ADVANCES IN
Immunology VOLUME 71
This Page Intentionally Left Blank
ADVANCES IN
Immu EDITED BY
FRANK J. DIXON Th...
10 downloads
2874 Views
23MB 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
ADVANCES IN
Immunology VOLUME 71
This Page Intentionally Left Blank
ADVANCES IN
Immu EDITED BY
FRANK J. DIXON The Scripps Research Institute La Jolla, California ASSOCIATE EDITORS
Frederick Alt K. Frank Austen Tadamitsu Kishimoto Fritz Melchers Jonathan W. Uhr
VOLUME 71
ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto
hhh
This book is printed on acid-free paper.
@
Copyright 0 1999 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923). for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-1999 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-2776/99 $30.00
Academic Press
a division of Harcourr Brace & Company 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.apnet.com Academic Press 24-28 Oval Road. London NWI 7DX, UK http://www.hbuk.co.uWap/ International Standard Book Number: 0-12-02247 1-2 PRINTED IN THE UNITED STATES OF AMERICA 98 99 0 0 0 1 02 0 3 E B 9 8 7 6 5
4
3
2
1
CONTENTS
ix
CONTRIBUTORS
apIy8 Lineage Commitment in the Thymus of Normal and Genetically Manipulated Mice
HANSJOHC I7EHI,IN(:,
SUSAN
GILFILIAN, AND RHOIIHICEREDIC:
1 16 Analysis of TCR Gene Rearran ements in c.P and y6 Lineage Cells 19 35 Analysis of TCR Transgenic an! Gene-Targeted Mice 52 Cell Culture Studies 53 Developmental Considerations 54 In Search of a Consensus Model for the &y8 Lineage Split 64 Refihrences
I. Introduction Models of crP/yG Lineage Chmmitnient and Lineage Maintenance
11. 111. I\’. V. VI. VII.
lmmunoregulatory Functions of y8 T Cells w1~1.i Ram, CAHOI. CA
I J~E,5 W A JoNEs-CAHSONAKlhO MICHAELL a m , A N D REBECYA O ’ B R I E ~
MUhASA,
I. Introduction 11. Origin, Lineage and Development, and Distribution III. Specificity IV. Functions V. Concluding Remarks References
77 78 83 93 123 124
STATs as Mediators of Cytokine-Induced Responses
TIMOTHY HOEYA N D MICIIAEI. J. G H U ~ H Y
I. Introduction 11. The STAT Gene Family
111. Structural and Functional Domains in STAT Proteins
145 145
146
vi
CONTEXTS
IV. V. VI. VII.
STAT-Deficient Mice STAT Function in Cellular Proliferation and Disease Regulation of STAT Function Suiniiiary and Perspective References
152 155 156 157 158
CD95(APO-1/Fas)-Mediated Apoptosis: Live and Let Die
PETERH. KRAMMER I. Introduction 11. Death Receptors and Ligands
111. The CD95/CD95L Systein IV. Gene Defects in the CDya5/CD95LSystem V. Role of the CD95/CD95L System ill Deletion of Peripheral T Cells VI. Role of the CD95/CD95L Systein in Liver Homeostasis VII. Signal Transduction of CD95-Mediated Apoptosis VIII. The Death Domain IX. CD95 Associated Si naling Molecules X. Other Signaling Mo ecules Invoked in CD95 signaling XI. Proteins of the Bcl-2 Family XII. The Death-Inducing Signaling Complex (DISC) XIII. Downstream Caspases in CD95 Death Receptor Signaling XIV. Type I and Type I1 Cells XV. FLIPS (FLICE Inhibitory Proteins) XVI. Sensitivity and Resistance of T Lyinphocytes toward CD95-Mediated Apoptosis XVII. The CD95 System and Chemotherapy XVIII. The CD95 Death System in AIDS XIX. Further Considerations on the Role of Apoptosis in the Clinic References
k
163 164 166 167 168 169 169 170 170 172 172 174 176 180 181 182 184 188 190 192
A CXC Chemokine SDF-l/PBSF: A Ligand for a HIV Coreceptor, CXCR4
TAKASHI NACASAWA, KALUNOBU TACHIBANA, A N D KENJIKAWABATA I. Introduction
211
11. Identification, Structure, and Expression of CXC Chemokine
SDF- 1/PBSF 111. Physiological Functions of SDF-l/PBSF
IV. A SDF-l/PBSF Receptor, CXCR4 V. HIV-1 Infection and CXCR4 VI. Perspectives Refirences
212 215 217 219 222 222
vii
CONTEKTS
T Lymphocyte Tolerance: From Thymic Deletion to Peripheral Control Mechanisms BRICITTASTOCKIN(;ER I. Introduction 11. Central Tolerance Induction in the Thymus 111. Peripheral Tolerance References
229 229 240 25 1
Confrontation between lntracellular Bacteria and the Immune System ULHICH
I. II. 111. IV. V. VI. VII. VIII. IX.
E. SCHAIBLE, HELENL. c O l , l , l N S ,
AND
STEFAN H. E. ~
Introduction What Is an Intracellular Pathogen? How to Enter the Host Cell Is How to Survive Induction of Nonspecific Immuni Phagosoine Maturation and Micro ,id Detours Antigen Processin and Presentation Pathways T-cell Subsets an Effector Mechanisms Host Genetics Influencing the Outcome of Infection Immune Intervention Strategies References
T
8
INDEX CONTENTS OF KECEIVTVOINMES
U F M A N N
267 268 269 273 283 292 307 322 326 336 379 385
This Page Intentionally Left Blank
CONTRIBUTORS
Willi Born (77),Departinent of Medicine, National Jewish Medical and Research Center, Denver, Colorado 80206; and Department of Immunology, University of Colorado Health Sciences Center, Denver, Colorado 80962 Carol Cady (Ti),Department o f Iinniunology, University of Colorado Health Sciences Center, Denver, Colorado 80262 Rhodri Ceredig (l),Centre de Recherche d’Iminunologie et Hematologie, F-67091 Strasbourg, France Helen L. Collins (267), M ~ L \Planck Institute for Infection Biology, D-10117 Berlin, Germany Hans Jorg Fehling (11, Basel Institute for Ininiunology, CH-4005 Basel, Switzerland Susan Gilfillan ( l),Basel Institute for Immunology, CH-4005 Basel, Switzerland Michael J. Grusby ( 145), Departinent of Iniinunology and Infectious Diseases, Harvard School of Public Health, and Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115 Timothy Hoey ( 145), Tularik, Inc., South San Francisco, California Jessica Jones-Carson ( T ) Department , of Imniunoloby, University of Colorado Health Sciences Center, Denver, Colorado 80262 Stefan H. E. Kaufmann (267),Max Planck Institute for Infection Biology, D-10117 Berlin, Gerinany Kenji Kawabata (211), Department of Immunology, Research Institute, Osaka Medical Center for Maternal and Child Health, Izumi, Osaka 594-1101, Japan Peter H. Krammer ( 163),Tumor Iuiniunology Prograin, German Cancer Research Center, D-69120 Heidelberg, Germany Michael Lahn (77),Department of Medicine, National Jewish Medical and Research Center, Denver, Colorado 80206 is
X
CONTRIRUTORS
Akiko Mukasa (77),Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts Takashi Nagasawa (21l),Department of Immunology, Research Institute, Osaka Medical Center for Maternal and Child Health, Izumi, Osaka 594-1101, Japan Rebecca O’Brien (77),Department of hiledicine, National Jewish Medical and Research Center, Denver, Colorado 80206; and Department of Immunology, University of Colorado Health Sciences Center, Denver, Colorado 80262 Ulrich E. Schaible (267), Max Planck Institute for Infection Biology, D-10117 Berlin, Germany Brigitta Stockinger (229), Division of Molecular Immunology, National Institute for Medical Research, London NW7 lAA, United Kingdom Kazunobu Tachibana (211), Department of Immunology, Research Institute, Osaka Medical Center for Maternal and Child Health, Izumi, Osaka 594-1101, Japan
\I>\\U( I \ I N l\!\!UhOl~X.\ LO1 i I
ap/ yS lineage Commitment in the Thymus of Normal and Genetically Manipulated Mice HANS JORG FEHUNG,’ SUSAN GILFILLAN,’ AND RHODRI CEREDlGt hsel Instihrte lor Immunology, CH-4005 h e / , Switzedond; ond t Centre de Recherche d’lmmunologie el Hemoblogie, F-6709I Slrosbourg, France
I. Introduction
The purpose of this article is to review the mechanism of @/yS T-cell lineage commitment in tlie tlwmus. As way of introduction, some basic and largely historical aspects ofT-cell development will be reviewed briefly with particular reference to differences that might exist between the generation of a0 versus y6 cells. A more detailed analysis of some of these points can be found elsewhere (Distelhorst and Dubyak, 1998; Kang and Raulet, 1997; Kisielow and von Boelimer, 1995; Robey and Fowlkes, 1998; Rodewald and Feliling, 1998). Although natural killer ( N K ) T cells are discussed with particular reference to their presence in the thymus, detailed analysis of their phenotypic development is beyond the scope of this review.
A. SOMEUNIQUEFLATUHP\ o~ T H E yS T-CELLLINLAGE 1 Identijkation of yS T Cclls
The profound immunodeficiency observed in neonatally thyinectornized mice (Miller, 1961) led to the realization that tlie thymus was a primary lyniphoid organ responsible for tlie generation of thymus-derived, or T cells (Miller, 19S9). Following these seminal observations, it was assumed that there was only one lineage of T cells. With the diwovery that T cells were clearly part of the adaptive inimune system, showing phenomena of specificity and memory, it seemed obvioiis that, like B cells, T cells should also express clonally distributed receptor molecules. However, the fact that T-cell receptor (TCR) molecules were not secreted meant that their identification was protracted and required the advent of inore sophisticated biochemical (Allison and Lanier, 198’i) and molecular biological (Hedrick et nl., 1984) approache5 in the 1980s. It must be emphasized that it was only by such bioclieniical and molecular approaches that yS receptor inolecules were first identified (Raulet, 1989). It would now appear that all jawed vertebrates (gnathostomes) generate two distinct types of T cells, which are characterized by the mutually exclusive expression of (YPor yS T-cell receptor isotypes (Rast et al , 1997). Cells expressing y6 receptors were first identified by flow cytoinetry with monoclonal antibodies to hu1
( op,nplit 0 1YYY h\ 4~id( m~ he,\ 411 ngl+ of rc pndii(tim i n u n l o r n n w r w d
2
I1ANS ]OR(:
FEHLINC: ct 01
man CD3 and ap TCR as CD3+,ap TCR- cells (Brenner et al., 1986). It was then shown that the mouse fetal thymus also contained yS cells that migrated to peripheral lymphoid organs prior to the generation of ap T cells (Fowlkes and Pardoll, 1989; Havran and Allison, 1988; Pardoll et al., 1987). 2. Obvious Functional Difierences between
ap and
yS T Cells
That neonatally thymectomized mice, which presumably contained peripheral y6 T cells, nevertheless showed immunodeficiency (Miller, 1961) raises the important issue of the biological role of 76 T cells. If one considers that the hallmark of cells of the classical adaptive immune system is (1)antigen specificity and (2) memory, then the position of 7 6 cells in such a scenario is perplexing. Recognition of antigen by yS cells is dependent neither on CD4/CD8 co-receptor expression by the T cells themselves nor on expression of classical MHC molecules on the cells being recognized (Haas et al., 1993; Schild et al., 1994). Antigen recognition by yS T cells is quite different from that of ap T cells. Indeed, crystallographic analysis of yS TCR (Li et al., 1998) appears to confirm the idea that recognition of antigen by yS TCR is more like that of antigen by immunoglobulins than that by ap TCR. This becomes particularly relevant when one considers the possible positive (Haas et al., 1993) or negative (Boismenu and Havran, 1997) selection events affecting y6 T-cell development in the thymus and the form, if any, of their selecting ligands. The second issue is that of meniory within the y6 T-cell compartment. Although there appears to be a proliferative response among y6 cells to viral (Cardinget al., 1990),bacterial (Skeen and Ziegler, 1993),and parasitic (Rosat et al., 1993) infections, whether a specific memory component is generated following this proliferative phase is less clear (Mombaerts et al., 1993). This is in striking contrast to that seen in ap T cells where elegant studies have shown selection and expansion of cells with characteristic TCR clonotypes following immunization (Brawand et al., 1998; MacDonald et al., 1993; McHeyzer-Williams and Davis, 1995). In addition, primed ap T cells show modified activation thresholds upon rechallenge (Iezzi et al., 1998), but such information for y6 cells is sparse (Carena et al., 1997). The question of the turnover and life span of yS cells in unimmunized normal and y6 TCR transgenic mice has been addressed (Tough and Sprent, 1998). Thus, most thymic emigrant yS T cells appeared to have a restricted life span as naive cells. However, some yS cells converted to a memory phenotype as judged by acquisition of the CD44'"gh,CD62L'"", HSA'"", CD45RB"ghphenotype. It will be interesting to see if naive and memory phenotype yS cells in normal mice show different activation characteristics.
aplyS IJNE?Z(:E COMMITMENT
3
3. Anatomical Considerations with Regard to Thymic T-Cell Developirient In the intervening period between tlie discovery of T cells and their corresponding receptor molecules, many studies addressed the issue of how the thymus generated T cells. In this regard, differences between ap and yS cells clearly exist. Conibined histological and [ 'Hlthyniidine labeling experiments indicated that the thymus was divided into two main anatomical regions (Metcalf, 1966). First, a predominantly outer cortex comprising 80-90% of cells and where both cell division and death occurred (McPliee et nl., 1979; Shortinan and Jackson, 1974); cortical cells were smaller in diameter and in a compact organization. Second, a predominantly inner medulla where cell division and death were rare (Egerton ct al., 1990); medullary cells were larger in size and inore widely spaced than in the cortex (Metcalf, 1966). From these combined studies, the notion was put forward that T cells were generated in the cortex and that cortical cells were the direct precursors of cells in the medulla (Shortman and Jackson, 1974). It was quickly realized that the vast majority of thyinocytes were destined to die in sitri, a fincling that seemed perplexing before it was realized that the process of apoptosis is a major feature of both B and T lymphocyte development and is linked to t1i.e requirement for receptor selection (Kisielow and von Boehmer, 1995). For the ap lineage, more recent refinements of this cortical to rnedullaiy differentiation model certainly corroborate these earlier findings. However, the situation for yS cells is less clear. By immunohistoclieinical analysis with antibodies to surface y6 TCR, the few y6 cells in the thymus are mostly found in clusters in the cortex (Farr et al., 1990). Little information is available as to their subsequent transit through the thymus, although from labeling experiments it would seem that they probably migrate to the periphery from the medulla (Kellyet nl., 1993).Importantly, as discussed by Tough and Sprent (1998),the kinetics of ap versus yS thymocyte selection may be quite different, with 76 cells being generated and exported from the thymus more rapidly than ap cells, which in tlie adult thymus require a prolonged sojourn in the rnedulla prior to emigration. 4. The Queytion
(fPositive mid Negatiue Selection
The question oftlie developmental site and movement ofy6 cells through the thymus during differentiation is not a trivial one because, for a@cells, the transit from the cortex to the medulla is associated with receptor selection events. Positive selection takes place in the cortex whereas negative selection may take place in both cortex and medulla (Anderson et nl., 1997; Kisielow and von Boehmer, 1995; Merkenschlager et ul., 1997; Punt
et al., 1997). The death of cells during intrathymic development is due either to absence of positive selection or to negative selection. Negative selection of developing y6 cells (Dent et al., 1990) seems to be generally accepted. Whether positive selection also takes place is less certain (Schweighoffer and Fowlkes, 1996; reviewed in Haas et nl., 1993; Robey and Fowlkes, 1998). The issue of positive selection is particularly pertinent with regard to those T-cell subsets expressing invariant 76 receptors. These subsets include so-called dendritic epidermal cells (DECs), which are located in the skin and mostly bear a canonical Vy3’NSl TCR, or y6 cells in the reproductive tract, which express predominantly invariant Vy4N61 TCRs. Experiments by Mallick-Wood et nl. (1998) have revealed that mice lacking the Vy3 chain due to targeted gene disruption are capable of generating almost normal numbers of DECs expressing a y6 TCR with a similar, conserved conformational determinant (idiotype)as found in wildtype mice, despite the use of another nondeleted Vy gene segment. This result provides convincing evidence for the positive selection of y6 cells, at least with regard to this particular y6 subset. At face value, these new findings seem to contradict earlier studies by Asarnow et al. (1993), who used transgenic TCRy minigenes as artificial recombination substates to demonstrate that directed gene rearrangements-even in the absence of the possibility for selection-resulted in efficient formation of the invariant Vy3 junctional sequence. However, both findings can be easily reconciled by assuming that the generation of the highly restricted TCR repertoire of dendritic epidermal cells is the result of two processes: (a) biased gene rearrangements mediated by the recombination machinery and (b) subsequent selection of cells bearing TCRs with the respective invariant determinant. 5. Sensitivity to Glucocorticoids and Cyclosporin A Administration of glucocorticoids to mice results in a dramatic depletion of 100% of cortical and about 50%of medullary thyinocytes 48 hr after drug administration (Blomberg and Andersson, 1971). From this observation, medullary cells, like peripheral “mature” T cells, are “resistant” to glucocorticoids and are therefore called “mature” thymocytes to distinguish them from their glucocortioid-sensitive cortical “immature” partners (Ceredig et al., 1982). The glucocorticoid-mediated death of thymocytes is by apoptosis, and evidence implicates the purinergic receptor P2XI and an inosito1 1,4,5-trisphosphate receptor ( IP3R) in mediating this process (Distelhorst and Dubyak, 1998). For ap cells, the transition from glucocorticoid “sensitive” to “resistant” occurs immediately post-TCR receptor selection
’ Nomenclature throughout this article is according to Garinan et 01.
(1986)
&yG
LINEAGE COM MlTM E NT
i5
(Crompton et al., 1992; Tolosa ct al., 1998). Little information is available on differences in glucocorticoid sensitivity between immature and mature y6 cells. Interestingly, the development of yS but not a0 cells is largely resistant to the administration of cyclosporin A (Robey and Fowlkes, 1998). However, cyclosporin A does have an effect on the phenotypic maturation of intrathymic yS cells (Leclercq et al., 1993).
B. ARRIVINC: AT THE DN + DP + SP MODELOF THYMOPOIESIS 1. CD4 and CD8 as Usefiil, Developnmtal Stage-Specijic Cell Sii $ace Marker,$ The expression of serologically detectable markers was soon found to provide an important parameter for following T-cell development within the thymus. In general, expression of surface markers is used to define different stages within cell lineages, usually a valuable approach (for potential pitfalls, see Section I,B,4). The first such serological marker was the Thy-1 (CD90) antigen (Reif and Allen, 19641, which was considered to define cells of the T lymphoid lineage. By several criteria, CD90 was found not to be uniformly expressed on thymocytes, with small cortical cells expressing more CD90 than their larger medullary descendants (Ceredig et al., 1982). CD90 expression is also low on the very earliest cells in the thyinus and can be practically absent on some cells with T-cell characteristics, notably among intraepithelial lymphocytes ( IEL) (Lefrancois and Goodman, 1989).This variation in CD90 antigen expression may be linked to the presence of multiple promoters within the CD90 gene (Spanopoulou et al., 1991). There are also large species variations in the expression of CD90, with peripheral T cells in rats being mostly CD90- (Hosseinzadeh and Golschneider, 1993). However, at the time, with anti-CD90 reagents no clear dichotomy of peripheral T-cell subsets was observed. The advent of serology identified a series of T lymphocyte (Lyt) alloantigens, which for the first time subdivided mouse peripheral T cells into two phenotypically and functionally distinct populations, namely ( 1) Lyt1 (CDS)+,Lyt-2 (CD8a)-, Lyt-3 (CD8P)- “helper” and (2) Lyt-1-, Lyt2+, Lyt-3’ “cytotoxic cells” (Cantor and Boyse, 1975, 1977).CD5 was later shown to be expressed by a subset of B lyniphocytes (Hardy et al., 1994). Both subsets of peripheral T cells were derived from thymic precursors expressing all three Lyt alloantigens (Kisielow et al., 1975). Application of monoclonal antibody technology to human T cells and thyinocytes resulted in the identification of two subpopulations of mature T cells expressing the antigens CD4 and CD8 in a mutually exclusive fashion (Reinherz et al., 1980; Reinherz and Schlossman, 1980). This dichotomy of T-cell phenotype was particularly attractive given that there
6
H A N S J o H G FEHLJNG rt 01
appeared to be an association between CD4/CD8 phenotype and the specificity of MHC antigen recognition (Swain, 1980). When the human thymus was analyzed, cortical cells were found to express both CD4 and CD8 and were thus called double positive (DP),whereas medullary cells, like peripheral T cells, expressed either CD4 or CD8 and were called single positive (SP) (Janossy et al., 1980). Combining the cortical to medullary anatomical model outlined earlier with the phenotypic results of CD4 and CD8 expression, the DP (cortical) to SP (medullary) transition of human thymocytes was proposed (Reinherz and Schlossman, 1980). With a monoclonal antibody (GK-1.5) to the mouse CD4 antigen (Dialynas et al., 1983), peripheral T cells and medullary thymocytes were found to be either CD4 or CD8 (SP) and cortical cells DP (Ceredig et al., 1983). In addition, these studies identified a small subpopulation of cells in the thymus that expressed neither CD4 nor CD8, so-called double negative (DN) cells. Based on the observation that 100% of cells in the day 15 mouse fetal thymus were DN and that DP cells first appeared at day 16, 2 days before SP cells at days 18 to 19, for ap T cells the DN to DP to SP model of mouse thyrnocyte development was proposed (Ceredig et al., 1983; Fowlkes and Pardoll, 1989) (see Fig. 1A). This scheme appears valid for "classical" ap TCR-expressing cells not bearing the NK-1.1 marker. The DN to DP transition of thymocytes was directly demonstrated by in vitro culture (Ceredig et nl., 1983) and in vivo transfer experiments
FIG.1 . (A) A Simplified schematic representation of adult mouse thyinocyte differentiation. Thyinocyte sul)popiilations are outlined based on their expression of CD4 and CD8 and their relative proportions indicated as a percentage. For conventional (non-NK1. l + ) aP T cells. development progresses from cells expressing neither CD4 nor CD8 (DN) (lower left) to DP cells expressing both antigens (upper right). Efficient transition from DN to DP is contingent on successful TCRP rearrangement and pre-TCR expression (see text for details). Most cells transit directly; howcver, some DN cells proceed to DP via a CD8+/CD4- intermediate; such cells have been called immature single positives (ISP). In some mouse strains, CD4' ISP can also be detected. Following aPTCR receptor selection, DP cells become either CD4 or CD8 single positives (SP). (B) Subpopulations of mouse CD4-/CD8- (DN)thymocytes. A schematic representation of DN thyinocyte subpopulations defined by their expression ofCD25 and CD44. CD44 expression varies from weakly positive (-/low) to bright ( + + ) a n d ,togetherwith CD25, helps define four subsets of DN thymocytes that have been called CD25-/CD44" DN#1, CD2St/CD44' DN#2, CD25'/CD44-""" DN#3, and CD25-/CD44-"""' DN#4. This scheme highlights the heterogeneity of the DN#1 subset. B cells can he distinguished by expression of CD19, NK. and NK T cells by their expression of N K 1 . l in appropriate mouse strains and by weak expression of CD117 (c-kit) and, finally, dendritic precursors by expression of C D l l c and MHC class 11. T precursor cells can be distinguished by bright expression of CD117. See text for further details.
