigen-specific CD4` T cells for direct FACS analysis, our understanding of the in vivo biology of these lymphocytes must lag behind the experiments with the CD8` set. The analysis is worth pursuing, especially for the large DNA viruses where CD4` T cell–mediated effector mechanisms seem to be of major importance. In addition, CD40 ligand–mediated CD4`

Figure 5 The acute and long-term cHV-68–specific CD4` T cell response. Frequencies expressed as reciprocals were determined for enriched CD4` T cell populations using a 72-h IFN-c ELISPOT assay. Some of the lymphocytes analyzed at later time points were sorted into CD4`CD62Lhi and CD4`CD62Llo sets prior to further stimulation and analysis. Cumulated from data presented in Reference 113.


583

T help is essential for the antibody response that constitutes the main mechanism for resistance to reinfection with most viruses.


CONCLUSIONS The tetramers have opened a new era in the quantitative and functional analysis of CMI to viruses. The first experimental studies were published in 1998. As a consequence, much that has been done so far has concentrated on the obvious and the accessible. Most of the discussion has been reactive to earlier conceptual frameworks, and many investigators have not yet had the opportunity to probe their own ideas and model systems with these reagents. A very real necessity is to make the tetramers broadly available as soon as possible, a process that is in hand (at least in the USA) with the NIH-funded program led by John Altman at Emory University. At this stage, the perceptions of the broader immunology community may be somewhat too focused on the very high T cell frequencies reported from tetramer studies with persistent infections, such as HIV, HTLV-1, EBV, SIV, and cHV-68, or from experiments with murine LCMV (10, 13, 15, 16, 57–61). The levels of 5–10% or so of virus-specific CD8` T cells in the PBL of persistently infected individuals may be surprising in the light of earlier LDA studies, but are they so extraordinary for T cells that may be continually driven by antigen? Such people (or experimental animals) still have .90% of their CD8` T cell repertoire available to deal with other pathogens. Analyzing the lymphoid tissue of mice infected with LCMV illustrates one extreme of a spectrum. The lymph nodes may be considered as sites of simultaneous response and pathology (43, 89), a conjunction that does not occur with more localized infections. The nonlytic nature of LCMV in the mouse (46) leads to a pathogenesis that may have no real equivalent in any human disease. If we look at the mouse infections caused by primary challenge with the influenza A viruses, TMV and MHV, we find that ,5.0% of the CD8` T cells in the lymph nodes and spleen are involved at the peak of the response, with these values decreasing to levels that are barely detectable (,0.5%) in a matter of weeks (6, 14, 123, 124). Primary infection with the intracellular bacterium L. monocytogenes gives similar results (142). Secondary challenge with the influenza A viruses does provoke enormous clonal expansion (.20% of the CD8` set) in the spleen, but the prevalence falls to 5–10% or so within a month or two (6, 14). The intent of this review has thus been to sketch a broad and fragmentary outline rather than to attempt what would be a very premature synthesis. Many more experimental systems are in the process of being analyzed, and new applications are in progress. Results can be generated very rapidly when the readout is three- or four-color FACS analysis. If this laboratory is typical, all those working with the tetramers are in the process of generating a mass of new data, much of which (at least in our case) needs to be more complete before it is inflicted on

584


the field at large. The possibilities are enormous but, to paraphrase Winston Churchill, we are not yet even at the end of the beginning and far from the beginning of the end. What is clear is that our capacity to access at least one aspect of immune complexity has been enhanced and is taking us into new territory. The analysis of phenotypic and functional diversity for CD8` effector and memory T cells has been greatly enhanced. We have, for the first time, some real numbers for the antigen-specific CD8` set. Developing a better understanding of rate effects and the T cell diaspora is, however, likely to be a continuing challenge for the ‘‘dynamicists.’’


ACKNOWLEDGMENTS The experiments with the tetramers would not have been possible without the support of Dr. John Altman and Dr. Rafi Ahmed, both of Emory University, Atlanta. Recombinant influenza A viruses were made by Dr. Maria Castrucci, of the Istituto Superiore di Sanita, Rome. Recent data from this laboratory that is discussed here is from published experiments done by Drs. Gabrielle Belz, James Brooks, Rhonda Cardin, Jan Christensen, Kirsten Flynn, Anne Marie HamiltonEaston, Janice Riberdy, Philip Stevenson, and David Topham. We thank Dr. David Woodland, Dr. Marcia Blackman, Dr. Julia Hurwitz, and Dr. Chris Coleclough for advice and discussion. Vicki Henderson helped with the preparation of the manuscript. These studies were funded by US Public Health Service Grants AI29579, AI38359, and CA 21765, and by the American Syrian Lebanese Associated Charities. J.P.C. is the recipient of a fellowship from the Alfred Benson Foundation, Denmark. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Sprent J, Tough DF. 1994. Lymphocyte life-span and memory. Science 265:1395–400 2. Tanchot C, Rocha B. 1995. The peripheral T cell repertoire: independent thomeostatic regulation of virgin and activated CD8` T cell pools. Eur. J. Immunol. 25:2127–36 3. Doherty PC, Topham DJ, Tripp RA. 1996. Establishment and persistence of virus-specific CD4` and CD8` T cell memory. Immunol. Rev. 150:23–44 4. Doherty PC, Hamilton-Easton AM,

Topham DJ, Riberdy J, Brooks JW, Cardin RD. 1997. Consequences of viral infections for lymphocyte compartmentalization and homeostasis. Semin. Immunol. 9:365–73 5. Tough DF, Sprent J. 1994. Turnover of naive-and memory-phenotype T cells. J. Exp. Med. 179:1127–35 6. Flynn KJ, Riberdy JM, Christensen JP, Altman JD, Doherty PC. 1999. In vivo proliferation of naı¨ve and memory influenza-specific CD8` T cells. Proc. Natl. Acad. Sci. USA 96:8597–602


DYNAMICS OF T CELL RESPONSES 7. De Boer RJ, Perelson AS. 1997. Competitive control of the self-renewing T cell repertoire. Int. Immunol. 9:779–90 8. Antia R, Pilyugin SS, Ahmed R. 1998. Models of immune memory: on the role of cross-reactive stimulation, competition, and homeostasis in maintaining immune memory. Proc. Natl. Acad. Sci. USA 95:14926–31 9. Nowak MA, Bangham CR. 1996. Population dynamics of immune responses to persistent viruses. Science 272:74–79 10. Altman JD, Moss PAH, Goulder PJR, Barouch DH, McHeyzer-Williams MG, Bell JI, McMichael AJ, Davis MM. 1996. Phenotypic analysis of antigenspecific T lymphocytes. Science 274:94– 96 11. Greten TF, Slansky JE, Kubota R, Soldan SS, Jaffee EM, Leist TP, Pardoll DM, Jacobson S, Schneck JP. 1998. Direct visualization of antigen-specific T cells: HTLV-1 Tax11-19-specific CD8` T cells are activated in peripheral blood and accumulate in cerebrospinal fluid from HAM/TSP patients. Proc. Natl. Acad. Sci. USA 95:7568–73 12. Crawford F, Kozono H, White J, Marrack P, Kappler J. 1998. Detection of antigenspecific T cells with multivalent soluble class II MHC covalent peptide complexes. Immunity 8:675–82 13. Murali-Krishna K, Altman JD, Suresh M, Sourdive DJ, Zajac AJ, Miller JD, Slansky J, Ahmed R. 1998. Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 8:177–87 14. Flynn KJ, Belz GT, Altman JD, Ahmed R, Woodland DL, Doherty PC. 1998. Virus-specific CD8` T cells in primary and secondary influenza pneumonia. Immunity 8:683–91 15. Kuroda MJ, Schmitz JE, Barouch DH, Craiu A, Allen TM, Sette A, Watkins DI, Forman MA, Letvin NL. 1998. Analysis of Gag-specific cytotoxic T lymphocytes in simian immunodeficiency virus-

16.

17.

18.

19.

20.

21.

22.

23.

585

infected rhesus monkeys by cell staining with a tetrameric major histocompatibility complex class I-peptide complex. J. Exp. Med. 187:1373–81 Callan MF, Tan L, Annels N, Ogg GS, Wilson JD, O’Callaghan CA, Steven N, McMichael AJ, Rickinson AB. 1998. Direct visualization of antigen-specific CD8` T cells during the primary immune response to Epstein-Barr virus in vivo. J. Exp. Med. 187:1395–402 Belz GT, Altman JD, Doherty PC. 1998. Characteristics of virus-specific CD8(`) T cells in the liver during the control and resolution phases of influenza pneumonia. Proc. Natl. Acad. Sci. USA 95:13812–17 Usherwood EJ, Hogan RJ, Crowther G, Surman SL, Hogg TL, Altman JD, Woodland DL. 1999. Functionally heterogeneous CD8` T-cell memory is induced by Sendai virus infection of mice. J. Virol. 73:7278–86 Walker JA, Sakaguchi T, Matsuda Y, Yoshida T, Kawaoka Y. 1992. Location and character of the cellular enzyme that cleaves the hemagglutinin of a virulent avian influenza virus. Virology 190:278– 87 Tashiro M, Seto JT, Choosakul S, Hegemann H, Klenk HD, Rott R. 1992. Changes in specific cleavability of the Sendai virus fusion protein: implications for pathogenicity in mice. J. Gen. Virol. 73:1575–79 Horimoto T, Kawaoka Y. 1995. The hemagglutinin cleavability of a virulent avian influenza virus by subtilisin-like endoproteases is influenced by the amino acid immediately downstream of the cleavage site. Virology 210:466–70 McWilliam AS, Napoli S, Marsh AM, Pemper FL, Nelson DJ, Pimm CL, Stumbles PA, Wells TNC, Holt PG. 1996. Dendritic cells are recruited into the airway epithelium during the inflammatory response to a broad spectrum of stimuli. J. Exp. Med. 184:2429–32 McWilliam AS, Marsh AM, Holt PG.

586

24.

25.


26.

27.

28.

29.

30.

31.

32.

33.

DOHERTY n CHRISTENSEN 1997. Inflammatory infiltration of the upper airway epithelium during Sendai virus infection: involvement of epithelial dendritic cells. J. Virol. 71:226–36 Hamilton-Easton AM, Eichelberger M. 1995. Virus-specific antigen presentation by different subsets of cells from lung and mediastinal lymph node tissues of influenza virus-infected mice. J. Virol. 69:6359–66 Usherwood EJ, Hogg TL, Woodland DL. 1999. Enumeration of antigen-presenting cells in mice infected with Sendai virus. J. Immunol. 162:3350–55 Bujdoso R, Hopkins J, Dutia BM, Young P, McConnell I. 1989. Characterization of sheep afferent lymph dendritic cells and their role in antigen carriage. J. Exp. Med. 170:1285–301 Eichelberger MC, Wang M, Allan W, Webster RG, Doherty PC. 1991. Influenza virus RNA in the lung and lymphoid tissue of immunologically intact and CD4-depleted mice. J. Gen. Virol. 72:1695–98 Kawaoka Y. 1991. Equine H7N7 influenza A viruses are highly pathogenic in mice without adaptation: potential use as an animal model. J. Virol. 65:3891–94 Riberdy JM, Flynn KJ, Stech J, Webster RG, Altman JD, Doherty PC. 1999. Protection against a lethal avian influenza A virus in a mammalian system. J. Virol. 73:1453–59 Webster RG. 1997. Influenza virus: transmission between species and relevance to emergence of the next human pandemic. Arch. Virol. 142:105–13 Yewdell JW, Bennink JR. 1992. Cell biology of antigen processing and presentation to major histocompatibility complex class I molecule-restricted T lymphocytes. Adv. Immunol. 52:1–123 Townsend AR, Gotch FM, Davey J. 1985. Cytotoxic T cells recognize fragments of the influenza nucleoprotein. Cell 42:457–67 Doherty PC, Effros RB, Bennink J. 1977.

34.

35.

36.

37.

38.

39.

40.

41.

Heterogeneity of the cytotoxic response of thymus-derived lymphocytes after immunization with influenza viruses. Proc. Natl. Acad. Sci. USA 74:1209–13 Effros RB, Bennink J, Doherty PC. 1978. Characteristics of secondary cytotoxic Tcell responses in mice infected with influenza A viruses. Cell Immunol. 36:345–53 Patick AK, Thiemann RL, O’Brien PC, Rodriguez N. 1991. Persistence of Theiler’s virus infection following promotion of central nervous system remyelination. J. Neuropathol. Exp. Neurol. 50:523–37 Gallagher TM, Buchmeier MJ, Perlman. 1992. Cell receptor-independent infection by a neurotropic murine coronavirus. Virology 191:517–22 Ruttkay-Nedecka S, Rajcani J, Eleckova E, Ruttkay-Nedecky G. 1987. Infection of the central nervous system of mice by standard Sendai virus, defective interfering Sendai virus and the mixture of both: comparison of virus multiplication and pathogenicity. Acta Virol. 31:78–82 Cattaneo R, Schmid A, Rebmann G, Baczko K, Ter Meulen V, Bellini WJ, Rozenblatt S, Billeter MA. 1986. Accumulated measles virus mutations in a case of subacute sclerosing panencephalitis: interrupted matrix protein reading frame and transcription alteration. Virology 154:97–107 Dhib-Jalbut S, McFarland HF, Mingioli ES, Sever JL, McFarlin DE. 1988. Humoral and cellular immune responses to matrix protein of measles virus in subacute sclerosing panencephalitis. J. Virol. 62:2483–89 Hotchin J. 1974. The role of transient infection in arenavirus persistence. Prog. Med. Virol. 18:81–93 Gilden DH, Cole GA, Nathanson N. 1972. Immunopathogenesis of acute central nervous system disease produced by lymphocytic choriomeningitis virus. II.


42.

43.


44.

45.

46.

47.

48.

49.

50.

51.

Adoptive immunization of virus carriers. J. Exp. Med. 135:874–89 Marker O, Volkert M. 1973. Studies on cell-mediated immunity to lymphocytic choriomeningitis virus in mice. J. Exp. Med. 137:1511–25 Zinkernagel RM, Doherty PC. 1979. MHC-restricted cytotoxic T cells: studies on the biological role of polymorphic major transplantation antigens determining T-cell restriction-specificity, function, and responsiveness. Adv. Immunol. 27:51–177 Moskophidis D, Assmann-Wischer U, Simon MM, Lehmann-Brube F. 1987. The immune response of the mouse to lymphocytic choriomeningitis virus. V. High numbers of cytolytic T lymphocytes are generated in the spleen during acute infection. Eur. J. Immunol. 17:937– 42 Ahmed R, Gary D. 1996. Immunological memory and protective immunity: Understanding their relation. Science 272:54–60 Buchmeier MJ, Welsh RM, Dutko FJ, Oldstone MB. 1980. The virology and immunobiology of lymphocytic choriomeningitis virus infection. Adv. Immunol. 30:275–331 Lau LL, Jamieson BD, Somasundaram T, Ahmed R. 1994. Cytotoxic T-cell memory without antigen. Nature 369:648–52 Klenerman P, Hengartner H, Zinkernagel RM. 1997. A non-retroviral RNA virus persists in DNA form. Nature 390:298– 301 Ka¨gi D, Seiler P, Pavlovic J, Ledermann B, Bu¨rki K, Zinkernagel RM, Hengartner H. 1995. The roles of perforin- and Fasdependent cytotoxicity in protection against cytopathic and noncytopathic viruses. Eur. J. Immunol. 25:3256–62 Suzuki S, Hotchin J. 1971. Initiation of persistent lymphocytic choriomeningitis infection in adult mice. J. Infect. Dis. 123:603–10 Moskophidis D, Lechner F, Pircher H,

52.

53.

54.

55.

56.

57.

58.

587

Zinkernagel RM. 1993. Virus persistence in acutely infected immunocompetent mice by exhaustion of antiviral cytotoxic effector T cells. Nature 362:758–61 Brodie SJ, Lewinsohn DA, Patterson BK, Jiyamapa D, Krieger J, Corey L, Greenberg PD, Riddell SR. 1999. In vivo migration and function of transferred HIV-1-specific cytotoxic T cells. Nat. Med. 5:34–41 Tan R, Xu X, Ogg GS, Hansasuta P, Dong T, Rostron T, Luzzi G, Conlon CP, Screaton GR, McMichael AJ, RowlandJones S. 1999. Rapid death of adoptively transferred T cells in acquired immunodeficiency syndrome. Blood 93:1506–10 Hay CM, Ruhl DJ, Basgoz NO, Wilson CC, Billingsley JM, DePasquale MP, D’Aquila RT, Wolinsky SM, Crawford JM, Montefiori DC, Walker BD. 1999. Lack of viral escape and defective in vivo activation of human immunodeficiency virus type 1-specific cytotoxic T lymphocytes in rapidly progressive infection. J. Virol. 73:5509–19 Schmitz JE, Kuroda MJ, Santra S, Sasseville VG, Simon MA, Lifton MA, Racz P, Tenner-Racz K, Dalesandro M, Scallon BJ, Ghrayeb J, Forman MA, Montefiori DC, Rieber EP, Letvin NL, Reimann KA. 1999. Control of viremia in simian immunodeficiency virus infection by CD8` lymphocytes. Science 283:857–60 Perelson AS, Neumann AU, Markowitz M, Leonard JM, Ho DD. 1996. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science 271:1582–86 Ogg GS, Jin X, Bonhoeffer S, Dunbar PR, Nowak MA, Monard S, Segal JP, Cao Y, Rowland-Jones SL, Cerundolo V, Hurley A, Markowitz M, Ho DD, Nixon DF, McMichael AJ. 1998. Quantitation of HIV-1-specific cytotoxic T lymphocytes and plasma load of viral RNA. Science 279:2103–6 Ogg GS, Jin X, Bonhoeffer S, Moss P,

588


59.

60.

61.

62.

63.

64.

65.

66.

DOHERTY n CHRISTENSEN Nowak MA, Monard S, Segal JP, Cao Y, Rowland-Jones SL, Hurley A, Markowitz M, Ho DD, McMichael AJ, Nixon DF. 1999. Decay kinetics of human immunodeficiency virus-specific effector cytotoxic T lymphocytes after combination antiretroviral therapy. J. Virol. 73:797–800 Jeffery KJ, Usuku K, Hall SE, Matsumoto W, Taylor GP, Procter J, Bunce, M Oggs GS, Welsh KI, Weber JN, Lloyd LA, Nowak MA, Nagai M, Kodama D, Izumo S, Osame M, Bangham CR. 1999. HLA alleles determine human T-lymphotropic virus-I (HTLV-I) proviral load and the risk of HTLV-I-associated myelopathy. Proc. Natl. Acad. Sci. USA 96:3848– 53 Wodarz D, Nowak MA, Bangham CR. 1999. The dynamics of HTLV-I and the CTL response. Immunol. Today 20:220– 27 Bieganowska K, Hollsberg P, Buckle GJ, Lim DG, Greten TF, Schneck J, Altman JD, Jacobson S, Ledis SL, Hanchard B, Chin J, Morgan O, Roth PA, Hafler DA. 1999. Direct analysis of viral-specific CD8` T cells with soluble HLA- A2/ Tax11-19 tetramer complexes in patients with human T cell lymphotropic virusassociated myelopathy. J. Immunol. 162:1765–71 Klein G. 1994. Epstein-Barr virus strategy in normal and neoplastic B cells. Cell 77 7:791–93 Nash AA, Sunil-Chandra NP. 1994. Interactions of the murine gammaherpesvirus with the immune system. Curr. Opin. Immunol. 6:560–63 Rickinson AB, Moss DJ. 1997. Human cytotoxic T lymphocyte responses to Epstein-Barr virus infection. Annu. Rev. Immunol. 15:405–31 Spriggs MK. 1996. One step ahead of the game: viral immunomodulatory molecules. Annu. Rev. Immunol. 14:101–30 Strockbine LD, Cohen JI, Farrah T, Lyman SD, Wagener F, DuBose RF,

67.

68.

69.

70.

71.

72.

73.

74.

75.

Armitage RJ, Spriggs MK. 1998. The Epstein-Barr virus BARF1 gene encodes a novel, soluble colony-stimulating factor-1 receptor. J. Virol. 72:4015–21 Virgin HW, Latreille P, Wamsley P, Hallsworth K, Weck KE, Dal CA, Speck SH. 1997. Complete sequence and genomic analysis of murine gammaherpesvirus 68. J. Virol. 71:5894–904 Ganem D. 1998. Human herpesirus 8 and its role in the genesis of Kaposi’s sarcoma. Curr. Clin. Top. Infect. Dis. 18:327–51 Orenstein JM, Alkan S, Blauvelt A, Jeang KT, Weinstein MD, Ganem D, Herndier B. 1997. Visualization of human herpesvirus type 8 in Kaposi’s sarcoma by light and transmission electron microscopy. AIDS 11:F35–45 Osman M, Kubo T, Gill J, Neipel F, Backer M, Smith G, Weiss R, Gazzard B, Boshoff C, Gotch F. 1999. Identification of human herpesvirus 8-specific cytotoxic T-cell responses. J. Virol. 73:6136–40 Young L, Alfieri C, Hennessy K, Evans H, O’Hara C, Anderson KC, Ritz J, Shapiro RS, Rickinson Kieff E, Cohen JI. 1989. Expression of Epstein-Barr virus transformation-associated genes in tissues of patients with EBV lymphoproliferative disease. N. Engl. J. Med. 321:1080–85 Sunil-Chandra NP, Efstathiou S, Nash AA. 1993. Interactions of murine gammaherpesvirus 68 with B and T cell lines. Virology 193:825–33 Walling DM, Raab-Traub N. 1994. Epstein-Barr virus intrastrain recombination in oral hairy leukoplakia. J. Virol. 68:7909–17 Cardin RD, Brooks JW, Sarawar SR, Doherty PC. 1996. Progressive loss of CD8` T cell-mediated control of a gamma-herpesvirus in the absence of CD4` T cells. J. Exp. Med. 184:863–71 Stevenson PG, Belz GT, Altman JD, Doherty PC. 1998. Virus-specific CD8`


76.

77.


78.

79.

80.

81.

82.

83.

T cell numbers are maintained during gamma-herpesvirus reactivation in CD4deficient mice. Proc. Natl. Acad. Sci. USA 95:15565–70 Doherty PC, Tripp RA, Hamilton-Easton AM, Cardin RD, Woodland DL, Blackman MA. 1997. Tuning into immunological dissonance: an experimental model for infectious mononucleosis. Curr. Opin. Immunol. 9:477–83 Henle G, Henle W, Diehl V. 1968. Relation of Burkitt’s tumor-associated herpes-type virus to infectious mononucleosis. Proc. Natl. Acad. Sci. USA 59:94– 101 Tripp RA, Hamilton-Easton AM, Cardin RD, Nguyen P, Behm FG, Woodland DL, Doherty PC, Blackman MA. 1997. Pathogenesis of an infectious mononucleosis-like disease induced by a murine gamma-herpesvirus: role for a viral superantigen? J. Exp. Med. 185:1641–50 Hardy CL, Silins SL, Woodland DL, Blackman MA. 1999. Murine gammaherpesvirus infection causes Vb4-specific CDR3-restricted clonal expansions within CD8` peripheral blood T lymphocytes. Submitted for publication Stevenson PG, Belz GT, Altman JD, Doherty PC. 1999. Changing patterns of dominance in the CD8` T cell response during acute and persistent murine gamma-herpesvirus infection. Eur. J. Immunol. 29:1059–67 Brooks JW, Hamilton-Easton AM, Christensen JP, Cardin RD, Hardy CL, Doherty PC. 1999. Requirement for CD40 ligand, CD4` T cells and B cells in an infectious mononucleosis-like syndrome. J. Virol. 73:9650–54 Flano E, Woodland DL, Blackman MA. 1999. Requirement for CD4` T cells in Vb4` CD8` T cell activation associated with latent murine gammaherpesvirus infection. J. Immunol. 163:3403–8 Coppola MA, Flano E, Nguyen P, Hardy CL, Cardin RD, Shastri N, Woodland DL, Blackman MA. 1999. Apparent

84.

85.

86.

87. 88.

89.

90.

91.

92.

93.

94. 95.

96.

97.

589

MHC-independent stimulation of CD8` T cells in vivo during latent murine gammaherpesvirus infection. J. Immunol. 163:1481–89 Miller JFAP. 1999. Discovering the origins of immunological competence. Annu. Rev. Immunol. 17:1–17 Tanchot C, Lemonnier FA, Pe´rarnau B, Freitas AA, Rocha B. 1997. Differential requirements for survival and proliferation of CD8 naive or memory T cells. Science 276:2057–62 Ford WL, Burr W, Simonsen M. 1970. A lymph node weight assay for the graftversus-host activity of rat lymphoid cells. Transplantation 10:258–66 Mackaness GB, Blanden RV. 1967. Cellular immunity. Prog. Allergy 11:89–40 Blanden RV. 1974. T cell response to viral and bacterial infection. Transplant. Rev. 19:56–88 Doherty PC, Zinkernagel RM. 1974. Tcell-mediated immunopathology in viral infections. Transplant. Rev. 19:89–120 Burnet FM, Burnet D. 1960. Graft versus host reactions on the chorioallantoic membrane of the chick embryo. Nature 188:376–79 Sexton C. 1991. The Seeds of Time: The Life of Sir Macfarlane Burnet. Melbourne: Oxford Univ. Press Doherty PC. 1999. Burnet Oration: living in the Burnet lineage. Immunol. Cell Biol. 77:167–76 Burnet FM. 1967. The impact of ideas on immunology. Cold Spring Harbor Symp. Quant. Biol. 32:1–8 Talmage DW. 1957. Allergy and immunology. Annu. Rev. Med. 8:239–56 Cerottini JC, Brunner KT. 1974. Cellmediated cytotoxicity, allograft rejection, and tumor immunity. Adv. Immunol. 18:67–132 Zinkernagel RM, Doherty PC. 1997. The discovery of MHC restriction. Immunol. Today 18:14–17 Doherty PC, Zinkernagel RM. 1975. A biological role for the major histocompatibility antigens. Lancet 1:1406–9


590


98. Brunner KT, Engers HD, Cerottini JC. 1976. The 51Cr release assay as used for the quantitative measurement of cellmediated cytolysis in vitro. In Vitro Methods Cell-Mediated Tumor Immun. 1:423–28 99. Ryser JE, Sordat B, Cerottini JC, Brunner KT. 1977. Mechanism of target cell lysis of cytolytic T lymphocytes. I. Characterization of specific lymphocytetarget cell conjugates separated by velocity sedimentation. Eur. J. Immunol. 7:110–17 100. Schick B, Berke G. 1979. Competitive inhibition of cytotoxic T lymphocyte-target cell conjugation. A direct evaluation of membrane antigens involved in cellmediated immunity. Transplantation 27:365–68 101. Owen JA, Allouche M, Doherty PC. 1982. Limiting dilution analysis of the specificity of influenza-immune cytotoxic T cells. Cell Immunol. 67:49–59 102. Maier B, Buhring HJ, Simon M, Eichmann K, Melchers I. 1986. Limiting dilution analysis of proliferating and helper T cells in the in vivo immune response to KLH: depression of helper T cells at moderately increased frequencies. J. Mol. Cell Immunol. 2:293–305 103. Tabi Z, Lynch F, Ceredig R, Allan JE, Doherty PC. 1988. Virus-specific memory T cells are Pgp-1` and can be selectively activated with phorbol ester and calcium ionophore. Cell Immunol. 113:268–77 104. Hou S, Doherty PC. 1993. Partitioning of responder CD8` T cells in lymph node and lung of mice with Sendai virus pneumonia by LECAM-1 and CD45RB phenotype. J. Immunol. 150:5494–500 105. Tripp RA, Hou S, Doherty PC. 1995. Temporal loss of the activated L-selectinlow phenotype for virus-specific CD8` memory T cells. J Immunol. 154:5870– 75 106. Doherty PC, Korngold R. 1983. Characteristics of poxvirus-induced meningi-

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

tis: virus-specific and non-specific cytotoxic effectors in the inflammatory exudate. Scand. J. Immunol. 18:1–7 Tripp RA, Hou S, McMickle A, Houston J, Doherty PC. 1995. Recruitment and proliferation of CD8` T cells in respiratory virus infections. J. Immunol. 154:6013–21 McMichael AJ, O’Callaghan CA. 1998. A new look at T cells. J. Exp. Med. 187:1367–71 Miller RA, Reiss CS. 1984. Limiting dilution cultures reveal latent influenza virus-specific helper T cells in virusprimed mice. J. Mol. Cell. Immunol. 1:357–68 Ewing C, Topham DJ, Doherty PC. 1995. Prevalence and activation phenotype of Sendai virus-specific CD4` T cells. Virology 210:179–85 Topham DJ, Doherty PC. 1998. Longitudinal analysis of the acute Sendai virus-specific CD4` T cell response and memory. J. Immunol. 161:4530–35 Whitmire JK, Asano MS, MuraliKrishna K, Suresh M, Ahmed R. 1998. Long-term CD4 Th1 and Th2 memory following acute lymphocytic choriomeningitis virus infection. J. Virol. 72:8281–88 Christensen JP, Doherty PC. 1999. Quantitative analysis of the acute and longterm CD4` T cell response to a persistent gammaherpesvirus. J. Virol. 73:4279–83 Butz EA, Bevan MJ. 1998. Massive expansion of antigen-specific CD8` T cells during an acute virus infection. Immunity 8:167–75 Niemialtowski MG, Rouse BT. 1992. Predominance of Th1 cells in ocular tissues during herpetic stromal keratitis. J. Immunol. 149:3035–39 Polic B, Hengel H, Krmpotic A, Trgovcich J, Pavic I, Luccaronin P, Jonjic S, Koszinowski UH. 1998. Hierarchical and redundant lymphocyte subset control


117.

118.


119.

120.

121.

122.

123.

124.

precludes cytomegalovirus replication during latent infection. J. Exp. Med. 188:1047–54 Christensen JP, Cardin RD, Branum KC, Doherty PC. 1999. CD4` T cell-mediated control of a c-herpesvirus in B celldeficient mice is mediated by IFN-c. Proc. Natl. Acad. Sci. USA 96:5135–40 Tan LC, Gudgeon N, Annels NE, Hansasuta P, O’Callaghan CA, RowlandJones S, McMichael AJ, Rickinson AB, Callan MF. 1999. A re-evaluation of the frequency of CD8` T cells specific for EBV in healthy virus carriers. J. Immunol. 162:1827–35 Stevenson PG, Belz GT, Altman JD, Doherty PC. 1999. Changing patterns of dominance in the CD8` T cell response during acute and persistent murine gamma-herpesvirus infection. Eur. J. Immunol. 29:1059–67 Sourdive DJ, Murali-Krishna K, Altman JD, Zajac AJ, Whitmire JK, Pannetier C, Kourilsky P, Evavold B, Sette A, Ahmed R. 1998. Conserved T cell receptor repertoire in primary and memory CD8 T cell responses to an acute viral infection. J. Exp. Med. 188:71–82 Jacob, J, Baltimore D. 1999. Modelling T-cell memory by genetic marking of memory T cells in vivo. Nature 399:593– 97 Yee C, Savage PA, Lee PP, Davis MM, Greenberg PD. 1999. Isolation of high avidity melanoma-reactive CTL from heterogeneous populations using peptide-MHC tetramers. J. Immunol. 162:2227–34 Bergmann CC, Altman JD, Hinton D, Stohlman SA. 1999. Inverted immunodominance and impaired cytoloytic function of CD8` T cells during viral persistence in the CNS. J. Immunol. 163:3379–87 Johnson AJ, Njenga MK, Kuhns ST, Chen L, Rodriguez M, Pease LR. 1999. Prevalent class I-restricted T-cell response to the Theiler’s virus epitope

125.

126.

127.

128.

129.

130.

131.

132.

133.

591

Db:VP2121-130 in the absence of endogenous CD4 help, tumor necrosis factor alpha, gamma interferon, perforin, or costimulation through CD28. J. Virol. 73:3702–8 Jordan HL, Kuroda MJ, Schmitz JE, Steenbeke T, Forman MA, Letvin NL. 1999. Detection of simian immunodeficiency virus Gag-specific CD8` T lymphocytes in semen of chronically infected rhesus monkeys by cell staining with a tetrameric major histocompatibility complex class I-peptide complex. J. Virol. 73:4508–12 Tripp RA, Lahti JM, Doherty PC. 1995. Laser light suicide of proliferating virusspecific CD8` T cells in an in vivo response. J. Immunol. 155:3719–21 Tough DF, Borrow P, Sprent J. 1996. Induction of bystander T cell prolifration by viruses and type I interferon in vivo. Science 272:1947–50 Zhang XH, Sun SQ, Hwang IK, Tough DF, Sprent J. 1998. Potent and selective stimulation of memory-phenotype CD8` T cells in vivo by IL-15. Immunity 8:591–99 Nahill SR, Welsh RM. 1993. High frequency of cross-reactive cytotoxic T lymphocytes elicited during the virusinduced polyclonal cytotoxic T lymphocyte response. J. Exp. Med. 177:317–27 Zarozinski CC, Welsh RM. 1997. Minimal bystander activation of CD8 T cells during the virus-induced polyclonal T cell response. J. Exp. Med. 185:1629–39 Vella AT, Dow S, Potter TA, Kappler J, Marrack P. 1998. Cytokine-induced survival of activated T cells in vitro and in vivo. Proc. Natl. Acad. Sci. USA 95:3810–15 Kuroda MJ, Schmitz JE, Charini WA, Nickerson CE, Lord CI, Forman MA, Letvin NL. 1999. Comparative analysis of cytotoxic T lymphocytes in lymph nodes and peripheral blood of simian immunodeficiency virus-infected rhesus monkeys. J. Virol. 73:1573–79 Cole GA, Hogg TL, Woodland DL.

592

134.


135.

136.

137.

DOHERTY n CHRISTENSEN 1994. The MHC class I-restricted T cell response to Sendai virus infection in C57BL/6 mice: a single immunodominant epitope elicits an extremely diverse repertoire of T cells. Int. Immunol. 6:1767–75 Hou S, Hyland L, Ryan KW, Portner A, Doherty PC. 1994. Virus-specific CD8` T-cell memory determined by clonal burst size. Nature 369:652–54 Gray CM, Lawrence J, Schapiro JM, Altman JD, Winters MA, Crompton M, Loi M, Kundu SK, Davis MM, Merigan TV. 1999. Frequency of class I HLArestricted anti-HIV CD8` T cells in individuals receiving highly active antiretroviral therapy (HAART). J. Immunol. 162:1780–88 Zajac AJ, Blattman JN, Murali-Krishna K, Sourdive DJ, Suresh M, Altman JD, Ahmed R. 1998. Viral immune evasion due to persistence of activated T cells without effector function. J. Exp. Med. 188:2205–13 Andrew ME, Coupar BE, Boyle DB, Ada GL. 1987. The roles of influenza virus

138.

139.

140.

141.

142.

haemagglutinin and nucleoprotein in protection: analysis using vaccinia virus recombinants. Scand. J. Immunol. 25: 21–28 Doherty PC, Allan W, Boyle DB, Coupar BE, Andrew ME. 1989. Recombinant vaccinia viruses and the development of immunization strategies using influenza virus. J. Infect Dis. 159:1119–22 Stevenson PG, Belz GT, Castrucci MR, Altman JD, Doherty PC. 1999. A c -herpesvirus sneaks through a CD8` T cell response primed to a lytic-phase epitope. Proc. Natl. Acad. Sci. USA 96:9281–86 Topham DJ, Tripp RA, Hamilton-Easton AM, Sarawar SR, Doherty PC. 1996. Quantitative analysis of the influenza virus-specific CD4` T cell memory in the absence of B cells and Ig. J. Immunol. 157:2947–52 Butcher EC Picker LJ. 1996. Lymphocyte homing and homeostasis. Science 272:60–66 Busch DH, Pilip IM, Vijh S, Pamer EG. 1998. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 8:353–62


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:593–620 Copyright q 2000 by Annual Review. All rights reserved

THE ROLE OF CHEMOKINE RECEPTORS IN PRIMARY, EFFECTOR, AND MEMORY IMMUNE RESPONSES


Federica Sallusto1, Charles R. Mackay2, and Antonio Lanzavecchia3 1 Basel Institute for Immunology, Grenzacherstrasse 487, CH-4005 Basel, Switzerland; e-mail: [email protected] 2 Garvan Institute of Medical Research, 384 Victoria St. Darlinghurst, NSW Australia; 2010; e-mail: [email protected] 3 Institute of Research in Biomedicine, Via Vela 6, CH-6500 Bellinzona, Switzerland; e-mail: [email protected]

Key Words chemokine, dendritic cell, Th1, Th2, memory T cell, immune response Abstract The immune system is composed of single cells, and its function is entirely dependent on the capacity of these cells to traffic, localize within tissues, and interact with each other in a precisely coordinated fashion. There is growing evidence that the large families of chemokines and chemokine receptors provide a flexible code for regulating cell traffic and positioning in both homeostatic and inflammatory conditions. The regulation of chemokine receptor expression during development and following cell activation explains the complex migratory pathways taken by dendritic cells, T and B lymphocytes, providing new insights into the mechanisms that control priming, effector function, and memory responses.

INTRODUCTION Immune responses require the timely interaction of multiple cell types within specific microenvironments. During primary responses, the rare antigen-specific T and B lymphocytes need to maximize the possibility of encounter with antigen and with each other. They do so by continuously recirculating through secondary lymphoid organs, where antigen is carried and displayed on antigen-presenting cells. In contrast, in the effector phase of the immune response, effector T cells must be able to enter any tissue where pathogens may be present, and the T cells need to interact with other leukocytes such as eosinophils, mast cells, and basophils (in the case of allergic reactions), or macrophages and neutrophils (in the case of DTH reactions). Finally, memory T cells provide immediate surveillance and protection in tissues against a secondary challenge but are also able to reach 0732–0582/00/0410–0593$14.00

593

594

SALLUSTO n MACKAY n LANZAVECCHIA

secondary lymphoid organs to undergo further clonal expansion. Thus, leukocyte traffic represents a key element in the regulation of the immune response. Recent evidence indicates that the selectivity and flexibility necessary to regulate cell traffic under homeostatic and inflammatory conditions is provided by a differential tissue distribution of chemokines and a regulated expression of chemokine receptors on different leukocyte subsets.


CHEMOKINES AND CHEMOKINE RECEPTORS IN HOMEOSTASIS AND INFLAMMATION Chemokines are chemotactic cytokines that signal through seven transmembrane receptors (7TMR) coupled to pertussis toxin–sensitive Gi-proteins (1–4). Chemokines and chemokine receptors are involved in the two distinct steps of leukocyte migration. The first is extravasation from the blood into lymph nodes, Peyer’s patches, and inflamed tissues. This step is determined by several sequentially acting receptor-ligand pairs (5, 6). Selectins and their carbohydrate ligands allow rolling of leukocytes on the vascular endothelium, leading to exposure of chemokine receptors to their ligands displayed on the surface of endothelial cells. The triggering of chemokine receptors results in the rapid activation of integrins leading to firm adhesion and transmigration. Thus, while the chemokine/ chemokine receptor interaction represents an essential component, the specificity of leukocyte extravasation rests on a combinatorial code resulting from the expression of three determinants: selectins, chemokine receptors, and integrins. The second step controlled by chemokines and chemokine receptors is the directional migration and positioning of leukocytes within secondary lymphoid organs and tissues. In these conditions, cells may encounter multiple chemoattractant signals in a complex spatial and temporal pattern. It has been shown that neutrophils can use chemokine receptors sequentially to move along different chemotactic gradients, a process defined as ‘‘multistep navigation’’ (7). The switch from the use of one type of receptor to another is determined by the capacity of the first to undergo ligand-induced desensitization. This process is mediated by specific GRK kinases that phosphorylate the engaged receptors, resulting in their binding to b-arrestin and their localization within clathrin coated pits (8, 9). Thus, it is not surprising that, as a general rule, leukocytes express several types of chemokine receptors. These are required to perform distinct steps in the process of extravasation and positioning within the tissue. There are approximately 40 chemokines identified to date, which are classified according to the configuration of cysteine residues near the N-terminus into four families: CC-, CXC-, C-, and CX3C (for the new nomenclature of chemokines, see reference 4). The large number of chemokines has resulted in a considerable difficulty in understanding their individual relevance and function. A more recent


CHEMOKINE RECEPTORS IN IMMUNE RESPONSES

595

classification into ‘‘inflammatory’’ and ‘‘lymphoid’’ chemokines based on the site of production and the eliciting stimuli represents a useful paradigm. Inflammatory chemokines were the first to be discovered and found to attract primarily neutrophils and other cells of the innate immune system (10–12). Subsequently they were shown to attract also different types of effector and memory lymphocytes. The inflammatory chemokines are produced by several cell types such as endothelial, epithelial, and stromal cells as well as leukocytes. They are induced to high level of expression by inflammatory stimuli such as LPS, IL-1, and TNF-a. Classical examples include IL-8 (13), RANTES (14), eotaxin (15– 17), MIP-1b (18), MCP-1 (19), and IP-10 (20). These chemokines are produced in different types of inflammatory diseases, possibly with distict kinetics. For instance eotaxin is produced at high levels in the lung undergoing allergic inflammation (15), while IP-10 production is regulated by IFN-c at sites of DTH reactions (21). Lymphoid chemokines are produced within lymphoid tissues and are involved in maintaining homeostatic leukocyte traffic and cell compartimentalization within these organs (3). Examples in this group are the B cell attracting chemokine-1 (BCA-1 and its mouse homologue BLC), which is produced by stromal cells in B cell follicles (22, 23); the secondary lymphoid chemokine (SLC; 24– 26), which is produced by endothelial cells of lymphatics and HEV and by stromal cells in the T cell area of lymph nodes (27); and the EBV1 ligand chemokine (ELC; 28), which is produced by interdigitating denolritic cells (DC) (29, 30). While most chemokines fit either the inflammatory or lymphoid type, some may play a dual role depending on the context in which they are produced. For instance, macrophage-derived chemokine (MDC; 31) can serve an inflammatory role when it is produced in the inflamed lung (32) but it can also participate in the regulation of cell-cell interaction in secondary lymphoid organs (33, 34). Some chemokines have restricted tissue distribution and may regulate local traffic. For instance, TECK production is restricted to thymus and intestine (35), while Lungkine has been reported to be made exclusively in the lung (36). With the notable exception of fractalkine, which is an integral membrane protein displayed on a long mucine stalk (37), all chemokines are secreted molecules. Because of their positive charge they bind to sulfated proteoglycans present on the cell surface or in the extracellular matrix (38). Thus, once released, they tend to remain concentrated locally, forming stable gradients. The immobilization of chemokines is especially important for endothelial cells, which are exposed to the blood flow. The chemokines displayed on the surface of endothelial cells can be either synthesized by the cell itself or taken up from the extracellular medium. Endothelial cells can endocytose chemokines present on the abluminal aspect and transport them to the luminal part, where they are then displayed to rolling leukocytes (39). It is possible that chemokines and cytokines are transported by the afferent lymph and funneled to the high endothelial venules (HEV) by fibroblastic reticular cell conduit system (40).

596


Chemokines can be stored in secretory granules to be mobilized rapidly when needed. Endothelial cells that have been exposed to inflammatory stimuli develop the Weibel Palade bodies, which contain chemokines and P selectin (41, 42). The content of these granules can be transferred to the cell surface following stimulation with histamine or thrombin, thus providing the endothelial cell with a sort of memory of previous challenge that allows a rapid recruitment of leukocytes without need for protein synthesis. Similarly, in the cytotoxic granules of CTL, chemokines are stored together with proteoglycans and perforin in a form that is particularly effective in inhibiting HIV infection (43).


CHEMOKINES AND RECEPTORS FOR T AND B CELL DEVELOPMENT The differentiation of hematopoietic progenitors into mature T and B lymphocytes takes place in the specialized microenvironment of thymus and bone marrow. Recruitment of precursors, development within these organs, and exit of mature cells into the blood are all steps in which chemokines are expected to play an important role. Maturing thymocytes migrate from cortex to medulla, while undergoing positive and negative selection before entering the circulating pool (44). The importance of chemokine receptors in this process is suggested by the ability of a pertussis toxin transgene to inhibit export of mature T cells from the thymus (45). The response to chemokines present in the thymus is developmentally regulated during thymocyte maturation (46, 47). Both cortical and medullary thymocytes respond to TECK, a ligand for CCR9 (35, 48), and lose this response at the late stages of maturation before leaving the thymus, concomitantly with upregulation of L-selectin. Reciprocally, the response to ELC and SLC, two CCR7 ligands (28, 49), is acquired only at the latest stage of mature single positive T cells and is maintained in mature peripheral blood cells. Interestingly, the response to MDC, a CCR4 ligand (50), is transiently upregulated in transitional cells that migrate from the cortex to the medulla (47). These findings suggest a role for CCR9 in retaining cells in the thymus until completion of their maturation process, and for CCR4 in driving the migration of developing cells from cortex to medulla. Finally, CCR7 and CXCR4 (that is expressed at all developmental stages) may play a role in migration of mature T cells. Intriguingly, mice lacking CCR4, CXCR4, SDF1, or CCR7 have been reported to have normal thymocyte development, a result suggesting that these may be redundant. Several studies have demonstrated a role for CXCR4 in B cell lymphopoiesis. SDF-1 is constitutively expressed in bone marrow–derived stromal cells and promotes proliferation of B cell progenitors (51). Mice deficient in SDF-1 or CXCR4 die perinatally, and the number of B cell progenitors in mutant embryos is severely reduced in both fetal liver and bone marrow. These mice have other severe abnor-


597

malities, such as a cardiac ventricular septal defect and abnormal cerebellum development, suggesting a role for SDF-1/CXCR4 in a number of developmental processes outside the hematopoietic system (52–54). CXCR4 is also involved in homing and mobilization of CD34 progenitors (55) and is required for the retention of B lineage and granulocytic precursors within bone marrow or fetal liver (56). Treatment of human stem cells with antibodies to CXCR4 prevents their engraftment into SCID mice (57). These reports are consistent with a role for CXCR4 and SDF-1 in stem cell migration to the bone marrow and the retention of various hematopoietic cells within the microenvironment. Some chemokines have been shown to suppress myelopoiesis by a yet unknown mechanism (58, 59).


CHEMOKINES AND CHEMOKINE RECEPTORS FOR PRIMARY IMMUNE RESPONSES In the course of a primary immune response, the rare antigen-specific T and B lymphocytes first need to interact with APC, dendritic cells (DC), and follicular dendritic cells (FDC), respectively, and subsequently with each other. These encounters and the exchange of partners take place in the organized structure of lymph nodes, spleen, and Peyer’s patches. In lymph nodes and Peyer’s patches, the port of entry for naive T and B cells is the HEV, while antigen is carried by the afferent lymph together with APC or transcytosed by intestinal M cells into the dome region. In contrast, lymphocytes and antigen enter the spleen via the same route in the marginal zone sinuses. We first consider the migration of DC, then that of naive T and B cells, and finally the interactions among these three cell types (Figure 1, see color insert).

Traffic of DC from Sites of Antigen Capture to Sites of Antigen Presentation DC and their precursors migrate from the blood into the tissue up to the source of antigen. Here the cells are exposed to inflammatory stimuli that redirect them into the lymphatics and subsequently into the T cell areas of secondary lymphoid organs, where they present antigen to T cells (60). Monocytes represent the immediate nonproliferating precursors of immature DC in culture supplemented with GM-CSF and IL-4 (61), and it has been suggested that in vivo this differentiation can be promoted by the process of transendothelial migration (62). Both monocytes and immature DC express receptors for inflammatory chemokines—CXCR1, CCR1, CCR2, and CCR5—which accounts for their capacity to extravasate and migrate into inflamed tissues (31, 63–65). CXCR1 and CCR2 may be particularly important because they mediate arrest of rolling monocytes under flow condition (66). Other chemoattractant receptors for complement and bacterial products, characteristic of the innate


598


immune response, are expressed on monocytes and immature DC and may be particularly important in the final steps of approach to the antigen. These include receptors for fmlp, C5a, and C3a as well as a novel 7TMR (63, 67, 68). The migration up a gradient of inflammatory chemokines allows immature DC to be increasingly exposed to maturation stimuli. These include the proinflammatory cytokines TNF-a and IL-1, and bacterial and viral products such as LPS or dsRNA (69–71). Maturation of DC results in a complete reprogramming of the cell, with downregulation of endocytic activity (69), upregulation of MHC, adhesion, and costimulatory molecules (71, 72) as well as a striking switch in chemokine receptor usage (29, 64, 65, 73–75). As soon as 1 h after induction of maturation, there is a complete loss of responsiveness to inflammatory chemokines due to a loss of the corresponding receptors from the cell surface, which are downregulated first at the protein and subsequently at the transcriptional level. At the same time maturing DC upregulate receptors for lymphoid chemokines such as CXCR4, CCR4, and especially CCR7. Upon maturation CCR7 was found to be expressed at increasingly high levels and to be, unlike other receptors, resistant to ligand-induced downregulation (76). CCR7 upregulation results in responsiveness to SLC and ELC, CCR7 ligands produced by lymphatic endothelial cells (27) and by interdigitating DC, respectively (30, 77). The striking CCR7 upregulation on DC and the precise anatomical localization of its ligands immediately suggested a prominent role for CCR7 in driving DC migration to the lymph nodes. The unique role of CCR7 as a regulator of DC migration is supported by the observation that in plt/plt mice, which lack SLC and have low levels of ELC as a result of a spontaneous mutation, maturing DC are not able to migrate from skin to lymph nodes (78). A similar defect in DC migration has been observed in CCR7–deficient mice, directly demonstrating an essential role of this receptor in DC migration (79). The migration of DC from the site of antigen capture to the site of antigen presentation represents a general principle with slight variations in the anatomical details. In the spleen, immature DC are strategically localized in the marginal zone, where blood-borne antigens enter and migrate to the T cell areas following exposure to maturation stimuli such as LPS or a toxoplasma extract (80, 81). In the tonsils, immature DC are present in the epithelial crypts that represent the site of antigen entry in these organs. MIP-3a, a CCR6 ligand (82), is selectively produced (29) in the epithelial crypts, and Langerhans cells express CCR6 in addition to other chemokine receptors characteristic of immature DC (29, 83). In maturing Langerhans cells CCR6 is downregulated, while CCR7 is upregulated concomitantly with their relocalization within the T cell areas of the tonsils. MIP3a and CCR6 represent a chemokine/receptor pair particularly relevant in mucosal immunity. In the Peyer’s patches, immature DC are present in the dome region, where they capture antigens transcytosed by the M cells and subsequently migrate to the T cell areas (84–86). Maturing DC are also an abundant and strategic source of chemokines, which are produced in a precise time-ordered fashion (76). Following stimulation with



599

LPS, DC have an initial burst of MIP-1a, MIP-1b, and IL-8 production, which ceases within a few hours. RANTES and MCP-1 are also induced, but in a more steady fashion. The amount of inflammatory chemokines released at these early time points by maturing DC is extremely high (1 pg of MIP-1b/h/cell). At later time points DC produce mainly lymphoid chemokines. ELC is induced after 10– 20 h, and TARC, MDC, and PARC, which are constitutively transcribed at low levels, are strongly upregulated with a similarly slow kinetics (76). In vivo mouse Langerhans cells were found to upregulate MDC once they reach the lymph node, thus becoming more attractive for recently activated T cells (34). The production of such high levels of chemokines in a precisely time-ordered fashion suggests an important role in regulating DC function. On the one hand, inflammatory chemokines produced early on enhance and sustain the recruitment of immature DC and other inflammatory cells while downregulating the cognate receptors in an autocrine fashion, a process which may allow maturing DC to follow new chemotactic gradients. On the other hand the late production of lymphoid chemokines may facilitate cell positioning and cell-cell interaction within the lymph node microenvironment. We proposed DC recruitment and activation as a central control mechanism that discriminates between effective T cell priming and tolerance (87). The recruitment of immature DC into tissues and their maturation and migration to draining lymph nodes can occur at different rates. In the presence of pathogens, a large number of immature DC will continuously enter the tissue, will be activated to a highly stimulatory state, and will migrate to the draining lymph node. This process will result in a strong and sustained stimulation of T cells, which is required to induce effector function (88, 89). There is evidence that the process of DC migration can also occur at a lower rate in the absence of inflammation. In transgenic mice expressing a model antigen under the control of the rat insulin promoter, bone marrow-derived APC (most likely DC) spontaneously migrate from pancreas and kidney to the draining lymph nodes, where they stimulate specific T cells (90). Strikingly, however, this stimulation results in tolerance rather than priming (91). A plausible explanation is that low numbers of DC, which are poorly stimulatory and short lived, may be insufficient to drive full T cell activation. Because of its central role in T cell priming, DC migration represents an ideal target for escape or immunomodulation. While most viruses induce DC maturation (71), herpes virus fails to do so and actually inhibits upregulation of CCR7, possibly preventing DC migration to the draining lymph nodes (92). It is possible that other pathogens as well as tumor cells may use the same strategy to decrease the efficiency of antigen presentation. Indeed, failure of tumor-infiltrating DC to migrate to the draining lymph nodes has been reported. (93). A novel DC-like cell, the so-called plasmocytoid monocyte, is present in very small numbers in peripheral blood and has been shown to accumulate around HEV in inflamed lymph nodes or tonsils (94, 95). These cells produce extremely high levels of type I IFN (96, 97) and can differentiate in vitro into so-called

600


DC2, which have been reported to induce Th2 responses (98). Plasmocytoid monocytes express L-selectin and CXCR3 (99), a receptor for Mig, IP-10, and ITAC, all chemokines upregulated by IFN-c, which may account for their homing capacity to inflamed lymph nodes (97).


Lymphocyte Homing to Secondary Lymphoid Organs and Organization of T and B Cell Areas Lymphocytes treated with pertussis toxin fail to enter lymph nodes and Peyer’s patches (100, 101) and, although able to enter the spleen, fail to localize to the T and B areas (102). These findings demonstrate the critical involvement of signaling via Gi-coupled receptors in both extravasation and positioning within secondary lymphoid organs. HEV are ports of entry for T and B lymphocytes but normally exclude other cell types such as neutrophils and monocytes. Two of the components of the HEV entry code, L selectin and LFA-1, have been known for a long time (5). However these molecules are also expressed on neutrophils which in fact roll on HEV, suggesting that the specificity for extravasation is determined by chemokine receptors. Several lines of evidence indicate that the chemokine receptor responsible for lymphocyte arrest on HEV is CCR7. First, CCR7 is selectively expressed on naive T and B lymphocytes (103–105). Second, its ligand SLC is produced by endothelial cells of HEV (27) and is efficiently retained on their surface where it is displayed to rolling cells (3). Third, SLC can mediate arrest of naive T cells under flow conditions (106). The most compelling evidence, however, came again from two strains of mutant mice. In plt/plt mice that fail to produce SLC, T cells do not enter the lymph nodes (78), and this defect can be restored by intradermal injection of SLC, which is transported to the HEV (107). Finally, mice lacking CCR7 have a profound defect of T and also B cell homing (79). Although the homeostatic traffic through HEV appears to be controlled by SLC and CCR7, it is possible that under inflammatory conditions additional chemokines may be displayed on HEV, thus enhancing the recruitment capacity and broadening its specificity (40). This is suggested by the dramatic changes in naive and memory T cell traffic that occur across the HEV after antigen challenge (108). Once having crossed the HEV, T and B lymphocytes take different routes marked by specific chemokines (102). T cells localize to the surrounding T cell area, where they scan the DC surface for specific antigen. With the exception of SLC, which is made by stromal cells (109), most of the landmark chemokines of this area are produced by mature DC (29, 34, 76, 110). The organizing role of DC-derived chemokines is underscored by the fact that relB-deficient mice, which lack mature DC, have profoundly disorganized T cell areas, a defect that can be partially restored by reconstitution with DC (111). Besides ELC, mature DC produce PARC (110), a chemokine that binds to a receptor that is expressed on naive T cells. In addition, they produce MDC and TARC (33, 34, 76), which bind



601

to CCR4, a receptor that is expressed on a subset of memory T cells but not on naive T cells and is rapidly upregulated following T cell activation (112, 113). The CCR4-MDC interaction may allow activated T cells to compete more effectively with naive cells for the limiting space on the DC surface and to be retained for a longer time to complete the polarization process. B lymphocytes that enter into the T cell area through HEV take one further step and migrate to the B cell follicle, where antigen is displayed on FDC. In a pioneering experiment, Lipp and coworkers showed that mice lacking the orphan receptor BLR1 (now called CXCR5) have a severe defect in development of B cell follicles in spleen and Peyer’s patches and lack inguinal lymph nodes (114). The picture has been recently completed by the finding that the CXCR5 ligand BLC/BCA-1 is produced by stromal cells, most likely FDC, within B cell follicles (22, 23). Interestingly, LTa/b and TNF are required both for FDC development and function and for BLC/BCA-1 expression (109, 115). Furthermore, cells producing BCA-1 are found in organized lymphoid structures in patients chronically infected with H. pilori (117). Altogether these observations suggest that the architecture of lymphoid organs is controlled by a cascade of signaling events between stromal cells and hematopoietic cells in which chemokines play a major role.

T Cell Activation and T Cell–B Cell Cooperation The segregation of T and B cells in distinct areas allows their separate stimulation by antigen displayed by the appropriate type of APC. However, once stimulated, these rare cells still need to come together. These movements have been visualized using adoptively transferred antigen-specific transgenic T and B cells labeled with fluorescent dyes. These cells localize initially in their respective areas and, following antigenic stimulation, move in a synchronous fashion out of these areas toward each other, meeting at the boundary between them (118). This rapid and reciprocal change in migratory pathways suggests a switch in chemokine receptor usage, and some evidence for this has been provided. Antigen-stimulated T cells downregulate CCR7 function, while upregulating CXCR5 and CCR4 and thus becoming sensitive to BLC/BCA-1 in the B cell areas (119) and to MDC that is produced by activated B cells (33). On the other hand, antigen-stimulated B cells acquire responsiveness to ELC, most likely by upregulating CCR7, which drive them to the T cell area (30). The reciprocal exchange of migratory capacity following activation can elegantly explain the efficiency and topology of T-B encounter (119a). This mechanism may actually depend on a subtle change in the balance between CXCR5 and CCR7 signaling. In the case of B cells this change can occur at any time during their migration and may result in their rerouting to the T cell area even before entering the B cell area, a phenomenon described as follicular exclusion (120). In the case of T cells, this balance can be influenced by the strength of signaling and by the nature of costimulatory molecules. Ligation of OX40 on T cells by OX40L expressed on activated DC is required for optimal

602



CXCR5 upregulation (119, 121). The dependence of these chemokine receptor changes on the strength of signaling may explain why administration of antigen in adjuvant—leading to presentation on activated DC—results in T cell proliferation and migration into B cell areas, while administration of antigen intravenously—resulting in presentation on nonactivated DC and B cells—leads to an abortive proliferation of T cells that fail to migrate to the B cell follicles (122). Once having interacted with specific T cells, some B cells proliferate and differentiate outside the follicle, while others give rise to the germinal center reaction. It is likely that changes in chemokine receptor are involved in this process as well, but the details are not known. Similarly there is no information on whether and how chemokines may control homing of memory B cells and plasma cells to medullary cords, splenic red pulp, mucosal-associated tissue, and bone marrow.

CHEMOKINES AND RECEPTORS FOR EFFECTOR AND MEMORY RESPONSES Antigen-primed T cells that exit from secondary lymphoid organs have acquired new migratory capacity, being capable of entering peripheral inflamed tissues. This migration has also been shown to depend on a pertussis toxin-sensitive signaling step (123). There is growing evidence that the accumulation of different types of effector cells in inflammatory lesions is a dynamic process orchestrated by the regulated expression of chemokines and chemokine receptors (124).

Flexibility of Chemokine Receptor Usage in Type 1 and Type 2 Responses Type 1 and type 2 polarized T cells mediate different types of protective or pathogenetic responses by secreting different cytokines and interacting with different types of leukocytes (125–127). Type 1 cells produce IFN-c and colocalize with macrophages and neutrophils in DTH lesions, while type 2 cells produce IL-4, IL-5, and IL-13 and are present with eosinophils and basophils at sites of allergic inflammation. The chemokines produced during the inflammatory process are expected to determine the extent, quality, and duration of the cellular infiltrate. A complex regulatory network of cytokine-chemokine interactions is emerging. Initial evidence indicates that this network may follow the basic rules of type 1/type 2 regulation. For instance IL-4 and IL-13 stimulate production of eotaxin and MDC, an effect that is counteracted by IFN-c (128, 129). Conversely IFN-c induces IP10 and Mig and upregulates RANTES, and this effect is antagonized by IL-4 (130). On the other hand, TNF-a, which is associated with both type 1 and type 2 responses, costimulates production of both type 1 or type 2 chemokines.



603

A striking example of chemokines and receptors associated with polarized responses is provided by CCR3 and its ligand eotaxin. Eotaxin is abundantly produced in mucosal tissues undergoing allergic inflammation (15, 128, 131, 132). CCR3 is expressed on eosinophils (133) and basophils (134) as well as on in vivo and in vitro polarized Th2 cells (135, 136). The sharing of CCR3 may allow these three cell types to colocalize at sites of eotaxin production (136). Here the IL-4, IL-5, and IL-13 produced by Th2 cells may stimulate the effector function of eosinophils and basophils and boost the production of eotaxin and MDC, thus amplifying and modulating the inflammatory reaction. Besides CCR3, other chemokine receptors are expressed on Th2 cells. These include CCR4 (105, 137, 138), a receptor for TARC and MDC, which is also expressed on basophils (139), CCR8 (112), the receptor for I-309 (140), and the orphan chemoattractant receptor CRTh2 (141). The relative importance of CCR3 and CCR4 as markers for Th2 cells and their role in allergic reactions has been a contentious issue. While CCR3 expression is confined to IL-4–producing type2 T cells (135), CCR4 expression is not. In fact, CCR4 is also expressed on activated Th1 cells (112), on TGF-b-dependent seminaive T cells (105), and on

Figure 2 The relationship between antigenic stimulation, homing capacity, and polarization. As a function of the duration of TCR and cytokine signaling, naı¨ve T cells undergo a linear differentiation process into a nonpolarized lymph node homing stage to a polarized tissue homing stage and eventually to apoptotic death. Cells with such properties are generated in the primary response and persist as central memory TCM and effector memory TEM for years.


604


skin homing T cells, which are mainly Th1 (142). The presence of CCR4 on such different functional subsets may reflect the multiple roles of TARC and MDC, which are expressed in a variety of lymphoid and nonlymphoid tissues and may behave as inflammatory, constitutive, or tissue-homing chemokines in different circumstances. CCR3 and CCR4 ligands are produced with different kinetics, suggesting that these receptors may play distinct roles at different times of the inflammatory process at the onset or chronic phase and either in cell recruitment or in retention (32). Preliminary reports indicate however that allergic inflammation is not impaired in CCR3 and CCR4 knockout mice, suggesting that these receptors play a redundant role (143, 144). This is perhaps not surprising since other receptors for inflammatory chemokines such as CCR2 are expressed on Th2 cells (105). The receptors expressed preferentially on Th1 cells are CCR5, CXCR3, and CCR1 (105, 137, 145). In rheumatoid arthritis and multiple sclerosis, thought to be Th1-associated diseases, virtually all T cells in the lesions express CCR5 and CXCR3, although usually only 5–15% of peripheral blood T cells have this phenotype (146, 147). CCR1 and CCR5 are expressed on monocytes and macrophages, which explains their colocalization with Th1 cells. However, CCR5 and CXCR3 are expressed also at lower levels on Th2 cells (105). As discussed for Th2 responses, the cytokines produced by Th1 cells, namely IFN-c, upregulate Th1-attracting chemokines while antagonizing Th2-attracting chemokines. For instance IP-10 is induced by IFN-c and is expressed abundantly in Th1 lesions (21, 130). Despite their association with Th1 responses, the relative importance of CCR5, CXCR3, and CCR1 is not clear. CCR5 is not essential for Th1 responses, since individuals homozygous for the D32/D32 CCR5 mutation (which results in loss of CCR5 expression) are healthy (148). This single point mutation must carry a considerable selective advantage besides resistance to HIV, considering its rapid spread in the northern european population (149). The pattern of chemokine receptor expression described so far applies to the resting state of Th1 or Th2 T cells, i.e. the cells that are bound to migrate from blood into tissues before encountering antigen. However, TCR triggering and IL2 have been shown to modulate some chemokine receptors, suggesting that, after antigenic stimulation, T cells might change their migratory capacity (112, 135, 150). A systematic analysis revealed a strikingly clear functional program (113). Within 6 h following TCR stimulation, the receptors for inflammatory chemokines—CCR1, CCR2, CCR3, CCR5, CCR6, and CXCR3—are downregulated at the mRNA and protein level, while CCR7, CCR4, CCR8, and CXCR5 are upregulated. This switch in chemokine receptors is transient, since the original pattern is regained within a few days as the cells go back to the resting state. We suggest that CCR7 upregulation may allow T cells that have been activated by antigen in the tissue to migrate further to the draining lymph nodes following the pathway of DC. Indeed, activated T cells often clustered with DC have been described in the afferent lymph (151–153). In addition, the new set of chemokine receptors expressed on recently activated T cells may relocalize them within the tissues to


605

sites where the corresponding chemokines, ELC, SLC, BCA, I-309, and TARC, are produced. This mechanism may explain why activated T cells are selectively enriched at sites of chronic inflammation (154, 155). These scenarios suggest a role for ectopic production of lymphoid chemokines in the organization of chronic inflammatory processes.


Subsets of Central Memory and Effector Memory T Cells Identified by Chemokine Receptors A fraction of activated T lymphocytes generated during the primary response persist as circulating memory cells, which are able to give a quantitatively enhanced and qualitatively different response upon secondary challenge (156– 158). Memory T cells migrate mainly through peripheral tissues (152) and possess rapid effector function, but can also undergo further clanal expansion in the lymph nodes. Until recently, memory T cells were difficult to distinguish from effector T cells, and little was understood about the connection between phenotype, function, and migratory pathways (159). An important distinction of circulating CD4 and CD8 memory cells can be made according to their expression of CCR7 (160). Memory cells that lack CCR7 express low levels of L-selectin and high levels of integrins, produce effector cytokines IFN-c and IL-4 with rapid kinetics, and, in the case of CD8, contain cytotoxic granules. In contrast, the memory T cells that express CCR7 also express high levels of L-selectin and do not produce effector cytokines, consistent with a lymph node homing, nonpolarized phenotype. However, both CCR71 and CCR7` memory cells are more sensitive than naive T cells to TCR stimulation and, consequently, stimulate DC more efficiently to produce IL-12. It is worth noting that CCR3 and CCR5 are expressed exclusively in different proportions of the CCR71 effector cells, while CCR4 is expressed both in a small subset of CCR71 cells and on approximately half of the CCR7` cells. The latter CCR7` CCR4` phenotype had been previously described on ‘‘seminaive’’ T cells generated in vitro in the presence of TGF-b (105) and may correspond to the uncommitted proliferating precursors described in mice (161). Importantly, both in vitro and ex vivo experiments suggest a linear differentiation from CCR7` naive to CCR7` lymph-node homing non-effector memory cells, to CCR71 tissuehoming effector memory cells. Preliminary evidence indicates that this linear differentiation is dependent on the duration and strength of T cell stimulation. These findings show that effector function and migratory capacity are coordinately regulated following T cell activation and establish a functional heterogeneity among memory cells. On the one hand, the CCR71 subset represents effector memory T cells (TEM) that can be rapidly recruited to inflamed peripheral tissues, where they perform immediate effector functions, such as DTH or cytotoxicity, required to rapidly contain invasive pathogens (162). On the other hand, the CCR7` subset represents a pool of central memory T cells (TCM) that home to lymph nodes to stimulate DC and to generate new waves of effector cells.

606


Helper memory cells for B and CTL responses (163–165) are likely to belong to the TCM subset.


Chemokine Receptors for Selective Homing to Skin and Gut The existence of subsets of memory T cells that preferentially migrate to gut or skin is well documented. Skin homing T cells can be identified by their expression of cutaneous lymphocyte associated antigen (CLA) (166), a glycosylated form of PSGL-1 produced by fucosyltransferase VII (167). CLA binds to E-selectin on endothelial cells of inflamed skin. However, E-selectin expression is not restricted to inflamed cutaneous endothelium, suggesting that there must be other determinants for skin trophism. This element of specificity is a chemokine receptor, CCR4. In fact, CCR4 is coexpressed together with CLA on a subset of memory/ effector T cells, and its ligand TARC is expressed on the endothelial cells of inflamed skin but not of inflamed gut (142). TARC has been shown to induce integrin-dependent firm adhesion of CLA` T cells consistent with its role in the extravasation process. Besides its role in extravasation through skin vessels, CCR4 may drive cell migration within several types of inflamed tissue, wherever MDC and TARC are produced by resident cells, for instance in the lung (32) or in the liver (168). Gut homing T cells express high levels of a4b7 integrin that mediates Lselectin independent rolling on MAdCAM-1, a vascular addressin that is expressed on lamina propria venules and Peyer’s patch HEV, but also in the marginal zone of the spleen and in areas of chronic inflammation (169). The obligatory role of G protein–coupled receptors in leukocyte migration has been demonstrated also for homing to the gut (170). The receptor involved has been recently identified as GPR-9-6, which has been accordingly defined as CCR9, a receptor for TECK, a chemokine selectively expressed in the endothelial cells of gut-associated tissues as well as thymus (171). CCR6 is also expressed on subsets of memory T cells and may drive their migration in response to MIP-3a to both skin and gut (172). Altogether, these findings provide a striking example of how selectivity of homing can be achieved via a combinatorial usage of relatively few components.

CHEMOKINE RECEPTORS AS TARGETS FOR IMMUNE MODULATION Because of their high degree of specificity and their role in inflammatory responses, chemokine receptors have been immediately recognized as attractive targets for drug development (173). The above discussion has provided examples on how chemokine receptors function in directing cell migration associated with specific immunological responses. It is conceivable that blocking a single che-



607

mokine receptor might inhibit one type of immune response, while leaving others unaffected. There are several ways to achieve a selective blockade of chemokine receptors (174). The first is represented by monoclonal antibodies that have been used as effective inhibitors of both chemokine and HIV binding (175–177). A second approach is based on the observation that N-terminal modifications of chemokines generate ligands that, although still capable of binding, have antagonistic properties (178). For instance, N-terminal modification of RANTES results in powerful inhibitors of CCR5 and other chemokine receptors (179–181), and a truncated antagonist of MCP-1 inhibits arthritis in the MRL-Ipr mouse model (182). The most promising approach to the development of selective chemokine receptor antagonists is represented by small organic molecules, which classically are able to disrupt ligand binding to G protein–coupled receptors (183–185). To date, a number of small molecule antagonists of chemokine receptors have been identified that are selective for CXCR4, CCR5, CXCR2, CCR1, and others (reviewed in 186). Some of the other receptors have proven more difficult to antagonize, and this may relate to the way the ligand binds to the receptor. The feasibility and the biological consequences of antagonizing chemokine receptors are illustrated by the fact that herpes and pox viruses target chemokines as part of their strategy to alter the immune response. The viral chemokines were probably pirated by viruses and modified so as to retain high-affinity binding, but not necessarily signaling capacity. The Kaposi’s sarcoma herpesvirus HHV8 encodes two chemokines, vMIPI and vMIPII, which display a broad range of binding specificities for both CC and CXC chemokine receptors as well as biological activities. These chemokines bind and stimulate CCR8, a receptor that may participate in Th2 responses, while possibly antagonizing Th1-associated receptors (187–189). In addition they have been reported to block HIV infection via CCR3 and CCR5 and to stimulate neoangiogenesis (190, 191). It has been suggested that these chemokines may promote neoangiogenesis and deviate the inflammatory response from a protective Th1 to a nonprotective Th2 type. The MC148 gene of the Molluscum contagiosum virus encodes a chemokine with a broad spectrum antagonistic activity for several CC and CXC chemokines with diverse receptor specificities (192). This chemokine may explain the prolonged absence of an inflammatory response in skin tumors that harbor replicating Molluscum contagiosum virus. Another strategy to block chemokines is used by some poxviruses that produce soluble chemokine binding molecules. These molecules show a broad specificity for inflammatory chemokines and modulate the influx of inflammatory cells into virus-infected tissues (193, 194).

CONCLUDING REMARKS The past few years have witnessed an explosive growth in the discovery of new chemokines and new chemokine receptors. While this process of discovery will soon be over, we are just starting to appreciate the complexity and intricacy of

608



the system. In the near future it should be possible to gain further insights into the hyerarchical organization of the chemokine system, which is an essential requirement for any therapeutic exploitation. Nonetheless some fundamental principles have already emerged, for instance, the critical contribution of chemokine receptors to the combinatorial code for extravasation, the serial usage of different chemokine receptors in leukocyte navigation, and the flexible regulation of chemokine receptor expression by the activation and differentiation processes. It is conceivable that chemokines will soon find therapeutic applications to selectively induce, modulate, or prevent cellular interactions in both the afferent and efferent limb of the immune response. We can foresee the possibility of using chemokines adjuvants to induce formation of new lymphoid tissue in ectopic sites. The feasibility of blocking chemokine receptors has already paved the way to their application as antiinflammatory drugs. It is certainly fair to say that the chemokine system has already gained a central position in all phases of the immune response. ACKNOWLEDGMENTS We thank Marco Baggiolini, Mariagrazia Uguccioni, and Chistoph Schaniel for critical reading end comments. The Basel Institute for Immunology was founded and is supported by F. Hoffmann-La Roche Ltd., Basel, Switzerland. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Murphy PM. 1994. The molecular biology of leukocyte chemoattractant receptors. Annu. Rev. Immunol. 12:593–633 2. Baggiolini M, Dewald B, Moser B. 1997. Human chemokines: an update. Annu. Rev. Immunol. 15:675–705 3. Yoshie O, Imai T, Nomiyama H. 1997. Novel lymphocyte-specific CC chemokines and their receptors. J. Leukoc. Biol. 62:634–44 4. Rossi D, Zlotnik A. 2000. The biology of chemokines and their receptors. Annu. Rev. Immunol. 18:000–00 5. Springer TA. 1994. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76:301–14 6. Butcher EC, Picker LJ. 1996. Lymphocyte homing and homeostasis. Science 272:60–66

7. Foxman EF, Campbell JJ, Butcher EC. 1997. Multistep navigation and the combinatorial control of leukocyte chemotaxis. J. Cell Biol. 139:1349–60 8. Premont RT, Inglese J, Lefkowitz RJ. 1995. Protein kinases that phosphorylate activated G protein-coupled receptors. FASEB J. 9:175–82 9. Daaka Y, Luttrell LM, Ahn S, Della Rocca, GJ, Ferguson SS, Caron MG, Lefkowitz RJ. 1998. Essential role for G protein-coupled receptor endocytosis in the activation of mitogen-activated protein kinase. J. Biol. Chem. 273:685–88 10. Oppenheim JJ, Zachariae CO, Mukaida N, Matsushima K. 1991. Properties of the novel proinflammatory supergene ‘‘intercrine’’ cytokine family. Annu. Rev. Immunol. 9:617–48 11. Baggiolini M, Dewald B, Moser B. 1994.


12.

13.


14.

15.

16.

17.

18.

IL-8 and related chemotactic cytokinesCXC and CC chemokines. Adv. Immunol. 55:97–179 Schall TJ, Bacon KB. 1994. Chemokines, leukocyte trafficking, and inflammation. Curr. Opin. Immunol. 6:865–73 Bazzoni F, Cassatella MA, Rossi F, Ceska M, Dewald B, Baggiolini M. 1991. Phagocytosing neutrophils produce and release high amounts of the neutrophil-activating peptide 1/interleukin 8. J. Exp. Med. 173:771–74 Schall TJ, Bacon K, Toy KJ, Goeddel DV. 1990. Selective attraction of monocytes and T lymphocytes of the memory phenotype by cytokine RANTES. Nature 347:669–71 Jose PJ, Griffiths-Johnson DA, Collins PD, Walsh DT, Moqbel R, Totty NF, Truong O, Hsuan JJ, Williams TJ. 1994. Eotaxin: a potent eosinophil chemoattractant cytokine detected in a guinea pig model of allergic airways inflammation. J. Exp. Med. 179:881–87 Rothenberg ME, Luster AD, Leder P. 1995. Murine eotaxin: an eosinophil chemoattractant inducible in endothelial cells and in interleukin 4-induced tumor suppression. Proc. Natl. Acad. Sci. USA 92:8960–64 Ponath PD, Qin S, Ringler DJ, ClarkLewis I, Wang J, Kassam N, Smith H, Shi X, Gonzalo J-A, Newman W, Gutierrez-Ramos J-C, Mackay CR. 1996. Cloning of the human eosinophil chemoattractant, eotaxin: expression, receptor binding and functional properties provide a mechanism for the selective recruitment of eosinophils. J. Clin. Invest. 97:604–12 Sherry B, Tekamp-Olson P, Gallegos C, Bauer D, Davatelis G, Wolpe SD, Masiarz F, Coit D, Cerami A. 1988. Resolution of the two components of macrophage inflammatory protein 1, and cloning and characterization of one of those components, macrophage inflam-

19.

20.

21.

22.

23.

24.

25.

26.

27.

609

matory protein 1 beta. J. Exp. Med. 168:2251–59 Leonard EJ, Yoshimura T. 1990. Human monocyte chemoattractant protein-1 (MCP-1). Immunol. Today 11:97–101 Luster AD, Ravetch JV. 1987. Biochemical characterization of a c-interferoninduced protein (IP-10). J. Exp. Med. 166:1084–97 Kaplan G, Luster AD, Hancock G, Cohn ZA. 1987. The expression of a gamma interferon-induced protein (IP-10) in delayed immune responses in human skin. J. Exp. Med. 166:1098–108 Legler DF, Loetscher M, Roos RS, Clark-Lewis I, Baggiolini M, Moser B. 1998. B cell-attracting chemokine 1, a human CXC chemokine expressed in lymphoid tissues, selectively attracts B lymphocytes via BLR1/CXCR5. J. Exp. Med. 187:655–60 Gunn MD, Ngo VN, Ansel KM, Ekland EH, Cyster JG, Williams LT. 1998. A Bcell-homing chemokine made in lymphoid follicles activates Burkitt’s lymphoma receptor-1. Nature 391:799– 803 Nagira M, Imai T, Hieshima K, Kusuda J, Ridanpaa M, Takagi S, Nishimura M, Kakizaki M, Nomiyama H, Yoshie O. 1997. Molecular cloning of a novel human CC chemokine secondary lymphoid-tissue chemokine that is a potent chemoattractant for lymphocytes and mapped to chromosome 9p13. J. Biol. Chem. 272:19518–24 Hedrick JA, Zlotnik A. 1997. Identification and characterization of a novel beta chemokine containing six conserved cysteines. J. Immunol. 159:1589–93 Hromas R, Kim CH, Klemsz M, Krathwohl M, Fife K, Cooper S, SchnizleinBick C, Broxmeyer HE. 1997. Isolation and characterization of Exodus-2, a novel C-C chemokine with a unique 37-amino acid carboxyl-terminal extension. J. Immunol. 159:2554–58 Gunn MD, Tangemann K, Tam C, Cyster

610

28.


29.

30.

31.

32.

33.

SALLUSTO n MACKAY n LANZAVECCHIA JG, Rosen SD, Williams LT. 1998. A chemokine expressed in lymphoid high endothelial venules promotes the adhesion and chemotaxis of naive T lymphocytes. Proc. Natl. Acad. Sci. USA 95:258–63 Yoshida R, Imai T, Hieshima K, Kusuda J, Baba M, Kitaura M, Nishimura M, Kakizaki M, Nomiyama H, Yoshie O. 1997. Molecular cloning of a novel human CC chemokine EBI1-ligand chemokine that is a specific functional ligand for EBI1, CCR7. J. Biol. Chem. 272:13803–9 Dieu M-C, Vanbervliet B, Vicari A, Bridon J-M, Oldham E, Ait-Yahia S, Briere F, Zlotnik A, Lebecque S, Caux C. 1998. Selective recruitment of immature and mature dendritic cells by distinct chemokines expressed in different anatomic sites. J. Exp. Med. 188:373–86 Ngo VN, Tang HL, Cyster JG. 1998. Epstein-Barr virus-induced molecule 1 ligand chemokine is expressed by dendritic cells in lymphoid tissues and strongly attracts naive T cells and activated B cells. J. Exp. Med. 188:181–91 Godiska R, Chantry D, Raport CJ, Sozzani S, Allavena P, Leviten D, Mantovani A, Gray PW. 1997. Human macrophage-derived chemokine (MDC), a novel chemoattractant for monocytes, monocyte-derived dendritic cells, and natural killer cells. J. Exp. Med. 185: 1595–604 Gonzalo JA, Pan Y, Lloyd CM, Jia GQ, Yu G, Dussault B, Powers CA, Proudfoot AE, Coyle AJ, Gearing D, GutierrezRamos JC. 1999. Mouse monocytederived chemokine is involved in airway hyperreactivity and lung inflammation. J. Immunol. 163:403–11 Schaniel C, Pardali E, Sallusto F, Speletas M, Ruedl C, Shimizu T, Seidl T, Andersson J, Melchers F, Rolink AG, Sideras P. 1998. Activated murine B lymphocytes and dendritic cells produce a novel CC chemokine which acts selec-

34.

35.

36.

37.

38.

39.

40.

41.

42.

tively on activated T cells. J. Exp. Med. 188:451–63 Tang HL, Cyster JG. 1999. Chemokine up-regulation and activated T cell attraction by maturing dendritic cells. Science 284:819–22 Vicari AP, Figueroa DJ, Hedrick JA, Foster JS, Singh KP, Menon S, Copeland NG, Gilbert DJ, Jenkins NA, Bacon KB, Zlotnik A. 1997. TECK: a novel CC chemokine specifically expressed by thymic dendritic cells and potentially involved in T cell development. Immunity 7:291– 301 Rossi DL, Hurst SD, Xu Y, Wang W, Menon S, Coffman RL, Zlotnik A. 1999. Lungkine, a novel CXC chemokine, specifically expressed by lung bronchoepithelial cells. J. Immunol. 162:5490–97 Bazan JF, Bacon KB, Hardiman G, Wang W, Soo K, Rossi D, Greaves DR, Zlotnik A, Schall TJ. 1997. A new class of membrane-bound chemokine with a CX3C motif. Nature 385:640–44 Webb LM, Ehrengruber MU, ClarkLewis I, Baggiolini M, Rot A. 1993. Binding to heparan sulfate or heparin enhances neutrophil responses to interleukin 8. Proc. Natl. Acad. Sci. USA 90:7158–62 Middleton J, Neil S, Wintle J, ClarkLewis I, Moore H, Lam C, Auer M, Hub E, Rot A. 1997. Transcytosis and surface presentation of IL-8 by venular endothelial cells. Cell 91:385–95 Gretz JE, Anderson AO, Shaw S. 1997. Cords, channels, corridors and conduits: critical architectural elements facilitating cell interactions in the lymph node cortex. Immunol. Rev. 156:11–24 Wolff B, Burns AR, Middleton J, Rot A. 1998. Endothelial cell ‘‘Memory’’ of inflammatory stimulation: human venular endothelial cells store interleukin 8 in weibel-palade bodies. J. Exp. Med. 188:1757–62 Utgaard JO, Jahnsen FL, Bakka A, Brandtzaeg P, Haraldsen G. 1998. Rapid


43.

44.


45.

46.

47.

48.

49.

50.

51.

secretion of prestored interleukin 8 from weibel-palade bodies of microvascular endothelial cells. J. Exp. Med. 188:1751– 56 Wagner L, Yang OO, Garcia-Zepeda EA, Ge Y, Kalams SA, Walker BD, Pasternack MS, Luster AD. 1998. Beta-chemokines are released from HIV-1specific cytolytic T-cell granules complexed to proteoglycans. Nature 391: 908–11 von Boehmer H. 1994. Positive selection of lymphocytes. Cell 76:219–28 Chaffin KE, Perlmutter RM. 1991. A pertussis toxin-sensitive process controls thymocyte emigration. Eur. J. Immunol. 21:2565–73 Kim CH, Pelus LM, White JR, Broxmeyer HE. 1998. Differential chemotactic behavior of developing T cells in response to thymic chemokines. Blood 91:4434–43 Campbell JJ, Pan J, Butcher EC. 1999. Cutting edge: developmental switches in chemokine responses during T cell maturation. J. Immunol. 163:2353–57 Zaballos A, Gutierrez J, Varona R, Ardavin C, Marquez G. 1999. Identification of the orphan chemokine receptor GPR9-6 as CCR9, the receptor for the chemokine TECK. J. Immunol. 162:5671– 75 Yoshida R, Nagira M, Kitaura M, Imagawa N, Imai T, Yoshie O. 1998. Secondary lymphoid-tissue chemokine is a functional ligand for the CC chemokine receptor CCR7. J. Biol. Chem. 273: 7118–22 Imai T, Chantry D, Raport CJ, Wood CL, Nishimura M, Godiska R, Yoshie O, Gray PW. 1998. Macrophage-derived chemokine is a functional ligand for the CC chemokine receptor 4. J. Biol. Chem. 273:1764–68 D’Apuzzo M, Rolink A, Loetscher M, Hoxie JA, Clark-Lewis I, Melchers F, Baggiolini M, Moser B. 1997. The chemokine SDF-1, stromal cell-derived

52.

53.

54.

55.

56.

57.

58.

59.

611

factor 1, attracts early stage B cell precursors via the chemokine receptor CXCR4. Eur. J. Immunol. 27:1788–93 Tachibana K, Hirota S, Iizasa H, Yoshida H, Kawabata K, Kataoka Y, Kitamura Y, Matsushima K, Yoshida N, Nishikawa S, Kishimoto T, Nagasawa T. 1998. The chemokine receptor CXCR4 is essential for vascularization of the gastrointestinal tract. Nature 393:591–94 Zou YR, Kottmann AH, Kuroda M, Taniuchi I, Littman DR. 1998. Function of the chemokine receptor CXCR4 in haematopoiesis and in cerebellar development. Nature 393:595–99 Ma Q, Jones D, Borghesani PR, Segal RA, Nagasawa T, Kishimoto T, Bronson RT, Springer TA. 1998. Impaired B-lymphopoiesis, myelopoiesis, and derailed cerebellar neuron migration in CXCR4and SDF-1-deficient mice. Proc. Natl. Acad. Sci. USA 95:9448–53 Aiuti A, Webb IJ, Bleul C, Springer T, Gutierrez-Ramos JC. 1997. The chemokine SDF-1 is a chemoattractant for human CD34` hematopoietic progenitor cells and provides a new mechanism to explain the mobilization of CD34` progenitors to peripheral blood. J. Exp. Med. 185:111–20 Ma Q, Jones D, Springer TA. 1999. The chemokine receptor CXCR4 is required for the retention of B lineage and granulocytic precursors within the bone marrow microenvironment. Immunity 10:463–71 Peled A, Petit I, Kollet O, Magid M, Ponomaryov T, Byk T, Nagler A, Ben-Hur H, Many A, Shultz L, Lider O, Alon R, Zipori D, Lapidot T. 1999. Dependence of human stem cell engraftment and repopulation of NOD/SCID mice on CXCR4. Science 283:845–48 Broxmeyer HE, Kim CH. 1999. Regulation of hematopoiesis in a sea of chemokine family members with a plethora of redundant activities. Exp. Hematol. 27:1113–23 Patel VP, Kreider BL, Li Y, Li H, Leung

612

60.


61.

62.

63.

64.

65.

66.

SALLUSTO n MACKAY n LANZAVECCHIA K, Salcedo T, Nardelli B, Pippalla V, Gentz S, Thotakura R, Parmelee D, Gentz R, Garotta G. 1997. Molecular and functional characterization of two novel human C-C chemokines as inhibitors of two distinct classes of myeloid progenitors. J. Exp. Med. 185:1163–72 Banchereau J, Steinman RM. 1998. Dendritic cells and the control of immunity. Nature 392:245–52 Sallusto F, Lanzavecchia A. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor alpha. J. Exp. Med. 179:1109–18 Randolph GJ, Beaulieu S, Lebecque S, Steinman RM, Muller WA. 1998. Differentiation of monocytes into dendritic cells in a model of transendothelial trafficking. Science 282:480–83 Sozzani S, Sallusto F, Luini W, Zhou D, Piemonti L, Allavena P, Van Damme J, Valitutti S, Lanzavecchia A, Mantovani A. 1995. Migration of dendritic cells in response to formyl peptides, C5a, and a distinct set of chemokines. J. Immunol. 155:3292–95 Sallusto F, Schaerli P, Loetscher P, Schaniel C, Lenig D, Mackay CR, Qin S, Lanzavecchia A. 1998. Rapid and coordinated switch in chemokine receptor expression during dendritic cell maturation. Eur. J. Immunol. 28:2760–69 Sozzani S, Allavena P, D’Amico G, Luini W, Bianchi G, Kataura M, Imai T, Yoshie O, Bonecchi R, Mantovani A. 1998. Differential regulation of chemokine receptors during dendritic cell maturation: a model for their trafficking properties. J. Immunol. 161:1083–86 Gerszten RE, Garcia-Zepeda EA, Lim YC, Yoshida M, Ding HA, Gimbrone MA Jr, Luster AD, Luscinskas FW, Rosenzweig A. 1999. MCP-1 and IL-8 trigger firm adhesion of monocytes to

67.

68.

69.

70.

71.

72.

73.

vascular endothelium under flow conditions. Nature 398:718–23 Morelli A, Larregina A, Chuluyan I, Kolkowski E, Fainboim L. 1996. Expression and modulation of C5a receptor (CD88) on skin dendritic cells. Chemotactic effect of C5a on skin migratory dendritic cells. Immunology 89:126–34 Samson M, Edinger AL, Stordeur P, Rucker J, Verhasselt V, Sharron M, Govaerts C, Mollreau C, Vassart G, Doms RW, Parameter M. 1998. ChemR23, a putative chemoattractant receptor, is expressed in monocytederived dendritic cells and macrophages and is a coreceptor for SIV and some primary HIV-1 strains. Eur. J. Immunol. 28:1689–700 Sallusto F, Cella M, Danieli C, Lanzavecchia A. 1995. Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J. Exp. Med. 182:839–40 Rescigno M, Citterio S, Thery C, Rittig M, Medaglini D, Pozzi G, Amigorena S, Ricciardi-Castagnoli P. 1998. Bacteriainduced neo-biosynthesis, stabilization, and surface expression of functional class I molecules in mouse dendritic cells. Proc. Natl. Acad. Sci. USA 95:5229–34 Cella M, Salio M, Sakakibara Y, Langen H, Julkunen I, Lanzavecchia A. 1999. Maturation, activation, and protection of dendritic cells induced by doublestranded RNA. J. Exp. Med. 189:821–29 Cella M, Engering A, Pinet V, Pieters J, Lanzavecchia A. 1997. Inflammatory stimuli induce accumulation of MHC class II complexes on dendritic cells. Nature 388:782–87 Delgado E, Finkel V, Baggiolini M, Mackay CR, Steinman RM, GranelliPiperno A. 1998. Mature dendritic cells respond to SDF-1, but not to several


74.

75.


76.

77.

78.

79.

80.

81.

beta-chemokines. Immunobiology 198: 490–500 Yanagihara S, Komura E, Nagafune J, Watarai H, Yamaguchi Y. 1998. EBI1/ CCR7 is a new member of dendritic cell chemokine receptor that is up-regulated upon maturation. J. Immunol. 161:3096– 102 Lin CL, Suri RM, Rahdon RA, Austyn JM, Roake JA. 1998. Dendritic cell chemotaxis and transendothelial migration are induced by distinct chemokines and are regulated on maturation. Eur. J. Immunol. 28:4114–22 Sallusto F, Palermo B, Lenig D, Miettinen M, Matikainen S, Julkunen I, Forster R, Brugstahler R, Lipp M, Lanzavecchia A. 1999. Distinct patterns and kinetics of chemokine production regulate dendritic cell function. Eur. J. Immunol. 29:1617– 25 Saeki H, Moore AM, Brown MJ, Hwang ST. 1999. Secondary lymphoid-tissue chemokine (SLC) and CC chemokine receptor 7 (CCR7) participate in the emigration pathway of mature dendritic cells from the skin to regional lymph nodes. J. Immunol. 162:2472–75 Gunn MD, Kyuwa S, Tam C, Kakiuchi T, Matsuzawa A, Williams LT, Nakano H. 1999. Mice lacking expression of secondary lymphoid organ chemokine have defects in lymphocyte homing and dendritic cell localization. J. Exp. Med. 189:451–60 Foerster R, Schubel A, Breitfeld D, Kremmer E, Renner-Muller I, Wolf E, Lipp M. 1999. CCR7 coordinates the primary immune response by establishing functional microenvironments in secondary lymphoid organs. Cell. 99:23–33 De Smedt T, Pajak TB, Muraille E, Lespagnard L, Heinen E, De Baetselier P, Urbain J, Leo O, Moser M. 1996. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184:1413–24 Sousa CR, Hieny S, Scharton-Kersten T,

82.

83.

84.

85.

86.

87.

88.

89.

90.

613

Jankovic D, Charest H, Germain RN, Sher A. 1997. In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas [see comments]. J. Exp. Med. 186:1819–29 Baba M, Imai T, Nishimura M, Kakizaki M, Takagi S, Hieshima K, Nomiyama H, Yoshie O. 1997. Identification of CCR6, the specific receptor for a novel lymphocyte-directed CC chemokine LARC. J. Biol. Chem. 272:14893–98 Yang D, Howard OMZ, Chen Q, Oppenheim JJ. 1999. Immature dendritic cells generated from monocytes in the presence of TGF-beta1 express functional CC chemokine receptor 6. J. Immunol. 163:1737–41 Kelsall BL, Strober W. 1996. Distinct populations of dendritic cells are present in the subepithelial dome and T cell regions of the murine Peyer’s patch. J. Exp. Med. 183:237–47 Neutra MR, Frey A, Kraehenbuhl JP. 1996. Epithelial M cells: gateways for mucosal infection and immunization. Cell 86:345–48 Ruedl C, Hubele S. 1997. Maturation of Peyer’s patch dendritic cells in vitro upon stimulation via cytokines or CD40 triggering. Eur. J. Immunol. 27:1325–30 Sallusto F, Lanzavecchia A. 1999. Mobilizing dendritic cells for tolerance, priming, and chronic inflammation. J. Exp. Med. 189:611–14 Iezzi G, Karjalainen K, Lanzavecchia A. 1998. The duration of antigenic stimulation determines the fate of naive and effector T cells. Immunity 8:89–95 Lanzavecchia A, Iezzi G, Viola A. 1999. From TCR engagement to T cell activation: a kinetic view of T cell behavior. Cell 96:1–4 Kurts C, Heath WR, Carbone FR, Allison J, Miller JF, Kosaka H. 1996. Constitutive class I-restricted exogenous

614

91.

92.


93.

94.

95.

96.

97.

98.

99.

SALLUSTO n MACKAY n LANZAVECCHIA presentation of self antigens in vivo. J. Exp. Med. 184:923–30 Kurts C, Kosaka H, Carbone FR, Miller JF, Heath WR. 1997. Class I-restricted cross-presentation of exogenous selfantigens leads to deletion of autoreactive CD8(`) T cells. J. Exp. Med. 186:239– 45 Salio M, Cella M, Suter M, Lanzavecchia A. 1999. Inhibition of dendritic cell maturation by herpes simplex virus. Eur. J. Immunol. 29:3245–53 Stoppacciaro A, Paglia P, Lombardi L, Parmiani G, Baroni C, Colombo MP. 1997. Genetic modification of a carcinoma with the IL-4 gene increases the influx of dendritic cells relative to other cytokines. Eur. J. Immunol. 27:2375–82 Facchetti F, de Wolf-Peeters C, van den Oord JJ, de Vos R, Destmet VJ. 1989. Plasmacytoid monocytes (so-called plasmacytoid T-cells) in Kikuchi’s lymphadenitis. An immunohistologic study. Am. J. Clin. Pathol. 92:42–50 Olweus J, BitMansour A, Warnke R, Thompson PA, Carballido J, Picker LJ, Lund-Johansen F. 1997. Dendritic cell ontogeny: a human dendritic cell lineage of myeloid origin. Proc. Natl. Acad. Sci. USA 94:12551–56 Siegel FP, Kadowski N, Schodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, Antonenko S, Liu YJ 1999. The nature of the principal type 1 interferon-producing cells in human blood. Science 284:1835–37 Cella M, Jarrossay D, Facchetti F, Alebardi O, Nakajima H, Lanzavecchia A, Colonna M. 1999. Plasmacytoid monocytes migrate to inflamed lymph nodes and produce large amounts of type I interferon. Nat. Med. 5:919–23 Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, Liu YJ. 1999. Reciprocal control of T helper cell and dendritic cell differentiation. Science 283:1183–86 Loetscher M, Gerber B, Loetscher P,

100.

101.

102.

103.

104.

105.

106.

107.

Jones SA, Piali L, Clark-Lewis I, Baggiolini M, Moser B. 1996. Chemokine receptor specific for IP10 and mig: structure, function, and expression in activated T-lymphocytes. J. Exp. Med. 184:963–69 Spangrude GJ, Braaten BA, Daynes RA. 1984. Molecular mechanisms of lymphocyte extravasation. I. Studies of two selective inhibitors of lymphocyte recirculation. J. Immunol. 132:354–62 Spangrude GJ, Sacchi F, Hill HR, Van Epps DE, Daynes RA. 1985. Inhibition of lymphocyte and neutrophil chemotaxis by pertussis toxin. J. Immunol. 135:4135–43 Cyster JG, Goodnow CC. 1995. Pertussis toxin inhibits migration of B and T lymphocytes into splenic white pulp cords. J. Exp. Med. 182:581–86 Burgstahler R, Kempkes B, Steube K, Lipp. 1995. Expression of the chemokine receptor BLR2/EBI1 is specifically transactivated by Epstein-Barr virus nuclear antigen 2. Biochem. Biophys. Res. Commun. 215:737–43 Campbell JJ, Bowman EP, Murphy K, Youngman KR, Siani MA, Thompson DA, Wu L, Zlotnik A, Butcher EC. 1998. 6-C-kine (SLC), a lymphocyte adhesiontriggering chemokine expressed by high endothelium, is an agonist for the MIP3beta receptor CCR7. J. Cell Biol. 141:1053–59 Sallusto F, Lenig D, Mackay Cr, Lanzavecchia A. 1998. Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes. J. Exp. Med. 187:875–83 Campell JJ, Hedrick J, Zlotnik A, Siani MA, Thompson DA, Butcher EC. 1998. Chemokines and the arrest of lymphocytes rolling under flow conditions. Science 279:381–84 Stein J, Rot A, Narasimhaswamy M, Nakano H, Luo Y, Gunn MD, Matsuzawa A, Quackenbush EJ. Dorf ME, von Andrian UH. 1999. The CC chemokine


108.


109.

110.

111.

112.

113.

114.

615

and specific anatomic compartments of TCA-4 (SLC, 6Ckine, Exodus-2) trigthe spleen. Cell 87:1037–47 gers LFA-1-mediated arrest of rolling T lymphocytes in peripheral lymph nodes 115. Fu YX, Chaplin DD. 1999. Development and maturation of secondary lymphoid high endothelial venules. J. Exp. Med. In tissues. Annu. Rev. Imunol. 17:399–433 press Mackay CR, Martson W, Dudler L. 1992. 116. Deleted in proof Altered patterns of T cell migration 117. Mazzucchelli L, Blaser A, Kappeller A. Schaerli P. Laissue JA, Baggiolini M. through lymph nodes and skin following Uguccioni M. 1999. BCA-1 is highly antigen challenge. Eur. J. Immunol. expressed in Helicobacter pylori-induced 22:2205–10 mucosa-associated lymphoid tissue and Ngo VN, Korner H, Gunn MD, Schmidt gastric lymphoma. J. Clin. Invest. KN, Riminton DS, Cooper MD, Brown104:R19–54 ing JL, Sedgwick JD, Cyster JG. 1999. Lymphotoxin alpha/beta and tumor 118. Garside P, Ingulli E, Merica RR, Johnson JG. Noelle RJ, Jenkins MK. 1998. Visunecrosis factor are required for stromal alization of specific B and T lymphocyte cell expression of homing chemokines in interactions in the lymph node. Science B and T cell areas of the spleen. J. Exp. 281:96–99 Med. 189:403–12 Adema J, Hartgers F, Verstraten R, de 119. Flynn S, Toellner KM, Raykundalia C, Goodall M, Lane P. 1998. CD4 T cell Vries E, Marland G, Menon S, Foster J, cytokine differentiation: the B cell actiXu Y, Nooyen P, McClanahan T, Bacon vation molecule, OX40 ligand, instructs KB, Figdor CG. 1997. A dendritic-cellCD4 T cells to express interleukin 4 and derived C-C chemokine that preferenupregulates expression of the chemokine tially attracts naive T cells. Nature receptor, Blr-1. J. Exp. Med. 188:297– 387:713–17 304 Burkly L, Hession C, Ogata L, Reilly C, Marconi LA, Olson D, Tizard R, Cate R, 119a. Ansel KM, McHeyzer-Williams LJ, Ngo VN, McHeyzer-Williams MG, Cyster Lo D. 1995. Expression of relB is JG. 1999. In vivo-activated CD4 T cells required for the development of thymic upregulate CXC chemokine receptor 5 medulla and dendritic cells. Nature and reprogram their response to lym373:531–36 phoid chemokines. J. Exp. Med. D’Ambrosio D, Iellem A, Bonecchi R, 190:1123–34 Mazzeo D, Sozzani S, Mantovani A, Sinigaglia F. 1998. Selective up-regulation 120. Cyster JG, Hartley SB, Goodnow CC. 1994. Competition for follicular niches of chemokine receptors CCR4 and CCR8 exludes self-reactive cells from the recirupon activation of polarized human type culating B-cell repertoire. Nature 2 Th cells. J. Immunol. 161:5111–15 371:389–95 Sallusto F, Kremmer E, Palermo B, Hoy A, Ponath P, Qin S, Forster R, Lipp M, 121. Brocker T, Gulbranson-Judge A, Flynn S, Riedinger M, Raykundalia C, Lane P. Lanzavecchia A. 1999. Switch inche1999. CD4 T cell traffic control: in vivo mokine receptor expression upon TCR evidence that ligation of OX40 on CD4 stimulation reveals novel horning potenT cells by OX40-ligand expressed on tial for recently activated T cells. Eur. J. dendritic cells leads to the accumulation Immunol. 29:2037–45 of CD4 T cells in B follicles. Eur. J. Forster R, Mattis AE, Kremmer E, Wolf Immunol. 29:1610–16 E, Brem G, Lipp M. 1996. A putative chemokine receptor, BLR1, directs B cell 122. Kearney ER, Pape KA, Loh DY, Jenkins MK. 1994. Visualizing of peptide-spemigration to defined lymphoid organs

616

123.

124.


125.

126.

127.

128.

129.

130.

131.

SALLUSTO n MACKAY n LANZAVECCHIA cific T cell immunity and peripheral tolerance induction in vivo. Immunity 1:327–39 Spangrude GJ, Araneo BA, Daynes RA. 1985. Site-selective homing of antigenprimed lymphocyte populations can play a crucial role in the efferent limb of cellmediated immune responses in vivo. J. Immunol. 134:2900–7 Sallusto F, Lanzavecchia A, Mackay CR. 1998. Chemokines and chemokine receptors in T-cell priming and Th1/Th2-mediated responses. Immunol. Today 19:568– 74 Mosmann TR, Coffman RL. 1989. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7:145–73 Seder RA, Paul WE. 1994. Acquisition of lymphokine-producing phenotype by CD4` T cells. Annu. Rev. Immunol. 12:635–73 Romagnani, S. 1994. Lymphokine production by human T cells in disease states. Annu. Rev. Immunol. 12:227–57 Li L, Xia Y, Nguyen A, Lai YH, Feng L, Mosmann TR, Lo D. 1999. Effects of Th2 cytokines on chemokine expression in the lung: IL-13 potently induces eotaxin expression by airway epithelial cells. J. Immunol. 162:2477–87 Bonecchi R, Sozzani S, Stine JT, Luini W, D’Amico G, Allavena P, Chantry D, Mantovani A. 1998. Divergent effects of interleukin-4 and interferon-gamma on macrophage- derived chemokine production: an amplification circuit of polarized T helper 2 responses. Blood 92:2668–71 Luster AD, Ravetch JV. 1987. Genomic characterization of a gamma-interferoninducible gene (IP-10) and identification of an interferon-inducible hypersensitive site. Mol. Cell. Biol. 7:3723–31 MacLean JA, Ownbey R, Luster AD. 1996. T cell-dependent regulation of eotaxin in antigen-induced pulmonary eosinophila. J. Exp. Med. 184:1461–69

132. Ganzalo JA, Jia GQ, Aguirre V, Friend D, Coyle AJ, Jenkins NA, Lin GS, Katz H, Lichtman A, Copeland N, Kopf M, Gutierrez-Ramos JC. 1996. Mouse Eotaxin expression parallels eosinophil accumulation during lung allergic inflammation but it is not restricted to a Th2-type response. Immunity 4:1–14 133. Ponath PD, Qin S, Post TW, Wang J, Wu L, Gerard NP, Newman W, Gerard C, Mackay CR. 1996. Molecular cloning and characterization of a human eotaxin receptor expressed selectively on eosinophils. J. Exp. Med. 183:2437–48 134. Uguccioni M, Mackay CR, Ochensberger B, Loetscher P, Rhis S, LaRosa GJ, Rao P, Ponath PD, Baggiolini M, Dahinden CA. 1997. High expression of the chemokine receptor CCR3 in human blood basophils. Role in activation by eotaxin, MCP-4, and other chemokines. J. Clin. Invest. 100:1137–43 135. Sallusto F, Mackay CR, Lanzavecchia A. 1997. Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Science 277:2005–7 136. Gerber B, Zanni MP, Uguccioni M, Loetscher M, Mackay CR, Pichler WJ, Yawalkar N, Baggiolini M, Moser B. 1997. Functional expression of the eotaxin receptor CCR3 in T lymphocytes co-localising with eosinophils. Curr. Biol. 7:836–43 137. Bonecchi R, Bianchi G, Bordignon PP, D’Ambrosio D, Lang R, Borsatti A, Sozzani S, Allavena P, Gray PA, Mantovani A, Sinigaglia F. 1998. Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (Th1s) and Th2s. J. Exp. Med. 187:129–34 138. Imai T, Nagira M, Takagi S, Kakizaki M, Nishimura M, Wang J, Gray PW, Matsushima K, Yoshie O. 1999. Selective recruitment of CCR4-bearing Th2 cells toward antigen-presenting cells by the CC chemokines thymus and activationregulated chemokine and macrophage-


139.


140.

141.

142.

143. 144. 145.

146.

147.

derived chemokine. Int. Immunol. 11:81– 88 Power CA, Meyer A, Nemeth K, Bacon KB, Hoogewerf AJ, Proudfoot AE, Wells TN. 1995. Molecular cloning and functional expression of a novel CC chemokine receptor cDNA from a human basophilic cell line. J. Biol. Chem. 270:19495–500 Goya I, Gutierrez J, Varona R, Kremer L, Zaballos A, Marquez G. 1998. Identification of CCR8 as the specific receptor for the human beta-chemokine I-309: cloning and molecular characterization of murine CCR8 as the receptor for TCA3. J. Immunol. 160:1975–81 Nagata K, Tanaka K, Ogawa K, Kemmotsu K, Imai T, Yoshie O, Abe H, Tada K, Nakamura M, Sugamura K, Takano S. 1999. Selective expression of a novel surface molecule by human Th2 cells in vivo. J. Immunol. 162:1278–86 Campbell JJ, Haraldsen G, Pan J, Rottman J, Qin S, Ponath P, Andrew DP, Warnke R, Ruffing N, Kassam N, Wu L, Butcher EC. 1999. The chemokine receptor CCR4 in vascular recognition by cutaneous but not intestinal memory T cells. Nature 400:776–80 Deleted in proof Deleted in proof Loetscher P, Uguccioni M, Bordoli L, Baggiolini M, Moser B, Chizzolini C, Dayer JM. 1998. CCR5 is characteristic of Th1 lymphocytes. Nature 391:344–45 Qin S, Rottman JB, Myers P, Kassam N, Weinblatt M, Loetscher M, Koch AE, Moser B, Mackay CR. 1997. The chemokine receptors CXCR3 and CCR5 mark subsets of T cells associated with certain inflammatory reactions. J. Clin. Invest. 101:746–54 Balashov KE, Rottman JB, Weiner HL, Hancock WW. 1999. CCR5(`) and CXCR3(`) T cells are increased in multiple sclerosis and their ligands MIP1alpha and IP-10 are expressed in

148.

149.

150.

151.

152.

153.

154.

617

demyelinating brain lesions. Proc. Natl. Acad. Sci. USA 96:6873–78 Samson M, Libert F, Doranz BJ, Rucker J, Liesnard C, Farber CM, Saragosti S, Lapoumeroulie C, Cognaux J, Forceille C, Muyldermans G, Verhofstede C, Burtonboy G, Georges M, Imai T, Rana S, Yi Y, Smyth RJ, Collman RG, Doms RW, Vassart G, Parmentier M, 1996. Resistance to HIV-1 infection in caucasian individuals bearing mutant alleles of the CCR-5 chemokine receptor gene. Nature 382:722–25 Libert F, Cochaux P, Beckman G, Samson M, Aksenova M, Cao A, Czeizel A, Claustres M, de la Rua C, Ferrari M, Ferrec C, Glover G, Grinde B, Guran S, Kucinskas V, Lavinha J, Mercier B, Ogur G, Peltonen L, Rosatelli C, Schwartz M, Spitsyn V, Timar L, Beckman L, Vassart G, et al. 1998. The deltaccr5 mutation conferring protection against HIV-1 in Caucasian populations has a single and recent origin in Northeastern Europe. Hum. Mol. Genet. 7:399–406 Loetscher P, Seitz M, Baggiolini M, Moser B. 1996. Interleukin-2 regulates CC chemokine receptor expression and chemotactic responsiveness in T lymphocytes. J. Exp. Med. 184:569–77 Mackay CR, Hein WR, Brown MH, Matzinger P. 1988. Unusual expression of CD2 in sheep: implications for T cell interactions. Eur. J. Immunol. 18:1681– 88 Mackay CR, Marston WL, Dudler L. 1990. Naive and memory T cells show distinct pathways of lymphocyte recirculation. J. Exp. Med. 171:801–17 Pope M, Betjes MG, Romani N, Hirmand H, Cameron PU, Hoffman L, Gezelter S, Schuler G, Steinman RM. 1994. Conjugates of dendritic cells and memory T lymphocytes from skin facilitate productive infection with HIV-1. Cell 78:389–98 Scotet E, David-Ameline J, Peyrat MA, Moreau-Aubry A, Pinczon D, Lim A,

618

155.

156.


157.

158.

159.

160.

161.

162.

163.

164.

SALLUSTO n MACKAY n LANZAVECCHIA Even J, Semana G, Berthelot JM, Breathnach R, Bonneville M, Houssaint E. 1996. T cell response to Epstein-Barr virus transactivators in chronic rheumatoid arthritis. J. Exp. Med. 184:1791–800 Ludewig B, Odermatt B, Landmann S, Hengartner H, Zinkernagel RM. 1998. Dendritic cells induce autoimmune diabetes and maintain disease via de novo formation of local lymphoid tissue. J. Exp. Med. 188:1493–501 Zinkernagel RM, Bachmann MF, Kundig TM, Oehen S, Pirchet H, Hengartner H. 1996. On immunological memory. Annu. Rev. Immunol. 14:333–67 Ahmed R, Gray D. 1996. Immunological memory and protective immunity: understanding their relation. Science 272:54– 60 Dutton RW, Bradley LM, Swain SL. 1998. T cell memory. Annu. Rev. Immunol. 16:201–23 Mackay CR. 1991. T-cell memory: the connection between function, phenotype and migration pathways. Immunol. Today 12:189–92 Sallusto F, Lenig D, Forster R, Lipp M, Lanzavecchia A. 1999. Two subsets of memory T lymphocytes with distinct homing potential and effector function. Nature 401:708–12 Sad S, Mosmann TR. 1994. Single IL-2secreting precursor CD4 T cell can develop into either Th1 or Th2 cytokine secretion phenotype. J. Immunol. 153: 3514–22 Zinkernagel RM, Hengartner H. 1997. Antiviral immunity. Immunol. Today 18:258–60 Ridge JP, Di Rosa F, Matzinger P. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4` Thelper and a T-killer cell. Nature 393:474–78 Bennett SR, Carbone FR, Karamalis F, Flavell RA, Miller JF, Heath WR. 1998. Help for cytotoxic-T-cell responses is

165.

166.

167.

168.

169.

170.

171.

mediated by CD40 signalling. Nature 393:478–80 Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 393:480–83 Berg EL, Yoshino T, Rott LS, Robinson MK, Warnock RA, Kishimoto TK, Picker LJ, Butcher EC. 1991. The cutaneous lymphocyte antigen is a skin lymphocyte homing receptor for the vascular lectin endothelial cell-leukocyte adhesion molecule 1. J. Exp. Med. 174:1461– 66 Fuhlbrigge RC, Kieffer JD, Armerding D, Kupper TS. 1997. Cutaneous lymphocyte antigen is a specialized form of PSGL-1 expressed on skin-homing T cells. Nature 389:978–81 Yoneyama H, Harada A, Imai T, Baba M, Yoshie O, Zhang Y, Higashi H, Murai M, Asakura H, Matsushima K. 1998. Pivotal role of TARC, a CC chemokine, in bacteria-induced fulminant hepatic failure in mice. J. Clin. Invest. 102:1933–41 Berlin C, Berg EL, Briskin MJ, Andrew DP, Kilshaw PJ, Holzmann B, Weissman IL, Hamann A, Butcher EC. 1993. Alpha 4 beta 7 integrin mediates lymphocyte binding to the mucosal vascular addressin MAdCAM-1. Cell 74:185–85 Bargatze RF, Jutila MA, Butcher EC. 1995. Distinct roles of L-selectin and integrins alpha 4 beta 7 and LFA-1 in lymphocyte homing to Peyer’s patchHEV in situ: the multistep model confirmed and refined. Immunity 3:99–108 Zabel BA, Agace WW, Campbell JJ, Heath HM, Parent D, Roberts AI, Ebert EC, Kassam N, Qin S, Zovko M, LaRosa GJ, Yang LL, Soler D, Butcher EC, Ponath PD, Parker CM, Andrew DP. 1999. Human G protein-coupled receptor GPR9-6/CC chemokine receptor 9 is selectively expressed on intestinal homing T lymphocytes, mucosal lymphocytes, and thymocytes and is required for thymus-


172.

173.


174.

175.

176.

177.

178.

179.

expressed chemokine-mediated chemotaxis. J. Exp. Med. 190:1241–56 Liao F, Rabin RL, Smith CS, Sharma G, Nutman TB, Farber JM. 1999. CC-chemokine receptor 6 is expressed on diverse memory subsets of T cells and determines responsiveness to macrophage inflammatory protein 3 alpha. J. Immunol. 162:186–94 Baggiolini M, Walz A, Kunkel SL. 1989. Neutrophil-activating peptide-1/interleukin 8, a novel cytokine that activates neutrophils. J. Clin. Invest. 84:1045–49 Baggiolini M, Moser B. 1997. Blocking chemokine receptors. J. Exp. Med. 186:1189–91 Heath H, Qin S, Rao P, Wu L, LaRosa G, Kassam N, Ponath PD, Mackay CR. 1997. Chemokine receptor usage by human eosinophils. The importance of CCR3 demonstrated using an antagonistic monoclonal antibody. J. Clin. Invest. 99:178–84 Wu L, LaRosa G, Kassam N, Gordon CJ, Heath H, Ruffing N, Chen H, Humblias J, Samson M, Parmentier M, Moore JP, Mackay CR. 1997. Interaction of chemokine receptor CCR5 with its ligands: multiple domains for HIV-1 gp120 binding and a single domain for chemokine binding. J. Exp. Med. 186:1373–81 Grimaldi JC, Yu NX, Grunig G, Seymour BW, Cottrez F, Robinson DS, Hosken N, Ferlin WG, Wu X, Soto H, O’Garra A, Howard MC, Coffman RL. 1999. Depletion of eosinophils in mice through the use of antibodies specific for C-C chemokine receptor 3 (CCR3). J. Leukocyte Biol. 65:846–53 Moser B, Dewald B, Barella L, Schumacher C, Baggiolini M, Clark-Lewis I. 1993. Interleukin-8 antagonists generated by N-terminal modification. J. Biol. Chem. 268:7125–28 Gong JH, Uguccioni M, Dewald B, Baggiolini M, Clark-Lewis I. 1996. RANTES and MCP-3 antagonists bind

180.

181.

182.

183.

184.

185.

186.

187.

619

multiple chemokine receptors. J. Biol. Chem. 271:10521–27 Simmons G, Clapham PR, Picard L, Offord RE, Rosenkilde MM, Schwartz TW, Buser R, Wells TNC, Proudfoot AE. 1997. Potent inhibition of HIV-1 infectivity in macrophages and lymphocytes by a novel CCR5 antagonist. Science 276:276–79 Mack M, Luckow B, Nelson PJ, Cihak J, Simmons G, Clapham PR, Signoret N, Marsh M, Stangassinger M, Borlat F, Wells TN, Schlondorff D, Proudfoot AE. 1998. Aminooxypentane-RANTES induces CCR5 internalization but inhibits recycling: a novel inhibitory mechanism of HIV infectivity. J. Exp. Med. 187:1215–24 Gong JH, Ratkay LG, Waterfield JD, Clark-Lewis I. 1997. An antagonist of monocyte chemoattractant protein 1 (MCP-1) inhibits arthritis in the MRL-lpr mouse model. J. Exp. Med. 186:131–37 Schols D, Struyf S, Van Damme J, Este JA, Henson G, De Clercq E. 1997. Inhibition of T-tropic HIV strains by selective antagonization of the chemokine receptor CXCR4. J. Exp. Med. 186: 1383–88 Murakami T, Nakajima T, Koyanagi Y, Tachibana K, Fujii N, Tamamura H, Yoshida N, Waki M, Matsumoto A, Yoshie O, Kishimoto T, Yamamoto N, Nagasawa T. 1997. A small molecule CXCR4 inhibitor that blocks T cell linetropic HIV-1 infection. J. Exp. Med. 186:1389–93 Doranz BJ, Grovit-Ferbas K, Sharron MP, Mao SH, Goetz MB, Daar ES, Doms RW, O’Brien WA. 1997. A small-molecule inhibitor directed against the chemokine receptor CXCR4 prevents its use as an HIV-1 coreceptor. J. Exp. Med. 186:1395–400 Mackay CR, Lanzavecchia A, Sallusto F. 1999. Chemoattractant receptors and immune responses. Immunologist 7:112– 18 Sozzani S, Luini W, Bianchi G, Allavena


620


P, Wells TN, Napolitano M, Bernardini G, Vecchi A, D’Ambrosio D, Mazzeo D, Sinigaglia F, Santoni A, Maggi E, Romagnani S, Mantovani A. 1998. The viral chemokine macrophage inflammatory protein-II is a selective Th2 chemoattractant. Blood 92:4036–39 188. Endres MJ, Garlisi CG, Xiao H, Shan L, Hedrick JA. 1999. The Kaposi’s sarcoma-related herpesvirus (KSHV)encoded chemokine vMIP- I is a specific agonist for the CC chemokine receptor (CCR)8. J. Exp. Med. 189:1993–98 189. Dairaghi DJ, Fan RA, McMaster BE, Hanley MR, Schall TJ. 1999. HHV8encoded vMIP-I selectively engages chemokine receptor CCR8. Agonist and antagonist profiles of viral chemokines. J. Biol. Chem. 274:21569–74 190. Moore PS, Boshoff C, Weiss RA, Chang Y. 1996. Molecular mimicry of human cytokine and cytokine response pathway genes by KSHV. Science 274:1739–44

191. Boshoff C, Endo Y, Collins PD, Takeuchi Y, Reeves JD, Schweickart VL, Siani MA, Sasaki T, Williams TJ, Gray PW, Moore PS, Chang Y, Weiss RA. 1997. Angiogenic and HIV-inhibitory functions of KSHV-encoded chemokines. Science 278:290–94 192. Damon I, Murphy PM, Moss B. 1998. Broad spectrum chemokine antagonistic activity of a human poxvirus chemokine homolog. Proc. Natl. Acad. Sci. USA 95:6403–7 193. Alcami A, Symons JA, Collins PD, Williams TJ, Smith GL. 1998. Blockade of chemokine activity by a soluble chemokine binding protein from vaccinia virus. J. Immunol. 160:624–33 194. Lalani AS, Masters J, Graham K, Liu L, Lucas A, McFadden G. 1999. Role of the myxoma virus soluble CC-chemokine inhibitor glycoprotein, M-T1, during myxoma virus pathogenesis. Virology 256:233–45

Figure 1 The role of chemokines and chemokine receptors in the afferent and efferent limb of the immune response. Left panel: na•ve T (green) and B (pink) cells enter the lymph node through HEV and localize in T and B cell areas. The few DC that migrate from noninflamed tissues are poorly stimulatory and, even when they carry tissue antigens, fail to prime T cells. Center panel: pathogens and inflammatory cytokines trigger DC maturation and migration from tissues to the lymph node and promote further monocyte recruitment into inflamed tissue, thus sustaining antigen sampling. The large number of stimulatory DC that reaches the lymph node induced T cell proliferation and differentiation into Th1 (red), Th2 (blue), or non-polarized T cells (yellow), depending on the duration of stimulation and the presence of polarizing cytokines. B cells (purple) are initially activated by antigen presented by FDC. Right panel: activated T and B cells migrate toward each other and interact in a cognate fashion. Th1 and Th2 cells enter the blood and exit at inflammatory sites together with effector leukocytes (not shown). Memory cells persist for years and comprise both tissue homing effector memory T cells (TEM) and lymph node homing central memory T xxxoooBurke,cells (TCM).


SALLUSTO ET AL C-1


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annual Review of Immunology 2000. 18:621–663 Copyright q 2000 by Annual Reviews. All rights reserved

PHOSPHORYLATION MEETS UBIQUITINATION: The Control of NF-jB Activity Michael Karin1 and Yinon Ben-Neriah2 1

Department of Pharmacology, Laboratory of Gene Regulation and Signal Transduction, University of California, San Diego, La Jolla, California 92093-0636; e-mail: [email protected] 2 Lautenberg Center for Immunology, Hadassah Medical School, Hebrew University, Jerusalem, Israel; 91120; e-mail: [email protected]


Key Words IKK, proteasome, ubiquitin, protein kinase, TNF-a, IL-1 Abstract NF-jB (nuclear factor-jB) is a collective name for inducible dimeric transcription factors composed of members of the Rel family of DNA-binding proteins that recognize a common sequence motif. NF-jB is found in essentially all cell types and is involved in activation of an exceptionally large number of genes in response to infections, inflammation, and other stressful situations requiring rapid reprogramming of gene expression. NF-jB is normally sequestered in the cytoplasm of nonstimulated cells and consequently must be translocated into the nucleus to function. The subcellular location of NF-jB is controlled by a family of inhibitory proteins, IjBs, which bind NF-jB and mask its nuclear localization signal, thereby preventing nuclear uptake. Exposure of cells to a variety of extracellular stimuli leads to the rapid phosphorylation, ubiquitination, and ultimately proteolytic degradation of IjB, which frees NF-jB to translocate to the nucleus where it regulates gene transcription. NF-jB activation represents a paradigm for controlling the function of a regulatory protein via ubiquitination-dependent proteolysis, as an integral part of a phosphorylationbased signaling cascade. Recently, considerable progress has been made in understanding the details of the signaling pathways that regulate NF-jB activity, particularly those responding to the proinflammatory cytokines tumor necrosis factor-a and interleukin-1. The multisubunit IjB kinase (IKK) responsible for inducible IjB phosphorylation is the point of convergence for most NF-jB–activating stimuli. IKK contains two catalytic subunits, IKKa and IKKb, both of which are able to correctly phosphorylate IjB. Gene knockout studies have shed light on the very different physiological functions of IKKa and IKKb. After phosphorylation, the IKK phosphoacceptor sites on IjB serve as an essential part of a specific recognition site for E3RSIjB/b-TrCP, an SCF-type E3 ubiquitin ligase, thereby explaining how IKK controls IjB ubiquitination and degradation. A variety of other signaling events, including phosphorylation of NF-jB, hyperphosphorylation of IKK, induction of IjB synthesis, and the processing of NF-jB precursors, provide additional mechanisms that modulate the level and duration of NF-jB activity.

0000–0000/00/0820–0621$14.00

621

622

KARIN n BEN-NERIAH


INTRODUCTION There are several well-characterized signaling pathways that lead to transcriptional activation. Of these, the nuclear factor-jB (NF-jB) pathway is unique in the rapidity of its activation, its unusual mechanism of regulation, and the comparatively high level of our current understanding, which spans the entire spectrum from detailed molecular structures to the physiological function of its components. NF-jB is maintained in an inactive form by sequestration in the cytoplasm through interaction with inhibitory proteins, the IjBs. Proteolytic degradation of IjB immediately precedes and is required for NF-jB nuclear translocation. This irreversible step in the signaling pathway constitutes a commitment to transcriptional activation. The signal is eventually terminated through cytoplasmic resequestration of NF-jB, which depends on IjBa synthesis, a process requiring NF-jB transcriptional activity, as well as on IjBa-dependent nuclear export. NF-jB activation often occurs in situations in which rapid and decisive action is required for cell survival, such as during activation of the innate immune response, our first line of defense against bacterial, viral, and fungal infections. NF-jB was identified in 1986 as a nuclear factor that bound to an enhancer element of the immunoglobulin (Ig)j light-chain gene and was believed to be specifically expressed in B cells (1). It quickly became apparent that NF-jB is present in virtually every cell type, but is retained in the cytoplasm in an inactive form bound to specific inhibitors, the IjBs (reviewed in 2–5). This review is not to enumerate all of the experimental data related to NF-jB, but to provide a unifying hypothesis regarding the signaling mechanisms involved in its regulation. Our intention is to avoid duplication of the many fine reviews that have been published over the years on various aspects of NF-jB composition and function. Moreover, because so many different stimuli activate NF-jB and, in many cases, warrant and have received specialized reviews, this review concentrates on the events and features that are common to the major NF-jB signaling pathways, those that are activated by proinflammatory stimuli and converge on the IjB kinase (IKK). At this time, it is this common pathway of NF-jB activation that is most clearly determined and where our understanding appears to be expanding most rapidly.

NF-jB AND ITS REGULATION VIA IjB PROTEINS NF-jB/Rel Proteins In its active DNA-binding form, NF-jB is a heterogeneous collection of dimers, composed of various combinations of members of the NF-jB/Rel family (Figure 1, see color insert). Five mammalian Rel proteins were identified, NF-jB1 (p50 and its precursor p105), NF-jB2 (p52 and its precursor p100), c-Rel, RelA (p65), and RelB (reviewed in 3–5). Although it is not certain whether additional NF-


CONTROL OF NF-jB VIA IjB DEGRADATION

623

jB/Rel genes are contained in mammalian genomes, no new family members have been identified within the last 7 years despite the considerable amount of new sequence information generated over that time period. Three Rel proteins were identified in Drosophila melanogaster, Dorsal, Dif, and Relish (reviewed in 3). All of these proteins share a highly conserved 300-amino-acid Rel homology region (RHR), composed of two immunoglobulin (Ig)-like domains. The RHR is responsible for dimerization, DNA binding, and interaction with the inhibitory IjB proteins. It also contains a nuclear localization sequence (NLS). Different NF-jB dimers exhibit different binding affinities for jB sites bearing the consensus sequence GGGRNNYYCC, where R is purine, Y is pyrimidine, and N is any base (reviewed in 4). The Rel proteins differ in their abilities to activate transcription, such that only p65/RelA and c-Rel were found to contain potent transcriptional-activation domains among the mammalian family members. It is believed that dimers composed solely of Rel proteins that lack transcriptionalactivation domains, such as p50, mediate transcriptional repression. To better understand the physiological function of the various NF-jB proteins, targeted disruption of individual Rel loci has been carried out in mice (reviewed in 3, 6). These ‘‘knockout’’ studies indicate specific roles for each NF-jB protein, although surprisingly only p65/RelA was found to be essential for survival, and there appears to be some compensation between other NF-jB members. p65:p50 heterodimers were the first form of NF-jB to be identified and are the most abundant in most cell types. Consequently, the term NF-jB most often is used to describe to the p50:p65 complex. Currently, it appears that all NF-jB complexes are regulated in the same manner—primarily through interactions with IjBs, although different IjBs may show preference for certain NF-jB dimers.

IjB Proteins The IjB family includes IjBa, IjBb, IjBc, IjBe, Bcl-3, the precursors of NFjB1 (p105), and NF-jB2 (p100), and the Drosophila protein Cactus (reviewed in 3, 7; Figure 1). Of these, the most important regulators of mammalian NF-jB are IjBa, IjBb, and IjBe. All IjBs contain either six or seven ankyrin repeats, and these stacked helical domains mediate binding to the RHR and masking the NLS of NF-jB. Only IjBa, IjBb, and IjBe contain N-terminal regulatory regions, which are required for stimulus-induced degradation, the key step in NFjB activation. IjBs also play an important role in termination of NF-jB activation. Newly synthesized IjBa enters the nucleus and binds NF-jB, thereby enhancing its dissociation from the DNA (the affinity of NF-jB to IjB appears to be higher than its affinity to jB sites on DNA) and causing its re-exportation to the cytoplasm by means of a nuclear export sequence (NES) present on IjBa (8). IjBa was the first IjB family member to be cloned and is still the best characterized. The physiological properties of the other IjBs are for the most part poorly understood. Mice lacking Bcl-3 and IjBa have been generated (reviewed

624

KARIN n BEN-NERIAH

in 3, 6). Bcl-31/1 mice exhibit specific defects in response to certain immunogenic agents. IjBa1/1 mice die 7–10 days after birth and exhibit a variety of inflammatory conditions consistent with increased NF-jB activity (9, 10). NFjB is constitutively elevated in hematopoietic cells of IjBa knockout mice, but not in IjBa-deficient embryonic fibroblasts. However, sustained activation of NFjB was observed in the latter cells after tumor necrosis factor-a (TNF-a) stimulation (9, 10). Although differences between the IjBs exist in their structures and apparent preferences for specific forms of NF-jB, IjBb is the only species whose expression is not regulated by NF-jB. Consequently, although IjBb does not compensate for the lack of IjBa in IjBa1/1 mice, when placed under control of the NF-jB-responsive IjBa promoter, IjBb is able to effectively substitute for IjBa (11).


Structure of NF-jB and IjB The butterfly-like crystal structures of various forms of NF-jB bound to DNA have revealed the structural origins of the transcription factor’s sequence and dimerization specificities (12–15). The RHR contains two Ig-like folds connected by a flexible linker region. Both folds contact the DNA; loops within the Nterminal fold are primarily responsible for sequence-specific recognition, whereas the C-terminal fold contains the dimer interface. Dimerization is mediated by extensive hydrophobic interactions along the interface surface, which is formed by a three-stranded b-sheet packing against a similar sheet in the opposing molecule. Recently, two independent structure determinations were reported for NFjB:IjB ternary complexes consisting of the RHR of p65, the C-terminal Ig-like fold of p50 (including the NLS), and the ankyrin repeats of IjBa (16, 17). The ankyrin repeats of IjBa form a curved a-helical stack, which binds to the Cterminal Ig-like folds of the RHRs in a discontinuous fashion. In both structures the NES on IjBa is exposed and disruption of NF-jB DNA binding appears to be primarily mediated by residues C-terminal to the ankyrin repeats of IjBa. Although these structures provide new insight into the regulation of NF-jB by IjB, the data, unfortunately, do not unambiguously address the manner in which IjBa masks the NLS of NF-jB. This is due in part to differences between the two structures in the p65 NLS and the lack of sufficient ordered structure in the p50 NLS. In addition, both structures lack the critical N-terminal regulatory domain of IjBa. Nevertheless, a likely assumption is that the N-terminal ankyrin repeats of IjBa sterically hinder the binding of karyopherins to the NLS of NFjB.

Biomedical Importance of NF-jB NF-jB regulates the transcription of an exceptionally large number of genes, particularly those involved in immune and inflammatory responses (reviewed in 5, 18, 19). NF-jB activation also plays an important role in the antiviral response through interferon gene induction (20, 21). Yet, through adaptation, many viruses,


625

including human immunodeficiency virus type 1 (HIV-1) (reviewed in 22) and human T-lymphotrophic virus type 1 (HTLV-1) (reviewed in 23), which do not cause interferon induction, exploit NF-jB to activate their own genes and to stimulate the survival and proliferation of lymphoid cells in which they replicate. In most cells, NF-jB activation protects from apoptosis, through induction of survival genes, although under certain conditions and in certain cell types it may also induce apoptosis (reviewed in 24, 25). Inappropriate regulation of NF-jB is directly involved in a wide range of human disorders, including a variety of cancers (reviewed in 26, 27), neurodegenerative diseases (reviewed in 28), ataxiatelangiectasia (29), arthritis (30), asthma (reviewed in 31), inflammatory bowel disease (reviewed in 32), and numerous other inflammatory conditions.


Activating Signals NF-jB is activated by a wide variety of different stimuli, including proinflammatory cytokines such as TNF-a and interleukin-1 (IL-1), T- and B-cell mitogens, bacteria, and bacterial lipopolysaccharide (LPS), viruses, viral proteins, doublestranded (ds) RNA, and physical and chemical stresses (reviewed in 3, 5). Although it is beyond the scope of this review to discuss all of the properties of these activation pathways, many of their common features are presented in later sections. Unfortunately, most NF-jB activators also stimulate other signaling pathways that are not directly related to NF-jB, often making it difficult to distinguish which early events are relevant to NF-jB activation and which are superfluous. Only after the late events that are common to most NF-jB activation pathways are thoroughly characterized can a detailed understanding of stimulusspecific early events that activate each of these pathways be readily obtained. A discussion of some of the better understood early signaling events, those involving proinflammatory cytokines, is offered below.

Overview of the Consensus NF-jB Activation Pathway Although there is no full understanding or agreement about how various extracellular and intracellular stimuli trigger NF-jB activation or at what point their signaling pathways converge, several common features do exist (see Figure 2, color insert). Potent activators, such as TNF-a, IL-1, or LPS, induce rapid degradation of the IjBs (especially IjBa) within minutes. For IjBa this degradation process consists of a series of well-characterized steps, which seem to be relevant to the other IjBs (reviewed in 2, 3). Inducible IjB phosphorylation, one of the earliest events in the common activation pathway, occurs at serines 32 and 36 in IjBa, and mutation of either serine (even to a threonine) greatly inhibits the degradation process (33–38). Phosphorylation leads to the immediate recognition of IjBa by the recently identified F-box/WD40 E3RSIjB/b-TrCP (39; reviewed in 40, 41), which consequently results in the polyubiquitinylation of IjBa primarily at lysines 21 and 22 (36, 42–44) by an SKp1–Cullin–F-box (SCF)type E3 (see below). This modification then targets IjBa for rapid degradation

626

KARIN n BEN-NERIAH

by the 26S proteasome. The degradation of its inhibitor exposes the NLS of NFjB resulting in binding to karyopherins and translocation of NF-jB to the nucleus. It is important that inhibitors of the 26S proteasome efficiently block NF-jB nuclear translocation, indicating that neither phosphorylation nor ubiquitinylation is sufficient to dissociate IjB from NF-jB (35, 45, 46).


Atypical Activation Pathways The mechanism of activation described above appears to be generally applicable to all potent NF-jB activators, although two additional activation pathways were reported. One is observed as a result of hypoxia or pervanadate treatment and is believed to require phosphorylation of IjBa at Tyr-42 (47, 48). The exact protein tyrosine kinases involved in this pathway are not known, but certain members of the Src family were proposed to be responsible for IjB phosphorylation. The subsequent dissociation of tyrosine-phosphorylated IjBa from NF-jB was suggested to be mediated by interaction with phosphoinositide-3 (PI3) kinase and not by degradation by the 26S proteasome (48). Because Tyr-42 is not conserved in other IjB family members, this pathway is specific for IjBa. It is, therefore, of interest to examine whether mice whose IjBa was replaced by IjBb (11) are capable of activating NF-jB in response to hypoxia. The second atypical activation pathway is observed in cells exposed to short-wavelength UV radiation (254 nm). UV radiation induces IjBa degradation via the 26S proteasome, but this process is not mediated by phosphorylation of Ser-32 and Ser-36 or Tyr-42 (49, 50). The mechanism of IjBa degradation in response to UV radiation is unknown (see below). In both of these alternative pathways, NF-jB activation is considerably slower and weaker than it is in response to the prototypical activators TNF-a, IL-1, or LPS.

IjB-KINASE (IKK)—THE KEY TO NF-jB ACTIVATION Identification and Composition Once it became firmly established that the critical step in NF-jB activation was IjB phosphorylation at Ser-32 and Ser-36 (of IjBa) or their equivalents, considerable effort was directed to identification of the responsible protein kinase(s). A cytokine-inducible protein kinase activity specific for the N-terminal regulatory serines of IjBa and IjBb was identified (51, 52). This activity, named IKK, could be related to an earlier constitutive IjB kinase activity, whose biochemical nature was not identified (53, 54). Although numerous other enzymes were suggested to mediate IjB phosphorylation, only IKK matches the characteristics of IjBa degradation, that is, rapid signal-induced activation, simultaneous phosphorylation of both serine residues, and the preference of serine over threonine as a phosphorylation target (51). The majority of IKK activity eluted at an apparent molecular mass of 700–900 kDa (51, 52, 55), suggesting that it was a multi-



627

component protein complex. However, a small amount of activity also eluted with an apparent molecular mass of ;300 kDa (55). By means of protein purification, microsequencing, and molecular cloning, three components of IKK were unambiguously identified. IKKa and IKKb (also referred to as IKK1 and IKK2) are 85- and 87-kDa proteins, respectively (51, 52, 55–57). The DNA sequence of IKKa was previously identified, by a reverse transcriptase-polymerase chain reaction (RT-PCR)-based cloning strategy searching for myc-like genes, as a putative serine/threonine kinase of unknown function called CHUK (58). IKKa/CHUK was also isolated through a two-hybrid screen as a protein that interacts with the NF-jB–inducing kinase (NIK) (56). IKKa and IKKb are highly homologous proteins (50% sequence identity; .70% sequence similarity) and contain N-terminal protein kinase domains as well as leucine zipper (LZ) and helix-loop-helix (HLH) motifs (see Figure 3, color insert). IKKa and IKKb serve as the catalytic subunits of the IKK complex. The third component of IKK is a 48-kDa regulatory subunit named IKKc (59), NFjB essential modulator (NEMO) (60), IKK-associated protein 1 (IKKAP1) (61), or 14.7 interacting protein (FIP-3) (62). IKKc was identified through the purification of the entire IKK complex (59, 61) and by genetic complementation of a cell line unresponsive to NF-jB–activating stimuli (60). Secondary-structure prediction algorithms indicate that IKKc is predominantly helical with large stretches of coiled-coil structure, including an LZ motif near the C terminus (59; see Figure 3). The last 23 residues of IKKc have a 70% sequence identity to the C terminus of FIP-2 (including the 3 cysteines and the histidine, a putative zinc finger motif) (62). IKKc interacts with the adenovirus protein Ad E3-14.7K, possibly through the FIP-2 homology region, and this interaction could account for the observed inhibition of TNF-a–induced cytolysis by the adenovirus protein (63). Other proteins suggested to be components of the 700- to 900-kDa IKK complex include various members of the IjB and Rel families (64, 65), MKP-1 and MEKK1 (52), and NIK (56, 64). In each of these cases, either these proteins were not detected by other purification schemes, or immunoblots of column profiles have shown that they do not coelute with the IKKa-IKKb-IKKc complex. Consequently, the question of whether these proteins are integral components of IKK remains unanswered. One group has reported that a 150-kDa regulatory subunit named IKAP and NIK are components of IKK (64). However, IKAP association with IKK could not be further confirmed, and this mammalian protein appears to be the homolog of a yeast protein that functions as a translation elongation factor (66, 67).

Stoichiometry and Complex Assembly Much of the difficulty in determining which proteins are legitimate IKK subunits is that most of the published biochemical purifications of IKK have been partial and, therefore, have yielded numerous proteins, most of which are probably not


628

KARIN n BEN-NERIAH

directly associated with IKK. The stoichiometry of the known components of IKK has not yet been determined. Thus far the majority of the 700- to 900-kDa complex appears to contain roughly equivalent amounts of IKKa and IKKb and an indeterminate amount of IKKc, which, from chemical crosslinking experiments, may exist in either a dimer or a trimer (59). The 300-kDa complex, which is minor in wild-type cells but is the only IKK complex in IKKc-deficient cell lines, appears to contain only IKKa/b dimers (59, 60). It is thus possible that IKKc, through its interactions with IKKa/b and self-association, accounts for the formation of the 700- to 900-kDa complex. Given that a single IKKa/b dimer elutes at 300 kDa (68), it is likely that the large IKKc-containing complex is composed of a dimer of dimers held together by IKKc. Several unique features distinguish IKK from other known kinases, such as the presence of HLH and LZ motifs in its catalytic subunit. Mutagenesis studies indicate that dimerization of IKKa/b is mediated by the LZ regions (55, 57, 68). Dimerization, which is essential for the activity, is direct and does not involve other proteins (55, 68). Although IKKc appears to be a major coordinator of IKK assembly, its precise role in this process is not yet clear. In vitro translation experiments have shown that IKKc can interact with IKKb but not with IKKa (59–61). However, IKK complexes in IKKb-deficient cells contain IKKa associated with IKKc (69–71). This could indicate that an additional component mediates the interaction between IKKa and IKKc. However, some of the contradictory results could be caused by using proteins which are N- or C-terminally epitope tagged, in which the presence of additional amino acid sequences could influence protein associations. Nterminal deletion mutants of IKKc have suggested that the critical site of interaction with IKKb is confined to residues 135–235 (59, 61, 72). Consistent with this is the observation that C-terminal deletion mutants of IKKc lacking the final 119 (59) or 154 (61) amino acids are able to interact with IKKb.

Kinase Domain The kinase domains of IKKa/b are similar to other known serine/threonine kinases. Consequently, the ATP binding site is highly conserved. The conserved lysine in this region (Lys-44) was mutated to generate kinase-defective forms of IKKa/b (52, 55, 57, 68). In vitro kinase assays carried out with purified recombinant IKKa/b indicated that homodimers containing two kinase-defective subunits were catalytically inactive, but heterodimers containing a single functional kinase domain still exhibit significant in vitro activity (68). In mammalian cells the kinase-defective forms of IKKa/b exhibit considerable TNF-a–induced activity when analyzed by in vitro kinase assays (55). This activity is caused by the ability of the ectopically expressed mutant form to heterodimerize with endogenous wild-type IKKa/b. However, overexpression of the kinase-defective forms can inhibit TNF-inducible nuclear translocation of p65 (52, 55) or activation of an NF-jB–dependent reporter gene (52, 57). These results suggest that IKK activ-



629

ity is rate limiting for NF-jB activation, because a small reduction in kinase activity can considerably decrease the extent of IjB degradation and NF-jB activation. Indeed, Ikkb`/1 cells were found to exhibit a twofold reduction in IKK activity but a 70%–80% decrease in NF-jB activation (70). Kinetic analyses were carried out with partially purified IKK obtained from mammalian cells (52, 73) or highly purified recombinant IKKa/b dimers produced in insect cells (61, 65, 68, 74). Studies carried out with recombinant proteins demonstrated that homodimers of IKKa and IKKb efficiently phosphorylate, with the correct specificity, various IjB substrates. Although the kinetic constants Km and Kcat were determined, the obtained values vary from one report to another such that no consensus can be drawn as to the actual values, primarily owing to differences in experimental conditions, particularly the choice of substrate and the manner in which the enzymes were prepared. However, several important conclusions can still be made. One is that the C-terminal region of IjBa lowers the apparent Km by an order of magnitude, based on two pieces of experimental data. One is comparison of the phosphorylation by partially purified IKK from mammalian cells of an N-terminal peptide with and without the addition of a peptide corresponding to residues 279–303 of IjBa (73). The other is a comparison of the phosphorylation of IjBa by recombinant IKKb, using full-length IjBa and an N-terminal peptide as substrates (61). Similar results were obtained when comparing phosphorylation of these substrates by recombinant IKKa. Vmax was increased when full-length IjBa complexed with p65:p50 was used as substrate. This preference of IKK for IjBs complexed with NF-jB over free IjBs probably contributes to the ability of newly synthesized IjB to accumulate in cells where IKK is still active (68).

IjB Kinase Activation Considerable evidence indicates that IKK activation depends on its phosphorylation. Purified IKK is inactivated upon incubation with protein phosphatase 2A (PP2A), whereas treatment of HeLa cells with the PP2A inhibitor, okadaic acid, results in IKK activation (51). Like other protein kinases, IKKa and IKKb contain an activation loop within their kinase domains, having a sequence that is identical between the two kinases (52, 75, 76). This region contains specific sites whose phosphorylation causes a conformational change that results in kinase activation (reviewed in 77). Replacement of serines 177 and 181 in the IKKb activation loop with alanines prevents IKK activation (52, 76). Both of these serines are phosphorylated in response to cell stimulation by TNF-a or IL-1 (76), and their replacement with glutamic acid (which mimics phosphoserine) results in constitutive IKK activity (52). Similarly, mutation of Ser-176 to alanine inactivates IKKa, whereas substitution with glutamic acid leads to full activation (75). However, a more recent study, in which exogenous IKKa and IKKb were expressed at low levels and were well incorporated into endogenous IKK complexes, concluded that conversion of Ser-176 and Ser-180 in IKKa to alanine, either indi-


630

KARIN n BEN-NERIAH

vidually or together, had no effect whatsoever on TNF-a– or IL-1–induced IKK activity (76), results that are entirely consistent with the genetic analysis of IKKa/ b function (see below). Therefore, although phosphorylation of the IKKa activation loop is undoubtedly important for its activation, it is not essential for activation of the IKK complex by proinflammatory stimuli. The latter, however, are entirely dependent on IKKb phosphorylation (76). Another region that plays a critical role in IKK activation is the HLH motif, mutations within which severely decreased IKK activity when assayed in mammalian cells by transient transfection (55) or as recombinant proteins (68). The HLH mutants are still able to dimerize via their LZ motif (55, 68) and to bind IKKc (D Rothwarf, unpublished data). Deletion of a C-terminal fragment containing the HLH motif also results in IKK inactivation, but coexpression of this fragment (amino acids 558–756) with the truncated kinase and LZ domains (amino acids 1–559) of IKKb restored kinase activity (76). The same mutations of the HLH motif that interfered with activity of the intact kinase prevented the complementation activity of the C-terminal fragment. It was therefore proposed that the HLH region of IKKb may serve as an endogenous activator of IKK in a manner similar to the function of the cyclin subunits of cyclin-dependent kinases [CDKs (76; Figure 4, see color insert)]. The activation by the HLH region, which physically interacts with the kinase domain, requires the phosphorylation of the latter. When overexpressed in mammalian cells, a number of different protein kinases were found to activate IKK in their wild-type forms and/or inhibit IKK activation in their kinase-defective forms. These include different protein kinase C (PKC) isozymes (78) and the mitogen-activated protein kinase (K) kinase kinase (MAPKKK) family members, NIK (57, 79), AKT/PKB (80, 81), MEKK1 (82, 83), MEKK2 (84), MEKK3 (84), COT/TPL-2 (85), and TAK1 (86, 87). IKK activation by AKT/PKB may be particularly significant, as it could play a role in the anti-apoptotic pathways involving NF-KB (24, 25). NIK was identified by means of its association with TNFR associated factor 2 (TRAF2) (88) and was shown to potently activate NF-jB when overexpressed (75, 88). Expression of kinase-defective forms of NIK blocks NF-jB activation (and IKK whenever tested) in response to most inducers, including TNF-a. Consequently, NIK has been assumed to be directly involved in TNF-a–induced activation of NF-jB and has been suggested to be involved in NF-jB activation in response to other stimuli, especially IL-1. However, NIK also interacts with other TRAF proteins including TRAF3, which appears not to be involved in NF-jB activation (89). Furthermore, the region of TRAF2 to which NIK binds can be replaced with heterologous oligomerization domains, and resulting chimeric proteins, which no longer bind NIK, can still activate IKK (90). Yeast two-hybrid and protein interaction studies have indicated that NIK strongly and preferentially interacts with IKKa (57, 75), and it was suggested that NIK activates IKK through direct phosphorylation of the IKKa subunit (75). However, results from IKKa knockout studies demonstrated that IKKa is not required for IKK activation by



631

TNF-a (91, 92). In addition, no convincing data exist regarding direct interaction between endogenous IKK complexes and NIK under physiological conditions. Recently, an understanding of NIK’s physiological role came from analysis of the alymphoplasia (aly) mutation of mice. Affected mice fail to develop lymphoid organs, and, when the mutant allele was positionally cloned, it was identified as a point mutation in the NIK locus (93). aly/aly mice exhibit a phenotype similar to that of mice that are doubly deficient in TNF-a and lymphotoxin-a (LTa, also called TNF-b) (94) and mice with knockout of the lymphotoxin b-receptor (LTbR) (93), a member of the TNF receptor family. The aly mutation occurs in the TRAF-interacting region and does not affect NIK’s catalytic activity or its ability to activate NF-jB when overexpressed (93). Consistent with these results, analysis of NIK-deficient cells derived from Nik1/1 mice has indicated that NIK is not involved in IKK activation by either TNF-a or IL-1 (D Goeddel, personal communication). It remains to be determined whether NIK is involved in IKK activation in response to LTa. Apart from NIK, MEKK1 has been implicated in NF-jB activation more frequently than other MAPKKKs, for several main reasons: (a) there is generally a good (but not perfect) correlation between the regulation of JNK activity, which is partially under MEKK1 control (95), and NF-jB activity (54); (b) dominant negative MEKK1 inhibits NF-jB activation upon overexpression (54, 80); and (c) MEKK1 catalytic activity is stimulated by TNF-a through interaction with TRAF2, and the latter is capable of activating NF-jB in transient transfection experiments (90). However, the inhibitory effect of catalytically inactive mutants is shared by many MAPKKKs and may not be specific. Furthermore, studies in transgenic and knockout mice suggest that TRAF2 is not essential for NF-jB activation in response to TNF-a (96, 97). A more effective method of providing pathway-specific information about MAPKKKs would require a catalytically inactive MAPKKK that is unable to bind IKK but is capable of interacting with upstream activators. More likely, progress in assigning particular kinases to specific IKK- or NF-jB–activating pathways will depend on the use of genetic approaches. For, example, it remains to be seen whether MEKK1-deficient cells display defects in IKK or NF-jB activation in response to TNF-a, IL-1, or other stimuli. Although it appears that IKK is the most likely point of convergence for many NF-jB–signaling pathways, it is safe to predict that several physiologically relevant IKK kinases exist. IKKc, which is required for activation by TNF-a, IL-1, LPS, dsRNA, and Tax [as determined by studies in IKKc/NEMO-deficient cells (60)], may function to recruit various upstream activators (IKK kinases) to the IKK complex. In addition to upstream kinases, the phosphorylation of the activation loop can be carried out by IKK itself. IKKb and IKKa exhibit strong activity in the absence of stimulation when overexpressed in insect (61, 65, 68, 74) or mammalian (55) cells. Although it is possible that some other kinase is responsible for their activation, the simplest explanation is that IKK autophosphorylates and autoactivates. Autophosphorylation at the activation loop may

632

KARIN n BEN-NERIAH

amplify the extent of IKK activation through an upstream kinase, which, owing to low abundance or affinity, can phosphorylate only a small number of IKK molecules. Autoactivation also provides a mechanism whereby the activation of a single catalytic subunit results in the activation of all catalytic subunits within an individual IKK complex. Autophosphorylation may play a more critical role in the response to certain activators, such as the Tax protein of HTLV-1 (23). Although Tax is not a protein kinase, it can function as a dimerizer that stabilizes dimer formation in other proteins (23). Consequently Tax, which has been shown to directly associate with IKK through IKKc (72), may bring two IKK catalytic domains into close proximity and thereby induce their autoactivation. Certain upstream kinases may be found to use a similar mechanism rather than cause activation of IKK through direct phosphorylation of the activation loops.


Inactivation of IjB Kinase The mechanism of IKK inactivation is not fully understood, but it is clear that most stimuli cause only its transient activation; the stronger the initial activation, the more transient the response seems to be. Transient IKK activation or down modulation of its activity is physiologically important because persistent NF-jB activity can result in deleterious or fatal conditions, such as septic shock or acute inflammation. After stimulation by TNF-a, in most cell types IKK activity peaks within 5–15 min, and by ;30 min it decreases to ;25% of its peak value and then decreases only slightly over the next 90 min (51, 55, 76). Consistent with these kinetics, IjBa degradation is generally complete within the first 15 min after TNF-a treatment, and resynthesized pools of cytoplasmic IjBa reappear within 60 min. As discussed above, small decreases in IKK activity result in large decreases in IjB degradation and NF-jB activation. Consequently, the ability of newly synthesized IjBa to escape degradation probably results from a combination of the apparent fourfold decrease in IKK activity and the lower catalytic activity of IKK towards free IjB (68). The initial down-regulation of IKK in TNF-a–stimulated cells appears to be caused by autophosphorylation of IKK. After TNF-a treatment, IKKa and IKKb are heavily autophosphorylated at their C-terminal regions (76). Replacement of 10 autophosphorylated serines by alanines in the C-terminal portion of IKKb prolonged IKK activation by nearly fourfold. Conversely, replacement of these serines by glutamic acid, which mimics phosphoserine, resulted in greatly diminished TNF-a–induced IKK activity (76). Because inhibition requires multiple phosphorylation events that may occur more readily after IjBs, which compete for IKK phosphorylation, have been depleted, the C-terminal autophosphorylation can serve as a timing mechanism to limit the duration of strong IKK activation. Hyperphosphorylation may result in a conformational change of the C-terminal region, which contains the HLH motif, leading to a decrease in kinase activity (76; see Figure 4). An alternative possibility is that IKK is down-regulated by a phosphatase that is recruited to the hyperphosphorylated C-terminal region. In


633

addition, it is possible that the actual mechanism of IKK down-regulation is a combination of these two processes. A conformational change decreases the efficiency of activation loop autophosphorylation, thus rendering the kinase more sensitive to phosphatase action. A significant role for constitutive phosphatases in IKK down-regulation is suggested by the ability of the phosphatase inhibitor okadaic acid to activate IKK in mammalian cells (51). Whether an inducible phosphatase could also play a role in IKK inactivation is a matter for future studies.


Physiological Function of IKK Given the high sequence similarity between IKKa and IKKb and the fact that recombinant forms of both kinases can effectively phosphorylate IjB in vitro (61, 65, 68, 74), it seemed likely that the two kinases may have functionally redundant roles in NF-jB activation. The presence of both protein kinases in a single complex makes it difficult to determine whether they have distinct functions. This issue seems less complex for IKKc/NEMO. It appears to be the only stoichiometric regulatory subunit of the IKK complex, and it is clear that IKKc/NEMOdeficient cells are fully defective in IKK and NF-jB activation in response to numerous stimuli, including TNF-a, IL-1, dsRNA, LPS, and Tax (60). Analysis of IKKa/b activation loop phosphorylation strongly suggested that only IKKb plays a critical role in the response to proinflammatory stimuli and that IKKa may serve a different function (76). However, the functional distinction between the two catalytic subunits became clear only after the generation of IKKa- and IKKb-deficient mice through gene-targeting experiments. The phenotype of the IKKa-deficient mice was completely unexpected. Ikka1/1 mice die within 4 hours after birth (91, 92). Superficially, newborn Ikka1/1 mice appear to lack limbs, tails, and ears. They also exhibit severe craniofacial deformities, and their skin is smooth, taut, and shiny. Further analysis revealed that Ikka1/1 mice contain most of the skeletal elements of the limbs and tail, as well as apparently normal limb musculature, but cannot emerge out of the thickened skin. Whereas the proximal limb bones were almost normal in size, some of the distal elements (digits) were missing, and others were fused. This aberration is probably caused by an earlier defect in interdigital apoptosis that results in formation of webbed limb buds (91). The most dramatic cellular defect in Ikka1/1 mice is in formation and organization of the epidermis. In the mutants the epidermis is #5- to 10-fold thicker than normal, a defect that is caused by excessive proliferation of the basal layer. In addition, there is almost a complete absence of epidermal differentiation, and, instead of having well-stratified epithelium composed of four cell types, the epidermis of Ikka1/1 mice is more-or-less uniform. Owing to the absence of the two outer layers, the stratum corneum and stratum granulosum, the mutant epidermis is very sticky, thereby explaining why the limbs are fused to the body in Ikka1/1 mice. Thus IKKa plays a major role in epidermal differentiation.


634

KARIN n BEN-NERIAH

A second major outcome of these studies was that IKKa is not essential for IKK activation by proinflammatory stimuli. Stimulation with TNF-a, IL-1, or LPS resulted in normal IKK activation and IjBa degradation in Ikka1/1 embryonic fibroblasts, primary keratinocytes, and liver tissue (91). NF-jB binding activity is, however, reduced by ;50% in Ikka1/1 embryonic fibroblasts, suggesting that IKKa plays some as yet undetermined role in NF-jB activation (91). Although these results indicate that IKKa is not essential for activation of NFjB by proinflammatory stimuli, it appears that IKKa is critical for IjB-dependent NF-jB activation in response to an as yet unidentified developmental signal that triggers keratinocyte differentiation. NF-jB appears to be activated during keratinocyte differentiation, such that p65/RelA is cytoplasmic in basal cells but nuclear in their differentiated derivatives (98) (Y Hu, unpublished data). This process could be coupled to IjBa degradation, because the abundance of IjBa is lower in differentiated keratinocytes compared with basal cells. In Ikka1/1 mice, however, IjBa is highly abundant throughout the epidermis, and p65 remains cytoplasmic (Y Hu, unpublished data). The signals that control this process are unknown, but this developmental control of NF-jB activity resembles the control of Dorsal activity in D. melanogaster (99). One of the target genes for Dorsal is twist (100, 101). Defects in mammalian twist lead to craniofacial, limb, and skeletal abnormalities in humans (102, 103). Reduced Twist expression was also observed in Ikka1/1 fetuses (92). However, it is not clear to what extent the decreased Twist expression contributes to any of the morphogenetic defects in Ikka1/1 mice. Despite the defects in developmental NF-jB activation, Ikka1/1 keratinocytes are highly responsive to IL-1 and TNF-a and mount a very efficient NF-jB activation response to these stimuli (Y Hu, unpublished data), which is most likely mediated by IKKb. Fewer surprises were generated by disruption of the Ikkb gene. IKKb-deficient embryos die at embryonic day (E) 12.5–14.5 as a result of massive liver apoptosis (69–71). An essentially identical phenotype is presented by mice deficient in p65/ RelA, which die at E14.5 (104) or mice that are doubly deficient in both p65 and p50, which die at E12.5 (105). p65 expression protects liver cells from TNF-a– induced apoptosis, and mice that lack both TNF-a and p65 are viable and have normal livers (106). For unknown reasons, embryonic liver expresses copious amounts of TNF-a, which in the absence of NF-jB activity triggers massive apoptosis. As expected, IKKb-deficient cells are more sensitive to TNF-a– induced apoptosis than are wild-type cells (69, 71), and cells exhibit very little IKK or NF-jB activation in response to TNF-a or IL-1 (69–71). Even Ikkb`/1 cells, which exhibit a twofold reduction in IKKb expression, exhibit a 50% decrease in kinase activity and a 70%–80% decrease in NF-jB activation (70). Undoubtedly, IKKb is the catalytic subunit that is required for NF-jB activation in response to proinflammatory stimuli. Collectively, the results of the IKKa and IKKb knockout studies demonstrate that, even though the predominant form of IKK is a heterocomplex containing both IKKa and IKKb, homodimeric complexes can function effectively.


635

Ikka1/1 mice have apparently normal livers and spleens, whereas Ikkb1/1 mice, although defective in response to proinflammatory stimuli, exhibit none of the skin or skeletal abnormalities of the Ikka1/1 mice. Thus far, Ikka1/1 and Ikkb1/1 mice appear to have no common phenotype that could provide a functional explanation for IKKa/b heterocomplex formation.