9
I
I
CD8 B DN#3
I I I
+
-/low
CD44-
++
(Fowlkes et al., 1985). Importantly, these in vitro experiments, when combined with the DNA-labeling technique (Ceredig and MacDonald, 1985; Ceredig et al., 1983; Sekaly et al., 1983),indicated that the differentiation to DP cells i n uitro was a process independent of cell division. Later experiments showed that the transition from DN to DP in vim at the population level was accompanied by a burst of rapid cell division (Hoffman et al., 1996; Howe and MacDonald, 1988). However, whether all cells undergoing this transition do so by dividing has not been determined. It should be recalled that in other cell differentiation systems, e.g., gut epithelial cell development, cellular differentiation, as defined by changes in cell phenotype, may occur independently of cell division (Simon and Gordon, 1995). The transition of thymocytes from DP to SP was initially difficult to directly demonstrate in vitro, but has been subsequently confirmed by many groups. Several phenotypic changes are associated with the transition from DP to SP, including changing cell size, downregulation of CD24 and CD90, and upregulation of CD69, and is a topic that has been adequately reviewed elsewhere (Kisielow and von Boehmer, 1995).
2. Heterogeneity cf DN Tliywwcytes Phenotypic analysis of purified DN cells indicated that they were themselves heterogeneous for the expression of several markers, including CD3, TcRyG, TcRaO, CD25, and CD44 (Fowlkes and Pardoll, 1989).DN thymocytes depleted of CD3' ap and yS cells are called triple negative (TN) cells (Godfrey and Zlotnik, 1993). In the adult but not fetal thymus (Antica et d ,1993), the earliest populations of TN cells are weakly CD4 positive, becoming negative at the CD25 stage (Wu et al., 1991). It has been suggested, however, that the CD4 molecules on such cells are passively acquired, presumably froin surrounding DP cells (Michie et al., 1998). Additional refinements to the TN developmental sequence have included c-kit (CD117),the stem cell factor receptor (Godfreyet al., 1992; Matsuzah et al., 1993). Thus, in both fetal and adult thymus, the earliest (TN#1) subset is CD117'/CD2,5-/CD44', which then progresses through a CD 117+/CD25+/CD44'(TN#2) stage to CD 117-/CD25'/CD44-"" (TN#3) and finally to CD117-/CD25-/CD44-"'" (TN#4) cells (Fig. 1B). With sensitive flow cytoinetric techniques, purified TN cells do not show completely biphasic profiles with any of these markers. Indeed, expression of CD117 byTN#l cells is quite heterogeneous in the thymus of recombination activating gene knock-out (RAG KO) mice, ranging from high on a small subset to low on a population of mature N K cells, which are found in the thymus of both RAG KO and normal mice (Carlyle et nl., 1998). In fact, CD117'"" mature NK cells constitute the majority of TN#1 cells in adult RAG KO mice (R. Ceredig, unpublished data).
Several important events take place at the CD25 stage of thymocyte development. For instance, it was demonstrated that the L7Pto (D)JP rearrangement of TCR genes occurs among CD117-/CD25+ TN#3 cells and that subsequent transition to the CD2Fi- TN#4 subset is contingent upon in-frame TcRP rearrangeinents (Mallick cf al., 1993). This process has been called “TCRP selection” and is mediated by the pre-TCR (reviewed by Fehling and von Boehiner, 1997; Kodewald and Fehling, 1998). It should be recalled that D to JP rearrangements are not unique to T cells and that the nioleciilar indicator of T-cell coniinitment is the VP to (D)JP rearrangement. This is equally valid for B cells where the V,, to (D)JHrearrangement inarks B-cell commitment. In the B lyinphocyte lineage, CD25 expression is chiuacteristic of pre-BII cells (Rolink et nl., 1994), a stage following successful Ig,, rearrangements at the CD1 lT/ CD25- pre-B1 cells (Osmond et al., 1998). Based on CD11’7 and CD25 expression, pre-B 1 cells resemble TN#1 thymocytes, cells that contain little, if any, TCR VP+(D)JP rearrangcnients ( Koyis~ict al., 1997). This differing pattern of CD25 expression by developing T and B cells indicates that there is, most likely, no physiologically relevant relationship between receptor gene rearrangement events and CD25 expression. I n contrast, activation of CD25 transcription may be a completely fortuitous event due to the presence of a particular combination of transcription factors at ii given developmental stage ( Ivanov and Ceredig, 1992; Rothenberg and Ward, 1996). Although changes in CD25 expression on niaturing thymocytes itre kipparently of no functional iniportance, CD25 clearly provides a very useful developmental marker, particularly when iised in combination with CD44. Heterogeneous CD25 and CD44 expression has therefore become the most frequently used marker system to subdivide the DN thymocyte population in a developmentally meaningful way. Figure 1 B represents a scheme based on these two markers that illustrates the developmental progression of CD3-CD4-CD8- (TN) cells in the adult thymiis along the four CD25/ CD44-defined stages (TN#l-TN#4). Tlie scheme also reveals the distinct heterogeneity of DN thyniocytes in the adult mouse. The inclusion of a few additional markers leads to further refinement, allowing the attribution of most CD2XD44-defined DN subsets to ii distinct developmental stage or lymphoid cell lineage. Apart from conventional a/3 and y6 T cells, the following cell types can be identified within the DN thymocyte subpopulation. a. B Cells. The thymus contains a distinct population of B cells that can be phenotypically distinguished from peripheral blood B cells transiting the thymus (R. Ceredig, unpublished observations). To exclude thymic B
10
H A N S J o R G FEHLINC, t,t nl
cells from DN or TN preparations, the most frequently used marker has been B220. This is problematic for two reasons. First, B220 expression on thymic B cells is very heterogeneous and can overlap with the negative control. This is most dramatically seen in IL-7 transgenic mice where the number of thymic B cells increases threefold (R. Ceredig, unpublished observations). Second, B220 expression is not unique to B cells and is induced on T cells undergoing apoptosis in vivo (Renno et al., 1996). Whether CD25t DN#3 cells undergoing apoptosis also express B220 is not clear. Some thymic B cells are weakly CD25' and their expression of CD44 is also variable (see Fig. 1B). Thus, B cells, which could potentially contain DP+JP TCR rearrangements, could contaminate all subsets of DN cells. This problem could be overcome either by using CD19 to define B cells or analyzing thymuses from B-less mice, such as membrane p-KO (pmT),animals. Alternatively, fetal mice, whose thymuses contain far fewer B cells, could be used.
b. N K T Cells. The presence of a subpopulation of CD3'"', TcRap"" CD24- CD44' DN cells in the thymus was initially a very puzzling observation (Budd et al., 1987;Cerediget al., 1987; Fowlkes et al., 1987).However, they are now known to belong to a separate lineage of so-called N K T cells (reviewed in Bendelac et al., 1997; MacDonald, 1995; Vicari and Zlotnik, 1996), which transit through the thyinus during their developmental program. N K T cell precursors appear to be present in the fetal mouse thymus (Ceredig, 1988).Importantly, human (Battistini et al., 1997) and mouse (Vicari et al., 1996;Williams et al., 1997) NK T cells expressing y6 TCR have been identified. In the mouse, like their ap partners, they are probably mostly CD24'""', but it will be of interest to determine their TCR receptor repertoire. In the thymus, NK TCRaP cells express CD3 and CD117 weakly but can be identified in appropriate mouse strains using the N K 1 . l marker (Koyasu et al., 1997). NK T cells in the thymus and in the periphery can express CD4 (Arase et al., 1993; Chen et al., 1997; Hoshimoto and Paul, 1994). TCR a/3 N K T cells express surface TCR molecules encoded for by the products of a single Va14Ja281TCRa gene combined with a particular repertoire of three, predominantly V08, but also V/37 and VP2 genes. Apart from being CD3t, NK T cells share many phenotypic properties with NK cells, which are also in the thymus of normal mice. RAG KO mice have been used as a source of DN#1 cells, but as mentioned earlier, these preparations are enriched for NK cells. From gene knockout experiments (Mendiratta et al., 1997),the development of a/3 N K T cells appears to depend on TCR receptor engagement by CD1. The crystal structure of mouse C D l d l has identified a hydrophobicbinding site occupied by glycosylphosphatidylinositol that could constitute
aplyS 1.1h'EhCE COMMITMENT
11
the natural ligand of CDldl-restricted NK T cells (Joyce et d., 1998). In addition, cytokine signaling through the IL-R cominon y (7') chain for commitment and through the IL-7Ra chain for expansion (Boesteanu rt al., 1997) of N K T cells has been demonstrated. These signals appear to 199i),the gene for which is involve interactions with IL-15 (Ohteki et d., regulated by the interferon regulatory factor-1 ( IRF-1) transcription activator. Analysis of IRF-l-deficient mice shows that this factor regulates IL-15 gene expression and thereby N K T-cell development (Ohteki et nl., 1998). Because of their DN and CD3"" phenotype, complete elimination of NK T cells from "TN" preparations may well be difficult to achieve. However, because of their unique TCRa chain and restricted Vp rearrangements, ap NK T cells can be distinguished easily from cells in the major pathway of ap T-cell development (Koyasu et al., 1997).For such analyses, measurement of rearrangements involving Vp8, 7 and 2 genes should be avoided. c. Dendritic Cells. DN#1 cells also contain precursors of dendritic cells. Thymic dendritic cells are CD44'"g'', CD117'"", MHC class 11' and CDllc'. Based on transfer experiments, it was proposed that early thymocytes contain cells capable of forming B/T/NK and dendritic cells (Ardavin et ul., 1993; Wu et al., 1991). None of these experiments were done at the clonal level. In mice with mutations in both CD117 and the IL-R yt chain, the thymus contains an apparently nornial population of antigen-presenting dendritic cells, despite the absence of thyniocyte progenitors. This finding seems more compatible with the concept of separate T/dendritic cell precursors (Rodewald and Fehling, 1998). The remarkable diversity of DN thyinocytes places severe constraints on attempts to study TCR rearrangeinents by single-cell polyinerase chain reactions (PCK) in this population because the heterogeneity of cells within the DN#1 compartment is problematic for the phenotypic identification of T precursor cells. Some contaminants (e.g., N K T cells) are mature T cells with distinct TCR rearrangements, whereas others could contain Dp+Jp rearrangements, yet not be in the T-cell lineage (e.g., B cells). DN#2 cells can be clearly identified as CD117""~1", CD44', CD25' cells, whereas DN#3 have become C D l l 7 and CD44 dull. At the population level, there appears to be a distinct quantitative increase in TCR Vp+(D)JP rearrangements at this DN#2 to DN#3 transition (Koyasu et d ,1997; 1997). As mentioned earlier, qualitative changes in TCR Tourigny et d., Vp+( D)JP rearrangements take place at the DN#3 to DN#4 transition (Mallick c't al., 1993).
12
3. Position of y6 Tliyiiwcytes within the C D 4 K D 8 Decelopnientul S c h i e In the adult mouse, the vast majority of76 cells are clearly part of the DN thyniocyte population, as they Fail to express CD4 and CD8 coreceptors. Initially, it was considered that developing yS cells were exclusively DN (Raulet, 1989). Although this holds true for the vast majority of TCRy6bearing cells in the adult thymus, y6expressing cells that coexpress CD4 and/or CD8 do exist. At day 16 of mouse fetal thymus development, about 50% of y6 cells express CD8 (Fisher and Ceredig, 1991). CD8 can be expressed by y6 cells upon activation (Goodman and Lefrancois, 1988) and is also seen on a subpopulation of intraepithelial y6 cells (Guy-Grand ct d.,1991). Because y6 cells in the fetal thymus are actively cycling (Ceredig, 1990; Fisher and Ceredig, 1991), this may explain why some are CD8’. Expression of CD4 by cloned y6 cell lines (Spits et d., 1991; Wen ct d., 1998), fetal thymocytes (Fisher and Ceredig, 1991), and y6 cells in pre-Ta KO mice (Fehling et al., 1997) has been reported. Additional markers, including CD5, CD45RB, CD62L (Mel-14), and CD24 (HSA), have been used to study the development of fetal intrathymic Vy3+ cells (Leclercq ct al., 1993). These authors concluded that in analoa with a/3 cell development, y6 cell development went froin TCR“””/CD24h’g’’ to TcR‘”g”/CD24‘””and that this transition was affected by cyclosporin A. Another point of view is that HSA”p” y6 cells represent newly formed and HSA“’” activated, or memory, cells (Tough and Sprent, 1998). Unfortunately, at present, the phenotypic analysis of 76 lineage cells is not sufficiently advanced to allow the design of a generally accepted developmental scheme, like the one for a/3 lineage cells. However, a large body of experimental data provides convincing evidence that the differentiation of y6 T cells does not follow a DN to DP to SP transition. For instance, the number of y6 thymocytes is normal or even augmented in many gene knockout mouse strains in which the generation of DP and SP thymocytes is severely hampered, strongly suggesting that CD4 and CD8 expressing thyinocytes are generally not obligatoryintermediates in y6 cell development. Teleologically this makes sense, as the physiologic function of CD4 and CD8 molecules is to interact with MHC, and the analysis of MHC-deficient inice has revealed that this interaction is dispensable for norinal y6 T-cell development. Notwithstanding CD4 and CD8 expression by y6 T cells in certain situations as mentioned earlier, the obligatory steps in y6 maturation therefore seem to occur exclusively within the DN compartment. It thus seems safe to confine a search for possible intermediary stages in the yS developinental pathway to the DN population.
4. A Note of Cnrttion Regnrditig the Us.e of S u f k c c hilurker Exprossioir to ~ c j CCU i Liningcs ~ The variability in cell surface marker expression by TN thymocytes raises the issue, perhaps extreme, of whether it is justified at all in using cell surface phenotyFe to define the lineage relationship of cells. This is particularly relevant for the expression of molecules which apparently play no functional role in thymocyte dcvelopment, like CD2,5, CD44, CD90, and even CD4 aiid CD8, as thr as their expression on the most immature thymocyte precursors is coiicerned. It shoiild also be recalled that detection of antigens by flow microfluoriinetry requires thc expression of several thousand cell surface molecules. For cell surface molecules with receptor function, expression of a few hundred molecules may be sufficient to traiisduce a biological response. Consequently, cells expressing sufficient receptors to respond to the corrcsponding ligand can nevertheless appear negative by flow microfluoriinetry. A particularly cogent example is CD25, the (Y chain of the IL-2 receptor complex. Following the demonstration that CD25 was expressed by DN cells (Ceredig et d . , 1985), much effort was made to demonstrate a role for the corresponding ligand, naniely IL-2, in thymocyte tlevelopment. This was despite the fact that the IL-2 “receptor” on DN cells was of lower affinity arid differed biochemically from I Id-2receptors o i i activated peripliera1 T cells (Lowenthal et d., 1986).It was later sliow~~ that other coniponents of the trimolecular IL-2 receptor coinplcv were not coordinately expressed on DN cells (Falk ct d ,1993). In particular, some IL-2RPexpressing “early” thyinocytes could in fact be y6 T cells (Leclercq et d., 1995). Together with the development of IL-2 KO mice (Schorle et d., 1991),it became apparent that IL-2 wus not an absolute requirement for thyinocyte differentiation. Finally, expression of both CD4 and CD8 needs to be considered, particularly the question of whether expression of both antigens is a unique characteristic of (YPlineage cells. First, several experimental manipulations of RAG KO mice, in which no receptor gene rearrangement can take place [i.e.,activation of thymocytes with anti-CD3 antibody either in t h o ( Jacobs et al., 1994) or itz Gitro (Levelt ct nl., l993), by y-irradiation in uivo (Zuniga-Pfluckeret d . ,1994),or introduction of an activated lck transgene (Mombaerts et nl., 1994)],result in the generation of DP cells. Although these manipulations are generally considered to mimick one or more physiologic functions of the pre-TCH, it has not been formally established that all the resulting DP thymocytes are really genuine ab lineage cells. Second, although the percentage and a l d u t e nunilwr of DP cells are drastically
reduced in TCRP KO mice, a significant and highly variable proportion of DP cells is still present (typically, the thymus of TCRP-I- inice still contains about 20% of DP thymocytes, but because the total thymic cellularity is decreased at least 10-fold in these animals, a proportion of 20% corresponds to less than 2% of the absolute nuinber found in wild-type mice) (Mombaerts et al., 1992).That the proportion of such DP cells was further reduced from -20% of total thymic cellularity in TCRP-I- inice to less than 1%when the S locus w a s also inactivated (Le., in TCRP-I- X TCR&/- mice) could be interpreted to indicate that some DP cells may belong to the yS lineage (Mombaerts et al., 1992; Robey and Fowlkes, 19%) (but also see Section IV,B,2). In conclusion there is no formal proof that the DP phenotype per se is sufficient to indicate ap lineage commitment. Conversely, absence of CD4 and CD8 expression on mature T lineage cells alone cannot be taken iis a reliable phenotype for the y6 lineage, a s a large proportion of mature N K T cells are DN as well. These considerations become particularly iinportant in situations where utilization of the crP or y6 TCR as a reliable lineage marker is compromised, for instance in TCR transgenic or certain TCR knockout mouse strains (see Section IV). C. DEVELOPMENTAL STACEOF apiyS LINEAGE DIVERGEN(:E Three main observations indicate that aP and y6 lineages are derived from a common T-lineage-committed precursor: First, ap and y6 T cells have strilclngly similar phenotypes and patterns of gene rearrangement. In fact, apart from the different TCRs, no single marker has been found to date that unequivocally distinguishes both types of lymphocytes. Second, both ap and y6 T cells differentiate arid mature inside the thymus (with the exception of gut-associated IEL, which will not tie considered here) and, importantly, both lineages develop from phenotypically identical precursor subsets. Finally, and y6 cells can differentiate from "developmentally advanced" T N subpopulations (see later) that have lost the potential to give rise to B cells, N K cells, or thymic DC (dendritic cells), suggesting that the emergence of both cell types defines a final branch point in the development of T-lineage-restricted precursors. Adoptive transfer studies in vivo and repopulation experiments in fetal thymic organ cultures (FTOC) have provided the most direct approach to determine at which developmental stage aP and 76 lineages diverge. Intravenous or intrathyinic injection of Iiiglily purified CD25-CD44+c-kit +CD4'""thyinocytes considered to represent the most immature developmental stage within the thymus (see Section I,B,2) resulted in the generation of mature cells of both lineages (Shortman et nl., 1991; Wu et al., 19911. Later stages (CD25+CD44+or CD2StCD44-""" T N thymocytes)
a/3/ylyS I,I NEAGE C O M M I T M E N T
15
could also generate both ap and yS T cells when incubated in cell culture medium in the presence of IL-7 (Suda and Zlotnik, 1993) or when transferred into deoxyguanosine-treated FTOC in uitro (Godfrey et al., 1993). The generally accepted findlng that all TN subsets at least up to the CD2SfCD44-””“stage, can generate both types of T cells has suggested that the branch point for lineage divergence might reside within the CD25+CD44-””“pre-T-cell population. Intuitively, this assumption seems to be supported by the observation that the CD2rj+CD44-’I”’b subset is the first in which TCRP, y, and 6 rearrangenients can be detected to a significant extent (Dudley et ul., 1995; Godfrey et ul., 1994; Passoni et al., 1997; Petrie et ul., 1995; Tourigny et ul., 1997). Because these molecular events are a prerequisite for the generation of lineage-specific receptors, they might represent, at least in theory, a good starting point for lineage divergence. Productive rearrangement of isotype-specific T-cell receptor genes within a given population would indeed be indicative of a developmental branch point, if such rearrangenients were restricted to the corresponding lineage. However, this i y clearly not the case, as productive y and 6 rearrangements can be found in ap and functional p rearrangements in y6 lineage cells (see later). The onset of 7 , 6, and p rearrangenients predominantly at the CD2Fit TN stage as such is therefore no conclusive evidence for tlie prevnce of a branch point at this stage. A stronger argument in Favor of the CD25’ pre-T-cell stage as the likely point of divergence in c@/yS cell fates is provided by the finding that an important developniental event occurs at this particular stage, namely preTCR-mediated ‘‘0 selection” (see Section I,B,2),which has lineage-specific features because it is required for the normal development of (rp lineage cells, but is completely dispensable for the generation of y6-expressing cells (reviewed in Fehling and von Boehmer, 1997; von Boehrner and Fehling, 1997). The differential dependence of a0 and y 6 cells on /3 selection could therefore represent a developmental division of the CD25+ pre-T-cell population into an cup-committed subset, which depends on pre-TCR signaling for survival, and a y6committed subset that can mature and differentiate in the absence of pre-TCR expression and “ p selection.” However, analyses of TCRP rearrangements in yS lineage cell5 can be interpreted to indicate that many y6-expressing cells in normal mice have actually been subject to p selection (Burtrum et al., 1996; Dudley et id., 1994, 1995) (see later), suggesting that ap and 7 6 cells may diverge at a developmentally later stage. Although CD25-””’ TN cells, which represent the developmental stage immediately following “ p selection,” failed to give rise to significant numbers of 7 6 TCR’ thymocytes in a study involving FTOCs (Godfrey et al., 1993),an earlier report found that C D 2 S CD44-’ h $ T N p ecursors are indeed capable of generating ap and y6 cells, both
16
HANS JORG FEHLINC: rt cd
in vivo after intrathymic injection and in vitro in simple culture medium or medium complemented with cytokines (Petrie et al., 1992). Thus, cup and y6 cells may diverge just prior to the onset of CD4 and CD8 expression. Unfortunately, cell transfer and repopulation assays are vulnerable to at least one serious caveat: they are unable to establish the clonality of precursor-product relationships. Therefore, it cannot be excluded that lineage divergence occurs at a relatively early developmental stage and that precoinmitted cells follow separate pathways, which are phenotypically indistinguishable until after expression of the appropriate lineage marker, i.e., the respective TCR isotype. The reported generation of 76 T cells from CD25-CD44-""" TN precursors is therefore not incompatible with a branch point (stage of commitment) at the CD2S' pre-T-cell stage. It was shown some time ago that single precursor cells obtained from the fetal thymus could reconstitute lymphocyte-depleted FTOC and that the lo5progeny cells that were generated contained multiple TcRP rearrange1986).Importantly, Anderson, Jenkinson, and Owen ments (Williamset d., have shown that upon such in vitro trans'fer, a single sorted CD2St thymocyte can generate both ap and yS cells (personal communication). At present, available data do not allow a precise definition of the developmental stage at which crp and yS lineages separate irreversibly. In fact, instead of being a sudden event that can be ascribed to a certain developmental stage, aP/yG lineage commitment may be a more gradual process involving two or more developmentally successive subpopulations that become increasingly unable to change their developmental fate. The apparent difficulty in demarcating a specific stage in ap/y6 lineage commitment may reflect this situation. In order to solve the issue, some knowledge about the molecular mechanisms that control lineage commitment is clearly necessary. The following section attempts to critically illuminate what is known about these mechanisms at present. II. Models of cup/yij Lineage Commitment and lineage Maintenance
A. SEQUENTIAL REARRANGEMENT MODEL
In the fetal thymus, TCRy, 6, and P gene rearrangements occur significantly before TCRa rearrangements (Fowlkes and Pardoll, 1989).The two T-cell receptor isotypes are expressed on the cell surface in a corresponding fashion: first the y6 TCR at around day 14 and then the crp TCR at day 17/18 (Hedrick and Eidelman, 1993). The discovery of a defined temporal order of TCR gene rearrangements during fetal mouse development led to speculation very early that TCR gene rearrangements may influence or even determine the aPIy6 lineage decision. A common precursor cell may first attempt to generate functional y and 6 rearrangements, and successful
uPlyS I.INEAGE COMMITMENT
17
assembly ofa y6 heterodimer would suppress further TCR rearrangements, forcing the respective cell to differentiate along the y6 lineage. Only when y or 6 gene rearrangements o n both chroinosoines were nonproductive would the cell get the opportunity to commit to the ap lineage and attempt the forination of a fhctional TCRP and eventually TCRa chain. This concept seemed to provide a good rationale for the different timing of TCR rearrangements during fetal tliymopoiesis, which lias become known as the “sequential rearrangement model” or the “model of Allison and Pardoll” (Allison and Lanier, 1987; Pardoll et nl., 1987).
B. COMPETITI\’E REAHRANCEMENT M(>L)EI. Subsequent studies of T-cell development in the adult thymus, however, have shown that y , P. and pro1)ably also 6 rearrangements are initiated and completed essentially at the same developmental stage (Godfrey et nl., 1994; Petrie et nl., 1995), suggestiiig that there is no strict temporal order with respect to these three types of rearrangements in postembryonic thymopoiesis. It is therefore conceivable that y , 6, and P gene segments compete with each other for thc forination of a signaling-competent receptor: if, in uncommitted thymocytes, y and 6 genes were rearranged in a productive fashion first, the cells might commit to the y6 lineage; and if a functional TCRP gene was assembled first, the cell might follow the ap pathway. This concept, known as the “competitive rearrangement niodel,” predicts that lineage commitment occurs during the developmental period at which the vast majority of 2111 y , 6, and P genes rearrange, i.e., at the CD2StCD44-“””pre-T-cell stage. As TCRa gene rearrangements are delayed, occurring during both fetal and postfetal thymopoiesis predominantly when lineage commitment lias already taken place (Burtrurn et nl., 1996; Petrie et nl., 199s; Raulet ct nZ,, 1985; Snodgrass et nl., 1985;Wilson et al., 1996),they are considered irrelevant for the ap/y6 lineage decision, but may participate in lineage maintainance (see later). The “competitive rearrangement model” necessarily implies that ii precursor cell perceives signals from a y6 TCR differently than from a TCRP containing TCR because the differential expression of these TCRs can only be utilized for a lineage decision if the precursor is able to distinguish y6 expression from pre-TCR expression. LINEAGE MODEL C. SEPARATE A very different view of the commitment process is provided by the “separate lineage model,” which, in its purest form, states that the outcome of TCR gene rearrangeinelits is conipletely irrelevant for the lineage decision, and comniitment to the aP or y6 pathway is brought about by some other a s yet unknown mechanism (Winoto and Baltimore, 1989a). The
18
HANS JORG FEMLING et a[.
separate lineage model as such is less satisfying because it does not explain how a lineage decision is achieved, but just points out how it is not achieved. Moreover, it does not make any prediction about the developmental stage at which ap and y6 cells diverge, as “branching” of the a@and y6 lineages can occur before, during, or after the completion of TCR rearrangements (although the separate lineage model is often cited as if it implied that lineage commitment preceded rearrangements). Because it is a very general concept, the separate lineage model is supported experimentally by all results not in line with the more explicit rearrangement model. Direct proof for the concept of separate lineages would require the demonstration that one or more of the TN thyinocyte populations can be sorted into two discrete subsets (separate lineages) with indistinguishable TCR gene rearrangement status but exclusive developmental potential.
D. MAINTENANCE OF THE LINEACE DECISION Two molecular mechanisms have been proposed to explain how a lineage decision, once made, may be implemented and maintained. The first proposal states that cells committing to the ap lineage activate a putative TCRy-specific silencer which then prevents the expression of rearranged TCRy genes, and thus of a y6 TCR (Haas and Tonegawa, 1992). The existence of a cis-acting y silencer element downstream of the TCRy constant region gene has been deduced from experiments involving TCRy transgenic mice (see later). A second proposal states that in cells following the cup pathway, a programmed excision event is triggered that deletes the DS, 16, and C6 regions, thereby preventing permanently the formation of a functional 6 chain (de Villartay and Cohen, 1990; Hockett et al., 1988). This proposal is based on the identification of recombinational elements, termed “6rec,” which are located upstream of the TCRG locus and frequently recombine with a pseudo Ja element ($]a)located immediately 5’ to the J a region (de Villartay et al., 1988; Hockett et al., 1989; Toda et al., 1988) or, in the mouse, also with certain Ja elements (Janowski et al., 199’7; Shutter et al., 1995). Another consequence of such a recombination is the apposition of V a and Jagene clusters, with a potentially advantageous effect on the generation of ap lineage cells, as it might enhance the efficiency of subsequent Va+ J a rearrangements. Although both the “silencer mechanism” and the “deletional mechanism” were originally proposed by proponents of the separate lineage model, these concepts are also fully compatible with the rearrangement models when considered as possible mechanisms for the implementation and maintenance of the lineage decision.
19 111. Analysis of TCR Gene Rearrangements in (YOand 76 lineage Cells
A. G E N E HAL CON 5ID ERATIO N 5 The three models for the a&8 lineage split just described make different predictions with regard to the rearrangement status of the T-cell receptor loci in mature a@and 7 6 lymphocytes. According to the sequential rearrangement model, aP T cells 4iould bear signs of failed y and 6 rearrangements. Consequently, as well as 6 loci must be rearranged extensively in all a6 T cells. Moreover, the frequency of nonproductive rearrangements should be significantly higher than one would expect if the outcome of the rearrangement did not influence the lineage decision, as either y or 6 must be out of frame to allow a cell to undergo subsequent a rearrangement. y6 T cells, however, should be essentially devoid of VP+( D)Jp and Va- J a rearrangements, as they are expected to commit to the y6 lineage before TCRP and TCRa loci have become available for the recombination machinery. The competitive rearrangement model, however, does not exclude the appearance of VP+(D)J@ rearrangements in y6 cells as long as they are out of frame. It is also compatible with a lack of y or 6 rearrangements in some ap T cells because this could reflect a situation in which a functional TCRP rearrangement and subsequent aP lineage commitment has occurred before a y or 6 rearrangement could be initiated. However, the competitive rearrangement model does not allow the complete absence of y or 6 rearrangements in all (or nearly all) CUPT cells. Most important, the key assumption of the competitive rearrangement model that a functional rearrangement at a given locus would promote commitment to the corresponding lineage predicts a bias against in-frame rearrangements of that locu~in the opposite lineage. In other words, if functional y and 6 rearrangements can promote the development of y6 T cells, they should be underrepresented in aP T cells (as predicted by the AllisodPardoll model), and, vice-versa, if productive TCRP rearrangements can direct development into the a@lineage, in-frame P-rearrangements should be underrepresented in y6 cells. In the absence of any kind of selection for or against functional rearrangements, one-third of all genes assembled by V(D)J recombination are expected to be in-frame. Because the outcome of TCR rearrangements is postulated to play no role for the commitment process in the separate lineage model, it does not impose any restrictions on the status of TCR rearrangements in mature aP or 76 cells, at least in its most generalized form, except that there should be no bias against in-frame or 6 rearrangements in cells of the a@lineage or against functional P rearrangements in cells of the y6 lineage.
B. METHODSOF MOLECULAR ANALYSIS The realization that the outcome of TCR gene rearrangements might lineage decision has incited numerous studies to deterinfluence the a p / ~ G mine the status of TCR loci in matiire T cells of both lineages. In the beginning, these analyses were largely confined to individual T cell clones, lines, or hybridomas with the drawback that the statistical significance of the observed frequencies of productive versus nonproductive rearrangements was difficult to assess. Several recent methodological advances have provided the tools for a statistically more solid analysis of rearrangements at the population level. Three sorts of approaches are commonly used. The first and historically earliest approach is based on Southern blotting of genomic DNA, where D N A fragments hybridizing within defined regions of particular T cell receptor loci serve as probes to visualize potential gene rearrangements. This method suffers from two limitations. First, the sensitivity of Southern blotting is not sufficiently high to allow detection of infrequent rearrangements in mixed cell populations. Specific rearrangements in less than 5-10% of cells in a population under study are unlikely to be detected. Low sensitivity is even more of a problem when Southern blotting is used in a “reverse” fashion by measuring decreasing intensities of germline bands as indication of rearrangement. Of course these limitations do not apply when rearrangements in homogeneous cell populations, such as hybridomas or T cell clones, are studied. A second drawback of the Southern blotting technique is its critical dependence on a good knowledge of both the genomic position of the respective probe(s) and the relative order of the rearranging gene fragments in the genome. Insufficient sequence information and the lack of detailed restriction maps have been serious constraints in the past, especially when a comprehensive and detailed analysis of gene rearrangements was required. In the era of genome research and large-scale sequencing, it is obviously just a matter of time before this limitation is completely overcome. The availability of more refined molecular maps and the rapid accumulation of exact sequence information of TCR loci in recent years have already led to a renaissance of Southern blotting for the analysis of TCR gene rearrangements. A dstinct advantage of this method is that it allows the quantitative measurement of multiple gene loci without the possibility of introducing any undue bias, which is always a risk when using amplification-based technologies, such as polymerase chain reaction. All other current approaches to study TCR gene rearrangements rely on PCR. One technique, termed polyrnerase chain reaction-restriction fragment length polymorphism (PCR-RFLP), provides information as to
aIJIyS LINEAGE C:OMMITMENT
21
the ratio of coding to noncoding sequences within a particular subpopulation (Mallick et nl., 1993).Starting with DNA purified from large numbers ( -lo4) of sorted cells, amplification of the rearranged fragment is carried out and the labeled, restriction enzyme-trimnied PCR product is elecrophoresed down a sequencing gel alongside fragments obtained from a control sequencing reaction. Thiis, PCR-amplified products from rearrangements coding for protein will be seen as major hands every three nucleotides. The ratio of coding to noncoding sequences can then be determined by densitometry. Importantly, without additional sequencing, the PCR-RFLP approach does not provide inforination as to whether coding-sized fragments will actually code for a protein nor does it determine what fraction of cells in the starting population contain rearranged fragments. A second, inore quantitative technique determines what fraction of a particular subpopulation of cells contains rearrangements (Mertsching and Ceredig, 1996). This is done by carrying out two parallel PCR reactions on each of two DNA preparations from cells with differing degrees of rearrangements. The control PCR reaction amplifies a DNA fragment within a single exon and i5 used to equilibrate the amount of PCRamplifiable DNA in the two samples. The second PCR reaction amplifies only rearranged fragments. The products of both PCR reactions are transferred to filters and probed with specific internal oligonucleotide probes. By comparing the densitornetiy curves from both reactions, an estimate as to dlfferences in the degree of rearrangement between two subpopulations can be made. Although sensitive enough to detect rearrangements in single cells (Mertsching and Ceredig, 1996),in pooled cells, differences of less than three-fold cannot be measured reliably by this technique. Again, sequencing is required to obtain the qualitative nature of the rearrangements. To gain further insight into the TcR rearrangement status of developing thyinocytes, it would be desirable to apply PCR on the single cell level, as has been done very successfully for B cells in both mice (Loffert et nl., 1996; ten Boekel et nl., 1995, 1997) and humans (Ghia et d.,1996), including subsequent sequence analysis of specific PCR fragments (ten Boekel et d.,1997). However, compared with single cell analysis of lg gene rearrangements in phenotypically defined subsets of developing B cells (Ehlich et aZ., 1994; Osinond et al., 1998),the structure of native TcR loci poses niiinerous additional problems, which make a coinprehensive PCR analysis in single T lineage cells virtually impossible. The mouse 7 locus contains multiple gene clnsters in different orientations (Raulet, 1989).The 6 locus is contained within a (Chien et al., 1987).In particular, VS genes are scattered across the VCYlocus, and although many ap T cells
22
HANS J()HG FEHLING et a!
contain two a rearrangements (Malissen et al., 1992), clearly some have only one and contain detectable 6 rearrangements on the other allele or on extrachromosomal circles (Nakajima et al., 1995). Little is known about allelic exclusion of the mouse 6 locus. The p locus has two DJCp clusters, and sampling of Vp rearrangements to one Jp locus means that no information on the other will be obtained. Some of this p locus complexity can be bypassed by using natural mouse mutants, such as the New Zealand Black (NZB) strain, in which one DJCp cluster is absent (Kotzin et al., 1985). Alternatively, TCR mini loci can be introduced as transgenes in the mouse genome, which then serve as simplified rearrangement substrates (Asarnow et al., 1993; van Meenvijk et al., 1990; Capone et d.,1993a, 1995; Lauzurica and Krangel, 1994; Kang et al., 1995). Targeting such TCR mini loci into their corresponding gene loci, so-called “knock-in” experiments, would refine this experimental approach and may eventually allow a more universal application of the single cell PCR approach, also to T lineage cells. C. RESULTSOF TCR GENEREARRANGEMENT ANALYSES 1. TCRyLocus
Analysis of a large number of T cell clones, T-cell lines, and T hybridomas (reviewed in Raulet et al., 1991) as well as total thymocytes and peripheral T cells (Garman et al., 1986;Kranz et al., 1985)has demonstrated unequivocally that Vy-.Jy gene rearrangements are very abundant in cells of the afl lineage. The fact that TCRy rearrangements are common in ap lineage cells suggests that either common progenitors commit only after the occurrence of Vy+ Jy rearrangements or commitment to the afl lineage does not notably suppress such rearrangements. However, Southern hybridization analysis of DNA from tissues enriched for a/? T cells also revealed that ap T cells can carry TCRy loci in germline configuration (Garman et al., 1986; Kranz et al., 1985). Moreover, T-cell clones with at least one TCRy locus in germline configuration have been found frequently (Heilig and Tonegawa, 1987; Moisan et al., 1989; Reilly et al., 1986). Thus, a/3 T-cell development can occur in the absence of saturating y rearrangements. The recurrent detection of unrearranged TCRy alleles in ap T cells does not support the successive rearrangement model, which advocates exhaustive TCRy rearrangements before commitment to the ap lineage. Whether or not TCRy rearrangements are biased against productive rearrangements in ap lineage cells has been more difficult to address. Sequence analysis of y rearrangements cloned from individual a0 T-cell clones did not allow a firm conclusion. In fact, data suggested that there might be differences with regard to different TCR loci. For instance, the
ap/yG LINEAGE COhlMITMENT
23
frequency of in-frame Vy1.2-Jy2 rearrangements in crp T cells seemed to be close to one-third, as expected for unselected rearrangements (Heilig and Tonegawa, 1987; Kranz et al., 1985; Traunecker et al., 1986),whereas productive Vy2+ Jyl rearrangements were rarely found (Heilig and Tonegawa, 1987; Traunecker et al., 1986). A major point of concern in this type of experimental approach is the statistically limited number of sequences analyzed and the uncertainty whether the particular crp T-cell clones or lines studied are representative for crp lineage cells in general. This problem has been overcome in an elegant experimental approach involving the generation of mice that carried a transgenic TCRy minilocus (Kang et al., 1995). The minilocus encompassed Vy2, Vy3, Vy4, Jyl, and Cyl gene segments in their natural, germline configuration. However, the transgenic Vy genes were engineered to contain a frame-shift mutation that would stop translation prematurely upstream of the V-J junction. These mo&fications were introduced to ascertain that cellular selection could not affect the frequency of rearrangements involving a certain reading frame. The transgenic minilocus therefore provided a reliable internal control for assessing the influence of selection on the frequency of productive TCRy rearrangements in ap T cells. PCR amplification and analysis of almost 100 independent V y 2 4Jrlrearrangements derived from endogeneous TCRy loci and, at the same time, from the transgenic minilocus revealed a significant bias against productive Vy2+ J y l rearrangements in c.0 T cells (Kang et al., 1995):in the absence of selective pressure, approximately 18% of rearrangements (transgenic minilocus-derived sequences) were productive in contrast to only about 4% of the endogeneous Vy2Jyl joints (Kang et al., 1995). Assuming a rearrangement mechanism that does not Favor a particular reading frame and absence of any kmd of selection for or against productive rearrangements, one would expect 33% of VyJr joints to be productive. Efficient selection against cells that can produce a yS TCR (bearing a functionally rearranged TCRy and TCRS allele) should reduce the proportion of productive rearrangements to slightly less than 20% (for calculations, see Dudley et al., 199s). The arithmetically unexpected low frequency of productive rearrangements at both transgenic and endogeneous loci was shown to be due to an in-frame stop codon near the 3’ end of the Vy2 gene segment. When this naturally occurring stop codon was ignored, 32% of transgene rearrangements were in-frame, corresponding to the value (33%) anticipated for random rearrangements. Endogeneous in-frame rearrangements accounted for just 18% of the V y 2 4 Jyl rearrangements, again indicating a significant underrepresentation ( 18%versus 32%) (Kang et al., 1995). Suppression of functional TCRy rearrangements has also been reported in a study involving the PCR-RFLP technique (Dudley et al.,
24
HANS J8RC FEHLINC et al.