UPSTREAM TO IKK KINASES—SIGNALING BY PROINFLAMMATORY CYTOKINES TNF-a and IL-1 are important mediators of inflammation and potent activators of NF-jB, exerting their effects through interaction with cell surface receptors. Although TNF-a has two distinct receptors, TNFR1 and TNFR2, in most cell types NF-jB activation occurs primarily through TNFR1 (reviewed in 107–109). Similarly, of the two known IL-1 receptors, only IL-1R1 participates in cellular signaling (110; reviewed in 111). Whereas the biological functions of TNF-a and IL-1 are remarkably similar, the two proinflammatory cytokines and their receptors are not related in sequence. Binding of ligand to either receptor induces receptor clustering and recruitment of signaling proteins to their cytoplasmic domains. TNFR1 interacts with TNFR1-associated death domain protein (TRADD), which functions as an adapter for the recruitment of other proteins, including receptor-interacting protein (RIP), a serine/threonine kinase, and TRAF2 (reviewed in 107–109). RIP is required for NF-jB activation in response to TNF-a but not IL-1 or LPS (112). IL-1R1 signaling requires its association with the membrane-spanning accessory protein (ACP) IL-1RACP (113, 114). This complex along with the adaptor protein MyD88 recruits a serine/threonine IL-1R–activated kinase (IRAK) (114–116). IRAK then participates in the recruitment of a protein related to TRAF2 called TRAF6 (116, 117). Thus the common feature in signaling through TNFR1 and IL-1R is the use of TRAF proteins. TRAF2 and TRAF6, like other TRAF family members (reviewed in 108, 118, 119), contain a common region, the TRAF domain, in their C-terminal half. The TRAF domain can be further divided into two subdomains: TRAF-C, which is highly conserved and has been found to mediate hetero- and homodimerization of TRAF proteins, as well as interactions with receptors, and TRAF-N, which is more variable and has a coiled-coiled structure (120). In addition, most TRAF proteins contain an N-terminal RING finger and several zinc finger motifs. Overexpression of either TRAF2 or TRAF6 results in NF-jB activation, and the Nterminal region is required for this activity (89, 117, 121–123). Recently, the N-terminal portions of TRAF2 and TRAF6 were shown to be sufficient for activation of downstream effectors when fused to a foreign protein domain that could be inducibly oligomerized (90). Oligomerization of these chimeras resulted in efficient IKK and NF-jB activation, which depended on integrity of the Nterminal region, especially the RING finger (90).


636

KARIN n BEN-NERIAH

The unique role of individual TRAF family members in specific pathways is not thoroughly defined. Cells isolated from TRAF2 knockout and transgenic mice indicate that TRAF2 is not required for NF-jB activation in response to TNF-a (96, 97). One possible explanation for this observation is that other TRAF proteins, such as TRAF5 (124), may substitute for TRAF2. In cells obtained from TRAF6 knockout mice, NF-jB activation was severely impaired in response to IL-1 or LPS treatment (125). However, no defects in NF-jB activation in response to TNF-a treatment were observed in these cells. The requirement of TRAF6 for LPS-induced activation of NF-jB suggests that TRAF6 is deployed by members of the Toll receptor family (125). The ability of individual TRAF proteins to mediate NF-jB activation through interaction with a variety of receptors is further demonstrated by studies carried out using Traf61/1 mice, which indicate that TRAF6 is required for signaling through CD40 and RANK, two members of the TNFR family (125). The mechanism by which TRAF activates IKK is not understood, although presumably clustering of the N-terminal domain forms a recognition surface for a downstream signaling protein. It was demonstrated that MEKK1 interacts with the effector domain of TRAF2 in an oligomerization-dependent manner, but this event was suggested to be important for activation of the JNK cascade rather than IKK (90). Other potential targets include germinal-center kinase (GCK) (126), GCK-related kinase (GCKR) (127), and ASK1 (128, 129). However, ASK1 and GCK were found not to interact with the N-terminal effector domain (90). Again, it is interesting that all of these kinases are members of the MAPKKK group. This fact, coupled with the observations, as discussed above, that many MAPKKKs are able to activate IKK and that the activating phosphoacceptor sites of IKKa and IKKb are similar to those in MAP2Ks (76), supports the notion that members of this group are the missing pieces in the cytokine-inducible NF-jB signaling pathways, connecting TRAF proteins directly to IKK.

REL/NF-jB ACTIVATION IN DROSOPHILA The NF-jB signaling pathway is evolutionarily conserved, and the role of Drosophila Rel/NF-jB proteins in development is well characterized (reviewed in 99, 130). However, understanding of the specific signaling events regulating activation of the Drosophila Rel/NF-jB proteins Dorsal, Relish, and Dif lags behind the knowledge of NF-jB activation pathways in mammalian systems. The signaling pathway for Dorsal activation is the most thoroughly studied (reviewed in 131, 132) and is remarkably similar to the IL-1 signaling pathway. An extracellular ligand binds to Toll, a transmembrane protein whose intracellular domain is homologous to that of the IL-1 receptor, resulting in the recruitment of the IRAK homolog Pelle, through a novel adapter protein, Tube. This process results in phosphorylation and degradation of the IjB protein Cactus, in a manner thought to be similar to the IjBa degradation process (133). A component of the E3



637

ubiquitin ligase that is involved in degradation of phosphorylated Cactus is encoded by the Slimb gene (134). As of yet no Drosophila homologs of IKK have been reported. It is also not clear which is the target for Pelle’s kinase activity, because the kinase activity of its mammalian homolog IRAK is not required for IL-1 signaling (62, 135, 136; N Suzuki, unpublished data). Recently, a family of mammalian Toll-like proteins has been identified (reviewed in 137). There is growing evidence that these proteins function as signaling coreceptors for LPS and other structural components of bacterial and fungal cell walls (137, 138). This function is related to the second function of Drosophila Toll, which is the activation of innate immune responses (138). The signaling pathway involved in activation of innate immune responses via Toll is essentially identical to the pathway that activates Dorsal in response to the developmental cue. Nuclear translocation of Dif, like Dorsal, depends on the degradation of Cactus (139). However, activation of Dif is independent of Pelle and Tube, because Dif translocates normally in fat-body cells from Toll1/1 and pelle1/1 larvae (139). Virtually nothing is known about the processing or activation of Relish, a p100/ p105 homolog (140).

UBIQUITINATION AND DEGRADATION OF IjB Ubiquitin-Mediated Protein Degradation The ubiquitin proteolysis system is probably the busiest proteolytic system in the body. It was originally thought to degrade old, damaged, misfolded, or misassembled proteins, but this system recently has been implicated in controlling the abundance of many functional regulatory proteins, including oncoproteins, transcription factors, cell growth modulators, signal transducers, and cell cycle proteins. The system has acquired its name from the small ubiquitinously expressed protein, ubiquitin, which is highly conserved in all eukaryotes, from yeasts to humans. A requisite step for substrate degradation in this system is the covalent attachment of multiple ubiquitin molecules to the selected substrate. Degradation of a protein via the ubiquitin pathway proceeds in three discrete and successive steps: (a) covalent attachment of one or several ubiquitin polypeptides to the protein substrate, (b) ubiquitin-ubiquitin conjugation to create ubiquitin polymers, and (c) degradation of the ubiquitin-tagged protein by the 26S proteasome complex. The ubiquitin pathway consists of several components acting sequentially in hierarchical mode. A concerted, two-step reaction results in a high-energy thioester linkage between ubiquitin and a conserved ubiquitin-activating enzyme, E1. Through a trans-acetylation reaction, the ubiquitin chain is then transferred to one of several ubiquitin-conjugating enzymes, called E2s. These enzymes collaborate with E3 protein-ubiquitin ligases to attach ubiquitin molecules to the eamino group of lysine residues of the substrate, creating a reversible isopeptide bond. In successive reactions, a polyubiquitin chain is synthesized through the


638

KARIN n BEN-NERIAH

transfer of additional activated ubiquitin moieties to the previously conjugated ubiquitin molecule. This chain then serves as a recognition marker for the proteasome (reviewed in 141, 144). A single E1 activates ubiquitin, whereas many species of E2 enzymes have been characterized in yeasts (11 members), plants, and animals (145). The multitude of E2 enzymes indicates that they specialize in distinct ubiquitination processes. However, the biochemical basis for this putative specialization has not been elucidated. Whereas the E2 proteins bear a significant homology to each other, by which an E2 can be distinguished from the relatively few E3s that have been described, the latter form a highly heterogenous group. Some E3s are associated with large multisubunit complexes, and it is not clear which subunit of these complexes is responsible for the ubiquitin-protein ligase activities. Even the functional definition of an E3 may be confusing because of the variety of mechanisms by which an E3 can promote substrate ubiquitination. By definition, substrate recognition is a dedicated task of an E3. In some cases E3 accepts activated ubiquitin from an E2 while creating a thioester intermediate before transferring it to the substrate, but, more commonly, an E3 assists in transferring ubiquitin directly from E2 to the substrate, by bringing the two into close proximity. Therefore, the emerging definition of an E3 is a mechanistic one: a protein or a protein complex that binds, directly or indirectly, specific protein substrates and promotes the transfer of ubiquitin from an enzyme-thioester intermediate (E2 or E3) to amide bonds within proteins. According to this definition, five major functional classes of E3 can be distinguished. E3a/Ubr1 E3a/Ubr1, the first discovered E3, recognizes ‘‘N-end rule’’ proteins with basic or bulky hydrophobic residues at their N terminus (reviewed in 146). Although E3a/Ubr1 is one of the better characterized E3s, little is known about its physiological substrates. Homologous with E6-AP Carboxy Terminus Domain E3s (Hect) A large protein family (at least 30 mammalian members) containing the 350-aa domain that is homologous with the E6-AP carboxy terminus (Hect) and a conserved active cystein near the C terminus. These E3 proteins have a variable N-terminal domain that, with the exception of E6-AP, anchors directly to the substrate (147). In some cases the recognition function of an Hect protein is attributable to a protein-protein interaction module, the WW domain, a 38–40 amino acid structure containing hydrophobic binding pocket for a PPxY peptide (the PY motif). This motif is found within several Hect-E3 substrates, such as the epithelial sodium channel complex, which is a substrate for Nedd4 (148). Interestingly, the prototypic HectE3, E6-AP, interacts with one of its targets, p53, indirectly through the papillomavirus oncoprotein E6 (147). In contrast to E3a and probably most other E3s, the action of E6-AP involves an intermediary thioester transfer reaction, in which the E3 protein first accepts ubiquitin from E2 and then transfers it to the substrate by facilitating an amide linkage between ubiquitin and the substrate protein. The


639


conserved carboxy-terminal cysteine forms the ubiquitin acceptor site, and its substitution abolishes thioester transfer of ubiquitin from the E2 and consequently the ubiquitination activity of the Hect-E3 ligase (149). Anaphase-Promoting Complex (APC) Anaphase-promoting complex (APC) is a large multisubunit complex also called the cyclosome (C), which is mainly responsible for degradation of proteins regulating the late events in mitosis. The vertebrate APC is a 20S core complex with a minimum of 8 subunits (reviewed in 150, 151). This complex, in contrast to Hect-E3, does not appear to have a catalytic function by itself, but operates in concert with a specific E2 in the ubiquitination of target proteins. APC/C seems to recognize and bind to its substrates, although the identity of the substrate binding subunit has not been determined. Furthermore, the nature of the APC/C substrate recognition motif is still elusive (152, 153). For some substrates (e.g. certain cyclins) a 9-aa degenerate peptide motif called the destruction box is essential, but whether it serves as a direct recognition motif is not clear. The APC/C complex, in contrast to other E3s, is inactive during interphase and becomes activated at metaphase and early anaphase. The mechanism(s) by which APC/C is activated is rather vague. It involves both phosphorylation and dephosphorylation events, affecting two different regulatory subunits. One of these APC/C subunits, Cdc20, is activated by mitotic CDKs, which are necessary for destroying budding yeast anaphase inhibitors like Pds1. On the other hand, CDK phosphorylation inactivates the other major APC/ C regulator, Hct1, which is necessary for the destruction of mitotic cyclins, thereby allowing exit from mitosis (150, 151). The activation of Hct1 is mediated by a phosphatase, Cdc14, which reverses CDK-mediated phosphorylation, allowing Hct1 to associate with APC/C and stimulate it (154). Skp1–cullin–F-Box (or Skp1–cullin–ROC1/Rbx1/Hrt1–F-Box) System The SCF (or SCRF) ligase family, another multisubunit ligase system, uses several common subunits and one variable component, an F-box protein, that functions as the substrate recognition subunit of the complex. The SCF-E3 ligases were first discovered and have been mainly characterized in yeast. Genes encoding certain SCF subunits are essential to cell cycle progression, and mutations in the different subunits result in a similar phenotype of cell cycle arrest, supporting the view that they are acting in concert (reviewed in 155, 156). Many substrates of these E3 ligases have one common feature, phosphorylation that is a prerequisite for being recognized by the ligase. Similarly to APC/C, the SCF-E3 ligases have no apparent catalytic function of their own, but promote substrate ubiquitination through a recruited E2. With the exception of the variable F-box proteins (Fsubunit), the function of other SCF subunits has been only partially resolved. At least one subunit, Skp1, likely serves as an adapter that links the F-box protein to the rest of the complex. The other subunits, Cull and the newly discovered subunit ROC1/Rbx1/Hrt1, may serve as adapters for recruiting an E2 to the substrate. Yet, it is possible that the latter subunits have other functions, associated

640

KARIN n BEN-NERIAH


with the polymerization of the ubiquitin chain, which requires E2, but not the Fbox protein (157, 158). Notably, two of the APC subunits, APC2 and APC11, are homologous to Cul1 and ROC1, respectively, and may therefore perform a similar function in the ubiquitination process. In fact, a particular motif called the R-box (a RING finger, small metal-binding domain) is common to APC2 and ROC1. This motif is also shared among other E3s that do not belong to the multisubunit ligases, such as the N-end rule Ubr1 ligase and Mdm2, which binds p53 and targets it for ubiquitin-mediated degradation (and therefore complies with the definition of an E3). In this respect, the only ligases devoid of an R-box are the Hect E3s. Von Hippel-Lindau–Associated Elongins C and B Related to the SCF ligases is the von Hippel-Lindau (VHL)–associated elongins C and B (VCB) complex, which has not formally been proven to function biochemically as an E3 but is likely to do so. Similarly to the SCF, it is composed of several subunits, one of which, ROC1/Rbx1, is shared with SCF (159). Its substrate recognition subunit is VHL, the product of a tumor suppressor gene that is missing in the von HippelLindau cancer syndrome and in some renal cell cancers (160). A recently recognized substrate for VHL is the hypoxia-inducible factor (Hif1a), which binds VHL and is degraded in vivo in the presence of the VHL complex, but escapes degradation in VHL-deficient cells or in hypoxia (160). VHL associates with an adapter system that is similar to that of SCF ligases, where the VHL-associated proteins elongin C and elongin B are homologous to Skp1 and ubiquitin, respectively, and the elongin B/C partner, Cul2, is homologous to Cul1 (161). Rbx1 also associates with Cul2 and VHL, as it does with Cul1 and the F-box protein of SCRF (162).

Inducible IjB Degradation NF-jB activation represents a striking example of the targeting of a regulatory protein, IjB, by the ubiquitin-proteasome system, as an integral step of a signal transduction cascade. With a few exceptions, such as UV irradiation, all stimuli that induce NF-jB activation, target the ‘‘professional IjBs’’ (IjBa, IjBb, IjBe) for IKK-regulated degradation through the ubiquitin system. Substitution of any of the IKK phosphorylation sites (Ser 32 or Ser 36 of IjBa or the equivalent sites of the other IjBs) renders IjB resistant to degradation in vitro and in vivo (33, 34, 36, 42, 163). Likewise, IjBa can be stabilized in vivo by mutating two specific lysine residues, Lys 21 and Lys 22, which likely serve as sites of ubiquitin conjugation (36, 43, 44). Under physiological conditions, IjB associates with NF-jB and maybe other proteins, forming a complex whose apparent molecular mass is ;450 kDa (164). This complex may protect IjB from fortuitous IKKindependent ubiquitination and degradation, as well as render it a better substrate for IKK phosphorylation (68). Indeed, whereas in vivo, the NF-jB–associated,



641

nonphosphorylated inhibitor is relatively stable, exogenously expressed IjBa, in excess of endogenous NF-jB, is rapidly degraded (165, 166). How then is IjB recognized by the ubiquitin system? Is it by a phosphorylation-induced conformational change, such as the exposure of a sequestered binding and/or conjugation site, or is it the consequence of direct recognition of the modification site? When present in a complex with NF-jB, phosphorylated but not unphosphorylated IjBa can be ubiquitinated in a cell-free extract. Using peptides corresponding to the N-terminal regulatory domain of IjBa, Yaron et al (164) showed that the targeting component of the IjBa-E3 can be titrated by peptides phosphorylated at both serine 32 and 36. The efficacy of singly phosphorylated peptides was reduced $20-fold, relative to the doubly phosphorylated ones, whereas nonphosphorylated peptides had no effect on the ubiquitin ligase. Peptides lacking the lysine residues that correspond to the major sites of ubiquitin conjugation were still effective inhibitors of ubiquitination, implying that the site of IjBa recognition by the ubiquitin ligase is distinct from the conjugation site. Furthermore, immobilizing the inhibitory peptides on a solid support creates an affinity resin that can deplete cell lysates of their IjBa-ubiquitinating activity, without affecting the ubiquitination of other cellular proteins. These studies therefore indicated that a specific ubiquitin ligase recognizes phosphorylated IjBa (pIjBa) through a short peptide stretch, centered around the two inducibly phosphorylated serine residues. This sequence, composed of 6 aas, DS(PO3)GXXS(PO3), is conserved among all of the IjBs from drosophila to humans (164), representing a well-defined E3 recognition motif. A similar motif is present in the short-lived signal-transducing protein b-catenin (167), and mutations of any of the conserved residues within this recognition site result in stabilization of both IjBs and b-catenin (33, 34, 36, 37, 42, 163, 168–177). A lysine residue, located 9–12 amino acids N-terminal to the recognition site, is also conserved between IjBs and b-catenin, suggesting that a single enzyme mediates both the recognition and the conjugation of ubiquitin to these substrates via two functional sites, which may reside in one or two distinct proteins.

Identification of the IjB-E3 Phosphorylated IjBa (pIjBa) complexes have an even larger apparent molecular mass, 700 kDa, than complexes containing nonphosphorylated IjBa (164). One intriguing explanation for the increased mass of the pIjBa complexes is the recruitment of components of the ubiquitin machinery. Indeed an essential component of the pIjBa-E3 is recruited onto the phosphorylated IjBa substrate through a specific and direct interaction with the E3 recognition motif described above (39). Using advanced protein purification and sequence determination technology, the recognition component of the pIjB-specific E3 was recognized as the human F-box/WD protein b-TrCP, renamed E3RSIjB (E3 receptor subunit of IjB) (39). b-TrCP was first identified in xenopus as an orphan WD protein (178). It was only recently identified as an F-box protein, bearing 30% similarity to yeast


642

KARIN n BEN-NERIAH

Cdc4p, a protein specializing in ubiquitin-dependent degradation of cell cycle inhibitors (reviewed in 157). A protein similar to b-TrCP, Slimb, was recently isolated by genetic screening as an enhancer of the Wingless and Hedgehog signaling pathways in drosophila (134). Slimb mutations stabilize armadillo/bcatenin and the transcriptional regulator Ci (134). Human E3RSIjB/b-TrCP interacts with the HIV protein Vpu, inducing the degradation of CD4 on HIVinfected T cells (179). From these initial findings, an avalanche of papers have described E3RSIjB/b-TrCP as an essential component of an IjBa ligase (180– 185). To qualify as the substrate recognition component of the IjB ligase, an E3RSIjB protein should bind pIjBa specifically and assist in its ubiquitination, parameters that were examined in vivo and in cell-free systems. E3RSIjB associates with phosphorylated IjBa, both in mammalian cells and yeast (by a twohybrid system), and the two proteins can be coimmunoprecipitated from stimulated cells (180–184). Human and mouse E3RSIjB and, to some extent, even drosophila Slimb, can bind to IKK-phosphorylated IjBa (39) and to pIjBa peptides (184). pIjBa binding is abrogated by a peptide representing the pIjBrecognition motif, but not by the nonphosphorylated peptide, showing that the site of pIjBa recognition is the conserved degradation motif (39). An F-box deletion mutant of E3RSIjB binds pIjBa even better than the original protein, indicating that the binding site of the E3 receptor is likely contained in the WD40 repeat region, a protein-protein interaction motif (39). This motif is present in .140 different proteins and, from the crystal structure of one, b-transducin (186– 187), is presumed to display a b-propeller structure. The fact that this motif may accommodate a short phosphopeptide suggests that it may unexpectedly contain a pocket adapted for interaction with phosphate moieties. As expected, the wild-type mouse and human E3RSIjB and its associated proteins, in the presence of E1 and E2, induced the ubiquitination of pIjBa, but not of the nonphosphorylated IjBa (39, 180, 183, 184). Slimb, the Drosophila homolog, facilitated some pIjBa ubiquitination as well, but was $10-fold less efficient than the mammalian E3RSIjB (39). An E3RSIjB mutant devoid of the F-box failed to promote ubiquitination, despite its pIjBa binding capacity, indicating that the F-box plays an essential role in the ubiquitination process, probably through recruitment of the other components of the SCF complex. Indeed, many groups documented the association of both Skp1 and Cul1 with E3RSIjB (180, 181, 183, 184). Furthermore, depletion of Skp1 from a cell lysate codepletes E3RSIjB and abolishes pIjBa ubiquitination, indicating that functional E3RSIjB is associated with Skp1 (184). Like other SCF complexes, E3RSIjB appears to have no catalytic function of its own, but rather directs a specific E2 ubiquitinconjugating enzyme (UBC5) toward the substrate (39, 183; Figure 5, see color insert). Nevertheless, the biological function of the SCF complex in IjBa ubiquitination is not entirely clear. Although it could serve as an adapter system for linking the E2 to the substrate, SCF is traditionally associated with Cdc34 (UBC3), which fails to collaborate with E3RSIjB in in vitro ubiquitination of



643

IjBa (A Hatzubai & A Yaron, unpublished data). Alternatively, SCF could promote the ubiquitin polymerization that is necessary for proteasomal degradation (Figure 5). Indeed, recently IjBa-E3 was found to associate with a novel RING finger-containing SCF component, ROC1, which may promote ubiquitinubiquitin conjugation in conjunction with UBC5 (158). The F-box deletion mutant E3RSIjB is a dominant negative inhibitor of IjBa degradation and NF-jB activation in vivo (39, 181–184). This mutant stabilizes IjBa and inhibits NF-jB activation, presumably by binding pIjBa and blocking the access of the endogenous protein to its substrate. However, it may also work as a dominant negative by other means, which are not necessarily linked to its binding specificity and thus may interfere with protein ubiquitination in a nonspecific manner. The ability to block E3RSIjB activity in vitro with specific pIjBa peptides suggested that, by introducing the peptides to intact cells, it would be possible to modulate cellular NF-jB functions. Indeed, microinjection of a pIjBa peptide abolished TNF-a–stimulated NF-jB translocation into the nucleus as well as the expression of an NF-jB–dependent gene, E-selectin (164). These examples provide a proof of principle, suggesting that in vivo titration of a distinct E3 could be harnessed to modulating a specific cellular process. Phosphopeptides are not practical tools for in vivo studies, because they do not penetrate cells on their own. However, peptidomimetic or small-molecular-weight compounds that block E3RSIjB could be useful in studying its role and those of related E3 ligases in different cells under various physiological conditions. Is the E3RSIjB complex the only IjBa-E3? The ultimate answer to this question will be provided by gene-targeting experiments. Such experiments have implicated other F-box proteins in distinct proteolysis events in yeasts (reviewed in 156, 157). So far, there are enough data to implicate the E3RSIjB complex as a very competent IjBa-E3, which speaks in favor of it being the ligase, but these data mostly rely on biochemical studies. An attempt to challenge the unique pIjBa-binding specificity of E3RSIjB showed that it was the only F-box protein among eight that could bind the phosphorylated IjBa peptide (184). Furthermore, a closely related protein, E3RS2/b-TrCP2, which bears 85% similarity to E3RSIjB, failed to bind pIjBa in vitro (A Hatzubai & A Yaron, unpublished data). The efficiency of pIjBa ubiquitination in vitro lends further support to the role of E3RSIjB (39, 158, 184). Nevertheless, the best proof for the role of E3RSIjB in the inducible degradation of pIjBa will come from genetic analysis. The available data are limited to the maternal effect of a partially characterized slimb mutant on the Dorsal pathway in Drosophila embryo development and are, therefore, nonconclusive (183). At this point it is worth mentioning two exceptional stimuli that induce proteasomal degradation of IjBa, but do not involve IKK or the N-terminal phosphorylation-dependent E3 recognition motif in IjB. UV irradiation and certain chemotherapeutic drugs induce the degradation of IjBa, although with slower kinetics than those induced by most other stimuli (49, 50; A Baldwin,


644

KARIN n BEN-NERIAH

personal communication). However, IKK is neither activated in response to UV irradiation (50), nor is the inducible degradation affected by mutations at the IjB recognition motif or by expression of dominant negative IKK mutants (49, 50). The peculiar degradation process is not necessarily linked to the slow kinetics, because IjBa degradation in response to c-irradiation proceeds at the same kinetics, but involves the N-terminal phosphorylation-dependent E3 recognition motif (50). The UV-induced degradation process is inhibited by proteasome blockers (49, 50), as well as at the nonpermissive temperature in an E1-mutant cell line (50), suggesting that the proteasome, possibly together with the ubiquitin system, is responsible for UV-inducible IjBa degradation. It would be interesting to examine the effect of a dominant negative E3RSIjB in this process. If it fails to inhibit degradation of IjBa in response to UV or chemotherapeutic agents, this result would indicate either that another E3 is involved in these responses or that the proteasome is degrading IjBa in a ubiquitin-independent manner. An example of the latter possibility is provided by the proteasomal degradation of ornithine decarboxylase (ODC), which is facilitated through the action of antizyme, an ancillary protein tethering ODC to the proteasome (188).

Specificity of E3RSIjB: Targeting Proteins Other Than IjB The E3 recognition motif of the IjBs DS(PO3)GXXS(PO3) is present in other proteins. The template of this sequence, DSGXXS, is shared among nearly 1250 protein entries in the database (M Davis, unpublished data). Presumably, not every protein having this template is phosphorylated at both serines, an essential feature for E3RSIjB recognition (164). Even the substitution of phosphoserines with phosphothreonines significantly compromises E3RSIjB binding (A Hatzubai & A Yaron, unpublished data), emphasizing the stringency of the recognition process. Therefore, it is interesting that both E3RSIjB and IKK prefer serines over threonines (36), a feature that could be linked to an apparent coevolution of the two signaling systems. Interestingly, one of the IKK genes (IKKa) is genetically linked to the E3RSIjB gene, both residing at human chromosome 10q-24 (184, 189).

HIV Vpu and b-Catenin as E3RSIjB Ligands It is hard to estimate the number of potential E3RSIjB targets that are actually phosphorylated at the two serines and are therefore recognized by it. Nevertheless, it is clear that E3RSIjB can bind to at least two proteins that are nonrelated to IjB; b-catenin (184, 185, 190–192) and Vpu (179). Vpu is a small polypeptide encoded by HIV that resides in the endoplasmic reticulum membrane of infected cells. Vpu can simultaneously interact with the CD4 protein and with E3RSIjB to form a ternary complex which targets CD4 for proteasomal degradation (179). Similarly to pIjBa recognition, the interaction of E3RSIjB with Vpu occurs via the WD40 domain of bTrCP and requires phosphorylation at the two serines of the shared motif. However, there is a notable difference between targeting of



645

pIjBa and Vpu by E3RSIjB. Whereas pIjBa is targeted for ubiquitination, Vpu is apparently too small a target, perhaps lacking an available lysine next to the recognition signal. Instead, it acts as an adaptor that directs the E3-E2 complex to an associated protein, CD4. This is an interesting example of how a pathogen adapts to its host. By assimilating a specific host-protein degradation motif, HIV recruits the host ubiquitination machinery to eliminate a cellular protein. Downregulation of CD4 that is retained at the endoplasmic reticulum through interaction with Vpu could help to release the CD4-associated viral gp160 protein, thus facilitating maturation of the viral envelope. Down-regulation of CD4 would also result in a reduced number of HIV receptors, helping to prevent superinfection, which may be counterproductive for the virus. Another ligand of E3RSIjB is b-catenin. Unlike IjBa, b-catenin ubiquitination has not been demonstrated in vitro, but there is ample genetic evidence implicating E3RSIjB in the regulation of b-catenin stability. Overexpression of the Fbox–deleted E3RSIjB in xenopus and mammalian cells stabilizes b-catenins in vivo (190, 193, 194). However, these data are mostly based on the effect of dominant negative E3RSIjB rather than on E3RSIjB-deficient organisms, and they are, therefore, not conclusive. Furthermore, stabilization of b-catenins in human tumors is more often the result of mutations at sites that are adjacent to but distinct from the E3RSIjB recognition motif than of mutations affecting the recognition motif itself (168–171, 173, 174, 176, 177). Hence, targeting b-catenin for ubiquitin-mediated proteolysis appears more complex than targeting IjBs. This issue is of considerable importance when examining the option of using selective E3RSIjB blockers to inhibit NF-jB activation. b-catenin stabilization plays a critical role in human cancer, particularly in colon cancer and melanoma (168, 169). Stabilization may be the consequence of mutations in the b-catenin gene itself or in genes coding for signaling molecules that regulate its phosphorylation or degradation. In fact, the most frequent mutations in colon cancer occur at the APC gene (reviewed in 195). So far, no colon cancer-linked mutations have been detected in the gene coding for E3RSIjB (175). Furthermore, genetic abnormalities in the chromosomal region of the human E3RSIjB gene are frequent in glioblastoma (196), prostatic cancer (197), and small-cell lung cancer (198), but not in colon cancer (175). However, the loss of E3RSIjB function would result in inhibition of NF-jB activation and, given the role of NF-jB in protecting cells from apoptosis (69, 122, 199, 200), it is conceivable that E3RSIjB deficiency may not promote tumor survival and therefore would be subject to negative selection during tumor development.

Processing of the NF-jB Precursor Proteins p105 and p100 Both NF-jB1 and NF-jB2 encode precursor proteins that are much larger than the mature functional products, p50 and p52. The precursors must therefore be processed to generate the mature forms, but the precise mechanism by which accurate processing is achieved is not entirely clear. Although one study claimed


646

KARIN n BEN-NERIAH

that the processing of the NF-jB1 precursor p105 occurs cotranslationally (201), most studies support a model involving a major role for post-translational processing via the ubiquitin-proteasome system. Fan & Maniatis (202) and Palombella et al (203) have shown that the in vitro processing of an NF-jB1 precursor requires ATP and ubiquitin and is blocked by a ubiquitin mutant that prohibits the polymerization of the ubiquitin chain. Likewise, proteasome inhibitors block the processing of p105 in intact cells (203). Similar observations were made by Orian et al, who have implicated a 320-kDa E3 and two E2s (UBCH5 and UBCH7) in the in vitro processing (204). Using a similar assay, Coux & Goldberg have demonstrated in vitro processing of p105 activity, which segregates with a putative ligase having a 50-kDa mass and requires a different E2, E2-25kD (205). Therefore, although the identity of the ubiquitin enzymes that are responsible for generating p50 and p52 is not clear and awaits their further purification, the requirement for ubiquitination as part of the precursor processing is supported by most studies. The processing of p100 and p105 is mainly a constitutive event. However, certain stimulatory conditions, such as phorbol ester stimulation, facilitate the in vivo processing of both precursors (206–208). This stimulation could serve two purposes: (a) to replenish the cytoplasmic pool of the mature NF-jB subunits p50 and p52, under conditions that promote their translocation into the nucleus, and (b) to release NF-jB subunits that are bound to the precursor proteins in the cytoplasm and therefore prohibited from acting in the nucleus (209). Several studies indicated that the facilitated turnover of the precursors is mediated through their phosphorylation (206–208). A MAPKKK TPL-2 that is highly related to the human oncoprotein COT was found to associate with the C-terminus of p105 and to promote the degradation of the precursor molecule (210). However, TPL-2 does not enhance the generation of p50 from p105, but merely its release from the precursor, which undergoes degradation (210). In addition TPL-2 does not phosphorylate p105 and recently was reported to be an upstream activator of IKK (85). Thus, the effect of TPL-2 on NF-jB activity may actually be mediated via IKK. Indeed, Scheidereit and colleagues (208) showed that IKK associates with p105 and phosphorylates the NF-jB precursor at a c-terminal site resembling the IjB degradation motif; a possible hint of involvement of E3RSIjB in p105 processing. Ubiquitin-regulated proteolysis is commonly a progressive event that yields small peptides, but no partial protein fragments (211). The generation of p50 and p52 and perhaps several other events, such as the processing of the drosophila Ci transcription factor (134), are exceptional. The mechanism involved in these processes and the identity of the structural motifs that dictate the processing outcome, as well as the biochemical events, are mostly elusive. A Glu-rich region (GRR) spanning the amino-acid residues 372–394 of the mouse p105 precursor is required for processing (212). In a few cases, insertion of the GRR into unrelated proteins was sufficient to impose partial proteolysis, similar to p105 processing, suggesting that GRR may halt complete proteasomal proteolysis (212).



647

A related Gly-Ala repeat derived from the EBNA-1 protein of the Epstein-Barr virus has a similar negative effect, preventing the degradation of EBNA-1 and, consequently, its presentation as an antigen to T cells (213). Transfer of this EBNA-1 repeat to proteins such as EBNA-4 and IjBa (214), which are normally proteolyzed to completion, prevented their degradation, indicating that these repeats may ‘‘poison’’ the proteolysis system. Orian et al recently showed that the GRR does not affect the ubiquitin conjugation process, but probably interferes with proteasome-mediated proteolysis (215). It appears that the primary function of the GRR is to protect the mature p50 fragment within which it is contained. However, accurate processing depends on another stretch of amino-acids, 441– 454, which is located ;40 residues downstream of the processing site. This region contains two lysine residues that are the major ubiquitin conjugation targets of p105 (215). These findings are consistent with two alternative models that may explain p105 and p100 processing. According to the first model, processing is initiated by ubiquitination, followed by proteasomal nibbling of the p105 precursor from the C-terminal end. The essential stretch downstream of the processing site may be required for directing the ubiquitination event to a specific site on the Cterminus of p105. Once the proteasome hits the GRR, it stops nibbling the p105, and mature p50 is released. Alternatively, the downstream stretch, perhaps in conjunction with nearby ubiquitination, may be responsible for an endoproteolytic cut by a different protease. The C-terminal fragment is then destroyed by the proteasome, whereas the N-terminal fragment is spared from proteasomal digestion by the GRR. It is interesting that p105 processing may be reproduced in yeasts (216). The generation of p50 in yeast is abrogated in proteasome mutants, and the processing block is often accompanied by accumulation of ubiquitinated p105. Therefore, processing in yeast, like in mammalian cells, appears to work via the ubiquitin-proteasome pathway. The failure to generate p50 in the yeast doa4 mutant, deficient in a ubiquitin-hydrolase, supports the notion that a ubiquitinated p105 intermediate is the species that is processed by the proteasome and must therefore undergo deubiquitination during processing (216). However, the regulation of the processing in yeast is somewhat different from that in mammalian cells, because the GRR is not essential to accurate processing in yeast (216). Whether this difference is caused by the unique properties of the yeast proteasome, by different folding of p105 in yeasts or is connected to different post-translation modifications (including ubiquitination differences of p105 or its processing products) is as yet unknown.

REGULATION OF NF-jB ACTIVITY SUBSEQUENT TO IjB DEGRADATION Although degradation of IjBs is sufficient to cause NF-jB translocation into the nucleus, subsequent events can affect NF-jB’s ability to activate transcription. Whereas the presence and activity of other nuclear proteins, including transcrip-


648

KARIN n BEN-NERIAH

tion factors such as AP-1 and components of the basal transcriptional machinery, certainly contribute to NF-jB’s transcriptional activity, such considerations are beyond the scope of this review. However, it appears that phosphorylation of NFjB proteins may modulate their transcriptional activity. There are several reports of NF-jB being phosphorylated in response to activating stimuli (61, 217, 218). It has been shown that LPS induces the phosphorylation of serine 276 of p65 by protein kinase A and that this process increases the transcriptional activity of p65 by strengthening its interaction with the transcriptional coactivator CBP/p300 in the nucleus (218). It was also demonstrated that stimulation by TNF-a results in phosphorylation of serine 529 of p65 (217). Although this phosphorylation increases p65 transcriptional activity, it has no effect on nuclear translocation or DNA binding. The kinase responsible for this phosphorylation has not been identified. It was reported that IKK purified from mammalian cells phosphorylated p65 in vitro and that this activity was distinct from IjB kinase activity in that it could be dissociated from IKK by stringent washing (52). Subsequently it was found that purified recombinant IKKa/b can phosphorylate p65 (61). However, the site(s) of phosphorylation was not identified, and it is not known whether the phosphorylation of p65 by IKK is physiologically relevant. An additional modulation of NF-jB activity is suggested by observations made on IKKa-deficient fibroblasts (91). Ikka1/1 embryonic fibroblasts exhibit significantly lower levels of both basal and induced NF-jB DNA-binding activity than do wild-type cells. However, TNF-a–induced IjB degradation and in vitro IjB kinase activity were similar in wild-type and Ikka1/1 fibroblasts. The basis for this difference in DNA-binding activity is not yet known, although it may result from IKKa-mediated phosphorylation of an NF-jB component or from an effect on the processing of p105 to p50. Recently, it was reported that serine phosphorylation in the central portion of Dorsal is required for its signal-induced nuclear localization (219). Signal-induced phosphorylation occurs on serine 317 of Dorsal, while still bound to Cactus in the cytoplasm. Substitution of Ser-317 by alanine does not appear to affect signalinduced degradation of Cactus or, somewhat surprisingly, basal nuclear translocation of Dorsal. Ser-317 is completely conserved among Rel/NF-jB members, and it was suggested that this phosphorylation may be a general mechanism for regulating signal-induced NF-jB nuclear translocation (219).

TERMINATION AND DOWNREGULATION OF NF-jB ACTIVITY Negative regulation of NF-jB activity is very complex, and a variety of mechanisms are involved in both termination of NF-jB activation and its downregulation in response to specific signals. Clearly, down-regulation of NF-jB



649

activity may occur at several different levels. IKK inactivation is required to prevent or delay its reactivation, and possible mechanisms have already been discussed above. The critical inhibitory step, however, is thought to involve binding of newly synthesized IjBa (and other IjBs) to NF-jB in the nucleus. One level of regulation is the nuclear localization of IjBa. Nuclear translocation of IjBa was originally considered to be a passive process, but potential NLS regions were identified (220), and it has been demonstrated that IjBa translocation is an energy-dependent process (221). In vitro studies suggest that an as yet unidentified IjBa-specific carrier protein may be required for translocation (221). After nuclear translocation, IjBa is capable of terminating the activity of NF-jB by another mechanism, transporting NF-jB back to the cytoplasm (8, 222). This function is conferred by a leucine-rich nuclear export sequence located both in the C- and N-terminal regions of IjBa, which are homologous to the export signal of HIV Rev and the protein kinase A inhibitor PKI, and this function is recognized by the nuclear protein CRM1 (exportin-1) (223, 224). Interestingly, it was recently shown that while trapped in the nucleus due to blocking of the export system, IjBa may be subject to signal-induced degradation (224); validation in a physiological setting may challenge the current model of NF-jB activation, which views the cytoplasmic degradation of IjBa as the major step in NF-jB activation. Another aspect of NF-jB down-regulation by IjBs is the phosphorylation of residues in the C-terminal region of IjB. Although it appears that constitutive phosphorylation of IjBa is involved primarily in regulating the half-life of the protein in unstimulated cells (225–227), constitutive C-terminal phosphorylation of IjBb appears to play a much more important role in NF-jB regulation, specifically in controlling persistent NF-jB activation. Prolonged exposure to certain stimuli, such as LPS, leads to long-term induction of NF-jB activity, despite high levels of newly synthesized IjBa. It was suggested that the presence of newly synthesized unphosphorylated IjBb in the nucleus, which, in contrast to Cterminally phosphorylated IjBb, does not inhibit the DNA-binding ability of NFjB or mask its NLS, may protect NF-jB from inhibition by IjBa (228, 229). For IjBb to function in this manner, it must remain bound to DNA-tethered NFjB complex, and experimental results support this conclusion (229–232). Structural studies of the NF-jB:IjBa complex suggest that sequence differences in the sixth ankyrin repeats of IjBa and IjBb are probably significant enough to account for this functional difference between IjBa and IjBb (16). It should be noted, however, that the replacement of IjBa coding sequences with those of IjBb did not result in any pathophysiological aberrations or changes in NF-jB regulation (11). Thus, there is little genetic support for any substantial difference in the physical properties of these two IjB proteins.

650

KARIN n BEN-NERIAH


IMMUNE AND MEDICAL IMPLICATIONS OF IjB SIGNALING The discovery of the two complementary signaling complexes, IKK and the IjB ligase complex, may be of considerable interest, not only for understanding the mechanism of regulated gene expression but for the opportunity to alter immune and inflammatory processes and intervene in serious medical conditions. Modulation of NF-jB activity may be a superior treatment modality for a whole range of diseases, from bronchial asthma to AIDS (reviewed in 19, 233, 234), but the advantage of such interventions should be assessed against the risk of compromising an essential function of NF-jB (such as protection of cells from apoptosis), as well as affecting other signaling systems that rely on a similar kinase or ubiquitin ligase. Gene-targeting experiments have indicated that NF-jB deficiencies could affect the immune response, but the most serious effect is on the embryonic liver, which is demolished in the absence of NF-jB or IKKb at a defined stage of embryonic development (69, 70, 104). Our, recent study indicates that, in contrast to the critical role of NF-jB in fetal liver, the adult liver may cope better with NF-jB deficiency (I Lavon & Y Ben-Neriah, unpublished data). NF-jB cannot be induced in repatocytes of transgenic mice that express a liver-specific, regulatable, dominant IjBa. However, unless challenged with a serious infection or an immune system activator such as concanavalin A, these mice are mostly healthy, showing minimal signs of liver dysfunction. The reason for the extreme sensitivity of fetal liver to the absence of NF-jB is the high level of localized TNF-a synthesis during this stage of development. The loss of TNF-a expression completely protects RelA-deficient mice from liver apoptosis during fetal development (106). However, owing to a severe reduction in NF-jB function, the resulting RelA1/1, Tnfa1/1 mice are extremely sensitive to opportunistic infections and die within weeks of birth. The phenotypic similarity between RelA and IKKb knockout mice indicates that IKKb is probably dedicated to phosphorylating IjB, and that inhibiting IKKb would probably not impose on the adult animal further risks than those of the NF-jB inhibition per se. However, the IKKa knockout phenotype is more problematic concerning the feasibility of targeting IKK by drugs. It is unclear whether the primary defect in these mice is linked to the failure to phosphorylate IjB or another protein that plays a role in the differentiation of epidermal progenitors and perhaps in limb development as well (see above). The turnover of epidermal keratinocytes is a daily process, and, if compromised by IKK inhibition, it would limit the clinical utility of IKK inhibitors. Obviously, E3RSIjB is an even less certain target for drug intervention, considering the limited knowledge about its specificity. As mentioned above, at the moment the most serious consideration is related to b-catenin stabilization when blocking E3RSIjB function. Furthermore, there are as yet no available data on the consequences of blocking E3RSIjB in


651

vivo. The only relevant data in this respect may be derived from future gene knockout experiments. Finally, in many instances brief inhibition of NF-jB might be sufficient to revert a pathological condition, such as during an acute asthma episode (19). In fact, glucocorticoids, which are the most efficacious antiasthma drugs, are potent inhibitors of NF-jB activity in lymphoid cells (235). In such cases, short-term b-catenin stabilization may not impose that much risk, because tumorigenicity most likely requires a prolonged increase in b-catenin expression.


ACKNOWLEDGMENTS Research was supported by grants from the National Institutes of Health (R37 ES04151, R01 AI43477, and R01 CA76188) and Department of Energy (DEFG03-86ER60429) to M. K., and the Israeli Science Foundation Centers for Excellence Program and the Deutch-Israelische Projektkooperation (DIP) to Y. B-N. We thank David Rothwarf for many helpful discussions and help with writing this manuscript, and Asne Bauskin, Irit Alkalay, Allan Bar-Sinai, and Avraham Yaron for thoughtful comments. M. K. is the Frank and Else SchillingAmerican Cancer Society Research Professor. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Sen R, Baltimore D. 1986. Inducibility of the immunoglobulin enhancer-binding protein NF-jB by a posttranslational mechanism. Cell 47:921–8 2. Baldwin AS. 1996. The NF-jB and IjB proteins: new discoveries and insights. Annu. Rev. Immunol. 14:649–81 3. Ghosh S, May MJ, Kopp EB. 1998. NFjB and Rel proteins: evolutionarily conserved mediators of immune responses. Annu. Rev. Immunol. 16:225–60 4. Miyamoto S, Verma IM. 1995. RE1/NFjB/IjB story. Adv. Cancer Res. 66:255– 92 5. Siebenlist U, Franzoso G, Brown K. 1994. Structure, regulation and function of NF-jB. Annu. Rev. Cell. Biol. 10:405– 55 6. Attar RM, Caamanõ J, Carrasco D, Iotsova V, Ishikawa H, Ryseck RP, Weih F, Bravo R. 1997. Genetic approaches to

7.

8.

9.

10.

study Rel/NF-j B/I j B function in mice. Semin. Cancer Biol. 8:93–101 Whiteside ST, Israe¨l A. 1997. I j B proteins: structure, function and regulation. Semin. Cancer Biol. 8:75–82 Arenzana-Seisdedos F, Turpin P, Rodriguez M, Thomas D, Hay RT, Virelizier JL, Dargemont C. 1997. Nuclear localization of IjBa promotes active transport of NF-jB from the nucleus to the cytoplasm. J. Cell Sci. 110:369–78 Klement JF, Rice NR, Car BD, Abbondanzo SJ, Powers GD, Bhatt PH, Chen CH, Rosen CA, Stewart CL. 1996. IjBa deficiency results in a sustained NF-jB response and severe widespread dermatitis in mice. Mol. Cell. Biol. 16:2341–49 Beg AA, Sha WC, Bronson RT, Baltimore D. 1995. Constitutive NF-jB activation, enhanced granulopoiesis, and neonatal lethality in IjBa-deficient mice. Genes Dev. 9:2736–46


652

KARIN n BEN-NERIAH

11. Cheng JD, Ryseck RP, Attar RM, Dambach D, Bravo R. 1998. Functional redundancy of the nuclear factor jB inhibitors (IjBa) and (IjBb). J. Exp. Med. 188:1055–62 12. Chen FE, Huang D-B, Chen Y-Q, Ghosh G. 1998. Crystal structure of p50/p65 heterodimer of transcription factor NFjB bound to DNA. Nature 391:410–13 13. Chen YQ, Ghosh S, Ghosh G. 1998. A novel DNA recognition mode by the NFjB p65 homodimer. Nat. Struct. Biol. 5:67–73 14. Ghosh G, Vanduyne G, Ghosh S, Sigler PB. 1995. Structure of NF-jB p50 homodimer bound to a IjB. Nature 373:303– 10 15. Muller CW, Rey FA, Sodeoka M, Verdine GL, Harrison SC. 1995. Structure of the NF-jB homodimer bound to DNA. Nature 373:311–17 16. Huxford T, Huang DB, Malek S, Ghosh G. 1998. The crystal structure of the IjBa/NF-jB complex reveals mechanisms of NF-jB inactivation. Cell 95:759–70 17. Jacobs MD, Harrison SC. 1998. Structure of an IjBa/NF-jB complex. Cell 95:749–58 18. Baeuerle PA, Baichwal VR. 1997. NF-j B as a frequent target for immunosuppressive and anti-inflammatory molecules. Adv. Immunol. 65:111–37 19. Barnes PJ, Karin M. 1997. Nuclear factor-jB—a pivotal transcription factor in chronic inflammatory diseases. N. Engl. J. Med. 336:1066–71 20. Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD. 1998. How cells respond to interferons. Annu. Rev. Biochem. 67:227–64 21. Boehm U, Klamp T, Groot M, Howard JC. 1997. Cellular responses to interferon-gamma. Annu. Rev. Immunol. 15: 749–95 22. Roulston A, Lin R, Beauparlant P, Wainberg MA, Hiscott J. 1995. Regulation of human immunodeficiency virus type 1

23.

24. 25.

26.

27.

28.

29.

30.

31.

32.

33.

and cytokine gene expression in myeloid cells by NF-j B/Rel transcription factors. Microbiol. Rev. 59:481–505 Bex F, Gaynor RB. 1998. Regulation of gene expression by HTLV-I Tax protein. Methods 16:83–94 Baichwal VR, Baeuerle PA. 1997. Activate NF-jB or die? Curr. Biol. 7:R94–96 Sonenshein GE. 1997. Rel/NF-jB transcription factors and the control of apoptosis. Semin. Cancer Biol. 8:113–19 Gilmore TD, Koedood M, Piffat KA, White DW. 1996. Rel/NF-jB/IjB proteins and cancer. Oncogene 13:1367–78 Luque I, Ge´linas C. 1997. Rel/NF-jB and IjB factors in oncogenesis. Semin. Cancer Biol. 8:103–11 Grilli M, Memo M. 1999. NF-jB/Rel proteins: a point of convergence of signalling pathways relevant in neuronal function and dysfunction. Biochem. Pharmacol. 57:1–7 Lee SJ, Dimtchev A, Lavin MF, Dritschilo A, Jung M. 1998. A novel ionizing radiation-induced signaling pathway that activates the transcription factor NF-jB. Oncogene 17:1821–26 Foxwell B, Browne K, Bondeson J, Clarke C, de Martin R, Brennan F, Feldmann M. 1998. Efficient adenoviral infection with IjBa reveals that macrophage tumor necrosis factor alpha production in rheumatoid arthritis is NF-jB dependent. Proc. Natl. Acad. Sci. USA 95:8211–15 Barnes PJ, Adcock IM. 1998. Transcription factors and asthma. Eur. Respir. J. 12:221–34 Neurath MF, Becker C, Barbulescu K. 1998. Role of NF-jB in immune and inflammatory responses in the gut. Gut 43:856–60 Brockman JA, Scherer DC, McKinsey TA, Hall SM, Qi X, Lee WY, Ballard DW. 1995. Coupling of a signal response domain in IjB to multiple pathways for NF-jB activation. Mol. Cell. Biol. 15:2809–18


CONTROL OF NF-jB VIA IjB DEGRADATION 34. Brown K, Gerstberger S, Carlson L, Franzoso G, Siebenlist U. 1995. Control of IjBa proteolysis by site-specific, signal-induced phosphorylation. Science 267:1485–91 35. DiDonato JA, Mercurio F, Karin M. 1995. Phosphorylation of IjBa precedes but is not sufficient for its dissociation from NF-jB. Mol. Cell. Biol. 15:1302– 11 36. DiDonato JA, Mercurio F, Rosette C, Wu-li J, Suyang H, Ghosh S, Karin M. 1996. Mapping of the inducible IjB phosphorylation sites that signal its ubiquitination and degradation. Mol. Cell. Biol. 16:1295–304 37. Traenckner EB, Pahl HL, Henkel T, Schmidt KN, Wilk S, Baeuerle PA. 1995. Phosphorylation of human IjB on serines 32 and 36 controls IjB proteolysis and NF-jB activation in response to diverse stimuli. EMBO J. 14:2876–83 38. Whiteside T, Ernst MK, LeBail O, Laurent-Winter C, Rice N, Israel A. 1995. N- and C-terminal sequences control degradation of MAD3/IjBa in response to inducers of NF-jB activity. Mol. Cell. Biol. 15:5339–45 39. Yaron A, Hatzubai A, Davis M, Lavon I, Amit S, Manning AM, Andersen JS, Mann M, Mercurio F, Ben-Neriah Y. 1998. Identification of the receptor component of the IjBa-ubiquitin ligase. Nature 396:590–4 40. Laney JD, Hochstrasser M. 1999. Substrate targeting in the ubiquitin system. Cell 97:427–30 41. Maniatis T. 1999. A ubiquitin ligase complex essential for the NF-jB, Wnt/ Wingless, and Hedgehog signaling pathways. Genes Dev. 13:505–10 42. Chen Z, Hagler J, Palombella VJ, Melandri F, Scherer D, Ballard D, Maniatis T. 1995. Signal-induced site-specific phosphorylation targets IjB to the ubiquitinproteasome pathway. Genes Dev. 9:1586–97 43. Scherer DC, Brockman J, Chen Z, Man-

44.

45.

46.

47.

48.

49.

50.

51.

653

iatis T, Ballard D. 1995. Signal-induced degradation of IjBa requires site-specific ubiquitination. Proc. Natl. Acad. Sci. USA 92:11259–63 Baldi L, Brown K, Franzoso G, Siebenlist U. 1996. Critical role for lysines 21 and 22 in signal-induced, ubiquitinmediated proteolysis of IjBa. J. Biol. Chem. 271:376–79 Alkalay I, Yaron A, Hatzubai A, Orian A, Ciechanover A, Ben-Neriah Y. 1995. Stimulation-dependent IjBa phosphorylation marks the NF-jB inhibitor for degradation via the ubiquitin-proteasome pathway. Proc. Natl. Acad. Sci. USA 92:10599–603 Lin Y-C, Brown K, Siebenlist U. 1995. Activation of NF-jB requires proteolysis of the inhibitor IjB: Signal-induced phosphorylation of IjB alone does not release active NF-jB. Proc. Natl. Acad. Sci. USA 92:3003–9 Imbert V, Rupec RA, Livolsi A, Pahl HL, Traenckner EB, Mueller-Dieckmann C, Farahifar D, Rossi B, Auberger P, Baeuerle PA, Peyron JF. 1996. Tyrosine phosphorylation of IjBa activates NFjB without proteolytic degradation of IjBa. Cell 86:787–98 Beraud C, Henzel WJ, Baeuerle PA. 1999. Involvement of regulatory and catalytic subunits of phosphoinositide 3kinase in NF-jB activation. Proc. Natl. Acad. Sci. USA 96:429–34 Bender K, Gottlicher M, Whiteside S, Rahmsdorf HJ, Herrlich P. 1998. Sequential DNA damage-independent and -dependent activation of NF-jB by UV. EMBO J. 17:5170–81 Li N, Karin M. 1998. Ionizing radiation and short wavelength UV activate NFjB through two distinct mechanisms. Proc. Natl. Acad. Sci. USA 95:13012–17 DiDonato JA, Hayakawa M, Rothwarf DM, Zandi E, Karin M. 1997. A cytokine-responsive IjB kinase that activates the transcription factor NF-jB. Nature 388:548–54


654

KARIN n BEN-NERIAH

52. Mercurio F, Zhu H, Murray BW, Shevchenko A, Bennett BL, Li J, Young DB, Barbosa M, Mann M, Manning A, Rao A. 1997. IKK-1 and IKK-2: cytokineactivated IjB kinases essential for NFjB activation. Science 278:860–66 53. Chen ZJ, Parent L, Maniatis T. 1996. Site-specific phosphorylation of IjBa by a novel ubiquitination-dependent protein kinase activity. Cell 84:853–62 54. Lee FS, Hagler J, Chen ZJ, Maniatis T. 1997. Activation of the IjBa kinase complex by MEKK1, a kinase of the JNK pathway. Cell 88:213–22 55. Zandi E, Rothwarf DM, Delhase M, Hayakawa M, Karin M. 1997. The IjB kinase complex (IKK) contains two kinase subunits, IKKa and IKKb, necessary for IjB phosphorylation and NFjB activation. Cell 91:243–52 56. Re´gnier CH, Song HY, Gao X, Goeddel DV, Cao Z, Rothe M. 1997. Identification and characterization of an IjB kinase. Cell 90:373–83 57. Woronicz JD, Gao X, Cao Z, Rothe M, Goeddel DV. 1997. IjB kinase-b; NF-jB activation and complex formation with IjB kinase-a and NIK. Science 278:866– 69 58. Connelly MA, Marcu KB. 1995. CHUK, a new member of the helix-loop-helix and leucine zipper families of interacting proteins, contains a serine-threonine kinase catalytic domain. Cell. Mol. Biol. Res. 41:537–49 59. Rothwarf DM, Zandi E, Natoli G, Karin M. 1998. IKKc is an essential regulatory subunit of the IjB kinase complex. Nature 395:297–300 60. Yamaoka S, Courtois G, Bessia C, Whiteside ST, Weil R, Agou F, Kirk HE, Kay RJ, Israel A. 1998. Complementation cloning of NEMO, a component of the IjB kinase complex essential for NFjB activation. Cell 93:1231–40 61. Mercurio F, Murray BW, Shevchenko A, Bennett BL, Young DB, Li JW, Pascual G, Motiwala A, Zhu H, Mann M, Man-

62.

63.

64.

65.

66.

67.

68.

69.

70.

ning AM. 1999. IjB kinase (IKK)-associated protein 1, a common component of the heterogeneous IKK complex. Mol. Cell. Biol. 19:1526–38 Li Y, Kang J, Friedman J, Tarassishin L, Ye J, Kovalenko A, Wallach D, Horwitz MS. 1999. Identification of a cell protein (FIP-3) as a modulator of NF-jB activity and as a target of an adenovirus inhibitor of tumor necrosis factor alpha-induced apoptosis. Proc. Natl. Acad. Sci. USA 96:1042–47 Gooding LR, Sofola IO, Tollefson AE, Duerksen-Hughes P, Wold WS. 1990. The adenovirus E3-14.7K protein is a general inhibitor of tumor necrosis factor-mediated cytolysis. J. Immunol. 145:3080–86 Cohen L, Henzel WJ, Baeuerle PA. 1998. IKAP is a scaffold protein of the IjB kinase complex. Nature 395:292–96 Heilker R, Freuler F, Pulfer R, Di Padova F, Eder J. 1999. All three IjB isoforms and most Rel family members are stably associated with the IjB kinase 1/2 complex. Eur. J. Biochem. 259:253–61 Otero G, Fellows J, Li Y, de Bizemont T, Dirac AM, Gustafsson CM, ErdjumentBromage H, Tempst P, Svejstrup JQ. 1999. Elongator, a multisubunit component of a novel RNA polymerase II holoenzyme for transcriptional elongation. Mol. Cell. 3:109–18 Database, YP. Yeast Protein Database http://www.com/database/YPD/blastsw/ yeast/human/IK13.html Zandi E, Chen Y, Karin M. 1998. Direct phosphorylation of IjB by IKKa and IKKb: discrimination between free and NF-jB-bound substrate. Science 281: 1360–63 Li Q, Van Antwerp D, Mercurio F, Lee K-F, Verma IM. 1999. Severe liver degeneration in mice lacking the IjB kinase 2 gene. Science 284:321–25 Li Z-W, Chu W, Hu Y, Delhase M, Deerinck T, Ellisman M, Johnson R, Karin M. 1999. The IKKb subunit of IjB kinase


71.

72.


73.

74.

75.

76.

77.

78.

79.

(IKK) is essential for NF-jB activation and prevention of apoptosis. J. Exp. Med. 189:1839–45 Tanaka M, Fuentes ME, Yamaguchi K, Durnin MH, Dalrymple SA, Hardy KL, Goeddel DV. 1999. Embryonic lethality, liver degeneration, and impaired NF-jB activation in IKK-b-deficient mice. Immunity 10:421–29 Chu Z-L, Shin Y-A, Yang J-M, DiDonato JA, Ballard DW. 1999. IKKc mediates the interaction of cellular IjB kinases with the tax transforming protein of human T cell leukemia virus type 1. J. Biol. Chem. 274:15297–300 Burke JR, Miller KR, Wood MK, Meyers CA. 1998. The multisubunit IjB kinase complex shows random sequential kinetics and is activated by the C-terminal domain of IjBa. J. Biol. Chem. 273:12041–46 Li J, Peet GW, Pullen SS, Schembri-King J, Warren TC, Marcu KB, Kehry MR, Barton R, Jakes S. 1998. Recombinant IjB kinases a and b are direct kinases of IjBa. J. Biol. Chem. 273:30736–41 Ling L, Cao Z, Goeddel DV. 1998. NFjB-inducing kinase activates IKK-a by phosphorylation of Ser-176. Proc. Natl. Acad. Sci. USA 95:2791–97 Delhase M, Hayakawa M, Chen Y, Karin M. 1999. Positive and negative regulation of IjB kinase activity through IKKb subunit phosphorylation. Science 284: 309–13 Johnson LN, Noble MEM, Owen DJ. 1996. Active and inactive protein kinases—structural basis for regulation. Cell 85:149–58 Lallena MJ, Diaz-Meco MT, Bren G, Paya´ CV, Moscat J. 1999. Activation of IjB kinase beta by protein kinase C isoforms. Mol. Cell. Biol. 19:2180–88 Lin X, Mu Y, Cunningham ET Jr, Marcu KB, Geleziunas R, Greene WC. 1998. Molecular determinants of NF-jBinducing kinase action. Mol. Cell. Biol. 18:5899–907

655

80. Romashkova JA, Makarov SS. 1999. NF-jB is a target of AKT in anti-apoptotic PDGF signaling. Nature 401:86–90 81. Ozes ON, Mayo LD, Gustin JA, Pfeffer SR, Pfeffer LM, Donner DB. 1999. NFjB activation by tumor necrosis factor requires the AKT serine-threonine kinase. Nature 401:82–85 82. Nemoto S, DiDonato JA, Lin A. 1998. Coordinate regulation of IjB kinases by mitogen-activated protein kinase kinase kinase 1 and NF-jB-inducing kinase. Mol. Cell. Biol. 18:7336–43 83. Lee FS, Peters RT, Dang LC, Maniatis T. 1998. MEKK1 activates both IjB kinase a and IjB kinase b. Proc. Natl. Acad. Sci. USA 95:9319–24 84. Zhao M, New L, Kravchenko VV, Kato Y, Gram H, di Padova F, Olson EN, Ulevitch RJ, Han J. 1999. Regulation of the MEF2 family of transcription factors by p38. Mol. Cell. Biol. 19:21–30 85. Lin X, Cunningham ET Jr, Mu Y, Geleziunas R, Greene WC. 1999. The protooncogene Cot kinase participates in CD3/ CD28 induction of NF-jB acting through the NF-jB-inducing kinase and IjB kinases. Immunity 10:271–80 86. Ninomiya-Tsuji J, Kishimoto K, Hiyama A, Inoue J, Cao Z, Matsumoto K. 1999. The kinase TAK1 can activate the NIKI jB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 398:252–56 87. Sakurai H, Miyoshi H, Toriumi W, Sugita T. 1999. Functional interactions of transforming growth factor beta-activated kinase 1 with IjB kinases to stimulate NF-jB activation. J. Biol. Chem. 274:10641–48 88. Malinin NL, Boldin MP, Kovalenko AV, Wallach D. 1997. MAP3K-related kinase involved in NF-jB induction by TNF, CD95 and IL-1. Nature 385:540–44 89. Song HY, Re´gnier CH, Kirschning CJ, Goeddel DV, Rothe M. 1997. Tumor necrosis factor (TNF)-mediated kinase cascades: bifurcation of nuclear factor-

656

90.


91.

92.

93.

94.

95.

96.

KARIN n BEN-NERIAH jB and c-jun N-terminal kinase (JNK/ SAPK) pathways at TNF receptorassociated factor 2. Proc. Natl. Acad. Sci. USA 94:9792–96 Baud V, Liu Z-G, Bennett B, Suzuki N, Xia Y, Karin M. 1999. Signaling by proinflammatory cytokines: oligomerization of TRAF2 and TRAF6 is sufficient for JNK and IKK activation and target gene induction via an N-terminal effector domain. Genes Dev. 13:1297– 308 Hu Y, Baud V, Delhase M, Zhang P, Deerinck T, Ellisman M, Johnson R, Karin M. 1999. Abnormal morphogenesis but intact IKK activation in mice lacking the IKKa subunit of the IjB kinase. Science 284:316–20 Takeda K, Takeuchi O, Tsujimura T, Itami S, Adachi O, Kawai T, Sanjo H, Yoshikawa K, Terada N, Akira S. 1999. Limb and skin abnormalities in mice lacking IKKa. Science 284:313–16 Shinkura R, Kitada K, Matsuda F, Tashiro K, Ikuta K, Suzuki M, Kogishi K, Serikawa T, Honjo T. 1999. Alymphoplasia is caused by a point mutation in the mouse gene encoding NF-jB-inducing kinase. Nat. Genet. 22:74–77 Eugster HP, Muller M, Karrer U, Car BD, Schnyder B, Eng VM, Woerly G, Le Hir M, di Padova F, Aguet M, Zinkernagel R, Bluethmann H, Ryffel B. 1996. Multiple immune abnormalities in tumor necrosis factor and lymphotoxin-alpha double-deficient mice. Int. Immunol. 8:23–36 Xia Y, Wu Z, Su B, Murray B, Karin M. 1998. JNKK1 organizes a MAP kinase module through specific and sequential interactions with upstream and downstream components mediated by its amino-terminal extension. Genes Dev. 12:3369–81 Lee SY, Reichlin A, Santana A, Sokol KA, Nussenzweig MD, Choi Y. 1997. TRAF2 is essential for JNK but not NFjB activation and regulates lymphocyte

97.

98.

99.

100.

101.

102.

103.

104.

proliferation and survival. Immunity 7:703–13 Yeh WC, Shahinian A, Speiser D, Kraunus J, Billia F, Wakeham A, de la Pompa JL, Ferrick D, Hum B, Iscove N, Ohashi P, Rothe M, Goeddel DV, Mak TW. 1997. Early lethality, functional NFjB activation, and increased sensitivity to TNF-induced cell death in TRAF2deficient mice. Immunity 7:715–25 Seitz CS, Lin Q, Deng H, Khavari PA. 1998. Alterations in NF-jB function in transgenic epithelial tissue demonstrate a growth inhibitory role for NF-jB. Proc. Natl. Acad. Sci. USA 95:2307–12 Morisato D, Anderson KV. 1995. Signaling pathways that establish the dorsalventral pattern of the Drosophila embryo. Annu. Rev. Genet. 29:371–99 Pan DJ, Huang JD, Courey AJ. 1991. Functional analysis of the Drosophila twist promoter reveals a dorsal-binding ventral activator region. Genes Dev. 5:1892–901 Thisse C, Perrin-Schmitt F, Stoetzel C, Thisse B. 1991. Sequence-specific transactivation of the Drosophila twist gene by the dorsal gene product. Cell 65:1191–201 el Ghouzzi V, Le Merrer M, PerrinSchmitt F, Lajeunie E, Benit P, Renier D, Bourgeois P, Bolcato-Bellemin AL, Munnich A, Bonaventure J. 1997. Mutations of the TWIST gene in the SaethreChotzen syndrome. Nat. Genet. 15:42– 46 Howard TD, Paznekas WA, Green ED, Chiang LC, Ma N, Ortiz de Luna RI, Garcia Delgado C, Gonzalez-Ramos M, Kline AD, Jabs EW. 1997. Mutations in TWIST, a basic helix-loop-helix transcription factor, in Saethre-Chotzen syndrome. Nat. Genet. 15:36–41 Beg AA, Sha WC, Bronson RT, Ghosh S, Baltimore D. 1995. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-jB. Nature 376:167–69


CONTROL OF NF-jB VIA IjB DEGRADATION 105. Horwitz BH, Scott ML, Cherry SR, Bronson RT, Baltimore D. 1997. Failure of lymphopoiesis after adoptive transfer of NF-j B-deficient fetal liver cells. Immunity 6:765–72 106. Doi TS, Marino MW, Takahashi T, Yoshida T, Sakakura T, Old LJ, Obata Y. 1999. Absence of tumor necrosis factor rescues RelA-deficient mice from embryonic lethality. Proc. Natl. Acad. Sci. USA 96:2994–99 107. Vandenabeele P, Declercg W, Beyaert R, Fiers W. 1995. Two tumour necrosis factor receptors: structure and function. Trends Cell. Biol. 5:392–99 108. Baker SJ, Reddy EP. 1998. Modulation of life and death by the TNF receptor superfamily. Oncogene 17:3261–70 109. Orlinick JR, Chao MV. 1998. TNFrelated ligands and their receptors. Cell Signal. 10:543–51 110. Sims JE, Gayle MA, Slack JL, Alderson MR, Bird TA, Giri JG, Colotta F, Re F, Mantovani A, Shanebeck K, Grabstein KH, Dower SK. 1993. Interleukin 1 signaling occurs exclusively via the type I receptor. Proc. Natl. Acad. Sci. USA 90:6155–59 111. Auron PE. 1998. The interleukin 1 receptor: ligand interactions and signal transduction. Cytokine Growth Factor Rev. 9:221–37 112. Kelliher MA, Grimm S, Ishida Y, Kuo F, Stanger BZ, Leder P. 1998. The death domain kinase RIP mediates the TNFinduced NF-jB signal. Immunity 8:297– 303 113. Greenfeder SA, Nunes P, Kwee L, Labow M, Chizzonite RA, Ju G. 1995. Molecular cloning and characterization of a second subunit of the interleukin 1 receptor complex. J. Biol. Chem. 270:13757–65 114. Wesche H, Henzel WJ, Shillinglaw W, Li S, Cao Z. 1997. MyD88: an adapter that recruits IRAK to the IL-1 receptor complex. Immunity 7:837–47 115. Cao Z, Henzel WJ, Gao X. 1996. IRAK:

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

657

a kinase associated with the interleukin1 receptor. Science 271:1128–31 Muzio M, Ni J, Feng P, Dixit VM. 1997. IRAK (Pelle) family member IRAK-2 and MyD88 as proximal mediators of IL1 signaling. Science 278:1612–15 Cao Z, Xiong J, Takeuchi M, Kurama T, Goeddel DV. 1996. TRAF6 is a signal transducer for interleukin-1. Nature 383:443–46 Arch RH, Gedrich RW, Thompson CB. 1998. Tumor necrosis factor receptorassociated factors (TRAFs)—a family of adapter proteins that regulates life and death. Genes Dev. 12:2821–30 Wajant H, Grell M, Scheurich P. 1999. TNF receptor associated factors in cytokine signaling. Cytokine Growth Factor Rev. 10:15–26 Park YC, Burkitt V, Villa AR, Tong L, Wu H. 1999. Structural basis for selfassociation and receptor recognition of human TRAF2. Nature 398:533–38 Rothe M, Sarma V, Dixit VM, Goeddel DV. 1995. TRAF-2-mediated activation of NF-jB by TNF receptor 2 and CD40. Science 269:1424–27 Liu Z-G, Hu H, Goeddel DV, Karin M. 1996. Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis, while NF-jB activation prevents cell death. Cell 87:565– 76 Takeuchi M, Rothe M, Goeddel DV. 1996. Anatomy of TRAF2: distinct domains for nuclear factor-jB activation and association with tumor necrosis factor signaling proteins. J. Biol. Chem. 16:19935–42 Nakano H, Oshima H, Chung W, Williams-Abbott L, Ware CF, Yagita H, Okumura K. 1996. TRAF5, an activator of NF-jB and putative signal transducer for the lymphotoxin-b receptor. J. Biol. Chem. 271:14661–64 Lomaga MA, Yeh WC, Sarosi I, Duncan GS, Furlonger C, Ho A, Morony S, Capparelli C, Van G, Kaufman S, van der

658

126.


127.

128.

129.

130.

131.

132.

133.

KARIN n BEN-NERIAH Heiden A, Itie A, Wakeham A, Khoo W, Sasaki T, Cao Z, Penninger JM, Paige CJ, Lacey DL, Dunstan CR, Boyle WJ, Goeddel DV, Mak TW. 1999. TRAF6 deficiency results in osteopetrosis and defective interleukin-1, CD40, and LPS signaling. Genes Dev. 13:1015–24 Yuasa T, Ohno S, Kehrl JH, Kyriakis JM. 1998. Tumor necrosis factor signaling to stress-activated protein kinase (SAPK)/ Jun NH2-terminal kinase (JNK) and p38. J. Biol. Chem. 273:22681–92 Shi CS, Kehrl JH. 1997. Activation of stress-activated protein kinase/c-Jun Nterminal kinase, but not NF-jB, by the tumor necrosis factor (TNF) receptor 1 through a TNF receptor-associated factor 2- and germinal center kinase relateddependent pathway. J. Biol. Chem. 272:32102–7 Ichijo H, Nishida E, Irie K, ten Dijke P, Saitoh M, Moriguchi T, Takagi M, Matsumoto K, Miyazono K, Gotoh Y. 1997. Induction of apoptosis by ASK1, a mammalian MAPKKK that activates SAPK/ JNK and p38 signaling pathways. Science 275:90–94 Nishitoh H, Saitoh M, Mochida Y, Takeda K, Nakano H, Rothe M, Miyazono K, Ichijo H. 1998. ASK1 is essential for JNK/SAPK activation by TRAF2. Mol. Cell. 2:389–95 Rusch J, Levine M. 1996. Threshold responses to the dorsal regulatory gradient and the subdivision of primary tissue territories in the Drosophila embryo. Curr. Opin. Genet. Dev. 6:416–23 Belvin MP, Anderson KV. 1996. A conserve signaling pathway: the Drosophila toll-dorsal pathway. Annu. Rev. Dev. Biol. 12:393–416 Drier EA, Steward R. 1997. The dorsoventral signal transduction pathway and the Rel-like transcription factors in Drosophila. Semin. Cancer Biol. 8:83–92 Reach M, Galindo RL, Towb P, Allen JL, Karin M, Wasserman SA. 1996. A gradient of cactus protein degradation estab-

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144. 145.

lishes dorsoventral polarity in the Drosophila embryo. Dev. Biol. 180:353– 64 Jiang J, Struhl G. 1998. Regulation of the Hedgehog and Wingless signalling pathways by the F-box/WD40-repeat protein Slimb. Nature 391:493–96 Knop J, Martin MU. 1999. Effects of IL1 receptor-associated kinase (IRAK) expression on IL-1 signaling are independent of its kinase activity. FEBS Lett. 448:81–85 Maschera B, Ray K, Burns K, Volpe F. 1999. Overexpression of an enzymically inactive interleukin-1-receptor-associated kinase activates nuclear factor-jB. Biochem. J. 339:227–31 Kopp EB, Medzhitov R. 1999. The Tollreceptor family and control of innate immunity. Curr. Opin. Immunol. 11:13– 18 Hoffmann JA, Kafatos FC, Janeway CAJ, Ezekowiz RAB. 1999. Phylogenetic perspectives in innate immunity. Science 284:1313–18 Wu LP, Anderson KV. 1998. Regulated nuclear import of Rel proteins in the Drosophila immune response. Nature 392:93–97 Dushay MS, Asling B, Hultmark D. 1996. Origins of immunity: Relish, a compound Rel-like gene in the antibacterial defense of Drosophila. Proc. Natl. Acad. Sci. USA 93:10343–47 Baumeister W, Walz J, Zuhl F, Seemuller E. 1998. The proteasome: paradigm of a self-compartmentalizing protease. Cell 92:367–80 Hochstrasser M. 1996. Ubiquitin-dependent protein degradation. Annu. Rev. Genet. 30:405–39 Hershko A, Ciechanover A. 1998. The ubiquitin system. Annu. Rev. Biochem. 67:425–79 Varshavsky A. 1997. The ubiquitin system. Trends Biochem. Sci. 22:383–87 Haas AL, Siepmann TJ. 1997. Pathways


146.

147.

148.


149.

150.

151.

152.

153.

154.

155.

156.

of ubiquitin conjugation. FASEB J. 11:1257–68 Varshavsky A. 1996. The N-end rule: functions, mysteries, uses. Proc. Natl. Acad. Sci. USA 93:12142–49 Scheffner M, Huibregtse JM, Vierstra RD, Howley PM. 1993. The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell 75:495–505 Staub O, Gautschi I, Ishikawa T, Breitschopf K, Ciechanover A, Schild L, Rotin D. 1997. Regulation of stability and function of the epithelial Na` channel (ENaC) by ubiquitination. EMBO J. 16:6325–36 Scheffner M, Nuber U, Huibregtse JM. 1995. Protein ubiquitination involving an E1-E2-E3 enzyme ubiquitin thioester cascade. Nature 373:81–83 Morgan DO. 1999. Regulation of the APC and the exit from mitosis. Nat. Cell Biol. 1:E47–E52 Zachariae W, Nosmyth K. 1999. Whose end is destruction: cell cycle division and the anaphase promoting complex. Genes Dev. 13:2039–58 Yamano H, Tsurumi C, Gannon J, Hunt T. 1998. The role of the destruction box and its neighbouring lysine residues in cyclin B for anaphase ubiquitin-dependent proteolysis in fission yeast: defining the D-box receptor. EMBO J. 17:5670– 78 Fang G, Yu H, Kirschner MW. 1998. Direct binding of CDC20 protein family members activates the anaphase-promoting complex in mitosis and G1. Mol. Cell. 2:163–71 Visintin R, Craig K, Hwang ES, Prinz S, Tyers M, Amon A. 1998. The phosphatase Cdc14 triggers mitotic exit by reversal of Cdk-dependent phosphorylation. Mol. Cell. 2:709–18 Peters JM. 1998. SCF and APC: the yin and yang of cell cycle regulated proteolysis. Curr. Opin. Cell. Biol. 10:759–68 Patton EE, Willems AR, Tyers M. 1998.

157.

158.

159.

160.

161.

162.

163.

659

Combinatorial control in ubiquitindependent proteolysis: Don’t Skp the Fbox hypothesis. Trends Genet. 14:236– 43 Seol JH, Feldman RM, Zachariae W, Shevchenko A, Correll CC, Lyapina S, Chi Y, Galova M, Claypool J, Sandmeyer S, Nasmyth K, Deshaies RJ. 1999. Cdc53/cullin and the essential Hrt1 RING-H2 subunit of SCF define a ubiquitin ligase module that activates the E2 enzyme Cdc34. Genes Dev. 13:1614–26 Ohta T, Michel JJ, Schottelius AJ, Xiong Y. 1999. ROC1, a homolog of APC11, represents a family of cullin partners with an associated ubiquitin ligase activity. Mol. Cell. 3:535–41 Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin WG Jr, Elledge SJ, Conaway RC, Harper JW, Conaway JW. 1999. Rbx1, a component of the VHL tumor suppressor complex and SCF ubiquitin ligase. Science 284:657–61 Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER, Ratcliffe PJ. 1999. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 399:271–75 Lonergan KM, Iliopoulos O, Ohh M, Kamura T, Conaway RC, Conaway JW, Kaelin WG Jr. 1998. Regulation of hypoxia-inducible mRNAs by the von Hippel-Lindau tumor suppressor protein requires binding to complexes containing elongins B/C and Cul2. Mol. Cell. Biol. 18:732–41 Skowyra D, Koepp DM, Kamura T, Conrad MN, Conaway RC, Conaway JW, Elledge SJ, Harper JW. 1999. Reconstitution of G1 cyclin ubiquitination with complexes containing SCFGrr1 and Rbx1. Science 284:662–65 Sun S, Elwood J, Greene WC. 1996. Both amino- and carboxyl-terminal sequences within I j B alpha regulate its

660

164.

165.

166.


167.

168.

169.

170.

171.

172.

173.

KARIN n BEN-NERIAH inducible degradation. Mol. Cell. Biol. 16:1058–65 Yaron A, Gonen H, Alkalay I, Hatzubai A, Jung S, Beyth S, Mercurio F, Manning AM, Ciechanover A, Ben-Neriah Y. 1997. Inhibition of NF-jB cellular function via specific targeting of the IjBubiquitin ligase. EMBO J. 16:6486–94 Scott ML, Fujita T, Liou HC, Nolan GP, Baltimore D. 1993. The p65 subunit of NF-j B regulates IjB by two distinct mechanisms. Genes Dev. 7:1266–76 Rice NR, Ernst MK. 1993. In vivo control of NF-j B activation by IjBa. EMBO J. 12:4685–95 Aberle H, Bauer A, Stappert J, Kispert A, Kemler R. 1997. Beta-catenin is a target for the ubiquitin-proteasome pathway. EMBO J. 16:3797–804 Rubinfeld B, Robbins P, El-Gamil M, Albert I, Porfiri E, Polakis P. 1997. Stabilization of beta-catenin by genetic defects in melanoma cell lines. Science 275:1790–92 Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, Kinzler KW. 1997. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275:1787– 90 Palacios J, Gamallo C. 1998. Mutations in the beta-catenin gene (CTNNB1) in endometrioid ovarian carcinomas. Cancer Res. 58:1344–47 Muller O, Nimmrich I, Finke U, Friedl W, Hoffmann I. 1998. A beta-catenin mutation in a sporadic colorectal tumor of the RER phenotype and absence of beta-catenin germline mutations in FAP patients. Genes Chrom. Cancer 22:37–41 Miyoshi Y, Iwao K, Nagasawa Y, Aihara T, Sasaki Y, Imaoka S, Murata M, Shimano T, Nakamura Y. 1998. Activation of the beta-catenin gene in primary hepatocellular carcinomas by somatic alterations involving exon 3. Cancer Res. 58:2524–27 Zurawel RH, Chiappa SA, Allen C, Raf-

174.

175.

176.

177.

178.

179.

180.

181.

fel C. 1998. Sporadic medulloblastomas contain oncogenic beta-catenin mutations. Cancer Res. 58:896–99 Voeller HJ, Truica CI, Gelmann EP. 1998. Beta-catenin mutations in human prostate cancer. Cancer Res. 58:2520–23 Sparks AB, Morin PJ, Vogelstein B, Kinzler KW. 1998. Mutational analysis of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res. 58:1130– 34 de La Coste A, Romagnolo B, Billuart P, Renard CA, Buendia MA, Soubrane O, Fabre M, Chelly J, Beldjord C, Kahn A, Perret C. 1998. Somatic mutations of the beta-catenin gene are frequent in mouse and human hepatocellular carcinomas. Proc. Natl. Acad. Sci. USA 95:8847–51 Chan EF, Gat U, McNiff JM, Fuchs E. 1999. A common human skin tumour is caused by activating mutations in betacatenin. Nat. Genet. 21:410–13 Spevak W, Keiper BD, Stratowa C, Castanon MJ. 1993. Saccharomyces cerevisiae cdc15 mutants arrested at a late stage in anaphase are rescued by Xenopus cDNAs encoding N-ras or a protein with beta-transducin repeats. Mol. Cell. Biol. 13:4953–66. Erratum. 1993. 13(11):7199 Margottin F, Bour SP, Durand H, Selig L, Benichou S, Richard V, Thomas D, Strebel K, Benarous R. 1998. A novel human WD protein, h-beta TrCP, that interacts with HIV-1 Vpu connects CD4 to the ER degradation pathway through an F-box motif. Mol. Cell. 1:565–74 Suzuki H, Chiba T, Kobayashi M, Takeuchi M, Suzuki T, Ichiyama A, Ikenoue T, Omata M, Furuichi K, Tanaka K. 1999. IjBa ubiquitination is catalyzed by an SCF-like complex containing Skp1, cullin-1, and two F-box/WD40repeat proteins, betaTrCP1 and betaTrCP2. Biochem. Biophys. Res. Commun. 256:127–32 Hatakeyama S, Kitagawa M, Nakayama K, Shirane M, Matsumoto M, Hattori K, Higashi H, Nakano H, Okumura K, Onoe


182.

183.


184.

185.

186.

187.

188.

189.

K, Good RA, Nakayama K. 1999. Ubiquitin-dependent degradation of IjBa is mediated by a ubiquitin ligase Skp1/Cu1 1/F-box protein FWD1. Proc. Natl. Acad. Sci. USA 96:3859–63 Kroll M, Margottin F, Kohl A, Renard P, Durand H, Concordet JP, Bachelerie F, Arenzana-Seisdedos F, Benarous R. 1999. Inducible degradation of IjBa by the proteasome requires interaction with the F-box protein h-betaTrCP. J. Biol. Chem. 274:7941–45 Spencer E, Jiang J, Chen ZJ. 1999. Signal-induced ubiquitination of IjBa by the F-box protein Slimb/b-TrCP. Genes Dev. 13:284–94 Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, Harper JW. 1999. The SCF b-TRCP-ubiquitin ligase complex associates specifically with phosphorylated destruction motifs in IjBa and b-catenin and stimulates IjBa ubiquitination in vitro. Genes Dev. 13:270– 83 Fuchs SY, Chen A, Xiong Y, Pan Z-Q, Ronai Z. 1999. HOS, a human homolog of Slimb, forms an SCF complex with Skp1 and Cullin1 and targets the phosphorylation-dependent degradation of IjB and b-catenin. Oncogene 18:2039–46 Gaudet R, Bohm A, Sigler PB. 1996. Crystal structure at 2.4 angstroms resolution of the complex of transducin betagamma and its regulator, phosducin. Cell 87:577–88 Sondek J, Bohm A, Lambright DG, Hamm HE, Sigler PB. 1996. Crystal structure of a G-protein beta gamma dimer at 2.1A resolution. Nature 379:369–74. Erratum. 1996. Nature 379(6568):847 Murakami Y, Matsufuji S, Kameji T, Hayashi S, Igarashi K, Tamura T, Tanaka K, Ichihara A. 1992. Ornithine decarboxylase is degraded by the 26S proteasome without ubiquitination. Nature 360:597– 99 Fujiwara T, Suzuki M, Tanigami A, Iken-

190.

191.

192.

193.

194.

195.

196.

197.

661

oue T, Omata M, Chiba T, Tanaka K. 1999. The BTRC gene, encoding a human F-Box/WD40-repeat protein, maps to chromosome 10q24-q25. Genomics 58:104–5 Hart M, Concordet JP, Lassot I, Albert I, del los Santos R, Durand H, Perret C, Rubinfeld B, Margottin F, Benarous R, Polakis P. 1999. The F-box protein betaTrCP associates with phosphorylated beta-catenin and regulates its activity in the cell. Curr. Biol. 9:207–10 Kitagawa M, Hatakeyama S, Shirane M, Matsumoto M, Ishida N, Hattori K, Nakamichi I, Kikuchi A, Nakayama K. 1999. An F-box protein, FWD1, mediates ubiquitin-dependent proteolysis of betacatenin. EMBO J. 18:2401–10 Latres E, Chiaur DS, Pagano M. 1999. The human F box protein beta-Trcp associates with the Cul1/Skp1 complex and regulates the stability of beta-catenin. Oncogene 18:849–54 Liu C, Kato Y, Zhang Z, Do VM, Yankner BA, He X. 1999. Beta-Trcp couples beta-catenin phosphorylation-degradation and regulates Xenopus axis formation. Proc. Natl. Acad. Sci. USA 96:6273–78 Marikawa Y, Elinson RP. 1998. BetaTrCP is a negative regulator of Wnt/betacatenin signaling pathway and dorsal axis formation in Xenopus embryos. Mech. Dev. 77:75–80 Polakis P. 1997. The adenomatous polyposis coli (APC) tumor suppressor. Biochim. Biophys. Acta 1332:F127–47 Chernova O, Cowell JK. 1998. Molecular definition of chromosome translocations involving 10q24 and 19q13 in human malignant glioma cells. Cancer Genet. Cytogenet. 105:60–6 Prochownik EV, Eagle Grove L, Deubler D, Zhu XL, Stephenson RA, Rohr LR, Yin X, Brothman AR. 1998. Commonly occurring loss and mutation of the MXI1 gene in prostate cancer. Genes Chrom. Cancer. 22:295–304


662

KARIN n BEN-NERIAH

198. Kim SK, Ro JY, Kemp BL, Lee JS, Kwon TJ, Hong WK, Mao L. 1998. Identification of two distinct tumor-suppressor loci on the long arm of chromosome 10 in small cell lung cancer. Oncogene 17:1749–53 199. Beg AA, Baltimore D. 1996. An essential role for NF-jB in preventing TNF-a induced cell death. Science 274:782–84 200. Wang C-Y, Mayo MW, Baldwin AS Jr. 1996. TNF- and cancer therapy-induced apoptosis: potentiation by inhibition of NF-jB. Science 274:784–87 201. Lin L, DeMartino GN, Greene WC. 1998. Cotranslational biogenesis of NFjB p50 by the 26S proteasome. Cell 92:819–28 202. Fan CM, Maniatis T. 1991. Generation of p50 subunit of NF-jB by processing of p105 through an ATP-dependent pathway. Nature 354:395–98 203. Palombella VJ, Rando OJ, Goldberg AL, Maniatis T. 1994. The ubiquitin-proteasome pathway is required for processing the NF-jB1 precursor protein and the activation of NF-jB. Cell 78:773–85 204. Orian A, Whiteside S, Israe¨l A, Stancovski I, Schwartz AL, Ciechanover A. 1995. Ubiquitin-mediated processing of NF-jB transcriptional activator precursor p105: reconstitution of a cell-free system and identification of the ubiquitin-carrier protein, E2, and a novel ubiquitin-protein ligase, E3, involved in conjugation. J. Biol. Chem. 270:21707– 14 205. Coux O, Goldberg AL. 1998. Enzymes catalyzing ubiquitination and proteolytic processing of the p105 precursor of nuclear factor-jB1. J. Biol. Chem. 273:8820–28 206. MacKichan ML, Logeat F, Israe¨l A. 1996. Phosphorylation of p105 PEST sequence via a redox-insensitive pathway up-regulates processing of p50 NF-jB. J. Biol. Chem. 271:6084–91 207. Mellits KH, Hay RT, Goodbourn S. 1993. Proteolytic degradation of MAD-3

208.

209.

210.

211.

212.

213.

214.

215.

IjBa and enhanced processing of the NF-jB precursor p105 are obligatory steps in the activation of NF-jB. Nucleic Acids Res. 21:5059–66 Heissmeyer V, Krappman D, Wulczyn FG, Scheidereit C. 1999. NF-jB p105 is a target for IjB kinases and controls signal induction of Bcl-3-p50 complexes. EMBO J. 18:4766–78 Mercurio F, DiDonato J, Rosette C, Karin M. 1993. p105 and p98 precursor proteins play an active role in NF-jB mediated signal transduction. Genes. Dev. 7:705–18 Belich MP, Salmeroń A, Johnston LH, Ley SC. 1999. TPL-2 kinase regulates the proteolysis of the NF-jB-inhibitory protein NF-jB1 p105. Nature 397:363– 68 Tamura N, Lottspeich F, Baumeister W, Tamura T. 1998. The role of tricorn protease and its aminopeptidase-interacting factors in cellular protein degradation. Cell 95:637–48 Lin L, Ghosh S. 1996. A glycine-rich region in NF-jB p105 functions as a processing signal for the generation of the p50 subunit. Mol. Cell. Biol. 16:2248–54 Levitskaya J, Sharipo A, Leonchiks A, Ciechanover A, Masucci MG. 1997. Inhibition of ubiquitin/proteasomedependent protein degradation by the Gly-Ala repeat domain of the EpsteinBarr virus nuclear antigen 1. Proc. Natl. Acad. Sci. USA 94:12616–21 Sharipo A, Imreh M, Leonchiks A, Imreh S, Masucci MG. 1998. A minimal glycine-alanine repeat prevents the interaction of ubiquitinated I jB alpha with the proteasome: a new mechanism for selective inhibition of proteolysis. Nat. Med. 4:939–44 Orian A, Schwartz AL, Israe¨l A, Whiteside S, Kahana C, Ciechanover A. 1999. Structural motifs involved in ubiquitinmediated processing of the NF-jB precursor p105: roles of the glycine-rich


216.

217.


218.

219.

220.

221.

222.

223.

224.

225.

region and a downstream ubiquitination domain. Mol. Cell. Biol. 19:3664–73 Sears C, Olesen J, Rubin D, Finley D, Maniatis T. 1998. NF-j B p105 processing via the ubiquitin-proteasome pathway. J. Biol. Chem. 273:1409–19 Wang D, Baldwin AS Jr. 1998. Activation of nuclear factor-jB-dependent transcription by tumor necrosis factor-alpha is mediated through phosphorylation of RelA/p65 on serine 529. J. Biol. Chem. 273:29411–16 Zhong H, SuYang H, Erdjument-Bromage H, Tempst P, Ghosh S. 1997. The transcriptional activity of NF-jB is regulated by the IjB-associated PKAc subunit through a cyclic AMP-independent mechanism. Cell 89:413–24 Drier EA, Huang LH, Steward R. 1999. Nuclear import of the Drosophila Rel protein dorsal is regulated by phosphorylation. Genes Dev. 13:556–68 Sachdev S, Hoffmann A, Hannink M. 1998. Nuclear localization of IjBa is mediated by the second ankyrin repeat: the IjBa ankyrin repeats define a novel class of cis-acting nuclear import sequences. Mol. Cell. Biol. 18:2524–34 Turpin P, Hay RT, Dargemont C. 1999. Characterization of IjBa nuclear import pathway. J. Biol. Chem. 274:6804–12 Arenzana-Seisdedos F, Thompson J, Rodriguez MS, Bachelerie F, Thomas D, Hay RT. 1995. Inducible nuclear expression of newly synthesized IjB alpha negatively regulates DNA-binding and transcriptional activities of NF-jB. Mol. Cell. Biol. 15:2689–96 Wen W, Meinkoth JL, Tsien RY, Taylor SS. 1995. Identification of a signal for rapid export of proteins from the nucleus. Cell 82:463–73 Johnson C, Van Antwerp D, Hope TJ. 1999. An N-terminal nuclear export signal is required for the nucleocytoplasmic shuttling of IjBa. EMBO J. 23:6682–93 Lin RT, Beauparlant P, Makris C, Meloche S, Hiscott J. 1996. Phosphorylation

226.

227.

228.

229.

230.

231.

232.

233.

234.

235.

663

of IjBa in the C-terminal pest domain by casein kinase Iu affects intrinsic protein stability. Mol. Cell. Biol. 16:1401–9 McElhinny JA, Trushin SA, Bren GD, Chester N, Paya CV. 1996. Casein kinase Ii phosphorylates IjBa at S-283, S-289, S-293, and T-291 and is required for its degradation. Mol. Cell. Biol. 16:899–906 Schwarz EM, Vanantwerp D, Verma IM. 1996. Constitutive phosphorylation of IjBa by casein kinase Ii occurs preferentially at serine 293—requirement for degradation of free IjBa. Mol. Cell. Biol. 16:3554–59 Suyang H, Phillips R, Douglas I, Ghosh S. 1996. Role of unphosphorylated, newly synthesized IjBb in persistent activation of NF-jB. Mol. Cell. Biol. 16:5444–49 Tran K, Merika M, Thanos D. 1997. Distinct functional properties of IjBa and IjBb. Mol. Cell. Biol. 17:5386–99 McKinsey TA, Chu ZL, Ballard DW. 1997. Phosphorylation of the PEST domain of IjBb regulates the function of NF-jB/IjBb complexes. J. Biol. Chem. 272:22377–80 Phillips RJ, Ghosh S. 1997. Regulation of IjB beta in WEHI 231 mature B cells. Mol. Cell. Biol. 17:4390–6 DeLuca C, Petropoulus L, Zmeureanu D, Hiscott J. 1999. Nuclear IjBb maintains persistent NF-jB activation in HIV-1infected myeloid cells. J. Biol. Chem. 274:13010–6 Baeuerle PA, Henkel T. 1994. Function and activation of NF-jB in the immune system. Annu. Rev. Immunol. 12:141–79 Chen F, Castranova V, Shi X, Demers LM. 1999. New insights into the role of nuclear factor-jB, a ubiquitous transcription factor in the initiation of diseases. Clin. Chem. 45:7–17 Auphan N, DiDonato JA, Rosette C, Helmberg A, Karin M. 1995. Immunosuppression by glucocorticoids: Inhibition of NF-jB activity through induction of IkB synthesis. Science 270:286–90


KARIN

■

BEN-NERIAH C-1

Figure 1 Schematic representation of members of the mammalian and Drosophila Rel/NF-κB and IκB families of proteins. The numbers of amino acids in each protein are listed on the right. The arrows point to the C-terminal residues of p50 and p52 (following processing of p105 and p100, respectively). The amino acid N-terminal motif required for inducible degradation of IκBα, IκBβ, IκBε, and Cactus is indicated, and the conserved residues are colored in red; LZ, leucine zipper; GRR, glycine-rich region; SRR, serine-rich region. The number and position of ankyrin repeats were determined based upon homology and the X-ray structure of IκBα.


C-2 KARIN

■

BEN-NERIAH

Figure 2 A schematic model of NF-κB activation. Various stimuli, including the proinflammatory cytokines TNFα and IL-1, activate IKK through the action of as-yet-unidentified components. Once activated, IKK phosphorylates IκBα, leading to its recognition by E3RSIκB, a receptor component of a SCF type E3, which results in the polyubiquitination of IκBα. This then targets IκBα for rapid degradation by the 26S proteasome. IκBα degradation exposes the nuclear localization sequence on NF-κB resulting in its translocation to the nucleus. In the nucleus NF-κB regulates transcription of target genes, including IκBα, which functions to terminate NF-κB activity. Some of the NF-κB target genes code for inflammatory mediators, such as TNFα and IL-1 and chemokines, which can lead to recruitment of additional cells to the inflammatory response.


KARIN

■

BEN-NERIAH C-3

Figure 3 Schematic diagram showing the known subunits of IKK and their putative functional and structural motifs. CC, coiled coil; Helix, α-helix; HLH, helix-loop-helix; LZ, leucine zipper; ZF, zinc finger.


C-4 KARIN

■

BEN-NERIAH

Figure 4 Schematic representation of proposed model of IKK regulation by phosphorylation. Two catalytic subunits of IKK (IKKα and IKKβ) dimerize via their leucine zippers (LZ). The C-terminal and helix-loop-helix (HLH) regions in IKK interact with the kinase domain (KD). Phosphorylation of the two serines (SS) in the activation loop of IKK results its activation. Once activated, in addition to phosphorylating IκBs, IKK autophosphorylates its C-terminal region. When 9 or more serines in the C-terminal region are phosphorylated, the interaction between the C-terminal region, which contains the HLH motif that functions as an intrinsic activator of the kinase, and the kinase domain is altered and the activity of IKK decreases. In this state IKK is more susceptible to further inhibition by phosphatase action.


KARIN

■

BEN-NERIAH C-5

Figure 5 A. The IκB degradation motif: alignment of a number of peptide sequences containing the recognition signal for E3RSIκB. The 1250 entries to the SWISSPROT/EMBL databanks, representing hundreds of different proteins, share this sequence motif but must be phosphorylated on both serine residues to act as the IκBα degradation motif. Beside the two serines, which are colored in red, the glycine and aspartic acid are conserved too and are colored in green. Substitution of any of the conserved residues results in stabilization of IκBα or β-catenin. A lysine residue positioned 8-12 amino-acids upstream of the first serine is not part of the E3RSIκB recognition signal but is apparently the target for ubiquitination. Vpu is devoid of the conserved lysine residue and may instead target a lysine residue of an associated protein (CD4) for E3RSIκB-mediated ubiquitination. B Schematic representation of the IκBα-E3 complex, interacting with the phosphorylated IκBα to form a predegradation complex. The 700 kD predegradation complex is composed of NF-κB, pIκBα, and other unknown proteins (possibly chaperons), which interact with a specific E3 complex. The E3 complex is composed of several proteins: Skp1, Cul1, Roc1/Rbx1/Hrt1 and an F-box protein, forming the SCF complex. The F-box E3RSIκB protein is the variable component of SCF complexes and functions as the pIκBα receptor component of the E3 complex. Following pIκBα binding, the E3 complex associates with a specific E2, UBC5, and in the presence of E1 and ubiquitin catalyses the conjugation of ubiquitin to IκBα, as well as ubiquitin-ubiquitin conjugation. The latter function may be attributed to the Cul1:Roc1 complex.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


RESERVOIRS FOR HIV-1: Mechanisms for Viral Persistence in the Presence of Antiviral Immune Responses and Antiretroviral Therapy Theodore Pierson, Justin McArthur, and Robert F. Siliciano


Departments of Medicine and Neurology, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205; e-mail: [email protected]

Key Words latency, viral reservoirs, HAART, antiretroviral therapy, memory T cells, eradication Abstract The success of combination antiretroviral therapy for HIV-1 infection has generated interest in mechanisms by which the virus can persist in the body despite the presence of drugs that effectively inhibit key steps in the virus life cycle. It is becoming clear that viral reservoirs established early in the infection not only prevent sterilizing immunity but also represent a major obstacle to curing the infection with the potent antiretroviral drugs currently in use. Mechanisms of viral persistence are best considered in the context of the dynamics of viral replication in vivo. Virus production in infected individuals is largely the result of a dynamic process involving continuous rounds of de novo infection of and replication in activated CD4` T cells with rapid turnover of both free virus and virus-producing cells. This process is largely, but not completely, interrupted by effective antiretroviral therapy. After a few months of therapy, plasma virus levels become undetectable in many patients. Analysis of viral decay rates initially suggested that eradication of the infection might be possible. However, there are several potential cellular and anatomical reservoirs for HIV-1 that may contribute to long-term persistence of HIV-1. These include infected cell in the central nervous system and the male urogenital tract. However, the most worrisome reservoir consists of latently infected resting memory CD4` T cells carrying integrated HIV-1 DNA. Definitive demonstration of the presence of this form of latency required development of methods for isolating extremely pure populations of resting CD4` T cells and for demonstrating that a small fraction of these cells contain integrated HIV-1 DNA that is competent for replication if the cells undergo antigen-driven activation. Most of the latent virus in resting CD4` T cells is found in cells of the memory phenotype. The half-life of this latent reservoir is extremely long (44 months). At this rate, eradication of this reservoir would require over 60 years of treatment. Thus, latently infected resting CD4` T cells provide a mechanism for life-long persistence of replication-competent forms of HIV-1, rendering unrealistic hopes of virus eradication with current antiretroviral regimens. The extraordinary stability of the reservoir may reflect gradual reseeding by a very low level of ongoing viral replication and/or mechanisms that contribute to the intrinsic stability of the 0732–0582/00/0410–0665$14.00

665

666

PIERSON n MCARTHUR n SILICIANO

memory T cell compartment. Given the substantial long-term toxicities of current combination therapy regimens, novel approaches to eradicating this latent reservoir are urgently needed.


INTRODUCTION Many viruses have evolved mechanisms for establishing persistence in the host. Typically, persistence involves viral reservoirs, cellular or anatomical sites in which the virus can persist despite antiviral immune responses. Some viruses establish a reversibly nonproductive (or latent) state of infection that allows escape from host immune mechanisms. For HIV-1, mechanisms of persistence and latency have been the subject of much recent interest. This interest reflects the fact that successes in antiretroviral therapy have raised, for the first time since the beginning of the AIDS epidemic, the possibility that the infection might be curable (1). Inhibitors of HIV-1 reverse transcriptase and protease, when used in multiple drug combinations, cause dramatic reductions in viremia in many patients. Plasma virus levels often fall to below the limit of detection of current assays, around 20 copies of genomic viral RNA/ml (equivalent to 10 diploid virus particles/ml) (1–3). Striking decreases in morbidity and mortality have been documented since the introduction of the protease inhibitors (4). However, the goal of curing HIV-1 infection has not yet been achieved, and it is becoming clear that the viral reservoirs established early in the infection not only prevent sterilizing immunity but also represent a major obstacle to curing the infection with the potent antiretroviral drugs currently in use. This review describes recent studies of reservoirs for HIV-1. For a discussion of earlier work on viral reservoirs and latency in HIV-1 infection, the reader is referred to several comprehensive reviews (5–12).

THE NATURAL HISTORY OF HIV-1 INFECTION As a prelude to the discussion of viral reservoirs, the natural history of HIV-1 infection and the dynamics of infection at the cellular level are briefly reviewed. Clinically, HIV-1 infection may be divided into three phases. During the initial phase, known as primary HIV-1 infection, virus present in the infecting inoculum replicates in the host, infecting cells that express both CD4 and the appropriate coreceptor, usually the chemokine receptor CCR5 (see 13 for a review of HIV-1 entry). Even early in infection, CD4` T lymphocytes appear to be the major target cells (11). Viremia develops within the first few weeks of exposure, accompanied by infectious mononucleosis-like symptoms in some patients (14). Studies of viral dynamics during primary HIV-1 and SIV infection suggest that virus populations double every 6–10 h in the initial stages of the infection, with each


RESERVOIRS FOR HIV-1

667

infected cell giving rise to approximately 20 new infected cells (229, 230). During symptomatic primary infection, levels of infectious virus and of infected cells in the circulation are both very high (15). The concentration of virus particles in the blood is often .106/ml and can be as high as 108/ml. The initial systemic seeding of the peripheral lymphoid organs with HIV-1 occurs as a result of the high levels of viremia that develop during primary HIV-1 infection. Interestingly, studies in the simian immunodeficiency virus (SIV) and simian/human immunodeficiency virus (SHIV) models of HIV-1 infection suggest that the large pool of CD4` T cells in the gut-associated lymphoid tissue (GALT) is an important site for viral replication in primary infection as well as during subsequent phases of the disease (16, 17). This likely reflects the fact that a relatively high fraction of CD4` T cells in the GALT are in an activated state and are therefore permissive for HIV1 replication. Within a few weeks, the level of virus in the blood declines. This decline coincides with the development of an immune response to HIV-1. Virus-specific cytolytic T lymphocytes (CTL) appear early and may represent a critical host factor in the control of primary HIV-1 infection (18–21). Recent studies in the SIV model have convincingly demonstrated the importance of CD8` CTL in the control of the viremia of primary infection. Monkeys depleted of CD8` T cells with monoclonal antibodies are unable to control primary infection (21). The combined effects of CTL and other elements of the immune response cause the amount of virus in the blood to decrease to a lower plateau level or set point. However, as is discussed below, sterilizing immunity is never achieved. The second phase of HIV-1 infection is the long asymptomatic period between primary infection and the development of clinical immunodeficiency (AIDS). There are two related pathophysiologic characteristics of the asymptomatic phase: ongoing viral replication in the peripheral lymphoid tissues and gradual loss of CD4` T cells. Although the asymptomatic phase may represent a period of clinical latency, the virus replicates continuously during this time. Free virus can be detected readily in the circulation by appropriate methods (22), although the levels are lower than those observed during symptomatic primary HIV-1 infection or, in patients with AIDS, the third and final stage of the infection. Infected cells are readily detectable in the peripheral lymphoid tissues, and their number correlates directly with plasma virus levels, suggesting that the lymphoid tissues are a major source of plasma virus (23–27). The rate of decline in CD4` T cells appears to be determined by the level of ongoing viral replication, as patients with higher plasma virus set points progress to AIDS more rapidly (28). The mechanisms underlying CD4 depletion are incompletely understood but probably include a decrease in T cell production by the thymus and an increase in the rate of destruction of T cells in the periphery (29). For a discussion of T cell dynamics in HIV1 infection, the reader is referred to several excellent reviews (11, 30). The final phase of the infection is characterized by the emergence of clinical immunodeficiency. In the year or two before AIDS develops, there is often a more rapid decline in CD4` T cells. This decline may be preceded by an increase in

668


viral load (31), with viral replication occurring in many sites in addition to the lymphoid tissue (32). In some cases disease progression is associated with the evolution of more pathogenic viral species that utilize the chemokine receptor CXCR4 instead of CCR5 (33–36). As the CD4 count falls below 200 cells/ul, opportunistic infections begin to occur. The degree of CD4 decline is an excellent predictor of the risk for particular infections, providing strong support for the notion that the loss of CD4` T cells is the central cause of immunodeficiency in this disease.


VIRAL DYNAMICS The pathophysiology of HIV-1 infection and the mechanisms of viral persistence are best considered in the context of the dynamics of viral replication in vivo. Understanding viral dynamics requires a steady-state analysis of the amount of free virus and the number of virally infected cells present in infected individuals (the viral load) and a dynamic analysis of the rates at which virus particles and virally infected cells are generated and cleared. Substantial progress has been made in understanding key elements of HIV-1 dynamics, and the paradigms developed have already proven useful in understanding the pathogenesis of other infectious diseases (37).

Extent of Infection of CD4` T Cells Different experimental approaches have been used to measure infected cells, including limiting dilution or focal analysis of the frequency of cells in the peripheral blood capable of producing infectious virus (38–40), PCR analysis of viral DNA in cells in peripheral blood (41–48), lymph nodes (23–25, 48) and the central nervous system (49), and analysis of viral mRNA species in productively infected cells (27, 32, 50–53). Unfortunately, there is still tremendous uncertainty over the fundamental question of what fraction of CD4` T cells are infected, with widely divergent answers coming from different experimental approaches. Part of the confusion results from the fact that many of the experimental approaches used to study this issue do not provide information about the distribution of virus between latent and active states, between defective and replicationcompetent forms, between CD4` T cell and macrophage compartments, and between circulating and tissue sites. All of these factors are important in understanding viral load. Determining the fraction of CD4` T cells infected in vivo is important because it bears on the critical question of what causes the loss of CD4` T cells that leads to immunodeficiency. Because HIV-1 is cytopathic for T cells in vitro, it was originally presumed that direct viral cytopathic effects were responsible for CD4 depletion. However, subsequent studies of viral load in HIV-1 infection have raised significant questions about whether the fraction of cells infected is high enough to account for the depletion of the entire CD4` T cell compartment. An



669

early in situ hybridization study suggested that even in patients with AIDS, only in a small fraction (, 0.01%) of the total pool of lymphocytes was productively infected, as judged by expression of high levels of viral mRNA (50). Subsequent PCR studies detected HIV-1 DNA in a larger fraction of CD4` T cells and were interpreted as indicating that most infected cells were in a state of latent infection. Nevertheless, in most conventional DNA PCR studies, cells carrying HIV-1 DNA have been detected only at low frequency (0.01–0.1%) among peripheral blood mononuclear cells (PBMC) of patients in the asymptomatic phase of the infection (42, 43). Studies using in situ PCR have detected HIV-1 DNA in a much higher fraction of cells (as high as 1–40% of PBMC or lymph node cells) (24, 45, 47). The reasons for the discrepancy are unclear. Both conventional and in situ methods may detect both cells with integrated HIV-1 provirus and recently infected cells in which the HIV-1 genome has been reverse transcribed but is not yet integrated (see below). The in situ PCR results are at odds not only with conventional PCR assays but also with virus culture studies in which a much lower frequency of virally infected cells is typically detected (38). Careful analysis of the fraction of PBMC from which virus can actually be cultured has shown that cells harboring replication-competent provirus are actually rare (,0.01%) (38, 40). Since productive infection requires integration of reverse transcribed viral DNA into the host genome (54, 55), the extent of infection can also be addressed by the analysis of the frequency of CD4` T cells carrying integrated provirus. Studies using inverse PCR to detect the junction between integrated HIV-1 proviruses and host chromosomal DNA have shown that in individuals in the asymptomatic phase of infection, integrated virus is found in ,0.01% of CD4` T cells in the peripheral blood and lymph nodes (40, 56). It is important to keep in mind that there may be differences in viral burden in different tissues. Lymphocytes in the circulation represent only a small fraction (,2%) of the total lymphocyte pool. During the asymptomatic phase of the infection, the proportion of cells carrying HIV-1 DNA is three- to tenfold higher in the lymph node than in the peripheral blood (23). This difference may be due to the fact that activated CD4` T cells, which are the primary targets for productive HIV-1 infection, tend to remain in the lymphoid tissues due to expression of high levels of adhesion molecules. The proportion of resting CD4` T cells harboring integrated virus is not dramatically different in the blood and lymph nodes, as is consistent with the continual recirculation of resting CD4` T cells (40). To summarize, although CD4` T cells productively infected with HIV-1 can be readily detected in the blood and lymphoid tissues of untreated patients with HIV-1 infection, the fraction of CD4` T cells that carry integrated HIV-1 DNA at any given time is very small.

The First Phase of Decay The development of rapid and quantitative RT-PCR methods for detecting genomic viral RNA in virions in the plasma has proven to be of extraordinary value in understanding viral load in HIV-1 infection. The critical initial obser-


670


vation was that HIV-1 RNA was present in the plasma of virtually all infected individuals, even during the asymptomatic phase (22). During this phase, levels ranging from 102 to 107 copies/ml are seen in different patients. Over the short term, levels are reasonably constant in individual patients, indicating a quasisteady state (22). Studies of the drug-induced perturbations of these steady-state levels (1, 57–61) have provided the critical insights upon which current models for viral dynamics are based. In 1995, pioneering studies by Wei et al (58) and Ho et al (57) showed that virus production in infected individuals is largely the result of a dynamic process involving continuous rounds of de novo infection of and replication in host cells, with rapid turnover of both free virus and virus-producing cells. In these studies, nevirapine and ritonavir, potent inhibitors of HIV-1 reverse transcriptase and protease, respectively, were given as single agents to patients with HIV-1 infection, and plasma virus levels were monitored closely. Both types of drugs act to prevent new rounds of infection without blocking virus production by cells that already carry integrated HIV-1 DNA. Both agents produced a dramatic (2 log) drop in plasma virus levels in the first two weeks of therapy (Figure 1). This result indicates that the half-lives of plasma virus and of the cells that produce the vast majority of the plasma virus are both very short. Perelson and colleagues then attempted to measure separately the two processes that contribute to the rapid initial decay of plasma virus: the clearance of free virions and the loss of the productively infected cells (60). They examined plasma viral levels early after the initiation of therapy and fit the experimental data with a model consisting of a set of differential equations describing the dynamics of infected cells and free virus particles. Because the logarithm of the plasma virus concentration fell linearly with time during the initial treatment period, the decay of plasma virus and of productively infected cells was described in terms of exponential processes. The simplest expression describing the dynamics of HIV-1 production and clearance is: dV/dt 4 NdT * 1 cV, where V is the concentration of free virus particles, N is the number of virus particles released by a productively infected cell, d is the decay rate constant for productively infected cells, T* is the concentration of productively infected T cells, and c is the decay rate constant for free virus. Using this approach, decay rate constants for free virus (c) and for the cells that produce most of the plasma virus (d) were calculated. The resulting values for c were strikingly similar in different patients, consistent with the idea that free virus in the plasma has a constant intrinsic decay rate (60). The same was true for d, the decay rate of productively infected cells. Mean values for c and d were 3 and 0.5 day11, respectively. These rate constants are most easily evaluated in the form of a half-lives (t1/2 4 ln 2/c, ln 2/d, respectively). For free virus, the half-life is ,6 h. Recent particle infusion studies in the SIV system suggest that virions may last in the plasma for only a matter of minutes before being cleared by unknown mechanisms



671

Figure 1 Hypothetical decay curve for plasma virus levels in a patient treated with highly active antiretroviral therapy (HAART). The first phase of decay reflects the short half-life of plasma virus and of the productively infected CD4` lymphoblasts that produce most of the plasma virus. The second phase of decay reflects the longer half-life of a second population of virus-producing cells. The plasma virus levels fall to a new set point that is generally below the level of detection of current assays. Virus replication continues at a low level, with only very occasional viremic episodes (‘‘blips’’). In the hypothetical case of treatment with regimens that completely prevent new infection of susceptible cells by extracellular virus, plasma virus levels would drop further until they reached the extremely low levels associated with activation of cells in the latent reservoir. The intrinsic decay rate of this reservoir remains to be determined.

(62). The most surprising result is that the half-life of the cells that produce most of the plasma virus is only about 1–2 days (60, 61). The lability of both the plasma virus and the cells that produce most of the plasma virus underscores the fact that virus replication is active and ongoing throughout the course of the disease. Direct confirmation of the rapid turnover of productively infected cells has been provided by Haase and colleagues using in situ hybridization analysis of tonsillar biopsies from patients starting antiretroviral therapy (61). During the asymptomatic phase of the infection, plasma HIV-1 RNA concentrations reach a quasi-steady state, with little change over a time frame of days. Thus, at steady state, dV/dt 4 0 and NdT* 4 cV, and the plasma virus concentration (V) directly reflects the number of productively infected cells (V 4 (Nd/c)T*) (60). Given that the half-life of the cells that produce most of the plasma


672


virus is 1–2 days, it follows that the steady-state level of viral RNA genomes in the plasma reflects very recent virus production. Thus, the measurement of plasma virus levels provides a real time measure of the rate of replication of the pathogen in the host, something not available in the case of most infectious diseases. As is discussed below, in vivo measurements of virus replication rates have provided important insights into the pathogenesis of the disease. It is generally assumed that activated CD4` T cells are responsible for producing most of the plasma virus and that the rapid initial decay in plasma virus levels following the initiation of potent antiretroviral therapy reflects the rapid turnover of these cells. Immunohistochemical studies suggest that in the secondary lymphoid tissues, most productively infected cells are T lymphocytes (11). The mechanisms involved in the rapid turnover of productively infected CD4` T lymphoblasts remain unclear. HIV-1 is cytopathic for CD4` T cells in vitro, and the presumption has been that cytopathic effects of HIV-1 on productively infected CD4` T cells are involved. A number of different mechanisms of cell killing have been invoked (63–67). Of particular interest is recent work showing that the product of the HIV-1 vpr gene can induce apoptosis (68, 69). Previous studies had shown that vpr induces arrest of the cell cycle in G2 (70–74). This arrest may increase the amount of virus produced by CD4` lymphoblasts (73, 75), but ultimately results in death of the infected cell. These findings on the proapoptotic effects of vpr add to previous studies that have focused on the HIV-1 env protein as a principal mediator of cytopathic effects (for a review, see reference 76). The rapid decay of productively infected cells may also reflect the killing of these cells by host cytolytic effector mechanisms, especially CTL (77, 78). Nowak and colleagues have argued that the frequency of activated effector CTL in vivo is high enough to account for this rapid turnover (79). It is now well established that CTL can actually lyse HIV-1 infected cells (80). The evidence for a beneficial role of CTL has been accumulating and is reviewed elsewhere (81, 82). Perhaps the most direct evidence comes from recent studies showing that viral load increases dramatically following experimental depletion of CD8` T cells in SIVinfected rhesus monkeys (21, 83). In addition, recent studies using new methods to enumerate antigen-specific CTL have confirmed that the frequency of CTL specific for HIV-1 is indeed very high in many infected individuals (84). On the other hand, it has recently been suggested that the nef protein may function to downregulate class I molecules, thereby protecting infected cells from lysis by CTL (85). Additional studies will be needed to determine whether the short halflife of most productively infected cells is due to lysis by CTL (79, 86), death resulting from cytopathic effects of viral proteins, or both. Measurement of the viral clearance rate c allows calculation of P, the rate of virus production, using the equation P 4 cV. Figures in excess of 1010 virions per day have been reported (60). The high levels of virus produced per day are instructive in several regards (see reference 59). The high rate of virus production, when considered in the context of the short half-life of the cells that produce most



673

of the plasma virus, suggests that the rate of new infection of CD4` T cells is substantial. Uncertainty over the number of virions produced per infected cell (the burst size, N) makes it difficult to calculate precisely the number of productively infected cells (T*) from the steady-state relationship NdT* 4 cV. Nevertheless, it is clear that in an untreated patient, large numbers of CD4` T cells become infected daily. The reverse transcription process that occurs in each of those newly infected cells has a sufficiently high error rate that viral genomes with every possible single point mutation arise daily (59). Perelson and colleagues have estimated that if 108 new cells are infected per day, then not only are all possible single point mutations generated daily, but almost 1% of all possible double mutations are produced each day (87). This finding has enormous implications for understanding the evolution of drug resistance mutations and viral escape from immune effector mechanisms. The high rate of virus production also has implications for the debate over mechanisms of CD4` T cell depletion. The characteristics of the first phase of decay suggest that productively infected cells die quickly, but it has been unclear whether this effect alone is sufficient to account for CD4` T cell depletion. The absolute number of CD4` T cells that die per day from infection is still unknown. Some of the death of CD4` T cells in vivo may result from indirect effects on uninfected CD4` T cells (88, 89). CD4 depletion may also be a consequence of decreased thymic production of T cells (90–92). Recent evidence suggests that thymopoesis continues in healthy adults (93), but this residual thymic function may not be enough to counterbalance the destructive mechanisms. An excellent review of these complex issues has recently been published (94).

The Second Phase of Decay Although resistance developed in a matter of weeks in the initial studies of nevirapine and ritonavir monotherapy (58), further studies showed that combinations of antiretroviral agents could produce a decline in plasma virus to undetectable levels in many patients (1–3). After the rapid initial decay during the first 1–2 weeks of treatment, plasma virus declined at a slower rate (1). This was interpreted as reflecting the turnover of a longer lived viral reservoir or infected cell population. Perelson and colleagues (1) modeled the second phase of decay by incorporating additional terms into the above equation to account for the low level of virus production by a longer lived population of virus producing cells, M*, which were assumed to produce virus at a constant rate p. Thus, dV/dt 4 NdT * ` pM * 1 cV This long-lived cell reservoir accounts for only a small portion of the total virus production in an untreated individual (NdT* .. pM*) and becomes evident only when the cells that produce most of the plasma virus have largely decayed (Figure 1). The half-life of the compartment responsible for this second phase of decay was estimated to be 1–4 weeks (1). A biphasic decay process was also observed directly in in situ hybridization studies of productively infected mononuclear cells


674


in the lymphoid tissues (61). Recent studies in children with perinatally acquired HIV-1 infection have shown that the same biphasic decay process occurs in pediatric patients (95). The nature of the cellular or anatomical reservoir responsible for the second phase in the classic viral decay curve is still unclear. These cells may be macrophages, which at least in vitro are less susceptible to the cytopathic effects of the virus than are CD4` lymphoblasts (96, 97). Because HIV-1 is less cytopathic for macrophages (97), infected macrophages can in principle continue to release virus for their normal life span. In uninfected individuals, macrophage turnover is balanced by continuous production of new monocytes in the bone marrow. Monocytes circulate for less than a day and enter the tissues where they differentiate into macrophages and persist with a t1/2 of about 2 weeks (98), consistent with the kinetics of the second phase. It is currently unclear what fraction of macrophages in different tissues are infected. In most published studies, the fraction of infected macrophages is very low (40, 99, 100). Thus a great deal remains to be learned about the potential role of infected macrophages as a long-term reservoir for HIV-1. It is also possible that the second phase is due to the turnover of CD4` T cells that are in a different state of activation. As is discussed below, the virus does not replicate in T cells that are in a resting state. However, there may be partially activated states that are permissive for lower levels of replication, levels that do not cause rapid destruction of the cells (232). The virus that appears during the second phase of decay may result from the remobilization of free virions trapped on follicular dendritic cells (FDC) in the germinal centers of the peripheral lymphoid tissue (24, 101–104). These cells do not appear to be productively infected (105) but can trap virus particles on their surfaces. This pool of trapped extracellular virions declines with a t1/2 on the order of two weeks in patients on effective combination therapy (61). While there is evidence that trapped virions bound to FDC can retain infectivity (106), the bound virus is likely to be complexed with antibody and complement, and it is not clear how this trapped virus can serve as a source for virus in the plasma during the second phase. Although initial viral dynamic models have been extremely useful in understanding pathogenesis and the response to therapy, some questions remain. Models of viral decay have been built upon the assumption that no new cells are infected once combination therapy has started. This notion has recently been challenged by Grossman and colleagues who have suggested that the death of infected CD4` T cells is more appropriately modeled as an aging process in which cells that have been infected for a longer period of time have a higher probability of dying. They have argued that the use of an exponential decay model is not appropriate because it does not incorporate this notion that the chance an infected cell will die is related to the length of time a cell has been infected (107, 231). In their model, combination therapy does not completely block new infection of cells. Rather it results in decrease in the amplitude of local bursts of HIV-1 rep-



675

lication. In support of this assertion, these authors have shown that treatment with five drugs results in a more rapid decay in plasma virus levels that does treatment with three drugs, consistent with the idea that current combination therapy is not stopping all new infection of susceptible cells (231). They go on to develop a model for HIV-1 persistence that involves local bursts of viral replication associated with immune activation that are attenuated but not fully suppressed by combination therapy. An increasing body of evidence suggests that there may be some low level of ongoing viral replication even in patients whose HIV-1 plasma RNA levels are below the limit of detection (see below). In another theoretical paper, Bucy has argued that distinct cell populations need not be invoked to explain the first and second phases of decay (86). He proposes that a change in the rate of immune clearance following the initiation of therapy is sufficient to explain the two phases. In his model, declining antigen load following the initiation of therapy leads to a waning of HIV-1-specific immune responses, a notion for which there is increasing experimental evidence (108, 109). This in turn reduces the rate of clearance of infected cells. Whether the decline in HIV-1specific immune responses is rapid enough to account for the two phases that are routinely observed is unclear. Although it is likely that viral dynamics can be effectively modeled in several different ways, it remains clear that patients starting a combination therapy regimen experience a very characteristic biphasic decay in plasma virus levels, consistent with the idea of distinct compartments with different kinetics.

The Eradication Hypothesis After two months of highly active antiretroviral therapy (HAART), the level of plasma virus falls to below the limit of detection (20–500 copies/ml, depending on the assay used) in a fraction of patients. This fraction can be as high as 80– 90% in some studies, particularly in previously untreated patients (2, 3). In the general patient population, it is more typically 50–60%. In patients whose plasma HIV-1 RNA levels are below the limit of detection, it becomes difficult to culture the virus from the blood. There has been hope that in these patients, prolonged treatment might lead to eradication. Perelson et al were the first to use an analysis of viral decay rates in treated patients to make rational predictions of treatment times required for virus eradication (1). These investigators attempted to extrapolate the second phase of decay to zero residual infected cells. Since the nature and total body number of cells responsible for virus production during the second phase were unknown, they assumed that the initial number of chronically infected cells could be as high as 1012, equivalent to the total number of lymphocytes in the body. Even if the number of long-lived infected cells is this high, the secondphase decay extrapolates to ,1 residual infected cell in about three years. However, the prediction of eradication was made with the caveat that there may be undetected compartments or viral reservoirs that are not measurable by standard techniques (1). It is now clear that the potential of prolonged combination therapy

676


to eradicate the virus depends on the characteristics of these compartments and long-lived reservoirs for HIV-1.


RESERVOIRS FOR HIV-1 Potential reservoirs for HIV-1 include various types of long-lived infected cells in various locations in the body. For example, replication-competent HIV-1 can be recovered from cells in the seminal fluid of men on combination therapy who had no detectable plasma virus (110). The nature of the cells harboring HIV-1, the extent of infection of these cells, and the half-life of the cells remain unclear. Nevertheless, this result has obvious implications for the possibility of transmission from patients whose plasma HIV-1 RNA levels are undetectable. Previous studies have demonstrated the presence of HIV-1 in T cells and macrophages in the semen (111) and have suggested that there is some degree of compartmentalization of virus in the male urogenital tract (112).

THE CENTRAL NERVOUS SYSTEM AS A RESERVOIR FOR HIV-1 Another potentially important reservoir for HIV-1 is the CNS (for a review, see 113). In the pre-HAART era, neurological problems were common among infected individuals, with a unique dementia syndrome, HIV-associated dementia (HAD), developing in 15–20% of patients (114). This syndrome was the result of direct effects of HIV-1 on the CNS and could be distinguished from CNS diseases caused by opportunistic infections. The capacity of HIV-1 to enter the CNS and cause disease raises important questions about whether the virus can persist there. HIV-1 probably gains access to the CNS from the blood stream, either by direct infection of capillary endothelial cells (115) or, more likely, by ingress of infected monocytes/macrophages. The triggering mechanisms for initial monocyte/macrophage recruitment to the brain are unknown but may involve the upregulation of chemoattractant b-chemokines such as MCP-1 (116) and the expression of adhesion molecules on endothelial cells. Studies using an artificial blood brain barrier demonstrated that upregulation of adhesion molecules and proinflammatory cytokines are critical for transendothelial migration (117). Heightened trafficking may occur with peripheral activation of monocytes in late stage HIV-1 infection, which is generally when HAD occurs. The current consensus is that the principal cellular target for HIV-1 in the CNS is the macrophage or microglial cell. A large study in clinically well-characterized adults found no convincing evidence for HIV-1 DNA in neurons, endothelial cells, or oligodendrocytes (118). Progress in the understanding of the extent of infection within the CNS has been hampered by the obvious difficulty in obtaining tissue and by uncertainties



677

with regard to the relationship between level of virus in the cerebrospinal fluid (CSF) and levels of virus in the brain parenchyma (119). The development of an excellent animal model for AIDS dementia (120) may facilitate analysis of this important problem. In an extensive study in the SIV system, SIV-infected cells were not detected in the CNS in monkeys during the asymptomatic stage of the infection (32). Following progression to AIDS, SIV infection in the CNS was detected by in situ hybridization (32, 120). Certain strains of HIV-1 might have an increased propensity to invade (neurotropism) and cause damage in the nervous system (neurovirulence). This is significant because the development of HAD is not universal in advanced AIDS, suggesting that there may be viral determinants of heightened risk. Indeed distinct strains of HIV-1 isolated from both peripheral blood and the nervous system of the same individual can have different biological characteristics and cellular tropisms (33, 121). Brain isolates tend to be more macrophage-tropic with specifically conserved regions in a portion of the envelope, the V3 domain (122, 123). One approach to evaluating viral load in the CNS is to measure virus in the CSF. There is a significant correlation between CSF HIV-1 load and the severity of neurological disease (119, 124–126). However, it remains unclear whether CSF HIV-1 RNA levels actually reflect brain tissue levels. Different brain regions tend to have similar levels of HIV-1 RNA, but there is only a weak correspondence between brain and CSF HIV-1 levels (119). Potential sources of CSF HIV-1 RNA include the meninges, choroid plexus, parenchyma, and trafficking lymphocytes and monocytes. Presumably, the parenchymal levels are the most relevant for the study of neurological disease. CSF HIV-1 RNA might derive from different sources at different stages of HIV infection. Price & Staprans (127) have suggested that transitory infection due to trafficking cells should be distinguished from autonomous infection involving parenchymal infection of macrophages and microglia. Given that HIV-1 can enter the CNS, there is concern that the virus might persist there and produce CNS disease even in treated patients who have no detectable plasma virus. This scenario is plausible because of the limited CNS penetration of certain protease inhibitors (128). Other important factors include the active efflux of antiretroviral drugs through transporters including pglycoprotein (129). Early case reports alerted clinicians to this possibility, describing patients who had undetectable or low plasma HIV RNA levels yet significantly higher CSF HIV-1 RNA levels (130). However, despite these concerns, there have been relatively few clinical examples of ‘‘CNS escape.’’ In fact, significant reductions in the incidence rates of HAD have been noted since 1996 (131–133). There is now accumulating evidence that HAART regimes can actually improve neuropsychological performance and radiological abnormalities in those with HAD (134–136). Given these findings, it remains unclear whether the CNS acts as a reservoir capable of reseeding the systemic compartment. Some perivascular macrophages return to the periphery after a sojourn within the brain (W Hickey, personal com-


678


munication). Theoretically, these could introduce CNS-derived strains of HIV into the systemic compartment. To date, however, there is no evidence that this occurs in vivo. There is still relatively little information about the effects of antiretroviral therapy on viral replication in the CNS. Several studies have reported successful suppression of HIV in CSF with the long-term use of protease inhibitor containing regimens (137–141). The kinetics of virologic suppression may differ in the CSF compartment, lagging behind the more rapid response seen in plasma (142, 143) and sometimes failing to suppress to undetectable levels despite the use of a CNSpenetrant regimen (144). In one study the half-life of HIV-1 RNA in the CSF compartment was significantly longer than in the blood, falling by 1 log over 20 days (t1/2 4 5 days) (143). In summary, the CNS clearly represents a relatively poorly understood site with respect to HIV-1 replication. The CNS is actually several different compartments including the brain parenchyma and the CSF in addition to trafficking monocytes. While HAART has had a measurable impact on the incidence rates of HAD, the possibility exists that CNS escape might occur, particularly in individuals with poor medication adherence or using antiretroviral regimes with poor CNS penetration. Whether the CNS can serve as a long-term reservoir for HIV1 in patients with good suppression of viral replication remains to be determined.

A STABLE RESERVOIR OF LATENTLY INFECTED RESTING CD4` T CELLS Physiologic Basis of Latent Infection of CD4` T Cells At the present time, the reservoir that appears to be the major barrier to achieving eradication of HIV-1 is the extremely stable reservoir composed of latently infected resting CD4` T cells with integrated provirus (40, 56). These cells are not found exclusively in any particular tissue location, but rather populate the peripheral lymphoid tissues and recirculate continuously between them. The existence of a latent reservoir for HIV-1 in resting CD4` T cells can be best explained by considering the infection of CD4` T cells in the context of the normal physiology of T cell activation (Figure 2). Naive CD4` T cells exit the thymus and enter the peripheral lymphoid tissues in a process that continues throughout life (93). These newly generated T cells persist in a resting state until they encounter the relevant antigen. Following initial exposure to antigen, naive T cells undergo blast transformation and enter the cell cycle. Activation results in metabolic changes including increases in nucleotide pools and in the upregulation of expression of sets of proteins that allow the cell to carry out its functions. These include transcription factors, effector molecules such as cytokines, and cell surface proteins such as cytokine receptors and adhesion molecules (145). Many of the activated T cells die within a few weeks after activation, either as a result of additional



679

Figure 2 Cellular dynamics of HIV-1 infection of CD4` T cells. Transitions between resting (small) and activated (large) CD4` T cells are illustrated by vertical arrows. The normal generation of memory CD4` T cells is illustrated on the left. These cells are derived from antigen (Ag)-activated CD4` T cells that revert back to a resting memory state. These memory cells survive for long periods of time, allowing responses to the same antigen Ag in the future. The memory cell compartment may be maintained by a process of proliferative renewal. Successive steps in the life cycle of the virus are indicated by horizontal arrows. R5 isolates can infect activated CD4` T cells but may infect only the subset of resting memory CD4` T cells that express sufficient amounts of CCR5. Following infection of resting memory CD4` T cells, there is a block in the virus life cycle, probably at the level of nuclear import of the preintegration complex containing the viral genome. Resting cells with unintegrated HIV-1 DNA are likely to represent a relatively labile reservoir for the virus (preintegration latency). Productive infection requires Agdriven activation of recently infected resting CD4` cells or, more commonly, direct infection of Ag-activated CD4` T cells. Productively infected cells generally die within a few days from cytopathic effects of the infection or host cytolytic effector mechanisms, but some infected lymphoblasts survive long enough to go back to a resting state (boxed), thereby establishing a stable latent reservoir of resting memory CD4` T cells with integrated HIV-1 DNA (postintegration latency).

antigen-induced activation, which can trigger expression of pro-apoptotic regulatory molecules such as FasL and TNF-a, or as a result of diminishing concentrations of cytokines needed to promote survival (reviewed in 146). Some of the activated cells escape these death pathways, exit the cell cycle, lose expression of activation markers such as HLA-DR, and revert to a resting state in which they


680


persist as memory T cells capable of responding to subsequent exposures to the initiating antigen. The return to quiescence may involve expression of the transcription factor LKLF, which functions in an unknown manner to decrease FasLassociated killing and to regulate T cell quiescence and survival (147). The survival of resting CD4` T cells may be dependent upon cytokines including IL4, IL-6, and IL-7 (148, 149). This linear differentiation model in which memory T cells are derived from activated effector cells that escape programmed cell death has been recently confirmed for CD8` T lymphocytes (150) and likely holds for CD4` T cells as well. The state of activation and history of antigen exposure of a given T cell can be discerned by analysis of the patterns of expression of certain cell surface proteins (151–153). Altered patterns of expression of some cell membrane proteins reflect differences between the resting and activated states. For example, T cell activation results in rapid and temporary expression of CD69 and of the a chain of the IL-2 receptor (CD25) as well as a slower and more prolonged upregulation of HLA-DR that persists while the cell is in an activated state. Antigendriven activation also leads to an essentially permanent change in the expression of other cell membrane proteins. These changes can be used to distinguish naive and memory T cells (151). For example, naive and memory T cells express different splice variants of the membrane tyrosine phosphatase CD45 (CD45RA and CD45RO, respectively) (153–156), and naive cells express the lymphocyte homing receptor L-selectin (CD62L), which is typically not found on memory T cells (157). Phenotypic analysis of naı¨ve and memory cells is somewhat hampered by the finding that memory cells can revert to a naı¨ve (CD45RA`) phenotype (158– 160). The ability of HIV-1 to bind to and enter resting and activated CD4` T cells is a function of the expression of sufficient levels of both CD4 and an appropriate chemokine receptor and depends upon the chemokine receptor utilization capacity of the env protein of the viral isolate in question (13). The most commonly transmitted forms of HIV-1 utilize the chemokine receptor CCR5 as a co-receptor. Because CCR5 is upregulated upon T cell activation, these viruses can bind to, infect, and replicate in activated CD4` T cells. The infection of resting CD4` T cells by viruses utilizing CCR5 appears both inefficient and restricted to those of the memory phenotype. Infection of resting memory cells by these viruses can be observed only in the presence of a large inoculum and likely represents the infection of the small fraction of cells with continued low level CCR5 expression (T Pierson & RF Siliciano, unpublished data). Even when CCR5-utilizing viruses enter resting memory CD4` T cells, there is a block in the replication cycle such that no virus is produced (see below). Viruses that utilize the chemokine receptor CXCR4 can bind to and enter resting naı¨ve and memory CD4` T cells and activated CD4` T cells because CXCR4 is expressed at sufficiently high levels on all of these cell types to mediate the entry. However, there is again a block in the replication cycle in resting CD4` T cells. Since at any given time the majority


681

of lymphocytes are in a resting state, it is important to understand the consequences of infection of resting CD4` T cells.


Preintegration Latency Although HIV-1 does not replicate in resting CD4` T cells, it can establish a state of latent infection in resting cells through two mechanisms, both of which are operative only in cells that are in a quiescent G0 state (5). A preintegration form of latency is observed immediately following direct infection of resting CD4` T cells. As discussed above, HIV-1 virions can bind to and fuse with resting CD4` T cells under some conditions. This is followed by reverse transcription of the genomic RNA, a process that occurs in a large preintegration complex containing the viral genome and virion proteins including reverse transcriptase, integrase, matrix, and vpr. Previous studies suggested that the post-entry block in the virus life cycle in resting CD4` T cells resulted from either an inability to complete the reverse transcription reaction (161, 162) due to low nucleotide pools (163), or failure to import the ribosome-sized preintegration complex into the nucleus (165). In short-term in vitro infections of resting CD4` T cells, reverse transcription does not proceed to completion. However, in longer term in vitro cultures and in in vivo studies, complete reverse transcripts can be found in resting CD4` T cells (40, 44, 164; T Pierson & RF Siliciano, manuscript in preparation). These results suggest that although reverse transcription is slow in resting CD4` T cells, an additional block exists, probably at the level of nuclear import (165). Interestingly, recent experiments using retroviral vectors to transduce primary CD4` lymphoblasts have shown that overexpression of the transcription factor NFATc is sufficient to allow previously activated CD4` T cells to remain completely permissive for productive HIV-1 infection at time points (17 days post activation) when control, nontransduced cells are no longer able to complete reverse transcription (166). Although infected resting cells do not normally produce infectious virus, pioneering studies by Zack et al (161) and Bukrinski et al (44) demonstrated that recently infected resting CD4` T cells can function as an inducible latent reservoir for HIV-1. If an infected resting T cell is activated by antigen before the preintegration complex becomes nonfunctional, then the subsequent steps of nuclear import, integration into host chromosomes, virus gene expression, and release of infectious virions can all occur (44, 161, 164). Although there is some controversy in the literature over exactly how long functional preintegration complexes can persist in resting CD4` T cells, this preintegration form of latency appears to be relatively labile, persisting a matter of days to weeks. It is important to keep in mind that, in untreated patients, resting CD4` T cells with this unintegrated form of HIV-1 DNA are much more prevalent than cells with stably integrated HIV-1 DNA and that this unstable form of latency is detected in virus culture assays in which the resting cells are subjected to activating stimuli (40, 44, 46).

682



Postintegration Latency A more stable form of latency may occur in CD4` T cells that have undergone integration of proviral DNA (40, 52, 56, 167–169). Because nuclear import is dependent upon T cell activation, postintegration latency can, in principle, result only from the return of an activated T cell with integrated provirus back to a resting state in which there is minimal expression of viral genes (Figure 2). In order for a productively infected CD4` T lymphoblast to enter a condition of postintegration latency, it must survive the homeostatic mechanisms that kill the majority of activated T cells, the cytopathic effects of the virus, and virus-specific cytolytic host effector mechanisms. The cell must survive for long enough to allow it to revert to a resting memory state in which there is minimal expression of HIV-1 genes. This form of latency may arise when there is infection of activated cells that are in the process of transitioning back to a resting state. In this situation, the cells may be permissive for reverse transcription, nuclear import, and integration, but not fully permissive for virus gene expression (7). The cells therefore escape the rapid destruction that is the usual fate of productively infected cells. Recent in vitro experiments using pseudotyped HIV-1 vectors have demonstrated that certain cytokines, including IL-2, IL-4, IL-7, and IL-15, can provide signals that render resting CD4` T cells permissive for nuclear import, integration, and virus gene expression (170). In the absence of cytokine stimulation, transduction was not seen. These results raise the possibility that infected resting cells may transit from a labile preintegration state to stable form of latency through both antigen-dependent and cytokine-dependent mechanisms.

Evidence for Postintegration Latency in Vivo Although it had been long presumed that the integration of HIV-1 DNA into the genomes of infected CD4` T lymphocytes would allow viral persistence, until recently little direct evidence suggested that resting CD4` T cells with integrated provirus function as a latent reservoir for HIV-1 in infected individuals. In important early experiments, transformed cell lines carrying integrated HIV-1 DNA were used to model latent infection (167, 168, 171). Cell lines were infected in vitro, and surviving cells that grew out and contained integrated HIV-1 DNA were analyzed. As is discussed below, studies of these cell lines provided information about the molecular mechanisms involved in the regulation of HIV-1 gene expression. The proof that resting CD4` T cells with latent, integrated provirus are present in vivo required the development of techniques for the isolation of extremely pure populations of resting CD4` T cells as well as methods for the unambiguous detection of integrated HIV-1 DNA in the presence of excess unintegrated HIV1 DNA and the isolation of replication-competent virus from the purified resting CD4` T cells. In 1995, Chun et al used a special inverse PCR method to demonstrate unambiguously the presence of cells with integrated HIV-1 DNA in



683

extremely pure populations of resting CD4` T cells isolated from infected donors (56). The purification procedure used yielded cells that were CD4` but lacked expression of early (CD25, CD69) and late (HLA-DR) activation markers. The PCR method used amplifies the junction between the HIV-1 provirus and the upstream genomic DNA and does not detect unintegrated HIV-1 DNA. In the same study, replication-competent virus was obtained from these cell populations by stimulating the cells with the mitogen PHA. In the absence of mitogenic stimulation, virus cannot be recovered from highly purified resting cell populations, and therefore virus isolated from these cells may be considered to have come from a latent reservoir. The frequencies of cells with integrated HIV-1 DNA among the resting CD4` T cell populations were similar, and extremely low (0.05%), in the blood and lymph nodes of infected individuals who were not yet on effective therapy consistent with the continual recirculation of resting T lymphocytes (40). The frequencies were not higher in patients with advanced disease, suggesting that a relatively stable steady state is established in which only a minute fraction of the resting CD4` T cell population carries integrated HIV-1 DNA (40). The total body number of resting CD4` T cells with integrated HIV1 DNA was estimated to be approximately 107 cells, only a fraction of which carried replication-competent forms of the virus. Among resting CD4` T cells, integrated HIV-1 DNA was present primarily among cells with a memory phenotype (40). Recent studies using a virus culture technique to detect postintegration latency have also demonstrated that most latently infected cells have a CD45RO` phenotype (J Siliciano, RF Siliciano, unpublished results), consistent with the mechanism proposed in Figure 2. It has often been presumed that the proviral DNA in T cells represents mainly defective archival sequences. However, with enhanced virus culture conditions designed to induce uniform activation of resting CD4` T cells, replicationcompetent virus could be recovered from highly purified resting CD4` T cells (40). The frequencies of resting CD4` T cells with replication-competent provirus are lower than the frequencies of resting CD4` T cells with integrated HIV-1 DNA (as detected by inverse PCR). This suggests that some of the integrated HIV-1 DNA in resting CD4` T cells is defective. Nonetheless, the finding that replication-competent virus can persist in a latent form in resting CD4` T cells raised the possibility that this reservoir might represent a major barrier to virus eradication in patients on combination therapy (see below).

Molecular Mechanisms for Postintegration Latency What accounts for the ability of HIV-1 to persist in a latent, integrated form in resting CD4` T cells? Three general mechanisms are described below. It should be pointed out the latency in vivo may involve multiple mechanisms. 1. Inefficient or absent initiation of transcription. The simplest mechanism for latency involves the absence in resting CD4` T cells of transcription from the


684


HIV-1 LTR. This could result from proviral integration into chromosomal sites that become inaccessible to the transcription machinery in quiescent cells (172) or from the absence in resting cells of requisite forms of host transcription factors that activate gene expression from the HIV-1 LTR (167, 173–178). The U3 region of the HIV-1 LTR functions as the viral promoter (for reviews, see 179 and 180) and contains binding sites for host transcription factors that function as positive regulators of T cell activation–specific gene expression in normal, uninfected T lymphoblasts; these include Ets, NFAT, and NFjB. The LTR actually contains two tandem, highly conserved binding sites for NFjB. Following the pioneering studies of Nabel & Baltimore (174) showing that this inducible host transcription factor regulated HIV-1 gene expression, Fauci and colleagues explored the notion that latent infection of T cell might involve the absence of the requisite host transcription factors (167, 173, 181). They showed that in transformed T cell lines carrying an integrated copy of the HIV1 genome, upregulation of HIV-1 gene expression following exposure to TNFa was mediated through NFjB (173). More recently, it has been shown that reporter viruses with mutations in the NFjB binding sites replicate poorly in activated CD4` T lymphoblasts, consistent with an important role for NFjB in HIV-1 gene expression (182) . In resting CD4` T cells, NFjB may be sequestered in the cytoplasm through interaction with IjB, and it has been suggested that the absence of a nuclear NFjB in resting CD4` T cell with integrated HIV-1 genomes prevents transcription of viral genes. Leiden and colleagues have suggested that HIV-1 gene expression is regulated by inducible T cell enhancons that are composed of Ets, NKjB, NFAT, and AP-1 proteins and that control the expression of a number of genes that are upregulated in activated T cells (183). However, the hypothesis that postintegration latency is regulated at the level of initiation of transcription has not been tested directly in primary CD4` T cells. Latently infected cells are rare in vivo, and it is therefore hard to carry out this kind of mechanistic study. An additional problem is that no in vitro model exists for a resting G0 CD4` T cell with integrated HIV-1 DNA. The transformed cell lines used in many of the published studies on latency may not mimic the transcriptional status of the extremely quiescent cells that constitute the latent reservoir for HIV-1 in vivo. 2. Failure of transcriptional elongation. A more recently identified mechanism for latency involves the absence in resting CD4` T cells of host factors that interact with the HIV-1 tat protein and allow tat-mediated upregulation of transcriptional elongation (53, 184–187). It is now clear that tat-mediated upregulation of HIV-1 gene expression is dependent upon association of tat with the tar sequence in the HIV-1 RNA in a process that involves the host proteins cyclin T1 and CDK9. These proteins are both components of the pTEFb CTD kinase complex, which phosphorylates the C-terminal domain of RNA polymerase II, a modification that increases the processivity of the enzyme. Expression of both cyclin T1 and CDK9 is upregulated upon T cell



685

activation (184), and levels in resting CD4` T cells may be too low to allow the tat-mediated enhancement of transcriptional elongation. 3. Failure to export unspliced viral mRNAs into the cytoplasm due to low levels of the HIV-1 rev protein (52, 168, 169, 188–190). In some cell lines carrying an integrated copy of the HIV-1 genome, upregulation of HIV-1 gene expression is associated with a shift in the predominant class of viral mRNAs made from the small 2-kb mRNAs encoding regulatory proteins tat, rev, and nef to the larger 4-kb singly spliced and 9-kb unspliced mRNAs coding env and gag/ pol, respectively. In the absence of sufficient threshold amounts of the rev protein, the 9-kb RNA that serves as a mRNA for gag and pol and as the viral genome may be spliced, thereby preventing the production of virus particles. This threshold effect may reflect a requirement for multimerization of rev on the rev responsive element in the viral RNA (190). Under conditions where only low amounts of viral mRNAs are made, the level of rev may be insufficient to prevent splicing of the full-length viral RNA molecules. According to this model, some multiply spiced transcripts may actually be made in latently infected cells. While numerous studies have detected multiply spliced HIV-1 RNA in unfractionated blood or lymph node cells from infected individuals, it has been unclear whether this RNA is present in latently infected cells since the studies have not been performed on purified resting CD4` T cells. Thus at the present time, the molecular mechanisms of postintegration latency remain unclear.

LATENTLY INFECTED CELLS AS A BARRIER TO VIRUS ERADICATION The demonstration that CD4` T cells in the postintegration state of latency were present in infected individuals (40, 56) raised the concern that these cells might function as a long-term reservoir for HIV-1. The importance of this reservoir derives from two fundamental aspects of the biology of these cells. First, as discussed in the previous section, the level of viral gene expression is likely to be very low in latently infected resting CD4` T cells, rendering the cells resistant to viral cytopathic effects and host cytolytic effector mechanisms. Second, resting CD4` T cells can live for long periods of time. The half-life of CD4` T cells in normal humans has not been extensively studied, but some estimates based on the survival of cells with DNA damage induced by high dose irradiation have been published (158, 159). In these studies, half-life of T cells expressing the RO isoform of CD45, found on memory cells, is in the range of 5–6 months (159). Because the forms of chromosomal damage studied cause death of the cell if it goes through mitosis, these studies actually measure only the intermitotic halflife of the cells, the time required for half of the cells to divide or die. Memory cells and their progeny may actually survive for much longer periods of time


686


(.20 years) as a result of occasional cell division stimulated by persistent antigen, cross-reacting antigen, or cytokines. The long-term survival of memory cells is a fundamental property of the immune system, essential for protection against previously encountered pathogens. The ability of HIV-1 to establish a state of latent infection in resting memory CD4` T cells thus provides a potentially important mechanism for viral persistence. Although reactivation of virus from this latent reservoir normally contributes only a minute fraction of the plasma virus in untreated individuals, this reservoir assumes tremendous significance in patients who are on HAART and in whom productively infected cells have decayed to the point where plasma virus is no longer measurable. The notion that resting CD4` T cells might represent a barrier to HIV-1 eradication in patients on combination therapy was dramatically illustrated by three studies that examined whether replication-competent virus could persist in the resting CD4` T cells of patients on HAART (191–193). These studies focused on a subset of patients who responded extremely well to therapy and who had been aviremic for as long as 2.5 years. Although it is generally very difficult to isolate virus by conventional methods from patients on long-term HAART, all three groups found that with enhanced culture techniques, replication-competent virus could be readily isolated from resting CD4` lymphocytes of these patients. The culture methods used were similar in that, in each case, the patient’s T cells were subjected to conditions that would efficiently activate resting CD4` T cells and thereby allow them to express latent virus. In one study, a cross-sectional analysis suggested that the frequencies of latently infected CD4` T cells did not decrease during the first two years of therapy (191). This is in marked contrast to other viral reservoirs examined to date, all of which show readily measurable decay rates (Figure 1). The existence of a persistent latent reservoir for HIV-1 required a reevaluation of the eradication hypothesis. Included in the original model developed by Perelson and colleagues was the idea that the dynamics of productively infected CD4` T cells would be governed by the following equation: dT */dt 4 kVT 1 dT * in which productively infected cells (T*) are generated from uninfected T cells (T) with a rate constant k in a process that depends on the plasma virus concentration (V) (60). The cells are cleared with a rate constant d. Later models (1) incorporated the notion that productively infected T cells could also arise by activation of latently infected cells (L) with a rate constant, a: dT */dt 4 kVT ` aL 1 dT * At high viral loads, the contribution of the latent reservoir is small {kVT*..aL). However, if the half-life of latently infected cells is very long, then this pool could clearly thwart efforts to eradicate the virus even under conditions where the plasma virus level (V) is very low. Accurate measurement of the half-life of the latent reservoir required the development of assays to detect latently infected cells.


687


Assays for Latently Infected Cells Both molecular and virologic approaches have been used. Detection of latent integrated provirus by molecular methods presents two serious problems. First, conventional PCR assays for HIV-1 DNA do not distinguish between unintegrated HIV-1 DNA and the stable integrated form. Second, unless the assays are performed on rigorously purified resting CD4` T cells, there is no way to tell whether the HIV-1 genomes detected are latent. In the original studies defining a latent reservoir for HIV-1, an inverse PCR method was used to detect unambiguously the integrated form of HIV-1 DNA in highly purified resting CD4` T cells (56). This method amplified the junction between the LTR and upstream host genomic DNA. An Alu PCR method has also been used (193). Although these methods allow detection of cells with integrated HIV-1 DNA, they are cumbersome and can only be made quantitative with the use of a limiting dilution format. In addition, these methods suffer from the problem that they detect defective as well as replication competent HIV-1 DNA. An alternative approach is to attempt to culture virus from purified resting CD4` T cells by activating them with mitogenic stimuli (40, 191). In conventional virus culture assays (38, 194), the patient’s cells are cocultured with PHAactivated lymphoblasts from a normal donor, but no specific step is included to induce uniform activation of resting cells from the patient. Some investigators have suggested that the detection of latently infected cells might be facilitated by the inclusion of the mitogen PHA, which can activate resting T cells (39). Quantitative analysis of latently infected cells in vivo was made possible through the development of enhanced culture techniques in which purified resting CD4` T cells are first isolated and then subjected to conditions that induce activation with high efficiency (40). Addition of PHA alone to resting CD4` T cells is insufficient as PHA activation is dependent upon the presence of monocytes/macrophages. Thus, purified resting CD4` T cells from the patient are mixed with an excess of irradiated PBMC from a normal donor in the presence of PHA. In subsequent steps, CD4` lymphoblasts from normal donors are added to amplify any virus released from latently infected cells. CD8` T cells, which can suppress virus replication (195–197), are deleted from both the patient and donor lymphoblast populations. The activation conditions used in these assays actually allow expansion of CD4` T cells with high cloning efficiency (198), and the approach can be made quantitative through the use of a limiting dilution format. This enhanced virus culture assay has the advantage that it will only detect replication-competent forms of HIV-1. However, these assays can detect both recently infected cells in the preintegration state of latency and cells in the postintegration state of latency. There are several ways around this problem. First, the preintegration state of latency is labile. If viral replication is stopped or dramatically reduced with antiretroviral drugs, then the number of recently infected resting cells should fall, and culture assays done at long time intervals (.3 months) after initiating effective therapy may preferentially detect integrated virus

688


(J Blankson, D Persaud, RF Siliciano, unpublished data). The number of cells in the preintegration state of latency is correlated with viral load, and analysis of this correlation has shown that as the viral load falls below 20 copies/ml, the number of cells in the preintegration state of latency should represent only a minor fraction of the total amount of latent virus detected in the enhanced culture assays. For this reason, enhanced virus culture assays done on patients with undetectable plasma virus are likely to detect mainly cells in the postintegration state of latency. An additional problem with this assay is that when the frequency of latently infected cells is low, detection requires addition of large numbers of purified resting CD4` T cells while maintaining appropriate conditions of cell density (194) and ratios of irradiated donor PBMC and of CD4` lymphoblasts.


Half-Life of the Latent Reservoir for HIV-1 The detection of latently infected cells in patients on HAART raised the question of how long these cells would persist. Definitive measurement of the half-life of the latent reservoir in resting CD4` T cells required longitudinal studies in individual patients over long periods of time. The first such measurements have recently been reported. It now appears that decay of the latent reservoir cannot be expected in the average patient who is responding well to HAART. In a study of 35 patients who were being treated with HAART according to current guidelines (199) and whose plasma virus levels were undetectable by a standard RTPCR assay with a sensitivity down to 200 copies/ml, the latent reservoir was found to be extremely stable, with a mean half-life of over 43 months (200). Even with conservative estimates of the total body number of latently infected cells, an average of . 60 years of treatment would be required to eradicate this compartment. In fact, the mean slope was not statistically different than zero, indicating that there may not be decay in the average patient. This study used an enhanced virus culture method to detect latently infected cells, and the viruses obtained from this reservoir were fully replication competent in vitro and therefore likely to be capable of rekindling the infection in patients who stop therapy. The extremely slow decay rate of this reservoir raises the disturbing prospect that in some patients the time required for HIV-1 eradication with current combination regimens may be so great that other intervening problems such as cumulative toxicities of antiretroviral drugs (201) may make eradication difficult if not impossible. The persistence of the virus in latently infected T cells is at least one reason that, with extremely rare exceptions (202), rapid rebound of plasma viremia has been noted in all patients who have interrupted therapy (203). The slow decay rate measured in the study cited above is best thought of as the observed decay of the reservoir in the average patient who has responded well to current standard of care therapy. The true intrinsic decay rate of this reservoir may be faster if there is still ongoing viral replication leading to the entry of new cells into the reservoir (see below). Evidence that the intrinsic decay rate is faster has come from several sources. In one patient in the above study, the latent



689

reservoir declined with t1/2 of only 3 months. In addition, more rapid decay had been observed in a recent study of the latent reservoir in patients who have been started on HAART during primary HIV-1 infection (204). The latent reservoir is established early in primary infection, and latently infected cells can be detected in patients who have been started as early as 48 h after presentation with acute retroviral syndrome (191, 200, 205). Zhang et al have used enhanced culture assays to measure the decay of resting CD4` T cells carrying replicationcompetent virus in a small highly selected group of patients who were started on therapy during primary HIV-1 infection. The latent reservoir was seen to decay with a mean half-life of about 6 months (204). In further studies by this group in a larger cohort of patients, a slower mean decay rate has been observed (t1/2 . 12 months), and slow decay has been associated with the occurrence of intermittent low level viremia (206). Thus, there is agreement that in only a small subset of patients on current standard of care therapy is there measurable decay of the latent reservoir (200, 204, 206). Long-term persistence of HIV-1 is supported by a recent study in which integrated HIV DNA, circular HIV-1 DNA, and various forms of HIV-1 RNA in PBMC were all found after the first year of therapy to reach a very stable plateau with little additional decay (207). Further studies with longer follow-up will be needed to determine whether this reservoir actually undergoes any decay in a clinically meaningful time frame. Another approach to measuring the half-life of the latent reservoir is to examine the decay of HIV-1 DNA in cells in the peripheral blood. Perelson et al reported that HIV-1 DNA in PBMC decayed with a mean t1/2 of about 5 months (1). A t1/2 of 10 months was recently reported for a small series of patients who had responded well to antiretroviral therapy (208). It is important to keep in mind that standard PCR assays do not distinguish integrated and unintegrated HIV-1 DNA and that such measurements cannot be used to define the turnover of cells in the postintegration state of latency unless the conditions are such that cells in the preintegration state of latency no longer contribute significantly to the measurement.

Factors Contributing to the Stability of the Latent Reservoir The remarkable stability of the latent reservoir for HIV-1 can be explained in several ways. The observed stability is consistent with the fact that the reservoir is composed at least in part of memory T cells carrying integrated HIV-1 DNA (40). The biological function of memory T cells is to persist and provide protection against previously encountered microorganisms. The long-term survival of antigen-specific memory T cell is accepted for CD8` T cells (209) and is becoming increasingly well documented for CD4` T cells (210). As discussed above, the intermitotic half-life of memory T cells has been estimated to be in the range of 5–6 months (158, 159). The actual life-span of a memory cell and its progeny may be much longer, consistent with the nearly life-long memory associated with many infections. Recent studies in the murine system indicate that T cell memory


690


may be maintained by occasional proliferation of memory cells. It is possible that latently infected cells may undergo occasional proliferation driven by stimuli that do not fully turn on HIV-1 gene expression. In this situation, the virus would be able to persist by taking advantage of the normal homeostatic mechanisms that maintain immunologic memory. Another explanation for the stability of the reservoir has been recently put forward by Bucy. Noting the decline in CTL responses in patients on suppressive antiretroviral therapy, he suggests that the rate of clearance of cells activated from the latent reservoir is reduced in treated patients due to the decline of HIV-1specific immune responses (86). A third explanation for the stability of the latent reservoir is that the reservoir is being reseeded by a low level of ongoing viral replication in patients on combination therapy. As discussed below, several lines of evidence suggest that a low level of viral replication may be occurring in patients with undetectable viral loads. The issue of ongoing replication is an important one because it is directly related to strategies for therapy. If ongoing replication sustains the latent reservoir, then intensification of therapy may stop residual replication and reveal the intrinsic decay rate, allowing viral eradication if this decay rate is fast enough. If the reservoir is intrinsically stable, then strategies to flush out latently infected cells must be considered.