1995). Amplification of Vyl. 1+ Jy4 rearrangements from double positive thymocytes and peripheral ap T cells revealed a much lower frequency of in-frame rearrangements (19 and 1896, respectively) than one would have expected from a rearrangement process not subjected to yS counterselection. PCR-RFLP data are convincing, provided the frequency of unselected Vyl.l+Jy4 rearrangements is really 33%.This could not be tested directly because an internal control (e.g., a transgenic Vyl.llCy4 minilocus with an engineered stop codon in the Vyl.1 segment or a mouse with a constant region knockout allele) was not available when the experiments were done. Taken together, current data strongly indicate that the outcome of TCRy rearrangements influences the ap/yS lineage decision in that functional TCRy rearrangements seem to disfavor development along the ap lineage. Combined with the observation that not all TCRy loci or alleles are rearranged in ap T cells, the findings are most compatible with the competitive rearrangement model. 2. TCRSLocus Analysis of TCRS rearrangements in ap T cells has been hampered severely by the fact that TCRS gene segments are located within the TCRa locus and are excised as a circular piece of DNA upon Va+Ja rearrangement (Chien et al., 1987b; Koop et al., 1992) because ap T cells usually rearrange both TCRa aIleles (reviewed by Malissen et al., 1992). TCRS sequences are deleted on both chromosomes and are no longer found in cell lines or hybridomas of the ap lineage. However, the byproducts of a deletional Va-Ja recombination event, circular DNA molecules containing the deleted DNA, are retained to a large degree in thymocytes and newly generated peripheral T cells that have not yet undergone extensive proliferation. These circles have been purified and used to construct “circle DNA libraries,” which were then analyzed for the presence or absence of 6 rearrangements (Takeshita et al., 1989; Winoto and Baltimore, 1989a). Hybridization of a large number of circle DNA-containing h clones with probes flanking the DS2 gene segment and the JSl gene segment, which is by far the most frequently used J element in yS expressing cells, revealed that the overwhelming majority of JSl containing circles hybridized to all probes, implying that DS2 and JSl gene segments were predominantly unrearranged (Winoto and Baltimore, 1989a). This finding seemed to put an end to the “rearrangement models” of lineage commitment. However, a subsequent study using a similar technique reported a different result: the S locus of cloned circle DNA was found to be frequently rearranged involving mainly the JS2 element (Takeshita et al., 1989). This second study was later criticized (Winoto, 1991) for having used a vector
that might have selectively excluded the larger-sized JSl germline fragments and thus introduced a bias for 562 rearrangements. The latter might have been generated by secondary rearrangements occurring after the excision of the TCRG locus, as J62 rearrangements are normally rather rare. The status of TCRG rearrangements in ap lineage cells therefore remained a moot point. The problem was again addressed several years later by Livak and colleagues (1995) who observed that excised TCRG sequences were inaintained in adult mice in both total thymocytes and peripheral T cells to such an extent that they could be studied directly by Southern blot analysis. Potential artifacts associated with the purification and cloning of excised circle DNA could thus be avoided. This observation, together with the fact that the sequence of the TCRdS locus and reliable phosphoimaging methods had become available, allowed a detailed and quantitative Southe m blot analysis of the status of the TCRG locus in DNA extracted from thymocytes or lymph node T cells without further manipulation. The analysis revealed that sequences between D61, D62, and JSl were extensively deleted, whereas sequences between J6l and J62 were largely retained, indicating numerous V( D)J6 rearrangements involving predominantly the JSl gene segment (Livak et nl., 1995), in good agreement with the pattern known from chromosomal TCRG rearrangements in y6 cells (Chien et al., 1987a; Elliott et al., 1988).These results were confirmed with additional probes that directly visualized specific rearrangements. Quantification of band intensities suggested that cup T cells contain at minimum 40% of the retained 6 sequences in a V( D)J rearranged configuration. Importantly, sequence analysis of about 100 independent 6 rearrangements, obtained by PCR amplification of DNA from total thymus and peripheral T cells with VWJ61- and L765/J6l-specificprimers, revealed that only about 20% of these rearrangements were in-frame, significantly less than the 1: 3 ratio expected for random rearrangements, indicating that functional TCRS rearrangements were depleted in cells ofthe ap lineage (Livak et al., 1995). These findings were confirmed and extended in an independent study. Using the PCR-RFLP technique, Dudley et al. (1995) fourid extensive V(D)JS rearrangements in CD25' pre-T cells, CD4'8' thymocytes, and peripheral ap T cells. Densitometric analysis of the PCR-RFLP banding pattern indicated a significant depletion of productive rearrangements in cells of the ap lineage, as only approximately 24 and 19% of the 6 rearrangements detected in DP thymocytes and peripheral T cells, respectively, were in-frame. Interestingly, this depletion was also apparent in the CD25+ pre-T population, suggesting that a set of cells with productive 6 rearrangements had been removed to become 76 T cells already at this early developmental stage. To exclude the possibility that functional TCRS
26
HANS JOHG F E H L N G ct nl
rearrangements are inherently underrepresented in ap lineage cells due to some unknown bias in the recombination mechanism, the analysis was repeated with thymocytes from mutant mice that lack the TCRG constant region gene. Although cells from these mice can undergo (V(D)JG rearrangements, they cannot generate a TCRG chain and thus are not subject to TCRG-mediated selection (Itohara et al., 1993). In CG-’- thymocytes, in-frame TCRG rearrangements were not depleted, constituting about 32%, as expected for random rearrangements (Dudley et al., 1995). A partly different result was reported in a third study. To obtain some information on the status of the TCRdG locus at distinct developmental stages in a normal adult thymus, Wilson et al. (1996) examined the nature and extent of TCRa and TCRG transcripts in developing thymocytes by Northern blotting and sequence analysis of RT-PCR-amplified transcripts. During development, TCRa-specific transcripts were found for the first time at the ISP stage. These transcripts were not full-length, but sterile C a message originating from the so-called TEA (transcription early alpha) region located 3’ of CG and 5’ of the most upstream J a element. TEA transcripts are thought to indicate the opening of the TCRa locus for subsequent rearrangement and thus to reflect commitment to the ap lineage. The detection of TEA transcripts in ISPs is in line with the view that these cells represent a transitional population between DN and DP thymocytes already committed to the ap lineage. Interestingly, ISPs were shown to express full-length TCRG transcripts at high levels, implying that chromosomal V( D)JG rearrangements occur frequently in a@ lineagecommitted cells, well before they are excised by Va- J a rearrangements (Wilson et al., 1996). This result demonstrated that at least a significant fraction of the observed 6 rearrangements in the extrachromosomal circle DNA of ap lineage cells had taken place before excision and thus before lineage commitment, which was not clear from previous studies. However, the study by Wilson et al. (1996) did not confirm a bias against in-frame V(D)JG rearrangements in ap lineage cells: sequencing of 22 PCRamplified V(D)JCG transcripts from ISP-derived hybridomas and of an additional 77 V(D)JCG PCR products amplified from cDNA of freshly sorted ISP thymocytes revealed an overall frequency of productive rearrangements of 29%, which is close to the theoretical value expected for random V(D)JGjoining. A potential problem in this study, however, is the fact that the sequences were obtained by analyzing RNA and not genomic DNA as in other studies. The result could therefore be biased in favor of in-frame rearrangements due to the preferential stability of functional message. In fact, drastically lower steady-state levels of mRNA from Ig and TCR genes harboring premature stop codons are well documented (reviewed by Li and Wilkinson, 1998).
uplyG LINEAGE COMMITMENT
27
Another convincingdemonstration of V( D)J rearrangements at the TCRG locus in thymocytes of the ap lineage was provided by Nakajima and colleagues ( 1995). Southern analysis of genomic DNA from thymocyte subsets of adult and newborn mice, which had been fractionated corresponding to descrete developmental stages, revealed that essentially all C6 genes in these cells were associated with D h J G or V(D)JG rearrangements. Such 6 rearrangements were found irrespective of the developmental stage, including those ap-committed thymocytes in which Va+ Ja rearrangements were not yet completed. Direct evidence for the occurrence of V( D)JG rearrangements in c.p lineage cells before the excision of the TCRG locus was provided by two-dimensional gel electrophoresis of large DNA fragments. Approximately 20% of the rearranged TCRG genes resided on chromosomal DNA in CD3-""" DP thymocytes, which was not the case in more mature subsets. The finding that V(D)JGrearrangements occur in ap-committed thymocytes to a significant extent before 6 loci are excised by Va+ J a rearrangements was confirmed in thymocytes of TCRaP transgenic mice, in which endogeneous a rearrangements are suppressed. In these mice, approximately 50% of rearrangements involving the JGl element were retained on the chromosome, even in mature subsets (Nakajima et al., 1995). The latter finding also demonstrates adventitiously that deletion of the TCRG locus is not a prerequisite for the generation of ap lineage cells. A similar conclusion had been reached in an earlier study, which showed that cytotoxic T-cell lines from ap transgenic mice had committed to the ap lineage without deleting the TCRG locus (Ohashi et al., 1990). Retention of a rearranged TCRG gene on one chromosome in the presence of an a rearrangement on the second has also been found in some T hybridomas generated from thymocytes of nontransgenic mice (Thompson et al., 1990, 1991). Although the presence of chromosomal or extrachromosomal V( D)JG rearrangements in cells of the ap lineage per se is compatible with the rearrangement models as well as the separate lineage model, the findings clearly show that there is no lineage-specific control of rearrangement at the TCRG locus. Two moleculdr mechanisms for lineage determination can therefore be excluded, namely recombinational silencing of the 6 locus in ap cells (Diaz et al., 1994; Winoto and Baltimore, 1989a) and obligatory excision of the 6 locus prior to Va+Ja rearrangement by site-specific recombination involving Grec and +JS recoinbinational elements (de Villartay and Cohen, 1990; Hockett et al., 1988). Experiments indicating a selective depletion of in-frame 6 rearrangements in ap thymocytes (Dudley et al., 1995; Livak et al., 1995) seem to provide strong support for a rearrangement model of lineage commitment.
28
HANS JORG F E H I J N G ct nl.
3. TCRPLOCUS The sequential rearrangement model predicts the absence of VP-( D)JP rearrangements in 76 lineage cells. In contrast, such rearrangements should be frequent, but generally out of frame, according to the competitive rearrangement model. Initial studies involving a limited number of 78 Tcell clones and hybridomas suggested that DO+ JP rearrangements are very common, whereas complete VP+( D)JP rearrangements are rather rare in cells of the y 8 lineage (see reviews by Raulet et al., 1991; Haas and Tonegawa, 1992). More recently, it has become clear that these earlier findings cannot be generalized. With the goal of studying TCRP rearrangements in a representative sample of 78-expressing cells, Dudley and colleagues (1994) analyzed 76 T-cell populations that were sorted from spleen and lymph nodes of TCRadeficient mice. In TCR& mice, y6e.upressing lymphocytes are the predominant population of CD3+ cells, which facilitates their purification and lowers the risk of undue contamination with T cells of the aP lineage. Quantitative Southern blotting with appropriate probes revealed that at least 90% of TCRP loci in sorted y6 cells had DP+JP rearrangements and, most important, at least 50% had VP+(D)JP rearrangements. Interestingly, analysis of the pattern of V( D)JP rearrangements in such y8 T cells with the PCR-RFLP technique and primers specific for VP134P2.2 and VP4iJP2.2 indicated that approximately 68 and 73%, respectively,were in-frame. This result was supported by sequence analysis of 20 Vp13+(D)JP2.2 joints derived from sorted splenic y6 T cells, revealing 70% productive rearrangements (Dudley et al., 1994). Notably, this value is as high as in aP T cells, which are known to be subject to P selection (see Section IV,C,l). The detection of functional VP+( D)JP rearrangements and their apparent overrepresentation in y6 T cells relative to the value expected for random rearrangements has serious implications for all models of aPly8 lineage commitment. It was therefore important to show that this finding was not limited to TCRa-deficient mice. In a subsequent study (Dudley et al., 1995), the analysis was extended to y6 cells sorted from spleen and lymph nodes of normal mice. The PCR-RFLP assay with VP13/JP2.2specific primers again revealed a frequency of approximately 70% in-frame rearrangements. VP-JP rearrangements were also detected in y6 IEL of the gut, which are thought to be partly of thymic origin, using VP4/JP2.5and VfllGIJP2.2-specific primers, and again most of these rearrangements were in-frame. A different pattern was observed in TCRP rearrangements from dendritic epidermal T cells (DECs), which are derived from fetal thymic
afllyS LINEAGE COMMITMENT
29
precursors and regarded as the earliest subset of the y6 lineage (Heyborne et al., 1993; Ikuta et al., 1990). Although rearrangements involving VP4/ JP2.2 and VP5lJP2.6 elements could be clearly detected, the PCR-RFLP patterns were irregular, suggesting fewer or less diverse rearrangements (Dudley et al., 1995). Moreover, in-frame joints were in the minority. The findings regarding DECs were complemented by an independent study focusing on VP6lJP2.5 and VPWJP2.5 rearrangements in day 15 fetal y6 thymocytes (Mertsching and Ceredig, 1996). Semiquantitative PCR analysis revealed a similar degree of TCRP rearrangements in fetal day 15 yS cells as in adult CD8+ ISPs or CD25-CD44-""" DN thymocytes. However, in fetal thymocytes, such rearrangements were mostly (58%)out of frame, as determined by sequencing of 104 PCR-amplified V( D)JP joints. In this context, it is interesting to note that fetal thymocytes provide the precursor cells for dendritic epidermal cells (Allison, 1993; Boismenu and Havran, 1995). The finding of approximately 70% in-frame VP+( D)JP rearrangements in splenic and lymph node-derived 76 T cells (Dudley et al., 1995) suggested that these cells had been selected for functional TCRP chains, possibly in a pre-TCR-mediated process during thymopoiesis. However, because peripheral y6 cells are many steps past primary differentiation events and are likely to have been exposed to a variety of selective events, such as homing or clonal expansion, it is unclear whether these cells reflect events in early differentiation correctly. To avoid this potential pitfall, Burtrum and colleagues (1996) studied TCRP rearrangements in purified 78-expressing cells from the thymus of normal adult mice. Quantitative Southern blot analysis revealed that both DJP clusters were extensively rearranged at levels essentially indistinguishable from those in mature (YP T cells. VP+( D)JP rearrangements were less frequent, but nonetheless substantial, occurring in approximately 15-20% of all alleles, versus approximately 75% of mature (YPT cells. Analysis of the reading frame of complete V( D)JP rearrangements by PCR-RFLP with VPUJP2.6- and VPYJP2.6specific primers revealed that approximately 55 and $5l%, respectively, were productive (Burtrum et al., 1996). Sequencing of 16 independent VP4lJP2.6 joints derived by PCR amplification from genomic DNA of 7 6 thymocytes confirmed the PCR-RFLP data, as 9 (56%) of the rearrangements were in-frame. The results therefore suggest that functional TCRP rearrangements are already overrepresented at an early (thymic) stage in y6 T-cell development, although the predominance of in-frame TCRP rearrangements seems to be less pronounced than in peripheral 76 T cells (50-55% versus 70%). An even lower percentage of productive TCRP rearrangements in 76 thymocytes was found in an independent study. Mertsching and colleagues
30
HANb JORG FEHLING et a/
(1997) analyzed PCR-amplified VP6-(D)JP2.5 rearrangements from CD24’90’ y6 thymocytes, which represent the most numerous subpopulation of thymic y6 cells and supposedly the only subset emigrating from the thymus. Sequencing of 43 distinct VP&(D)JP2.5 joints showed that only 42% were in-frame, which was the same frequency as had been found previously in fetal y6 cells (Mertsching and Ceredig, 1996). Although this value is still 9% higher than the 33% that would be predicted by random joining, it no longer supports the view that ‘‘P selection” is a very common event in y6 T-cell development, as suggested by the previous reports. The reason for the discrepancy in the percentage of productive V( D)JP joints in this study (42%) and the one of Burtrum et al. (50-55%) is unclear at present, but may be related to the analysis of different VP+JP elements or the fact that one analysis (Mertschinget al., 1997) focused on a subpopulation of y6 thymocytes whereas the other (Burtrum et al., 1996) dealt with unfractionated y6 thymocytes. The examination of TCRP rearrangements in 76-expressing cells has provided a number of important findings with relevance for the mechanism of the crPly6 lineage split. The fact that VP+(D)JP rearrangements are common in y6 lineage cells effectively excludes the simple successive rearrangement model, which postulates that TCRP rearrangements are initiated only in those cells that fail to generate a functional yS TCR. A modified version of the successive rearrangement model may still be valid, if it incorporates two assumptions. First, expression of a y6 TCR relegates precursor cells irreversibly into the y6 lineage, but this does not prevent subsequent TCRP rearrangements. Second, P rearrangements that occur after the formation of a functional y6 TCR come too late to influence the lineage decision. The second assumption implies that either the lineagedetermining signal provided by the yS TCR cannot be overridden by signals from a pre-TCR that forms subsequently or, alternatively, a signalingcompetent pre-TCR can no longer form in cells that received a signal by a y6 TCR, e.g., because pTa expression has been switched off concomitantly. Data on TCRP rearrangements in 78-expressing cells also pose a serious threat to the simple version of the competitive rearrangement model. As outlined in detail earlier, this model presumes a competition between y and 6 rearrangements on the one hand and P rearrangements on the other, with the lineage fate being determined by the type of receptor (a y6 TCR or a TCRP-containing pre-TCR) that is generated first. The fact that productive TCRP rearrangements are found frequently in y6 cells excludes the possibility that thymic y6 cells might be derived from a salvage pathway for thymocytes that have failed to produce a TCRP chain. Moreover, it strongly suggests that the formation of a pre-TCR per se does not preclude differentiation along the y6 lineage. This incongruency with the competi-
aO/y6 LINEAGE COMMITMENT
31
tive rearrangement model can be circumnavigated again by evoking tlie same two assumptions made earlier to rescue tlie successive rearrangement model. However, with these two “amendments,” both models become virtually identical. An alternative explanation for the occurrence of productive V( D)JP rearrangements in cells of the y6 lineage is suggested by studies of early B-cell development. The ability of IgH proteins to pair with surrogate light chain plays a critical role in controlling membrane expression of the pre-BCK and hence subsequent differentiation (ten Boekel et nl., 1997, 1998).Keyna and colleagues (1995)identified two p heavy chains expressed in prwursor B cells that were unable to form a pre-BCK, and an analysis of the earliest B-cell precursors in bone marrow revealed that about half of tlie productive V( D)JH rearrangements in these cells encoded p heavy chains unable to pair with surrogate light chains (ten Boekel et al., 1997). In analogy, it remains possible that the in-frame V( D)JP rearrangements observed in y6 cells encode TCRP chains that are unable to associate with pTa, an hypothesis that has not yet been tested experimentally. Failure to form a pre-TCK would provide an explanation why some TCKPexpressing cells are not drawn into the aP differentiation pathway. However, the apparent overrepresentation of productive TCKP rearrangements in y6 cells seems to argue against sucli an interpretation, unless this overrepresentation is brought about by a pre-TCK-independent mechanism (see later). An overrepresentation of in-frame TCKP rearrangements in cells of the y6 lineage has been found repeatedly (Burtnim et al., 1996; Dudley et nl., 1994, 1995), although the extent of this overrepresentation remains a 1997).The high matter of debate (Burtrum et nl., 1996; Mertsching et d., frequency of in-frame V( D)JP rearrangements in y6 cells has come as a surprise because the biologic pressure for this selection is difficult to envisage. From gene knockout experiments it is clear that a functional TCKP chain is not required for normal y6 development (see later). Although the possibility that certain in-frame TCRP gene rearrangements confer a survival advantage on y6 T cells has not yet been excluded, another explanation is currently more popular. It is assumed that tlie accumulation of productive TCKP rearrangements in 76 cells is most likely due to a coincidental expansion of thymic yS precursors by tlie same mechanism that is responsible for the expansion of aP precursors after in-frame TCRPrearrangement. In other words, a varying but substantial number of 76 cells is thought to have been subjected to pre-TCK-mediated P selection (Burtruin et nl., 1996; Mertsching et d., 1997).If correct, this interpretation has two important implications. First, expression of a pre-TCR alone cannot be tlie decisive event for a commitment to the ap lineage, and second,
32
HANS JORC FEIlLlNC et a[
the split between the aP and 76 lineages must occur after the stage at which /3 selection is operating, i.e., at the CD25-44-""" DN stage. Alternatively, y6 cells may be able to undergo P selection well after lineage commitment has taken place. Although pre-TCR-mediated selection certainly provides an attractive and popular explanation for the buildup of in-frame rearrangements in yG cells, alternative scenarios should not yet be neglected. A functional TCRP chain may, for instance, be able to act in a pre-TCR-independent fashion. Experiments have shown that transgenic TCRP chains can signal in the absence of pTa and other T-cell receptor chains, mediating a number of effects, e.g., the generation of small numbers of DP thymocytes in RAG-'- X pTa? mice (Krotkova et al., 1997).Whether the overrepresentation of productive TCRP rearrangements in cells of the yG lineage is indeed due to pre-TCR-mediated selection or some other mechanism can be tested by analyzing rearrangements in y6 cells of pTa-deficient mice (Fehling et al., 1995a), which lack a functional pre-TCR, but possess normal TCRP alleles.
4 . TCRaLocus The presence of Va+ Jarearrangements is widely accepted as a hallmark of cells committed to the aP lineage. This view is based on the following observations. First, Va+ J a rearrangements occur late in thymopoiesis relative to TCRy, 8, and P rearrangements. For instance, significant transcriptional activity of the TCRa locus, which is thought to indicate the opening of the TCRa locus for subsequent rearrangements, can be detected by Northern blotting only at the ISP stage (Wilson et al., 1996). Moreover, the TCRa-specific mRNA in ISP thymocytes consists predominantly of sterile C a transcripts, indicating that the majority of TCRa loci are still in germline eonfiguration at this developmental stage (Wilson et al., 1996). Northern blotting analysis suggests that significant amounts of full-length (VJCa) transcripts are generated only at the DP stage (Wilson et al., 1996). These findings are in good agreement with an earlier study based on quantitative Southern blotting, which demonstrated that substantial TCRa rearrangements do not occur before the CD3-"" DP stage (Petrie et al., 1995).Because ISP and DP thymocytes are generally regarded as cells that belong to the a0 lineage, Va+Ja rearrangements seem to occur predominantly after lineage commitment has taken place and are therefore expected only in aP lineage cells. Second, the particular chromosomal organization of the TCRdG genes results inevitably in the deletion of the TCRG locus upon Va+Ja recombination (Chien et al., 198%; Koop et al., 1992).Continued expression of a yGTCR, and hence the functionality of yG T cells, therefore depends critically on efficient prevention of Va+ J a
aplyG LINEACE COhlMlTMENT
33
rearrangements. Third, transfection studies in ap and y6 T-cell lines have shown that the TCRa enhancer, located 3’ of the C a gene, contains a silencer sequence that prevents enhancer-mediated transcription in y6 but not in ap lineage cells (Winoto and Baltimore, 198913).These findings are supported by experiments with transgenic mice showing that y6 thymocytes selectively fail to express rearranged TCRa transgenes driven by an autologous locus control region (LCR), including the a enhancer and silencers (Diaz et nl., 1994). Other experiments with transgenic mice have demonstrated that the TCRa enhancer also restricts TCR rearrangements to cells of the ap lineage, indicating that the TCRa locus is most likely not only transcriptionally, but also recombinationally silent in y6 lineage cells (Cupone et d . , 1993; Lauz~iricaand Krangel, 1994). Taken together, these considerations suggest that cells of the y6 lineage do not rearrange the TCRa lociis and usually retain both alleles of the 6 locus. This is exactly what has been found in most of the y6 cell lines that have been analyzed (reviewed in Raulet ~t nl., 1991). These findings are supported by an analysis of total y6 thymocytes: quantitative Southern blotting with Ja-specific probes hybridizing to selected intronic regions spaced equilstantly across the Ja gene cluster did not reveal any TCRa rearrangements in the DNA of pools of sorted y6 thymocytes (Burtruni et nl., 1996). Although this type of analysis is not sensitive enough to exclude a low level of Va-Ja rearrangements, it demonstrates that such rearrangements are certainly not common in thymic yS lineage cells. In contrast, the vast majority of ab-expressing cells have rearranged both TCRa loci (Casanova et al., 1991; Malissen et d., 1992). The paucity of Va+ Ja rearrangements in y Sexpressing lymphocytes can in fact provide a useful molecular marker for the identification of y6 lineage cells in situations where a determination of the relevant TCR isotype is either not possible or not informative, e.g., in mice expressing TCR transgenes (Bruno et al., 1996; see later). However, not all y6 cells seem to be completely devoid of Va-Ja rearrangements. Using a semiquantitative PCR assay and primers specific for a single J a element and three distinct V a families, Mertsching et nl. ( 1997) clearly detected Va- J a rearrangements in sorted 76-expressing thyinocytes of adult but not fetal mice. That y6 cells from T C R P P mice contained a similar level of rearrangements suggested that a rearrangement in y6 cells could be generated without expression of a conventional preTCR composed of pTa and TCRP proteins (Mertsching et nl., 1997). Quantitation of the respective PCR bands and comparison with the corresponding bands obtained from SP thymocytes of the c.up lineage suggested that approximately 7% of thymic y6 cells harbored a Va-Ja rearrangement. Further analysis of these y6 thymocytes by RT-PCR revealed a
34
MANS JORC FEIILING el nl
roughly equivalent amount of VJCa transcripts, indicating that all the rearranged TCRa genes were most likely expressed. These data do not necessarily contradict previous findings. Although the percentage of Va+Ja rearrangements reported (7%) is clearly too high to be accounted for by contaminating ap lineage cells, the percentage is sufficiently low to escape detection in Southern blotting analyses or when studying TCRa rearrangements in a limited number of 76 T cell clones or hybridomas. However, why should y6 cells allow TCRa rearrangements when these rearrangements threaten their very existence? The answer might be that such rearrangements occur before lineage commitment is established. In fact, data from the analysis of pTa-’- X TCRa-I- and pTa? X T C R P doubly deficient mice provide clear, albeit indirect, evidence that Va+Ja rearrangements do occur in a few CD25+CD44-“”” pre-T cells (Buer et al., 1997; Mertsching et al., 1997). Some of these cells may eventually enter the y6 pathway. Whatever the reason for low-level TCRa rearrangements in y6 thymocytes may be, the vast majority of y6 lineage cells (>go%) is clearly devoid of any Va+Ja rearrangements and the absence of these rearrangements in mature T cells can be seen as a distinctive feature of the y6 lineage.
5. Concluding Evaluation of T C R Rearrangement Data with Regard to the DifSerent Mode1.s of Lineage Commitment The analysis of TCR gene rearrangements has proven that Vy+Jy, V&( D)J6and VP+( D)JP rearrangements are not lineage-specificevents. Because many aP lineage cells have one or more TCRy loci in germline configuration, it is also clear that a0 T-cell development can proceed in the absence of exhaustive TCRy rearrangements. Moreover, the analysis of the adult thymus did not provide any evidence for a specific temporal order of TCRyIG versus TCRP rearrangements. Taken together, these findings are incompatible with the sequential rearrangement model, at least in its historical, simplest form. This model will therefore not be considered any further. The finding that productive Vy-Jy and V&( D)JS rearrangements are underrepresented in cells of the aP lineage seems to provide strong support for the competitive rearrangement model because the most straightforward interpretation would suggest that progenitors with functional y6 rearrangements are diverted into the y6 lineage. Ilowever, underrepresentation of in-frame y and 6 rearrangements in ap lineage cells does not prove that successful TCRyIS rearrangements guarantee y6 T-cell development. The observed paucity of functional yI6 joins in a0 lineage cells could also be a result of selective cell death or deficits in proliferation, affecting those precursors that express the “wrong” TCR isotype (see later). Moreover,
aply8 LINEAGE COMMITMENT
35
the frequent occurrence of in-frame TCRP rearrangements in 76expressing cells can only be integrated in a competitive rearrangement model by invoking some as yet unproven, ad hoc assumptions. In summary, data obtained from the analysis of TCR gene rearrangements do not allow a distinction between the competitive rearrangement and the separate lineage model; rather, they are compatible with both models, provided some modifications are introduced. That yl8 rearrangements in CUPlineage cells are disproportionately out of frame is certainly a key finding because it demonstrates that these rearrangements are not neutral events but do influence cell fate, a conclusion that must be incorporated in any viable model of aPly8 lineage commitment. IV. Analysis of TCR Transgenic and Gene-Targeted Mice
A. TCR TRANSGENIC MICE In theory, TCR transgenic mice should be a perfect tool for studying the molecular mechanism of the aPIy8 lineage split, as the separate lineage and the rearrangement models make distinct predictions regarding the influence of TCR transgenes on lineage commitment. According to the rearrangement model, introduction of functionally rearranged receptor genes encoding a specific receptor isotype should direct all developing thymocytes into the corresponding lineage and prevent the formation of cells of the opposite lineage. However, if the separate lineage model were correct, the presence of productively rearranged TCR transgenes should not significantly interfere with the formation of either lineage. Although these predictions in theory seem sufficiently straightforward to allow an easy verification, the phenotypes of TCR transgenic mice regarding the development of aP and 76 lineages have turned out to be highly variable and often difficult to interpret. 1 , TCRaP Traiisgcizic Mice
Concerning the effects of TCRaP transgenes on 76 T-cell development, the most comprehensive set of data has been derived from the analysis of HY transgenic mice, which were generated with large genomic fragments encoding a and /3 T-cell receptor chains recognizing a male-specificpeptide (HY) in the context of H2-D” (Kisielow ct nl., 1988). Expression of the TCRP transgene in the presence or absence of the TCRa transgene was shown to suppress endogeneous V y 2 4Jyl rearrangements and to prevent the generation of 78-expressing cells in thymus and lymph nodes (von Boehmer et al., 1988). Together with similar data from another laboratory (Fenton et al., 1988), this result seemed to support the competitive rearrangement model, as it suggested that early expression of a functional
TCRP chain would relegate all T-cell precursors into the ap lineage. However, data from an analysis of highly unusual CD4-8- and CD4-8"" lymphocytes present in HY-TCR transgenic mice advocate an alternative explanation for the absence of y6-expressing cells. The new findings directly support the separate lineage model because these unusual cells were shown to have several features typical for y6 lineage cells (Bruno et d., 1996). First, they were characterized as DN or CD8'"" cells, which is unusual for mature aP lymphocytes. Second, they were not dependent on positive selection, as they were also found in HY-TCR transgenic mice with a nonselecting MHC background. Third, although they expressed the transgenic HY-specific TCR and could be induced to proliferate by antireceptor antibodies, they did not respond to male antigen. Moreover, the fact that these unusual cells accumulated in peripheral lymphoid organs of male mice demonstrated that their precursors were not subject to thymic negative selection, in contrast to HY-TCR-bearing precursors of conventional ap lineage cells. Fourth and most suggestive, they did not rearrange their endogeneous TCRa genes and retained TCRG alleles on both chromosomes, although conventional SP cells, including CD4-8' cells, from HYTCR transgenic mice showed extensive rearangement of endogeneous TCRa genes. Finally, these cells were absent in mice lacking the common cytokine receptor y chain (DiSanto et al., 1996), which has been shown to be specifically required for the development of inature y6 lineage cells (Cao et al., 1995; DiSanto et al., 1995). These particular features strongly suggested that CD4-8- and CD4-8'"" cells expressing the transgenic TCR actually represent cells of the y6 lineage. This interpretation was endorsed by the finding that a significant percentage of these cells coexpressed a y6 TCR when derived from HY-TCR transgenic mice lacking pTa (Bruno et nl., 1996). Therefore, it seems that y6 lineage cells are not really absent in HY and possibly other aPTCR transgenic mice, rather, they are disguised as aP cells that express the transgenic TCR. This interpretation is in line with earlier findings from another TCRaP transgenic mouse model (Capone et al., 1995). Capone and colleagues (1995) generated transgenic mice carrying an unrearranged TCR minilocus that was under the control of either the TCRa or the TCRP enhancer. This minilocus served as an artificial rearrangement substrate. It was found that rearrangement and expression of the /3 enhancer-containing transgenes occurred during thymopoiesis before those containing the a enhancer, with a pattern superimposable on the patterns of endogeneous TCRP or TCRa rearrangements and expression, respectively. These mice were then crossed with TCRaP transgenic mice expressing an alloreactive receptor specific for the MHC class I H2-K" molecule. The thymus of all K"-TCR transgenic animals contained a significant proportion of unusual I P - T C R ~ ~ ~ ~
aD/y/yfi LINEAGE COMMITMENT
37
cells that were CD4-8-. Analogous to the findings in HY-TCR transgenic mice mentioned earlier, these DN, K”-TCR’”gl’cells did not undergo negative selection on a deleting (H2-K”)background. Most interesting, although the transgenic rearrangement substrates under the control of the (Y enhancer were extensively rearranged in all other K”-TCR-bearing thymocytes, rearrangements were barely detectable in the unusual TCR’”zhDN population (Capone et ul., 1995). Although bearing a transgenic (YPTCR, these cells therefore resembled yS tliymocytes, in which V(Y+J(Y rearrangements are known to be very rare (see Section III,C,3). Taken together, data obtained with TCRaP transgenic mice therefore strongly suggest that early expression of an (YP TCR does not prevent tlie formation of y6 lineage cells. Rather, the transgenic ap TCR seems to be able to functionally replace the y6 TCR, allowing yS development, despite tlie absence of the “correct” receptor-a view fully in line with the separate lineage model.
2. TCRyS Trunsgenic Mice y6 transgenic mice should provide a inore sensitive system to distinguish between the different models of aply6 lineage commitment, as a block in the development of ap thymocytes, as predicted by the rearrangement, but not the separate lineage model, would produce a most obvious phenotype. Unfortunately and contrary to expectation, the analysis of TCRyS transgenic mice has revealed very complex and seemingly inconsistent phenotypes. Two distinct sets of TCRyS transgenic mice have been described by Tonegawa’s laboratory. The first set of mice was generated with large cosmid-based fragments encoding rearranged Vy2Jyl and Va5DJSl receptor chains that were derived from the alloreactive yS T-cell hybridoma KN6 specific for an allelic MHC class I molecule encoded in the T l a region of the MHC (T22”).Coinjection of these fragments yielded three transgenic lines with apparently similar phenotypes (Ishida et al., 1990). Thymic cellularity was nornial, as was the number and subset distribution of ap thymocytes. There was a modest increase in thymic 76-expressing cells from 0.,5% in nontransgenic mice to inaxirnally 5% in 76 transgenic mice. Notably, no transgenic y or S transcripts could be detected in cells of the ap lineage, although the respective transgenes were shown to be present in such cells. To investigate the mechanism of this specific block in TCRy and 6 transgene expression, another group of transgenic mice was generated using a genomic DNA fragment that contained essentially the same Vy2Jyl gene as before, but much less flanking sequence. This time, the 6 construct was not injected. Interestingly, the shortened y transgene was strongly expressed in ap lineage cells in all three independent lines of TCRy transgenic mice studied, whereas expression of the
38
H A N S JORC. F E l K I N G ct nl
corresponding endogeneous y genes was still suppressed (Ishida et d., 1990). These results suggested that the C y l gene, and by inference other Cy genes, carried a cis-acting DNA element in their flanking regions that silenced y transcription in ap lineage cells. Based on this interpretation, a new model for the differentiation of ap and y6 T cells was proposed. According to this model, generation of ap and y6 cells would be independent of the outcome of TCRy,S, or /3 rearrangements. Instead, commitment of a given T-cell precursor to the ap lineage would induce a silencing mechanism, which effectively blocks expression of rearranged TCRy genes, thereby preventing the formation of a functional y6 TCR in the inappropriate lineage. This model was tested in a second set of TCRyG transgenic mice constructed with short y transgenes lacking the putative silencer region. Possibly of relevance, this time the transgenes encoded a completely different y STCR, namely the nonpolymorphic “canonical” Vy3Jyl and VSlDJ62 receptor chains known to be expressed exclusively on the surface of embryonic y6 thymocytes that are generated during the first wave of thymopoiesis and give rise to DECs (reviewed in Allison, 1993; Boismenu and Havran, 1995). In both lines of “DEC transgenic” mice that were analyzed, T-cell development was severely perturbed (Bonneville et al., 1989).The absolute number of thymocytes was markedly reduced, and in young mice, ap-expressing thymocytes were virtually absent. In older mice, some ap-expressing thymocytes could be found but their number never reached more than 5% of the level observed in nontransgenic mice. Southe m blot analysis of thymic DNA revealed that D+Jp and Vp+(D)Jp rearrangements were severely blocked, providing a likely explanation for the paucity of ap-expressing cells in DEC transgenic mice. These results were interpreted in favor of the “silencer model” (Bonneville et al., 1989). It was argued that the absence of the silencer in the transgenic construct permitted early expression of both y and 6 transgenes in ap-committed cells, thereby blocking the development of such cells. However, other scenarios are conceivable. In uitru cell culture experiments have suggested that DEC receptor-expressing y6 thymocytes may be selectively depleted in the adult thymus, depending on the particular strain of mice (Iwashima et al., 1991). To explore this directly, Iwashima and colleagues (1991)generated DEC transgenic mice with hstinct genetic backgrounds. The genomic Vy3N61 constructs used for transgenesis were identical to those in Bonneville’s et al. (1989). DEC transgenic mice with a B6 genetic background revealed a similar phenotype as described earlier. However, subsequent experiments established that this phenotype was largely due to the presence of a negatively selecting factor in the B6, but not the C3H strain. For instance, in mice of the latter strain, there was no strong reduction in the total number of thymocytes, and the development of
DP thymocytes was barely inhibited, in sharp contrast to DEC transgenic animals with a BG back$-ound. Fiirtlier expeiiments demonstrated that this depletion was genetically dominant, mapped to chromosome 18, and was mediated by bone-marrow-derived, radiosensitive cells in the thymus. The inhibition of (YOthyrnopoiesis observed in DEC tranvgenic mice therefore seems to be more a function of the genetic background than being related to the presence or absence of a y gene silencer. Wiether a y silencer is important for an unperturbed development of cup lineage cells has been investigated in a third laboratory with yet another set of y6 transgenic mice (Sin1ct al., 1995). The 6 construct in these mice consisted of a VSlJ82CS cDNA, cloned into an expression vector that contained the heterologous H2-K” promoter and Ig heavy chain enhancer in order to support strong expression in lymphoid cells. The y construct consisted of a short geiiornic fragment carrying a productively rearranged Vy4Jyl gene segment with its own regulatory sequences, all Cyl exons, but, most important, not the putative y silencer. Analysis of nine transgenic lines revealed a wide range of phenotypes (Sim et al., 1995).The frequency of 76-expressing cells in the thymus varied between 1.8 and 93% with a spectrum of intermediate values depending on the particular line under investigation. Unfortunately, the influence of the 76 transgenes on the thymic subset composition and on the absolute number of thymocytes was not reported. Analysis of splenocytes, however, revealed that the number of a0 T cells in two lines was essentially normal, whereas in the remaining seven lines the frequency of a/3 T cells was depressed to various extents, ranging from 25 to 75% of normal. The extent of c.p T-cell suppression in the periphery did not correlate wit11 the frequency of 76-expressing cells in the thymus or with transgene copy numbers. Interestingly, in two lines a distinct population of splenic T cells was found to coexpress an a/3 TCK and the transgenic y6 TCR. These results were taken as evidence that expression of a y6 TCR in a0 lineage cells w7as not inconipatible with c.upT-cell development, implying that functioning of the putative y silencer was not essential for a commitment to the ap lineage (Sim et al., 1995). The study also indicated that the expression of functional TCRy and 6 genes was insufficient to direct the differentiation of a precursor T cell into the y6 lineage, clearly contradicting the rearrangement model. Unfortunately, the study did not provide any data on the RNA expression levels of the transgenes during thymopoiesis in the different transgenic lines, raising concerns that at least some of the phenotypes seen might have been due to weak, delayed, or \wiegated transgene expression. Although these are potential problems in TCR transgenic mice in general, they are more common when using highly synthetic, cDNA-based constructs as transgenes.
40
IlANS JOHC, FEIILING ct a/
Another striking example of the vanability in phenotypes of y6 TCR transgenic inice has been provided by Hedrick’s group. These investigators introduced into the mouse gerinline functionally rearranged TCRy and 6 genes derived froin another alloreactive y6 T-cell clone, called G8 (Dent et nl., 1990).The G8 receptor, like the K N 6 receptor in Tonegawa’s mice, was specific for the b (and to a lesser degree the k) allele ofthe nonclassical MHC class I antigen T22. The transgenic constructs used were unaltered genoinic fragments carrying productive Vy2Jyl and V68(V a l 1)JSl rearrangements isolated froin genomic DNA of the G8 hybridoma. The transgenic y construct was relatively short and most likely devoid of the putative y silencer. Surprisingly, mice within a single transgenic line had two rather different phenotypes (Dent et nl., 1990). In some animals, the thymus consisted of approximately 50% of cells expressing the transgenic y6 TCR, whereas cells bearing an a/3 TCR were virtually absent. In these mice, referred to as type I, the absolute iiiirnber of thymocytes was reduced 7- to 10-fold and the residual thyinocytes were predominantly CD4-8-, suggesting that expression of a y6 TCR had blocked the development of a/3 lineage cells. In other mice of the same line, referred to as type 11, inhibition of afl T-cell developinent was not as severe. Although thymic cellularity in these aniinals was also decreased 7- to 10-fold, the percentage of 76-expressing cells was not as strongly elevated as in type I transgenics and a norinal ratio of a/3 to y6 thynocytes was essentially maintained. Moreover, CD4/CD8 profiles and the expression pattern of the a/3 TCR in different thymic subsets were comparable to normal, nontransgenic mice. Why the same line of transgenic mice gave rise to two apparently quite different phenotypes remained unclear. The presence of different genetic background genes could be responsible. Also, a later study by other investigators demonstrated incidentally that the copy number of the transgenes has a critical influence on the degree of inhibition of a/3 development, as a/3 TCR-expressing cells were shown to be, on average, 8- to 9-fold rarer in hornozygous G8 transgenic inice than in heterozygous litterinates (Livhk et al., 1997). G8 transgenic inice with an H-2” background, in which the ligand for the G8 receptor, the MHC class I T22b “autoantigen” was expressed, exhibited a third phenotype (Dent et al., 1990): cells expressing the transgenic y6 TCR were mostly rare, evident only in relatively small numbers with significantly reduced surface expression levels of the G8 TCR. The analysis established that G8-expressing, self-reactive y6 cells were subject to thymic negative selection. Remarkably, the development of a/3 T cells in thymi of such mice was quite efficient. The relative proportions of DP and SP thymocytes and the expression pattern of the aPTCR were similar to nontransgenic mice. In the original description of G8 transgenic/H-2”
aplyS LINEAGE (:OMM lThlENT
41
mice, almost normal numbers of thymocytes were reported (Dent et al., 1990), although later studies revealed that total tliymic cellularity was significantly (4- to -5-fold) reduced in most animals (Livak et al., 1997). Nevertheless, the fact that large numbers of ap T cells were generated in a thymus with a grossly nonnal subset composition demonstrated that self-reactive transgenic y6 T cells could be eliminated by clonal deletion without precluding ap T-cell development. This implied that the ap T cells did not go through a precursor stage at which the transgenic y6 TCR was expressed in a functional form. These findings were in contrast to the phenotype of DEC transgenic mice, in which a deleting background almost completely blocked ap T-cell development ( Iwashiina et nl., 1991). In a separate approach, Hedrick and colleagues created an additional set of y6TCR transgenic niice (Kersh et al., 1995).The receptor, composed of a Vyl.lJy4Cy4 and a VBD62J61C6 chain, was derived from a hybridoina termed “BAS” and appeared to be specific for a ligand expressed by the hybridoina itself. Possibly of importance, the transgenes were constructed in such a way that their expression could not be extinguished in cells of the ap lineage: the genomic y construct was coinpleinented with an extra enhancer element inserted between two Cy4 constant region exons, and the y transgene consisted of the VDJC6 cDNA driven by regulatory elements of the human CD2 gene known to confer T-cellspecific, copy number-dependent, and insertion-independeiit expression. Based on the phenotype, all foiinder lines could be divided into two groups (Kersh et al., 199rj).In one group, there was low expression ofthe transgenic y cliain and little to no expression of transgenic 6. Not surprising, in these mice, ap T-cell development was normal. The other group was characterized by extremely high levels of transgenic y and 6 message in the thymus, a block in Vp+( D)Jp rearrangements, and an essentially complete absence of cells expressing an ap TCR. Surprisingly,even though Vp+( D)Jp rearrangements were almost fully blocked, these mice had virtually iiorinal numbers of DP thyinocytes, which were devoid of CD25 and expressed low levels of the transgenic y6 TCR. Moreover, these y6 expressing thyinocytes had deleted their endogeneous 6 loci and expressed full-length TCRa transcripts, strongly suggesting that they belonged to the a/3 lineage. However, development beyond the DP stage was coinpletely blocked, most likely because the requirement of ap lineage cells for positive selection could not be met by a y6 TCR complex. Collectively, these data indicate that high level and persistent expression of y and 6 (trans)genes in early thymocytes of the a0 lineage can lead to all consequences of pre-TCR expression: downregulation of CD25, proliferative expansion, differentiation into DP thymocytes, and inhibition of complete rearrangements at the entlogeneous TCRP locus, imitating allelic exclusion. To-
42
IlANS J O R G FEII1.INC et n/
gether with the identification of “disguised” y6 cells in HY-TCR transgenic mice described earlier (Bruno et al., 1996), the findings strongly suggest that CD2StCD44-””” pre-T cells do not distinguish among a@, y6-, or pre-TCR, implying that expression of a specific receptor isotype does not determine lineage commitment. Why do TCRy6 transgenic mice exhibit such a medley of distinct phenotypes? At least three parameters are likely to contribute to this rather unexpected variability. First, the outcome of experiments involving TCRtransgenic mice is almost certainly influenced by the strength and timing of TCR transgene expression. For instance, too low or delayed expression of a y or 6 transgene during the period critical for lineage commitment niay fail to reveal any effects, whereas overexpression of the transgenic TCR at a given stage could lead to excessive signaling and unphysiologic responses. The marked phenotypic differences between heterozygous and hornozygous G8 transgenic mice of the same line illustrate how varymg transgene copy numbers, and thus expression levels, can influence the pattern of thymopoiesis (LivAk et al., 1997). Second, the potential presence of undefined thymic autoantigens recognized by the transgenic TCR may contribute to unusual phenotypes. DEC transgenic C57/B16 ( Iwashima et ul., 1991) and G8 transgenic mice with a H2b background (Dent et al., 1990) provide a good example for the profound effect of a deleting autoantigen on thymopoiesis and highlight the potential for a misinterpretation of results regarding the role of the transgenic receptor in the aP/yS lineage decision. In this context, it may be worth pointing out that the phenotypes of G8 transgenic mice on the deleting background and of KN6 transgenic mice described by Tonegawa’s group (Ishida et al., 1990) are not exceedingly different. The possibility that the developmental pattern seen in mice transgenic for the KN6 receptor (also an alloreactive receptor) may be brought about by the presence of a deleting autoantigen rather than the proposed presence of a y silencer should therefore not be completely neglected. Finally, it cannot be excluded that the influence of a transgenic y6 TCR on the crPIy6 lineage decision depends to some extent on which particular kind of y6 rearrangement has been chosen for transgenesis. For instance, y6 T-cell development involves at least two distinct stein cell populations and occurs in, at minimum, three distinct waves, each associated with discrete types of Vy and V6 rearrangements (reviewed in Boismenu and Havran, 1995). The mechanism of lineage Commitment may therefore vary during different stages in ontogeny (see also Section VII). The use of autologously regulated y and 6 transgenes encoding receptors typical for one of the early waves may elicit results differing from those obtained with receptor genes representative for a later wave. The distinct effects of a deleting background on DEC transgenic (Bonneville et nl.,
a/3/yS LINEACE COMMITMEYT
43
1989) versus G8 transgenic mice (Dent et al., 1990) could be a case in point. Because the genes encoding the DEC receptor are representative for the first wave of thymopoiesis, the corresponding transgenes are most likely expressed much earlier in ontogeny than G8-encoding transgenes. Deletion of DEC-expressing thymocytes may therefore include most precursor cells of the ap lineage, whereas a relative delay in the expression of G8 might leave a significant number of ap precursors unperturbed, explaining why on a deleting background ap thyniopoiesis is completely blocked in DEC, but not in G8 transgenic inice. The massive expansion of ap lineage cells expressing a transgenic Vyl. 1JyW&(D)JSl-TCR in one transgenic model (Kersh et al., 1995) may likewise be related to the particular yl6 genes used. It is interesting to note that the V&(D)J61 rearrangement used in that model exhibits random distribution of in-frame joins in normal ap T cells (Livak et al., 1997; Wilson et al., 1996), whereas most other 6 rearrangements are depleted of in-frame joins, suggesting that many V&( D)JSl containing TCRs may behave differently from the most common TCRy6 heterodimers. Moreover, the observation that infranie Vyl.1Jy4 rearrangements are strongly selected in a subset of CD25-44-""\'cells has led to the suggestion that y chains with this particular rearrangement might act, in analoq to pTa, as a pre-Ty (Passoni et al., 1997). In contrast to expectation, the enormous variability in the phenotypes of different TCR transgenic mouse strains demonstrates that this experimental system does not provide the ideal tool to distinguish between different models of lineage determination. A major problem is to recognize which of the observed phenotypes reflect a physiologically relevant situation. However, the experiments clearly show that thymopoiesis is strongly influenced by the onset and level of TCR expression and possibly also by the type of V, D, or J elements of a particular TCR, but it is not clear to what extent if at all, these parameters influence the aply6 lineage decision in the unmanipulated animal. Although TCR transgenic mice at present cannot provide a definitive answer with respect to the different models of lineage commitment, their analysis has given important clues that can be used as a platform for further discussion and experimentation.