ONGOING VIRAL REPLICATION IN PATIENTS ON HAART Although the postintegration state of latency clearly represents one mechanism for the persistence of HIV-1, there is another interrelated mechanism that involves low levels of ongoing replication. The persistence of virus in treated individuals whose plasma HIV-1 RNA levels are below the limit of detection may reflect the fact that therapy is not suppressing all ongoing viral replication. Original models of viral dynamics in HIV-1 infection postulated that potent antiretroviral regimens would stop all new infection of susceptible cells types in vivo. However, several lines of evidence (reviewed below) suggest that the current regimens are not completely effective in preventing any additional cells from being infected. New cycles of infection may be occurring, but at a set point that is generally below the limit of detection (Figure 1). Ongoing viral replication might involve the generation of new infected cells via de novo infection from extracellular virus released from other infected cells. Each new round of infection is of course subject to inhibition by the protease inhibitors at the stage of virus maturation and by RT inhibitors at the stage of reverse transcription. If the drugs do not completely block new infections, then persistence could be explained by low level of replication as long as new cells are successfully infected at a rate that balances the rate at which infected cells die. In this model, the virus might persist even without the aid of latent reservoir (Figure 1). A more likely scenario is that occasional


691


activation of the latent reservoir feeds this ongoing replication and that conversely this ongoing replication contributes to the stability of the latent reservoir by generating new latently infected cells. Evidence for ongoing replication in patients on HAART comes from several sources: 1. Many patients who have plasma virus levels below that limit of detection of conventional assays (200–500 copies/ml) have occasional positive determinations (termed ‘‘blips’’) when more sensitive assays for viral RNA are used (206, 211). These ‘‘ultrasensitive’’ assays currently detect HIV-1 RNA in the plasma down to a level of 10–50 copies/ml. The source of these blips is unclear. 2. When patients are switched from three-drug regimens to simpler two-drug ‘‘maintenance’’ regimens, breakthrough viremia is observed in a significant fraction of patients (212–214). The failure of the maintenance therapy trials suggests that current three-drug regimens barely contain viral replication. 3. Several recent studies suggest that productively infected cells can be detected in individuals who are aviremic on HAART. Detection methods include in situ hybridization or RT-PCR assays for cells expressing HIV-1 RNA and immunohistochemical detection of cells expressing HIV-1 proteins (27, 207, 211, 215). In an elegant study by Hockett et al, cells expressing full-length viral RNA were quantitated in lymph node biopsy specimens. Quantitative analysis of the relationship between the number of RNA positive cells and plasma viremia suggested that in patients with a viral load below 50 copies/ml, there could be as many as 100,000 productively infected cells at any given time. Of course, the detection of RNA` cells does not in itself prove an ongoing viral replication. It is possible that these cells are cells reactivated from latency. However, 100,000 cells RNA` cells would represent a substantial fraction of the latent reservoir (estimated to be 106 – 107 cells), inconsistent with the slow turnover of the reservoir. A more plausible idea is that viruses released following the activation of latently infected resting CD4` T cells occasionally spread and infect other cells, fueling a low level of ongoing replication in patients on HAART. 4. The detection of recently infected cells suggests active replication. Chun et al (205) measured total and integrated HIV-1 DNA in resting CD4` T cells from patients on HAART. They concluded that the difference between the two measurements was due to the presence of recently infected cells with unintegrated HIV-1 DNA (193). Similarly, 2-LTR circles can be detected in PBMC from patients on HAART (207). These are formed from complete linear reverse transcripts that undergo an aberrant end-to-end ligation reaction instead of integration into the host chromosome. They appear to be unstable and may prove to be an important marker of recent infection. 5. A recent study suggests that in some aviremic patients, viruses with altered env sequences arise, consistent with ongoing replication (204). However, env

692



evolution was noted only in a subset of patients (2/8). Other patients on HAART show no env evolution, consistent with preservation of earlier viral sequences in the latent reservoir (204, 216). Interestingly, there is only limited evidence for evolution in the pol gene despite the stringent selection enforced by antiretroviral drug regimens. Most of the viruses isolated from the latent reservoir show a drug sensitive genotype (191, 217, 218), and direct sequencing of PBMC DNA has shown no evidence for the evolution of drug resistance among patients who were aviremic on therapy (208). It is possible that in patients on HAART, the rate of replication is so low that the ordered accumulation of multiple pol mutations needed for high-level resistance rarely occurs unless the patients have persistent low-level viremia (192). Env mutations may accumulate as a result of the fact that changes in some regions of the env protein have little negative effect on viral replication. In summary, several lines of evidence suggest that new cells may become infected in patients on HAART whose plasma HIV-1 RNA measurements are below the limit of detection of current assays. A major unresolved issue is the relationship between the pool of latently infected cells and the low-level ongoing replication that is observed in patients on HAART. Are the productively infected cells in these individuals cells that have been activated from latency? It is likely that there is some communication between the pool of latently infected cells and the cells that are actively replicating, with the latent pool seeding the smoldering ongoing replication and the ongoing replication generating new latently infected cells (see Figure 1).

APPROACHES FOR CLEARING OR CONTAINING THE LATENT RESERVOIR Given that the latent reservoir for HIV-1 appears to represent a major barrier to virus eradication, several investigators have explored approaches for eliminating this reservoir. One approach involves intensification of combination therapy. It is possible that the stability of the latent reservoir reflects a low level of viral replication that ‘‘reseeds’’ the reservoir. It follows that intensification of antiretroviral therapy may completely inhibit viral replication and thus result in a quicker turnover of the reservoir (206). Given the long half-life of the reservoir and the toxicity associated with HAART, other approaches for the elimination of this reservoir are also being considered. One strategy involves the use of cytokines to activate latently infected cells and ‘‘flush out’’ the reservoir (12, 219). The goal is to promote viral transcription and replication in cells that harbor latent virus. These cells are then likely to die either from cytopathic effects of the virus or from recognition and lysis by virus-specific CTL. Virus produced from these cells should be unable to



693

infect other cells due to the presence of the antiretroviral agents. A recent study has shown that a combination of IL-2, IL-6, and TNF-a can induce activation and viral replication in latently infected resting cells in vitro (220). This same combination of cytokines was shown previously to be able to drive resting T cells to proliferate in the absence of antigen (221). IL-2 is one of the few T cell– activating agents approved for use in humans for the treatment of other diseases. Prior to the widespread availability of the protease inhibitors, several groups demonstrated that IL-2 could produce increases in CD4 counts (222–224). While the CD4 count increases observed in these studies were promising, no decreases in viral load were seen. In the era of HAART, a new approach involving the use of IL-2 to activate HIV-1 gene expression in latently infected cells has been explored by Fauci and colleagues (12, 225). Although latently infected cells do not express the high-affinity IL-2 receptor (40, 56), IL-2 may activate these cells through the low-affinity IL-2 receptor or indirectly by inducing the release of other cytokines. In a recent study, patients on HAART plus subcutaneous IL-2 were tested for the presence of latently infected cells. In a subset of the patients examined (3/14), virus could not be isolated from purified resting CD4` T cells, suggesting a very low frequency of latently infected cells (225). However, the occurrence of viral rebound observed following interruption of therapy in these patients suggests that levels of latently infected cells below the limit of detection of current assays can persist or that other forms of the virus may persist (12). Therefore, it remains unclear whether IL-2 will have clinical benefit in this situation. While this strategy is potentially promising, its in vivo application is limited by the fact that treatment with IL-2 causes the nonspecific activation of a large number of T cells resulting in significant toxicity. This would also hold true for treatment with the anti-CD3 antibody, although this reagent is sometimes used in the treatment of transplant rejection. The combination of a regimen of HAART with an agent that would specifically activate only those resting CD4` T cells that are infected would be ideal. However, no such agents are currently available. An alternative approach is to attempt to eliminate all memory T cells. This is based on the notion that latent virus will be found in the memory CD4` T cell compartment. In a provocative in vitro study, Vitetta and colleagues have shown that an anti-CD45RO immunotoxin can eliminate latently infected cells (226). In this case, the latently infected cells were generated by in vitro infection of PBMC from normal donors followed by an anti-CD25 treatment to kill activated cells. The latently infected cells under study were most likely cells in the preintegration state of latency. Impressive reductions in the amount of recoverable virus were achieved by treating the cells with an anti-CD45RO-toxin conjugate. The application of such an approach in vivo is complicated by the fact that it would kill activated and memory cells from both the CD4` and CD8` T cell subsets. In addition, a small portion of the latent virus may be present in naı¨ve CD45RO1 T cells (227). These latently infected cells may arise through phenotypic reversion of memory cell, infection of developing thymocytes, or through other unknown mechanisms.

694


Finally, it is possible that strategies to enhance immune responses to HIV-1 may allow immunologic control of the small amount of virus that emerges from the latent reservoir. The capacity of the immune system to control latent HIV-1 is illustrated by the fact that some long-term nonprogressors have no detectable plasma virus, possibly as a result of a strong immune response to HIV-1 (228). Recent studies suggest that latent virus is present in such individuals but is being controlled (202; J Blankson, R Siliciano, unpublished results).


SUMMARY Recent studies have identified an extraordinarily stable reservoir in which HIV-1 can potentially persist for life even in patients on effective antiretroviral therapy. This reservoir consists of a small pool of latently infected resting CD4` T cells. These cells carry an integrated copy of the HIV-1 genome and can produce infectious virus when the cells are activated by the appropriate antigen. The stability of the reservoir may reflect the fact that it is continuously reseeded by a low level of ongoing viral replication. Alternatively, it may reflect normal mechanisms that maintain the level of memory T cells relatively constant. In any event, elimination of this reservoir by novel therapeutic approaches will likely be required before eradication can be achieved. ACKNOWLEDGMENTS We thank Drs. Anthony Fauci, Tae-Wook Chun, Douglas Richman, Joseph Wong, David Ho, Bharat Ramratnam, Mario Stevenson, Pat Bucy, Zvi Grossman, and Ellen Vitetta for sharing unpublished data. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Perelson AS, Essunger P, Cao Y, Vesanen M , Hurley A, Saksela K, Markowitz M, Ho DD. 1997. Decay characteristics of HIV-1-infected compartments during combination therapy. Nature 387: 188–91 2. Gulick RM, Mellors JW, Havlir D, Eron JJ, Gonzalez C, McMahon D, Richman DD, Valentine FT, Jonas L, Meibohm A, Emini EA, Chodakewitz JA. 1997. Treatment with indinavir, zidovudine, and lamivudine in adults with human immu-

nodeficiency virus infection and prior antiretroviral therapy. N. Engl. J. Med. 337:734–39 3. Hammer SM, Squires KE, Hughes MD, Grimes JM, Demeter LM, Currier JS, Eron JJ Jr, Feinberg JE, Balfour HH Jr, Deyton LR, Chodakewitz JA, Fischl MA. 1997. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. AIDS


4.

5.

6.


7.

8.

9. 10.

11.

12.

13.

14. 15.

Clinical Trials Group 320 Study Team. N. Engl. J. Med. 337:725–33 Palella FJ, Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten GA, Aschman DJ, Holmberg SD. 1998. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N. Engl. J. Med. 338:853–60 Garcia-Blanco MA, Cullen BR. 1991. Molecular basis of latency in pathogenic human viruses. Science. 254:815–20 Havlir DV, Richman DD. 1996. Viral dynamics of HIV: implications for drug development and therapeutic strategies. Ann. Int. Med. 124:984–94 Stevenson M. 1997. Molecular mechanisms for the regulation of HIV replication, persistence and latency AIDS. 11(Suppl. A):S25–S33 Schrager LK, D’Souza PM. 1998. Cellular and anatomical reservoirs of HIV-1 in patients receiving potent antiretroviral combination therapy. JAMA 280:67–71 Pantaleo G, Perrin L. 1998. Can HIV be eradicated? AIDS 12:S175-S180 Finzi D, Siliciano RF. 1998. Viral dynamics in HIV-1 infection. Cell 93:665–71 Haase AT. 1999. Population biology of HIV-1 infection: viral and CD4` T cell demographics and dynamiocs in lymphatic tissues. Annu. Rev. Immunol. 17:625–56 Chun T-W, Fauci AS. 1999. Latent reservoirs of HIV: obstacles to the eradication of virus. Proc. Natl. Acad. Sci. USA. 96:10958–61 Berger EA, Murphy PM, Farber JM. 1999. Chemokine receptors as HIV-1 coreceptors: roles in viral entry, tropism and disease. Annu. Rev. Immunol. 17:657–700 Quinn TC. 1997. Acute primary HIV infection. JAMA 278:58–62 Daar ES, Moudgil T, Meyer RD, Ho DD. 1991. Transient high levels of viremia in patients with primary human immuno-

16.

17.

18.

19.

20.

21.

22.

695

deficiency virus type 1 infection. N. Engl. J. Med. 324:961–64 Veazey RS, Demaria MA, Chalifoux LV, Shvetz DE, Pauley DR, Knight HL, Rosenzweig M, Johnson RP, Desrosiers RC, Lackner AA. 1998. Gastrointestinal tract as a major site of CD4` T cell depletion and viral replication in SIV infection. Science 280:427–31 Harouse JM, Gettie A, Tan RCH, Blanchard J, Cheng-Mayer C. 1999. Distinct pathogenic sequela in rhesus macaques infected with CCR5 or CXCR4 utilizing SHIVs. Science 284:816–19 Borrow P, Lewicki H, Hahn BH, Shaw GM, Oldstone MB. 1994. Virus-specific CD8` cytotoxic T-lymphocyte activity associated with control of viremia in primary human immunodeficiency virus type 1 infection. J. Virol. 68:6103–10 Koup RA, Safrit JA, Cao Y, Andrews CA, McLeod G, Borkowsky W, Farthing C, Ho DD. 1994. Temporal association of cellular immune responses with the initial control of viremia in primary human immunodeficiency virus type 1 syndrome. J. Virol. 68:4650–55 Pantaleo G, Demarest JF, Soudeyns H, Graziosi C, Denis F, Adelsberger JW, Borrow P, Saag MS, Shaw GM, Sekaly RP, Fauci AS. 1994. Major expansion of CD8` T cells with a predominant Vb during the primary immune response to HIV. Nature. 370:463–67 Schmitz JE, Kuroda MJ, Santra S, Sasseville VG, Simon MA, Lifton MA, Racz P, Tenner-Racz K, Dalesandro M, Scallon BJ, Ghrayeb J, Forman MA, Montefiori DC, Rieber EP, Letvin NL, Reimann KA. 1999. Control of viremia in simian immunodeficiency virus infection by CD8` lymphocytes. Science 283:857–60 Piatak M Jr, Saag MS, Yang LC, Clark SJ, Kappes JC, Luk KC, Hahn BH, Shaw GM, Lifson JD. 1993. High levels of HIV-1 in plasma during all stages of

696

23.

24.


25.

26.

27.

28.

29.

30.

PIERSON n MCARTHUR n SILICIANO infection determined by competitive PCR. Science 259:1749–54 Pantaleo G, Graziosi C, Butini L, Pizzo PA, Schnittman SM, Kotler DP, Fauci AS. 1991. Lymphoid organs function as major reservoirs for human immunodeficiency virus. Proc. Natl. Acad. Sci. USA 88:9838–42 Embretson J, Zupancic M, Ribas JL, Burke A, Racz P, Tenner-Racz K, Haase AT. 1993. Massive covert infection of helper T lymphocytes and macrophages by HIV during the incubation period of AIDS. Nature 362:359–62 Pantaleo G, Graziosi C, Demarest JF, Butini L, Montroni M, Fox CH, Orenstein JM, Kotler DP, Fauci AS. 1993. HIV infection is active and progressive in lymphoid tissue during the clinically latent stage of disease. Nature 362:355– 58 Harris M, Patendaude P, Copperberg P, Filipenko D, Thorn A, Raboud Jeal. 1997. Correlation of virus load in plasma and lymph node tissue in human immunodeficiency virus infection. J. Infect. Dis. 176:1388–92 Hockett RD, Michael KJ, Derdeyn CA, Saag MS, Sillers M, Squires K, Chiz S, Nowak MA, Shaw GM, Bucy RP. 1999. Constant mean viral copy number per infected cell in tissues regardless of high, low, or undetectable plasma HIV RNA. J. Exp. Med. 189:1545–54 Mellors JW, Rinaldo CW Jr, Gupta P, White RM, Todd JA, Kingsley LA. 1996. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 272:1167–70 Hellerstein M, Hanley MB, Cesar D, Siler S, Papageorgopoulos C, Wieder E, Schmidt D, Hoh R, Neese R, Macallan D, Deeks S, McCune JM. 1999. Directly measured kinetics of circulating T lymphocytes in normal and HIV-1-infected humans Nat. Med. 5:83–89 Hellerstein MK, McCune JM. 1997. T

31.

32.

33.

34.

35.

36.

37.

38.

cell turnover in HIV-1 disease. Immunity 7:583–89 Connor RI, Mohri H, Cao Y, Ho DD. 1993. Increased viral burden and cytopathicity correlate temporally with CD4` T-lymphocyte decline and clinical progression in human immunodeficiency virus type 1-infected individuals. J. Virol. 67:1772–77 Reinhart TA, Rogan MJ, Huddleston D, Rausch DM, Eiden LE, Haase AT. 1997. Simian immunodeficiency virus burden in tissues and cellular compartments during clinical latency and AIDS. J. Infect. Dis. 176:1198–1208 Cheng-Mayer C, Seto D, Tateno M, Levy JA . 1988. Biological features of HIV-1 that correlate with virulence in the host. Science 240:80–82 Fenyo EM, Morfeldt-Manson L, Chiodi F, Lind B, von Gegerfelt A, Albert J, Olausson E, Asjo B. 1988. Distinctive replicative and cytopathic characterisitics of human immunodeficiency virus isolates. J. Virol. 62:4414–19 Tersmette M, Gruters RA, de Wolf F, de Goede RE, Lange JM, Schellekens PT, Goudsmit J, Huisman HG, Miedema F. 1989. Evidence for a role of virulent human immunodeficiency virus (HIV) variants in the pathogenesis of acquired immunodeficiency syndrome: studies on sequential HIV isolates. J. Virol. 63:2118–25 Kimata JT, Kuller L, Anderson DB, Dailey P, Overbaugh J. 1999. Emerging cytopathic and antigenic simian immunodeficiency virus variants influence AIDS progression. Nat. Med. 5:535–41 Neumann AU, Lam NP, Dahari H, Gretch DR, Wiley TE, Layden TJ, Perelson AS. 1998. Hepatitis C viral dynamics in vivo and the antiviral efficacy of interferon-alpha therapy. Science 282: 103–7 Ho DD, Moudgil T, Alam M. 1989. Quantitation of human immunodeficiency virus type 1 in the blood of


39.

40.


41.

42.

43.

44.

45.

46.

infected persons. N. Engl. J. Med. 321:1621–25 Chesebro B, Wehrly K, Metcalf J, Griffin DE. 1991. Use of a new CD4-positive HeLa cell clone for direct quantitation of infectious human immunodeficiency virus from blood cells of AIDS patients. J. Infect. Dis. 163:64–70 Chun T-W, Carruth L, Finzi D, Shen X, Digiuseppe JA, Taylor H, Hermankova M, Chadwick K, Margolick J, Quinn TC, Kuo Y-H, Brookmeyer R, Zeiger MA, Barditch-Crovo P, Siliciano RF. 1997. Quantitation of latent tissue reservoirs and total body load in HIV-1 infection. Nature 387:183–88 Schnittman SM, Psallidopoulos MC, Lane HC, Thompson L, Baseler M, Massari F, Fox CH, Salzman NP, Fauci AS. 1989. The reservoir for HIV-1 in human peripheral blood is a T cell that maintains expression of CD4. Science 245:305–8 Psallidopoulus MC, Schnittman SM, Thompson LM, Baseler M, Fauci AS, Lane HC, Salzman NP. 1989. Integrated proviral human immunodeficiency virus type 1 is present in CD4` peripheral blood lymphocytes in healthy seropositive individuals J. Virol. 63:4626–31 Simmonds P, Balfe P, Peutherer JF, Ludlam CA, Bishop JO, Brown AJL. 1990. Human immunodeficiency virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol. 64:864–72 Bukrinsky MI, Stanwick TL, Dempsey MP, Stevenson M. 1991. Quiescent T lymphocytes as an inducible virus reservoir in HIV-1 infection. Science 254:423–27 Bagasra O, Hauptman SP, Lischner HW, Sachs M, Pomerantz RJ. 1992. Detection of human immunodeficiency virus type 1 provirus in mononuclear cells by in situ polymerase chain reaction. N. Engl. J. Med. 326:1385–91 Bush CE, Donovan RM, Smereck SM,

47.

48.

49.

50.

51.

52.

53.

697

Strang D, Markowitz N, Saravolatz LD. 1993. Quantitation of unintegrated HIV1 DNA in asymptomatic patients in the presence or absence of antiretroviral therapy. AIDS Res. Hum. Retroviruses 9:183–87 Patterson BK, Till M, Otto P, Goolsby C, Furtado MR, McBride LJ, Wolinsky SM. 1993. Detection of HIV-1 DNA and messenger RNA in individual cells by PCRdriven in situ hybridization and flow cytometry. Science 260:976–79 Saksela K, Muchmore E, Girard M, Fultz P , Baltimore D. 1993. High viral load in lymph nodes and latent human immunodeficiency virus (HIV) in peripheral blood cells of HIV-1-infected chimpanzees J. Virol. 67:7423–27 Pang S, Koyanagi Y, Miles S, Wiley C, Vinters HV, Chen ISY. 1990. High levels of unintegrated HIV-1 DNA in brain tissue of AIDS dementia patients. Nature 343:85–89 Harper ME, Marselle LM, Gallo RC, Wong-Staal F. 1986. Detection of lymphocytes expressing human T-lymphotropic virus type III in lymph nodes and periphernal blood from infected individuals by in situ hybridization. Proc. Natl. Acad. Sci. USA 83:772–76 Michael NL, Vahey M, Burke DS, Redfield RR. 1992. Viral DNA and mRNA expression correlate with the stage of human immunodeficiency virus (HIV) type 1 infection in humans: evidence for viral replication in all stages of HIV disease. J. Virol. 66:310–16 Seshamma T, Bagasra O, Trono D, Baltimore D, Pomerantz RJ. 1992. Blocked early-stage latency in the peripheral blood cells of certain individuals infected with human immunodeficiency virus type 1. Proc. Natl. Acad. Sci. USA 89:10663–67 Adams M, Sharmeen L, Kimpton J, Romeo JM , Garcia JV, Peterlin BM, Groudine M, Emerman M. 1994. Cellular latency in human immunodeficiency

698

54.


55.

56.

57.

58.

59.

60.

61.

PIERSON n MCARTHUR n SILICIANO virus-infected individuals with high CD4 levels can be detected by the presence of promoter-proximal transcripts. Proc. Natl. Acad. Sci. USA 91:3862–66 LaFemina RL, Schneider CL, Robbins HL, Callahan PL, LeGrow K, Roth E, Schleif WA, Emini EA. 1992. Requirement of active human immunodeficiency virus type 1 integrase enzyme for productive infection of human T-lymphoid cells J. Virol. 66:7414–19 Englund G, Theodore TS, Freed EO, Engleman A, Martin MA. 1995. Integration is required for productive infection of monocyte- derived macrophages by human immunodeficiency virus type 1. J. Virol. 69:3216–19 Chun T-W, Finzi D, Margolick J, Chadwich K, Schwartz D, Siliciano RF. 1995. Fate of HIV-1-infected T cells in vivo: Rates of transition to stable latency. Nat. Med. 1:1284–90 Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M. 1995. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373:123–26 Wei X, Ghosh SK, Taylor ME, Johnson VA , Emini EA, Deutsch P, Lifson JD, Bonhoeffer S, Nowak MA, Hahn BH, Shaw GM. 1995. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373:117–22 Coffin JM. 1995. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 267:483–89 Perelson AS, Neumann AU, Markowitz M, Leonard JM, Ho DD. 1996. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, viral generation time. Science 271:1582–86 Cavert W, Notermans DW, Staskus K, Wietgrefe SW, Zupancic M, Gebhard K, Henry K, Zhang S-C, Mills R, McDade H, Goudsmit J, Danner SA, Haase AT. 1997. Kinetics of response in lymphoid

62.

63.

64.

65.

66.

67.

68.

69.

70.

tissue to antiretroviral therapy of HIV-1 infection. Science 276:960–64 Zhang LQ, Dailey PJ, He T, Gettie A, Bonhoeffer S, Perelson AS, Ho DD. 1999. Rapid clearance of simian immunodeficiency virus particles from plasma of rhesus macaques. J. Virol. 73:855–60 Koga Y, Sasaki M, Yoshida H, Oh-Tsu M, Kimura G, Nomoto K. 1991. Disturbance of nuclear transport of proteins in CD4` cells expressing gp160 of human immunodeficiency virus. J. Virol. 65: 5609–12 Laurent-Crawford AG, Krust B, Muller S, Riviere Y, Rey-Cuille MA, Bechet JM, Montagnier L, Hovanessian AG. 1991. The cytopathic effect of HIV is associated with apoptosis. Virology 185:829– 39 Terai C, Kornbluth RS, Pauza CD, Richman DD, Carson DA. 1991. Apoptosis as a mechanism of cell death in cultured T lymphoblasts acutely infected with HIV1. J. Clin. Invest. 87:1710–15 Todd BJ, Kedar P, Pope JH. 1995. Syncytium induction in primary CD4` T-cell lines from normal donors by human immunodeficiency virus type 1 isolates with non-syncytium-inducing genotype and phenotype in MT-2 cells. J. Virol. 69:7099–7105 Cao J, Park IW, Cooper A, Sodroski J. 1996. Molecular determinants of acute single-cell lysis by human immunodeficiency virus type 1. J. Virol. 70:1340–54 Stewart SA, Poon B, Jowett JB, Chen IS. 1997. Human immunodeficiency virus type 1 Vpr induces apoptosis following cell cycle arrest. J. Virol. 71:5579–92 Yao XJ, Mouland AJ, Subbramanian RA, Forget J, Rougeau N, Bergeron D, Cohen EA. 1998. Vpr stimulates viral expression and induces cell killing in human immunodeficiency virus type 1-infected dividing Jurkat T cells. J. Virol. 72:4686– 93 Rogel ME, Wu LI, Emerman M. 1995. The human immunodeficiency virus type


71.

72.


73.

74.

75.

76.

77.

78.

79.

1 vpr gene prevents cell proliferation during chronic infection. J. Virol. 69:882–88 Jowett JB, Planelles V, Poon B, Shah NP, Chen ML, Chen IS. 1995. The human immunodeficiency virus type 1 vpr gene arrests infected T cells in the G2 ` M phase of the cell cycle. J. Virol. 69:6304– 13 Bartz SR, Rogel ME, Emerman M. 1996. Human immunodeficiency virus type 1 cell cycle control: Vpr is cytostatic and mediates G2 accumulation by a mechanism which differs from DNA damage checkpoint control. J. Virol. 70:2324–31 Goh WC, Rogel ME, Kinsey CM, Michael SF , Fultz PN, Nowak MA, Hahn BH, Emerman M. 1998. HIV-1 Vpr increases viral expression by manipulation of the cell cycle: a mechanism for selection of Vpr in vivo. Nat. Med. 4:65– 71 Poon B, Grovit-Ferbas K, Stewart SA, Chen IY. 1998. Cell cycle arrest by Vpr in HIV-1 virions and insensitivity to antiretroviral agents. Science 281:266–69 Gummuluru S, Emerman M. 1999. Cellcycle and vpr-mediated regulation of human immunodeficiency virus type 1 expression in primary and transformed Tcell lines. J. Virol. 73:5422–30 Siliciano RF. 1995. The role of CD4 in HIV envelope-mediated pathogenesis Curr. Top. Microbiol. Immunol. 205: 159–79 Walker BD, Chakrabarti S, Moss B, Paradis TJ, Flynn T, Durno AG, Blumberg RS, Kaplan JC, Hirsch MS, Schooley RT. 1987. HIV-specific cytotoxic T lymphocytes in seropositive individuals. Nature 328:345–48 Plata F, Autran B, Martins LP, WainHobson S, Raphael M, Mayaud C, Denis M, Guillon JM, Debre P. 1987. AIDS virus-specific cytotoxic T lymphocytes in lung disorders. Nature 328:348–51 Klenerman P, Phillips RE, Rinaldo CR, Wahl LM, Ogg G, May RM, McMichael

80.

81.

82.

83.

84.

85.

86.

87.

88.

699

AJ, Nowak MA. 1996. Cytotoxic T lymphocytes and viral turnover in HIV type 1 infection. Proc. Natl. Acad. Sci. USA 93:15323–28 Yang OO, Kalams SA, Rosenzweig M, Tocha A, Jones N, Koziel M, Walker BD, Johnson RP. 1996. Efficient lysis of human immunodeficiency virus type 1infected cells by cytotoxic T lymphocytes. J. Virol. 70:5799–7806 Johnson RP, Walker BD. 1994. Cytotoxic T lymphocytes in human immunodeficiency virus infection: responses to structural proteins. Curr. Top. Microbiol. Immunol. 189:35–63 Johnson RP, Siliciano RF, McElrath MJ. 1998. Cellular immune responses to HIV-1. [Review] [91 refs] AIDS 1998. 12 Suppl A:S113-S120 Jin X, Bauer DE, Tuttleton SE, Lewin S, Gettie A, Blanchard J, Irwin CE, Safrit JT, Mittler J, Weinberger L, Kostrikis LG, Zhang L, Perelson AS, Ho DD. 1999. Dramatic rise in plasma viremia after CD8(`) T cell depletion in simian immunodeficiency virus-infected macaques. J. Exp. Med. 189:991–98 Altman JD, Moss PAH, Goulder PJR, Barouch DH, McHeyzer-Williams MG, Bell JI, McMichael AJ, Davis MM. 1996. Phenotypic analysis of antigenspecific T lymphocytes. Science 274:94– 96 Collins KL, Chen BK, Kalams SA, Walker BD, Baltimore D. 1998. HIV-1 Nef protein protects infected primary cells against killing by cytotoxic T lymphocytes. Nature 391:397–401 Bucy RP. 1999. Immune clearance of HIV type 1 replication active cells: a model of two patterns of steady state HIV infection. AIDS Res. Hum. Retroviruses 15:223–27 Perelson AS, Essunger P, Ho DD. 1998. Dynamics of HIV-1 and CD4` lymphocytes in vivo AIDS. 11 (suppl A):S17– S34 Banda NK, Bernier J, Kurahara DK, Kur-

700

89.


90.

91.

92.

93.

94.

95.

96.

PIERSON n MCARTHUR n SILICIANO rie R, Haigwood N, Sekaly R-P, Finkel TH. 1992. Cross-linking CD4 by human immunodeficiency virus gp120 primes T cells for activation-induced apoptosis J. Exp. Med. 176:1099–1106 Finkel TH, Tudor-Williams G, Banda NK, Cotton MF, Curiel T, Monks C, Baba TW, Ruprecht RM, Kupfer A. 1995. Apoptosis occurs predominantly in bystander cells and not in productively infected cells of HIV- and SIV-infected lymph nodes Nat. Med. 1:129–34 Su L, Kaneshima H, Bonyhadi M, Salimi S, Kraft D, Rabin L, McCune JM. 1995. HIV-1-induced thymocyte depletion is associated with indirect cytopathicity and infection of progenitor cells in vivo. Immunity 2:25–36 Rabin RL, Roederer M, Maldonado Y, Petru A, Herzenberg LA. 1995. Altered representation of naive and memory CD8 T cell subsets in HIV-infected children. J. Clin. Invest. 95:2054–60 Roederer M, Dubs JG, Anderson MT, Raju PA, Herzenberg LA. 1995. CD8 naive T cell counts decrease progressively in HIV-infected adults. J. Clin. Invest. 95:2061–66 Douek DC, McFarland RD, Keiser PH, Gage EA, Massey JM, Haynes BF, Polis MA, Haase AT, Feinberg MA, Sullivan JL, Jamieson BD, Zack JA, Picker LJ, Koup RA. 1998. Changes in thymic function with age and during the treatment of HIV infection. Nature 396:690– 95 Hellerstein MK, McCune JM. 1997. T cell turnover in HIV-1 disease. Immunity 7:583–89 Luzuriaga K, Wu H, McManus M, Britto P, Borkowsky W, Burchett S, Smith B, Mofenson L, Sullivan JL. 1999. Dynamics of human immunodeficiency virus type 1 replication in vertically infected infants. J. Virol. 73:362–67 Ho DD, Rota TR, Hirsh MS. 1986. Infection of monocyte/macrophages by

97.

98.

99.

100.

101.

102.

103.

104.

105.

human T lymphotropic virus type III. J. Clin. Invest. 77:1712–20 Nicolson JKA, Gross GD, Callaway CS, McDougal JS. 1986. In vitro infection of human monocytes with T lymphotrophic virus type III/lymphadenopathy-associated virus (HTLV-III/LAV). J. Immunol. 137:323–29 Van Furth R. 1989. Origin and turnover of monocytes and macrophages. Curr. Top. Pathol. 79:125–50 Sierra-Madero JC, Toossi Z, Hom DL, Finegan CK, Hoenig E, Rich EA. 1994. Relationship between load of virus in alveolar macrophages from human immunodeficiency virus type 1-infected persons, production of cytokines, and clinical status. J. Infect. Dis. 169:18–27 McIlroy D, Autran B, Cheynier R, Clauvel JP, Oksenhendler E, Debre P, Hosmalin A. 1996. Low infection frequency of macrophages in the spleens of HIV` patients. Res. Virol. 147:115–21 Fox CH, Tenner-Racz K, Racz P, Firpo A, Pizzo PA, Fauci AS. 1991. Lymphoid germinal centers are reservoirs of human immunodeficiency virus type 1 RNA. J. Infect. Dis. 164:1051–57 Yasutomi Y, Reimann KA, Lord CI, Miller MD, Letvin NL. 1993. Simian immunodeficiency virus-specific CD8` lymphocyte response in acutely infected rhesus monkeys. J. Virol. 67:1707–11 Burton GF, Masuda A, Heath SL, Smith BA , Tew JG, Szakal AK. 1997. Follicular dendritic cells (FDC) in retroviral infection: host/pathogen perspectives. Immunol. Rev. 156:185–97 Grouard G, Clark EA. 1997. Role of dendritic and follicular dendritic cells in HIV infection and pathogenesis. Curr. Opin. Immunol. 9:563–67 Reinhart TA, Rogan MJ, Viglianti GA, Rausch DM, Eiden LE, Haase AT. 1997. A new approach to investigating the relationship between productive infection and cytopathicity in vivo. Nat. Med. 3:218–21


RESERVOIRS FOR HIV-1 106. Heath SL, Tew JG, Szakal AK, Burton GF. 1995. Follicular dendritic cells and human immunodeficiency virus infectivity. Nature 377:740–44 107. Grossman Z, Feinberg M, Kuznetsov V, Dimitrov D, Paul W. 1998. HIV infection: How effective is drug combination treatment? Immunol. Today 19:528–32 108. Morris L, Binley JM, Clas BA, Bonhoeffer S, Astill TP, Kost R, Hurley A, Cao Y, Markowitz M, Ho DD, Moore JP. 1998. HIV-1 antigen-specific and -nonspecific B cell responses are sensitive to combination antiretroviral therapy. J. Exp. Med. 188:233–45 109. Markowitz M, Vesanen M, Tenner-Racz K, Cao Y, Binley JM, Talal A, Hurley A, Chaudhry MR, Yaman M, Frankel S, Heath-Chiozzi M, Leonard JM, Moore JP, Racz P, Nixon DF, Ho DD, Ji X. 1999. The effect of commencing combination antiretroviral therapy soon after human immunodeficiency virus type 1 infection on viral replication and antiviral immune responses. J. Infect. Dis. 179:527–37 110. Zhang H, Dornadula G, Beumont M, Livornese L, Van Uitert B, Henning K, Pomerantz RJ. 1998. HIV-1 in the semen of men receiving highly active anti-retroviral therapy N. Engl. J. Med. 339:1803–0 111. Quayle AJ, Xu C, Mayer KH, Anderson DJ . 1997. T lymphocytes and macrophages, but not motile spermatozoa, are a significant source of human immunodeficiency virus in semen. J. Infect. Dis. 176:960–68 112. Zhu T, Wang N, Carr A, Nam DS, MoorJankowski R, Cooper DA, Ho DD. 1996. Genetic characterization of human immunodeficiency virus type 1 in blood and genital secretions: evidence for viral compartmentalization and selection during sexual transmission. J. Virol. 70:3098–3107 113. Glass JD, Johnson RT. 1996. Human immunodeficiency virus and the brain. Annu. Rev. Neurosci. 19:1–26

701

114. Bacellar H, Munoz A, Miller EN, Cohen BA, Besley D, Selnes OA, Becker JT, McArthur JC. 1994. Temporal trends in the incidence of HIV-1 related neurologic diseases: Multicenter AIDS Cohort Study, 1985–1992. Neurology 44:1892– 1900 115. Johnson RT, McArthur JC, Narayan O. 1988. The neurobiology of human immunodeficiency virus infections. FASEB J. 2:2970–81 116. Conant K, Garzinodemo A, Nath A, McArthur JC, Halliday W, Power C, Gallo RC, Major EO. 1998. Induction of monocyte chemoattractant protein-1 in HIV-1 tat-stimulated astrocytes and elevation in AIDS dementia. Proc. Natl. Acad. Sci. USA 95:3117–21 117. Persidsky Y, Stins M, Way D, Witte MH, Weinand M, Kim KS, Bock P, Gendelman HE, Fiala M. 1997. A model for monocyte migration through the bloodbrain barrier during HIV-1 encephalitis. J. Immunol. 158:3499–3510 118. Takahashi K, Wesselingh SL, Griffin DE, McArthur JC, Johnson RT, Glass JD. 1996. Localization of HIV-1 in human brain using polymerase chain reaction/in situ hybridization and immunocytochemistry. Ann. Neurol. 39:705–11 119. McArthur JC, McClernon DR, Cronin MF, Nance-Sproson TE, Saah AJ, StClair M, Lanier ER. 1997. Relationship between human immunodeficiency virus-associated dementia and viral load in the cerebrospinal fluid and brain. Ann. Neurol. 42:689–98 120. Zink MC, Spelman JP, Robinson RB, Clements JE. 1998. SIV infection of macaques–modeling the progression to AIDS dementia. J. Neurovirol. 4:249–59 121. Cheng-Mayer C, Weiss C, Seto D, Levy JA . 1989. Isolates of human immunodeficiency virus type 1 from the brain may constitute a special group of the AIDS virus. Proc. Natl. Acad. Sci. USA 86:8575–79 122. Gartner S, Markovits P, Markovitz DM,

702

123.


124.

125.

126.

127.

128.

129.

PIERSON n MCARTHUR n SILICIANO Kaplan MH, Gallo RC, Popovic M. 1986. The role of mononuclear phagocytes in HTLV-III/LAV infection. Science 233:215–19 Chesebro B, Wehrly K, Nishio J, Perryman S. 1992. Macrophage-tropic human immunodeficiency virus isolates from different patients exhibit unusual V3 envelope sequence homogeneity in comparison with T-cell-tropic isolates: definition of critical amino acids involved in cell tropism. J. Virol. 66:6547–54 Sei S, Saito K, Stewart SK, Crowley JS, Brouwers P, Kleiner DE, Katz DA, Pizzo PA, Heyes MP. 1995. Increased human immunodeficiency virus (HIV) type 1 DNA content and quinolinic acid concentration in brain tissues from patients with HIV encephalopathy. J. Infect. Dis. 172:638–47 Ellis RJ, Hsia K, Spector SA, Nelson JA, Heaton RK, Wallace MR, Abramson I, Atkinson JH, Grant I, McCutchan JA, the HIV Neurobehavioral Research Center Group. 1997. Cerebrospinal fluid human immunodeficiency virus type 1 RNA levels are elevated in neurocognitively impaired individuals with acquired immunodeficiency syndrome Ann. Neurol. 42:679–88 Gisslen M, Norkrans G, Svennerholm B, Hagberg L. 1997. The effect of human immunodeficiency virus type 1 RNA levels in cerebrospinal fluid after initiation of zidovudine or didanosine. J. Infect. Dis. 175:434–37 Price RW, Staprans S. 1997. Measuring the ‘‘viral load’’ in cerebrospinal fluid in human immunodeficiency virus infection: window into brain infection? Ann. Neurol. 42:675–78 Groothuis DR, Levy RM. 1997. The entry into antiviral and antiretroviral drugs into the central nervous system. J. Neurovirol. 3:387–400 Kim RB, Fromm MF, Wandel C, Leake B, Wood AJJ, Roden DM, Wilkinson GR. 1998. The drug transporter P-gly-

130.

131.

132.

133.

134.

135.

136.

coprotein limits oral absorption and brain entry of HIV-1 protease inhibitors J. Clin. Invest. 101:289–94 Pialoux G, Fournier S, Moulignier A, Poveda JD, Clavel F, Dupont B. 1997. Central nervous system (CNS) as sanctuary of HIV-1 in a patient treated with AZT`3TCìndinavir (abstract 233) 4th Conference on Retroviruses and Opportunistic Infections. January 22–26, Washington, DC Brodt HR, Kamps BS, Gute P, Knupp B, Staszewski S, Helm EB. 1997. Changing incidence of AIDS-defining illnesses in the era of antiretroviral combination therapy. AIDS 11:1731–38 Sacktor N, Lyles RH, McFarlane G, Skolasky RL, Kleeberger C, Becker JT, Cohen B, Wesch J, Miller EN, McArthur JC, The Multicenter AIDS Cohort Study. 1999. HIV-1-related neurological disease incidence changes in the era of highly active antiretroviral therapy (abstract S31. 001) Neurology 52 (suppl 2):A252– A253 Moore RD, Keruly JC, Bartlett JG, Chaisson RE. 1999. Differential effect of HAART on development of opportunistic illness and death in clinical practice (abstract 691) 6th Conf. on Retroviruses and Opportunistic Infections, Jan 31– Feb 4 Brew BJ, Brown SJ, Catalan J, Sacktor N , Halman M, Symonds WT, Romero C, CNAB 001 Study Team. 1998. Safety and efficacy of abacavir (ABC, 1592) in AIDS dementia complex (Study CNAB 3001) (abstract #32192) 12th World AIDS Conf., Geneva, June 28–July 3 Filippi CG, Sze G, Farber SJ, Shahmanesh M, Selwyn PA. 1998. Regression of HIV encephalopathy and basal ganglia signal intensity abnormality at MR imaging in patients with AIDS after the initiation of protease inhibitor therapy. Radiology 206:491–98 Sacktor NC, Lyles RH, Skolasky RL, Anderson DE, McArthur JC, McFarlane


137.


138.

139.

140.

141.

142.

G, Selnes OA, Becker JT, Cohen B, Wesch J, Miller EN, for the Multicenter AIDS Cohort Study (MACS). 1999. Combination antiretroviral therapy including protease inhibitors improves HIV-associated cognitive impairment. Neurology 52:1640–47 Letendre SL, Caparelli E, Ellis RJ. 1999. Levels of serum and cerebrospinal fluid (CSF) indinavir (IDV) and HIV RNA in HIV-infected individuals (abstract 407) 6th Conf. on Retroviruses and Opportunistic Infections, Jan 31–Feb 4 Collier A, Marra C, Coombs RW. 1997. Cerebrospinal fluid (CSF) HIV RNA levels in patients on chronic indinavir therapy. (Abst. 22) Abstracts of the Infectious Disease Soc. of Am., 35th annu. meet., San Francisco. Sept 13–16 Marra CM, Coombs RW, Collier AC. 1999. Changes in CSF and plasma HIV1 RNA and in neuropsychological test performance after starting HAART (abstract 408) 6th Conf. on Retroviruses and Opportunistic Infections, Jan 31– Feb 4 Sension M, Farthing C, Siemon-Hryczyk P, Pilson R, Fischer L. 1997. Saquinavir soft gel capsule (SGC) in combination with AZT and 3TC in treatment naive patients (abstract I-190) Intersci. Conf. on Antimicrobial Agents and Chemotherapy (ICAAC), Toronto. Sept 28–Oct. 1:278–78 Gisolf E, Jurrians S, Van Der enda M. 1999. The Prometheus Study: double protease inhibitor (PI) treatment only is unable to suppress detectable viral load in the CSF (abstract 403). 6th Conf. on Retroviruses and Opportunistic Infections, Jan 31–Feb 4 Staprans S, Inkina N, NovakovicAgopian T, Deeks S, Heyes M, Aweeka F, Grant R, Price RW. 1998. Kinetics of CSF HIV-1 response to antiviral therapy. (Abstr. 454) 5th Conf. on Retroviruses and Opportunistic Infections, Feb 1–5, Chicago. 164

703

143. Ellis RJ, Spector SA, Hsia K, Deutsch R, Gamst A, Grant I, McCutchan JA, The HNRC Group. 1998. HIV-1 viral dynamics differ in cerebrospinal fluid and plasma. (Abstr. 453) 5th Conf. on Retroviruses and Opportunistic Infections, Feb 1–5, Chicago. 164 144. Polis M, Yoder C, Mican J, Schrager L, Metcalf J, Dewar R, Feinberg M, Piscitelli S. 1999. More than 2 months of an aggressive 4-drug antiretroviral regimen is required to suppress CSF HIV viral burden in previously antiretroviral naive patients. (Abstr. 404) 6th Conf. on Retroviruses and Opportunistic Infections, Jan 31–Feb 4 145. Crabtree GR. 1989. Contingent genetic regulatory events in T lymphocyte activation. Science 243:355–61 146. Lenardo M, Chan FKM, Horung F, McFarland H, Siegel R, Wang J, Zheng L. 1999. Mature T lymphocyte apoptosis–immune regulation in a dynamic and unpredictable antigenic environment. Annu. Rev. Immunol. 17:221–55 147. Kuo CT, Veselitis ML, Leiden JM. 1997. LKLF: A trancriptional regulator of single-positive T cell quiescence and survival. Science 277:1986–90 148. Boise LH, Minn AJ, June CH, Lindsten T, Thompson CB. 1995. Growth factors can enhance lymphocyte survival without committing the cell to undergo cell division. Proc. Natl. Acad. Sci. USA 92:5491–95 149. Marrack P, Mitchell T, Bender J, Hildeman D, Kedl R, Teague K, Kappler J. 1998. T-cell survival. Immunol. Rev. 165:279–85 150. Opferman JT, Ober BT, Ashton-Rickardt PG. 1999. Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 283:1745–48 151. Mackay CR. 1992. Migration pathways and immunologic memory among T lymphocytes. Semin. Immunol. 4:51–58 152. Young AJ, Hay JB, Mackay CR. 1993. Lymphocyte recirculation and life span

704

153.

154.


155.

156.

157.

158.

159.

160.

161.

PIERSON n MCARTHUR n SILICIANO in vivo. Curr. Top. Microbiol. Immunol. 184:161–73 Akbar AN, Terry L, Timms A, Beverley PCL, Janossy G. 1988. Loss of CD45R and gain of UCHL1 reactivity is a feature of primed T cells J. Immunol. 140:2171– 78 Sanders ME, Makgoba MW, Sharrow SO, et al. 1988. Human memory T lymphocytes express increased levels of three cell adhesion molecules (LFA-3, CD2, and LFA-1) and three other molecules (UCHL1, CDw29, and Pgp-1) and have enhanced IFN-gamma production. J. Immunol. 140:1401–7 Birkland ML, Johnson P, Trowbridge IS, Pure E. 1989. Changes in CD45 isoform expression accompany antigen-induced murine T-cell activation. Proc. Natl. Acad. Sci. USA 86:6734–38 Lee WT, Yin XM, Vitetta ES. 1990. Functional and ontogenetic analysis of murine CD45Rhi and CD45Rlo CD4` T cells. J. Immunol. 144:3288–95 Picker LJ, Treer JR, Ferguson DB, Collins PA, Buck D, Terstappen LW. 1993. Control of lymphocyte recirculation in man. I. Differential regulation of the peripheral lymph node homing receptor L-selectin on T cells during the virgin to memory cell transition. J. Immunol. 150:1105–21 Michie CA, McLean A, Alcock C, Beverley PC. 1992. Lifespan of human lymphocyte subsets defined by CD45 isoforms. Nature 360:264–65 Mclean AR, Michie CA. 1995. In vivo estimates of division and death rates of human T lymphocytes. Proc. Natl. Acad. Sci. USA 92:3707–11 Bunce C, Bell EB. 1997. CD45RC isoforms define two types of CD4 memory T cells, one of which depends on persisting antigen. J. Exp. Med. 185:767–76 Zack JA, Arrigo SJ, Weitsman SR, Go AS, Haislip A, Chen ISY. 1990. HIV-1 entry into quiescent primary lymphocytes: molecular analysis reveals a labile,

latent viral structure. Cell 61:213–22 162. Zack JA, Haislip AM, Krogstand P, Chen ISY. 1992. Incompletely reverse-transcribed human immunodeficiency virus type I genomes function as intermediates in the retroviral life cycle. J. Virol. 66:1717–25 163. Korin YD, Zack JA. 1999. Nonproductive human immunodeficiency virus type 1 infection in nucleoside-treated G0 lymphocytes. J. Virol. 73:6526–32 164. Spina CA, Guatelli JC, Richman DD. 1995. Establishment of a stable, inducible form of human immunodeficiency virus type 1 DNA in quiescent CD4 lymphocytes in vitro. J. Virol. 69:2977–88 165. Bukrinsky MI, Sharova N, Dempsey MP, Stanwick TL, Bukrinskaya AG, Haggerty S, Stevenson M. 1992. Active nuclear import of human immunodeficiency virus type 1 preintegration complexes. Proc. Natl. Acad. Sci. USA 89:6580–84 166. Kinoshita S, Chen BK, Kaneshima H, Nolan GP. 1998. Host control of HIV-1 parasitism in T cells by the nuclear factor of activated T cells. Cell 95:595–604 167. Folks TM, Clouse KA, Justement J, Rabson A, Duh E, Kehrl JH, Fauci AS. 1989. Tumor necrosis factor alpha induces expression of human immunodeficiency virus in a chronically infected T-cell clone. Proc. Natl. Acad. Sci. USA 86:2365–68 168. Pomerantz RJ, Trono D, Feinberg MB, Baltimore D. 1990. Cells nonproductively infected with HIV-1 exhibit an aberrant pattern of viral RNA expression: a molecular model for latency. Cell 61:1271–76 169. Pomerantz RJ, Seshamma T, Trono D. 1992. Efficient replication of human immunodeficiency virus type 1 requires a threshold level of Rev: potential implications for latency. J. Virol. 66:1809–13 170. Unutmaz D, KewalRamani VN, Marmon S, Littman DR. 1999. Cytokine signals are sufficient for HIV-1 infection of rest-


171.

172.


173.

174.

175.

176.

177.

178.

179.

180.

ing human T lymphocytes. J. Exp. Med. 189:1735–46 Folks T, Powell DM, Lightfoote MM, Benn S, Martin MA, Fauci AS. 1986. Induction of HTLV-III/LAV from a nonvirus-producing T-cell line: implications for latency. Science 231:600–2 Winslow BJ, Pomerantz RJ, Bagasra O, Trono D. 1993. HIV-1 latency due to the site of proviral integration. Virology. 196:849–54 Duh EJ, Maury WJ, Folks TM, Fauci AS, Rabson AB. 1989. Tumor necrosis factor alpha activates human immunodeficiency virus type 1 through induction of nuclear factor binding to the NF-kappa B sites in the long terminal repeat. Proc. Natl. Acad. Sci. USA 86:5974–78 Nabel G, Baltimore D. 1987. An inducible transcription factor activates expression of human immunodeficiency virus in T cells. Nature 326:711–13 Siekevitz M, Josephs SF, Dukovich M, Peffer N, Wong-Staal F, Greene WC. 1987. Activation of the HIV-1 LTR by T cell mitogens and the transactivator protein of HTLV-I. Science 238:1575–78 Tong-Starksen SE, Luciw PA, Peterlin BM. 1987. Human immunodeficiency virus long terminal repeat responds to Tcell activation signals. Proc. Natl. Acad. Sci. USA 84:6845–49 Bohnlein E, Lowenthal JW, Siekevitz M, Ballard DW, Franza BR, Greene WC. 1988. The same inducible nuclear proteins regulates mitogen activation of both the interleukin-2 receptor-alpha gene and type 1 HIV. Cell 53:827–36 Greene WC. 1990. Regulation of HIV-1 gene expression. [Review] Annu. Rev. Immunol. 8:453–75 Gaynor R. 1992. Cellular transcription factors involved in the regulation of HIV1 gene expression. AIDS 1992 6:347–63 Jones KA, Peterlin BM. 1994. Control of RNA initiation and elongation at the HIV-1 promoter. Annu. Rev. Biochem. 63:713–43

705

181. Folks TM, Justement J, Kinter A, Dinarello CA, Fauci AS. 1987. Cytokineinduced expression of HIV-1 in a chronically infected promonocyte cell line. Science 238:800–2 182. Chen BK, Feinberg MB, Baltimore D. 1997. The kB sites in the human immunodeficiency virus type 1 long terminal repeat enhance virus replication yet are not absolutely required for viral growth. J. Virol. 71:5495–504 183. Bassuk AG, Anandappa RT, Leiden JM. 1997. Physical interactions between Ets and NFkB/NFAT proteins play an important role in their cooperative activation of the human immunodeficiency virus enhancer in T cells. J. Virol. 71:3563–73 184. Wei P, Garber ME, Fang SM, Fischer WH, Jones KA. 1998. A novel CDK9associated C-type cyclin interacts directly with HIV-1Tat and mediates its high-affinity, loop-specific binding to TAR RNA. Cell 92:451–62 185. Herrmann CH, Carroll RG, Wei P, Jones KA, Rice AP. 1998. Tat-associated kinase, TAK, activity is regulated by distinct mechanisms in peripheral blood lymphocytes and promonocytic cell lines. J. Virol. 72:9881–88 186. Garber ME, Jones KA. 1999. HIV-1 tat: coping with negative elongation factors. Curr. Opin. Immunol. 11:460–65 187. Garber ME, Wei P, Jones KA. 1998. HIV1 Tat interacts with cyclin T1 to direct the P-TEFb CTD kinase complex to TAR RNA. Cold Spring Harb. Symp. Quant. Biol. 63:371–80 188. Kim SY, Byrn R, Groopman J, Baltimore D. 1989. Temporal aspects of DNA and RNA synthesis during human immunodeficiency virus infection: evidence for differential gene expression. J. Virol. 63:3708–13 189. Felber BK, Drysdale CM, Pavlakis GN. 1990. Feedback regulation of human immunodeficiency virus type 1 expression by the Rev protein. J. Virol. 64:3734–41


706


190. Malim MH, Cullen BR. 1991. HIV-1 structural gene expression requires the binding of multiple rev monomers to the viral RRE: implications for HIV-1 latency. Cell 65:241–48 191. Finzi D, Hermankova M, Pierson T, Carruth LM, Buck C, Chaisson RE, Quinn TC, Chadwick K, Margolick J, Brookmeyer R, Gallant J, Markowitz M, Ho DD, Richman D, Siliciano RF. 1997. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science 278:1295–300 192. Wong JK, Gunthard HF, Havlir DV, Zhang ZQ, Haase AT, Ignacio CC, Kwok, Emini E, Richman DD. 1997. Reduction of HIV-1 in blood and lymph nodes following potent antiretroviral therapy and the virologic correlates of treatment failure. Proc. Natl. Acad. Sci. USA 94:12574–79 193. Chun TW, Stuyver L, Mizell SB, Ehler LA , Mican JM, Baseler M, Lloyd, AL, Nowak MA, Fauci AS. 1997. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc. Natl. Acad. Sci. USA 94: 13193–97 194. Castro BA, Weiss CD, Wiviott LD, Levy JA. 1999. Optimal conditions for recovery of the human immunodeficiency virus from peripheral blood mononuclear cells. J. Clin. Micro. 26:2371–76 195. Walker CM, Moody DJ, Stites DP, Levy JA. 1986. CD8` lymphocytes can control HIV infection in vitro by suppressing virus replication. Science 234:1563–66 196. Tsubota H, Lord CI, Watkins DI, Morimoto C, Letvin NL. 1989. A cytotoxic T lymphocyte inhibits acquired immunodeficiency syndrome virus replication in peripheral blood lymphocytes. J. Exp. Med. 169:1421–34 197. Cocchi F, DeVico AL, Garzino-Demo A, Arya SK, Gallo RC, Lusso P. 1995. Identification of RANTES, MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive

198.

199.

200.

201. 202.

203.

204.

205.

factors produced by CD8` T cells. Science 270:1811–15 Patel SS, Duby AD, Thiele DL, Lipsky PE. 1988. Phenotypic and functional characterization of human T cell clones. J. Immunol. 141:3726–36 Carpenter CC, Fischl MA, Hammer SM, Hirsch MS, Jacobsen DM, Katzenstein DA, Montaner JS, Richman DD, Saag MS, Schooley RT, Thompson MA, Vella S, Yeni PG, Volberding PA. 1998. Antiretroviral therapy for HIV infection in 1998: undated recommendatiosn of the International AIDS Society–USA Panel. JAMA 280:78–86 Finzi D, Blankson J, Siliciano JD, Margolick JB, Chadwick K, Pierson T, Smith K, Lisziewicz J, Lori F, Flexner C, Quinn TC, Chaisson RE, Rosenberg E, Walker B, Gange S, Gallant J, Siliciano RF. 1999. Latent infection of CD4` T cells provides a mechanism for lifelong persistence of HIV-1, even in patients on effective combination therapy. Nat. Med. 5:512–17 Flexner C. 1998. HIV-1 protease inhibitors. N. Engl. J. Med. 338:1281–92 Lisziewicz J, Rosenberg E, Lieberman J, Jessen H, Lopalco L, Siliciano R, Walker B, Lori F. 1999. Control of HIV despite the discontinuation of antiretroviral therapy [letter]. N. Engl. J. Med. 340:1683– 84 Harrigan PR, Whaley M, Montaner JSG. 1999. Rate of HIV-1 RNA rebound upon stopping antiretroviral therapy. AIDS. 13:F59–F62 Zhang L, Ramratnam B, Tenner-Racz K, He Y, Vesanen M, Lewin S, Talal A, Racz P, Perelson AS, Korber B, Markowitz M, Ho DD. 1999. Quantifying residual HIV1 replication in patients receiving combination antiretroviral therapy. N. Engl. J. Med. 340:1605–13 Chun T-W, Engel D, Berrey MM, Shea T, Corey L, Fauci AS. 1998. Early establishment of a pool of latently infected, resting CD4` T cells during primary


206.


207.

208.

209.

210.

211.

212.

213.

HIV-1 infection. Proc. Natl. Acad. Sci. USA 95:8869–73 Ramratnam B, Mittler JE, Zhang L, Boden D, Hurley A, Fang F, Macken CA, Perelson AS, Markowitz M, Ho DD. 1999. The decay of the latent reservoir of replication competent HIV-1 is inversely correlated with the extent of residual viral replication during prolonged antitretroviral therapy. Ms in prep. Furtaldo MR, Callaway DS, Phair JP, Kunstman BS, Stanton JL, Macken CA, Perelson AS, Wolinsky SM. 1999. Persistence of HIV-1 transcription in peripheral blood mononuclear cells in patients receiving potent antiretroviral therapy. N. Engl. J. Med. 340:1614–22 Izopet J, Salama G, Pasquier C, Sandres K, Marchou B, Massip P, Puel J. 1998. Decay of HIV-1 DNA in patients receiving suppressive antiretroviral therapy. J. Acquir. Immune Defic. Syndr. 19:478–83 Murali-Krishna K, Altman JD, Suresh M, Sourdive DJD, Zajac AJ, Miller JD, Slansky J, Ahmed R. 1998. Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 8:177–87 Topham DJ, Doherty PC. 1998. Longitudinal analysis of acute Sendai virus -specific CD4` T cell response and memory. J. Immunol. 161:4530–35 Natarajan V, Bosche M, Metcalf JA, Ward DJ, Lane HC, Kovacs JA. 1999. HIV-1 replication in patients with undetectable plasma virus receiving HAART. Highly active antiretroviral therapy. Lancet 353:119–20 Reijers MH, Weverling GJ, Jurriaans S, Wit FW, Weigel HM, Ten Kate RW, Mulder JW, Frissen PH, van Leeuwen R, Reiss P, et al. 1998. Maintenance therapy after quadruple induction therapy in HIV-1 infected individuals: Amsterdam Duration of Antiretroviral Medication (ADAM) study. Lancet 352:185–90 Havlir DV, Marscher IC, Hirsch MS, Collier AC, Tebas P, Basset RL, Ioanni-

214.

215.

216.

217.

218.

219.

707

dis JPA, Holohan BA, Leavitt R, Boobe G, Richman DD. 1998. Maintenance antiretroviral therapies in HIV-infected subjects with undetectable plasma HIV RNA after triple drug therapy. N. Engl. J. Med. 339:1261–68 Pialoux G, Raffi F, Brun-Vezinet R, Meiffredy V, Flandre P, Gastaut JA, Dellamonica P, Yeni P, Delfraissy JF, Aboulker JP. 1998. Trilege study team: a randomized trial of three maintenance regimens given after three months of induction therapy with zidovudine, lamivudine, and indinavir in previously treated patients. N. Engl. J. Med. 338:1269–76 Lewin SR, Vesanen M, Kostrikis L, Hurley A, Duran A, Zhang L, Ho DD, Markowitz M. 1999. Use of real-time PCR and molecular beacons to detect virus replication in human immunodeficiency virus type 1-infected individuals on prolonged effective antiviral therapy. J. Virol. 73:6099–6103 Simmonds P, Zhang LQ, McOmish F, Balfe P, Ludlam CA, Brown AJL. 1991. Discontinuous sequence changes of human immunodeficiency virus (HIV) type 1 env sequences in plasma viral and lyphocyte associated proviral populations in vivo: implications for models of HIV pathogenesis. J. Virol. 65:6266–76 Gunthard HF, Wong JK, Ignacio CC, Guatelli JC, Riggs NL, Havlir DV, Richman DD. 1998. Human immunodeficiency virus replication and genotypic resistance in blood and lymph nodes after a year of potent antiretroviral therapy. J. Virol. 72:2422–28 Wong JK, Hezareh M, Gunthard HF, Havlir DV, Ignacio CC, Spina CA, Richman DD. 1997. Recovery of replicationcompetent HIV despite prolonged suppression of plasma viremia. Science 278:1291–95 Chun T-W, Engel D, Mizell S, Metcalf J, Hallahan J, Kovacs J, Davey R, Dybul M, Mican J, Lane C, Fauci A. 1999.

708

220.

221.


222.

223.

224.

225.

226.

PIERSON n MCARTHUR n SILICIANO Effect of interleukin-2 in diminution of a pool of latently infected, resting CD4` T cells in HIV-1 infected patients receiving highly active antiretroviral therapy. 6th Conf. on Retroviruses and Opportunistic Infections, Chicago, IL Chun T-W, Engel D, Mizell SB, Ehler LA, Fauci AS. 1998. Induction of HIV-1 replication in latently infected CD4` T cells using a combination of cytokines. J. Exp. Med. 188:83–91 Unutmaz D, Pileri P, Abrignani S. 1994. Antigen-independent activation of naive and memory resting T cells by a cytokine combination. J. Exp. Med. 180:1159–64 Kovacs JA, Baseler M, Dewar RJ, Vogel S, Davey RT, Falloon J, Polis MA, Walker RE, Stevens R, Salzman NP, Metcalf JA, Masur H, Lane HC. 1995. Increases in CD4 T lymphocytes with intermittent courses of interleukin-2 in patients with human immunodeficiency virus infection. N. Engl. J. Med. 332:567–75 Davey RTJr, Chaitt DG, Piscitelli SC, Wells M, Kovacs JA, Walker RE, Falloon J, Polis MA, Metcalf JA, Masur H, Fyfe G, Lane HC. 1997. Subcutaneous administration of interleukin-2 in human immunodeficiency virus type 1-infected persons. J. Infect. Dis. 175:781–89 Jacobson EL, Pilaro F, Smith KA. 1996. Rational interleukin 2 therapy for HIV positive individuals: daily low doses enhance immune function without toxicity. Proc. Natl. Acad. Sci. USA 93: 10405–10 Chun TW, Engel D, Mizell SB, Hallahan CW, Fischette M, Park S, Davey RTJ, Dybul M, Kovacs JA, Metcalf JA, Mican JM, Berrey MM, Corey L, Lane HC, Fauci AS. 1999. Effect of interleukin-2 on the pool of latently infected, resting CD4` T cells in HIV-1-infected patients receiving highly active antiretroviral therapy. Nat. Med. 5:651–55 McCoig C, Van Dyke G, Chou C-S,

227.

228.

229.

230.

231.

232.

Picker L, Ramilo O, Vitetta ES. 1999. An anti-CD45RO immunotoxin eliminates T cells latently infected with HIV-1 in vitro. Proc. Natl. Acad. Sci. USA 96:11482–85 Ostrowski MA, Chun TW, Justement SJ, Motola I, Spinelli MA, Adelsberger J, Ehler LA, Mizell SB, Hallahan CW, Fauci AS. 1999. Both memory and CD45RA`/CD62L` naive CD4(`) T cells are infected in human immunodeficiency virus type 1-infected individuals. J. Virol. 73:6430–35 Rosenberg ES, Billingsley JM, Caliendo AM, Boswell SL, Sax PE, Kalams SA, Walker BD. 1997. Vigorous HIV-1-specific CD4` T cell responses associated with the control of viremia. Science 278:1447–50 Nowak MA, Lloyd AL, Vasquez GM, Wiltrout TA, Wahl LM, Bishofberger N, Williams J, Kinter A, Fauci AS, Hirsch VM, Lifson JD. 1997. Viral dynamics of primary viremia and antiretroviral therapy in simian immunodeficiency virus infection. J. Virol. 71:7518–25 Little SJ, McLean AF, Spina CF, Richman DD, Havlir DV. 1999. Viral dynamics of acute HIV-1 infection. J. Exp. Med. 190:841–50 Grossman Z, Polis M, Feinbrg MB, Grossman Z, Levi I, Jankelevich S, Yarchoan R, Boon J, de Wolf F, Lange JMA, Goudsmit J, Dimitrov DS, Paul WE. 1999. Ongoing HIV dissemination during HAART. Nat. Med. 5:1099–1104 Zhang Z, Schuler T, Zupancic M, Wietgrefe S, Staskus KA, Reimann KA, Reihart TA, Rogan M, Cavert W, Miller CJ, Veazey RS, Notermans D, Little S, Danner SA, Richman DD, Havlir D, Wong J, Jordan HL, Schacker TW, Racz P, Tenner-Racz K, Letvin NL, Wolinsky S, Haase AT. 1999. Sexual transmission and propagation of SIV and HIV in resting and activated CD4(`) T cells. Science 286:1353–57


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


REGULATION OF ANTIBODY RESPONSES VIA ANTIBODIES, COMPLEMENT, AND FC RECEPTORS Birgitta Heyman Department of Genetics and Pathology, Uppsala University, Uppsala, Sweden; e-mail: [email protected]


Key Words immune complex, CD23, IgE, IgG, IgM Abstract Antibodies can completely suppress or enhance the antibody response to their specific antigen by several hundredfold. Immunoglobulin M (IgM) enhances antibody responses via the complement system, and complement activation by IgM probably starts the chain of events leading to antibody responses to suboptimal antigen doses. IgG can enhance primary antibody responses in the absence of the complement system and seems to be dependent on Fc receptors for IgG (FccRs). IgE enhances antibody responses via the low-affinity receptor for IgE (FceRII/CD23). The precise effector mechanisms that cause enhancement are not known, but direct B-cell signaling, antigen presentation, and increased follicular localization are all possibilities. IgG, IgE, and IgM may also suppress antibody responses when used in certain immunization regimes, and it seems reasonable that an important mechanism behind suppression is the masking of antigenic epitopes by antibodies. In addition, FccRIIB, which contains a cytoplasmic inhibitory motif, acts as a negative regulator of antibody responses. This receptor, however, may prevent the antibody responses from exceeding a certain level rather than causing complete suppression.

INTRODUCTION The components of immune complexes include antigens, antibodies, and, when an antibody is of a complement-activating isotype, complement. Such complexes interact with the B-cell receptor (BCR) and Fc and complement receptors and are known to have dramatic immunoregulatory functions. Administration of small amounts of specific immunoglobulin G (IgG) with erythrocytes completely inhibits erythrocyte-specific antibody responses (1–4). This capacity of IgG has been used successfully in clinical practice to prevent women who lack the Rhesus D antigen on their erythrocytes (i.e., are RhD1) from becoming immunized against fetal RhD` erythrocytes transferred via transplacental hemorrhage (5). Treatment of mothers with IgG anti-RhD has signifi0732–0582/00/0410–0709$14.00

709


710

HEYMAN

cantly lowered the incidence of hemolytic disease in newborns (reviewed in 6) and is one of the major clinical achievements of modern immunology. In other situations, the presence of an antibody may act to enhance antibody responses. The classic example is the ability of IgM to enhance the response to sheep erythrocytes (SRBCs) (1), a response that was later shown to depend on the ability of IgM to activate complement (7). Recent data indicate that the binding to antigen and activating of complement by endogenous, natural IgM start the chain of events that lead to antibody responses to physiological concentrations of antigen (8, 9). This finding may explain why patients and animals with deficiencies in the early components of the complement system (C2, C3, and C4) or in complement receptor 2 (CR2/CD21) have poor immune responses (reviewed in 10–12). The enhancement of the antibody response by specific IgM suggests that vaccinations might be improved by targeting antigens to complement receptors. Indeed, immunization of animals with a fusion protein of antigen and C3d (13), with antigen covalently coupled to C3b (14) or to monoclonal antibodies (mAbs) specific for complement receptor 1 (CR1/CD35) and CR2 (15), results in potent antigen-specific responses. Not only IgM, but also IgG and IgE, can enhance antibody responses. In contrast to IgM, these isotypes normally do not work with particulate antigens but with soluble proteins. For example, administration of 2,4,6-trinitrophenyl (TNP)-specific IgG (16, 17) or IgE (18, 19) with TNP-bovine serum albumin (TNP-BSA) or TNP-keyhole limpet hemocyanine (TNP-KLH) induces antibody responses that are several hundredfold higher than administration of antigen alone. These phenomena require Fc receptors for IgG (FccRs) (17) and lowaffinity Fc receptors for IgE (FceRIIs) (18–20), respectively, and are thought to be significant in the development of autoimmune and allergic diseases (21–23). In recent years, the understanding of antibody-mediated regulation of antibody responses has advanced considerably, owing to mAb technology, allowing identification of the class or subclass of antibodies involved. In addition, mAbs have been altered and the importance of their specific effector functions compared with those of the native antibody. The use of mAbs that block Fc or complement receptors and gene-targeted mice has, in some cases, made it possible to determine which receptors are involved. Despite these technical advances, the precise molecular mechanisms by which an antibody can upregulate or downregulate the antibody response are still largely unknown. Co-cross-linking the BCR with the CR2-receptor complex (known to enhance B-cell activation in vitro), efficient receptor-mediated endocytosis by antigen-presenting cells followed by presention to T cells, and increased localization of immune complexes in lymphoid follicles on follicular dendritic cells (FDCs) carrying complement receptors and Fc receptors (FcRs) have been proposed as explanations for enhancement. Suppression of antibody responses may result from elimination of complexes through receptormediated phagocytosis, prevention of B-cell recognition of antigen through epitope masking, or via co-cross-linking of FccRIIB and BCR, known to inhibit B-cell activation in vitro (24–27) as well as in vivo (17, 28).

ANTIBODY FEEDBACK REGULATION

711


This review summarizes experimental data on how antibodies regulate antibody responses to their specific antigens. Unless stated otherwise, only antibodymediated regulation in vivo is discussed. There are considerable differences between in vitro and in vivo systems, such as the lack of lymphoid organ structure in vitro, and therefore comparisons are difficult. Extensive reviews of antibodymediated regulation have been published previously (29, 30). A condensed view of the field is presented in Figure 1. A summary of the mAbs used to define the antibody classes involved and references for more information can be found in Table 1, whereas Table 2 summarizes current knowledge about which Fc and complement receptors are involved and provides references for more information.

Figure 1 Antibody-mediated regulation of antibody responses. The frequency of antibody-mediated suppression of antibody responses is the greatest by far when IgG is administered with particulate antigens like erythrocytes (1–4). However, it has also been observed that IgM (32, 40, 51) and IgE (4) can induce suppression of antibody responses to erythrocytes and that IgG can suppress responses to soluble antigen (16, 34) and to antigens administered in adjuvants (60–62). Enhancement of antibody responses is efficiently induced by IgM with particulate antigens (erythrocytes or malaria parasites) (1, 38, 104) or by IgE (18–20) and IgG (16, 39) with soluble protein antigens. IgA was shown in one report to enhance induction of immunological memory to soluble antigen (39) but has otherwise not been studied. (Adapted from 30a)

712

HEYMAN

TABLE 1 Monoclonal antibodies and antibody-mediated regulation mAb classa

Suppressionb

IgM IgG1 IgG2a IgG2b IgG3 IgA IgE

(2) (2, 3, 47) (2, 3, 47) (4, 47) (2, 3) NTe (4)

Enhancementc (7, 38, 104) (16, 17, 39, 41, 45) (16, 17, 41, 45, 85) (4, 16, 17, 41, 45, 85) d

(39) (18–20, 42, 141, 142)

a

mAb administered together with antigen. Suppression was studied by using erythrocytes as antigen. c IgM-mediated enhancement was studied by using erythrocytes as antigen, whereas IgG-, IgA-, and IgE-mediated enhancement were studied by using soluble proteins. d J Dahlstro¨m, unpublished data. e NT, Not tested.


b

TABLE 2 Receptor-deficient mice and antibody-mediated regulation

Mouse strain FccRIIB1/1 FccR1/1c FccRIII1/1 FccRIIB1/1 2 FcRc1/1d b2-Microglobulin1/1e FceRII1/1 CR1/21/1

IgG-mediated suppressiona (reference)

IgG-mediated enhancementb (reference)

IgM-mediated enhancementa

IgE-mediated enhancementb (reference)

Normal (4) Normal (4) NT (4) Normal (4)

Enhanced (17) Decreased (17) Normal (17) NT

NTh NT NT NT

Normal (17) Normal (17) NT NT

Normal (4) NT NT

NT Normalf Normalg

NT NT Absentg

NT Absent (20,173) Normalg

a

Studied using SRBC or TNP-SRBC as antigen. Studied using TNP-BSA as antigen. c Deficient in functional FccRI and FccRIII. d Double knockouts deficient in FccRI, FccRIIB, and FccRIII. e Deficient in the neonatal FcR, FcRn. f S Gustavsson, S Wernersson, B Heyman (173) g SE Applequist, J Dahlstro¨m, H Molina, B Heyman (manuscript in preparation). h NT, Not tested. b

GENERAL EXPERIMENTAL CONSIDERATIONS Antibody feedback regulation is a general phenomenon in mammals and has been seen in many species, including rabbits, horses, pigs, guinea pigs, rats, and mice. IgG-mediated suppression and IgM-mediated enhancement have also been observed in humans (5). In the experimental systems used to study antibody-



713

mediated regulation, antigen is administered with, shortly before, or soon after a preformed specific antibody, and the active antibody production is then compared with antibody production in animals given the antigen alone. To induce enhancement, the antibody must be present at the time of antigen encounter or be administered within a few hours after the antigen (19, 31–34). Suppression by IgG has been shown to be most efficient when IgG is administered within 24 h of the antigen (1). However, RhD prophylaxis works well when IgG anti-RhD is administered 72 h after the antigen (35), and IgG given 5 days after the antigen is able to terminate ongoing antibody responses (36). Enhancement by IgM (1, 7, 16, 32, 37, 38), IgG (16, 34, 39–41), and IgE (18, 19, 42) is antigen specific but not epitope specific. This means that an antibody specific for one epitope of the antigen enhances the response to all epitopes present on that antigen, but it has no effect on responses to antigens not present in the complex. ‘‘Bystander enhancement’’ against a nonspecific antigen, administered with the specific IgM/antigen or IgG/antigen complex, has occasionally been observed (38, 41), but it is not a reproducible finding (SE Applequist, J Dahlstro¨m, H Molina, B Heyman, manuscript in preparation; see also 17). It may be that immunostimulatory cytokines generated by the interaction between specific IgG/ antigen complexes and the immune system enhance responses to other antigens in the vicinity. Suppression by IgG (2–4) and IgE (4) is also antigen specific and is usually not epitope specific (2, 3, 43–47), although epitope-specific suppression has been observed (48, 49).

IGM- AND IGE-MEDIATED SUPPRESSION Both IgM and IgE, which are generally known for their ability to enhance antibody responses, can also act as inhibitors, although such reports are rare. High doses of IgM suppress the response to SRBCs, whereas low doses of the same preparation enhance the response (40, 51). IgM, administered 2–48 h after SRBCs, suppressed antibody responses, whereas the same antibody preparation administered before the antigen was given enhanced antibody responses (32). The same IgM preparation was shown to enhance antibody responses in vivo, whereas it suppressed the in vitro response (52). Finally, monoclonal IgM, specific for 4hydroxy-3-nitrophenylacetyl (NP), was shown to suppress 98% of the response to haptenated SRBCs (2). Recently, three monoclonal IgE anti-TNP antibodies known to enhance antibody responses to TNP-BSA (18) were reported to suppress the response to TNPSRBCs (4). IgE binds FceRI, FceRII, FccRIIB, and FccRIII (53), but the suppressive effects of IgE were unperturbed in mice lacking functional FccRI, FccRIII, and FceRI [owing to deletion of the common FcRc-chain subunit (54)] and in mice lacking FccRIIB (MCI Karlsson, T Diaz de Sta˚hl, B Heyman, manuscript in preparation). It seems reasonable that IgM- and IgE-mediated suppres-

714

HEYMAN

sion, as well as IgG-mediated suppression (see below), can be caused by passive antibodies masking antigenic epitopes.


IGG-MEDIATED SUPPRESSION The most widely recognized feedback regulation is probably the ability of IgG to inhibit antibody responses. Most studies of IgG-mediated suppression have been performed in animals immunized with heterologous erythrocytes, usually haptenated or native SRBCs (1–4, 50, 51, 55), and the responses have been assayed as serum titers or at the single-cell level in hemolytic PFC assays. In such systems, the primary IgM (1–4, 49, 55) and IgG (55) production is severely impaired, often to ,1% of control responses. Both passively administered and endogenously produced antibodies down-regulate antibody responses (3, 56). Although primary antibody responses are completely inhibited by IgG, other parameters of the immune response are less severely impaired. Priming of Thelper cells appears to be normal in vivo (4) and in vitro (57). Induction of anamnestic antibody responses as well as secondary antibody responses can be inhibited, but the effects are less pronounced than in primary responses (55, 58, 59). In addition to its effects in systems using erythrocytes as antigens, IgG suppresses responses to proteins administered in adjuvants (60–62). Although IgG usually enhances the response to soluble proteins, inhibition of the response to KLH, diphtheria, and tetanus toxoid has occasionally been observed (16, 34). Monoclonal TNP-specific IgG administered with TNP-SRBCs suppressed the response to SRBCs, whereas the same antibodies administered with TNP-KLH enhanced the response to KLH (45, 47), illustrating the dual effects of IgG. IgG-mediated suppression is dose dependent, and high concentrations of IgG can suppress the response to optimal doses of erythrocytes (1, 59). Administration of $0.4 lg of polyclonal or monoclonal IgG can inhibit .90% of a primary IgM response (3, 4), whereas $2 lg is required to obtain .99% suppression (2–4). Monoclonal IgG antibodies of all subclasses, including IgG3, suppress antibody responses (2–4, 47), and the suppressive ability correlates closely with affinity (2, 47, 59, 63). Comparison of the suppressive effects of F(ab’)2 fragments and intact IgG on responses to SRBCs in vivo has produced conflicting results. In some studies F(ab’)2 fragments were shown to suppress responses (4, 60, 64), whereas in other studies suppression by the fragments was much less efficient than by intact IgG (2, 36, 45, 58, 65).

Epitope Masking and Fcc Receptors in IgG-Mediated Suppression At least three different explanations for IgG-mediated suppression have been proposed: (a) IgG could mask antigenic epitopes and thereby prevent B cells from recognizing and responding to the antigen, (b) IgG/antigen complexes may be


715


eliminated by FccR` phagocytic cells before they can activate specific B cells, and (c) IgG/antigen may inhibit B-cell activation by co-crosslinking the BCR and the inhibitory FccRIIB. FccRIIB is the only FccR expressed on B cells and contains a cytoplasmic inhibitory motif, the immune-receptor tyrosine-based inhibitory motif, which, when brought into proximity with receptors containing a specific activation motif, that is, the immune-receptor tyrosin-based activation motif, inhibits cell activation through the latter [reviewed in Dae¨ron (66)]. The BCR contains immune-receptor tyrosin-based activation motifs, and it has been well documented that FccRIIB inhibits B-cell activation in vitro (24–27). Whereas epitope masking would function independently of the Fc portion of IgG, the second and third mechanism would require an intact Fc portion. The question of whether IgG-mediated suppression is Fc dependent is therefore of considerable interest but has been difficult to answer unequivocally. Role of the IgG(Fc) Region In vivo and in vitro studies demonstrating that F(ab’)2 fragments were poor suppressors (2, 45, 58, 65, 67) and that IgG-mediated suppression was not epitope specific (2, 3, 43–47) suggested that the IgG(Fc) region was required. The ability of IgG to suppress antibody responses independently of the complement system, tested by using non-complement-activating IgG (68) and in mice depleted of C3 via treatment with cobra venom factor (B Heyman, unpublished observation), implied that the Fc-mediated effector function necessary for suppression was the ability of IgG to bind to FccRs. This conclusion was strengthened by the in vitro findings that deglycosylated monoclonal IgG, which could neither bind FccRs nor activate complement, was a poor suppressor (69) and that suppression by normal IgG could be reversed by a mAb binding to FccRIIB and III (70). However, it was recently shown that monoclonal IgG antiTNP suppressed the primary response to TNP-SRBCs in mice that lacked FccRIIB as well as FccRI ` III [owing to the lack of the common FcRc-chain subunit (FcRc)], FccRI ` IIB ` III (double knockouts of FcRc and FccRIIB), or the neonatal FcR (FcRn) (owing to the lack of b2-microglobulin, which constitutes part of the receptor) (4). Suppression by this preparation was as efficient as in the wild-type controls with IgG concentrations that ranged from 0.4 to 100 lg/mouse. Possible Mechanisms in IgG-Mediated Suppression Efficient suppression in FccR-deficient mice is hard to reconcile with a significant involvement of FccRs and is best explained by the epitope-masking model. This model is also compatible with the observations that IgG must bind to the antigen at a high epitope density (3, 47–50) and with a high affinity (2, 47, 59, 63) to be suppressive and that IgG3 (2, 3), which is unable to bind FccRIIB, IgM (2, 32, 40, 51), IgE (4), and F(ab’)2 (4, 60, 64) fragments can be suppressive. Because epitope masking would not prevent uptake of IgG/antigen complexes by antigen-presenting cells, presentation to T cells would be expected to be normal. This is indeed the case


716

HEYMAN

(4), and it may explain the relative difficulty in inhibiting induction of memory and secondary responses. How can the epitope-masking model for IgG-mediated suppression be resolved, taking into account the abundance of apparently conflicting data that support Fc dependence? This has been discussed in detail previously (4, 71), but a short summary is presented here. First, the existence of non-epitope-specific suppression cannot be taken as evidence for Fc dependence, because TNP-specific F(ab’)2 injected with TNP-SRBCs suppressed the SRBC-specific response (4). In this situation, Fc-mediated effects by IgG are excluded, and the interpretation of the results is that antibodies binding to one epitope on an antigen can sterically hinder recognition of neighboring epitopes. It is known that IgG, to induce nonepitope-specific suppression, must bind to the antigen at a high density, that is, the antigen must have a high epitope density (2, 3, 47, 50). Therefore, most likely, non-epitope-specific suppression will occur in situations of high epitope density, whereas epitope-specific suppression will result when the epitope is less abundant. Second, inefficient suppression by F(ab1)2 could be a result of more rapid elimination of F(ab1)2-fragments than of intact IgG, because protection of IgG from degradation by FcRn requires an intact Fc-portion (72). In this sense, suppression can be said to be Fe-dependent. Finally, it is important to distinguish between suppression of primary/secondary and in vivo/in vitro antibody responses. It should be emphasized that unperturbed IgG-mediated suppression in FccR-deficient mice so far has been demonstrated only in primary in vivo antibody response when IgG is administered in close temporal relationship to erythrocytes. Activated B cells express levels of FccRIIB that are tenfold higher than levels from resting B cells (73), and primed B cells have a higher affinity for antigen than naı¨ve B cells. These factors could increase the likelihood of co-cross-linking between the BCR and FccRIIB in for example secondary responses. FccRIIB has been shown to regulate negatively many immune functions in vitro, for example, activation of B cells (24–27), T cells (74), and mast cells (74, 75), as well as antigen presentation (76, 77) and phagocytosis (78). Studies of FccRIIB-deficient mice have confirmed the inhibitory role of the receptor in vivo. Such animals exhibit augmented antibody and anaphylactic responses (17, 28, 79), increased inflammation (80), and more severe collagen-induced arthritis (81) (S Kleinau, P Martinsson, & B Heyman, submitted for publication) than wildtype animals. Why, then, does FccRIIB not play a significant role in the suppression of primary SRBC responses? A likely explanation is that this receptor does not completely prevent antibody responses, but merely prevents antibody levels from extending above an appropriate ceiling. This point is illustrated by an experiment where FccRIIB-deficient and wild-type mice were immunized with IgG and soluble antigen or with antigen alone (17). IgG is known to enhance antibody responses to soluble antigen and, as expected, wild-type animals given, e.g., monoclonal IgG1 anti-TNP/TNP-BSA had significantly higher titers of BSAspecific IgG than wild-type animals given antigen alone (28 lg/ml versus 0.1 lg/ ml). FccRIIB-deficient mice given the same IgG1/antigen complexes produced 990 lg/ml of IgG anti-BSA. Thus, the function of FccRIIB in normal mice was


717

not to prevent responses to IgG/antigen altogether, but merely to prevent them from reaching abnormal levels. If FccRIIB works in this manner, it would be expected that the antibody response to optimal doses of noncomplexed antigen would be higher in FccRIIB-deficient than in normal mice, which is the case (28), but that the response to suboptimal doses of antigen would be the same for the two types of mice. In conclusion, most data point toward epitope masking as the most important mechanism explaining complete suppression of primary antibody responses. Negative regulation via FccRIIB probably plays a role in preventing responses to highly immunogenic antigen from becoming too high. Its role in terminating ongoing antibody responses, in memory responses, and in antibody responses in vitro remains to be investigated.


IGG-MEDIATED ENHANCEMENT In addition to its suppressive effect, IgG is a potent enhancer of antibody responses. Early studies have shown that complexes of antigen and heterologous antisera often elicit stronger immune responses than antigen alone (see 82 for references). Complexes formed with hyperimmune mouse serum (82–84) were also found to enhance primary antibody responses in mice, thus excluding the foreign nature of heterologous antisera as the reason for enhancement. The demonstration that monoclonal IgG antibodies enhanced primary serum IgG (16, 17, 34, 41, 45, 85) and IgM titers (34), as well as the numbers of specific antibodyproducing B cells (34, 45), definitely proved that IgG antibodies can induce an augmentation of the antibody response. The magnitude of enhancement is impressive, sometimes resulting in IgG levels in mice given IgG/KLH complexes that were .1000-fold higher than in mice given KLH alone (16). Hyperimmune antisera also prime memory responses efficiently (39, 83, 84, 86–88), and administration of $0.1 lg of 2,4, dinitrophenyl (DNP)-KLH in complex with IgG was as efficient as a 100-fold higher dose of antigen alone (84). In addition, monoclonal IgG has been shown to enhance secondary immune responses (16, 39). Priming can occur even though primary responses are undetectable or even suppressed (83, 89, 90). IgG enhances responses in several species and in most mouse strains, but for unknown reasons it is unable to enhance a primary response to TNP-BSA in mice homozygous for MHC class II Ab (85). IgG-mediated enhancement has been described primarily when the antigen is a soluble protein. Monoclonal TNP-specific IgG given with TNP-KLH, TNPtransferrin, or TNP-BSA enhances the carrier-specific IgG responses (16, 17, 34, 41, 47, 85). In similar systems, the IgG anti-DNP response (45) and the IgM antiKLH response (34) were also enhanced. IgG seems to be more efficient when KLH with a low, rather than a high, TNP substitution ratio is used (34). A moderate (2- to 10-fold) upregulation of primary responses to erythrocytes has been observed in rabbits (40) as well as in mice (44, 90–92).


718

HEYMAN

In studies with hyperimmune sera, optimal responses are obtained with complexes formed with a slight antigen excess or with antigen at an equivalent level (83, 84, 93, 94). Monoclonal IgG has an optimal effect at an antigen/antibody ratio of 1:5 (350-fold enhancement) but is still effective at a ratio of 25:1 (11fold enhancement) (16). Primary responses at the single-cell level are significantly enhanced 5–6 days after priming with IgG anti-TNP and TNP-KLH (34, 45), and serum titers remain high for several months (16, 41). Priming of B memory cells is already detectable at day 5 (87), and evidence for T-cell priming on day 3 has been reported (86). IgG anti-4-hydroxy-3-nitrophenyl (NP) administered with NP-KLH also increased the number of somatic mutations in NP-specific germinal center B cells (86a). Monoclonal IgG1, IgG2a, and IgG2b (16, 17, 41, 45, 47, 85), as well as IgG3 (J Dahlstro¨m, T Diaz de Sta˚hl, B Heyman, unpublished data), enhance primary responses in mice. Monoclonal IgG1 (39), IgG2a (16), or IgG2b (16) primes for memory, whereas IgG3 has not been tested in this respect. The affinity of IgG and antibodies does not seem to correlate with the enhancing effect (16, 47).

Fcc Receptors and Complement in IgG-Mediated Enhancement IgG-mediated enhancement of the antibody response is not caused by simple aggregation of the antigen (84), implying involvement of either the complement system, FccRs, or both. Role of Complement Whether IgG antibodies need to be able to activate complement to enhance antibody responses has been a controversial issue. Complement depletion of mice treated with cobra venom factor prevented trapping of aggregated human IgG (95) and IgG/DNP-KLH complexes (84, 90) in mouse spleens. The same treatment abolished the ability of IgG to prime DNP-specific B cells (84, 90) and implied that IgG-mediated enhancement was complement dependent. This conclusion received support when the capacity of monoclonal DNP-specific IgG, IgM, and IgA, complexed to 125I-labeled DNP-KLH, to localize in splenic follicles was found to correlate with the ability of these antibodies to enhance memory cell induction (39). In addition, the ability of a panel of IgG mAbs to activate complement in vitro correlated with their ability to enhance primary antibody responses (16, 47). Other studies suggest that IgG can enhance primary antibody responses in the absence of the complement system. Mutant IgG2a and monoclonal IgG1, both unable to activate complement, efficiently enhance the antibody responses to KLH (41). Monoclonal IgG can enhance antibody responses in complement-depleted mice (41) and in mice deficient in CR1/ 2 (SE Applequist, J Dahlstro¨m, H Molina, B Heyman, manuscript in preparation). A strong indication that IgG-mediated enhancement is primarily dependent on FccRs rather than on complement came from the observation of an almost com-


719


plete lack of response to IgG1/antigen and IgG2a/antigen in FcRc-deficient mice (17), although these animals have a normally functioning complement system. From the results described above, it seems clear that the complement system is not required for the ability of IgG to enhance primary responses, although the possibility cannot be excluded that the complement system plays a role in the ability of IgG to prime for memory. However, as discussed in detail below, it is now known that complement is also crucial for the generation of antibody responses to low doses of antigen administered in noncomplexed form (reviewed in 12). Therefore, the lack of responses to IgG/antigen in complement-depleted mice may reflect a general impairment of antibody responses that is not necessarily related to the IgG in the complex. Role of Fcc-Receptors The impaired responses to IgG1/antigen and IgG2a/antigen in FcRc-deficient mice (17), which are shown to lack functional FccRI and FccRIII (54), suggest the involvement of one or both of these receptors. The dependence on FcRc for IgG2b- and IgG3-mediated enhancement is not yet clearly established. Mice that selectively lack FccRIII have normal responses to complexes between IgG1, IgG2a, and IgGb and antigen (17). Therefore, FccRI (perhaps with FccRIII) most likely mediates enhanced responses to IgG/antigen complexes. FccRI has been identified as the murine IgG3-receptor (96), and the observation that IgG3 can enhance antibody responses in normal mice (J Dahlstro¨m, T Diaz de Sta˚hl, B Heyman, unpublished data) is consistent with a FccRI role in IgG-mediated enhancement. FcRc is a subunit not only of FccRI and FccRIII, but also of FceRI, FcaR, and a subset of CD3 (reviewed in 97). However, the involvement of FceRI or FcaR is unlikely because IgG does not bind to these receptors. Impaired T-cell help via CD3 malfunction is unlikely because antibody responses to IgE/antigen complexes are normal in FcRc-deficient mice (the antibody response to IgE/antigen is exclusively dependent on FceRII but presumably needs T-cell help in a manner similar to IgG/antigen complexes) (17). In FcRc-deficient mice, FccRIIB is present in normal amounts. Therefore, the lack of response to IgG1 and IgG2a/antigen complexes excludes FcRIIB as a major player in these situations. Possible Mechanisms in IgG-Mediated Enhancement In summary, FccRmediated, rather than complement-mediated, effects seem to be of major importance to the ability of IgG to enhance primary responses, but complement system influence is still a possibility, especially in the induction of memory. It is not known which cell type or types are involved in IgG-mediated enhancement. Murine FccRI is expressed on macrophages, monocytes, and neutrophils (reviewed in 97) and probably also on dendritic cells (98). In vitro, dendritic cells and macrophages present antigenic peptides from IgG/antigen complexes efficiently to T cells (99–101), and it is likely that this mechanism also operates in vivo. Another possibility is that IgG/antigen complexes are captured by FccR`

720

HEYMAN

FDC in lymphoid follicles, which would further enhance the presentation of antigen to B cells.


IGM-MEDIATED ENHANCEMENT The most commonly observed immunoregulatory function of IgM is its ability to enhance antibody responses. A vast majority of investigators have studied regulation of the primary IgM response to erythrocytes (as direct PFCs) administered in close temporal relationship to the IgM antibodies (1, 32, 37, 38, 52, 91). However, primary IgG responses [indirect PFCs or serum titers (38, 55)], including all IgG subclasses (102), IgE production (103), and memory cell (55), as well as T-helper cell induction (104), are also upregulated by IgM. IgM-mediated enhancement has been observed in humans (5) and most laboratory mouse strains with the exceptions of nude mice (37, 105, 106) and old MRL/lpr mice (107). IgM enhances the response to particulate antigens like erythrocytes (1, 32, 37, 38, 52, 91) and malaria parasites (104), but generally does not enhance responses to soluble antigen. However, a monoclonal IgM anti-DNP was able to enhance the IgG response to DNP-KLH (45), and one of four tested monoclonal IgM antiTNPs enhanced the response to TNP-KLH (16), whereas two monoclonal IgM anti-DNPs used in another study had no effect (39). For IgM to enhance a response, it must be administered within a few hours before or after the antigen (31–33). The same IgM preparation that augments a response when administered before SRBCs causes suppression when given 1 or 2 days after SRBCs (32). Enhancement is seen only with suboptimal SRBC doses (1, 33, 106) but requires the presence of the antigen (33, 37, 38, 104). IgM preparations are effective over a wide range of doses (38, 91), but enhancing preparations may turn suppressive with increasing IgM concentrations (40, 51). The enhancing effect of IgM is observed as increased direct PFCs as early as 3 days after immunization (38, 44, 91), and IgG levels are higher than those in control groups for at least 3 months (55).

Complement and Antibody Responses Role of Complement in the Antibody Response to IgM/Antigen Complexes IgM-mediated enhancement of antibody responses does not function without the complement system. Monoclonal IgM anti-TNP, which, owing to a point mutation, is unable to activate complement, loses its ability to enhance the response to TNP-SRBCs (7). Moreover, IgM cannot enhance antibody responses in mice depleted of complement via treatment with cobra venom factor (7) or in mice without functional CR1/2 (SE Applequist, J Dahlstro¨m, H Molina, B Heyman, manuscript in preparation). Possibly the conformational changes that the IgM molecule must go through before it can activate complement (108, 109) cannot take place after binding to small soluble antigens, explaining why it is difficult for IgM to enhance antibody responses to anything but particulate antigens (16, 39).



721

Both IgG and IgM are capable of classical pathway complement activation. Nevertheless, IgG enhances antibody responses in the absence of complement (see above). The complement dependence of enhancement mediated by IgM was studied using erythrocytes, whereas IgG was studied using soluble proteins, and the observed differences could be explained by this fact. For IgG to fix C1q, at least two IgG molecules are required (110). In contrast, one IgM molecule is sufficient to activate complement and to lyse an erythrocyte (111). The more efficient complement activation by IgM is illustrated by the ability of B cells secreting IgM anti-SRBCs to form direct PFCs (112), whereas B cells secreting IgG anti-SRBCs are unable to do so without the addition of anti-IgG antiserum to aid in the complement activation (indirect PFCs). The most likely explanation for the dominating role of complement in IgM- but not IgG-mediated enhancement is that the IgG in these in vivo situations does not activate sufficient amounts of complement. The importance of endogenous natural IgM for antibody responses was predicted in 1967 (40), and the question was recently addressed in studies with mice carrying a targeted deletion of secretory IgM. Such animals have normal numbers of B cells expressing surface IgM and normal levels of serum IgA and IgG, but they completely lack serum IgM. Immunization with soluble NP-KLH resulted in production of less specific IgG, a defect that could be overcome with a transfusion of IgM from normal mouse serum (8, 9). These findings emphasize the role of IgM in initiating antibody responses, and they may explain why complement and CR1/2 are crucial in responses to antigen administered also in the absence of passive IgM (see below). Role of Classical Pathway Complement Activation in the Antibody Response to Noncomplexed Antigen A connection between the complement system and a normal humoral immune response was first made when severely diminished antibody responses were found in mice depleted of C3 via treatment with cobra venom factor (113). Subsequently, guinea pigs (114–116), humans (117), and dogs (118) with hereditary deficiencies in C2, C3, or C4, as well as mice with gene-targeted deletions in C3 or C4 (119), were shown to have poor antibody responses. The fact that deficiencies in the classical pathway components C4 (115–117, 119) and C2 (114) lead to impaired antibody responses suggests that alternative pathway activation is not sufficient. This conclusion is supported by the observation that mice without a functioning alternative pathway, owing to factor B gene targeting, have normal antibody responses (120). Therefore, the primary pathway supporting antibody responses is the classical pathway, which can be activated by antibodies as well as by mannose-binding lectin (121). Role of CR2 in the Antibody Response to Noncomplexed Antigen Very poor antibody responses to erythrocytes and bacteriophage UV174 are elicited in mice lacking CR1/2 (122–124) or when ligation of CR2, but not CR1, was prevented (125–128; reviewed in 12). It is therefore reasonable to assume that the target


722

HEYMAN

receptor for complement-mediated immune regulation is CR2 (125, 128), although distinguishing between murine CR1 and CR2 is difficult because they are derived from the same gene by alternative splicing (129). CR2 is associated on the surface of the B cell with the signaling molecules CD19, TAPA-1, and Leu-13 [reviewed in 130 and D Fearon & MC Carroll (this volume)]. Nevertheless, ligation of CR2 seems to be of primary importance for the function of the receptor complex in inducing antibody production because CR1/2-deficient mice, expressing normal amounts of CD19, still have poor antibody responses (122– 124). Murine CR1/2 is expressed mainly on B cells and FDCs. Studies of CR1/2deficient mice in bone marrow transfer models and in RAG-2–deficient blastocyst complementation suggest that B-cell expression of CR1/2 is of primary importance for normal antibody responses (122, 123). However, a role for FDCs in long-term responses can be inferred, because mice expressing CR1/2 on B cells but not on FDCs have a lower secondary response than mice expressing CR1/2 on both cell types (131). Possible Mechanisms in IgM-Mediated Enhancement From the data summarized above, it is apparent that classical pathway complement activation by IgM is a crucial factor in the chain of events leading to normal antibody production. The exact mechanism by which IgM/antigen/complement complexes initiate antibody responses is not known, although a number of possibilities have been proposed (reviewed in 12). The pathway leading from IgM/antigen/complement complexes to antibody production may involve CR2 exclusively, but it could also involve additional complement receptors. The finding that IgM-mediated, but not IgG- or IgE-mediated, enhancement of antibody responses was abolished in CR1/2-deficient mice suggests an important role for CR2 (SE Applequist, J Dahlstro¨m, H Molina, B Heyman, manuscript in preparation). Co-crosslinking the BCR and the CR2/CD19/TAPA-1/Leu-13 complex lowers the threshold for B-cell activation in vitro (132). A similar mechanism in vivo, in which the receptors are coligated by IgM/antigen/complement complexes, may render the B cell responsive to lower doses of antigen than it would be without such coligation. Antigen can be taken up by CR1/2 on B cells and presented to T cells in vitro (15, 133, 134), but the in vivo relevance of these findings is not clear. T-cell priming is normal in mice with severely impaired B-cell responses caused by C3 or C4 deficiency (119) or by treatment with anti-CR1/2 mAbs (135). Localization of antigen in lymphoid follicles and development of germinal centers are dependent on a functioning complement system (84, 90, 95, 119, 122, 131, 136). Nevertheless, expression of CR1/2 on B cells, but not on FDCs, seems to play a major role for normal antibody responses (122, 123). This apparent paradox can be explained if retention and/or survival of CR1/2` B cells in lymphoid follicles were complement dependent. Indeed, interaction between a complement-derived CR1/2 ligand on FDCs and CR1/2 on B cells was shown to be


723

required for optimal antibody responses in vitro (137). The importance of adhesion molecules is also suggested by the observation that patients with leukocyte adhesion deficiency, resulting in the absence of CD11b/CD18 (CR3) and CD11c/ CD18 (CR4), have an impairment of antibody responses to bacteriophage UV174 similar to that in patients who are deficient in C2 or C3 (138). The explanations above are consistent with the finding that the complement system is crucial only for generating antibody responses to suboptimal immunization (124, 125), whereas the requirement for complement or CR1/2 can be overcome with high doses of erythrocytes (125), with antigen in adjuvants (15, 139, 140), or with antigen in complex with specific IgG or IgE (SE Applequist, J Dahlstro¨m, B Heyman, manuscript in preparation).


IGE-MEDIATED ENHANCEMENT The immunoregulatory capacity of IgE antibodies was recognized relatively recently, and a limited number of reports have been published (18–20, 42, 141, 142). Monoclonal IgE anti-TNP, administered intravenously with TNP-BSA in physiological solutions to mice, up-regulates primary IgM, IgG (both IgG1 and IgG2a), and IgE anti-BSA (18, 19, 141). Priming with IgE/antigen also induces a more efficient immunological memory than priming with antigen alone, demonstrating that IgE/antigen does not only have temporary effects on the immune response (19). Only murine experimental systems have been used, and the response can be enhanced in all tested mouse strains except nude mice (19) and mice expressing MHC class II Ab (42). The lack of responsiveness in C57BL/6 mice, which carry Ab, could not be restored with transgenic responder Ak, suggesting that the molecule responsible for the low level of responsiveness was not the A molecule itself, but another gene product encoded within the A region (42). TNP- or DNP-specific monoclonal IgE enhances the primary carrier-specific IgG response to haptenated BSA, ovalbumin (OVA), tetanus, and diphtheria toxoid (18–20, 42, 141, 142). No enhancement of the response to TNP-SRBCs or TNP-KLH was observed (19). In the majority of studies, IgE and antigen are administered intravenously within 1 h of each other (18–20, 141). Intraperitoneal administration of antigen is equally efficient, whereas subcutaneous immunization induces enhancement only early in the response period (142). Maximal enhancement is seen when IgE and antigen are administered simultaneously, and IgE administered $24 h before or 24 h after the antigen has no effect (19). Usually, doses of $10 lg IgE/mouse are required (18). The primary response to $2 lg TNP-BSA can be dramatically augmented (2690-fold), and responses to 200 lg are still enhanced (50-fold) (42). IgE induces a rapid increase in specific IgG-producing B cells, with a peak occurring 6 days after primary immunization (142). The serum titers of carrierspecific IgG and IgE reach a peak within 7 days, and the IgG levels remain high for at least 7–8 weeks (18, 19, 142).

724

HEYMAN


FceRII and Antibody Responses Role of FceRII in the Antibody Response to IgE/Antigen Complexes The receptor involved in IgE-mediated enhancement has been defined unambiguously as FceRII, also known as CD23. The structural and functional aspects of this receptor have been reviewed previously (143–145, 173). IgE has no effect on antibody responses in mice pretreated with rat anti-FceRII mAb (18, 19) or in FceRII-deficient mice (20) (S Gustavsson, S Wernersson, B Heyman, submitted for publication). The complete abrogation of IgE-mediated enhancement in these animals excludes the involvement of FceRI, FccRIIB, or FccRIII, which are present in normal amounts in FceRII-deficient mice, and all three bind IgE/antigen complexes (53). The fact that IgE-mediated enhancement was normal in mice deficient in either FccRIIB or FcRc supports this conclusion (17). In addition, immunization with covalent conjugates of anti-FceRII and OVA led to higher production of OVA-specific IgG1 and IgE compared to immunization with OVA alone (146), indicating that targeting of the antigen to FceRII, and not the presence of specific IgE in the immune complex, is critical. Two isoforms of human FceRII, a and b, which differ only in their cytoplasmic regions, have been described. In mice, a homolog of FceRIIa is expressed constitutively on B cells (147) and FDCs (148). Whether other murine isoforms exist has been controversial, but mRNA for an isoform analogous to human FceRIIb has been detected in non-B spleen cells after stimulation with lipopolysaccharide and interleukin-4 (149). It is, however, uncertain whether this isoform is expressed at the protein level (150). To determine whether B cells or FDCs were involved in IgE-mediated enhancement, an adoptive transfer system was used (S Gustavsson, S Wernersson, B Heyman, J. Immunol., in press). The response to IgE/antigen was normal in mice with FceRII` B cells and FceRII1 FDCs (irradiated FceRIIdeficient mice reconstituted with FceRII` spleen or bone marrow cells). In contrast, no response to IgE/antigen was detected in mice with FceRII1 B cells and FceRII` FDCs (irradiated wild-type mice reconstituted with FceRII1 spleen or bone marrow cells). Based on the assumptions that FDCs are radio resistant and that host FDCs are not replaced by donor cells after the transfer of bone marrow (122), these results suggest that the B cell is the effector cell in FceRII-mediated responses to IgE/antigen.

Possible Mechanisms in IgE-Mediated Enhancement In summary, IgEmediated enhancement affects primary responses as well as induction of memory, is antigen specific, and requires FceRII` B cells and the presence of T cells (19). The majority of FceRII` murine spleen cells bind IgE/antigen complexes ex vivo, regardless of BCR specificity (85). Therefore, to explain the antigen specificity of IgE-mediated enhancement, a mechanism preventing nonspecific B cells from being activated must exist. The most straightforward model would be that FceRII and BCR must be co-crosslinked to start the chain of events leading to the dra-



725

matically enhanced antibody responses. The precise mechanism by which binding of IgE/antigen to FceRII leads to enhanced antibody production is not known. Increased uptake of antigen via FceRII and subsequent presentation to T cells is one possibility. FceRII` B cells are capable of endocytosing and presenting IgE/antigen complexes to T cells via FceRII in vitro (151–155). These studies are, however, difficult to compare with the in vivo studies because Epstein-Barr virus-transformed B cells or total spleen cells without specificity for the antigen were used. An alternative explanation for IgE-mediated enhancement would be that cocrosslinking of FceRII and BCR initiates a signaling cascade that directly activates the B cell. Low levels of co-crosslinking of surface immunoglobulin and FceRII on murine splenic B cells in vitro enhance proliferation and production of polyclonal IgM and IgG1, whereas high levels of co-crosslinking inhibit these responses (156). Administration of IgE/antigen to naı¨ve mice would be expected to induce low levels of crosslinking, because these B cells express only background levels of FceRII. Accordingly, interleukin-4–deficient mice, expressing levels of FceRII fivefold lower than wild-type animals, were still efficiently enhanced by IgE/antigen (141). Direct B-cell signaling would, however, not circumvent the need for T-cell help because no IgE-mediated enhancement takes place without T cells (19). There is no evidence that IgE/antigen would localize more efficiently in lymphoid follicles than would antigen alone, and IgE did not induce a response in chimeric mice with FceRII` FDCs and FceRII- B cells (S Gustavsson, S Wernersson, B Heyman, submitted for publication). However, both human and murine FceRIIs interact with adhesion molecules. Human FceRII is a receptor for CR2, CR3, and CR4 (157, 158), and murine FceRII interacts with CR3 (159) but not with CR2 (160). IgE-mediated enhancement is normal in CR1/2-deficient mice (SE Applequist, J Dahlstro¨m, H Molina, B Heyman, manuscript in preparation), demonstrating that CR1/2 does not play a significant role in IgE-mediated enhancement. Role of FceRII in the Antibody Responses to Noncomplexed Antigen In the studies described above, ligation of FceRII was achieved by IgE/antigen or antiFceRII/antigen complexes. The effects of FceRII on in vivo antibody responses have also been investigated in situations in which the ligand is unknown, that is, without administration of IgE. In two FceRII-deficient mouse lines, the effects of immunization with noncomplexed thymus-dependent antigens in alum were studied. Fujiwara et al (20) found similar antibody responses in FceRII-deficient and wild-type mice, whereas Yu et al (161) reported #12-fold-higher levels of antigen-specific and total serum IgE in FceRII-deficient mice. Overexpression of FceRII at the start of immunization, either in FceRII-transgenic mice (162, 163) or after treatment with an inhibitor of proteolytic processing (164), led to decreased production of IgE and IgG1 after immunization with OVA in alum. The possibility that a change in serum half-life of IgE explained the changes in IgE levels was excluded (161, 163). Finally, treatment of rats with polyclonal

726

HEYMAN


FceRII-specific rabbit antibodies inhibited the IgE anti-OVA response after immunization with OVA in pertussis toxin, suggesting an enhancing effect of FceRII (165). In summary, responses to alum-precipitated noncomplexed antigen are usually regulated negatively by FceRII, whereas responses to soluble IgE/antigen complexes are enhanced via FceRII. Although these two sets of data appear to conflict, it is more likely that different phenomena are being studied. Many questions remain to be answered regarding the down-regulatory role of FceRII. What is the nature of the FceRII ligand in these situations? Is the B cell the target cell or has the inflammatory response, caused by the adjuvant, induced expression of the putative FceRIIb on non-B cells (149)? In that case, other cells such as macrophages, able to produce inflammatory cytokines, may be involved. This scenario could more easily explain the regulatory effects of FceRII on polyclonal IgE production and on the development of arthritis (166, 167).

IGA-MEDIATED ENHANCEMENT Studies on the ability of IgA antibodies to regulate the antibody response are scarce. A monoclonal IgA anti-DNP antibody was as efficient as IgG purified from hyperimmune serum in inducing B memory cells (39). No effect on primary antibody responses was seen with the same IgA antibody (16).

PHYSIOLOGICAL ROLE OF ANTIBODY-MEDIATED REGULATION Epitope masking by antibodies most likely plays a physiological role in modulating the B-cell repertoire. Masking of epitopes that have already elicited a response would increase the chances of recognizing other, less immunogenic epitopes, thereby ensuring a broad spectrum of antibody specificities. By competing with B cells for epitopes, such masking would also increase the affinity of B cells (168). Complete suppression of antibody responses by epitope masking, mimicking the 99.9% suppression seen in experimental systems with IgG and heterologous erythrocytes, is not likely to take place in physiological situations. Negative regulation via FccRIIB inhibits antibody responses from extending to abnormal levels and may play a role in preventing autoimmunity. The most obvious potential physiological role of IgG-mediated enhancement is to aid in the generation of an efficient immunological memory. IgG resulting from a primary immunization and still remaining in the circulatory system at the time of the second encounter with antigen can be envisaged to capture small quantities of antigen and target them to antigen-presenting cells and FDCs, thereby causing a potent restimulation. IgG antibodies are present in high con-



727

centrations, often have high affinity, and are the only antibodies protected from catabolism by the IgG-specific receptor FcRn (72). These characteristics make IgG a suitable antibody for long-term immunoregulation. It is evident that the complement system via IgM plays a vital role in the generation of normal antibody responses. The most important physiological role may be in the immune response against bacterial infections, which has been shown to require IgM and complement (169, 170). Mice deficient in FceRII lead an apparently healthy life, although immune responses to infections [apart from those of Nippostrongylus brasiliensis, which was shown to induce normal antibody responses (20, 161, 171)], have not been investigated. Antigens inducing high local levels of IgE may use FceRII to further upregulate antibody production. An interesting idea is that allergen-specific IgE in atopic patients targets low doses of inhaled allergen to FceRII, resulting in increased production of allergen-specific IgE and IgG and perpetuation of the atopic disease (23). It may appear that administration to naı¨ve animals of preformed specific antibodies is an unnatural situation. However, a physiological equivalent is the transfer of IgG (and IgA) via placenta or milk from mother to offspring, which is likely to influence the immune response of the offspring in a way similar to that seen in the experimental models described above.

CONCLUDING REMARKS In summary, IgM, IgG, IgE, and IgA are able to enhance antibody responses to suboptimal doses of antigen when antigen and antibody are administered in close temporal relationship and can form a complex in the circulatory system. Enhancement by IgM is complement dependent and is primarily seen against particulate antigens whose cell membrane presumably provides the surface required for the conformational change of the IgM molecule that allows it to activate complement. It is known that IgE-mediated enhancement requires FceRII and IgG-mediated enhancement requires FccRs, but the molecular mechanisms behind the respective enhancing pathways are still unknown. Current data suggest that all antibody classes, provided that they have a sufficiently high affinity, can suppress the response to the epitope to which they bind. In experimental situations, antibody can also completely suppress the response against all determinants on the antigen. This requires that antibodies bind to the antigen with a high density, allowing them to sterically hinder B-cell recognition of responses to neighboring epitopes. Suppression is most commonly observed against erythrocytes. When both IgM and IgG are administered with erythrocytes, a competition between enhancing and suppressing activity will ensue, and with increasing doses of IgG, suppression will overtake enhancement (1). The exact role of FccRIIB in suppression of antibody responses is not known. IgG can induce .99% suppression of primary responses to erythrocytes in FccRIIB-

728

HEYMAN


deficient mice. However, such animals have augmented responses both to high doses of erythrocytes administered in noncomplexed form and complexes of IgG and soluble antigen. These data suggest that FccRIIB does not cause complete suppression but sets the upper limit for specific antibody levels. An interesting analogy is the FcRn-mediated protection of IgG from degradation, which sets the lower limit for IgG levels. The major challenge for future studies on antibody-mediated regulation of antibody responses will be to define the effector mechanisms used by the various Fc and complement receptors in vivo. A better understanding of these mechanisms will be important for the understanding of physiological as well as pathological immune responses. Antibody/antigen complexes play a central role in, for example, autoimmune and atopic diseases. The capacity of microgram amounts of syngeneic antibody, administered without adjuvants, to completely suppress or enhance specific antibody responses by several hundredfold has potential applications in future immunotherapy and vaccinations. ACKNOWLEDGMENTS I thank Imma Brogren, Erik Wiersma, Susanne Gustavsson, Susanna Chomez, Sandra Kleinau, Sara Wernersson, Mikael Karlsson, Steven Applequist, Jo¨rgen Dahlstro¨m, and Teresita Diaz de Sta˚hl for their important intellectual and technical contributions to this work over the years. I am also indebted to many of them, as well as to Drs. Daniel Conrad and Marc Shulman, for valuable feedback on the manuscript. Work in the author’s laboratory was supported by Ellen, Walter and Lennart Hesselman’s Foundation, Hans von Kantzow’s Foundation, King Gustaf V’s 80-Year Foundation, The Swedish Medical Research Council, and The Swedish Foundation for Health Care Sciences and Allergy Research. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Henry C, Jerne N. 1968. Competition of 19S and 7S antigen receptors in the regulation of the primary immune response. J. Exp. Med. 128:133–52 2. Bru¨ggemann M, Rajewsky K. 1982. Regulation of the antibody response against hapten-coupled erythrocytes by monoclonal anti-hapten antibodies of various isotypes. Cell. Immunol. 71:365– 73 3. Heyman B, Wigzell H. 1984. Immunoregulation by monoclonal sheep eryth-

rocyte specific IgG antibodies: Suppression is correlated to level of antigen binding and not to isotype. J. Immunol. 132:1136–43 4. Karlsson MCI, Wernersson S, Diaz de Sta˚hl T, Gustavsson S, Heyman B. 1999. Efficient IgG-mediated suppression of primary antibody responses in Fcc receptor-deficient mice. Proc. Natl. Acad. Sci. USA 96:2244–49 5. Clarke CA, Donohoe WTA, Woodrow JC, Finn R, Krevans JR, Kulke W,


6.

7.


8.

9.

10.

11.

12.

13.

14.

15.

Lehane D, Sheppard PM. 1963. Further experimental studies on the prevention of Rh haemolytic disease. Br. Med. J. 1:979–84 Bowman JM. 1988. The prevention of Rh immunization. Transfus. Med. Rev. 2:129–50 Heyman B, Pilstro¨m L, Shulman MJ. 1988. Complement activation is required for IgM-mediated enhancement of the antibody response. J. Exp. Med. 167:1999–2004 Ehrenstein MR, O’Keefe TL, Davies SL, Neuberger MS. 1998. Targeted gene disruption reveals a role for natural secretory IgM in the maturation of the primary immune response. Proc. Natl. Acad. Sci. USA 95:10089–93 Boes M, Esau C, Fischer MB, Schmidt T, Carroll M, Chen J. 1998. Enhanced B1 cell development, but impaired IgG antibody responses in mice deficient in secreted IgM. J. Immunol. 160:4776–87 Bitter-Suermann D, Burger R. 1989. C3 deficiencies. Curr. Top. Microbiol. Immunol. 153:223–33 Heyman B. 1994. The role of complement receptors in the regulation of the immune response. In New Aspects of Complement Structure and Function, ed. A Erdei, pp. 59–69. Austin, TX: Landes Carroll MC. 1998. The role of complement and complement receptors in induction and regulation of immunity. Annu. Rev. Immunol. 16:545–68 Dempsey PW, Allison MED, Akkaraju S, Goodnow CC, Fearon DT. 1996. C3d of complement as a molecular adjuvant: bridging innate and acquired immunity. Science 271:348–50 Villiers M-B, Villiers CL, Laharie A-M, Marche PN. 1999. Amplification of the antibody response by C3b complexed to antigen through an ester link. J. Immunol. 162:3647–52 Baiu DC, Prechl J, Tchorbanov A, Molina HD, Erdei A, Sulica A, Capel PJA, Hazenbos WLW. 1999. Modulation

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

729

of the humoral immune response by antibody-mediated antigen targeting to complement receptors and Fc receptors. J. Immunol. 162:3125–30 Coulie P, Van Snick J. 1985. Enhancement of IgG anti-carrier responses by IgG2-anti-hapten antibodies in mice. Eur. J. Immunol. 15:793–98 Wernersson S, Karlsson M, Dahlstro¨m J, Mattsson R, Verbeek JS, Heyman B. 1999. IgG-mediated enhancement of Ab responses is low in FcRc chain deficient mice and increased in FccRII deficient mice. J. Immunol. 163:618–22 Heyman B, Liu T, Gustavsson S. 1993. In vivo enhancement of the specific antibody response via the low affinity receptor for IgE. Eur. J. Immunol. 23:1739–42 Gustavsson S, Hjulstro¨m S, Liu T, Heyman B. 1994. CD23/IgE-mediated regulation of the specific antibody response in vivo. J. Immunol. 152:4793–800 Fujiwara H, Kikutani H, Suematsu S, Naka T, Yoshida K, Yoshida K, Tanaka T, Suemura M, Matsumoto N, Kojima S, Kishimoto T, Yoshida N. 1994. The absence of IgE antibody-mediated augmentation of immune responses in CD23-deficient mice. Proc. Natl. Acad. Sci. USA 91:6835–39 Gessner JE, Heiken H, Tamm A, Schmidt RE. 1998. The IgG Fc receptor family. Ann. Hematol. 76:231–48 Mossalayi MD, Arock M, Debre P. 1997. CD23/FceRII: signaling and clinical implication. Int. Rev. Immunol. 16:129– 46 Mudde GC, Bheekha R, BruijnzeelKoomen CAFM. 1995. Consequences of IgE/CD23-mediated antigen presentation in allergy. Immunol. Today 16:380–83 Phillips NE, Parker DC. 1984. Crosslinking of B lymphocyte Fcc receptors and membrane immunoglobulin inhibits antiimmunoglobulin-induced blastogenesis. J. Immunol. 132:627–32 Bijsterbosch MK, Klaus GGB. 1985. Crosslinking of surface immunoglobulin

730

26.

27.


28.

29.

30.

30a.

31.

32.

33.

34.

HEYMAN and Fc receptors on B lymphocytes inhibits stimulation of inositol phospholipid breakdown via the antigen receptors. J. Exp. Med. 162:1825–27 Amigorena S, Bonnerot C, Drake JR, Choquet D, Hunziker W, Guillet J-G, Webster P, Sautes C, Mellman I, Fridman WH. 1992. Cytoplasmic domain heterogeneity and functions of IgG Fc receptors in B lymphocytes. Science 256:1808–12 Muta T, Kurosaki T, Misulovin Z, Sanchez M, Nussenzweig MC, Ravetch JV. 1994. A 13-amino-acid motif in the cytoplasmic domain of FccRIIB modulates B-cell receptor signalling. Nature 368:70–73 Takai T, Ono M, Hikida M, Ohmori H, Ravetch JV. 1996. Augmented humoral and anaphylactic responses in FccRIIdeficient mice. Nature 379:346–49 Uhr JW, Mo¨ller G. 1968. Regulatory effect of antibody on the immune response. Adv. Immunol. 8:81–127 Wiersma EJ. 1991. Effects of IgG antibodies and complement on the immune response. PhD thesis. Uppsala Univ., Uppsala, Sweden. 54 pp. Heyman B. 1996. Complement and Fc receptors in regulation of the antibody response. Immunol. Lett. 54:195–99 Dennert G. 1973. Effects of IgM on the in vivo and in vitro immune response. Proc. Soc. Exp. Biol. Med. 143:889–93 Wason WM. 1973. Regulation of the immune response with antigen specific IgM antibody: a dual role. J. Immunol. 110:1245–52 Powell R, Hutchings P, Cooke A, Lydyard PM. 1982. Antibody mediated regulation of immune responses. I. Enhancement of specific antibody responses through IgM antibodies. Immunol. Lett. 4:253–58 Wiersma EJ. 1992. Enhancement of the antibody response to protein antigens by specific IgG under different experimental conditions. Scand. J. Immunol. 36:193– 200

35. Pollack W, Gorman JG, Freda VJ, Ascari WQ, Allen AE, Baker WJ. 1968. Results of clinical trials of RhoGAM in women. Transfusion 8:151–53 36. Chan PL, Sinclair NR. 1973. Regulation of the immune response. VI. Inability of F(ab’)2 antibody to terminate established immune responses and its ability to interfere with IgG antibody-mediated immunosuppression. Immunology 24:289–310 37. Whited Collisson E, Andersson B, Ro¨nnholm M, Lamon E. 1983. Potentiation of antibody responses by specific IgM: specificity and thymus dependency. Cell. Immunol. 79:44–55 38. Heyman B, Andrighetto S, Wigzell H. 1982. Antigen dependent IgM-mediated enhancement of the sheep erythrocyte response in mice: evidence for induction of B cells with specificities other than that of the injected antibodies. J. Exp. Med. 155:994–1009 39. Klaus GGB. 1979. Generation of memory cells. III. Antibody class requirements for the generation of B-memory cells by antigen-antibody complexes. Immunology 37:345–51 40. Pearlman DS. 1967. The influence of antibodies on immunologic responses: the effect on the response to particulate antigen in the rabbit. J. Exp. Med. 126:127–48 41. Wiersma EJ, Nose M, Heyman B. 1990. Evidence of IgG-mediated enhancement of the antibody response in mice without classical pathway complement activation. Eur. J. Immunol. 20:2585–89 42. Gustavsson S, Hjulstro¨m-Chomez S, Lidstro¨m B-M, Ahlborg N, Andersson R, Heyman B. 1998. Impaired antibody responses in H-2Ab mice. J. Immunol. 161:1765–71 43. Greenbury CL, Moore DH. 1968. Nonspecific antibody-induced suppression of the immune response. Nature 219:526– 27 44. Whited Collisson E, Andersson B, Lamon EW. 1979. Modulation of hapten-


45.

46.


47.

48.

49.

50.

51.

52.

specific responses with anticarrier antibody. I. Differential effects of IgM and IgG anticarrier on primary direct and indirect hapten-specific plaque-forming cells. Proc. Soc. Exp. Biol. Med. 162:194–98 Enriques-Rincon F, Klaus GGB. 1984. Differing effects of monoclonal antihapten antibodies on humoral responses to soluble or particulate antigens. Immunology 52:129–36 Whited Collisson E, Andersson B, Lamon EW. 1984. Avidities of haptenspecific antibodies when the responses are modulated by anti-carrier antibodies. Immunology 53:443–49 Wiersma EJ, Coulie PG, Heyman B. 1989. Dual immunoregulatory effects of monoclonal IgG-antibodies: suppression and enhancement of the antibody response. Scand. J. Immunol. 29:439–48 Brody NI, Walker JG, Siskind GW. 1967. Studies on the control of antibody synthesis: interaction of antigenic competition and suppression of antibody formation by passive antibody on the immune response. J. Exp. Med. 126:81– 91 Mo¨ller G. 1985. Antibody-mediated suppression of the immune response is determinant specific. Eur. J. Immunol. 15:409–12 Quintana IZ, Silveira AV, Mo¨ller G. 1987. Regulation of the antibody response to sheep erythrocytes by monoclonal IgG antibodies. Eur. J. Immunol. 17:1343–49 Mo¨ller G, Wigzell H. 1965. Antibody synthesis at the cellular level: antibodyinduced suppression of 19S and 7S antibody responses. J. Exp. Med. 121:969– 89 Schrader JW. 1973. Regulation of the immune response by IgM antibody: a paradoxical suppression of the in vitro primary immune response to sheep erythrocytes by IgM. Aust. J. Exp. Biol. Med. Sci. 51:333–46

731

53. Takizawa F, Adamczewski M, Kinet J-P. 1992. Identification of the low affinity receptor for immunoglobulin E on mouse mast cells and macrophages as FccRII and FccRIII. J. Exp. Med. 176:469–76 54. Takai T, Li M, Sylvestre D, Clynes R, Ravetch JV. 1994. FcRc chain deletion results in pleiotrophic effector cell defects. Cell 76:519–29 55. Heyman B, Wigzell H. 1985. IgM enhances and IgG suppresses immunological memory in mice. Scand. J. Immunol. 21:255–66 56. Bystryn J-C, Graf MW, Uhr JW. 1970. Regulation of antibody formation by serum antibody. II. Removal of specific antibody by means of exchange transfusion. J. Exp. Med. 132:1279–87 57. Kappler JW, Hoffman M, Dutton RW. 1971. Regulation of the immune response. I. Differential effect of passively administered antibody on the thymus-derived and bone marrow-derived lymphocytes. J. Exp. Med. 134:577–87 58. Sinclair NR. 1969. Regulation of the immune response. I. Reduction in ability of specific antibody to inhibit longlasting IgG immunological priming after removal of the Fc fragment. J. Exp. Med. 129:1183–201 59. Dresser DW. 1990. Feedback by early and late primary antisera on the primary and secondary adoptive immune responses of mice to burro erythrocytes. Cell. Immunol. 127:405–19 60. Cerottini JC, McConahey PJ, Dixon FJ. 1969. The immunosuppressive effect of passively administered antibody IgG fragments. J. Immunol. 102:1008–15 ¨ , Belin L. 1970. Suppres61. Strannega˚rd O sion of reagin synthesis in rabbits by passively administered antibody. Immunology 20:775–85 62. Krieger N, Pesce A, Michael JG. 1983. Immunoregulation of the anti-bovine serum albumin response by polyclonal and monoclonal antibodies. Cell. Immunol. 80:279–87


732

HEYMAN

63. Walker JG, Siskind GW. 1968. Studies on the control of antibody synthesis: effect of antibody affinity upon its ability to suppress antibody formation. Immunology 14:21–28 64. Tao TW, Uhr JW. 1966. Capacity of pepsin-digested antibody to inhibit antibody formation. Nature 212:208–9 65. Sinclair NR, Lees RK, Elliott EV. 1968. Role of the Fc fragment in the regulation of the primary immune response. Nature 220:1048–49 66. Dae¨ron M. 1997. Fc receptor biology. Annu. Rev. Immunol. 15:203–34 67. Heyman B. 1990. Fc-dependent IgGmediated suppression of the antibody response: fact or artefact? Scand. J. Immunol. 31:601–7 68. Heyman B, Wiersma E, Nose M. 1988. Complement activation is not required for IgG-mediated suppression of the antibody response. Eur. J. Immunol. 18: 1739–43 69. Heyman B, Nose M, Weigle WO. 1985. Carbohydrate chains on IgG2b: a requirement for efficient feedback immunosuppression. J. Immunol. 134:4018–23 70. Heyman B. 1989. Inhibition of IgGmediated immunosuppression by a monoclonal anti-Fc-receptor antibody. Scand. J. Immunol. 29:121–26 71. Heyman B. 1999. Antibody feedback suppression: towards a unifying concept? Immunol. Lett. 68:41–45 72. Junghans RP, Anderson CL. 1996. The protection receptor for IgG catabolism is the beta-2-microglobulin-containing neonatal intestinal transport receptor. Proc. Natl. Acad. Sci. USA 93:5512–16 73. Amigorena S, Bonnerot C, Choquet D, Fridman WF, Teillaud J-L. 1989. FccRII expression in resting and activated B lymphocytes. Eur. J. Immunol. 19:1379– 85 74. Dae¨ron M, Latour S, Malbec O, Espinosa E, Pina P, Pasmans S, Fridman WH. 1995. The same tyrosine-based inhibition motif, in the intracytoplasmic domain of

75.

76.

77.

78.

79.

80.

81.

82.

83.

FccRIIB, regulates negatively BCR-, TCR-, and FcR-dependent cell activation. Immunity 3:635–46 Dae¨ron M, Malbec O, Latour S, Arock M, Fridman WH. 1995. Regulation of high-affinity IgE receptor-mediated mast cell activation by murine low-affinity IgG receptors. J. Clin. Invest. 95:577–85 Minskoff SA, Matter K, Mellman I. 1998. FccRII-B1 regulates the presentation of B cell receptor-bound antigens. J. Immunol. 161:2079–83 Wagle NM, Faassen AE, Kim JH, Pierce SK. 1999. Regulation of B cell receptormediated MHC class II antigen processing by FccRIIB1. J. Immunol. 162:2732– 40 Hunter S, Indik ZK, Kim MK, Cauley MD, Park JG, Schreiber AD. 1998. Inhibition of Fcc receptor-mediated phagocytosis by a nonphagocytic Fcc receptor. Blood 91:1762–68 Ujike A, Ishikawa Y, Ono M, Yuasa T, Yoshino T, Fukumoto M, Ravetch JV, Takai T. 1999. Modulation of immunoglobulin (Ig)E-mediated systemic anaphylaxis by low-affinity Fc receptors for IgG. J. Exp. Med. 189:1573–79 Clynes R, Maizes JS, Guinamard R, Ono M, Takai T, Ravetch JV. 1999. Modulation of immune complex-induced inflammation in vivo by the coordinate expression of activation and inhibitory Fc receptors. J. Exp. Med. 189:179–85 Yuasa T. 1999. Deletion of Fcc receptor IIB renders H-2b mice susceptible to collagen-induced arthritis. J. Exp. Med. 189:187–94 Stoner RD, Terres G. 1963. Enhanced antitoxin responses in irradiated mice elicited by complexes of tetanus toxoid and specific antibody. J. Immunol. 91:761–70 Stoner RD, Terres G, Hess MW. 1975. Early and enhanced antitoxin responses elicited with complexes of tetanus toxoid and specific mouse and human antibodies. J. Infect. Dis. 1975:230–38


ANTIBODY FEEDBACK REGULATION 84. Klaus GGB. 1978. The generation of memory cells. II. Generation of B memory cells with preformed antigen-antibody complexes. Immunology 34:643– 52 85. Gustavsson S, Chomez S, Heyman B. 1999. Low responsiveness to immunization with immunoglobulin E/antigen and immunoglobulin G/antigen complexes in H-2Ab mice. Scand. J. Immunol. 50:45–51 86. Terres G, Habicht GS, Stoner RD. 1974. Carrier-specific enhancement of the immune response using antigen-antibody complexes. J. Immunol. 112:804–11 86a. Nie X, Basu S, Cerny J. 1997. Immunization with immune complex alters the repertoire of antigen-reactive B cells in the germinal centers. Eur. J. Immunol. 27:3517–25 87. Kunkl A, Klaus GGB. 1981. The generation of memory cells. IV. Immunization with antigen-antibody complexes accelerates the development of B-memory cells, the formation of germinal centres and the maturation of antibody affinity in the secondary responses. Immunology 43:371–77 88. Hamaoka T, Kitagawa M. 1971. Antibody production in mice. II. The mechanism of antigenic stimulation in the secondary immune response. Immunology 20:191–203 89. Safford JJW, Tokuda S. 1971. Antibodymediated suppression of the immune response: effect on the development of immunologic memory. J. Immunol. 107:1213–25 90. Enriquez-Rincon F, Klaus GGB. 1984. Follicular trapping of hapten-erythrocyte-antibody complexes in mouse spleen. Immunology 52:107–16 91. Dennert G. 1971. The mechanism of antibody-induced stimulation and inhibition of the immune response. J. Immunol. 106:951–55 92. Murgita RA, Vas SI. 1972. Specific antibody-mediated effect on the immune

93.

94.

95.

96.

97.

98.

99.

100.

733

response: suppression and augmentation of the primary immune response in mice by different classes of antibodies. Immunology 22:319–31 Terres G, Wolins W. 1961. Enhanced immunological sensitization of mice by the simultaneous injection of antigen and specific antiserum. J. Immunol. 86:361– 68 Terres G, Stoner RD. 1962. Specificity of enhanced immunological sensitization of mice following injections of antigens and specific antisera. Proc. Soc. Exp. Biol. Med. 109:88–91 Papamichail M, Gutierrez C, Embling P, Johnson P, Holborow EJ, Pepys MB. 1975. Complement dependence of localisation of aggregated IgG in germinal centers. Scand. J. Immunol. 4:343–47 Gavin AL, Barnes N, Dijstelbloem HM, Hogarth PM. 1998. Identification of the mouse IgG3 receptor: implications for antibody effector function at the interface between innate and adaptive immunity. J. Immunol. 160:20–23 Hulett MD, Hogarth PM. 1994. Molecular basis of Fc receptor function. Adv. Immunol. 57:1–127 Regnault A, Lankar D, Lacabanne V, Rodriguez A, Thery C, Rescigno M, Saito T, Verbeek S, Bonnerot C, Ricciardi-Castagnoli P, Amigorena S. 1999. Fcc-receptor-mediated induction of dendritic cell maturation and major histocompatibility complex class I-restricted antigen presentation after immune complex internalization. J. Exp. Med. 189:371–80 Ukkonen P, Lewis V, Marsh M, Helenius A, Mellman I. 1986. Transport of macrophage Fc receptors and Fc receptorbound ligands to lysosomes. J. Exp. Med. 163:952–71 Manca F, Fenoglio D, Li Pira G, Kunkl S, Celada F. 1991. Effect of antigen/antibody ratio on macrophage uptake, processing, and presentation to T cells of

734

101.

102.


103.

104.

105.

106.

107.

108.

109.

HEYMAN antigen complexed with polyclonal antibodies. J. Exp. Med. 173:37–48 Sallusto F, Lanzavecchia A. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor a. J. Exp. Med. 179:1109–18 Heyman B, Hobbs M, Weigle WO. 1985. IgM-mediated enhancement of in vivo anti-sheep erythrocyte antibody responses: isotype analysis of the enhanced responses. Cell. Immunol. 92: 134–41 ¨ , Belin L. 1971. EnhanceStrannega˚rd O ment of reagin formation in rabbits by passively administered 19S antibody. Immunology 20:427–31 Harte PG, Cooke A, Playfair JHL. 1983. Specific monoclonal IgM is a potent adjuvant in murine malaria vaccination. Nature 302:256–58 Coutinho A, Forni L. 1981. The enhancement of antibody response by IgM antibodies is dependent on antigen-specific T helper cells. Immunobiology 158:182– 90 Lehner P, Hutchings P, Lydyard PM, Cooke A. 1983. Regulation of the immune response by antibody. II. IgMmediated enhancement: dependency on antigen dose, T-cell requirement and lack of evidence for an idiotype-related mechanism. Immunology 50:503–9 Heyman B, Gustavsson S, Kusakari C, Nose M, Kinoshita T, Tarkowski A, Klareskog L. 1993. Antibody feedback regulation in MRL/lpr mice. J. Autoimmun. 6:437–48 Beale D, Feinstein A. 1976. Structure and function of the constant regions of immunoglobulins. Q. Rev. Biophys. 9: 135–80 Karush F, Chua M-M, Rodwell JD. 1979. Interaction of a bivalent ligand with IgM anti-lactose antibody. Biochemistry 18: 2226–32

110. Wright JK, Tschopp J, Jaton J-C, Engel J. 1980. Dimeric, trimeric and tetrameric complexes of immunoglobulin G fix complement. Biochem. J. 187:775–80 111. Borsos T, Rapp HJ. 1965. Complement fixation on cell surfaces by 19S and 7S antibodies. Science 150:505–6 112. Jerne NK, Nordin AA. 1963. Plaque formation in agar by single antibodyproducing cells. Science 140:405 113. Pepys MB. 1974. Role of complement in induction of antibody production in vivo: effect of cobra factor and other C3reactive agents on thymus-dependent and thymus-independent antibody responses. J. Exp. Med. 140:126–45 114. Bitter-Suermann D, Hoffmann T, Burger R, Hadding U. 1981. Linkage of total deficiency of the second component of the complement system and of genetic C2-polymorphism to the major histocompatibility complex of the guinea pig. J. Immunol. 127:608–12 115. Ochs HD, Wedgwood RJ, Frank MM, Heller SR, Hosea SW. 1983. The role of complement in the induction of antibody responses. Clin. Exp. Immunol. 53:208– 16 116. Bo¨ttger EC, Hoffmann T, Hadding U, Bitter-Suermann D. 1985. Influence of genetically inherited complement deficiencies on humoral immune response in guinea pigs. J. Immunol. 135:4100–7 117. Jackson CG, Ochs HD, Wedgwood RJ. 1979. Immune response of a patient with deficiency of the fourth component of complement and systemic lupus erythematosus. N. Engl. J. Med. 300:1124– 29 118. OŃeil KM, Ochs SR, Heller SR, Cork LC, Morris JM, Winkelstein JA. 1988. Role of C3 in humoral immunity: defective antibody production in C3-deficient dogs. J. Immunol. 140:1939–45 119. Fischer MB, Ma M, Goerg S, Zhou X, Xia J, Finco O, Han S, Kelsoe G, Howard RG, Rothstein TL, Kremmer E, Rosen FS, Carroll MC. 1996. Regulation of the


120.

121.


122.

123.

124.

125.

126.

127.

B cell response to T-dependent Ags by classical pathway complement. J. Immunol. 157:549–56 Matsumoto M, Fukuda W, Circolo A, Goellner J, Strauss-Schoenberger J, Wang X, Fujita S, Hidvegi T, Chaplin DD, Colten HR. 1997. Abrogation of the alternative complement pathway by targeted deletion of murine factor B. Proc. Natl. Acad. Sci. USA 94:8720–25 Turner MW. 1996. Mannose-binding lectin: the pluripotent molecule of the innate immune system. Immunol. Today 17: 532–40 Ahearn JM, Fischer MB, Croix D, Goerg S, Ma M, Xia J, Zhou X, Howard RG, Rothstein TL, Carroll MC. 1996. Disruption of the Cr2 locus results in a reduction in B-1a cells and in an impaired B cell response to T-dependent antigen. Immunity 4:251–62 Croix DA, Ahearn JM, Rosengard AM, Han S, Kelsoe G, Ma M, Carroll MC. 1996. Antibody response to a T-dependent antigen requires B cell expression of complement receptors. J. Exp. Med. 183:1857–64 Molina H, Holers VM, Li B, Fang Y-F, Mariathasan S, Goellner J, StraussSchoenberger J, Karr RW, Chaplin DD. 1996. Markedly impaired humoral immune responses in mice deficient in complement receptors 1 and 2. Proc. Natl. Acad. Sci. USA 93:3357–61 Heyman B, Wiersma EJ, Kinoshita T. 1990. In vivo inhibition of the antibody response by a monoclonal complement receptor specific antibody. J. Exp. Med. 172: 665–68 Wiersma EJ, Kinoshita T, Heyman B. 1991. Inhibition of immunological memory and T-independent humoral responses by monoclonal antibodies specific for murine complement receptors. Eur. J. Immunol. 21:2501–6 Thyphronitis G, Kinoshita T, Inoue K, Schweinle JE, Tsokos GC, Metcalf ES, Finkelman FD, Balow JE. 1991. Modu-

128.

129.

130.

131.

132.

133.

134.

135.

136.

735

lation of mouse complement receptors 1 and 2 suppresses antibody responses in vivo. J. Immunol. 147:224–30 Hebell T, Ahearn JM, Fearon DT. 1991. Suppression of the immune response by a soluble complement receptor of B lymphocytes. Science 254:102–5 Kurtz CB, O’Toole E, Christensen SM, Weis JH. 1990. The murine complement receptor gene family. IV. Alternative splicing of Cr2 gene transcripts predicts two distinct gene products that share homologous domains with both human CR2 and CR1. J. Immunol. 144:3581–91 Fearon DT, Carter RH. 1995. The CD19/ CR2/TAPA-1 complex of B lymphocytes: linking natural to acquired immunity. Annu. Rev. Immunol. 13:127– 49 Fang Y, Xu C, Fu Y-X, Holers VM, Molina H. 1998. Expression of complement receptors 1 and 2 on follicular dendritic cells is necessary for the generation of a strong antigen-specific IgG response. J. Immunol. 160:5273–79 Carter RH, Fearon DT. 1992. CD19: lowering the threshold for antigen receptor stimulation of B lymphocytes. Science 256:105–7 Thornton BP, Vetvicka V, Ross GD. 1994. Natural antibody and complementmediated antigen processing and presentation by B-lymphocytes. J. Immunol. 152:1727–37 Boackle SA, Morris MA, Holers VM. 1998. Complement opsonization is required for presentation of immune complexes by resting peripheral blood B cells. J. Immunol. 161:6537–43 Gustavsson S, Kinoshita T, Heyman B. 1995. Antibodies to murine complement receptor 1 and 2 can inhibit the antibody response in vivo without inhibiting Thelper cell induction. J. Immunol. 154:6524–28 Davies KA, Erlendsson K, Beynon HLC, Peters AM, Steinsson K, Valdimarsson H, Walport MJ. 1993. Splenic uptake of

736

137.

138.


139.

140.

141.

142.

143.

144. 145.

HEYMAN immune complexes in man is complement-dependent. J. Immunol. 151:3866– 73 Qin D, Wu J, Carroll MC, Burton GF, Szakal AK, Tew JG. 1998. Evidence for an important interaction between a complement-derived CD21 ligand on follicular dendritic cells and CD21 on B cells in the initiation of IgG responses. J. Immunol. 161:4549–54 Ochs HD, Nonoyama S, Zhu Q, Farrington M, Wedgwood RJ. 1993. Regulation of antibody responses: the role of complement and adhesion molecules. Clin. Immunol. Immunopathol. 67:S33–40 Martinelli GP, Matsuda T, Osler AG. 1978. Studies of immunosuppression by cobra venom factor. I. On early IgG and IgM responses to sheep erythrocytes and DNP-protein conjugates. J. Immunol. 121:2043–47 Bo¨ttger EC, Hoffmann T, Metzger S, Hadding U, Bitter-Suermann D. 1986. The role and mechanism of cobra venom factor-induced suppression of the humoral immune response in guinea pigs. J. Immunol. 137:1280–85 Hjulstro¨m S, Landin A, Jansson L, Holmdahl R, Heyman B. 1995. No role of interleukin-4 in CD23/IgE-mediated enhancement of the murine antibody response in vivo. Eur. J. Immunol. 25:1469–72 Westman S, Gustavsson S, Heyman B. 1997. Early expansion of secondary B cells after primary immunization with antigen complexed with IgE. Scand. J. Immunol. 46:10–15 Delespesse G, Sarfati M, Wu CY, Fournier S, Letellier M. 1992. The lowaffinity receptor for IgE. Immunol. Rev. 125:77–97 Sutton BJ, Gould HJ. 1993. The human IgE network. Nature 366:421–28 Bonnefoy JY, Lecoanet-Henchoz S, Gauchat JF, Graber P, Aubry JP, Jeannin P, Plater-Zyberk C. 1997. Structure and

146.

147.

148.

149.

150.

151.

152.

153.

154.

functions of CD23. Int. Rev. Immunol. 16:113–28 Squire CM, Studer EJ, Lees A, Finkelman FD, Conrad DH. 1994. Antigen presentation is enhanced by targeting antigen to the FceRII by antigen-antiFceRII conjugates. J. Immunol. 152: 4388–96 Rao M, Lee WT, Conrad DH. 1987. Characterization of a monoclonal antibody directed against the murine B lymphocyte receptor for IgE. J. Immunology 138:1845–51 Maeda K, Burton GF, Padgett DA, Conrad DH, Huff TF, Masuda A, Szakal AK, Tew JG. 1992. Murine follicular dendritic cells and low affinity Fc receptors for IgE (FceRII). J. Immunol. 148:2340– 47 Kondo H, Ichikawa Y, Nakamura K, Tsuchiya S. 1994. Cloning of cDNAs for new subtypes of murine low-affinity Fc receptor for IgE (FceRII/CD23). Int. Arch. Allergy Immunol. 105:38–48 Conrad DH, Kozak CA, Vernachio J, Squire CM, Rao M, Eicher EM. 1993. Chromosomal location and isoform analysis of mouse FceRII/CD23. Mol. Immunol. 30:27–33 Kehry MR, Yamashita LC. 1989. Lowaffinity IgE receptor (CD23) function on mouse B cells: role in IgE-dependent antigen focusing. Proc. Natl. Acad. Sci. USA 86:7556–60 Pirron U, Schlunck T, Prinz JC, Rieber EP. 1990. IgE-dependent antigen focusing by human B lymphocytes is mediated by low-affinity receptor for IgE. Eur. J. Immunol. 20:1547–51 Santamaria LF, Bheekha R, van Reijsen FC, Perez Soler MT, Suter M, Bruijnzeel-Koomen CAFM, Mudde GC. 1993. Antigen focusing by specific monomeric immunoglobulin E bound to CD23 on Epstein-Barr virus-transformed B cells. Hum. Immunol. 37:23–30 van der Heijden FL, van Neerven RJJ, van Katwijk M, Bos JD, Kapsenberg


155.

156.


157.

158.

159.

160.

161.

162.

163.

ML. 1993. Serum-IgE-facilitated allergen presentation in atopic disease. J. Immunol. 150:1–8 Bheekha Escura R, Wasserbauer E, Hammerschmid F, Pearce A, Kidd P, Mudde GC. 1995. Regulation and targeting of Tcell immune responses by IgE and IgG antibodies. Immunology 86:343–50 Campbell KA, Lees A, Finkelman FD, Conrad DH. 1992. Co-crosslinking FceRII/CD23 and B cell surface immunoglobulin modulates B cell activation. Eur. J. Immunol. 22:2107–12 Aubry J-P, Pochon S, Graber P, Jansen KU, Bonnefoy J-Y. 1992. CD21 is a ligand for CD23 and regulates IgE production. Nature 358:505–7 Lecoanet-Henchoz S, Gauchat J-F, Aubry J-P, Graber P, Life P, Paul-Eugene N, Ferrua B, Corbi AL, Dugas B, PlaterZyberk C, Bonnefoy J-Y. 1995. CD23 regulates monocyte activation through a novel interaction with adhesion molecules CD11b-CD18 and CD11c-CD18. Immunity 3:119–25 Lecoanet-Henchoz S, Plater-Zyberk C, Graber P, Gretener D, Aubry J-P, Conrad D, Bonnefoy J-Y. 1997. Mouse CD23 regulates monocyte activation through an interaction with the adhesion molecule CD11b/CD18. Eur. J. Immunol. 27:2290–94 Davey EJ, Bartlett WC, Kikutani H, Fujiwara H, Kishimoto T, Conrad DH, Severinson E. 1995. Homotypic aggregation of murine B lymphocytes is independent of CD23. Eur. J. Immunol. 25:1224–29 Yu P, Kosco-Vilbois M, Richards M, Ko¨hler G, Lamers MC. 1994. Negative feedback regulation of IgE synthesis by murine CD23. Nature 369:753–56 Texido G, Eibel H, Le Gros G, van der Putten H. 1994. Transgene CD23 expression on lymphoid cells modulates IgE and IgG1 responses. J. Immunol. 153: 3028–42 Payet ME, Woodward EC, Conrad DH. 1999. Humoral response suppression

164.

165.

166.

167.

168.

169.

170.

171.

737

observed with CD23 transgenics. J. Immunol. 163:217–23 Christie G, Barton A, Bolognese B, Buckle DR, Cook RM, Hansbury MJ, Harper GP, Marshall LA, McCord ME, Moulder K, Murdock PR, Seal SM, Spackman VM, Weston BJ, Mayer RJ. 1997. IgE secretion is attenuated by an inhibitor of proteolytic processing of CD23 (FceRII). Eur. J. Immunol. 27:3228–35 Flores-Romo L, Shields J, Humbert Y, Graber P, Aubry J-P, Gauchat J-F, Ayala G, Allet B, Chavez M, Bazin H, Capron M, Bonnefoy J-Y. 1993. Inhibition of an in vivo antigen-specific IgE response by antibodies to CD23. Science 261:1038– 41 Plater-Zyberk C, Bonnefoy J-Y. 1995. Marked amelioration of established collagen-induced arthritis by treatment with antibodies to CD23 in vivo. Nat. Med. 1:781–85 Kleinau S, Martinsson P, Gustavsson S, Heyman B. 1999. Importance of CD23 for collagen-induced arthritis: delayed onset and reduced severity in CD23-deficient mice. J. Immunol. 162:4266–70 Siskind GW, Dunn P, Walker JG. 1968. Studies on the control of antibody synthesis. II. Effect of antigen dose and of suppression by passive antibody on the affinity of antibody synthesized. J. Exp. Med. 127:55–66 Shigeoka AO, Jensen CL, Pincus SH, Hills HR. 1984. Absolute requirement for complement in monoclonal IgM antibody-mediated protection against experimental infection with type III group B streptococci. J. Infect. Dis. 150:63–70 Wessels MR, Butko P, Ma M, Warren HB, Lage AL, Carroll MC. 1995. Studies of group B streptococcal infection in mice deficient in complement component C3 and C4 demonstrate an essential role for complement in both innate and acquired immunity. Proc. Natl. Acad. Sci. USA 92:11490–94 Stief A, Texido G, Sansig G, Eibel H, Le

738

HEYMAN


Gros G, van der Putten H. 1994. Mice deficient in CD23 reveal its modulatory role in IgE production but no role in T and B cell development. J. Immunol. 152:3378–90 172. Carter RH, Spycher MO, Ng YC, Hoffmann R, Fearon DT. 1988. Synergistic interaction between complement recep-

tor type 2 and membrane IgM on B-lymphocytes. J. Immunol. 141:457–63 173. Gustavson S, Wernersson S, Hayman B. 2000. Restoration of the antibody response to IgE/antigen complexes in CD23-deficient mice by CD23` spleen or bone marrow cells. J. Immunol. In press


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


MULTIPLE ROLES FOR THE MAJOR HISTOCOMPATIBILITY COMPLEX CLASS I– RELATED RECEPTOR FCRN Victor Ghetie and E. Sally Ward


Center for Immunology and Cancer Immunobiology Center, University of Texas Southwestern Medical Center, Dallas, Texas 75235–8576; e-mail: [email protected]

Key Words neonatal Fc receptor, gammaglobulin, transcytosis, serum half-life, maternofetal/intestinal transfer of IgGs Abstract Multiple functions have recently been identified for the neonatal Fc receptor FcRn. In addition, a human homolog of the rodent forms of FcRn has been identified and characterized. This major histocompatibility complex class I-related receptor plays a role in the passive delivery of immunoglobulin (Ig)Gs from mother to young and the regulation of serum IgG levels. In addition, FcRn expression in tissues such as liver, mammary gland, and adult intestine suggests that it may modulate IgG transport at these sites. These diverse functions are apparently brought about by the ability of FcRn to bind IgGs and transport them within and across cells. However, the molecular details as to how FcRn traffics within cells have yet to be fully understood, although in vitro systems have been developed for this purpose. The molecular nature of the FcRn-IgG interaction has been studied extensively and encompasses residues located at the CH2-CH3 domain interface of the Fc region of IgG. These Fc amino acids are highly conserved in rodents and man and interact with residues primarily located on the a2 domain of FcRn. Thus, it is now possible to engineer IgGs with altered affinities for FcRn, and this has relevance to the modulation of IgG serum half-life and maternofetal IgG transport for therapeutic applications.

INTRODUCTION The neonatal Fc receptor FcRn was first identified in rodents as the receptor that transfers maternal gammaglobulins (IgGs) from mother to young via the neonatal intestine (1–6). However, more recent data have indicated that this receptor not only delivers IgGs across the maternofetal barrier during gestation (7–9), but is also responsible for the maintenance of serum IgG levels (10–13). How does FcRn carry out these apparently diverse functions? It appears that the ability of FcRn to transport IgGs in intact form both within and across cells is exploited in 0732–0582/00/0410–0739$14.00

739

740

GHETIE n WARD

each of these activities (reviewed in 14). However, the molecular details as to how FcRn functions in distinct cellular environments have yet to be unraveled. Recent developments (15, 16) have resulted in model systems that could provide valuable tools for both cellular and molecular studies directed towards a better understanding of FcRn function. The role of FcRn as an IgG transporter offers novel routes towards the generation of therapeutics for use in situations in which IgG delivery is required (e.g. fetal immunization) or a modulation of IgG levels is desirable (e.g. autoimmunity). Thus, understanding how this protein functions has a multitude of potential applications in the treatment of human disease.


FCRN: THE GENES AND PROTEIN The gene encoding rat FcRn was first isolated by Simister & Mostov in 1989 (17). Unexpectedly, FcRn was shown to comprise a heterodimer of b2microglobulin and a 45- to 53-kDa protein, suggesting that the heavy chain might be a major histocompatibility complex (MHC) class I homolog. The cloning and sequencing of the heavy-chain gene confirmed this possibility (17) and expanded the function of this class of molecules beyond their known role in antigen presentation. All three extracellular and transmembrane domains of FcRn share homology with the corresponding regions of MHC class I molecules, with much less homology between the cytoplasmic domains (17). The divergence in the cytoplasmic regions is consistent with the different functional activities of the two types of proteins. The X-ray crystallographic structure of the extracellular domains of FcRn confirmed that it is structurally similar to MHC class I molecules (18). Notably, the peptide groove that is occupied by peptide or, in some cases, glycolipid ligand in classical and nonclassical MHC class I molecules (19–22) is occluded in FcRn. This occlusion is primarily caused by the presence of proline at position 165 of the a2 domain helix, which introduces a kink. However, it is interesting that the introduction of a proline at the corresponding position in the MHC class I molecule H-2Dd does not affect peptide presentation to T cells (23), suggesting that additional structural features of FcRn may result in the closed groove. More recently the genes encoding both mouse and human FcRn alpha chains have been isolated (24, 25). The rodent and human genes share homology, with mouse and rat FcRn being highly related and the human form more divergent. The identification of human FcRn in human syncytiotrophoblast (25–28) led to the suggestion that it plays a role in the maternofetal transfer of IgGs, which is discussed in more detail below. The isolation and characterization of human FcRn provided an important link between studies of FcRn in rodents and humans.

741

MHC CLASS I–RELATED RECEPTOR FcRn

THE MOLECULAR DETAILS OF THE FCRNIMMUNOGLOBULIN G INTERACTION


Mapping the Binding Site for FcRn on Immunoglobulin G A combination of approaches has been used to localize the interaction site for FcRn on IgG. This has involved the analysis of mutated derivatives of human and mouse IgGs/Fc fragments in both in vivo and in vitro binding studies (9, 29– 32). The in vivo studies include the analysis of the IgGs or Fc fragments using assays in which FcRn is known to play a role, namely transfer across the neonatal intestine, maternofetal transfer, and serum half-life determination (discussed further below). Activity in these assays almost invariably provides an indication of FcRn affinity (9, 29, 31). The studies have resulted in the identification of several conserved amino acids located at the CH2-CH3 domain interface that play a central role in the interaction of rat or mouse FcRn with human or mouse IgGs. Ile253 and His310 are key players in the interaction (9, 29–31). These residues are highly conserved across species (Table 1), and their location on the threedimensional structure of human IgG1 (hIgG1)-derived Fc (33) is shown in Figure 1 (see color insert). His436 plays a minor but significant role in the mouse FcRnIgG interaction (31). The lack of conservation of this amino acid across species (Table 1) is consistent with more limited involvement. TABLE 1 Variations of IgG sequences in the region involved in the binding of FcRn Amino Acid Sequence at Position: 252 253 254 255 256 257

307 308 309 310 311

433 434 435 436

mouse IgG1 IgG2a IgG2b IgG3

Thr Met Met Met

Ile Ile Ile Ile

Thr Ser Ser Ser

Leu Leu Leu Leu

Thr Thr Thr Thr

Pro Pro Pro Pro

Pro Pro Pro Pro

Ile Ile Ile Ile

Met Gln Gln Gln

His His His His

Gln Gln Gln Gln

His His Lys His

Asn Asn Asn Asn

His His Tyr His

His His Tyr His

rat

IgG1 IgGa IgG2b IgG2c

Thr Thr Leu Met

Ile Ile Ile Ile

Thr Thr Ser Thr

Leu Leu Gln Leu

Thr Thr Asn Thr

Pro Pro Ala Pro

Pro Pro Pro His

Ile Ile Ile Ile

Leu Val Gln Gln

His His His His

Gln Arg Gln Gln

His His His His

Asn Asn Asn Asn

His His His His

His His His His

human IgG1 IgG2 IgG3* IgG4

Met Met Met Met

Ile Ile Ile Ile

Ser Ser Ser Ser

Arg Arg Arg Arg

Thr Thr Thr Thr

Pro Pro Pro Pro

Thr Thr Thr Thr

Val Val Val Val

Leu Val Leu Leu

His His His His

Gln Gln Gln Gln

His His His His

Asn Asn Asn Asn

His His Arg His

Tyr Tyr Phe Tyr

*Allotype containing Arg instead of His at Position 435


742

GHETIE n WARD

In functional studies in mice, mutation of His433 to alanine in recombinant Fc fragments derived from both human and mouse IgG1 does not affect the activity of the protein (31, 32), although simultaneous mutation of His433 and Asn434 does have a moderate effect (29). In contrast, mutation of His435 to alanine results in a dramatic loss of function (31, 32). Thus, the effect of the double His433/Asn434 mutation is most likely caused by perturbation of the conformation of the critical residue His435. The spatial location of these amino acids is shown in Figure 1. A more recent study in our laboratory has indicated that the sequence difference of hIgG1 and hIgG3 (G3m,s1,t1 allotype) at position 435 (His in IgG1; Arg in G3m,s1,t1 allotype of IgG3) is responsible for the lower activity of IgG3 in FcRn-mediated functions, providing further support for a role for residue 435 in the FcRn interaction (32). In contrast to our in vivo and in vitro analyses, surface plasmon resonance analyses of chimeric human-mouse IgGs (containing the hIgG4 constant region) in which His435 was mutated to arginine detected no significant effect on the pH dependence of the FcRn-IgG interaction, and, in the same study, His433 was implicated in FcRn binding (30). There is therefore a discrepancy between the relative roles of His433 and His435 in the FcRn (rodent)-IgG interaction, and this is possibly a reflection of the different systems and approaches used. However, the available data unequivocally demonstrate the involvement of IgG histidines in FcRn binding, and Table 1 shows that these histidines are highly conserved in mouse, rat, and human. This provides an explanation for the strict pH dependence (binding at pH 6–6.5; very weak or undetectable binding at pH 7.2) of the FcRn-IgG (or Fc) interaction that is observed by using soluble recombinant FcRn in binding studies (30, 34, 35) and cell-binding assays (3, 4, 6, 36). This mechanism of achieving pH-dependent binding in ranges of pH that are physiologically relevant avoids the need to invoke conformational changes and could be exploited in the engineering of other proteins. Indeed, the crystallographic analysis of rat FcRn at pH 6.5 and pH 8 shows no major conformational differences between the two forms (37). By analogy, other proteins, such as the hemochromatosis protein HFE, bind to the transferrin receptor in a pH-dependent way with binding at slightly basic pH and release at acidic pH, that is, the reverse pH dependence to that seen with FcRn, and this is again mediated by histidines (38). Further analyses of the activity of rat IgGs in FcRn-mediated functions in mice revealed a role for several additional residues in proximity to the CH2-CH3 domain interface (39). The putative involvement of these amino acids was probed by transplanting the rat IgG sequences (Table 1) onto the corresponding positions in the highly homologous mouse IgG1 Fc region. Analysis of the resulting recombinant Fc fragments demonstrated that amino acids at position 257 and, to a lesser extent, positions 307 and 309 play a role in the FcRn-Fc interaction. The location of residue 257 is shown in Figure 1. The study also excluded the involvement of residues 386 and 387, which are located in an exposed loop in the vicinity of the FcRn interaction site (40). The role of residues 257, 307, and 309 is less marked than that of Ile253, His310, and His435. However, sequence variations at 257,



743

307, and 309 can be used to explain the different affinities of rat IgGs for FcRn [affinities decreasing in the order IgG2a . IgG1 . IgG2c . IgG2b (39)]. This ranking of affinities is consistent with the binding activities of rat IgGs to isolated neonatal rat brush borders, with the exception of rat IgG2c, which was reported to have undetectable activity (41). It is possible that the discrepancies between the two studies (39, 41) are caused by functional differences between mouse and rat FcRn and/or variations in the sources of the ligands used. ˚ resolution structure of the rat Consistent with the functional data, the 6-A FcRn-Fc complex indicated that FcRn interacts with Fc residues located at the CH2-CH3 interface (40). Furthermore, fragment B of staphylococcal protein A (SpA), which, from the three-dimensional structure of an SpA (fragment B)human Fc complex (33), is known to bind to amino acids at this interdomain interface, competes with IgG for binding to FcRn (42). Earlier studies indicated that blockade of the FcRn interaction site on rabbit IgG by SpA (or fragment B) resulted in a dramatic reduction in serum half-life (43), consistent with the more recent data in mice supporting a role for FcRn in serum IgG homeostasis (10– 13).