B. GENE-TARGETED MICE 1. Mice Unable to Geiierute TCRa, p, or 6 Chains Targeted disruption of the TCRa, p, or 6 constant region genes in embryonic stein cells has been used to generate mice that fail to express specific T-cell receptor chains. The effects of such mutations on T-cell development are unequivocal. TCRa- (Mombaerts et al., 1992b; Philpott
al., 1992) and TCRP-deficient mice (Mombaerts et al., 1992b) lack mature aP lineage cells, but maintain a full complement of 76 cells. Conversely, TCRG knockout mice are deficient in yG-expressing cells, but this deficit does not affect the development and number of aP lineage cells (Itohara et al., 1993).The results demonstrate that cells of one lineage can develop norinally in the complete absence of mature cells of the other lineage. Unfortunately, these findings alone do not provide any further insight into the mechanism of how the ap/yS lineage decision is made, as the fact that both lineages can develop independent of each other is compatible with both models of lineage commitment presented earlier.
r>t
2. Mornentous Clues: Analysis of TCRy and 6 Rearrangements in TCRP Knockout Mice The development of aP thymocytes is severely blocked in TCRP-deficient inice at the transition from the CD25+44-"" to the CD25-44-""" DN stage (Godfrey et al., 1994; Mombaerts et al., 1992b). It is now clear that this transition is controlled by the pre-TCR consisting of a conventional TCRP chain, an invariant molecule called the pre-TCRa (pTa) chain, and certain CD3 subunits (reviewed in Fehling and von Boehmer, 1997; Rodewald and Fehling, 1998; von Boehmer and Fehling, 1997).The cornpetitive rearrangement model suggests that if the generation of a functional TCRP chain and the subsequent formation of a signaling-competent preTCR precede formation of a ySTCR, the respective cell will commit to the ap lineage. Conversely, generation of a productive ySTCR before the generation of a pre-TCR is predicted to result in a decision favoring the yG fate. If this model were correct, one would expect that in the absence of a TCRP chain, precursor cells will fail to enter the ap lineage. This does not seem to be the case, as the thymus of TCRP knockout mice contains a significant number of DP thymocytes (Mombaerts et al., 1992b). An answer to the question of how these residual DP thymocytes arise in the absence of a TCRP chain, and whether they are genuine ap lineage cells, should therefore provide important clues about the molecular mechanism controlling the aP/yG lineage split. In TCRP-'- mice, DP cells constitute typically only about 20% of all thymocytes, which corresponds to approximately 1-5% of the absolute number found in wild-type littermates (although the frequency of these cells varies widely between individual animals). Importantly, in T C R P P X T C R V doubly deficient mice, DP thymocytes are almost absent, comprising fewer than 1% of all thymocytes, which corresponds to less than 0.085%of the absolute number found in normal, unmanipulated animals. The almost complete lack of DP thymocytes in TCRP-'- X TCRGP mice implies that the vast majority of residual DP thymocytes in
T C R P P singly deficient inice is generated by pathways dependent on functional TCRG (and most likelv TCRy) protein. Earlier experiments in severe combined immuno defic‘ient (SCID) mice have shown that the presence of y6expressing cells can induce other DN thymocytes in trans, in some unknown way, to upregulate CD4 and CD8, resulting in the generation of small nurnbers of TCR negative DP thymocytes (Lynch and Shevach, 1993; Shores et al., 1990). In theory, a similar mechanism could operate in the thymus of T C R P P mice, as it contains large numbers of y6 thymocytes. However, data from two independent laboratories clearly demonstrate that expression of a ysTCR can promote the generation of DP thymocytes not only in trans but also in cis. In the first study, Passoni and colleagues (1997) used the PCR-RFLP technique (Section II1,B) to analyze the quality of V&( D)JG rearrangements in the DP thymocyte population of TCRP-I- mice. Similar studies in normal mice (summarized in Sections III,C,l and III,C,2)had previously shown that in-frame TCRyIG rearrangements are generally underrepresented in DP thymocytes, a finding that had been proclaimed as strong evidence for rearrangement models of lineage commitment (Dudley et al., 1995; Livak et al., 1995). Interestingly, the analysis of rearranged 6 loci in residual DP thymocytes of TCRP-’- inice gave the opposite result: all three V&(D)JS rearrangements studied [VM, V65, V&(D)JSl] revealed inore than 75% of in-frame joins, indicating selection of productive TCRG rearrangements (6 selection). The overrepresentation of functional V&( D)J6 rearrangements in DP thymocytes of TCRP-’- mice was shown to be accompanied by y selection, involving primarily Vy2+ Jrljoins. These findings were confirmed and extended by another study ( LivBk et al., 1997). LivBk and colleagues (1997) first reexamined crP versus y6 T-cell development in G8 TCRyG transgenic mice that had been generated and described some years before (Dent et nl., 1990) (see Section IV,A,2). The G8 ySTCR is specific for the nonclassical MHC class I antigen T22”. Although the development of mature c.P T cells was shown to be severely blocked in honiozygous G8 transgenic mice, largely in line with previous data (Dent et nl., 1990),a significant number of DP thymocytes was found. Quantitative Southern blotting revealed that these DP cells were obviously generated by a TCRP-independent pathway, as they were virtually devoid of VP+( D)JP rearrangements. This was confirmed with G8 transgenic, TCRP-I- mice, which also contained a significant number of DP thymocytes. However, the generation of DP cells was shown to be completely blocked when the G8 transgenes, along with the TCRP-I- mutation, were bred on a negatively selecting H-2” (T22”-expressing)background. These data indicated that the development of DP thymocytes in G8 transgenic, T C R P P mice was contingent on expression of the G8 ySTCR, as the
development of DP cells was abrogated in the presence of a negatively selecting ligand. Importantly, although the proportion of TCRy 6’ cells were strongly reduced in mice where negative selection occurred, significant numbers of cells expressing reduced levels of the G8 TCR were still present. Obviously these cells were unable to induce the generation of DP thymocytes in trans. These findings prompted the same researchers to investigate the quality of TCRy and 6 rearrangements in DP thymocytes of nontransgenic TCRP-I- mice, also using the PCR-RFLP technique (LivBk et d., 1997). The results were very similar to those reported by Passoni et al. (1997), in that in-frame V h (D)J6 and in-frame Vy+ Jy rearrangements were found to be strongly overrepresented in sorted DP thymocytes from TCR0-I- mice. Interestingly, the proportion of in-frame TCRS and TCRy rearrangements at all examined loci was nearly as high as in bona fide y6 cells. In line with these findings, additional experiments involving in situ hybridization and Northern blotting demonstrated that the proportion of DP thymocytes expressing TCRG message was much higher in T C R P P than in wild-type mice. Taken together, these data indicated that, in the absence of a TCRP chain, a sizable fraction of DP thymocytes was generated from precursor cells expressing a functional y6 TCR. To confirm that the DP thymocytes under study really represented lineage cells, additional experiments were performed. First, it was shown that DP thymocytes in G8 transgenic, TCRP-’- mice not only expressed CD4 and CD8 at normal levels, but also had downregulated surface expression of CD25 and the y6 TCR. Second, and more significantly, DP cells had initiated TCRa gene rearrangement at multiple J a genes. Initiation of TCRa gene rearrangements was also demonstrated in DP thymocytes from nontransgenic TCRPY mice (Mombaerts et nl., 1992). All these features are typical for ap lineage cells. The two studies just presented make a very important point. The finding that in-frame TCRy and S rearrangements are overrepresented in sorted DP thymocytes of T C R P P mice provides strong evidence that expression of a functional y6 TCR in a precursor cell is not only compatible with a developmental progression to the DP stage, but even necessary for the generation of a significant fraction of DP thymocytes when a TCRP chain is not available. This finding is so important, as it completely contradicts the competitive rearrangement model. However, analysis also shows that this $-driven pathway to the DP stage is very inefficient, as it becomes apparent only when the major, pre-TCR-dependent pathway is blocked. At least three potential reasons for this inefficiency are conceivable, which are not mutually exclusive: First, in contrast to the pre-TCR-driven pathway, which requires only a productive TCRP rearrangement to be opera-
ol/3/yG LINEAGE COMMITMENT
47
tional, the ySinediated generation of DP thymocytes is thought to need both a functional TCRy and a functional TCRG chain. The probability of generating in-frame rearrangements for both genes within one cell may be relatively low. Second, expression of a y6 TCR in DN precursor cells may not be associated with cellular proliferation, the hallmark of preTCR-mediated differentiation to the DP stage. Cell cycle analysis of DN thymocyte subsets in TCRP-’- mice seems to support this hypothesis (Passoni et al., 1997). Finally, the 76-driven pathway may be accessible for a tiny subset of DN T lineage precursors only. One could hypothesize that such cells belong to a special lineage of unknown function. Expression of a productive yS TCR in all other, more conventional precursor cells might then be expected to give rise to y6 lineage cells, fully in line with the predictions of the rearrangement model of lineage commitment. The demonstration that only a relatively insignificant fraction of precursor cells can follow the y6-driven pathway to the DP stage would rescue the competitive rearrangement model.
3. Lineage Commitinent in pTa-De$cient Mice TCRa genes are rearranged and expressed relatively late in T-cell development, and the availability of functional TCRa chains is therefore unlikely to influence the a@yS lineage decision. In accordance with this view, a/? T-cell development proceeds norinally up to the DP stage in TCRadeficient mice, and a change in tlie ratio of ap- to y6-committed thyinocytes cannot be detected (Mombaerts et nl., 1992b; Philpott et al., 1992). A potential role of TCRP rearrangements in the aP/yS lineage decision, as postulated by the rearrangement models, is thus expected to be mediated predominantly by the pre-TCR rather than a mature aP TCR, intimating an important role of pTa in the aP/y6 lineage switch. The phenotype of pTa-deficient mice seeins consistent with this view (Fehling et al., 1995a, 1997):whereas the generation of CUPlineage cells is severely impaired, the development of cells with a y6 TCR proceeds normally. As a result, the proportion of y Sexpressing cells in tlie thymus of pTa? mice is strongly elevated, comprising typically 10 to 20% of the total number of thymocytes. In some mice the percentage of thymic 7 6 cells can be as high as 30%.More important, however, the absolute number of 76-expressing thymocytes is also augmented and is, in general, 3- to 10-fold higher than in pTa’ littermates (Fehling et al., 1997).This can be interpreted as evidence for a direct role of pTa and the pre-TCR in inhibiting the formation of y6 cells in normal mice, as suggested by the rearrangement models of lineage commitment. Interestingly, a pTa-deficient thymus not only harbors larger numbers of 76 cells, but also features a novel population of y6expressing cells that
48
HANS JORC FEIILING ef a[
coexpress CD4-a population that cannot be detected in the adult thymus of normal mice (Fehling et al., 1997). Although these cells have not yet been characterized in detail, the peculiar surface phenotype suggests that the absence of pTa affects y6 T-cell development not only quantitatively, but can lead to a qualitatively different thymic subpopulation-a finding that again might be suggestive of a direct role of pTa in aPly6 lineage decisions. However, pTa-deficient mice are not completely devoid of apexpressing cells. Similar to TCRP-’- mice, they generate significant numbers of DP thymocytes. At least some of these DP cells express an aPTCR and give rise to apparently normal aP T cells (Fehling et al., 1995a), excluding an absolute requirement for pTa and the pre-TCR in directing cells into the ap pathway. The analysis of pTa? X TCRa-’- mice strongly suggests that some cells in pTa-’- mice follow the aP lineage due to functional replacement of the pre-TCR by a prematurely expressed a/3 TCR (Buer et al., 1997). In fact, the majority of mature ap-expressing cells in pTa+ mice are most likely derived from such precursors that, for as yet unknown reasons, happen to rearrange and express TCRa genes unusually early at the CD25+ DN stage. However, ap lineage cells seem to arise even in pTa? X TCRa-’- doubly deficient mice, as indicated by the presence of DP thymocytes in numbers equivalent to those in pTa? single mutant mice (Buer et al., 1997). These cells might be generated in the same way as DP thymocytes in TCRP-’- mice (see Section IV,B,2). If correct, one would predict that PCR-RFLP analysis will reveal an overrepresentation of in-frame y and 6 rearrangements in these pTa+ X TCRa-’- DP thymocytes. The failure to detect 6 chains in pTa-’- X TCRa-/- DP thymocytes by cytoplasmic staining with anti-TCR6 antibodies (Buer et al., 1997) might just signify that these cells expressed a y6 TCR at an earlier stage in their development, but downregulated expression of the y and/or 6 chain upon entering the DP stage. The presence of DP thymocytes in pTa-’- and especially in pTa-’- X TCRa? mice argues against an obligatory role of pTa and the preTCR in lineage commitment. A similar conclusion had been reached after analysis of TCRP-deficient mice (see Section IV,B,2). The function of the pre-TCR could therefore be limited to trigger expansion of a0 lineage cells and not to influence the lineage decision itself. The significant increase in the absolute number of y&expressing thymocytes in pTa-’- mice could be the result of an indirect effect of pTa deficiency, e.g., the provision of additional space in the thymus due to the reduced number of ab lineage cells. Notably, even in pTa-’- and TCRP-’- mice, y6 cells cannot compensate for the loss of aP lineage cells, as total thymic cellularity on the whole remains below 10% of that found in normal mice (Fehling et al., 1995a),
confirming the largely independent regulation of cell number in both lineages, as already described for TCRy6 transgenic mouse strains exhibiting ii block in aP tliymopoiesis. The assumption that pTa is not directly involved in the lineage decision does not exclude a potential negative influence of the pre-TCR on the rearrangement and/or expression of the TCRyIG loci, suggested by two kinds of observations. First, inhibition of endogeneous Vy2+Jyl rearrangements in TCRP transgenic inice (von Boehmer et al., 1988) is much less pronounced in mice carrymg the same P transgene but lacking pTa (Krotkova et al., 1997). A similar phenomenon is seen with regard to V(D)J6rearrangements (A. Krotkova, H. von Boehmer, and H. J. Fehling, unpublished observation). Second, whereas thyniocytes and lymph node cells from HY-TCR transgenic mice with a pTa+ background are essentially devoid of y6 TCRs, a significant fraction of these cells coexpress the transgenic a0 receptor and the endogeneous y6 TCRs when derived from transgenic litterrnates lacking pTa (Bruno et al., 1996). These findings clearly suggest that expression of a functionally rearranged TCRP chain and subsequent forination of a pre-TCR can strongly inhibit not only further rearrangements at the second TCRP allele, but also rearrangements and perhaps expression of y and 6 loci, at least in TCR transgenic mice. It inust be pointed out that this does not automatically imply that the preTCR can iriliibit the forination of y6 lineage cells under physiological conditions, i.e., in normal mice. For instance, it is conceivable that forination of a pre-TCR and inhibition of y and 6 rearrangements occur only after lineage commitment has taken place, just reducing the number of aP lineage cells with y6 rearrangements. In such a scenario, the pre-TCRmediated inhibition of y6 rearrangements would be fully compatible with the separate lineage model because the inhibitory effects of the pre-TCR would be irrelevant for the lineage decision itself. The fact that y6 T-cell development does not depend at all on the presence of pTa may allow cells that coininit to the 76 lineage to inimediately switch off pTa expression. This may, in fact, be necessary to avoid inappropriate expansion of 76-committed cells coincidentally expressing a functional TCRP chain. These considerations lead to an interesting hypothesis: cells that are destined to follow the y6 lineage may no longer express pTa. This hypothesis would predict the existence of two separate precursor populations for aP and y6 lineage cells that can be distinguished based on the presence or absence of pTa. Targeting of a marker gene that can give rise to an easily identifiable cell surface protein, into the pTa locus of embryonic stem cells by hoinologous recombination, should allow the generation of a mouse strain in which such a presumed lineage-specific expression patteiii of pTa can be easily detected, even in the absence
50
HANS JORC FEHL.INC: rt al
of TCRP and other components of the T-cell receptor complex. Such experiments are in progress.