The Conformational Dependence of the FcRn Interaction Site on Immunoglobulin G It is interesting that analyses of chimeric IgGs comprising mouse variable regions linked to human constant-region domains indicate that amino acids distal to the CH2-CH3 domain interface may also be involved in regulating serum half-life in mice and, by extension, in binding to FcRn (44). It will be of interest to determine the affinities for FcRn of these recombinant IgGs in which the constant-region domains of IgG2 and IgG1 were shuffled with the corresponding regions of IgG3 and IgG4, respectively. The data obtained for these engineered IgGs led to the suggestion that the FcRn interaction can be affected by long-range conformational effects of amino acids that are distal to the FcRn footprint (44). In this respect, critical IgG residues for the FcRn interaction (9, 29–31) are located on three loops that are spatially close but distal in primary amino acid sequence (Figure 1). This suggests that they might be highly dependent on the conformation of the b strands that support them and also on the relative disposition of the CH2 and CH3 domains. Several observations are consistent with this explanation. (a) The effect of mutation (Pro to Ala) at position 257 of a recombinant mouse IgG1-derived Fc fragment on serum half-life is most likely from perturbation of the conformation of the loop encompassing Ile253 (39). This same sequence difference is probably responsible for the shorter serum half-life of rat IgG2b relative to rat IgG1/IgG2a in mice (39). (b) Mutation of Glu333 to Ala results in a significant decrease in serum persistence of a recombinant Fc fragment [mouse IgG1-derived (V Ghetie & ES Ward, unpublished observations)]. This amino acid is located in a b strand in the CH2 domain that is on the ‘interior’ of the CH2 domain and therefore not suitably positioned for direct interaction with FcRn (33). (c)


744

GHETIE n WARD

Removal of the hinge region by recombinant techniques from an Fc fragment results in a shorter serum half-life (45). The role of the hinge appears to be to constrain the CH2 domains such that the configuration of the FcRn interaction site at the CH2-CH3 domain interface is optimal, rather than through direct FcRn:hinge interactions. This suggestion was validated by replacing the wildtype hinge sequence with a synthetic hinge that has a distinct sequence but allows -S-S-linked homodimerization of the Fc (CH2-CH3) domain monomers. This engineered, disulfide-constrained Fc fragment has the same half-life as an Fc fragment containing the wild-type hinge sequence (45). In this context, the longer and more flexible hinge region of hIgG3 compared with hIgG1 (46) might account, in part at least, for the shorter serum half-lives of IgG3 hinge-containing chimeras compared with those containing IgG2-derived hinge sequences (44). The data from these mutagenesis studies suggest that the FcRn interaction site on IgG or Fc is highly conformational dependent. This may make it a challenging task to mimick FcRn binding by an IgG-derived peptide, in contrast to the demonstrated ability of a CH2 domain-derived peptide to compete with IgGs for binding to C1q (47). However, the demonstration that an Fc region-derived peptide encompassing residues 308–317 blocks the binding of SpA to IgG (48) suggests that individual strands encompassing key FcRn interaction residues may be effective mimics.

Is There a Relationship Between the FcRn and FccR Interaction Sites on Immunoglobulin G? Mutagenesis studies from several different laboratories have indicated that residues in the lower hinge region of IgG are important for binding to FccRI, FccRII, and/or FccRIII (49–55). However, each receptor sees this site with a slightly different footprint, and, in addition, there are two other regions that play a role in binding to FccRI (Pro331) and FccRII (Glu318) (50). These regions are spatially close to the lower hinge (33). The available data indicate that FccR binding is also dependent on glycosylation of IgG (56–58) and that even minor alterations in carbohydrate structure can have an effect on this activity (59, 60). Thus, the FcRn interaction site appears to be distinct from the region of IgG involved in FccR binding. Furthermore, at least for some species/isotypes, FcRn binding is not affected by the absence of CH2 domain glycosylation (29, 35). It is interesting that the recent crystal structures of human FccRIIa (61) and FccRIIb (62) resulted in two interaction models for FccRII-IgG complex formation that are fundamentally different. FccRIIa was solved as a crystallographic dimer in which the binding regions of each monomer were brought into proximity to generate a single IgG interaction site, and the IgG-FccRIIa interaction was postulated to involve the lower hinge region of IgG (61). This model is consistent with data from mutagenesis studies (reviewed in 54). In contrast, FccRIIb was suggested to bind to IgG or Fc in the region encompassing the CH2-CH3 domain interface, with both sites on IgG occupied to form a symmetric 1:2 IgG:FccRIIb


745

complex (62). The validity of the model proposed for FccRIIa has been investigated by carrying out surface plasmon resonance-binding studies with recombinant, soluble human FccRIIa or mouse FccRI and mouse/hIgGs (M Hogarth, personal communication). For both FccRs, binding to IgG was not inhibited by either SpA or recombinant soluble FcRn, providing clear evidence that the FccR interaction site does not encompass the CH2-CH3 domain interface. Consistent with this are earlier observations that mutations of IgGs affecting FccR binding do not alter the activity of the IgG in a function that is known to be mediated by FcRn, namely serum persistence (63).


Mapping the Interaction Site for Immunoglobulin G on FcRn Although the X-ray crystallographic structure of rat Fc-FcRn complexes indicated the potential contacts for Fc on FcRn (40), the resolution was not sufficiently high to unequivocally define the amino acids involved. The binding site of mouse/ rat IgG on rat FcRn has therefore been mapped using site-directed mutagenesis followed by analysis of the mutants by in vitro binding assays (42, 64). This has resulted in the identification of several a2 domain and one b2-microglobulin residue as playing a direct role in the interaction. The a2 domain residues (Glu117, Glu132, Trp133, Glu135, and Asp137) are reasonably well conserved across species (17, 24, 25), and their location on the rat FcRn structure (18) is shown in Figure 2 (see color insert). The acidic nature of four of these amino acids, together with the role of IgG histidines in binding, suggests that electrostatic forces may play a predominant role in mediating the Fc/IgG-FcRn interaction. Hydrophobic effects may also be involved, and it is plausible that Trp133 interacts with the conserved Ile253 of IgGs (64). Residue 1 (Ile) of b2-microgobulin, which is conserved in rats, mice, and humans (46), has also been shown to be involved in binding to IgG but to a lesser extent than the a2 domain residues (64). This amino acid has been proposed to contact IgG via hydrophobic interactions in the vicinity of residue 309 and/or 311. Alternatively, this N-terminal residue may play a role through more indirect effects on Glu117 in the a2 domain (64). The a2 domain residues that are involved in binding to IgG are located in close proximity on a region at the end of the a2 domain helix (18; Figure 2). Residues 84–86 of the a1 domain are spatially close to this region of FcRn, but, from mutagenesis studies, have been shown not to contact IgG (64). The majority of contacts for IgG on FcRn therefore entail amino acids in the a2 domain. Furthermore, the interaction of IgG with FcRn occurs in a mode distinct from that reported for T-cell receptor-peptide-MHC class I interactions (65, 66), in which the T-cell receptor footprint spans the surface of the two MHC helices and binds antigenic peptide in a diagonal orientation. Much evidence from in vitro studies has supported a role for FcRn dimerization in high-afffinity binding to IgG, resulting in the formation of ‘‘lying-down’’ complexes comprising an FcRn dimer bound to one IgG/Fc molecule (42, 64, 67,

746

GHETIE n WARD


68; Figure 3). First, FcRn dimers mediated via a3 domain contacts have been observed in several different crystal forms of rat FcRn (18, 40). Second, mutation of Gly191, His250, His251, and a loop encompassing residues 219–224 (all a3 domain residues) results in a reduction in the affinity of FcRn for IgG (42, 64). The effects of these mutations have been interpreted to be caused by destabilization of FcRn dimerization and/or loss of contacts to an IgG molecule bound to an adjacent FcRn molecule in a lying-down complex (42, 64, 69). Third, immobilization of FcRn molecules on a biosensor chip favors the formation of the dimeric, high-affinity form of FcRn (68). In addition, oriented coupling of FcRn via exposed, engineered cysteines located at different sites indicates that the ori-

Figure 3 A. Schematic representation of ‘‘lying-down’’ and ‘‘standing-up’’ FcRn-Fc complexes (40). Both complexes comprise a stoichiometry of 1 Fc:2 FcRn molecules, but they show fundamental differences in the configuration of the constituent molecules. Lying-down complexes comprise Fc bound to an FcRn-FcRn dimer, with the dimer mediated primarily via a3 domain contacts and associated carbohydrate (37). In contrast, the standing-up complex comprises a symmetric FcRn-FcFcRn configuration. However, in the lying-down complex, additional contacts may be made between the Fc molecule and the secondary, that is, distal, FcRn molecule (64), but for simplicity these are not shown. B. Schematic representation of oligomeric ribbons (82) in which both standing-up and lying-down complexes coexist. For simplicity the transmembrane regions (TM) of FcRn are not shown. The arrows indicate the contacts that mediate FcRn dimerization.


747

entation of FcRn immobilization affects the extent of dimerization and, in turn, the affinity of its interaction with IgG (67). More recent X-ray crystallographic studies have demonstrated that a3 domainassociated carbohydrate (at Asn225) interacts to form much of the contacts at the FcRn dimer interface via a ‘‘carbohydrate handshake’’ (37). This indicates that, depending on the expression host used to make recombinant FcRn, the type of carbohydrate could affect the stability of FcRn dimerization with consequent effects on affinity measurements of the Fc/IgG-FcRn interaction (37). However, neither the carbohydrate addition site nor Gly191 in the a3 domain is conserved in human FcRn (25), suggesting that, if this FcRn species forms an analogous dimer, then other molecular mechanisms must be involved.


The Stoichiometry of the Fc/Immunoglobulin G-FcRn Interaction The presence of two potential binding sites for FcRn on IgG suggested that FcRn might bind to ligand in a symmetric 2:1 ‘‘standing up’’ complex (Figure 3A). For rat FcRn-Fc complexation, this possibility was supported by isothermal-titration calorimetry and column-binding assays (70). However, more recent studies have indicated that this symmetric complex may not always be formed (35). Although the need for two functional FcRn interaction sites on a recombinant Fc fragment was observed using in vivo studies in mice to assess FcRn function (71), the stoichiometry of the complex formed between recombinant mouse FcRn and mouse Fc was observed to be 1:1 (35). This apparent discrepancy prompted us to propose a model analogous to the situation described as ‘‘half-sites reactivity’’ in enzymology (14, 72, 73). In this model, we predicted that the FcRn interaction sites on IgG might not be equivalent and that, when binding occurs on one side of the Fc region, the affinity of the other site decreases. The segmental flexibility of IgG (74–76) would be consistent with such a model. This is also reminiscent of data obtained for the FceRI-IgE interaction, in which a 1:1 complex is observed despite the presence of two potential interaction sites on IgE (77, 78). Furthermore, by fluorescence resonance energy transfer methods, evidence to support a bent configuration for IgE in solution has been obtained (76). In the same study (76), the flexibility of IgG was observed to be greater than that of IgE. Furthermore, for IgE the asymmetry is so extreme that it is not possible to detect the binding of a second FceRI molecule to IgE. An asymmetric model for the FcRn-IgG interaction has recently been supported by modeling studies (69) and equilibrium gel filtration (79) or sedimentation equilibrium analyses (80). The high degree of flexibility of IgG in solution (76) is consistent with the concept that FcRn binding induces asymmetry (69) rather than the asymmetry being preformed before FcRn docking, and Fc distortion after FcRn binding was suggested in the earlier crystallographic analysis of rat Fc with FcRn (40). It has also been reported that the stoichiometry of the IgGFcRn interaction varies depending on the IgG species/isotype and glycosylation


748

GHETIE n WARD

state of the recombinant FcRn (37, 79). More generally, it is probable that the binding observed in solution with soluble (or biosensor chip-bound) FcRn might not be representative of the physiological situation when FcRn has both transmembrane regions and a cytoplasmic tail that might mediate direct or indirect interactions via adaptor proteins etc. Thus, although valuable data concerning the IgG-FcRn interaction have been obtained from in vitro binding studies with recombinant FcRn, the validity of extrapolating these findings to the in vivo situation may be questioned. It is clear, however, that in vitro binding affinity of an IgG or Fc for recombinant FcRn almost invariably correlates with the in vivo activity in FcRn-mediated functions (31, 81). A model that incorporates both standing-up FcRn-IgG:FcRn complexes and lying-down IgG-FcRn dimer (through a3 domain contacts) complexes observed in the FcRn-Fc cocrystal has been proposed by Bjorkman and colleagues (82). This embraces data supporting the existence of both types of complex (40, 67, 68) and proposes the formation of oligomeric ribbons (Figure 3B). Such ribbons have been suggested to mediate membrane vesicularization by bridging adjacent membranes (83). Some in vitro data support the model (40, 67, 68), but to date the existence of oligomeric ribbons in vivo has not been proven owing to obvious technical limitations. However, this model explains the effects of mutations at both the FcRn-Fc interaction site and the FcRn-FcRn dimer interface (42, 64).

MULTIPLE FUNCTIONS FOR FCRN FcRn: the Neonatal Fc Receptor FcRn was first identified as the Fc receptor responsible for transferring maternal IgGs from mothers’ milk across the intestinal epithelial cells of the neonatal gut of rodents (1–6). In rodents, this is the major route by which IgGs are transferred, whereas, in humans, essentially all IgGs are transferred prenatally across the placenta (see below). Early studies of rodent FcRn, using isolated rat brush border membranes, indicated that FcRn binds to IgG in a pH-dependent way, with binding at pH 6–6.5 and an undetectable interaction at pH 7.2 (3, 4, 6, 36). This result, combined with histochemical analyses, was used to build a picture of how FcRn functions as an IgG transporter (84). IgG binds to the luminal surface of the epithelial cells via FcRn and the FcRn-IgG complexes are taken up by receptormediated endocytosis. The complexes are then transcytosed across the cells and delivered via exocytosis at the basolateral surface of the cells. FcRn-IgG dissociation occurs at this site owing to the instability of the interaction at pH 7.4 (3, 4, 6, 36). This model has formed the basis of our current understanding of FcRn function, but variations on this theme need to be invoked when this receptor operates in different cellular environments (discussed below). Several observations indicate that FcRn is the only IgG transporter involved in the delivery of maternal IgGs from mothers’ milk. (a) In b2-microglobulin-deficient (b2m1/1)


749

mice that do not express functional FcRn, maternal IgGs are not transferred (8, 85). (b) Analyses of mutated murine IgG1-derived Fc fragments in assays directed towards determining their ability to be transferred across the neonatal gut indicate a direct correlation between binding affinity for FcRn and the activity in inhibiting transfer of wild-type IgG1 (29, 31).


FcRn and the Maternofetal Transfer of Immunoglobulin Gs The isolation and characterization of rat FcRn from the yolk sac endoderm led to the proposal by Roberts et al (7) that this Fc receptor is involved in the maternofetal transfer of IgGs. Quantitatively, this represents the minor route of IgG delivery from mother to young in rodents. The studies demonstrated that the FcRn-IgG interaction occurred in apical vesicles and not on the yolk sac cell surface (7). In addition, cell surface expression of FcRn could not be detected. This led to the suggestion that FcRn binding to IgG occurs only after nonspecific uptake via fluid-phase pinocytosis by yolk sac. Bound IgG is then transported across the cell and delivered to the basolateral surface, where it is released at the slightly basic pH. Binding in acidic endosomes is consistent with the known pH dependence of the FcRn-IgG interaction. There is therefore an important mechanistic difference between yolk sac and intestinal transfer. This difference may result in yolk sac transfer being less efficient (particularly at low IgG concentrations), because IgGs are not captured by cell surface receptors, but the difference is apparently a necessary tradeoff for the ability of FcRn to bind ligand in a strictly pH-dependent way. Interestingly, in vitro assays have shown that the neonatal intestine does not need to be bathed in an acidic medium on the apical surface to carry out transcytosis (86), indicating that, even for this tissue, receptor-mediated cell surface binding of IgG is not a prerequisite for uptake. By analogy with intestinal transfer, in neonates, maternofetal IgG transfer is ablated in b2m1/1 mice (8, 85). Furthermore, the efficiency of transfer of different IgG isotypes or mutated Fc fragments in mice correlates with binding affinity for FcRn (9). FcRn therefore appears to be solely responsible for IgG transport. Maternofetal transfer of IgGs in humans shows some specificity insofar as there is preferential transport of some isotypes over others (see below). In addition, the Fc region of IgG is known to be essential (87; reviewed in 88). The isolation of FcRn cDNA from human syncytiotrophoblast suggested that it is responsible for maternofetal transfer (25), consistent with the criteria that this transfer show some specificity and be mediated through the Fc region. In humans, the different organization of the fetal membranes results in IgG transfer across the syncytiotrophoblast of the chorioallantoic placenta rather than the yolk sac, and it is the major, if not sole, route of IgG delivery to offspring. Human FcRn has subsequently been identified at the protein level in syncytiotrophoblast (26– 28) and in one study has been reported to be undetectable on the apical membrane (28). Regardless of whether it is on the apical surface, the consensus


750

GHETIE n WARD

appears to be that, in the apical region of the cells, the majority is located in vesicles that are most likely acidic endosomes. However, it is controversial whether human FcRn is expressed in placental endothelial cells (26–28). If not expressed in this location, how IgGs transfer across this cellular barrier to the fetal circulation remains an open question. Colocalization studies indicate that FcRn associates with IgG in endosomes in syncytiotrophoblast cells (28). Thus, IgG-FcRn association most likely occurs after IgG uptake in acidic, apical endosomes in an analogous way to that proposed for rodent yolk sac. Consistent with this result, IgG uptake by syncytiotrophoblast is coincident with uptake of a fluid-phase marker (89). In this respect, other proteins such as placental alkaline phosphatase and annexin II have been suggested to play a role in mediating IgG uptake, but recent data argue against this (90; reviewed in 88). Furthermore, the expression of FccRs in placenta (87, 91, 92) suggested that these may play a role in transferring IgGs. However, this is not supported by the selectivity of binding of these FcRs to different IgG isotypes. For example, FccRIIb binds preferentially to IgG3 compared with IgG4 (54), but this preference does not impinge on the relative efficiency of maternofetal transfer of these isotypes (see below). The function of placental FccR expression is most likely in the clearance of immune complexes, particularly on Hofbauer cells, in which all three forms are expressed (87, 91, 92; reviewed in 88). The relative transfer efficiencies of hIgGs across the placenta remain a controversial issue, with some studies indicating poor transport of IgG2 (93–95) and others showing equivalent passage of all subclasses (96, 97). The discrepancies are most likely caused by the systems used and obvious ethical limitations on experimentation. The use of an in vitro perfused placental assay that gives a good indication of the ability of a protein/drug to cross the placenta during pregnancy indicated that the hierarchy of transfer is hIgG4 . hIgG1 . hIgG3 . hIgG2 (98). The development of in vitro systems to assess IgG transport across trophoblast cells (15, 99) should provide the opportunity to obtain quantitative data concerning FcRn-mediated transport, and this would have obvious practical implications. However, limitations of such cell systems may be that only one cell type is being investigated, in contrast to the situation during in vivo transfer, when several distinct cellular barriers must be traversed.

A Role for FcRn in Immunoglobulin G Homeostasis Over 30 years ago, Brambell and colleagues proposed that the receptors mediating transfer of maternal IgGs might be related to the protective receptors that regulate serum IgG half-life (100, 101). These receptors were suggested to function by binding and protecting IgGs against lysosomal degradation; that is, they act as salvage receptors. Consistent with this hypothesis, analyses of the binding affinity for murine FcRn of mutated mouse IgG1-derived Fc fragments with reduced serum persistence indicated that the sites of IgG or Fc that binds to FcRn and



751

regulates serum half-life closely overlap (29, 31, 102). This result prompted us to analyze the serum half-lives of murine IgG1 and IgG1-derived Fc fragments in mice that are b2m deficient (b2m1/1) and, as a result, do not express functional MHC class I molecules or homologs including FcRn (10). The half-lives of these proteins were abnormally short in these mice (10), and these results have been independently confirmed by others (11, 12). In fact, in b2m1/1 mice, the half-life of the wild-type Fc fragment is essentially the same as that of a mutated Fc with no detectable binding affinity for FcRn (10). The short serum half-lives of IgGs in b2m1/1 mice are consistent with observations that, in these mice, the circulating IgG levels are abnormally low despite an apparently normal B-cell compartment (103). The rates of IgG synthesis in these mice are insignificantly different from those in wild-type mice (11). Although we reported differences in IgG synthetic rates between wild-type and b2m1/1 mice when using the b phase half-life in the calculation (10), this difference becomes insignificant if the catabolic rate is used (V Ghetie & ES Ward, unpublished data). Consistent with the hypercatabolism that is observed in b2m1/1 mice, mice crossed onto a b2m1/1 background are also resistant to the IgG-mediated disease, bullous pemphigoid (104). Furthermore, mice crossed onto a b2m1/1 background are resistant to the induction of systemic lupus erythematosus after idiotype (16/6Id) immunization (105). This was originally believed to be caused by the lack of classical MHC class I molecules in these mice, but an alternative explanation could be that these mice have abnormally low serum IgG levels. However, recent studies in systemic lupus erythematosus– susceptible mice in which the B cells are engineered to express only surface immunoglobulin have revealed an IgG-independent mechanism for renal and vascular disease (106). In all studies to date, a good correlation between affinity for FcRn binding and serum half-life has been observed for Fc mutants or IgG variants (29, 31). Significantly, this correlation can be extended to engineered Fc fragments with higher affinity than their parent wild-type molecule. This was carried out by randomly mutating a recombinant murine Fc fragment at three residues located in proximity to the FcRn interaction site (81). These amino acids (Thr252, Thr254, and Thr256) were chosen because they are exposed and not highly conserved across species. The library of mutated fragments was expressed on the surface of bacteriophage and Fc fragments with higher affinity for recombinant mouse FcRn selected (81). One of these Fc fragments with mutations of Thr252, Thr254, and Thr256 to Leu, Ser, and Phe, respectively, had a significantly longer serum halflife than the wild-type Fc. This suggests that it should be possible to combine protein design with selection to generate therapeutic antibodies with increased serum persistence. In this context, a chimeric antibody (mouse variable domains linked to human constant regions) in which the constant region domains were shuffled to generate a hybrid IgG1/IgG4 molecule had a significantly longer halflife in mice than either of the parent IgG1 or IgG4 molecules (44). It will be of interest to unravel the molecular basis of this effect.

752

GHETIE n WARD


Where and How Does Immunoglobulin G Homeostasis Occur? Much evidence supports the concept that FcRn is ubiquitously expressed in adult tissues (10, 25, 107), in addition to its earlier localization to neonatal gut and yolk sac/placenta. The expression of FcRn in almost every tissue analyzed led us and others to propose that the endothelial cells might be the site at which IgG homeostasis occurs (10, 108). This would be consistent with the close apposition of these cells to the bloodstream and earlier suggestions of Mariani & Strober (109) concerning the site of regulation of serum IgG levels. Distribution studies of murine IgG1, Fc fragments, and anti-FcRn antibodies indicate that the major sites of FcRn function in adult, nonpregnant mice are the skin and muscle, with a lesser involvement of liver and adipose (110). Furthermore, histochemical analyses of mouse muscle and liver indicate that FcRn is expressed in the endothelium of small arterioles and capillaries but cannot be detected in larger vessels such as the central vein and portal vasculature. Functional FcRn can also be isolated from a murine endothelial cell line (110). Interestingly, immunohistochemistry indicates that the steady-state distribution of FcRn in these cells is primarily intracellular, which is reminiscent of the data for rat FcRn expression in yolk sac (7). Analysis of FcRn expression in humans also indicates that it is expressed in endothelial cells (108), and this is consistent with earlier observations that human FcRn is widely expressed throughout adult tissues (25). A model for how FcRn might function in its role as an IgG homeostat is shown in Figure 4. IgGs are taken up by endothelial cells by nonspecific pinocytosis and then enter acidic endosomes. If the IgGs bind to FcRn, they are salvaged from lysosomal degradation. Paradoxically, the endothelial cells are therefore involved in both the breakdown and salvaging of IgGs. The route by which FcRn takes in salvaging IgGs is far from clear; whether it acts as a recycling or transcytotic receptor or both in its role as the protective receptor is as yet unknown. By analogy with FcRn as an IgG homeostat, it is interesting that, when FcRn is apparently functioning as a transcytotic receptor during maternofetal transfer, significant amounts of IgGs are degraded (3, 111, 112). It is tempting to speculate that FcRn can act in recycling or transcytotic mode (or both), depending on the cell type in which it is expressed. The model (Figure 4) explains how FcRn acts as a homeostat that is finely tuned to regulate serum IgG levels. When IgG levels decrease, more FcRn is available for IgG binding so that an increased amount of IgG is salvaged. Conversely, if serum IgG levels rise, FcRn becomes saturated, and this results in an increase in the proportion of pinocytosed IgGs that are degraded.

A Possible Role for FcRn at Other Sites In addition to its role in regulating serum IgG levels and transferring maternal IgGs, additional studies have indicated that the transporting function of FcRn is



753

Figure 4 Models for the trafficking of FcRn in its role as an IgG homeostat in endothelial cells. IgGs are initially taken up by nonspecific pinocytosis because the pH of the bloodstream is not permissive for the IgG-FcRn interaction. Pinocytosed IgG molecules (Y) bind to free FcRn (1) in acidified endosomes. IgG that does not bind to FcRn is destined for degradation in lysosomes. The available evidence with in vitro cell lines indicates that FcRn can act as either a transcytotic or recycling receptor, and it is as yet unknown in which mode it operates during IgG homeostasis. However, if FcRn transcytoses IgG molecules across endothelial cells, then bidirectional transcytosis by necessity occurs for the IgG to be returned to the serum. Alternatively, IgG may traffic in one or both directions via transcellular channels (142), in addition to being transcytosed by an FcRn-mediated process. Surface expression of FcRn is not shown on this figure, because the available data indicate that the steady-state distribution in endothelial cells is primarily intracellular (110).

exploited for other purposes. For example, the detection of FcRn in human intestinal epithelial cells (113), and the bidirectional transcytosis of IgG across human intestinal T84 cell monolayers (113a) has led to several suggestions concerning its possible function. (a) It might be involved in the transfer of passive immunity via oral feeding of IgG. (b) FcRn expression at this site might act to detect pathogenic antibody in the lumen of the intestine. (c) By analogy with the transfer of immune complexes by the intestinal cells of the neonatal rodent, human FcRn might serve to deliver antigens to induce either immune activation or tolerance.


754

GHETIE n WARD

(d) By analogy with the cow (114), transfer of IgG into the gut via an FcRnmediated process might play a role in the clearance of IgGs. FcRn has also been identified in the adult rodent liver (110, 115; R Junghans, personal communication) and lactating mammary gland (116). For the liver, hepatocytic expression was suggested to play a role in delivering immune complexes from the canalicular space to the Kupffer cells and the bile (115). However, an inverse correlation between binding affinity for FcRn and transfer into bile has been observed for recombinant Fc fragments in mice (110). In addition, no significant differences were observed for serum-to-bile transport of IgGs in wildtype and b2m1/1 deficient mice, and all IgG transport could be accounted for by passive transfer (Junghans, personal communication). These observations suggest that the function of FcRn in hepatocytes might be to act as a recycling receptor whereby it salvages IgGs back into the bloodstream. In contrast, lack of binding to FcRn appears to result in an Fc fragment transiting into the bile by an unknown mechanism (110). By analogy with the activity of FcRn in liver, a similar function has been attributed to FcRn in the murine lactating mammary gland (116). For both Fc fragments and complete IgGs, an inverse correlation between delivery into milk and affinity for FcRn was observed. This unexpected observation led us to propose that this relationship might result in a balance of IgGs in the neonatal serum that reflects the composition in the mother’s bloodstream. More specifically, for transfer of maternal IgGs into neonatal serum, an IgG has to pass through two barriers: the mammary gland and the neonatal gut. Censorship by FcRn at both barriers would result in dramatic decreases of the levels of IgG isotypes that bind relatively weakly to FcRn at the expense of higher-affinity binders. Hence, we envisage that the process of transfer may occur in two steps, the first (maternal blood to milk) at which an inverse correlation between affinity and transfer is observed and the second (gut lumen to neonatal blood) at which a direct correlation is observed (Figure 5). This hypothesis would suggest that FcRn operates primarily in recycling mode in the mammary gland and in transcytotic mode in the neonatal gut. What relevance does this have to humans, in whom maternofetal transfer is the major, if not sole, route of passive IgG delivery? Interestingly, on day 2 postpartum, the levels of IgGs in colostrum range from ;30% (IgG1) to 2% (IgG4) of the levels in the maternal serum (117), suggesting that some maternal IgG transfer to the neonate may occur in the first few days of life. However, the levels of IgG in the colostrum/milk drop precipitously over the next few days post partum, resulting in IgG concentrations that are ;100-fold lower than in serum (117). This suggests that, if FcRn is expressed in the human lactating mammary gland, it may result in IgGs being recycled away from the milk to exclude this Ig class. In contrast, during colostrum production, it may be that FcRn is not as effective in this recycling function. Although ethical considerations make testing of this hypothesis difficult, it may be possible to address using in vitro cell lines.



755

Figure 5 Schematic representation of a model for the transfer of IgGs from the maternal to neonatal bloodstream via mothers’ milk, indicating the barriers that must be traversed and their relative transfer efficiencies. In mice, rat IgG1 accumulates in milk less efficiently than rat IgG2b (116), that is, in inverse correlation with affinity for FcRn (39). In contrast, the available data indicate that the transfer of an IgG molecule across the neonatal intestine correlates closely with its affinity for FcRn (29, 31). The outcome of this two-step transfer would be that the IgG subclass ratio in the neonatal serum would closely resemble that in the maternal blood.

ANALYSIS OF THE TRAFFICKING OF FCRN The analysis of transcytosis of human IgGs across in vitro monolayers of ex vivo trophoblast cells has been reported (99). More recently this has been extended to the use of the BeWo choriocarcinoma cell line (15). For both systems, IgG transport occurs preferentially in the apical to basolateral direction, rather than the opposite direction. Although preferential binding of IgG to the apical surface of BeWo cells at pH 6.0 indicates that human FcRn is present on the cell membrane, this mode of binding is unlikely to be operative in vivo, where the cells are bathed at pH 7.4 [at which only nonspecific binding could be detected (15)]. The data from these analyses are consistent with the proposed model for transcytosis across the rat yolk sac (7) and also with IgG-human colocalization studies in isolated human syncytiotrophoblast (28). Thus, this in vitro system should be a valuable tool to study human FcRn function in the placental trophoblast. To analyze the role of the cytoplasmic tail of FcRn in intracellular sorting, polarized Madin-Darby canine kidney (MDCK) cells have been transfected with FccRIIb-rat FcRn cytoplasmic tail chimeras (16). Several types of constructs were made, but the most informative are the wild type, a variant in which the di-leucine


756

GHETIE n WARD

motif was mutated to di-alanine, and a truncated mutant with essentially no cytoplasmic tail. The di-leucine motif has been shown to mediate basolateral sorting and receptor internalization for many different receptors [e.g. FccRII (118) and the CD3c chain of the T-cell receptor (119)]. However, for rat FcRn the di-leucine motif appears to play no role in basolateral sorting, although truncation of almost all of the cytoplasmic tail resulted in reduced transport to the basolateral surface. This suggests that there may be an as yet unidentified motif involved in sorting of the FccRII-FcRn chimera to the basolateral surface. In contrast to its lack of involvement in basolateral sorting, the di-leucine motif mediates efficient internalization of FcRn (16). It is interesting that receptor internalization was not completely ablated in the di-alanine variant, leading to the suggestion that a casein kinase II phosphorylation site in the cytoplasmic domain (N-terminal to the dileucine motif) might also regulate the surface expression of rat FcRn. Such casein kinase II sites have been reported to play a role in receptor internalization in other systems (120). In contrast to the di-leucine motif, this site is not conserved in human FcRn (17, 25), suggesting that there may be subtle differences in FcRn trafficking across species. In MDCK cell transfectants expressing the FccRIIb-FcRn chimera, transcytosis occurred in both directions but was more efficient in the basolateral-to-apical direction than vice versa (16). Moreover, the apical surface was more active in recycling than the basolateral surface. However, these studies were carried out using anti-FccRIIb Fab fragments, and whether these biases in transcytosis and recycling are seen with natural ligand remains to be analyzed. The MDCK system has provided much valuable information concerning the trafficking of other receptors (121, 122) and should be useful for the elucidation of the molecular mechanisms involved in FcRn function.

PROSPECTS: DOES FCRN HAVE RELEVANCE IN THE CLINIC? The apparently diverse roles of FcRn are linked by the ability of FcRn to bind and traffic IgGs within and across cells. The observation that the catabolic rate of IgG is dependent on its plasma concentration (123), known as the concentration-catabolism phenomenon, was until recently without explanation. However, the identification of FcRn as the receptor that regulates serum IgG homeostasis gives a molecular explanation for this phenomenon. This leads to the question of how FcRn expression is regulated, because fluctuations in FcRn levels would in turn affect serum IgG concentrations. To date, little is known about this regulation, but the presence of motifs for two cytokine-inducible factors [nuclear factor (NF)– interleukin (IL)-6 and NF-1)] upstream of mouse FcRn (124) suggests that it may be upregulated during the acute phase of an immune response. It is tempting to suggest that dysregulation of FcRn expression may be involved in situations in



757

which hypercatabolism is observed, such as after burns (125) and in myotonic dystrophy (126). Furthermore, it is possible that some types of IgG deficiencies such as familial idiopathic hypercatabolism (127) may be caused by abnormalities in FcRn expression or function. The role of FcRn as an IgG homeostat suggests that the modulation of FcRn function and/or expression might be an effective approach for the treatment of IgG-mediated disease. For example, in autoimmune diseases in which pathogenic IgGs are involved, blockade of FcRn function might be an effective treatment modality (128). Indeed, the use of intravenous immunoglobulin has been shown to be efficacious in such situations through a mechanism that involves the induction of hypercatabolism of endogenous IgGs (129). The synergism that is observed between intravenous immunoglobulin therapy and methylprednisone in the treatment of Guillain-Barre syndrome (130) is most likely caused by the ability of glucocorticoids to decrease FcRn expression at the transcriptional level (131). This explanation is also consistent with the earlier report that hydrocortisone treatment results in hypogammaglobulinemia (132). An alternative way of regulating FcRn function would be to generate novel FcRn ligands that block IgG binding or affect FcRn trafficking, and our expanding knowledge of FcRn at the molecular level may facilitate this task. Understanding the molecular basis by which IgGs persist in the serum is of relevance to the engineering of improved antibodies for use in therapy (81, 133). For example, for the passive delivery of antitumor antibodies [e.g. anti-HER-2 (134) and anti-CD20 (135)], a longer half-life of the IgG could result in improved efficacy and the need for fewer doses. Alternatively, for use in imaging, it is desirable to have a short half-life and, although this can be achieved by using Fab fragments, it is now also possible to engineer complete antibodies with single amino acid substitutions [e.g. Ile253 to Ala (31, 32)], which would be predicted to have reduced serum persistence. An additional FcRn function that could be exploited in the clinic is the passive delivery of therapeutic antibodies in maternofetal medicine. There are many situations in which this might be useful, but one may be in the specific delivery of antipathogen IgGs. In this respect, the engineering of IgGs with higher affinities for FcRn is attractive (81), because these would be expected to be transferred more efficiently. Although it has been reported that maternally derived IgGs can adversely affect the neonatal response to immunization (136), less is known about the effects of IgGs on T-cell responses. In fact, in vitro analyses have indicated that some IgGs can enhance (137, 138) or suppress (139, 140) T-cell responses to specific epitopes, most likely via effects on antigen processing. It is important that a recent study in mice has demonstrated that maternal IgG does not appear to have a detrimental effect on helper T-cell responses (141). Furthermore, early prime-boosting regimens can result in normal secondary responses even in the presence of passively transferred IgG (141). This may allay concerns about interference effects of maternally derived antibodies on subsequent vaccination of the

758

GHETIE n WARD

neonate, but has yet to be tested for the panoply of pathogens for which vaccines exist. Finally, the expression of FcRn and its possible function at other sites such as the adult intestine (113) and liver (110, 115) (Junghans, personal communication) remain to be more fully investigated and could result in yet more roles being defined for this intriguing MHC class I relative. Note Added in Proof


Recent studies by Martin and Bjorkman (143) have demonstrated that the ‘‘lying down’’ FcRn dimer-Fc complex proposed earlier by Bjorkman and colleagues (42, 64, 67, 68) does not exist in solution. However, this does not exclude the possibility that these complexes exist under physiological conditions when FcRn is membrane bound. ACKNOWLEDGMENTS We are indebted to Jin-Kyoo Kim, Corneliu Medesan, Caius Radu, Petru Cianga, Sergei Popov, Bertram Ober, Jozef Borvak, Mihail Firan, Felicia Antohe, and Maya Simionescu for their contributions to our understanding of FcRn. We are also grateful to Caius Radu for assistance with the figures. Our FcRn-related studies are supported by grants from the National Institutes of Health (R01AI39167), the Robert A. Welch Foundation (I-1333), and the Texas Coordinating Board. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Jones EA, Waldmann TA. 1972. The mechanism of intestinal uptake and transcellular transport of IgG in the neonatal rat. J. Clin. Invest. 51:2916–27 2. Rodewald R. 1973. Intestinal transport of antibodies in the newborn rat. J. Cell Biol. 58:189–211 3. Rodewald R. 1976. pH-dependent binding of immunoglobulins to intestinal cells of the neonatal rat. J. Cell Biol. 71:666–69 4. Rodewald R. 1980. Distribution of immunoglobulin G receptors in the small intestine of the young rat. J. Cell Biol. 85:18–32

5. Rodewald R, Abrahamson DR. 1982. Receptor-mediated transport of IgG across the intestinal epithelium of the neonatal rat. Ciba Found. Symp. 94:209– 32 6. Wallace KH, Rees AR. 1980. Studies on the immunoglobulin-G Fc-fragment receptor from neonatal rat small intestine. Biochem. J. 188:9–16 7. Roberts DM, Guenthert M, Rodewald R. 1990. Isolation and characterization of the Fc receptor from the fetal yolk sac of the rat. J. Cell Biol. 111:1867–76 8. Zijlstra M, Bix M, Simister NE, Loring JM, Raulet DH, Jaenisch R. 1990. Beta


9.

10.


11.

12.

13.

14.

15.

16.

2-microglobulin deficient mice lack CD4–8` cytolytic T cells. Nature 344:742–46 Medesan C, Radu C, Kim JK, Ghetie V, Ward ES. 1996. Localization of the site of the IgG molecule that regulates maternofetal transmission in mice. Eur. J. Immunol. 26:2533–36 Ghetie V, Hubbard JG, Kim JK, Tsen MF, Lee Y, Ward ES. 1996. Abnormally short serum half-lives of IgG in beta 2microglobulin-deficient mice. Eur. J. Immunol. 26:690–96 Junghans RP, Anderson CL. 1996. The protection receptor for IgG catabolism is the beta2-microglobulin-containing neonatal intestinal transport receptor. Proc. Natl. Acad. Sci. USA 93:5512–16 Israel EJ, Wilsker DF, Hayes KC, Schoenfeld D, Simister NE. 1996. Increased clearance of IgG in mice that lack beta 2-microglobulin: possible protective role of FcRn. Immunology 89:573–78 Christianson GJ, Brooks W, Vekasi S, Manolfi EA, Niles J, Roopenian SL, Roths JB, Rothlein R, Roopenian DC. 1997. Beta 2-microglobulin-deficient mice are protected from hypergammaglobulinemia and have defective antibody responses because of increased IgG catabolism. J. Immunol. 159:4781–92 Ghetie V, Ward ES. 1997. FcRn: the MHC class I-related receptor that is more than an IgG transporter. Immunol. Today 18:592–98 Ellinger I, Schwab M, Stefanescu A, Hunziker W, Fuchs R. 1999. IgG transport across trophoblast-derived BeWo cells: a model system to study IgG transport in the placenta. Eur. J. Immunol. 29:733–44 Stefaner I, Praetor A, Hunziker W. 1999. Nonvectorial surface transport, endocytosis via a Di-leucine-based motif, and bidirectional transcytosis of chimera encoding the cytosolic tail of rat FcRn

17.

18.

19.

20.

21.

22.

23.

24.

25.

759

expressed in Madin-Darby canine kidney cells. J. Biol. Chem. 274:8998–9005 Simister NE, Mostov KE. 1989. An Fc receptor structurally related to MHC class I antigens. Nature 337:184–87 Burmeister WP, Gastinel LN, Simister NE, Blum ML, Bjorkman PJ. 1994. ˚ resolution of Crystal structure at 2.2 A the MHC-related neonatal Fc receptor. Nature 372:336–43 Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC. 1987. Structure of the human class I histocompatibility antigen, HLA-A2. Nature 329:506–12 Bjorkman PJ, Saper MA, Samraoui B, Bennett WS, Strominger JL, Wiley DC. 1987. The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens. Nature 329: 512–18 Zeng Z, Castano AR, Segelke BW, Stura EA, Peterson PA, Wilson IA. 1997. Crystal structure of mouse CD1: an MHC-like fold with a large hydrophobic binding groove. Science 277:339–45 Joyce S, Woods AS, Yewdell JW, Bennink JR, De Silva AD, Boesteanu A, Balk SP, Cotter RJ, Brutkiewicz RR. 1998. Natural ligand of mouse CD1d1: cellular glycosylphosphatidylinositol. Science 279:1541–44 Plaksin D, Polakova K, Mage MG, Margulies DH. 1997. Rigidification of the alpha2 helix of an MHC class I molecule by a valine to proline mutation in position 165 does not prevent peptide- specific antigen presentation. J. Immunol. 159:4408–14 Ahouse JJ, Hagerman CL, Mittal P, Gilbert DJ, Copeland NG, Jenkins NA, Simister NE. 1993. Mouse MHC class I-like Fc receptor encoded outside the MHC. J. Immunol. 151:6076–88 Story CM, Mikulska JE, Simister NE. 1994. A major histocompatibility complex class I-like Fc receptor cloned from human placenta: possible role in transfer

760

26.

27.


28.

29.

30.

31.

32.

33.

34.

GHETIE n WARD of immunoglobulin G from mother to fetus. J. Exp. Med. 180:2377–81 Simister NE, Story CM, Chen HL, Hunt JS. 1996. An IgG-transporting Fc receptor expressed in the syncytiotrophoblast of human placenta. Eur. J. Immunol. 26:1527–31 Leach JL, Sedmak DD, Osborne JM, Rahill B, Lairmore MD, Anderson CL. 1996. Isolation from human placenta of the IgG transporter, FcRn, and localization to the syncytiotrophoblast: implications for maternal-fetal antibody transport. J. Immunol. 157:3317–22 Kristoffersen EK, Matre R. 1996. Colocalization of the neonatal Fc gamma receptor and IgG in human placental term syncytiotrophoblasts. Eur. J. Immunol. 26:1668–71 Kim JK, Tsen MF, Ghetie V, Ward ES. 1994. Localization of the site of the murine IgG1 molecule that is involved in binding to the murine intestinal Fc receptor. Eur. J. Immunol. 24:2429–34 Raghavan M, Bonagura VR, Morrison SL, Bjorkman PJ. 1995. Analysis of the pH dependence of the neonatal Fc receptor/immunoglobulin G interaction using antibody and receptor variants. Biochemistry 34:14649–57 Medesan C, Matesoi D, Radu C, Ghetie V, Ward ES. 1997. Delineation of the amino acid residues involved in transcytosis and catabolism of mouse IgG1. J. Immunol. 158:2211–17 Kim J-K, Firan M, Radu CG, Kim C-H, Ghetie V, Ward ES. 1999. Mapping the site on human IgG for binding of the MHC class I related receptor, FcRn. Eur. J. Immunol. 29:2819–25 Deisenhofer J. 1981. Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylo˚ resolucoccus aureus at 2.9- and 2.8-A tion. Biochemistry 20:2361–70 Gastinel L, Simister NE, Bjorkman PJ. 1992. Expression and crystalization of a

35.

36.

37.

38.

39.

40.

41.

42.

43.

soluble and functional form of an Fc receptor related to class I histocompatibility molecules. Proc. Natl. Acad. Sci. USA 89:638–42 Popov S, Hubbard JG, Kim J, Ober B, Ghetie V, Ward ES. 1996. The stoichiometry and affinity of the interaction of murine Fc fragments with the MHC class I-related receptor, FcRn. Mol. Immunol. 33:521–30 Guyer RL, Koshland ME, Knopf PM. 1976. Immunoglobulin binding by mouse intestinal epithelial cell receptors. J. Immunol. 117:587–93 Vaughn DE, Bjorkman PJ. 1998. Structural basis of pH-dependent antibody binding by the neonatal Fc receptor. Structure 6:63–73 Lebron JA, Bennett MJ, Vaughn DE, Chirino AJ, Snow PM, Mintier GA, Feder JN, Bjorkman PJ. 1998. Crystal structure of the hemochromatosis protein HFE and characterization of its interaction with transferrin receptor. Cell 93:111–23 Medesan C, Cianga P, Mummert M, Stanescu D, Ghetie V, Ward ES. 1998. Comparative studies of rat IgG to further delineate the Fc:FcRn interaction site. Eur. J. Immunol. 28:2092–100 Burmeister WP, Huber AH, Bjorkman PJ. 1994. Crystal structure of the complex of rat neonatal Fc receptor with Fc. Nature 372:379–83 Hobbs SM, Jackson LE, Peppard JV. 1987. Binding of subclasses of rat immunoglobulin G to detergent-isolated Fc receptor from neonatal rat intestine. J. Biol. Chem. 262:8041–46 Raghavan M, Chen MY, Gastinel LN, Bjorkman PJ. 1994. Investigation of the interaction between the class I MHCrelated Fc receptor and its immunoglobulin G ligand. Immunity 1:303–15 Dima S, Medesan C, Mota G, Moraru I, Sjoquist J, Ghetie V. 1983. Effect of protein A and its fragment B on the catabolic


44.

45.


46.

47.

48.

49.

50.

51.

52.

and Fc receptor sites of IgG. Eur. J. Immunol. 13:605–14 Zuckier LS, Chang CJ, Scharff MD, Morrison SL. 1998. Chimeric humanmouse IgG antibodies with shuffled constant region exons demonstrate that multiple domains contribute to in vivo half-life. Cancer Res. 58:3905–8 Kim JK, Tsen MF, Ghetie V, Ward ES. 1995. Evidence that the hinge region plays a role in maintaining serum levels of the murine IgG1 molecule. Mol. Immunol. 32:467–75 Kabat EA, Wu TT, Perry HM, Gottesman KS, Foeller C. 1991. Sequences of Proteins of Immunological Interest. Washington, DC: U.S. Dep. Health Hum. Serv. 5th ed. Duncan AR, Winter G. 1988. The binding site for C1q on IgG. Nature 332:738– 40 Frick IM, Wikstrom M, Forsen S, Drakenberg T, Gomi H, Sjobring U, Bjorck L. 1992. Convergent evolution among immunoglobulin G-binding bacterial proteins. Proc. Natl. Acad. Sci. USA 89:8532–36 Duncan AR, Woof JM, Partridge LJ, Burton DR, Winter G. 1988. Localization of the binding site for the human highaffinity Fc receptor on IgG. Nature 332:563–64 Canfield SM, Morrison SL. 1991. The binding affinity of human IgG for its high affinity Fc receptor is determined by multiple amino acids in the CH2 domain and is modulated by the hinge region. J. Exp. Med. 173:1483–91 Lund J, Pound JD, Jones PT, Duncan AR, Bentley T, Goodall M, Levine BA, Jefferis R, Winter G. 1992. Multiple binding sites on the CH2 domain of IgG for mouse Fc gamma R11. Mol. Immunol. 29:53–59 Lund J, Winter G, Jones PT, Pound JD, Tanaka T, Walker MR, Artymiuk PJ, Arata Y, Burton DR, Jefferis R. 1991. Human Fc gamma RI and Fc gamma RII

53.

54.

55.

56.

57.

58.

59.

60.

761

interact with distinct but overlapping sites on human IgG. J. Immunol. 147:2657–62 Hogarth PM, Ierino FL, Hulett MD. 1994. Characterization of FcR Ig-binding sites and epitope mapping. Immunomethods 4:17–24 Hulett MD, Hogarth PM. 1994. Molecular basis of Fc receptor function. Adv. Immunol. 57:1–127 Gavin AL, Barnes N, Dijstelbloem HM, Hogarth PM. 1998. Identification of the mouse IgG3 receptor: implications for antibody effector function at the interface between innate and adaptive immunity. J. Immunol. 160:20–23 Tao MH, Morrison SL. 1989. Studies of aglycosylated chimeric mouse-human IgG: role of carbohydrate in the structure and effector functions mediated by the human IgG constant region. J. Immunol. 143:2595–601 Tsuchiya N, Endo T, Matsuta K, Yoshinoya S, Aikawa T, Kosuge E, Takeuchi F, Miyamoto T, Kobata A. 1989. Effects of galactose depletion from oligosaccharide chains on immunological activities of human IgG. J. Rheumatol. 16:285–90 Walker MR, Lund J, Thompson KM, Jefferis R. 1989. Aglycosylation of human IgG1 and IgG3 monoclonal antibodies can eliminate recognition by human cells expressing Fc gamma RI and/or Fc gamma RII receptors. Biochem. J. 259:347–53 Wright A, Morrison SL. 1994. Effect of altered CH2-associated carbohydrate structure on the functional properties and in vivo fate of chimeric mouse-human immunoglobulin G1. J. Exp. Med. 180:1087–96 Lund J, Takahashi N, Pound JD, Goodall M, Jefferis R. 1996. Multiple interactions of IgG with its core oligosaccharide can modulate recognition by complement and human Fc gamma receptor I and

762

61.

62.


63.

64.

65.

66.

67.

68.

GHETIE n WARD influence the synthesis of its oligosaccharide chains. J. Immunol. 157:4963–69 Maxwell KF, Powell MS, Hulett MD, Barton PA, McKenzie IF, Garrett TP, Hogarth PM. 1999. Crystal structure of the human leukocyte Fc receptor, Fc gammaRIIa. Nat. Struct. Biol. 6:437–42 Sondermann P, Huber R, Jacob U. 1999. Crystal structure of the soluble form of the human Fc gamma-receptor IIb: a new member of the immunoglobulin super˚ resolution. EMBO J. family at 1.7 A 18:1095–103 Wawrzynczak EJ, Denham S, Parnell GD, Cumber AJ, Jones PT, Winter G. 1992. Recombinant mouse monoclonal antibodies with single amino acid substitutions affecting Clq and high affinity Fc receptor binding have identical serum half-lives in the BALB/c mouse. Mol. Immunol. 29:221–27 Vaughn DE, Milburn CM, Penny DM, Martin WL, Johnson JL, Bjorkman PJ. 1997. Identification of critical IgG binding epitopes on the neonatal Fc receptor. J. Mol. Biol. 274:597–607 Garcia KC, Degano M, Stanfield RL, Brunmark A, Jackson MR, Peterson PA, Teyton L, Wilson IA. 1996. An alphabeta ˚ and its T cell receptor structure at 2.5 A orientation in the TCR-MHC complex. Science 274:209–19 Garboczi DN, Ghosh P, Utz U, Fan QR, Biddison WE, Wiley DC. 1996. Structure of the complex between human T-cell receptor, viral peptide and HLA-A2. Nature 384:134–41 Raghavan M, Wang Y, Bjorkman PJ. 1995. Effects of receptor dimerization on the interaction between the class I major histocompatibility complex-related Fc receptor and IgG. Proc. Natl. Acad. Sci. USA 92:11200–4 Vaughn DE, Bjorkman PJ. 1997. Highaffinity binding of the neonatal Fc receptor to its IgG ligand requires receptor immobilization. Biochemistry 36:9374– 80

69. Weng Z, Gulukota K, Vaughn DE, Bjorkman PJ, DeLisi C. 1998. Computational determination of the structure of rat Fc bound to the neonatal Fc receptor. J. Mol. Biol. 282:217–25 70. Huber AH, Kelley RF, Gastinel LN, Bjorkman PJ. 1993. Crystallization and stoichiometry of binding of a complex between a rat intestinal Fc receptor and Fc. J. Mol. Biol. 230:1077–83 71. Kim JK, Tsen MF, Ghetie V, Ward ES. 1994. Catabolism of the murine IgG1 molecule: evidence that both CH2-CH3 domain interfaces are required for persistence of IgG1 in the circulation of mice. Scand. J. Immunol. 40:457–65 72. Fersht AR. 1975. Demonstration of two active sites on a monomeric aminoacyltRNA synthetase: possible roles of negative cooperativity and half-of-the-sites reactivity in oligomeric enzymes. Biochemistry 14:5–12 73. Bosshard HR, Koch LE, Hartley BS. 1975. Aminoacyl-tRNA synthetases from Bacillus stearothermophilus: asymmetry of substrate binding to tyrosyltRNA synthetase. Eur. J. Biochem. 53:493–98 74. Nezlin R. 1990. Internal movements in immunoglobulin molecules. Adv. Immunol. 48:1–40 75. Burton DR. 1990. Antibody: the flexible adaptor molecule. Trends Biochem. Sci. 15:64–69 76. Zheng Y, Shopes B, Holowka D, Baird B. 1992. Dynamic conformations compared for IgE and IgG1 in solution and bound to receptors. Biochemistry 31:7446–56 77. Keown MB, Ghirlando R, Young RJ, Beavil AJ, Owens RJ, Perkins SJ, Sutton BJ, Gould HJ. 1995. Hydrodynamic studies of a complex between the Fc fragment of human IgE and a soluble fragment of the Fc epsilon RI alpha chain. Proc. Natl. Acad. Sci. USA 92:1841–45 78. Robertson MW. 1993. Phage and Escherichia coli expression of the human high


79.

80.


81.

82.

83.

84.

85.

86.

87.

88.

89.

affinity immunoglobulin E receptor alpha-subunit ectodomain: domain localization of the IgE-binding site. J. Biol. Chem. 268:12736–43 Sanchez LM, Penny DM, Bjorkman PJ. 1999. Stoichiometry of the interaction between the MHC-related Fc receptor and its Fc ligand. Biochemistry 38:9471– 76. Schuck P, Radu CG, Ward ES. 1999. Sedimentation equilibrium analysis of recombinant mouse FcRn with murine IgG1. Mol. Immunol. 36:1117–25 Ghetie V, Popov S, Borvak J, Radu C, Matesoi D, Medesan C, Ober RJ, Ward ES. 1997. Increasing the serum persistence of an IgG fragment by random mutagenesis. Nat. Biotechnol. 15:637–40 Raghavan M, Bjorkman PJ. 1996. Fc receptors and their interactions with immunoglobulins. Annu. Rev. Cell Dev. Biol. 12:181–220 Bjorkman PJ. 1998. Antibodies: Molecular, Cellular and Clinical Aspects. Cancer Res. Inst. (Abstr.), April 1998 Rodewald R, Kraehenbuhl JP. 1984. Receptor-mediated transport of IgG. J. Cell Biol. 99:159s–64s Israel EJ, Patel VK, Taylor SF, MarshakRothstein A, Simister NE. 1995. Requirement for a beta 2-microglobulinassociated Fc receptor for acquisition of maternal IgG by fetal and neonatal mice. J. Immunol. 154:6246–51 Benlounes N, Chedid R, Thuillier F, Desjeux JF, Rousselet F, Heyman M. 1995. Intestinal transport and processing of immunoglobulin G in the neonatal and adult rat. Biol. Neonate 67:254–63 Kristoffersen EK, Ulvestad E, Vedeler CA, Matre R. 1990. Fc gamma receptor heterogeneity in the human placenta. Scand. J. Immunol. 32:561–64 Simister NE, Story CM. 1997. Human placental Fc receptors and the transmission of antibodies from mother to fetus. J. Reprod. Immunol. 37:1–23 Leach L, Eaton BM, Firth JA, Contractor

90.

91.

92.

93.

94.

95.

96.

97.

98.

763

SF. 1991. Immunocytochemical and labelled tracer approaches to uptake and intracellular routing of immunoglobulinG (IgG) in the human placenta. Histochem. J. 23:444–49 Stefaner I, Stefanescu A, Hunziker W, Fuchs R. 1997. Expression of placental alkaline phosphatase does not correlate with IgG binding, internalization and transcytosis. Biochem. J. 327:585–92 Kameda T, Koyama M, Matsuzaki N, Taniguchi T, Saji F, Tanizawa O. 1991. Localization of three subtypes of Fc gamma receptors in human placenta by immunohistochemical analysis. Placenta 12:15–26 Sedmak DD, Davis DH, Singh U, van de Winkel JG, Anderson CL. 1991. Expression of IgG Fc receptor antigens in placenta and on endothelial cells in humans. An immunohistochemical study. Am. J. Pathol. 138:175–81 Wang AC, Faulk WP, Stuckey MA, Fudenberg HH. 1970. Chemical differences of adult, fetal and hypogammaglobulinemic IgG immunoglobulins. Immunochemistry 7:703–8 Hay FC, Hull MG, Torrigiani G. 1971. The transfer of human IgG subclasses from mother to foetus. Clin. Exp. Immunol. 9:355–58 Malek A, Sager R, Schneider H. 1994. Maternal-fetal transport of immunoglobulin G and its subclasses during the third trimester of human pregnancy. Am. J. Reprod. Immunol. 32:8–14 Sidiripoulos D, Hermann U, Morell A, Muratt G, Barundun S. 1986. Transplacental passage of intravenous immunoglobulin in the last trimester of pregnancy. J. Pedriatr. 109:505–8 Smith CIE, Hammarstrom L. 1985. Intravenous immunoglobulin in pregnancy. Obstet. Gynecol. 66:39S–40S Malek A, Sager R, Sakher A, Schneider H. 1995. Transport of immunoglobulin G and its subclasses across the in vitro per-

764

99.

100.

101.


102.

103.

104.

105.

106.

107.

GHETIE n WARD fused human placenta. Am. J. Obstet. 108. Junghans RP. 1997. Finally! The Brambell receptor (FcRB). Mediator of transGynecol. 173:760–67 mission of immunity and protection from Sooranna SR Contractor SF. 1991. Veccatabolism for IgG. Immunol. Res. torial transcytosis of immunoglobulin G 16:29–57 by human term trophoblast cells in cul109. Mariani G, Strober W. 1990. Immunoture. Exp. Cell Res. 192:41–45 globulin Metabolism, Fc Receptors, and Brambell FWR, Hemmings WA, Morris the Action of Antibodies, ed. H Metzger, IG. 1964. A theoretical model of gampp. 94–177. Washington, DC: Am. Soc. maglobulin catabolism. Nature 203: Microbiol. 1352–55 Brambell FWR. 1970. The Transmission 110. Borvak J, Richardson J, Medesan C, Antohe F, Radu C, Simionescu M, Ghetie of Passive Immunity from Mother to V, Ward ES. 1998. Functional expression Young. Amsterdam: North Holland of the MHC class I-related receptor, Kim JK, Tsen MF, Ghetie V, Ward ES. FcRn, in endothelial cells of mice. Int. 1994. Identifying amino acid residues Immunol. 10:1289–98 that influence plasma clearance of murine IgG1 fragments by site-directed 111. Gitlin D, Kumate J, Urrusti J, Morales C. 1964. The selectivity of the human plamutagenesis. Eur. J. Immunol. 24:542– centa in the transfer of plasma proteins 48 from mother to fetus. J. Clin. Invest. Spriggs MK, Koller BH, Sato T, Morris43:1938–51 sey PJ, Fanslow WC, Smithies O, Voice RF, Widmer MB, Maliszewski CR. 1992. 112. Contractor SF, Eaton BM, Stannard PJ. 1983. Uptake and fate of exogenous Beta 2-microglobulin-, CD8` T-cellimmunoglobulin G in the perfused deficient mice survive inoculation with human placenta. J. Reprod. Immunol. high doses of vaccinia virus and exhibit 5:265–73 altered IgG responses. Proc. Natl. Acad. 113. Israel EJ, Taylor S, Wu Z, Mizoguchi E, Sci. USA 89:6070–74 Blumberg RS, Bhan A, Simister NE. Liu Z, Roopenian DC, Zhou X, Chris1997. Expression of the neonatal Fc tianson GJ, Diaz LA, Sedmak DD, receptor, FcRn, on human intestinal epiAnderson CL. 1997. Beta2-microglobuthelial cells. Immunology 92:69–74 lin-deficient mice are resistant to bullous 113a. Dickinson BL, Badizadegan K, Wu Z, pemphigoid. J. Exp. Med. 186:777–83 Ahouse X, Simister NE, Blumberg RS, Mozes E, Kohn LD, Hakim F, Singer DS. Lencer WI. 1999. Bidirectional FcRn1993. Resistance of MHC class I-defidependent IgG transport in a polarized cient mice to experimental systemic human intestinal epithelial cell line. J. lupus erythematosus. Science 261:91–93 Clin. Invest. 104:903–11 Chan OT, Hannum LG, Haberman AM, Madaio MP, Shlomchik MJ. 1999. A 114. Besser TE, McGuire TC, Gay CC, Pritchett LC. 1988. Transfer of functional novel mouse with B cells but lacking immunoglobulin G (IgG) antibody into serum antibody reveals an antibodythe gastrointestinal tract accounts for IgG independent role for B cells in murine clearance in calves. J. Virol. 62:2234–37 lupus. J. Exp. Med. 189:1639–48 Simister NE, Mostov KE. 1989. Cloning 115. Blumberg RS, Koss T, Story CM, Barisani D, Polischuk J, Lipin A, Pablo L, and expression of the neonatal rat intesGreen R, Simister NE. 1995. A major tinal Fc receptor, a major histocompatihistocompatibility complex class Ibility complex class I antigen homolog. related Fc receptor for IgG on rat hepaCold Spring Harb. Symp. Quant. Biol. tocytes. J. Clin. Invest. 95:2397–402 54(1):571–80


MHC CLASS I–RELATED RECEPTOR FcRn 116. Cianga P, Medesan C, Richardson J, Ghetie V, Ward ES. 1999. Identification and function of neonatal Fc receptor in mammary gland of lactating mice. Eur. J. Immunol. 29:2515–23 117. Mehta PD, Isaacs CE, Coyle PK. 1991. Immunoglobulin G subclasses in human colostrum and milk. In Immunology of Milk and the Neonate, ed. J Mestecky, C Blair, PL Ogra, pp. 223–26. New York and London: Plenum 118. Hunziker W, Fumey C. 1994. A di-leucine motif mediates endocytosis and basolateral sorting of macrophage IgG Fc receptors in MDCK cells. EMBO J. 13:2963–67 119. Letourneur F, Klausner RD. 1992. A novel di-leucine motif and a tyrosinebased motif independently mediate lysosomal targeting and endocytosis of CD3 chains. Cell 69:1143–57 120. Simmen T, Nobile M, Bonifacino JS, Hunziker W. 1999. Basolateral sorting of furin in MDCK cells requires a phenylalanine-isoleucine motif together with an acidic amino acid cluster. Mol. Cell Biol. 19:3136–44 121. Mostov KE, Deitcher DL. 1986. Polymeric immunoglobulin receptor expressed in MDCK cells transcytoses IgA. Cell 46:613–21 122. Hunziker W, Mellman I. 1989. Expression of macrophage-lymphocyte Fc receptors in Madin-Darby canine kidney cells: polarity and transcytosis differ for isoforms with or without coated pit localization domains. J. Cell Biol. 109:3291– 302 123. Sell S, Fahey JL. 1964. Relationship between gamma-globulin metabolism and low serum gamma-globulin in germfree mice. J. Immunol. 93:81–87 124. Kandil E, Noguchi M, Ishibashi T, Kasahara M. 1995. Structural and phylogenetic analysis of the MHC class I-like Fc receptor gene. J. Immunol. 154:5907–18 125. Hansborough JR, Miller LN, Field TO, Gadd NA. 1988. High dose intravenous

126.

127.

128.

129.

130.

131.

132.

133.

134.

765

immunoglobulin therapy in burn patients. pharmacokinetics and effects on microbial opsonization and phagocytosis. Pediatr. Infect. Dis. 7:S49–S46 Wochner RD, Drews G, Strober W, Waldmann TA. 1966. Accelerated breakdown of immunoglobulin G (IgG) in myotonic dystrophy: a hereditary error of immunoglobulin catabolism. J. Clin. Invest. 45:321–29 Waldmann TA, Miller EJ, Terry WD. 1968. Hypercatabolism of IgG and albumin: a new familial disorder. Clin. Res. 16:45 Yu Z, Lennon VA. 1999. Mechanism of intravenous immune globulin therapy in antibody-mediated autoimmune diseases. N. Engl. J. Med. 340:227–28 Masson PL. 1993. Elimination of infectious antigens and increase of IgG catabolism as possible modes of action of IVIg. J. Autoimmun. 6:683–89 Geha RS, Rosen FS. 1996. Intravenous immune globulin therapy. In Therapeutic Immunology, ed. KF Austen, SJ Burakoff, FS Rosen, TB Strom, pp. 280–96. Cambridge, MA: Blackwell Science Martin MG, Wu SV, Walsh JH. 1993. Hormonal control of intestinal Fc receptor gene expression and immunoglobulin transport in suckling rats. J. Clin. Invest. 91:2844–49 Levy AL, Waldmann TA. 1970. The effect of hydrocortisone on immunoglobulin metabolism. J. Clin. Invest. 49: 1679–84 Riechmann L, Holliger P. 1997. Antibodies live long and prosper. Nat. Biotechnol. 15:617 Baselga JD, Tripathy D, Mendelsohn J, Baughman S, Benz CC, Dantis L, Sklarin NT, Seidman AD, Hudis CA, Moore J, Rosen PP, Twaddell T, Henderson IC, Norton L. 1996. Phase II study of weekly intravenous recombinant humanized anti-p185HER2 monoclonal Ab in patients with HER2/neu-overexpressing

766

135.

136.


137.

138.

GHETIE n WARD metastatic breast cancer. J. Clin. Oncol. 14:737–44 Berinstein NL, Grillo-Lopez AJ, White CA, Bence-Bruckler I, Maloney D, Czuczman M, Green D, Rosenberg J, McLaughlin P, Shen D. 1998. Association of serum Rituximab (IDEC-C2B8) concentration and anti-tumor response in the treatment of recurrent low-grade or follicular non-Hodgkin’s lymphoma. Ann. Oncol. 9:995–1001 Uhr JW, Baumann J. 1961. Antibody formation: I. The suppression of antibody formation by passively administered antibody. J. Exp. Med. 113:395–97 Celis E, Zurawski VRJ, Chang TW. 1984. Regulation of T-cell function by antibodies: enhancement of the response of human T-cell clones to hepatitis B surface antigen by antigen-specific monoclonal antibodies. Proc. Natl. Acad. Sci. USA 81:6846–50 Schalke BC, Klinkert WE, Wekerle H, Dwyer DS. 1985. Enhanced activation of a T cell line specific for acetylcholine receptor (AChR) by using anti-AChR monoclonal antibodies plus receptors. J. Immunol. 134:3643–48

139. Jemmerson R, Johnson JG, Burrell E, Taylor PS, Jenkins MK. 1991. A monoclonal antibody specific for a cytochrome c T cell stimulatory peptide inhibits T cell responses and affects the way the peptide associates with antigen-presenting cells. Eur. J. Immunol. 21:143–51 140. Watts C, Lanzavecchia A. 1993. Suppressive effect of antibody on processing of T cell epitopes. J. Exp. Med. 178:1459–63 141. Siegrist CA, Barrios C, Martinez X, Brandt C, Berney M, Cordova M, Kovarik J, Lambert PH. 1998. Influence of maternal antibodies on vaccine responses: inhibition of antibody but not T cell responses allows successful early prime-boost strategies in mice. Eur. J. Immunol. 28:4138–48 142. Simionescu M, Simionescu N. 1991. Endothelial transport of macromolecules: transcytosis and endocytosis. Cell Biol. Rev. 25:1–78 143. Martin WL, Bjorkman PJ. 1999. Characterization of the 2:1 complex between the class I MHC related Fc receptor and its Fc ligand in solution. Biochemistry 38:12639–47


GHETIE

■

WARD C-1

Figure 1 Alpha-carbon trace of the Fc region of human IgG1 (33). Residues shown to play a role in the FcRn:IgG interaction using both binding and in vivo assays (29-32, 39) are indicated (Ile253, green; Pro257, red; His310, blue; His435-Tyr436, black; note that the role of Tyr436 in human IgG1 has not been investigated, but His436 in mouse IgG1 is involved). Residues 307 and 309 play a less significant role (39) and are therefore not highlighted. The figure was drawn using RASMOL (Roger Sayle, Bioinformatics Research Institute, University of Edinburgh, Edinburgh, U.K.).


C-2 GHETIE

■

WARD

Figure 2 Alpha-carbon trace of rat FcRn with α1 and α2 domain helices perpendicular to the page (18). Indicated are the α2 domain residues shown by site-directed mutagenesis studies (64) to be involved in binding to rodent IgGs (Glu117, blue; Glu132-Trp133, red; Glu135, black; Asp137, green). The figure was drawn with RASMOL (Roger Sayle, Bioinformatics Research Institute, University of Edinburgh, Edinburgh, U.K.).


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:767–811 Copyright q 2000 by annual Reviews. All rights reserved

IMMUNOBIOLOGY OF DENDRITIC CELLS


Jacques Banchereau1, Francine Briere2, Christophe Caux2, Jean Davoust1, Serge Lebecque2, Yong-Jun Liu3, Bali Pulendran1, and Karolina Palucka1 1 Baylor Institute for Immunology Research, Dallas, Texas 75204; e-mail: [email protected], [email protected], [email protected], [email protected] 2 Laboratory of Immunological Research, Schering-Plough, Dardilly 69572, France; e-mail: [email protected], [email protected], [email protected] 3 Department of Immunobiology, DNAX Research Institute, Palo Alto, California 94304; e-mail: [email protected]

Key Words dendritic cell subsets, innate immunity, antigen presentation, chemokines, T cell priming, B cell differentiation Abstract Dendritic cells (DCs) are antigen-presenting cells with a unique ability to induce primary immune responses. DCs capture and transfer information from the outside world to the cells of the adaptive immune system. DCs are not only critical for the induction of primary immune responses, but may also be important for the induction of immunological tolerance, as well as for the regulation of the type of T cell–mediated immune response. Although our understanding of DC biology is still in its infancy, we are now beginning to use DC-based immunotherapy protocols to elicit immunity against cancer and infectious diseases. Host defense relies on a concerted action of both antigen (Ag)–nonspecific innate immunity and Ag-specific adaptive immunity (1–3). Key features of the mammalian innate immune system include (a) the ability to rapidly recognize pathogen and/or tissue injury and (b) the ability to signal the presence of danger to cells of the adaptive immune system (4). The innate system includes phagocytic cells, natural killer (NK) cells, complement, and interferons (IFNs). Cells of the innate system use a variety of pattern recognition receptors to recognize patterns shared between pathogens, for instance bacterial lipopolysaccharide (LPS), carbohydrates, and double-stranded viral RNA (5–7). Evolutionary pressure has led to development of adaptive immunity, the key features of which are (a) the ability to rearrange genes of the immunoglobulin family, permitting creation of a large diversity of Ag-specific clones and (b) immunological memory. Yet this highly sophisticated and potent system needs to be instructed and regulated by Ag-presenting cells (APCs). Dendritic cells (DCs) are unique APCs because they are the only ones that are able to induce primary immune responses, thus permitting establishment of immunological memory (8–11). DC progenitors in the bone marrow give rise to circulating precursors that home to tissues, where they reside as immature cells with high phagocytic capacity (Figure 1). Fol0732–0582/00/0410–0767/$14.00

767

768

BANCHEREAU ET AL


lowing tissue damage, immature DCs capture Ag and subsequently migrate to the lymphoid organs, where they select rare Ag-specific T cells, thereby initiating immune responses. DCs present Ag to CD4` T-helper cells, which in turn regulate the immune effectors, including Ag-specific CD8` cytotoxic T cells and B cells, as well as non– Ag-specific macrophages, eosinophils (12), and NK cells. Moreover, DCs educate effector cells to home to the site of tissue injury. Four stages of development have

Figure 1 The life cycle of dendritic cells (DC). Circulating precursor DCs enter tissues as immature DCs. They can also directly encounter pathogens (e.g. viruses) that induce secretion of cytokines (e.g. IFNa), which in turn can activate eosinophils, macrophages (MF), and natural killer (NK) cells. After antigen capture, immature DCs migrate to lymphoid organs where, after maturation, they display peptide-major histocompatibility complexes, which allow selection of rare circulating antigen-specific lymphocytes. These activated T cells help DCs in terminal maturation, which allows lymphocyte expansion and differentiation. Activated T lymphocytes migrate and can reach the injured tissue, because they can traverse inflamed epithelia. Helper T cells secrete cytokines, which permit activation of macrophages, NK cells, and eosinophils. Cytotoxic T cells eventually lyse the infected cells. B cells become activated after contact with T cells and DCs and then migrate into various areas where they mature into plasma cells, which produce antibodies that neutralize the initial pathogen. It is believed that, after interaction with lymphocytes, DCs die by apoptosis.


769

been delineated, including (a) bone marrow progenitors; (b) precursor DCs that are patrolling through blood and lymphatics as well as lymphoid tissues, and that, upon pathogen recognition, release large amounts of cytokines, e.g. IFN-a, thereby limiting the spread of infection; (c) tissue-residing immature DCs, which possess high endocytic and phagocytic capacity permitting Ag capture; and (d) mature DCs, present within secondary lymphoid organs, that express high levels of costimulatory molecules permitting Ag presentation (Figure 1). DCs constitute a complex system of cells which, under different microenvironmental conditions, can induce such contrasting states as immunity and tolerance. This review summarizes recent progress in our understanding of DC development and immunoregulatory functions. For earlier references regarding DC biology, the reader is invited to consult the most recent reviews (8–11).

HETEROGENEITY OF DENDRITIC CELL SUBSETS Annu. Rev. Immunol. 2000.18:767-811. Downloaded from arjournals.annualreviews.org by HINARI on 09/01/07. For personal use only.

Mice At least two distinct pathways of DC development have been identified in mice, myeloid and lymphoid. Evidence for the myeloid origin of DCs comes mainly from in vitro studies in which myeloid-committed precursors give rise to both granulocytes/monocytes and myeloid DCs under the influence of granulocyte/ macrophage colony-stimulating factor (GM-CSF) (13, 14). DCs can also arise from lymphoid-committed precursors (15–19). Transfer of a population of purified lymphoid precursors into irradiated hosts results in the development of T cells, B cells, NK cells, and DCs that express CD8a, but not cells of the myeloid lineage (15, 19). The ability to generate DC but not NK or B cells is maintained by the downstream CD44`CD25`CD31 pre–T cell population, suggesting that DCs are more closely linked to T cells than to NK or B cells (19, 20). However, a strict clonal analysis showing that DCs arise from the same precursor cells as NK cells or T cells has not yet been done. Although the lymphoid origin of DCs has been demonstrated only for CD8a`, the similarity in phenotype of thymic DCs to CD8a` splenic and lymph node DCs (21) suggests a common origin. Lymphoid and myeloid DCs differ in phenotype, localization, and function. Both subsets express high levels of CD11c, class II major histocompatibility complex (MHC), and the costimulatory molecules CD86 and CD40. To date, the most reliable marker to distinguishing these two subsets is CD8a, which is expressed as a homodimer on the lymphoid DC, but is absent from the myeloid subset (19, 21–23). Other markers such as DEC-205 and CD1d are expressed at higher levels on lymphoid DCs, but they can be upregulated on myeloid DCs by in vitro culture (19, 21–23) or LPS treatment (B Pulendran et al, unpublished observations). Lymphoid DCs are localized in the T cell–rich areas of the periarteriolar lymphatic sheaths (PALS) in the spleen and lymph nodes (17, 23–25). In contrast, myeloid DCs are in the marginal zone bridging channels of the spleen (17, 23–25) but can be induced to migrate to the PALS under the influence of proinflammatory signals such as LPS (24) or parasite extracts (26). The lymphoid


770

BANCHEREAU ET AL

DCs make higher levels of interleukin (IL)-12 (23, 26–28) and are less phagocytic than myeloid DCs (23, 25). IL-12 induces production of IFNc in lymphoid but not in myeloid DCs (28). In vitro, the lymphoid DCs were reported to prime allogeneic CD4 and CD8 T cells less efficiently than myeloid DCs (29, 30). Although the interpretation could be that lymphoid DCs may downregulate T cell responses, the in vivo demonstration of this hypothesis is still pending. In vivo, both lymphoid and myeloid DCs appear to prime Ag-specific CD4` T cells efficiently (27, 31; see below). Flt3 ligand (Flt3-L) and GM-CSF can expand mature DC in mice. Flt3-L targets primitive hematopoietic progenitors in the bone marrow, inducing their expansion and differentiation (32), and both lymphoid and myeloid DC numbers increase dramatically upon Flt3-L injection (22, 23, 33). Flt3-L treatment leads to an increase in DC numbers in multiple organs in mice, including spleen, lymph nodes, blood, thymus, Peyer’s patch, liver, and lungs. In contrast, GM-CSF preferentially expands the myeloid DC subset in vivo (31). In mice, the relationship between Langerhans cells (LCs) and lymphoid and myeloid DCs is not completely understood. LCs do share many common markers with myeloid DCs in the dermis (34). Studies of skin sensitization suggest that a subset of myeloid DC found in the draining lymph nodes represents LCs that have migrated there from the skin (35). LC development seems to be critically dependent on transforming growth factor (TGF)-b because TGF-b knockout mice are devoid of LCs, but not of their precursors (36). Genetic evidence for separate pathways of DC development comes from the study of mice that are deficient in certain genes. Thus RelB 1/1 mice are deficient in myeloid DCs; mice bearing a mutant Ikaros gene are deficient in lymphoid DCs (37–40).

Humans DC heterogeneity in humans is reflected at four levels. (a) Precursor Populations. For instance, in humans, at least two subsets of DC precursors circulate in the blood: CD14` CD11c` monocytes and lineage-negative (LINneg) CD11c1 IL3Ra` precursor DCs (41–44; Figure 2). The LINneg CD11c` cells may represent a third precursor, although these cells are more committed because they can spontaneously differentiate into DCs when put into culture. (b) Anatomical Localization. The level of heterogeneity reflected by anatomical localization includes skin epidermal LCs, dermal (interstitial) DCs (intDCs), splenic marginal DCs, Tzone interdigitating cells, germinal-center DCs, thymic DCs, liver DCs, and blood DCs. Although certain phenotypic differences have been observed among these different DC subsets, their lineage origins, maturation stages, and functional differences have not been clearly established. (c) Function. Both murine and human DC subsets exert different functions, particularly in the regulation of B cell proliferation and differentiation of T cells toward type 1 or type 2, as discussed later. (d) The Final Outcome of Immune Response. The final outcome of immune response refers to the induction of tolerance or immunity.



771

Figure 2 Three subsets of human dendritic cells (DC) and related macrophages. Myeloid CD34` progenitors differentiate into monocytes (CD14` CD11c` DC precursors) that yield the immature DCs in response to granulocyte/macrophage colony-stimulating factor– positive (GM-CSF) interleukin (IL)-4 and macrophages in response to macrophage colony stimulating factor (M-CSF) (the interstitial pathway). Myeloid progenitors also differentiate into CD11c` CD141 precursors, which yield Langerhans cells in response to GMCSF and IL-4 and transforming growth factor (TGF) b, and macrophages in response to M-CSF. Note that these later precursors can spontaneously differentiate into DCs in cultures. The CD141 CD11c1 IL-3Ra` DC precursor (also called pDC2, IFNa-producing cell, or plasmacytoid T cell; a possible equivalent to the murine lymphoid DCs) may originate from the lymphoid CD34` progenitor. A blood cell population with a comparable phenotype has been shown to yield T cells in fetal thymic organ cultures. CD11c1 IL3Ra` DC precursors differentiate into immature DCs in response to IL-3. The immature cells differentiate to mature cells in response to cytokines (MCM, monocyte-conditioned medium) or pathogen products [lipopolysaccharide (LPS) or DNA].

A large body of literature has recently accumulated concerning the origin and in vitro differentiation pathways of DCs (9, 10, 45). Overall, lymphoid and myeloid DCs have been characterized, although the existence of human lymphoid DCs is somewhat controversial (46, 47). The CD11c` CD14` monocytes (48– 50), as well as LINneg CD11c` blood DCs (51) give rise to immature DCs under the influence of GM-CSF and IL-4 or tumor necrosis factor (TNF). Furthermore, CD11c` blood DCs can differentiate into LCs in the presence of TGF-b (51). In


772

BANCHEREAU ET AL

fact, TGF-b appears to represent a major factor for LC differentiation both in mice and humans (36, 52). Both monocytes and CD11c` blood DCs can give rise to macrophages if cultured with GM-CSF or M-CSF, suggesting the plasticity of the DC system (53, 54). Distinct factors regulate the survival and differentiation of CD11c1 DC precursors, originally described as plasmacytoid T cells or plasmacytoid monocytes (41, 42, 55–58). These cells die rapidly after isolation and are critically dependent on IL-3 for survival and CD40-L for maturation. Phenotypically similar populations of adult blood cells have been shown to express the pre–T cell receptor and to contain precursors of mature CD4` TCRa/b` T cells, hence the presumption of the lymphoid origin of CD11c1 DC (59). Monocytes and CD11c1 IL-3Ra` DC precursors display many phenotypic differences, and monocytes, but not CD11c1 DC precursors, express significant levels of CD11b, CD13, CD14, CD33, and CD45RO. Whereas monocytes express high GM-CSFRa and low IL-3Ra, CD11c1 DC precursors display reciprocal patterns of cytokine receptor expression, low GM-CSFRa, and high IL-3Ra, consistent with in vitro cytokine responsiveness. Finally, immature monocyte-derived DCs display high endocytic/phagocytic capacity, contrary to immature CD11c1 DCs. The generation of intDCs (60) and LCs from CD34` hematopoietic progenitors is regulated by the same cytokines that drive differentiation of blood precursors (53, 61, 62). Whereas the corresponding mature DC progeny are equally potent in stimulating the proliferation of naive T cells, only intDCs induce the IL-2–driven differentiation of naı¨ve B cells in vitro (62). Although both subsets express IL-12 upon CD40 ligation, intDCs exclusively express IL-10 (63). Finally, intDCs demonstrate a high efficiency of Ag capture, that is ;10-fold higher than that of LCs. intDCs also express high levels of nonspecific esterases, whereas LCs do not. Currently, no biological function specific to LCs has been formally identified. LCs lack functional mannose receptors and are poor stimulators of Ag-specific CD4` T cell clones when compared with monocyte-derived DCs (64). This characteristic, together with observations that CD34` HPC1 derived DCs, composed of both LCs and intDCs, are more potent in the priming of Melan-A/MART-1–specific cytotoxic T lymphocytes (CTLs) than DCs generated from monocytes (65), prompts the hypothesis that the primary function of LCs may be the priming of CD8` T cells. Our current view of the differentiation pathways and maturation stages of DCs and their precursors is summarized in Figure 2. Besides replenishing the pool of tissue-residing immature DCs, circulating DC precursors play a critical role in the immediate reaction to pathogens and in the shaping of immune response. Monocytes have long been recognized as initial effectors of LPS-related inflammatory responses, as well as a limited source of IFNa. The exact nature of IFNa-producing cells has been, however, enigmatic until the recent identification of CD11c1 IL-3Ra` blood DC precursors as a major source of IFNa in response to virus (55, 56, 58, 66). The emerging finding is again the plasticity of the system, illustrated by (a) the specialization of DC precursors to respond to different pathogens, virus, or bacteria; and (b) the dual


773

function of these cells at two distinct stages of differentiation, as exemplified by (i) the ability of precursor DCs to secrete large amounts of proinflammatory and/ or antiviral cytokines and (ii) ability of mature DCs to activate and modulate T cell responses. Thus, to efficiently combat pathogen invasion, DCs link the two branches of the immune system: the Ag-nonspecific innate immunity and the Agspecific adaptive immunity (Figure 1).


INTIMATE LINK BETWEEN ANTIGEN CAPTURE, MIGRATION, AND MATURATION An important attribute of DCs at various differentiation stages is their mobility (67). DCs migrate from bone marrow to peripheral tissue, where their encounter with Ags triggers their migration to the secondary lymphoid organs. There, Agbearing DCs select the Ag-specific lymphocytes from the pool of recirculating T cells. The selective migration of DCs and their residence in nonlymphoid as well as lymphoid organs are tightly regulated events whose molecular control is being rapidly unraveled (Figure 3, see color insert).

Recruitment of Dendritic Cell Precursors Newly generated DCs migrate, presumably through the blood stream, from the bone marrow to nonlymphoid tissues, where they eventually become resident cells. DCs accumulate rapidly (within an hour) at the sites of Ag deposition as demonstrated in bronchial epithelium after Ag inhalation (68, 69). This accumulation likely represents recruitment of circulating DC precursors, in response to the production of chemokines upon local inflammation. In vitro, immature DCs respond to a large spectrum of chemokines through specific receptors (Table 1). Different DC subsets display unique sensitivity to certain chemokines. For instance, CD34` HPC-derived immature DCs express CCR6, whose ligand, MIP3a (also identified as LARC, Exodus-1), appears to be the most powerful chemokine guiding their migration (70, 71). However, MIP-3a has no effect on monocyte-derived immature DCs (70–72), a difference that may be linked either to a putative inhibitory effect of IL-4 on CCR6 expression or to a specific activity of MIP-3a on LCs present within CD34` HPC-derived DCs. MIP-3a expression is restricted to epithelium, as observed in tonsils and gut. Its induction during inflammatory processes (70, 73) might represent a fundamental mechanism for the chemoattraction of immature LCs or their precursors to inflammatory epithelial sites. In this context, MIP-3a displays selective activity for other leukocyte populations (memory T cells, c/d T cells) with skin or gut epithelial tropisms (73, 74). The accumulation of immature DCs, mostly LCs, in the breast carcinoma bed is also associated with the production of MIP-3a by tumor cells (75). During their migration, DCs are involved in several adhesion events. For instance, E-cadherin, uniquely expressed by LCs, permits, through homotypic

774

BANCHEREAU ET AL

TABLE 1 Expression of chemokine receptors by dendritic cellsa Receptor

Ligands

Immature DC CCR1 CCR2 CCR4 CCR5 CCR6 (LC only) CXCR1 CXCR4

MIP-1a, RANTES, MCP-3, MIP-5 MCPs TARC, MDC MIP-1a, MIP-1b, RANTES MIP-3a IL-8 SDF-1

Mature DC CCR7

MIP-3b, SLC (6Ckine)


a

Abbreviations: DC, dendric cells; MIP, macrophage inflammatory protein; RANTES, regulated on activation, normal T expressed and secreted; MCP, monocyte chemoattractant protein; TARC, thymus and activation-regulated chemokine; MDC, monophage derived chemokine; SDF, stromal derived factor; IL, interleukion; SLC, secondary lymphoid-tissue chemokine.

interactions, the residence of LCs in epidermis (75, 76). Ag encounter results in downregulation of E-cadherin that allows LC migration out of the skin (77). The release of type IV collagenase by LCs may facilitate their migration through the basement membranes (78). Likewise, human macrophage elastase, which degrades several components of the extracellular matrix, is highly expressed by DCs and may thus contribute to their migration (S Lebecque, unpublished observation). Analysis of 17,000 genes from a complementary-DNA library constructed from immature monocyte-derived DCs identified the expression of many genes presumably involved in cell migration, including a metalloproteinase with elastolytic activity as well as a DC-specific HAI-2 gene, a serine protease inhibitor of hepatocyte growth factor activator (79). Overall, the differentiation of monocytes to DCs was accompanied by significant changes in the expression of genes related to cell structure and motility.

Antigen Capture Immature DCs are very efficient in Ag capture and can use several pathways, such as (a) macropinocytosis; (b) receptor-mediated endocytosis via C-type lectin receptors (mannose receptor, DEC-205) (64, 80–84) or Fcc receptor types I (CD64) and II (CD32) [uptake of immune complexes or opsonized particles (85)]; and (c) phagocytosis of particles such as latex beads (86), apoptotic and necrotic cell fragments (involving CD36 and avb3 or avb5 integrins) (87–89), viruses, and bacteria including mycobacteria (90, 91), as well as intracellular parasites such as Leishmania major (92). DCs can also internalize the peptide loaded heat shock proteins gp96 and Hsp70 through presently unknown mechanisms (93, 94).


775

Whereas mannosylated Ags are rapidly internalized by intDCs and selectively targeted to a class II processing/presentation pathway, the capacity of LCs to capture mannosylated Ags is somewhat controversial. Freshly isolated murine epidermal LCs can uptake both mannosylated and nonmannosylated Ags (84). However, human LCs lack classical mannose receptors and have poor endocytic capacity and low levels of lysosome markers (64). A novel C-type lectin (called Langerin), recognized by LC-specific monoclonal antibody DCGM4 and seemingly involved in the formation of Birbeck granules, has been recently cloned (95, 96). These differences between LCs and intDCs extend to differential expression of Fce (97) and Fcc receptors, further strengthening the notion of DC functional heterogeneity and its effect on the type of immune response induced.


Migration to Draining Lymphoid Organs and Maturation The antigen/pathogen induces the immature DC to undergo phenotypic and functional changes that culminate in the complete transition from Ag-capturing cell to APC. DC maturation is intimately linked with their migration from the peripheral tissue to the draining lymphoid organs, and therefore these two key events in the life span of DCs are discussed together. Dendritic-Cell Activation and Maturation Several molecules including CD40, TNF-R, and IL-1R have been shown to activate DCs and to trigger their transition from immature Ag-capturing cells to mature Ag-presenting DCs. DC maturation is a continuous process initiated in the periphery upon Ag encounter and/or inflammatory cytokines and completed during the DC–T cell interaction. The molecules involved in T cell–mediated DC maturation are discussed later. Numerous factors induce and/or regulate DC maturation (Figure 4), including (a) pathogen-related molecules such as LPS (91), bacterial DNA (98–100), and double-stranded RNA (101); (b) the balance between proinflammatory and antiinflammatory signals in the local microenvironment, including TNF, IL-1, IL-6, IL-10, TGF-b, and prostaglandins; and (c) T cell–derived signals. The maturation process is associated with several coordinated events such as (a) loss of endocytic/ phagocytic receptors; (b) upregulation of costimulatory molecules CD40, CD58, CD80, and CD86; (c) change in morphology, (d) shift in lysosomal compartments with downregulation of CD68 and upregulation of DC–lysosome-associated membrane protein (DC-LAMP, as discussed later); and (e) change in class II MHC compartments. Morphological changes accompanying DC maturation include a loss of adhesive structures, cytoskeleton reorganization, and acquisition of high cellular motility (102). An important controller of cytoskeleton remodeling may be the actin-bundling protein p55 fascin, expressed at high levels in blood DCs and in interdigitating DCs located in the T cell areas of lymph nodes (103). Indeed, the formation of dendritic projections in maturing LCs can be inhibited by fascin antisense oligonucleotides (104). DC cytoskeleton abnormalities and reduced


776

BANCHEREAU ET AL

Figure 4 Maturation of dendritic cells (DCs). The left side of the scheme shows the factors inducing progression from one stage to another (GM-CSF, granulocyte/macrophage colony-stimulating factor; IL, interleukin; LPS, lipopolysaccaride; TNF, tumor necrosis factor; dsRNA, double-stranded RNA); the right side shows the main properties of each differentiation/maturation stage (IFN, interferon; MHCII, major histocompatibility complex II; MIIC, MHCII-rich compartment; LAMP, lysosome-associated membrane protein).


777


mobility in vitro have been detected in Wiskott-Aldrich syndrome (WAS) (105), an X-linked recessive disorder characterized by thrombocytopenia, eczema, and immunodeficiency. The WAS gene encodes a 502-amino-acid proline-rich protein (WASp) whose transcripts are detectable throughout differentiation from early hematopoietic progenitors to DCs (106). Because the Cdc42/Rac-binding motif of WASp can control cytoskeleton rearrangement, this molecule may be of importance during DC capture of pathogens and establishment of DC–T cell contacts. Among molecules regulating DC activation/maturation, DC immunoreceptor is a calcium-dependent (C-type) lectin (S Lebecque, manuscript in preparation) that displays an intracellular domain containing an immunoregulatory tyrosinebased inhibitory motif (ITIM) characteristic of immunoglobulin superfamily members and membrane lectins (107, 108). Migration of Antigen-Bearing Dendritic Cells Allogeneic skin transplantation models (109–111), as well as injection of labeled DCs (112) or Leishmaniainfected LCs (92) demonstrated that DCs leave the nonlymphoid organs through the afferent lymph. Pathogen products such as LPS and the local production of TNFa or IL-1 (113), all mediators of DC maturation, trigger peripheral DC migration into the T cell area of lymphoid organs. This migration of maturing DCs also involves a coordinated action of several chemokines. After Ag uptake, inflammatory stimuli turn off the response of immature DCs to MIP-3a (and other chemokines specific for immature DCs) through either receptor downregulation or receptor desensitization dependent on autocrine chemokine production (70, 114, 115). Consequently, maturing DCs escape from the local gradient of MIP3a. Upon maturation DCs upregulate a single known chemokine receptor, CCR7 (116), and accordingly acquire responsiveness to MIP-3b (ELC, Exodus 3) and 6Ckine [secondary lymphoid-tissue chemokine (SLC), Exodus 2] (70, 117). Consequently, maturing DCs will leave the inflamed tissues and enter the lymph stream, potentially directed by 6Ckine that is expressed on lymphatic vessels (118, 119). Mature DCs entering the draining lymph nodes will be driven into the paracortical area in response to the production of MIP-3b and/or 6Ckine by cells spread over the T cell zone (70, 120). The newly arriving DCs might themselves become a source of MIP-3b and 6Ckine (70, 114, 120), allowing an amplification and/or a persistence of the chemotactic signal. Because these two chemokines can attract mature DCs and naive T lymphocytes (118, 120, 121), they are likely to play a key role in helping Ag-bearing DCs to encounter specific T cells. The role of MIP-3b and 6Ckine is supported by a natural mutant mouse for 6Ckine (SLC) (122–124) and CCR7-deficient mice, both of which have a specific deficiency in T cell and DC homing into lymph nodes (125). Upon encounters with T cells, which can take place not only in the secondary lymphoid organ but also at the site of tissue injury, DCs receive additional maturation signals from CD40 ligand, RANK/TRANCE, 4–1BB, and OX40 ligand molecules, which induce the release of chemokines such as IL-8, fractalkine (126), and macrophage derived chemokines that attract lymphocytes (127, 128).

778

BANCHEREAU ET AL

ANTIGEN PROCESSING AND PRESENTATION DCs are well equipped to capture and process Ags, and a number of molecules involved in these processes have been identified and are discussed below. Furthermore, evaluation of 3000 sequences randomly cloned out of a library constructed from CD34-derived DCs showed that 20% of the sequences were related to Ag presentation (MHC molecules constituting 7%), and 6% of sequences constituted enzymes including several members of the serine- and metalloprotease families (S Lebecque unpublished observations).


Major Histocompatibility Complex Class II Soluble and particulate Ags are efficiently captured by immature DCs and targeted to MHC class II compartments (49, 80, 83, 129, 130; Figure 5). Immature DCs constantly accumulate MHC class II molecules in lysosome-related intracellular compartments identified as MHC class II–rich compartments (MIICs), with multivesicular and multilamelar structure (131, 132). The captured Ag is directed towards MIICs containing HLA-DM that promotes the catalytic removal of class II–associated invariant chain peptide and enhances peptide binding to MHC class II molecules (133, 134). However, the loading of class II Ags within DC can also occur in the absence of the invariant chain (135). In immature DCs, Ags and macromolecules gain access to mildly acidic prelysosomal MIICs (136), where MHC class II-Ii chain complexes accumulate. The proteolytic degradation of Ii is regulated by the ratio between cathepsin S and its endogeneous inhibitor cystatin C (137). Upon maturation, cystatin C is downregulated, and the activity of cathepsin S increases, promoting Ii degradation and allowing the export of peptide-loaded class II molecules to the cell surface. Whereas, in immature DCs, class II molecules are rapidly internalized and have a short half-life, maturation/inflammatory stimuli lead to a burst of class II synthesis and translocation of the MHC II-peptide complexes to the cell surface where they remain stable for days and are available for recognition by CD4` T cells (102, 130, 138, 139). DC-LAMP, a novel marker exquisitely induced in mature DCs, is localized in lysosomes as well as in MHC class II compartments immediately before the translocation of MHC class II molecules to the cell surface (140), which is coordinated with DC maturation. It is interesting to note that IL10 can block translocation of peptide-class II complexes to DC plasma membrane, in parallel with inhibition of DC maturation (141). The different surface receptors used by DC subsets to capture Ags, as well as subtle differences in proteolytic machinery, may determine the nature of immunodominant peptides presented by MHC class II molecules (142). These variations in Ag processing may permit recruitment of CD4` T cells with diverse TCR specificities and spreading of the response.



779

Figure 5 Dendritic cells load their major histocompatibility complex (MHC) molecules in multiple ways. (Left) MHC class I. Besides the classical endogenous pathway that loads peptides from self and intracellular pathogens, dendritic cells can also load MHC class I antigens through exogenous pathways with peptides originating from phagocytosed particulate antigens or immune complexes. Peptides are generated in the proteasome, transferred into the endoplasmic reticulum (ER), and loaded onto the nascent MHC class I molecules. (Right) MHC class II. Dendritic cells capture soluble antigen (Ag) either through macropinocytosis or receptor-mediated endocytosis. They also capture particles through phagocytosis. The antigens are subsequently degraded in endosomes, and the generated polypeptides are transported into the MHC class II–rich compartments (MIIC) for their loading onto the nascent MHC class II molecules while DCs mature. The invariant chain, associated with nascent MHC class II, is cleaved by cathepsin S (Cath.S), which in immature DCs is inhibited by cystatin C (Cyst.C). After maturation, cystatin C is downregulated, thereby releasing active cathepsin S. The HLA-DM molecules help the loading of peptides onto MHC class II molecules. A fraction of the peptides are loaded onto empty MHC class II molecules recycled from the cell surface (cycle on the right).

Major Histocompatibility Complex Class I To generate CD8` cytotoxic killer cells, DCs present antigenic peptides on MHC class I molecules, which can be loaded through both an endogenous and an exogenous pathway (143, 144; Figure 5). Endogenous Major Histocompatibility Complex Class I Pathway The endogenous MHC class I pathway operates through the degradation of cytosolic proteins and the loading of peptides onto newly synthesized MHC class I molecules within the endoplasmic reticulum. Ag processing occurs first in the cytosol through an

780

BANCHEREAU ET AL


ATP-dependent proteolytic system, which starts by ubiquitin conjugation. DCs, similarly to B cells, constitutively express di-ubiquitin, which could permit more efficient Ag processing (145). This gene, also known as FAT10 (146), encodes a di-ubiquitin protein containing tandem head to tail ubiquitin-like domains, with the conservation of key functional residues. The ubiquitinylated proteins are directed to the proteasome, which cleaves the protein into peptides. The peptides are then translocated into the ER via ATP-dependent TAP1/2 transmembrane transporters and are trimmed into 8–10 mers, which accommodate the MHC class I-binding groove. Cross-Priming and Class I Presentation of Exogenous Antigens Over 20 years ago Bevan observed priming of host MHC class I-restricted CTLs for minor Ags after immunization with cells that lacked the cognate MHC class I molecules (147, 148). This property, termed cross-priming, suggested that minor Ags could be transferred to host cells for presentation by host MHC class I molecules (149). This observation and many others have led to the conclusion that DCs and, to a lesser extent, macrophages have an alternative MHC class I pathway that can present peptides derived from extracellular Ags. Two routes for the exogenous MHC class I pathway have been described, a TAP-independent pathway in which Ag is most likely hydrolyzed in endosomes (150) and a phagosome-to-cytosol pathway (151, 152) that is TAP dependent. This pathway is thought to be involved in immune responses against transplantation Ags (147), particulate Ags (151), tumors (153), and viruses (154). It is also operative in the development of tolerance (155). The engulfment and processing of cell bodies by DCs represent a possible pathway for the loading of MHC class I (87, 88, 156). Indeed, monocytederived DCs loaded with apoptotic bodies obtained from either macrophages or HeLa cells infected with influenza virus stimulate the proliferation of influenzaspecific T cells and the generation of class I–restricted influenza-specific CD8` CTLs (88). Most recently, FccR-mediated capture of immune complexes (157) and exosomes derived either from tumor cells or from tumor-peptide–pulsed DCs (158) were demonstrated as another pathway permitting access to DC MHC class I presentation. Finally, transfer of peptides carried by heat shock proteins, including hsp70 and gp96, allows in vivo development of protective immunity and tumor rejection in murine models (94, 159). Thus, manipulation of the exogenous class I presentation pathway may permit priming of T cells with desired Ag specificity, for instance in tumor immunotherapy.

CD1 Molecules Recent studies have identified the CD1 family as nonclassical, Ag-presenting molecules involved in regulation of T cell responses to microbial lipids and glycolipids-containing Ag (for reviews, see 160, 161). Both endogenous and exogenous lipids can be presented, and this pathway may contribute not only to microbial immunity but to autoimmunity and antitumor responses. CD1 mole-


781


cules, a hallmark of the DC phenotype, constitute a family of b2-microglobulin– associated nonpolymorphic glycoproteins that assemble with a nonprocessed Ag in the endosomal/lysosomal compartments and present Ag in a TAP-independent manner. In humans, four CD1 proteins (CD1a–d) are expressed by myeloid DCs, whereas in mice only CD1d has been identified. CD1 proteins are functionally heterogeneous, and two subgroups can be identified. Subgroup I, including human CD1b–c, can present glycolipids to a large repertoire of T cells. Indeed, mycobacteria-specific, CD1b-restricted CD8à/b TCR T cells have been demonstrated. Binding of the lipids to these CD1 molecules requires endosomal acidification. Subgroup II includes mouse and human CD1d and binds a limited set of Ags (a-galactosyloceramide) and activates a restricted set of T cells as well as NK T cells (162). CD1-restricted presentation appears to also regulate c/d T cells and intestinal intraepithelial lymphocytes. Much remains to be learned about this presentation pathway and the possibilities of its use in vaccine protocols.

ANTIGEN PRESENTATION AND T CELL ACTIVATION T Cell Priming The ability to prime naive CD4` T cells constitutes a unique and critical function of DCs both in vitro and in vivo. Soluble Ag-pulsed DCs elicit potent Ag-specific T-helper responses when injected into mice (163). Demonstration of DC–Thelper-cell interactions in the PALS by immunohistology suggests direct Ag presentation by pulsed DCs (164). However, an alternative indirect pathway may exist whereby apoptotic fragments of exogenous DCs can be phagocytosed, processed, and presented by resident DCs in the PALS (129). The relative contribution of both Ag-presentation pathways to T cell priming in vivo remains to be investigated. In the presence of free Ag, T-helper cells primed by DCs can interact with B cells and stimulate Ag-specific antibody production (165). The extent of CD4 T cell and antibody responses can be dramatically enhanced in vivo by increased DC numbers, as shown recently with Flt3-L (166). Furthermore, the potent immunogenicity of DCs can result in the abrogation of peripheral T cell tolerance against soluble Ags (166), viral Ags (167), tumors (168), and transplant Ags (169), as well as in neonates (170). DCs are equally important in priming naı¨ve CD8` T cells. In vitro, DCs can stimulate the proliferation of allogeneic CD8` T cells (171), directly in the absence of T cell help (172, 173). They can also generate Ag-specific CTLs from naı¨ve precursors (174–176). Strong CTL responses can be induced in vivo by injection of mice with Ag-bearing DCs, including (a) allogeneic DCs (177), (b) peptide-pulsed DCs (178), (c) protein-loaded DCs (179, 180), (d) DCs transfected with DNA (181), (e) DCs expressing virally encoded Ags (182, 183), and ( f ) DCs pulsed with RNA (184). Although DCs can activate CD8` T cells directly (172, 185, 186), they often require CD4–T cell help. In traditional models of CTL

782

BANCHEREAU ET AL

activation, the CD4` T cells and CD8` T cells were thought to recognize Ag on the same Ag-presenting cell. However, in the current model (187–189), the APCs are licensed to activate T-killer cells by T helpers via upregulation of CD40-L on the DCs. Thus, a conditioned DC becomes a temporal bridge between a CD4` T-helper cell and a T-killer cell. In addition to priming, DCs appear essential to maintain survival of naive CD4 T cells (190) and immune T cell memory (191).


Molecules Involved in Dendritic-Cell/T Cell Interaction It remains to be determined whether the unique ability of DCs to prime T cells results from the expression of molecules unique to DCs or from the high density of molecules involved in DC–T cell interactions. MHC products and MHCpeptide complexes are 10- to 100-fold higher on DCs than on other APCs like B cells and monocytes (130). Recognition of MHC-peptide complexes on DCs by Ag-specific TCRs constitutes ‘‘signal one’’ in DC–T cell interaction. DC–T cell clustering is mediated by several adhesion molecules, like integrins b1 and b2 and members of the immunoglobulin superfamily (CD2, CD50, CD54, and CD58) (9, 10). Recently, a high-affinity receptor for intercellular adhesion molecule 3 (with no homology to LFA1) was found specifically expressed on monocyte-derived DCs (Figdor, personal communication). The crucial factor, that constitutes ‘‘signal two,’’ required to sustain T cell activation, is the interaction between costimulatory molecules expressed by DCs and their ligands expressed by T cells. CD86 on DCs is so far the most critical molecule for amplification of T cell responses (192, 193). T cells can activate DCs via CD40 ligand (CD40-L)-CD40 signaling leading to increased expression of CD80/CD86 and cytokine release (IL1, TNF, chemokines, and IL-12) (49, 187–189, 194). Triggering of CD40 on DCs results in upregulation of OX40 ligand (195), which then signals naive T cells to express IL-4 (196) and upregulates the chemokine receptor CXCR-5, whose ligand directs B lymphocytes into follicles (197). Accordingly, expression of the OX40-L transgene into DCs leads to accumulation of CD4 T cells in B follicles. Mature DCs also express 4–1BB ligand (198), which complements the function of OX40-L. 4–1BB is a costimulator expressed primarily on activated CD4` and CD8` T cells (199). 4–1BB costimulatory signals preferentially induce CD8` T cell proliferation and production of IFNc (but not of IL-4) (200), leading to the amplification of in vivo cytotoxic T cell responses in graft-vs-host disease as well as allograft rejection (201). Whether OX40-L and 4–1BB are expressed simultaneously or exclusively by the same DCs remains to be established. Engagement of RANK, a member of the TNFR family, by its ligand (RANKL/TRANCE) expressed on activated T cells, stimulates the secretion of cytokines like IL-1, IL6, and IL-12 by DCs. This results in increased DC survival, by inhibition of DC apoptosis and, in turn, in enhanced proliferative T cell responses in mixed lymphocyte reactions. The demonstration that TRANCE is responsible for the CD40-


783

L–independent T-helper cell activation during viral infection suggests an important and specific role for this molecule during infection (202–204).


Distinct Dendritic-Cell Subsets Elicit Distinct T Helper Cell Responses DC subsets may provide T cells with the different cytokine/molecule microenvironments that determine the classes of immune response, for example, type 1 vs type 2 CD4 helper cell profile. In humans, monocyte-derived CD11c` DCs polarize naive T cells predominantly towards a Th1 profile, whereas the CD11c1 DC subset induces T cells to predominantly produce Th2 cytokines (57). The extent of T cell polarization by CD11c1 DCs may be related to their differentiation/maturation stages (55). Thus, CD11c1 DC precursors may be prone to elicit more of the Th0 cytokine profile, whereas their mature progeny may induce Th2 differentiation. The induced pattern of T cell cytokine secretion is dependent on the DC production of IL-12 (205). Indeed, CD11c`, but not the CD11c1 DC subset, can be induced to secrete IL-12 (57). Skewing is not restricted to CD4 T cells and also applies to CD8 T lymphocytes and NK T cells (206). In mice, the splenic CD8a` lymphoid DC subset primes naive CD4 T cells to make Th1 cytokines, whereas the splenic CD8a1 myeloid DC subset primes naive CD4 T cells to make Th2 cytokines (27, 31). Consistent with this, GM-CSF, which preferentially mobilizes myeloid DCs in mice, elicits mainly IgG1 antibodies in response to soluble Ag, whereas Flt3-L, which mobilizes both lymphoid and myeloid DC subsets (31), also elicits IgG2a antibodies, a Th1 signature. DCs from IL-12–deficient mice fail to induce Th1 responses, suggesting the critical role of IL-12 in lymphoid DCs-induced Th1 responses (27). Lymphoid but not myeloid DCs can be induced to make large amounts of IL-12 (23, 26–28) and IFNc (28). The mechanism by which myeloid DCs induce Th2 cytokines is not established, although IL-13 (207), IL-6 (208), and OX40-L (196) are good candidates. The involvement of CD80 and CD86 in Th1/Th2 polarization remains unclear, but, in some experimental systems, B7.1/CD80 was shown to promote Th1 responses, whereas B7.2/CD86 ligation tended to skew toward Th2 responses (209, 210). Overall, distinct DC subsets exist in mice and humans that differentially skew Th responses.

DC Functions Exhibit Considerable Plasticity Although it is clear that distinct DC subpopulations exhibit distinct functions, there is also evidence that these DC functions can be altered by the cytokine environment. In particular, DCs exhibit considerable plasticity in their ability to skew Th responses, and DCs that normally induce Th1 profiles can be converted to Th2-skewing cells when treated with anti-inflammatory cytokines such as IL-10 and TGFb or with steroids (211) or prostaglandin E2 (212–217). In this context, DC subsets isolated from different organs differently affect Th responses. Thus,

784

BANCHEREAU ET AL


mouse and rat DCs from Peyers patches elicit Th2 responses, whereas those from spleen induce Th1 responses (218). Although the mechanisms underlying these functional differences are currently unknown, these observations may offer an explanation for the distinct immunological outcomes of oral vs systemic administration of Ags. Thus, adjuvants such as LPS or Flt3-L enhance immunological tolerance to orally administered Ags (219), but abrogate tolerance to systemic injections of Ags (166, 220). DC plasticity is also reflected in their differentiation, which may determine the fate of Ag, that is, processing and presentation or degradation. This aspect is exemplified by the potential of macrophage to differentiate to DCs (54), a pathway that may permit high Ag capture (macrophage) and presentation (dendritic cell). The final signal for DC differentiation and maturation may be provided during the migration of DCs across endothelial barriers between the inflamed tissue and lymphatics (221, 222).

Dendritic Cells and Tolerance In mice, thymic DCs are capable of mediating negative selection of T cells in fetal organ cultures (223) and against superantigens in vivo (224, 225). In addition, thymic DCs can induce tolerance to myelin basic protein and limit the development of experimental autoimmune encephalomyelitis (226). The role of thymic DCs in negative selection (but not positive selection) was confirmed by targeted expression of MHC class II molecules on DCs (227). In the periphery, a role of DCs in establishing peripheral T cell tolerance has not yet been formally demonstrated. In fact, the available evidence suggests that DCs can abrogate T cell tolerance against soluble Ags (166), viral Ags (167), tumors (168), and transplant Ags (169) and in neonates (170). However, in vitro work from Shortman’s group suggests that, in mice, both lymphoid and myeloid DCs can stimulate T cells, but that the lymphoid DC subset can limit the proliferation of T cells (29, 30). The lymphoid DCs appear to kill a proportion of the activated CD4` T cells (30), whereas they limit cytokine production of CD8` T cells (29). The relevance of these in vitro findings for the in vivo tolerance induction remains to be established. Finally, DCs are also considered to play an important role in the establishment of transplantation tolerance through the development of microchimerism (228– 230).

REGULATION OF B LYMPHOCYTES Beside activating naive T cells, DCs can directly activate naive and memory B cells. DCs enhance differentiation of CD40-activated memory B cells towards IgG-secreting cells through secretion of the soluble IL6Ra gp80, which complexes to IL-6 (231, 232). DCs also help the differentiation of activated-naive B cells to plasma cells. This help is mediated by IL-12 in synergy with IL-6/soluble



785

IL6Ra. DCs induce surface IgA expression on CD40-activated naive B cells, which is partially mediated by TGFb, but neither IL-10 nor IL-12 appears to be involved (233). The presence of IgA switch circles in CD40-activated B cells cultured with DCs demonstrates the occurrence of DNA recombination. While DCs alone are able to induce CD40-activated naive B cells to express surface IgA, IL-10 is essential for further differentiation into IgA-secreting cells. Moreover, in the presence of IL10 and TGFb, DCs skew CD40-activated naive B cells towards the secretion of both IgA1 and IgA2 subclasses (233). These results suggest the DC-mediated direct activation of naive B cell during the initiation of the immune response and the involvement of DC in the development of mucosal/ humoral immune responses (Figure 6). The germinal center (GC) is the microenvironment that allows the generation of B cell memory. There B cells proliferate and undergo somatic mutation, isotype switching, affinity selection, and differentiation into memory B cells or plasmablasts. The GC also contains T cells, follicular DCs, and GC DCs (GCDCs). It is now clear that GCDCs are quite different from follicular DCs in phenotype and function (234, 235). GCDCs stimulate, in an IL-12–dependent manner, CD40-activated germinal-center B cell proliferation and drive their differentiation towards plasma cells. In addition, GCDCs induce IL10-independent isotype switching towards IgG1. Thus, DC subsets have the capacity to directly regulate B cell responses. To generate a humoral immune response, Ag-specific CD4` T helper and Ag-

Figure 6 Dendritic cells (DCs) directly signal B cells at the time of the ‘‘menage a` trois.’’ CD40-activated DCs produce IL-X (BAFF/BLys?), which enhances the proliferation of CD40-activated B cells. CD40-activated DCs also secrete IL-12 and sgp80, which binds interleukin (IL)-6 (secreted by the B cells and some DCs). These factors, together with IL-2, induce CD40-activated naive B cells to differentiate into plasma cells secreting IgM. CD40-activated DCs also provide B cells the uncharacterized IL-Y, which together with IL-10 and TGFb permits isotype switching towards IgA1 and IgA2.


786

BANCHEREAU ET AL

specific B cells must interact. Within paracortical areas of the secondary lymphoid organs, interdigitating DCs select the rare Ag-specific T and B cells. As recently demonstrated in vivo in the rat, DCs can also capture and retain unprocessed Ag, then transfer it to naive B cells to initiate a specific Th2-associated antibody response (236). This could be the role of the GCDC population localized within GCs and originally described as an ‘‘antigen-transporting cell’’ (237, 238), which could display the Ag to both T and B cells in a meńage a` trois. However, one could consider that a conditioned DC can be a temporal bridge between a CD4` T helper and a B lymphocyte by analogy to recent models in which DCs offer costimulatory signals to CD4 helper T cells and CD8 T cells (187–189). During the extrafollicular reaction, interdigitating DCs could play a role in the induction of an IL-2–dependent IgM plasma cell differentiation. GC formation starts with the migration of GC founder cells in the follicles and involves Th2 CD4` T cells. CD40 activation upregulates OX40-L expression on DC and B cells (117, 195, 239) and early OX40 ligation promotes Th2 cytokine secretion (196) and causes CD4 T cell migration within B cell follicles (197). Thus, GCDC may contribute to the GC reaction and the role of OX40–OX40-L needs to be analyzed. A novel member of the TNF family, designated BAFF/Blys-L (for B cellactivating factor belonging to the TNF family) and found on DCs and T cells, binds to a receptor restricted to B cells (185, 240, 241) and induces both proliferation and immunoglobulin secretion by different B cell subsets. BAFF may represent an important costimulator through which DCs regulate B cell proliferation and function. Like TNFa and FasL, BAFF/Blys is a transmembrane molecule that is processed and secreted by a protease yet to be identified. Decysin, a novel disintegrin-metalloproteinase isolated from GCDC and specific to mature DCs, represents a candidate for the cleavage of molecules of the TNF family and may thus play an important role in the regulation of T- and B cell functions (242).

DENDRITIC CELLS AND EFFECTORS OF INNATE IMMUNITY DCs at different stages of differentiation can regulate effectors of innate immunity such as NK cells and NK T cells. Both direct cell-cell interactions and indirect cytokine-mediated interactions have been implicated (Figure 7). Precursors of CD11c1 DCs may activate NK cells through the release of IFN-a, thereby leading to enhanced antiviral and antitumor activity of NK cells (56, 58, 138). DCs at later stages of differentiation may regulate the activity of NK/NK T cells through the release of IL-12, IL-15 and IL-18 (243, 244). Both murine and human NK T cells produce high amounts of IL-4 or IFN-c and thus may determine the type of induced immune responses. On recognition of the a-galactosyloceramide (a-GalCer)/CD1d complex, NK T cells release IFNc (162). However, DC subsets differentially regulate NK T cytokine profiles, with



787

Figure 7 Dendritic cells as a link between innate immunity and adaptive immunity in antitumor immune responses (a hypothesis). Precursor DCs recognize tumor pathogenassociated molecular patterns (PAMP) through their pattern recognition receptors (PRR). Consequently, DCs release interferon (IFN) a, which activates macrophages (MF), natural killer (NK) T cells, and NK cells that kill tumors leading to the release of tumor cell bodies. The cell bodies are captured by immature DCs (which may be the progeny of the initial precursor), which will mature and display tumor antigens for selection of tumorspecific lymphocytes. CD8` T cells will further directly kill the tumor while selected CD4 T cells will activate macrophages, NK cells, and eosinophils as discussed in Figure 1. The tumor may affect this process at various stages, by either preventing DC maturation or skewing the T cell responses towards the type 2.

monocyte-derived DCs promoting IFN-c release, whereas plasmacytoid DCs polarize NK T cells to IL-4 production (206). Because a-GalCer or related glycolipids are expressed in bacteria and tumors and because tumor-protective responses can be induced by a-GalCer–activated NK T cells, NK T cells may constitute an important effector mechanism of innate immunity (245). NK cells can also be activated, directly or indirectly, by DCs (246, 247). In murine tumor models, DC transfer or in vivo mobilization with FLT3-L resulted in the NK-mediated rejection of MHC class I-negative tumors. Thus, it is likely that in vivo expansion of both NK cells and DCs by FLT3-L may account for the potent antitumor activity of FLT3-L (248, 249). This synergism could also apply to viral infections or transplantation. DCs can trigger NK cells, through the release of IL-12, to CD28-dependent and -independent cytotoxicity (250). This may result in elimination of B7-expressing cells, including autologous DCs, thus permitting either downregulation of response or amplification of the response via

788

BANCHEREAU ET AL

cross-presentation of Ags released from dying DCs. Finally, activated NK cells may also elicit positive regulatory signals towards immature DCs, by promoting DC maturation in the spleen marginal zones (244, 249).


DENDRITIC CELLS IN TUMOR IMMUNOLOGY The immune system has the potential to eliminate neoplastic cells, as evidenced by occasional spontaneous remissions in renal-cell carcinomas and melanomas (9, 251, 252). Perhaps the most compelling evidence of active in vivo tumorrelated immune responses arises from the study of paraneoplastic neurologic disorders that led to the discovery of onconeural Ags (253, 254). Paraneoplastic neurologic disorders are a rare group of neuronal degenerative diseases that develop as remote effects of systemic malignancies (253, 254). The discovery of onconeural antibodies led to the proposal that paraneoplastic cerebellar degeneration, associated with breast and ovarian cancer, is an autoimmune disorder mediated by the humoral arm of the immune system. These antibodies permitted the cloning of the cdr2 Ag, a protein with a coil/leucine zipper domain. Furthermore, the presence of cdr2-specific CD8` CTLs circulating in the blood of these patients has been demonstrated (88). The list of onconeural Ags is growing (254). The induction of tumor immunity can be initiated by the effectors of innate immunity and further developed by cells of adaptive immunity, with DCs playing a central regulatory role (Figure 7). Several steps are involved, including (a) recognition of tumor molecules by DC precursors, (b) direct and IFN-c–mediated killing of transformed cells by NK/NK T cells activated by DCs, (c) capture and cross-presentation of released-tumor-associated Ags (TAAs) by immature DCs, (d) selection and activation of TAA-specific T cells, as well as nonspecific effectors including macrophages and eosinophils, and (e) homing of TAA-specific T cells to the tumor site and recognition of restriction elements leading to the elimination of tumor cells. Tumors may escape immune surveillance owing to alterations at each of these steps (9, 252). Thus, by release of cytokines such as IL-6, IL-10, M-CSF, and vascular endothelial growth factor, tumors can prevent DC differentiation and/or APC function (255). Indeed, tumor-associated DCs are usually of a low allostimulatory capacity, particularly if isolated from the progressing metastatic lesions, as in malignant melanoma, or from blood, as in patients with advanced breast cancer. Furthermore, IL-10 is capable of converting DC-APC function to the induction of Ag-specific anergy, thus leading to the state of tolerance against tumor tissue (256–258). Analysis of tumor tissue distribution of DCs in breast carcinoma revealed two levels of heterogeneity: (a) immature CD1a` DCs, mostly of the LC type (Langerin`) are retained within the tumor bed in .90% samples, (b) mature DCs— CD83` DC-LAMP` DCs present in 60% of samples—are confined to peritumoral areas. The high numbers of immature DCs found in the tumor may best be explained by high levels of MIP3a expression by virtually all tumor cells,



789

as discussed earlier. In some cases, T cells cluster around the mature DCs in peritumoral areas, thus resembling the DC/T cell clusters of secondary lymphoid organs, which are characteristic of ongoing immune reactions (75). The unique ability of DCs to induce and sustain primary immune responses makes them optimal candidates for vaccination protocols in cancer (17, 251, 259). DCs loaded with appropriate TAAs can induce protective/rejection-immune responses in animal models (182, 260–266), and promising preliminary data are reported in humans (267–270). Several systems have been used to deliver TAA to DCs, including (a) defined peptides of known sequences, (b) undefined acideluted peptides from autologous tumors, (c) whole tumor lysates, (d) retroviral and adenoviral vectors, (e) tumor cell-derived RNA, (f) fusion of DCs with tumor cells, and (g) exosomes derived from DCs pulsed with tumor peptides (subcellular structures containing high levels of MHC molecules and peptides) (9, 251, 252). However, DC-mediated induction of immunity represents a challenge, and several parameters need to be considered to ensure the optimal outcome of DC-based vaccination protocols including (a) the source of TAA, (b) the methods for TAA preparation and loading, and (c) the diversity of DC subsets. These aspects are discussed elsewhere in this same volume (L Fong & E Engelman, ARI 18:245– 73).

DENDRITIC CELLS AND PATHOGENS DCs have evolved to identify danger represented either as tissue damage or as a microbial invasion. Microbial products such as LPS or CpG DNA may activate innate immune mechanisms, including DCs, leading to responses beneficial to the host. However, pathogens have devised multiple strategies to evade immune responses by altering each step in the response, including inhibition of APC maturation and function (271), interference with MHC class I and class II processing/ presentation pathways (viruses and bacteria), and hyperactivation of T cells (bacterial superantigens) to name a few. Recent studies are uncovering how pathogens can escape Ag-presenting functions of macrophages and DCs. Bacterial LPS is a major molecule recognized by the innate immune system (6, 7). Ligation of membrane CD14 by complexes of LPS and soluble LPSbinding protein leads to proinflammatory signals including TNF and IL-1 secretion, which, when released at the site of tissue damage, increase the turnover of local APCs as well as recruitment of precursor cells (272). Toll-like receptor-4 (TLR4) transduces signal in LPS responses leading to NF-jB activation (5), and TLR4-deficient mice are hyporesponsive to LPS. TLRs contain a region in the intracellular domain that is homologous to part of the IL-1R and that is involved in the activation of IL-1R-associated kinase (273), leading to activation of several down-stream effectors. Furthermore, TLR2 but not TLR4 mediates responses elicited by components of gram-positive bacteria, such as peptidoglycan and lipoteichoic acid (274, 275). Although human monocytes transcribe Toll RNA and


790

BANCHEREAU ET AL

Figure 8 The plasticity of dendritic cells (DCs). According to microenvironmental instructions (e.g. cytokines), DC functions can be altered. For instance DCs can become macrophages (MF) with higher phagocytic functions in response to macrophage–colonystimulating factor (M-CSF). Furthermore, interleukin (IL)-12 secreting DCs that induce type 1 T cell differentiation can switch, in response to IL-10 and prostaglandin E2, to DCs inducing T cell anergy or T cell differentiation into type 2 cells or regulatory T cells.

human macrophages release IL-12 on Toll-2 signaling, little is know regarding expression or function of Toll-like receptors on DC subsets (5). After in vitro or in vivo exposure to LPS or other bacterial products, DCs undergo activation and maturation (91). In vitro, bacteria-induced DC maturation involves two signaling pathways: (a) ERK kinase, allowing for DC survival, and (b) NF-jB, allowing for DC maturation characterized by increased expression of costimulatory and MHC-class II molecules, release of chemokines, and migration. This coordinated process leads to high T cell–stimulatory capacity as well as IL12 release (91), all of which result in the induction of protective immune responses. DCs also initiate immune responses against parasites such as Leishmania. Immature DCs can phagocytose the organism, and LCs infected by Leishmania. are present in the dermal infiltrate of skin lesions (92). Leishmania-infected LCs can migrate into the draining lymph nodes where they mature and activate Leishmania-specific T cells. Another parasite, Toxoplasma gondii, can induce the redistribution of DCs to T cell areas and activate the secretion of IL-12 by DCs but not by macrophages (26). Parasites can also subvert DC function to promote their own survival. A good example comes from the malaria parasite Plasmodium



791

falciparum, where P. falciparum–infected erythrocytes adhere to DCs and inhibit their maturation and capacity to stimulate T cells (276). DCs are implicated in the pathogenesis/response to a variety of viruses, such as cytomegalovirus (CMV), human immunodeficiency virus (HIV), measles, herpes viruses, influenza virus, and most recently respiratory syncytial virus. DCs may be affected by viruses in several ways, including the following: (a) Because of their distribution throughout the body surfaces, DCs provide a means for viruses to access other cells; (b) persistent viruses may be sequestered within the DCs and may subvert DC function and thus escape immune surveillance, for instance CMV or HIV (277–279); (c) DCs may be susceptible to cytopathic effects of viruses, as shown in measles and HIV (280–282); and (d) viral doublestranded DNA can induce DC maturation and resistance to cytopathic effects of viruses, as shown recently in influenza. The acquisition of viral Ags by DCs may happen via (a) capture of virus-infected apoptotic cells, as for influenza (88); (b) expression of receptors as for HIV, in which DCs express both CD4, the receptor for HIV, and chemokine receptors that act as coreceptors for HIV (283); and (c) internalization of nonclathrin-coated caveolae as in respiratory syncytial virus infection (284). For the last two mechanisms, it remains to be determined whether the pathways of viral entry permit the access of viral Ags to DC processing/ presentation machinery. DCs contribute to the development of both nonclonal and Ag-specific antiviral responses. Interactions of blood CD11c1 DC precursors with viruses leads to IFN-a release, initiating the cascade of antiviral response (55, 58, 66) mediated primarily by direct and indirect (IFNc) cytotoxicity of NK cells, NK T cells, and macrophages (Figure 2). Development of subsequent clonal immunity may differ depending on the virus. For instance, DCs infected with wild-type measles virus, as well as the vaccine strains, eventually undergo apoptosis and are unable to stimulate proliferation of alloreactive T cells. Although this can explain the profound immunosuppression caused by measles, it becomes unclear how immunity against measles is ever established. One possibility is that noninfected DCs may capture measles virus–induced apoptotic bodies, as occurs with influenza virus, and subsequently initiate CTL responses. Alternatively, measles virus may differentially affect the various DC subsets or maturational stages, as evidenced by the fact that measles virus–infected immature DCs induce T cell death, whereas infected mature DCs do not (280–282). Viruses can alter DC functions by interaction with inhibitory leukocyte Ig-like receptors (LIR) (285). Indeed, the UL18 glycoprotein of CMV, homologous to MHC class I, is a decoy ligand for LIR1 expressed on myelomonocytic cells including monocytes and likely DCs (285). Several functional consequences are possible, including both inhibition of differentiation and cytokine release and downregulation of antiviral response or subversion of the negative regulatory function of LIR, ultimately allowing viral replication. The DC system is also involved during HIV infection (286, 287). Thus, DCs can act as (a) transporters of the HIV, initially deposited on the mucosa, to activated T cells in secondary

792

BANCHEREAU ET AL


lymphoid organs and (b) permissive sites for virus replication. Indeed, cocultures of DCs and T cells permit HIV replication that seems to occur within syncytia that are heterokaryons of DCs and T cells. Such HIV-expressing syncytia have been found in vivo at the surfaces of mucosal lymphoid tissues like tonsils and adenoids (280). HIV-induced cell fusion of DCs and memory T cells brings together at least two transcription factors, such as NF-jB and Sp1, the coexpression of which, in heterologous syncytia, permits virus transcription and chronic replication (288). Furthermore, patients with high viral loads have decreased proportions of DC precursors in the blood, which may contribute to immunodeficiency during HIV infection (289). Much remains to be learned about the interactions of DCs and viruses and how persistent viruses like CMV or HIV subvert in vivo DC function and/or maturation. This is of particular importance because manipulation of the DC system could permit control of viral infections, for instance by increasing the activity of IFNa-producing CD11c1 DCs.

CONCLUDING REMARKS Dendritic cells induce, sustain, and regulate immune responses. Several key features of dendritic cells can be highlighted: (a) the existence of different DC subsets that share biological functions, yet display unique ones such as polarization of T cell responses towards type 1 or type 2 or regulation of B cell responses; (b) the functional specialization of DCs in relation to their differentiation/maturation stages, including (i) ability to secrete large amounts of pro-inflammatory and/or anti-viral cytokines at their precursor stage, (ii) high Ag capture capacity at their immature stage, and (iii) ability to activate and modulate T cell responses at their mature stage; (c) the plasticity of DC functions, which is determined by the microenvironment (e.g. cytokines) and may manifest as (i) the final differentiation into either DC (enhanced Ag presentation) or macrophage (enhanced Ag degradation), (ii) the induction of immunity or tolerance, and (iii) the polarization of T cell responses towards type 1 or type 2. The next few years will undoubtedly increase our understanding of the pathophysiology of DCs. We expect the genomic studies to yield molecules that will permit better definition of DC heterogeneity and explain the unique biological functions of DCs. The current clinical trials with ex vivo–generated DCs will yield precious information regarding their potential as vectors for immunotherapy. Ultimately we predict that DCs will be targeted in vivo by ‘‘intelligent missiles,’’ man-made viruses composed of a lipid envelope expressing specific ligands that can bind to either all DCs or to a specific subset. The missile may be loaded with (a) DC modulators (activators or inhibitors) to induce or suppress a given immune response or (b) Ags together with DC modulators for vaccination. This is enough to keep us busy for a while.


793

ACKNOWLEDGMENTS The authors from the Baylor Institute for Immunology Research have been supported by grants from the Baylor Health Care System Foundation, Cap CURE, Ligue Nationale contre le Cancer, axe Immunologie, and NIH CA78846A. DNAX Research Institute is supported by the Schering-Plough Corp. JD is a Directeur de Recherche au CNRS. The authors express their gratitude to the numerous collaborators who contributed to their studies on dendritic cells. Due to limited space we could only cite a fraction of the published work, which does not undermine the great value of noncited studies.



LITERATURE CITED 1. Fearon, D.T., R.M. Locksley. 1996. The instructive role of innate immunity in the acquired immune response. Science 272:50–53 2. Hoffmann, J.A., F.C. Kafatos, C.A. Janeway, R.A. Ezekowitz. 1999. Phylogenetic perspectives in innate immunity. Science 284:1313–18 3. Medzhitov, R., C.A. Janeway, Jr. 1997. Innate immunity: the virtues of a nonclonal system of recognition. Cell 91: 295–98 4. Matzinger P. 1994. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12:991–1045 5. Muzio, M., G. Natoli, S. Saccani, M. Levrero, A. Mantovani. 1998. The human toll signaling pathway: divergence of nuclear factor jB and JNK/ SAPK activation upstream of tumor necrosis factor receptor-associated factor 6 (TRAF6). J. Exp. Med. 187:2097–101 6. Brightbill, H.D., D.H. Libraty, S.R. Krutzik, R.B. Yang, J.T. Belisle, J.R. Bleharski, M. Maitland, M.V. Norgard, S.E. Plevy, S.T. Smale, P.J. Brennan, B.R. Bloom, P.J. Godowski, R.L. Modlin. 1999. Host defense mechanisms trig-

7.

8.

9.

10.

11.

12.

13.

gered by microbial lipoproteins through toll-like receptors. Science 285:732–36 Aliprantis, A.O., R.B. Yang, M.R. Mark, S. Suggett, B. Devaux, J.D. Radolf, G.R. Klimpel, P. Godowski, A. Zychlinsky. 1999. Cell activation and apoptosis by bacterial lipoproteins through toll-like receptor-2. Science 285:736–39 Banchereau, J., R.M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245–52 Bell, D., J.W. Young, J. Banchereau. 1999. Dendritic cells. Adv. Immunol. 72:255–324 Hart DN. 1997. Dendritic cells: unique leukocyte populations which control the primary immune response. Blood 90: 3245–87 Steinman RM. 1991. The dendritic cell system and its role in immunogenicity. Annu. Rev. Immunol. 9:271–96 Lambrecht, B.N., B. Salomon, D. Klatzmann, R.A. Pauwels. 1998. Dendritic cells are required for the development of chronic eosinophilic airway inflammation in response to inhaled antigen in sensitized mice. J. Immunol. 160:4090–97 Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu,

794

14.


15.

16.

17.

18.

19.

20.

21.

BANCHEREAU ET AL R.M. Steinman. 1992. Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 176:1693–702 Scheicher, C., M. Mehlig, R. Zecher, K. Reske. 1992. Dendritic cells from mouse bone marrow: in vitro differentiation using low doses of recombinant granulocyte-macrophage colony-stimulating factor. J. Immunol. Methods 154:253–64 Ardavin, C., L. Wu, C.L. Li, K. Shortman. 1993. Thymic dendritic cells and T cells develop simultaneously in the thymus from a common precursor population. Nature 362:761–63 Saunders, D., K. Lucas, J. Ismaili, L. Wu, E. Maraskovsky, A. Dunn, K. Shortman. 1996. Dendritic cell development in culture from thymic precursor cells in the absence of granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 184:2185–96 Steinman, R.M., M. Pack, K. Inaba. 1997. Dendritic cells in the T-cell areas of lymphoid organs. Immunol. Rev. 156:25–37 Vremec, D., M. Zorbas, R. Scollay, D.J. Saunders, C.F. Ardavin, L. Wu, K. Shortman. 1992. The surface phenotype of dendritic cells purified from mouse thymus and spleen: investigation of the CD8 expression by a subpopulation of dendritic cells. J. Exp. Med. 176:47–58 Wu, L., C.L. Li, K. Shortman. 1996. Thymic dendritic cell precursors: relationship to the T lymphocyte lineage and phenotype of the dendritic cell progeny. J. Exp. Med. 184:903–11 Shortman, K., D. Vremec, L.M. Corcoran, K. Georgopoulos, K. Lucas, L. Wu. 1998. The linkage between T-cell and dendritic cell development in the mouse thymus. Immunol. Rev. 165:39–46 Vremec D, K. Shortman. 1997. Dendritic cell subtypes in mouse lymphoid organs: cross-correlation of surface markers,

22.

23.

24.

25.

26.

27.

28.

changes with incubation, and differences among thymus, spleen, and lymph nodes. J. Immunol. 159:565–73 Maraskovsky E., K. Brasel, M. Teepe, E.R. Roux, S.D. Lyman, K. Shortman, H.J. McKenna. 1996. Dramatic increase in the numbers of functionally mature dendritic cells in Flt3 ligand-treated mice: multiple dendritic cell subpopulations identified. J. Exp. Med. 184:1953– 62 Pulendran B., J. Lingappa, M.K. Kennedy, J. Smith, M. Teepe, A. Rudensky, C.R. Maliszewski, E. Maraskovsky. 1997. Developmental pathways of dendritic cells in vivo: distinct function, phenotype, and localization of dendritic cell subsets in FLT3 ligand-treated mice. J. Immunol. 159:2222–31 De Smedt T., B. Pajak, E. Muraille, L. Lespagnard, E. Heinen, P. De Baetselier, J. Urbain, O. Leo, M. Moser. 1996. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184:1413–24 Leenen P.J., K. Radosevic, J.S. Voerman, B. Salomon, N. van Rooijen, D. Klatzmann, W. van Ewijk. 1998. Heterogeneity of mouse spleen dendritic cells: in vivo phagocytic activity, expression of macrophage markers, and subpopulation turnover. J. Immunol. 160:2166–73 Reis e Sousa C.R., S. Hieny, T. SchartonKersten, D. Jankovic, H. Charest, R.N. Germain, A. Sher. 1997. In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas. J. Exp. Med. 186:1819–29 Maldonado-Lopez R, T. De Smedt, P. Michel, J. Godfroid, B. Pajak, C. Heirman, K. Thielemans, O. Leo, J. Urbain, M. Moser. 1999. CD8a` and CD8a— subclasses of dendritic cells direct the development of distinct T helper cells in vivo. J. Exp. Med. 189:587–92 Ohteki, T., T. Fukao, K. Suzue, C. Maki,


29.

30.


31.

32.

33.

34.

35.

36.

M. Ito, M. Nakamura, S. Koyasu. 1999. Interleukin 12-dependent interferon c production by CD8a` lymphoid dendritic cells. J. Exp. Med. 189:1981–86 Kronin, V., K. Winkel, G. Suss, A. Kelso, W. Heath, J. Kirberg, H. von Boehmer, K. Shortman. 1996. A subclass of dendritic cells regulates the response of naive CD8 T cells by limiting their IL-2 production. J. Immunol. 157:3819–27 Suss G., K. Shortman. 1996. A subclass of dendritic cells kills CD4 T cells via Fas/Fas-ligand-induced apoptosis. J. Exp. Med. 183:1789–96 Pulendran B., J.L. Smith, G. Caspary, K. Brasel, D. Pettit, E. Maraskovsky, C.R. Maliszewski. 1999. Distinct dendritic cell subsets differentially regulate the class of immune response in vivo. Proc. Natl. Acad. Sci. USA 96:1036–41 Lyman S.D., S.E. Jacobsen. 1998. c-kit ligand and Flt3 ligand: stem/progenitor cell factors with overlapping yet distinct activities. Blood 91:1101–34 Shurin M.R., P.P. Pandharipande, T.D. Zorina, C. Haluszczak, V.M. Subbotin, O. Hunter, A. Brumfield, W.J. Storkus, E. Maraskovsky, M.T. Lotze. 1997. FLT3 ligand induces the generation of functionally active dendritic cells in mice. Cell Immunol. 179:174–84 Mohamadzadeh M., H. Jonuleit, G. Kolde, A. Pavlidou, E. Schmitt, J. Knop. 1993. Functional and morphological characterization of 4F7` spleen accessory dendritic cells. Int. Immunol. 5:615– 24 Salomon B., J.L. Cohen, C. Masurier, D. Klatzmann. 1998. Three populations of mouse lymph node dendritic cells with different origins and dynamics. J. Immunol. 160:708–17 Borkowski T.A., J.J. Letterio, A.G. Farr, M.C. Udey. 1996. A role for endogenous transforming growth factor b1 in Langerhans cell biology: the skin of transforming growth factor b1 null mice is

37.

38.

39.

40.

41.

42.

43.

795

devoid of epidermal Langerhans cells. J. Exp. Med. 184:2417–22 Burkly L., C. Hession, L. Ogata, C. Reilly, L.A. Marconi, D. Olson, R. Tizard, R. Cate, D. Lo. 1995. Expression of relB is required for the development of thymic medulla and dendritic cells. Nature 373:531–36 Wu L., A. Nichogiannopoulou, K. Shortman, K. Georgopoulos. 1997. Cellautonomous defects in dendritic cell populations of Ikaros mutant mice point to a developmental relationship with the lymphoid lineage. Immunity 7:483–92 Weih, F., D. Carrasco, S.K. Durham, D.S. Barton, C.A. Rizzo, R.P. Ryseck, S.A. Lira, R. Bravo. 1995. Multiorgan inflammation and hematopoietic abnormalities in mice with a targeted disruption of RelB, a member of the NF-c B/Rel family. Cell 80:331–40 Boehmelt, G., J. Madruga, P. Dorfler, K. Briegel, H. Schwarz, P.J. Enrietto, M. Zenke. 1995. Dendritic cell progenitor is transformed by a conditional v-Rel estrogen receptor fusion protein v-RelER. Cell 80:341–52 Grouard, G., M.C. Rissoan, L. Filgueira, I. Durand, J. Banchereau, Y.J. Liu. 1997. The enigmatic plasmacytoid T cells develop into dendritic cells with interleukin (IL)-3 and CD40-ligand. J. Exp. Med. 185:1101–11 Olweus, J., A. BitMansour, R. Warnke, P.A. Thompson, J. Carballido, L.J. Picker, F. Lund-Johansen. 1997. Dendritic cell ontogeny: a human dendritic cell lineage of myeloid origin. Proc. Natl. Acad. Sci. USA 94:12551–56 O’Doherty, U., R.M. Steinman, M. Peng, P.U. Cameron, S. Gezelter, I. Kopeloff, W.J. Swiggard, M. Pope, N. Bhardwaj. 1993. Dendritic cells freshly isolated from human blood express CD4 and mature into typical immunostimulatory dendritic cells after culture in monocyteconditioned medium. J. Exp. Med. 178:1067–76


796

BANCHEREAU ET AL

44. Strobl, H., C. Scheinecker, E. Riedl, B. Csmarits, C. Bello-Fernandez, W.F. Pickl, O. Majdic, W. Knapp. 1998. Identification of CD68`lin-peripheral blood cells with dendritic precursor characteristics. J. Immunol. 161:740–48 45. Strunk, D., C. Egger, G. Leitner, D. Hanau, G. Stingl. 1997. A skin homing molecule defines the langerhans cell progenitor in human peripheral blood. J. Exp. Med. 185:1131–36 46. Galy, A., M. Travis, D. Cen, B. Chen. 1995. Human T, B, natural killer, and dendritic cells arise from a common bone marrow progenitor cell subset. Immunity 3:459–73 47. Res, P., E. Martinez-Caceres, A. Cristina Jaleco, F. Staal, E. Noteboom, K. Weijer, H. Spits. 1996. CD34`CD38dim cells in the human thymus can differentiate into T, natural killer, and dendritic cells but are distinct from pluripotent stem cells. Blood 87:5196–206 48. Romani, N., S. Gruner, D. Brang, E. Kampgen, A. Lenz, B. Trockenbacher, G. Konwalinka, P.O. Fritsch, R.M. Steinman, G. Schuler. 1994. Proliferating dendritic cell progenitors in human blood. J. Exp. Med. 180:83–93 49. Sallusto, F., A. Lanzavecchia. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor a. J. Exp. Med. 179:1109–18 50. Zhou, L.J., T.F. Tedder. 1996. CD14` blood monocytes can differentiate into functionally mature CD83` dendritic cells. Proc. Natl. Acad. Sci. USA 93:2588–92 51. Ito, T., M. Inaba, K. Inaba, J. Toki, S. Sogo, T. Iguchi, Y. Adachi, K. Yamaguchi, R. Amakawa, J. Valladeau, S. Saeland, S. Fukuhara, S. Ikehara. 1999. A CD1a`/CD11c` subset of human blood dendritic cells is a direct precursor

52.

53.

54.

55.

56.

57.

58.

59.

of langerhans cells. J. Immunol. 163: 1409–19 Strobl, H., C. Bello-Fernandez, E. Riedl, W.F. Pickl, O. Majdic, S.D. Lyman, W. Knapp. 1997. flt3 ligand in cooperation with transforming growth factor-b1 potentiates in vitro development of Langerhans-type dendritic cells and allows single-cell dendritic cell cluster formation under serum-free conditions. Blood 90:1425–34 Caux, C., B. Vanbervliet, C. Massacrier, C. Dezutter-Dambuyant, B. de Saint-Vis, C. Jacquet, K. Yoneda, S. Imamura, D. Schmitt, J. Banchereau. 1996. CD34` hematopoietic progenitors from human cord blood differentiate along two independent dendritic cell pathways in response to GM-CSF`TNFa. J. Exp. Med. 184:695–706 Palucka, K.A., N. Taquet, F. SanchezChapuis, J.C. Gluckman. 1998. Dendritic cells as the terminal stage of monocyte differentiation. J. Immunol. 160:4587–95 Cella, M., D. Jarrossay, F. Facchetti, O. Alebardi, H. Nakajima, A. Lanzavecchia, M. Colonna. 1999. Plasmacytoid monocytes migrate to inflamed lymph nodes and produce large amounts of type I interferon. Nat. Med. 5:919–23 Palucka, K., J. Banchereau. 1999. Linking innate and adaptive immunity. Nat. Med. 5:868–70 Rissoan, M.C., V. Soumelis, N. Kadowaki, G. Grouard, F. Briere, R. de Waal Malefyt, Y.J. Liu. 1999. Reciprocal control of T helper cell and dendritic cell differentiation. Science 283:1183–86 Siegal, F.P., N. Kadowaki, M. Shodell, P.A. Fitzgerald-Bocarsly, K. Shah, S. Ho, S. Antonenko, Y.J. Liu. 1999. The nature of the principal type 1 interferon-producing cells in human blood. Science 284:1835–37 Bruno, L., P. Res, M. Dessing, M. Cella, H. Spits. 1997. Identification of a committed T cell precursor population in


60.

61.


62.

63.

64.

65.

adult human peripheral blood. J. Exp. Med. 185:875–84 Nestle, F.O., X.G. Zheng, C.B. Thompson, L.A. Turka, B.J. Nickoloff. 1993. Characterization of dermal dendritic cells obtained from normal human skin reveals phenotypic and functionally distinctive subsets. J. Immunol. 151:6535– 45. Erratum. 1994. J. Immunol. 152(1): 376 Caux, C., C. Dezutter-Dambuyant, D. Schmitt, J. Banchereau. 1992. GM-CSF and TNF-a cooperate in the generation of dendritic Langerhans cells. Nature 360:258–61 Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, I. Durand, M. Cella, A. Lanzavecchia, J. Banchereau. 1997. CD34` hematopoietic progenitors from human cord blood differentiate along two independent dendritic cell pathways in response to granulocyte-macrophage colony-stimulating factor plus tumor necrosis factor a: II. Functional analysis. Blood 90:1458–70 de Saint-Vis, B., I. Fugier-Vivier, C. Massacrier, C. Gaillard, B. Vanbervliet, S. Ait-Yahia, J. Banchereau, Y.J. Liu, S. Lebecque, C. Caux. 1998. The cytokine profile expressed by human dendritic cells is dependent on cell subtype and mode of activation. J. Immunol. 160:1666–76 Mommaas, A.M., A.A. Mulder, R. Jordens, C. Out, M.C. Tan, P. Cresswell, P.M. Kluin, F. Koning. 1999. Human epidermal Langerhans cells lack functional mannose receptors and a fully developed endosomal/lysosomal compartment for loading of HLA class II molecules. Eur. J. Immunol. 29:571–80 Mortarini, R., A. Anichini, M. Di Nicola, S. Siena, M. Bregni, F. Belli, A. Molla, A.M. Gianni, G. Parmiani. 1997. Autologous dendritic cells derived from CD34` progenitors and from monocytes are not functionally equivalent antigen-presenting cells in the induction of

66.

67.

68.

69.

70.

71.

72.

797

melan-A/Mart-1(27–35)-specific CTLs from peripheral blood lymphocytes of melanoma patients with low frequency of CTL precursors. Cancer Res. 57:5534– 41 Ferbas, J.J., J.F. Toso, A.J. Logar, J.S. Navratil, C.R. Rinaldo, Jr. 1994. CD4` blood dendritic cells are potent producers of IFN-a in response to in vitro HIV-1 infection. J. Immunol. 152:649–62. Erratum. 1994. J. Immunol. 153(2):910 Austyn, J.M., J.W. Kupiec-Weglinski, D.F. Hankins, P.J. Morris. 1988. Migration patterns of dendritic cells in the mouse: homing to T cell-dependent areas of spleen, and binding within marginal zone. J. Exp. Med. 167:646–51 McWilliam, A.S., S. Napoli, A.M. Marsh, F.L. Pemper, D.J. Nelson, C.L. Pimm, P.A. Stumbles, T.N. Wells, P.G. Holt. 1996. Dendritic cells are recruited into the airway epithelium during the inflammatory response to a broad spectrum of stimuli. J. Exp. Med. 184:2429– 32 McWilliam, A.S., D. Nelson, J.A. Thomas, P.G. Holt. 1994. Rapid dendritic cell recruitment is a hallmark of the acute inflammatory response at mucosal surfaces. J. Exp. Med. 179:1331–36 Dieu, M.C., B. Vanbervliet, A. Vicari, J.M. Bridon, E. Oldham, S. Ait-Yahia, F. Briere, A. Zlotnik, S. Lebecque, C. Caux. 1998. Selective recruitment of immature and mature dendritic cells by distinct chemokines expressed in different anatomic sites. J. Exp. Med. 188:373–86 Power, C.A., D.J. Church, A. Meyer, S. Alouani, A.E. Proudfoot, I. Clark-Lewis, S. Sozzani, A. Mantovani, T.N. Wells. 1997. Cloning and characterization of a specific receptor for the novel CC chemokine MIP-3a from lung dendritic cells. J. Exp. Med. 186:825–35 Greaves, D.R., W. Wang, D.J. Dairaghi, M.C. Dieu, B. Saint-Vis, K. FranzBacon, D. Rossi, C. Caux, T. McClanahan, S. Gordon, A. Zlotnik, T.J.

798

73.


74.

75.

76.

77.

78.

79.

80.

BANCHEREAU ET AL Schall. 1997. CCR6, a CC chemokine receptor that interacts with macrophage inflammatory protein 3a and is highly expressed in human dendritic cells. J. Exp. Med. 186:837–44 Tanaka, Y., T. Imai, M. Baba, I. Ishikawa, M. Uehira, H. Nomiyama, O. Yoshie. 1999. Selective expression of liver and activation-regulated chemokine (LARC) in intestinal epithelium in mice and humans. Eur. J. Immunol. 29:633–42 Liao, F., R.L. Rabin, C.S. Smith, G. Sharma, T.B. Nutman, J.M. Farber. 1999. CC-chemokine receptor 6 is expressed on diverse memory subsets of T cells and determines responsiveness to macrophage inflammatory protein 3a. J. Immunol. 162:186–94 Bell, D., D. Chomarat, D. Broyles, G. Netto, G.M. Harb, S. Lebecque, J. Valladeau, J. Davoust, A. Palucka, J. Banchereau. 1999. In breast carcinoma tissue, immature DC reside within the tumor while mature DC are located in peritumoral area. J. Exp. Med. 190:1417–26 Jakob, T., M.C. Udey. 1998. Regulation of E-cadherin-mediated adhesion in Langerhans cell-like dendritic cells by inflammatory mediators that mobilize Langerhans cells in vivo. J. Immunol. 160:4067–73 Tang, A., M. Amagai, L.G. Granger, J.R. Stanley, M.C. Udey. 1993. Adhesion of epidermal Langerhans cells to keratinocytes mediated by E-cadherin. Nature 361:82–85 Kobayashi, Y. 1997. Langerhans’ cells produce type IV collagenase (MMP-9) following epicutaneous stimulation with haptens. Immunology 90:496–501 Hashimoto, S., T. Suzuki, H.Y. Dong, S. Nagai, N. Yamazaki, K. Matsushima. 1999. Serial analysis of gene expression in human monocyte-derived dendritic cells. Blood 94:845–52 Engering, A.J., M. Cella, D. Fluitsma, M. Brockhaus, E.C. Hoefsmit, A. Lanzavecchia, J. Pieters. 1997. The mannose

81.

82.

83.

84.

85.

86.

87.

receptor functions as a high capacity and broad specificity antigen receptor in human dendritic cells. Eur. J. Immunol. 27:2417–25 Jiang, W., W.J. Swiggard, C. Heufler, M. Peng, A. Mirza, R.M. Steinman, M.C. Nussenzweig. 1995. The receptor DEC205 expressed by dendritic cells and thymic epithelial cells is involved in antigen processing. Nature 375:151–55 Sallusto, F., M. Cella, C. Danieli, A. Lanzavecchia. 1995. Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class II compartment: downregulation by cytokines and bacterial products. J. Exp. Med. 182:389–400 Tan, M.C., A.M. Mommaas, J.W. Drijfhout, R. Jordens, J.J. Onderwater, D. Verwoerd, A.A. Mulder, A.N. van der Heiden, D. Scheidegger, L.C. Oomen, T.H. Ottenhoff, A. Tulp, J.J. Neefjes, F. Koning. 1997. Mannose receptormediated uptake of antigens strongly enhances HLA class II-restricted antigen presentation by cultured dendritic cells. Eur. J. Immunol. 27:2426–35 Reis e Sousa, C., P.D. Stahl, J.M. Austyn. 1993. Phagocytosis of antigens by Langerhans cells in vitro. J. Exp. Med. 178:509–19 Fanger, N.A., K. Wardwell, L. Shen, T.F. Tedder, P.M. Guyre. 1996. Type I (CD64) and type II (CD32) Fc c receptormediated phagocytosis by human blood dendritic cells. J. Immunol. 157:541–48 Matsuno, K., T. Ezaki, S. Kudo, Y. Uehara. 1996. A life stage of particleladen rat dendritic cells in vivo: their terminal division, active phagocytosis, and translocation from the liver to the draining lymph. J. Exp. Med. 183:1865–78 Albert, M.L., S.F. Pearce, L.M. Francisco, B. Sauter, P. Roy, R.L. Silverstein, N. Bhardwaj. 1998. Immature dendritic cells phagocytose apoptotic cells via avb5 and CD36, and cross-present anti-


88.

89.


90.

91.

92.

93.

94.

95.

gens to cytotoxic T lymphocytes. J. Exp. Med. 188:1359–68 Albert, M.L., B. Sauter, N. Bhardwaj. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class Irestricted CTLs. Nature 392:86–89 Rubartelli, A., A. Poggi, M.R. Zocchi. 1997. The selective engulfment of apoptotic bodies by dendritic cells is mediated by the a(v)b3 integrin and requires intracellular and extracellular calcium. Eur. J. Immunol. 27:1893–900 Inaba, K., M. Inaba, M. Naito, R.M. Steinman. 1993. Dendritic cell progenitors phagocytose particulates, including bacillus Calmette-Guerin organisms, and sensitize mice to mycobacterial antigens in vivo. J. Exp. Med. 178:479–88 Rescigno, M., F. Granucci, S. Citterio, M. Foti, P. Ricciardi-Castagnoli. 1999. Coordinated events during bacteriainduced DC maturation. Immunol. Today 20:200–3 Moll, H. 1993. Epidermal Langerhans cells are critical for immunoregulation of cutaneous leishmaniasis. Immunol. Today 14:383–87 Arnold-Schild, D., D. Hanau, D. Spehner, C. Schmid, H.G. Rammensee, H. de la Salle, H. Schild. 1999. Cutting edge: receptor-mediated endocytosis of heat shock proteins by professional antigenpresenting cells. J. Immunol. 162:3757– 60 Todryk, S., A.A. Melcher, N. Hardwick, E. Linardakis, A. Bateman, M.P. Colombo, A. Stoppacciaro, R.G. Vile. 1999. Heat shock protein 70 induced during tumor cell killing induces Th1 cytokines and targets immature dendritic cell precursors to enhance antigen uptake. J. Immunol. 163:1398–408 Valladeau, J., V. Duvert-Frances, J.J. Pin, C. Massacrier, J. Vincent, K. Yoneda, C. Dezutter-Dambuyant, C. Vincent, J. Banchereau, C. Caux, J. Davoust, S. Saeland. 1999. The monoclonal antibody DCGM4 recognizes langerin, a protein

96.

97.

98.

99.

100.

101.

102.

799

specific of Langherans cells, and is rapidly internalized from the cell-surface. Eur. J. Immunol. 29:2695–2704 Valladeau, J., O. Ravel, C. DezutterDambuyant, K. Moore, M. Kleijmeer, V. Duvert-Frances, C. Vincent, D. Schmitt, J. Davoust, C. Caux, S. Lebecque, S. Saeland. 1999. Langerin, a novel transmembrane C-type lectin specific of Langherans cells, induces the formation of Birbeck granules. Immunity. In press Maurer, D., E. Fiebiger, B. Reininger, C. Ebner, P. Petzelbauer, G.P. Shi, H.A. Chapman, G. Stingl. 1998. Fc epsilon receptor I on dendritic cells delivers IgEbound multivalent antigens into a cathepsin S-dependent pathway of MHC class II presentation. J. Immunol. 161:2731–39 Akbari, O., N. Panjwani, S. Garcia, R. Tascon, D. Lowrie, B. Stockinger. 1999. DNA vaccination: transfection and activation of dendritic cells as key events for immunity. J. Exp. Med. 189:169–78 Hacker, H., H. Mischak, T. Miethke, S. Liptay, R. Schmid, T. Sparwasser, K. Heeg, G.B. Lipford, H. Wagner. 1998. CpG-DNA-specific activation of antigen-presenting cells requires stress kinase activity and is preceded by nonspecific endocytosis and endosomal maturation. EMBO J. 17:6230–40 Hartmann, G., G.J. Weiner, A.M. Krieg. 1999. CpG DNA: a potent signal for growth, activation, and maturation of human dendritic cells. Proc. Natl. Acad. Sci. USA 96:9305–10 Cella, M., M. Salio, Y. Sakakibara, H. Langen, I. Julkunen, A. Lanzavecchia. 1999. Maturation, activation, and protection of dendritic cells induced by doublestranded RNA. J. Exp. Med. 189:821–29 Winzler, C., P. Rovere, M. Rescigno, F. Granucci, G. Penna, L. Adorini, V.S. Zimmermann, J. Davoust, P. RicciardiCastagnoli. 1997. Maturation stages of mouse dendritic cells in growth factordependent long-term cultures. J. Exp. Med. 185:317–28


800

BANCHEREAU ET AL

103. Mosialos, G., M. Birkenbach, S. Ayehunie, F. Matsumura, G.S. Pinkus, E. Kieff, E. Langhoff. 1996. Circulating human dendritic cells differentially express high levels of a 55-kd actin-bundling protein. Am. J. Pathol. 148:593–600 104. Ross, R., X.L. Ross, J. Schwing, T. Langin, A.B. Reske-Kunz. 1998. The actinbundling protein fascin is involved in the formation of dendritic processes in maturing epidermal Langerhans cells. J. Immunol. 160:3776–82 105. Binks, M., G.E. Jones, P.M. Brickell, C. Kinnon, D.R. Katz, A.J. Thrasher. 1998. Intrinsic dendritic cell abnormalities in Wiskott-Aldrich syndrome. Eur. J. Immunol. 28:3259–67 106. Parolini, O., S. Berardelli, E. Riedl, C. Bello-Fernandez, H. Strobl, O. Majdic, W. Knapp. 1997. Expression of WiskottAldrich syndrome protein (WASP) gene during hematopoietic differentiation. Blood 90:70–75 107. Cella, M., C. Dohring, J. Samaridis, M. Dessing, M. Brockhaus, A. Lanzavecchia, M. Colonna. 1997. A novel inhibitory receptor (ILT3) expressed on monocytes, macrophages, and dendritic cells involved in antigen processing. J. Exp. Med. 185:1743–51 108. Long, E.O. 1999. Regulation of immune responses through inhibitory receptors. Annu. Rev. Immunol. 17:875–904 109. Kripke, M.L., C.G. Munn, A. Jeevan, J.M. Tang, C. Bucana. 1990. Evidence that cutaneous antigen-presenting cells migrate to regional lymph nodes during contact sensitization. J. Immunol. 145:2833–38 110. Larsen, C.P., R.M. Steinman, M. WitmerPack, D.F. Hankins, P.J. Morris, J.M. Austyn. 1990. Migration and maturation of Langerhans cells in skin transplants and explants. J. Exp. Med. 172:1483–93 111. Macatonia, S.E., S.C. Knight, A.J. Edwards, S. Griffiths, P. Fryer. 1987. Localization of antigen on lymph node dendritic cells after exposure to the con-

112.

113.

114.

115.

116.

117.

118.

119.

tact sensitizer fluorescein isothiocyanate: functional and morphological studies. J. Exp. Med. 166:1654–67 Fossum, S. 1988. Lymph-borne dendritic leucocytes do not recirculate, but enter the lymph node paracortex to become interdigitating cells. Scand. J. Immunol. 27:97–105 Cumberbatch, M., I. Kimber. 1995. Tumour necrosis factor-a is required for accumulation of dendritic cells in draining lymph nodes and for optimal contact sensitization. Immunology 84:31–35 Sallusto, F., P. Schaerli, P. Loetscher, C. Schaniel, D. Lenig, C.R. Mackay, S. Qin, A. Lanzavecchia. 1998. Rapid and coordinated switch in chemokine receptor expression during dendritic cell maturation. Eur. J. Immunol. 28:2760–69 Sozzani, S., P. Allavena, A. Vecchi, A. Mantovani. 1999. The role of chemokines in the regulation of dendritic cell trafficking. J. Leukocyte Biol. 66:1–9 Yoshida, R., T. Imai, K. Hieshima, J. Kusuda, M. Baba, M. Kitaura, M. Nishimura, M. Kakizaki, H. Nomiyama, O. Yoshie. 1997. Molecular cloning of a novel human CC chemokine EBI1ligand chemokine that is a specific functional ligand for EBI1, CCR7. J. Biol. Chem. 272:13803–9 Chan, V.W., S. Kothakota, M.C. Rohan, L. Panganiban-Lustan, J.P. Gardner, M.S. Wachowicz, J.A. Winter, L.T. Williams. 1999. Secondary lymphoid-tissue chemokine (SLC) is chemotactic for mature dendritic cells. Blood 93:3610–16 Gunn, M.D., K. Tangemann, C. Tam, J.G. Cyster, S.D. Rosen, L.T. Williams. 1998. A chemokine expressed in lymphoid high endothelial venules promotes the adhesion and chemotaxis of naive T lymphocytes. Proc. Natl. Acad. Sci. USA 95:258–63 Saeki, H., A.M. Moore, M.J. Brown, S.T. Hwang. 1999. Cutting edge: secondary lymphoid-tissue chemokine (SLC) and CC chemokine receptor 7 (CCR7) par-


120.

121.


122.

123.

124.

125.

126.

127.

ticipate in the emigration pathway of mature dendritic cells from the skin to regional lymph nodes. J. Immunol. 162:2472–75 Ngo, V.N., H.L. Tang, J.G. Cyster. 1998. Epstein-Barr virus-induced molecule 1 ligand chemokine is expressed by dendritic cells in lymphoid tissues and strongly attracts naive T cells and activated B cells. J. Exp. Med. 188:181–91 Campbell, J.J., J. Hedrick, A. Zlotnik, M.A. Siani, D.A. Thompson, E.C. Butcher. 1998. Chemokines and the arrest of lymphocytes rolling under flow conditions. Science 279:381–84 Gunn, M.D., S. Kyuwa, C. Tam, T. Kakiuchi, A. Matsuzawa, L.T. Williams, H. Nakano. 1999. Mice lacking expression of secondary lymphoid organ chemokine have defects in lymphocyte homing and dendritic cell localization. J. Exp. Med. 189:451–60 Nakano, H., T. Tamura, T. Yoshimoto, H. Yagita, M. Miyasaka, E.C. Butcher, H. Nariuchi, T. Kakiuchi, A. Matsuzawa. 1997. Genetic defect in T lymphocytespecific homing into peripheral lymph nodes. Eur. J. Immunol. 27:215–21 Nakano, H., S. Mori, H. Yonekawa, H. Nariuchi, A. Matsuzawa, T. Kakiuchi. 1998. A novel mutant gene involved in T-lymphocyte-specific homing into peripheral lymphoid organs on mouse chromosome 4. Blood 91:2886–95 Forster, R., A.E. Mattis, E. Kremmer, E. Wolf, G. Brem, M. Lipp. 1996. A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell 87:1037–47 Kanazawa, N., T. Nakamura, K. Tashiro, M. Muramatsu, K. Morita, K. Yoneda, K. Inaba, S. Imamura, T. Honjo. 1999. Fractalkine and macrophage-derived chemokine: T cell-attracting chemokines expressed in T cell area dendritic cells. Eur. J. Immunol. 29:1925–32 Adema, G. J, F. Hartgers, R. Verstraten,

128.

129.

130.

131.

132.

133.

134.

801

E. de Vries, G. Marland, S. Menon, J. Foster, Y. Xu, P. Nooyen, T. McClanahan, K.B. Bacon, C.G. Figdor. 1997. A dendritic-cell-derived C-C chemokine that preferentially attracts naive T cells. Nature 387:713–17 Tang, H.L., J.G. Cyster. 1999. Chemokine up-regulation and activated T cell attraction by maturing dendritic cells. Science 284:819–22 Inaba, K., S. Turley, F. Yamaide, T. Iyoda, K. Mahnke, M. Inaba, M. Pack, M. Subklewe, B. Sauter, D. Sheff, M. Albert, N. Bhardwaj, I. Mellman, R.M. Steinman. 1998. Efficient presentation of phagocytosed cellular fragments on the major histocompatibility complex class II products of dendritic cells. J. Exp. Med. 188:2163–73 Inaba, K., M. Pack, M. Inaba, H. Sakuta, F. Isdell, R.M. Steinman. 1997. High levels of a major histocompatibility complex II-self peptide complex on dendritic cells from the T cell areas of lymph nodes. J. Exp. Med. 186:665–72 Kleijmeer, M.J., M.A. Ossevoort, C.J. van Veen, J.J. van Hellemond, J.J. Neefjes, W.M. Kast, C.J. Melief, H.J. Geuze. 1995. MHC class II compartments and the kinetics of antigen presentation in activated mouse spleen dendritic cells. J. Immunol. 154:5715–24 Nijman, H.W., M.J. Kleijmeer, M.A. Ossevoort, V.M. Oorschot, M.P. Vierboom, M. van de Keur, P. Kenemans, W.M. Kast, H.J. Geuze, C.J. Melief. 1995. Antigen capture and major histocompatibility class II compartments of freshly isolated and cultured human blood dendritic cells. J. Exp. Med. 182:163–74 Cresswell, P. 1996. Invariant chain structure and MHC class II function. Cell 84:505–7 Castellino, F., G. Zhong, R.N. Germain. 1997. Antigen presentation by MHC class II molecules: invariant chain function, protein trafficking, and the molec-

802

135.


136.

137.

138.

139.

140.

141.

BANCHEREAU ET AL ular basis of diverse determinant capture. Hum. Immunol. 54:159–69 Rovere, P., V.S. Zimmermann, F. Forquet, D. Demandolx, J. Trucy, P. Ricciardi-Castagnoli, J. Davoust. 1998. Dendritic cell maturation and antigen presentation in the absence of invariant chain. Proc. Natl. Acad. Sci. USA 95:1067–72 Lutz, M.B., P. Rovere, M.J. Kleijmeer, M. Rescigno, C.U. Assmann, V.M. Oorschot, H.J. Geuze, J. Trucy, D. Demandolx, J. Davoust, P. Ricciardi-Castagnoli. 1997. Intracellular routes and selective retention of antigens in mildly acidic cathepsin D/lysosome-associated membrane protein-1/MHC class II-positive vesicles in immature dendritic cells. J. Immunol. 159:3707–16 Pierre, P., I. Mellman. 1998. Developmental regulation of invariant chain proteolysis controls MHC class II trafficking in mouse dendritic cells. Cell 93:1135– 45 Cella, M., A. Engering, V. Pinet, J. Pieters, A. Lanzavecchia. 1997. Inflammatory stimuli induce accumulation of MHC class II complexes on dendritic cells. Nature 388:782–87 Pierre, P., S.J. Turley, E. Gatti, M. Hull, J. Meltzer, A. Mirza, K. Inaba, R.M. Steinman, I. Mellman. 1997. Developmental regulation of MHC class II transport in mouse dendritic cells. Nature 388:787–92 de Saint-Vis, B., J. Vincent, S. Vandenabeele, B. Vanbervliet, J.J. Pin, S. AitYahia, S. Patel, M.G. Mattei, J. Banchereau, S. Zurawski, J. Davoust, C. Caux, S. Lebecque. 1998. A novel lysosome-associated membrane glycoprotein, DC-LAMP, induced upon DC maturation, is transiently expressed in MHC class II compartment. Immunity 9:325–36 Koppelman, B., J.J. Neefjes, J.E. de Vries, R. de Waal Malefyt. 1997. Interleukin-10 down-regulates MHC class II

142.

143.

144.

145.

146.

147.

148.

149.

ab peptide complexes at the plasma membrane of monocytes by affecting arrival and recycling. Immunity 7:861–71 Butz, E. A, M.J. Bevan. 1998. Differential presentation of the same MHC class I epitopes by fibroblasts and dendritic cells. J. Immunol. 160:2139–44 Pamer, E., P. Cresswell. 1998. Mechanisms of MHC class I–restricted antigen processing. Annu. Re.v Immunol. 16: 323–58 Rock, K.L., A.L. Goldberg. 1999. Degradation of cell proteins and the generation of MHC class I-presented peptides. Annu. Rev. Immunol. 17:739–79 Bates, E.E., O. Ravel, M.C. Dieu, S. Ho, C. Guret, J.M. Bridon, S. Ait-Yahia, F. Briere, C. Caux, J. Banchereau, S. Lebecque. 1997. Identification and analysis of a novel member of the ubiquitin family expressed in dendritic cells and mature B cells. Eur. J. Immunol. 27:2471–77 Liu, Y.C., J. Pan, C. Zhang, W. Fan, M. Collinge, J.R. Bender, S.M. Weissman. 1999. A MHC-encoded ubiquitin-like protein (FAT10) binds noncovalently to the spindle assembly checkpoint protein MAD2. Proc. Natl. Acad. Sci. USA 96:4313–18 Bevan, M.J. 1976. Cross-priming for a secondary cytotoxic response to minor H antigens with H-2 congenic cells which do not cross-react in the cytotoxic assay. J. Exp. Med. 143:1283–88 Bevan, M.J. 1977. Priming for a cytotoxic response to minor histocompatibility antigens: antigen specificity and failure to demonstrate a carrier effect. J. Immunol. 118:1370–74 Norbury, C.C., B.J. Chambers, A.R. Prescott, H.G. Ljunggren, C. Watts. 1997. Constitutive macropinocytosis allows TAP-dependent major histocompatibility complex class I presentation of exogenous soluble antigen by bone marrow-derived dendritic cells. Eur. J. Immunol. 27:280–88


IMMUNOBIOLOGY OF DENDRITIC CELLS 150. Pfeifer, J.D., M.J. Wick, R.L. Roberts, K. Findlay, S.J. Normark, C.V. Harding. 1993. Phagocytic processing of bacterial antigens for class I MHC presentation to T cells. Nature 361:359–62 151. Kovacsovics-Bankowski, M., K. Clark, B. Benacerraf, K.L. Rock. 1993. Efficient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl. Acad. Sci. USA 90:4942–46 152. Kovacsovics-Bankowski, M., K.L. Rock. 1995. A phagosome-to-cytosol pathway for exogenous antigens presented on MHC class I molecules. Science 267: 243–46 153. Huang, A.Y., P. Golumbek, M. Ahmadzadeh, E. Jaffee, D. Pardoll, H. Levitsky. 1994. Role of bone marrow-derived cells in presenting MHC class I-restricted tumor antigens. Science 264:961–65 154. Sigal, L.J., S. Crotty, R. Andino, K.L. Rock. 1999. Cytotoxic T-cell immunity to virus-infected non-haematopoietic cells requires presentation of exogenous antigen. Nature 398:77–80 155. Kurts, C., F.R. Carbone, M. Barnden, E. Blanas, J. Allison, W.R. Heath, J.F. Miller. 1997. CD4` T cell help impairs CD8` T cell deletion induced by crosspresentation of self-antigens and favors autoimmunity. J. Exp. Med. 186:2057– 62 156. Nouri-Shirazi, M., D. Bell, S. Burkeholder, E. Kraus, J. Davoust, J. Banchereau, A. Palucka. 1999. Dendritic cells capture tumor cell bodies and process/ present their antigens to elicit primary immune responses. Submitted for publication 157. Regnault, A., D. Lankar, V. Lacabanne, A. Rodriguez, C. Thery, M. Rescigno, T. Saito, S. Verbeek, C. Bonnerot, P. Ricciardi-Castagnoli, S. Amigorena. 1999. Fcc receptor-mediated induction of dendritic cell maturation and major histocompatibility complex class I-restricted antigen presentation after immune com-

158.

159.

160.

161.

162.

163.

164.

165.

803

plex internalization. J. Exp. Med. 189:371–80 Zitvogel, L., A. Regnault, A. Lozier, J. Wolfers, C. Flament, D. Tenza, P. Ricciardi-Castagnoli, G. Raposo, S. Amigorena. 1998. Eradication of established murine tumors using a novel cell-free vaccine: dendritic cell-derived exosomes. Nat. Med. 4:594–600 Chandawarkar, R.Y., M.S. Wagh, P.K. Srivastava. 1999. The dual nature of specific immunological activity of tumorderived gp96 preparations. J. Exp. Med. 189:1437–42 Burdin, N., M. Kronenberg. 1999. CD1mediated immune responses to glycolipids. Curr. Opin. Immunol. 11:326–31 Porcelli, S.A., R.L. Modlin. 1999. The CD1 system: antigen-presenting molecules for T cell recognition of lipids and glycolipids. Annu. Rev. Immunol. 17:297–329 Kitamura, H., K. Iwakabe, T. Yahata, S. Nishimura, A. Ohta, Y. Ohmi, M. Sato, K. Takeda, K. Okumura, L. Van Kaer, T. Kawano, M. Taniguchi, T. Nishimura. 1999. The natural killer T (NKT) cell ligand a-galactosylceramide demonstrates its immunopotentiating effect by inducing interleukin (IL)-12 production by dendritic cells and IL-12 receptor expression on NKT cells. J. Exp. Med. 189:1121–28 Inaba, K., J.P. Metlay, M.T. Crowley, R.M. Steinman. 1990. Dendritic cells pulsed with protein antigens in vitro can prime antigen-specific, MHC-restricted T cells in situ. J. Exp. Med. 172:631–40. Erratum. 1990. J. Exp. Med. 172(4):1275 Ingulli, E., A. Mondino, A. Khoruts, M.K. Jenkins. 1997. In vivo detection of dendritic cell antigen presentation to CD4(`) T cells. J. Exp. Med. 185:2133– 41 Sornasse, T., V. Flamand, G. De Becker, H. Bazin, F. Tielemans, K. Thielemans, J. Urbain, O. Leo, M. Moser. 1992. Antigen-pulsed dendritic cells can efficiently

804

166.

167.


168.

169.

170.

171.

172.

173.

BANCHEREAU ET AL induce an antibody response in vivo. J. Exp. Med. 175:15–21 Pulendran, B., J.L. Smith, M. Jenkins, M. Schoenborn, E. Maraskovsky, C.R. Maliszewski. 1998. Prevention of peripheral tolerance by a dendritic cell growth factor: flt3 ligand as an adjuvant. J. Exp. Med. 188:2075–82 Shimizu, Y., L.G. Guidotti, P. Fowler, F.V. Chisari. 1998. Dendritic cell immunization breaks cytotoxic T lymphocyte tolerance in hepatitis B virus transgenic mice. J. Immunol. 161:4520–29 Gong, J., D. Chen, M. Kashiwaba, Y. Li, L. Chen, H. Takeuchi, H. Qu, G.J. Rowse, S.J. Gendler, D. Kufe. 1998. Reversal of tolerance to human MUC1 antigen in MUC1 transgenic mice immunized with fusions of dendritic and carcinoma cells. Proc. Natl. Acad. Sci. USA 95:6279–83 Steptoe, R.J., F. Fu, W. Li, M.L. Drakes, L. Lu, A.J. Demetris, S. Qian, H.J. McKenna, A.W. Thomson. 1997. Augmentation of dendritic cells in murine organ donors by Flt3 ligand alters the balance between transplant tolerance and immunity. J. Immunol. 159:5483–91 Ridge, J.P., E.J. Fuchs, P. Matzinger. 1996. Neonatal tolerance revisited: turning on newborn T cells with dendritic cells. Science 271:1723–26 Inaba, K., J.W. Young, R.M. Steinman. 1987. Direct activation of CD8` cytotoxic T lymphocytes by dendritic cells. J. Exp. Med. 166:182–94 Young, J.W., R.M. Steinman. 1990. Dendritic cells stimulate primary human cytolytic lymphocyte responses in the absence of CD4` helper T cells. J. Exp. Med. 171:1315–32 McCoy, K.D., I.F. Hermans, J.H. Fraser, G. Le Gros, F. Ronchese. 1999. Cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) can regulate dendritic cellinduced activation and cytotoxicity of CD8(`) T cells independently of

174.

175.

176.

177.

178.

179.

180.

181.

182.

CD4(`) T cell help. J. Exp. Med. 189:1157–62 Macatonia, S.E., P.M. Taylor, S.C. Knight, B.A. Askonas. 1989. Primary stimulation by dendritic cells induces antiviral proliferative and cytotoxic T cell responses in vitro. J. Exp. Med. 169:1255–64 Nair, S., F. Zhou, R. Reddy, L. Huang, B.T. Rouse. 1992. Soluble proteins delivered to dendritic cells via pH-sensitive liposomes induce primary cytotoxic T lymphocyte responses in vitro. J. Exp. Med. 175:609–12 Mehta-Damani, A., S. Markowicz, E.G. Engleman. 1994. Generation of antigenspecific CD8` CTLs from naive precursors. J. Immunol. 153:996–1003 McKinney, E.C., J.W. Streilein. 1989. On the extraordinary capacity of allogeneic epidermal Langerhans cells to prime cytotoxic T cells in vivo. J. Immunol. 143:1560–64 Takahashi, H., Y. Nakagawa, K. Yokomuro, J.A. Berzofsky. 1993. Induction of CD8` cytotoxic T lymphocytes by immunization with syngeneic irradiated HIV-1 envelope derived peptide-pulsed dendritic cells. Int. Immunol. 5:849–57 Nair, S., J.S. Babu, R.G. Dunham, P. Kanda, R.L. Burke, B.T. Rouse. 1993. Induction of primary, antiviral cytotoxic, and proliferative responses with antigens administered via dendritic cells. J. Virol. 67:4062–69 Brossart, P., M.J. Bevan. 1997. Presentation of exogenous protein antigens on major histocompatibility complex class I molecules by dendritic cells: pathway of presentation and regulation by cytokines. Blood 90:1594–99 Rouse, R.J., S.K. Nair, S.L. Lydy, J.C. Bowen, B.T. Rouse. 1994. Induction in vitro of primary cytotoxic T-lymphocyte responses with DNA encoding herpes simplex virus proteins. J. Virol. 68:5685– 89 Specht, J.M., G. Wang, M.T. Do, J.S.


183.


184.

185.

186.

187.

188.

189.

Lam, R.E. Royal, M.E. Reeves, S.A. Rosenberg, P. Hwu. 1997. Dendritic cells retrovirally transduced with a model antigen gene are therapeutically effective against established pulmonary metastases. J. Exp. Med. 186:1213–21 Kaplan, J.M., Q. Yu, S.T. Piraino, S.E. Pennington, S. Shankara, L.A. Woodworth, B.L. Roberts. 1999. Induction of antitumor immunity with dendritic cells transduced with adenovirus vectorencoding endogenous tumor-associated antigens. J. Immunol. 163:699–707 Ashley, D.M., B. Faiola, S. Nair, L.P. Hale, D.D. Bigner, E. Gilboa. 1997. Bone marrow-generated dendritic cells pulsed with tumor extracts or tumor RNA induce antitumor immunity against central nervous system tumors. J. Exp. Med. 186:1177–82 Bhardwaj, N., A. Bender, N. Gonzalez, L.K. Bui, M.C. Garrett, R.M. Steinman. 1994. Influenza virus-infected dendritic cells stimulate strong proliferative and cytolytic responses from human CD8` T cells. J. Clin. Invest. 94:797–807 Bohm, W., R. Schirmbeck, A. Elbe, K. Melber, D. Diminky, G. Kraal, N. van Rooijen, Y. Barenholz, J. Reimann. 1995. Exogenous hepatitis B surface antigen particles processed by dendritic cells or macrophages prime murine MHC class Irestricted cytotoxic T lymphocytes in vivo. J. Immunol. 155:3313–21 Ridge, J.P., F. Di Rosa, P. Matzinger. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4` Thelper and a T-killer cell. Nature 393:474–78 Bennett, S.R., F.R. Carbone, F. Karamalis, R.A. Flavell, J.F. Miller, W.R. Heath. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393:478–80 Schoenberger, S.P., R.E. Toes, E.I. van der Voort, R. Offringa, C.J. Melief. 1998. T-cell help for cytotoxic T lymphocytes

190.

191.

192.

193.

194.

195.

196.

805

is mediated by CD40-CD40L interactions. Nature 393:480–83 Brocker, T. 1997. Survival of mature CD4 T lymphocytes is dependent on major histocompatibility complex class II-expressing dendritic cells. J. Exp. Med. 186:1223–32 Ludewig, B., S. Oehen, F. Barchiesi, R.A. Schwendener, H. Hengartner, R.M. Zinkernagel. 1999. Protective antiviral cytotoxic T cell memory is most efficiently maintained by restimulation via dendritic cells. J. Immunol. 163:1839–44 Caux, C., B. Vanbervliet, C. Massacrier, M. Azuma, K. Okumura, L.L. Lanier, J. Banchereau. 1994. B70/B7–2 is identical to CD86 and is the major functional ligand for CD28 expressed on human dendritic cells. J. Exp. Med. 180:1841– 47 Inaba, K., M. Witmer-Pack, M. Inaba, K.S. Hathcock, H. Sakuta, M. Azuma, H. Yagita, K. Okumura, P.S. Linsley, S. Ikehara, S. Murramatsu, R.J. Hodes, R.M. Steinman. 1994. The tissue distribution of the B7–2 costimulator in mice: abundant expression on dendritic cells in situ and during maturation in vitro. J. Exp. Med. 180:1849–60 Caux, C., C. Massacrier, B. Vanbervliet, B. Dubois, C. Van Kooten, I. Durand, J. Banchereau. 1994. Activation of human dendritic cells through CD40 cross-linking. J. Exp. Med. 180:1263–72 Stuber, E., M. Neurath, D. Calderhead, H.P. Fell, W. Strober. 1995. Cross-linking of OX40 ligand, a member of the TNF/NGF cytokine family, induces proliferation and differentiation in murine splenic B cells. Immunity 2:507–21 Flynn, S., K.M. Toellner, C. Raykundalia, M. Goodall, P. Lane. 1998. CD4 T cell cytokine differentiation: the B cell activation molecule, OX40 ligand, instructs CD4 T cells to express interleukin 4 and upregulates expression of the chemokine receptor, Blr-1. J. Exp. Med. 188:297–304


806

BANCHEREAU ET AL

197. Brocker, T., A. Gulbranson-Judge, S. Flynn, M. Riedinger, C. Raykundalia, P. Lane. 1999. CD4 T cell traffic control: in vivo evidence that ligation of OX40 on CD4 T cells by OX40-ligand expressed on dendritic cells leads to the accumulation of CD4 T cells in B follicles. Eur. J. Immunol. 29:1610–16 198. DeBenedette, M.A., A. Shahinian, T.W. Mak, T.H. Watts. 1997. Costimulation of CD28-T lymphocytes by 4–1BB ligand. J. Immunol. 158:551–59 199. Saoulli, K., S.Y. Lee, J.L. Cannons, W.C. Yeh, A. Santana, M.D. Goldstein, N. Bangia, M.A. DeBenedette, T.W. Mak, Y. Choi, T.H. Watts. 1998. CD28-independent, TRAF2-dependent costimulation of resting T cells by 4–1BB ligand. J. Exp. Med. 187:1849–62 200. Kim, Y.J., S.H. Kim, P. Mantel, B.S. Kwon. 1998. Human 4–1BB regulates CD28 co-stimulation to promote Th1 cell responses. Eur. J. Immunol. 28:881–90 201. Shuford, W.W., K. Klussman, D.D. Tritchler, D.T. Loo, J. Chalupny, A.W. Siadak, T.J. Brown, J. Emswiler, H. Raecho, C.P. Larsen, T.C. Pearson, J.A. Ledbetter, A. Aruffo, R.S. Mittler. 1997. 4– 1BB costimulatory signals preferentially induce CD8` T cell proliferation and lead to the amplification in vivo of cytotoxic T cell responses. J. Exp. Med. 186:47–55 202. Anderson, D.M., E. Maraskovsky, W.L. Billingsley, W.C. Dougall, M.E. Tometsko, E.R. Roux, M.C. Teepe, R.F. DuBose, D. Cosman, L. Galibert. 1997. A homologue of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell function. Nature 390:175–79 203. Wong, B.R., R. Josien, S.Y. Lee, B. Sauter, H.L. Li, R.M. Steinman, Y. Choi. 1997. TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantly expressed in T cells, is a dendritic cell-specific survival factor. J. Exp. Med. 186:2075–80

204. Josien, R., B.R. Wong, H.L. Li, R.M. Steinman, Y. Choi. 1999. TRANCE, a TNF family member, is differentially expressed on T cell subsets and induces cytokine production in dendritic cells. J. Immunol. 162:2562–68 205. Macatonia, S.E., N.A. Hosken, M. Litton, P. Vieira, C.S. Hsieh, J.A. Culpepper, M. Wysocka, G. Trinchieri, K.M. Murphy, A. O’Garra. 1995. Dendritic cells produce IL-12 and direct the development of Th1 cells from naive CD4` T cells. J. Immunol. 154:5071–79 206. Kadowaki, N., S. Antonenko, S. Ho, M.C. Rissoan, V. Soumelis, S.A. Porcelli, L.L. Lanier, Y.J. Liu. 1999. Human dendritic cell subsets DC1 and DC2 induce NKT cell differentiation into IFN-c-producing NKT1 versus IL-4-producing NKT2 cells. Submitted for publication 207. McKenzie, G.J., C.L. Emson, S.E. Bell, S. Anderson, P. Fallon, G. Zurawski, R. Murray, R. Grencis, A.N. McKenzie. 1998. Impaired development of Th2 cells in IL-13-deficient mice. Immunity 9:423– 32 208. Rincon, M., J. Anguita, T. Nakamura, E. Fikrig, R.A. Flavell. 1997. Interleukin (IL)-6 directs the differentiation of IL-4producing CD4` T cells. J. Exp. Med. 185:461–69 209. Freeman, G.J., V.A. Boussiotis, A. Anumanthan, G.M. Bernstein, X.Y. Ke, P.D. Rennert, G.S. Gray, J.G. Gribben, L.M. Nadler. 1995. B7–1 and B7–2 do not deliver identical costimulatory signals, since B7–2 but not B7–1 preferentially costimulates the initial production of IL4. Immunity 2:523–32 210. Kuchroo, V.K., M.P. Das, J.A. Brown, A.M. Ranger, S.S. Zamvil, R.A. Sobel, H.L. Weiner, N. Nabavi, L.H. Glimcher. 1995. B7–1 and B7–2 costimulatory molecules activate differentially the Th1/ Th2 developmental pathways: application to autoimmune disease therapy. Cell 80:707–18


IMMUNOBIOLOGY OF DENDRITIC CELLS 211. Piemonti, L., P. Monti, P. Allavena, M. Sironi, L. Soldini, B.E. Leone, C. Socci, V. Di Carlo. 1999. Glucocorticoids affect human dendritic cell differentiation and maturation. J. Immunol. 162:6473–81 212. Buelens, C., F. Willems, A. Delvaux, G. Pierard, J.P. Delville, T. Velu, M. Goldman. 1995. Interleukin-10 differentially regulates B7–1 (CD80) and B7–2 (CD86) expression on human peripheral blood dendritic cells. Eur. J. Immunol. 25:2668–72 213. Caux, C., C. Massacrier, B. Vanbervliet, C. Barthelemy, Y.J. Liu, J. Banchereau. 1994. Interleukin 10 inhibits T cell alloreaction induced by human dendritic cells. Int. Immunol. 6:1177–85 214. De Smedt, T., M. Van Mechelen, G. De Becker, J. Urbain, O. Leo, M. Moser. 1997. Effect of interleukin-10 on dendritic cell maturation and function. Eur. J. Immunol. 27:1229–35 215. Kalinski, P., J.H. Schuitemaker, C.M. Hilkens, M.L. Kapsenberg. 1998. Prostaglandin E2 induces the final maturation of IL-12-deficient CD1a`CD83` dendritic cells: the levels of IL-12 are determined during the final dendritic cell maturation and are resistant to further modulation. J. Immunol. 161:2804–9 216. Liu, L., B.E. Rich, J. Inobe, W. Chen, H.L. Weiner. 1998. Induction of Th2 cell differentiation in the primary immune response: dendritic cells isolated from adherent cell culture treated with IL-10 prime naive CD4` T cells to secrete IL4. Int. Immunol. 10:1017–26 217. Takeuchi, M., M.M. Kosiewicz, P. Alard, J.W. Streilein. 1997. On the mechanisms by which transforming growth factor-b 2 alters antigen-presenting abilities of macrophages on T cell activation. Eur. J. Immunol. 27:1648–56 218. Iwasaki, A., B.L. Kelsall. 1999. Freshly isolated Peyer’s patch, but not spleen, dendritic cells produce interleukin 10 and induce the differentiation of T helper type 2 cells. J. Exp. Med. 190:229–40

807

219. Viney, J.L., A.M. Mowat, J.M. O’Malley, E. Williamson, N.A. Fanger. 1998. Expanding dendritic cells in vivo enhances the induction of oral tolerance. J. Immunol. 160:5815–25 220. Pape, K.A., A. Khoruts, A. Mondino, M.K. Jenkins. 1997. Inflammatory cytokines enhance the in vivo clonal expansion and differentiation of antigenactivated CD4` T cells. J. Immunol. 159:591–98 221. D’Amico, G., G. Bianchi, S. Bernasconi, L. Bersani, L. Piemonti, S. Sozzani, A. Mantovani, P. Allavena. 1998. Adhesion, transendothelial migration, and reverse transmigration of in vitro cultured dendritic cells. Blood 92:207–14 222. Randolph, G.J., S. Beaulieu, S. Lebecque, R.M. Steinman, W.A. Muller. 1998. Differentiation of monocytes into dendritic cells in a model of transendothelial trafficking. Science 282:480–83 223. Matzinger, P., S. Guerder. 1989. Does Tcell tolerance require a dedicated antigen-presenting cell? Nature 338:74–76 224. Inaba, M., K. Inaba, M. Hosono, T. Kumamoto, T. Ishida, S. Muramatsu, T. Masuda, S. Ikehara. 1991. Distinct mechanisms of neonatal tolerance induced by dendritic cells and thymic B cells. J. Exp. Med. 173:549–59 225. Mazda, O., Y. Watanabe, J. Gyotoku, Y. Katsura. 1991. Requirement of dendritic cells and B cells in the clonal deletion of Mls-reactive T cells in the thymus. J. Exp. Med. 173:539–47 226. Khoury, S.J., L. Gallon, W. Chen, K. Betres, M.E. Russell, W.W. Hancock, C.B. Carpenter, M.H. Sayegh, H.L. Weiner. 1995. Mechanisms of acquired thymic tolerance in experimental autoimmune encephalomyelitis: thymic dendritic-enriched cells induce specific peripheral T cell unresponsiveness in vivo. J. Exp. Med. 182:357–66 227. Brocker, T., M. Riedinger, K. Karjalainen. 1997. Targeted expression of major histocompatibility complex (MHC) class

808

228.

229.


230.

231.

232.

233.

234.

235.

236.

BANCHEREAU ET AL II molecules demonstrates that dendritic cells can induce negative but not positive selection of thymocytes in vivo. J. Exp. Med. 185:541–50 Shlomchik, W.D., M.S. Couzens, C.B. Tang, J. McNiff, M.E. Robert, J. Liu, M.J. Shlomchik, S.G. Emerson. 1999. Prevention of graft versus host disease by inactivation of host antigen-presenting cells. Science 285:412–15 Starzl, T.E., R.M. Zinkernagel. 1998. Antigen localization and migration in immunity and tolerance. N. Engl. J. Med. 339:1905–13 Thomson, A.W., L. Lu, N. Murase, A.J. Demetris, A.S. Rao, T.E. Starzl. 1995. Microchimerism, dendritic cell progenitors and transplantation tolerance. Stem Cells 13:622–39 Dubois, B., C. Massacrier, B. Vanbervliet, J. Fayette, F. Briere, J. Banchereau, C. Caux. 1998. Critical role of IL-12 in dendritic cell-induced differentiation of naive B lymphocytes. J. Immunol. 161:2223–31 Dubois, B., B. Vanbervliet, J. Fayette, C. Massacrier, C. Van Kooten, F. Briere, J. Banchereau, C. Caux. 1997. Dendritic cells enhance growth and differentiation of CD40-activated B lymphocytes. J. Exp. Med. 185:941–51 Fayette, J., B. Dubois, S. Vandenabeele, J.M. Bridon, B. Vanbervliet, I. Durand, J. Banchereau, C. Caux, F. Briere. 1997. Human dendritic cells skew isotype switching of CD40-activated naive B cells towards IgA1 and IgA2. J. Exp. Med. 185:1909–18 Liu, Y.J., G. Grouard, O. de Bouteiller, J. Banchereau. 1996. Follicular dendritic cells and germinal centers. Int. Rev. Cytol. 166:139–79 Chaplin, D.D., Y. Fu. 1998. Cytokine regulation of secondary lymphoid organ development. Curr. Opin. Immunol. 10:289–97 Wykes, M., A. Pombo, C. Jenkins, G.G. MacPherson. 1998. Dendritic cells inter-

237.

238.

239.

240.

241.

242.

243.

act directly with naive B lymphocytes to transfer antigen and initiate class switching in a primary T-dependent response. J. Immunol. 161:1313–19 Grouard, G., I. Durand, L. Filgueira, J. Banchereau, Y.J. Liu. 1996. Dendritic cells capable of stimulating T cells in germinal centres. Nature 384:364–67 Szakal, A.K., M.H. Kosco, J.G. Tew. 1989. Microanatomy of lymphoid tissue during humoral immune responses: structure function relationships. Ann. Rev. Immunol. 7:91–109 Ohshima, Y., Y. Tanaka, H. Tozawa, Y. Takahashi, C. Maliszewski, G. Delespesse. 1997. Expression and function of OX40 ligand on human dendritic cells. J. Immunol. 159:3838–48 Moore, P.A., O. Belvedere, A. Orr, K. Pieri, D.W. LaFleur, P. Feng, D. Soppet, M. Charters, R. Gentz, D. Parmelee, Y. Li, O. Galperina, J. Giri, V. Roschke, B. Nardelli, J. Carrell, S. Sosnovtseva, W. Greenfield, S.M. Ruben, H.S. Olsen, J. Fikes, D.M. Hilbert. 1999. BLyS: member of the tumor necrosis factor family and B lymphocyte stimulator. Science 285:260–63 Schneider, P., F. MacKay, V. Steiner, K. Hofmann, J.L. Bodmer, N. Holler, C. Ambrose, P. Lawton, S. Bixler, H. AchaOrbea, D. Valmori, P. Romero, C. Werner-Favre, R.H. Zubler, J.L. Browning, J. Tschopp. 1999. BAFF, a novel ligand of the tumor necrosis factor family, stimulates B cell growth. J. Exp. Med. 189:1747–56 Mueller, C.G., M.C. Rissoan, B. Salinas, S. Ait-Yahia, O. Ravel, J.M. Bridon, F. Briere, S. Lebecque, Y.J. Liu. 1997. Polymerase chain reaction selects a novel disintegrin proteinase from CD40activated germinal center dendritic cells. J. Exp. Med. 186:655–63 Shah, P.D. 1987. Dendritic cells but not macrophages are targets for immune regulation by natural killer cells. Cell Immunol. 104:440–45


IMMUNOBIOLOGY OF DENDRITIC CELLS 244. Geldhof, A.B., M. Moser, L. Lespagnard, K. Thielemans, P. De Baetselier. 1998. Interleukin-12-activated natural killer cells recognize B7 costimulatory molecules on tumor cells and autologous dendritic cells. Blood 91:196–206 245. Kawano, T., J. Cui, Y. Koezuka, I. Toura, Y. Kaneko, H. Sato, E. Kondo, M. Harada, H. Koseki, T. Nakayama, Y. Tanaka, M. Taniguchi. 1998. Natural killer-like nonspecific tumor cell lysis mediated by specific ligand-activated Va14 NKT cells. Proc. Natl. Acad. Sci. USA 95:5690–93 246. Fernandez, N.C., A. Lozier, C. Flament, P. Ricciardi-Castagnoli, D. Bellet, M. Suter, M. Perricaudet, T. Tursz, E. Maraskovsky, L. Zitvogel. 1999. Dendritic cells directly trigger NK cell functions: cross-talk relevant in innate anti-tumor immune responses in vivo. Nat. Med. 5:405–11 247. Cayeux, S., G. Richter, C. Becker, A. Pezzutto, B. Dorken, T. Blankenstein. 1999. Direct and indirect T cell priming by dendritic cell vaccines. Eur. J. Immunol. 29:225–34 248. Lynch, D.H., A. Andreasen, E. Maraskovsky, J. Whitmore, R.E. Miller, J.C. Schuh. 1997. Flt3 ligand induces tumor regression and antitumor immune responses in vivo. Nat. Med. 3:625–31 249. Shaw, S.G., A.A. Maung, R.J. Steptoe, A.W. Thomson, N.L. Vujanovic. 1998. Expansion of functional NK cells in multiple tissue compartments of mice treated with Flt3-ligand: implications for anticancer and anti-viral therapy. J. Immunol. 161:2817–24 250. Chambers, B.J., M. Salcedo, H.G. Ljunggren. 1996. Triggering of natural killer cells by the costimulatory molecule CD80 (B7–1). Immunity 5:311–17 251. Timmerman, J.M., R. Levy. 1999. Dendritic cell vaccines for cancer immunotherapy. Annu. Rev. Med. 50:507–29 252. Palucka, K., J. Fay, J. Banchereau. 1999. Dendritic cells and tumor immunity.

253.

254.

255.

256.

257.

258.

259.

260.

809

Curr. Opin. Oncol. Endocrine Metabolic Invest. Drug. 1:282–90 Darnell, R.B. 1996. Onconeural antigens and the paraneoplastic neurologic disorders: at the intersection of cancer, immunity, and the brain. Proc. Natl. Acad. Sci. USA 93:4529–36 Darnell, R.B. 1999. The importance of defining the paraneoplastic neurologic disorders. N. Engl. J. Med. 340:1831–33 Gabrilovich, D.I., H.L. Chen, K.R. Girgis, H.T. Cunningham, G.M. Meny, S. Nadaf, D. Kavanaugh, D.P. Carbone. 1996. Production of vascular endothelial growth factor by human tumors inhibits the functional maturation of dendritic cells. Nat. Med. 2:1096–103. Erratum. 1996. Nat. Med. 2(11):1267 Enk, A.H., H. Jonuleit, J. Saloga, J. Knop. 1997. Dendritic cells as mediators of tumor-induced tolerance in metastatic melanoma. Int. J. Cancer 73:309–16 Enk, A.H., J. Saloga, D. Becker, B.P. m. M, J. Knop. 1994. Induction of haptenspecific tolerance by interleukin 10 in vivo. J. Exp. Med. 179:1397–402 Steinbrink, K., H. Jonuleit, G. Muller, G. Schuler, J. Knop, A.H. Enk. 1999. Interleukin-10-treated human dendritic cells induce a melanoma-antigen-specific anergy in CD8(`) T cells resulting in a failure to lyse tumor cells. Blood 93:1634–42 Nair, S.K., S. Hull, D. Coleman, E. Gilboa, H.K. Lyerly, M.A. Morse. 1999. Induction of carcinoembryonic antigen (CEA)-specific cytotoxic T-lymphocyte responses in vitro using autologous dendritic cells loaded with CEA peptide or CEA RNA in patients with metastatic malignancies expressing CEA. Int. J. Cancer 82:121–24 Flamand, V., T. Sornasse, K. Thielemans, C. Demanet, M. Bakkus, H. Bazin, F. Tielemans, O. Leo, J. Urbain, M. Moser. 1994. Murine dendritic cells pulsed in vitro with tumor antigen induce tumor

810

261.

262.


263.

264.

265.

266.

267.

BANCHEREAU ET AL resistance in vivo. Eur. J. Immunol. 24:605–10 Mayordomo, J.I., T. Zorina, W.J. Storkus, L. Zitvogel, C. Celluzzi, L.D. Falo, C.J. Melief, S.T. Ildstad, W.M. Kast, A.B. Deleo, M.T. Lotze. 1995. Bone marrow-derived dendritic cells pulsed with synthetic tumour peptides elicit protective and therapeutic antitumour immunity. Nat. Med. 1:1297–302 Paglia, P., C. Chiodoni, M. Rodolfo, M.P. Colombo. 1996. Murine dendritic cells loaded in vitro with soluble protein prime cytotoxic T lymphocytes against tumor antigen in vivo. J. Exp. Med. 183:317– 22 Porgador, A., E. Gilboa. 1995. Bone marrow-generated dendritic cells pulsed with a class I-restricted peptide are potent inducers of cytotoxic T lymphocytes. J. Exp. Med. 182:255–60 Song, W., H.L. Kong, H. Carpenter, H. Torii, R. Granstein, S. Rafii, M.A. Moore, R.G. Crystal. 1997. Dendritic cells genetically modified with an adenovirus vector encoding the cDNA for a model antigen induce protective and therapeutic antitumor immunity. J. Exp. Med. 186:1247–56 Boczkowski, D., S.K. Nair, D. Snyder, E. Gilboa. 1996. Dendritic cells pulsed with RNA are potent antigen-presenting cells in vitro and in vivo. J. Exp. Med. 184:465–72 Toes, R.E., E.I. van der Voort, S.P. Schoenberger, J.W. Drijfhout, L. van Bloois, G. Storm, W.M. Kast, R. Offringa, C.J. Melief. 1998. Enhancement of tumor outgrowth through CTL tolerization after peptide vaccination is avoided by peptide presentation on dendritic cells. J. Immunol. 160:4449–56 Hsu, F.J., C. Benike, F. Fagnoni, T.M. Liles, D. Czerwinski, B. Taidi, E.G. Engleman, R. Levy. 1996. Vaccination of patients with B-cell lymphoma using autologous antigen-pulsed dendritic cells. Nat. Med. 2:52–58

268. Nestle, F.O., S. Alijagic, M. Gilliet, Y. Sun, S. Grabbe, R. Dummer, G. Burg, D. Schadendorf. 1998. Vaccination of melanoma patients with peptide- or tumor lysate-pulsed dendritic cells. Nat. Med. 4:328–32 269. Murphy, G.P., B.A. Tjoa, S.J. Simmons, J. Jarisch, V.A. Bowes, H. Ragde, M. Rogers, A. Elgamal, G.M. Kenny, O.E. Cobb, R.C. Ireton, M.J. Troychak, M.L. Salgaller, A.L. Boynton. 1999. Infusion of dendritic cells pulsed with HLA-A2specific prostate-specific membrane antigen peptides: a phase II prostate cancer vaccine trial involving patients with hormone-refractory metastatic disease. Prostate 38:73–78 270. Dhodapkar, M.V., R.M. Steinman, M. Sapp, H. Desai, C. Fossella, J. Krasovsky, S.M. Donahoe, P.R. Dunbar, V. Cerundolo, D.F. Nixon, N. Bhardwaj. 1999. Rapid generation of broad T-cell immunity in humans after a single injection of mature dendritic cells. J. Clin. Invest. 104:173–80 271. Bachmann, M.F., R.M. Zinkernagel, A. Oxenius. 1998. Immune responses in the absence of costimulation: viruses know the trick. J. Immunol. 161:5791–94 272. Roake, J.A., A.S. Rao, P.J. Morris, C.P. Larsen, D.F. Hankins, J.M. Austyn. 1995. Dendritic cell loss from nonlymphoid tissues after systemic administration of lipopolysaccharide, tumor necrosis factor, and interleukin 1. J. Exp. Med. 181:2237–47 273. Yang, R.B., M.R. Mark, A. Gray, A. Huang, M.H. Xie, M. Zhang, A. Goddard, W.I. Wood, A.L. Gurney, P.J. Godowski. 1998. Toll-like receptor-2 mediates lipopolysaccharide-induced cellular signalling. Nature 395:284–88 274. Yoshimura, A., E. Lien, R.R. Ingalls, E. Tuomanen, R. Dziarski, D. Golenbock. 1999. Cutting edge: recognition of grampositive bacterial cell wall components by the innate immune system occurs via


275.

276.


277.

278.

279.

280.

281.

282.

Toll-like receptor 2. J. Immunol. 163: 1–5 Schwandner, R., R. Dziarski, H. Wesche, M. Rothe, C.J. Kirschning. 1999. Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll-like receptor 2. J. Biol. Chem. 274:17406–9 Urban, B.C., D.J. Ferguson, A. Pain, N. Willcox, M. Plebanski, J.M. Austyn, D.J. Roberts. 1999. Plasmodium falciparuminfected erythrocytes modulate the maturation of dendritic cells. Nature 400:73– 77 Hahn, G., R. Jores, E.S. Mocarski. 1998. Cytomegalovirus remains latent in a common precursor of dendritic and myeloid cells. Proc. Natl. Acad. Sci. USA 95:3937–42 Redpath, S., A. Angulo, N.R. Gascoigne, P. Ghazal. 1999. Murine cytomegalovirus infection down-regulates MHC class II expression on macrophages by induction of IL-10. J. Immunol. 162:6701–7 Soderberg-Naucler, C., K.N. Fish, J.A. Nelson. 1997. Reactivation of latent human cytomegalovirus by allogeneic stimulation of blood cells from healthy donors. Cell 91:119–26 Frankel, S.S., B.M. Wenig, A.P. Burke, P. Mannan, L.D. Thompson, S.L. Abbondanzo, A.M. Nelson, M. Pope, R.M. Steinman. 1996. Replication of HIV-1 in dendritic cell-derived syncytia at the mucosal surface of the adenoid. Science 272:115–17 Fugier-Vivier, I., C. Servet-Delprat, P. Rivailler, M.C. Rissoan, Y.J. Liu, C. Rabourdin-Combe. 1997. Measles virus suppresses cell-mediated immunity by interfering with the survival and functions of dendritic and T cells. J. Exp. Med. 186:813–23 Grosjean, I., C. Caux, C. Bella, I. Berger, F. Wild, J. Banchereau, D. Kaiserlian.

283.

284.

285.

286.

287.

288.

289.

811

1997. Measles virus infects human dendritic cells and blocks their allostimulatory properties for CD4` T cells. J. Exp. Med. 186:801–12 Choe, H., M. Farzan, Y. Sun, N. Sullivan, B. Rollins, P.D. Ponath, L. Wu, C.R. Mackay, G. LaRosa, W. Newman, N. Gerard, C. Gerard, J. Sodroski. 1996. The b-chemokine receptors CCR3 and CCR5 facilitate infection by primary HIV-1 isolates. Cell 85:1135–48 Werling, D., J.C. Hope, P. Chaplin, R.A. Collins, G. Taylor, C.J. Howard. 1999. Involvement of caveolae in the uptake of respiratory syncytial virus antigen by dendritic cells. J. Leukocyte Biol. 66:50– 58 Cosman, D., N. Fanger, L. Borges. 1999. Human cytomegalovirus, MHC class I and inhibitory signalling receptors: more questions than answers. Immunol. Rev. 168:177–85 Pinchuk, L.M., P.S. Polacino, M.B. Agy, S.J. Klaus, E.A. Clark. 1994. The role of CD40 and CD80 accessory cell molecules in dendritic cell-dependent HIV-1 infection. Immunity 1:317–25 Pope, M., M.G. Betjes, N. Romani, H. Hirmand, P.U. Cameron, L. Hoffman, S. Gezelter, G. Schuler, R.M. Steinman. 1994. Conjugates of dendritic cells and memory T lymphocytes from skin facilitate productive infection with HIV-1. Cell 78:389–98 Granelli-Piperno, A., M. Pope, K. Inaba, R.M. Steinman. 1995. Coexpression of NF-c B/Rel and Sp1 transcription factors in human immunodeficiency virus 1induced, dendritic cell-T-cell syncytia. Proc. Natl. Acad. Sci. USA 92:10944–48 Grassi, F., A. Hosmalin, D. McIlroy, V. Calvez, P. Debre, B. Autran. 1999. Depletion in blood CD11c-positive dendritic cells from HIV-infected patients. AIDS 13:759–66

BANCHEREAU ET AL C-1

BLOOD CCR6+ CCR7-

CD34+ Progenitors

TISSUES

MIP-3α PATHOGEN ENTRY

BONE MARROW

immature DC


B

CCR6CCR7++ mature DC

6Ckine MIP-3β

CCR6CCR7+

T DRAINING LYMPH NODES

Figure 3 The critical role of chemokines in a dendritic-cell odyssey. Precursor and immature Langerhans cells that display CCR6 are attracted by the epithelium that expresses the specific ligand MIP-3α. Upon antigen capture and activation, Langerhans cells detach from keratinocytes by downregulating E-cadherin, and they traverse the basal membrane by secreting proteases such as collagenase. Meanwhile CCR6 is replaced by CCR7, whose ligands are (a) 6Ckine, which is expressed on lymphatic vessel walls, and (b) MIP-3β, which is expressed in the T cell areas of lymphoid organs. This may guide the maturing DCs to the T cell areas where they will start to produce chemokines that attract lymphocytes.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:813–827 Copyright q 2000 by Annual Reviews. Al rights reserved

AN ADDRESS SYSTEM IN THE VASCULATURE OF NORMAL TISSUES AND TUMORS E. Ruoslahti and D. Rajotte Cancer Research Center, The Burnham Institute, La Jolla, California 92037; e-mail: [email protected]


Key Words homing, metastasis, angiogenesis, peptides, endothelial cells Abstract The vasculature of individual tissues is highly specialized. The endothelium in lymphoid tissues expresses tissue-specific receptors for lymphocyte homing, and recent work utilizing phage homing has revealed an unprecedented degree of specialization in the vasculature of other normal tissues. In vivo screening of libraries of phage that displace random peptide sequences on their surfaces has yielded specific homing peptides for a large number of normal tissues. The tissue-specific endothelial molecules to which the phage peptides home may serve as receptors for metastasizing malignant cells. Probing of tumor vasculature has yielded peptides that home to endothelial receptors expressed selectively in angiogenic neovasculature. These receptors, and those specific for the vasculature of individual normal tissues, are likely to be useful in targeting therapies to specific sites.

INTRODUCTION The vascular bed of an individual tissue can express molecules specific for that particular tissue. In particular, various lymphoid tissues possess such selectively expressed markers, and these markers are used by circulating cells as they home to a specific tissue. This is particularly true for the lymphoid tissues, where the high endothelium is composed of cells that express unique adhesion molecules for lymphocyte homing (1–3). In another type of homing phenomenon, leukocytes bind to endothelia at inflammatory sites by recognizing adhesion molecules induced in the endothelium by the inflammation (4). Tumor metastasis into preferred organs may represent yet another cell homing phenomenon that depends on an adhesive interaction between the tumor cells and organ-specific endothelial markers. Indeed, endothelial cell membrane vesicles bind preferentially those tumor cells that metastasize to the tissue of origin of the endothelial vesicles (5, 6). Quite recently, pluripotent stem cells from the bone marrow (7) have been shown to home to sites of tissue injury, where they appear to contribute to tissue repair by differentiating into a variety of cell types, including bone, cartilage, muscle, and endothelial cells (8–10). The signals that bring 0732–0582/00/0410–0813/$14.00

813

814

RUOSLAHTI n RAJOTTE

the bone marrow–derived cells to the sites of tissue repair are not known, but blood vessels undergoing angiogenesis express a number of cell surface molecules that are not found in normal vasculature (see below) and that could serve as homing receptors for the bone marrow–derived cells. This review focuses on specific features of the vasculature in various nonlymphoid tissues and in angiogenesis.


DETECTION OF ENDOTHELIAL SPECIALIZATION IN TISSUES Identification of organ-specific vascular markers outside the lymphoid system has progressed slowly, at least partly because of difficulties in isolating pure populations of endothelial cells from tissues. Moreover, isolated and cultured cells may lose their tissue-specific traits when the cells are removed from their microenvironment for in vitro culture (11–13). The potential instability of the endothelial cell phenotype in vitro has led to the development of in situ and in vivo methods for the identification of endothelial cell surface markers. The first homing molecules for lymphocytes were identified by using lymphocyte binding to tissue sections in a procedure known as the Stamper-Woodruff assay (1). Other types of tissues have been analyzed by isolating endothelial membranes from individual tissues through binding onto microbeads, followed by gel electrophoretic analysis and monoclonal antibody production (14, 15). Lectin staining of tissue sections (16, 17) and monoclonal antibodies prepared after immunization with whole tissues (14) have also revealed specific features in individual vascular beds. The new genetic and proteomics techniques, such as comparison of mRNA species expressed by endothelial cells from different tissues in microchip displays (18, 19), should prove very useful in the discovery of further markers of tissue-specific vascular specialization. Phage display peptide libraries are commonly used to obtain defined peptide sequences interacting with a particular molecule. We have used phage as a surrogate for homing cells to explore the specialization of various vascular beds in vivo (20–22). The method makes use of peptide libraries expressed on the surface of phage and containing as many as 109 different peptides (23). The peptides are expressed as a fusion with one of the phage surface proteins, and the phage carrying peptides with the desired properties are selected on the basis of binding to a target molecule. The strength of this technology is its ability to identify interactive regions of proteins and other molecules without preexisting notions about the nature of the interaction. Phage libraries have been used to select for peptides that bind immobilized proteins (24–27), carbohydrates (28), and for peptides that bind to cultured cells (29). This method, in vivo screening of peptide libraries displayed on phage, allows detection of peptides capable of homing to selected target tissues through the circulation (Figure 1). Our strategy has been to screen the phage libraries in vivo for peptides that home to specific sites in the vasculature. It may seem surprising that in vivo phage


VASCULAR ADDRESS SYSTEM

815

Figure 1 Schematic illustration of the role of cell adhesion to tissue-specific endothelial cell surface proteins in the homing of metastatic tumor cells.

libraries would yield tissue-specific homing peptides, rather than peptides that home nonselectively to any tissue. We believe that the ability of the method to select for tissue-specificity results from a combination of negative selection against phage capable of binding at nontargeted sites and positive selection for the target. This would leave only those phage capable of binding selectively to the intended target tissue to be enriched there.

VASCULAR SPECIALIZATION IN NONLYMPHOID TISSUES Several tissue-selective molecular differences in endothelia have been reported. Investigators have prepared antibodies that specifically recognize endothelial cell surface components in one tissue only (14, 30). Endothelial cell surface differ-


816

RUOSLAHTI n RAJOTTE

ences have also been detected by Dolichos biflorus lectin binding (16) and by in situ glycoprotein labeling of the vasculature in liver, kidney, and brain (17). Organ-specific transcriptional regulation of von Willebrand factor expression by endothelial cells (31, 32) and differential sensitivity of endothelial cell in various organs to endotoxin-induced injury (33) provide additional indications of individuality among endothelia at different sites. Selective binding of metastatic tumor cells to blood vessels in lungs (5, 34) and other sites of preferential metastasis (6, 35) also supports the existence of tissue-specific endothelial individuality. In vivo phage library screening has allowed a systematic analysis of tissuespecific vascular specificities. Phage that are capable of homing selectively to the vasculature of the brain and kidney (20), to lung, skin, pancreas, intestine, uterus, adrenal gland, and retina (22) have been described. Moreover, a peptide that homes to muscle tissue has been isolated from a phage library by using a combination of in vitro selection on cultured muscle fibers and in vivo homing (36). In this case, the authors’ interpretation is that the peptide binds to muscle cells, but vascular targeting does not seem excluded by their data. Additionally, several other organs have been targeted in recent work from our laboratory (R Pasqualini, W Arap, D MacKenna, D Rajotte, J Hoffman, E Ruoslahti, unpublished). One interesting feature revealed by these sequences is that the binding motif in the targeting peptides is often a tripeptide or tetrapeptide that appears in different sequence contexts. In vivo screens have yielded the SRL tripeptide as a brain-homing motif (20), an RDV motif in peptides that home to the retina (22), and a CGFE motif in lung homing peptides (22, 37; see below). In a different but related situation, the RGD motif is known to be important for integrin binding in distinct molecular contexts (38). Thus, many adhesive interactions seem to derive their specificity from a very small recognition motif. These findings establish a system of tissue-specific individuality of vascular beds. The homing peptides bind to vascular structures, most likely to the endothelial cells, in the target tissues (20–22, 39). Thus, the vasculature of many, perhaps all, organs express specific features in the form of markers that can be detected by phage screening. In one case, pancreas and uterus, such a signature is shared by two unrelated tissues (22). The total number of organs targeted so far is more than 10. Nonspecific binding of phage to the tissues that contain reticuloendothelial tissue (liver and spleen; 39, 40) initially prevented us from screening for these organs. Quite recently, including noninfective phage as a competitor to prevent the nonspecific phage binding has allowed the isolation of liverhoming peptides (R Pasqualini, E Koivunen, E Ruoslahti, unpublished results). These results show that endothelia from individual tissues are not alike; phage library screening, in particular, has revealed an unprecedented extent of tissuespecific molecular individuality in vascular beds. Moreover, the phage screening has been successful with every organ chosen for targeting in our laboratory so


817

far, raising the possibility that every tissue carries a tissue-specific signature in its vasculature that makes it different from the vasculature of other tissues. The potential of in vivo phage display and other methods in the identification of tissue-specific vascular markers is obviously not exhausted. For example, phage display has often yielded peptides with unrelated sequences that can home to the same vascular bed (20, 22). This suggests the presence of more than one tissue-specific marker molecule in these vascular locations. However, in some cases, the same motif has appeared repeatedly in independent experiments (20). This result indicates that, while a given vascular bed may have multiple specific markers, the number of such markers is finite.


TISSUE-SPECIFIC ENDOTHELIAL MARKER MOLECULES While probing of the vasculature with antibodies, lectins, tumor cells, and phage has revealed much heterogeneity in the vasculature, only a few of the differences that make vascular beds in individual tissues unique have been elucidated at the molecular level. The lung vasculature has been analyzed in this regard in more detail than any other normal tissue. A study of endothelial surface molecules capable of mediating adhesion of metastasizing tumor cells to lung endothelium identified two proteins that are selectively expressed in lung vasculature, Lu-ECAM-1, a protein with sequence homology to chloride channels (41), and dipeptidyl peptidase IV (34). Phage screening performed in our laboratory has established membrane dipeptidase (MDP), a cell surface zinc metalloprotease involved in the metabolism of glutathione, leukotriene D4, and certain b-lactam antibiotics (42), as a marker of lung endothelium (37). MDP recognizes and cleaves the Cys-Gly (CG) dipeptide (42). In agreement with this specificity, a set of lung-homing peptides that share the sequence CGFE home to lung vasculature and bind to MDP (37). At least one of these peptides, CGFECVRQCPERC, is an inhibitor of MDP enzymatic activity. MDP is expressed mainly in the lung and kidney; low level of MDP expression has been detected in several other tissues (42). In the kidney, MDP expression is concentrated in the brush border region of the proximal tubules, and it is not expressed in the blood vessels. In contrast, the endothelium in lung blood vessels, both the in alveolar capillaries and in microvessels of the tracheal submucosa, expresses MDP (42–44). These results illustrate a characteristic of in vivo phage display that is important to note: It can select for molecules that, while specific for a given vascular bed within the vasculature, may be expressed by other types of cells in a manner that is not tissue specific. An interesting feature of peptide library screening in general is that it tends to yield peptides that bind to functionally important regions of their target proteins.


818

RUOSLAHTI n RAJOTTE

As discussed above, the lung-homing CGFE peptide is an inhibitor of the MDP peptidase activity, indicating that the peptide binds to the catalytic site in the enzyme. The majority of peptides obtained by screening on integrins bind to the ligand-binding site of the integrin (27, 45). Antibodies as targets primarily yield peptides that bind to the antigen-binding site (23). Because functionally important sites in proteins tend to be well conserved, the binding of peptides to these sites makes peptides likely to be reactive across species, a characteristic that may in some situations offer advantages over antibodies. The origin of the tissue-specific endothelial cell gene expression revealed by the variation detected in the phage screening and by other methods poses some intriguing questions. Is the expression of the tissue-specific markers intrinsic to the endothelial cells in a given location? Or is it induced by the tissue the endothelial cells reside in? Would the markers be retained in tumors? Do the tissuespecific vascular markers provide some tissue-specific metabolic function, or might they play a role in cell trafficking? Full exploration of these questions will have to await further research, but some preliminary answers are available. A von Willebrand factor promoter-reporter gene construct that is expressed in the endothelium of the heart, but not of the lungs, has been used to study the ability of various tissue environments to induce the expression of this gene (31). Blood vessels growing into subcutaneously transplanted heart tissue acquired the ability to express the gene, whereas those growing into similarly transplanted lung tissue did not. The new gene expression in the ingrowing endothelial cells was dependent on the induction of platelet-derived growth factor (PDGF) AB heterodimer expression. The PDGF-A subunit was constitutively expressed by the endothelial cells, and the PDGF-B subunit expression was induced in them by the cardiac myocytes (32). These results indicate that, at least in some cases, the expression of tissue-specific endothelial markers may be induced under the influence of the parenchymal tissue. At least some tissue-specific endothelial markers are expressed both by the endothelial and parenchymal cells in a given tissue (30, 37). MDP, which is expressed in the lung by endothelial cells as well as various types of epithelial cells (42–44), is an example of this expression pattern. It may be that the agent that induces the expression of these markers acts on more than one cell type.

MARKERS OF BLOOD VESSELS UNDERGOING ANGIOGENESIS Tumors have a specialized vasculature. A tumor generally cannot grow beyond the size of about 1 mm in diameter without the support of blood vessels. Because the tumor is new tissue, its blood vessels will also have to be new. These vessels form through a process known as angiogenesis, the sprouting of new blood vessels



819

from existing ones (46). The new vessels are biochemically and structurally different from normal resting blood vessels. One prominent difference is the expression of certain integrins in vasculature that is undergoing angiogenesis. The avb3 and avb5 integrins are upregulated in angiogenic tumor endothelial cells (47, 48). To what extent the av integrins are restricted to sites of angiogenesis remains to be analyzed in detail; avb5 integrin expression has been reported in the thymic vasculature (49). The b1 integrins a1b1 and a2b1 may also be important in angiogenesis; in one study, vascular endothelial growth factor (VEGF) upregulated the expression of these integins in cultured endothelial cells, and antibodies against these integrins inhibited VEGFdependent angiogenesis in vivo (50). In vivo screening of phage libraries for tumor-homing phage has yielded a panel of peptide motifs that recognize markers in tumor vasculature. These motif include RGD in the cyclic peptide CDCRGDCFC (termed RGD-4C), and NGR in a number or peptides, e.g. CNGRC (21). A third motif, GSL, was found frequently in the screenings with various types of tumors. The homing of these phage is highly selective and can inhibited by coinjecting the appropriate soluble peptide with the phage. The inhibition is specific; the RGD-4C peptide inhibits the homing of the RGD4-C phage and has a slight effect on the homing of the CNGRC phage, whereas the CNGRC peptide is an efficient inhibitor of the cognate phage but has no effect on the RGD4-C phage. The RGD-4C peptide has been previously identified as a selective binder of the avb3 and avb5 integrins (51) and shown to home to the vasculature of various kinds of tumors (39). Given that these integrins are expressed in tumor vasculature at elevated levels, the selection of a peptide that binds to them provides a validation of the tumor-screening procedure. With its four cysteine residues, the RGD-4C can assume three possible disulfide-bonded configurations. It is likely that only one of these is active in the av integrin binding because the presentation of the RGD sequence in the context of a protein or peptide determines its integrinbinding specificity (38). Future work will have to establish the binding properties of individually prepared RGD-4C conformers. The tumor homing of the phage carrying the RGD, NGR, and GSL motif peptides is independent of the tumor’s origin. It depends on the angiogenic characteristics of tumor vasculature because some of these phage also home to neovasculature in the eye (M Hagedorn, W Arap, E Ruoslahti, unpublished). We are currently quantitating the phage homing in various types of angiogenesis models to determine whether any of the peptides might favor one type of angiogenesis over another. The av integrins are known to be expressed in angiogenesis caused by various types of stimuli. Thus, arthritic synovium, the vasculature of healing wounds, and retinal neovasculature all express these integrins (52, 53). Surprisingly, the receptor for the NGR tumor-homing peptides, despite the similarity of NGR and RGD, is not an integrin (R Pasqualini, E Koivunen, R Kain, J Lahden-


820

RUOSLAHTI n RAJOTTE

ranta, M Sakamoto, A Struhn, RA Ashmun, L Shapiro, W Arap, E Ruoslahti, submitted). Angiogenesis is driven by various receptor tyrosine kinases and their ligands. VEGF receptors play a critical role in angiogenesis (54–56). They also are expressed at elevated levels in tumor blood vessels and in other types of angiogenesis (57). The receptors Flt-1 (VEGFR-1) and KDR/Flk-1 (VEGF-2) are expressed in endothelial cells of blood vessels (54), while the expression of Flt4 (VEGFR-3; 58) is restricted to lymphatic endothelium in the adult (59). The Tie receptors are another class of tyrosine kinase receptors expressed almost exclusively in endothelial cells; the Tie receptors and their ligands also have an essential role in vascular development and play a role in angiogenesis (60, 61). Enhanced expression of the Tie receptors (Tie-1 and Tie-2/Tek) has been found in the blood vessels of a variety of cancers (56, 60). The Eph receptors and their ligands, ephrins, also play a role in angiogenesis, as well as in the development of blood vessels (62–66). Interestingly, an ephrin has been found to be expressed on the arterial side, whereas one of its Eph receptors marks the venous side of developing vasculature in the embryo (65, 66). The expression of these proteins in the adult vasculature remains to be studied in detail. Significantly, inhibitory anti-ephrin antibodies can have an anti-angiogenic effects in corneal neovascularization tests (63). Matrix metalloproteases (MMPs) are involved in cell motility and invasiveness, including that of endothelial cells (67). Phage displaying a peptide that binds to MMP-2 and MMP-9 specifically homes to tumor vasculature (68). This result indicates that one, or both, of these MMPs are specifically expressed in tumor vasculature and available for phage binding from the circulation. Oncofetal fibronectin is an alternatively spliced form of fibronectin that contains an extra fibronectin type III domain. This form of fibronectin is present in angiogenic but not in mature blood vessels. A monoclonal antibody that recognizes this protein was isolated from a phage library that expresses human antibodies and was then shown to accumulate in the vasculature of F9 cell murine teratocarcinomas after an intravenous injection (69). Owing to the human origin of this antibody, it is unlikely to be immunogenic in human subjects, making it potentially useful for therapeutic applications. Aminopeptidase A (APA) and the NG2 proteoglycan are markers of pericytes, rather than endothelial cells. They are both highly and selectively expressed in vasculature undergoing angiogenesis (70, 71). Aminopeptidase A is a zinc metalloenzyme that removes amino acids from the N terminus of proteins and peptides. Its physiological role may be to modify the activities of various biologically active peptides. NG2 is a transmembrane protein that carries chondroitin sulfate side chains. The function of NG2 is not well understood, but it binds platelet-derived growth factor (PDGF) and enhances the ability of PDGF-AA to signal through its receptor, PDGFRa (72). Interestingly, even though NG2 is thought to be


821

expressed only in pericytes and not in endothelial cells, phage that display NG2binding peptides home specifically into tumors in vivo (73). The apparent explanation is that the pericytes are available for circulating phage in tumor vasculature.


SPECIAL FEATURES OF LYMPHATIC VESSELS The endothelial cells in the lymphatic system differ from blood vessel endothelial cells in many ways; relevant to the topic of this article, the lymphatic endothelial cells have their own cell surface markers. The specialization of these cells in lymph nodes is well appreciated; specific cell surface markers serve as homing receptors in lymphocyte trafficking to specific parts of the lymphatic tissues. This aspect of the lymphatic system has been extensively covered in other reviews (e.g. 1). The endothelial cells in the lymphatic vessels themselves are also specialized. The recent discoveries that VEGF-R3 (Flt-4) is primarily expressed in lymphatic endothelial cells in the adult and that its ligand, VEGF-C, enhances the growth of lymphatic vessels have been major advances in the field (56, 58). VEGF-C also binds to VEGF-R2 in blood vessels and is angiogenic (56). This should not be a problem if VEGF-C is used to stimulate lymphangiogenesis, which may now for the first time be possible as a therapy for conditions involving perturbed lymphatic drainage. In addition to VEGF-R3, another cell surface marker, podoplanin, has been reported to be specific for lymphatic endothelial cells (74). It will be interesting to see to what extent lymphatic vessels might share the tissue-specific ‘‘addresses’’ of the blood vessels, or whether they might possess their own parallel address system. In vivo phage display may lend itself to the analysis of lymphatic vessel homing receptors, even though the delivery of the phage to the lymphatic system and the recovery of phage bound to the lymphatic vessels will be more difficult than has been the case with blood vessels.

VASCULAR ADDRESSES IN CELL TRAFFICKING The ability of lymphocytes to home to specific sites in the lymphatic system is reflected in the same tendencies in tumors. Transfecting the a4 integrin subunit into tumor cells, which results in the expression of the a4b1 integrin, can change the metastatic pattern of the transfected cells; having favored the lungs as the natural site of metastasis, intravenously injected Chinese hamster ovary cells and K562 human erythroleukemia cells metastasized into the lungs and bones after a4 transfection (75). LuECAM-1, a protein with sequence homology to chloride channels, is responsible for the adhesion of B16 murine melanoma cells to lung endothelia


822

RUOSLAHTI n RAJOTTE

(41). Dipeptidyl peptidase IV (DPP IV), also selectively expressed in lung vasculature, promotes the metastasis of breast and prostate carcinoma cells into the lungs (34). Like DPP IV, MDP may serve as a receptor for tumor cells metastasizing into the lungs (D Rajotte, R Pasqualini, W Arap, E Ruoslahti, submitted). The tumor cells that use the peptidases as homing receptors in the lung vasculature may express a cell surface protein containing a sequence that binds to the peptidases. That sequence would probably be a poor substrate for the enzyme, allowing stable binding to the endothelium. Fibronectin present on the surface of the metastatic cells is thought to be the ligand for DPP IV-dependent homing of the breast cancer cells to lung vasculature (76); the molecules responsible for the binding of tumor cells to MDP are unknown. The identification of two proteolytic enzymes and a presumed ion channel as receptors for metastasizing tumor cells in the lungs shows that adhesive events resulting in cell homing can depend on molecules that are distinct from classical adhesion molecules. It is also interesting that metastasis into the lungs, which is usually thought to be common because the lung vasculature is the first vascular bed encountered by tumor cells originating for most tissue, may nevertheless be dependent on selective adhesion of the tumor cells to lung vasculature. Tumor cells that express the Lewis x carbohydrate at their surface are more metastatic into the lungs than control cells lacking it (77, 78). In this case, the receptor is likely to be an Lewis x-recognizing lectin. As receptors for additional vascular homing peptides are identified, the molecular basis of tissue-specific cell homing will become better understood.

VASCULAR ADDRESSES IN THERAPEUTIC TARGETING The tissue-specific vascular markers provide new opportunities for the targeting of therapeutic compounds, such as genes and drugs. The phage itself is one of the gene therapy vectors under development (79); its successful targeting into many normal tissues and tumors shows that the distribution of a vector into a chosen target can be substantially enhanced. More conventional gene therapy vectors have also been modified with homing peptides and antibodies. Several groups have shown that inserting an RGD sequence that binds to av integrins changes the tropism of the virus such that the virus acquires the ability to infect cells expressing these integrins (80, 81). Based on the phage results, these inserted peptides may also function as homing devices in vivo. Drug targeting with homing peptides shows promise. Coupling of doxorubicin or a pro-apoptotic peptide to an integrin-binding RGD peptide or NGR peptide yields compounds that have enhanced antitumor activities and reduced side effects (21; 82). Given that a vascular address system encompasses most, perhaps all organs, much wider application of targeted therapies is likely to be possible.


823

ACKNOWLEDGMENTS The authors’ work was supported by NIH grants CA 74238 and CA 78804 and Cancer Center Support Grant CA 30199. DR is a Terry Fox Research Fellow of the National Cancer Institute of Canada. Visit the Annual Reviews home page at www.AnnualReviews.org.


LITERATURE CITED 1. Butcher EC, Picker LJ. 1996. Lymphocyte homing and homeostasis. Science 272:60–66 2. Lasky LA. 1992. Selectins: interpreters of cell-specific carbohydrate information during inflammation. Science 258:964– 69 3. Salmi M, Jalkanen S. 1997. How do lymphocytes know where to go: current concepts and enigmas of lymphocyte homing. Adv. Immunol. 64:139–218 4. Springer TA. 1994. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76:301–4 5. Johnson RC, Augustin-Voss HG, Zhu D, Pauli BU. 1991. Endothelial cell membrane vesicles in the study of organ preference of metastasis. Cancer Res. 51:394–99 6. Auerbach R, Lu WC, Pardon E, Gumkowski F, Kaminska G, Kaminski M. 1987. Specificity of adhesion between murine tumor cells and capillary endothelium: an in vitro correlate of preferential metastasis in vivo. Cancer Res. 47:1492–96 7. Caplan AI. 1991. Mesenchymal stem cells. J. Orthopaedic Res. 9:641–50 8. Asahara T, Murohara T, Sullivan M, van der Zee R, Li T, Witzenbichler B, Schatteman G, Isner JM. 1997. Isolation of putative endothelial cells for angiogenesis. Science 275:964–67

9. Pereira RF, O’Hara MD, Laptev AV, Halford KW, Pollard MD, Class R, Simon D, Livezey K, Prockop DJ. 1998. Marrow stromal cells as a source of progenitor cells for nonhematopoietic tissues in transgenic mice with a phenotype of osteogenesis imperfecta. Proc. Nat. Acad. Sci. USA 95:1142–47 10. Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD, Moorman MA, Simonetti DW, Craig S, Marshak DR. 1999. Multilineage potential of adult human mesenchymal stem cells. Science 284:143–47 11. Augustin HG, Kozian DH, Johnson RC. 1994. Differentiation of endothelial cells: analysis of the constitutive and activated endothelial cell phenotypes. BioEssays 16:901–6 12. Borsum T, Hagen I, Henriksen T, Carlander B. 1982. Alterations in the protein composition and surface structure of human endothelial cells during growth in primary culture. Atherosclerosis 44:367– 78 13. Bono DP, Green C. 1984. The adhesion of different cell types to cultured vascular endothelium: effects of culture density and age. Br. J. Exp. Pathol. 65:145–54 14. Ghitescu L D, Crine P, Jacobson BS. 1997. Antibodies specific to the plasma membrane of rat lung microvascular endothelium. Exp. Cell Res. 232:47–55 15. Jacobson BS, Stolz DB, Schnitzer JE.

824

16.

17.


18.

19.

20.

21.

22.

23.

24.

RUOSLAHTI n RAJOTTE 1996. Identification of endothelial cellsurface proteins as targets for diagnosis and treatment of disease. Nat. Med. 2:482–84 Ponder BAJ, Wilkinson MM. 1983. Organ-related differences in binding of Dolichos biflorus agglutinin to vascular endothelium. Dev. Biol. 96: 535–41 Belloni, PN, Nicolson GL. 1988. Differential expression of cell surface glycoproteins on various organ-derived microvascular endothelia and endothelial cultures. J. Cell. Physiol. 136:398–410 Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JCR, Trent JM, Staudt LM, Hudson J Jr, Boguski MS, Lashkari D, Shalon D, Botstein D, Brown PO. 1999. The transcriptional program in the response of human fibroblasts to serum. Science 283:83–87 Alon U, Barkai N, Notterman DA, Gish K, Ybarra S, Mack D, Levine AJ. 1999. Broad patterns of gene expression revealed by clustering analysis of tumor and normal colon tissues probed by oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 96:6745–50 Pasqualini R, Ruoslahti E. 1996. Organ targeting in vivo using phage display peptide libraries. Nature 380:364–66 Arap W, Pasqualini R, Ruoslahti E. 1998. Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science 279:377–80. Rajotte D, Arap W, Hagedorn M, Koivunen E, Pasqualini R, Ruoslahti E. 1998. Molecular heterogeneity of the vascular endothelium revealed by in vivo phage display. J. Clin. Invest. 102:430– 37 Smith GP, Scott JK. 1993. Libraries of peptides and proteins displayed on filamentous phage. Meth. Enzymol. 217: 228–57 Burritt JB, Bond CW, Doss KW, Jesaitis AJ. 1996. Filamentous phage display of oligopeptide libraries. Analyt. Biochem. 238:1–13

25. Cwirla S.E, Peters EA, Barrett RW, Dower WJ. 1990. Peptides on phage: a vast library of peptides for identifying ligands. Proc. Natl. Acad. Sci. USA 87:6378–82 26. Wrighton N C, Farrell FX, Chang R, Kashyap AK, Barbone FP, Mulcahy LS, Johnson DL, Barrett RW, Jolliffe LK, Dower WJ. 1996. Small peptides as potent mimetics of the protein hormone erythropoietin. Science 273:458–63 27. Koivunen E, Wang B, Dickinson CD, Ruoslahti E. 1994. Peptides in cell adhesion research. Meth. Enzymol. 245:346– 69 28. Peletskaya EN, Glinsky VV, Glinsky GV, Deutscher SL, Quinn TP. 1997. Characterization of peptides that bind the tumor-associated Thomsen-Friedenreich antigen selected from bacteriophage display libraries. J. Mol. Biol. 270: 374–84 29. Barry MA, Dower WJ, Johnston SA. 1996. Towards cell-targeting gene therapy vectors: selection of cell-binding peptides from random peptide-presenting phage libraries. Nat. Med. 2:299–305 30. Auerbach R, Alby L, Morrissey LW, Tu M, Joseph J. 1985. Expression of organspecific antigens on capillary endothelial cells. Microvasc. Res. 29:401–11 31. Aird WC, Edelberg JM, Weiler-Guettler H, Simmons WW, Smith TW, Rosenberg RD. 1997. Vascular bed-specific expression of an endothelial cell gene is programmed by the tissue microenvironment. J. Cell. Biol. 138:1117–24 32. Edelberg JM, Aird WC, Wu W, Rayburn H, Mamuya WS, Mercola M, Rosenberg RD 1998. PDGF mediates cardiac microvascular communication. J. Clin. Invest. 102:837–43 33. Maeda K, Abello PA, Abraham MR, Wetzel RC, Robotham JL, Buchman TG. 1995. Endotoxin induces organ-specific endothelial cell injury. Shock 3:46–50 34. Johnson RC, Zhu D, Augustin-Voss HG, Pauli BU. 1993. Lung endothelial dipeptidyl peptidase IV is an adhesion mole-


35.

36.

37.


38.

39.

40.

41.

42.

43.

44.

cule for lung-metastatic rat breast and prostate carcinoma cells. J. Cell Biol. 121:1423–32 Netland PA, Zetter BR. 1984. Organspecific adhesion of metastatic tumor cells in vitro. Science 224:1113–15 Samoylova TI, Smith BF 1999. Elucidation of muscle-binding peptides by phage display screening. Muscle Nerve 22:460–66 Rajotte D, Ruoslahti E. 1999. Membrane dipeptidase is the receptor for a lung targeting peptide identified by in vivo phage display. J. Biol. Chem. 274:11593–98 Ruoslahti E. 1996. RGD and other recognition sequences for integrins. Annu. Rev. Cell Dev. Biol. 12:697–715 Pasqualini R, Koivunen E, Ruoslahti E. 1997. AlphaV integrins as receptors for tumor targeting by circulating ligands. Nat. Biotech. 15:542–46 Geter MR, Trigg ME, Merril CR. 1973. Fate of bacteriophage lambda in nonimmune germ-free mice. Nature 246: 221–23 Elble RC, Widom J, Gruber AD, AbdelGhany M, Levine R, Goodwin A, Cheng HC, Pauli BU. 1997. Cloning and characterization of lung endothelial cell adhesion molecule-1 suggest it is an endothelial chloride channel. J. Biol. Chem. 272:27853–61 Keynan S, Hooper NM, Turner AJ. 1996. Molecular aspects of membrane dipeptidase. In Molecular and Functional Aspects of Membrane Dipeptidase in Zinc Metalloproteases in Health and Disease, ed. NM Hooper. pp. 285–309. London: Taylor & Francis Habib GM, Barrios R, Shi ZZ, Lieberman MW. 1996. Four distinct membranebound dipeptidase RNAs are differentially expressed and show discordant regulation with gamma-glutamyl transpeptidase J. Biol. Chem. 271:16273–80 Inamura T, Pardridge WM, Kumagai Y, Black KL. 1994. Differential tissue expression of immunoreactive dehydro-

45.

46.

47.

48.

49.

50.

51.

52.

53.

825

peptidase. I, a peptidyl leukotriene metabolizing enzyme. Prostaglandins Leukotrienes Essential Fatty Acids 50: 85–92 O’Neil KT, Hoess RH, Jackson SA, Ramachandran NS, Mousa SA, DeGrado WF. 1992. Identification of novel peptide antagonists for GPIIb/IIIa from a conformationally constrained phage peptide library. Proteins 14:509–15 Hanahan D, Folkman J. 1996. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353–64 Brooks PC, Clark RA, Cheresh DA. 1994. Requirement of vascular integrin avb3 for angiogenesis. Science 264:569– 71 Brooks PC, Montgomery AM, Rosenfeld M, Reisfeld RA, Hu T, Klier G, Cheresh DA. 1994. Integrin avb3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels. Cell 79:1157–64 Pasqualini R, Bodorova J, Ye S, Hemler ME. 1993. Monoclonal antibodies for study of the structure, function and distribution of b5 containing integrins. J. Cell Sci. 105:101–11 Senger DR, Claffey KP, Benes JE, Perruzzi CA, Sergiou AP, Detmar M. 1997. Angiogenesis promoted by vascular endothelial growth factor: regulation through a1b1 and a2b1 integrins. Proc. Natl. Acad. Sci. USA 94:13612–17 Koivunen E, Wang B, Ruoslahti E. 1995. Phage libraries displaying cyclic peptides with different ring sizes: ligand specificities of the RGD-directed integrins. Bio/technology 13:265–70 Eliceiri BP, Cheresh DA. 1999. The role of av integrins during angiogenesis: insights into potential mechanisms of action and clinical development. J. Clin. Invest. 103:1227–30 Hammes HP, Brownlee M, Jonczyk A, Sutter A, Preissner KT. 1996. Subcutaneous injection of a cyclic peptide

826

54.

55.

56.

57.


58.

59.

60.

61.

62.

63.

RUOSLAHTI n RAJOTTE antagonist of vitronectin receptor-type integrins inhibits retinal neovascularization. Nat. Med. 2:529–33 Ferrara N, Davis-Smyth T. 1997. The biology of vascular endothelial growth factor. Endocrine Rev. 18:4–25 Hanahan D. 1997. Signaling vascular morphogenesis and maintenance. Science 277:48–50 Korpelainen EI, Alitalo K. 1998. Signaling angiogenesis and lymphangiogenesis.Curr. Opin. Cell Biol. 10:159–64 Brekken RA, Huang X, King SW, Thorpe PE. 1998. Vascular endothelial growth factor as a marker of tumor endothelium. Cancer Res. 58:1952–59 Jeltsch M, Kaipainen A, Joukov V, Meng XJ, Lakso M, Rauvala H, Swartz M, Fukumura D, Jain RK, Alitalo K. 1997. Hyperplasia of lymphatic vessels in VEGF-C transgenic mice. Science 276:1423–25 Jussila L, Valtola R, Partanen TA, Salven P, Heikkila P, Matikainen MT, Renkonen R, Kaipainen A, Detmar M, Tschachler E, Alitalo R, Alitalo K. 1998. Lymphatic endothelium and Kaposi’s sarcoma spindle cells detected by antibodies against the vascular endothelial growth factor receptor-3. Cancer Res. 58:1599–1604 Peters KG, Coogan A, Berry D, Marks J, Iglehart JD, Kontos CD, Rao P, Sankar S, Trogan E. 1998. Expression of Tie/Tek in breast tumour vasculature provides a new marker for evaluation of tumour angiogenesis. Br. J. Cancer 77:51–56 Holash J, Maisonpierre PC, Compton D, Boland P, Alexander CR, Zagzag D, Yancopoulos GD, Wiegand SJ. 1999. Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF. Science 284:1994–98 Pasquale EB. 1997. The Eph family of receptors. Curr. Opin. Cell Biol. 9:608– 15 Pandey, A, Shao, H, Marks, RM, Polverini PJ, Dixit VM. 1995. Role of B61, the ligand for the Eck receptor tyrosine

64.

65.

66.

67.

68.

69.

70.

71.

72.

kinase, inTNF-alpha-induced angiogenesis. Science 268:567–69 Yancopoulos GD, Klagsbrun M, Folkman J. 1998. Vasculogenesis, angiogenesis, and growth factors: ephrins enter the fray at the border. Cell 93:661–64 Adams RH, Wilkinson GA, Weiss C, Diella F, Gale NW, Deutsch U, Risau W, Klein R. 1999. Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial/ venous domains, vascular morphogenesis, and sprouting angiogenesis. Genes Dev. 13:295–306 Wang HU, Chen ZF, Anderson DJ. 1998. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell 93:741–53 Birkedal-Hansen H. 1995. Proteolytic remodeling of extracellular matrix. Curr. Opin. Cell Biol. 7:728–35 Koivunen E, Arap W, Valtanen H, Rainisalo A, Medina OP, Heikkila P, Kantor C, Gahmberg CG, Salo T, Konttinen YT, Sorsa T, Ruoslahti E, Pasqualini R. 1999. Tumor targeting with a selective gelatinase inhibitor. Nat. Biotech. 17:768–74 Neri D, Carnemolla B, Nissim A, Leprini A, Querze G, Balza E, Pini A, Tarli L, Halin C, Neri P, Zardi L,Winter G. 1997. Targeting by affinity-matured recombinant antibody fragments of an angiogenesis associated fibronectin isoform. Nat. Biotech. 15:1271–75 Schlingemann RO, Rietveld FJ, de Waal RM, Ferrone S, Ruiter DJ. 1990. Expression of the high molecular weight melanoma-associated antigen by pericytes during angiogenesis in tumors and in healing wounds. Am. J. Pathol. 136: 1393–405 Schlingemann RO, Oosterwijk E, Wesseling P, Rietveld FJ, Ruiter DJ. 1996. Aminopeptidase A is a constituent of activated pericytes in angiogenesis. J. Pathol. 179:436–42 Grako KA, Stallcup WB. 1995. Partici-


73.


74.

75.

76.

77.

pation of the NG2 proteoglycan in rat aortic smooth muscle cell responses to platelet-derived growth factor. Exp. Cell Res. 221:231–40 Burg MA, Pasqualini R, Arap W, Ruoslahti E, Stallcup B. 1999. NG2 proteoglycan-binding peptides home to tumor neovasculature. Cancer Res. 59: 2869– 74 Breiteneder-Geleff S, Soleiman A, Kowalski H, Horvat R, Amann G, Kriehuber E, Diem K, Weninger W, Tschachler E, Alitalo K, Kerjaschki D. 1999. Angiosarcomas express mixed endothelial phenotypes of blood and lymphatic capillaries: podoplanin as a specific marker for lymphatic endothelium. Am. J. Pathol. 154:385–94 Matsuura N, Puzon-McLaughlin W, Irie A, Morikawa Y, Kakudo K, Takada Y. 1996. Induction of experimental bone metastasis in mice by transfection of integrin alpha 4 beta 1 into tumor cells Am. J. Pathol. 148:55–61 Cheng HC, Abdel-Ghany M, Elble RC, Pauli BU. 1998. Lung endothelial dipeptidyl peptidase IV promotes adhesion and metastasis of rat breast cancer cells via tumor cell surface-associated fibronectin. J. Biol. Chem. 273:24207–15 Takada A, Ohmori K, Yoneda T, Tsuyuoka K, Hasegawa A, Kiso M, Kannagi R. 1993. Contribution of carbohydrate anti-

78.

79.

80.

81.

82.

827

gens sialyl Lewis A and sialyl Lewis X to adhesion of human cancer cells to vascular endothelium. Cancer Res. 53:354– 61 Ohyama C, Tsuboi S, Fukuda M. 1999. Dual roles of sialyl Lewis X oligosaccharides in tumor metastasis and rejection by natural killer cells. EMBO J. 18:1516–25 Larocca D, Kassner PD, Witte A, Ladner RC, Pierce GF, Baird A. 1999. Gene transfer to mammalian cells using genetically targeted filamentous bacteriophage. FASEB J. 13:727–34 Wickham TJ, Tzeng E, Shears LL, Roelvink PW, Li Y, Lee GM, Brough DE, Lizonova A, Kovesdi I. 1997. Increased in vitro and in vivo gene transfer by adenovirus vectors containing chimeric fiber proteins. J. Virol. 71:8221–29 Dmitriev I, Krasnykh V, Miller CR, Wang M, Kashentseva E, Mikheeva G, Belousova N, Curiel DT. 1998. An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. J. Virol. 72:9706–13 Ellerby HM, Arap W, Andrusiak R, Ellerby LM, Kain R, Del Rio G, Ruoslahti E, Bredesen DE, Pasqualini R. 1999. Anti-tumor activity of targeted pro-apoptotic peptides. Nat. Med. 5: 1032–38



CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


GENOMIC VIEWS OF THE IMMUNE SYSTEM* Louis M. Staudt1 and Patrick O. Brown2 1 Metabolism Branch, Division of Clinical Sciences, National Cancer Institute, National Institutes of Health, Bethesda, Maryland; e-mail: [email protected] 2 Dept. of Biochemistry and Howard Hughes Medical Institute, Stanford University School of Medicine; e-mail: [email protected]


Key Words genomics, microarray, gene expression Abstract Genomic-scale experimentation aims to view biological processes as a whole, yet with molecular precision. Using massively parallel DNA microarray technology, the mRNA expression of tens of thousands of genes can be measured simultaneously. Mathematical distillation of this flood of gene expression data reveals a deep molecular and biological logic underlying gene expression programs during cellular differentiation and activation. Genes that encode components of the same multi-subunit protein complex are often coordinately regulated. Coordinate regulation is also observed among genes whose products function in a common differentiation program or in the same physiological response pathway. Recent application of gene expression profiling to the immune system has shown that lymphocyte differentiation and activation are accompanied by changes of hundreds of genes in parallel. The databases of gene expression emerging from these studies of normal immune responses will be used to interpret the pathological changes in gene expression that accompany autoimmunity, immune deficiencies, and cancers of immune cells.

INTRODUCTION The established, model-driven field of immunology is about to collide with the upstart, discovery-driven field of genomics. Traditional research in molecular biology and molecular immunology can be likened to trying to understand a movie by successively examining a few pixels (genes) at a time from each frame. Genomic approaches allow the scientist to view the entire movie in one sitting and discover complex interrelationships among the plot, characters, and recurring themes. The tension between genomic approaches and the more traditional single gene orientation of molecular biology often leads to criticism of genomic approaches as non-hypothesis-driven. Those who favor a genomic approach embrace this characterization, noting that genomic approaches are deliberately not hypothesis-limited and are instead discovery-driven. When the powerful *The US government has the right to retain a nonexclusive, royalty-free license in and to any copyright covering this paper. 0732–0582/00/0410–0829$14.00

829


830

STAUDT n BROWN

molecular tools of genomics are applied to a new biological question, discoveries will almost certainly be made that will generate new hypotheses and necessitate a reworking of existing models. The field of immunology is especially primed to receive the new insights that genomics can provide. Numerous immune cell types have been defined with high precision, and methods to culture and manipulate these cells are well developed. Such experimental systems are ideal settings in which to study genome-wide phenomena under very well controlled circumstances. Powerful techniques for the analysis of single gene mutations in lymphocytes have been developed in the mouse, yielding a plethora of precise genetic models that are ideal substrates for genomic approaches. Finally, malfunctions of the immune system give rise to a host of autoimmune diseases, immune deficiency diseases, and malignancies in need of fresh insights that may be supplied by genomic views of the pathological processes. The young field of genomics has already been somewhat arbitrarily subdivided into two separate disciplines. One branch of genomics, structural genomics, has as its immediate goal to determine complete genomic DNA sequences of the major model organisms. To date, the complete genomes of the yeast Saccharomyces cerevisiae (1), the worm Caenorhabditis elegans (2), and numerous prokaryotes have been sequenced (3). The complete genomes of these simple organisms have yielded a plethora of orthologues of human and mouse genes. New insights into the function of these evolutionarily conserved gene families are thus made possible using the more tractable genetics of these model organisms. Much of this review focuses, however, on the newly coined field of functional genomics. Broadly construed, functional genomics encompasses any experimental approach that uses genomic structural information to view and understand biological processes in a systematic and comprehensive fashion. This vast frontier, opened up by the genome sequencing projects, is just beginning to be explored. Even at this early stage, a diversity of approaches have been developed for exploring the living genome. In this review, however, we focus primarily on one of them: the genome-wide analysis of mRNA expression using DNA microarrays. Because of the central role played by regulation of mRNA levels in development and physiology and because of the deep, logical connection between the function of a gene’s product and its pattern of expression, this specific area of functional genomics research has been the richest source of new biological insights. One of the defining characteristics of functional genomic approaches is that they generate data streams that overwhelm the traditional analytical methods of biology and indeed make possible entirely new ways of looking at living systems. Throughout this review, we discuss how the field of bioinformatics has faced the challenge of organizing, distilling, and visualizing the information provided by genomic data in ways that allow biological insights to be found. The field of genomics naturally intersects with classical genetics in the study of complex genetic diseases. In polygenic disorders, the contribution of any one


GENOMIC VIEWS OF THE IMMUNE SYSTEM

831

locus to the disease phenotype is small and may be apparent only in the context of specific alleles in other genes. The current race to define allelic variants of genes in human populations is largely fueled by the desire to understand their contribution to differential disease susceptibility. Millions of single nucleotide polymorphisms exist in the human population, and recognizing the linkage or association of a single polymorphism with a disease state is a considerable challenge (4). Techniques in functional genomics provide information that can complement linkage and association methods in making the connection between genes and disease risks. For virtually every gene, variation in its expression, as a function of cell specialization, physiology, or disease, is much richer than allelic variation in that gene. Because the pattern in which each gene is expressed is so closely connected to the biological role and effects of its product, systematic studies of variation in gene expression can provide an alternative approach to linking specific genes with specific diseases and to recognizing heritable variation in genes important for immune function. For example, allelic differences in the regulatory regions of cytokine genes may influence the expression levels of cytokines during particular immune responses. An appreciation for such quantitative traits in the immune system may help unravel the genetics of autoimmune diseases and lymphoproliferative disorders.

STRUCTURAL GENOMICS AND THE IMMUNE SYSTEM Systematic studies of genomic expression programs are best pursued in two independent steps. The first step is to obtain as complete a catalog as possible of all the expressed genes in the genome. The second step is to use parallel methods, such as DNA microarray hybridization, to measure the expression of each gene in the genome over the range of conditions and cell types under investigation. Our still incomplete knowledge of the human and mouse genomic sequence and the incomplete catalog of genes in these genomes present an important challenge in functional exploration of mammalian genomes. Even when a full mammalian genomic sequence is known, it will not immediately be possible to identify all of the segments that are expressed as mRNA. Computer algorithms such as GRAIL (5) use machine learning techniques to identify putative coding regions in genomic sequences. In practice, however, these algorithms need to be supplemented by cDNA sequence data to completely annotate the exon-intron structure of a mammalian genome. Indeed, even in microbial genomes with few or no introns and much higher densities of transcribed and protein-coding sequences than are found in mammalian genomes, current algorithms for identifying genes in genomic sequences have significant false positive and false negative rates. Therefore, an indispensable component of any mammalian genome project is high-throughput, single-pass sequencing of cDNA libraries to generate expressed sequence tags (ESTs) (6), which provides a systematic set of unique labels for identifying the mRNAs that can be expressed from a genome. The current release


832

STAUDT n BROWN

of the EST database dbEST [release 070999 (7)] contains 1,476,380 human ESTs and 658,511 mouse ESTs. These numbers are much larger than the numbers of distinct transcripts represented in each set because a very large fraction of the ESTs in each set is composed of multiple representations of mRNAs that are widely or highly expressed in the cells from which the source libraries were obtained. Indeed, despite these large numbers, it is clear that that not all of the human genes are represented by an EST in this public database. To illustrate this deficiency, consider the representation of interleukin sequences in the dbEST. ESTs representing about half of the known human interleukins can be found in this database, but no ESTs representing interleukins 2, 3, 5, 9, 11, 12 beta, 14, and 17 have yet been encountered. By contrast, of the 8963 known human genes with full-length cDNA sequences, 89% are represented by an EST in the dbEST database. This discrepancy reflects the bias in the dbEST database toward genes that are widely or highly expressed and the fact that very few of the EST sequences in the public domain have come from cDNA libraries made from activated cells of the immune system. Given this example from the immune system, one wonders how many inducible genes in other specialized or rare cell types have yet to be identified. For the present, filling the gaps in our catalog of human expressed genes is a practical problem for which simple, though technically challenging, incremental solutions can often be found. Several years ago, the public EST database was strikingly deficient in sequences from B lymphocytes. This void was a serious impediment not only to the study of normal B cell development and physiology but also to the study of human lymphoid malignancies, the majority of which are derived from B cells. In order to fill this void, several libraries were created from normal and malignant human B cells and sequenced under the auspices of the Cancer Genome Anatomy Project (8, 9). As shown in Table 1, each of these cDNA libraries yielded a large number of novel ESTs, ranging from 12% to 22% of the total ESTs sequenced, most presumably representing genes never previously identified or studied. In part, this apparent high rate of gene discovery can be attributed to the paucity of previous EST sequences from B cell libraries and to the normalization process used in creating the NCI_CGAP_GCB1 library (10). Among the non-unique ESTs, some represented genes that were observed only in B cell libraries or only in other lymphoid libraries (Table 1). This example illustrates the challenge that will be faced in trying to discover the complete set of expressed human genes, including all the genes expressed at low levels or in highly specialized cells or conditions. The Unigene project at the National Center for Biotechnology Information has attempted to provide a systematic classification of EST sequences (11). Unigene uses sequence alignment methods to group overlapping cDNA sequences into clusters, each of which provisionally corresponds to a unique gene. The Unigene analysis of the B cell library ESTs also reveals a high rate of gene discovery: 1652 of the 83,240 Unigene clusters at the time of this writing are defined only by ESTs derived from B cells. Viewed in another way, approximately 10% of the

TABLE 1 High-throughout sequencing of human B cell cDNA libraries

CDNA Library Name

mRNA Source

3* ESTs 3* ESTs Only in 3* ESTs Unique to B Cell Total Library Libraries 7388

233

Unigene Analysis Unigene Clusters Unigene 3* ESTs Uniquely Clusters Only in Defined by Lymphoid Containing Libraries Library Clone Library Clone

NCI_CGAP_GCB1

Tonsillar germinal center/memory B cells 40428

NCI_CGAP_GCB0

Tonsillar germinal center/memory B cells

907

139

1

4

495

4

NCI_CGAP_Lym12

Follicular mixed small and large cell

4038

480

21

18

2381

200

NCI_CGAP_Lym5

Follicular lymphoma

1293

182

17

10

859

40

NCI_CGAP_Lym6

Mantle cell lymphoma

621

135

3

3

316

16

NCI_CGAP_CLL1

Chronic lymphocytic leukemia

8628

1085

127

54

4612

277

55915

9409

402

532

15992

1652

All B cell libraries

833


BLAST Analysis

443

13078

1058

834

STAUDT n BROWN

genes (Unigene clusters) that were sampled during the sequencing of B cell libraries were B cell–restricted. This result dramatically demonstrates our relative ignorance of the molecular biology of B lymphocytes and the need for systematic, genomic approaches to determine the expression patterns and functions of these novel genes in immune responses and other physiological processes.


GENOMIC-SCALE ANALYSIS OF GENE EXPRESSION Although posttranscriptional mechanisms are important in regulating the expression of many genes, most cellular regulation is achieved by changes in mRNA levels. Consequently, systematic studies of gene expression patterns have proven to be remarkably powerful sources of insight into gene function and biological processes. Four aspects of genome-wide gene expression analysis are particularly appealing. First is its feasibility: DNA microarrays make it easy to measure, in a single hybridization, the mRNA abundance of every gene for which either a clone or sufficient DNA sequence information exists. Second, there is a biologically rational connection between the function of a gene product and its expression pattern. Natural selection has acted to optimize simultaneously the functional properties of the product encoded by a gene and the program that dictates where, when, and in what amounts the product is made. As a rule, each gene is expressed in the specific cells and under the specific conditions in which its product makes a contribution to fitness. The richness and precision with which mRNA levels can be controlled is such that virtually every gene in the yeast genome can be distinguished from every other gene based on its pattern of expression. Therefore, even subtle variations in the expression patterns of genes can be related to corresponding differences in the functions of the products they encode. Third, promoters and the regulatory systems that act upon them function as transducers, integrating diverse kinds of information about the identity, environment, and internal state of a cell. Thus, a diversity of information that is difficult or impossible to measure is transformed into a signal that can readily be measured systematically using DNA microarrays. Learning to decode this transduced information is one of the immediate priorities of functional genomics. Finally, the set of genes expressed in a cell determines how the cell is built, what biochemical and regulatory systems are operative, and what it can and cannot do. Thus, as we learn to infer the biological consequences of gene expression patterns, using our growing knowledge of the functions of individual genes, we can use microarrays as microscopes to see a comprehensive, dynamic molecular picture of the living cell. Several methods have been developed over the last several years to quantitate the mRNA expression of thousands of genes in parallel. One method, termed serial analysis of gene expression (SAGE), relies on high-throughput sequencing of 14-bp, gene-specific cDNA tags to enumerate the expression of individual genes in a cell (12). Because of its reliance on DNA sequencing, SAGE can identify novel transcripts that have not been observed in other high-throughput



835

sequencing projects. On the other hand, it is difficult to analyze large numbers of samples, or to measure changes in the abundance of rare transcripts, using SAGE, and thus this method is most suited to binary questions in which the transcriptional response to a particular cellular stimulus or to a single transcription factor is assessed. Within the immune system, SAGE has recently been used to analyze gene expression in mast cells before and after stimulation through the high-affinity IgE receptor (13). An interesting and unanticipated finding was the expression in resting mast cells of macrophage inhibition factor (MIF), a cytokine that was previously known to be constituitively expressed only in macrophages and anterior pituitary cells. MIF is an important mediator of delayed-type hypersensitivity (DTH) reactions, and this observation suggests an important role for mast cells in some forms of DTH. Despite extensive prior study of cytokine production by mast cells, the expression of MIF had not been reported, pointing to the value of unbiased, genome-wide gene expression surveys. In the other common methods of genomic expression analysis, DNA fragments derived from individual genes are placed in an ordered array on a solid support. These arrays are then hybridized with radioactive or fluorescent cDNA probes prepared from total cellular mRNA by reverse transcription. Following washing, the hybridization of the cDNA probes to each array element is quantitated using either a phosphorimager for radioactive probes or a scanning confocal microscope for fluorescent probes. Three styles of arrays are used most commonly. Nitrocellulose filter arrays are prepared by robotic spotting of purified DNA fragments or lysates of bacteria containing cDNA clones, and the filter arrays are hybridized with radioactive cDNA probes (14–17). Oligonucleotide arrays can be produced by in situ oligonucleotide synthesis in conjunction with photolithographic masking techniques and are hybridized with fluorescent cDNA probes (18–22).These two array formats are typically hybridized with a single cDNA probe at a time. In order to compare the mRNA expression profiles of two samples, therefore, two probes are generated and hybridized to separate arrays. The relative hybridization of the two probes to each array element is determined indirectly by mathematical normalization of the two data sets. A third type of microarray is fabricated by robotic spotting of PCR fragments from cDNA clones onto glass microscope slides (23–29). These cDNA microarrays are simultaneously hybridized with two fluorescent cDNA probes, each labeled with a different fluorescent dye (typically Cy3 or Cy5). In this format, therefore, the relative mRNA expression in two samples is directly compared for each gene on the microarray (Figure 1A, see color insert). For a given gene, the fluorescence ratio corresponds well with more conventional measures of relative gene expression including Northern blot hybridization and quantitative RT-PCR (23, 29, 30). Scanning and interpreting large bodies of relative gene expression data is a formidable task, which is greatly facilitated by algorithms designed to organize the results in ways that highlight systematic features and by visualization tools that represent the differential expression of each gene as varying intensities and hues of color (Figure 1B, see color insert) (31).

836

STAUDT n BROWN


Mathematical Analysis of Gene Expression Data The ability to produce large systematic sets of measurements of gene expression on a genomic scale using DNA microarrays is becoming commonplace. A single group, in a year, can print several thousand microarrays with a single microarraying robot and can produce tens of millions of individual measurements of gene expression. The mathematical analysis of the resulting data is a rapidly evolving science that is nevertheless based on a rich mathematics of pattern recognition developed in other contexts (32). Typical goals of these analyses are to identify groups of genes that are coregulated within a biological system, to recognize and interpret similarities between biological samples on the basis of similarities in gene expression patterns, and to recognize features of gene expression patterns that can be related to distinct biological processes or phenotypes. In other words, the biologist wishes to identify systematic features in the data that can be understood as a molecular picture of a biological system. The expression pattern for each gene on an array across n experimental samples can be represented by a point in n-dimensional space, with each coordinate specified by an expression measurement in one of the n samples. In order to determine the proximity of points in this gene expression space (a measure of the similarity in the expression patterns of the corresponding genes), one must first define a metric that quantitates the distance between any two of these points. In the clustering algorithms that have been implemented thus far, the most commonly used metric is essentially the standard correlation coefficient of the two data vectors (31). Although there are other possible ways of measuring distance in gene expression space, this metric is well suited to gene expression data because it corresponds well to the intuitive idea of coordinate regulation of two genes (31). The second step in the mathematical treatment of array data is to apply one of many clustering algorithms that use the distance metrics to find clusters of genes in this n-dimensional space, corresponding to genes with similar patterns of variation in expression over a series of experiments. The clustering methods that have been applied to array data thus far are hierarchical clustering (31), self-organizing maps (SOMs) (33), k-means (34), and deterministic annealing (35). Each of these algorithms easily captures the main biological features within data sets. For example, hierarchical clustering, SOMs, and k-means algorithms have all been applied to cell cycle data in yeast and have each revealed several broad classes of cell cycle–regulated genes (33, 34, 36). Nonetheless, the differences in the various algorithms produce views of the data that differ in detail with respect to the assignment of genes or samples to particular clusters. There is no ideal approach to the problem that these clustering methods address, namely the projection of a very high dimensional body of data to a lower-dimensional space (often just a one-dimensional ordered list). A reasonable approach, therefore, is to use a variety of different algorithms, each emphasizing distinct orderly features of the data, in order to glean the maximal biological insight.



837

Figure 2 (see color insert) presents a simple example of hierarchical clustering applied to data from T cell and fibroblast activation experiments (30, 37). Hierarchical clustering begins by determining the gene expression correlation coefficients for each pair of the n genes studied. The two genes with the most correlated expression across all of the samples are fused into a node that is subsequently represented by the average expression of the two genes. This clustering process is then repeated on the n11 genes/nodes that remain. After n11 iterations, all genes are incorporated into a dendrogram that connects each of the nodes generated during the clustering (Figure 2, see color insert). The length of each fork in the dendrogram is inversely proportional to the similarity of the two nodes or genes that it connects. The data in Figure 2 are taken from one experiment with human peripheral blood T cells activated by phytohemagglutinin (PHA) and phorbol-myristoyl-acetate (PMA) and another experiment with human serumstarved fibroblasts activated by readdition of serum. In both experiments, the cells were initially in the G0 stage of the cell cycle and synchronously entered G1 and S phase following stimulation. Each experiment used microarrays containing the same set of 9000 human cDNAs to monitor changes in gene expression over time, comparing mRNA from each stimulated culture with mRNA from resting cells. Figure 2 (see color insert) shows data from a subset of the induced and repressed genes, presented at the left in an unclustered form and, at the right, arranged by hierarchical clustering to reveal coordinately expressed genes. In this example, the clustering algorithm identified three broad clusters that contain genes activated (a) in T cells only, (b) in both T cells and fibroblasts, or (c) in fibroblasts only. The genes upregulated in both T cells and fibroblasts include c-myc, a gene known to be important for progression from G0 to S phase, and genes involved in energy metabolism, presumably reflecting the increased energy requirements of activated cells. Within the T cell–specific cluster are the chemokines MIP-1-alpha and MIP-1-beta, which are known to be coordinately regulated during T cell activation and are important for recruitment of monocytes to regions of immune activation. Interestingly, the aryl-hydrocarbon receptor, the molecular target of dioxin, is specifically induced during T cell activation, perhaps accounting for the ability of dioxin to induce apoptosis in activated, but not resting, mouse T cells (38). The SH2- and SH3-containing protein SLAP (srclike adapter protein) was preferentially induced in T cells. This is noteworthy because SLAP has recently been shown to inhibit cell cycle progression in fibroblasts (39). These microarray data may thus have revealed an unsuspected differential function of SLAP in T cell and fibroblast mitogenesis. Notable among the genes induced preferentially in fibroblasts are basic fibroblast growth factor (basic FGF) and vascular endothelial growth factor (VEGF), both of which are involved in a wound healing response (see below) (30). In addition to these three broad gene expression clusters, there is biologically important fine structure. For example, c-fos, jun B, and MAP kinase phosphatase were all downregulated in late T cell activation, whereas they were induced during the serum response of fibroblasts. The above example highlights several general principles that can

838

STAUDT n BROWN

emerge from clustering of gene expression data. As described in the following section, studies of global gene expression patterns in yeast have shown that genes with related biological roles are often tightly coregulated (28, 31, 36, 40, 41). A corollary is that novel genes of unknown function that are clustered with a large group of functionally related genes are likely to participate in the same biological process. In this light, it is interesting to note that several novel genes were selectively induced in T cells rather than fibroblasts (Figure 2, see color insert). Cluster analysis provides a systematic way to focus attention on subsets of the novel genes represented in a survey of gene expression patterns that warrant further investigation in relation to specific biological processes. Finally, Figure 2 demonstrates the usefulness of systematic databases of gene expression measurements that allow fresh biological insights to be made by juxtaposing and comparing data sets from disparate biological systems.


Genomic-Scale Gene Expression Analysis in Model Systems Whole Genome Gene Expression Analysis in Yeast The most extensive and systematic studies of global gene expression patterns to date have been carried out in Saccharomyces cerevisiae. The yeast genome was the first genome of a free-living organism to be completely sequenced, and it has thus been the first model used for development of many functional genomic approaches that can now be applied to mammalian genomes. Over the past three years, several groups have reported studies of genome-wide patterns of gene expression in response to physiological stimuli, drugs, developmental programs or specific mutations in yeast (28, 36, 40, 42–45). Each such study has provided a wealth of new information and insight into a specific process: the switch from glycolysis to respiration, progression through the cell division cycle, the program of gametogenesis and spore formation, and the targets of specific and global transcriptional regulators. Trivially, these studies provide comprehensive catalogs of the genes whose expression varies in each specific process or in response to each specific perturbation, and the studies define the temporal pattern of each gene’s response. But the systematic nature of these observations, involving comprehensive, quantitative measurements of variation in each transcript from the yeast genome, makes it possible to view the entire set of data as one large and expanding survey of the expression program of the yeast genome. A new and remarkably useful kind of gene expression map emerges from this approach. In contrast to conventional genetic maps based on the physical order of genes in the genome, gene expression maps derive their order from the logic underlying the expression program of a genome. Gene expression maps are constructed by first organizing the gene expression data using any of the various clustering algorithms outlined above. The ordered tables of data are then displayed graphically in a way that allows biologists to assimilate the choreography of gene expression on a broad scale as well as the fine distinctions in expression of individual genes. The large panel on the left of



839

Figure 3 (see color insert) shows one example of such a map: Each row represents the expression pattern of one of the 6220 known or predicted genes of yeast, and each column represents the results of one of 204 genome-wide microarray gene expression experiments. The expression measurements in this figure were derived from yeast that were placed under 28 distinct physiological or nutritional conditions and assayed multiply over time. For this map, the hierarchical clustering strategy was used to group genes on the basis of similarity in gene expression patterns (31). It is worth noting that this set of over one million measurements of gene expression represents considerably less than 10% of the genome-wide expression data that has been collected over the past two years for this one organism. A simple evolutionary logic emerges from an analysis of yeast gene expression maps: Genes with similar expression patterns under a particular set of conditions encode protein products that play related roles in the physiological adaptation to those conditions. The extent and precision with which this simple organizing principle determines the geography represented in this map of the genome is unexpected and remarkable (31, 41). Genes encoding products that invariably function together in a stoichiometric complex are virtually always among the most highly coregulated groups in the genome. For example, the vertical bar at the upper right of this map (Figure 3) marks the position of a cluster comprising about 2% of the genes in the yeast genome, including, almost exclusively, all the genes that encode ribosomal proteins. Similar coregulated clusters identify the histones, the subunits of the proteosome, and subunits of numerous other multimeric enzymes (31, 41). Genes whose products work together in a metabolic pathway or a discrete physiological or developmental program are typically less tightly coregulated than components of stoichiometric complexes, but they are sufficiently similar in their expression patterns to cluster together in this genomic expression map. Expanded views of two such clusters are shown at the right of Figure 3 (see color insert): One cluster is composed of genes encoding components of the mitochondrial electron transport and ATP synthase complexes (labeled ‘‘respiration’’), and the other is composed of genes that play key roles in chromosome synapsis and meiotic recombination (labeled ‘‘meiosis’’). Each cluster includes genes without a presently identified function. The consistent relationship between a gene’s expression pattern and its function, reflected in this map, provides the basis for imputing functions to these previously uncharacterized genes. Indeed, the essential role in sporulation for one of the previously uncharacterized genes in the sporulation cluster, YPR007C, which is predicted to encode a putative chromosome cohesion protein, was established following its identification by this cluster analysis method (40). Conversely, each of the conditions represented in this gene expression map (the vertical columns) is characterized by a unique and recognizable signature in its gene expression pattern. Each cell transduces variation in its environment, internal state, and developmental state into readily measured and recognizable variation in gene expression patterns.

840

STAUDT n BROWN


Thus the global pattern of gene expression provides a distinctive and accessible molecular picture of the state and identity of biological samples. The prospects for mapping the regulatory networks that control gene expression programs and connecting them to the corresponding environmental stimuli and the physiological processes that they mediate are already apparent from studies in yeast. These studies have revealed unsuspected complexity in the relationships among regulatory proteins and the genes they control and, at the same time, have provided compelling evidence for the experimental tractability of this problem to systematic dissection (28, 36, 40, 43, 45).

Unconventional Pictures of Biological Responses from Genome-Scale Gene Expression Profiles One of the most useful qualities of the systematic characterization of gene expression programs is that the results are much less constrained by preconceived models than traditional, ‘‘hypothesis-limited,’’ experimental approaches. A vivid example of this feature was provided by a genome-scale survey of gene expression changes during the response of serum-deprived cultured human fibroblasts to serum (30). A cDNA microarray representing approximately 9000 different human genes was used to measure gene expression changes at 14 time points following the readdition of serum, beginning 15 min after stimulation and continuing for 24 h. The experiment was intended to provide new insights into the transition from the G0 cell cycle state to a proliferating state since, historically, the serum response of fibroblasts had been viewed as a simple model for this transition. However, the proliferation-related changes in gene expression accounted for only a small fraction of the program of gene expression that was observed in this experiment. The gene expression program of serum-stimulated fibroblasts was far richer than anticipated and pointed to an important physiological role of fibroblasts in the wound healing response. Serum, the soluble fraction of clotted blood, is normally encountered by cells in vivo in the context of a wound. Indeed, the expression program that was observed in response to serum suggested that fibroblasts are programmed to interpret the abrupt exposure to serum not as a general mitogenic stimulus but as a specific physiological signal signifying a wound. Numerous genes with known roles in processes relevant to wound healing were induced by the serum stimulus. These included genes involved in the direct role of fibroblasts in remodeling the clot and the extracellular matrix as well as genes encoding intercellular signaling proteins that promote inflammation, angiogenesis, and reepithelialization. Although this study focused exclusively on the fibroblast and was not intended or expected to address any aspect of immunity, the observed expression program pointed to an important role for fibroblasts in orchestrating the immune response to a wound. The serum-induced genes encoded proteins that promote chemotaxis and activation of neutrophils, monocytes and macrophages, T lymphocytes and B lymphocytes, thus providing innate and antigen-specific defenses against



841

wound infection. In addition, the recruitment of phagocytic cells is required to clear out the debris during wound remodeling. The results, unexpectedly, remind us of the importance of viewing an immune response as a concerted physiological program, involving not only cells normally regarded as components of the immune system per se but also virtually any cell that finds itself in a setting where an immune response is called for. The picture painted by the transcriptional response to serum suggests that the fibroblast is an active participant in a conversation among the diverse cells that work together in wound repair, interpreting, amplifying, modifying, and broadcasting signals that control inflammation, angiogenesis, and epithelial regrowth during the response to an injury. Another implication of this experiment is that fibroblasts, and very likely many other cells, are programmed to recognize exposure to serum as a signal representing a serious injury. Inclusion of serum in mammalian cell culture medium has become a common, almost ubiquitous, practice. Yet, this experiment suggests that trying to study the normal behavior of cells in the presence of serum may be analogous to trying to study normal human behavior in a burning building. Signal Transduction One of the natural arenas for genomic-scale gene expression analysis in mammalian systems is signal transduction. It is clear from studies of protein-protein interactions and inducible phosphorylation events that proximal signaling pathways are considerably interwoven. However, not yet known is the extent to which the downstream transcriptional targets of different signaling pathways are overlapping or distinct. For one class of target genes, the immediate early genes, the answer appears to be that disparate signaling pathways converge on virtually identical immediate early target genes (46). Oligonucleotide microarrays were used to compare the immediate early gene response (i.e. genes induced within 4 h of stimulation) of fibroblasts to platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and epidermal growth factor (EGF), all of which signal through distinct tyrosine kinase receptors. Out of 5938 genes on the array, 66 genes displayed an immediate early response to PDGF. Almost all of these genes were also induced by FGF to the same degree as by PDGF. Correspondingly, these two growth factors cause a quantitatively similar mitogenic response in fibroblasts. Although EGF induced many, but not all, of the same immediate early genes, the magnitude of the induction was quantitatively lower than observed with PDGF and FGF. In this experimental system, therefore, the immediate early genes behave as a transcriptional ‘‘module’’ that is invoked to a greater or lesser degree by different cellular stimuli. A second important conclusion from this study was that none of the tyrosines in the cytoplasmic tail of the PDGF receptor was absolutely required for any discrete feature of the immediate early response. This was a surprise because previous work had shown that each tyrosine serves as the docking site for a different signal transduction protein. The results therefore suggest that signal transduction networks must be extensively ramified proximal to the membrane tyrosine kinase receptors, converging on a common set of nuclear immediate early responses.


842

STAUDT n BROWN

This is, of course, only one snapshot of the genomic response to membrane signaling events. Since the serum response of fibroblasts induced a stereotypical set of genes beyond the immediate early time frame, it is quite plausible that different receptor kinases will cause distinct delayed transcriptional responses (30). The signaling events through other cell surface receptors certainly will lead to receptor-specific transcriptional responses in some cases. Microarray analysis of cytokine responses, for example, reveals both cytokine-specific and generic transcriptional responses (37). This is not surprising given the direct docking of distinct STAT family transcription factors to the various cytokine receptors (47). Thus, when membrane signaling events lead more directly to the activation and/ or nuclear translocation of transcription factors without invoking extensively interconnected proximal signaling networks, signature transcriptional responses may be elicited. Finally, the cell type chosen for signaling experiments will inevitably influence the genomic transcriptional response. For example, a microarray analysis of PMA-responsive genes in myeloid and lymphoid cell lines revealed sets of induced genes that were cell line–specific as well as genes that were PMAresponsive in all myeloid cell lines but not in Jurkat T cells (33). The developmental history of a cell, preserved within heritable chromatin structure or by DNA methylation, will shape the outcome of signaling, as will the different repertoires of transcription factors that are available to various cell types. The direct target genes of transcription factors can be revealed by genomicscale gene expression analysis, as illustrated by studies of p53 and BRCA1 (48– 51). Inducible overexpression of transcription factors is the experimental design that is currently adopted in most cases. Although valuable, this approach is somewhat risky in that artificial overexpression can lead to nonphysiological titration of protein-protein interactions and binding of transcription factors to inappropriate sites within the genome. Genomic studies of loss-of-function mutants will be an important goal in this field. Analysis of cells taken from knockout animals will be helpful, particularly in cases in which the developmental program has not been overtly altered by the engineered mutation. Large-scale loss-of-function studies in somatic cells in culture await the development of robust methods of gene disruption or interference.

Genomic-Scale Gene Expression Analysis in the Immune System Ultimately, studies of gene expression in the immune system will examine the entire genomic repertoire of genes in each sample investigated. Although this complete repertoire is not yet available, many insights into the gene expression programs evoked during immune responses can be made using large DNA microarrays that deliberately include many genes known to be expressed in immune cells. An example of such a specialized subgenomic microarray is the Lymphochip, a specialized human cDNA microarray that is enriched for genes related to immune function (8). The Lymphochip microarray is composed of



843

17,853 cDNA clones derived from three sources. The majority of clones (;80%) were derived from the lymphoid cDNA libraries that were subjected to highthroughput EST sequencing (Table 1). The selection of these clones was based on bioinformatics algorithms that identified ESTs that were either unique or enriched in lymphoid cDNA libraries (8). A second set of Lymphochip clones was identified during the course of previous microarray analyses of immune responses using first-generation microarrays of ;10,000 human genes (37). Last, a curated collection of 3183 ‘‘named’’ genes that are of known or suspected importance to immune function, cell proliferation, apoptosis, or oncogenesis and 57 open reading frames from the pathogenic human viruses HIV-1, HTLV-I, EBV, and HHV-6, 7, and 8 were incorporated into the Lymphochip. One of the virtues of mechanically printed microarrays like the Lymphochip, in this era of continuing gene discovery, is that they can be readily upgraded: New genes that are discovered during further high-throughput sequencing or as a result of directed molecular biology experiments can be added to new editions of the Lymphochip in days. The Genomic Expression Program in Lymphocyte Differentiation Systematic exploration of gene expression programs during human lymphocyte development and activation is under way. Early work has focused on late-stage B cell differentiation, following mature, naive B cells from the resting state through the germinal center reaction and into terminal differentiation. The germinal center is an inducible microenvironment formed during an immune response by the concerted action of antigen-specific B and T cells together with follicular dendritic antigenpresenting cells (FDCs) (52, 53). The germinal center reaction is initiated when the surface immunoglobulin receptor on a B cell encounters its cognate antigen, and activated T cells signal the B cell through CD40. FDCs secrete a gradient of the chemokine BLC, which signals the activated B cell through the BLR1/CXCR5 receptor to migrate toward the FDC (54). Activated T cells also migrate to the nascent germinal center where they continue to interact with germinal center B cells. The germinal center becomes polarized, with highly proliferative centroblast B cells in the ‘‘dark’’ zone and less proliferative centrocytes in the ‘‘light’’ zone. The process of somatic hypermutation of immunoglobulin genes is initiated in centroblasts, which then migrate to the light zone to become centrocytes. If the hypermutation process has improved, or at least preserved, the ability of the B cell to bind antigen on the surface of the FDC, the B cell is rescued from programmed cell death. The B cell may then migrate back to the dark zone and continue somatic hypermutation or may terminally differentiate into a memory B cell or plasma cell. B cells at each of these stages of differentiation were purified from human tonsils or peripheral blood, and their transcript patterns were characterized using the Lymphochip microarray (8). As important controls, B cells were activated polyclonally in vitro by ligation of the antigen receptor and activation with CD40 ligand, with and without IL-4. Additionally, T cells were mitogenically activated


844

STAUDT n BROWN

with phorbol ester and ionomycin. The gene expression profiles shown in Figure 4 (see color insert) reveal that germinal center B cells represent a distinct stage of B cell differentiation that activates a broad gene expression program that is not observed in mitogenically activated peripheral blood B cells. Germinal center B cells not only express scores of genes that are missing in activated peripheral blood B cells but also lack expression of many genes that are induced during in vitro B cell activation. Thus, coligation of undefined B cell surface receptors, together with stimulation through the antigen receptor and CD40, may be needed to generate the germinal center gene expression profile. Indeed, no convincing in vitro culture system has yet been developed that is able to induce resting peripheral B cells to adopt a full germinal center phenotype. The large set of germinal center B cell–specific genes discovered by microarray analysis can therefore serve as a yardstick to measure the success of in vitro cultures in mimicking the germinal center state. Mitogenically activated B and T cells shared a common set of activation genes (Figure 4, see color insert), which may reflect the convergence of multiple signaling pathways on common nuclear targets (46) and the fact that the cell cycle gene expression program was activated in both cell types. However, mitogenically activated T cells expressed a distinct set of genes not observed in resting T cells or in activated B cells (not shown). This set of genes includes, of course, various cytokines such as IL-2 and TNF alpha but also a number of novel genes. Based on the coordinate expression of these novel genes with cytokines and the lineage specificity of their expression, they are attractive candidates for functional analysis in the future. The Relationship of Lymphoid Malignancies to Normal Lymphocyte Differentiation Genomic-scale gene expression profiling is certain to illuminate many aspects of cancer pathogenesis, cancer diagnosis, and the mechanisms underlying treatment resistance and susceptibility. Traditionally, studies of mutations, amplifications, and deletions in the genomic DNA of cancer cells have revealed many of the key genetic events that occur during the progression to cancer. Many of these genetic alterations may have acted for many years prior to diagnosis to bypass key checkpoints and allow cell cycle progression. On the other hand, gene expression profiling of cancer cells reflects the molecular phenotype of the cancer cell at diagnosis. As a consequence, the detailed picture provided by the genomic expression pattern may provide the basis for a new systematic classification of cancers and more accurate predictions of the responses of a cancer to treatment. A major determinant of the biological potential of a cancer cell is likely to be the normal cell from which it was derived. About 90% of human lymphoid malignancies are derived from B cells, and each of these malignancies has been provisionally assigned to a particular stage of B cell differentiation based on analysis of immunoglobulin gene rearrangement and mutation together with cell surface phenotyping. However, the extent to which the gene expression program of nor-



845

mal B cells is retained in the cancer cell is best addressed by genomic-scale gene expression analysis. A particular breeding ground for human lymphomas is thought to be the germinal center reaction. This notion is based on analysis of rearranged immunoglobulin genes in these malignancies, which often show extensive somatic hypermutation (55). Indeed, in two categories of non-Hodgkin’s lymphoma, follicular lymphoma and MALT lymphoma, the immunoglobulin sequences from a single biopsy specimen show evidence of ongoing mutation (56–59). In other malignancies in which the immunoglobulin sequences are mutated but invariant, the cell of origin could as well be a postgerminal center B cell. Even the presence of immunoglobulin mutations in a B cell malignancy is not conclusive evidence that the cell of origin passed through the germinal center microenvironment, since in some mutant mouse models, somatic hypermutation of immunoglobulin genes can occur in the absence of detectable germinal centers (60). The most common form of non-Hodgkin’s lymphoma is diffuse large cell lymphoma (DLCL), comprising ;40% of all cases. The immunoglobulin genes in DLCL are invariably mutated. Furthermore, a recurrent translocation in this malignancy involves the BCL-6 gene, a gene also required for normal germinal center development (61–63). However, this translocation occurs in only ;32% of DLCLs, thus revealing potential heterogeneity in this diagnostic category. Patterns of gene expression in a large number of DLCLs were therefore analyzed, using the Lymphochip microarray, to determine the relationship of this malignancy to normal germinal center cells and to investigate the possibility that this diagnostic category may harbor more than one disease entity. Figure 5 (see color insert) shows the expression of a subset of 60 genes from the Lymphochip in 25 different lymph node biopsies of DLCL and in a variety of normal B cell preparations. It is evident that the gene expression patterns in DLCLs are strikingly heterogeneous and that a subset of DLCLs shows a pattern with a strong resemblance to the pattern seen in normal germinal center B cells. Distinct patterns of gene expression identify at least two different subtypes in what has previously been considered a single disease. The similarities in gene expression patterns strongly imply that the cell of origin of one DLCL subtype is the germinal center B cell, but the origin of the other cases is enigmatic. These cases could be derived from a postgerminal center B cell that had extinguished the germinal center gene expression program. Alternatively, the oncogenic transforming event(s) may have disrupted signaling pathways that are critical to maintain the germinal center phenotype. Preliminary surveys of other B cell malignancies demonstrate that each diagnostic category has its own gene expression signature. Gene expression patterns observed in follicular lymphomas share significant features with the patterns seen in germinal center B cells, whereas the expression patterns in chronic lymphocytic leukemia cells do not resemble those in germinal center cells but instead are reminiscent of resting peripheral blood lymphocytes. Within each of these diagnostic categories, however, the molecular heterogeneity reflected in the gene

846

STAUDT n BROWN


expression profiles suggests the existence of disease subtypes, as were revealed in DLCL. The stratification of patients according to gene expression signatures could ultimately contribute to clinical decisions directing the patient to the most appropriate therapy.

Gene Expression Changes During Immune Responses Oligonucleotide arrays have been used to discover gene expression correlates of antigen-induced anergy and activation in B lymphocytes (R Glynne, C Goodnow, personal communication). Transgenic animals expressing heavy and light chains for anti-HEL (hen egg lysozyme) antibody provide B cells of near monoclonality that can be either anergized or activated depending on the method and form of antigen administration (64). Anergic/tolerant B cells are profoundly resistant to subsequent exposure to antigen under activation conditions. B cell anergy involves activation of some but not all of the signaling pathways that are engaged during lymphocyte activation: NF-AT and erk MAP kinase pathways are activated in tolerant cells, whereas NF-jB and jnk pathways are not (65). Microarray analysis of gene expression in antigen-stimulated naı¨ve B cells demonstrated that 59 genes were significantly induced or repressed after 1 h of stimulation, whereas more than 300 genes were altered in expression after 6 h (R Glynne, C Goodnow, personal communication). By contrast, only 8 of these genes were regulated in tolerant B cells. Instead, tolerant B cells displayed a distinct gene expression signature consisting of 20 upregulated genes and 8 downregulated genes that were not altered during activation of naı¨ve B cells. Interestingly, pharmacological inhibition of NF-AT by the immunosuppresive drug FK506 was less efficient than tolerance in blocking B cell activation responses: One third of the antigen-induced gene expression changes in naı¨ve B cells were unaffected by FK506. These findings could have important implications for the discovery of novel immunosuppresive drugs. An ideal immunosuppressive drug would have all of the functional effects of natural tolerance without eliciting the side effects that limit the utility of FK506 and cyclosporin in some patients. The gene expression signature of tolerant B cells could by used as a surrogate marker in drug screens for compounds that might mimic the anergic state (R Glynne, C Goodnow, personal communication). Furthermore, if a novel compound in a drug screen induces gene expression changes not found in tolerant cells, this might signal an unwanted ‘‘off-target’’ effect of that compound (42). T cell responses to antigenic and mitogenic stimulation have also been analyzed by cDNA and oligonucleotide microarray analysis (37) (P Marrack, personal communication). Since T cell activation is a well-trodden path, many of the induced genes are well known, including some depicted in Figure 2 (see color insert). Interestingly, an equal number of genes were repressed as were induced during T cell activation, leaving the total diversity of mRNAs roughly equivalent between resting and activated cells. Immunologists have evidently spent less


847

effort investigating the genes that are downregulated during lymphocyte activation because this class contained more novel genes than the upregulated class.


FUNCTIONAL GENOMICS AND THE GENETICS OF COMPLEX IMMUNOLOGICAL DISEASES Positional cloning of susceptibility genes for diseases with simple Mendelian inheritance is now routine. However, the majority of medically important genetic diseases show familial clustering with an indeterminant inheritance pattern. The heritable risk for these diseases may be determined by many genes, each of which can affect relative risk for the disease phenotype. In human autoimmune diseases, a sibling of an affected individual has a relative risk of developing the same autoimmune disease of 6–100-fold, compared with the prevalence of the disease within the general population (66). The genetics of autoimmune mouse models has been particularly illustrative of the complexity of some immune-mediated diseases. For example, autoimmune diabetes in the NOD mouse may be controlled by 15 genes on 11 chromosomes (66). The genetic complexity of these diseases is most likely a reflection of their complex pathophysiology. In most autoimmune diseases, the major histocompatibility complex (MHC) plays a dominant role, presumably by dictating which autoantigens can be presented to the immune system. Nevertheless, MHC alleles confer a relative disease risk of only 1.3–8.3fold (66). Other genetic loci may control the breaking of immunological tolerance, the repertoire of autoimmune T and B cells, the expansion of pathogenic CD4, CD8, and/or B lymphocyte subsets, and the skewing of immune responses by cytokines. One approach to such complex diseases is to artificially simplify the genetics. In the NOD mouse diabetes model, transgenic expression of a single pathogenic T cell receptor has been used to short-circuit some of the disease pathogenesis and to reduce the number of disease susceptibility loci to five genomic intervals (67). Further breeding of this simplified mouse model to knockout animals has revealed a role for interferon gamma in the development of diabetes (68). Recent reviews have focused on the ways in which genome-wide application of polymorphic markers can identify which genomic intervals may harbor the disease susceptibility genes (66), and therefore we do not review this genomic arena extensively here. A large number of genome-wide screens in immunemediated diseases of humans and animals have been conducted (67, 69–94). In most cases, however, the genomic interval containing the susceptibility gene has not been narrowed to , 1 centiMorgan (;1 2 106 base pairs) by these methods. Interestingly, the susceptibility loci in these various diseases often coincide, suggesting that some common genes may influence many autoimmune and inflammatory diseases (66, 95).


848

STAUDT n BROWN

The molecular definition of specific susceptibility alleles in complex immunemediated diseases will clearly require new strategies that complement genetic linkage analysis. One functional genomics strategy that holds promise is the ‘‘positional candidate gene’’ approach (reviewed in 96). Having narrowed the susceptibility interval by linkage analysis, the known genes that map within the interval can be identified. Since mutation detection is still technically cumbersome and a 1-centiMorgan susceptibility region could contain 30 or more genes, it may be helpful to first focus attention on candidate genes with functions that can be plausibly connected to the disease phenotype. One potential identifying characteristic of a candidate susceptibility gene would be expression in the cells or tissues presumed to be at fault in the disease. Soon, public databases of gene expression measurements will make this analysis routine. Even when genetic linkage has not been performed, the candidate gene approach may quickly reveal potential disease susceptibility loci. For example, the candidate gene approach was used to examine the genetic differences between chronic granulomatous disease (CGD) patients who differed in susceptibility to immune-mediated complications (97). CGD results from a primary defect in genes for NADPH oxidase that control superoxide production in phagocytes. CGD patients differ dramatically in the frequency with which they develop a variety of chronic complications, including granulomatous diseases of the gastrointestinal and urinary systems as well as autoimmune and rheumatological disorders. A priori, such differences may be due to differences in the mutations present in the 4-NAPDH oxidase subunit genes found in different patients. Alternatively, polymorphisms in other disease-modifying genes could contribute to the risk of immune complications. Foster et al examined polymorphisms in seven candidate genes encoding myeloperoxidase, mannose binding lectin, TNF alpha, IL-1 receptor antagonist, and the Fc gamma receptors IIa, IIIa, and IIIb (97). Alleles of myeloperoxidase and Fc gamma receptor IIIb were significantly associated with an enhanced risk for gastrointestinal complications. Alleles of myeloperoxidase were associated with increased risk of autoimmune and rheumatological disorders. Combinations of specific alleles of different genes conferred an even greater relative risk for chronic immune-mediated complications. In this relatively rare genetic disease it would be difficult, if not impossible, to enroll enough patients to conduct a standard linkage analysis of immune complications; thus the candidate gene approach provides a tractable alternative. cDNA microarray analysis of gene expression promises to aid significantly in the search for disease susceptibility loci. One example of this approach comes from analysis of the spontaneously hypertensive SHR rat, which is a model for human diabetes, hyperlipidemia, obesity, and hypertension (98). Genetic linkage analysis of this disorder had focused attention on an interval from rat chromosome 4, but the causative gene had not been identified. cDNA microarrays were used to compare gene expression in adipose tissue from the SHR strain and control, nonhypertensive rat strains. SHR cDNA probes hybridized poorly to a microarray spot representing CD36, a gene that maps to regions of mouse and human chro-



849

mosomes that are syntenic to rat chromosome 4. This microarray finding prompted a further analysis of the CD36 gene, which revealed multiple coding region mutations in the CD36 gene of SHR rats. CD36 is a fatty acid receptor and transporter whose overexpression in transgenic mice or deletion by gene disruption in mice leads to alterations in blood lipid levels (98, 99); thus it is a strong candidate gene for the hyperlipidemic quantitative trait in SHR rats. The apparently diminished expression of CD36 in SHR rats, detected by cDNA microarray hybridization, was traced to a genomic deletion in this strain within the CD36 38 untranslated region, the only region of the CD36 gene represented on the array. Although this may appear to be an exceptional case, in that the genetic lesion in this example directly affected the ability to measure the expression of the gene, it is likely that many disease-causing mutations will be found to affect transcript levels. Nonsense mutations and mutations that disrupt normal splicing can lead to reduced mRNA levels via nonsense-mediated decay mechanisms. Many mutations, including many classical genetic disease-causing mutations (e.g. many thalassemias), directly alter transcription of the affected gene. Indeed, the evolutionary constraints on mutation in perigenic noncoding regions, reflected in limited sequence polymorphism observed in these regions as compared to degenerate positions in coding sequences, argue that the potential for deleterious consequences from mutations in regulatory sequences of genes rivals that of mutations in protein-coding sequences (4, 100). Perhaps a more common use of cDNA microarrays in the investigation of genetic diseases will be to detect quantitative differences in the expression of genes between different animal strains or different human individuals. Quantitative traits that distinguish individuals of the same species undoubtedly arise as a result of both coding region polymorphisms that alter the function of a gene product and regulatory region polymorphisms that affect the expression level of the mRNA or protein. The relative contribution of these two types of allelic differences to genetic diversity is unclear at present, but cDNA microarray analysis may soon reveal a broad range of quantitative gene expression traits within the immune system. Polymorphisms in the mouse TNF alpha gene have been described that affect TNF alpha levels and can modulate the development of nephritis in animals predisposed to systemic lupus erythematosis (101). Similarly, regulatory mutations in the human TNF alpha gene have been associated with a wide variety of diseases, but the interpretations of such studies in humans is complicated by the location of the TNF alpha gene in the MHC and the difficulty in teasing apart the contributions of allelic differences in the TNF alpha gene and the MHC genes. An interesting quantitative gene expression trait was recently described involving FRIP, a gene that encodes an adapter protein involved in IL-4 signaling (102). The FRIP gene was mapped to a region of mouse chromosome 14 very close to the gene for the hairless mutation. The hairless mutation results from the insertion of an endogenous mouse retrovirus into the mouse hairless locus (103). The hair-


850

STAUDT n BROWN

less gene is also mutated in human alopecia universalis and encodes a pioneer protein of unknown function (104). The hairless mouse also has immune abnormalities, including lymphadenopathy and augmented response of anti-CD3 stimulated T cells to IL-4 (102). The FRIP gene is expressed at significantly lower levels in hairless mice compared with wild-type mice, possibly as a result of the same retroviral insertion event. Given the proximity of FRIP to the hairless gene and the IL-4-related abnormalities of hairless mice, the FRIP quantitative gene expression trait may well account for the lymphoproliferative disorder in these mice. Genomic-scale gene expression analysis may help to unravel complex genetic diseases by defining more precisely the disease ‘‘phenotype.’’ As a hypothetical example, suppose one of the disease susceptibility genes involved in a complex immunological disease regulates responsiveness of T lymphocytes to IL-2. cDNA microarray analysis of IL-2-stimulated peripheral blood T cells might therefore reveal a gene expression profile that correlates with the presence of this susceptibility allele. This gene expression correlate of the susceptibility gene might be observed in family members of affected individuals who are clinically ‘‘normal’’ due to segregation of other susceptibility alleles. Linkage analysis using polymorphic markers could be applied to this gene expression phenotype rather than to the whole clinical syndrome, thereby isolating one component of the complex disease phenotype.

GENE EXPRESSION PROFILES OF PERIPHERAL BLOOD AS SENTINELS OF DISEASE Immune cells circulate throughout the body responding to internal and external threats to homeostasis. Circulating white blood cells are charged with the task of seeking out, recognizing, and mounting a suitable response to the earliest signs of an infection or injury. The sensitive and diverse repertoire of receptors and signal transduction systems that cells use to monitor and respond to trouble at any site in the body may well give rise to signature patterns of altered gene expression in peripheral blood cells reflecting the nature and site of an infection or injury. It is plausible that gene expression patterns in specific subsets of peripheral blood cells might be altered in characteristic ways in response to the presence of specific occult infectious agents. As a consequence, peripheral blood mononuclear cells might display a pathognomonic gene expression signature that could be used to diagnose occult disease. Gene expression changes induced by cytomegalovirus (CMV) infection of fibroblasts and HTLV-I infection of T lymphocytes were studied using microarrays and revealed, not surprisingly, largely distinct gene expression profiles (105, 106). Infection with CMV invoked a strong interferon response that was not observed with HTLV-I, whereas HTLV-1 induced a number of NF-jB target genes as a consequence of the nuclear translocation of NF-jB induced by the HTLV-I tax protein. Recently, the response of monocytic



851

cells to bacterial exposure in vitro was monitored using Lymphochip cDNA microarrays (D Relman, in preparation). B. pertussis, H. pylori, and S. typhimurium each induced a distinctive gene expression profile in the monocytes. Further, mutant strains of B. pertussis that lacked individual toxin genes elicited gene expression changes that differed from the response to the wild-type strain. Thus, gene expression profiles could be used not only to recognize exposure to an infectious agent but perhaps to identify the agent or category of agent, based on specific characteristics of the response. This ability would clearly be especially useful in cases in which the agent cannot be readily cultured from the host. Since gene expression responses to infectious agents take place within the first few hours after exposure, gene expression profiling might be useful in diagnosing infectious exposure in advance of clinical symptoms, allowing exposed patients to be rapidly triaged for treatment. Finally, the course of infection and the ensuing host response could potentially be monitored by changes in peripheral blood gene expression. This approach could aid in the management of sepsis, which is a disease characterized by an orderly progression of pathophysiological events (107). Gene expression profiles could thus be used to stratify patients into distinct pathophysiological groups and, ultimately, treat each group with a therapy tailored to the disease stage. It is not difficult to imagine a wider range of clinical settings in which peripheral blood gene expression profiles might aid in patient management. Exposure to toxic xenobiotic compounds such as dioxin should be readily detectable by virtue of the expression of the aryl hydrocarbon receptor in activated T cells (Figure 2, see color insert). T cells from cancer patients display an anergic phenotype, partly due to loss of the zeta chain of the T cell receptor (108, 109), which should result in a gene expression signature in peripheral blood cells. In cases of occult malignancy, such as often occurs in ovarian cancer, this gene expression signature might be detectable in advance of clinical symptoms. In autoimmune diseases such as multiple sclerosis, changes in peripheral blood gene expression may precede a clinical exacerbation, allowing clinicians to time immunosuppressive treatment optimally. Recognition of characteristic patterns of gene expression in circulating peripheral blood cells may thus prove broadly useful as an approach to noninvasive diagnostics, in effect recruiting these readily accessible cells as ‘‘spies’’ to report the presence of occult infection or injury they have encountered during their surveillance of the integrity of the body. Of course, as with any clinical test, gene expression profiling of blood mononuclear cells must be both sensitive and specific to be useful. If only a small fraction of peripheral blood cells responds to a pathological event, microarray analysis of gene expression may not be a sufficiently sensitive test. Furthermore, we do not yet know the extent of normal variation in gene expression patterns in peripheral blood cells, nor the extent to which they are altered by everyday, non–life threatening events. For example, it will be necessary (and very interesting in its own right) to catalog the effects of upper respiratory viral infections, stress hormones, age, sex, and even circadian rhythms on gene expression in peripheral blood cells. As already mentioned, genetic variation in immune regulatory genes may be associated with

852

STAUDT n BROWN

quantitative gene expression traits that will need to be considered in interpreting gene expression profiles of blood cells. Given the rich insights that genomic-scale gene expression analysis has already provided, we can be optimistic that this new mode of biological discovery will illuminate many issues in clinical pathophysiology.


ACKNOWLEDGMENTS The microarray work summarized in this paper results from a collaborative effort of the authors’ laboratories with a variety of researchers at the National Cancer Institute, NIH; Stanford University; Center for Information Technology, NIH; CBER, FDA, University of Nebraska Medical Center, and Research Genetics. Key individuals who contributed to the Lymphochip project are Ash Alizadeh, Mike Eisen, R Eric Davis, Chi Ma, Hajeer Sabet, Truc Tran, John Powell, Liming Yang, Gerry Marti, Troy Moore, Jim Hudson, John Chan, Tim Greiner, Denny Weisenburger, Jim Armitage, Izadore Lossos, Ron Levy, and David Botstein. The authors thank Richard Glynne, Chris Goodnow, Philippa Marrack, and David Relman for communicating results prior to publication. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Goffeau A, Barrell BG, Bussey H, Davis RW, Dujon B, Feldmann H, Galibert F, Hoheisel JD, Jacq C, Johnston M, Louis EJ, Mewes HW, Murakami Y, Philippsen P, Tettelin H, Oliver SG. 1996. Life with 6000 genes. Science 274:546:563–67 2. Consortium TC e S. 1998. Genome sequence of the nematode C. elegans: a platform for investigating biology. Science 282:2012–18 3. Clayton RA, White O, Fraser CM. 1998. Findings emerging from complete microbial genome sequences. Curr. Opin. Microbiol. 1:562–66 4. Cargill M, Altshuler D, Ireland J, Sklar P, Ardlie K, Patil N, Lane CR, Lim EP, Kalayanaraman N, Nemesh J, Ziaugra L, Friedland L, Rolfe A, Warrington J, Lipshutz R, Daley GQ, Lander ES. 1999. Characterization of single-nucleotide polymorphisms in coding regions of human genes. Nat. Genet. 22:231–38

5. Uberbacher EC, Xu Y, Mural RJ. 1996. Discovering and understanding genes in human DNA sequence using GRAIL. Meth. Enzymol. 266:259–81 6. Adams MD, Dubnick M, Kerlavage AR, Moreno R, Kelley JM, Utterback TR, Nagle JW, Fields C, Venter JC. 1992. Sequence identification of 2,375 human brain genes. Nature 355:632–34 7. Boguski MS, Lowe TM, Tolstoshev CM. 1993. dbEST—database for ‘‘expressed sequence tags.’’ Nat. Genet. 4:332–33 8. Alizadeh A, Eisen M, Davis RE, Ma C, Sabet H, Tran T, Powell J, Yang L, Marti G, Moore T, Hudson J, Chan WC, Greiner T, Weisenburger D, Armitage JO, Lossos I, Levy R, Botstein D, Brown PO, Staudt LM. 1999. The Lymphochip: a specialized cDNA microarray for the genomic-scale analysis of gene expression in normal and malignant lympho-


9.

10.

11.


12.

13.

14.

15.

16.

17.

18.

cytes. Cold Spring Harbor Symp. Quant. Biol. In press Strausberg RL, Dahl CA, Klausner RD. 1997. New opportunities for uncovering the molecular basis of cancer. Nat. Genet. 15(Spec. No.):415–16 Soares MB, Bonaldo MF, Jelene P, Su L, Lawton L, Efstratiadis A. 1994. Construction and characterization of a normalized cDNA library. Proc. Natl. Acad. Sci. USA 91:9228–32 Schuler GD. 1997. Pieces of the puzzle: expressed sequence tags and the catalog of human genes. J. Mol. Med. 75:694–98 Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. 1995. Serial analysis of gene expression. Science 270:484–87 Chen H, Centola M, Altschul SF, Metzger H. 1998. Characterization of gene expression in resting and activated mast cells. J. Exp. Med. 188:1657–68 Southern EM, Maskos U, Elder JK. 1992. Analyzing and comparing nucleic acid sequences by hybridization to arrays of oligonucleotides: evaluation using experimental models. Genomics 13: 1008–17 Southern EM, Case-Green SC, Elder JK, Johnson M, Mir KU, Wang L, Williams JC. 1994. Arrays of complementary oligonucleotides for analysing the hybridisation behaviour of nucleic acids. Nucleic Acids Res. 22:1368–73 Pietu G, Alibert O, Guichard V, Lamy B, Bois F, Leroy E, Mariage-Sampson R, Houlgatte R, Soularue P, Auffray C. 1996. Novel gene transcripts preferentially expressed in human muscles revealed by quantitative hybridization of a high density cDNA array. Genome Res. 6:492–503 Gress TM, Muller-Pillasch F, Geng M, Zimmerhackl F, Zehetner G, Friess H, Buchler M, Adler G, Lehrach H. 1996. A pancreatic cancer-specific expression profile. Oncogene 13:1819–30 Pease AC, Solas D, Sullivan EJ, Cronin MT, Holmes CP, Fodor SP. 1994. Light-

19.

20.

21.

22.

23.

24.

25.

26.

27.

853

generated oligonucleotide arrays for rapid DNA sequence analysis. Proc. Natl. Acad. Sci. USA 91:5022–26 Chee M, Yang R, Hubbell E, Berno A, Huang XC, Stern D, Winkler J, Lockhart DJ, Morris MS, Fodor SP. 1996. Accessing genetic information with highdensity DNA arrays. Science 274:610–14 Lipshutz RJ, Morris D, Chee M, Hubbell E, Kozal MJ, Shah N, Shen N, Yang R, Fodor SP. 1995. Using oligonucleotide probe arrays to access genetic diversity. Biotechniques 19:442–47 Lockhart DJ, Dong H, Byrne MC, Follettie MT, Gallo MV, Chee MS, Mittmann M, Wang C, Kobayashi M, Horton H, Brown EL. 1996. Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat. Biotechnol. 14:1675–80 Wodicka L, Dong H, Mittmann M, Ho MH, Lockhart DJ. 1997. Genome-wide expression monitoring in Saccharomyces cerevisiae. Nat. Biotechnol. 15:1359–67 Schena M, Shalon D, Davis RW, Brown PO. 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270:467–70 Schena M, Shalon D, Heller R, Chai A, Brown PO, Davis RW. 1996. Parallel human genome analysis: microarraybased expression monitoring of 1000 genes. Proc. Natl. Acad. Sci. USA 93:10614–19 Shalon D, Smith SJ, Brown PO. 1996. A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Res. 6:639–45 Lashkari DA, DeRisi JL, McCusker H, Namath AF, Gentile C, Hwang SY, Brown PO, Davis RW. 1997. Yeast microarrays for genome wide parallel genetic and gene expression analysis. Proc. Natl. Acad. Sci. USA 94:13057–62 Heller RA, Schena M, Chai A, Shalon D, Bedilion T, Gilmore J, Woolley DE,

854

28.

29.


30.

31.

32. 33.

34.

35.

36.

STAUDT n BROWN Davis RW. 1997. Discovery and analysis of inflammatory disease-related genes using cDNA microarrays. Proc. Natl. Acad. Sci. USA 94:2150–55 DeRisi JL, Iyer VR, Brown PO. 1997. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278:680–86 DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen Y, Su YA, Trent JM. 1996. Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat. Genet. 14:457–60 Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JCF, Trent JM, Staudt LM, Hudson J Jr, Boguski MS, Lashkari D, Shalon D, Botstein D, Brown PO. 1999. The transcriptional program in the response of human fibroblasts to serum. Science 283:83–87 Eisen MB, Spellman PT, Brown PO, Botstein D. 1998. Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95:14863–68 Kohonen T. 1997. Self-Organizing Maps. Berlin: Springer Tamayo P, Slonim D, Mesirov J, Zhu Q, Kitareewan S, Dmitrovsky E, Lander ES, Golub TR. 1999. Interpreting patterns of gene expression with self-organizing maps: methods and application to hematopoietic differentiation. Proc. Natl. Acad. Sci. USA 96:2907–12 Tavazoie S, Hughes JD, Campbell MJ, Cho RJ, Church GM. 1999. Systematic determination of genetic network architecture. Nat. Genet. 22:281–85 Alon U, Barkai N, Notterman DA, Gish K, Ybarra S, Mack D, Levine AJ. 1999. Broad patterns of gene expression revealed by clustering analysis of tumor and normal colon tissues probed by oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 96:6745–50 Spellman PT, Sherlock G, Zhang MQ, Iyer VR, Anders K, Eisen MB, Brown PO, Botstein D, Futcher B. 1998. Com-

37.

38.

39.

40.

41.

42.

43.

44.

prehensive identification of cell cycleregulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization. Mol. Biol. Cell. 9:3273–97 Alizadeh A, Eisen M, Botstein D, Brown PO, Staudt LM. 1998. Probing lymphocyte biology by genomic-scale gene expression analysis. J. Clin. Immunol. 18:373–79 Pryputniewicz SJ, Nagarkatti M, Nagarkatti PS. 1998. Differential induction of apoptosis in activated and resting T cells by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and its repercussion on T cell responsiveness. Toxicology 129:211–26 Roche S, Alonso G, Kazlauskas A, Dixit VM, Courtneidge SA, Pandey A. 1998. Src-like adaptor protein (SLAP) is a negative regulator of mitogenesis. Curr. Biol. 8:975–78 Chu S, DeRisi J, Eisen M, Mulholland J, Botstein D, Brown PO, Herskowitz I. 1998. The transcriptional program of sporulation in budding yeast. Science 282:699–705 Brown PO, Botstein D. 1999. Exploring the new world of the genome with DNA microarrays. Nat. Genet. 21:33–37 Marton MJ, DeRisi JL, Bennett HA, Iyer VR, Meyer MR, Roberts CJ, Stoughton R, Burchard J, Slade D, Dai H, Bassett DE Jr, Hartwell LH, Brown PO, Friend SH. 1998. Drug target validation and identification of secondary drug target effects using DNA microarrays. Nat. Med. 4:1293–1301 Cho RJ, Campbell MJ, Winzeler EA, Steinmetz L, Conway A, Wodicka L, Wolfsberg TG, Gabrielian AE, Landsman D, Lockhart DJ, Davis RW. 1998. A genome-wide transcriptional analysis of the mitotic cell cycle. Mol. Cell. 2:65–73 Gray NS, Wodicka L, Thunnissen AM, Norman TC, Kwon S, Espinoza FH, Morgan DO, Barnes G, LeClerc S, Meijer L, Kim SH, Lockhart DJ, Schultz PG. 1998. Exploiting chemical libraries,


45.

46.


47.

48.

49.

50.

51.

52.

53. 54.

structure, and genomics in the search for kinase inhibitors. Science 281:533–38 Holstege FC, Jennings EG, Wyrick JJ, Lee TI, Hengartner CJ, Green MR, Golub TR, Lander ES, Young RA. 1998. Dissecting the regulatory circuitry of a eukaryotic genome. Cell 95:717–28 Fambrough D, McClure K, Kazlauskas A, Lander ES. 1999. Diverse signaling pathways activated by growth factor receptors induce broadly overlapping, rather than independent, sets of genes. Cell 97:727–41 Darnell JE, Kerr IM, Stark GR. 1994. Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins. Science 264:1415–21 Harkin DP, Bean JM, Miklos D, Song YH, Truong VB, Englert C, Christians FC, Ellisen LW, Maheswaran S, Oliner JD, Haber DA. 1999. Induction of GADD45 and JNK/SAPK-dependent apoptosis following inducible expression of BRCA1. Cell 97:575–86 Amundson SA, Bittner M, Chen Y, Trent J, Meltzer P, Fornace AJ Jr. 1999. Fluorescent cDNA microarray hybridization reveals complexity and heterogeneity of cellular genotoxic stress responses. Oncogene 18:3666–72 Hermeking H, Lengauer C, Polyak K, He TC, Zhang L, Thiagalingam S, Kinzler KW, Vogelstein B. 1997. 14–3-3 sigma is a p53-regulated inhibitor of G2/M progression. Mol. Cell. 1:3–11 Polyak K, Xia Y, Zweier JL, Kinzler KW, Vogelstein B. 1997. A model for p53induced apoptosis. Nature 389:300–5 Kelsoe G. 1996. The germinal center: a crucible for lymphocyte selection. Semin. Immunol. 8:179–84 MacLennan ICM. 1994. Germinal centers. Annu. Rev. Immunol. 12:117–39 Gunn MD, Ngo VN, Ansel KM, Ekland EH, Cyster JG, Williams LT. 1998. A Bcell-homing chemokine made in lymphoid follicles activates Burkitt’s

55.

56.

57.

58.

59.

60.

61.

62.

855

lymphoma receptor-1. Nature 391:799– 803 Klein U, Goossens T, Fischer M, Kanzler H, Braeuninger A, Rajewsky K, Kuppers R. 1998. Somatic hypermutation in normal and transformed human B cells. Immunol. Rev. 162:261–80 Du M, Diss TC, Xu C, Peng H, Isaacson PG, Pan L. 1996. Ongoing mutation in MALT lymphoma immunoglobulin gene suggests that antigen stimulation plays a role in the clonal expansion. Leukemia 10:1190–97 Bahler DW, Levy R. 1992. Clonal evolution of a follicular lymphoma: evidence for antigen selection. Proc. Natl. Acad. Sci. USA 89:6770–74 Bahler DW, Miklos JA, Swerdlow SH. 1997. Ongoing Ig gene hypermutation in salivary gland mucosa-associated lymphoid tissue-type lymphomas. Blood 89:3335–44 Qin Y, Greiner A, Hallas C, Haedicke W, Muller-Hermelink HK. 1997. Intraclonal offspring expansion of gastric low-grade MALT-type lymphoma: evidence for the role of antigen-driven high-affinity mutation in lymphomagenesis. Lab. Invest. 76:477–85 Matsumoto M, Lo SF, Carruthers CJ, Min J, Mariathasan S, Huang G, Plas DR, Martin SM, Geha RS, Nahm MH, Chaplin DD. 1996. Affinity maturation without germinal centres in lymphotoxin-alpha-deficient mice. Nature 382: 462–66 Fukuda T, Yoshida T, Okada S, Hatano M, Miki T, Ishibashi K, Okabe S, Koseki H, Hirosawa S, Taniguchi M, Miyasaka N, Tokuhisa T. 1997. Disruption of the Bcl6 gene results in an impaired germinal center formation. J. Exp. Med. 186:439–48 Ye BH, Cattoretti G, Shen Q, Zhang J, Hawe N, de Waard R, Leung C, NouriShirazi M, Orazi A, Chaganti RS, Rothman P, Stall AM, Pandolfi PP, Dalla-Favera R. 1997. The BCL-6 proto-

856

63.