4 . Potential Role uf “Notch” in Lineage Coniinitment Developmental studies in invertebrates, such as Drosophila or Caenorhabditis elegans, have revealed an important role of the transmembrane receptor protein ‘Notch’ and its ligand ‘Delta’ in a number of distinct cell fate decisions (reviewed in Artavanis-Tsakonas et al., 1995; Greenwald and Rubin, 1992; Simpson, 1995). Signaling through ‘Notch’ can divert cells that would otherwise follow a default pathway of differentiation to adopt an alternative, secondary fate. Quantitative differences of ‘Notch’ expression between adjacent cells appear to play a central role in establishing the lineage split, in that initially small differences in ‘Notch’expression between neighboring cells (possibly due to random fluctuations in gene expression) are amplified over time by the ability of ‘Notch’to boost its own expression and inhibit expression of its ligand. This feedback loop results eventually in the formation of two cellular subsets, which express ‘Notch’ and its ligand ‘Delta’in a mutually exclusive fashion, forcing them to follow distinct developmental fates. The finding that at least one of several murine notch homologues is expressed at high levels in developing thymocytes (Hasserjian et al., 1996; Weinmaster et al., 1991) has prompted a group of researches to investigate whether ‘Notch’ might also play a role in lineage decisions in the thymus. A first study revealed that transgenic ‘Notch’can alter the CD4CD8 ratio of developing thymocytes in favor of CD8-bearing cells when expressed in an activated form (Robey et al., 1996). Along with other data, this finding suggested that ‘Notch,’ in concert with the specificity of the TCR, influenced the choice between CD4 and CD8 T-cell lineages. A second study specifically addressed the role of ‘Notch’ in the aPly6 lineage split, and several observations suggested that ‘Notch’signaling might promote preferential development along the a/3 lineage (Washburn et al., 1997). First, reconstitution of R A G - P mice with a 50 : 50 mixture of bone marrow cells from wild-type mice and from mice in which one ‘notch’ allele had been inactivated by targeted mutagenesis revealed a significantly lower ratio of notcht’- versus notcht’+ donor cells among a0 T cells than among y6 T cells, indicating that cells heterozygeous for ‘notch’ contribute less to the a@ than to the y6 population. Second, in reconstituted mice, the proportion of cells heterozygous for ‘notch’was much lower among C D 4 CD8 DP thymocytes, which are considered to belong to the aP lineage, than among their CD25+ DN precursors, which are bipotential. Third, although the absolute number of y6 thymocytes was not decreased in transgenic mice expressing an activated form of ‘Notch,’an almost fourfold
a@yS LINEAGE COMMITMENT
51
increase in the proportion of 76 TCR’ thymocytes expressing the CD8aP heterodimer was found. Introduction of the ‘notch’transgene into ySTCR transgenic mice expressing the G8 receptor resulted in a similar effect, in that the absolute number of 76-expressing cells was not reduced, but the proportion of CD8aP bearing y6 cells was raised from less than 20% to more than 50% of the 76 population. In fact, a significant fraction of these cells was reported to coexpress CD4 along with CD8aP. Because CD4 and CD8aP expression are generally considered markers for the ab lineage, the observed phenotype was taken as a possible hint that many +expressing thymocytes in ‘notch’ transgenic mice might have actually adopted an aP T-cell fate. Fourth, introduction of the ‘notch’ transgene into TCRP-’mice resulted in essentially complete restoration of the DP thymocyte compartment, indicating-according to the authors-a selective promotion of aP T-cell development. The fact that activated ‘Notch’ had no effect on thymic development when introduced in RAG-deficient mice suggested that ‘Notch’could not act alone but required the presence of a y6 receptor. This was strongly supported by the demonstration that productive rearrangements of the TCR6. and :dso y lociis, wcrc ovcrrepresented in DP thymocytcs from TCWfl-’- mice expressing activated ‘Notch,’in contrast to DP thymocytes from normal mice, which carry predominantly nonproductive y and 6 rearrangements. The authors of this study interpreted their results in the framework of rearrangement models of lineage commitment (Washburn et al., 1997). They proposed that the generation of productive TCRP or TCRy6 chains would influence in some unknown way whether a precursor cell could receive a ‘Notch’ signal. The signals from ‘Notch’ and the respective TCR would then be integrated to lead to a specific lineage decision, in that reduced signaling through ‘Notch’would favor the 76 and enhanced signaling the ap pathway. Although the reported results are highly suggestive and consistent with a role of ‘Notch’ in the aP/yS commitment process, alternative explanations are conceivable. For instance, a similar bias in the ratio of y6 versus aP cells, as the one reported for notch+’-/notch+’+ bone marrow reconstituted mice, would be expected if ‘Notch’ provided a tliymocyte-specific survival rather than a differentiation signal and if y6 cells tolerated a reduction in ‘Notch’ signaling more readily then aP cells. The overrepresentation of in-frame y and 6 rearrangements in DP thymocytes of notch-transgenic TCRP KO mice does not provide conclusive evidence for an involvement of ‘Notch’ signaling in lineage commitment decisions either, because newer studies have found such an overrepresentation also in T C R P P mice lacking a ‘notch’ transgene (Passoni et al., 1997; Livlik et al., 1997; see also section IV.B.2). A ‘notch’ transgene is thus not required to allow thymocyte precursors with in-frame y and 6 rearrange-
ments to develop along the ap pathway; it might just boost their proliferative expansion. Clearly, more details about the effects of ‘Notch’ signaling on defined molecular events in thymocyte development are required before a crucial role of ‘Notch’ in ap/yS lineage commitment can be regarded as firmly established. Studying the course of thymopoiesis in the complete absence of ‘Notch’ might actually allow a more conclusive definition of its role in T-cell development in general and the ap/y6 lineage decision in particular. Because conventional ‘notch’ knockout mice die at a stage well before the initiation of thymopoiesis (Conlon et al., 1995; Swiatek et al., 1994), conditional gene disniption should be considered. V. Cell Culture Studies
Cell culture studies have been carried out in order to determine if the commitment to the cup or yS lineage can be influenced in vitro (Schleussner et al., 1991). Briefly, culturing undisrupted mouse fetal thymus lobes at an air-liquid interface, as so-called fetal thymus organ cultures, results in the development of predominantly ap T cells. However, in optimal cultures, the absolute number of recovered ap lineage cells in FTOC is considerably less than that in the corresponding thymus in vivo (Ceredig, 1988). Conversely, the proportion of yS cells in FTOC is higher than in the corresponding thymus in vivo and can be further increased by the addition of IL-7 to FTOC (Plum et al., 1993). The addition of IL-7 also results in the decreased recovery of a/3 cells. In FTOC, the development of cup but not y6 T cells can be inhibited by activation of the CAMPdependent signal transduction pathway and is associated with a dramatic inhibition of TcRa rearrangements (Lalli et al., 1996). When whole fetal thymus lobes are cultured submerged in tissue culture medium gassed with 10% C02,y6 T cells grow preferentially (Ceredig et al., 1989). Again, a significant inhibition of TcRa rearrangement is seen (E. Mertsching and R. Ceredig, unpublished observation). Importantly, the generation of ap T cells in submersion culture can be restored by elevating oxygen concentrations (Ivanov et al., 1993). Although the exact molecular mechanism(s) responsible for the differential growth of (up versus y6 T cells in these cultures has not been defined, these systems demonstrate that a variety of factors usually not considered in lineage commitment models potentially influence the developmental fate. IL-7 is the only cytokine for which a nonredundant function in thymopoiesis has been clearly established. Its potential role regarding the apl y6 lineage decision therefore deserves special comment. In uitro studies have suggested that IL-7 has an important role in promoting the development of y6 cells. These findings are strongly supported by data obtained
aP/y6 LINEAGE COMMITMENT
53
with gene-targeted mice. Notably, TCRy6+ cells are absent in IL-7 (Moore et al., 1996), IL-7Ra (He and Malek, 1996; Maki et al., 1996b), and ILRy ( y o (Cao et nl., 1995; DiSanto et al., 1995, 1996) knockout animals, whereas the development of ap lineage cells in these mouse mutants is impaired h i t not abrugiitcd. The differential sensitivity of up a r i d y6 cells with respect to 1L-7 deficiency could suggest a direct role of IL-7 in tlie process of lineage commitment. This view is supported by studies in IL7Ra-deficient mice indicating that IL-’7 is selectively required for the initiation of Vy+Jy rearrangements (Maki et nl., 1996a). The arrest of y6 but not ap T-cell development could therefore be due to the inability of IL-7Ra chain-deficient mice to rearrange TCRy loci (Maki et al., 1996a). However, an analysis of TCRy and 6 rearrangements in 7,-deficient mice has revealed rearrangements of both types with normal levels of diversity (Rodewald and Haller, 1998). These latter findings clearly showed that y(mediated signaling was not required to obtain diverse TCRy rearrangements. These apparently discrepant results can be reconciled by assuming that signals can be transduced via the IL-7Ra chain in 7,-deficient mice but not by yc in IL-7Ra-deficient mice. It is also possible that ligands other than IL-7, that can bind the IL-7Ra but not yc chain, are required for TCRy rearrangements. Thymic stroma-derived lymphopoietin (TSLP) is an obvious candidate (Peschon et al., 1994). Moreover, it cannot be excluded that the apparent increase in the level of rearrangements in cells cultured in the presence of IL-7 is due to the preferential outgrowth of cells already containing rearrangements. The role of IL-7 in TCR (Candeias et al., 1997; Muegge et al., 1993) and also IgH (Corcoran et al., 1998) rearrangements therefore remains controversial. VI. Developmental Considerations
Does the mechanism of lineage dwergence differ between the fetal and the adult thymus and is this dependent on tlie stem cell population developing at that particular time? It was shown some time ago that the thymus in birds (Jotereau and Le Douarin, 1982) and mice was colonized more than once during the fetal period (Jotereau et al., 1987). There are several differences between tliymocyte development in fetal mice compared with the adult. a. In the fetal thymus, the kinetics of development are more rapid, so that transition from earliest precursors to mature T-cell migration from the thymus takes only 6-7 days. In the embryo and neonatal mouse, mature T cells do not appear to require a prolonged sojourn in the medulla prior to migration to the periphery. Mature ap and y6 T cells are cycling in
54
HANS JOKG FEIILING et nl
the fetal but not the adult thymus (Ceredig, 1990).The cohorts of precursor cells developing in the fetal and adult thymus may have different origins. As reviewed by Rodewald and Fehling (l998), data obtained with Tcf-l (Verbeek et al., 1995) and Zkaros (Wang et al., 1996) KO mice indicate that inactivation of these genes may have different effects on fetal versus adult T-cell progenitors. b. The efficient generation of Vy3+ cells in the fetal mouse thymus, cells that migrate to the skin to form DECs, seems to depend on a combination of heniatopoietic stem cells and fetal thymus stroma (Uchida et al., 1993). Data (Mallick-Woodet al., 1998) from the fetal thymus argue strongly for the presence of a positive selecting ligand for these canonical y6 receptors. Sequence data from day 14 mouse fetal thymus cDNA indicated that TcRy receptors with noncanonical sequences were expressed at this time (Schleussner et al., 1992), supporting a positive selection model for the generation of y6 cells expressing canonical receptors. c. In addition to V gene repertoire (Havran and Allison, 1988) and junctional diversity (reviewed in Gilfillan et al., 1995), differences in the occurrence of Va-Ja rearrangements between fetal and adult y6 cells can be noted. Such rearrangements are detectable at low levels in adult but not in fetal y6 cells (Mertsching et al., 1997). Fetal but not adult y6 cells are generated as cycling cells (Ceredig, 1990), and the cycling status of y6 cells may reduce RAG protein activity, thereby influencing subsequent TcRa and/or p rearrangements. Alternatively, the efficiency of a rearrangements may be developmentally regulated. d. The fetal thymus contains a greatly increased proportion of y6 cells compared with the adult (Havran and Allison, 1988; Pardoll et al., 1987). In the adult, the presence of a@ cells may have a negative influence on the development of y6s, which may involve the Notch signaling pathway (Robey and Fowlkes, 1998) (see Section IV,B,4). It is unclear if there are differences in Notch signaling between the fetal and the adult thymus. VII. In Search of a Consensus Model for the @ / y 6 lineage Split
A. SUMMARY OF EXPERIMENTAL DATA A decade of research into the mechanism of aPIy6 lineage commitment has not yet provided a generally accepted scheme that would explain how a bipotential precursor decides which of the two alternative T-cell pathways it will choose. However, the large collection of diverse experiments described in detail on the previous pages has given nse to a defined set of data that can serve as building blocks for any future model of aPIy6 lineage commitment. These data can be summarized as follows.
afily8 LINEAGE COMhlITMENT
*55
1. The three most immature CD4-8- thymic subsets, defined as CD25-44 'c-kit +, CD25 '44 'c-kit +,and CD2Fj+44-"""c-kit-'""'TN subpopulations, can develop into both ap and 76 lineage cells in thymic organ culture or upon adoptive transfer in RAG-deficient mice. Tlie most advanced TN subset (CD2i3-44-'"''c-kit-''"" cells) seeins to be able to generate y G-expressing thyinocytes as well, when transferred intrathymically (Petrie et al., 1992), although an independent study involving FTOCs failed to confirm this finding (Godfrey et nl., 1993). 2. Vy-Jy and V&(D)JS rearraiigeinents are coininon in a0 lineage cells, and DP-JP as well as complete VP-(D)JP rearrangements are common in yG lineage cells, demonstrating the absence of tight lineagespecific controls with regard to TCRy, 6, and /3 rearrangements. TCRS rearrangements in aP lineage cells occur to a significant extent on chromosomal DNA, before V a 4Ja-mediated excision of the TCRG locus. 3 . The vast majority of mature a0 T lymphocytes have Va-Ja rearrangements on both chromosomes, resulting in the complete deletion of the TCRG locus. However, deletion of the TCRG locus is not necessary to coininit cells to the UP lineage. In 76 lineage cells, Va-Ja rearrangements occur 011 less than 10% of all available alleles and are therefore rather rare. The paucity of Va- J a rearrangements in iriature T cells can be taken as a marker for cells that have developed along the yS pathway. The biological significance of the residual Va- Ja rearrangements in 78 cells is unclear. 4. In-frame Vy-Jy and V&( D)JGrearrangements are generally underrepresented in aP lineage cells. The depletion of productive TCRy and TCRG joints is the result of a selection process and not of a bias introduced by the rearrangement mechanisin. Tlie depletion is detectable in mature aP T lymphocytes,in DP thyniocytes, and, with respect to TCRG rearrangements, apparently also in CD25+44-""" DN thymocytes, suggesting that diversion of yG lineage cells begins at this early developinental stage. 5 . y6 T lymphocytes and yG thymocytes are clearly not devoid of functional TCKP rearrangements. In fact, in-frame VP+( D)JP rearrangements appear to be overrepresented in most peripheral y6 T lymphocytes, except in DECs. An overrepresentation of productive P joints has not been found in day 15 embryonic yS thymocytes, and whether there is a significant overrepresentation of functional TCRP rearrangements in adult 76 thymocytes is still a matter of debate. The apparent dominance of y6 T cells, and possibly yG thymocytes, with in-fi-ame TCRP rearrangements is unexplained, both mechanisi-ically and teleologically. 6. The effect of functionally rearranged TCRyG transgeiles on the development of a0 lineage cells in TCR transgenic inice is extremely variable between different transgenic inoiise lines and sometimes even within a
56
H A N S JCjHG FEIlLlNG et NI
given line. In many TCRy6 transgenic mice, the development o f a p lineage cells is not abrogated. However, development is blocked in DEC transgenic mice, which express a y6 TCR typical for the first wave of tliymic yS cells, and in some mice of the G8 transgenic strain expressing a y6 TCR specific for certain alleles of the nonclassical class I antigen T22. In some situations, where aP development is blocked by y6 transgenes, it is difficult to exclude that negative selection rather than altered lineage decisions are responsible for the observed phenotype. When the development of afi lineage cells is blocked by 76 transgenes, the observed increase in 76-expressing cells never compensates for more than 10% of the total number of cells in a normal thymus, indicating autonomous control of homeostasis for both lineages. TCR can inhibit the formation of y6 7. Expression of a transgenic TCRs, but at the same time rescue the development of y6 lineage cells (Bnino et nl., 1996).Conversely, expression of certain transgenic constructs encoding a y6 TCR can block the rearrangement of endogeneous TCRP loci, and thus the generation of a functional TCRP chain, but at the same time fully rescue the development of aP lineage cells up to the DP stage (Kersh et nl., 1995). 8. The development of y6 lymphocytes proceeds normally in the absence of TCRa, TCRP, and pTa chains. Conversely, the absence of TCRS chains does not affect the generation of lineage cells. These findings imply that the differentiation of cells of one lineage is independent of the presence of mature cells of the opposite lineage. 9. The formation of a functional pre-TCR is not obligatory for thymic precursor cells to enter the aP pathway, as documented in pTa- and TCRP-deficient mice, which contain small but significant numbers of a$ lineage cells. 10. A significant portion of the residual DP thymocytes in TCRP-’mice represent cells selected for in-frame TCRy and 6 rearrangements. A similar situation may exist in pTa-deficient mice. Provided that the respective DP thymocytes in TCRP-I- mice are genuine a0 lineage cells, as many parameters suggest, this seminal finding implies that expression of a ysTCR allows precursor cells to follow an aP fate. This result discounts the rearrangement model of lineage commitment because it demonstrates that the isotype of the TCR is not the determinant of intrathymic lineage commitment, at least in TCRP-deficient mice. The question is now to what extent this observation can be generalized, i.e., whether it applies to most T-lineage-committed precursor cells or only to some atypical subset(s ) that may be dominating in the absence of a functional TCRP gene. 11. The absolute number of y6-expressing thymocytes is 3- to 10-fold increased in both pTa-’- and T C R P P mice. This increase cannot compen-
sate for the dramatic loss of aP lineage cells in these mouse mutants, indicating again an aiitoiioinous control of homeostasis for both T-cell lineages. 12. The expression level of the traiisismembrane receptor ‘Notch’ might influence tlie aply6 lineage decision, as cells with relatively higher levels of ‘Notch’ fivor the a0 pathway and those with relatively lower levels the y6 pathway. 13. Cell culture studies suggest that microeiivironinental factors, such as oxygen or carbon dioxide concentration, presence of cytokines and hormones, and cell-cell contacts, might have the potential to influence the ap versus y6 lineage decision. 14. The fact that fetal and adult thyinopoiesis differ to some degree raises concerns that certain aspects of lineage commitment might change at or around birth. This possibilit\, sliould he kept in mind when interpreting results obtained with young ‘id ti 1t mice. A major problem in finding a unifying sclieme of aP/yS lineage commitmet based on these data is posed by the confusingly diverse phenotypes observed in TCR transgenic inice (suniniaiy:points 6 and 7 ) .If‘one assumes that at least some of these phenotypes represent transgenic-specific peculiarities that do not reflect a physiologically relevant situation, one might decide to disregard the results obtained with TCR transgenic inice altogether and to focus, for the sake of simplicity, first of all on data obtained with norinal and gene knockout mice. If‘this approach led to an inherently concordant model, one coiild then, in a second step, assess to wliat extent the various data derived from transgenic inice support or discredit this conselisus. Following this approach, the following sections present inodified versions of the classical competitive rearrangement and separate lineage models, taking into accoiiiit tlie findings listed earlier. The presented models may help pinpoint “loose ends” and “moot points,” thus providing a platform -for further experimentation and discussion.
B. R E\T, E 11COMPETITIVE RRAHHAN c Based on the competitive rearrangement inodel, the following scheme of a/3/y6 lineage commitment can be devised (Fig. 2 A ) . If a functional TCRP chain is generated in ii bipotential T-cell precursor before the formation of ;I fiinctional ySTCR, the TCKP chain will associated with pTa and form a signaling-conrpetent pre-TCR complex. Signaling through the pre-TCR will then result in coinpletc inhibition of further rearrangemerits at all TCRy, 6. and P loci in tlic rcspective cell, upregulation of ‘Notch’ activity, commitment to the aP lineage, escape from prograiriined cell death, and massive cellular proliferation, followed by lineage-specific
bipotential precursor cells
A
-
tive B rearrangement first
productive y and 6 rearra formation of a y first
formation of a pre-TCR
pTa-
pTa+
* shutdown of further Y
3 shutdown
and S rearrangements of pTa expression I
3 commitment
3 shutdown
to the ve
B
0
of further rearrangements notch activity the ap lineage
3 upregulation of 3 commitment to
rearrangement
+ downregulation of CD25
bipotential precursor cells
ap-committed
*shutdown of pTa expression
no productive p T a + P S ) O rearrangements
rearrangements
tJB
productive p rearrangement reyangement
no productive rearrangement
formation of a yG-TCR
/
productive 1and 6 rearrangements 3 shutdown of further rearrangements
3 shutdown
NO or VERY LITTLE
/
mature yS thymocyte
t
inappropriate TCRF failed positive selection
3
=1formation
of a pre-TCR
of further rearran ements
P
aplyG IJNEAGE COMMITMENT
59
differentiation. If, however, the bipotential T-cell precursor manages to express a functional ySTCR before a pre-TCR, it will stop further rearrangements at the remaining TCRy and 6 loci, rapidly shut off pTa expression, downregulate ‘Notch’ activity, and commit to the y6 lineage. Subsequent TCRP rearrangements may be impaired, but the generation of a functional TCRP chain will have little consequence for the respective cell due to the postulated absence of pTa. Nevertheless, the presence of a functional TCRP protein may give rise to some weak, pre-TCR-independent signal that could result in a veiy modest expansion of the respective cells, similar to that seen in TCRP-transgenic mice lacking pTa (Krotkova et d., 1997),which might explain a potential overrepresentation of productive P rearrangements in y6 thymocytes (summary: point 5). A critical prediction of this modified version of the original competitive rearrangement model remains, namely, that a bipotential precursor must perceive and interpret signals from a pre-TCR very differently from those generated by a y6 TCR.
FIG. 2 . Revised versions of the "competitive) rearrangeinent” (A) arid the “separate lineage” ( B ) models. Tlie classical “separate lineage model” was updated to be compatible with the finding that productive TCRy a r i d 6 rearrangernents are clearly underrepresented in cells of tlie (YPlineage. This obsenution has been incorporated into the revised model (B) by assuming that cells committed to the CUPlirieage but express a y6 TCR will fail to significantly proliferate and thus become a minor subset, in contrast to those precnrsors expressing a pre-TCR. Moreover, y8-expressing ab lineage cells are expected to perish, either on their way toward tlie D P stage because expression of tlie y and/or 6 genes is turned off ( e g , through excision of the S locus aiicl/or activation of a y silencer) or at the D P stage itself because of the presnined inability of a y6 TCR to mediate positive selection of (YPlineage cells. Further modifications of both classical models were necessary to account for the fact that y6-expressing cells are not devoid of functional TCRP rearrangeinents. This is explained in the revised models by assuming that coniniitment to the yS lineage will shut-off pre-TCR signaling, either 1)y extinguishing p Ta expression (as shown) or by sonic other as yet unknown niechanisnt. Alternatively, tlie productive rearrangements seen in y6 lineage cells may give rise exclusively to TCHP chains incapable of forming a preTCR due to pairing problems with pTa and/or the hypothetical VpreT subunit. The demonstration that a significant fraction of‘DP tliyniocytcs in TCRP-’- mice is ( y ) 6selected appears incompatible d i the competitive rearrangeinent inodel depicted in A. To rescue this model, one has to invoke at least wie of’tlie following three ad hoc assumptions: First, the (y)S-selected DP thyinocytes in TCRP-deficient mice are not genuine aP lineage cells. Second, the 76-dependent pathway Ieatling to D P thymocytes in TCRP-deficient mice is a very minor, physiologically irrelevant route, which is not available for tlie vast majority of T-cell precursors. Third, the determining role of tlie TCR isotype for the aPly8 lineage commitment decision is not absolute and can be overruled in some (rare?) situations by other extracellular or cell-autonomous factors.