64.


65.

66. 67.

68.

69.

70.

71.

STAUDT n BROWN oncogene controls germinal-centre formation and Th2- type inflammation. Nat. Genet. 16:161–70 Dent AL, Shaffer AL, Yu X, Allman D, Staudt LM. 1997. Control of inflammation, cytokine expression, and germinal center formation by BCL-6. Science 276:589–92 Goodnow CC, Cyster JG, Hartley SB, Bell SE, Cooke MP, Healy JI, Akkaraju S, Rathmell JC, Pogue SL, Shokat KP. 1995. Self-tolerance checkpoints in B lymphocyte development. Adv. Immunol. 59:279–368 Healy JI, Goodnow CC. 1998. Positive versus negative signaling by lymphocyte antigen receptors. Annu. Rev. Immunol. 16:645–70 Vyse TJ, Todd JA. 1996. Genetic analysis of autoimmune disease. Cell 85:311–18 Gonzalez A, Katz JD, Mattei MG, Kikutani H, Benoist C, Mathis D. 1997. Genetic control of diabetes progression. Immunity 7:873–83 Wang B, Andre I, Gonzalez A, Katz JD, Aguet M, Benoist C, Mathis D. 1997. Interferon-gamma impacts at multiple points during the progression of autoimmune diabetes. Proc. Natl. Acad. Sci. USA 94:13844–49 Moser KL, Neas BR, Salmon JE, Yu H, Gray-McGuire C, Asundi N, Bruner GR, Fox J, Kelly J, Henshall S, Bacino D, Dietz M, Hogue R, Koelsch G, Nightingale L, Shaver T, Abdou NI, Albert DA, Carson C, Petri M, Treadwell EL, James JA, Harley JB. 1998. Genome scan of human systemic lupus erythematosus: evidence for linkage on chromosome 1q in African-American pedigrees. Proc. Natl. Acad. Sci. USA 95:14869–74 Todd JA. 1995. Genetic analysis of type 1 diabetes using whole genome approaches. Proc. Natl. Acad. Sci. USA 92:8560–65 Concannon P, Gogolin-Ewens KJ, Hinds DA, Wapelhorst B, Morrison VA, Stirling B, Mitra M, Farmer J, Williams SR,

72.

73.

74.

75.

76.

Cox NJ, Bell GI, Risch N, Spielman RS. 1998. A second-generation screen of the human genome for susceptibility to insulin-dependent diabetes mellitus. Nat. Genet. 19:292–96 Satsangi J, Parkes M, Louis E, Hashimoto L, Kato N, Welsh K, Terwilliger JD, Lathrop GM, Bell JI, Jewell DP. 1996. Two-stage genome-wide search in inflammatory bowel disease provides evidence for susceptibility loci on chromosomes 3, 7 and 12. Nat. Genet. 14:199–202 Sundvall M, Jirholt J, Yang HT, Jansson L, Engstrom A, Pettersson U, Holmdahl R. 1995. Identification of murine loci associated with susceptibility to chronic experimental autoimmune encephalomyelitis. Nat. Genet. 10:313–17 Ebers GC, Kukay K, Bulman DE, Sadovnick AD, Rice G, Anderson C, Armstrong H, Cousin K, Bell RB, Hader W, Paty DW, Hashimoto S, Oger J, Duquette P, Warren S, Gray T, O’Connor P, Nath A, Auty A, Metz L, Francis G, Paulseth JE, Murray TJ, Pryse-Phillips W, Risch R, et al. 1996. A full genome search in multiple sclerosis. Nat. Genet. 13:472– 76 Sawcer S, Jones HB, Feakes R, Gray J, Smaldon N, Chataway J, Robertson N, Clayton D, Goodfellow PN, Compston A. 1996. A genome screen in multiple sclerosis reveals susceptibility loci on chromosome 6p21 and 17q22. Nat. Genet. 13:464–68 Haines JL, Ter-Minassian M, Bazyk A, Gusella JF, Kim DJ, Terwedow H, Pericak-Vance MA, JB Rimmler, Haynes CS, Roses AD, Lee A, Shaner B, Menold M, Seboun E, Fitoussi RP, Gartioux C, Reyes C, Ribierre F, Gyapay G, Weissenbach J, Hauser SL, Goodkin DE, Lincoln R, Usuku K, Oksenberg JR, et al. 1996. A complete genomic screen for multiple sclerosis underscores a role for the major histocompatability complex. The Multi-


77.

78.


79.

80.

81.

82.

83.

ple Sclerosis Genetics Group. Nat. Genet. 13:469–71 Kuokkanen S, Sundvall M, Terwilliger JD, Tienari PJ, Wikstrom J, Holmdahl R, Pettersson U, Peltonen L. 1996. A putative vulnerability locus to multiple sclerosis maps to 5p14-p12 in a region syntenic to the murine locus Eae2. Nat. Genet. 13:477–80 Davies JL, Kawaguchi Y, Bennett ST, Copeman JB, Cordell HJ, Pritchard LE, Reed PW, Gough SC, Jenkins SC, Palmer SM, et al. 1994. A genome-wide search for human type 1 diabetes susceptibility genes. Nature 371:130–36 Jacob HJ, Pettersson A, Wilson D, Mao Y, Lernmark A, Lander ES. 1992. Genetic dissection of autoimmune type I diabetes in the BB rat. Nat. Genet. 2:56– 60 Hashimoto L, Habita C, Beressi JP, Delepine M, Besse C, Cambon-Thomsen A, Deschamps I, Rotter JI, Djoulah S, James MR, et al. 1994. Genetic mapping of a susceptibility locus for insulin-dependent diabetes mellitus on chromosome 11q. Nature 371:161–64 Hugot JP, Laurent-Puig P, Gower-Rousseau C, Olson JM, Lee JC, Beaugerie L, Naom I, Dupas JL, Van Gossum A, Orholm M, Bonaiti-Pellie C, Weissenbach J, Mathew CG, Lennard-Jones JE, Cortot A, Colombel JF, Thomas G. 1996. Mapping of a susceptibility locus for Crohn’s disease on chromosome 16. Nature 379:821–23 Matthews D, Fry L, Powles A, Weber J, McCarthy M, Fisher E, Davies K, Williamson R. 1996. Evidence that a locus for familial psoriasis maps to chromosome 4q. Nat. Genet. 14:231–33 Tomfohrde J, Silverman A, Barnes R, Fernandez-Vina MA, Young M, Lory D, Morris L, Wuepper KD, Stastny P, Menter A, et al. 1994. Gene for familial psoriasis susceptibility mapped to the distal end of human chromosome 17q. Science 264:1141–45

857

84. Daniels SE, Bhattacharrya S, James A, Leaves NI, Young A, Hill MR, Faux JA, Ryan GF, le Souef PN, Lathrop GM, Musk AW, Cookson WO. 1996. A genome-wide search for quantitative trait loci underlying asthma. Nature 383:247– 50 85. Remmers EF, Longman RE, Du Y, O’Hare A, Cannon GW, Griffiths MM, Wilder RL. 1996. A genome scan localizes five non-MHC loci controlling collagen-induced arthritis in rats. Nat. Genet. 14:82–85 86. Kono DH, Burlingame RW, Owens DG, Kuramochi A, Balderas RS, Balomenos D, Theofilopoulos AN. 1994. Lupus susceptibility loci in New Zealand mice. Proc. Natl. Acad. Sci. USA 91:10168–72 87. Morel L, Rudofsky UH, Longmate JA, Schiffenbauer J, Wakeland EK. 1994. Polygenic control of susceptibility to murine systemic lupus erythematosus. Immunity 1:219–29 88. Baker D, Rosenwasser OA, O’Neill JK, Turk JL. 1995. Genetic analysis of experimental allergic encephalomyelitis in mice. J. Immunol. 155:4046–51 89. Jirholt J, Cook A, Emahazion T, Sundvall M, Jansson L, Nordquist N, Pettersson U, Holmdahl R. 1998. Genetic linkage analysis of collagen-induced arthritis in the mouse. Eur. J. Immunol. 28:3321–28 90. Ghosh S, Palmer SM, Rodrigues NR, Cordell HJ, Hearne CM, Cornall RJ, Prins JB, McShane P, Lathrop GM, Peterson LB, et al. 1993. Polygenic control of autoimmune diabetes in nonobese diabetic mice. Nat. Genet. 4:404–9 91. Dallas-Pedretti A, McDuffie M, Haskins K. 1995. A diabetes-associated T-cell autoantigen maps to a telomeric locus on mouse chromosome 6. Proc. Natl. Acad. Sci. USA 92:1386–90 92. Rowe RE, Wapelhorst B, Bell GI, Risch N, Spielman RS, Concannon P. 1995. Linkage and association between insulindependent diabetes mellitus (IDDM) sus-

858

93.


94.

95.

96.

97.

98.

STAUDT n BROWN ceptibility and markers near the glucokinase gene on chromosome 7. Nat. Genet. 10:240–42 Field LL, Tobias R, Magnus T. 1994. A locus on chromosome 15q26 (IDDM3) produces susceptibility to insulindependent diabetes mellitus. Nat. Genet. 8:189–94 Gaffney PM, Kearns GM, Shark KB, Ortmann WA, Selby SA, Malmgren ML, Rohlf KE, Ockenden TC, Messner RP, King RA, Rich SS, Behrens TW. 1998. A genome-wide search for susceptibility genes in human systemic lupus erythematosus sib-pair families. Proc. Natl. Acad. Sci. USA 95:14875–79 Becker KG, Simon RM, Bailey-Wilson JE, Freidlin B, Biddison WE, McFarland HF, Trent JM. 1998. Clustering of nonmajor histocompatibility complex susceptibility candidate loci in human autoimmune diseases. Proc. Natl. Acad. Sci. USA 95:9979–84 Odunsi K, Kidd KK. 1999. A paradigm for finding genes for a complex human trait: polycystic ovary syndrome and follistatin. Proc. Natl. Acad. Sci. USA 96:8315–17 Foster CB, Lehrnbecher T, Mol F, Steinberg SM, Venzon DJ, Walsh TJ, Noack D, Rae J, Winkelstein JA, Curnutte JT, Chanock SJ. 1998. Host defense molecule polymorphisms influence the risk for immune-mediated complications in chronic granulomatous disease. J. Clin. Invest. 102:2146–55 Aitman TJ, Glazier AM, Wallace CA, Cooper LD, Norsworthy PF, Wahid FN, Al-Majali KM, Trembling PM, Mann CJ, Shoulders CC, Graf D, St. Lezin E, Kurtz TW, Kren V, Pravenec M, Ibrahimi A, Abumrad NA, Stanton LW, Scott J. 1999. Identification of Cd36 (Fat) as an insulinresistance gene causing defective fatty acid and glucose metabolism in hypertensive rats [see comments]. Nat. Genet. 21:76–83

99. Febbraio M, Abumrad NA, Hajjar DP, Sharma K, Cheng W, Pearce SF, Silverstein RL. 1999. A null mutation in murine CD36 reveals an important role in fatty acid and lipoprotein metabolism. J. Biol. Chem. 274:19055–62 100. Halushka MK, Fan JB, Bentley K, Hsie L, Shen N, Weder A, Cooper R, Lipshutz R, Chakravarti A. 1999. Patterns of single-nucleotide polymorphisms in candidate genes for blood-pressure homeostasis. Nat. Genet. 22:239–47 101. Jacob CO, McDevitt HO. 1988. Tumour necrosis factor-alpha in murine autoimmune ‘lupus’ nephritis. Nature 331:356– 58 102. Nelms K, Snow AL, Hu-Li J, Paul WE. 1998. FRIP, a hematopoietic cell-specific rasGAP-interacting protein phosphorylated in response to cytokine stimulation. Immunity 9:13–24 103. Stoye JP, Fenner S, Greenoak GE, Moran C, Coffin JM. 1988. Role of endogenous retroviruses as mutagens: the hairless mutation of mice. Cell 54:383–91 104. Ahmad W, Faiyaz ul Haque M, Brancolini V, Tsou HC, ul Haque S, Lam H, Aita VM, Owen J, deBlaquiere M, Frank J, Cserhalmi-Friedman PB, Leask A, McGrath JA, Peacocke M, Ahmad M, Ott J, Christiano AM. 1998. Alopecia universalis associated with a mutation in the human hairless gene. Science 279:720–24 105. Zhu H, Cong JP, Mamtora G, Gingeras T, Shenk T. 1998. Cellular gene expression altered by human cytomegalovirus: global monitoring with oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:14470–75 106. Harhaj EW, Good L, Xiao G, Sun SC. 1999. Gene expression profiles in HTLVI-immortalized T cells: deregulated expression of genes involved in apoptosis regulation. Oncogene 18:1341–49 107. Natanson C, Hoffman WD, Suffredini AF, Eichacker PQ, Danner RL. 1994.



Selected treatment strategies for septic shock based on proposed mechanisms of pathogenesis. Ann. Intern. Med. 120: 771–83 108. Mizoguchi H, O’Shea JJ, Longo DL, Loeffler CM, McVicar DW, Ochoa AC. 1992. Alterations in signal transduction

859

molecules in T lymphocytes from tumorbearing mice. Science 258:1795–98 109. Whiteside TL. 1998. Immune cells in the tumor microenvironment. Mechanisms responsible for functional and signaling defects. Adv. Exp. Med. Biol. 451:167– 71


STAUDT

■

BROWN C-1

Figure 1 A Schematic of cDNA microarray gene expression analysis. In this illustration, the relative gene expression in a mature B cell and a plasma cell is compared. Gene X represents a gene more highly expressed in the plasma cell. See text for details. B Quantitative analysis of relative gene expression using cDNA microarrays. For each spot on the microarray, the fluorescence intensities of hybridized Cy3- and Cy5-labelled cDNA probes are separately quantitated, as shown in the middle panel. The Cy5/Cy3 fluorescence intensity ratio is a measure of relative gene expression in the two starting mRNA samples. The fluorescence ratios are divided into numerical bins and depicted visually using the color scale shown at the right.


C-2 STAUDT

■

BROWN

Figure 2 Gene expression profiles identified using hierarchical clustering algorithms. DNA microarray measurements of gene expression were taken from mitogenically activated T lymphocytes and serum-stimulated fibroblasts over a time course. Gene expression at each time point was measured relative to gene expression in unstimulated cells. The relative gene expression data is depicted using the color scheme of Figure 1B with the brightest red and green boxes representing eightfold induced or repressed genes, respectively. The left panel displays the genes in random order, while the right panel displays the genes in the order determined by hierarchical clustering.

■

Figure legend appears on the following page.


STAUDT BROWN C-3

Figure 3 (see preceding page) A gene expression map of Saccharomyces cerevisiae. The left panel depicts the results from 204 genome-wide microarray gene expression experiments in Saccharomyces cerevisiae. Yeast were placed under 28 distinct physiological or nutritional conditions (delineated by the vertical black stripes) and assayed multiply over time. Each column represents one microarray experiment, and each row represents one of the 6220 known or predicted genes of yeast. The coordinate regulation of genes encoding ribosomal subunits is indicated. Genes involved in meiosis and respiration form separate clusters which are expanded at the right.


C-4 STAUDT ■

BROWN


STAUDT

■

BROWN C-5

Figure 4 Gene expression signatures in lymphocyte differentiation. Gene expression measurements were taken from four categories of lymphocyte differentiation/activation: resting peripheral blood B cells (both naive and memory), in vitro activated peripheral blood B cells (anti-IgM +/Ð CD40 ligand +/Ð IL-4), tonsillar germinal center and memory B cells, and resting or activated (PMA + ionomycin) T cells. The genes were chosen to highlight the difference between germinal center B cells and in vitro activated peripheral blood B cells.


C-6 STAUDT

■

BROWN

Figure 5 Gene expression in normal and malignant B cells. Gene expression in diffuse large B cell lymphoma (DLCL) lymph node biopsies was compared with gene expression in tonsillar germinal center B cells, tonsillar memory B cells, resting peripheral blood B cells, and peripheral blood B cells activated in vitro with anti-IgM + CD40 ligand + IL-4 for 6 and 24 h. A subset of diffuse large B cell lymphomas resembles normal germinal center B cells.


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


VIRAL SUBVERSION OF THE IMMUNE SYSTEM Domenico Tortorella, Benjamin E. Gewurz, Margo H. Furman, Danny J. Schust, and Hidde L. Ploegh Department of Pathology, Harvard Medical School, Boston, Massachusetts 02115; e-mail: [email protected], [email protected], [email protected], [email protected], [email protected]


Key Words antigen presentation, apoptosis, cytokines, complement, viral evasion Abstract This review describes the diverse array of pathways and molecular targets that are used by viruses to elude immune detection and destruction. These include targeting of pathways for major histocompatibility complex-restricted antigen presentation, apoptosis, cytokine-mediated signaling, and humoral immune responses. The continuous interactions between host and pathogens during their coevolution have shaped the immune system, but also the counter measures used by pathogens. Further study of their interactions should improve our ability to manipulate and exploit the various pathogens.

INTRODUCTION Given their short generation times, a proper host defense against viral pathogens is essential and is aimed at several levels. Mechanical protection afforded by skin and epithelia can keep out many intruders, but viruses often possess specialized mechanisms to breach epithelial barriers. Rapid deployment of cells and molecules of the innate immune system is the next layer of defense. As the evolution of successful virulent pathogens attests, these defenses, too, can be penetrated. Finally, in many cases, lasting protection is afforded through acquired immunity. The trait of immunological memory allows re-exposure to many pathogens to be experienced without significant morbidity. Still, the lifetime persistence and repeated reactivation of many viruses must be enabled by specific evasion of adaptive immunity. The coexistence of pathogens with their host should be viewed as a precarious balance. Without a susceptible reservoir, pathogens could not survive. Elimination of the host by an extremely virulent version of a pathogen would not, at first glance, appear to be a viable evolutionary strategy from the pathogen’s perspective. This poses the interesting question of how this balance is maintained and continues to evolve. More detailed insight into interactions of viruses with the 0732–0582/00/0410–0861/$14.00

861

862

TORTORELLA ET AL


immune system stands to generate concepts for more adequate vaccine strategies and will also help us to better understand the immune system. This review recapitulates important means of escape from immune detection and also recounts how certain aspects of a normally protective host response have been redirected or modified to the advantage of the virus. Concepts such as antigenic variation as the means to escape an antibody and cytotoxic T lymphocyte (CTL) responses were among the first of these strategies to be described; although of immediate relevance, these mechanisms of structural variation are beyond the scope of this review. Instead, we focus on the following areas. 1. The expression of cell surface molecules such as major histocompatibility complex (MHC) products and the coreceptors CD4 and CD8, as well as countless other surface structures, is required to initiate and sustain an immune response. Ablation of their expression is a prevalent theme exploited by viruses and is accomplished at every conceivable level of their synthesis; modulation of transcription, protein synthesis, accelerated destruction, intracellular retention, and increased clearance from the cell surface are common occurrences. With the near certainty that new viruses remain to be discovered and that many known viruses are incompletely understood, it is a safe bet that the list of target proteins in the cross hairs of viruses is far from complete. 2. Natural killer (NK) cells operate where mechanisms of CD8` T-cell–mediated killing fail, for example because of the removal of the class I molecules themselves. NK cells rely on the presence of class I molecules for silencing or inhibition. Remarkably, certain viruses contain class I homologs that may obstruct the signals required for NK activation. In addition, immunoevasive strategies aimed at elimination of class I molecules may avoid elimination of those class I products that silence NK cells. 3. Programmed cell death or apoptosis is increasingly perceived as an important aspect of homeostasis and serves to eliminate cells that have outlived their useful life span or that have sustained injuries considered beyond repair. These situations would include the hijacking of the cell’s machinery by viruses to accomplish their multiplication. Not surprisingly, many viruses have defused this pathway to create a favorable environment. Mechanisms that are presently known are reviewed. 4. Cytokines are the messenger molecules by which cells of the immune system communicate, both at short range and at a distance. The ability to slow cytokine production has protective value, and, from the perspective of the virus, is a thing to be neutralized or counteracted. Examples of this type are particularly prevalent among the large DNA viruses such as poxviruses or herpesviruses. 5. The humoral immune response relies on the ability to effectively process and eliminate immune complexes, a process in which both complement and Fc receptors play key roles. We discuss several examples of viral proteins that manipulate this response.

IMMUNE EVASION BY VIRUSES

863

VIRAL INHIBITION OF MHC CLASS I–RESTRICTED ANTIGEN PRESENTATION


Class I Antigen Presentation The T-cell receptor complex of CD8` CTLs recognizes MHC class I molecules on infected and professional antigen-presenting cells (1, 2). MHC class I molecules present virus derived antigenic peptides to CTLs that can activate an antiviral response and attract other immune system components (3, 4). Hence, class I surface expression is essential for antiviral immunity. Viruses have evolved several strategies to downregulate the surface expression of class I molecules at the transcriptional and post-translational levels. Here we focus on the viral proteins that can interact with class I molecules and interfere with antigen presentation (Table 1; Figure 1). The mature class I molecule exists as a trimeric complex comprised of a heavy chain, light chain, and peptide (5). The heavy chain is a 43-kDa type I membrane glycoprotein and consists of an immunoglobulin (Ig)-like domain that supports a peptide groove formed by two antiparallel a-helices and an eight-stranded b-sheet (6). The light chain is the 12-kDa soluble protein b2-microglobulin (b2m). Peptide ligands for class I molecules are generated in the cytoplasm mostly by the proteasome. Presentation of endocytosed material may be far more prevalent than originally recognized and, of particular relevance for so-called cross-priming or cross-presentation, pathways that are key in initiating an antiviral CTL response (7). Peptides of 8–10 residues are translocated from the cytosol into the endoplasmic reticulum (ER) lumen in an ATP-dependent manner through the MHCencoded complex for the transporter associated with antigen processing (TAP) (5). In the course of and shortly after synthesis of the class I heavy chain and light chain, their folding and peptide loading are assisted by ER-resident chaperones such as calnexin, calreticulin, and the thiol-reductase ERp57 (8–10). Tapasin, an ER-resident protein that interacts with TAP, retains the peptide-free class I dimer/chaperone complex in the ER and may also guide it to the TAP complex to be loaded with peptide (11, 12). The trimeric class I complex exits the ER, proceeds through the Golgi, and is displayed at the cell surface, where it may be sampled by CTLs for detection of foreign peptides of viral or pathogenic origin.

Inhibition of Generation of Antigenic Peptide During virus infection, viral gene products expressed in the cytosol may be targeted for degradation and presented by class I molecules (13). In this manner, CTL can act early to eliminate the infected cell. To remain undetected, viruses may interfere with proteolysis. During latency, the Epstein-Barr virus (EBV or human herpesvirus 5)-infected B cell expresses the EBV nuclear antigen-1 (EBNA-1) and the latent membrane protein-1 (LMP-1). CTLs specific for EBNA-1 are generated during T-cell devel-

864

TORTORELLA ET AL

TABLE 1 Viral gene products that interfere with major histocompatibility complex Virus Adenovirus

EBV


HCMV

Gene product

Function

E3/19K

Retains class I molecules in ER

(235)

E1A

Interferes with MHC class II upregulation (IFN-c signal transduction cascade)

(87)

EBNA-1

Refractory to proteolysis

(15)

BZLF2

May interfere with MHC class II antigen presentation

(13, 72)

Pp65

Inhibits generation of antigenic peptides of a 72kDa transcription factor

(20)

US2

Targets MHC class I heavy chains (class II DR, DM a chain) for degradation

(48, 71)

US3

Retains class I molecules in the ER

(45, 46)

US6

Inhibits TAP

(34–37)

US11

Targets MHC class I heavy chains for degradation

(367)

UL18

Inhibits NK cell lysis (MHC class I homolog)

(142)

IE/E product Interferes with MHC class II upregulation (IFN-c signal transduction cascade) HIV

Reference(s)

(85, 86)

Nef

Rapid endocytosis of cell-surface MHC class I and (58) CD4, may interfere with MHC class II processing (acidification of endosomal vesicles)

Vpu

Destabilizes newly synthesized class I molecules and targets CD4 for degradation

(66)

E5

May interfere with MHC class II processing (acidification of endosomal vesicles)

(78)

E6

May interfere with MHC class II processing (interaction with AP complex)

(77)

HSV

ICP-47

Inhibits TAP

(26–28)

HSV-1 (KOS)

Unknown

May interfere with MHC class II antigen presentation

(73)

MCV

MC080R

May inhibit NK cell lysis (MHC class I homolog)

(139)

MCMV

m152

Retains MHC class I in ERGIC

(55)

m04

Binds MHC class I molecules

(57)

m06

Targets MHC class I molecules for degradation in lysosomes

(56)

m144

Inhibits NK cell lysis (MHC class I homolog)

(141)

r144

May inhibit NK cell lysis (MHC class I homolog)

(140)

HPV/BPV

RCMV

865



Figure 1 Viral proteins that interfere with the processing steps of major histocompatibility complex (MHC) class I antigen presentation. EBV, Epstein-Barr virus; ERGIC, ERGolgi intermediate compartment; HCMV, human cytomegalovirus; HIV, human immunodeficiency virus; HSV, herpes simplex virus; MCMV, mouse cytomegalovirus.


866

TORTORELLA ET AL

opment, but fail to be activated during an EBV infection, suggesting that the appropriate EBNA-1–derived peptides are not presented by class I molecules (14). EBNA-1 is a phosphoprotein that contains a 239-residue stretch of Gly-Ala repeats, implicated in inhibition of its degradation (15). Other EBV proteins such as EBNA-2, -3, and -4 do not contain a Gly-Ala domain, and their degradation products are presented by class I molecules (16, 17). Chimeric proteins consisting of EBNA-4 and either a 200-amino-acid (aa) or 17-aa Gly-Ala stretch in cis prevent the presentation of epitopes from EBNA-4 (15). A chimeric molecule in which an eight-residue fragment of the Gly-Ala repeat was inserted into the transcriptional repressor IjB, a protein targeted for proteasomal degradation after phosphorylation in response to various external stimuli, was ubiquitinated, but not degraded (18). Whether the Gly-Ala repeat evolved specifically for evasion of an EBNA-1–specific CTL response or whether the long half-life of EBNA-1 is essential for maintenance of the latent state is not known. How does the GlyAla repeat prevent proteasomal proteolysis? Does it assume a unique secondary structure that would prevent EBNA-1’s productive association with the proteasome? Structural analysis of the Gly-Ala domain suggests that it is a random coil, making any such prediction difficult (19). EBNA-1’s resistance to proteolysis remains poorly understood. During the immediate early phase of human cytomegalovirus (HCMV) infection, a CTL response is directed against antigenic peptides derived from a 72kDa transcription factor (20). When the 72-kDa protein is expressed together with the matrix protein phosphoprotein 65 (pp65), which has kinase activity (21, 22), there is no CTL response against the 72-kDa protein (23). pp65 may phosphorylate the 72-kDa protein and inhibit the generation of the 72-kDa–derived antigenic peptides, a process that appears temporally controlled. How a phosphorylated protein can avoid protein degradation is not known, but phosphorylation might affect p72’s cleavage pattern or result in its failure to be appropriately Ub conjugated. Human immunodeficiency virus (HIV) and influenza virus, to name but two examples, can modify antigenic peptides through random mutations. We do not discuss the phenomena of antigenic drift/shift whereby recognition by antibodies or T cells may be avoided easily by rapid, substantial alterations in the peptide that is recognized. However, structural variation in antigenic proteins affects not only direct recognition by antibodies, but can also interfere with proteolysis, subsequent transport of peptides, or their association with MHC products. An antigenic peptide derived from HIV can even act as an antagonist peptide (24, 25) and thus prevent full activation of specific T cells.

Antigenic Peptides Transported via the TAP Complex Several herpesviruses express proteins that inhibit the TAP complex. Herpes simplex virus 1 (HSV-1) and HSV-2 express a soluble 9-kDa immediate early (IE) gene product, infected-cell protein 47 (ICP47), that prevents peptide transport



867

(26–28). ICP47 prevents peptide binding by interacting with both TAP1 and TAP2 on the cytosolic side of the ER (29, 30). Residues 2–35 of ICP47 are the minimum region required for TAP interaction and inhibition of peptide transport (29), which ICP47 accomplishes by weakening the association between TAP1 and TAP2 (31). ICP47 is moderately species specific; it inhibits human, porcine, canine, bovine, and simian TAP, but not murine, rat, guinea pig, and rabbit TAP (30, 32, 33). HCMV contains within its unique short (US) region the US6 gene, the product of which inhibits TAP function (34–37). US6 is 21-kDa type I membrane glycoprotein expressed during both early and late phases of viral infection. Unlike ICP47, it interacts with the TAP complex on the lumenal side of the ER. A soluble version of US6 that lacks the transmembrane region and cytoplasmic tail also prevents peptide transport. Interaction of US6 with TAP does not prevent peptide or ATP binding. Hence, the US6-TAP interaction may prevent peptide transport by occluding the exit pore of the TAP complex. It is likely that other herpesviruses have evolved distinct, yet to be discovered mechanisms of TAP inhibition, with various degrees of host specificity.

Inhibition of MHC Class I Surface Expression The adenovirus E3/19-kDa protein (E19) retains human as well as murine class I molecules in the ER and, in so doing, prevents the display of MHC molecules complexed with viral peptides (38). E19 binds to the a1 and a2 regions (39, 40) of class I heavy chains with allelic preference (41). Such ER retention by E19 is attributable to a double lysine motif within its cytoplasmic tail. This dilysine motif interacts with coat protein-I (COP-I; see below), involved in vesicle retrieval between the ER and Golgi, and is a common ER retention motif (42, 43). E19 or a truncated version without its ER retention signal can also bind to TAP, independently of its association with MHC class I molecules and tapasin. E19 may therefore inhibit peptide loading of the MHC class I molecules (44). At the immediate early phase of virus infection, HCMV expresses US3, a 23kDa glycoprotein that retains class I molecules in the ER (45, 46). US3 interacts with peptide-loaded MHC class I molecules, but its mechanism of ER retention is not clear. There is no obvious ER retention motif within US3 that would facilitate its association with proteins involved in ER retention. However, the US3 molecule strongly and persistently interacts with calnexin (P Stern, H Ploegh, K Fruh, unpublished observation), itself retained in the ER through a dilysine motif. HCMV encodes two gene products, US2 and US11, that selectively target class I heavy chains for degradation by the proteasome. US2 or US11 induces the destruction of the human leukocyte antigen-A (HLA-A) and -B locus products, but not the HLA-C and -G locus products (47). In cells that express US2 or US11, class I heavy chains are dislocated from the ER through the translocon (Sec61p complex) and into the cytosol (48). The dislocation reaction is dependent on ATP and is sensitive to changes in the redox potential (48, 49). During dislocation, the


868

TORTORELLA ET AL

class I heavy chains are deglycosylated—probably on the cytosolic face (50) of the ER—by N-glycanase, followed by their degradation. The US2 and US11 gene products are type I membrane glycoproteins (51). US2 associates with class I molecules through its lumenal region both in vivo and in vitro. (B Gewurz & H Ploegh, unpublished results). The extreme COOH end of US2 targets class I heavy chains for destruction; deletion of only 5 residues suffices to inactivate dislocation, but not the class I binding function of US2 (D Tortorella, & H Ploegh, unpublished results). The COOH end of US11 is dispensable for inducing class I heavy-chain degradation. The cytoplasmic tail of the class I heavy chain is required for its degradation, despite the fact that ‘‘tailless’’ class I molecules still interact with the viral gene products (52). Small amounts of ubiquitinated class I heavy chains can be detected in US2- and US11expressing cells exposed to proteasome inhibitors. Some of those ubiquitinatedconjugated class I heavy chains can be detected in membrane-associated form. The bulk of the Ub class I heavy-chain intermediates have lost their N-linked glycan (53). Class I heavy chains may therefore be ubiquitinated largely after extraction from the ER membrane. The mechanism of degradation of class I heavy chains by US2 or US11 is strikingly similar to the degradation of misfolded and abnormal proteins in the ER (54). US2 and US11 may function merely by accelerating the rate of an otherwise constitutive process and confer specificity by selective interaction with class I products. Mouse cytomegalovirus (MCMV) and HCMV display 70% sequence similarity, roughly the same as their natural hosts. MCMV also displays immune evasive behavior, but it is rather different from that seen for HCMV. For example, MCMV possesses no region homologous to the US region of HCMV. Conversely, MCMV uses a set of genes without clear homologs in HCMV. The MCMV gene m152 encodes a 37-kDa type I glycoprotein expressed early during infection. The m152 gene product retains class I molecules within the ER-Golgi intermediate compartment, which prevents surface expression of class I molecules (55). The complex is held within the ER-Golgi intermediate compartment through a retention signal within the lumenal region of m152. The MCMV gene m06 encodes a 48kDa glycoprotein that also prevents the class I complex from reaching the cell surface (56). The m06 gene product binds tightly through its lumenal/transmembrane region to the class I molecules and targets class I molecules to the lysosomes. The m06 product contains two dileucine motifs that can act as lysosomal targeting signals. MCMV gene m04 encodes a 34-kDa glycoprotein (gp34) that interacts with class I molecules within the ER. The gp34/MHC class I complex is detectable in the ER and at the cell surface (57). The function of gp34 is unclear, but it is expressed as the levels of the m152 transcript are waning, suggesting that gp34 may be required when m152‘s role is played out. Because the m04 gp34 product is associated with class I molecules at the cell surface, it could alter the interaction with CD8` T cells or affect the induction of a full CD8` T-cell response.



869

MCMV, like HCMV, has evolved multiple and distinct strategies that interfere with class I surface expression. Whereas some of the strategies used by human pathogens have a certain appeal because they make immunological sense, pathogens such as MCMV have the decided advantage of allowing experiments in animals. The best evidence in support of immune evasion having a role in virulence, not surprisingly, comes from studies that exploit rodent models for pathogenesis. Mice infected with wild-type MCMV or a recombinant MCMV mutant lacking the m152 gene showed an attenuated infection for the mutant virus and a requirement for CD8` T cells to clear the virus. HIV expresses two proteins that downregulate the expression of surface MHC class I molecules, Nef and Vpu (58). Nef is a 206-aa myristoylated protein that localizes to the plasma membrane and accelerates endocytosis of class I complexes. The class I/Nef complex is then targeted to the lysosomes. Nef targets the HLA-A and -B locus products, but not the -C and -E locus products for lysosomal degradation (59, 60). Residues in the cytoplasmic tail of the HLA-A and HLAB heavy chain facilitate the interaction between class I molecules and Nef (Figure 2). The extent of downregulation accomplished by Nef is sufficient to abolish

Figure 2 The critical residues within the cytoplasmic tails of CD4 and major histocompatibility complex (MHC) class I A2 locus product that mediate Nef internalization are underlined. The residues within CD4 vary slightly among human immunodeficiency virus (HIV)-1, HIV-2, and simian immunodeficiency virus (SIV) Nef.


870

TORTORELLA ET AL

recognition by CD8` CTLs (61). The specific targeting of the HLA-A and -B locus products and not the -C and -E locus products may be relevant for NK cell recognition (see section below on ‘‘Viral Inhibition of Natural Killer Cell Lysis’’) (62–64). The sites of interaction of Nef with the endocytic machinery and class I molecules have been mapped (Figure 3). This interaction may facilitate the rapid endocytosis of cell-surface class I molecules, followed by accumulation of the complex in the trans-Golgi network (TGN) and delivery to lysosomes by the display of a lysosomal targeting motif within Nef. However, Nef mutants that do not associate with the endocytic machinery continue to downregulate class I molecules. The Nef protein contains an SH3-binding domain that allows interaction with different kinases (65). Whether the interactions between Nef and cellular kinases assist in the endocytosis function of Nef has yet to be determined. The HIV protein Vpu prevents the cell surface expression of class I molecules (66). Vpu attacks only newly synthesized class I molecules and induces their destabilization, as judged by a conformation-specific anti-class I antibody. Phosphorylation of Vpu may be required because a nonphosphorylatable Vpu mutant

Figure 3 Regions of Nef that facilitate its many functions. Cell surface localization, aa1– 10; CD4 binding, aa WLE(57–59), GGL(95–97), R106, and L110; adaptor protein (AP) complex binding, human immunodeficiency virus-1: aa WL(57–58), 143–170, LL(164– 165), E174, ERE(177–179), *HIV-2: aa Y39, SIV: aa YY28/39, 58–88; CD4 endocytosis, aaG29, D36, A56, WLE(57–59), SS175; MHC class I endocytosis, aa 17–26, EEEE(62– 65), P72, P78; Lysosomal targeting, EE(154–155). The amino acids are designated by the standard single letter code.


871

does not affect class I stability. How Vpu disrupts the processing of class I molecules is unknown, but Vpu’s downregulation of CD4 could provide important clues (see below).

VIRAL INHIBITION OF MAJOR HISTOCOMPATIBILITY COMPLEX CLASS II RESTRICTED-ANTIGEN PRESENTATION


Class II Antigen Presentation When pathogens enter the endocytic pathway by phagocytosis or receptormediated endocytosis, many of their proteins are degraded into antigenic peptides, mostly by endosomal proteases. These peptides are presented by class II molecules to CD4` T cells (67, 68). Although class II molecules are expressed on the cell surface of B cells, macrophages, and dendritic cells, they can be induced on a variety of other cell types. Activated CD4` T cells stimulate the development of CTLs and help coordinate an antiviral response against the pathogen (69). Preventing class II antigen presentation would thus interfere with helper-T-cell activation. A mature class II molecule is composed of an ab heterodimer and an antigenic peptide. In the ER, the ab dimer associates tightly with the invariant chain (Ii). A portion of the Ii chain, class II-associated invariant chain peptide, binds within the class II peptide groove to prevent endogenous peptides from binding to the class II molecules while in the ER. Two leucines within the cytoplasmic domain of the Ii chain direct the a/b/Ii complex from the TGN to acidic compartments. Proteases within these compartments cleave the invariant chain and leave a class II-associated invariant chain peptide remnant (70). The class II-associated invariant chain peptide is then exchanged for the antigenic peptide within the acidic compartment with the aid of HLA-DM. Peptide stabilizes the class II molecules and allows the complexes to proceed to the cell surface where they can interact with the T-cell receptor of CD4` T cells.

Inhibition of Surface Expression of Class II Molecules The HCMV gene product US2 targets class I molecules for degradation shortly after synthesis, but US2 can also target the class II DR-a and DM-a molecules for degradation by proteasomes (71). Even though the class I and class II molecules share little primary sequence homology, their secondary and tertiary structures are quite similar, and these shared structural features may contribute the target for attack by US2. Additional viral gene products interact with class II molecules and may interfere with antigen presentation. The EBV gene BZLF2 encodes for a 42-kDa type II membrane glycoprotein that interacts with cell surface and intracellular class

872

TORTORELLA ET AL

II molecules. The BZLF2 protein interferes with T-cell activation and possibly the presentation of antigenic peptides (13, 72), but the underlying mechanism has not been clarified. Infection of class II` cells with the HSV-1 KOS strain redistributes class II molecules away from the endocytic compartment and would prevent antigen presentation (73). Even though trafficking of class II through the endocytic pathway was not explored directly, it is possible that viruses may have evolved ways to prevent class II molecules from generating an immune response, such as stimulating the production of cytokines such as interleukin-10 (IL-10; 74, 75; see section on ‘‘Interference with Receipt of Cytokine Signal’’). At first glance, the unique sets of polypeptides (Ii, DM, and DO) involved in class IIrestricted presentation constitute attractive targets for specific viral or microbial interference in a manner that is unlikely to perturb more general aspects of physiology.


Interference in the Endocytic Pathway Vesicle transport to and from the cell surface is important for delivering cargo to its proper destination (76), a process both complex and highly regulated. Vesicles fuse and bud from the different cellular organelles by creating a ‘‘coat’’ around the forming vesicle, often composed of clathrin and additional factors called adaptor protein (AP) complexes. One AP complex consists of a heterotetramer of two ;100-kDa a and b subunits, one ;50-kDa l subunit, and one 19-kDa r subunit. Distinct AP complexes mediate different transport steps; AP-1 is involved in vesicle formation at the TGN and targets vesicles for early endosomes, whereas AP-2 initiates early endocytic vesicle formation at the plasma membrane. The AP-3 complex drives vesicle formation at the TGN and targets vesicles to late endocytic compartments/lysosomes. Non–clathrin-coated vesicles contain a different set of coat proteins, termed COP proteins. COP complexes are involved in vesicle trafficking early in the secretory pathway between the ER and Golgi. The COP-I–coated vesicles consist of COP proteins (a, b, b8, c, d, e, and f ) and the ADP-ribosylation factor-1 that mediates retrograde transport of vesicles within the Golgi and between the Golgi and ER. COP-I vesicles have also been found in late endosomes and may be involved in vesicle transport of early to late endosomes, but these vesicles lack the c and d subunits. COP II vesicles are found in the ER and contain the COP proteins and additional components such as sec23p/24p complex, sec13/31p complex, and GTP-binding protein sar1p. Because class II molecules must exploit both the secretory and endocytic pathways to perform their function, the complexity of the transport machinery affords many possible targets for interference by microbes. Papillomavirus is a DNA virus that infects keratinocytes as well as fibroblasts. The bovine (BPV) and human (HPV) papillomaviruses encode proteins that directly interact with components of the endocytic pathway. BPV E6 interacts with AP-1, the TGN-specific clathrin adaptor complex (77). The association of



873

E6 with the endocytic machinery may alter the processing of cellular proteins and could affect their processing in the endocytic pathway, although there is currently no direct experimental support for this suggestion. HPV and BPV express E5, which, at least for BPV, is known to interact with the 16K subunit of the vacuolar proton pump or vacuolar H` ATPase (78). The vacuolar H` ATPase is composed of 5–10 proteins, and the 16K protein is believed to be the main component of the pore. The presence of E5 inhibits the degradation of cell surface receptors (79–81), which suggests that it may influence acidification of endocytic compartments and consequently processing of antigenic proteins. The HIV Nef protein also interacts with the endocytic machinery and triggers the rapid internalization of CD4 and MHC class I molecules from the cell surface (see below), placing Nef in close proximity to the class II molecule-containing compartments. Nef associates with a protein termed the Nef-binding protein 1 (82). Nef-binding protein 1 is homologous to the 56-kDa subunit of the vacuolar H` ATPase complex. This interaction with the vacuolar H` ATPase is believed to aid in the downregulationof CD4 from the cell surface by an unknown mechanism. However, Nef’s ability to associate with the proton pump further suggests that it may interfere with the acidification of MHC class II-containing vesicles and could affect presentation of antigenic peptides.

Inhibition of Interferon-Induced Expression of MHC Class II Molecules Signal transduction cascades initiated by extracellular ligands modulate the expression and regulation of MHC class II molecules (83). Exposure of susceptible cells to cytokines such as interferon-c (IFN-c) activates transcription of the class II transcription factor, CIITA, via the Jak/Stat pathway and thereby drives expression of class II molecules. Viruses can interfere with IFN-induced MHC expression (84). Proteins expressed at the immediate early and early phase of an HCMV infection tamper with the IFN-c signaling cascade and prevent the expression of CIITA (85). HCMV destabilizes Jak1 and p48, which are required transduction intermediates for proper activation of CIITA expression (86). The adenovirus E1A protein alters the STAT1 and p48 levels within the cell, which interferes with the IFN-a and -c signaling pathway (87, 88). MCMV also inhibits IFN-c–induced expression of class II molecules (89). Unlike HCMV, MCMV affects signal transduction subsequent to STAT1a activation (90).

DOWNREGULATION OF CD4 BY HUMAN IMMUNODEFICIENCY VIRUS Downregulation of Cell Surface CD4 by Nef The HIV accessory proteins Nef and Vpu are not required for viral replication, but are important for viral pathogenesis (58, 91). Both Nef and Vpu are multifunctional; they aid in virion release from the infected cell and downregulate


874

TORTORELLA ET AL

immunologically relevant molecules such as MHC class I molecules and CD4. HIV primarily infects CD4-bearing T cells and macrophages (92–94). HIV uses CD4 and the chemokine receptor CXCR4 or CCR5 (for HIV-1) to dock to the cell membrane before viral entry (95, 96). HIV downregulates cell surface and newly synthesized CD4 within the infected cell. The presence of CD4 within an infected cell inhibits the release of virions and interferes with gp120 incorporation into the virion, which suggests that HIV would not be able to generate infectious virions in the presence of CD4 (97, 98). Downregulation of CD4 may also prevent activation of infected T-helper cells via the MHC class II antigen presentation pathway and thus help the virus evade immune detection. HIV-Nef interacts with CD4 at the cell surface and induces a 5- to 10-fold increase in CD4 endocytosis, followed by transport of CD4 to the lysosomes (58, 65, 91). Deletion mutants of Nef that lack the myristoylation site or the 10 aminoterminal residues are inactive. At the cell surface, Nef interacts with CD4 through the dileucine motif within 20-aa proximal to the transmembrane region (residues 397–415) of CD4 (99, 100). Nef from HIV-1 and -2 and simian immunodeficiency virus (SIV) selectively interact with different residues within the CD4 cytoplasmic tail (Figure 2). Transferring the 20 proximal amino acids of CD4 to CD8 results in rapid endocytosis of the chimeric CD8 protein (101, 102). The tight interaction between Nef and CD4 ensures that the CD4 will be endocytosed efficiently. Rapid endocytosis of CD4 by Nef relies on Nef’s ability to interact with the endocytic machinery (58, 91), in particular, the AP complexes. Nef colocalizes with clathrin and the b subunit of the AP-2 complex in Jurkat T cells expressed alone or together with CD4 (103). HIV-1 and -2 and SIV Nef interact with the l chains of the AP-1 and AP-2 complexes, but not the a, b, or c subunits (59, 104). The regions of Nef that interact with the AP complex and are required for CD4 downregulation are known (Figure 3) (59, 62–64, 82, 103–108). These regions supposedly lack a defined secondary structure and contain a dileucine endocytosis motif for HIV-1 Nef, and a tyrosine-based (YXXf) motif within HIV-2 and SIV Nef. The presence of these endocytic motifs within Nef may explain endocytosis of the Nef/CD4 complex (106, 107, 109, 110). The Tyr-based endocytosis motif within SIV Nef probably functions differently (108). The SIV Nef-CD4 interaction could expose the endocytosis motif within CD4 itself and thus induce rapid endocytosis. The rapid endocytosis of CD4 by Nef may not be as simple as linking CD4 to the endocytic machinery, because Nef can associate with auxiliary cellular proteins involved in endocytosis. Nef interacts with a subunit of the vacuolar ATPase (Nef-binding protein 1) involved in endocytosis of CD4 (vide supra) (82). The significance of the interaction between Nef and Nef-binding protein 1 for endocytosis is not clear, but suggests the involvement of additional cellular factors in the rapid endocytosis of CD4. After endocytosis, CD4 is targeted to the lysosomes. Nef associates with early endosomes through the C-terminal portion of the bCOP protein (aa 476–925) of the COPI complex (111, 112). In addition to the dileucine endocytosis signal,


875


Nef possesses a diacidic motif thought to be involved in targeting of internalized CD4 to lysosomes. Note that poxviruses also target cell surface CD4 for lysosomal degradation. After myxoma infection of a T cell, the cell surface level of CD4 decreases, and CD4 is directed to the lysosome (113). In human herpes virus-7 (HHV-7)-infected cells, there is also a downregulation of CD4 (114), but in either case the molecular players remain to be identified. Thus, the CD4 coreceptor is targeted by a number of unrelated viruses. Given that MHC class I products and coreceptors are targets for virus-induced downregulation, other coreceptors (e.g. CD8) and even costimulatory molecules may also be attractive targets for such devious behavior. In fact, adenovirus encodes the receptor internalization and degradation (RID) complex (see section on ‘‘Viral Interference with Apoptosis’’) involved in internalization of Fas, presumably to avoid engagement of the apoptotic cascade.

Vpu Induced Degradation of Newly Synthesized CD4 Selective pressure to eliminate CD4 from within an HIV-infected cell induces the degradation of newly synthesized CD4. The HIV gene product Vpu, together with gp160, induces the proteasomal degradation of CD4 in the cytosol (58). Vpu is found in HIV-1 and SIVcpz strains, but is absent from HIV-2 and other SIV strains (115–117). Vpu is an 81-aa type I membrane polypeptide; its aminoterminal hydrophobic domain of 27 aa anchors Vpu to the membrane, whereas the charged cytoplasmic domain is formed by two amphipathic a-helices (118, 119). Within the connecting loop, residues 52 and 56 must be phosphorylated for induction of CD4 degradation (120, 121). Vpu interacts with aa 411–419 of the cytoplasmic tail of CD4 (122). In addition, the carboxyl-terminal end of Vpu and the integrity of the overall structure of Vpu are essential for interaction with and for targeting CD4 for destruction (122, 123). The degradation of CD4 is inhibited by replacement of the lysines in its cytoplasmic tail, which eliminates potential ubiquitin-conjugating sites (124). A functional ubiquitin-conjugating system is required for CD4 degradation. Vpu may connect CD4 to the degradation machinery by interaction with a b-transducin repeat-containing protein (h-bTrCP), a 579-aa cytosolic protein whose primary sequence contains an F box and seven WD repeats (125, 126). The F box motif can interact with Skp1p and cullin to form the Skp1p-cullin-F box complex responsible for the ubiquitination of certain substrates targeted for proteasomal destruction (127). The Vpu/h-bTrCP interaction may therefore induce the ubiquitination of CD4 and target it for destruction. However, it is not clear at which step Vpu recruits h-bTrCP to the ER membrane—before removal of CD4 from the ER or after CD4 has been dislocated into the cytosol. These observations raise the important question of whether Skp1p-cullin-F box-type complexes more generally control dislocation and subsequent degradation of membrane proteins from the ER.

876

TORTORELLA ET AL

VIRAL INHIBITION OF NATURAL KILLER CELL LYSIS


Natural Killer Cell Lysis NK cells are significant effector components of the innate immune system, which aid in the initial defense against viral infection both via direct cellular cytotoxicity and by the production of inflammatory cytokines. In mice, NK cell-mediated cytotoxicity and IFN-c production occur early in the course of infection by respiratory syncytial virus, and the presence of NK cells is associated with the subsequent influx of CD8` T cells (128). NK cells also play an important role in the control of MCMV infection, as shown by studies with SCID mice and NK celldefective beige mice (129–132), in conjunction with NK cell depletion and reconstitution experiments in nonmutant murine models (133, 134). Further, organ-specific resistance of some murine strains to infection by MCMV is associated with a non-MHC-encoded gene, Cmv-1, which maps to the NK complex of murine chromosome 6. The effect of Cmv-1 is mediated by NK cells (135). In humans, the specific congenital lack of NK cells is associated with severe herpesvirus infections (136), whereas low absolute numbers and percentages of NK cells correlate with more rapid progression toward AIDS in HIV-seropositive patients (137). The missing-self hypothesis proposes that NK cells recognize the absence of cell surface-expressed self-MHC class I product as a signal for attack and destruction of a target cell (138). This recognition is mediated by the presence of both activating and inhibitory receptors on the surface of NK cells. Will NK cells attack MHC-deficient targets that arise as a consequence of viral infection? Having cultivated an effective strategy for eluding CD8` T-cell–mediated cytotoxicity, a rapidly evolving immunoevasive virus must also develop a strategy to evade detection and destruction by NK cells. Some pathogenic viruses might escape detection by NK cells via the production of surface-expressed MHC class I homologs.

Viral MHC Class I Homologs The human poxvirus, molluscum contagiosum (MCV) contains a gene encoding an MHC class I homolog that is unlikely to be able to bind antigenic peptide (139). Herpesviruses, including rat CMV (140), MCMV (141), and HCMV (142), respectively, also encode MHC class I homologs (143). MCMV contains a class I gene, m144, the deletion of which results in a severe replication deficit during early stages (days 2–6) of in vivo infection. No similar m144-associated replication deficit was observed in vitro by using infected fibroblasts. NK cells are responsible for the replication-defective phenotype of the m144-deletion mutant virus. m144 is a 383-aa membrane protein with 25% homology to murine MHC class I products (141). m144 binds b2m; but the structure of the a2 domain of m144 suggests an inability to bind peptide. The known murine NK cell C-lectin–



877

type inhibitory receptors in the Ly49 family are unlikely to be important in the regulation of MCMV infection (144, 145), and an alternate receptor for m144 is hypothesized. NK cell-mediated lysis of the class I-deficient RMA-S tumor cell line is blocked when transfected with m144. In vivo, the presence of m144 was associated with a decrease in the accumulation of leukocytes and NK cells at the point of tumor challenge. NK cells recruited to those sites were less active when m144 was present in the tumor inoculum (146). The functional significance of the MHC class I homolog encoded by human cytomegalovirus, UL-18, remains controversial. UL-18 is a 348-residue type I transmembrane protein with approximately 25% aa sequence homology to the human MHC class I heavy chain and only 18% homology to m144 (147). Like MHC class I and m144, UL-18 binds b2m (148). Class I downregulation in HCMV-infected cells was originally proposed to be caused by sequestration of b2m by UL-18. However, HCMV-infected cells retain substantial levels of free b2m (149). Unlike m144, UL-18 does bind endogenous peptides (150). Transfection of UL-18 into a class I-deficient cell line inhibited lysis of the UL-18– expressing cells by two distinct NK cell clones (151). This UL-18–associated protection from NK-mediated lysis was abolished by the addition of antibodies against the C-type lectin CD94, a component of some human and murine NK inhibitory receptors (152–155). Because HCMV-infected cells are MHC class I deficient, expression of UL-18 in HCMV-infected cells might thus protect cells from NK cell-mediated recognition and attack. The extracellular domain of UL-18, when fused to the Fc region of human IgG1, interacts with a novel immunoglobulin superfamily receptor, leukocyte immunoglobulin-like receptor -1 (156) and may be the true UL-18 counter structure. Leukocyte immunoglobulin-like receptor is highly expressed on the cell surface of B lymphoblastoid and monocytic cell lines, as well as some subsets of NK cells (157). Expression of leukocyte immunoglobulin-like receptor-1 on NK cells partially overlaps with expression of both human immunoglobulin-like and C-lectin-type (CD94/NKG2A) NK cell inhibitory receptors; however, binding of the UL-18–Fc fusion protein to a variety of NK cell subsets revealed selectivity for leukocyte immunoglobulin-like receptor` cells, but not killer cell inhibitory receptor` or CD94/NKG2A` cells (157). The hypothetical involvement of UL18 in downregulating NK cell lysis of HCMV-infected cells is further questioned by reports with UL-18 knockout HCMV. Expression of UL-18 is in fact associated with enhanced NK cell destruction of infected target cells. Killer cell inhibitory receptors and CD94/NKG2A play little role in NK lysis of CMV-infected fibroblasts (157). Similarly, fibroblasts infected with UL-18-mutant HCMV display a downregulation of MHC class I similar to that seen in wild-type HCMV-infected cells (158). Finally, NK cell lysis of HCMV-infected fibroblasts does not correlate with MHC class I downregulation, but rather with the cell surface expression of the lymphocyte function-associated antigen-3 (159). To date, in vivo expression of UL-18 in HCMV-infected cells remains to be demonstrated. The possibility


878

TORTORELLA ET AL

that UL-18 may compete intracellularly with components of class I assembly and peptide loading likewise deserves to be explored. HLA-A and -B locus products function primarily in presentation of endogenous peptide to cytotoxic T lymphocytes, whereas HLA-C, -E, and-G locus products may function primarily in the inhibition of NK cell-mediated cellular lysis. If this hypothesis is correct, selective allele-specific MHC class I downregulation by the immunoevasive virus could protect the infected cell from both CTLmediated and NK cell-mediated lysis. In other words, if the virus could selectively downregulate HLA-A and -B, but leave HLA-C, -E, and/or -G unaffected, neither CTLs nor NK cells would be alerted to an infection-associated change in surface expression of MHC class I. Selective downregulation of HLA locus products by HIV Nef may allow for continued expression of those alleles that protect the infected cell from NK cell lysis (59–61). Viral immunoevasion might indeed be allele specific (160). E3/19K, an adenoviral protein, differentially interacts with murine MHC class I products as was subsequently confirmed for human class I heavy chains as well (41; see section on ‘‘Inhibition of MHC class I Surface Expression’’ above). Whereas all studied murine class I products were degraded in the presence of US11, a more limited repertoire was attacked by US2. With vaccinia-driven expression of US2 and US11 either in a placentally derived cell line or in xenogeneic cells stably transfected with human MHC class I alleles, HLA-C and HLA-G were resistant to MHC class I degradation. HLA-E appears likewise resistant to degradation associated with both US2 and US11 (D Schust, H Ploegh, unpublished data). It has been suggested that HSV is able to completely disarm immune-mediated clearance (140), because both lymphokine-activated killer (161) and CTL (162) effector functions can be blocked in the presence of HSV infection. Because HLA-C and HLA-G require TAP-dependent peptides for stable expression at the cell surface, they are retained in the ER of HSV-infected cells (163). To date, the effects of HSV infection on surface expression of HLA-E have not been described. One would predict that HLA-E would similarly fall prey to the downregulation associated with HSV-encoded ICP47, as HLA-E binds TAP-dependent, MHC class I heavy-chain signal sequence-derived peptides. This global downregulation of MHC class I products in the presence of HSV infection, however, would alert NK cells to attack and destroy the infected MHC class I-null cell. No HSV-encoded MHC class I homolog has been described, whereas MCMV and HCMV may rely on it. We do not have an explanation for this dichotomy.

VIRAL INTERFERENCE WITH APOPTOSIS Replicating viruses may stimulate suicide of the host cell directly or provoke recognition by cytolytic T cells and NK cells. These immune effector cells induce apoptosis by secretion of cytotoxic cytokines such as tumor necrosis factors (TNFs) and by processes requiring direct cell-cell contact, such as targeted release


879

of perforin and granzyme proteins or delivery of FasL to Fas on the target cell (164). Premature cell death would limit the time available for production of new virions and interrupt cycles of latency and reactivation used by persistent viruses. As with antigen presentation, viral gene products can interfere at multiple points in the orderly execution of apoptosis. Knowledge of the steps blocked by viruses highlights pathways used by the immune system to purge these viruses from the host. The control points in the initiation and effector phase of apoptosis as induced by antiviral immune defenses, as well as examples of viral molecules that interfere with these steps, are discussed below. Several recent reviews provide further references (165–170). See Table 2 and Figure 4 for a summary of known viral genes that interfere with apoptosis.


Death Receptors Cytotoxic T and NK cells can induce apoptosis by secretion of cytotoxic cytokines such as TNFs or by delivery of FasL to infected cells. TNFs and FasL interactions with TNF receptor (TNFR) family molecules initiate a death (or activation) signal when they oligomerize in the presence of their ligand (reviewed in 171, 172). A number of pox viruses encode proteins that manipulate TNF signaling. Two independent modes of death inhibition have been delineated for a TNFR homolog encoded by the rabbit myxomavirus, a member of the poxvirus family. A secreted form of M-T2 binds and blocks the action of extracellular TNF-a, whereas an intracellular version protects infected T cells from apoptosis by an as yet uncharacterized mechanism (173). Cowpox virus encodes three distinct TNFR homologs, CrmB, CrmC, and CrmD, each with differential abilities to neutralize cytotoxic cytokines, including lymphotoxin-a and TNF (174–177). A subset of vaccinia virus strains produces functional TNFR homologs which, interestingly, are found in both secreted and cell surface-associated forms (177). In addition, HCMV encodes a TNFR homolog, UL-144, which appears to be retained intracellularly, like M-T2 of myxoma, but its function is unclear (178). Adenovirus encodes a set of proteins (E3–10.4K, -14.5K, or RID complex) that induce internalization and lysosomal degradation of cell surface Fas and another TNFR/nerve growth factor receptor family molecule, epidermal growth factor receptor (179–181). RID proteins form a multimeric, membrane-associated complex that contains endosomal targeting motifs and specifically triggers endocytosis of the host molecules (180, 182, 183). RID may also induce internalization of other TNFR subtypes (169). Perforin-competent CTLs lyse cells infected with wild-type and E3-deficient adenovirus with equal efficiency. Perforin-deficient CTLs are unable to lyse adenovirus-infected cells but can lyse cells infected with a strain deficient for E3–10.4K/14.5K (181). Such specific inhibition of Fasdependent killing must confer some advantage to the virus. Genetic studies indicate that Fas is critical in maintaining homeostasis of lymphocyte populations, so perhaps this perturbation of Fas activity is particularly advantageous in cell types in which adenovirus can establish persistent infections. Separate anti-TNF activ-

880

TORTORELLA ET AL

TABLE 2 Viral antiapoptotic genes Cellular target or homolog Fas/TNF

Virus Myxoma

Gene MT-2

Description TNFR homolog, secreted form blocks

Reference(s) (173)


TNF-a cytolysis, intracellular form protects T cells from apoptosis Adenovirus

E3-10.4/ 14.5K

Multimerec complex that forces internalization and lysosomal destruction of Fas

(179–181)

Cowpox

crmB

Neutralizes TNF and LTa

(176)

Cowpox

crmC

Neutralizes TNF but not LTa

(174)

Cowpox

crmD

Secreted as disulfide-linked complexes, blocks LTa and TNF activity in vitro

(175)

HCMV

UL144

Retained intracellularly, function (187) unclear (may not prevent apoptosis at all)

K13

vFLIPs contain 2 DEDs and prevent activation of caspases by death receptors

HHV8 vFLIPS (DEDcontaining proteins)

Caspases

(188)

HVS

ORF71

(188)

EHV-2

E8

(188,189)

MCV

MC159, MC160

(188,189)

BHV4

BORFE2

(188,368)

Cowpox

crmA/SPI-2

Vaccinia

SPI-2/B13R2 Serpin family, highly (194) homologous to cowpox CrmA

Baculoviruses

p35

Inhibits multiple caspases

(197)

Baculoviruses

IAP

Inhibits multiple caspases

(197, 198)

AFSV

A224L or 5HL

IAP homolog, may only function in arthopod vector

(201, 369)

Serpin family, inhibits caspase8, caspase-l, and granzyme B

(192)

(continued )

881

IMMUNE EVASION BY VIRUSES TABLE 2 (continued) Viral antiapoptotic genes Cellular target or homolog

Gene

Description

Reference(s)

Adenovirus

14.7K

(202, 203) Interacts with caspase-8, preventing activation of downstream caspases; unlike vFLIPs, contains no DED, can be inhibited by host FIP-2

EBV

BHRF1

Bcl-2 homolog

(211)

EBV

BALF1

Bcl-2 homolog

(216)

EBV

LMP-1

Up-regulates Bcl-2 and other (217, 218) cell survival proteins, also mimics CD40/TNFR signaling

HHV8

ksbcl-2 (ORF16)

Bcl-2 homolog

(208, 213)

AFSV

A179L

Bcl-2 homolog

(209, 370)

HVS

ORF16

Bcl-2 homolog

(212)

MHV

ORFM11

Bcl-2 homolog

(196)

Adenovirus

E1B 19K

Bcl-2 homolog

(223, 224)

Oxidative stress

MCV

MC066L

Inhibits UV and peroxideinduced apoptosis, homologous to human glutathione peroxidase

(227)

Cell cycle

HPV

E6

Targets p53 for degradation

(228)

Adenovirus

E1B 55K

Binds and inactivates p53

(371)

Adenovirus

E4 orf6

Binds and inactivates p53

(372)

SV40

Large T antigen

Binds and inactivates p53 and pRb, may also promote apoptosis

(168, 373, 374)

Transcription factor which blocks apoptosis by TNFa, ceramide, UV irradiation, serum withdrawal

(375)

Bcl-2


Virus

Transcription Marek’s disease MEQ virus

Unknown

HCMV

IE1, IE2

Inhibit TNFa- but not UVinduced apoptosis

(230)

HSV

Us3

Serine/threonine kinase

(229)

Myxoma virus

M-T2

ER-retained protein with RDEL retention sequence

(231)

TORTORELLA ET AL


882

Figure 4 Viral interference with apoptosis. Schematic diagram of apoptosis as induced by several immune effector components and viral molecules that interfere with these pathways. Note that many additional host proteins involved in apoptosis have been omitted for clarity.


883

ities have been documented for RID as well. RID can prevent TNF-induced translocation of cytosolic phospholipase A2, which may dampen the release of inflammatory mediators and suppress TNF-induced cell death (184–186). Adenovirus uses a number of other proteins to inhibit TNF-induced death, including E1B-19K and -14.7K (see below).


Death Effector Domain-Containing Proteins The cytosolic domains of oligomerized/trimerized TNFRs assemble with the adaptor molecules Fas-associated and TNFR-associated death domains (FADDs and TRADDs, respectively) to transduce activation signals to the initiating caspases required to trigger apoptosis. The FADD/caspase contacts rely on interactions between homologous protein interaction motifs present in each, referred to as death-effector domains (DEDs). Several herpesviruses (HHV-8, bovine herpesvirus-4, herpesvirus saimiri (HVS), and equine herpesvirus-2) and a poxvirus (MCV) encode molecules that directly interfere with the FADD-caspase interaction: viral-FLICE inhibitory proteins (vFLIPs), which contain two DEDs (187–190). The DEDs in the vFLIPs bind FADD and/or caspases-8 and -10 and prevent proper recruitment and activation of these initiator caspases upon ligation of Fas, TNFR1, TRAIL, and TRAMP. One cellular FLIP homolog resembles vFLIPs structurally, whereas a long variant contains a proteolytically inactive caspase-like domain in addition to two DEDS (187). These cellular homologs are certain to be critical to the sensitivity of normal cells to apoptosis.

Caspase Inhibitors Caspase activation is a critical event in the induction and execution of the apoptotic program (191). Caspases comprise a family of aspartate-specific cysteine proteases that orchestrate the apoptotic program by proteolytically activating other caspases and other cellular targets to enact the final stages of cell death. Fas and TNFR ligation relay death signals through FADDs (see above) to upstream or initiator caspases (caspase-8 and -10). Downstream caspases (caspases-3, -6, and -7) are stimulated by a variety of conditions and are active during the more terminal stages of cell death. Although the specific intracellular events that govern activation of each caspase are unclear and may vary between cell types, most apoptotic stimuli converge at the level of caspase activation. Thus it is not surprising that these enzymes are preferred targets of viral inhibitors. One well-studied caspase inhibitor is the cytokine response modifier A (CrmA) of cowpox virus (192). Despite its homology to the serpin family of serine protease inhibitors, CrmA blocks caspase activity. Specifically, CrmA is a potent inhibitor of caspase-1 [also known as IL-1b-converting enzyme (ICE)] and caspase-8 (also known as ‘‘FLICE’’). Inhibition of proteolytic maturation of IL-1b by CrmA greatly reduces the inflammatory response triggered by this cytokine. Blockade of caspase-8 by CrmA successfully stalls apoptosis in response to many apoptotic stimuli, including TNF, FasL, and—less potently—granzyme B. Rab-


884

TORTORELLA ET AL

bitpox virus encodes a CrmA-like protein that can inhibit caspase-1 but does not block apoptosis as potently as CrmA, consistent with the role of caspase-1 in promoting IL-1 maturation rather than its involvement in cell death (193). Vaccinia virus and murine herpesvirus 68 also have CrmA homologs (194–196). Two distinct families of caspase-inhibitory proteins have been identified in baculoviruses, IAP and p35. The IAP proteins have led to the discovery of mammalian homologs. IAPs block cell death by an incompletely characterized mechanism and share no homology with p35 proteins (197, 198). Human homologs of inhibitors of apoptosis (IAPs) inhibit caspases-3, -7, -9 and/ or bind TNFRassociated factors. Overexpression of the IAPs can inhibit apoptosis induced by a variety of stimuli (199, 200). African swine fever virus has an IAP homolog that may function only in the arthropod vector for this virus (201). IAPs generally contain a baculoviral IAP repeat domain coupled with RING finger domains. Baculoviral IAP repeat domains are characterized by a conserved cysteine-rich motif and are conserved from yeasts to humans. The p35 protein from Autographa californica polyhedrosis virus inhibits the insect Sf-1 caspase as well as multiple mammalian and nematode caspases (197). Strains of virus lacking the p35 gene cause premature apoptosis of infected cells, resulting in reduced virus production. Mammalian homologs to baculoviral p35 proteins have not been discovered so far, but the ability of these proteins to block death in cells from divergent species emphasizes the conservation of apoptotic pathways at the level of caspases. Adenovirus encodes a 14.7-kDa protein (14.7K) that inhibits caspase-8 (202). Unlike vFLIPs, 14.7K does not inhibit the caspase with a DED. 14.7K also interacts with a novel host protein, FIP-2 (14.7K-interacting protein) which may regulate a signaling pathway downstream of Fas and TNFRs (203). FIP-2 inhibits the protective effects of 14.7K on death induced by TNFR through receptor interacting protein (RIP). Two other 14.7K-interacting proteins have been identified, FIP-1 and FIP-3. FIP-1 is a low-molecular-weight GTP-binding protein with homology to metalloproteinases (204), and FIP-3 inhibits nuclear factor-jB (NFjB) activity and activates apoptosis (205). In addition, 14.7K together with E3 10.4/14.5K blocks phospholipase A2 translocation in response to TNFR ligation (see above).

Bcl-2 Homologs The Bcl-2 family of proteins has emerged as crucial regulators of programmed cell death (206, 207). Precise placement of these proteins in the cascade of events that lead to apoptosis has been complex. Their ability to alter susceptibility to apoptosis varies between cell types and with apoptotic stimuli. Bcl-2 proteins regulate ion permeability and membrane potential of mitochondria, and they also inhibit the release of cytochrome c from the mitochondria, an event that powerfully activates effector caspases. Localization within the cytosol or mitochondrial



885

membrane and the dimerization state of Bcl-2 proteins with other family members affect their ability to function as regulators of apoptosis. Numerous viruses modulate activity of Bcl proteins in the cell, either by use of a virally encoded homolog or by alteration of host Bcl levels. As with caspase inhibitors, these molecules may protect infected cells from apoptosis induced by multiple pathways. EBV, HHV-8, African swine fever virus, HVS, murine herpesvirus-68, alcelaphine herpesvirus-1, and bovine herpesvirus-4 contain Bcl family homologs (196, 208–214; Table 2). Homology between host Bcls and viral homologs is concentrated in Bcl-2 homology (BH) domains, which comprise putative pore-forming or heterodimerization domains. Viral Bcl-2 homologs generally contain highly conserved BH-1 and BH-2 domains but lack conservation in BH-3 domains, which mediate contacts between pro- and antiapoptotic Bcls. Expression of many of these Bcl homologs is detected only during lytic replication of the viruses. EBV encodes two Bcl homologs, BHRF-1 and BALF-1 (211, 215, 216) . A third protein from EBV, LMP-1, upregulates the expression of cellular Bcl-2 as well as other proteins that promote cell survival (217). It also interacts with components of TNFR signaling, namely TNFR-associated factors and TRADD, and activates NF-jB and AP-1 transcription factors (reviewed in 218, 219). Therefore LMP-1 clearly has multiple, complex functions that ensure survival of the virus. It is expressed during both latent and lytic phases of the viral life cycle and is required for lymphocyte transformation by EBV (218). Why does EBV possess two Bcl homologs in addition to LMP-1? BHRF-1 is functionally conserved in $15 distinct isolates of EBV, implying significance in vivo, although it is not required for transformation of B cells in vitro (220–222). Adenovirus E1B-19K binds and inhibits numerous proapoptotic Bcls such as Bax. (223, 224). E1B-19K also interacts with a novel transcriptional repressor Btf, which associates with antiapoptotic host Bcl proteins and can prevent transformation in vitro with adenovirus E1A (225). Lymphocytes are particularly sensitive to regulation of apoptosis by Bcls, consistent with the presence of so many Bcl-modulatory proteins in lymphotropic herpesviruses. Many of the viruses that contain Bcls also encode vFLIPs (HHV8, HVS, and bovine herpesvirus), in agreement with the observation that Bcl-2 and Fas regulate distinct apoptotic pathways in lymphocytes (226). The precise role of these gene products in replication and persistence of the virus remains to be defined. Also unclear is the degree to which such products promote oncogenicity of these viruses.

Oxidative Stress MCV encodes a glutathione peroxidase, MC066L, that uses a rare selenocysteine and protects MC066L transfectants from UV- and hydrogen-peroxide–induced apoptosis (227). Such a protein would be useful in a virus that exclusively infects epidermal cells and would constantly be exposed to UV irradiation. It is inter-

886

TORTORELLA ET AL

esting that MCV encodes a vFLIP as well and thus is susceptible to interference with several forms of apoptotic stimuli, despite the limited tissue specificity of the virus.


Cell Cycle The mechanisms by which viruses manipulate the cell cycle continue to shed light on the regulation of cell survival. Many viruses encode molecules that target the central coordinators of cell growth, such as p53 and pRb (Table 2). E6 of HPV binds p53 and targets it for ubiquitin-dependent proteolysis (228). HPV is not the only virus that directs host molecules to the ubiquitin-proteasome system. HCMV US2 and US11 target MHC class I molecules for destruction by the proteasome. Similarly, HIV Vpu targets host CD4 molecules for destruction by the proteasome. The EBV-encoded EBNA-1 product is also thought to exploit the ubiquitinproteasome system such that it is essentially refractory to proteasome-mediated proteolysis (see section on ‘‘Viral Inhibition of Major Histocompatibility Complex Class I-Restricted Antigen Presentation’’).

Additional Viral Proteins That Interfere with Apoptosis More anti-apoptotic viral proteins await characterization. The US3-encoded serine/threonine kinase from HSV-1 has been implicated in inhibition of apoptosis (229). The IE-1 and IE-2 genes of HCMV also are required to prevent apoptosis by that virus (230). The M-T4 protein from myxoma virus is required to block apoptosis in infected T cells. This protein contains a carboxy-terminal RDEL amino acid sequence that retains it in the ER of infected cells (231). The apoptotic pathways discussed above are not regulated by ER-retained proteins. Thus MT4’s mode of action may reveal yet more surprises. The above examples display both diversity and conservation of strategies by which a pathogen subverts the defenses of the host immune system. Inhibition of apoptosis may not be essential for the life cycle of all viruses. In fact, virusinduced apoptosis is a hallmark of many virus infections, HIV being a prominent example. Adenovirus encodes genes (not discussed here) that induce apoptosis in addition to the array that block it. Genome size and tissue specificity may influence the forms of apoptosis that must be blocked. The immune-subversive molecules encoded by adenovirus are notable for their relative lack of homology to host proteins, as compared with vFLIPs, for example, and for the use of multiple targets to achieve the same end.

INHIBITION OF CYTOKINE ACTION Cytokines are secreted polypeptides that coordinate inflammation, cellular activation, proliferation, differentiation, and chemotaxis (232). Among the earliest immune mediators produced upon virus infection, cytokines orchestrate the


887

induction and maintenance of innate and adaptive antiviral responses. Indeed, cytokines are directly responsible for the flulike symptoms of myalgia, fever, headache, and drowsiness that are common manifestations of many acute virus infections. Cytokine responses themselves are powerful antiviral mediators, allowing clearance of virus infection in the majority of cases. Consequently, targeting their function is of immediate value and contributes to virulence of viral pathogens. Viruses display remarkable inventiveness when it comes to diverting the potent antiviral cytokine responses to their benefit. Several cytokines are of particular importance and are frequently targeted by viruses. These include IL-1, IL-12, TNF, the type I interferons (IFN a/b), IFN-c, and a number of chemokines.


Interruption of Cytokine Production The most direct route to avoid the consequences of antiviral cytokines is to inhibit their production in the first place. Viruses have indeed devised ways to accomplish this goal, a fact that is particularly evident in the prevention of interferon synthesis. Infected cells are capable of recognizing double-stranded RNA produced during viral infection and respond by synthesis of type I IFN-a and IFN-b. Type I IFNs induce a multitiered defense system against acute viral infection; they arrest protein translation and cellular proliferation and upregulate enzymes that degrade viral RNA (232). In virus-infected cells, IFN regulatory factor-3 (IRF3), a widely expressed transcription factor, becomes phosphorylated, translocates to the nucleus, and binds to the transcription coactivator CBP/p300 to stimulate IFN a/b transcription. To circumvent this response, adenovirus E1A disrupts the IRF-3/CBP interaction (233–236). Viruses likewise disrupt chemokine synthesis to alter antivirus immune responses. For instance, HCMV inhibits transcription of the chemokine monocyte chemoattractant protein-1 upon infection of fibroblasts in vitro, although the mechanism responsible remains unknown (237). Several viruses halt maturation of cytokines. Members of the poxvirus family synthesize polypeptide inhibitors of ICE, thereby preventing secretion of mature IL-1b (vide supra) (238, 239). IL-1b is a proinflammatory cytokine that contributes to fever responses. Although viral serpins may primarily serve to block apoptosis by inhibiting other caspases (see section on ‘‘Viral Interference of Apoptosis’’), preliminary evidence suggests that a poxvirus-imposed blockade of ICE has physiological significance. Compared with infection with wild-type virus, myxoma viruses that lack the ICE inhibitor SERP-2 demonstrate accelerated inflammatory responses in rabbit infection models, suggesting a role for inhibition of ICE in the course of myxoma virus infection (239). Cowpox CrmA and vaccinia virus B13-R also inhibit ICE and may similarly alter IL-1b production (240). Rather than inhibiting the machinery required for cytokine expression directly, several viruses exploit regulatory pathways to suppress cytokine synthesis. Measles virus (MV) inhibits the production of IL-12 by monocytes/macrophages and dendritic cells. IL-12 induces development of Th1 responses, promotes IFN-c

888

TORTORELLA ET AL

production by T and NK cells, and is required for delayed-type hypersensitivity responses (241). The measles virus hemagglutinin binds specifically to the cellular complement receptor CD46; this interaction promotes viral uptake and suppresses IL-12 production (242). Monocytes/macrophages are the predominant cells infected during MV infection, and CD46 ligation appears to contribute to immunosuppression of cell-mediated responses observed after MV infection. Endogenous complement ligands similarly downregulate monocyte IL-12 production (242). In the same manner that Th1 cytokines antagonize Th2 responses and vice versa, several viruses likewise repress cytokine production by altering the Th1 or Th2 polarization of the antiviral immune response.


Interference with Receipt of Cytokine Signal Viruses come equipped with a variety of strategies to alter the interaction of host cytokines with their receptors. Perhaps taking an evolutionary cue from soluble cytokine receptors that arise from alternative splicing or proteolytic cleavage from the cell surface, there are many examples of virus-encoded, secreted cytokine receptors (13, 243), which are particularly prevalent among the large-DNA herpesviruses and poxviruses (Table 3). It appears that many of these ‘‘viroceptor” homologs of host cytokine receptors serve to compete with cellular receptors for cytokine binding (244). Multiple poxviruses interfere with TNF signaling by instructing infected cells to secrete TNFR homologs. TNF is secreted by activated macrophages and T cells and has potent antiviral effects (232). These TNFR homologs are likely to inhibit the effector functions of TNF, including induction of apoptosis (see ‘‘Inhibition of Apoptosis’’). Cowpox virus encodes two such TNFR molecules, CrmB and CrmC, which each bind TNF-a and are thought to impede interaction with cellular receptors (174, 176). Myxoma virus likewise encodes the TNFR homolog M-T2, which binds TNF-a with similar affinity to that of endogenously secreted cytokine receptors. MT-2 is a virulence factor in rabbit infection models (244). HCMV (Toledo strain) open reading frame (ORF) UL-144 encodes a TNFR homolog. UL-144 is a type I transmembrane glycoprotein expressed with early kinetics in fibroblast infection models, yet it is retained intracellularly and likely has a mechanism of action distinct from that of other viral TNFR homologs (178, 245). Viruses similarly secrete soluble IFN receptor homologs, which are generally species specific, and neutralize cytokines principally from the natural host of the virus. Several orthopoxviruses possess soluble IFN receptors that interact with IFNs from many species, consistent with the capacity of these viruses to infect a broad species range. For instance, the type I IFN receptor homolog B18-R secreted by vaccinia virus interacts with IFN-a from several species (246). B18R mutants are attenuated in mouse infection models (246). Likewise, B8-R is a vaccinia virus IFN-c receptor that can block the biological activity of cow, human, rabbit, and rat IFN-c (247–249). An orthopoxvirus ancestor with ability to infect many species must have possessed an IFN-c receptor that promoted replication

889

IMMUNE EVASION BY VIRUSES TABLE 3 Virus manipulation of cytokine signaling Cellular target or homolog TNF receptor


IL-1b

Type I IFNs

Virus

Determinant

Reference(s)

Cowpox

CrmB

Secreted, inhibits TNF-a and LTa

(174, 176)

Cowpox

CrmC

Secreted, inhibits LTa

(174, 176)

Shope fibroma virus

SFV T2

Secreted, binds and inhibits TNF

(244)

Myxoma

MT-2

Secreted, binds and inhibits TNF

(376)

HCMV

UL144

Retained intracellularly, unknown function

(178, 245)

Vaccinia, cowpox

B15R

Secreted, binds and inhibits IL-1b

(377, 378)

Vaccinia

B13R

Inhibits several caspases, including the IL-1b converting enzyme (ICE)

(195)

Cowpox

CrmA/Spi-2

Inhibits several caspases, including ICE

(379)

Vaccinia

B18R

Secreted type I IFN receptor homolog, binds and inhibits IFN-a

(246, 380)

Adenovirus

E1A

Blocks IFN-induced JAK/STAT pathway; blocks transriptional activation by ISGF3

(87)

Hepatitis B virus

Terminal protein Blocks IFN signaling

HHV-8 (KSHV) vIRF K9

PKR and RNAse L

Function

(381, 382)

Blocks transcription activation in response to IFN

(299)

HSV

c134.5

Reverses IFN-induced translation block

(300)

HHV-8

vIRF-2

May modulate expression of early inflammatory genes

(383)

EBV

EBNA-2

Down-regulates IFN-stimulated transcription

(384, 385)

Adenovirus

VAI RNA

Blocks PKR activity

(304, 386)

EBV

EBER I

Blocks PKR activity

(303, 305)

Hepatitis C virus

E2

Inhibits PKR activation in response to type I interferon

(301)

HIV

TAR RNA

Recruits cellular PKR inhibitor TRBP

(387, 388)

Vaccinia

K3L

eIF2a homolog, inhibits PKR

(302)

HSV

28-58(A)

RNA analog, inhibits RnaseL

(292)

Vaccinia, ORF virus

E3L

Sequesters double stranded RNA and prevents PKR activation

(391) (continued )

890

TORTORELLA ET AL

TABLE 3 (continued) Virus manipulation of cytokine signaling Cellular target or homolog


IFN-c

Virus

Determinant

Function

Reference(s)

Myxoma

M-T7

Secreted type II IFN R homolog, sequesters IFN-c and disrupts solid-phase chemokine gradients

(250, 251, 392)

Vaccinia

B8R

Secreted type II IFN R homolog, sequesters IFN-c

(247–249)

Tanapox

35-kDa

Secreted protein sequesters IFN-c (IL-2 and IL-5 also)

(256)

IL-12

VM

Hemagglutinin

Binds CD46, inhibits IL-12 production

(242, 393)

Chemokine receptors

Capripox virus

Q2/3L

Putative G-protein–coupled chemokine receptor

(394)

Cowpox, p35 variola, rabbit pox, vaccinia

Secreted, binds CC chemokines

(254, 255)

EBV

BARF1

Secreted, sequesters CSF-1

(395)

HCMV

US28

CC chemokine receptor, sequesters (259, 260) chemokines and initiates signaling

HCMV

US27

G-coupled protein, function unknown

(396)

HCMV

UL33

Chemokine receptor homolog, function unknown

(396)

RCMV, MCMV R33, M33

UL33 homolog, role in salivary (269, 270) gland dissemination or replication

HHV-6,7

U12

CC chemokine receptor

HHV-6,7

U51

Putative chemokine receptor

(397)

HHV-8

ORF74

Constitutively active receptor, inhibited by SDF, IP-10 and vMIP-II

(265, 266)

HVS

ECRF3

CXC-chemokine receptor

(262, 398)

Myxoma

M-T1

Secreted, sequesters CC chemokines (399, 400)

Swinepox virus

(397)

K2R

Chemokine receptor

(401)

BCRF1

IL-10 homolog, antagonizes Th1 responses

(258, 296)

Equine herpesvirus-2

EHVIL-10

IL-10 homolog, may antagonize Th1 (402) responses

Poxvirus orf virus

OV IL10

IL-10 homolog, may antagonize Th1 (403) responses

HHV-6

U83

CC or CX3C chemokine

HHV-8

vMIP-1

CCR8 agonist, Th2 chemoattractant

(286, 287)

HHV-8

vMIP-II

Th2 chemoattractant, chemokine receptor antagonist

(263, 288)

Virus-encoded EBV cytokines and chemokines

(276)

(continued )

891

IMMUNE EVASION BY VIRUSES TABLE 3 (continued) Virus manipulation of cytokine signaling


Cellular target or homolog

Virus

Determinant

Function

Reference(s)

HHV-8

vMIP-III

Function unknown

(263)

HHV-6

vIL6

Increases angiogenesis and hematopoiesis

(289)

M. contagiosum

MC148P

MIP-1b homolog, chemokine receptor antagonist

(139, 273, 274)

MCMV

MCK-1/MCK-2 Chemokine homolog, promotes (m131/129) virus dissemination

HVS

vIL-17

Possible T-cell mitogen

(404)

Vaccinia

A39R

Binds semaphorin receptor VESPR, up-regulates monocyte ICAM-1 and cytokine production

(405)

(277–279)

in multiple species, perhaps affording an evolutionary edge to these viruses (247). In contrast, the myxoma virus B8-R homolog, M-T7, exhibits species-specific inhibition of rabbit IFN-c, consistent with the narrow host range of myxoma virus (250). Studies with myxoma strains that lack M-T7 reveal a mechanism by which viral receptors can obstruct cytokine interactions with their cellular receptors: the disruption of solid-phase chemokine gradients (251). Chemokines are a proinflammatory cytokine subset that activate and regulate leukocyte trafficking. They facilitate rapid recruitment to sites of inflammation; chemokines can associate with the extracellular matrix via glycosaminoglycan-binding domains and establish gradients that guide leukocytes to infected and injured tissue (252). Chemokine binding also facilitates diapedesis by upregulation of leukocyte adhesion molecules on the recruited leukocytes (253). M-T7 is heavily secreted from myxoma virus-infected cells, exceeding 107 molecules/cell per hour, and binds to the a-helix heparin-binding regions of CC, CXC, and C chemokine subfamilies. Binding of M-T7 in turn prevents chemokine association with heparin present in the extracellular matrix and thereby diminishes the invasion of infected tissues by activated leukocytes (251). Table 3 lists additional examples of viral proteins known to scavenge chemokines. Viruses also encode novel molecules that neutralize cytokines. Such convergent evolution is exemplified by 35K (p35), a chemokine-binding protein of vaccinia and related poxviruses. Host chemokine receptors are homologous seven-membrane-spanning G-protein–coupled receptors with a membranestabilized ligand-binding site. Although p35 shows no sequence homology to chemokine receptors or other known proteins, p35 binds to CC chemokines with significantly greater affinity than cellular chemokine receptors (254, 255). Such high binding affinities may allow p35 to competitively inhibit proinflammatory activity generally across the b-chemokine family (254, 255). Similarly, tanapox


892

TORTORELLA ET AL

virus 38-kDA glycoprotein may use novel binding properties to inhibit three dissimilar host cytokines: IFN-c, IL-2, and IL-5 (256). EBV encodes a secreted protein that neutralizes human colony stimulating factor-1 (257). In contrast to murine colony stimulating factor-1, this cytokine does not cause differentiation of macrophages from their bone marrow precursors (258). Viruses can use membrane-bound cytokine receptors to interfere with target cell recognition of cytokines. In several instances, virus-encoded cytokine receptors retain signaling function and alter cell physiology in response to cytokine signals. Such signals may force the cells to create an environment that is of direct benefit to the virus, for example by blocking apoptosis (see section on ‘‘Viral Interference with Apoptosis’’) or by allowing more efficient replication. HCMV chemokine receptor homolog US28 induces second-messenger signaling, including calcium flux in response to the CC chemokines RANTES, MIP-1a/b, and MCP-1 (259). US28 depletes the CC chemokines MCP-1 and RANTES that surround virus-infected cells, at least in part via stimulation of continuous internalization (260). Such chemokine theft could prevent both leukocyte chemotaxis and activation of effector function. US28 is distantly related to CCR5 and CXCR4, and HIV can use US28 as a coreceptor for membrane fusion in lieu of host cytokine receptors in cells that express the HIV coreceptor CD4 (261). US28 may likewise allow HCMV to respond to host chemokine signals. Several c-herpesviruses, including HHV-8 (Kaposi’s sarcoma-associated herpesvirus) and HVS, encode a transmembrane chemokine receptor, ORF-74, which shows sequence homology with an IL-8 receptor, CXCR2 (262). ORF-74 activity is thought to contribute to the formation of highly vascularized tumors (263, 264). ORF-74 is constitutively active and stimulates signal transduction cascades similar to those induced by inflammatory cytokines, including both the JNK/SAPK and p38MAPK pathways (265). However, HHV-8 ORF-74 uses G proteins that are infrequently associated with other chemokine receptors (266). ORF-74 can interact with both CC and CXC chemokines, binding IL-8 with high affinity (263). IFN-c, CXC chemokine IP-10, vMIP-II (see below), and stromal cell-derived factor-1a each inhibit signaling by ORF-74 (267, 268). HCMV encodes several additional proteins with homology to chemokine receptors, including UL-33. Although little is known about UL-33 function, MCMV and rat CMV homologs R-33 and M-33, respectively, may provide insight. Infection models with strains that lack M-33 or R-33 replicate inefficiently in salivary gland epithelial cells, an important site for rat CMV and MCMV persistence and transmission. This defect suggests a role for M-33/R-33 in salivary gland cell entry or replication (269, 270). HHV-6 and HHV-7 U12 ORFs are 30% homologous to UL-33, and a homologous role might be inferred for these products (m90–92). Viruses have devised several means to inactivate host cytokine receptors directly. T-lymphotropic HHV-6 and HHV-7 downregulate surface expression of the chemokine receptor CXCR4 upon infection of CD4` T cells (271). Downregulation of CXCR4 causes diminished calcium flux and chemotaxis in response



893

to the CXCR4 ligand, stromal cell-derived factor-1 (271, 272). In contrast, several viruses secrete cytokines (virokines) that block host receptors (244). Kaposi’s sarcoma-associated herpesvirus synthesizes several chemokines, including vMIPII, a broad-spectrum chemokine receptor antagonist. vMIP-II binds with high affinity to both CC and CXC receptors and inhibits calcium mobilization upon subsequent chemokine stimulation; vMIP-II thus blocks monocyte chemotactic responses to RANTES, MIP-1a, or MIP-1b (263). Likewise, the poxvirus MCV secretes a CC chemokine homolog, MC-148P, which potently inhibits chemotaxis by neutrophils, macrophages, and lymphocytes in response to many CC and CXC chemokines. MCV infection causes small, papular skin tumors characterized by a striking absence of inflammatory cells, despite protracted MCV replication. MC148’s antagonistic activity likely contributes to this property. The amino terminus of MC-148P is truncated in comparison with that of other chemokines, and chemokine receptor activation domains map to this N-terminal domain. The MC148P antagonist activity may thus result from receptor binding without resultant triggering (273, 274). The broad-spectrum chemokine receptor binding of MC148P and vMIP-II contrasts with the binding specificities of nonviral chemokines; CC and CXC chemokines do not use the same receptors, and chemokine specificity allows selective recruitment and activation of particular cell types (275). Viruses also use host cytokine receptors to their advantage. The HHV-6 chemokine U83 binds to CC or CX3C chemokine receptors and induces a transient calcium flux. In vivo, U83 may recruit mononuclear cells to sites of viral replication, facilitating dissemination (276). Similarly, MCMV produces a CC chemokine homolog, m131/129 (277, 278). Murine infection models with wild-type MCMV and m131/129-deficient strains suggest a role for this viral chemokine in attracting inflammatory cells to infected tissue. Viruses that lack m131/129 show impaired salivary gland infection, suggesting that m131/129 may promote viral dissemination in the course of infection (277, 279). Perhaps m131/129 recruits leukocytes to infected tissue that can be subsequently infected and support further MCMV replication. In a similar fashion, HHV-8 encodes an IL-6 homolog with 24.7% amino acid identity with human IL-6, with the highest degree of sequence conservation in the receptor-binding domain (280). vIL-6 appears biologically active; Kaposi’s sarcoma spindle cells express the high-affinity IL-6 receptor, and vIL-6 may contribute to HHV-8–related disease by preventing apoptosis (280). vIL-6 may play a direct role in HHV-8–associated pathogenesis, which presumably includes Kaposi’s sarcoma, body cavity-based lymphoma, and Castleman’s disease (281–283). T helper cell subsets differentially express chemokine receptors (284, 285), and several viruses may take advantage of this divergence. The HHV-8 CC chemokine vMIP-1 interacts with CCR8, a chemokine receptor preferentially expressed by Th2 cells. Thus, HHV-8 may use vMIP-1 to selectively recruit and activate Th2 cells to areas of viral infection, biasing the Th1/Th2 balance of the antiviral immune response (286, 287). We must presume that the outcome is favorable for the virus. Likewise, the HHV-8 chemokine vMIP-II can function as


894

TORTORELLA ET AL

a selective Th2 chemoattractant via CCR8 agonist activity (288). vMIP-1 and vMIP-II are highly homologous, displaying 50% amino acid identity, and vMIPII antagonist activity towards CCR5 and CXCR3, which are frequently expressed by Th1 cells, could synergize with selective Th2 recruitment (263, 289). Cells transfected with CCR8 show chemotaxis in response to both vMIP-1 and vMIPII in vitro (286, 288). The natural CCR8 ligand I-309 similarly attracts Th2polarized T cells in vitro (290, 291). In a similar fashion, MCV encodes several secreted glycoproteins with homology to the IL-18–binding protein: MC-51L, MC-53L, and MC-54L (292, 293). IL-18–binding protein sequesters IL-18, which in turn is involved in NK cell activation and induction of Th1 responses. Recombinant IL-18–binding protein suppresses murine IFN-c responses to LPS (.90%), and NK cells from IL-18–deficient mice demonstrate impaired cytotoxic responses in vitro (293, 294). Because IL-18 induces IL-8 production, the MCV proteins may dampen influx of Th1 cells into infected sites (295). Several viruses encode homologs of IL-10, a Th2 cytokine that antagonizes Th1 responses, in part by negatively regulating IL-12 production (Table 3). The IL-10 homolog produced by EBV shows 84% sequence identity to its human counterpart (258). Whereas the human cytokine binds to the IL-10 receptor with 1000-fold-higher affinity than the EBV-derived cytokine, a second subunit of the IL-10 receptor has recently been identified, and comparison of human and viral IL-10 binding to this compound receptor awaits testing (296, 297). Interestingly, an allelic trait associated with an inclination for high IL-10 production is protective against EBV infection, whereas low IL-10–producing capacity was found to heighten susceptibility for severe EBV infection (298). Thus, much remains to be learned about the function of vIL-10.

Interference with Cytokine Effector Function Because of their intracellular lifestyle, viruses are uniquely positioned to intercept and alter signals emanating from cytokine receptors. Viruses frequently exploit this opportunity, skillfully manipulating multiple steps in cytokine response pathways. The numerous viral tactics that counteract type I IFN signaling best exemplify this strategy. Viruses can blunt responses from type I IFN receptors by downregulating receptor-proximal signaling components. Signal transduction from the type I IFN receptor is mediated by the Janus kinase/signal transducers and activators of transcription pathway, culminating in gene activation by IFN-stimulated gene factors, hetero-oligomeric complexes of signal transducers and activators of transcription, and other nuclear factors. HCMV infection diminishes cellular levels of p48 and Janus kinase-1, which disables IFN-a–stimulated gene expression (85, 86). Likewise, adenovirus E1A downregulates expression of STAT-1 (87). In contrast, HHV-8 encodes two proteins with homology to IRFs, proteins that bind conserved cis-acting IFN-stimulated response elements and regulate IFN-induced transcription. HHV-8 K9 (vIRF) contains a DNA-binding domain and appears to mimic



895

cellular regulatory factors (i.e. IRF-2) that compete for IFN-stimulated response element binding, without stimulating transcriptional activation. In this manner, vIRF represses transcriptional activation induced by IFN-a/b and IFN-c (299). Type I IFNs induce proteins with potent antiviral activity, and a range of viruses possess products that counteract such intracellular immunity (Table 3). Viruses rely on the host translational apparatus for viral replication. Hence, the interferon-induced synthesis of double-stranded RNA-dependent protein kinase (PKR), which prevents further protein synthesis by phosphorylating the a subunit of the translation initiation factor eIF2, poses a challenge that must be circumvented for further viral biosynthesis. HSV-1 c134.5 directly counteracts this mechanism, activating protein phosphatase 1a to dephosphorylate eIF2a (300). Hepatitis C virus envelope protein E2 contains a twelve amino acid stretch that is highly homologous to both the PKR autophosphorylation sequence and the eIF2a phosphorylation motif. This mimicry allows E2 to competitively inhibit PKR (301). Vaccinia virus K3L similarly inhibits eIF2a (302). PKR requires the common viral byproduct double-stranded RNA (or single-stranded RNA with extensive secondary structure) for activation, and both adenovirus and EBV encode RNAs that inhibit PKR activation, Ad VAI and EBER-1, respectively (303–305). The enzyme 2’5’-oligoadenylate synthetase [2–5(A)], a second double-stranded–RNA-dependent enzyme induced by type I IFN, polymerizes ATP into 2’-5’–linked oligoadenylates of various lengths. 2’-5’(A) activates the latent endoribonuclease, RNaseL, which degrades both viral and cellular RNAs, further inhibiting protein synthesis (303). HSV thwarts RNaseA activation by producing 2’-5’(A) inhibitory analogs (292).

VIRAL EVASION OF HUMORAL IMMUNITY Although cell-mediated immune responses are often necessary to clear virus infection, antibody and complement nonetheless contribute significantly to antiviral immunity. Humoral responses prevent viruses from establishing primary infection, particularly when previous exposure has stimulated production of protective antibody. Immunoglobulins bind virus surface structures and can block interaction with cellular receptors. Phagocytic cells express IgG Fc receptors and can subsequently clear antibody-tagged virus from the circulation. Fc receptors likewise instruct NK cells to lyse antibody-coated infected cells (antibody-dependent cellular cytotoxicity). The complement system also attacks viruses upon infection. IgM- and IgGbound viruses can be neutralized by activation of the classical complement pathway. Viruses activate the alternate and lectin complement pathways directly, where complement lyses enveloped viruses. Covalently bound complement components C3 and C4 facilitate virus clearance by cells that express complement receptors (CRs). Linking the innate with the adaptive immune system, C3d-tagged virus particles cross-link the CD19/CD21/CD81 complex with membrane Ig and

896

TORTORELLA ET AL

augment B-cell activation (306–308). Phagocytosis of opsonized antigen by professional antigen-presenting cells stimulates antigen processing and inflammatory cytokine release and promotes T-cell activation. Byproducts of complement activation attract leukocytes and further amplify the inflammatory process. Follicular dendritic cells (FDCs) bearing Fc receptors CR1 and CR2 trap opsonized virus particles and present antigen to germinal-center B lymphocytes, which promotes virus-specific antibody production (309, 310). Thus, antibody and complement coordinate immune responses to virus.


Subversion of Antibody Responses Viruses skillfully use Ig Fc receptors to avoid the effector consequences of antibody binding. Viral Fc receptors may hide antigenic structures by encoating the virus or infected cell surface with IgG. Alternatively, such receptors may scavenge exposed Fc domains after Fab binding to virus antigen. This antibody bipolar bridging would prevent subsequent Fc-dependent immune activation (311). The HSV-1 gE and gI glycoproteins form a hetero-oligomer complex that binds both monomeric and aggregated IgG with high affinity on HSV-1 virions and on infected cells. gE functions as a low-affinity Fc receptor in the absence of gI, and it binds IgG aggregates only (312). Soluble gE-gI complexes bind IgG with a 1:1 stoichiometry and target a histidine residue at the CH2-CH3 domain interface of IgG (313). Staphylococcus protein A, Streptococcus protein G, and neonatal Fc receptors also bind to the CH2-CH3 domain interface (314–316). gE/gI function may account for the relative ineffectiveness of anti-HSV Ig against HSV infection (317). In contrast, maternal antibody protects neonates from acquiring herpes during delivery (318). The basis for this discrepancy remains unknown. Pseudorabies virus (PRV) contains a homologous gE-gI complex. The HSV1, HSV-2, and PRV gE cytoplasmic tail contains two YXXL internalization motifs, activated by phosphorylation upon antibody-induced gE-gI capping in PRV-infected cells. Antibody-induced shedding of surface antigens plays an important role in Entamoeba histolytica immune evasion (319). Perhaps HSVand PRV-infected cells internalize gE-gI after bipolar bridging, likewise removing the immune complex. Indeed, PRV-infected monocytes treated with PRV-specific antibody demonstrate rapid endocytosis of antibody-antigen complexes in vitro (320). The dual roles of gE and gI complicate studies of the FccR’s role in animal infection models. Aside from their FcR function, HSV gE and gI facilitate intercellular spread. Subtle disruption of the gE domain most homologous to human FcR allows construction of mutant HSV-1 that lacks FccR activity. Such mutants maintain other gE functions. Although wild-type and mutant HSV-1 cause similar disease in murine infection models, administration of human anti-HSV-IgG before infection reduces mutant, but not wild-type, virus titer and pathology (321). In vitro, gE-gI protects HSV-1–infected cells from antibody-dependent cellular cyto-



897

toxicity, granulocyte recognition of attached anti-HSV Fc, and classical pathway complement activation (312, 322, 323). Why does FccR function appear so prevalent in the herpesvirus family? HSV1, HSV-2, CMV, and varicella-zoster virus each establishes persistent infection after resolution of the acute disease and targets various host cell reservoirs. After primary exposure, viruses establish latency, from which they may reactivate periodically. Thus, herpesviruses must replicate and spread in the context of an immunized host, where high-affinity antiviral IgG is present as a result of prior infection. Perhaps acquisition of FccR activity affords reactivated virus a greater opportunity to replicate and establish lesions at physiologically relevant sites, thereby increasing the probability of transmission to a new host. Viruses use immune complex receptors to promote spread in the lymphoid compartment. FDCs express high levels of complement receptors and FcR and retain antibody-antigen complexes in the follicles of secondary lymphoid tissue (324). For instance, villous processes of FDCs trap numerous particles of opsonized HIV without subsequent virus uptake. Immobilized virus remains highly infectious, and FDCs readily transmit neutralized HIV to CD4` T cells in coculture experiments (325). Neutralized HIV in the absence of FDCs does not infect T cells in similar experiments. Thus, antibody- and complement-opsonized viruses may take advantage of immune complex receptors to infect CD4` T cells that travel past the FDC network of germinal centers (326, 327). Opsonized HIV may also enter phagocytic cells that express Fc receptors and B cells with membrane-bound Ig specific for virus surface antigen. However, antigen-presenting cells may destroy such internalized virus in the phagolysosome. FccRII (CD32) antagonizes signals from the B-cell antigen receptor. FccRII binds immune complexes and delivers negative signals through immunoreceptor tyrosine-based inhibitory motifs in its cytosolic tail (328–331). In this fashion, FccRII fine-tunes B-cell antibody production. MV subverts this negative feedback system. The MV nucleocapsid protein binds to both FccRIIb1* and FccRIIb2 isoforms and attenuates B-cell antibody production in vitro (332). NP downregulates B cell antibody production even in the absence of BCR co-ligation, as NP inhibits the response of CD40-activated B cells (332). As discussed elsewhere in the review, MV gH ligation of CD46 suppresses IL-12 production. Nucleocapsid protein and gH function may both contribute to profound immune suppression that coincides with MV infection.

Subversion of Complement Responses The complement cascade is poised to attack viruses upon initial exposure and throughout the infectious cycle. Positive amplification steps accelerate complement deposition, and unrestrained complement activation would cause severe bystander damage to the host. A family of abundant membrane-bound and soluble control proteins protects the host from such cytotoxicity. The regulators of complement activation (RCA) dismantle the classical and alternative pathway C3 and


898

TORTORELLA ET AL

C5 convertases (CD35 and CD55), facilitate factor I cleavage of C3b and C4b (CD35 and CD46), and prohibit formation of the membrane attack complex on host cells (CD55). Many pathogens benefit from expression of complement control proteins, and viruses are no exception. Viruses possess homologous control proteins or seize host control proteins directly. Members of the orthopoxvirus, herpesvirus, and retrovirus families mimic RCA proteins. Vaccinia virus, cowpox, and variola viruses encode homologous secreted proteins that block C3 convertase assembly and accelerate decay. These short consensus repeat (SCR)-containing proteins exhibit cofactor activity for C3b and C4b cleavage by factor I (333, 334). The resultant iC3b and iC4b are unable to promote C3 convertase formation (335). Human complement is inhibited more effectively by variola protein smallpox inhibitor of complement enzymes than by vaccinia control protein. Whereas vaccinia infects several species of animal, humans are the only variola reservoir. Thus, smallpox inhibitor of complement enzymes illustrates the principle of homologous restriction (336). The cowpox control inflammation modulatory protein (IMP) reduces the inflammatory response at the site of cowpox infection in murine models. Compared with wildtype virus, mutant cowpox viruses that lack IMP cause greater swelling responses and histopathology at the site of injection (337–339). This discrepancy persists in mice that lack complement component C5. IMP may reduce inflammation by limiting the activation of complement receptors. In the absence of IMP, increased complement activation produces chemotactic peptides C3a and C5a, both of which recruit macrophages to sites of infection. Viruses would likewise be more densely opsonized in the absence of IMP, and subsequent CR ligation would further activate infiltrating mononuclear cells (339). Thus, IMP minimizes tissue damage and may support high-level poxvirus replication, enhancing the probability of virus transmission to a new host. At least five members of the herpesvirus family inhibit complement activation (Table 4). Of these, the HSV-1/2 transmemberane glycoprotein (gC-1/2) has been most extensively studied. The gC is expressed on the virion envelope and on the surface of HSV-infected cells. gC1 and gC2 protect against complement neutralization of infectious virus (340–342). Although the mechanism of gC2 function remains unclear, gC1 inhibits C3b interaction with the serum protein properdin. This accelerates decay of the alternate pathway C3 convertase, because properdin stabilizes the C3bBb complex. gC-1 also prevents interaction between C3b and C5, binding of factor H to C3, and binding of CR2 to iC3b (340, 343). The rare clinical isolates that fail to express gC-1 lose resistance to complement (344, 345). Human RCA proteins interact only with C3 cleavage products, whereas gC-1 and gC-2 also bind native C3 (340). The lymphotropic HVS encodes two proteins that limit complement activation. Along with other herpesvirus complement evasion proteins, complement control protein homolog targets the C3 convertase (346). In addition, HVS encodes a CD59 homolog. HVS CD59 displays 48% sequence identity with human CD59 and 69% identity with CD59 from the squirrel monkey Saimiri sciureus (347). A

899

IMMUNE EVASION BY VIRUSES TABLE 4 Virus mimicry of humoral control proteinsa Virus

Gene product


Inactivation of C3 and C3 convertase EBV HSV 1/2 HVS KSHV McHV68 Cowpox Vaccinia virus Variola virus Inhibition of C9 polymerization HVS Function unknown Vaccinia virus IgG Fc receptors HCMV HSV-1 HSV-2 MCMV MHV, BC, PTGV PRV VZV

Homolog

Reference(s)

Factor unknown gC-1/2 ORF-4 CCPH ORF-4 IMP VCP SPICE

(406) CR1 (340, 342) C4bp, CD46, CD55 (346, 407) C4bp, CD46, CD55 (280) CD46, CD55 (196, 408) C4bp (339) C4bp (333) C4bp (409)

HVS CD59

CD59

(347)

B5R

contains 4 SCRs

(410, 411)

Factor unknown gE-gI gE Fcr1 S peplomer gE-gI gE-gI

FccR FccR FccR FccR FccR FccR

(412) (413, 414) (415) (416) (417) (317) (418)

Ligation of FccRII (inhibitory receptor) MV NP

(332)

a

mcHV68, murine gamma herprsvirus 68; BC, Bovine coronavirus; PTGV, porcine transmissible gastroenteritis virus.

glycosylphosphatidylinositol anchor attaches HVS CD59 to the membrane, analogous to human CD59. Viruses that bud through the plasma membrane partially camouflage themselves with host cell membrane. Such viruses thereby acquire host integral membrane proteins that commonly include RCA family members, most notably CD46, CD55, and CD59. The vaccinia virus life cycle demonstrates the importance of such incorporation. Vaccinia virus produces two distinct virus configurations during infection, intracellular mature virus (IMV) and extracellular enveloped virus (EEV). Whereas cell lysis releases IMV, EEV arises by several routes, for instance when IMVs bud through the plasma membrane (49). IMV resists physical disruption, but is highly sensitive to attack by antibody and complement. EEV is fragile, although resistant to neutralization by complement. EEV derives comple-


900

TORTORELLA ET AL

ment resistance from host RCA proteins, because EEV produced in vitro by cells that do not express RCA proteins loses protection (348). CD55 imparts much of the protection (349). Complement resistance suits EEV to disseminate virus within the host, whereas IMV transmits virus between hosts (350). In addition, theft of host RCA permits vaccinia virus to infect a broad species range. Upon infection of a different species, vaccinia virus acquires homologous complement regulatory proteins during infection, quickly adapting progeny virus for survival in a new environment (348). Incorporation of host RCA increases serum resistance for many other viruses, including HIV, human T cell leukemia virus, and HCMV (351–354). Several viruses alter cellular expression of host complement control proteins. HCMV upregulates CD46 and CD55 expression on infected cells in vitro. HCMVinfected cells survive in the presence of antivirus antibody, and heightened RCA expression may account for this resistance (355). Increased CD55 expression may also benefit progeny virions, which could use additional regulatory protein to offset complement attack. HIV downregulates cell surface complement receptors after infection (356, 357). HIV infection diminishes CR1 expression on B cells and erythrocytes, perhaps by proteolytic cleavage of the receptor (358). The causal link between HIV infection and protease activation remains unknown. HIV infection also downregulates CR2 expression, perhaps by altered transcription (359). Exposure to gp120 decreases C5a receptor expression, which may impair monocyte chemotactic responses to inflammatory stimuli (360, 361). Factor H is an abundant soluble protein that supports factor I cleavage of C3b to iC3b and downregulates the activation loop of complement activation (362). HIV gp41 and gp120 recruit factor H to multiple binding sites. This secondary factor H attachment provides viruses and infected cells with a second route to minimize complement deposition (363, 364). Indeed, depletion of factor H from human serum increases complement-dependent lysis of HIV-infected cells and free virus in vitro (365). Streptococcus spp. have developed a similar serum resistance strategy, because the M protein similarly binds factor H (366).

CONCLUDING REMARKS Viruses acquire many of their genes through hijacking of cellular functions and, in doing so, necessarily avail themselves of basic cellular machinery. The selective pressure exerted by the immune system has led to the evolution of the diverse array of strategies surveyed in this review. Are such tricks limited to the virus world? Bacteria face many of the same problems common to viruses, but their synthetic autonomy places them in a separate category. Nonetheless, the highly sophisticated types of secretory apparatus used by bacteria to release or inject virulence-associated proteins into host cells illustrate the lengths to which pathogenic microbes will and must go to be successful and survive. What about cancer cells? If the immune system engages in surveillance to weed out malignant or cancerous cells at an early stage, they face the same selective pressures exerted


901

by the immune system as those experienced by viruses. This could favor the outgrowth of cells that escape these immune recognition mechanisms. Indeed, many tumors show a reduction in class I expression. The failure of many tumors to undergo apoptosis, when compared with their untransformed counterparts, could be taken as a page straight out of the playbook of some of the viruses described in this review. We finish with the suggestion that these very different mechanisms of immune evasion could be linked at least conceptually. Further studies of host-pathogens interactions could therefore have ramifications far beyond the area of infectious disease.


ACKNOWLEDGMENTS This work was supported by National Institutes of Health grant 5R37-A133456 and a grant from Boehringer Ingelheim. D.T. is an Irvington Institute for Immunological Research fellow. B.E.G. is a predoctoral fellow of the Howard Hughes Medical Institute. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITED 1. Townsend A, H Bodmer. 1989. Antigen recognition by class I-restricted T lymphocytes. Annu. Rev. Immunol. 7:601–24 2. Brodsky FM, L Lem, A Solache, EM Bennett. 1999. Human pathogen subversion of antigen presentation. Immunol. Rev. 168:199–215 3. Doherty P. 1995. Immunity to viruses. Immunologist 3:231–33 4. Guidotti LG, FV Chisari. 1996. To kill or to cure: options in host defense against viral infection. Curr. Opin. Immunol. 8:478–83 5. Heemels MT, H Ploegh. 1995. Generation, translocation, and presentation of MHC class I-restricted peptides. Annu. Rev. Biochem. 64:463–91 6. Bjorkman PJ, MA Saper, B Samraoui, WS Bennett, JL Strominger, DC Wiley. 1987. Structure of the human class I histocompatibility antigen, HLA-A2. Nature 329:506–12 7. Rock KL. 1996. A new foreign policy:

8.

9.

10.

11.

12.

MHC class I molecules monitor the outside world. Immunol. Today 17:131–37 Williams DB, TH Watts. 1995. Molecular chaperones in antigen presentation. Curr. Opin. Immunol. 7:77–84 Hughes EA, P Cresswell. 1998. The thiol oxidoreductase ERp57 is a component of the MHC class I peptide-loading complex. Curr. Biol. 8:709–12 Lindquist JA, ON Jensen, M Mann, GJ Hammerling. 1998. ER-60, a chaperone with thiol-dependent reductase activity involved in MHC class I assembly. EMBO J. 17:2186–95 Ortmann B, J Copeman, PJ Lehner, B Sadasivan, JA Herberg, AG Grandea, SR Riddell, R Tampe, T Spies, J Trowsdale, P Cresswell. 1997. A critical role for tapasin in the assembly and function of multimeric MHC class I-TAP complexes. Science 277:1306–9 Fruh K, A Gruhler, RM Krishna, GJ Schoenhals. 1999. A comparison of viral

902

13.

14.


15.

16.

17.

18.

19.

TORTORELLA ET AL immune escape strategies targeting the MHC class I assembly pathway. Immunol. Rev. 168:157–66 Spriggs MK 1996. One step ahead of the game: viral immunomodulatory molecules. Annu. Rev. Immunol. 14:101–30 Khanna R, SR Burrows, PM Steigerwald-Mullen, SA Thomson, MG Kurilla, DJ Moss. 1995. Isolation of cytotoxic T lymphocytes from healthy seropositive individuals specific for peptide epitopes from Epstein-Barr virus nuclear antigen 1: implications for viral persistence and tumor surveillance. Virology 214:633–37 Levitskaya J, A Sharipo, A Leonchiks, A Ciechanover, MG Masucci. 1997. Inhibition of ubiquitin/proteasome-dependent protein degradation by the Gly-Ala repeat domain of the Epstein-Barr virus nuclear antigen 1. Proc. Natl. Acad. Sci. USA 94:12616–21 Murray, RJ, MG Kurilla, HM Griffin, JM Brooks, M Mackett, JR Arrand, M Rowe, SR Burrows, DJ Moss, E Kieff, AB Rickinson. 1990. Human cytotoxic T-cell responses against Epstein-Barr virus nuclear antigens demonstrated by using recombinant vaccinia viruses. Proc. Natl. Acad. Sci. USA 87:2906–10 Khanna R, SR Burrows, MG Kurilla, CA Jacob, IS Misko, TB Sculley, E Kieff, DJ Moss. 1992. Localization of EpsteinBarr virus cytotoxic T cell epitopes using recombinant vaccinia: implications for vaccine development. J. Exp. Med. 176:169–76 Sharipo A, M Imreh, A Leonchiks, S Imreh, MG Masucci. 1998. A minimal glycine-alanine repeat prevents the interaction of ubiquitinated I kappaB alpha with the proteasome: a new mechanism for selective inhibition of proteolysis. Nat. Med. 4:939–44 Leonchiks A, E Liepinsh, M Barishev, A Sharipo, MG Masucci, G Otting. 1998. Random coil conformation of a Gly/Alarich insert in IkappaB alpha excludes structural stabilization as the mechanism

20.

21.

22.

23.

24.

25.

26.

27.

for protection against proteasomal degradation. FEBS Lett. 440:365–69 Gilbert MJ, SR Riddell, CR Li, PD Greenberg. 1993. Selective interference with class I major histocompatibility complex presentation of the major immediate-early protein following infection with human cytomegalovirus. J. Virol. 67:3461–69 Jahn G, BC Scholl, B Traupe, B Fleckenstein. 1987. The two major structural phosphoproteins (pp65 and pp150) of human cytomegalovirus and their antigenic properties. J. Gen. Virol. 68:1327– 37 Ruger B, S Klages, B Walla, J Albrecht, B Fleckenstein, P Tomlinson, B Barrell. 1987. Primary structure and transcription of the genes coding for the two virion phosphoproteins pp65 and pp71 of human cytomegalovirus. J. Virol. 61:446–53 Gilbert MJ, SR Riddell, B Plachter, PD Greenberg. 1996. Cytomegalovirus selectively blocks antigen processing and presentation of its immediate-early gene product. Nature 383:720–22 Klenerman P, S Rowland-Jones, S McAdam, J Edwards, S Daenke, D Lalloo, B Koppe, W Rosenberg, D Boyd, A Edwards, et al. 1994. Cytotoxic T-cell activity antagonized by naturally occurring HIV-1 Gag variants. Nature 369:403–7 Bertoletti A, A Sette, FV Chisari, A Penna, M Levrero, M De Carli, F Fiaccadori, C Ferrari. 1994. Natural variants of cytotoxic epitopes are T-cell receptor antagonists for antiviral cytotoxic T cells. Nature 369:407–10 York IA, C Roop, DW Andrews, SR Riddell, FL Graham, DC Johnson. 1994. A cytosolic herpes simplex virus protein inhibits antigen presentation to CD8` T lymphocytes. Cell 77:525–35 Hill A, P Jugovic, I York, G Russ, J Bennink, J Yewdell, H Ploegh, D Johnson. 1995. Herpes simplex virus turns off the


28.


29.

30.

31.

32.

33.

34.

35.

TAP to evade host immunity. Nature 375:411–15 Fruh K, K Ahn, H Djaballah, P Sempe, PM van Endert, R Tampe, PA Peterson, Y Yang. 1995. A viral inhibitor of peptide transporters for antigen presentation. Nature 375:415–18 Galocha B, A Hill, BC Barnett, A Dolan, A Raimondi, RF Cook, J Brunner, DJ McGeoch, HL Ploegh. 1997. The active site of ICP47, a herpes simplex virusencoded inhibitor of the major histocompatibility complex (MHC)-encoded peptide transporter associated with antigen processing (TAP), maps to the NH2terminal 35 residues. J. Exp. Med. 185:1565–72 Ahn K, TH Meyer, S Uebel, P Sempe, H Djaballah, Y Yang, PA Peterson, K Fruh, R Tampe. 1996. Molecular mechanism and species specificity of TAP inhibition by herpes simplex virus ICP47. EMBO J. 15:3247–55 Lacaille VG, MJ Androlewicz. 1998. Herpes simplex virus inhibitor ICP47 destabilizes the transporter associated with antigen processing (TAP) heterodimer. J. Biol. Chem. 273:17386–90 Jugovic P, AM Hill, R Tomazin, H Ploegh, DC Johnson. 1998. Inhibition of major histocompatibility complex class I antigen presentation in pig and primate cells by herpes simplex virus type 1 and 2 ICP47. J. Virol. 72:5076–84 Tomazin R, NE van Schoot, K Goldsmith, P Jugovic, P Sempe, K Fruh, DC Johnson. 1998. Herpes simplex virus type 2 ICP47 inhibits human TAP but not mouse TAP. J. Virol. 72:2560–63 Hengel H, T Flohr, GJ Hammerling, UH Koszinowski, F Momburg. 1996. Human cytomegalovirus inhibits peptide translocation into the endoplasmic reticulum for MHC class I assembly. J. Gen. Virol. 77:2287–96 Hengel H, JO Koopmann, T Flohr, W Muranyi, E Goulmy, GJ Hammerling, UH Koszinowski, F Momburg. 1997. A

36.

37.

38.

39.

40.

41.

42.

43.

44.

903

viral ER-resident glycoprotein inactivates the MHC-encoded peptide transporter. Immunity 6:623–32 Ahn K, A Gruhler, B Galocha, TR Jones, EJ Wiertz, HL Ploegh, PA Peterson, Y Yang, K Fruh. 1997. The ER-luminal domain of the HCMV glycoprotein US6 inhibits peptide translocation by TAP. Immunity 6:613–21 Lehner PJ, JT Karttunen, GW Wilkinson, P Cresswell. 1997. The human cytomegalovirus US6 glycoprotein inhibits transporter associated with antigen processing-dependent peptide translocation. Proc. Natl. Acad. Sci. USA 94:6904–9 Cox JH, JR Bennink, JW Yewdell. 1991. Retention of adenovirus E19 glycoprotein in the endoplasmic reticulum is essential to its ability to block antigen presentation. J. Exp. Med. 174:1629–37 Burgert HG, S Kvist. 1987. The E3/19K protein of adenovirus type 2 binds to the domains of histocompatibility antigens required for CTL recognition. EMBO J. 6:2019–26 Jefferies WA, HG Burgert. 1990. E3/19K from adenovirus 2 is an immunosubversive protein that binds to a structural motif regulating the intracellular transport of major histocompatibility complex class I proteins. J. Exp. Med. 172:1653– 64 Beier DC, JH Cox, DR Vining, P Cresswell, VH Engelhard. 1994. Association of human class I MHC alleles with the adenovirus E3/19K protein. J. Immunol. 152:3862–72 Paabo S, BM Bhat, WS Wold, PA Peterson. 1987. A short sequence in the COOH-terminus makes an adenovirus membrane glycoprotein a resident of the endoplasmic reticulum. Cell 50:311–17 Jackson MR, T Nilsson, PA Peterson. 1990. Identification of a consensus motif for retention of transmembrane proteins in the endoplasmic reticulum. EMBO J. 9:3153–62 Bennett EM, JR Bennink, JW Yewdell,

904

45.

46.


47.

48.

49.

50.

51.

TORTORELLA ET AL FM Brodsky. 1999. Cutting edge: adenovirus E19 has two mechanisms for affecting class I MHC expression. J. Immunol. 162:5049–52 Ahn K, A Angulo, P Ghazal, PA Peterson, Y Yang, K Fruh. 1996. Human cytomegalovirus inhibits antigen presentation by a sequential multistep process. Proc. Natl. Acad. Sci. USA 93:10990–95 Jones TR, EJ Wiertz, L Sun, KN Fish, JA Nelson, HL Ploegh. 1996. Human cytomegalovirus US3 impairs transport and maturation of major histocompatibility complex class I heavy chains. Proc. Natl. Acad. Sci. USA 93:11327–33 Schust DJ, D Tortorella, J Seebach, C Phan, HL Ploegh. 1998. Trophoblast class I major histocompatibility complex (MHC) products are resistant to rapid degradation imposed by the human cytomegalovirus (HCMV) gene products US2 and US11. J. Exp. Med. 188:497– 503 Wiertz EJ, D Tortorella, M Bogyo, J Yu, W Mothes, TR Jones, TA Rapoport, HL Ploegh. 1996. Sec61-mediated transfer of a membrane protein from the endoplasmic reticulum to the proteasome for destruction. Nature 384:432–38 Tortorella D, CM Story, JB Huppa, EJ Wiertz, TR Jones, I Bacik, JR Bennink, JW Yewdell, HL Ploegh. 1998. Dislocation of type I membrane proteins from the ER to the cytosol is sensitive to changes in redox potential. J. Cell Biol. 142:365–76 Suzuki T, A Seko, K Kitajima, Y Inoue, S Inoue. 1994. Purification and enzymatic properties of peptide:N-glycanase from C3H mouse-derived L-929 fibroblast cells: possible widespread occurrence of post-translational remodification of proteins by N-deglycosylation. J. Biol. Chem. 269:17611–18 Bankier AT, S Beck, R Bohni, CM Brown, R Cerny, MS Chee, CA Hutchison, T Kouzarides, JA Martignetti, E Preddie, et al. 1991. The DNA sequence

52.

53.

54.

55.

56.

57.

58.

59.

of the human cytomegalovirus genome. DNA Seq. 2:1–12 Story CM, MH Furman, HL Ploegh. 1999. The cytosolic tail of class I MHC heavy chain is required for its dislocation by the human cytomegalovirus US2 and US11 gene products. Proc. Natl. Acad. Sci. USA 96:8516–21 Shamu CE, CM Story, TA Rapoport, HL Ploegh. 1999. The pathway of US11dependent degradation of MHC class I heavy chains involves a ubiquitin-conjugated intermediate. J. Cell. Biol. 147:45–58 Plemper RK, DH Wolf. 1999. Retrograde protein translocation: eradication of secretory proteins in health and disease. Trends Biochem. Sci. 24:266–70 Ziegler H, R Thale, P Lucin, W Muranyi, T Flohr, H Hengel, H Farrell, W Rawlinson, UH Koszinowski. 1997. A mouse cytomegalovirus glycoprotein retains MHC class I complexes in the ERGIC/ cis-Golgi compartments. Immunity 6:57– 66 Reusch U, W Muranyi, P Lucin, HG Burgert, H Hengel, UH Koszinowski. 1999. A cytomegalovirus glycoprotein reroutes MHC class I complexes to lysosomes for degradation. EMBO J. 18:1081–91 Kleijnen MF, JB Huppa, P Lucin, S Mukherjee, H Farrell, AE Campbell, UH Koszinowski, AB Hill, HL Ploegh. 1997. A mouse cytomegalovirus glycoprotein, gp34, forms a complex with folded class I MHC molecules in the ER which is not retained but is transported to the cell surface. EMBO J. 16:685–94 Piguet V, O Schwartz, S Le Gall, D Trono. 1999. The downregulation of CD4 and MHC-I by primate lentiviruses: a paradigm for the modulation of cell surface receptors. Immunol. Rev. 168:51–63 Le Gall S, L Erdtmann, S Benichou, C Berlioz-Torrent, L Liu, R Benarous, JM Heard, O Schwartz. 1998. Nef interacts with the mu subunit of clathrin adaptor


60.

61.


62.

63.

64.

65.

66.

67.

complexes and reveals a cryptic sorting signal in MHC I molecules. Immunity 8:483–95 Cohen GB, RT Gandhi, DM Davis, O Mandelboim, BK Chen, JL Strominger, D Baltimore. 1999. The selective downregulation of class I major histocompatibility complex proteins by HIV-1 protects HIV-infected cells from NK cells. Immunity 10:661–71 Collins KL, BK Chen, SA Kalams, BD Walker, D Baltimore. 1998. HIV-1 Nef protein protects infected primary cells against killing by cytotoxic T lymphocytes. Nature 391:397–401 Greenberg ME, AJ Iafrate, J Skowronski. 1998. The SH3 domain-binding surface and an acidic motif in HIV-1 Nef regulate trafficking of class I MHC complexes. EMBO J. 17:2777–89 Mangasarian A, V Piguet, JK Wang, YL Chen, D Trono. 1999. Nef-induced CD4 and major histocompatibility complex class I (MHC-I) down-regulation are governed by distinct determinants: N-terminal alpha helix and proline repeat of Nef selectively regulate MHC-I trafficking. J. Virol. 73:1964–73 Riggs NL, HM Craig, MW Pandori, JC Guatelli. 1999. The dileucine-based sorting motif in HIV-1 Nef is not required for down-regulation of class I MHC. Virology 258:203–7 Marsh JW. 1999. The numerous effector functions of Nef. Arch. Biochem. Biophys. 365:192–98 Kerkau T, I Bacik, JR Bennink, JW Yewdell, T Hunig, A Schimpl, U Schubert. 1997. The human immunodeficiency virus type 1 (HIV-1) Vpu protein interferes with an early step in the biosynthesis of major histocompatibility complex (MHC) class I molecules. J. Exp. Med. 185:1295–305 Wolf PR, HL Ploegh. 1995. How MHC class II molecules acquire peptide cargo: biosynthesis and trafficking through the

68.

69.

70.

71.

72.

73.

74.

75.

905

endocytic pathway. Annu. Rev. Cell. Dev. Biol. 11:267–306 Pieters J. 1997. MHC class II restricted antigen presentation. Curr. Opin. Immunol. 9:89–96 Jonjic S, W Mutter, F Weiland, MJ Reddehase, UH Koszinowski. 1989. Siterestricted persistent cytomegalovirus infection after selective long-term depletion of CD4` T lymphocytes. J. Exp. Med. 169:1199–212 Chapman HA. 1998. Endosomal proteolysis and MHC class II function. Curr. Opin. Immunol. 10:93–102 Tomazin R, J Boname, NR Hegde, DM Lewinsohn, Y Altschuler, TR Jones, P Cresswell, JA Nelson, SR Riddell, DC Johnson. 1999. Cytomegalovirus US2 destroys two components of the MHC class II pathway, preventing recognition by CD4` T cells. Nat. Med. 5:1039–43 Spriggs MK, RJ Armitage, MR Comeau, L Strockbine, T Farrah, B Macduff, D Ulrich, MR Alderson, J Mullberg, JI Cohen. 1996. The extracellular domain of the Epstein-Barr virus BZLF2 protein binds the HLA-DR beta chain and inhibits antigen presentation. J. Virol. 70:5557–63 Lewandowski GA, D Lo, FE Bloom. 1993. Interference with major histocompatibility complex class II-restricted antigen presentation in the brain by herpes simplex virus type 1: a possible mechanism of evasion of the immune response. Proc. Natl. Acad. Sci. USA 90:2005–9 de Waal Malefyt R, J Haanen, H Spits, MG Roncarolo, A te Velde, C Figdor, K Johnson, R Kastelein, H Yssel, JE de Vries. 1991. Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigenspecific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via downregulation of class II major histocompatibility complex expression. J. Exp. Med. 174:915– 24 Koppelman B, JJ Neefjes, JE de Vries, R

906

76.

77.


78.

79.

80.

81.

82.

83.

84.

TORTORELLA ET AL de Waal Malefyt. 1997. Interleukin-10 down-regulates MHC class II alphabeta peptide complexes at the plasma membrane of monocytes by affecting arrival and recycling. Immunity 7:861–71 Jackson T. 1998. Transport vesicles: coats of many colours. Curr. Biol. 8:R609–12 Tong X, W Boll, T Kirchhausen, PM Howley. 1998. Interaction of the bovine papillomavirus E6 protein with the clathrin adaptor complex AP-1. J. Virol. 72:476–82 Andresson T, J Sparkowski, DJ Goldstein, R Schlegel. 1995. Vacuolar H(`)ATPase mutants transform cells and define a binding site for the papillomavirus E5 oncoprotein. J. Biol. Chem. 270:6830–37 Martin P, WC Vass, JT Schiller, DR Lowy, TJ Velu. 1989. The bovine papillomavirus E5 transforming protein can stimulate the transforming activity of EGF and CSF-1 receptors. Cell 59:21– 32 Straight SW, PM Hinkle, RJ Jewers, DJ McCance. 1993. The E5 oncoprotein of human papillomavirus type 16 transforms fibroblasts and effects the downregulation of the epidermal growth factor receptor in keratinocytes. J. Virol. 67:4521–32 Waters CM, KA Overholser, A Sorkin, G Carpenter. 1992. Analysis of the influences of the E5 transforming protein on kinetic parameters of epidermal growth factor binding and metabolism. J. Cell Physiol. 152:253–63 Lu X, H Yu, SH Liu, FM Brodsky, BM Peterlin. 1998. Interactions between HIV1 Nef and vacuolar ATPase facilitate the internalization of CD4. Immunity 8:647–56 Boss JM. 1997. Regulation of transcription of MHC class II genes. Curr. Opin. Immunol. 9:107–13 Rinaldo CR Jr. 1994. Modulation of major histocompatibility complex anti-

85.

86.

87.

88.

89.

90.

91.

92.

93.

gen expression by viral infection. Am. J. Pathol. 144:637–50 Miller DM, BM Rahill, JM Boss, MD Lairmore, JE Durbin, JW Waldman, DD Sedmak. 1998. Human cytomegalovirus inhibits major histocompatibility complex class II expression by disruption of the Jak/Stat pathway. J. Exp. Med. 187:675–83 Miller DM, Y Zhang, BM Rahill, WJ Waldman, DD Sedmak. 1999. Human cytomegalovirus inhibits IFN-alphastimulated antiviral and immunoregulatory responses by blocking multiple levels of IFN-alpha signal transduction. J. Immunol. 162:6107–13 Leonard GT, GC Sen. 1996. Effects of adenovirus E1A protein on interferonsignaling. Virology 224:25–33 Leonard GT, GC Sen. 1997. Restoration of interferon responses of adenovirus E1A-expressing HT1080 cell lines by overexpression of p48 protein. J. Virol. 71:5095–101 Heise MT, HW Virgin. 1995. The T-cellindependent role of gamma interferon and tumor necrosis factor alpha in macrophage activation during murine cytomegalovirus and herpes simplex virus infections. J. Virol. 69:904–9 Heise MT, M Connick, HW Virgin. 1998. Murine cytomegalovirus inhibits interferon gamma-induced antigen presentation to CD4 T cells by macrophages via regulation of expression of major histocompatibility complex class II-associated genes. J. Exp. Med. 187:1037–46 Oldridge J, M Marsh. 1998. Nef—an adaptor adaptor? Trends Cell. Biol. 8:302–5 McDougal JS, JK Nicholson, GD Cross, SP Cort, MS Kennedy, AC Mawle. 1986. Binding of the human retrovirus HTLVIII/LAV/ARV/HIV to the CD4 (T4) molecule: conformation dependence, epitope mapping, antibody inhibition, and potential for idiotypic mimicry. J. Immunol. 137:2937–44 Maddon PJ, AG Dalgleish, JS Mc-


94.

95.


96.

97.

98.

99.

100.

101.

Dougal, PR Clapham, RA Weiss, R Axel. 1986. The T4 gene encodes the AIDS virus receptor and is expressed in the immune system and the brain. Cell 47:333–48 Klatzmann D, E Champagne, S Chamaret, J Gruest, D Guetard, T Hercend, JC Gluckman, L Montagnier. 1984. Tlymphocyte T4 molecule behaves as the receptor for human retrovirus LAV. Nature 312:767–68 Dittmar MT, A McKnight, G Simmons, PR Clapham, RA Weiss, P Simmonds. 1997. HIV-1 tropism and co-receptor use. Nature 385:495–96 Connor RI, KE Sheridan, D Ceradini, S Choe, NR Landau. 1997. Change in coreceptor use coreceptor use correlates with disease progression in HIV-1–infected individuals. J. Exp. Med. 185:621–28 Lama J, A Mangasarian, D Trono. 1999. Cell-surface expression of CD4 reduces HIV-1 infectivity by blocking Env incorporation in a Nef- and Vpu-inhibitable manner. Curr. Biol. 9:622–31 Ross TM, AE Oran, BR Cullen. 1999. Inhibition of HIV-1 progeny virion release by cell-surface CD4 is relieved by expression of the viral Nef protein. Curr. Biol. 9:613–21 Aiken C, J Konner, NR Landau, ME Lenburg, D Trono. 1994. Nef induces CD4 endocytosis: requirement for a critical dileucine motif in the membrane-proximal CD4 cytoplasmic domain. Cell 76:853–64 Hua J, BR Cullen. 1997. Human immunodeficiency virus types 1 and 2 and simian immunodeficiency virus Nef use distinct but overlapping target sites for downregulation of cell surface CD4. J. Virol. 71:6742–48 Grzesiek S, SJ Stahl, PT Wingfield, A Bax. 1996. The CD4 determinant for downregulation by HIV-1 Nef directly binds to Nef: mapping of the Nef binding surface by NMR. Biochemistry 35: 10256–61

907

102. Grzesiek S, A Bax, GM Clore, AM Gronenborn, JS Hu, J Kaufman, I Palmer, SJ Stahl, PT Wingfield. 1996. The solution structure of HIV-1 Nef reveals an unexpected fold and permits delineation of the binding surface for the SH3 domain of Hck tyrosine protein kinase. Nat. Struct. Biol. 3:340–45 103. Greenberg ME, S Bronson, M Lock, M Neumann, GN Pavlakis, J Skowronski. 1997. Co-localization of HIV-1 Nef with the AP-2 adaptor protein complex correlates with Nef-induced CD4 downregulation. EMBO J. 16:6964–76 104. Piguet V, YL Chen, A Mangasarian, M Foti, JL Carpentier, D Trono. 1998. Mechanism of Nef-induced CD4 endocytosis: Nef connects CD4 with the mu chain of adaptor complexes. EMBO J. 17:2472–81 105. Iafrate AJ, S Bronson, J Skowronski. 1997. Separable functions of Nef disrupt two aspects of T cell receptor machinery: CD4 expression and CD3 signaling. EMBO J. 16:673–84 106. Bresnahan PA, W Yonemoto, S Ferrell, D Williams-Herman, R Geleziunas, WC Greene. 1998. A dileucine motif in HIV1 Nef acts as an internalization signal for CD4 downregulation and binds the AP1 clathrin adaptor. Curr. Biol. 8:1235–38 107. Greenberg M, L DeTulleo, I Rapoport, J Skowronski, T Kirchhausen. 1998. A dileucine motif in HIV-1 Nef is essential for sorting into clathrin-coated pits and for downregulation of CD4. Curr. Biol. 8:1239–42 108. Lock M, ME Greenberg, AJ Iafrate, T Swigut, J Muench, F Kirchhoff, N Shohdy, J Skowronski. 1999. Two elements target SIV Nef to the AP-2 clathrin adaptor complex, but only one is required for the induction of CD4 endocytosis. EMBO J. 18:2722–33 109. Craig HM, MW Pandori, JC Guatelli. 1998. Interaction of HIV-1 Nef with the

908

110.


111.

112.

113.

114.

115.

116.

117.

TORTORELLA ET AL cellular dileucine-based sorting pathway is required for CD4 down-regulation and optimal viral infectivity. Proc. Natl. Acad. Sci. USA 95:11229–34 Mangasarian A, M Foti, C Aiken, D Chin, JL Carpentier, D Trono. 1997. The HIV-1 Nef protein acts as a connector with sorting pathways in the Golgi and at the plasma membrane. Immunity 6:67– 77 Benichou S, M Bomsel, M Bodeus, H Durand, M Doute, F Letourneur, J Camonis, R Benarous. 1994. Physical interaction of the HIV-1 Nef protein with beta-COP, a component of non-clathrincoated vesicles essential for membrane traffic. J. Biol. Chem. 269:30073–76 Piguet V, F Gu, M Foti, N Demaurex, J Gruenberg, JL Carpentier, D Trono. 1999. Nef-induced CD4 degradation: a diacidic-based motif in Nef functions as a lysosomal targeting signal through the binding of beta-COP in endosomes. Cell 97:63–73 Barry M, SF Lee, L Boshkov, G McFadden. 1995. Myxoma virus induces extensive CD4 downregulation and dissociation of p56lck in infected rabbit CD4` T lymphocytes. J. Virol. 69: 5243–51 Secchiero P, D Gibellini, L Flamand, I Robuffo, M Marchisio, S Capitani, RC Gallo, G Zauli. 1997. Human herpesvirus 7 induces the down-regulation of CD4 antigen in lymphoid T cells without affecting p56lck levels. J. Immunol. 159:3412–23 Strebel K, T Klimkait, MA Martin. 1988. A novel gene of HIV-1, vpu, and its 16kilodalton product. Science 241:1221–23 Matsuda Z, MJ Chou, M Matsuda, JH Huang, YM Chen, R Redfield, K Mayer, M Essex, TH Lee. 1988. Human immunodeficiency virus type 1 has an additional coding sequence in the central region of the genome. Proc. Natl. Acad. Sci. USA 85:6968–72 Cohen EA, EF Terwilliger, JG Sodroski,

118.

119.

120.

121.

122.

123.

124.

125.

WA Haseltine. 1988. Identification of a protein encoded by the vpu gene of HIV1. Nature 334:532–34 Maldarelli F, MY Chen, RL Willey, K Strebel. 1993. Human immunodeficiency virus type 1 Vpu protein is an oligomeric type I integral membrane protein. J. Virol. 67:5056–61 Wray V, T Federau, P Henklein, S Klabunde, O Kunert, D Schomburg, U Schubert. 1995. Solution structure of the hydrophilic region of HIV-1 encoded virus protein U (Vpu) by CD and 1H NMR spectroscopy. Int. J. Pept. Protein Res. 45:35–43 Schubert U, K Strebel. 1994. Differential activities of the human immunodeficiency virus type 1-encoded Vpu protein are regulated by phosphorylation and occur in different cellular compartments. J. Virol. 68:2260–71 Paul M, MA Jabbar. 1997. Phosphorylation of both phosphoacceptor sites in the HIV-1 Vpu cytoplasmic domain is essential for Vpu-mediated ER degradation of CD4. Virology 232:207–16 Tiganos E, XJ Yao, J Friborg, N Daniel, EA Cohen. 1997. Putative alpha-helical structures in the human immunodeficiency virus type 1 Vpu protein and CD4 are involved in binding and degradation of the CD4 molecule. J. Virol. 71:4452– 60 Margottin F, S Benichou, H Durand, V Richard, LX Liu, E Gomas, R Benarous. 1996. Interaction between the cytoplasmic domains of HIV-1 Vpu and CD4: role of Vpu residues involved in CD4 interaction and in vitro CD4 degradation. Virology 223:381–86 Schubert U, LC Anton, I Bacik, JH Cox, S Bour, JR Bennink, Morlowski, K Strebel, JW Yewdell. 1998. CD4 glycoprotein degradation induced by human immunodeficiency virus type 1 Vpu protein requires the function of proteasomes and the ubiquitin-conjugating pathway. J. Virol. 72:2280–88 van der Voorn L, HL Ploegh. 1992. The


IMMUNE EVASION BY VIRUSES WD-40 repeat. FEBS Lett. 307:131–34 126. Margottin F, SP Bour, H Durand, L Selig, S Benichou, V Richard, D Thomas, K Strebel, R Benarous. 1998. A novel human WD protein, h-beta TrCp, that interacts with HIV-1 Vpu connects CD4 to the ER degradation pathway through an F-box motif. Mol. Cell. 1:565–74 127. Bai C, P Sen, K Hofmann, L Ma, M Goebl, JW Harper, SJ Elledge. 1996. SKP1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box. Cell 86:263–74 128. Hussell T, PJ Openshaw. 1998. Intracellular IFN-gamma expression in natural killer cells precedes lung CD8` T cell recruitment during respiratory syncytial virus infection. J. Gen. Virol.79:2593– 601 129. Bukowski JF, BA Woda, S Habu, K Okumura, RM Welsh. 1983. Natural killer cell depletion enhances virus synthesis and virus- induced hepatitis in vivo. J. Immunol. 131:1531–38 130. Shellam GR, JE Allan, JM Papadimitriou, GJ Bancroft. 1981. Increased susceptibility to cytomegalovirus infection in beige mutant mice. Proc. Natl. Acad. Sci. USA 78:5104–8 131. Welsh RM, JO Brubaker, M VargasCortes, CL O’Donnell. 1991. Natural killer (NK) cell response to virus infections in mice with severe combined immunodeficiency: the stimulation of NK cells and the NK cell-dependent control of virus infections occur independently of T and B cell function. J. Exp. Med. 173:1053–63 132. Welsh RM, CL O’Donnell, LD Shultz. 1994. Antiviral activity of NK 1.1` natural killer cells in C57BL/6 SCID mice infected with murine cytomegalovirus. Nat. Immun. 13:239–45 133. Bukowski JF, BA Woda, RM Welsh. 1984. Pathogenesis of murine cytomegalovirus infection in natural killer celldepleted mice. J. Virol. 52:119–28

909

134. Bukowski JF, JF Warner, G Dennert, RM Welsh. 1985. Adoptive transfer studies demonstrating the antiviral effect of natural killer cells in vivo. J. Exp. Med. 161:40–52 135. Scalzo AA, NA Fitzgerald, CR Wallace, AE Gibbons, YC Smart, RC Burton, GR Shellam. 1992. The effect of the Cmv-1 resistance gene, which is linked to the natural killer cell gene complex, is mediated by natural killer cells. J. Immunol. 149:581–89 136. Biron CA, KS Byron, JL Sullivan. 1989. Severe herpesvirus infections in an adolescent without natural killer cells. N. Engl. J. Med. 320:1731–35 137. Bruunsgaard H, C Pedersen, P Skinhoj, BK Pedersen. 1997. Clinical progression of HIV infection: role of NK cells. Scand. J. Immunol. 46:91–95 138. Ljunggren HG, K Karre. 1990. In search of the missing self: MHC molecule and NK cell recognition. Immunol. Today 11:237–44 139. Senkevich TG, JJ Bugert, JR Sisler, EV Koonin, G Darai, B Moss. 1996. Genome sequence of a human tumorigenic poxvirus: prediction of specific host response-evasion genes. Science 273: 813–16 140. Farrell HE, NJ Davis-Poynter. 1998. From sabotage to camouflage: viral evasion of cytotoxic T lymphocyte and natural killer cell-mediated immunity. Semin. Cell. Dev. Biol. 9:369–78 141. Rawlinson WD, HE Farrell, BG Barrell. 1996. Analysis of the complete DNA sequence of murine cytomegalovirus. J. Virol. 70:8833–49 142. Beck S, BG Barrell. 1988. Human cytomegalovirus encodes a glycoprotein homologous to MHC class-I antigens. Nature 331:269–72 143. Farrell HE, H Vally, DM Lynch, P Fleming, GR Shellam, AA Scalzo, NJ DavisPoynter. 1997. Inhibition of natural killer

910

144.

145.


146.

147.

148.

149.

150.

151.

TORTORELLA ET AL cells by a cytomegalovirus MHC class I homologue in vivo. Nature 386:510–14 Tay CH, LY Yu, V Kumar, L Mason, JR Ortaldo, RM Welsh. 1999. The role of LY49 NK cell subsets in the regulation of murine cytomegalovirus infections. J. Immunol. 162:718–26 Kubota A, S Kubota, HE Farrell, N Davis-Poynter, F Takei. 1999. Inhibition of NK cells by murine CMV-encoded class I MHC homologue m144. Cell Immunol. 191:145–51 Cretney E, MA Degli-Esposti, EH Densley, HE Farrell, NJ Davis-Poynter, MJ Smyth. 1999. m144, a murine cytomegalovirus (MCMV)-encoded major histocompatibility complex class I homologue, confers tumor resistance to natural killer cell-mediated rejection. J. Exp. Med. 190:435–44 Chapman TL, PJ Bjorkman. 1998. Characterization of a murine cytomegalovirus class I major histocompatibility complex (MHC) homolog: comparison to MHC molecules and to the human cytomegalovirus MHC homolog. J. Virol. 72:460– 66 Browne H, G Smith, S Beck, T Minson. 1990. A complex between the MHC class I homologue encoded by human cytomegalovirus and beta 2 microglobulin. Nature 347:770–72 Beersma MF, MJ Bijlmakers, HL Ploegh. 1993. Human cytomegalovirus down-regulates HLA class I expression by reducing the stability of class I H chains. J. Immunol. 151:4455–64 Fahnestock ML, JL Johnson, RM Feldman, JM Neveu, WS Lane, PJ Bjorkman. 1995. The MHC class I homolog encoded by human cytomegalovirus binds endogenous peptides. Immunity 3:583–90 Reyburn HT, O Mandelboim, M ValesGomez, DM Davis, L Pazmany, JL Strominger. 1997. The class I MHC homologue of human cytomegalovirus

152.

153.

154.

155.

156.

157.

158.

159.

inhibits attack by natural killer cells. Nature 386:514–17 Ho EL, JW Heusel, MG Brown, K Matsumoto, AA Scalzo, WM Yokoyama. 1998. Murine Nkg2d and Cd94 are clustered within the natural killer complex and are expressed independently in natural killer cells. Proc. Natl. Acad. Sci. USA 95:6320–25 Lazetic S, C Chang, JP Houchins, LL Lanier, JH Phillips. 1996. Human natural killer cell receptors involved in MHC class I recognition are disulfide-linked heterodimers of CD94 and NKG2 subunits. J. Immunol. 157:4741–45 Vance RE, DM Tanamachi, T Hanke, DH Raulet. 1997. Cloning of a mouse homolog of CD94 extends the family of Ctype lectins on murine natural killer cells. Eur. J. Immunol. 27:3236–41 Vance RE, JR Kraft, JD Altman, PE Jensen, DH Raulet. 1998. Mouse CD94/ NKG2A is a natural killer cell receptor for the nonclassical major histocompatibility complex (MHC) class I molecule Qa-1(b). J. Exp. Med. 188:1841–48 Cosman D, N Fanger, L Borges, M Kubin, W Chin, L Peterson, ML Hsu. 1997. A novel immunoglobulin superfamily receptor for cellular and viral MHC class I molecules. Immunity 7:273–82 Vitale M, R Castriconi, S Parolini, D Pende, ML Hsu, L Moretta, D Cosman, A Moretta. 1999. The leukocyte Ig-like receptor (LIR)-1 for the cytomegalovirus UL18 protein displays a broad specificity for different HLA class I alleles: analysis of LIR-1 ` NK cell clones. Int. Immunol. 11:29–35 Browne H, M Churcher, T Minson. 1992. Construction and characterization of a human cytomegalovirus mutant with the UL18 (class I homolog) gene deleted. J. Virol. 66:6784–87 Fletcher JM, HG Prentice, JE Grundy. 1998. Natural killer cell lysis of cytomegalovirus (CMV)-infected cells cor-


160.


161.

162.

163.

164.

165.

166.

167.

168.

169.

relates with virally induced changes in cell surface lymphocyte function-associated antigen-3 (LFA-3) expression and not with the CMV- induced down-regulation of cell surface class I HLA. J. Immunol. 161:2365–74 Tanaka Y, SS Tevethia. 1988. Differential effect of adenovirus 2 E3/19K glycoprotein on the expression of H-2Kb and H2Db class I antigens and H-2Kb- and H-2Db- restricted SV40-specific CTLmediated lysis. Virology 165:357–66 York IA, DC Johnson. 1993. Direct contact with herpes simplex virus-infected cells results in inhibition of lymphokineactivated killer cells because of cell-tocell spread of virus. J. Infect. Dis. 168:1127–32 Posavad CM, JJ Newton, KL Rosenthal. 1993. Inhibition of human CTL-mediated lysis by fibroblasts infected with herpes simplex virus. J. Immunol. 151:4865–73 Schust DJ, AB Hill, HL Ploegh. 1996. Herpes simplex virus blocks intracellular transport of HLA-G in placentally derived human cells. J. Immunol. 157:3375–80 Shresta S, CT Pham, DA Thomas, TA Graubert, TJ Ley. 1998. How do cytotoxic lymphocytes kill their targets? Curr. Opin. Immunol. 10:581–87 Cuff S, J Ruby. 1996. Evasion of apoptosis by DNA viruses. Immunol. Cell. Biol. 74:527–37 Barry M, G McFadden. 1998. Apoptosis regulators from DNA viruses. Curr. Opin. Immunol. 10:422–30 Tschopp J, M Thome, K Hofmann, E Meinl. 1998. The fight of viruses against apoptosis. Curr. Opin. Genet. Dev. 8:82– 87 Teodoro JG, PE Branton. 1997. Regulation of apoptosis by viral gene products. J. Virol. 71:1739–46 Wold WS, K Doronin, K Toth, M Kuppuswamy, DL Lichtenstein, AE Tollefson. 1999. Immune responses to

170.

171.

172. 173.

174.

175.

176.

177.

178.

911

adenoviruses: viral evasion mechanisms and their implications for the clinic. Curr. Opin. Immunol. 11:380–86 Meinl E, H Fickenscher, M Thome, J Tschopp, B Fleckenstein. 1998. Antiapoptotic strategies of lymphotropic viruses. Immunol. Today 19:474–79 Ashkenazi A, VM Dixit. 1998. Death receptors: signaling and modulation. Science 281:1305–8 Nagata S. 1997. Apoptosis by death factor. Cell 88:355–65 Schreiber M, L Sedger, G McFadden. 1997. Distinct domains of M-T2, the myxoma virus tumor necrosis factor (TNF) receptor homolog, mediate extracellular TNF binding and intracellular apoptosis inhibition. J. Virol. 71:2171– 81 Smith CA, FQ Hu, TD Smith, CL Richards, P Smolak, RG Goodwin, DJ Pickup. 1996. Cowpox virus genome encodes a second soluble homologue of cellular TNF receptors, distinct from CrmB, that binds TNF but not LT alpha. Virology 223:132–47 Loparev VN, JM Parsons, JC Knight, JF Panus, CA Ray, RM Buller, DJ Pickup, JJ Esposito. 1998. A third distinct tumor necrosis factor receptor of orthopoxviruses. Proc. Natl. Acad. Sci. USA 95:3786–91 Hu FQ, CA Smith, DJ Pickup. 1994. Cowpox virus contains two copies of an early gene encoding a soluble secreted form of the type II TNF receptor. Virology 204:343–56 Alcami A, A Khanna, NL Paul, GL Smith. 1999. Vaccinia virus strains Lister, USSR and Evans express soluble and cell-surface tumour necrosis factor receptors. J. Gen. Virol. 80:949–59 Benedict CA, KD Butrovich, NS Lurain, J Corbeil, I Rooney, P Schneider, J Tschopp, CF Ware. 1999. Cutting edge: a novel viral TNF receptor superfamily member in virulent strains of human

912

179.

180.

181.


182.

183.

184.

185.

186.

TORTORELLA ET AL cytomegalovirus. J. Immunol. 162:6967– 70 Elsing A, HG Burgert. 1998. The adenovirus E3/10.4K-14.5K proteins downmodulate the apoptosis receptor Fas/ Apo-1 by inducing its internalization. Proc. Natl. Acad. Sci. USA 95:10072–77 Shisler J, C Yang, B Walter, CF Ware, LR Gooding. 1997. The adenovirus E3– 10.4K/14.5K complex mediates loss of cell surface Fas (CD95) and resistance to Fas-induced apoptosis. J. Virol. 71:8299– 306 Tollefson AE. 1998. Forced degradation of Fas inhibits apoptosis in adenovirusinfected cells. Nature 392:726–30 Carlin CR, AE Tollefson, HA Brady, BL Hoffman, WS Wold. 1989. Epidermal growth factor receptor is down-regulated by a 10,400 MW protein encoded by the E3 region of adenovirus. Cell 57:135–44 Tollefson AE, AR Stewart, SP Yei, SK Saha, WS Wold. 1991. The 10,400- and 14,500-dalton proteins encoded by region E3 of adenovirus form a complex and function together to down-regulate the epidermal growth factor receptor. J. Virol. 65:3095–105 Dimitrov T, P Krajcsi, TW Hermiston, AE Tollefson, M Hannink, WS Wold. 1997. Adenovirus E3–10.4K/14.5K protein complex inhibits tumor necrosis factor-induced translocation of cytosolic phospholipase A2 to membranes. J. Virol. 71:2830–37 Krajcsi P, T Dimitrov, TW Hermiston, AE Tollefson, TS Ranheim, SB Vande Pol, AH Stephenson, WS Wold. 1996. The adenovirus E3–14.7K protein and the E3–10.4K/14.5K complex of proteins, which independently inhibit tumor necrosis factor (TNF)-induced apoptosis, also independently inhibit TNF-induced release of arachidonic acid. J. Virol. 70:4904–13 Thorne TE, C Voelkel-Johnson, WM Casey, LW Parks, SM Laster. 1996. The activity of cytosolic phospholipase A2 is

187.

188.

189.

190.

191.

192.

193.

194.

195.

required for the lysis of adenovirusinfected cells by tumor necrosis factor. J. Virol. 70:8502–7 Tschopp J, M Irmler, M Thome. 1998. Inhibition of fas death signals by FLIPs. Curr. Opin. Immunol. 10:552–58 Thome M, P Schneider, K Hofmann, H Fickenscher, E Meinl, F Neipel, C Mattmann, K Burns, JL Bodmer, M Schroter, C Scaffidi, PH Krammer, ME Peter, J Tschopp. 1997. Viral FLICE-inhibitory proteins (FLIPs) prevent apoptosis induced by death receptors. Nature 386:517–21 Bertin J, RC Armstrong, S Ottilie, DA Martin, Y Wang, S Banks, GH Wang, TG Senkevich, ES Alnemri, B Moss, MJ Lenardo, KJ Tomaselli, JI Cohen. 1997. Death effector domain-containing herpesvirus and poxvirus proteins inhibit both Fas- and TNFR1-induced apoptosis. Proc. Natl. Acad. Sci. USA 94:1172–76 Hu S, C Vincenz, M Buller, VM Dixit. 1997. A novel family of viral death effector domain-containing molecules that inhibit both CD-95- and tumor necrosis factor receptor-1-induced apoptosis. J. Biol. Chem. 272:9621–24 Thornberry NA, Y Lazebnik. 1998. Caspases: enemies within. Science 281: 1312–16 Dbaibo GS, YA Hannun. 1998. Cytokine response modifier A (CrmA): a strategically deployed viral weapon. Clin. Immunol. Immunopathol. 86:134–40 Macen J, A Takahashi, KB Moon, R Nathaniel, PC Turner, RW Moyer. 1998. Activation of caspases in pig kidney cells infected with wild-type and CrmA/SPI-2 mutants of cowpox and rabbitpox viruses. J. Virol. 72:3524–33 Dobbelstein M, T Shenk. 1996. Protection against apoptosis by the vaccinia virus SPI-2 (B13R) gene product. J. Virol. 70:6479–85 Kettle S, A Alcami, A Khanna, R Ehret, C Jassoy, GL Smith. 1997. Vaccinia virus serpin B13R (SPI-2) inhibits interleukin-


196.

197.


198.

199.

200.

201.

202.

203.

204.

1beta-converting enzyme and protects virus-infected cells from TNF- and Fasmediated apoptosis, but does not prevent IL-1beta-induced fever. J. Gen. Virol. 78:677–85 Virgin HW, P Latreille, P Wamsley, K Hallsworth, KE Weck, AJ Dal Canto, SH Speck. 1997. Complete sequence and genomic analysis of murine gammaherpesvirus 68. J. Virol. 71:5894–904 Miller LK. 1997. Baculovirus interaction with host apoptotic pathways. J. Cell. Physiol. 173:178–82 Deveraux QL, JC Reed. 1999. IAP family proteins—suppressors of apoptosis. Genes Dev. 13:239–52 Rothe M, MG Pan, WJ Henzel, TM Ayres, DV Goeddel. 1995. The TNFR2TRAF signaling complex contains two novel proteins related to baculoviral inhibitor of apoptosis proteins. Cell 83:1243–52 Deveraux QL, R Takahashi, GS Salvesen, JC Reed. 1997. X-linked IAP is a direct inhibitor of cell-death proteases. Nature 388:300–4 Neilan JG, Z Lu, GF Kutish, L Zsak, TG Burrage, MV Borca, C Carrillo, DL Rock. 1997. A BIR motif containing gene of African swine fever virus, 4CL, is nonessential for growth in vitro and viral virulence. Virology 230:252–64 Chen P, J Tian, I Kovesdi, JT Bruder. 1998. Interaction of the adenovirus 14.7kDa protein with FLICE inhibits Fas ligand-induced apoptosis. J. Biol. Chem. 273:5815–20 Li Y, J Kang, MS Horwitz. 1998. Interaction of an adenovirus E3 14.7-kilodalton protein with a novel tumor necrosis factor alpha-inducible cellular protein containing leucine zipper domains. Mol. Cell. Biol. 18:1601–10 Li Y, J Kang, MS Horwitz. 1997. Interaction of an adenovirus 14.7-kilodalton protein inhibitor of tumor necrosis factor alpha cytolysis with a new member of the

205.

206.

207.

208.

209.

210.

211.

212.

213.

913

GTPase superfamily of signal transducers. J. Virol. 71:1576–82 Li Y, J Kang, J Friedman, L Tarassishin, J Ye, A Kovalenko, D Wallach, MS Horwitz. 1999. Identification of a cell protein (FIP-3) as a modulator of NF-kappaB activity and as a target of an adenovirus inhibitor of tumor necrosis factor alphainduced apoptosis. Proc. Natl. Acad. Sci. USA 96:1042–47 Adams JM, S Cory. 1998. The Bcl-2 protein family: arbiters of cell survival. Science 281:1322–26 Chao DT, SJ Korsmeyer. 1998. BCL-2 family: regulators of cell death. Annu. Rev. Immunol. 16:395–419 Cheng EH, J Nicholas, DS Bellows, GS Hayward, HG Guo, MS Reitz, JM Hardwick. 1997. A Bcl-2 homolog encoded by Kaposi sarcoma-associated virus, human herpesvirus 8, inhibits apoptosis but does not heterodimerize with Bax or Bak. Proc. Natl. Acad. Sci. USA 94:690– 94 Brun A, C Rivas, M Esteban, JM Escribano, C Alonso. 1996. African swine fever virus gene A179L, a viral homologue of bcl-2, protects cells from programmed cell death. Virology 225:227– 30 Ensser A, R Pflanz, B Fleckenstein. 1997. Primary structure of the alcelaphine herpesvirus 1 genome. J. Virol. 71:6517–25 Henderson S, D Huen, M Rowe, C Dawson, G Johnson, A Rickinson. 1993. Epstein-Barr virus-coded BHRF1 protein, a viral homologue of Bcl-2, protects human B cells from programmed cell death. Proc. Natl. Acad. Sci. USA 90:8479–83 Nava VE, EH Cheng, M Veliuona, S Zou, RJ Clem, ML Mayer, JM Hardwick. 1997. Herpesvirus saimiri encodes a functional homolog of the human bcl-2 oncogene. J. Virol. 71:4118–22 Sarid R, T Sato, RA Bohenzky, JJ Russo, Y Chang. 1997. Kaposi’s sarcoma-asso-

914

214.

215.


216.

217.

218.

219.

220.

221.

TORTORELLA ET AL ciated herpesvirus encodes a functional bcl-2 homologue. Nat. Med. 3:293–98 Telford EA, MS Watson, HC Aird, J Perry, AJ Davison. 1995. The DNA sequence of equine herpesvirus 2. J. Mol. Biol. 249:520–28 Tarodi B, T Subramanian, G Chinnadurai. 1994. Epstein-Barr virus BHRF1 protein protects against cell death induced by DNA-damaging agents and heterologous viral infection. Virology 201:404–7 Marshall WL, C Yim, E Gustafson, T Graf, DR Sage, K Hanify, L Williams, J Fingeroth, RW Finberg. 1999. EpsteinBarr virus encodes a novel homolog of the bcl-2 oncogene that inhibits apoptosis and associates with Bax and Bak. J. Virol. 73:5181–85 Henderson S, M Rowe, C Gregory, D Croom-Carter, F Wang, R Longnecker, E Kieff, A Rickinson. 1991. Induction of bcl-2 expression by Epstein-Barr virus latent membrane protein 1 protects infected B cells from programmed cell death. Cell 65:1107–15 Kieff E. 1996. Epstein-Barr virus and its replication. In Virology, ed. BC Fields, DM Knipe, PM Howley, Vol. 2, pp. 2343–96. New York: Lippincott-Raven. 3rd ed. Farrell PJ. 1998. Signal transduction from the Epstein-Barr virus LMP-1 transforming protein. Trends Microbiol. 6:175–78 Khanim F, C Dawson, CA Meseda, J Dawson, M Mackett, LS Young. 1997. BHRF1, a viral homologue of the Bcl-2 oncogene, is conserved at both the sequence and functional level in different Epstein-Barr virus isolates. J. Gen. Virol.78:2987–99 Lee MA, JL Yates. 1992. BHRF1 of Epstein-Barr virus, which is homologous to human proto-oncogene bcl2, is not essential for transformation of B cells or for virus replication in vitro. J. Virol. 66:1899–906

222. Marchini A, B Tomkinson, JI Cohen, E Kieff. 1991. BHRF1, the Epstein-Barr virus gene with homology to Bc12, is dispensable for B-lymphocyte transformation and virus replication. J. Virol. 65:5991–6000 223. Huang DC, S Cory, A Strasser. 1997. Bcl-2, Bcl-XL and adenovirus protein E1B19kD are functionally equivalent in their ability to inhibit cell death. Oncogene 14:405–14 224. Chen G, PE Branton, E Yang, SJ Korsmeyer, GC Shore. 1996. Adenovirus E1B 19-kDa death suppressor protein interacts with Bax but not with Bad. J. Biol. Chem. 271:24221–25 225. Kasof GM, L Goyal, E White. 1999. Btf, a novel death-promoting transcriptional repressor that interacts with Bcl-2-related proteins. Mol. Cell. Biol. 19:4390–404 226. Van Parijs L, A Biuckians, AK Abbas. 1998. Functional roles of Fas and Bcl-2regulated apoptosis of T lymphocytes. J. Immunol. 160:2065–71 227. Shisler JL, TG Senkevich, MJ Berry, B Moss. 1998. Ultraviolet-induced cell death blocked by a selenoprotein from a human dermatotropic poxvirus. Science 279:102–5 228. Scheffner M, JM Huibregtse, RD Vierstra, PM Howley. 1993. The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell 75:495–505 229. Leopardi R, C Van Sant, B Roizman. 1997. The herpes simplex virus 1 protein kinase US3 is required for protection from apoptosis induced by the virus. Proc. Natl. Acad. Sci. USA 94:7891–96 230. Zhu H, Y Shen, T Shenk. 1995. Human cytomegalovirus IE1 and IE2 proteins block apoptosis. J. Virol. 69:7960–70 231. Barry M, S Hnatiuk, K Mossman, SF Lee, L Boshkov, G McFadden. 1997. The myxoma virus M-T4 gene encodes a novel RDEL-containing protein that is retained within the endoplasmic reticulum and is important for the productive


232.

233.


234.

235.

236.

237.

238.

239.

240.

infection of lymphocytes. Virology 239: 360–77 Paul WE, ed. 1999. Fundamental Immunology. New York: Lippincott-Raven, 3rd ed. Juang Y, W Lowther, M Kellum, WC Au, R Lin, J Hiscott, PM Pitha. 1998. Primary activation of interferon A and interferon B gene transcription by interferon regulatory factor 3. Proc. Natl. Acad. Sci. USA 95:9837–42 Lin R, C Heylbroeck, PM Pitha, J Hiscott. 1998. Virus-dependent phosphorylation of the IRF-3 transcription factor regulates nuclear translocation, transactivation potential, and proteasome-mediated degradation. Mol. Cell. Biol. 18:2986–96 Mahr JA, LR Gooding. 1999. Immune evasion by adenoviruses. Immunol. Rev. 168:121–30 Hiscott J, P Pitha, P Genin, H Nguyen, C Heylbroeck, Y Mamane, M Algarte, R Lin. 1999. Triggering the interferon response: the role of IRF-3 transcription factor. J. Interferon Cytokine Res. 19:1– 13 Hirsch AJ, T Shenk. 1999. Human cytomegalovirus inhibits transcription of the CC chemokine MCP-1 gene. J. Virol. 73:404–10 Petit F, S Bertagnoli, J Gelfi, F Fassy, C Boucraut-Baralon, A Milon. 1996. Characterization of a myxoma virus-encoded serpin-like protein with activity against interleukin-1 beta-converting enzyme. J. Virol. 70:5860–66 Messud-Petit F, J Gelfi, M Delverdier, MF Amardeilh, R Py, G Sutter, S Bertagnoli. 1998. Serp2, an inhibitor of the interleukin-1beta-converting enzyme, is critical in the pathobiology of myxoma virus. J. Virol. 72:7830–39 Nash P, J Barrett, JX Cao, S Hota-Mitchell, AS Lalani, H Everett, XM Xu, J Robichaud, S Hnatiuk, C Ainslie, BT Seet, G McFadden. 1999. Immunomodulation by

241.

242.

243.

244.

245.

246.

247.

248.

249.

250.

251.

915

viruses: the myxoma virus story. Immunol. Rev. 168:103–20 Muller G, J Saloga, T Germann, G Schuler, J Knop, AH Enk. 1995. IL-12 as mediator and adjuvant for the induction of contact sensitivity in vivo. J. Immunol. 155:4661–68 Karp CL 1999. Measles: immunosuppression, interleukin-12, and complement receptors. Immunol. Rev. 168:91– 101 Novick D, H Engelmann, D Wallach, M Rubinstein. 1989. Soluble cytokine receptors are present in normal human urine. J. Exp. Med. 170:1409–14 McFadden G, A Lalani, H Everett, P Nash, X Xu. 1998. Virus-encoded receptors for cytokines and chemokines. Semin. Cell. Dev. Biol. 9:359–68 Cha TA, E Tom, GW Kemble, GM Duke, ES Mocarski, RR Spaete. 1996. Human cytomegalovirus clinical isolates carry at least 19 genes not found in laboratory strains. J. Virol. 70:78–83 Symons JA, A Alcami, GL Smith. 1995. Vaccinia virus encodes a soluble type I interferon receptor of novel structure and broad species specificity. Cell 81:551–60 Alcami A, GL Smith. 1995. Vaccinia, cowpox, and camelpox viruses encode soluble gamma interferon receptors with novel broad species specificity. J. Virol. 69:4633–39 Mossman K, C Upton, RM Buller, G McFadden. 1995. Species specificity of ectromelia virus and vaccinia virus interferon-gamma binding proteins. Virology 208:762–69 Smith GL 1996. Virus proteins that bind cytokines, chemokines or interferons. Curr. Opin. Immunol. 8:467–71 Mossman K, C Upton, G McFadden. 1995. The myxoma virus-soluble interferon-gamma receptor homolog, M-T7, inhibits interferon-gamma in a speciesspecific manner. J. Biol. Chem. 270: 3031–38 Lalani AS, K Graham, K Mossman, K

916

252. 253.


254.

255.

256.

257.

258.

259.

260.

TORTORELLA ET AL Rajarathnam, I Clark-Lewis, D Kelvin, G McFadden. 1997. The purified myxoma virus gamma interferon receptor homolog M-T7 interacts with the heparinbinding domains of chemokines. J. Virol. 71:4356–63 Baggiolini M. 1998. Chemokines and leukocyte traffic. Nature 392:565–68 Schall TJ, KB Bacon. 1994. Chemokines, leukocyte trafficking, and inflammation. Curr. Opin. Immunol. 6:865–73 Smith CA, TD Smith, PJ Smolak, D Friend, H Hagen, M Gerhart, L Park, DJ Pickup, D Torrance, K Mohler, K Schooley, RG Goodwin. 1997. Poxvirus genomes encode a secreted, soluble protein that preferentially inhibits beta chemokine activity yet lacks sequence homology to known chemokine receptors. Virology 236:316–27 Alcami A, JA Symons, PD Collins, TJ Williams, GL Smith. 1998. Blockade of chemokine activity by a soluble chemokine binding protein from vaccinia virus. J. Immunol. 160:624–33 Essani K, S Chalasani, R Eversole, L Beuving, L Birmingham. 1994. Multiple anti-cytokine activities secreted from tanapox virus-infected cells. Microb. Pathog. 17:347–53 Strockbine LD, JI Cohen, T Farrah, SD Lyman, F Wagener, RF DuBose, RJ Armitage, MK Spriggs. 1998. The Epstein-Barr virus BARF1 gene encodes a novel, soluble colony-stimulating factor-1 receptor. J. Virol. 72:4015–21 Spriggs MK. 1999. Virus-encoded modulators of cytokines and growth factors. Cytokine Growth Factor Rev. 10:1–4 Billstrom MA, GL Johnson, NJ Avdi, GS Worthen. 1998. Intracellular signaling by the chemokine receptor US28 during human cytomegalovirus infection. J. Virol. 72:5535–44 Bodaghi B, TR Jones, D Zipeto, C Vita, L Sun, L Laurent, F Arenzana-Seisdedos, JL Virelizier, S Michelson. 1998. Chemokine sequestration by viral chemore-

261.

262.

263.

264.

265.

266.

267.

ceptors as a novel viral escape strategy: withdrawal of chemokines from the environment of cytomegalovirus-infected cells. J. Exp. Med. 188:855–66 Pleskoff O, C Treboute, A Brelot, N Heveker, M Seman, M Alizon. 1997. Identification of a chemokine receptor encoded by human cytomegalovirus as a cofactor for HIV-1 entry. Science 276:1874–78 Nicholas J, KR Cameron, RW Honess. 1992. Herpesvirus saimiri encodes homologues of G protein-coupled receptors and cyclins. Nature 355:362–65 Kledal TN, MM Rosenkilde, F Coulin, G Simmons, AH Johnsen, S Alouani, CA Power, HR Luttichau, J Gerstoft, PR Clapham, I Clark-Lewis, TNC Wells, TW Schwartz. 1997. A broad-spectrum chemokine antagonist encoded by Kaposi’s sarcoma-associated herpesvirus. Science 277:1656–59 Cesarman E, RG Nador, F Bai, RA Bohenzky, JJ Russo, PS Moore, Y Chang, DM Knowles. 1996. Kaposi’s sarcomaassociated herpesvirus contains G protein-coupled receptor and cyclin D homologs which are expressed in Kaposi’s sarcoma and malignant lymphoma. J. Virol. 70:8218–23 Bais C, B Santomasso, O Coso, L Arvanitakis, EG Raaka, JS Gutkind, AS Asch, E Cesarman, MC Gershengorn, EA Mesri, MC Gerhengorn. 1998. G-protein-coupled receptor of Kaposi’s sarcoma-associated herpesvirus is a viral oncogene and angiogenesis activator. Nature 391:86–89 Arvanitakis L, E Geras-Raaka, A Varma, MC Gershengorn, E Cesarman. 1997. Human herpesvirus KSHV encodes a constitutively active G-protein-coupled receptor linked to cell proliferation. Nature 385:347–50 Geras-Raaka E, A Varma, H Ho, I ClarkLewis, MC Gershengorn. 1998. Human interferon-gamma-inducible protein 10 (IP-10) inhibits constitutive signaling of Kaposi’s sarcoma-associated herpesvirus


268.

269.


270.

271.

272.

273.

274.

275.

G protein-coupled receptor. J. Exp. Med. 188:405–8 Geras-Raaka E, A Varma, I Clark-Lewis, MC Gershengorn. 1998. Kaposi’s sarcoma-associated herpesvirus (KSHV) chemokine vMIP-II and human SDF1alpha inhibit signaling by KSHV G protein-coupled receptor. Biochem. Biophys. Res. Commun. 253:725–27 Beisser PS, C Vink, JG Van Dam, G Grauls, SJ Vanherle, CA Bruggeman. 1998. The R33 G protein-coupled receptor gene of rat cytomegalovirus plays an essential role in the pathogenesis of viral infection. J. Virol. 72:2352–63 Davis-Poynter NJ, DM Lynch, H Vally, GR Shellam, WD Rawlinson, BG Barrell, HE Farrell. 1997. Identification and characterization of a G protein-coupled receptor homolog encoded by murine cytomegalovirus. J. Virol. 71:1521–29 Yasukawa M, A Hasegawa, I Sakai, H Ohminami, J Arai, S Kaneko, Y Yakushijin, K Maeyama, H Nakashima, R Arakaki, S Fujita. 1999. Down-regulation of CXCR4 by human herpesvirus 6 (HHV6) and HHV-7. J. Immunol. 162:5417–22 Secchiero P, D Zella, O Barabitskaja, MS Reitz, S Capitani, RC Gallo, G Zauli. 1998. Progressive and persistent downregulation of surface CXCR4 in CD4(`) T cells infected with human herpesvirus 7.Blood 92:4521–28 Damon I, PM Murphy, B Moss. 1998. Broad spectrum chemokine antagonistic activity of a human poxvirus chemokine homolog. Proc. Natl. Acad. Sci. USA 95:6403–7 Krathwohl MD, R Hromas, DR Brown, HE Broxmeyer, KH Fife. 1997. Functional characterization of the C-C chemokine-like molecules encoded by molluscum contagiosum virus types 1 and 2. Proc. Natl. Acad. Sci. USA 94:9875–80 Baggiolini M, B Dewald, B Moser. 1997. Human chemokines: an update. Annu. Rev. Immunol. 15:675–705

917

276. Zou P, Y Isegawa, K Nakano, M Haque, Y Horiguchi, K Yamanishi. 1999. Human herpesvirus 6 open reading frame U83 encodes a functional chemokine. J. Virol. 73:5926–33 277. Fleming P, N Davis-Poynter, M DegliEsposti, E Densley, J Papadimitriou, G Shellam, H Farrell. 1999. The murine cytomegalovirus chemokine homolog, m131/129, is a determinant of viral pathogenicity. J. Virol. 73:6800–9 278. MacDonald MR, MW Burney, SB Resnick, HI Virgin. 1999. Spliced mRNA encoding the murine cytomegalovirus chemokine homolog predicts a beta chemokine of novel structure. J. Virol. 73:3682–91 279. Saederup N, Y Lin, DJ Dairaghi, TJ Schall, ES Mocarski. 1999. Cytomegalovirus-encoded beta chemokine promotes monocyte-associated viremia in the host. Proc. Natl. Acad. Sci. USA 96:10881–86 280. Neipel F, JC Albrecht, B Fleckenstein. 1997. Cell-homologous genes in the Kaposi’s sarcoma-associated rhadinovirus human herpesvirus 8: determinants of its pathogenicity? J. Virol. 71:4187–92 281. Aoki Y, ES Jaffe, Y Chang, K Jones, J Teruya-Feldstein, PS Moore, G Tosato. 1999. Angiogenesis and hematopoiesis induced by Kaposi’s sarcoma-associated herpesvirus-encoded interleukin-6. Blood 93:4034–43 282. Nicholas J, VR Ruvolo, WH Burns, G Sandford, X Wan, D Ciufo, SB Hendrickson, HG Guo, GS Hayward, MS Reitz. 1997. Kaposi’s sarcoma-associated human herpesvirus-8 encodes homologues of macrophage inflammatory protein-1 and interleukin-6. Nat. Med. 3:287–92 283. Rettig MB, HJ Ma, RA Vescio, M Pold, G Schiller, D Belson, A Savage, C Nishikubo, C Wu, J Fraser, JW Said, JR Berenson. 1997. Kaposi’s sarcoma-associated herpesvirus infection of bone mar-

918

284.

285.


286.

287.

288.

289.

290.

291.

TORTORELLA ET AL row dendritic cells from multiple myeloma patients. Science 276:1851–54 Sallusto F, CR Mackay, A Lanzavecchia. 1997. Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Science 277:2005–7 Bonecchi R, G Bianchi, PP Bordignon, D D’Ambrosio, R Lang, A Borsatti, S Sozzani, P Allavena, PA Gray, A Mantovani, F Sinigaglia. 1998. Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (Th1s) and Th2s. J. Exp. Med. 187:129–34 Dairaghi DJ, RA Fan, BE McMaster, MR Hanley, TJ Schall. 1999. HHV8-encoded vMIP-I selectively engages chemokine receptor CCR8. Agonist and antagonist profiles of viral chemokines. J. Biol. Chem. 274:21569–74 Endres MJ, CG Garlisi, H Xiao, L Shan, JA Hedrick. 1999. The Kaposi’s sarcoma-related herpesvirus (KSHV)encoded chemokine vMIP- I is a specific agonist for the CC chemokine receptor (CCR)8. J. Exp. Med. 189:1993–98 Sozzani S, W Luini, G Bianchi, P Allavena, TN Wells, M Napolitano, G Bernardini, A Vecchi, D D’Ambrosio, D Mazzeo, F Sinigaglia, A Santoni, E Maggi, S Romagnani, A Mantovani. 1998. The viral chemokine macrophage inflammatory protein-II is a selective Th2 chemoattractant. Blood 92:4036–39 Moore PS, C Boshoff, RA Weiss, Y Chang. 1996. Molecular mimicry of human cytokine and cytokine response pathway genes by KSHV. Science 274:1739–44 D’Ambrosio D, A Iellem, R Bonecchi, D Mazzeo, S Sozzani, A Mantovani, F Sinigaglia. 1998. Selective up-regulation of chemokine receptors CCR4 and CCR8 upon activation of polarized human type 2 Th cells. J. Immunol. 161:5111–15 Zingoni A, H Soto, JA Hedrick, A Stoppacciaro, CT Storlazzi, F Sinigaglia, D D’Ambrosio, A O’Garra, D Robinson, M

292.

293.

294.

295.

296.

297.

298.

299.

Rocchi, A Santoni, A Zlotnik, M Napolitano. 1998. The chemokine receptor CCR8 is preferentially expressed in Th2 but not Th1 cells. J. Immunol. 161:547– 51 Cayley PJ, JA Davies, KG McCullagh, IM Kerr. 1984. Activation of the ppp(A2’p)nA system in interferontreated, herpes simplex virus-infected cells and evidence for novel inhibitors of the ppp(A2’p)nA-dependent RNase. Eur. J. Biochem. 143:165–74 Novick D, SH Kim, G Fantuzzi, LL Reznikov, CA Dinarello, M Rubinstein. 1999. Interleukin-18 binding protein: a novel modulator of the Th1 cytokine response. Immunity 10:127–36 Takeda K, H Tsutsui, T Yoshimoto, O Adachi, N Yoshida, T Kishimoto, H Okamura, K Nakanishi, S Akira. 1998. Defective NK cell activity and Th1 response in IL-18-deficient mice. Immunity 8:383–90 Puren AJ, G Fantuzzi, Y Gu, MS Su, CA Dinarello. 1998. Interleukin-18 (IFNgamma-inducing factor) induces IL-8 and IL-1beta via TNFalpha production from non-CD14` human blood mononuclear cells. J. Clin. Invest. 101:711–21 Liu Y, R de Waal Malefyt, F Briere, C Parham, JM Bridon, J Banchereau, KW Moore, J Xu. 1997. The EBV IL-10 homologue is a selective agonist with impaired binding to the IL-10 receptor. J. Immunol. 158:604–13 Kotenko SV, CD Krause, LS Izotova, BP Pollack, W Wu, S Pestka. 1997. Identification and functional characterization of a second chain of the interleukin-10 receptor complex. EMBO J. 16:5894– 903 Helminen M, N Lahdenpohja, M Hurme. 1999. Polymorphism of the interleukin10 gene is associated with susceptibility to Epstein-Barr virus infection. J. Infect. Dis. 180:496–99 Li M, H Lee, J Guo, F Neipel, B Fleckenstein, K Ozato, JU Jung. 1998.


300.

301.


302.

303.

304.

305.

306.

307.

308.

Kaposi’s sarcoma-associated herpesvirus viral interferon regulatory factor. J. Virol. 72:5433–40 He B, M Gross, B Roizman. 1997. The gamma(1)34.5 protein of herpes simplex virus 1 complexes with protein phosphatase 1alpha to dephosphorylate the alpha subunit of the eukaryotic translation initiation factor 2 and preclude the shutoff of protein synthesis by double-stranded RNA-activated protein kinase. Proc. Natl. Acad. Sci. USA 94:843–48 Taylor DR, ST Shi, PR Romano, GN Barber, MM Lai. 1999. Inhibition of the interferon-inducible protein kinase PKR by HCV E2 protein. Science 285:107–10 Davies MV, HW Chang, BL Jacobs, RJ Kaufman. 1993. The E3L and K3L vaccinia virus gene products stimulate translation through inhibition of the double-stranded RNA-dependent protein kinase by different mechanisms. J. Virol. 67:1688–92 Sharp TV, DA Raine, DR Gewert, B Joshi, R Jagus, MJ Clemens. 1999. Activation of the interferon-inducible (2’-5’) oligoadenylate synthetase by the Epstein-Barr virus RNA, EBER-1. Virology 257:303–13 Ma Y, MB Mathews. 1996. Structure, function, and evolution of adenovirusassociated RNA: a phylogenetic approach. J. Virol. 70:5083–99 Clemens MJ 1993. The small RNAs of Epstein-Barr virus. Mol. Biol. Rep. 17:81–92 Dempsey PW, ME Allison, S Akkaraju, CC Goodnow, DT Fearon. 1996. C3d of complement as a molecular adjuvant: bridging innate and acquired immunity. Science 271:348–50 Fearon DT 1998. The complement system and adaptive immunity. Semin. Immunol. 10:355–61 Carroll MC 1998. The role of complement and complement receptors in induction and regulation of immunity. Annu. Rev. Immunol. 16:545–68

919

309. Anderson DR, CM Carthy, JE Wilson, D Yang, DV Devine, BM McManus. 1997. Complement component 3 interactions with coxsackievirus B3 capsid proteins: innate immunity and the rapid formation of splenic antiviral germinal centers. J. Virol. 71:8841–45 310. Reitter JN, RE Means, RC Desrosiers. 1998. A role for carbohydrates in immune evasion in AIDS. Nat. Med. 4:679–84 311. Lubinski J, T Nagashunmugam, HM Friedman. 1998. Viral interference with antibody and complement. Semin. Cell. Dev. Biol. 9:329–37 312. Dubin G, E Socolof, I Frank, HM Friedman. 1991. Herpes simplex virus type 1 Fc receptor protects infected cells from antibody-dependent cellular cytotoxicity. J. Virol. 65:7046–50 313. Chapman TL, I You, IM Joseph, PJ Bjorkman, SL Morrison, M Raghavan. 1999. Characterization of the interaction between the herpes simplex virus type I Fc receptor and immunoglobulin G. J. Biol. Chem. 274:6911–19 314. Deisenhofer J. 1981. Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A resolution. Biochemistry 20:2361–70 315. Raghavan M, MY Chen, LN Gastinel, PJ Bjorkman. 1994. Investigation of the interaction between the class I MHCrelated Fc receptor and its immunoglobulin G ligand. Immunity 1:303–15 316. Stone GC, U Sjobring, L Bjorck, J Sjoquist, CV Barber, FA Nardella. 1989. The Fc binding site for streptococcal protein G is in the C gamma 2-C gamma 3 interface region of IgG and is related to the sites that bind staphylococcal protein A and human rheumatoid factors. J. Immunol. 143:565–70 317. Favoreel HW, HJ Nauwynck, P Van Oostveldt, TC Mettenleiter, MB Pensaert. 1997. Antibody-induced and cyto-

920

318.

319.


320.

321.

322.

323.

324.

325.

TORTORELLA ET AL skeleton-mediated redistribution and shedding of viral glycoproteins, expressed on pseudorabies virus-infected cells. J. Virol. 71:8254–61 Brown ZA, S Selke, J Zeh, J Kopelman, A Maslow, RL Ashley, DH Watts, S Berry, M Herd, L Corey. 1997. The acquisition of herpes simplex virus during pregnancy. N. Engl. J. Med. 337:509– 15 Calderon J, EE Avila. 1986. Antibodyinduced caps in Entamoeba histolytica: isolation and electrophoretic analysis. J. Infect. Dis. 153:927–32 Favoreel HW, HJ Nauwynck, HM Halewyck, P Van Oostveldt, TC Mettenleiter, MB Pensaert. 1999. Antibodyinduced endocytosis of viral glycoproteins and major histocompatibility complex class I on pseudorabies virusinfected monocytes. J. Gen. Virol. 80:1283–91 Nagashunmugam T, J Lubinski, L Wang, LT Goldstein, BS Weeks, P Sundaresan, EH Kang, G Dubin, HM Friedman. 1998. In vivo immune evasion mediated by the herpes simplex virus type 1 immunoglobulin G Fc receptor. J. Virol. 72:5351– 59 Frank I, HM Friedman. 1989. A novel function of the herpes simplex virus type 1 Fc receptor: participation in bipolar bridging of antiviral immunoglobulin G. J. Virol. 63:4479–88 Van Vliet KE, LA De Graaf-Miltenburg, J Verhoef, JA Van Strijp. 1992. Direct evidence for antibody bipolar bridging on herpes simplex virus-infected cells. Immunology 77:109–15 Burton GF, A Masuda, SL Heath, BA Smith, JG Tew, AK Szakal. 1997. Follicular dendritic cells (FDC) in retroviral infection: host/pathogen perspectives. Immunol. Rev. 156:185–97 Heath SL, JG Tew, JG Tew, AK Szakal, GF Burton. 1995. Follicular dendritic cells and human immunodeficiency virus infectivity. Nature 377:740–44

326. Embretson J, M Zupancic, JL Ribas, A Burke, P Racz, K Tenner-Racz, AT Haase. 1993. Massive covert infection of helper T lymphocytes and macrophages by HIV during the incubation period of AIDS. Nature 362:359–62 327. Pantaleo G, C Graziosi, JF Demarest, L Butini, M Montroni, CH Fox, JM Orenstein, DP Kotler, AS Fauci. 1993. HIV infection is active and progressive in lymphoid tissue during the clinically latent stage of disease. Nature 362:355– 58 328. Ravetch JV. 1994. Fc receptors: rubor redux. Cell 78:553–60 329. Daeron M. 1997. Fc receptor biology. Annu. Rev. Immunol. 15:203–34 330. D’Ambrosio D, KL Hippen, SA Minskoff, I Mellman, G Pani, KA Siminovitch, JC Cambier. 1995. Recruitment and activation of PTP1C in negative regulation of antigen receptor signaling by Fc gamma RIIB1. Science 268:293–97 331. Wagle NM, AE Faassen, JH Kim, SK Pierce. 1999. Regulation of B cell receptor-mediated MHC class II antigen processing by FcgammaRIIB1. J. Immunol. 162:2732–40 332. Ravanel K, C Castelle, T Defrance, TF Wild, D Charron, V Lotteau, C Rabourdin-Combe. 1997. Measles virus nucleocapsid protein binds to FcgammaRII and inhibits human B cell antibody production. J. Exp. Med. 186:269–78 333. Kotwal GJ, SN Isaacs, R McKenzie, MM Frank, B Moss. 1990. Inhibition of the complement cascade by the major secretory protein of vaccinia virus. Science 250:827–30 334. McKenzie R, GJ Kotwal, B Moss, CH Hammer, MM Frank. 1992. Regulation of complement activity by vaccinia virus complement-control protein. J. Infect. Dis. 166:1245–50 335. Sahu A, SN Isaacs, AM Soulika, JD Lambris. 1998. Interaction of vaccinia virus complement control protein with human complement proteins: factor I-


336.

337.


338.

339.

340.

341.

342.

343.

mediated degradation of C3b to iC3b1 inactivates the alternative complement pathway. J. Immunol. 160:5596–604 Rosengard AM, JM Ahearn. 1998. Creation and functional characterization of SPICE, the smallpox inhibitor of complement enzymes. Mol. Immunol. 35:397 Kotwal GJ, CG Miller, DE Justus. 1998. The inflammation modulatory protein (IMP) of cowpox virus drastically diminishes the tissue damage by downregulating cellular infiltration resulting from complement activation. Mol. Cell. Biochem. 185:39–46 Howard J, DE Justus, AV Totmenin, S Shchelkunov, GJ Kotwal. 1998. Molecular mimicry of the inflammation modulatory proteins (IMPs) of poxviruses: evasion of the inflammatory response to preserve viral habitat. J. Leukocyte Biol. 64:68–71 Miller CG, SN Shchelkunov, GJ Kotwal. 1997. The cowpox virus-encoded homolog of the vaccinia virus complement control protein is an inflammation modulatory protein. Virology 229:126–33 Kostavasili I, A Sahu, HM Friedman, RJ Eisenberg, GH Cohen, JD Lambris. 1997. Mechanism of complement inactivation by glycoprotein C of herpes simplex virus. J. Immunol. 158:1763–71 McNearney TA, C Odell, VM Holers, PG Spear, JP Atkinson. 1987. Herpes simplex virus glycoproteins gC-1 and gC-2 bind to the third component of complement and provide protection against complement-mediated neutralization of viral infectivity. J. Exp. Med. 166:1525–35 Friedman HM, L Wang, NO Fishman, JD Lambris, RJ Eisenberg, GH Cohen, J Lubinski. 1996. Immune evasion properties of herpes simplex virus type 1 glycoprotein gC. J. Virol. 70:4253–60 Huemer HP, Y Wang, P Garred, V Koistinen, S Oppermann. 1993. Herpes simplex virus glycoprotein C: molecular mimicry of complement regulatory pro-

344.

345.

346.

347.

348.

349. 350.

351.

352.

921

teins by a viral protein. Immunology 79:639–47 Lachmann PJ, A Davies. 1997. Complement and immunity to viruses. Immunol. Rev. 159:69–77 Hidaka Y, S Sakuma, Y Kumano, H Minagawa, R Mori. 1990. Characterization of glycoprotein C-negative mutants of herpes simplex virus type 1 isolated from a patient with keratitis. Arch. Virol. 113:195–207 Fodor WL, SA Rollins, S Bianco-Caron, RP Rother, ER Guilmette, WV Burton, JC Albrecht, B Fleckenstein, SP Squinto. 1995. The complement control protein homolog of herpesvirus saimiri regulates serum complement by inhibiting C3 convertase activity. J. Virol. 69:3889–92 Rother RP, SA Rollins, WL Fodor, JC Albrecht, E Setter, B Fleckenstein, SP Squinto. 1994. Inhibition of complement-mediated cytolysis by the terminal complement inhibitor of herpesvirus saimiri. J. Virol. 68:730–37 Vanderplasschen A, E Mathew, M Hollinshead, RB Sim, GL Smith. 1998. Extracellular enveloped vaccinia virus is resistant to complement because of incorporation of host complement control proteins into its envelope. Proc. Natl. Acad. Sci. USA 95:7544–49 Smith GL. 1999. Vaccinia virus immune evasion. Immunol. Lett. 65:55–62 Payne LG. 1980. Significance of extracellular enveloped virus in the in vitro and in vivo dissemination of vaccinia. J. Gen. Virol. 50:89–100 Saifuddin M, CJ Parker, ME Peeples, MK Gorny, S Zolla-Pazner, M Ghassemi, IA Rooney, JP Atkinson, GT Spear. 1995. Role of virion-associated glycosylphosphatidylinositol-linked proteins CD55 and CD59 in complement resistance of cell line-derived and primary isolates of HIV-1. J. Exp. Med. 182:501– 9 Spear GT, NS Lurain, CJ Parker, M Ghassemi, GH Payne, M Saifuddin.

922

353.


354.

355.

356.

357.

358.

359.

TORTORELLA ET AL 1995. Host cell-derived complement control proteins CD55 and CD59 are incorporated into the virions of two unrelated enveloped viruses: human T cell leukemia/lymphoma virus type I (HTLVI) and human cytomegalovirus (HCMV). J. Immunol. 155:4376–81 Schmitz J, JP Zimmer, B Kluxen, S Aries, M Bogel, I Gigli, H Schmitz. 1995. Antibody-dependent complementmediated cytotoxicity in sera from patients with HIV-1 infection is controlled by CD55 and CD59. J. Clin. Invest. 96:1520–26 Marschang P, J Sodroski, R Wurzner, MP Dierich. 1995. Decay-accelerating factor (CD55) protects human immunodeficiency virus type 1 from inactivation by human complement. Eur. J. Immunol. 25:285–90 Spiller OB, BP Morgan, F Tufaro, DV Devine. 1996. Altered expression of host-encoded complement regulators on human cytomegalovirus-infected cells. Eur. J. Immunol. 26:1532–38 Pascual M, C Danielsson, G Steiger, JA Schifferli. 1994. Proteolytic cleavage of CR1 on human erythrocytes in vivo: evidence for enhanced cleavage in AIDS. Eur. J. Immunol. 24:702–8 Munson LG, ME Scott, AL Landay, GT Spear. 1995. Decreased levels of complement receptor 1 (CD35) on B lymphocytes in persons with HIV infection. Clin. Immunol. Immunopathol. 75:20–25 Jouvin MH, W Rozenbaum, R Russo, MD Kazatchkine. 1987. Decreased expression of the C3b/C4b complement receptor (CR1) in AIDS and AIDSrelated syndromes correlates with clinical subpopulations of patients with HIV infection. AIDS 1:89–94 Larcher C, TF Schulz, J Hofbauer, P Hengster, N Romani, H Wachter, MP Dierich. 1990. Expression of the C3d/ EBV receptor and of other cell membrane surface markers is altered upon HIV-1 infection of myeloid, T, and B

360.

361.

362.

363.

364.

365.

366.

367.

cells. J. Acquired Immune Defic. Syndr. 3:103–8 Wahl SM, JB Allen, S Gartner, JM Orenstein, M Popovic, DE Chenoweth, LO Arthur, WL Farrar, LM Wahl. 1989. HIV1 and its envelope glycoprotein downregulate chemotactic ligand receptors and chemotactic function of peripheral blood monocytes. J. Immunol. 142: 3553–59 Speth C, MP Dierich. 1999. Modulation of cell surface protein expression by infection with HIV-1. Leukemia 13(Suppl. 1):99–105 Pangburn MK, RD Schreiber, HJ MullerEberhard. 1981. Formation of the initial C3 convertase of the alternative complement pathway: acquisition of C3b-like activities by spontaneous hydrolysis of the putative thioester in native C3. J. Exp. Med. 154:856–67 Stoiber H, C Ebenbichler, R Schneider, J Janatova, MP Dierich. 1995. Interaction of several complement proteins with gp120 and gp41, the two envelope glycoproteins of HIV-1. AIDS 9:19–26 Pinter C, AG Siccardi, L Lopalco, R Longhi, A Clivio. 1995. HIV glycoprotein 41 and complement factor H interact with each other and share functional as well as antigenic homology. AIDS Res. Hum. Retrovir. 11:971–80 Stoiber H, C Pinter, AG Siccardi, A Clivio, MP Dierich. 1996. Efficient destruction of human immunodeficiency virus in human serum by inhibiting the protective action of complement factor H and decay accelerating factor (DAF, CD55). J. Exp. Med. 183:307–10 Horstmann RD, HJ Sievertsen, J Knobloch, VA Fischetti. 1988. Antiphagocytic activity of streptococcal M protein: selective binding of complement control protein factor H. Proc. Natl. Acad. Sci. USA 85:1657–61 Wiertz EJ, TR Jones, L Sun, M Bogyo, HJ Geuze, HL Ploegh. 1996. The human cytomegalovirus US11 gene product dis-


368.

369.


370.

371.

372.

373.

374.

375.

376.

locates MHC class I heavy chains from the endoplasmic reticulum to the cytosol. Cell 84:769–79 Wang GH, J Bertin, Y Wang, DA Martin, J Wang, KJ Tomaselli, RC Armstrong, JI Cohen. 1997. Bovine herpesvirus 4 BORFE2 protein inhibits Fas- and tumor necrosis factor receptor 1-induced apoptosis and contains death effector domains shared with other gamma-2 herpesviruses. J. Virol. 71:8928–32 Chacon MR, F Almazan, ML Nogal, E Vinuela, JF Rodriguez. 1995. The African swine fever virus IAP homolog is a late structural polypeptide. Virology 214:670–74 Revilla Y, A Cebrian, E Baixeras, C Martinez, E Vinuela, ML Salas. 1997. Inhibition of apoptosis by the African swine fever virus Bcl-2 homologue: role of the BH1 domain. Virology 228:400–4 Yew PR, AJ Berk. 1992. Inhibition of p53 transactivation required for transformation by adenovirus early 1B protein. Nature 357:82–85 Moore M, N Horikoshi, T Shenk. 1996. Oncogenic potential of the adenovirus E4orf6 protein. Proc. Natl. Acad. Sci. USA 93:11295–301 Mietz JA, T Unger, JM Huibregtse, PM Howley. 1992. The transcriptional transactivation function of wild-type p53 is inhibited by SV40 large T-antigen and by HPV-16 E6 oncoprotein. EMBO J. 11:5013–20 DeCaprio JA, JW Ludlow, J Figge, JY Shew, CM Huang, WH Lee, E Marsilio, E Paucha, DM Livingston. 1988. SV40 large tumor antigen forms a specific complex with the product of the retinoblastoma susceptibility gene. Cell 54:275–83 Liu JL, Y Ye, LF Lee, HJ Kung. 1998. Transforming potential of the herpesvirus oncoprotein MEQ: morphological transformation, serum-independent growth, and inhibition of apoptosis. J. Virol. 72:388–95 Upton C, JL Macen, M Schreiber, G

377.

378.

379.

380.

381.

382.

383.

384.

923

McFadden. 1991. Myxoma virus expresses a secreted protein with homology to the tumor necrosis factor receptor gene family that contributes to viral virulence. Virology 184:370–82 Spriggs MK, DE Hruby, CR Maliszewski, DJ Pickup, JE Sims, RM Buller, J VanSlyke. 1992. Vaccinia and cowpox viruses encode a novel secreted interleukin-1-binding protein. Cell 71:145–52 Alcami A, GL Smith. 1992. A soluble receptor for interleukin-1 beta encoded by vaccinia virus: a novel mechanism of virus modulation of the host response to infection. Cell 71:153–67 Ray CA, RA Black, SR Kronheim, TA Greenstreet, PR Sleath, GS Salvesen, DJ Pickup. 1992. Viral inhibition of inflammation: cowpox virus encodes an inhibitor of the interleukin-1 beta converting enzyme. Cell 69:597–604 Smith GL, YS Chan. 1991. Two vaccinia virus proteins structurally related to the interleukin-1 receptor and the immunoglobulin superfamily. J. Gen. Virol. 72:511–18 Foster GR, AM Ackrill, RD Goldin, IM Kerr, HC Thomas, GR Stark. 1991. Expression of the terminal protein region of hepatitis B virus inhibits cellular responses to interferons alpha and gamma and double-stranded RNA. Proc. Natl. Acad. Sci. USA 88:2888–92. Erratum. 1995. Proc. Natl. Acad. Sci. USA 92(8):3632 Foster GR, RD Goldin, A Hay, MJ McGarvey, GR Stark, HC Thomas. 1993. Expression of the terminal protein of hepatitis B virus is associated with failure to respond to interferon therapy. Hepatology 17:757–62 Burysek L, WS Yeow, PM Pitha. 1999. Unique properties of a second human herpesvirus 8-encoded interferon regulatory factor (vIRF-2). J. Hum. Virol. 2:19–32 Kanda K, T Decker, P Aman, M Wahlstrom, A von Gabain, B Kallin. 1992. The

924

385.


386.

387.

388.

389.

390.

391.

TORTORELLA ET AL EBNA2-related resistance towards alpha interferon (IFN-alpha) in Burkitt’s lymphoma cells effects induction of IFNinduced genes but not the activation of transcription factor ISGF-3. Mol. Cell. Biol. 12:4930–36. Erratum. 1993. Mol. Cell. Biol. 13(3):1981 Aman P, A von Gabain. 1990. An Epstein-Barr virus immortalization as sociated gene segment interferes specifically with the IFN-induced anti-proliferative response in human B-lymphoid cell lines. EMBO J. 9:147–52 Clarke PA, MB Mathews. 1995. Interactions between the double-stranded RNA binding motif and RNA: definition of the binding site for the interferoninduced protein kinase DAI (PKR) on adenovirus VA RNA. RNA 1:7–20 Park H, MV Davies, JO Langland, HWChang, YS Nam, J Tartaglia, E Paoletti, BL Jacobs, RJ Kaufman, S Venkatesan. 1994. TAR RNA-binding protein is an inhibitor of the interferoninduced protein kinase PKR. Proc. Natl. Acad. Sci. USA 91:4713–17 Benkirane M, C Neuveut, RF Chun, SM Smith, CE Samuel, A Gatignol, KT Jeang. 1997. Oncogenic potential of TAR RNA binding protein TRBP and its regulatory interaction with RNA-dependent protein kinase PKR. EMBO J. 16:611–24 Chang HW, JC Watson, BL Jacobs. 1992. The E3L gene of vaccinia virus encodes an inhibitor of the interferon-induced, double-stranded RNA-dependent protein kinase. Proc. Natl. Acad. Sci. USA 89:4825–29 Romano PR, F Zhang, SL Tan, MT Garcia-Barrio, MG Katze, TE Dever, AG Hinnebusch. 1998. Inhibition of doublestranded RNA-dependent protein kinase PKR by vaccinia virus E3: role of complex formation and the E3 N-terminal domain. Mol. Cell. Biol. 18:7304–16 McInnes CJ, AR Wood, AA Mercer. 1998. Orf virus encodes a homolog of the

392.

393.

394.

395.

396.

397.

398.

399.

400.

vaccinia virus interferon-resistance gene E3L. Virus Genes 17:107–15 Mossman K, P Nation, J Macen, M Garbutt, A Lucas, G McFadden. 1996. Myxoma virus M-T7, a secreted homolog of the interferon-gamma receptor, is a critical virulence factor for the development of myxomatosis in European rabbits. Virology 215:17–30 Karp CL, M Wysocka, LM Wahl, JMmAhearn, PJ Cuomo, B Sherry, G Trinchieri, DE Griffin. 1996. Mechanism of suppression of cell-mediated immunity by measles virus. Science 273:228– 31. Erratum. 1997. Science 275:1053 Cao JX, PD Gershon, DN Black. 1995. Sequence analysis of HindIII Q2 fragment of capripoxvirus reveals a putative gene encoding a G-protein-coupled chemokine receptor homologue. Virology 209:207–12 Strockbine LD, JI Cohen, T Farrah, SD Lyman, F Wagener, RF DuBose, RJ Armitage, MK Spriggs. 1998. The Epstein-Barr virus BARF1 gene encodes a novel, soluble colony-stimulating factor-1 receptor. J. Virol. 72:4015–21 Chee MS, SC Satchwell, E Preddie, KM Weston, BG Barrell. 1990. Human cytomegalovirus encodes three G proteincoupled receptor homologues. Nature 344:774–77 Isegawa Y, Z Ping, K Nakano, N Sugimoto, K Yamanishi. 1998. Human herpesvirus 6 open reading frame U12 encodes a functional beta-chemokine receptor. J. Virol. 72:6104–12 Ahuja SK, PM Murphy. 1993. Molecular piracy of mammalian interleukin-8 receptor type B by herpesvirus saimiri. J. Biol. Chem. 268:20691–94 Graham KA, AS Lalani, JL Macen, TL Ness, M Barry, LY Liu, A Lucas, I ClarkLewis, RW Moyer, G McFadden. 1997. The T1/35kDa family of poxvirussecreted proteins bind chemokines and modulate leukocyte influx into virusinfected tissues. Virology 229:12–24 Lalani AS, J Masters, K Graham, L Liu,


401.


402.

403.

404.

405.

406.

407.

A Lucas, G McFadden. 1999. Role of the myxoma virus soluble CC-chemokine inhibitor glycoprotein, M-T1, during myxoma virus pathogenesis. Virology 256:233–45 Massung RF, V Jayarama, RW Moyer. 1993. DNA sequence analysis of conserved and unique regions of swinepox virus: identification of genetic elements supporting phenotypic observations including a novel G protein-coupled receptor homologue. Virology 197:511– 28 Rode HJ, W Janssen, A Rosen-Wolff, JJ Bugert, P Thein, Y Becker, G Darai. 1993. The genome of equine herpesvirus type 2 harbors an interleukin 10 (IL10)like gene. Virus Genes 7:111–16 Fleming SB, CA McCaughan, AE Andrews, AD Nash, AA Mercer. 1997. A homolog of interleukin-10 is encoded by the poxvirus orf virus. J. Virol. 71:4857–61 Yao Z, WC Fanslow, MF Seldin, AM Rousseau, SL Painter, MR Comeau, JI Cohen, MK Spriggs. 1995. Herpesvirus Saimiri encodes a new cytokine, IL-17, which binds to a novel cytokine receptor. Immunity 3:811–21 Comeau MR, R Johnson, RF DuBose, M Petersen, P Gearing, T VandenBos, L Park, T Farrah, RM Buller, JI Cohen, LD Strockbine, C Rauch, MK Spriggs. 1998. A poxvirus-encoded semaphorin induces cytokine production from monocytes and binds to a novel cellular semaphorin receptor, VESPR. Immunity 8:473–82 Mold C, BM Bradt, GR Nemerow, NR Cooper. 1988. Epstein-Barr virus regulates activation and processing of the third component of complement. J. Exp. Med. 168:949–69 Albrecht JC, J Nicholas, KR Cameron, C Newman, B Fleckenstein, RW Honess. 1992. Herpesvirus saimiri has a gene specifying a homologue of the cellular membrane glycoprotein CD59. Virology 190:527–30

925

408. Kapadia SB, H Molina, V van Berkel, SH Speck, HW Virgin. 1999. Murine gammaherpesvirus 68 encodes a functional regulator of complement activation. J. Virol. 73:7658–70 409. Shchelkunov SN 1995. Functional organization of variola major and vaccinia virus genomes. Virus Genes 10:53–71 410. Isaacs SN, EJ Wolffe, LG Payne, B Moss. 1992. Characterization of a vaccinia virus-encoded 42-kilodalton class I membrane glycoprotein component of the extracellular virus envelope. J. Virol. 66:7217–24 411. Engelstad M, GL Smith. 1993. The vaccinia virus 42-kDa envelope protein is required for the envelopment and egress of extracellular virus and for virus virulence. Virology 194:627–37 412. MacCormac LP, JE Grundy. 1996. Human cytomegalovirus induces an Fc gamma receptor (Fc gammaR) in endothelial cells and fibroblasts that is distinct from the human cellular Fc gammaRs. J. Infect. Dis. 174:1151–61 413. Johnson DC, MC Frame, MW Ligas, AM Cross, ND Stow. 1988. Herpes simplex virus immunoglobulin G Fc receptor activity depends on a complex of two viral glycoproteins, gE and gI. J. Virol. 62:1347–54 414. Baucke RB, PG Spear. 1979. Membrane proteins specified by herpes simplex viruses. V. Identification of an Fc-binding glycoprotein. J. Virol. 32:779–89 415. Para MF, L Goldstein, PG Spear. 1982. Similarities and differences in the Fcbinding glycoprotein (gE) of herpes simplex virus types 1 and 2 and tentative mapping of the viral gene for this glycoprotein. J. Virol. 41:137–44 416. Thale R, P Lucin, K Schneider, M Eggers, UH Koszinowski. 1994. Identification and expression of a murine cytomegalovirus early gene coding for an Fc receptor. J. Virol. 68:7757–65 417. Oleszak EL, J Kuzmak, B Hogue, R Parr, EW Collisson, LS Rodkey, JL Leibow-

926

TORTORELLA ET AL


itz.. 1995. Molecular mimicry between Fc receptor and S peplomer protein of mouse hepatitis virus, bovine corona virus, and transmissible gastroenteritis virus. Hybridoma 14:1–8

418. Litwin V, W Jackson, C Grose. 1992. Receptor properties of two varicellazoster virus glycoproteins, gpI and gpIV, homologous to herpes simplex virus gE and gI. J. Virol. 66:3643–51


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annu. Rev. Immunol. 2000. 18:927–974

DNA VACCINES: Immunology, Application, and Optimization* Sanjay Gurunathan1, Dennis M. Klinman2, and Robert A. Seder1 1


Laboratory of Clinical Investigation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892 (Corresponding author); 2 Section of Retroviral Immunology, Center for Biologics Evaluation and Research, US Food and Drug Administration, Bethesda, Maryland 20892; e-mail: [email protected]

Key Words CpG sequences, plasmid DNA Abstract The development and widespread use of vaccines against infectious agents have been a great triumph of medical science. One reason for the success of currently available vaccines is that they are capable of inducing long-lived antibody responses, which are the principal agents of immune protection against most viruses and bacteria. Despite these successes, vaccination against intracellular organisms that require cell-mediated immunity, such as the agents of tuberculosis, malaria, leishmaniasis, and human immunodeficiency virus infection, are either not available or not uniformly effective. Owing to the substantial morbidity and mortality associated with these diseases worldwide, an understanding of the mechanisms involved in generating long-lived cellular immune responses has tremendous practical importance. For these reasons, a new form of vaccination, using DNA that contains the gene for the antigen of interest, is under intensive investigation, because it can engender both humoral and cellular immune responses. This review focuses on the mechanisms by which DNA vaccines elicit immune responses. In addition, a list of potential applications in a variety of preclinical models is provided.

INTRODUCTION The concept of vaccination was demonstrated over 200 years ago when Jenner showed that prior exposure to cowpox could prevent infection by smallpox. Over the last century, the development and widespread use of vaccines against a variety of infectious agents have been a great triumph of medical science. Despite these successes, vaccines for many pathogens throughout the world, including human immunodeficiency virus (HIV) and the agents of malaria and tuberculosis, are *The US government has the right to retain a nonexclusive, royalty-free license in and to any copyright covering this paper.

927


928

GURUNATHAN n KLINMAN n SEDER

either ineffective or unavailable. One of the impediments to successful vaccination against the aforementioned infectious agents is that they likely require a cellular immune response for protection. In this regard, although all currently licensed vaccines are efficient at inducing antibody responses, only vaccines derived from live attenuated organisms induce cellular immunity efficiently. It should be noted, however, that widespread use of live attenuated vaccines might be precluded by practical constraints such as manufacturing and safety concerns. Thus, the demonstration over the last decade that plasmid DNA vaccines can induce both humoral and cellular immune responses in a variety of murine and primate disease models has engendered considerable excitement in the vaccine community. The historical basis for DNA vaccines rests on the observation that direct in vitro and in vivo gene transfer of recombinant DNA by a variety of techniques resulted in expression of protein. These approaches included retroviral gene transfer, using formulations of DNA with liposomes or proteoliposomes (1–3), calcium phosphate-coprecipitated DNA (4), and polylysine-glycoprotein carrier complex (5). In the seminal study by Wolff et al of ‘‘plasmid or naked’’ DNA vaccination in vivo, it was shown that direct intramuscular inoculation of plasmid DNA encoding several different reporter genes could induce protein expression within the muscle cells (6). This study provided a strong basis for the notion that purified/ recombinant nucleic acids (‘‘naked DNA’’) can be delivered in vivo and can direct protein expression. These observations were further extended in a study by Tang et al (7), who demonstrated that mice injected with plasmid DNA encoding hGH could elicit antigen-specific antibody responses. Subsequently, demonstrations by Ulmer et al (8) and Robinson et al (9) that DNA vaccines could protect mice or chickens, respectively, from influenza infection provided a remarkable example of how DNA vaccination could mediate protective immunity. The mouse study further documented that both antibody and CD8` cytotoxic T-lymphocyte (CTL) responses were elicited (8), consistent with DNA vaccines stimulating both humoral and cellular immunity. DNA vaccination might provide several important advantages over current vaccines (Table 1). (a) DNA vaccines mimic the effects of live attenuated vaccines in their ability to induce major histocompatibility complex (MHC) class Irestricted CD8` T-cell responses, which may be advantageous compared with conventional protein-based vaccines, while mitigating some of the safety concerns associated with live vaccines. (b) DNA vaccines can be manufactured in a relatively cost-effective manner and stored with relative ease, eliminating the need for a ‘‘cold chain’’ (the series of refrigerators required to maintain the stability of a vaccine during its distribution). In light of these potential advantages, this review focuses on the mechanisms by which DNA vaccines induce immune responses. In addition, we have provided a table of diseases for which DNA vaccines are effective in animal models. For additional information on DNA vaccination with an emphasis on viral infections, we refer to the recent review

929

DNA VACCINES TABLE 1 Comparative analysis of various vaccine formulations

Immune response Humoral Cellular

B cells CD4` CD8`

Antigen presentation Memory

Humoral Cellular

DNA vaccine

Live attenuated

Killed/protein subunit

``` ```Th1a `` MHC class I & II ``` ``

``` `/1Th1 ``` MHC class I & II ``` ```

``` `/1 Th1 1 MHC class II ``` `/1

````

`

``

``` ``` ```b

` ` ``c

` ``` ````


Manufacturing Ease of development and production Cost Transport/Storage Safety a

Th2 responses can be induced by gene gun immunization in mice. Data available only from Phase 1 trials. Live/attenuated vaccines may be precluded for use in immunocompromised patients and certain infections such as HIV.

b c

by Robinson & Pertmer (10). Finally, a comprehensive web site on DNA established by Whalen (10a) can be found at www.genweb.com/dnavax.html.

REQUIREMENTS FOR A DNA VACCINE VECTOR There are several factors that influence the type of immune response induced by DNA vaccination. This section outlines the two basic elements of a DNA vaccine that influence the transcription and modulation of the immune responses. A more comprehensive discussion of how the plasmid DNA can be optimized for a specific type of immune response is presented in a later section.

Expression Plasmid Backbone DNA vaccines consist of the foreign gene of interest cloned into a bacterial plasmid (Figure 1). The plasmid DNA is engineered for optimal expression in eukaroytic cells. Requisites include (a) an origin of replication allowing for growth in bacteria (the E. coli:Co1EI origin of replication in PUC plasmids is most commonly used for this purpose, because it provides large copy numbers in bacteria with high yields on purification); (b) a bacterial antibiotic resistance gene (this allows for plasmid selection during bacterial culture; the ampicillin resistance gene, the most common resistance gene used for studies in mice, is precluded for


930


Figure 1 Schematic for the basic requirements of a plasmid DNA vector. The essential features for a plasmid DNA vector include a transcriptional unit, which consists of a viral promoter (i.e. cytomegalovirus), an insert containing the antigen, and transcription/termination sequences (Poly A). The other essential components include a bacterial origin of replication and antibiotic resistance gene, allowing for growth and selection in bacteria. The adjuvant properties of a plasmid vector are highly influenced by the number of CpG motifs within the plasmid backbone.

use in humans, and kanamycin is often used); (c) a strong promoter for optimal expression in mammalian cells (for this, virally derived promoters such as from cytomegalovirus (CMV) or simian virus 40 provide the greatest gene expression); and (d) stabilization of mRNA transcripts, achieved by incorporation of polyadenylation sequences such as bovine growth hormone (BGH) or simian virus 40.

Contribution of Immunostimulatory Cytidine-PhosphateGuanosine Motifs In addition to the requirements outlined above, DNA vaccines also contain specific nucleotide sequences that play an important role in the immunogenicity of these vaccines. Yamamoto et al were the first to report that synthetic oligodeoxynucleotides (ODNs) with sequences patterned after those found in bacterial DNA could activate natural killer cells to secrete interferon (IFN)-c (11). They hypothesized that palindromic sequences present in the synthetic ODNs were


DNA VACCINES

931

responsible for this stimulation. More recently, it was shown that a specific sequence motif present in bacterial DNA elicited innate immune responses characterized by the production of interleukin (IL)-6, IL-12, tumor necrosis factor (TNF)-a, IFN-c, and IFN-a (12–14). This motif consists of an unmethylated cytidine-phosphate-guanosine (CpG) dinucleotide with appropriate flanking regions. In mice, the optimal flanking region is composed of two 5’ purines and two 3’ pyrimidines (14, 15). Such motifs are 20-fold more common in microbial than mammalian DNA, owing to differences in frequency of use and the methylation pattern of CpG dinucleotides in prokaryotes vs eukaryotes (16, 17). CpG motifs directly activate B cells to proliferate or secrete antibody (15). In addition, they directly induce professional antigen-presenting cells [APCs; i.e. macrophages and dendritic cells (DCs)] to secrete cytokines (12, 18, 19). Natural killer (NK) cells are indirectly activated by CpG motifs through cytokines induced by APCs (20). Finally, T cells are also stimulated directly or indirectly by CpG motifs, depending on their baseline activation state (21). Because CpG motifs have such a prominent role in enhancing the immune response after DNA vaccination, a more detailed summary of their role is highlighted below in the section discussing approaches to vaccine optimization.

IMMUNOLOGY OF DNA VACCINATION An important first step in the rational design of a vaccine is to understand the immune correlates of protection. For most viral and bacterial infections, primary protection is mediated by a humoral immune response (production of antibodies). For intracellular infections such as Mycobacterium tuberculosis, Leishmania major, and other parasites, protection is mediated by cellular immunity. Moreover, for some diseases [e.g. human immunodeficiency virus (HIV) infection, herpes, and malaria], both humoral and cellular responses are likely to be required. The cellular immune response comprises primarily CD4` and CD8` T cells. These cells recognize foreign antigens that have been processed and presented by APCs in the context of MHC class II or class I molecules, respectively. Exogenous antigens provided by killed/inactivated pathogens, recombinant protein, or protein derived from live vaccines are taken up by APCs by phagocytosis or endocytosis and presented by MHC class II molecules to stimulate CD4` T cells, which can help generate effective antibody responses. In contrast, MHC class I molecules associate with antigens synthesized within the cytoplasm of the cell (with rare exceptions) and are generally elicited by live or DNA vaccines. From an immunologic standpoint, based on the broad range of effector cells generated and the memory responses they induce, live attenuated vaccines represent the vaccines of choice for those diseases requiring both humoral and cellular responses (Table 1). From a practical and safety standpoint, however, live or live attenuated vaccines raise several issues that can preclude their widespread use. In this regard, DNA vaccines—which resemble live attenuated vaccines in their ability to induce

932


both humoral and cellular responses—may prove to be useful alternatives. In the next section, the mechanism by which DNA vaccines induce specific types of immunity is discussed.


Mechanism of Antigen Presentation One intriguing aspect of DNA vaccination involves the mechanism by which the antigen encoded by the foreign gene introduced into the bacterial plasmid is processed and presented to the immune system. Studies demonstrate that the quantity of antigen produced in vivo after DNA inoculation is in the picogram to nanogram range. Given the relatively small amounts of protein synthesized by DNA vaccination, the most likely explanation for the efficient induction of a broad-based and sustained immune response is the immune-enhancing properties of the DNA itself (i.e. CpG motifs) and/or the type of APC transfected. There are at least three mechanisms by which the antigen encoded by plasmid DNA is processed and presented to elicit an immune response: (a) direct priming by somatic cells (myocytes, keratinocytes, or any MHC class II-negative cells); (b) direct transfection of professional APCs (i.e. DCs); and (c) cross-priming in which plasmid DNA transfects a somatic cell and/or professional APC and the secreted protein is taken up by other professional APCs and presented to T cells. These three mechanisms are highlighted in Figure 2. Direct Transfection of Professional Antigen-Presenting Cells—Bone MarrowDerived Cells Directly Mediate Cellular Immune Responses after DNA Vaccination Several elegant studies with bone marrow-chimeric mice have conclusively demonstrated that bone marrow-derived APCs play a key role in the induction of the immune response after DNA vaccination. In these studies, parent into F1 bone marrow-reconstituted mice created a mismatch between the haplotypes of somatic cells and bone marrow-derived cells. The immune response generated on subsequent DNA immunization was found to be restricted to the haplotype of reconstituted bone marrow, providing conclusive evidence that bone marrow-derived cells were responsible for priming immune responses after DNA vaccination (22–24). Dendritic Cells Are the Principal Cells Initiating the Immune Response after DNA Vaccination The above findings were further extended to evaluate the cellular mechanisms responsible for the activation of T cells after DNA immunization. In particular, studies were aimed at defining the specific type of APCs regulating the immune response after DNA vaccination. The first study to address this question showed that isolated DCs but not B cells or keratinocytes from DNAvaccinated mice were able to efficiently present antigen to T cells in vitro (25). Moreover, in the same study it was estimated that only a small proportion of the DCs (0.4%) was transfected with plasmid DNA (25). Similar results were obtained in two additional studies in which the injection of DNA led to direct

DNA VACCINES


A. Direct Transfection of Bone-Marrow Derived APCs

B. Direct Transfection of Somatic Cells

Figure 2 Legend under Figure 2c.

933

934



C. Cross Priming

Figure 2 Mechanisms of antigen presentation after DNA immunization. A. Bone marrowderived antigen-presenting cells (APCs)s mediate immune responses after DNA vaccination. Injection of plasmid DNA leads to direct transfection of a small number of DCs that present antigen to T cells. B. Direct priming of immune responses by somatic cells (myocytes, keratinocytes, or any major histocompatibility complex class-II–negative cells). This result could occur after injection of plasmid DNA into muscle or skin, leading to protein production and presentation to T cells by the somatic cells themselves. Alternatively, C. Protein production by transfected somatic cells may be taken up by professional APCs, leading to T-cell activation (cross-priming).

transfection of small numbers of DCs (26, 27). It is notable that in both of these studies there was general activation and migration of large numbers of DCs that were not transfected. Finally, direct in vivo visualization of antigen-expressing DCs from draining lymph nodes after gene gun vaccination was demonstrated in a separate study in which gold particles and protein expression from a reporter gene could be co-localized within a cell that had morphologic indices consistent with a DC (28). Taken together, the preponderance of data clearly demonstrates that DCs play a key role in induction of the immune response after DNA vaccination. Furthermore, these data suggest that the predominant contribution to priming immune responses after DNA vaccination involves a small number of directly transfected DCs. Additionally, as noted above, the question arises whether the enhancement in the number of migrating DCs not directly transfected with DNA, seen in many studies, could also present antigen via additional mechanisms such as cross-priming (see below).


DNA VACCINES

935

Direct Priming of Somatic Cells—Skin vs Muscle The initial seminal study by Wolff et al (6) demonstrating the success of ‘‘plasmid or naked’’ DNA vaccination in vivo involved the direct intramuscular inoculation of plasmid DNA, leading to expression of protein within the transfected cell. The other important study by Ulmer et al (8), showing that direct intramuscular inoculation of plasmid DNA induced a strong CD8` CTL to influenza nucleoprotein, provided the first evidence that cellular responses could be induced in vivo and have a potentially important protective role. These and several additional studies suggested that muscle cells were critically involved in the initiation of immune responses after DNA vaccination. One conceptual difficulty with this premise is that, although muscle cells express MHC class I molecules, they do not express other cell surface molecules (i.e. CD80 and CD86) that are critical in optimizing T-cell priming. Therefore, they are not likely to be as efficient at presenting antigen as are DCs. This difficulty raised a question about the exact role that muscle cells play in the induction of cellular immune responses after intramuscular DNA vaccination. To address whether expression of antigen by myocytes was sufficient to induce protective immunity in vivo, it was shown that transfer of stably transfected myoblasts expressing an influenza nucleoprotein protected mice from infectious challenge (29). Although these data suggested that expression of viral protein by muscle cells in vivo is sufficient for CTL-mediated protection, the question of whether CTLs were induced directly by myocytes expressing protein directly or by transfer of protein from myocytes to professional APCs (cross-priming) remained open. Experiments were undertaken to directly test whether muscle cells alone are sufficient to prime immune responses. In one study, using bone marrow chimeras to examine the contribution of bone marrow- and non-bone marrow–derived cells to CTL priming, it was shown that antigen-specific CTL responses could be induced by non-bone marrow–derived (muscle) cells only when mice were vaccinated with DNA encoding the antigen and CD86 (30). By contrast, in a separate study with a plasmid DNA encoding a different antigen, it was shown that plasmids encoding CD86, IL-12, or granulocyte/macrophage colony-stimulating factor DNA failed to induce muscle cells to prime for CTL responses (31). Taken together, although these studies both show that muscle cells alone are not efficient at priming immune responses, one study does suggest that muscle cells expressing CD86 are sufficient to induce a response. Finally, the finding that removing the muscle immediately (within 10 min) after immunization does not alter the subsequent immune response (32) provides additional evidence that injected plasmid DNA is likely to gain access to the lymphatic or circulatory system, thus obviating the need for transfection of muscle cells at the site of injection. For other somatic cells, it has been shown that keratinocytes and Langerhans cells constitute the major cell types transfected by plasmid DNA after injection into the skin (33, 34). In contrast to the data mentioned above regarding removing muscle, immediate removal of skin after DNA vaccination prevented development of immune responses (32). Moreover, in a separate study, it was shown that

936



transplantation of vaccinated skin ,12 h postvaccination could elicit an immune response in naive animals (35). By contrast, little or no immune response could be initiated when the period of transplantation exceeded 24 h. These data suggest that cells that migrated from the epidermis within 24 h of immunization induced the primary immune response after DNA vaccination. Finally, it was shown that the magnitude of the primary immune response increased when the vaccination site was left intact (35). Taken together, these data suggest that antigen-expressing nonmigratory cells such as keratinocytes may continue to produce antigen to augment the immune response (27, 35). Cross-Priming As discussed above, secreted or exogenous proteins undergo endocytosis or phagocytosis to enter the MHC class II pathway of antigen processing to stimulate CD4` T cells. Endogenously produced proteins/peptides (e.g. viral antigens) are presented to the immune system through an MHC class Idependent pathway to stimulate naive CD8` T cells. Although peptides derived from exogenous sources are generally excluded from presentation on MHC class I molecules, there are now several examples showing that this can occur in vivo (36–40). Moreover, the concept of cross-priming, in which triggering of CD8` T-cell responses can occur without de novo antigen synthesis within the APCs, provides an additional mechanism by which DNA immunization can enhance immune responses. During cross-priming, antigen or peptides (both MHC class I and II) generated by somatic cells (myocytes or keratinocytes) can be taken up by professional APCs to prime T-cell responses. The demonstration that transfer of myoblasts expressing an influenza nucleoprotein into F1 hybrid mice induced CTL responses restricted by the MHC haplotype of the recipient mice provided the first evidence that transfer of antigen from myocytes to professional APCs can occur in vivo in the absence of direct transfection of bone marrow–derived cells (29, 41). In addition, cross-priming can occur when professional APCs process secreted peptides or proteins from somatic cells and/or other APCs by phagocytosis of either apoptotic or necrotic bodies (42, 43). This is supported by a study showing that cross-priming of DCs occurred when keratinocytes expressing antigen were exposed to irradiation in vitro, leading to cell death (27). In summary, the overwhelming evidence suggests that bone marrow-derived APCs, but not somatic cells, directly induce immune responses after DNA vaccination; however, because somatic cells such as myocytes or keratinocytes constitute the predominant cells transfected after DNA inoculation via muscle or skin injection, respectively, these cells may serve as a reservoir for antigen. Thus, somatic cells can be important in the induction of immune responses via crosspriming and may play a role in augmenting and/or maintaining the response.

Cellular Immunity CD4~ T Helper Cell Responses CD4` T cells play a central role in immune homeostasis. There are at least three major functions that CD4` T cells can mediate. First, activated CD4` T cells have a critical role in promoting B-cell survival


DNA VACCINES

937

and antibody production through CD40L-CD40 interactions (44). Second, CD4` T cells, through production of IL-2 and/or through CD40L-CD40 costimulation, provide helper function to CD8` T cells (45–47). Finally, CD4` T cells secrete a myriad of cytokines that have profound immunoregulatory effects in many disease states. In this regard, it has been demonstrated that activated CD4` T cells can be segregated into two distinct subsets based on their production of certain cytokines (48). For example, T-helper-1 (Th1) cells exclusively produce IFN-c, whereas CD4` T cells that exclusively produce IL-4, IL-5, and IL-13 are designated as T-helper-2 (Th2) cells. Although several factors have been shown to influence the differentiation of Th1- and Th2-type cells, the cytokine milieu present at the time of initial T-cell priming appears to be the most important (48, 49). Thus, the presence of IL-12 facilitates differentiation toward a Th1 phenotype, whereas the presence of IL-4 allows for Th2 differentiation. Because CpG motifs present in bacterial DNA can trigger the immune system to induce a variety of proinflammatory cytokines including IL-12, it would follow that the generation of Th1 responses may be a general property of DNA vaccines. Indeed, DNA vaccination has been successfully applied to several animal models of infection in which induction of a Th1 response correlates with protection (e.g. tuberculosis and leishmaniasis). DNA vaccination has also proven to be successful in a mouse model of respiratory syncytial virus infection, in which it is likely that antibodies correlate with protection. It is important that, in this infection, killed/inactivated vaccines induced a Th2-type response, which was associated with unfavorable pathology and outcome (50). This is a striking example in which DNA vaccination (by preferentially inducing a Th1 response) has a definite advantage over a formalin-inactivated respiratory syncytial virus vaccine by changing the qualitative immune response (51). Additional evidence that DNA vaccination favors a Th1 response stems from the observation that the predominant immunoglobulin (Ig) isotype detected after DNA vaccination is IgG2a (52). Of note, however, is that, under certain circumstances, DNA vaccines can also induce Th2 responses. Perhaps the best example of this involves using the gene gun method of immunization. Pertmer et al (53) first demonstrated increased IL4 production in mice repetitively immunized by gene gun, while production of IFN-c concomitantly decreased. As a further correlate, it was shown that the predominant Ig isotype generated after repetitive gene gun immunization was IgG1, whereas the predominant Ig isotype generated after intramuscular immunization was IgG2a (53). These observations were extended by the work of Feltquate et al (54), who substantiated the finding that different predominant T helper-type cytokines were generated by gene gun versus intramuscular DNA immunization. Preferential Th2 responses occurred whether DNA plasmids on gold beads ‘‘shot’’ into the skin or into surgically exposed muscle. Taken together, these data suggest that the use of the gene gun has a powerful influence on the induction of Th2 response regardless of the route of immunization. A potential explanation for why Th2-type responses are induced by gene gun is that the gun delivers plasmid DNA directly into cells, thus bypassing surface interaction of


938


CpG motifs, present in the plasmid backbone, with the APCs to mediate the release of proinflammatory Th1-type cytokines. Moreover, because DNA vaccines target DCs, it is possible that different methods of immunization could target different subsets of DCs that have been shown to preferentially bias helper T cell responses (55–59). Finally, there is some evidence that the nature of the antigen used (secreted vs intracellular) can preferentially bias T-helper responses (60, 61). It should be noted, however, that these studies used antibody subtypes rather than direct measurement of cytokine production as a surrogate for T-helper responses. The potential of DNA vaccines to strongly influence CD4` T-helper cell responses has several practical implications. For infectious diseases, the ability of DNA vaccines to preferentially generate Th1 responses may be particularly useful for preventing intracellular infections requiring Th1 immunity or for modulating ongoing immune responses to optimize intracellular killing. First, in terms of influencing an ongoing response, it was shown that CpG ODN treatment could strikingly enhance IFN-c and diminish IL-4 production in BALB/c mice that were already infected with Leishmania major, suggesting that Th1-type responses could be induced in the course of inducing ongoing Th2 response (62). Additionally, a Th1 response generated by DNA immunization may prevent or limit an ongoing Th2 response, for example, in allergic or asthmatic diseases. This was demonstrated in a study by Raz et al (63), who showed that a Th2 response (reflected by antigen-specific production of IL-4 and IgE antibody) generated by vaccination with b-galactosidase (b-gal) protein plus alum immunization could be altered (decreased IL-5 and IgE production) when these mice were boosted with plasmid DNA encoding b-gal. These data raise the broader question of whether immunotherapy with DNA vaccines affects already differentiated Th2 cells at a singlecell level or influences naive and/or activated but uncommitted CD4` T cells toward Th1 cytokine production at the population level. In addition to its ability to influence an established Th2 response, in a rat model of allergic hyperresponsiveness, it was demonstrated that injection of plasmid DNA encoding a house dust mite allergen prevented the induction of IgE and reduced airway hyperreactivity (64). In that study, the suppression of allergic responses could be transferred by CD4-depleted T cells. These findings raise the possibility that CD8` T cells can suppress IgE production and confirm the ability of DNA vaccines to induce both MHC class I- and class II-restricted responses (64). Finally, in experimental models of autoimmune disease, it appears that type I cytokine production (IL-12 or IFN-c) correlates with disease progression. Therefore, it might be expected that DNA vaccination would not be useful for preventing or limiting ongoing autoimmune diseases associated with Th1 responses (65), yet it was demonstrated that vaccination of mice with DNA encoding a gene for a pathogenic T-cell receptor (Vb8.2) for experimental autoimmune encephalitis (EAE) actually protected mice. Protection was associated with a reduction in the Th1 response and increase in the Th2 response (66). The mechanism for this novel observation remains to be elucidated.


DNA VACCINES

939

Cytotoxic T-Lymphocyte Responses As noted above, one of the major advantages of DNA vaccines is the ability to generate antigens endogenously, making them accessible to CD8` T cells via an MHC class I pathway (8). Although CD8 responses are also generated by live vaccines, they are difficult to induce with conventional protein-based vaccines. Moreover, owing to the potential safety concerns about certain live viral vaccinations, the induction of CD8` CTL responses after DNA vaccination may represent a principal advantage of this type of vaccine approach. In addition, because plasmid DNA encoding an antigen can be easily modified, this method of vaccination allows for optimization of both the qualitative and quantitative aspects of CTL responses (see section on vaccine optimization below). While it is clear that DNA vaccination is an effective method of inducing CD8` T-cell responses, there are at least two critical issues concerning the ability of this vaccine approach to mimic the responses of those achieved with live viral vaccines. The first point relates to the magnitude of the CTL response, whereas the second relates to the generation of CTL responses against dominant and subdominant epitopes. In three separate studies of DNA vaccination against lymphocytic choriomeningitis virus (LCMV) infection, mice inoculated with DNA encoding an LCMV protein generated no detectable CTL responses before infectious challenge (67–69). DNA-vaccinated mice, however, were protected from challenge with LCMV. Furthermore, in one of these studies, it was shown that mice inoculated with live LCMV had CTL activity that was immediately detectable ex vivo (69). Taken together, these data show that, in the LCMV mouse model, although live viral infection but not DNA vaccination induced a detectable frequency of effector CTLs immediately ex vivo, DNA vaccination did induce low numbers of precursor CTLs that expanded in vivo after infectious challenge sufficient for protection. In contrast to the LCMV model, the frequency of CTL precursors from cells of mice that were vaccinated with plasmid DNA encoding a Sendai virus nucleoprotein were comparable to those elicited by live Sendai virus infection in a previous report (70). Similarly, in a separate study, it was shown that CTLs generated from mice vaccinated with plasmid DNA encoding influenza nucleoprotein were comparable to those derived from mice that were infected with influenza virus (71). Thus, depending on the antigen and viral model system used, DNA vaccination can elicit CTL responses that are similar to live viral infection after short-term in vitro culture. Perhaps the critical issue of whether DNA vaccination is similar to live viral infection will be resolved by comparing the effector CTL responses immediately ex vivo without any further in vitro culturing. Current techniques, using MHC class I-specific tetramers and intracellular cytokine staining, should clarify this question. These issues are relevant to the optimization of vaccines for infections such as HIV infection or malaria, in which a high precursor frequency of effector CTLs at the time of infection may be required to limit dissemination of infection. Although the magnitude of CTLs induced by DNA vaccination may be sufficient for protection after infectious challenge, an additional consideration is

940


whether DNA vaccines can elicit the same breadth of response as that induced by natural infection. In this regard, although the number of epitopes available in a primary CTL response is relatively large after viral infection, the effector CTL responses selected by the host are often limited to a few dominant epitopes. Additionally, responses to subdominant epitopes may be important in mediating an effector role in the absence of CTL responses to a dominant epitope. For instance, two separate studies showed that DNA vaccines encoding an influenza or Sendai virus nucleoprotein were able to elicit CTL responses against both dominant and subdominant epitopes (70, 71). These data suggest that DNA vaccines can elicit broad memory responses to multiple epitopes. In this aspect, DNA vaccines resemble live viral vaccines by inducing a broad precursor CTL frequency and memory.


Humoral Responses Immunization with plasmid DNA can induce a strong antibody response to a variety of proteins in animal species, including mice, non-human primates, and, most recently, human subjects. Moreover, the humoral response generated by DNA vaccination has been shown to be protective in several animal models in vivo. Because conventional protein vaccines also induce protective antibody responses, it is useful to not only review the mechanism by which DNA vaccines induce antibody responses but also highlight potential differences and determine whether DNA vaccination offers any advantages compared with other types of vaccination. Dose Response and Kinetics of Antibody Response Induced by DNA Vaccination With regard to the quantitative aspect of antibody production after DNA vaccination, it was shown in two separate studies with DNA-encoding influenza hemagglutinin antigen that the antibody responses peaked and reached a plateau between 4 and 12 weeks after a single DNA immunization in mice (72, 73). Furthermore, antibody production is increased in a dose-responsive manner with either a single injection or multiple injections of DNA by various routes of immunization (72, 73). Although dosage and frequency of immunizations may affect the kinetics and magnitude of the response, it is interesting that single or multiple injections with an optimal dose of DNA did not significantly affect the amount of antibody produced once a plateau had been reached (72, 73). Finally, although the duration of the antibody response can be long lived [significant serum levels were present up to 1.5 years postvaccination (34, 72)], this duration is highly variable and depends on the model system and vaccine used. Comparison of Antibody Responses between DNA Vaccination and Protein or Live Infection: Effects on Avidity, Magnitude, Isotype, and Induction of Neutralizing Antibody As noted above, peak antibody responses after DNA vaccination occur 4–12 weeks postvaccination. Most studies comparing antibody production after DNA, protein, and live virus immunization use this time range.


DNA VACCINES

941

In comparing humoral responses in mice vaccinated with DNA H1 hemagglutinin and mice immunized with a sublethal viral challenge with H1N1 influenza, the amount of antibody produced was substantially greater and peaked more rapidly in the sublethally infected mice than in the DNA-vaccinated mice (74). Similarly, in a separate study comparing the antibody response to DNA encoding the hemagglutinin (HA) antigen and live influenza infection, the antibody titers in mice vaccinated with live influenza were higher than in DNA-vaccinated mice, although this result was seen with only certain antibody isotypes (72). In comparing the antibody response elicited by vaccination with DNA and protein, it was shown that antibody titers and avidity were significantly lower in mice vaccinated with DNA encoding a malarial surface protein than in those vaccinated with the protein alone (75). By contrast, in one study directly comparing the kinetics of antibody response after vaccination with DNA-encoding ovalbumin (OVA) and OVA protein, there did not appear to be a difference in total OVAspecific antibody production when DNA was administered intradermally at 2 or 4 weeks postvaccination (76). In this study, antibody induced by DNA had a higher avidity than that induced by protein. The antibody subtypes induced by DNA vaccination include IgG, IgM, and IgA. Moreover, as noted in the previous section, DNA vaccination generally enhances Th1 cytokine production. Because cytokines such as IL-4 and IFN-c can direct IgG1, and IgG2a production, respectively, it follows that the subclass of antibodies generated by pDNA vaccination will be biased toward IgG2a production. While this appears to be a general property of DNA vaccination in mice, it has been shown that DNA encoding secreted antigen generated higher levels of IgG1 than did membrane-bound antigen (60). Moreover, as noted above, the route of DNA vaccination (gene gun) can also preferentially bias toward IgG1 production (54). Finally, the ability of DNA vaccines to generate neutralizing antibodies suggests that antigen expressed in vivo after DNA vaccination can assume a native configuration. In this regard, the ability of plasmid DNA encoding influenza HA to generate neutralizing antibody suggests that HA was present in its native form, because the epitopes of HA that are recognized by these antibodies are formed by noncontiguous regions within HA. Thus, DNA vaccines may generate antibody responses that more closely resemble those seen after natural infection and provide a potential advantage over conventional protein vaccines. Since some recombinant proteins may lack linear determinants or conformational epitopes required for efficient generation of neutralizing antibodies. Data to support this were shown for mice immunized with a DNA vaccine encoding HIV gp120, in which sera contained antibodies reactive to linear epitopes within the V3 region of gp120 whereas sera from mice immunized with recombinant gp120 contained much lower levels of V3-specific antibodies (77). Similar results were observed with a rabbit model of papilloma virus (CRPV) infection. In this model, immunization with plasmid DNA encoding a major capsid protein L1 induced neutralizing antibody. In that study, adsorption experiments with native L1 or

942


denatured LI protein suggested that vaccination with plasmid DNA encoding LI elicited conformationally specific neutralizing antibody (78).


Memory Immunity The hallmark of any successful vaccine is the ability to induce long-term memory. Current vaccines—whether live attenuated or protein subunit—are successful at generating durable humoral responses. For diseases requiring cellular immunity such as parasitic, mycobacterial, and certain viral infections, however, it is not yet clear how memory responses are generated and maintained after vaccination. Regarding humoral immunity, it has been shown that mice vaccinated with DNA encoding an HA antigen had levels of anti-HA antibodies comparable with or greater than those from convalescent sera of previously infected mice that persisted over 1 year (34, 72). In other studies, however, plasmid DNA encoding a nucleoprotein of the LCMV virus administered by gene gun (69) or intramuscularly (67) either failed to give appreciable antibody responses before challenge or the responses had waned by 4 months postimmunization. Taken together, these results indicate that, while DNA vaccination can be effective at inducing longterm antibody responses, this effect may depend on the type of antigen used in the vaccine. In terms of cellular immunity, it was recently shown that the frequency of antigen-specific CD4` T cells as measured by proliferation remained elevated for #40 weeks postvaccination. Of interest, antigen was detectable only for 2 weeks postvaccination in DCs in the draining lymph nodes but for #12 weeks in keratinocytes (27). Moreover, a functional assay performed in vivo appeared to demonstrate no source of antigen present in the spleen or lymph nodes 20 days postvaccination. Taken together, these data showed that antigen-specific CD4` T cells are activated in the draining lymph nodes and migrate to the spleen, where they can persist for up to 40 weeks in the absence of detectable antigen (27). More definitive evidence showing that DNA can induce long-lived Th1 effector responses in vivo involved a mouse model of L. major infection. This study demonstrated that vaccination with plasmid DNA encoding a specific leishmanial antigen is more effective than vaccination with leishmanial protein plus IL-12 protein in maintaining antigen-specific Th1 cells capable of controlling L. major infection (79). These data provided evidence that DNA vaccination can induce long-term Th1 responses and suggested that DNA vaccination may be more effective than vaccination with protein plus adjuvant (i.e. IL-12). Reasons for the enhanced efficacy of DNA vaccination over protein and adjuvant may include low levels of persistent antigen and/or IL-12 induced by the CpG in the plasmid DNA. Induction of Long-Term Cytotoxic–T-Lymphocyte Responses after DNA Vaccination Although few studies have assessed the induction of CD8` T-cell responses for prolonged periods after DNA vaccination, there is a report showing

DNA VACCINES

943


that CTL responses could be observed #68 weeks after intradermal injection of DNA encoding a nucleoprotein from influenza virus (34). In a separate study, DNA-primed CTL responses to hepatitis B virus envelope proteins could be detected for #4 months post-DNA injection (80). It should be noted that these responses were detected only after cells were re-cultured in vitro for several days and then tested. Thus, the relative effectiveness of DNA vaccination for generating fresh, memory effector CTL responses remains to be determined. One other potentially important finding relating to DNA vaccination and induction of memory CD8` T-cell responses is that CpG motifs are potent stimulators of type-1 interferons. It was originally reported by Tough et al that IFN-a enhances the proliferation of CD8` T cells expressing a surface marker, consistent with a memory phenotype (81). More recent work showed that CpG DNA appeared to stimulate T cells by inducing type-1 interferons from APCs (82). Although these studies are important in establishing a role for IFN-a in regulating activation of CD8` T cells, a functional in vivo role for these cells remains to be elucidated. Mechanisms by Which DNA Vaccinations Induce Sustained Humoral and Cellular Immune Responses There is evidence that long-lived antigen-specific proliferative responses that are induced by DNA vaccination are maintained in the absence of detectable antigen (27). In contrast, one of the original studies on DNA vaccination by Wolff et al showed that intramuscular inoculation of plasmid DNA encoding several different reporter genes resulted in protein expression for .1 year (83). These data raise several possibilities as to how DNA vaccines induce long-term responses: (a)antigen is continuously present at low levels sufficient for antigen presentation but below the limit of detection as assessed by polymerase chain reaction or currently available functional assays. Alternatively, plasmid DNA may not be detectable, but synthesized antigen could persist in vivo (i.e. follicular DCs), providing a reservoir to maintain the immune response; (b) plasmid DNA as well as antigen are completely gone, and responses are antigen independent (27); and/or (c) memory cells generated by DNA vaccines differ qualitatively from those achieved by other forms of vaccination such as protein plus adjuvant.

APPROACHES TO VACCINE OPTIMIZATION Because different diseases have specific requirements for protective immunity, a rational approach to vaccine optimization would reflect these distinct requirements. Thus, one of the principal advantages of DNA vaccination is the ease with which plasmid DNA can be manipulated to alter the quantitative and qualitative aspects of the immune response. In this section, we discuss the factors that affect the efficiency of DNA vaccines and highlight how DNA vaccines can be influenced or tailored to generate the desired immune response.

944



Vector Optimization Role of Gene Regulatory Elements and Multicistronic Vectors One of the most important considerations in optimizing a DNA vaccine is the appropriate choice of a vector. The basic requirements for a plasmid vector are described above. It is generally believed that the level of gene expression in vivo obtained after DNA vaccination correlates with the immune response generated. Therefore, several laboratories sought to improve gene expression and immune responses after plasmid DNA vaccination. These approaches included optimizing gene regulatory elements within the plasmid backbone (e.g. promoter-enhancer complex or transcription termination signals) or modifying the plasmid backbone itself to enhance gene expression. As noted above, a requisite for a DNA vaccine vector is a promoter that stimulates a high level of gene expression within mammalian cells. Virally derived promoters have generally provided the greatest gene expression in vivo, whereas eukaryotic promoters are weaker (84, 85). The CMV immediate early enhancer-promoter produced the highest transgene expression in various tissues when compared with other promoters (84, 85). Furthermore, because optimal expression of certain mammalian genes depends on splicing of the mRNA transcript, inclusion of the first intron (intron A)of the immediate early gene from CMV in the promoter-enhancer complex further enhanced expression (86). To study the effects of manipulating transcriptional termination elements on gene expression, several different kinds of termination sequences have been studied. In one study, replacing the BGH transcriptional termination element with a transcriptonal terminator derived from the rabbit b-globin gene improved gene expression (87). Several other modifications that enhance gene expression have been examined. To express two or multiple genes in the same cell, dicistronic or multicistronic vectors with internal ribosome entry sites were studied. These vectors could be particularly useful in constructing multivalent vaccines from two or more different antigens from the same or different pathogens (88). Effects of Manipulating Heterologous Genes on the Immune Response Optimizing codon usage for eukaryotic cells can also enhance expression of antigens. Codon bias has been observed in several species, and the use of selective codons in a particular gene correlates with efficiency of gene expression (89). This correlation was shown by using a plasmid expressing listeriolysin O, in which codons frequently used in murine genes were substituted for the native codons for the encoded antigen. This substitution led to enhanced CTL and protective immunity (90). Similar results were noted in mice, by using the HIV-1 gp120 sequence (91) or gp160 sequence (92). A plasmid may also be engineered so that the encoded protein is either secreted or localized to the interior of the cell. Several studies show that the type and magnitude of the immune response depend on whether an antigen is secreted, bound on the surface of the cell, or retained within the cell. For example, secreted proteins induced higher IgG titers than the same antigen localized either on the


DNA VACCINES

945

cell membrane or within the cell (60, 93–95). It is unclear from these studies how DNA immunization induces antibody production against intracellular, noncytopathic proteins, because B cells require free or membrane-bound linear determinants or conformational epitopes to initiate the process of clonal expansion for efficient antibody production. These concepts suggest that a nonsecreted intracellular antigen would not elicit antibody production (27). The evidence that the nature of the antigen used (secreted vs intracellular) can preferentially bias Thelper responses is less clear. In two separate studies, it was demonstrated that secreted antigens induced a higher IgG1:IgG2a ratio (suggesting a Th2 bias) than did antigens that remained cell associated (membrane anchored or cytosolic); however, these studies analyzed antibody subtypes rather than directly measuring cytokines and thus provide only a surrogate for T-helper responses (60, 93, 95). In a separate study, plasmid DNA expressing either secreted or intracellular antigen induced comparable levels of antigen-specific IFN-c on in vitro stimulation (94). Taken together, these data suggest that cellular localization of the antigen after DNA immunization may play a role in modulating immune responses, although this role may depend on the nature of the antigen and model system used (95).

Optimizing Cytotoxic–T-Lymphocyte Responses Enhancing Delivery into the Major Histocompatibility Complex Class I Pathway CTL responses can be enhanced by engineering the antigen to target specific cellular compartments. An example for this engineering is the use of Nterminal ubiquitination signals, which target the protein to proteosomes, leading to rapid cytoplasmic degradation and presentation via the MHC class-I pathway. In this regard, it was demonstrated that a DNA vaccine encoding b-gal that was fused with ubiquitin was more effective at inducing CTL responses than was a plasmid encoding b-gal alone. The latter construct was also less efficient at inducing antibody responses, suggesting that the transfected gene product was rapidly degraded intracellularly and that processing precluded the release of native polypeptides or proteins for efficient antibody production (96). These results are in agreement with studies in other model systems targeting HIV Nef (97) and LCMV nucleoprotein (98, 99). Another approach is to design vectors that use the E3 leader sequence from adenovirus, which facilitates transport of antigens directly into the endoplasmic reticulum for binding to MHC class-I molecules, bypassing the need for the TAP transporter. The addition of the E3 leader sequence appeared to improve CTL responses for certain antigens (100, 101) but did not improve CTL in other model systems (100). These data suggest that endoplasmic reticulum-targeting of T-cell epitope DNA vaccines may not enhance the immune response for all antigens. Epitope-Specific Responses: Minigenes and Multiple Epitopes Another interesting approach for improving the ability of DNA vaccines to generate cell-


946


mediated responses is to engineer vaccines that elicit epitope-specific CTL responses. Several groups have successfully used minimal-epitope vaccines to induce CTL responses (100–107). Furthermore, it was demonstrated that these minimal-epitope vaccines could function in isolation and when linked to other epitopes in a ‘‘string-of-beads’’ vaccine. This approach may be advantageous, because a combination of antigenic epitopes can generate a broader immune response than a DNA vaccine encoding for a single antigen. Moreover, this approach may be effective in developing a single vaccine against multiple pathogens. In this regard, epitopes from several different pathogens could be combined in a single plasmid DNA vaccine, providing an advantage over a conventional DNA vaccine strategy with a plasmid-encoding antigen(s) against a single pathogen. In a study by Thomson et al, mice vaccinated with a DNA plasmid, encoding multiple contiguous minimal-CTL epitopes derived from five separate viruses and a parasite epitope derived from malarial protein, generated MHC class I–restricted CTL responses to each of these epitopes. Furthermore, these CTLs were protective after infectious viral challenge (104). In a separate study, a novel vector containing a polyepitope construct from HIV and Plasmodium falciparum was also effective in generating CTL responses in mice (103). Inclusion of a helper epitope can also enhance CTL activity after DNA vaccination (108). In a study designed to ascertain whether CTL responses generated by DNA vaccines are dependent on MHC class-II/CD4 help, CTL responses generated against a minimal epitope class-I–restricted OVA peptide were compared with those of a similar construct with the adjacent MHC class-II–restricted epitope. Very low or negligible CTL responses were observed in mice vaccinated with a minimal-epitope MHC class-I–restricted DNA construct. In contrast, mice vaccinated with either a full-length ovalbumin construct or a DNA construct with both MHC class-I and class-II epitopes induced a robust CTL response (108). These observations are in contrast to several studies in which minimal-epitope DNA vaccines generated robust CTL responses. Potential explanations for these differences include the following: (a)the polyepitope vaccines could lead to the assembly of neoepitopes that served to generate MHC class-II help; (b) CpG sequences can potently activate DCs in a nonspecific manner (27) and prime CD8` T cells in the absence of CD4 help; and (c) CpG motifs induce IFN-a, a cytokine shown to be important in expansion of CD8` T cells (82).

Role of Cytosine-Phosphate-Guanosine Motifs Over the past decade, portrayals of DNA as immunologically inert have been challenged. New data indicate that bacterial DNA can trigger and instruct the immune system to respond to danger and plays an important role in host defense. This role includes B-cell activation resulting in antibody production, stimulation of cytokine-producing cells, and activation of the innate immune system. The subsequent identification of CpG motifs present in bacterial DNA as potent immu-

DNA VACCINES

947


nostimulatory molecules has spurred tremendous interest in the development of immune-based therapies and of a new generation of experimental vaccines. Immunostimulatory Properties of Cytosine-Phosphate-Guanosine DNA As noted above, it was recently shown that a specific sequence motif present in bacterial DNA elicits an innate immune response characterized by the production of IL-6, IL-12, TNF-a, and IFN-c. Several lines of evidence suggest that CpG motifs in plasmid vectors contribute to the immunogenicity of DNA vaccines. First, vectors lacking protein-encoding inserts induce cytokine production in vitro in a manner indistinguishable from bacterial DNA (109). Second, when the cytosine of the CpG dinucleotides present in plasmid vectors is selectively methylated with Sss-I CpG methylase, the vaccine’s ability to stimulate cytokine production in vitro and antibody or CTL production in vivo is concomitantly reduced (14, 109). Third, coadministering ODN that contains CpG motifs with an antigenic protein boosts the antibody and cellular response similar to that achieved by DNA vaccination with a plasmid encoding the same antigen (110–112). Indeed, coadministering vector alone (without the antigen-encoding insert) also improves the immune response elicited by DNA vaccines. Presumably, CpG motifs present in the vector act as adjuvants in a fashion similar to CpG ODNs. This observation raises the interesting possibility that higher doses of a DNA vaccine or the coadministration of multiple antigen-encoding plasmids might synergistically boost the immune response to each element of a multicomponent vaccine. Perhaps the strongest evidence the CpG motifs contribute to the immunogenicity of DNA vaccines was provided by Sato et al, who substituted a CpGcontaining ampR gene for a kanR-selectable marker in a b-gal–encoding plasmid. They found that the reengineered plasmid elicited a higher IgG antibody response, more CTLs, and greater IFN-c production than did the original vector (14). The same effect was observed when additional CpG motifs were introduced into the plasmid backbone of the kanR-containing vector, a result subsequently confirmed in several other vectors by other laboratories (52, 109, 113, 114). As noted above, this effect is most apparent when low doses of DNA vaccine are administered, presumably because, at high dose, endogenous CpG motifs in plasmid vectors perform the same function. Thus, additional CpG motifs may decrease the amount of vaccine required to induce an immune response rather than increase the absolute magnitude of that response. Indeed, CpG motifs appear to be limited in their ability to augment antibody and cytokine production in vivo such that too many CpG motifs may actually reduce immunogenicity (114). For example, introducing 16 additional CpG motifs into the plasmid backbone improved the humoral immune response by the DNA vaccines, whereas introducing 50 such motifs was detrimental. The above studies were performed in mice, the animals in which the effects of CpG ODNs were first described. Of interest, the 6-base-pair motif that induces optimal stimulation in mice is less effective when tested on cells of primate origin (human, monkey, or chimpanzee). Thus, efforts to improve the efficacy of DNA vaccines intended for human use would require identification of

948



those sequence motifs that are optimally immunostimulatory in humans. Toward this end, Liang et al (115) identified several ODNs that induced proliferation and Ig secretion of human B cells. They did not, however, systematically examine the size or sequence of the CpG motif that provided optimal immune activation. Ballas et al (116) reported that an AACGTT motif embedded in an ODN at least 15 base pairs in length stimulated the proliferation of human NK cells ; however, it is unclear whether this motif is optimally immunostimulatory. In this context, recent evidence suggests that at least two different human cell types respond to ODN stimulation and that different CpG motifs are required to stimulate these distinct cell populations. These findings suggest that it may be possible to tailor the type of immune response elicited by a DNA vaccine by selectively engineering one, the other, or both types of stimulatory motifs into a vector. Immonosuppressive DNA Motifs Whereas CpG-containing bacterial DNA causes immune stimulation in vivo and in vitro, coadministration of mammalian DNA can block such activation. This suppression may account for the inability of mammalian DNA, which contains CpG motifs (albeit at much lower frequency than bacterial DNA), to stimulate the immune system. Several laboratories have shown that a subset of nonstimulatory ODNs can suppress the immune activation induced by ODNs that contain CpG motifs. Hacker et al (117) showed that an excess of non-CpG ODNs could inhibit the uptake of fluorescein-isothiocyanate– labeled CpG ODNs. This inhibition abrogated the ability of CpG ODNs to induce immune stimulation, interfering with cytokine production and stress kinase activation (117). Recent work by Krieg et al (114) confirmed that the immunostimulatory activity of CpG ODNs could be blocked by certain non-CpG motifs. They showed that eliminating suppressive motifs (tandem repeats of GpC) from the plasmid backbone of a DNA vaccine improved immunogenicity up to threefold. These observations demonstrate the complexity of the interaction between DNA sequence motifs and the immune system. An important feature of CpG motifs is their ability to stimulate multiple types of immune cells. They improve antigen-presenting function by monocytes, macrophages, and DCs, induce proliferation of B cells, and boost antibody production by antigen-activated lymphocytes. Efforts are under way to identify the sequence motifs that are optimally active in humans, to determine whether different motifs can be used to regulate discrete elements of the immune system, and to establish where in the plasmid these immunostimulatory sequences can be introduced to greatest benefit. Presumably, this will include the elimination of suppressive motifs present in the plasmid backbone. These efforts are likely to yield vectors with significantly improved immunostimulatory capacities for clinical use.

Role of Cytokines and Costimulatory DNA Adjuvants Because cytokines or costimulatory cell surface molecules play a crucial role in generation of the effector T-cell subsets and in determining the magnitude of the response, several groups have used plasmid DNA encoding various cytokine or

949

DNA VACCINES

costimulatory molecules to enhance or bias the immune response generated by DNA vaccination. The studies with cytokine-encoding DNA and their effects on humoral and cellular immunity are summarized in Table 2 (211–244). TABLE 2 Cytokine and costimulatory DNA adjuvants Cytokine

Antibody

Cellular response

CTL

References

IL-1

FIgG FIgG2a FIgG FIgG2a FIgG1a FIgG FIgG1

Fproliferation FIFN-c Fproliferation FIFN-c

FCTL

211-213

FCTL

212-219

IL-2


IL-4

IL-5 IL-6 IL-7 IL-8 IL-10 IL-12

IL-15 IL-18 TNF GM-CSF

TGF-b

IFN-c IFN-a B7-1 (CD80) B7-2 (CD86)

FIgG FIgG2a FIgG1 FG fIgG2a FIg2a F or f IgG1a F or f IgGa ?FIgGa FIgG FIgG FIgG FIgG2a and IgG1a FIgG1

FIgG2a F or fIgGa

Fproliferation fDTH FIL-4 5proliferation

156, 212, 213, 216, 218, 219 213 220 217

FIFN-c FNeutrophils fDTH fproliferation FDTH Fproliferation FIFN-c 5Fproliferation Fproliferation Fproliferation Fproliferation FIFN-c FIL-4 fDTH fproliferation fcytokines F or fproliferation FIFN-c, fIL-5

FDTH FProliferation

FCTL

FCTL FCTL FCTL FCTL

FCTL

FCTLa FCTLa

221 213, 219, 222-224 31, 79, 156, 212, 213, 217-219, 225-229 213 213 93, 213 31, 93, 156, 211, 212, 216, 218, 219, 230-233 157, 234

93, 157, 212, 218, 233, 235 236, 237 31, 238-241 31, 238-241 (continued )


950

TABLE 2 (continued) Cytokine and costimulatory DNA adjuvants Cytokine

Antibody

Cellular response

CTL

References

CD40L

FIgG2a FIgG, FIgG1a

FIFN-c

FCTL

226, 242

Fproliferation FIFN-c Fb-chemokines Fproliferation FIFN-c Fb-chemokines Fproliferation

FCTL

243

FCTL

243

ICAM-1 (CD54)

LFA-3


L-selectin CTLA4

FIgG FIgG2a.IgG1 FIgG FIgG1.IgG2a

Fproliferation

244 244

a

Changes in antibody or cellular responses are not in agreement between studies.

Alternative Boost Although DNA vaccination alone can elicit potent humoral and cellular responses to many antigens, it appears that for certain antigens (e.g. HIV envelope proteins and malarial proteins), the immune response generated by DNA vaccination may be suboptimal for protection. In such instances, alternative booster regimens have been shown to be helpful. The most common of these booster regimens have used either recombinant protein or poxviruses. Thus, for HIV, because multiple DNA vaccinations elicit only modest and transient titers of neutralizing antibody (118– 124), there have been many studies evaluating the effects of peptide (125) or protein boosting after DNA vaccination (119, 126–128). In two separate studies that used rhesus macaques, it was shown that antibody production could be substantially increased in monkeys vaccinated with DNA encoding an HIV-1 envelope protein followed by a protein boost (127, 128). In one of these reports, monkeys were protected after an infectious challenge (127). In a separate study, rabbits primed with various HIV-1 env-expressing plasmids had a rapid increase in the titer of antibody after a protein boost; however, the avidity and neutralizing activity rose more slowly. In contrast to HIV, high titers of antibody with good avidity and persistence were induced after DNA vaccination encoding an influenza virus HA glycoprotein without any protein boost (119). Taken together, these data underscore a potential difference between HIV and other viral proteins in requiring a protein boost after DNA vaccination to optimize both the qualitative and quantitative aspects of the humoral response.


DNA VACCINES

951

As noted above, although protein boosting enhanced the antibody response after DNA vaccination, there is little evidence that it affects the cellular immune response. Because cellular immunity might be required for protection against diseases such as HIV infection and malaria, there have been several studies attempting to increase cellular responses after DNA vaccination by using recombinant poxviruses. In several such studies, boosting with recombinant poxvirus substantially enhanced CTL and/or IFN-c responses in mice primed with DNA encoding either a malarial (129–131) or HIV envelope protein (132, 133). It should be noted that, whereas antibody production was also increased after DNA and poxvirus boosting in mice, by using vectors encoding malarial proteins (130), antibody production was not enhanced and eventually declined in rhesus macaques vaccinated with DNA and boosted with fowl pox-encoding HIV proteins (132). To conclude, in both malaria- and HIV-infected rodent and nonhuman primate models, DNA vaccination followed by poxvirus boosting gave consistent and striking increases in cellular immunity. One caveat that may be important with regard to vaccination against diseases requiring both humoral and cellular immunity (i.e. HIV infection) is whether this type of boosting also limits antibody responses.

Modes of Administration Route and Dosage A variety of routes of DNA injection, including intramuscular, intradermal, intravenous, intraperitoneal (134), epidermal delivery by scarification (34), oral (135–138), intranasal (134, 139–144), vaginal (145, 146), and, more recently, noninvasive vaccination to the skin (147) have been studied. The most common immunization routes studied have been intramuscular and, to a lesser extent, subcutaneous or intradermal. DNA is administered in a variety of diluents including distilled water, saline, and sucrose. For intramuscular injections, although some investigators have used agents such as cardiotoxin, bupivacaine, or hypertonic solutions (148, 149) to pretreat the muscle tissue to improve responses, additional studies suggest little benefit. Whereas the optimal dose depends on the particular antigen and model system used, typically, 10 to 100 lg of plasmid DNA is required to elicit responses when administered intramuscularly or subcutaneously. By contrast, immunization of DNA by gene gun often requires 0.1–1 lg of plasmid DNA to induce antibody or CTL responses. Thus, in terms of the amount of DNA used, immunization of plasmid DNA with a gene gun is the most efficient mode of delivery (134); however, as noted above, DNA immunization via gene gun can qualitatively alter the type of immune response that is generated. Although doses of 25 to 100 lg per injection (intramuscularly) are usually sufficient in mice, higher doses appear to be required in primates or humans. In a study of human volunteers given a DNA vaccine encoding a malarial antigen, doses of plasmid DNA in the 500- to 2500-lg range gave enhanced CTL responses (150). Whereas a single vaccination with DNA can induce both an antibody and CTL response in several model systems, both cellular

952



and humoral immune responses are increased by successive boosting (one or two additional immunizations). This requirement for multiple immunizations is well documented for the induction of humoral responses to HIV envelope proteins (118–124). It should be noted, however, that, in one study, antibody responses to HIV-1 gp120 were actually enhanced in rhesus macaques when the number of DNA vaccinations (as delivered by gene gun) was reduced but the interval between immunizations increased, suggesting the importance of a rest period between immunizations (128). Similar results were noted after DNA vaccination (given either by gene gun or intramuscularly) that expressed the circumsporozoite protein from Plasmodium berghei in mice (151). Mucosal Immunization Induction of mucosal immunity by DNA immunization after immunization by several different mucosal routes has been studied. These include application of plasmid DNA intranasally (134, 139–144), intratracheally (134, 152, 153), by aerosol (154, 155), by genital-tract immunization (145, 146), and by oral administration (135–138). In addition, in several studies, plasmid DNA was combined with various immunity-enhancing regimens such as cholera toxin (140, 141), plasmid-encoding cytokines (156), liposomes (139, 152, 154, 155), or other adjuvants (142, 143). There has been great interest in generating specific types of immune responses after mucosal immunization. In autoimmune models of disease, oral administration of protein can lead to immune tolerance. By contrast, for viral infections such as HIV and herpes simplex virus, the generation of potent antibody and/or cellular responses may be critical in mediating protection. Owing to the importance of the mucosal immune response for these diseases, studies were undertaken to compare the immune responses elicited by mucosal immunization with those achieved after systemic immunization. First, for antibody production, although many studies showed that serum IgG responses after mucosal immunization were comparable with those elicited after systemic immunization with the same plasmid constructs (142, 143, 146), other groups have demonstrated that mucosal immunization did not lead to an efficient induction of serum IgG responses (134, 140, 141). Of note is that mucosal immunization was superior to systemic immunization at inducing and sustaining mucosal IgA responses in all studies in which data examining this effect were available (142, 143, 146). Whereas mucosal DNA vaccination was advantageous in generating mucosal IgA responses, it was demonstrated in a murine model of herpes simplex virus infection that, despite the presence of virus-specific IgA at the time of challenge, virus could persist and replicate at the mucosal site of challenge (157). These results suggest either a failure of these immunization regimens to induce an adequate IgA response or the requirement for additional immune mechanisms to control viral replication at the mucosal site. For cellular immunity, the ability of DNA vaccines given mucosally or systemically to induce local mucosal T-cell responses has not been directly demonstrated; however, cellular responses have been studied from spleen cells of mice


DNA VACCINES

953

after vaccination via systemic or mucosal routes. In two of these studies, delayedtype hypersensitivity (DTH) responses and specific cytolytic activity from spleen cells was comparable between the two routes (142, 143). To conclude, determination of whether mucosal or systemic vaccination with DNA affects the cellular response at specific mucosal sites will likely be important for designing vaccines against such pathogens as HSV and HIV. A potentially exciting means of mucosal DNA delivery is the use of microparticles. Plasmid DNA trapped in these biodegradable microparticles, composed of polymers such as polylactice-coglycolides or chitosan, can be administered orally and has been shown to induce both mucosal and systemic immune responses (135, 138). The ability of polylactice-coglycolide–entrapped DNA vaccines to induce protective immune responses to rotavirus challenge after oral administration was demonstrated in two separate studies (136, 137). In addition to uses for infectious pathogens, oral administration of DNA vaccines has also been shown to be useful in treating allergic diseases. Recently, an oral DNA vaccine containing the gene for the main peanut allergen (Arah2) protected mice against peanut-induced anaphylaxis. This protection was correlated with a reduction of IgE (a surrogate for a Th2 response), providing further evidence that DNA vaccination by its preferences to stimulate Th1 responses may have broad clinical applications (138). Finally, delivery of plasmid DNA orally with attenuated enteric bacteria such as Salmonella or Shigella spp. is an active area of investigation (see below). Carrier-Mediated Approaches to Optimizing DNA Vaccines It appears that a majority of the DNA injected intramuscularly is degraded by extracellular deoxyribonucleases (158, 159). It follows that protecting plasmid DNA from extracellular degradation by introducing it directly into target cells should optimize DNA uptake. Several methods of carrier-mediated DNA transfection have been successful. Gene gun Gene gun technology uses a gas-driven biolistic bombardment device that propels gold particles coated with plasmid DNA directly into the skin (7, 33, 134, 160). These gold particles are propelled directly into the cytosol of target cells, resulting in transgene expression levels higher than those obtained by comparable doses of ‘‘naked DNA.’’ This mode of immunization induces protective immunity in several animal models of disease. Liposomes Liposomes are bilayered membranes consisting of amphipathic molecules (polar and nonpolar portions) such as phospholipids, forming unilayered or multilayered (lamellar) vesicles. Unilamellar vesicles have a single bilayer membrane surrounding an aqueous core and are characterized by either being small or large unilamellar vesicles, whereas multilayered vesicles have several lipid bilayers separated by a thin aqueous phase. Because liposomes can be prepared with significant structural versatility based on vesicle surface charge, size,

954



lipid content, and coentrapment of adjuvants, they offer considerable flexibility toward vaccine optimization. The full scope of the use of liposomes to increase the effect of DNA vaccines is currently an active area of investigation. Intramuscular injection of plasmid DNA (hepatitis B surface antigen) entrapped in liposomes elicited 100-fold increased antibody titers and increased levels of both IFN-c and IL-4 when compared with those in animals injected with ‘‘naked DNA’’ (161). A similar result on antibody augmentation was seen when DNA/liposome complexes were administered intranasally (139). Cochleates Cochleates are rigid calcium-induced spiral bilayers of anionic phosolipids. They have a unique structure that is different from that of liposomes. They are relatively stable after lyophilization or in harsh environments. It is believed that, on contact with target cell membrane, a fusion event occurs between the membrane and the outer layer of the cochleate leading to delivery of the contents of the cochleate into the cytosol. It has been reported that DNA/cochleate formulations were able to induce strong CTL and antibody responses after parenteral or oral administration (162–164). Microparticle encapsulation Another potentially exciting means of DNA delivery is the use of biodegradable polymeric microparticles. Plasmid DNA trapped in these polymers (e.g. polylactice-coglycolides or chitosan) can be given systemically or to mucosal surfaces (orally or via the respiratory tract). The ability of polylactice-coglycolide–entrapped DNA vaccines to induce asystemic and mucosal immune responses after oral or intraperitoneal administration has been demonstrated (see above). Attenuated organisms Delivery of DNA can also be accomplished by attenuated intracellular bacteria. Intracellular bacteria, carrying the DNA, undergo phagocytosis by APCs, delivering plasmid DNA into the host cell phagosome or cytosol. The released DNA is then transcribed, resulting in expression of encoded antigens. Attentuated strains of invasive bacteria Shigella flexneri (165, 166), Salmonella typhimurium (167, 168), and Listeria monocytogenes (169) have been used for the delivery of plasmid DNA. For S. typhimurium, the bacteria are lysed within the phagosome, releasing plasmid DNA from this compartment into the cytoplasm via an unknown mechanism. Vaccination of mice with attenuated S. typhimurium transformed with plasmid DNA encoding lysteriolysin induced specific antibody as well as T-cell responses (167). Moreover, in a separate study, fluorescent DCs were demonstrated after oral administration of S. typhimurium harboring plasmid DNA encoding green fluorescent protein. These data provided evidence that this delivery system could target relevant immune cells, leading to efficient induction of an immune response (168). For Shigella infection, after phagocytosis and lysis within host cells, antigenic material is released directly into the cytoplasm. Immunization by using attenuated S. flexneri transformed with a bacterial plasmids encoding b-gal led to induction of a strong antigen-specific


DNA VACCINES

955

humoral and cellular response (166). In a separate study, it was shown that mice vaccinated with attenuated Shigella vaccine harboring measles virus genes induced a vigorous antigen-specific response (170). Finally, delivery of eukaryotic expression vectors in murine macrophage cell lines by attenuated suicide L. monocytogenes has also been reported (169). Whereas immunization with naked DNA has not been reported to lead to genomic integration with a significant frequency (see below), delivery of DNA by L. monocytogenes has resulted in chromosomal integration in vitro (169, 171), raising safety concerns with this technology. Alphaviruses are arthropod-borne togaviruses with a positive-polarity and single-stranded RNA genome that can replicate in a large number of animal hosts. Development of a variety of expression strategies has made it possible to deliver foreign genes in vivo by using alphaviruses (reviewed in 172). During infection, viral RNA replication is initiated by translation of viral nonstructural replicase proteins directly from the viral genome. During replication, both full-length genomic RNA and RNA initiated from an internal viral subgenomic promoter are synthesized. These subgenomic RNA transcripts are produced in excess relative to the genomic RNA and serve as mRNA for viral structural proteins. Thus, the natural viral life cycle permits striking amplification of mRNA. It has been shown that substitution of a heterologous gene for a viral structural gene results in highlevel expression of the heterologous gene. Recently, the development of a layered plasmid DNA-based expression system by using alphaviruses has been described (173–176). The mode of heterologous gene expression from alphavirus-derived expression vectors differs from that of conventional DNA vaccine plasmids in that transcription of heterologous genes is achieved in multiple steps. The first step involves the generation of viral genomic RNA that functions as a template for mRNA synthesis. Second, taking advantage of the virus life cycle, amplification of mRNA is achieved to drive the synthesis of antigen-encoding sequences. As the virus encodes machinery required for RNA replication and amplification in the host cell cytoplasm, high levels of protein production can be obtained, thus circumventing many problems associated with nuclear gene expression (such as limitation of transcription factors, RNA transport, etc). This method of gene delivery provides an exciting advance in the field of DNA vaccines, because these vectors can express heterologous proteins at higher levels than can conventional DNA vaccines (177). Somatic transgene immunization The concept of expressing T-cell epitopes in Ig has been demonstrated in foreign genes inserted into one or more of the complementary determining regions in the Ig heavy-chain molecule (antigenized antibodies), leading to induction of an immune response against the heterologous epitopes. A DNA-based approach as an alternative to the above has recently been described. In two separate studies, it was demonstrated that antigenized antibodyDNA constructs containing either a B-cell epitope or a B- and T-cell epitope engineered to different complement-determining regions led to the production of an antibody response directed against both epitopes (178, 179). Unlike conven-


956

tional DNA vaccines, immunization with these constructs led to an efficient detection of both transgene expression in vivo and transgene product in the serum.

APPLICATION For details on models for specific applications of DNA vaccines, see Tables 3A– 3D, which are produced in their entirety at the Annual Reviews world-wide-web site (www.AnnualReviews.org). These tables provide data on models for allergic diseases (Table 3A), autoimmune diseases (Table 3B), infectious diseases (Table 3C), and tumors (Table 3D).


SAFETY A number of safety concerns have been raised about the use of DNA vaccines. These include the possibility that such vaccines may (a)integrate into the host genome, thereby increasing the risk of malignancy (by activating oncogenes or inactivating tumor suppressor genes); (b) induce responses against transfected cells, thereby triggering the development of autoimmune disease; (c) induce tolerance rather than immunity; and/or (d) stimulate the production of cytokines that alter the host’s ability to respond to other vaccines and resist infection (180). Plasmids can persist at the site of injection for many months. They can also be found far from the original site of injection (including the gonads), perhaps carried by transfected lymphocytes or macrophages. Long-term persistence may be especially common for plasmids that encode self-antigens, because these do not induce an immune response against the cell they transfect. To date, there is no clear evidence that plasmids integrate, yet neither has this possibility been eliminated. Efforts to prove that high-molecular-weight (genomic) DNA does not contain plasmids (proof that integration has not taken place) have failed, in part because of contamination of the high-molecular-weight fraction by plasmid concatamers combined with the enormous sensitivity of the polymerase chain reaction. To overcome this problem, investigators digested genomic DNA with a restriction enzyme that is specific for a single site within the plasmid. By repeatedly digesting and isolating high-molecular-weight DNA, most but not all of the plasmid can be eliminated (181). Whether the few remaining copies of plasmids represent integration events remains to be determined. Concern that DNA vaccines might promote the development of autoimmune diseases arises from the immunostimulatory activity of CpG motifs in the plasmid backbone. It has been known for many years that bacterial DNA can induce the production of anti–double-stranded-DNA autoantibodies in normal mice and accelerate the development of autoimmune disease in lupus-prone animals (182– 184). The CpG motifs present in bacterial DNA and DNA vaccines stimulate the production of IL-6 and block the apoptotic death of activated lymphocytes, both


DNA VACCINES

957

functions that predispose to the development of systemic lupus erythematosus by facilitating persistent B-cell activation (185–190). These findings led several groups to investigate whether systemic autoimmune disease was induced or accelerated by the CpG motifs (191). With sensitive spot enzyme-linked immuno spot (ELIspot) assays, the absolute number of B cells secreting autoantibodies was studied in normal mice repeatedly immunized with a DNA vaccine. Shortly after vaccination, the number of IgG anti-DNA–secreting cells rose by two- to threefold (192). This was accompanied by a 35%–60% increase in serum IgG anti-DNA antibody titer. This modest rise in autoantibody level did not, however, result in the development of disease in normal mice or accelerate disease in lupus-prone animals (191–194). Thus, although the theoretical possibility remains that a subset of DNA vaccines (particularly those encoding determinants cross-reactive with self) may induce or accelerate autoimmune disease, findings to date suggest that the level of autoantibody production elicited by DNA vaccines is insufficient to induce such an outcome. The situation is somewhat more complex for organ-specific autoimmune disease, whose induction is promoted by strong type I immune responses. In an IL12–dependent model of experimental allergic encephalomyelitis, animals treated with CpG motifs and then challenged with myelin basic protein developed autoreactive Th1 effector cells that caused experimental allergic encephalomyelitis (65). In a molecular mimicry model, CpG motifs acted as potent immunoactivators, inducing autoimmune myocarditis when coinjected with Chlamydiaderived antigen (195). These findings indicate that CpG motifs may trigger deleterious autoimmune reactions under certain circumstances. Balancing these safety concerns is the observation that toxicity has not been reported among normal animals treated with therapeutic doses of DNA vaccines or CpG ODNs. In addition, hundreds of human volunteers have been exposed to plasmid DNA vaccines without serious adverse consequences. Most vaccines intended for human use are administered to infants and children. Owing to the immaturity of their immune systems, newborns exposed to foreign antigens are at risk for developing tolerance rather than immunity (196). A number of factors influence the development of neonatal tolerance, including the nature, concentration, and mode of antigen presentation to the immune system as well as the age of the host (197–199). Because the protein encoded by a DNA vaccine is produced endogenously and expressed in the context of self-MHC, the potential exists for the neonatal immune system to recognize it as ‘‘self,’’ resulting in tolerance rather than immunity. Consistent with such a possibility, a DNA vaccine encoding the circumsporozoite protein of malaria was found to induce tolerance rather than immunity in newborn mice (200). Neonatal animals treated with this vaccine were unable to generate T- or B-cell responses when challenged with pCSP as adults, thereby remaining at increased risk from infection despite immunization (200, 201). In this system, the induction of tolerance was critically dependent on the age at which the vaccine was administered. Tolerance was observed only when vaccine was administered to mice ,8 days of age; however,


958


decreased protection was also observed in geriatric mice (.2 years of age), raising concern that DNA vaccines might be less immunogenic in the elderly as well as in the very young (202). Efforts are under way to improve the overall immunogenicity of DNA vaccines by coadministering plasmids encoding cytokines or costimulatory molecules. Recent results suggest that these approaches can improve immunization of neonates and the elderly (199, 203–209). Safety concerns also arise from the use of CpG motifs or cytokine-encoding plasmids as adjuvants to improve the in vivo response elicited by DNA vaccines. An important component of immune homeostasis is through a balance in the production of Th1 cytokines (which promote cell-mediated immunity) and Th2 cytokines (which facilitate humoral immune responses or counterregulate Th1 responses). These two classes of cytokine-producing cells form a dynamic and mutually inhibitory network, because Th1 cytokines can block the maturation of Th2-type cells and vice versa. The overproduction of one type of cytokine can disrupt immune homeostasis, thereby altering the host’s response to other vaccines, susceptibility to infection, and predisposition to develop autoimmune disease. Although the use of cytokine-encoding plasmids is growing in popularity, relatively little information is available on their long-term safety. Although no serious side effects have been reported after the administration of cytokine-encoding plasmids in animals, it is unclear whether systematic efforts to detect such events were undertaken. Studies indicate that cytokine-encoding plasmids given in conjunction with antigen do alter the cytokine milieu (ratio of Th1-:Th2-secreting cells) and ultimately bias the immune response. In contrast, cytokine DNA given alone did not appear to alter immune reactivity against unrelated antigens and did not lead to the development of autoimmunity (203, 210). Indeed, no change in the frequency of Th1 or Th2 cytokine-secreting precursors was detected in mice treated multiple times with IFN-c-, IL-4-, or granulocyte/macrophage colony-stimulating factor–encoding plasmids. It thus appears that the cytokine released by transfected cells primarily affects local rather than systemic immunity, leaving serum cytokine levels generally unchanged.

CONCLUSIONS DNA vaccines moved very rapidly from laboratory phenomena into clinical trials. This transition was sustained by our ability to harness the tools of molecular biology to design antigen-encoding plasmids capable of inducing immune responses against pathogens for which no conventional vaccine was available, yet enthusiasm for this new technology must be tempered by an appreciation of its potential risks. The long-term sequelae of DNA vaccination have received little attention despite the capacity of these plasmids to persist in vivo for months or years. As multicomponent DNA vaccines and DNA vaccines encoding both cytokines and antigens become more common, the possibility for detrimental side effects will increase. DNA vaccines have the potential to be administered to mil-

DNA VACCINES

959

lions of children and/or adults. Thus, adverse events occurring even at low frequency (,1 /1000) could affect many thousands of otherwise healthy individuals. Adequate preclinical studies coupled with large-scale human trials will still be needed to establish the risk of this new vaccine approach. To aid in this effort, the Food and Drug Administration has published ‘‘Points to Consider” a document that provides valuable suggestions for the evaluation of the safety, potency, and immunogenicity of candidate DNA vaccines (210a). ACKNOWLEDGMENTS We thank Brenda Rae Marshall for editing the manuscript.



LITERATURE CITED 1. Nicolau C, Le Pape A, Soriano P, Fargette F, Juhel MF. 1983. In vivo expression of rat insulin after intravenous administration of the liposome-entrapped gene for rat insulin I. Proc. Natl. Acad. Sci. USA 80:1068–72 2. Kaneda Y, Iwai K, Uchida T. 1989. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. Science 243:375–78 3. Mannino RJ, Gould-Fogerite S. 1988. Liposome mediated gene transfer. Biotechniques 6:682–90 4. Benvenisty N, Reshef L. 1986. Direct introduction of genes into rats and expression of the genes. Proc. Natl. Acad. Sci. USA 83:9551–55 5. Wu GY, Wu CH. 1988. Receptor-mediated gene delivery and expression in vivo. J. Biol. Chem. 263:14621–24 6. Wolff JA, Malone RW, Williams P, Chong W, Acsadi G, Jani A, Felgner PL. 1990. Direct gene transfer into mouse muscle in vivo. Science 247:1465–68 7. Tang DC, De Vit M, Johnston SA. 1992. Genetic immunization is a simple method for eliciting an immune response. Nature 356:152–54 8. Ulmer JB, Donnelly JJ, Parker SE,

9.

10.

10a. 11.

12.

Rhodes GH, Felgner PL, Dwarki VJ, Gromkowski SH, Deck RR, De Witt DM, Friedman A, Hawe LA, Leander KR, Martinez D, Perry HC, Shiver JW, Montgomery DC, Liu MA. 1993. Heterologous protection against influenza by injection of DNA encoding a viral protein. Science 259:1745–49 Robinson HL, Hunt LA, Webster RG. 1993. Protection against a lethal influenza virus challenge by immunization with a haemagglutinin-expressing plasmid DNA. Vaccine 11:957–60 Robinson HL, Pertmer TM. 1999. DNA vaccines for viral infection: basic studies and applications. Adv. Virus Res. In press Whalen R. http://www.genweb.com/ dnavax/dnavax.html Yamamoto S, Yamamoto T, Shimada S, Kuramoto E, Yano O, Kataoka T, Tokunaga T. 1992. DNA from bacteria, but not from vertebrates, induces interferons, activates natural killer cells and inhibits tumor growth. Microbiol. Immunol. 36:983–97 Klinman DM, Yi AK, Beaucage SL, Conover J, Krieg AM. 1996. CpG motifs present in bacteria DNA rapidly induce lymphocytes to secrete interleukin 6,

960

13.

14.


15.

16.

17.

18.

19.

20.

21.

22.

GURUNATHAN n KLINMAN n SEDER interleukin 12, and interferon c. Proc. Natl. Acad. Sci. USA 93:2879–83 Halpern MD, Kurlander RJ, Pisetsky DS. 1996. Bacterial DNA induces murine interferon-c production by stimulation of interleukin-12 and tumor necrosis factora. Cell Immunol. 167:72–78 Sato Y, Roman M, Tighe J, Lee D, Corr M, Nguyen MD, Silverman GJ, Lotz M, Carson DA, Raz E. 1996. Immunostimulatory DNA sequences necessary for effective intradermal gene immunization. Science 273:352–54 Krieg AM, Yi AK, Matson S, Waldschmidt TJ, Bishop GA, Teasdale R, Koretzky GA, Klinman DM. 1995. CpG motifs in bacterial DNA trigger direct Bcell activation. Nature 374:546–49 Cardon LR, Burge C, Clayton DA, Karlin S. 1994. Pervasive CpG suppression in animal mitochondrial genomes. Proc. Natl. Acad. Sci. USA 91:3799–803 Razin A, Friedman J. 1981. DNA methylation and its possible biological roles. Prog. Nucl. Acid Res. Mol. Biol. 25:33– 52 Stacey KJ, Sweet MJ, Hume DA. 1996. Macrophages ingest and are activated by bacterial DNA. J. Immunol. 157:2116– 22 Jakob T, Walker PS, Krieg AM, Udey MC, Vogel JC. 1998. Activation of cutaneous dendritic cells by CpG-containing oligodeoxynucleotides: a role for dendritic cells in the augmentation of Th1 responses by immunostimulatory DNA. J. Immunol. 161:3042–49 Cowdery JS, Chace JH, Yi AK, Krieg AM. 1996. Bacterial DNA induces NK cells to produce IFN-c in vivo and increases the toxicity of lipopolysaccharides. J. Immunol. 156:4570–75 Bendigs S, Salzer U, Lipford GB, Wagner H, Heeg K. 1999. CpG-oligodeoxynucleotides co-stimulate primary T cells in the absence of antigen-presenting cells. Eur. J. Immunol. 29:1209–18 Corr M, Lee DJ, Carson DA, Tighe H.

23.

24.

25.

26.

27.

28.

29.

30.

1996. Gene vaccination with naked plasmid DNA: mechanism of CTL priming. J. Exp. Med. 184:1555–60 Iwasaki A, Torres CA, Ohashi PS, Robinson HL, Barber BH. 1997. The dominant role of bone marrow-derived cells in CTL induction following plasmid DNA immunization at different sites. J. Immunol. 159:11–14 Doe B, Selby M, Barnett S, Baenziger J, Walker CM. 1996. Induction of cytotoxic T lymphocytes by intramuscular immunization with plasmid DNA is facilitated by bone marrow-derived cells. Proc. Natl. Acad. Sci. USA 93:8578–83 Casares S, Inaba K, Brumeanu TD, Steinman RM, Bona CA. 1997. Antigen presentation by dendritic cells after immunization with DNA encoding a major histocompatibility complex class II-restricted viral epitope. J. Exp. Med. 186:1481–86 Porgador A, Irvine KR, Iwasaki A, Barber BH, Restifo NP, Germain RN. 1998. Predominant role for directly transfected dendritic cells in antigen presentation to CD8` T cells after gene gun immunization. J. Exp. Med. 188:1075–82 Akbari O, Panjwani N, Garcia S, Tascon R, Lowrie D, Stockinger B. 1999. DNA vaccination: transfection and activation of dendritic cells as key events for immunity. J. Exp. Med. 189:169–78 Condon C, Watkins SC, Celluzzi CM, Thompson K, Falo LD Jr. 1996. DNAbased immunization by in vivo transfection of dendritic cells. Nat. Med. 2:1122– 28 Ulmer JB, Deck RR, Dewitt CM, Donnhly JI, Liu MA. 1996. Generation of MHC class I-restricted cytotoxic T lymphocytes by expression of a viral protein in muscle cells: antigen presentation by non-muscle cells. Immunology 89:59– 67 Agadjanyan MG, Kim JJ, Trivedi N, Wilson DM, Monzavi-Karbassi B, Morrison LD, Nottingham LK, Dentchev T,

DNA VACCINES

31.


32.

33.

34.

35.

36.

37.

38.

Tsai A, Dang K, Chalian AA, Maldonado MA, Williams WV, Weiner DB. 1999. CD86 (B7–2) can function to drive MHC-restricted antigen-specific CTL responses in vivo. J. Immunol. 162: 3417–27 Iwasaki A, Stiernholm BJ, Chan AK, Berinstein NL, Barber BH. 1997. Enhanced CTL responses mediated by plasmid DNA immunogens encoding costimulatory molecules and cytokines. J. Immunol. 158:4591–601 Torres CA, Iwasaki A, Barber BH, Robinson HL. 1997. Differential dependence on target site tissue for gene gun and intramuscular DNA immunizations. J. Immunol. 158:4529–32 Yang NS, Burkholder J, Roberts B, Martinell B, McCabe D. 1990. In vivo and in vitro gene transfer to mammalian somatic cells by particle bombardment. Proc. Natl. Acad. Sci. USA 87:9568–72 Raz E, Carson DA, Parker SE, Parr TB, Abai AM, Aichinger G, Gromkowski SH, Singh M, Lew D, Yankauckas MA, Baird SM, Rhodes GH. 1994. Intradermal gene immunization: the possible role of DNA uptake in the induction of cellular immunity to viruses. Proc. Natl. Acad. Sci. USA 91:9519–23 Klinman DM, Sechler JM, Conover J, Gu M, Rosenberg AS. 1998. Contribution of cells at the site of DNA vaccination to the generation of antigen-specific immunity and memory. J. Immunol. 160:2388–92 Wraith DC, Vessey AE, Askonas BA. 1987. Purified influenza virus nucleoprotein protects mice from lethal infection. J. Gen. Virol. 68:433–40 Staerz UD, Karasuyama H, Garner AM. 1987. Cytotoxic T lymphocytes against a soluble protein. Nature 329:449–51 Harding CV, Song R. 1994. Phagocytic processing of exogenous particulate antigens by macrophages for presentation by class I MHC molecules. J. Immunol. 153:4925–33

961

39. Falo LD Jr, Kovacsovics-Akowski M, Thompson K, Rock KL. 1995. Targeting antigen into the phagocytic pathway in vivo induces protective tumour immunity. Nat. Med. 1:649–53 40. Udono H, Srivastava PK. 1993. Heat shock protein 70-associated peptides elicit specific cancer immunity. J. Exp. Med. 178:1391–96 41. Fu TM, Ulmer JB, Caulfield MJ, Deck RR, Friedman A, Wang S, Liu X, Donnelly JJ, Liu MA. 1997. Priming of cytotoxic T lymphocytes by DNA vaccines: requirement for professional antigen presenting cells and evidence for antigen transfer from myocytes. Mol. Med. 3:362–71 42. Albert ML, Sauter B, Bhardwaj N. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class Irestricted CTLs. Nature 392:86–89 43. Albert ML, Pearce SF, Francisco LM, Sauter B, Roy P, Silverstein RL, Bhardwaj N. 1998. Immature dendritic cells phagocytose apoptotic cells via aVb5 and CD36, and cross-present antigens to cytotoxic T lymphocytes. J. Exp. Med. 188:1359–68 44. Banchereau J, Bazan F, Blanchard D, Briere F, Galizzi LP, van Kooten C, Liu YJ, Rousset F, Saeland S. 1994. The CD40 antigen and its ligand. Annu. Rev. Immunol. 12:881–922 45. Ridge JP, Di Rosa F, Matzinger P. 1998. A conditioned dendritic cell can be a temporal bridge between a CD4` Thelper and a T-killer cell. Nature 393:474–78 46. Bennett SRM, Carbone FR, Karamalis F, Flavell RA, Miller J, Heath WR. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393:478–80 47. Schoenberger SP, Toes REM, van der Voort EIH, Offringa R, Melief CJM. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 393:480–83 48. Seder RA, Paul WE. 1994. Acquisition

962

49.

50.


51.

52.

53.

54.

55.

56.

GURUNATHAN n KLINMAN n SEDER of lymphokine-producing phenotype by CD4` T cells. Annu. Rev. Immunol. 12:635–73 O’Garra A. 1998. Cytokines induce the development of functionally heterogeneous T helper cell subsets. Immunity 8:275–83 Waris ME, Tsou C, Erdman DD, Zaki SR, Anderson LJ. 1996. Respiratory synctial virus infection in BALB/c mice previously immunized with formalininactivated virus induces enhanced pulmonary inflammatory response with a predominant Th2-like cytokine pattern. J. Virol. 70:2852–60 Li X, Sambhara S, Li CX, Ewasyshyn M, Parrington M, Caterini J, James O, Cates G, Du RP, Klein M. 1998. Protection against respiratory syncytial virus infection by DNA immunization. J. Exp. Med. 188:681–88 Roman M, Martin-Orozco E, Goodman JS, Nguyen MD, Sato Y, Ronaghy A, Kornbluth RS, Richman DD, Carson DA, Raz E. 1997. Immunostimulatory DNA sequences function as T helper-1promoting adjuvants. Nat. Med. 3:849– 54 Pertmer TM, Roberts TR, Haynes JR. 1996. Influenza virus nucleoproteinspecific immunoglobulin G subclass and cytokine responses elicited by DNA vaccination are dependent on the route of vector DNA delivery. J. Virol. 70:6119– 25 Feltquate DM, Heaney S, Webster RG, Robinson HL. 1997. Different T helper cell types and antibody isotypes generated by saline and gene gun DNA immunization. J. Immunol. 158:2278–84 Suss G, Shortman K. 1996. A subclass of dendritic cells kills CD4 T cells via Fas/Fas ligand-induced apoptosis. J. Exp. Med. 183:1789–96 Pulendran B, Smith JL, Caspary G, Brasel K, Pettit D, Maraskovsky E, Maliszewski CR. 1999. Distinct dendritic cell subsets differentially regulate the class of

57.

58.

59.

60.

61.

62.

63.

immune response in vivo. Proc. Natl. Acad. Sci. USA 96:1036–41 Smith AL, de St Groth BF. 1999. Antigen-pulsed CD8a` dendritic cells generate an immune response after subcutaneous injection without homing to the draining lymph node. J. Exp. Med. 189:593–98 Rissoan MC, Soumelis V, Kadowaki N, Grouard G, Briere F, de Waal Malefyt R, Liu YJ. 1999. Reciprocal control of T helper cell and dendritic cell differentiation. Science 283:1183–86 Maldonado-Lopez R, De Smedt T, Michel P, Godfroid J, Pajak B, Heirman C, Thielemans K, Leo O, Urbain J, Moser M. 1999. CD8a` and CD8a– subclasses of dendritic cells direct the development of distinct T helper cells in vivo. J. Exp. Med. 189:587–92 Boyle JS, Koniaras C, Lew AM. 1997. Influence of cellular location of expressed antigen on the efficacy of DNA vaccination: cytotoxic T lymphocyte and antibody responses are suboptimal when antigen is cytoplasmic after intramuscular DNA immunization. Int. Immunol. 9:1897–906 Haddad D, Liljeqvist S, Stahl S, Andersson I, Perlmann P, Berzins K, Ahlborg N. 1997. Comparative study of DNA-based immunization vectors: effect of secretion signals on the antibody responses in mice. FEMS Immunol. Med. Microbiol. 18:193–202 Zimmermann S, Egeter O, Hausmann S, Lipford GB, Rocken M, Wagner H, Heeg K. 1998. CpG oligodeoxynucleotides trigger protective and curative Th1 responses in lethal murine leishmaniasis. J. Immunol. 160:3627–30 Raz E, Tighe H, Sato Y, Corr M, Dudler JA, Roman M, Swain SL, Spiegelberg HL, Carson DA. 1996. Preferential induction of a Th1 immune response and inhibition of specific IgE antibody for-

DNA VACCINES

64.

65.


66.

67.

68.

69.

70.

71.

mation by plasmid DNA immunization. Proc. Natl. Acad. Sci. USA 93:5141–45 Hsu CH, Chua KY, Tao MH, Lai YL, Wu HD, Huang SK, Hsieh KH. 1996. Immunoprophylaxis of allergen-induced immunoglobulin E synthesis and airway hyperresponsiveness in vivo by genetic immunization. Nat. Med. 2:540–44 Segal BM, Klinman DM, Shevach EM. 1997. Microbial products induce autoimmune disease by an IL-12-dependent pathway. J. Immunol. 158:5087–90 Waisman A, Ruiz PJ, Hirschberg DL, Gelman A, Oksenberg JR, Brocke S, Mor F, Cohen IR, Steinman L. 1996. Suppressive vaccination with DNA encoding a variable region gene of the T-cell receptor prevents autoimmune encephalomyelitis and activates Th2 immunity. Nat. Med. 2:899–905 Martins LP, Lau LL, Asano MS, Ahmed R. 1995. DNA vaccination against persistent viral infection. J. Virol. 69:2574– 82 Yokoyama M, Zhang J, Whitton JL. 1995. DNA immunization confers protection against lethal lymphocytic choriomeningitis virus infection. J. Virol. 69:2684–88 Zarozinski CC, Fynan EF, Selin LK, Robinson HL, Welsh RM. 1995. Protective CTL-dependent immunity and enhanced immunopathology in mice immunized by particle bombardment with DNA encoding an internal virion protein. J. Immunol. 154:4010–17 Chen Y, Webster RG, Woodland DL. 1998. Induction of CD8` T cell responses to dominant and subdominant epitopes and protective immunity to Sendai virus infection by DNA vaccination. J. Immunol. 160:2425–32 Fu TM, Friedman A, Ulmer JB, Liu MA, Donnelly JJ. 1997. Protective cellular immunity: cytotoxic T-lymphocyte responses against dominant and recessive epitopes of influenza virus nucleo-

72.

73.

74.

75.

76.

77.

78.

79.

80.

963

protein induced by DNA immunization. J. Virol. 71:2715–21 Deck RR, DeWitt CM, Donnelly JJ, Liu MA, Ulmer JB. 1997. Characterization of humoral immune responses induced by an influenza hemagglutinin DNA vaccine. Vaccine 15:71–78 Robinson HL, Boyle CA, Feltquate DM, Morin MJ, Santoro JC, Webster RG. 1997. DNA immunization for influenza virus: studies using hemagglutinin- and nucleoprotein-expressing DNAs. J. Infect. Dis. 176:S50–55 (Suppl) Boyle CM, Morin M, Webster RG, Robinson HL. 1996. Role of different lymphoid tissues in the initiation and maintenance of DNA-raised antibody responses to the influenza virus H1 glycoprotein. J. Virol. 70:9074–78 Kang Y, Calvo PA, Daly TM, Long CA. 1998. Comparison of humoral immune responses elicited by DNA and protein vaccines based on merozoite surface protein-1 from Plasmodium yoelii, a rodent malaria parasite. J. Immunol. 161:4211– 19 Boyle JS, Silva A, Brady JL, Lew AM. 1997. DNA immunization: induction of higher avidity antibody and effect of route on T-cell cytotoxicity. Proc. Natl. Acad. Sci. USA 94:14626–31 Peet NM, McKeating JA, Ramos B, Klonisch T, De Souza JB, Delves PJ, Lund T. 1997. Comparison of nucleic acid and protein immunization for induction of antibodies specific for HIV-1 gp120. Clin. Exp. Immunol. 109:226–32 Donnelly JJ, Martinez D, Jansen KU, Ellis RW, Montgomery DL, Liu MA. 1996. Protection against papillomavirus with a polynucleotide vaccine. J. Infect. Dis. 173:314–20 Gurunathan S, Prussin C, Sacks DL, Seder RA. 1998. Vaccine requirements for sustained cellular immunity to an intracellular parasitic infection. Nat. Med. 4:1409–15 Davis HL, Schirmbeck R, Reimann J,

964

81.

82.


83.

84.

85.

86.

87.

88.

GURUNATHAN n KLINMAN n SEDER Whalen RG. 1995. DNA-mediated immunization in mice induces a potent MHC class I-restricted cytotoxic T lymphocyte response to the hepatitis B envelope protein. Hum. Gene Ther. 6:1447– 56 Tough DF, Borrow P, Sprent J. 1996. Induction of bystander T cell proliferation by viruses and type I interferon in vivo. Science 272:1947–50 Sun S, Zhang X, Tough DF, Sprent J. 1998. Type I interferon-mediated stimulation of T cells by CpG DNA. J. Exp. Med. 188:2335–42 Wolff JA, Ludtke JJ, Acsadi G, Williams P, Jani A. 1992. Long-term persistence of plasmid DNA and foreign gene expression in mouse muscle. Hum. Mol. Genet. 1:363–69 Manthorpe M, Cornefert-Jensen F, Hartikka J, Felgner J, Rundell A, Margalith M, Dwarki V. 1993. Gene therapy by intramuscular injection of plasmid DNA: studies on firefly luciferase gene expression in mice. Hum. Gene Ther. 4:419–31 Cheng L, Ziegelhoffer PR, Yang NS. 1993. In vivo promoter activity and transgene expression in mammalian somatic tissues evaluated by using particle bombardment. Proc. Natl. Acad. Sci. USA 90:4455–59 Chapman BS, Thayer RM, Vincent KA, Haigwood NL. 1991. Effect of intron A from human cytomegalovirus (Towne) immediate-early gene on heterologous expression in mammalian cells. Nucl. Acids Res. 19:3979–86 Norman JA, Hobart P, Manthorpe M, Felgner P, Wheeler C. 1997. Development of improved vectors for DNAbased immunization and other gene therapy applications. Vaccine 15:801–3 Wild J, Gruner B, Metzger K, Kuhrober A, Pudollek HP, Hauser H, Schirmbeck R, Reimann J. 1998. Polyvalent vaccination against hepatitis B surface and core antigen using a dicistronic expression plasmid. Vaccine 16:353–60

89. Lewin B, ed. 1994. Genes V. Oxford Univ. Press/Cell. 5th ed. 90. Uchijima M, Yoshida A, Nagata T, Koide Y. 1998. Optimization of codon usage of plasmid DNA vaccine is required for the effective MHC class I-restricted T cell responses against an intracellular bacterium. J. Immunol. 161:5594–99 91. Andre S, Seed B, Eberle J, Schraut W, Bultmann A, Haas J. 1998. Increased immune response elicited by DNA vaccination with a synthetic gp120 sequence with optimized codon usage. J. Virol. 72:1497–503 92. Vinner L, Nielsen HV, Bryder K, Corbet S, Nielsen C, Fomsgaard A. 1999. Gene gun DNA vaccination with Rev-independent synthetic HIV-1 gp160 envelope gene using mammalian codons. Vaccine 17:2166–75 93. Lewis PJ, Cox GJ, van Drunen Littel-van den Hurk S, Babiuk LA. 1997. Polynucleotide vaccines in animals: enhancing and modulating responses. Vaccine 15:861–64 94. Inchauspe G, Vitvitski L, Major ME, Jung G, Spengler U, Maisonnas M, Trepo C. 1997. Plasmid DNA expressing a secreted or a nonsecreted form of hepatitis C virus nucleocapsid: comparative studies of antibody and T-helper responses following genetic immunization. DNA Cell Biol. 16:185–95 95. Rice J, King CA, Spellerberg MB, Fairweather N, Stevenson FK. 1999. Manipulation of pathogen-derived genes to influence antigen presentation via DNA vaccines. Vaccine 17:3030–38 96. Wu Y, Kipps TJ. 1997. Deoxyribonucleic acid vaccines encoding antigens with rapid proteasome-dependent degradation are highly efficient inducers of cytolytic T lymphocytes. J. Immunol. 159:6037– 43 97. Tobery TW, Siliciano RF. 1997. Targeting of HIV-1 antigens for rapid intracellular degradation enhances cytotoxic T lymphocyte (CTL) recognition and the

DNA VACCINES

98.

99.


100.

101.

102.

103.

104.

105. 106.

induction of de novo CTL responses in vivo after immunization. J. Exp. Med. 185:909–20 Rodriguez F, Zhang J, Whitton JL. 1997. DNA immunization: ubiquitination of a viral protein enhances cytotoxic T-lymphocyte induction and antiviral protection but abrogates antibody induction. J. Virol. 71:8497–503 Whitton JL, Rodriguez F, Zhang J, Hassett DE. 1999. DNA immunization: mechanistic studies. Vaccine 17:1612–19 Ciernik IF, Berzofsky JA, Carbone DP. 1996. Induction of cytotoxic T-lymphocytes and antitumor immunity with DNA vaccines expressing single T-cell epitopes. J. Immunol. 156:2369–75 Iwasaki A, Dela Cruz CS, Young AR, Barber BH. 1999. Epitope-specific cytotoxic T lymphocyte induction by minigene DNA immunization. Vaccine 17:2081–88 Yu Z, Karem KL, Kanangat S, Manickan E, Rouse BT. 1998. Protection by minigenes: a novel approach of DNA vaccines. Vaccine 16:1660–67 Hanke T, Schneider J, Gilbert SC, Hill AV, McMichael A. 1998. DNA multiCTL epitope vaccines for HIV and Plasmodium falciparum: immunogenicity in mice. Vaccine 16:426–35 Thomson SA, Sherritt MA, Medveczky J, Elliott SL, Moss DJ, Fernando JG, Brown LE, Suhrbier A. 1998. Delivery of multiple CD8 cytotoxic T cell epitopes by DNA vaccination. J. Immunol. 160:1717–23 Suhrbier A. 1997. Multi-epitope DNA vaccines. Immunol. Cell Biol. 75:402–8 Wang R, Doolan DL, Charoenvit Y, Hedstrom RC, Gardner NJ, Hobart P, Tine J, Sedegah M, Fallarme V, Sacci JB Jr, Kaur M, Klinman DM, Hoffman SL, Weiss WR. 1998. Simultaneous induction of multiple antigen-specific cytotoxic T lymphocytes in nonhuman primates by immunization with a mixture

107.

108.

109.

110.

111.

112.

113.

114.

115.

965

of four Plasmodium falciparum DNA plasmids. Infect. Immun. 66:4193–202 Shi YP, Hasnain SE, Sacci JB, Holloway BP, Fujioka H, Kumar N, Wohlhueter R, Hoffman SL, Collins WE, Lal AA. 1999. Immunogenicity and in vitro protective efficacy of a recombinant multistage Plasmodium falciparum candidate vaccine. Proc. Natl. Acad. Sci. USA 96:1615–20 Maecker HT, Umetsu DT, DeKruyff RH, Levy S. 1998. Cytotoxic T cell responses to DNA vaccination: dependence on antigen presentation via class II MHC. J. Immunol. 161:6532–36 Klinman DM, Yamshchikov G, Ishigatsubo Y. 1997. Contribution of CpG motifs to the immunogenicity of DNA vaccines. J. Immunol. 158:3635–39 Brazolot Millan CL, Weeratna R, Krieg AM, Siegrist CA, Davis HL. 1998. CpG DNA can induce strong Th1 humoral and cell-mediated immune responses against hepatitis B surface antigen in young mice. Proc. Natl. Acad. Sci. USA 95:15553–58 Davis HL, Weeranta R, Waldschmidt TJ, Tygrett L, Schorr J, Krieg AM. 1998. CpG DNA is a potent enhancer of specific immunity in mice immunized with recombinant hepatitis B surface antigen. J. Immunol. 160:870–76 Klinman DM. 1998. Therapeutic applications of CpG-containing oligodeoxynucleotides. Antisense Nucl. Acid Drug Dev. 8:181–84 Klinman DM, Barnhart KM, Conover J. 1999. CpG motifs as immune adjuvants. Vaccine 17:19–25 Krieg AM, Wu T, Weeratna R, Efler SM, Love-Homan L, Yang L, Yi AK, Short D, Davis HL. 1998. Sequence motifs in adenoviral DNA block immune activation by stimulatory CpG motifs. Proc. Natl. Acad. Sci. USA 95:12631–36 Liang H, Nishioka Y, Reich CF, Pisetsky DS, Lipsky PE. 1996. Activation of human B cells by phosphorothioate oli-

966

116.

117.


118.

119.

120.

121.

122.

GURUNATHAN n KLINMAN n SEDER godeoxynucleotides. J. Clin. Invest. 98: 1119–29 Ballas ZK, Rasmussen WL, Krieg AM. 1996. Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA. J. Immunol. 157:1840–45 Hacker H, Mischak H, Miethke T, Liptay S, Schmid R, Sparwasser T, Heeg K, Lipford GB, Wagner H. 1998. CpG-DNAspecific activation of antigen-presenting cells requires stress kinase activity and is preceded by non-specific endocytosis and endosomal maturation. EMBO J. 17:6230–40 Richmond JF, Mustafa F, Lu S, Santoro JC, Weng J, O’Connell M, Fenyo EM, Hurwitz JL, Montefiori DC, Robinson HL. 1997. Screening of HIV-1 Env glycoproteins for the ability to raise neutralizing antibody using DNA immunization and recombinant vaccinia virus boosting. Virology 230:265–74 Richmond JF, Lu S, Santoro JC, Weng J, Hu SL, Montefiori DC, Robinson HL. 1998. Studies of the neutralizing activity and avidity of anti-human immunodeficiency virus type 1 Env antibody elicited by DNA priming and protein boosting. J. Virol. 72:9092–100 Mustafa F, Richmond JF, FernandezLarsson R, Lu S, Fredriksson R, Fenyo EM, O’Connell M, Johnson E, Weng J, Santoro JC, Robinson HL. 1997. HIV-1 Env glycoproteins from two series of primary isolates: replication phenotype and immunogenicity. Virology 229:269–78 Fuller DH, Murphey-Corb M, Clements J, Barnett S, Haynes JR. 1996. Induction of immunodeficiency virus-specific immune responses in rhesus monkeys following gene gun-mediated DNA vaccination. J. Med. Primatol. 25:236–41 Fuller DH, Haynes JR. 1994. A qualitative progression in HIV type 1 glycoprotein 120-specific cytotoxic cellular and humoral immune responses in mice receiving a DNA-based glycoprotein 120

123.

124.

125.

126.

127.

128.

129.

vaccine. AIDS Res. Hum. Retrovir. 10:1433–41 Lu S, Wyatt R, Richmond JF, Mustafa F, Wang S, Weng J, Montefiori DC, Sodroski J, Robinson HL. 1998. Immunogenicity of DNA vaccines expressing human immunodeficiency virus type 1 envelope glycoprotein with and without deletions in the V1/2 and V3 regions. AIDS Res. Hum. Retrovir. 14:151–55 Lu S, Manson K, Wyand M, Robinson HL. 1997. SIV DNA vaccine trial in macaques: post-challenge necropsy in vaccine and control groups. Vaccine 15:920–23 Okuda K, Xin KO, Tsuji T, Bukawa H, Tanaka S, Koff WC, Tani K, Honma K, Kawamoto S, Hamajima K, Fukushima J. 1997. DNA vaccination followed by macromolecular multicomponent peptide vaccination against HIV-1 induces strong antigen-specific immunity. Vaccine 15:1049–56 Barnett SW, Rajasekar S, Legg H, Doe B, Fuller DH, Haynes JR, Walker CM, Steimer KS. 1997. Vaccination with HIV-1 gp120 DNA induces immune responses that are boosted by a recombinant gp120 protein subunit. Vaccine 15:869–73 Letvin NL, Montefiori DC, Yasutomi Y, Perry HC, Davies ME, Lekutis C, Alroy M, Freed DC, Lord CI, Handt LK, Liu MA, Shiver JW. 1997. Potent, protective anti-HIV immune responses generated by bimodal HIV envelope DNA plus protein vaccination. Proc. Natl. Acad. Sci. USA 94:9378–83 Fuller DH, Corb MM, Barnett S, Steimer K, Haynes JR. 1997. Enhancement of immunodeficiency virus-specific immune responses in DNA-immunized rhesus macaques. Vaccine 15:924–26 Hanke T, Blanchard TJ, Schneider J, Hannan CM, Becker M, Gilbert SC, Hill AV, Smith GL, McMichael A. 1998. Enhancement of MHC class I-restricted peptide-specific T cell induction by a

DNA VACCINES

130.

131.


132.

133.

134.

135.

136.

DNA prime/MVA boost vaccination regime. Vaccine 16:439–45 Sedegah M, Jones TR, Kaur M, Hedstrom R, Hobart P, Tine JA, Hoffman SL. 1998. Boosting with recombinant vaccinia increases immunogenicity and protective efficacy of malaria DNA vaccine. Proc. Natl. Acad. Sci. USA 95:7648–53 Schneider J, Gilbert SC, Blanchard TJ, Hanke T, Robson KJ, Hannan CM, Becker M, Sinden R, Smith GL, Hill AV. 1998. Enhanced immunogenicity for CD8` T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara. Nat. Med. 4:397–402 Kent SJ, Zhao A, Best SJ, Chandler JD, Boyle DB, Ramshaw IA. 1998. Enhanced T-cell immunogenicity and protective efficacy of a human immunodeficiency virus type 1 vaccine regimen consisting of consecutive priming with DNA and boosting with recombinant fowlpox virus. J. Virol. 72:10180–88 Robinson HL, Montefiori DC, Johnson RP, Manson KH, Kalish ML, Lifson JD, Rizvi TA, Lu S, Hu SL, Mazzara GP, Panicali DL, Herndon JG, Glickman R, Candido MA, Lydy SL, Wyand MS, McClure HM. 1999. Neutralizing antibody-independent containment of immunodeficiency virus challenges by DNA priming and recombinant pox virus booster immunizations. Nat. Med. 5:526–34 Fynan EF, Webster RG, Fuller DH, Haynes JR, Santoro JC, Robinson HL. 1993. DNA vaccines: protective immunizations by parenteral, mucosal, and gene-gun inoculations. Proc. Natl. Acad. Sci. USA 90:11478–82 Jones DH, Corris S, McDonald S, Clegg JC, Farrar GH. 1997. Poly(DL-lactideco-glycolide)-encapsulated plasmid DNA elicits systemic and mucosal antibody responses to encoded protein after oral administration. Vaccine 15:814–17 Chen SC, Jones DH, Fynan EF, Farrar

137.

138.

139.

140.

141.

142.

143.

144.

967

GH, Clegg JC, Greenberg HB, Herrmann JE. 1998. Protective immunity induced by oral immunization with a rotavirus DNA vaccine encapsulated in microparticles. J. Virol. 72:5757–61 Herrmann JE, Chen SC, Jones DH, Tinsley-Bown A, Fynan EF, Greenberg HB, Farrar GH. 1999. Immune responses and protection obtained by oral immunization with rotavirus VP4 and VP7 DNA vaccines encapsulated in microparticles. Virology 259:148–53 Roy K, Mao HQ, Huang SK, Leong KW. 1999. Oral gene delivery with chitosanDNA nanoparticles generates immunologic protection in a murine model of peanut allergy. Nat. Med. 5:387–91 Klavinskis LS, Gao L, Barnfield C, Lehner T, Parker S. 1997. Mucosal immunization with DNA-liposome complexes. Vaccine 15:818–20 Ban EM, van Ginkel FW, Simecka JW, Kiyono H, Robinson HL, McGhee JR. 1997. Mucosal immunization with DNA encoding influenza hemagglutinin. Vaccine 15:811–13 Kuklin N, Daheshia M, Karem K, Manickan E, Rouse BT. 1997. Induction of mucosal immunity against herpes simplex virus by plasmid DNA immunization. J. Virol. 71:3138–45 Sasaki S, Hamajima K, Fukushima J, Ihata A, Ishii N, Gorai I, Hirahara F, Mohri H, Okuda K. 1998. Comparison of intranasal and intramuscular immunization against human immunodeficiency virus type 1 with a DNA-monophosphoryl lipid A adjuvant vaccine. Infect. Immun. 66:823–26 Sasaki S, Sumino K, Hamajima K, Fukushima J, Ishii N, Kawamoto S, Mohri H, Kensil CR, Okuda K. 1998. Induction of systemic and mucosal immune responses to human immunodeficiency virus type 1 by a DNA vaccine formulated with QS21 saponin adjuvant via intramuscular and intranasal routes. J. Virol. 72:4931– 39 Sasaki S, Fukushima J, Hamajima K,

968

145.

146.


147.

148.

149.

150.

151.

152.

GURUNATHAN n KLINMAN n SEDER Ishii N, Tsuji T, Xin KQ, Mohri H, Okuda K. 1998. Adjuvant effect of Ubenimex on a DNA vaccine for HIV-1. Clin. Exp. Immunol. 111:30–35 Wang B, Dang K, Agadjanyan MG, Srikantan V, Li F, Ugen KE, Boyer J, Merva M, Williams WV, Weiner DB. 1997. Mucosal immunization with a DNA vaccine induces immune responses against HIV-1 at a mucosal site. Vaccine 15:821– 25 Livingston JB, Lu S, Robinson H, Anderson DJ. 1998. Immunization of the female genital tract with a DNA-based vaccine. Infect. Immun. 66:322–29 Shi Z, Curiel DT, Tang DC. 1999. DNAbased non-invasive vaccination onto the skin. Vaccine 17:2136–41 Davis HL, Jasmin BJ. 1993. Direct gene transfer into mouse diaphragm. FEBS Lett. 333:146–50 Davis HL, Michel ML, Mancini M, Schleef M, Whalen RG. 1994. Direct gene transfer in skeletal muscle: plasmid DNA-based immunization against the hepatitis B virus surface antigen. Vaccine 12:1503–9 Wang R, Doolan DL, Le TP, Hedstrom RC, Coonan KM, Charoenvit Y, Jones TR, Hobart P, Margalith M, Ng J, Weiss WR, Sedegah M, de Taisne C, Norman JA, Hoffman SL. 1998. Induction of antigen-specific cytotoxic T lymphocytes in humans by a malaria DNA vaccine. Science 282:476–80 Leitner WW, Seguin MC, Ballou WR, Seitz JP, Schultz AM, Sheehy MJ, Lyon JA. 1997. Immune responses induced by intramuscular or gene gun injection of protective deoxyribonucleic acid vaccines that express the circumsporozoite protein from Plasmodium berghei malaria parasites. J. Immunol. 159:6112– 19 Tsan MF, White JE, Shepard B. 1995. Lung-specific direct in vivo gene transfer with recombinant plasmid DNA. Am. J. Physiol. 268:L1052–56

153. Meyer KB, Thompson MM, Levy MY, Barron LG, Szoka FC Jr. 1995. Intratracheal gene delivery to the mouse airway: characterization of plasmid DNA expression and pharmacokinetics. Gene Ther. 2:450–60 154. Stribling R, Brunette E, Liggitt D, Gaensler K, Debs R. 1992. Aerosol gene delivery in vivo. Proc. Natl. Acad. Sci. USA 89:11277–81 155. McDonald RJ, Liggitt HD, Roche L, Nguyen HT, Pearlman R, Raabe OG, Bussey LB, Gorman CM. 1998. Aerosol delivery of lipid:DNA complexes to lungs of rhesus monkeys. Pharm. Res. 15:671–79 156. Okada E, Sasaki S, Ishii N, Aoki I, Yasuda T, Nishioka K, Fukushima J, Miyazaki J, Wahren B, Okuda K. 1997. Intranasal immunization of a DNA vaccine with IL-12- and granulocytemacrophage colony-stimulating factor (GM-CSF)–expressing plasmids in liposomes induces strong mucosal and cellmediated immune responses against HIV-1 antigens. J. Immunol. 159:3638– 47 157. Kuklin NA, Daheshia M, Chun S, Rouse BT. 1998. Immunomodulation by mucosal gene transfer using TGF-b DNA. J. Clin. Invest. 102:438–44 158. Kawabata K, Takakura Y, Hashida M. 1995. The fate of plasmid DNA after intravenous injection in mice: involvement of scavenger receptors in its hepatic uptake. Pharm. Res. 12:825–30 159. Lew D, Parker S, Latimer T, Abai A, Kuwahara-Rundell A, Doh S, Yang Z-Y, Gromkowski S, Nabel G, Manthrope M, Norman J. 1995. Cancer gene therapy using plasmid DNA: pharmacokinectic study of DNA following injection in mice. Hum. Gene Ther. 6:553–64 160. Williams RS, Johnston SA, Riedy M, De Vit MJ, McElligott SG, Sanford JC. 1991. Introduction of foreign genes into

DNA VACCINES

161.

162.


163.

164.

165.

166.

167.

168.

169.

tissues of living mice by DNA-coated microprojectiles. Proc. Natl. Acad. Sci. USA 88:2726–30 Gregoriadis G, Saffie R, de Souza JB. 1997. Liposome-mediated DNA vaccination. FEBS Lett. 402:107–10 Gould-Fogerite S, Kheiri MT, Zang F, Wang Z, Scolpino AJ, Feketeova E, Canki M, Mannino RJ. 1998. Targetting immune response induction with cochleate and liposomal-based vaccines. Adv. Drug Del. Rev. 32:273–87 Mannino RJ, Gould-Fogerite S. 1997. Antigen cochleate preparations for oral and systemic vaccination. In New Generation Vaccines, eds. ML Levine, GC Woodrow, JB Kaper, GS Cobon. New York: Marcel Dekker. 2nd ed. Gould-Fogerite S, Mannino RJ. 1996. Mucosal and systemic immunization using cochleate and liposome vaccines. J. Liposome Res. 2:357–79 Sizemore DR, Branstrom AA, Sadoff JC. 1995. Attenuated Shigella as a DNA delivery vehicle for DNA-mediated immunization. Science 270:299–302 Sizemore DR, Branstrom AA, Sadoff JC. 1997. Attenuated bacteria as a DNA delivery vehicle for DNA-mediated immunization. Vaccine 15:804–7 Darji A, Guzman CA, Gerstel B, Wachholz P, Timmis KN, Wehland J, Chakraborty T, Weiss S. 1997. Oral somatic transgene vaccination using attenuated S. typhimurium. Cell 91:765–75 Paglia P, Medina E, Arioli I, Guzman CA, Colombo MP. 1998. Gene transfer in dendritic cells, induced by oral DNA vaccination with Salmonella typhimurium, results in protective immunity against a murine fibrosarcoma. Blood 92:3172–76 Dietrich G, Bubert A, Gentschev I, Sokolovic Z, Simm A, Catic A, Kaufmann SH, Hess J, Szalay AA, Goebel W. 1998. Delivery of antigen-encoding plasmid DNA into the cytosol of macrophages by

170.

171.

172.

173.

174.

175.

176.

177.

178.

969

attenuated suicide Listeria monocytogenes. Nat. Biotechnol. 16:181–85 Fennelly GJ, Khan SA, Abadi MA, Wild TF, Bloom BR. 1999. Mucosal DNA vaccine immunization against measles with a highly attenuated Shigella flexneri vector. J. Immunol. 162:1603–10 Courvalin P, Goussard S, Grillot-Courvalin C. 1995. Gene transfer from bacteria to mammalian cells. C.R. Acad. Sci. III 318:1207–12 Tubulekas I, Berglund P, Fleeton M, Liljestrom P. 1997. Alphavirus expression vectors and their use as recombinant vaccines: a minireview. Gene 190:191–95 Driver DA, Latham EM, Polo JM, Belli BA, Banks TA, Chada S, Brumm D, Chang SMW, Mento SJ, Jolly DJ, Dubensky TW. 1995. Layered amplification of gene expression with a DNA gene delivery system. Ann. NY Acad. Sci. 772:261–64 Dubensky TW Jr, Driver DA, Polo JM, Belli BA, Latham EM, Ibanez CE, Chada S, Brumm D, Banks TA, Mento SJ, Jolly DJ, Chang SM. 1996. Sindbis virus DNA-based expression vectors: utility for in vitro and in vivo gene transfer. J. Virol. 70:508–19 Herweijer H, Latendresse JS, Williams P, Zhang G, Danko I, Schlesinger S, Wolff JA. 1995. A plasmid-based self-amplifying Sindbis virus vector. Hum. Gene Ther. 6:1161–67 Berglund P, Smerdou C, Fleeton MN, Tubulekas I, Liljestrom P. 1998. Enhancing immune responses using suicidal DNA vaccines. Nat. Biotechnol. 16:562– 65 Hariharan MJ, Driver DA, Townsend K, Brumm D, Polo JM, Belli BA, Catton DJ, Hsu D, Mittelstaedt D, McCormack JE, Karavodin L, Dubensky TW Jr, Chang SM, Banks TA. 1998. DNA immunization against herpes simplex virus: enhanced efficacy using a Sindbis virus-based vector. J. Virol. 72:950–58 Gerloni M, Baliou WR, Billetta R,

970

179.

180.


181.

182.

183.

184.

185.

186.

187.

GURUNATHAN n KLINMAN n SEDER Zanetti M. 1997. Immunity to Plasmodium falciparum malaria sporozoites by somatic transgene immunization. Nat. Biotechnol. 15:876–81 Xiong S, Gerloni M, Zanetti M. 1997. Engineering vaccines with heterologous B and T cell epitopes using immunoglobulin genes. Nat. Biotechnol. 15:882– 86 Klinman DM, Takeno M, Ichino M, Gu M, Yamshchikov G, Mor G, Conover J. 1997. DNA vaccines: safety and efficacy issues. Springer Semin. Immunopathol. 19:245–56 Martin T, Parker SE, Hedstrom R, Le T, Hoffman SL, Norman J, Hobart P, Lew D. 1999. Plasmid DNA malaria vaccine: the potential for genomic integration after intramuscular injection. Hum. Gene Ther. 10:759–68 Gilkeson GS, Ruiz P, Howell D, Lefkowith JB, Pisetsky DS. 1993. Induction of immune-mediated glomerulonephritis in normal mice immunized with bacterial DNA. Clin. Immunol. Immunopathol. 68:283–92 Gilkeson GS, Pippen AM, Pisetsky DS. 1995. Induction of cross-reactive antidsDNA antibodies in preautoimmune NZB/NZW mice by immunization with bacterial DNA. J. Clin. Invest. 95:1398– 402 Steinberg AD, Krieg AM, Gourley MF, Klinman DM. 1990. Theoretical and experimental approaches to generalized autoimmunity. Immunol. Rev. 118:129– 63 Klinman DM. 1990. Polyclonal B cell activation in lupus-prone mice precedes and predicts the development of autoimmune disease. J. Clin. Invest. 86: 1249–54 Klinman DM, Steinberg AD. 1987. Systemic autoimmune disease arises from polyclonal B cell activation. J. Exp. Med. 165:1755–60 Linker-Israeli M, Deans RJ, Wallace DJ, Prehn J, Ozeri-Chen T, Klinenberg JR.

188.

189.

190.

191.

192.

193.

194.

195.

196.

1991. Elevated levels of endogenous IL6 in systemic lupus erythematosus: a putative role in pathogenesis. J. Immunol. 147:117–23 Watanabe-Fukunaga R, Brannan CI, Copeland NG, Jenkins NA, Nagata S. 1992. Lymphoproliferation disorder in mice explained by defects in Fas antigen that mediates apoptosis. Nature 356: 314–17 Krieg AM. 1995. CpG DNA: a pathogenic factor in systemic lupus erythematosus? J. Clin. Immunol. 15:284–92 Yi AK, Hornbeck P, Lafrenz DE, Krieg AM. 1996. CpG DNA rescue of murine B lymphoma cells from anti-IgMinduced growth arrest and programmed cell death is associated with increased expression of c-myc and bcl-xL. J. Immunol. 157:4918–25 Katsumi A, Emi N, Abe A, Hasegawa Y, Ito M, Saito H. 1994. Humoral and cellular immunity to an encoded protein induced by direct DNA injection. Hum.Gene Ther. 5:1335–39 Mor G, Singla M, Steinberg AD, Hoffman SL, Okuda K, Klinman DM. 1997. Do DNA vaccines induce autoimmune disease? Hum. Gene Ther. 8:293–300 Xiang ZQ, Spitalnik SL, Cheng J, Erikson J, Wojczyk B, Ertl HC. 1995. Immune responses to nucleic acid vaccines to rabies virus. Virology 209:569– 79 Gilkeson GS, Conover J, Halpern M, Pisetsky DS, Feagin A, Klinman DM. 1998. Effects of bacterial DNA on cytokine production by (NZB/NZW)F1 mice. J. Immunol. 161:3890–95 Bachmaier K, Neu N, de la Maza LM, Pal S, Hessel A, Penninger JM. 1999. Chlamydia infections and heart disease linked through antigenic mimicry. Science 283:1335–39 Silverstein AM, Segal S. 1975. The ontogeny of antigen-specific T cells. J. Exp. Med. 142:802–4


DNA VACCINES

971

197. Marodon G, Rocha B. 1994. Activation 207. Davis HL, Brazolot Millan CL. 1997. and ‘deletion’ of self-reactive mature and DNA-based immunization against hepaimmature T cells during ontogeny of titis B virus. Springer Semin. ImmunoMls-1a mice: implications for neonatal pathol. 19:195–209 tolerance induction. Int. Immunol. 208. Bot A, Bot S, Bona C. 1998. Enhanced 6:1899–904 protection against influenza virus of mice 198. Silverstein AM. 1977. Ontogeny of the immunized as newborns with a mixture of plasmids expressing hemagglutinin immune response: a perspective. In and nucleoprotein. Vaccine 16:1675–82 Development of Host Defenses, eds. MD Cooper, DH Dayton, pp. 1–10. New 209. Manickan E, Yu Z, Rouse BT. 1997. York: Raven DNA immunization of neonates induces 199. Sarzotti M, Robbins DS, Hoffman PM. immunity despite the presence of maternal antibody. J. Clin. Invest. 100:2371– 1996. Induction of protective CTL 75 responses in newborn mice by a murine retrovirus. Science 271:1726–28 210. Ishii KJ, Weiss WR, Ichino D, Verthelyi 200. Mor G, Yamshchikov G, Sedegah M, D, Klinman DM. 1999. Activity and Takeno M, Wang R, Houghten RA, Hoffsafety of DNA plasmids encoding IL-4 man S, Klinman DM. 1996. Induction of and IFN-c. Gene Ther. 6:237–44 neonatal tolerance by plasmid DNA vac- 210a. U.S. Food and Drug Administration. cination of mice. J. Clin. Invest. Points to Consider. http://www.fda.gov/ 98:2700–5 cber/points.html 201. Ichino M, Mor G, Conover J, Weiss WR, 211. Hakim I, Levy S, Levy R. 1996. A nineTakeno M, Ishii KJ, Klinman DM. 1999. amino acid peptide from IL-1b augments Factors associated with the development antitumor immune responses induced by of neonatal tolerance after the adminisprotein and DNA vaccines. J. Immunol. tration of a plasmid DNA vaccine. J. 157:5503–11 Immunol. 162:3814–18 212. Maecker HT, Umetsu DT, DeKruyff RH, 202. Klinman DM, Conover J, Bloom ET, Levy S. 1997. DNA vaccination with Weiss W. 1998. Immunogenicity and cytokine fusion constructs biases the efficacy of DNA vaccination in aged immune response to ovalbumin. Vaccine mice. J. Gerontol. 53:B281–86 15:1687–96 203. Ishii KJ, Weiss W, Klinman DM. 1999. 213. Kim JJ, Trivedi NN, Nottingham LK, Prevention of neonatal tolerance by plasMorrison L, Tsai A, Hu Y, Mahalingam mid encoding GM-CSF. Vaccine. In press S, Dang K, Ahn L, Doyle NK, Wilson 204. Sarzotti M, Dean TA, Remington MP, Ly DM, Chattergoon MA, Chalian AA, CD, Furth PA, Robbins DS. 1997. InducBoyer JD, Agadjanyan MG, Weiner DB. tion of cytotoxic T cell responses in new1998. Modulation of amplitude and born mice by DNA immunization. direction of in vivo immune responses by Vaccine 15:795–97 co-administration of cytokine gene 205. Wang Y, Xiang Z, Pasquini S, Ertl HC. expression cassettes with DNA immu1997. Immune response to neonatal nogens. Eur. J. Immunol. 28:1089–103 genetic immunization. Virology 228: 214. Raz E, Watanabe A, Baird SM, Eisenberg 278–84 RA, Parr TB, Lotz M, Kipps TJ, Carson 206. Prince AM, Whalen R, Brotman B. 1997. DA. 1993. Systemic immunological effects of cytokine genes injected into Successful nucleic acid based immuniskeletal muscle. Proc. Natl. Acad. Sci. zation of newborn chimpanzees against USA 90:4523–27 hepatitis B virus. Vaccine 15:916–19


972


215. Chow YH, Huang WL, Chi WK, Chu YD, Tao MH. 1997. Improvement of hepatitis B virus DNA vaccines by plasmids coexpressing hepatitis B surface antigen and interleukin-2. J. Virol. 71:169–78 216. Geissler M, Gesien A, Tokushige K, Wands JR. 1997. Enhancement of cellular and humoral immune responses to hepatitis C virus core protein using DNA-based vaccines augmented with cytokine-expressing plasmids. J. Immunol. 158:1231–37 217. Prayaga SK, Ford MJ, Haynes JR. 1997. Manipulation of HIV-1 gp120-specific immune responses elicited via gene gunbased DNA immunization. Vaccine 15:1349–52 218. Chow YH, Chiang BL, Lee YL, Chi WK, Lin WC, Chen YT, Tao MH. 1998. Development of Th1 and Th2 populations and the nature of immune responses to hepatitis B virus DNA vaccines can be modulated by codelivery of various cytokine genes. J. Immunol. 160:1320–29 219. Kim JJ, Simbiri KA, Sin JI, Dang K, Oh J, Dentchev T, Lee D, Nottingham LK, Chalian AA, McCallus D, Ciccarelli E, Agadjanyan MG, Weiner DB. 1999. Cytokine molecular adjuvants modulate immune responses induced by DNA vaccine constructs for HIV-1 and SIV. J. Interferon Cytokine Res. 19:77–84 220. Larsen DL, Dybdahl-Sissoko N, McGregor MW, Drape R, Neumann V, Swain WF, Lunn DP, Olsen CW. 1998. Coadministration of DNA encoding interleukin-6 and hemagglutinin confers protection from influenza virus challenge in mice. J. Virol. 72:1704–8 221. Hengge UR, Chan EF, Foster RA, Walker PS, Vogel JC. 1995. Cytokine gene expression in epidermis with biological effects following injection of naked DNA. Nat. Genet. 10:161–66 222. Rogy MA, Auffenberg T, Espat NJ, Philip R, Remick D, Wollenberg GK,

223.

224.

225.

226.

227.

228.

Copeland EM III, Moldawer LL. 1995. Human tumor necrosis factor receptor (p55) and interleukin 10 gene transfer in the mouse reduces mortality to lethal endotoxemia and also attenuates local inflammatory responses. J. Exp. Med. 181:2289–93 Daheshia M, Kuklin N, Kanangat S, Manickan E, Rouse BT. 1997. Suppression of ongoing ocular inflammatory disease by topical administration of plasmid DNA encoding IL-10. J. Immunol. 159:1945–52 Manickan E, Daheshia M, Kuklin N, Chun S, Rouse BT. 1998. Modulation of virus-induced delayed-type hypersensitivity by plasmid DNA encoding the cytokine interleukin-10. Immunology 94:129–34 Kim JJ, Ayyavoo V, Bagarazzi ML, Chattergoon MA, Dang K, Wang B, Boyer JD, Weiner DB. 1997. In vivo engineering of a cellular immune response by coadministration of IL-12 expression vector with a DNA immunogen. J. Immunol. 158:816–26 Gurunathan S, Irvine KR, Wu CY, Cohen JI, Thomas E, Prussin C, Restifo NP, Seder RA. 1998. CD40 ligand/trimer DNA enhances both humoral and cellular immune responses and induces protective immunity to infectious and tumor challenge. J. Immunol. 161:4563–71 Tsuji T, Hamajima K, Fukushima J, Xin KQ, Ishii N, Aoki I, Ishigatsubo Y, Tani K, Kawamoto S, Nitta Y, Miyazaki J, Koff WC, Okubo T, Okuda K. 1997. Enhancement of cell-mediated immunity against HIV-1 induced by coinoculation of plasmid-encoded HIV-1 antigen with plasmid expressing IL-12. J. Immunol. 158:4008–13 Rakhmilevich AL, Turner J, Ford MJ, McCabe D, Sun WH, Sondel PM, Grota K, Yang NS. 1996. Gene gun-mediated skin transfection with interleukin 12 gene results in regression of established pri-

DNA VACCINES

229.

230.


231.

232.

233.

234.

235.

236.

mary and metastatic murine tumors. Proc. Natl. Acad. Sci. USA 93: 6291–96 Kim JJ, Ayyavoo V, Bagarazzi ML, Chattergoon M, Boyer JD, Wang B, Weiner DB. 1997. Development of a multicomponent candidate vaccine for HIV-1. Vaccine 15:879–83 Svanholm C, Lowenadler B, Wigzell H. 1997. Amplification of T-cell and antibody responses in DNA-based immunization with HIV-1 Nef by co-injection with a GM-CSF expression vector. Scand. J. Immunol. 46:298–303 Weiss WR, Ishii KJ, Hedstrom RC, Sedegah M, Ichino M, Barnhart K, Klinman DM, Hoffman SL. 1998. A plasmid encoding murine granulocyte-macrophage colony-stimulating factor increases protection conferred by a malaria DNA vaccine. J. Immunol. 161:2325–32 Lee AH, Suh YS, Sung YC. 1999. DNA inoculations with HIV-1 recombinant genomes that express cytokine genes enhance HIV-1 specific immune responses. Vaccine 17:473–79 Xiang Z, Ertl HC. 1995. Manipulation of the immune response to a plasmidencoded viral antigen by coinoculation with plasmids expressing cytokines. Immunity 2:129–35 Song XY, Gu M, Jin WW, Klinman DM, Wahl SM. 1998. Plasmid DNA encoding transforming growth factor-b1 suppresses chronic disease in a streptococcal cell wall-induced arthritis model. J. Clin. Invest. 101:2615–21 Lim YS, Kang BY, Kim EJ, Kim SH, Hwang SY, Kim TS. 1998. Potentiation of antigen specific, Th1 immune responses by multiple DNA vaccination with ovalbumin/interferon-c hybrid construct. Immunology 94:135–41 Yeow WS, Lawson CM, Beilharz MW. 1998. Antiviral activities of individual murine IFN-a subtypes in vivo: intra-

237.

238.

239.

240.

241.

242.

243.

973

muscular injection of IFN expression constructs reduces cytomegalovirus replication. J. Immunol. 160:2932–39 Horton HM, Anderson D, Hernandez P, Barnhart KM, Norman JA, Parker SE. 1999. A gene therapy for cancer using intramuscular injection of plasmid DNA encoding interferon a. Proc. Natl. Acad. Sci. USA 96:1553–58 Tsuji T, Hamajima K, Ishii N, Aoki I, Fukushima J, Xin KQ, Kawamoto S, Sasaki S, Matsunaga K, Ishigatsubo Y, Tani K, Okubo T, Okuda K. 1997. Immunomodulatory effects of a plasmid expressing B7–2 on human immunodeficiency virus-1-specific cell-mediated immunity induced by a plasmid encoding the viral antigen. Eur. J. Immunol. 27:782–87 Kim JJ, Bagarazzi ML, Trivedi N, Hu Y, Kazahaya K, Wilson DM, Ciccarelli R, Chattergoon MA, Dang K, Mahalingam S, Chalian AA, Agadjanyan MG, Boyer JD, Wang B, Weiner DB. 1997. Engineering of in vivo immune responses to DNA immunization via codelivery of costimulatory molecule genes. Nat. Biotechnol. 15:641–46 Corr M, Tighe H, Lee D, Dudler J, Trieu M, Brinson DC, Carson DA. 1997. Costimulation provided by DNA immunization enhances antitumor immunity. J. Immunol. 159:4999–5004 Horspool JH, Perrin PJ, Woodcock JB, Cox JH, King CL, June CH, Harlan DM, St. Louis DC, Lee KP. 1998. Nucleic acid vaccine-induced immune responses require CD28 costimulation and are regulated by CTLA4. J. Immunol. 160:2706–14 Mendoza RB, Cantwell MJ, Kipps TJ. 1997. Immunostimulatory effects of a plasmid expressing CD40 ligand (CD154) on gene immunization. J. Immunol. 159:5777–81 Kim JJ, Tsai A, Nottingham LK, Morrison L, Cunning DM, Oh J, Lee DJ, Dang

974



K, Dentchev T, Chalian AA, Agadjanyan MG, Weiner DB. 1999. Intracellular adhesion molecule-1 modulates b-chemokines and directly costimulates T cells in vivo. J. Clin. Invest. 103:869–77

244. Boyle JS, Brady JL, Lew AM. 1998. Enhanced responses to a DNA vaccine encoding a fusion antigen that is directed to sites of immune induction. Nature 392:408–11


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975


cd CELLS: A Right Time and a Right Place for a Conserved Third Way of Protection Adrian C. Hayday Department of Immunobiology, Guy’s King’s St. Thomas’ Medical School, King’s College, University of London, London, SE1 9RT, United Kingdom; e-mail: [email protected]


Key Words evolutionary conservation, cd-deficient mice, immunoprotection, regulatory functions of cd cells Abstract The tripartite subdivision of lymphocytes into B cells, ab T cells, and cd cells has been conserved seemingly since the emergence of jawed vertebrates, more than 450 million years ago. Yet, while we understand much about B cells and ab T cells, we lack a compelling explanation for the evolutionary conservation of cd cells. Such an explanation may soon be forthcoming as advances in unraveling the biochemistry of cd cell interactions are reconciled with the abnormal phenotypes of cd-deficient mice and with the striking differences in cd cell activities in different strains and species. In this review, the properties of cd cells form a basis for understanding cd cell interactions with antigens and other cells that in turn form a basis for understanding immunoprotective and regulatory functions of cd cells in vivo. We conclude by considering which cd cell functions may be most critical.

INTRODUCTION cd cells have been expertly reviewed before (e.g. 1–10). Hence, this manuscript focuses on experimental systems and findings that illustrate the signatory properties of cd cells. Those properties are: (a) (b) (c) (d) (e) (f) (g) (h)

Expression of an unique, conserved T cell receptor (TCR) Specialized anatomical distribution Characteristic cell phenotypes A distinct pathway of developmental maturation Unique antigen specificities A broad spectrum of cell-cell interactions A unique capacity to protect the host against specific pathogens A nonredundant capacity to regulate the course and consequences of immune responses (i) Uniquely age-dependent activities—an hypothesis. 0732–0582/00/0410–0975/$14.00

975

976

HAYDAY

If there is another cell that shares some of these features, it may be the lipidreactive CD41, CD81 (‘‘double negative’’, DN) ab T cell that, like the cd cell, is incompletely understood (11).

cd CELLS EXPRESS AN UNIQUE ANTIGEN RECEPTOR


Strong Evolutionary Conservation Conserved Gene Families with Potential for Diversity The four TCR chain gene families (a, b, c, d) appear strongly conserved across 400–500 million years of evolution of the jawed vertebrates. They have been identified in cartilaginous fish, e.g. Raja erinacea (skate), but not in extant jawless fish such as the lamprey. In all cases the TCRa and TCRd loci are closely linked and possibly interspersed, as they are in contemporary mammals (12–17). The c chain genes are always assembled via VJ rearrangement and the TCRd chain genes by V-D-J rearrangement, commonly involving D-D joining (Figure 1). This represents a more strongly conserved basis for the generation of diversity than is true for immunoglobulin (Ig) genes. For example, Ig light chain genes of Raja erinacea show no junctional diversity, because they do not somatically rearrange (18). Likewise, diversity in chicken Ig genes is driven primarily by gene conversion of a series of germ-line w-genes (19). TCR cd Is an Antigen Receptor TCRcd conserves structural features common to all TCRs, e.g. the organization of each V and C domain into ‘‘Ig folds’’ (;7 b strands packed face to face in two antiparallel b sheets), constrained by intradomain disulphide bonding. Moreover, Vc and Vd ‘‘Ig folds’’ would be predicted to associate (20). Likewise conserved are the sequence-based subdivision of V regions into framework and hypervariable regions (HVs), the latter encoding surface-exposed complementary determining regions (CDRs) (Figure 2). The relatedness of TCRc/d to other antigen receptor structures has also been conserved: c chain framework sequences are more similar to b chain sequences, and d chain framework sequences are more similar to a chain sequences. TCR cd Is a Unique Antigen Receptor Superimposed on generic TCR properties are cd signatures. In Vc there is often a serine residue at position 8, an IHWY motif at positions 34–37, and a tryptophan residue three amino acids Cterminal to the second conserved cysteine (21–23). Also conserved are the highly charged cytoplasmic tail of Cc (;15-residues) and the uncharged tail of Cd (;4 residues). Likewise, several criteria distinguish Vd DNA sequences from Va. For example, whereas murine, human, and skate Va sequences most commonly encode Q at framework residue 37, this is never the case in Vd. Instead, residue 37 in human Vd2 is positively charged, while Vc2, the chain with which Vd2 commonly pairs, is negatively charged at the corresponding position. Hence, posi-

977


Figure 1 Schematic representation of mouse and human TCRc/d gene families.TCR-V gene segments are shown as solid boxes; -D as striped boxes; -J as hatched boxes; and -C as stippled boxes. TCR gene segments indicate pseudogenes. Only representative Vc and Va are shown; some can be used as either (see text).


978

HAYDAY

Figure 2 Signatory features of TCRd as an antigen receptor, compared to TCRa and b. Bracketed are CDRs; for full description, see Ref. 20.

tion 37 may be a unique determinant of c/d chain pairing. Emphasizing the distinction of Vd and Va, structural data from a human Vd3 domain reveal a root ˚ from nine IgVH regions, over 89 mean square deviation (r.m.s.) of only 1.1A ˚ over 85 residues from three framework residues, compared to an r.m.s. of 1.4A Va domains (20). CDR1 and CDR2 CDR1 or CDR2 in commonly utilized Vd segments are larger than those in Va segments by two amino acids, the second one of which is always hydrophobic (21–23) (Figure 2). There are complementary extensions in Vc chains to which the Vd segments pair: Thus, human Vc2 and Vd2 have two extra amino acids in CDR1, whereas Vc1.3 and Vd3 (two other chains that commonly pair) have two extra residues in CDR2. Such corresponding alignments predict the restricted chain pairing that proves to be a feature of TCRcd (see below).


cd CELLS

979

Additionally, the characteristic c-d pairings show significant CDR1/2 divergence one from another: for example, murine Vc5 and Vc6 genes, located less than 1 kb apart, are only ;30% similar (24). Such divergence contrasts with the constraint imposed on TCRab CDR1 and CDR2 by the requirements for binding to major histocompatibility complex (MHC) antigens (25–27). If TCRcd reactivity is not constrained by a single class of antigen-presenting element (see below), CDR1/2cd sequences may diverge liberally, contributing significantly to antigen specificity. Emphasizing this, the canonical Vc5-Vd1 TCR expressed by .90% of murine skin–associated cd cells and the canonical Vc6-Vd1 TCR expressed by .90% of uterine epithelial, cd cells display an identical CDR3, leaving CDR1/ 2 as the sole sources of diversity. From the available human Vd structure, CDR2d may adopt an Ig VH-like conformation, whereas CDR1 lies in a Va-like conformation (20) (Figure 2). The importance of CDR1/2 in binding antigen has been best characterized for Ig, where simple mutations destroy or alter antigen specificities (28–31). In cartilaginous fish, where use is made of light chain genes prejoined in the germline, CDR1/2 must assume the burden of diversity (18). Nevertheless, CDR1/2 in Ig is diversified by somatic hypermutation, for which there is no evidence in TCRcd. Rather, TCRcd CDR1/2 diversity is developed and sustained in the germline, akin to diversity in the receptors of the innate immune system. This utilization of germline diversification of CDR1/2 may reflect the difficulty for epithelial cd cell subsets to comply with certain prerequisites of using somatic gene rearrangement to diversify the repertoire. Among these are the capacity (and need) to sample immense antigen diversity, the capacity to support clonal expansion of cells with ‘‘correct’’ receptors, and the capacity to select against cells with useless or deleterious self-reactive receptors (see below) (18). CDR3 Vd also differs from Va in the putative CDR3 regions, that through DD joining and flexible reading frame usage, can be highly diverse in length and composition (16, 32, 33) (Figure 2). Human TCRd junctional lengths are 8–21 amino acids, similar to a range of 3–25 amino acids for IgH. In skate, a range of 2–13 amino acids has been reported, albeit from a limited data set (16). CDR3c lengths are shorter but also variable (1–12 amino acids). For comparison, CDR3a/ b lengths are limited to 6–12 amino acids (33), constrained by limitations on the size of peptide antigens that bind to MHC. In this light, the variation in CDR3c/ d suggests that this part of the TCR might, like Ig, accommodate diverse antigens (34). Nonetheless, the available structural determination displays CDR3d in a Va-like conformation (20). In sum, TCRcd is a broadly conserved but unique structure, likely used for antigen recognition for more than 450 million years. Its evolutionary age places it within the same window as the recombination activating genes, which may represent a specialized form of a more ancient transposition system (35). Of TCRcd, TCRab, Ig, and the MHC genes, none is more obviously ancient than the other. Since TCRa and d most likely arose by gene duplication, TCRd may

980

HAYDAY

have preceded TCRa, originally pairing with either c or b. A single human bd heterodimer has been described (36), and the unique developmental timing of TCRa rearrangement during late thymocyte development (see below) distinguishes it from TCRb, c, and d gene families that rearrange earlier.


Species-Specific Evolution and Different Deployment of cd Cells in Different Organisms V Regions Despite the organizational conservation of TCRc/d genes, Vc genes are unusually diverged between species. Ordinarily, Ig or TCR V gene segments can be sorted into families sharing .75% identity. Such families of Va/d/b genes in mice and humans show higher similarity to each other than to other V gene families from within either species (23). Thus, single branches in dendograms are reliably populated by both mouse and human genes. This is not true for Vc, where both interspecies and intraspecies divergence are strikingly high (only human Vc1 and murine Vc7 can be clustered in a dendogram) (Tables 1, 2) (23). Genetic Complexity and Cell Numbers The interspecies differences in Vc genes reflect larger differences in TCRc/d gene families in different organisms (Figure 1). The TCRc locus in mice contains six commonly utilized genes; two within one family; the other four broadly diverged. There also appear to be six functional genes in humans, five in the Vc1 family and another more distantly related Vc2 gene. By contrast, there are 20 to .30 chicken Vc chain gene segments, and .6 Vc families in skate (the latter is probably a minimum estimate) (15, 16). TABLE 1 Proposed official nomenclature for murine Vc gene segments compared to two commonly used designations Official designationa

Heilig & Tonegawa (1986)b

Garman et al (1986)c

GV1S1

Vc5

Vc3

GV2S1

Vc6

Vc4

GV3S1

Vc4

Vc2

GV4S1

Vc7

Vc5

GV5S1

Vc1

Vc1.1

GV5S2

Vc2

Vc1.2

GV5S3

Vc3

Vc1.3

In the proposed designation, G refers to c; V to variable; 1 to the allocated number for the particular V region family (in this case, numbers 1–4 are based on increasing distance from Jc); s refers to the designation of a V segment as a member of a family if highly homologous V gene segments. GV1-V4 are not strongly homologous to any other V segments, thus they only have one member set. By contrast, there are three closely related members (S1–S3) of the V5 family. a Reference 22 (this review) b Heilig JS, Tonegawa S. 1986. Nature 322: 836. c Reference 24 (this review).

cd CELLS

981

TABLE 2 Proposed official nomenclature for humane Vc gene segments compared to commonly used designations Official designation GV1S1P GV1S2A1T


GV1S2A2T GV1S3A1N1T GV1S3A1N2T GV1S4A1N1T GV1S4A1N2T GV1S5 GV1S5P GV1S6P GV1S7P GV1S8 GV2S1A1 GV2S1A2 GV3S1P GV4S1P GV5S1P GV6S1P

Previous designations Vc1 Vc 2 Vc1.9 Vc1.9 Vc3 Vc1.3 Vc1.1 Vc4 Vc4 Vc5 Vc5P Vc6 Vc7 Vc8 Vc2.2 Vc9 Vc9 Vc10 Vc3.1 Vc11 fVA fVB

See footnote to Table 1 for basis of designation, with the added use of A1/A2 to denote alleles at a single locus; P, pseudogene; N, sequence substitutions that conserve the encoded polypeptide; T, allelic variants that are currently tentative. See Reference 21 (this volume)

Because TCRd genes are embedded within the TCRa locus, and some V gene segments are used both as Va and Vd (‘‘ADV’’ segments), it is difficult to define Vd diversity. However, in sequence divergence trees, the human ADV segments occupy their own branch, one that in mouse is occupied uniquely by Vd segments. Thus, we can estimate that there are ;16 murine Vd genes (of which 6 are homologous, and the other 10 are distinct) and ;8–10 distinct human Vd genes (21–23). By contrast, there are 20–30 chicken Vd genes, belonging to a single family, added to which there is widespread rearrangement of Va to DJ-Cd (17). A large number of Vd genes have also been described for cattle, sheep, and pigs (37), and there appear to be at least four Vd families in skate (16). TCRcd complexity shows an inverse correlation with TCRab complexity; while mice have

982

HAYDAY

;23 Vb genes and ;75 Va, and humans have ;47Vb genes and ;42Va genes, chickens harbor only two subfamilies of TCRa and b genes. Strikingly, TCRc/d gene complexity correlates approximately with the abundance of cd cells. Whereas cd cells commonly account for 0.5%–5% of adult murine or human CD3(`) cells, they are very common in chickens, and they can account for .70% of peripheral CD3(`) cells in young ruminants, declining to 5–25% with age (38, 39).

cd CELLS DISPLAY A UNIQUE ANATOMICAL DISTRIBUTION: IMPLICATIONS FOR ANTIGEN SAMPLING


cd Cells Localize to Tissues Although cd cells are well represented among peripheral blood mononuclear cells (PBMC), and in afferent and efferent lymph, they are rarely found in lymph node parenchyma, spleen, Peyer’s patches, and thymus (39, 40). This is true even in animals (e.g. calves) in which .70% of CD3(`) cells are TCRcd(`). Such splenic cd cells as do exist are found in the red pulp and the marginal zones— regions of cell trafficking—rather than in conventional T cell areas (40). Instead, cd cells are localized in tissues (39). They are disproportionately abundant in the intestine, commonly found as intraepithelial lymphocytes (IELs), interspersed between enterocytes (39–42). In humans, the cd : ab ratio among intestinal IELs is ;1:5, compared with ;1:50 in the lymph node. A striking expansion of the cd(`) IEL compartment is a diagnostic indicator for celiac disease (see below). In mice, cd(`) IELs are common in the skin [as ‘‘dendritic epidermal T cells’’ (DETC)] (43, 44) and in the uterine and vaginal epithelia, and they occur in tongue, lung, and mammary tissue (45). In sheep, tissues rich in cd cells include the skin (primarily the dermis), tongue, esophagus, trachea, and bladder. The circulation patterns of tissue-associated cd cells are only beginning to be clarified (46–48). Murine cd cells sharing signatory features with uterine cd(`) IELs (including a canonical TCR) appear in other tissues, including the lung following sensitization with mycobacterial aerosols (49), and in the liver and testicle, after infection with Listeria, or after the induction of inflammation using autologous cell grafts (50; see below). Whether these cells migrate from the genital epithelium, or from some other reservoir remains to be clarified.

cd Cells and Antigen Sampling The lymph nodes and the T cell areas of the spleen are the only known anatomical structures in which a great diversity of antigens can be presented and the consequent clonal expansion and differentiation of cognate lymphocytes supported. Therefore, the relative absence of cd cells from these areas is consistent with the

cd CELLS

983

hypothesis that cd cells do not routinely rely on professional APC for antigen recognition. Instead, they may recognize antigens directly in tissues, sometimes using extremely limited TCRcd diversity (Figure 3) (51, 52). The assignment to epithelial surfaces of cd cells with extremely limited diversity is consistent with this and provoked the ‘‘first line of defense’’ hypothesis (51), proposing that cd cells respond not to a diversity of microbial antigens, but to unique ‘‘stress antigens’’ that are markers of cell infection or transformation.

UNIQUE ASPECTS OF cd CELL PHENOTYPE: IMPLICATIONS FOR CELL FUNCTION AND DEVELOPMENT


Distribution of Cell Surface Marker Proteins Much can be inferred about function from the surface molecules that a cell expresses. Most peripheral cd cells are DN (double negative) for CD4 and CD8. Thus, in 3–4 month-old lambs, up to 75% of circulating T cells are DN cd(`) cells (39). While CD4(`) cd cells are seemingly conserved across species, they are rare. Additionally CD8(`) cd cells usually express CD8aa, not the CD8ab heterodimer expressed by conventional ab T cells (53). CD8 and CD4 are components of MHC recognition. Not surprisingly then, cd cells do not comply with

Figure 3 cd cells are tissue lymphocytes, posing questions about antigen sampling.


984

HAYDAY

the ‘‘ab T cell paradigm of MHC recognition’’ (54; see below). cd cells can be found in normal numbers in MHC class I and class II1/1 mice (55). CD28, the ligand for B7 (CD80/86) and the primary costimulator for ab T cells, is expressed on some cd cells. Although it was shown to regulate the response of mouse cd cells to TCR cross-linking, and of human cd cells to allogeneic dendritic cells (56–58), its expression is variable, and it is not clear that as a general rule cd cells utilize CD28 for interaction with B7(`) professional APCs. Nevertheless, with recent work showing that even simple B cell antigens are modified so as to coengage both the B cell receptor (BCR) and CR2/CD21 (59), it seems certain that cd cell activation will require more than stimulation through the antigen receptor alone. Several candidates, including the vitronectin receptor, have been explored as cd cell costimulators (60). Some cd cells express CD40L, implying a capacity to interact with CD40(`) cells, notably B cells (see below). Also, human and murine cd cells can express inhibitory NK receptors (61) as well as NKG2D, an NK-type, activating receptor that is also expressed by cytolytic CD8(`) ab cells and NK cells (62). Whether NKG2D activation accompanies activation through the TCR is under study (see below). Other molecules on cd cells (e.g CD2, CD5, CD6, and 2B4) show interspecies variation (63, 64). Importantly, there is no unequivocal, species-wide cd cellspecific surface marker other than TCRcd itself. The scavenger-receptor-like dimer, WC1, is a candidate for a cd cell-specific marker on bovine and ovine cells, although there are both WC1(`) and WC1(1)cd subsets. WC1(`) cd cells also express an unrelated high-molecular-mass antigen, GD3.5 (63, 65–67). Further characterization will determine whether they or their homologs will define cd cells in mice or humans.

Effector Molecules and Growth Regulation It is unclear whether cd cells en masse have a prevalent effector function. The cytolytic effectors, perforins, granulysin, fas/fasL, and the production of IFNc, are characteristics of cd cells and ab T cells alike, and in many contexts cd cells appear as cytolytic cells capable of eradicating infected cells and bacteria therein. Yet cd cells have also been portrayed as growth-factor producing cells, maintaining epithelial integrity—the primary barrier to infection. Thus, activated DETC and intestinal cd(`) IELs [but neither ab(`) IELs, nor systemic ab cells and cd cells] were reported to produce fibroblast growth factor VII (keratinocyte growth factor—KGF), which can induce growth or differentiation in myriad epithelial cells (68). Therefore, the potential exists to identify a unique, cd-cell specific genetic program that regulates KGF expression. cd cells may likewise produce epidermal growth factor. Th1/Th2 Similarly, the differential production of IL-2 and IFNc, and IL-4, IL5, IL-6, and IL-10, respectively, defines Th1 and Th2 cd cells, just as it categorizes ab cells (69–72). As for ab T cells, CD4 expression is a strong correlate of Th2

cd CELLS

985


cd cell activity and may contribute to Th2 commitment (71, 73). Since cd cells rarely express CD4, one can correctly infer that cd cells are predominantly Th1. It was recently found that whereas Th2 cells represent the default pathway of CD4(`) ab T cell differentiation, the Th1 state is the default pathway of splenic cd cell differentiation, just as it is for cytolytic ab T cells (74). Thus GATA3, which in part determines the Th2 status of ab T cells, fails to downregulate the IL-12 receptor b2 subunit, a defining component of Th1 cells in cd cells (74). Cytokine Responsiveness Peripheral cd cells can be strongly activated by a combination of IL-1 and IL-7, and are significantly more vulnerable than ab T cells or NK cells to IL-7-deficiency (75, 76). This is interesting, given that epithelial tissues can be sources of IL-7. cd cells also show characteristic responses to IL-15, and IL-12. In sum, cd cell effector capabilities and responses suggest there may be intracellular signaling webs unique to cd cells. This might be attributable to the development of cd cells from a distinct lineage of progenitors (see below). Additionally, there are undoubtedly distinct subsets of cd cells (just as there are for ab T cells and B cells) that are defined by unique developmental pathways (77).

cd CELLS FOLLOW A DISTINCT DEVELOPMENTAL PATHWAY OF MATURATION cd Cells Develop Early in Ontogeny and Are Stimulated in the Periphery Probably in all species, cd cells are the first T cells to develop. TCRcd gene rearrangements were detected in the murine thymus by embryonic day (E) 13 (78), in the chicken by E10 (15), and in the human by 8 weeks of fetal development (79). Some subsets of cd cells develop largely or exclusively prenatally (80–83), and these require peripheral stimulation for subsequent expansion (83). The early murine fetal thymus is dominated by DETC progenitors, whereafter come progenitors of vaginal IELs and gut IELs, respectively (1, 80). The earliest waves of cd cells develop before thymic terminal transferase is activated, accounting for the simplicity of V-D-J recombination joins in canonical epitheliaassociated cdTCRs. The schedule of the waves is due in part to sequential activation of gene expression and rearrangement at distinct Vc/Vd genes; this activation has in turn been attributed to programmed developmental potential of successively emerging hematopoietic progenitors (84, 85). But this potential additionally requires properties of the fetal/newborn stroma, since DETC development cannot be supported by an adult thymus (86, 87). The early development of cd(`) IELs also means that their signaling machinery may be different from that of other T cells. Specifically, there is little ZAP70 activity in the early fetal thymus, and reflecting this, cd(`) IELs are susceptible to syk kinase deficiency (88).

986

HAYDAY


Signals from the DETC TCR may be transduced to syk via TCR-associated FceRIc rather than via CD3f. Early rearrangements in chicken involve many Vc and Vd genes, plus P and N nucleotides; hence, emerging cd repertoires display extensive combinatorial and junctional diversity that contrasts with the restricted subsets of cd cells that appear early in mouse development. Similar broad diversity may be true in cattle and sheep (39). Nonetheless, the early ontogenetic development of cd cells is strikingly conserved. Hence, dominant cd repertoires in adult animals are most likely due to extensive peripheral expansion. That such expansion can be driven by abundant, low-molecular-mass antigens is consistent with this (see below). So too is the finding that murine cd cells commonly display activation/memory markers and have high turnover rates (89). Gut cd cells expand excessively in fasdeficient mice, indicating that their outgrowth in response to chronic antigen exposure is ordinarily limited by a fas-dependent mechanism (90).

cd Cells Do Not Pass Through a Preantigen ReceptorRegulated Checkpoint Lymphocyte maturation passes through a number of checkpoints, at which progression of one precursor cell type to become another is tightly regulated. One such checkpoint is regulated by preantigen receptors, the pre-BCR and pre-TCR, respectively (91, 92). The pre-TCR pairs TCRb with the surrogate a chain, pTa (92). The functional expression of the pre-BCR and pre-TCR is associated with rescue of cells from apoptosis and with active clonal expansion (93, 94). For ab T cell progenitors, this MHC-independent b-selection occurs as DN cells acquire CD4 and CD8 to become DP cells, in which TCRa rearrangement occurs. CD4 and CD8 are not commonly expressed by cd cells, and DP cells show reduced activity of the TCRd transcriptional enhancer (as activity at the TCRa chain gene enhancer increases) (95). TCRc chain expression is likewise decreased (24, 96). Hence, b-selection reflects a boundary point for cd cell and ab T cell divergence; whether this divergence is determined by lineage commitment events prior to this point remains uncertain (see below). Importantly, there is no evidence for a preantigen receptor in cd cell development. Any pairing of TCRc with pTa seems not important physiologically (97). An absence of a pre-TCR-cd is commonly linked to the fact that cd cell progenitors do not undergo developmental clonal expansion of the kind associated with b-selection, but rather they undergo complete maturation as DN cells.

Signals from TCRcd Regulate cd Cell Development Some late-stage DN cells are characterized by in-frame rearrangements of TCRd/ c chain genes (87). Therefore, even in the absence of clonal expansion, TCRcd would appear to transmit selective signals to some thymocytes. Those signals likely determine survival and further differentiation events, such as allelic exclu-

cd CELLS

987


sion (which is evident, albeit leaky, in cd cells), and possibly signals for cells to exit from the thymus. Such d selection precedes the downregulation of the heatstable antigen (87). The completion of maturation of cd cells as DN thymocytes, prior to major clonal expansion, perhaps depicts an ancient form of T cell development that existed prior to the duplication of the TCRd chain and the emergence of the ab T cell development pathway (98). Signals from TCRcd Perturb ab T Cell Development Cells maturing along the cd lineage so resemble those developing along the ab lineage that distinct precursors unique to the cd cell lineage have yet to be identified. TCRc/d gene rearrangement occurs in ab T cell progenitors, and some cd cells harbor rearranged TCRb chain genes (99–102). This excludes the possibility that early commitment of T cell progenitors to the ab lineage precludes TCRc/d chain gene rearrangement and vice versa. Moreover, TCRc/d expression in ab T cell progenitors is not neutral, as would be predicted to be the case in an entirely independent lineage. Several examples exist of TCRc/d transgenic mice in which ab T cell development is inhibited (103, 104). Such effects might be artefacts of transgenic mice because ordinarily expression of rearranged TCRc and d chains would be silenced in ab-lineage DP cells (95, 96). However, in-frame TCRc/d rearrangements seemingly prevent cells from entering the ab lineage in normal mice too, since ab T cell progenitors are depleted of in-frame TCRc/d chain gene rearrangements (101, 102). Most likely, selection against ab T cell development is imposed by expression of a complete TCRcd (101, 104). TCRcd, the Pre-TCR, and T Cell Lineage That TCRcd can actively signal in progenitors that might otherwise have become ab T cells is indicated by the fact that, in the absence of the pre-TCR, some DN cells are provoked to the DP stage, contingent on d selection (87, 104): Essentially no maturation is seen in the absence of either TCRcd or pre-TCR signaling (87). However, the provocation of DN to DP development by TCRcd in normal mice is markedly less efficient than that provoked by the pre-TCR (104). This has fueled the suggestion that TCRcd expression is a direct determinant of cell fate toward cd maturation and away from the DP stage (101). Likewise, the pre-TCR may be a dominant determinant of the DN to DP transition and consequent commitment to the ab lineage. Alternatively, neither TCRcd nor the pre-TCR may determine fate. Instead, they may facilitate maturation along pathways determined by other molecules. By this scenario, expression of TCRcd in a committed ab T cell progenitor is incompatible with the precommitted developmental pathway of that cell, as a result of which the cell dies, leaving viable ab T cells depleted of in-frame c/d rearrangements. The contrary would hold for pre-TCR expression in cd progenitors (105). Were lineage to be determined largely by TCR-independent factors, the genetic programs of the respective T cell progenitors might diverge from an early time point, consistent perhaps with differences in effector programs of the two cell types (section on Unique Aspects of cd Cell Phenotype). These different genetic


988

HAYDAY

programs could then determine intrinsically different, cell-specific responses to signaling through the pre-TCR/TCR. Certainly, prior expression of Notch can alter a thymocyte’s response to d-selection (106). Transgenic Notch 1 favors ab T cell development and can confer characteristics such as CD8ab expression on thymocytes developing in TCRcd transgenic mice (106). These effects require inductive interactions between thymocytes, since they are only seen when cells differing in Notch activity develop in competition; when either Notch 1`/` or Notch 1`/1 cells are examined in isolation, the balance of ab and cd cell development is normal. The hypothesis that TCRcd is at minimum a codeterminant of cell fate (101, 105) is supported by the correlation of genetic complexity at the TCRc/d loci with the abundance of cd cells (see above). In short, the greater the number of TCRc/d genes (as in chickens and in cattle), the greater the chance of TCRcd expression in early thymocytes, and hence the greater the number of cd T cells that develop. How signals from TCRcd and the pre-TCR might influence fate divergence will be determined biochemically. Recent experiments have indicated a capacity of the extended cytoplasmic tail of the pTa chain (unique to the preTCR) to activate NFjB, which is not activated in thymocytes maturing as cd cells (107).

A Ligand-Dependence of Developmental TCRcd Signaling Remains to be Defined There is currently no identification of functional thymic ligands for either TCRcd or the pre-TCR (or even the pre-BCR). ab(`) thymocyte development can be sustained by mutant pTa and TCRb proteins that lack ectodomains (108, 109). Thus, it is hypothesized that the pre-TCR (and presumably TCRcd) naturally aggregates into signaling rafts, from which ligand-independent signals are transduced. Nonetheless, there is no clear paradigm for this, and the pre-TCR has yet to be shown spontaneously to form super-molecular activation complexes (SMACs), the active signaling form of TCRab (110).

Some Evidence for Clonotypic Selection of cd Cells If somatic cells are allowed to develop high-affinity receptors with novel specificities, there must be a mechanism to clonally enrich for the appropriate ones (positive selection) and to inactivate undesirable ones (negative selection). The extent to which cd cells are selected has remained controversial. Negative Selection of Systemic cd Cells By use of TCRcd transgenic mice specific for the nonclassical, b2-microglobulin (b2m)-dependent–MHC class I molecule, TL (see below), negative selection was reported to occur, like that for ab T cells, through direct intrathymic engagement of the ligand on the appropriate haplotype (103). These data demonstrate that TCRcd-mediated signals can pro-

cd CELLS

989

mote apoptosis. Nonetheless, the affected thymocytes may not have been physiologic cd-progenitors, but rather surrogate ab-lineage thymocytes in which TCRcd transgene expression had excluded rearrangement of TCRa/b genes. Hence, the degree to which cd progenitors vie with TCR-dependent central deletion will remain uncertain until the regulated thymic expression of natural cd cell ligands is defined.


Negative Selection of Intraepithelial cd Cells The limitation of DETC development to the fetal/neonatal period may also be attributed to clonotypic negative selection. In DETC TCR(`) transgenic C57.BL/6 mice, DN DETC TCR(`) cells were depleted 1–2 weeks postpartum (111). Deletion did not occur in C3H mice. Use of recombinant inbred strains allowed assignment of a dominant deletion determinant to chromosome 18. Again, clonotypic IEL selection will be clarified when selecting and/or activating ligands for the DETC TCR are defined. Positive Selection of Intraepithelial cd Cells Positive selection would be an obvious mechanism to account for the limited diversity of canonical murine cd IEL repertoires. However, since these cells develop early in ontogeny, when gene rearrangement mechanisms are simple, the TCR repertoires will be limited by short sequence homologies that dictate recombination sites. In a normal fetal thymus, ;80% to .90% of in-frame Vc5–Jc1 recombinations are homogeneous, encoding the canonical DETC TCR (112). Molecular constraints are also evident among nonproductive gene rearrangements, but these are generally more diverse than productive ones, opening up the possibility that there is additional selection on the latter set. Similarly, when the potential for N nucleotide addition in DETC progenitors was provided by use of transgenic terminal transferase, canonical rearrangements still accounted for .50% of in-frame joins (113). Either, as the authors of the study stated, the directing force of short homologies is resistant to N nucleotide addition, or there are extra factors (e.g. positive selection) favoring DETC TCR(`) cells. Formally, the inability of adult thymuses to support DETC development (86) may reflect a time-dependent failure to positively select DETC progenitors. Thus, although the generation of canonical DETC joins declines somewhat, 1–2 weeks postpartum (85), there is a more acute decline in the selection of canonical joins from among others made (87). Additional evidence for the importance of the canonical DETC TCR was derived from Vc51/1 mice. DETC developed in Vc51/1 mice, demonstrating that the canonical TCR is not required to form a skin IEL repertoire (114). Nevertheless, a significant, albeit variable, fraction of the ‘‘replacement’’ DETC showed conservation of an epitope defined by antibody 17.D1 (115). In vivo 17.D1 detects the canonical DETC TCR expressed by mature DETC or TCR(`) fetal DETC progenitors and reacts with no other cells (114, 115). Thus, 17.D1 retention in Vc51/1 mice indicates a greater conservation of TCR conformation than of the linear sequence determinants. Hence, the bias of gene rearrangements toward the canonical DETC TCR might be an evolutionary adaptation to a selec-

990

HAYDAY


tive advantage offered by the DETC TCR. Such an idea—namely, that beneficial antigen receptors can be hard-wired in the genome—is manifest elsewhere. Thus, Ig light chain gene families in cartilaginous fishes include prejoined V-J segments (18), while in mice, there is bias in the rearrangement of Vb to particular Jb segments (116). Vd4(`) CD8a(`) intestinal IELs were also reported to be positively selected by I-Ek (117). Subsequent studies using MHC class II1/1 mice failed to confirm a direct role for I-E (118), although recent data have reignited the possibility that I-E may be a regulatory locus for cd function in the periphery (see below). Positive Selection of Systemic cd Cells Despite study, there is no unequivocal resolution of whether systemic cd cells require ligand-dependent, clonotypic positive selection (119,120). Such evidence as exists for positive selection primarily pertains to peripheral events. In C57.BL/6 mice, Vc4(`)Vd7(`) cells with variable CDR3 regions expand dramatically within 4 weeks postpartum, independent of either the thymus or microbes (121). By contrast, the cells fail to expand in DBA/2 mice in the same housing. The prevalence in human peripheral blood of the Vc9Vd2 subset is also largely explained by antigen-driven extrathymic selection, since their thymic development is limited to the fetus (83). Extrathymic Development Although clonotype-specific positive selection of cd IELs has not yet been defined, it is tempting to think that such a process could be sustained by cortical epithelial cell ligands that resemble those recognized by mature cd cells on peripheral epithelia. Unlike ab T cells, cd cells develop seemingly normally in abnormal thymi with disorganized medulla. This idea has been extended to suggest that cd cell development might be sustained entirely by extrathymic epithelium (122, 123). Indeed, distinct follicles of the small intestine— ‘‘cryptopatches’’—have been described that can apparently support cd cell development (124). Cryptopatches from IL-71/1 mice could not do this, consistent with the IL-7 dependence of cd cells (125). We do not know, however, whether cd(`) IELs normally develop extrathymically (126).

cd CELLS DISPLAY A UNIQUE REPERTOIRE OF ANTIGEN SPECIFICITIES, LIKELY PRESENTED BY A UNIQUE MECHANISM TCR specificities have been reported for human systemic cd cells, human cd(`) IELs, and systemic murine cd cells. The molecular targets of human systemic cd cells are so widespread in nature that they would seem to confer on cd cells a capacity to ‘‘see’’ myriad bacteria, protozoa, and infected host cells. Likewise, human cd(`) IELs appear to target a broad range of infected or stressed autologous epithelial cells. It is therefore remarkable that no human cd cell specificities are obviously conserved in mice: This requires urgent reconciliation.

cd CELLS

991

Likewise unclear is the natural origin of cd cell ligands. Many putative microbial ligands (outlined below) have autologous counterparts expressed under conditions of infection or cell stress. It is therefore possible that cd cell responses to microbes are cross-reactivities of cells selected primarily for their recognition of self (51). The latter hypothesis is consistent with the response of cd cells to nonmicrobial tissue damage (50; see below).


Systemic Human cd Cells Recognize Varied Low-MolecularMass, Nonpeptidic Products Recognition and Presentation of Phosphoantigens for Vc2Vd2(~) Cells Following the observations that human peripheral blood Vc2Vd2(`) cells are enriched in Mycobacterium-induced lepromatous lesions, an intensive search was undertaken for the nature of mycobacterial antigens. In 1990, it was reported that the ligand was low molecular mass and nonpeptidic (127). From this point, attempts to purify mycobacterial ligands were continued in parallel with assays of synthetic, nonpeptidic compounds. It was reported in 1994 that monoalkyl phosphates of ,4 carbons induced in mycobacteria-reactive Vc2Vd2(`) clones stimulate indices (SI) of 4x to 80x, compared to a mean SI of 47.3 induced by phytohemagglutinin (PHA) (128). Numerous carbohydrates, phospho-aminoacids, nucleotides, and carboxylic acids failed to activate. The most potent compound was monoethylphosphate (Figure 4A), which promoted cytolytic activity in Vc2Vd2(`) cells among a mixed primary cd cell population. The TCR-dependence of reactivity was demonstrated by inhibition by anti-TCRd antibodies, and by the requirement for a Vc2,Vd2 TCR that could confer reactivity on transfected Jurkat T cells. Stringent chemical requirements were soon placed on the ligand (Figure 4A). Simple modifications, e.g. the addition of a b-OH group, a b-carboxylic group, or a b-amino group, reduced antigenicity by 100 times; likewise, diethyl- and triethylphosphates were essentially inactive, as were molecules containing aromatic rings. Some natural metabolites with related chemical structures (e.g. glycerol-3-phosphate, and ribose-1-phosphate) showed very weak reactivity (Figure 4B). By contrast, pyrophosphates [e.g. ethylpyrophosphate and allylpyrophosphate (unsaturated C3)] (Figure 4C) were highly antigenic, although simple modifications (e.g. phenylethylpyrophosphate) again abrogated activity. The inactivity of isoamylpyrophosphate stresses the importance of the unsaturated bond in the longer chain pyrophosphates (see below). The importance of the phosphate group is demonstrated by the phosphatase-sensitivity of responses. Significantly, these chemical properties were paralleled by the description of isopentenyl pyrophosphate (IPP) and its hydroxymethyl derivative as natural ligands (Figure 4D) (129). These compounds are highly abundant in mycobacterial supernatants. Additionally, IPP is an intermediate in the synthesis of terpenoids that modify signaling molecules in actively growing mammalian cells. This work followed the independent elucidation of natural ligands as chemically


cd CELLS

Figure 4 A. Synthetic antigens for human Va2Vd2 cells.

992


993

HAYDAY

Figure 4 B. Weak natural antigens.

unique c-derivatives of uridine triphosphate (X-cUTP) and thymidine triphosphate (X-cTTP) (TUBag 3 and 4), which spontaneously convert to the pyrophosphate X-PP (Figure 4D) (130). These antigens are primarily found in mycobacterial cytoplasm. Further studies indicated that reactivity was a property of several alkenyl or prenyl derivatives of phosphate, pyrophosphate, or nucleoside triphosphates, the latter two classes being particularly potent (Figure 4E). Nonetheless, the diversity of active compounds has confounded a straightforward chemical definition of antigenicity. For the phosphates, short C2 chains were most active, while the longer C5 saturated pyrophosphate, IAPP was inactive (Figure 4C). However, longer chains were tolerated in the highly active unsaturated prenyl pyrophosphates [e.g. geranyl geranyl (C20) pyrophosphate] (Figure 4D). The


994

HAYDAY

Figure 4 C. Synthetic active ligands.

nucleotide groups in TUBags 3 and 4 and their analogs contributed to antigen strength but were not necessary. All low-molecular-mass cd cell antigens seem to require cell surface presentation. Therefore, differences in structure may critically affect any of three parameters: TCR binding, binding to an antigen presentation molecule, and stability (the latter two may be linked). Evidence that distinct structural components contribute to independent processes has been derived using several IPP derivatives (M Bonneville, personal communication; see 131). The double bond that renders IPP active compared to saturated IAPP most likely reflects critical susceptibility to nucleophilic attack, since modification of the alkyl chain with more susceptible groups (e.g. halogens) produces exponentially more potent compounds. However, important topological constraints were revealed by the inactivity of modified but shorter alkyl chains. Third, reactivity was lost by substitution of nonhydrolyzable


cd CELLS

995

Figure 4 D. Strong natural antigens.

pyrophosphate analogs. Combined with the fact that inorganic pyrophosphate inhibited antigen recognition, this indicated that the phosphate moiety is a substrate for enzymatic processing. Such ‘‘processing and presentation’’ is MHCindependent and probably does not require access to intracellular compartments. Many Vc2Vd2(`) clones and cells are responsive to a single antigen, leading to speculation that the reactivities are CDR1/CDR2 based (see above). However, Jurkat cells transfected with TCRc/d chimerae (132, 133) identified CDR3 substitutions that destroyed EPP reactivity while retaining anti-TCR responses. The authors concluded that TCR junctional diversity is at least a part of the mechanism that distinguishes between numerous low-molecular-mass antigens. Whether TCRcd junctional diversity distinguishes between foreign and self-prenyl phosphates remains to be determined. This could clarify whether cd cells are selected on self-antigens, to be subsequently activated by the same or related antigens in an infectious or inflammatory context.

HAYDAY


996

Figure 4 Highly active synthetic ligands; active and inactive non-phosphoantigens. E and F, putative low-molecular-mass cd cell ligands.

cd CELLS

997


Identification of Nonphosphorylated Antigens for Vc9Vd2(~) Cells Despite the high phosphatase sensitivity of Vc9Vd2(`) reactivity to mycobacterial fractions, a further class of widespread, mycobacterial nonphosphorylated, alkyl amines was reported to stimulate Vc9Vd2(`) cells albeit at much higher concentrations (Figure 4F) (134). Treatment of PBMCs with isobutylamine and isoamylamine (plus IL-2) induced . ten-fold enrichments in Vc9Vd2(`) cells. Again, an involvement of the TCR was deduced from the responsiveness of Jurkat cells transfected with the appropriate TCRc/d chains, and cell-cell contact was required, likely reflecting surface presentation. Mono-amino derivatives of straight or branched C2-C5 alkanes are active, but potency varies (compare active isoamyl amine with inactive n-amyl amine) (Figure 4F). The Biological Context and Implications for Function of Such Reactivities Are Unknown Both the phosphoantigens and the antigenic alkyl amines are very abundant among bacteria and beyond, the alkyl amines being found in mammalian catabolites of plant compounds. Thus, the reactivities of cd cells toward these molecules will likely extend to recognition of Listeria, Borrelia, plasmodium, coccidia, and several other procaryotic and eucaryotic pathogens. They may also extend to self-recognition. In that vein, we are formally uncertain of whether ligands such as IPP, TUBag 3, or butyl amine are low-affinity ligands, akin to positively selecting ligands for ab T cells, or genuine activating ligands that if expressed in the thymus might negatively select Vc9Vd2(`) cells. Second, we do not know whether reactivity to low molecular mass foreign antigens is automatically accompanied by recognition of homologous self-antigens, or whether cd cells are somehow tolerized to self. A resolution to this will come from understanding antigen presentation. Although several putative antigens are secreted products, the precedent set by antigen presentation to B cells (59) suggests they may be modified to be effectively presented to Vc9Vd2(`) cells. To date, however, there is no evidence for either heterologous presenting cells or a clear costimulation mechanism. Because no murine cd cell reactivities to lowmolecular-mass antigens have yet been demonstrated, a rodent system has not been used to establish the biological importance of such responses.

Intraepithelial Human cd Cells Recognize MHC Class I Chain-Related (MIC) Antigens—A and B First Line of Defense To explain their limited diversity, we proposed that cd(`) IELs responded to self ‘‘stress antigens,’’ not diverse microbial ligands. cd cells would be tolerant to low-level expression of such molecules but would be activated when expression of stress antigens exceeded a threshold. We suggested that epithelial stress antigens might be nonconventional class I MHC-related molecules, such as CD1, Qa, or TL (51). Conversely, Allison and colleagues proposed that the stress antigens would be heat shock proteins (hsp) (52). These two hypotheses were married by


998

HAYDAY

the discovery of MHC class I chain-related genes, MICA, B, C, D, and E (the latter three of which are pseudogenes) located ;50 kb from HLA-B in the MHC locus and expressed from heat shock–responsive promoters by intestinal epithelial cells (135). This paper considers briefly only those issues relevant to cd cells. MICA/B are three-domain (a1–3), b2m-independent molecules, ;16–35% homologous to conventional class I MHC. The equivalent of the peptide binding groove created by the a1 and a2 domains on class I MHC is narrow and not obviously bound to electron-dense material. Moreover, it appears to be oriented away from the solvent-accessible face where it could readily interact with ligand (136). TCR cd(`) IELs explanted from a human MICA(`) colon carcinoma were stimulated by the tumor itself, and by human B cells transfected with a MICA construct containing only the a1 and a2 domains (137). Reactivity was inhibitable with anti-TCR antibodies and was exclusively a property of Vc1, Vd1(`) cells. There was no obvious limitation to junctional diversity, and reactive cells also responded to diverged, nonhuman primate MICA proteins (138). Hence, the nature of any direct MICA-TCR interaction and its relative insensitivity to junctional sequence variation remains unclear. Possibly it is driven by CDR1/2 (see above). The MICA locus is deleted from the murine genome. Hence, no small animal model has been used to clarify whether MICA-cd cell interactions contribute critically to tumor surveillance or to the response to infection. MICA Polymorphisms 16 MICA and 11 MICB alleles with variations in the ectodomain and the transmembrane domain have been described thus far. The effects of polymorphism on cd cell recognition have not been elucidated. The most common caucasoid allele, MICA 8, encodes a premature termination codon that may give rise to a secreted protein. The significance of this is unclear; possibly MICB is the more important gene in such haplotypes. MICA 8 homozygosity may associate with primary sclerosing cholangitis, a serious liver lesion from which cd cells can be readily isolated. Another allele, MICA 9, together with HLA-B51 composes a haplotype associated with Behcet’s disease, a multisystem inflammatory disease focused on mucosal surfaces that again features high numbers of cd cells (139). The Bw4 domain of HLA-B51 can be engaged by the killer inhibitory receptor (KIR) p70 that may be expressed by some cd cells. It is therefore possible that cd cell reactivity toward MICA(`) targets is influenced positively and negatively by the expression levels of TCRcd, MICA, conventional HLA-B, and NK-receptors. NKG2D Another NK receptor, NKG2D, was recently identified as a MICA ligand (62). NKG2D engagement activates DAP10, one of a growing family of NK receptor–linked activator proteins (140, 141). NKG2D is expressed on cd cells, NK cells, and cytolytic CD8(`) ab T cells. Thus, TCRcd is not obligatory for MICA recognition but may, on TCRcd(`) cells, assemble with NKG2D in a MICA-binding complex. Perhaps engagement of MICA by either TCRcd alone,

cd CELLS

999

or by TCRcd plus NKG2D, elicits different responses, akin to the importance of costimulation for the activation/tolerization of ab T cells.


Systemic Murine cd Cells Recognize Proteins, Including MHC, Differently to ab T Cells Recognition of Class II MHC The recognition of IEk, IEb, and IEs but not IEd by the Vc2(`) clone, LBK5, is independent of peptide and of intracellular antigen presentation and focuses on an epitope, IEb 67 and 70, close to the end of the groove. Reactivity is probably low affinity, akin to IgM recognition of native protein (6, 54, 142). LBK5 cells can be stimulated by a high density of immobilized, E. coli-produced IEk (142); hence, glycosylation is not required. Nonetheless, it carries an important influence: While IEk(`) CHO cells mutant in endosomal acidification stimulate well, cells mutant in the N-glycosylation pathway from the endoplasmic reticulum to the Golgi body do not. Likewise, recognition was inhibited by a charge reversal in position a79 that affects glycosylation. From this and other results, the important glycosylation site was ˚ away from b67–70, and therefore unlikely mapped to a82, probably about 30A to sit within the primary epitope (142). Thus, as discussed by Chien and colleagues, the effect of glycosylation may be to establish peripheral electrostatic interactions that are germane to specificity, rather than to net binding energy, analogous to those interactions proposed for human growth hormone and its receptor (143), and for residues peripheral to the CDR in hapten binding by Ig (34). Provocatively, glycosylation patterns are substantially altered by cellular transformation or infection. Recognition of Class IB MHC Two independently derived cd cell hybridomas, G8 and KN6, recognize two closely related MHC class IB gene products, T10 and T22 (94% identity), independently of the antigen-processing requirements for presentation to ab T cells (144, 145). Like MICA, T10/T22 is inducible on activated cells, and cd cell recognition may require a threshhold density of T10/ T22 expression akin to the activation of human cd cells by MICA overexpression. T10/T22 expression on B and T cells might explain the capacity of cd cells to recognize activated but not naı¨ve lymphocytes and to regulate them (see below). Also like MICA, a1 and a2 helices are significantly altered, precluding conventional peptide binding. Thus, G8 can be stimulated by a complex of T10 and b2m produced in E. coli and refolded without peptide (146). This material can interact with soluble G8 TCR with an estimated KD 4 0.13lM, with a fast on-rate and slow off-rate compared to TCRab. G8 can also be stained with a T10/T22 tetramer, which has allowed the frequency of T10/T22 reactive cells in vivo to be estimated as between 1 in 100 and 1 in 1000 of cd cells (Y-h Chien, unpublished). Recognition of Herpes simplex Virus, Glycoprotein I (gI) TgI4.4 is another Vc2(`) DN cd clone, derived from the popliteal lymph nodes of a C3H mouse

1000

HAYDAY


inoculated with HSV-1 (147). It recognizes HSV1-gI expressed by transfected fibroblasts or as a purified recombinant protein immobilized on plastic, independent of MHC (148). The distinction between TgI4.4 recognition of gI and ab T cell recognition of processed protein was highlighted by the inability of TgI4.4 to respond to cells transfected with gI mutants that could not reach the cell surface (148). MHC unrestricted, HSV-specific cd cells have been reported in PBLs of infected people (149). Based on this and the characterization of TgI4.4, cd cells may make important responses to viral infections of cells (e.g. neurons) that express little MHC, and to viruses, e.g. CMV, that downregulate MHC expression (see below). Other Reactivities of cd T Cells Murine Vc1(`) hybridomas could be stimulated by mycobacterial and mammalian homologs of GroEL and their peptides (150). Likewise, human Vc9Vd2(`) cells could recognize the human GroEL homolog (hsp58) expressed on the surface of Daudi, a human Burkitt lymphoma cell line (151). Additionally, human cd T cells reactive to B cell lymphomas reportedly recognize idiotypic determinants of Ig presented by grp75, another hsp (152). Thus, there seems no doubt that hsps can stimulate cd cells, but further biochemical experiments are required to define any direct interactions with TCRcd.

cd CELLS DISPLAY BROAD SPECTRUM OF CELL-CELL INTERACTIONS; IMPLICATIONS FOR cd CELL INVOLVEMENT IN IMMUNE RESPONSES AND AUTOIMMUNE DISEASE The context in which a cd cell sees antigen is not clearly elucidated, in either the afferent or efferent phase. However, consistent with their distribution in the circulation and in tissues, cd cells can interact functionally with many cell types, examples of which are reviewed below.

The Interaction of cd(~) IELs with Epithelial Cells For IELs, the epithelial cell is an obvious candidate for both antigen presenting cell and target of cd cell effector function. Assays of IL-2 or IFNc release and of cytolysis have demonstrated that murine DETC can respond to keratinocytes in a TCR-dependent fashion in the absence of overt antigen (114, 115, 153, 154). The responses (that may seem unimpressive compared to conventional ab T cell responses) are increased by stressing the keratinocytes, e.g. by heat shock. Nonetheless, no known hsp is obviously involved. There are obvious parallels to the MICA reactivity of human cd(`) IELs, but the molecules involved remain to be clarified. cd cell–epithelial cell interactions may also be mediated by other molecules, including NKG2D (62; our unpublished data) and the integrin aEb7 that

cd CELLS

1001

interacts with E-cadherin expressed on epithelial cells (155). It also becoming increasingly clear that cells within epithelia produce cytokines, such as IL-7, that can strongly activate cd cells (Figure 5). On activation, DETC and intestinal cd(`) IELs secrete IL-2 and become cytolytic. They also express biologically active KGF (68; see above) and upregulate lymphotactic chemokines, MIP1a, MIP1b, RANTES, and particularly lymphotactin, that can recruit CD8(`) T cells (156). These data indicate that in response to stimuli expressed by epithelial cells, cd cells can exert effector function directly on target epithelial cells (via cytolysis or KGF) and indirectly on other T cells (via chemokines and cytokines) (Figure 5).


The Interaction of cd Cells with Myeloid Cells Antigen and Cytokine Activation of cd Cells Myriad studies have demonstrated cd cell responses to monocytes, macrophages (Mf), and dendritic cells. Murine, human, and ovine cd cells proliferate strongly in allogeneic mixed lymphocyte reactions (MLR) (39, 57, 157–159), although such interactions may not be classically MHC-restricted. In cattle, autologous MLRs driven by self-antigens elicit the highest levels of activation in cd cells (159). Thus, myeloid cells may (in some form) present antigens to cd cells. Mf also provide cytokines to cd cells. Thus, peritoneal cd cells from Listeria-infected mice secrete IFNc in response to heat-killed Listeria and IL-12 ` IL-1, produced by activated Mf (160). This is a strong effector but weak proliferative response. Conversely, there is a striking

Figure 5 A hypothetical afferent and efferent interaction of cd (`) IELs with epithelial cells.

1002

HAYDAY


proliferative enrichment for IFNc-producing, IPP-reactive human cd cells in the presence of IPP and Mf1produced IL-12 ` IL-15 (161). Likewise, murine cd cell proliferation is driven by a combination of Mf–produced IL-1 and IL-7 (possibly produced by different cell types) (75). Together these studies demonstrate that different cytokines produced by activated Mf elicit powerful yet distinct responses from cd cells. cd Cell Suppression Since cd cells are distributed throughout tissues, their potential to interact with resident tissue macrophages is of interest. Alveolar macrophages (AMf) form the first line of defense against inhaled Mycobacterium tuberculosis, directly by controlling bacterial growth and indirectly by regulating T cell responses. In the presence of mycobacterial antigens, bronchoalveolar lavage AMf from healthy volunteers promoted the expansion of Vc9Vd2(`) cells in mixed cultures of small resting T cells, but only so long as the AMf concentration remained below a threshold (162). Above this, AMf inhibited resting (but not activated) cd cells, in a cell-contact dependent fashion. The inhibition could be partially overcome with exogenous IL-2 (162). Although the basis for these antigen-dependent interactions requires characterization, the implication is clear: cd cells within tissues may be suppressed by tissue Mf until there is a strong, activating local immune response, such as that provoked by mycobacterial infection. This may constitute a form of peripheral cd cell tolerance.

The Interactions of cd Cells with Other Lymphocytes cd Cells Help B Cells ab T cell–deficient mice fail to generate antigen-specific, class-switched antibodies following challenge with myriad animate or inanimate T-dependent antigens. Nonetheless, repeatedly infected TCRb1/1 mice often develop high titers of IgG, IgA and IgE and form germinal centers (GC) (163, 164). Neither occur as frequently in TCR(bxd)1/1 mice. GC formation requires cd cell–B cell contact and is abrogated in CD40L1/1 mice, whereas high-titer IgG is not CD40-dependent and probably reflects extrafollicular cd cell help (72). Th1 and Th2 cd clones that reconstitute B cell activation and GC formation after adoptive transfer to T cell–deficient mice have been characterized (71), consistent with which, cd cell help has been documented to drive the production of Th1associated and Th2-associated isotypes, including very high levels of IgE. Interestingly, TCRd1/1 mice showed measurable deficiencies in lamina propria and Peyer’s Patch IgA (but not IgM or IgG) production in response to oral administration of tetanus toxoid plus cholera toxin (165). The emphasis on IgE and IgA production is provocative given that those isotypes, like cd cells, are associated with body surfaces. In humans, too, activated cd cells could help IgE synthesis in a CD40dependent fashion (166). Likewise, cd cells helped B cells stimulated with Paracoccidioides brasiliensis (167), and cd clones that help autoreactive B cells were derived from a lupus nephritis patient (168). Moreover, small numbers of

cd CELLS

1003


CD4(`) cd cells found in the GCs of normal animals, including humans, were readily apparent in draining lymph node GCs of sheep infected with Taenia hydatigena (39). Thus, cd cell–B cell interactions contribute to numerous cases of immune responses and immunopathology in several species. These interactions are not obviously driven by cognate antigen specificities, since the antibody reactivities that form are commonly self-reactive (69, 163, 164). The interaction may be related to the hsp-mediated response of cd cells to Daudi B cell lymphoma (see above). Interestingly, Daudi cells and other B cells that stimulate cd cells are commonly MHC-class I–deficient. This may help overcome inhibition of cd cell activity mediated by cd cell expression of NKG2A (61). cd Cells Regulate ab T Cells Notwithstanding evidence that cd cells contribute to a first line of defense, prior to activation of the conventional adaptive response, there are many cases where increases in cd cell numbers follow (and are dependent on) ab T cell activation. These cases exemplify cd cell–ab T cell crossregulation. A case in point is Borrelia burgdorferi–induced Lyme arthritis (169). Budd and colleagues reported on seven patients in whom cd cells (mostly Vd1`) in synovial membranes expanded to 18.9 ` 6.8% of CD3(`) cells (170). Over 6 days in culture, the cd cells expanded threefold in response to Borrelia sonicates. Interestingly, there was a striking loss of CD4(`) ab Th2 cells. Further studies of cd cells and 18 cd clones demonstrated their high, prolonged expression of FasL and their capacity to kill fas(`) CD4(`) ab Th2 cells. Whether this TT interaction involved TCRcd was unclear: Probably the TCR was primarily engaged in the initial activation of cd cells by Borrelia antigens (see above).

ab Cells Regulate cd T Cells In a detailed study of human PBMC that would ‘‘present’’ crude mycobacterial antigens to Vc2Vd2(`) cells, negligible activity was found in purified monocytes, B cells, or DC. Likewise, reactivity was unaffected by depletion of monocytes, B cells, DC, cd cells, or CD8(`) T cells. Conversely, it was severely reduced by CD4(`) T cell depletion, and the only purified cell population with potent presenting capacity was the CD4(`) CD45R0` TCRab(`) memory-type helper subset (171). This stimulation of cd cells by ab T cells was abrogated by a semipermeable membrane, thus excluding cytokines as the sole cause. Conversely, cytokine stimulation largely explained the capacity of more than 1000 irradiated, activated ab T cells from Listeriainfected mice to drive the proliferation of 10,000 cd cells (the contrary was not true). The regulation of cd cell responses by CD4(`) ab T cells was also noted after Leishmania and Plasmodium infections(172, 173). The capacity of activated ab T cells to assist cd cell activation has significant implications for immunopathology in inflammatory sites such as the rheumatoid joint and inflamed skin, where the two cell types colocalize. cd cell recognition of induced class IB MHC might explain interactions of cd cells and T cells. However, we regard antigen presentation as the focus of the interactions of DC, Mf, and B cells with ab T cells, the question arises as to

1004

HAYDAY

whether cd cells interact with ab T cells by presenting to them antigen internalized via the TCR and re-presented on MHC or CD1. It is striking that cd cells commonly express CD1 (71), the only known function of which is to present antigens to T cells.

A UNIQUE CAPACITY OF cd CELLS TO PROVIDE PRIMARY PROTECTION AGAINST SPECIFIC PATHOGENS: RELEVANCE TO HUMAN DISEASE


cd Cells Are Activated by Infection We need to understand the biological contexts in which cd cells interact with other cells. One such context is infection. Oligoclonal expansion, changes of cell surface markers, and expression of effector function have all been used to demonstrate cd cell responses in vivo to myriad challenges, including Mycobacteria, influenza, Listeria monocytogenes, Epstein Barr virus, Plasmodium, visceral Leishmania, Toxoplasma, and Salmonella (174–181). Some infections involve specific TCRs. Thus, human Vc2Vd2(`) cells expand in tuberculosis, leprosy, malaria, toxoplasmosis, and EBV infections, while Vd1(`) cells expand in Lyme disease, CMV, and HIV. Some infections also involve a specific cd cell phenotype. Thus, although cd cells are most often Th1-like, infection with an organism that elicits a Th2 response (e.g. Nippostrongylus) may elicit a Th2 cd cell response (70). This has provoked the idea that cd cell–produced cytokines, together with cd cell interactions, participate in skewing the ab T cell effector response (70, 71, 74, 182). The use of ab T cell–deficient mice has defined ab T cell–independent cd cell responses, which likely reflect a similar situation in humans.Thus, in immunosuppressed renal allograft recipients, dramatic increases were reported in Vd1(`) or Vd3(`) cd cells that correlated both retrospectively and prospectively with cytomegalovirus (CMV) infection and that on some occasions lasted for more than one to three years (183, 184). The cdTCR CDR3 regions were limited in diversity, suggestive of an antigen-driven response, consistent with which explanted cells responded to free virus or to CMV-infected cells. The authors hypothesized that the primary stimulus was in the gut, and it is conceivable that MICA induction by CMV was a driving force. Long-term changes were likewise reported for cd cell activation markers in patients with mycobacterial and acute pulmonary disease (185). Kinetics and Localization Infection of experimental animals has allowed the kinetics and the localization of cd cell responses to be measured. While rapid responses are common, particularly in the context of the proposed first line of defense function, these are not exclusive of later responses, that may relate to epithelial repair and/or immunoregulation (see below). Thus, the major increase

cd CELLS

1005


in bronchial lavage cd cells following influenza infection occurs after the first week, following the ab T cell response (175). A biphasic response of intestinal cd(`)IEL expansion was noted after infection of mice with the natural coccidian parasite Eimeria vermiformis, delivered naturally to the epithelium via gavage (186). Following injection of Listeria into the peritoneal cavity of mice, there was a tenfold, highly localized expansion of cd cells; nothing was apparent in the spleen or lymph nodes. Interestingly, cd cell expansion peaked 2–3 days after the resolution of infection, and the increased percentage of CD3(`) cells that were TCRcd (`) was sustained for ;6–7 weeks (187). Listeria strains lacking listeriolysin show weaker pathogenicity and fail to induce cd cells. From this and other studies, the cd cell response was attributed to the infectious process (‘‘danger’’), rather than to the injection of either microbes or inanimate nominal antigens (163, 187). It remains plausible that self-‘‘stress’’ antigens are as important as foreign antigens in evoking the cd cell response to danger.

cd Cell Contributions to the Primary Response Can Be Essential and Nonredundant Responses in the Lung In some striking cases, no other lymphoid or myeloid cells can substitute for cd cells in providing primary protection. One example is the infection of mice with Nocardia asteroides, an intracellular gram(`) bacterium, ordinarily cleared from infected airway epithelial cells by an inflammatory neutrophil response. Following intranasal inoculation, 100% of cd1/1 C57.BL/6 died within two weeks, compared to no deaths among controls (188). The cd1/1 mice failed to develop inflammatory infiltrates of polymorphonuclear leukocytes and Mf. Ozone provokes a nonpathogenic insult to approximately the same region, although different cells types are the primary targets of damage. The insult is again accompanied by neutrophil influx that clears necrotic epithelial cells. In TCRd1/1 mice, this influx was again reduced, leading to a build-up of damaged tissue (188). Three important points are made: First, cd cell contributions to primary protection can be essential; second, those contributions can involve collaborations with other cell types; and third, there is no overt requirement for ‘‘foreign-ness’’ for cd cell activation in vivo; tissue damage will suffice. Whether these effects reflect unique capacities of cd cells will be revealed when abT cell–deficient mice are tested in the same system. Responses to Listeria Perhaps the best-studied example of cd cell contribution to the primary response has been the infection of naı¨ve mice with L. monocytogenes. The response to this challenge is biphasic, with Mf and NK cells being critical ‘‘innate’’ effectors in the first phase–-a period of 3–6 days post infection (D3PI-D6PI), with ab T cells responsible for the second phase of clearance, and the establishment of immunity (189, 190). Depletion of mice with anti-cd antibodies revealed a substantial role for cd cells in the first phase of the response:

1006

HAYDAY


Bacterial titers were increased 100- to 1000-fold by D8PI and persisted beyond the usual period of clearance (187, 191, 192). The role of cd cells is attributable, at least in part, to the regulation of IFNc production by NK cells, which is significantly reduced in TCRd1/1 mice (193, 194). NK IFNc production is largely stimulated by Mf-produced IL12 ` TNFa, rather than IL12 and IL1, as is the case for cd cells (160). IL12 levels are unaffected in infected TCRd1/1 mice, but TNFa levels are reduced, consistent with which cd cell-produced IFNc ordinarily promotes Mf production of TNFa in vivo (193, 194). Hence cd cells activated by antigens and by cytokines may be a significant source of IFNc, complementing and indirectly promoting IFNc production by NK cells. cd cells are expanded in the peritoneum and other tissues, where NK cells can be less abundant. In sum, a protective effect of cd cells is mediated via a collaboration with cells of the innate immune system. Viral Infections Biphasic protective responses also characterize infections by viruses, including vaccinia (VV), ectromelia, and CMV. The early control of virus growth is largely attributable to perforin or IFNc-dependent NK cell activity, whereas the late phase is largely ab(`) CD8` T cell dependent. Intraperitoneal infection of TCRb1/1 mice with 106 pfu of vaccinia virus (VV) [WR strain], elicited 20-fold increases in peritoneal exudate cd cells, and measurable, albeit smaller, increases in peripheral blood cd(`) cells (195). These cd cells included IFNc-producers and cells with cytolytic activity. Their functional role was indicated by increases of splenic and footpad virus titers of 10–100-fold in TCRd1/1 C57.BL/6 mice at D3PI- D4PI (195)—strikingly similar kinetics to the involvement of cd cells in Listeria responses. Moreover, the increased susceptibility was similar to that obtained by treatment with anti-NK1.1 antibody. By contrast, ab T cell-deficient mice showed weakened control of the virus until D12PI, when viral titers increased uncontrollably and fatally. Other Infections Ordinarily, the numbers of cd cells and Mf both sharply increase following infection with Candida albicans. Co-culture of the two cell types elicits striking increases in production of nitric oxide (NO). When cd-cell depleted mice were infected, the levels of inducible NO-synthetase RNA were reduced, and there was an increased susceptibility to Candida infection (196). It has also been claimed that cd cells are essential for the immuno-protective containment of M. tuberculosis (197). However, the pathogen was delivered to TCRd1/1mice as a massive inoculum i.v. No such role for cd cells was reported by other groups, particularly when natural aerosols were used to deliver even highly virulent isolates (198).

cd Cell Contributions to the Primary Response Are Commonly Nonessential or Redundant Commonly, the antipathogen effector function that cd cells provide upon activation in vivo, is easily substituted for (and hence masked by) ab T cells. Such

cd CELLS

1007

‘‘cryptic’’ cd cell effector responses are unveiled in TCR(bxd)1/1 mice. Thus TCR(bxd)1/1 mice are measurably more susceptible than TCRb1/1 mice to primary infection by the gut pathogen Eimeria vermiformis (AL Smith & AC Hayday, unpublished). Antibodies to IFNc abrogated the effect of cd cells, consistent with the cd cell bias toward Th1. Following HSV-1 infection, TCRab1/1 and TCRcd1/1 mice are both essentially immunocompetent, whereas infection is fatal in TCR(bxd)1/1 mice (199). The likelihood that cd cells ordinarily contribute to immunoprotection in normal mice was indicated by an increased viral load in the trigeminal ganglion of cd cell-depleted Balb/c mice early after corneal infection with HSV-1. As mentioned above, HSV-1 reactive cd(`) clones have been identified and isolated (147).


cd Cells Are Seldom Important in the Recall Response—A Lack of Memory? TCRd1/1 and TCRb1/1 mice have been assayed for long-term immunoprotection against bacterial, viral, protozoal, and fungal pathogens. While immune responses inevitably vary among challenges, a general conclusion could be drawn: ab T cells not cd cells provide effective, cognate immunity to rechallenge (187, 195, 196, 200). Thus, TCRb1/1 mice, irrespective of their cd cell repertoires, are severely immuno-compromised. By contrast, immunity to myriad challenges develops essentially normally in TCRd1/1 mice. Why cd cells rarely contribute to functional memory is unclear. Possibly the context in which they engage antigen, and/or putative differences in their intracellular signaling programs, vis-a`-vis ab T cells (see above) preclude commitment to a functional memory state, irrespective of the expression of ‘‘memory markers’’ such as CD45R0 (83). Alternatively, most cd cells may exist de facto as circulating memory cells following activation by foreign or self-ligands, early in the host animal’s development, and hence may already be responding rapidly. This would be consistent with the widespread occurrence of cd cell ligands, and their mammalian homologs. However, such ‘‘memory cd cells’’ may have limited effector capability, or, for reasons of specificity, may fail to establish reservoirs of high-affinity, pathogen-specific, recall cells. Hence, they fail to provide complete protection against pathogen rechallenge. Nonetheless, there may be some exceptions, e.g. two cd lines isolated from Plasmodium yoelii (rodent malaria) sporozoite-immunized TCRa1/1 mice that could restrict the liver stages of malaria growth after adoptive transfer to naı¨ve recipients (201). However, subsequent studies have questioned the similarity of such cd cells to conventional, pathogen-specific, memory cells (202).

cd Cells Regulate Infection-Associated Inflammation While there are variable effects of cd cell deficiency on the control of different pathogens, the course of immune responses seems always to be altered. Intraperitoneal inoculation of mice with Listeria is followed by bacterial growth in hepa-


1008

HAYDAY

tocytes and splenic macrophages. Following infection, ab T cell-depleted and cd T cell-depleted mice each showed exacerbated liver lesions (190, 192, 203). However, whereas lesions in the former were rich in bacteria and almost devoid of neutrophils, those in cd cell-depleted mice were abscessed and rich in polymorphonuclear leukocytes. The abscesses were surrounded by circles of necrotic hepatocytes infiltrated by neutrophils, and there was extensive parenchymal damage (192). The phenotype was exacerbated upon secondary infection. Consistent with their overall control of pathogen growth, the cd-deficient mice did not show excessive bacterial accumulation. Eimeria vermiformis targets intestinal epithelial cells in the crypts and villi. Both TCRb1/1 and TCRd1/1 mice show abnormal but distinct phenotypes following gavage infection (200). TCRb1/1 mice are highly susceptible to parasitic infestation, but overall villus architecture rapidly recovers. By contrast, infected TCRd1/1 mice show villus tip and capillary damage, and some evidence of neutrophil infiltration. The phenotype was reduced by adoptive transfer of cd(`) cells to TCRd1/1 mice. However, the phenotype was not shown by TCR(bxd)1/1 mice, indicating that the uncontrolled pathology in TCRd1/1 mice is due (directly or indirectly) to ab T cells (200). A similar phenomenon was evident in the exaggerated pathology of some Listeria-infected TCRd1/1 mice (203). However, unlike with Listeria, the heightened pathology in Eimeriainfected TCRd1/1 mice was apparent only during primary infection. Memory responses to Eimeria are extremely effective and secondary infections resolve quickly, presumably before significant immunopathology is induced. When low doses of M. tuberculosis Erdman, or the highly virulent strain, CSU46, were delivered intranasally to TCRd1/1 mice, containment of the pathogen was unaffected, but there was a highly exaggerated pyogenic granulomatous response (198). Compared to granulomas in normal mice, formed at ;D21PI, those in TCRd1/1 mice were enlarged, rich in neutrophils and foamy Mf, but depleted of lymphocytes. The authors proposed that cd cells triggered by mycobacterial Ags promote, directly or indirectly, the necessary local cytokine/chemokine signals for a protective lymphocytic/monocytic granuloma (198). From the cases we have considered, this ‘‘shaping of the immune response’’ may be a general property of cd cells. In certain mice, Coxsackievirus B3 (CVB3) elicits T cell-dependent, Th1-type myocardial inflammation (204). MHC class II is a predisposition locus: IE(1) C57.BL/6 mice are resistant to myocarditis, despite high viral titers in the heart, while IE(`) transgenic C57.BL/6 mice are susceptible. Likewise, HLA-DR4 /1 and a histidine polymorphism in HLA-DQ b1 are susceptibility alleles in humans (205). Another predisposing factor is TCRcd. cd cell depletion in IE(`) strains abrogated myocarditis, while the IFNc:IL4 balance shifted from Th1 to Th2 (204). This may be explained by the cytolytic interaction of activated FasL(`) cd cells with Fas(`) Th2 cells (see above). The linkage to MHC class II is intriguing. Mice congenic at MHC class II may harbor different peripheral cd cell repertoires (117). More concretely, TCRcd recognition of MHC class II has been described

cd CELLS

1009

(54; see above). Alternatively, the class II MHC locus may determine the activation of CD4(`) ab Th2 cells that are then targets of cd cell regulation. Myocarditis is a serious human disease, twofold more prevalent in males. Incidence in female mice was increased by injection of 10 mg of testosterone; following infection with CVB3, CD69(`), cd(`) T cells increased in the spleen approximately threefold above normal (206).

A NONREDUNDANT CAPACITY TO REGULATE THE COURSE OF IMMUNE RESPONSES: IMPLICATIONS FOR HUMAN DISEASE


cd Cells and Inflammation The capacity of cd cells to interact with various target cells and to regulate infection-associated inflammation provokes the question as to whether cd cells also regulate noninfectious ‘‘stress.’’ Infection of a single testicle with Listeria induces orchitis, a lymphoid infiltration of the contralateral testicle, in the absence of bacterial spread. Prior to insult, the testicle is essentially lymphocyte-free. Hence, orchitis represents a breakdown in immunological tolerance (207). The cellular infiltrate includes ab T cells and canonical Vc6(`) cells, ordinarily associated with the genital epithelium (208, 209). Interestingly, orchitis, albeit weaker, can be mimicked by repeated injection of syngeneic testicular cells. In this case, infiltrates of the same ‘‘invariant’’ Vc6(`) subset were noted. Strikingly, the inflammatory spread is accelerated and exaggerated in TCRd1/1 mice (209), indicating that cd cells ordinarily downregulate inflammation. Such downregulation of inflammation by cd may apply more generally. One of the earliest descriptions of human cd cell expansion under nonmicrobial conditions was the abundance of cd(`) intestinal IELs in celiac disease (CD) (210). CD is an inflammatory gut allergy to gliadin peptides in cereal-containing diets, which develops into an autoimmune response to transglutaminase and other autoantigens (211). The disease is seemingly driven by gliadin-reactive, MHC class II–restricted ab T cells (212). In its full presentation, gut villi are effaced, and the crypts of Lieberkuhn are massively hyperplastic. Prior to this, CD progresses through silent or latent phases during which patients are asymptomatic and cd cells are particularly abundant (up to one cell per five enterocytes). Their representation declines as disease worsens. Hence it was hypothesized that cd cells are immunoregulatory, responding either to stressed epithelial cells and/or to activated T cells and limiting inflammatory damage (213).

cd Cells and Epithelial Integrity There is no known microbial involvement in Crohn’s disease (CD), a form of inflammatory bowel disease (IBD). And yet, there are reports of increased Vd1(`) cells (214), possibly driven by MICA misexpression. IBD can be mod-

1010

HAYDAY

eled in mice by dextran ingestion. In following the disease in T cell–deficient mice, a decrease in local KGF production in TCRd1/1 mice was correlated with prolonged tissue damage (215). Predictably, such regulation of epithelial tissue by cd cells may apply in other sites also. The invariant DETC subset expands in response to active hair growth or chemical sensitization of the skin (154, 216). It has furthermore been proposed that these are developmental effects: The turnover and maturation of enterocytes were reported to be dysregulated in TCRd1/1 mice (217), and, perhaps not unrelated, one cd cell subset expands ;100-fold in the reproductive tract during pregnancy (218). However, it is very difficult to segregate purely developmental effects from effects of microbes to which mucosal surfaces are chronically exposed.


cd Cells and Autoimmunity The capacity of TCRcd to engage antigens that, under particular circumstances, are expressed by the host (219) may underlie cd cell help for autoantibody production and the development, albeit variably, of autoimmune diseases in ab T cell–deficient mice (220, 221). cd cells can also downregulate autoimmune disease. Autoantibodies titers, CD4(`) ab T cell numbers, and Ig deposition in the kidneys are all dramatically increased in TCRd1/1 MRL.lpr mice relative to ‘‘normal’’ MRL.lpr mice, possibly because of a failure of cd cell–ab T cell regulation (see above). Significantly, TCRd1/1 MRL.lpr mice show a threefold increase in mortality at 24 weeks, from about 23% in MRL.lpr mice to 68% (222), providing clear evidence that cd cell dysfunction on a particular genetic background can have a major impact on life expectancy. Further studies of cd cell deficiency on different genetic backgrounds need to be undertaken. In this vein, cd cells were reported to suppress anti-islet cell reactivity in NOD mice following repeated intranasal administration of insulin or glutamic acid decarboxylase peptides (223).

cd Cells and Transplantation The capacity of an autoreactive ab T cell clone to attack Langerhans cells in the skin of mice into which it was adoptively transferred progressed uncontrolled in TCRd1/1 mice (224, 225). Grafts of putative DETC progenitors from the fetal thymus ameliorated the autoaggressive reaction (225). This downregulation of ab T cells finds parallels in the prolongation of allogeneic grafts, mediated by cd cell suppression of IL-2 and IFNc production by mesenteric lymph node Th1 cells (226). These results raise the issue of whether deliberate regulation of cd cells might be beneficial during transplantation.

cd Cells and Hypersensitivity The contrast between effector ab T cells and regulatory cd cells is illustrated in contact sensitivity (CS). CS to picryl choride and oxazolone is ab T cell dependent. However, high-dose tolerance induced by prior systemic inoculation of the

cd CELLS

1011

antigens was achieved in TCRa1/1 mice. In mixing experiments, cd cells from tolerized mice downregulated the capacity of effector ab(`) T cells to transfer CS to recipients (227). Downregulatory effects of cd cells on IgE production to ovalbumin (a Th2 response) have been described following intranasal tolerization (228). Nonetheless, while a general pattern of cd cell–mediated regulation would seem to be emerging, the generality of the latter study has been subsequently challenged (229).


cd Cells and Tumor Surveillance Surprisingly, TCR(bxd)1/1 MRL/lpr mice also showed a shorter life span than MRL/lpr counterparts (above), not because of autoimmune disease but because of B cell lymphoma. With neither T cell subset present, B cell expansion was fatally uncontrolled (230). The corollary of this is that cd cells suppress B cell expansion in vivo, possibly by promoting B cell differentiation (see above). Hence, some B cell lymphomas may be a manifestation of defective cd cell function. There is likewise evidence that cd cells can regulate T cell leukemias (231) and possibly carcinomas via engagement of MICA upregulated on transformed epithelial cells.

UNIQUELY AGE-DEPENDENT ACTIVITIES cd—AN HYPOTHESIS Clearly cd cells respond to infection and to noninfectious stress. Hence, many immune responses in normal animals will feature a cd-cell component. Nonetheless, the common redundancy of cd cell effector function in the control of pathogens makes it difficult to conceive of the selective forces that would have conserved cd cells across 450 million years. Possibly, a critical, selected phenotype is an immunoregulatory one (see above). Unregulated immunopathology and/or deficiencies in epithelial tissue repair would presumably decrease reproductive fitness, and ruminants such as cattle that have high numbers of cd cells must presumably avoid the chronic inflammation that would otherwise accompany relentless exposure to mucosal pathogens. This notwithstanding, we have examined an additional hypothesis, namely that cd cells provide immunoprotective functions in instances where ab T cells do not. Properties of cd cells that underlie this hypothesis have been outlined in this review. For example, eradication of MHC function, either by failure of peptides to bind or by intervention in MHC synthesis, is a primary route of immunoevasion, e.g. by CMV. By recognizing antigen independent of MHC, cd cells might control such infections (183, 184). ab T cell function is also compromised in early life. Young animals (including children) show defective ab(`) Th1 responses. Provocatively, cd cells are Th1biased, are disproportionately abundant in young individuals (1, 39, 83), and populate the body surfaces that are vulnerable to infections de novo. Moreover,


1012

HAYDAY

cd cells are disproportionately common in young animals (e.g. birds and ruminants) that are precocious from birth yet receive no transplacental maternal antibody. Added to this, cd cells can functionally collaborate with antimicrobial cells (e.g. Mf and NK cells) (see section cd Cells Display a Broad Spectrum of CellCell Interactions) and require neither intracellular antigen processing nor a sophisticated antigen sampling anatomy, neither of which may be fully developed at birth. Finally, cd cells (directly or indirectly) can restrict pathogen growth, and/ or promote tissue repair, but do not contribute to long-lived memory (see above). To test the hypothesis, we used gavage needles to deliver E. vermiformis to mice 1–8 days post weaning, and prior to sexual maturity. Three points became clear. First, in complete contrast to adult mice, TCRab deficiency did not affect susceptiblity to primary infection. Second, also in contrast to adult mice, TCRd1/1 mice commonly showed significantly increased susceptibility. Third, consistent with studies in adult mice, the memory response still relied on ab T cells (E Ramsburg, A Hayday, unpublished). A functional capacity of cd cells in young animals was further suggested by their highly activated surface marker profile (CD44hi, CD62Lhi) compared to ab T cells at the same age. Interestingly, spontaneously autoreactive Vc1(`) cells could be derived from the gut of weanling but not older mice (232). If cd cells prove more generally to offer immunoprotection to young animals, it would suggest a selective pressure acting for their conservation. Moreover, the different degrees and types of challenge experienced by different young animals might explain species-specific deployment of cd cells.

CONCLUDING REMARKS AND FUTURE DIRECTIONS cd cells are unique. They are tissue lymphocytes, but they are not limited to providing a first line of defense. Sometimes their activation follows that of ab T cells, and in such contexts, cd cells probably contribute more to immunoregulation and tissue repair than to immunoprotection. Like all lymphocytes, cd cells have inherent autoreactivity. Nonetheless, the common overlap of pathogenencoded cd cell antigens with molecules expressed by ‘‘stressed’’ host cells has fueled the view that cd cell responses are driven more by imbalance in the host (inflammation, cell transformation, tissue damage) than by specific pathogen challenge. cd cell autoreactivities must presumably be tightly controlled. The complex cytokine requirements of cd cells and the capacity of some cells to downregulate them have all been reviewed. Nevertheless, the capacity to clonotypically expand or to inactivate cells with different specificities is key to the adaptive response, and our failure so far to clarify mechanisms underlying cd cell selection/activation/tolerization has encouraged a view that cd cells might belong to the innate response. The unique specificities of the cells, their liberation from any requirement for conventional antigen presentation, their anatomical distribution, and

cd CELLS

1013


their early development all point to an absolutely critical role for cd cells, at the right time and in the right place. We would propose that this is at the body surfaces of young animals, but this is not to exclude contributions to adult immune responsiveness. In the coming years, cd cell–specific marker genes should allow characterization of cd cell–specific progenitors, while their associated regulatory elements could drive cd cell-specific knockouts of effector molecules (e.g IFNc or KGF), so that the contribution of cd cells to effector function in vivo can be defined unequivocally. With the implication of cd cells in immunoregulation, it would be invaluable to link polymorphisms in the human c/d loci to disease susceptibility. Likewise, mouse genetics could be used to define modifier genes for cd function. Finally, one wants to ascertain the potential contribution of cd cell activation to vaccine and immunotherapy regimens. At such a point, cd cells would be established as a biological phenomenon and possibly as a force in the pharmacologist’s armamentarium. ACKNOWLEDGMENTS The author extends thanks to Carrie Steele, Eric Hoffman, Jeremy Cridland, and David Oppenheim for help with figures and with sections of the text; to Elizabeth Ramsburg for permission to quote unpublished data; to many valued colleagues including Marc Bonneville, J-J. Fournie, Willi Born, Becky O’Brien, Max Cooper, Y-h. Chien, Mike Krangel, Stefan Kaufmann, Richard Boismenu, Lisa Selin, Sally Huber, Ralph Budd, and Craig Morita for provision of prepublication data, and particularly to Suzanne Creighton for her work on the text, and to the Wellcome Trust and the NIH for support. Visit the Annual Reviews home page at www.AnnualReviews.org.

LITERATURE CITATIONS 1. Allison JP, Havran WL. 1991. The immunobiology of T cells with invariant cd antigen receptors. Annu. Rev. Immunol. 9:679–705 2. Raulet DH. 1989. The structure, function, molecular genetics of the c/d T cell receptor. Annu. Rev. Immunol. 7:175– 207 3. Haas W, Pereira P, Tonegawa S. 1993. c/d cells. Annu. Rev. Immunol. 11:637– 85 4. Raulet DH. 1994. How cd T cells make a living. Curr. Biol. 4:246–48

5. Hayday AC. 1995. cd T cell specificity and function, In The T Cell Receptor, ed. J Bell, MJ Owen, E Simpson. Oxford, UK: Oxford Univ. Press 6. Chien Y-H, Jores R, Crowley MP. 1996. Recognition of c/d T cells. Annu. Rev. Immunol. 14:511–32 7. Kaufmann SHE. 1996. cd and other unconventional T lymphocytes: What do they see and what do they do? Proc. Natl. Acad. Sci. USA 93: 2272–79 8. de Libero G. 1997. A sentinel function

1014

9.

10.

11.


12.

13.

14.

15.

16.

17.

HAYDAY of broadly-reactive human. c/d T cells. Immunol. Today 18:22–26 Hayday AC, Pao W. 1998. T cell receptor, cd. In Encyclopedia of Immunology, ed. IM Roitt, PJ lves, pp. 2268–78. London: Academic Born W, Cady C, Jones-Carson J, Mukasa A, Lahn M, O’Brien R. 1999. Immunoregulatory functions of cd T cells. Adv. Immunol. 71:77–144 Beckman EM, Porcelli SA, Morita CT, Behar SM, Furlong ST, Brenner MB. 1994. Recognition of a lipid antigen by CD1-restricted ab T cells. Nature 372:691–94 Hayday AC, Saito H, Gillies SD, Kranz DM, Tanigawa G, Eisen HN, Tonegawa S. 1985. Structure, organization and somatic rearrangement of T cell gamma genes. Cell 40:259–69 Chien Y-H, Iwashima M, Kaplan K, Elliott JF, Davis MM. 1987. A new T cell receptor gene located within the alpha locus and expressed early in T cell differentiation. Nature 327:677–82 Lefranc M-P, Chuchana P, Dariavach P, Nguyen C, Huck S, Brockly F, Jordan B, Lefranc G. 1989. Molecular mapping of the human T cell receptor gamma (TRG) genes and linkage of the variable and constant regions. Eur. J. Immunol. 19:989–94 Six A, Rast JP, McCormack WT, Dunon D, Coutois D, Li Y, Chen C-L, Copper MD. 1996. Characterisation of avian T cell receptor gamma genes Proc. Natl. Acad. Sci. USA 93:15329–34 Rast JP, Anderson M, Strong SJ, Luer C, Litman RT, Litman GW. 1997. a,b, c and d T cell antigen receptor genes arose early in vertebrate phylogeny. Immunity 6:1–11 Chen C-L, Kubota T, Wang J-Y, Gobel T, Hockett R, Cooper MD, Chen C-L. 1999. Characterization of an avian (Gallus gallus domesticus) T cell receptor a/d gene locus. J. Immunol. 163:3858– 66

18. Anderson MK, Shamblott MJ, Litman RT, Litman GW. 1995. The generation of immunoglobulin light chain gene diversity in Raja erinacea is not associated with somatic rearrangement, an exception to the central paradigm of B cell immunity. J. Exp. Med. 181:109–19 19. Reynaud C-A, Anquez V, Grimal H, Weill J-C. 1987. A hyperconversion mechanism generates the chicken light chain pre-immune repertoire. Cell 48: 379–88 20. Li H, Lebedeva MI, Llera AS, Fields BA, Brenner MB, Mariuzza RA. 1998. Structure of the Vd domain of a human cd Tcell antigen receptor. Nature 391:502 21. Arden B, Clark SP, Kabelitz D, Mak TW. 1995. Human T-cell receptor variable gene segments. Immunogenetics 42:455– 501 22. Arden B, Clark SP, Kabelitz D, Mak TW. 1995. Mouse T-cell receptor variable gene segment families. Immunogenetics 42:501–30 23. Clark SP, Arden B, Kabelitz D, Mak TW. 1995. Comparison of human and mouse T-cell receptor variable gene segment subfamilies. Immunogenetics 42:531–40 24. Garman RD, Doherty PJ, Raulet DH. 1986. Diversity, rearrangement, expression of murine T cell gamma genes. Cell 45:733–42 25. Chien Y-H, Davis MM. 1993. How ab T cell receptors ‘see’ peptide/MHC complexes. Immunol. Today 14:597–602 26. Garcia KC, Degano M, Stanfiled RL, Brunmark A, Jackson MR, Peterson PA, Teyton L, Wilson IA. 1996. An ab T cell receptor structure at 2.5 angstrom and its orientation in the TCR-MHC complex. Science 274:209–19 27. Garboczi DN, Ghosh P, Utz U, Fan QR, Biddison WE, Wiley DC. 1996. Structure of the complex between human T-cell receptor, viral peptide, HLA-A2. Nature 384:134–41 28. Chen C, Roberts VA, Rittenberg MB. 1992. Generation and analysis of random

cd CELLS

29.

30.


31.

32.

33.

34.

35.

36.

37.

38.

point mutations in an antibody CDR2 sequence: Many mutated antibodies lose their ability to bind antigen. J. Exp. Med. 176:855–66 Brown M, Rittenberg MB, Chen C, Roberts VA. 1996. Tolerance of single but not multiple amino acid replacements in antibody VH CDR2: a means of minimizing B cell wastage from somatic hypermutation? J. Immunol. 156:3285–91 Diamond B, Scharff MD. 1984. Somatic mutation of the T15 heavy chain gives rise to an antibody with autoantibody specificity. Proc. Natl. Acad. Sci. USA 81:5841–45 Casson LP, Manser T. 1995. Evaluation of loss and change of specificity resulting from random mutagenesis of an antibody VH region. J. Immunol. 155:5647–54 Davis MM, Bjorkman P. 1988. T cell antigen receptor genes and T cell recognition Nature 334:395–402 Rock EP, Sibbald PR, Davis MM, Chien Y-H. 1994. CDR3 length in antigenspecific immune receptors. J. Exp. Med. 179:323–28 Arevalo JH, Hassig CA, Stura EA, Sims MJ, Taussig MJ, Wilson IA. 1994. Structural analysis of antibody specificity: detailed comparison of five ‘‘Fab’’-steroid complexes. J. Mol. Biol. 241:663– 80 Spanopoulou E, Zaitseva F, Wang F-U, Santagata S, Baltimore D, Panayotou, G. 1996. The homeodomain region of RAG1 reveals the parallel mechanisms of bacterial and V(D)J recombination. Cell 87:263–76 Hochstenbach F, Brenner MB. 1989. Tcell receptor d-chain can substitute for a to form a bd heterdimer. Nature 340:562–65 Hein WR, Dudler L. 1993. Divergent evolution of T-cells repertoires: extensive diversity and developmentally regulated expression of the sheep c-d T-cell receptor. EMBO J. 12:715–24 Hein WR, Dudler L, Morris B. 1990.

39.

40.

41.

42.

43.

44.

45.

46.

1015

Differential peripheral expansion and in vivo antigen reactivity of ab and cd T cells emigrating from the early lamb thymus. Eur. J. Immunol. 20:1805–13 Hein WR, Mackay CR, 1991. Prominence of c-d T-cells in the ruminant immune system. Immunol. Today 12:30– 34 Bucy P, Chen CL, Cihak J, Losch U, Cooper M. 1988. Avian T cells expressing cd receptors localize in the splenic sinusoids and the intestinal epithelium. J. Immunol. 141:2200–5 Goodman T, Lefrancois L. 1988. Expression of the cd T cell receptor on intestinal CD8(`) intraepithelial lymphocytes. Nature 33:855–58 Deusch K, Luling F, Reich K, Classen M, Wagner H, Pfeffer K. 1991. A major fraction of human intra-epithelial lymphocytes simultaneously expresses the gamma delta T cell receptor, the CD8 accessory molecule, and preferentially uses the Vd1 gene segment. Eur. J. Immunol. 21:1053–59 Kuziel WA, Takashima A, Bonyhadi M, Bergstresser PR, Allison JP, Tigelaar RE, Tucker PW. 1987. Regulation of T cell receptor c chain expression in murine thy-1(`) dendritic epidermal cells. Nature 328:263–66 Stingl G, Koning F, Yamada H, Yokoyama WM, Tschachler E, Bluestone JA, et al. 1987. Thy-1(`) dendritic epidermal cells express T3 antigen and the T-cell receptor c chain. Proc. Natl. Acad. Sci. USA 84:4586–89 Itohara S, Farr A, Lafaille JJ, Bonneville M, Takagaki Y, Hass W, et al. 1990. Homing of a gamma delta thymocyte subset with homogeneous T-cell receptors to mucosal epithelia. Nature 343:754–57 Galeá P, Brezinschek R, Lipsky PE, Oppenheimer-Marks N. 1994. Phenotypic characterization of CD41/ab TCR` and cd TCR ` T cells with a

1016

47.


48.

49.

50.

51.

52.

53.

54.

55.

HAYDAY transendothelial migratory capacity. J. Immunol. 153:529 Roth SJ, Diacovo TG, Brenner MB, Rosat J-P, Buccola J, Morita CT, Springer TA. 1998. Transendothelial chemotaxis of human ab and cd T lymphocytes to chemokines. Eur. J. Immunol. 28:104–13 Wilson E, Aydintug MK, Jutila MA. 1999. A circulating bovine cd T cell subset which is found in large numbers in the spleen accumulates inefficiently in an artificial site of inflammation: correlation with lack of expression of E-selectin ligands, L-selectin. J. Immunol. 162: 4914–19 Augustin A, Kubo RT, Sim G-K. 1989. Resident pulmonary lymphocytes expressing the cd T cell receptor. Nature 340:239–41 Mukasa A, Lahn M, Pflum EK, Born W, O’ Brien RL. 1997. Evidence that the same cd T cells respond during infectioninduced and autoimmune inflammation. J. Immunol. 159:5787–94 Janeway CA Jr, Jones B, Hayday AC. 1988. Specificity and function of cells bearing cd T cell receptors. Immunol. Today 9:73–76 Asarnow DM, Kuziel WA, Bonyhadi M, Tigelaar RE, Tucker PW, Alison JP. 1988. Limited diversity of cd antigen receptor genes of Thy-1(`) dendritic epidermal cells. Cell 55:837–47 Lefrancois L. 1991. Phenotypic complexity of intraepithelial lymphocytes of the small intestine. J. Immunol. 147: 1746–51 Schild H, Mavaddat N, Litzenberger C, Ehrlich EW, Davis MM, Bluestone JA, Matis L, Draper RK, Chien Y-H. 1994. The nature of major histocompatibility complex recognition by cd T-cells. Cell 76:29–37 Grusby MJ, Auchincloss H Jr, Lee R, Johnson RS, Spencer JP, Zijlstra M, Jaenisch R, Papaioannou VE, Glimcher LH. 1993. Mice lacking major histocom-

56.

57.

58.

59.

60.

61.

62.

63.

64.

patibility complex class I and class II molecules. Proc. Natl. Acad. Sci. USA 90:3913–17 Sperling AI, Linsley PS, Barrett TA, Bluestone JA. 1993. CD28-mediated costimulation is necessary for the activation of T-cell receptor-c-d ` T-lymphocytes. J. Immunol. 151:6043–50 Takamizawa M, Fagnoni F, MehtaDamani A, Rivas A, Engleman EG. 1995. Cellular and molecular basis of human cd T cell activation. J. Clin. Invest. 95:296–303 Penninger JM, Timms E, Shahinian A, Jezo-Bremond A, Nishina H, Ionescu J, Hedrick SM, Mak TW. 1995. Alloreactive cd thymocytes utilize distinct costimulatory signals from peripheral T cells. J. Immunol. 155:3847–55 Fearon DT, Carter RH. 1995. The CD19/ CR2/TAPA-1 complex of lymphocytes: linking natural to acquired immunity. Annu. Rev. Immunol. 13:127–49 Lynch F, Shevach EM. 1992. Activation requirements of newborn thymic c/d T cells. J. Immunol. 149:1649 Boullier S, Poquet Y, Halary F, Bonneville M, Fournie JJ, Gougeon ML. 1998. Phosphoantigen activation induces surface translocation of intracellular CD94/ NKG2A class I receptor on CD94peripheral V gamma 9 V delta 2 T cells but not on CD94- thymic or mature gamma delta T cells clones. Eur. J. Immunol. 28:3399–3410 Bauer S, Groh V, Wu J, Steinle A, Phillips JH, Lanier L, Spies T. 1999. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 285:727–29 Hanby-Flarida MD, Trask OJ, Yang TJ, Baldwin CL. 1996. Modulation of WC1, a lineage specific cell surface molecule of c/d T cells augments cellular proliferation. Immunology 88:116–23 Brown MH, Boles PK, van der Merwe A, Kumar V, Mathew PA, Barclay AN. 1998. 2B4, the NK and T cell immuno-

cd CELLS

65.


66.

67.

68.

69.

70.

71.

72.

globulin superfamily surface protein is a ligand for CD48. J. Exp. Med. 118:2083– 90 Wijngaard PLJ, Metzelaar MJ, MacHugh ND, Morrison WI, Clevers HC. 1992. Molecular characterisation of the WC1 antigen expressed specifically on bovine CD4-CD8- cd T lymphocytes. J. Immunol. 149:3273 Jones WM, Walcheck B, Jutila MA. 1996. Generation of a new cd T cell specific monoclonal antibody (GD3.5): biochemical comparison of GD3.5 antigen with the previously described workshop cluster (WC1) antigen. J. Immunol. 156:3772 Tuo W, Bazer FW, Davis WC, Zhu D, Brown WC. 1999. Differential effects of type I IFNs on the growth of WC1CD8` cd T cells and WC1` CD8- cd T cells in vitro. J. Immunol. 162:245–53 Boismenu R, Havran W. 1994. Modulation of epithelial cell growth by intraepithelial cd T cells. Science 266:1253–55 Wen L, Roberts SJ, Viney JL, Wong FS, Mallick C, Findly RC, Peng QS, Craft JE, Owen MJ, Hayday AC. 1994. Immunoglobulin synthesis and generalized autoimmunity in mice congenitally deficient in a-b(`) T cells. Nature 369:654– 59 Ferrick D, Schrenzel M, Mulvania T, Hsieh B, Ferlin W, Lepper H. 1995. Differential production of interferon-gamma and interleukin-4 in response to Th1- and Th2-stimulating pathogens by gamma delta T cells in vivo. Nature 373:255–57 Wen L, Barber DF, Pao W, Wong FS, Owen MJ, Hayday AC. 1998. Primary cd cell clones can be defined phenotypically and functionally as Th1/Th2 cells and illustrate the association of CD4 with Th2 differentiation. J. Immunol. 160: 1965–74 Wen L, Hayday AC. 1997. cd S cell help in responses to pathogens and in the development of systemic autoimmunity. Immunol. Res. 16:229–41

1017

73. Fowell DJ, Magram J, Turck CW, Killeen N, Locksley RM. 1997. Impaired Th2 subset development in the absence of CD4. Immunity 6:559–69 74. Yin Z, Zhang D, Welte T, Bahtiyar G, Jung S, Liu L, Fu X-Y, Ray A, Craft J. 1999. Dominance of IL-12 over IL-4 in cd T cell differentiation. In press 75. Skeen MJ, Ziegler HK. 1993. Intercellular interactions and cytokine responsiveness of peritoneal a/b and c/d T Cells from Listeria-infected mice: synergistic effects of interleukin 1 and 7 on c/d T cells. J. Exp. Med. 178:985–96 76. He Y-W, Malek TR. 1996. Interleukin–7 receptor alpha is essential for the development of cd T cells but not natural killer cells. J. Exp. Med 184:289–93 77. Azura V, Levraud JP, Lembezat MP, Pereira P. 1997. A novel subset of adult cd thymocytes that secretes a distinct pattern of cytokines, expresses a very restricted T cell receptor repertoire. Eur. J. Immunol. 27:544–53 78. Carding S, Kyes S, Jenkinson E, Kingston R, Bottomly K, Owen J, Hayday AC. 1990. Developmentally regulated fetal thymic and extrathymic T-cell receptor cd gene expression Genes Dev. 4:1304–15 79. McVay LD, Jaswal SS, Kennedy C, Hayday A, Carding SR. 1998. The generation of human gamma delta T cell repertoires during fetal development. J. Immunol. 160:5851–60 80. Havran W, Allison JP. 1990. Origin of Thy-1` dendritic epidermal cells of adult mice from fetal thymic precursors. Nature 344:68–70 81. Krangel MS, Yssel H, Brocklehurst C, Spits H. 1990. A distinct wave of human T cell receptor gamma/delta lymphocytes in the early fetal thymus: evidence for controlled gene rearrangement and cytokine production. J. Exp. Med. 172:847– 59 82. McVay LM, Carding SR, Bottomly K, Hayday AC. 1991. Regulated expression

1018

83.

84.


85.

86.

87.

88.

89. 90.

91.

92.

HAYDAY and structure of T cell antigen receptor cd transcripts in human thymic ontogeny. EMBO J. 10:83–91 Parker CM, Groh V, Band H, Porcelli SA, Morita C, Fabbi M, et al. 1990. Evidence for extrathymic changes in the T cell receptor cd repertoire. J. Exp. Med. 171:1597–612 Raulet DH, Spencer DM, Hsiang YH, Goldman JP, Bix M, Liao NS, Zijlstra M, Jaenisch R, Correa I. 1991. Control of cd T cell development. Immunol. Rev. 120:185–204 Aguilar LK, Belmont JW. 1991. Vc3 T cell receptor rearrangement and expression in the adult thymus. J. Immunol. 146:1348–52 Ikuta K, Kina T, MacNeil I, Uchida N, Peault B, Chien Y-H, et al. 1990. A developmental switch in thymic lymphocyte maturation potential occurs at the level of hematopoietic stem cells. Cell 62:863–74 Passoni L, Hoffman E, Kim S, Crompton T, Pao W, Dong M-Q, Owen MJ, Hayday AC. 1997. cd T cell development dependent on intrathymic d selection. Immunity 7:83–95 Mallick-Wood C, Pao W, Cheng A, Kulkarni S, Bolen J, Rowley B, Lewis JM, Tigelaar RE, Pawson T, Hayday AC. 1996. Disruption of intraepithelial cd T cell repertoires by mutation in the Syk tyrosine kinase. Proc. Natl. Acad. Sci. USA 93:9704–9 Tough DF, Sprent J. 1998. Lifespan of c/d T cells. J. Exp. Med. 187:357–65 Hughes DPM, Hayday AC, Craft JE, Owen MJ, Crispe IN. 1995. T cells with cd T cell receptors of intestinal type are preferentially expanded in TCRa-deficient lpr mice. J. Exp. Med. 182:233–41 Kitamura D, Kudo A, Schaal S, Muller W, Melchers F, Rawjewsky K. 1992. A critical role of k5 protein in B cell development. Cell 69:823–31 Groettrup M, Ungewiss K, Azogui O, Palacios R, Owen MJ, Hayday A, von

93.

94.

95.

96.

97.

98.

99.

100.

Boehmer H. 1993. A novel disulfidelinked heterodimer on pre-T cells consists of the T cell receptor b chain and a 33-kD glycoprotein. Cell 75:283–94 Grawunder U, Leu TM, Schatz DG, Werner A, Rolink AG, Melchers F, Winkler H. 1995. Down-regulation of RAG1 and RAG2 gene expression in preB cells after functional immunoglobulin heavy chain rearrangement. Immunity 3:601–8 Hoffman E, Passoni L, Crompton T, Leu T, Schatz D, Koff A, Owen MJ, Hayday AC. 1996. Productive T cell receptor b chain gene rearrangement: close interplay of cell cycle and cell differentiation during T cell development in vivo. Genes Dev. 10:948–62 Hernandez-Munain C, Sleckman BP, Krangel MS. 1999. A developmental switch from TCRd enhancer to TCRa enhancer function during thymocyte maturation. Immunity 10:723–33 Ishida I, Berbeek S, Bonneville M, Itohara S, Berns A, Tonegawa S. 1990. Tcell receptor cd and c transgenic mice suggest a role of a c gene silencer in the generation of ab T cells. Proc. Natl. Acad. Sci. USA 87:3067–71 Kang J, Fehling HJ, Laplace C, Malissen M, Cado D, Raulet DH. 1998. T cell receptor c gene regulatory sequences prevent the function of a novel TCRc/ pTa pre T cell receptor. Immunity 8:713– 21 Buer J, Aifantis I, di Santo JP, Fehling HJ, von Boehmer H. 1997. Role of different T cell receptors in the development of pre-T cells. J. Exp. Med. 185:1541–47 Takeshita S, Toda M, Yamagishi H. 1989. Excision products of the T cell receptor gene support a progressive rearrangement model of the a/d locus. EMBO J. 8:3261–70 Dudley EC, Petrie HT, Shah LM, Owen MJ, Hayday AC. 1994. T cell receptor beta chain gene rearrangement and selection during thymocyte development in adult mice. Immunity 1:83–93


cd CELLS 101. Dudley EC, Girardi M, Owen MJ, Hayday AC. 1995. ab T cells and cd T cells can share a late common precursor. Curr. Biol. 5:659–69 102. Livak F, Petrie HT, Crispe IN, Schatz DG. 1995. In-frame d TCR gene rearrangements play a critical role in the ab/ cd T cell lineage decision. Immunity 2:617–27 103. Dent AL, Matis LA, Hooshmand F, Widacki SM, Bluestone JA, Hedrick SM. 1990. Self reactive cd T cells are eliminated in the thymus. Nature 343:714–19 104. Kang J, Coles M, Cado D, Raulet DH. 1998. The developmental fate of T cells is critically influenced by TCR cd expression. Immunity 8:427–38 105. Hoffman ES, Passoni L, Dudley EC, Girardi M, Hayday AC. 1998. The ab/cd lineage decision. In The Molecular Biology of B-Cell and T-Cell Development, ed. JG Monroe, E Rothenberg, pp. 367– 96. Totowa, NJ: Humana 106. Washburn T, Schweighoffer E, Gridley T, Chang D, Fowlkes BJ, Cado D, Robey E. 1997. Notch activity influences the ab versus cd T cell lineage decision. Cell 88:833–43 107. Barber DF, Hoffman E, Murga C, Douglas N, Voll R, Ghosh S, Hayday AC. Ms submitted 108. Ossendorp F, Jacobs H, van der Horst G, de Vries E, Berns A, Borst J. 1992. T cell receptor ab lacking the b chain V domain can be expressed at the cell surface but prohibits T cell maturation. J. Immunol. 148:3714–22 109. Irving BA, Alt FW, Killeen N. 1998. Thymocyte development in the absence of pre-T cell receptor extracellular immunoglobulin domains. Science 280:905–8 110. Monks CRF, Freiberg BA, Kupfer H, Sciaky N, Kupfer A. 1998. Three dimensional segregation of supramolecular activation clusters in T cells. Nature 395:82–86 111. Iwashima M, Green A, Bonyhadi M,

112.

113.

114.

115. 116.

117.

118.

119.

120.

1019

Davis MM, Allison JP, Chien Y. 1991. Expression of a fetal c-d T-cell receptor in adult mice triggers a non-MHC-linked form of selective depletion. Int. Immunol. 3:385–93 Asarnow DM, Cado D, Raulet DH. 1993. Selection is not required to produce invariant T cell receptor c gene junctional sequences. Nature 362:158–60 Zhang Y, Cado D, Asarnow DM, Komori T, Alt FW, Raulet DH, Allison JP. 1995. The role of short homology reports and TdT in generation of the invariant cd antigen receptor repertoire in the fetal thymus. Immunity 3:439–47 Mallick-Wood CA, Lewis J, Richie L, Owen MJ, Tigelaar RE, Hayday AC. 1998. Conservation of T cell receptor conformation in epidermal cd cells with disrupted primary Vc gene usage. Science 279:1729–33 Lewis J. 1991. PhD thesis, Univ. Texas Southwestern Burtrum DB, Livak F, Rowen L, Schatz DG, Wayne J, Petrie HT. 1999. Control of TCR-beta gene segment utilization at the DNA level. In press Lefrancois L, LeCorre R, Mayo J, Bluestone J, Goodman T. 1990. Extrathymic selection of TCR gamma delta` T cells by class II major histocompatibility complex molecules. Cell 63:333–40 Schleussner C, Ceredig R. 1993. Analysis of intraepithelial lymphocytes from major histocompatibility complex (MHC)-deficient mice: no evidence for a role of MHC class II antigens in the positive selection of Vd4(`) cd T cells. Eur. J. Immunol. 23:1615–22 Wells FB, Gahm SJ, Hedrick SM, Bluestone JA, Dent A, Matis LA. 1991. Requirement for positive selection of cd receptor-bearing T cells. Science 253: 903–5 Schweighoffer E, Fowlkes BJ. 1996. Positive selection is not required for thymic maturation of transgenic cd T cells. J. Exp. Med. 183: 2033–41


1020

HAYDAY

121. Sperling AI, Decker DC, DiPaolo RJ, Stern DA, Shum A, Bluestone JA. 1997. Selective expansion of V gamma 2-V delta 7 TCR cd cells in C57BL/6 mice is postnatal and extrathymic. J. Immunol. 159:86–91 122. Bandeira A, Itohara S, Bonneville M, Burlen-Defranoux O, Mota-Santos T, Coutinho A, Tonegawa S. 1991. Extrathymic origin of intestinal intraepithelial T cells bearing the cd T cell receptor. Proc. Natl. Acad. Sci. USA 88:43–47 123. McVay LD, Carding SR. 1996. Extrathymic origin of human gamma delta T cells during fetal development. J. Immunol. 157: 2873–82 124. Saito H, Kanamori Y, Takemori T, Nariuchi H, Kubota E, Takahashi-Iwanaga H, Iwanaga T, Ishikawa H. 1998. Generation of intestinal T cells from progenitors residing in gut crytopatches. Science 280:275–78 125. Kanamori Y, Ishimaru K, Nanno M, Maki K, Ikuta K, Nariuchi H, Ishikawa H. 1996. Identification of novel lymphoid tissues in murine intestinal mucosa where clusters of c-kit(`) IL-7R(`) Thy1(`) lympho-hemopoietic progenitors develop. J. Exp. Med. 184:1449–59 126. Dunon D, Cooper MD, Imhof BA. 1993. Thymic origin of embryonic intestinal cd T cells. J. Exp. Med. 177:257–63 127. Pfeffer K, Schoel B, Gulle H, Kaufmann SHE, Wagner H. 1990. Primary responses of human T cells to mycobacteria: a frequent set of cd T cells are stimulated by protease resistant ligands. Eur. J. Immunol. 20:1175–80 128. Tanaka Y, Sano S, Nieves E, de Libero G, Rosa D, Modlin RL, Brenner MB, Bloom BR, Morita CT. 1994. Nonpeptide ligands for human c-d T-cells. Proc. Natl. Acad. Sci. USA 91:8175–79 129. Tanaka Y, Morita C, Nieves E, Brenner M, Bloom B. 1995. Natural and synthetic non-peptide antigens recognized by human cd T cells. Nature 375:155–58 130. Constant P, Davodeau F, Peyrat MA,

131.

132.

133.

134.

135.

136.

137.

138.

139.

Poquet Y, Puzo G, Bonneville M, Fournie JJ. 1994. Stimulation of human c-d T-cells by nonpeptidic mycobacterial ligands. Science 264:267–70 Hayday A. 1999. cd cells and natural killer cells—the edge of the picture is coming into focus. Immunologist 7:16– 18 Bukowski J, Morita C, Tanaka Y, Bloom B, Brenner M, Band H. 1995. Vc2 Vd2 TCR-dependent recognition of non-peptide antigens and Daudi cells analyzed by TCR gene transfer. J. Immunol. 154:998– 1006 Bukowski JF, Morita CT, Band H, Brenner MB. 1998. Crucial role of TCRc chain junctional region in prenyl pyrophosphate antigen recognition by cd T cells. J. Immunol. 161:286–93 Bukowski JF, Morita CT, Brenner MB. 1999. Human cd T cells recognise alkylamines derived from microbes, edible plants, tea: implications for innate immunity. Immunity 11:57–65 Groh V, Bahram S, Bauer S, Herman A, Beauchamp M, Spies T. 1996. Cell stress-regulated human major histocompatibility complex class I gene expressed in gastrointestinal epithelium. Proc. Natl. Acad. Sci. USA 93:12445–50 Li P, Sirkku TW, Bauer S, Morris DL, Spies T, Strong RK. 1999. Crystal structure of the MHC class I homolog, MICA, a cd T cell ligand. Immunity 10:577– 84 Groh V, Steinle A, Bauer S, Spies T. 1998. Recognition of stress-induced MHC molecules by intestinal epithelial cd T cells. Science 279:1737–40 Steinle A, Groh V, Spies T. 1998. Diversification, expression and cd T cell recognition of evolutionarily distant members of the MIC family of major histocompatibility class I-related molecules. Proc. Natl. Acad. Sci. USA 95:12510–15 Wallace GR, Verity DH, Delamaine LJ, Ohno S, Inoko H, Ota M, Mizuki N, Yabuki K, Kondiatis E, Stephens HA,

cd CELLS

140.


141.

142.

143.

144.

145.

146.

147.

Madanat W, Kanawati CA, Stanford MR, Vaughan RW. 1999. MIC-A allele profiles and HLA class I associations in Behcet’s disease. Immunogenetics 49: 613–17 Wu J, Song Y, Bakker ABH, Bauer S, Spies T, Lanier LL, Phillips JH. 1999. An activating immunoreceptor complex formed by NKG2D and DAP10. Science 285:730–33 Chang C, Dietrich J, Harpur AG, Lindquist JA, Haude A, Loke YW, King A, Colonna M, Trowsdale J, Wilson MJ. 1999. KAP10, a novel transmembrane adaptor protein genetically linked to DAP12 but with unique signaling properties. J. Immunol. Cutting Edge 163:4651–54 Hampl J, Schild H, Litzenberger C, Baron M, Crowley M P, Chien Y-H. The specificity of a weak cd TCR interaction can be modulated by the glycosylation of the ligand. J. Immunol. 163:288–94 Clackson T, Wells JA. 1995. A hot spot of binding energy in an hormone-receptor interface. Science 267:383 Weintraub BC, Jackson MR, Hedrick SM. 1994. Gamma delta T cells can recognize nonclassical MHC in the absence of conventional antigenic peptides. J. Immunol. 153:3051–58 Bonneville M, Ito K, Krecko EG, Itohara S, Kappes D, Ishida I, Kanagawa O, Janeway CA, Murphy DB, Tonegawa S. 1989. Recognition of a self major histocompatibility complex TL region product by cd T cell receptors. Proc. Natl. Acad. Sci. USA 86:5928–32 Crowley MP, Reich Z, Mavaddat N, Altman JD, Chien Y-H. 1997. The recognition of the non-classical major histocompatibility complex (MHC) class I molecule, T10, by the cd T cell, G8. J. Exp. Med. 185:1223–30 Johnson RM, Lancki DW, Sperling AL, Dick RF, Spear PG, Fitch FW, Bluestone JA. 1992. A murine CD4-CD8-T cell receptor-cd T lymphocyte clone specific

148.

149.

150.

151.

152.

153.

154.

155.

156.

1021

for herpes simplex virus glycoprotein I. J. Immunol. 148:983–88 Sciammas R, Johnson RM, Sperling AI, Brady W, Linsley PS, Spear PG, Fitch FW, Bluestone JA. 1994. Unique antigen recognition by a herpesvirus-specific TCR-cd cell. J. Immunol. 152:5392 Bukowski JF, Morita CT, Brenner MB. 1994. Recognition and destruction of virus-infected cells by human cd CTL. J. Immunol. 153:5133–40 O’Brien RL, Happ MP, Dallas A, Palmer E, Kubo R, Born WK. 1989. Stimulation of a major subset of lymphocytes expressing T cell receptor gamma delta by an antigen derived from Mycobacterium tuberculosis. Cell 57:667 Fisch P, Malkovsky M, Kovats S, Sturm E, Braakman E, Klein BS, et al. 1990. Recognition by human Vc9/Vd2 T cells of a GroEL homolog on Daudi Burkitts lymphoma cells. Science 250:1269–73 Kim HT, Nelson EL, Clayberger C, Sanjanwala M, Sklar J, Krensky AM. 1995. cd T cell recognition of tumor immunoglobulin peptide. J. Immunol. 154: 1614– 23 Havran W, Chien Y, Allison J. 1991. Recognition of self antigens by skin derived T cells with invariant cd antigen receptors. Science 252:1430–32 Tigelaar RE, Lewis JM. 1994. In Basic Mechanisms of Physiologic and Aberrant Lymphoproliferation in the Skin, ed. WC Lambert, B Gianotti, W van Vloten, pp. 39–55. New York: Plenum Schon MP, Arya A, Murphy EA, Adams CM, Strauch UG, Agace WW, Marsal J, Donohue JP, Her H, Beier DR, Olson S, Lefrancois L, Brenner MB, Grusby MJ, Parker CM. 1999. Mucosal T lymphocyte numbers are selectively reduced in integrin alpha E (CD103) deficient mice. J. Immunol. 162:6641–49 Boismenu R, Feng L, Xia YY, Chang CC, Havran WL. 1996. Chemokine expression by intraepithelial cd T cells. J. Immunol. 157:985–92


1022

HAYDAY

157. Jones B, Mjolsness S, Janeway CA, Hayday AC. 1986. Transcription of functionally rearranged gamma genes in primary T cells of adult immuno-competent mice. Nature 323:535–38 158. Jones B, Carding S, Kyes S, Mjolsness S, Janeway C, Hayday A. 1988. Molecular analyses of TCR gamma gene expression in allo-activated splenic T cells of adult mice. Eur. J. Immunol. 18:1907–15 159. Okragly AJ, Hanby-Flarida M, Mann D, Baldwin CL. 1996. Bovine cd T cell proliferation is associated with self-derived molecules constitutively expressed in vivo on mononuclear phagocytes. Immunology 87:71–79 160. Skeen MJ, Ziegler HK. 1995. Activation of cd T cells for IFN-c production is mediated by bacteria via macrophagederived cytokines IL-1 and IL-12. J. Immunol. 154:5832–41 161. Garcia VE, Jullien D, Song M, Uyemura K, Shuai K, Morita CT, Modlin RL. 1998. IL-15 enhances the response of human cd T cells to nonpeptide microbial antigens. J. Immunol. 160:4322–29 162. Balaji KN, Schwander SK, Rich EA, Boom WH. 1995. Alveolar macrophages as accessory cells for human cd T cells activated by Mycobacterium tuberculosis. J. Immunol. 22:1767 163. Pao W, Wen L, Smith AL, GulbransonJudge A, Zheng B, Kelsoe G, MacLennan ICM, Owen MJ, Hayday AC. 1996. cd T cell help of B cells is induced by repeated parasitic infection, in the absence of other T cells. Curr. Biol. 6:1317–25 164. Wen L, Pao W, Wong FS, Peng Q, Craft J, Zheng B, Kelsoe G, Dianda L, Owen MJ, Hayday AC. 1996. Germinal center formation, immunoglobulin class switching, autoantibody production driven by ‘‘non a/b’’ T cells. J. Exp. Med. 183:2271–82 165. Fujihashi K, Mcghee JR, Kweon MN, Cooper MD, Tonegawa S, Takahashi I,

166.

167.

168.

169.

170.

171.

172.

173.

Hiroi T, Mestecky J, Kiyono H. 1996. cd T cell deficient mice have impaired mucosal immunoglobulin A responses. J. Exp. Med. 183:1929–35 Horner A, Jabara H, Ramesh N, Geha R. 1995. c/d T lymphocytes express CD40 ligands and induce isotype-switching in B lymphocytes. J. Exp. Med. 181:1239– 44 Munk ME, Fazioli RA, Calich VLG, Kaufmann SHE. 1995. Paracoccidioides brasiliensis–stimulated human c/d T cells support antibody production by B cells. Infect. Immun. 63:1608–10 Rajogopolan S, Zordan T, Tsokos GC, Datta SK. 1990. Pathogenic anti-DNA autoantibody-inducing T helper cell lines from patients with active lupus nephritis: isolation of CD4-CD8-T helper lines that express the cd T cell antigen receptor. Proc. Natl. Acad. Sci. USA 87:7020–24 Holoshitz J, Koning F, Coligan JE, Bruyn JD, Strober S. 1989. Isolation of CD41CD81 mycobacteria-reactive T lymphocyte clones from rheumatoid arthritis synovial fluid. Nature 339:226– 29 Vincent MS, Roessner K, Lynch D, Wilson D, Cooper SM, Tschopp J, Sigal LH, Budd RC. 1996. Apoptosis of Fashigh CD4` synovial cells by Borrelia-reactive Fas-ligandhigh cd T cells in lyme arthritis. J. Exp. Med 184:2109–17 Vila LM, Haftel HM, Park H-S, Lin MS, Romzek NC, Hanash SM, Holoshitz J. 1995. Expansion of mycobacteriumreactive cd T cells by a subset of memory helper T cells. Infect. Immun. 63:1211– 17 Rosat JP, Waanders GA, Beermann F, Wilson A, Owen MJ, Hayday AC, Huang S, Aguet M, MacDonald HR, Louis A. 1995. Th2 type CD4` T cell dependent expansion of cd T cells in Balb/c mice infected with Leishmania major. Infect. Immun. 63:3000–4 Elloso MM, van der Heyde HC, Troutt A, Manning DD, Weidanz WP. 1996.

cd CELLS

174.

175.


176.

177.

178.

179.

180.

181.

Human cd T cell subset-proliferative response to malarial antigen in vitro depends on CD4` T cells or cytokines that signal through components of the IL2R. J. Immunol. 157:2096–102 Janis EM, Kaufmann SHE, Schwartz RH, Pardoll DM. 1989. Activation of cd T cells in the primary immune response to mycobacterium tuberculosis. Science 244:713–15 Carding SR, Allan W, Kyes S, Hayday A, Bottomly K, Doherty PC. 1990. Late dominance of the inflammatory process in murine influenza by c-d ` T cells. J. Exp. Med. 172:1225–31 Ohga S, Yoshikai Y, Takeda Y, Hiromatsu K, Nomoto K. 1990. Sequential appearance of cd and ab bearing T cells in the peritoneal cavity during an intraperitoneal infection with Listeria monocytogenes. Eur. J. Immunol. 20:533–38 De Paoli P, Gennari D, Martelli P, Cavarzerani V, Comoretto R, Santini G. 1990. cd T cell receptor-bearing lymphocytes during Epstein-Barr virus infection. J. Infect. Dis. 161:1013–16 Ho M, Webster HK, Tongtawe P, Pattanpanyasat K, Weidanz WP. 1990. Increased cd T cells in acute Plasmodium falciparum malaria. Immunol. Lett. 140:2171–75 Raziuddin S, Telmasani AW, El-Awad ME-H, Al-Amari O, Al-Janadi M. 1992. cd T cells and the immune response in visceral Leishmaniasis. Eur. J. Immunol. 22:1143–48 De Paoli P, Basaglia G, Gennari D, Crovatto M, Modolo ML, Santini G. 1992. Phenotypic profile and functional characterisation of human cd T cells during acute toxoplasmosis. J. Clin. Microbiol. 30:729 Hara T, Mizuno Y, Takaki K, Takada H, Akeda H, Aoki T, Nagata M, Ueda K, Matsuzaski G, Yoshikai Y, Nomoto K. 1992. Predominant activation and expansion of Vc9-bearing c-d T-cells in vivo

182.

183.

184.

185.

186.

187.

188.

189.

190.

1023

as well as in vitro in Salmonella infection. J. Clin. Invest. 90:204–10 Hsieh B, Schrenzel MD, Mulvania T, Lepper HD, DiMolfetto-Landon L, Ferrick DA. 1996. In vivo cytokine production in murine listeriosis. Evidence for immunoregulation by cd T cells. J. Immunol. 156:232–37 Dećhanet J, Merville P, Berge´ F, BoneMane G, Taupin J-L, Michel P, Joly P, Bonneville M, Potaux L, Moreau J-F. 1999. Major expansion of cd T lymphocytes following cytomegalovirus infection in kidney allograft recipients. J. Infect. Dis. 179:1–8 Dećhanet J, Merville P, Retie`rre C, Lim A, Pitard V, Lafarge X, Michelson S, Me´ric C, Hallet M-M, Kourilsky P, Potaux L, Bonneville M, Moreau J-F. 1999. Implication of cd T cells in the human immune response to cytomegalovirus. J. Clin. Invest. 103:1437–49 Behr-Perst SI, Munk ME, Schaberg T, Ulrichs T, Schulz R-J, Kaufmann SHE. 1999. Phenotypically activated cd T lymphocytes in the peripheral blood of tuberculosis patients. J. Infect. Dis. 180:141– 49 Findly RC, Roberts SJ, Hayday AC. 1993. Dynamic response of murine gut intraepithelial T cells after infection by the coccidian parasite Eimeria. Eur. J. Immunol. 23:2557–64 Skeen MJ, Ziegler HK. 1993. Induction of murine peritoneal c/d T cells and their role in resistance to bacterial infection. J. Exp. Med 178:971–84 King DP, Hyde DM, Jackson KA, Novosad DM, Ellis TN, Putney L, Stovall MY, Van Winkle LS, Beaman BL, Ferrick DA. 1999. Protective response to pulmonary injury requires cd T lymphocytes. J. Immunol. Cutting Edge 162:5033–36 Bancroft GJ. 1993. The role of natural killer cells in innate resistance to infection. Curr. Opin. Immunol. 5:503–10 Mombaerts P, Arnoldi J, Russ F, Tone-

1024

191.

192.


193.

194.

195.

196.

197.

198.

199.

HAYDAY gawa S, Kaufmann SHE. 1993. Different roles of ab and cd T cells in immunity against an intracellular pathogen. Nature 365:53–56 Hiromatsu K, Yoshikai Y, Matsuzaki G, Ohga S, Muramori K, Matsumoto K, Bluestone JA, Nomoto K. 1992. A protective role of c/d T cells in primary infection with Listeria monocytogenes in mice. J. Exp. Med. 175:49–56 Fu Y-X, Roark CE, Kelly K, Drevets D, Campbell P, O’Brien R, Born W. 1994. Immune protection and control of inflammatory tissue necrosis by cd T cells. J. Immunol. 153:3101–15 Nishimura H, Emoto M, Hiromatus K, Yamamoto S, Matsura K, Gomi H, Ikeda T, Itohara S, Yoshikai Y. 1995. The role of c/d T cells in priming macrophages to produce tumor necrosis factor a. Eur. J. Immunol 25:1465–68 Ladel CH, Blum C, Kaufmann SHE. 1996. Control of natural killer cellmediated innate resistance against the intracellular pathogen Listeria monocytogenes by c/d T lymphocytes. Infect. Immun. 64:5:1744–49 Welsh RM, Lin M-Y, Lohman BL, Varga SM, Zaronainski CC, Selin LK. 1997. ab and cd T-cell networks and their roles in natural resistance to viral infections. Immunol. Rev. 159:79–93 Jones-Carson J, Vazques-Torres A, van der Heyde H, Warner T, Wagner R, Balish E. 1995. cd T cell-induced nitric oxide production enhances resistance to mucosal candidiasis. Nat. Med. 1:552–57 Ladel CH, Blum C, Dreher A, Reifenberg K, Kaufman SHE. 1995. Protective role of c/d T cells and a/b T cells in tuberculosis. Eur. J. Immunol. 25:2877 D’Souza CD, Copper AM, Frank AA, Mazzaccaro RJ, Bloom BR, Orme IM. 1997. An anti-inflammatory role for cd T cells in acquired immunity to Mycobacterium tuberculosis. J. Immunol. 158:1217–21 Sciammas R, Johnson RM, Sperling AI,

200.

201.

202.

203.

204.

205.

206.

207.

Brady W, Linsley PS, Spear PG, Fitch FW, Bluestone JA. 1994. T cell receptorc/d cells protect mice from herpes simplex virus type 1-induced lethal encephalitis. J. Exp. Med. 185:1969–75 Roberts SJ, Smith AL, West AB, Wen L, Findly RC, Owen MJ, Hayday AC. 1996. T-cell ab and cd` deficient mice display abnormal but distinct phenotypes toward a natural, widespread infection of the intestinal epithelium. Proc. Natl. Acad. Sci. USA 93:11774–79 Tsuji M, Mombaerts P, Lefrancois L, Nussenzweig RS, Zavala F, Tonegawa S. 1994. c-d T-cells contribute to immunity against the liver stages of malaria in a-b T-cell-deficient mice. Proc. Natl. Acad. Sci. USA 91:345–49 Tsuji M, Eyster CL, O’Brien R, Born WK, Bapna M, Reichel M, Nussenzweig RS, Zavala F. 1996. Phenotypic and functional properties of murine cd T cell clones derived from malaria immunized, ab T cell deficient mice. Int. Immunol. 8:359–66 Thoma-Uszynski S, Ladel CH, Kaufmann SHE. 1997. Abscess formation in Listeria monocytogenes-infected cd T cell deficient mouse mutants involves ab T cells. Microbiol. Pathogen. 22:123–28 Huber SA, Stone JE, Wagner DH Jr, Kupperman J, Pfeiffer L, David C, O’Brien RL, Davis GS, Newell MK. 1999. cd` T cells regulate major histocompatibility complex class II (IA and IE)-dependent susceptibility to coxsackievirus B3-induced autoimmune myocarditis. J. Virol. 73:5630–36 Limas CJ. 1996. Autoimmunity in dilated cardiomyopathy and the major histocompatibility complex. Int. J. Cardiol. 54:113–16 Huber SA, Kupperman J, Newell MK. 1999. Hormonal regulation of CD4` Tcell responses in coxsackievirus B3induced myocarditis in mice. J. Virol. 73:4689–95 Sanui H, Yoshida S, Himeno K, Nomoto

cd CELLS

208.

209.


210.

211.

212.

213.

214.

215.

K. 1983. Experimental allergic orchitis induced by unilateral intratesticular bacterial infection in guinea-pigs. Immunology 49:45 Mukasa A, Hioromatsu K, Matsuzaki G, O’Brien R, Born W, Nomoto K. 1995. Bacterial infection of the testis leading to autoaggressive immunity triggers apparently opposed responses of ab and cd T cells. J. Immunol. 155:2047–56 Mukasa A, Born W, O’Brien RL. 1999. Inflammation alone evokes the response of a TCR-invariant mouse cd T cell subset. J. Immunol. 162:4910–13 Halstensen TS, Scott H, Brandtzaeg P. 1989. Intraepithelial T cells of the TCR cd CD81 and Vd1/Jd1` phenotypes are increased in coeliac disease. Scan. J. Immunol. 30:665–72 Dieterich W, Ehnis T, Bauer M, Donner P, Volta U, Riecken EO, Schuppen D. 1997. Identification of tissue transglutaminase as the autoantigen of coeliac disease. Nat. Med. 3:797–801 Sollid LM, Thorsby E. 1993. HLA susceptibility genes in coeliac disease; genetic mapping and role in pathogenesis. Gastroenterology 105:910–22 Hayday AC, Findly RC, Plehn-Dujowich D, Viney JL, Owen MJ, Roberts SJ. 1994. Intraepithelial cd` T cells in natural infection and in coeliac disease: protectors of epithelial integrity and mediators of immune regulation—an hypothesis. Int. Proc. Coeliac Dis., 6th. pp 46–57, Dublin, 1992 Giacomelli R, Parzanese I, Frieri G, Passacantando S, Pizzuto F, Pimpo T, Cipriani P, Viscido A, Caprilli R, Tonietti G. 1994. Increase of circulating cd T lymphocytes in the peripheral blood of patients affected by active inflammatory bowel disease. Clin. Exp. Immunol. 98:83–88 Chen Y, Fuchs E, Havran WL, Boismenu R. 1999. The maintenance and repair of the intestinal mucosa by intraepithelial cd cells. In press

1025

216. Paus R, Hofmann U, Eichmuller S, Czarnetzki BM. 1994. Distribution and changing density of gamma-delta T cells in the murine skin during the induced hair cycle. Br. J. Dermatol. 130:281–89 217. Komano H, Fuijura Y, Kawaguchi M, Matsumoto S, Hashimoto Y, Obana S, Mobaerts P, Tonegawa S, Yamamoto H, Itohara S, Nanno M, Ishikawa H. 1995. Homeostatic regulation of intestinal epithelia by intraepithelial c-d T-cells. Proc. Natl. Acad. Sci. USA 92:6147–51 218. Heyborne KD, Cranfill RL, Carding SR, Born WK, O’Brien RL. 1992. Characterization of cd T lymphocytes at the maternal fetal interface. J. Immunol. 149: 2872–78 219. Kikuchi GE, Roberts K, Shevach EM, Coligan JE. 1992. Gene transfer demonstrates that the Vc1.1. Cc4 Vd6 Cd T cell receptor is essential for autoreactivity. J. Immunol. 148:1302–7 220. Peng SL, Madaio MP, Hughes DPM, Crispe IN, Owen MJ, Wen L, Hayday AC, Craft J. 1996. Murine lupus in the absence of ab T cells. J. Immunol. 156:4041–49 221. Peng SL, Cappadona J, McNiff JM, Madaio MP, Owen MJ, Hayday AC, Craft J. 1998. Pathogenesis of autoimmunity in ab T cell-deficient lupus-prone mice Clin. Exp. Immunol. 111:107–16 222. Peng S, Madaio M, Hayday AC, Craft J. 1996. Propagation and regulation of systemic autoimmunity by cd T cells. J. Immunol. 157:5689–98 223. Harrison LC, Dempsey-Collier M, Dramer DR, Takahashi K. 1996. Aerosol insulin induces regulatory CD8 cd T cells that prevent murine insulin-dependent diabetes. J. Exp. Med. 184:2167–74 224. Shiohara T, Moriya N, Gotoh C, Hayakawa J, Ishikawa H, Siazawa K, et al. 1990. Loss of epidermal integrity by T cell mediated attack induces long term local resistance to subsequent attack. Thymus dependency in the induction of

1026

225.

226.


227.

228.

HAYDAY resistance to graft versus host disease. J. Immunol. 145:2482–88 Shiohara T, Moriya N, Hayakawa J, Itohara S, Ishikawa H. 1996. Resistance to cutaneous graft-vs.-host disease is not induced in T cell receptor d gene-mutant mice. J. Exp. Med. 183:1483–89 Gorczynski RM, Chen Z, Hoang Y, Rossi-Bergman B. 1996. A subset of cd T-cell receptor positive cells produce Thelper type-2 cytokines and regulate mouse skin graft rejection following portal venous pretransplant preimmunization. Immunology 87:381–89 Szczepnik M, Anderson LR, Ushio H, Ptak W, Owen MJ, Hayday AC, Askenase PW. 1996. cd T cells from tolerized ab TCR deficient mice inhibit contact sensitivity effector T cells in vivo, and their interferon c production in vitro. J. Exp. Med. 184:2129–39 McMenamin C, Pimm C, McKersey M, Holt P. 1994. Regulation of lgE responses to inhaled antigen in mice by

229.

230.

231.

232.

antigen-specific gamma-delta ` T cells. Science 265:1869–73 Seymour BWP, Gershwin LJ, Coffman RL. 1998. Aerosol-induced immunoglobulin (Ig)-E unresponsiveness to ovalbumin does not require CD8(`) or T cell receptor (TCR) cd(`) T cells or Interferon (IFN) c in a murine model of allergen sensitization. J. Exp. Med. 187:721–31 Peng S, Robert ME, Hayday AC, Craft J. 1996. A tumor-suppressor function for fas (CD95) revealed in T cell deficient mice. J. Exp. Med. 184:1149–54 Penninger J, Wen T, Timms E, Potter J, Wallaca V, Matsuyama T, Ferrick D, Sydora B, Kronenberg M, Mak T. 1995. Spontaneous resistance to acute T-cell leukemias in TCRVc1.1Jc4Cc4 transgenic mice. Nature 375:241–44 Nagler-Anderson C, McNair LA, Cradock A. 1992. Self-reactive, T cell receptor-cd`, lymphocytes from the intestinal epithelium of weanling mice. J. Immunol. 149:2315–22


CONTENTS



1 19 53 83 113 143 165 185 217 245 275 309 347 367 393 423 451

495

529

561 593 621 665



767


861


709 739

813 829

927 975

Annual Review of Immunology Volume 18 2000

Annual Review of Immunology Volume 2 1984

Annual Review of Immunology Volume 27 2009

Annual Review of Immunology Volume 16 1998

Annual Review of Immunology Volume 17 1999

Annual Review of Immunology Volume 22 2004

Annual Review of Immunology Volume 24 2006

Annual Review of Immunology Volume 14 1996

Annual Review of Immunology Volume 20 2002

Annual Review of Immunology Volume 23 2005

Annual Review of Immunology Volume 19 2001

Annual Review of Immunology Volume 21 2003

Annual Review of Immunology Volume 26 2008

Annual Review of Immunology Volume 1 1983

Annual Review of Immunology Volume 4 1986

Annual Review of Immunology Volume 3 1985

Annual Review of Psychology, vol 51 2000

Annual Review of Nursing Research, Volume 18, 2000: Focus on Chronic Illness

Biotechnology Annual Review, Volume 12 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 14

Biotechnology Annual Review, Volume 9 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 10 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 2

Biotechnology Annual Review, Volume 3

Biotechnology Annual Review, Volume 1

Annual Review of Medicine Volume 59 2008

Annual Review of Nano Research, Volume 3

Annual Review of Medicine Volume 55 2004

Annual Review of Jazz Studies, Volume 14

Annual review of nano research Volume 1

Annual Review of Immunology Volume 18 2000

Annual Review of Immunology Volume 2 1984

Annual Review of Immunology Volume 27 2009

Annual Review of Immunology Volume 16 1998

Annual Review of Immunology Volume 17 1999

Annual Review of Immunology Volume 22 2004

Annual Review of Immunology Volume 24 2006

Annual Review of Immunology Volume 14 1996

Annual Review of Immunology Volume 20 2002

Annual Review of Immunology Volume 23 2005

Annual Review of Immunology Volume 19 2001

Annual Review of Immunology Volume 21 2003

Annual Review of Immunology Volume 26 2008

Annual Review of Immunology Volume 1 1983

Annual Review of Immunology Volume 4 1986

Annual Review of Immunology Volume 3 1985

Annual Review of Psychology, vol 51 2000

Annual Review of Nursing Research, Volume 18, 2000: Focus on Chronic Illness

Biotechnology Annual Review, Volume 12 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 14

Biotechnology Annual Review, Volume 9 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 10 (Biotechnology Annual Review)

Biotechnology Annual Review, Volume 2

Biotechnology Annual Review, Volume 3

Biotechnology Annual Review, Volume 1

Annual Review of Medicine Volume 59 2008

Annual Review of Nano Research, Volume 3

Annual Review of Medicine Volume 55 2004

Annual Review of Jazz Studies, Volume 14

Annual review of nano research Volume 1

Recommend Documents