C. REVISEDSEPARATE LINEAGE MODEL Data obtained in normal and gene knockout mice can also be incorporated into a separate lineage model, as depicted in Fig. 2B. A bipotential precursor commits to the a/3 or yS lineage independent of tlie status of TCR rearrangements, but influenced by patterns of ‘Notch’expression and by other, as yet unidentified, factors and/or cellular interactions. Commitment to the y6 lineage results in an immediate shutdown of pTa expression or, alternatively, in a silencing of the pre-TCR signaling pathway by some other mechanism. Failure to generate a 76 TCR will lead to cell death, whereas expression of a functional y6 TCR will allow the cell to unfold its developmental program and to differentiate along the 76 pathway, albeit without extensive proliferation. Coincidental expression of a functional TCRP chain will not have the same strong effects as in cells committed to the crP lineage, as the formation and/or function of the pre-TCR is coinpromised after commitment to the 7 6 lineage. However, the presence of a functional TCRP chain may give rise to some weak pre-TCR-independent signals, as described earlier, which in turn may allow such cells to proliferate weakly or to survive somewhat longer than their TCRP-negative neighbors, giving them a greater chance to achieve productive y and 6 rearrangements before programmed cell death occurs. Expression of a functional TCRP chain in cells that have committed to the CXPlineage, however, will result in formation of a pre-TCR, inhibition of further rearrangements at TCRy, 6, and @ loci, massive proliferation, and preprogramnied differentiation along the c.P pathway. Inappropriate expression of a y6 TCR in @-committed precursors will also lead to inhibition of further rearrangements and possibly some developmental progression, however, and most importantly, not to extensive cellular proliferation. Expression of a y8 TCR in the a0 lineage is predicted to result eventually in cell death caused by one of several reasons. For instance, cominitinent to tlie ap lineage is most likely associated with a shutdown of tlie transcriptional activity of TCRy and/or TCRG genes and excision of the 6 loci (at a stage when the P locus can no longer be reactivated), depriving developing thyrnocytes of their receptor. Alternatively, or in addition, maturation beyond the DP stage may be blocked, as a y s TCR does not allow positive selection. The distinctive feature of this modified version of the separate lineage model is the assumption that the observed underrepresentation of productive TCRy and 6 rearrangements in CYPlineage cells is not due to a diversion of uncommitted precursors into the y8 lineage, as suggested by the rearrangement model. Rather, the paucity of in-frame yl6 rearrangements in a/3 lineage cells is regarded as the combined result of two effects: (1) elimination of ap-committed cells with in-frame y and 6 gene re-
arrangements, because they are developmentally arrested at the DP stage in the absence an aPTCR; and ( 2 )selective, pre-TCR-mediated expansion of cells with a functional TCRP chain lacking in-frame y/S rearrangements.
D. INTEGRATION OF CONFLICTING: DATA The uncontested finding that productive y and 6 rearrangements are underrepresented in ab lineage cells (suinmary: point 4)strongly supports a model that assumes that the oiitcoine of TCR rearrangements influences the lineage decision. However, the equally uncontested findmg of productive rearrangements at TCRP loci in y6 cells is difficult to incorporate into such a model because it demonstrates that expression of a functional TCRP chain is not sufficient to direct cells into the c.P pathway (summary: point 9). One way to reconcile these apparently contradicting findings is suggested by studies focusing on B-cell development, which have shown that not all functional Ig heavy chains are able to form a pre-B-cell receptor, obviously because of pairing problems with AS (the B-cell analogue of pTa) and/or VpreB chains (ten Boekel et a l , 1998). Whether functionally rearranged TCRP genes in y6 cells encode P polypeptides with similar deficits has not yet been tested and thus remains an interesting possibility. Another way to account for the presence of productive TCRP rearrangements in y6 lineage cells without discarding the competitive rearrangement model would be to amend this model with two ad hoc assumptions: (1)all functional P rearrangements observed in 76 cells have been generated after the formation of a yS TCR, implying that expression of a 76 receptor does not always block further rearrangements at the TCRP lociis; and (2) forination of a functional TCRP chain in cells already expressing a y6 TCR cannot override the lineage decision brought about by earlier signals from the 76 TCR. In fact, one may want to extend this second assumption by postulating that formation of a functional P chain in cells already committed to the y6 lineage also fails to induce massive cellular proliferation or any other effects usually associated with the formation of a pre-TCR. This extension seeins necessary to account for the low frequency of y6expressing cells compared with a p lineage cells in normal mice, and for the fact that the absence of a pre-TCR does not result in a fiirther reduction ofy6expressing cells, but instead in a significant increase (summary: point 11).Unless future studies demonstrate a selective, inherent deficit of y6 cell-specific TCRP chains in pre-TCR formation, the most straightforward explanation for the proposed failure of a TCRP chain to induce pre-TCR-mediated effects seeins to be the absence of pTa in y6committed cells. The most serious threat for the competitive rearrangement model, as depicted in Fig. 2A, arises from the finchng that DP thyinocytes in
62
HANS JOHC FEHLING ct ol
TCRP-I- mice are selected for in-frame TCRy and 6 rearrangements (summary: point 10).The corollary that expression of a yS TCR can stimulate precursor cells to differentiate along the a/3lineage completely contradicts the competitive rearrangement model. To rescue this model one would need to show that the yS-driven pathway to the DP stage is a peciiliarity of TCRP-deficient mice with little physiological relevance. The key question is therefore whether this pathway is available to all T-lineagecommitted precursor cells in a normal thymus or whether it is confined to some rare, atypical cells that may dominate in TCRP-deficient mice. The competitive rearrangement model could also be rescued by arguing that TCRP-’- DP thymocytes with functional ySTCRs are not “genuine” aP lineage cells. Rather, they may represent cells of the yS lineage that express CD4 and CD8 and downregulate CD25 (see also Section I,B,4). It is not easy to rigorously exclude this possibility. The demonstration of Va-+Ja rearrangements in DP thymocytes of TCRP? mice seems to address this issue (Mombaerts et al., 1992; Passoni et al., 1997), but it is hard to prove that these rearrangements occur in the same cells that have undergone (y)6 selection, but not in unselected bystander cells. In an attempt to reconcile experimental data with the competitive rearrangement model, the “codeterminant model” of lineage commitment has been proposed (Passoni et al., 1997). According to this model, fate is determined predominantly, but not exclusively, by the isotype of the TCR. Other factors in the extra- and intracellular environment, for instance, cytokines, cell-cell contacts, or “Notch,” can overrule decisions based on a particular TCR isotype. This adjustment of the competitive rearrangement model would account for the occasional appearance of ap or YC? lineage cells with a “wrong” receptor. However, unless specific “codeterminants” are identified and molecularly defined, this model appears somewhat abstract, especially in light of the fact that all experimental data derived from normal and gene-targeted mice can equally well be accommodated by the “separate lineage” model as presented in Fig. 2B, without postulating a determining role of the TCR isotype in the lineage decision. Which of the models is in best agreement with the findings in TCR transgenic mice? Data summarized under point 7 directly support the separate lineage model and the assumption that expression of a y6 TCR in a@lineage cells will eventually lead to a block of further a/3 lineagespecific development. Conversely, the same data are essentially incompatible with any simple version of the rearrangement model. The separate lineage model is also strongly supported by the finding that a pre-TCR is not obligatory for the formation of aP lineage cells (summary: point 9). Interestingly, both models presented here make similar predictions with regard to the phenotype of TCRy6 transgenic mice, in that early expression
aply8 IJh’EAGE C O M M I T M E N T
63
of a transgenic y6 TCR will obstruct the development along the aP pathway, albeit for different reasons and at different developmental stages (see earlier discussion). The block of (YPT-cell development in some G8 transgenic mice (Dent et nl., 1990)and the partial block observed in seven lines of TCRy6-transgenic mice described by Siin et nl. (1995)are therefore not in disagreement with the separate lineage model. Rather, the question is why does ~$3 T-cell development proceed at all in some TCRyS transgenic lines? Too weak, delayed, or variegated expression of one of the two transgenes in the particular transgenic lines is a possible and commonly presented answer. To avoid such problems and rigorously assess the effect of functionally rearranged TCR transgenes on the aPIy6 lineage decision, it may be necessary to target gene fragments encoding rearranged Tcell receptor chains directly into the respective endogenous TCR loci via homologous recombination in ES cells, which should prevent variation in the phenotype due to transgene-specific effects. While still being technically demanding, such experiments are clearly feasible and promise more reproducible results. Both models presented in Fig. 2 make a number of predictions that are at present not easy to verify experimentally. A major obstacle in studying the mechanism(s) of the a/I/yG lineage split is the absence of lineagespecific markers other than the TCR isotypes themselves. If such markers were available, one could, for instance, assess whether a lineage split exists before and irrespective of TCR rearrangements in rearrangement-deficient mice, which could yield direct proof for the separate lineage model. Conversely, monitoring the effect of targeted TCR transgenes on the expression of such putative lineage-specific markers should providc a conclii+c. answer to the question of whether expression of a specific TCR isotype gives rise to cells of the corresponding lineage and how universal this effect would be. Modern methods of subtractive screening may soon lead to the identification of genes that are expressed in an aP- or 76-specific fashion. Some of these dlfferentially expressed genes might be expected to encode transcription factors directly involved in the in-rplementationof the lineage decision, which would render such a search particularly rewarding. Whether pTa can be used as a reliable molecular marker to distinguish aj3- from y6committed thyinocytes in the absence of TCR expression should soon be known.
ACKNO\VI,EDC~.IENTS Rhodri Ceredig thanks INSERM for support during his leave of absence at the Basel Institute for Immunology. We thank Hans-Reimer Rodewdd ( Basel) for inany helpful coniinents and suggestions. The Basel Institute for Immuno~ogywas founded and is supported by F. Hoffniann La-Roche, Basel.
64
H A N S J O R G FEHLING et 01
REFERENCES Allison, J. P. (1993). y8 T-cell devekpnent. Crin-. Opin. Zrttnwno!. 5, 241-246. Allison, J. P., and Lanier, L. I,. (1987).Structure, function, and serology ofthe T-cell antigen receptor complex. Atmu. R P ~Znintunol. . 5, 503-540. Anderson, G.. Anderson, K. L., Tchilian, E. Z., Owen, J. J. T., and Jenkinson, E. J. (1997). Fil~roblastdependency during early thymocyte development maps to the CD25' CD44t stage and involves interactions with fibroblast matrix molecules. Eur. ]. Znmrztnol. 27, 1200-1206. Antica, M., Wn, L., Shortman, K., and Scollay, R. (1993). Intrathymic lyinphoid precursor cells during fetal thymus development. /. Ztritrtttnol. 151, 5887-5895. Arase, H., Arase. N., Nakagawa, K.. Good, R. A,, and Onoe, K. (1993). NK1.1tCD4tCD8thymocytes with specific lyinpliokine secretion. Eur. J. Ztrtn~iinol.23, 307-310. Ardaivin, C., Wu, L., Li, C.-L., and Shortinan, K. (1993).Thymic dendritic cells and T cells develop siinnltaneously in the thymus from a coininon precursor population. Nature 362, 761-763. Artavanis-Tsak(trias,S., Matsiino, K., and Foitini. M. E. (1995). Notch signaling. Science 268,225-232. Asamow. D. M., Cado, D., and Raulet, D. H. (1993). Selection is not required to produce invnriant T-cell receptor y-gene junctional sequences. Nature 362, 158-160. 13attistini. L., Horsellino, G . . Sawicki, G.. Poccia, F., Salvetti, M., Ristori, G., and Brosnan, C. F. (1997). Phenotypic and cytokine analysis of human peripheral blood y6 T cells expressing N K cell recrptors. /. Zttitntitiol. 159, 3723-3730. Bendelac, A., Rivera, M. N., Park, S. H., and Roark, J. H. (1997). Mouse CDl-specific N K I T cells: development, specificity, and fnnction. A;UUJ.Rev. Zttttnutd 15, 535-562. Blomberg, H., kind Andersson, B. (1971). Characterisation of the immnnoconipetent cells in the niouse thymus: Cell population changes during cortisone-induced atrophy and subsequent regeneration. Cell. Z i t m i i r t o L . 1, 545-560. Boesteaiiu, A., Silva, A., Nakajima, €I., Leonard. W. J., Peschon, J. J.. and Joyce, S. (1997). Distinct roles for signals relayed through the common cytokine receptor y chain and interleukin 7 receptor a chain in natural T cell development.]. Exp. Med. 186, 331-336. Boismenn, R., and Havran, W. L. (1995). T-cell clevelopment: T-cell lineage commitment reLisitrd. Ctrrr. B i d 5, 829-831. Boisinenu. R., and Havrm, W. I,. (1997). An innate view of y8T cells. Cum. Opin. Ittttmrtiol. 9, 57-63. Bonnevilk, M., Ishida, I., Moml)aerts, P., Katsuki, M.,Verbeek, S., Bems, A,, andTonegawa, S. (1989). Blockage of aP T-cell development by TCR y 6 transgeaes. Nature 342, 931-934. Brawand, P.. Biasi, G., Horvath, C., Cerottini, J . C., and MacDonald, H. R. (1998). Flowinicrofluorometricnionitoring of oligoclonal CD8' T cell responses to an immunodominant Moloney leukemia virus-encoded epitope it1 oiuo. /. Ztrimmol. 160, 1659-1665. Brenner, M. B., McLean, J., Dialynas. D. P., Strominger, J. L., Smith, J. A,, Owen, F. L., Seidniiin, J. G . , Ip, S., Rosen, F., and Krangel, M. S. (1986). Identification of a putative second T-cell receptor. Nature 322, 145-149. Bruno, L., Fehling, €1. J., and von Boehmer, H.(1996).The crp T cell receptor can replace the y6 receptor in the tkvehpinent of yS lineage cells. Ziriititinity 5, 343-352. Budtl, R. C., Miescher, G. C., Howe, R. C., Lees, R. K., Bron, C., and MacDonald, H. R. (1987). Developmentally regulated expression of T cell receptor P chain variable domains in immature thymocytes. ]. Exp. Med. 166, 577-582. Buer, J., Aifantis. I., DiSanto, J . P., Fehling, H. J.. and von Boehmer, H. (1997). Role of different T cell receptors in the development of pre-T ce1ls.J. Exp. Med. 185, 1541-1547.
Burtnini. D. B., Kim, S., Dudley, E. C . , Hayday, ‘4.C.. and Petrie, H. T. (1996). TCR gene recombination and a@-ySlineage divrrgence: Productive TCR-@ rearrangenlent is neither exclusive nor preclusive of y6 cell devehpmeut. I. I t i n n i c n o ! . 157, 4293-4296. Candrias. S., Mueggc. K., and Dnruni, S. K. (1997).IL-7 receptor and VDJ recombination: Tropliic versus meclranistic actions. Itrmiutiify 6, 501-508. Cantor, H., antl Boyse, E. A. ( 1975). Functional subclasses of T-lymphocytes 1)eariiigdifferent Ly antigens. I. The generation of finictionally dIstinct T-cell subclasses is a diffcrentiative process independent of antigen. I . Exp. A4rtl. 141, ~ 3 7 6 -1389. Cantor. H., and Boyse. E. A. (1977). Lyniphocytes a s models for the study of ~namnialian cellrilar difkwntiation. Ztnnurno/. Z h . 33, 105- 124. Cao. X., Shores, E. W., Hu Li, J., Anver, M. R., Kelsall, B. L., Russell, S. M.. Drago, J.. Noguclii, M., Crinherg, A., Blooni, E. T., et 01. (1995). Defective lymphoid development in mice lacking expression of the coinnion cytohne receptor y chain. Ztrmzrnity 2,223-238, Capone, M., Cuniow, J.. Bouvier. G., Ferrier, P., and Horvat, B. (1995).T cell development in TCR a@transgenic mice: Andysis iising \’(D)J recombination suhstrates. J . Zrtiniunol. 154, 5165-5172. Capone, M.. Wattin. F.. Fernex, C., Iloivat, B., Krippl. B., ~VII.L., Scollay, R . , and Ferrier, P. (1993). T C R @ and TCR a gene enliaricers confer tissue- antl stage-specificity on V(D)] recombination events. EMBO /. 12, 4335-4346. Carding, S. R.. Allan. W., Kyes, S., IIavday, A . , Bottondy, K., and Doherty, P. C. (1990). Late dominance of tlre inHarnmatory process in niiirine inflnenza by y/S+ T cells. 1.Exp. M d . 172, 1225-1231. Carena, I., Shamshiev, A,. Donda,A,, (:olonna, M., and Libero, ingeffects iipm traiisgcmic 6-tleletillg , 186, 91-100. elenients. J . E . Y ~hlcd. Joterpmi, F., Heuze, F., S;ilonion-Vie, \'., and Giscari, II. ( 1987). Cell kiwtics in the fetal inonse thyniris: Precursor cell input, proliferation, and eniignition. J . Z t t i t t t t r t i o l . 138, 1026- 1030. Jotereau, F. \I.. antl Le Douarin. N . M. (1982). Drnionstration of ii cyclic renewul of t h Iyinplioc>?e precnrsor cells i n the qiiail tllymlis duriiig enil)r).onic antl perinatal life. /. ~ t t t t t t u t i o l . 129, 1869-1877. M'oods, A. S.., Yewtlell, J. \V.. Bcnnink. J. H., Dcs Silva, A. D., Boesteanu. A., , P., Cotter, R. J.. arid Bi-iitkie\vicz,I{. R. (1998). Natural ligand of iiioiise CDldl: r glyccisylpliospliatitlylinositol. Scioicc~279, 1.541- 1541. King. J., Baker. J.. d Raiikt. I). H. ( 1 W S ) . Evitleiicts that protluctivcs rearrangrnicwts of TCR y genes infliieiicc. the conlniitment of progcwitor cc~llsto diffi.rentiilte into ap o r yS T cells. Eirr. J . f t t t t t t i i t i o l . 25, 2706-2709. Kang, 1.. and Karilet, D. 13. ( 1997). E\wits that rcAgrilutc, differentiation of ap TcH' and y 6 TcR' T cells frorii a coniiiioii p r c w n w r . S m i t t . r t w i i t t w ~ 9, . 171-179 Kelly. K , A,, Pearsr, hl., Lc>francois,I,.. and Scollay, 1%. (1993). Einigtxtiol S the addt I I I O I I S ~thynins. I t i f . r t t i t i i i m d . 5, :3O1sthsets of y 6 T C C ~ froin Kersh, G,J., l3oosliin;iild. F. F., and Iletkick, S. M. (1995). Efficieiit niatu l i n e q tliymocytes t o the CD4' CD8' stage in the ahsencr of TCK-@ rc,arrangeiiiriit. 1. r t t ~ t t ~ t t t ~ 154, ~ ) ! . ~706-5711. K e p i . U.. Beck-Engeser. G. B.. Jongstra, J.. Applvcpiist. S. E.. and Jack, H. M . (1YOij). light chain-tlepenclent selectioii o f Ig lieit\,? chain V rcsgions. J . Z t t r t t i t r r r o l .
70
HANS JOHC: FEHLINC: el cil.
Kisielow, P., Blnthniann, II., Staerz, U. D., Steinmetz, M., and von Boehmer, H. (1988). Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4'8' thymocytes. Nnttire 333, 742-746. Kisielow, P., Hirst, J. A,, Shiku, H., Beverley, P., Hoffman, M. K., Boyse, E. A,, and Oettgen, H. F. (1975). Ly antigens as markers for functionally distinct subpopulations of thymus derived lymphocytes of the mouse. Nature 253, 219-221. Kisielow, P., and von Boehmer, H. (1995). Development and selection of T cells: Facts and puzzles. Ado. Inirr~urud.58, 87-209. Koop, B. F., Wilson, R. K., Wang, K., Veniooij, B., Zallwer, D., Kuo. C. L., Seto, D., Toda, M., and Hood, L. (1992).Organization, stnicture, and function of95 kh of DNA spanning the inurine T-cell receptor C d C S region. Geiioinics 13, 1209-1230. Kotzin, B. L., Barr, V. L., and Palmer, E. (1985). A large deletion within the T-cell receptor p-chain gene complex in New Zealand white mice. Science 229, 167-171. Koyasu, S., Clayton, L. K., Heiken, H., Parkes, A., and Reinhertz, E. L. (1997). PreTCR signaling components trigger transcriptional activation of a rearranged TCR a gene locris and silencing of pre-TCR a locus: Implications for intrathymic differentiation. Int. ItnnmTICJ~. 9, 1475-1480. Kranz, D. M., Saito, H., Heller, M., Takagaki, Y., Haas, W.. Eisen, H. N., and Tonegawa, S. (1985). Limited diversity of the rearranged T-cell y gene. Nature 313, 752-755. Krotkova, A,, von Boehmer, H.. and Fehling, H. J. (1997).Allelic exclusion in pTa-deficient mice: No evidence for cell surface expression of two T cell receptor (TCR)-P chains, but less efficient inhibition of endogeneous Vp+( D)Jp rearrangements in the presence of a functional TCR-/3 transgene. /. Exp. Med 186, 767-775. Lalli, E., Sassone-Corsi, P., and Ceredig, R. (19%). Block of T lymphocyte dfferentiation by activation of the CAMP-dependentsignal transduction pathway. EMBO J . 15,528-537. Lauzurica, P.. and Krangel, M. S. (1994). Temporal and lineage-specific control of T cell receptor dS gene rearrangement by T cell receptor a and 6 enhancers. J . Ex?. Med. 179, 1913-1921. Leclercy, G., De Smedt, M., and Plum, J. (l9ga5).Cytokine dependence of Vy3 thymocytes: Matiire but not immature Vy3 cells require eiidogenous IL-2 and IL-7 to survive-evidence for cytokine redundancy. Int. I i n i m m d , 7, 84:3-851. Leclercq, G.. Plum, J., Nandi, D., De Smedt, M., and Allison, J. P. (1993). Intrathymic differentiation of Vy3 T cells. /. E x p . Med. 178, 309-315. Lefrancois, L., and Goodman, T. (1989). I n uioo modulation of cytolytic activity and Thy1 expression in TCR-yS' intraepithelial lymphocytes. Science 243, 1716-1718. Levek C. N., Ehrfeld, A., and Eichmann, K. (1993).Regulation of thytnocyte development through CD3. I. Timepoint of ligation of CD:3&determines clonal deletion or induction of developmental program. J . Exp. Merl. 177, 707-716. Li, H.. Lebedeva. M. I., Llera, A. S., Fields, B. A., Brenner, M . B., and Mariuzza, R. A. (1998). Structiire of the VS domain of a human y6 T-cell antigen receptor. Nature 391,502-506. Li, S., and Wilkinson, M. F. (1998). Nonseiise surveillance in lymphocytes? Immunity 8, 135-141. Livik, F., Petrie, 11. T., Crispe, I . N.. and Schatz, D. G. (1995). In-frame TCR S gene rearrangements play a critical role in the a,f3/y6 T cell lineage decision. Zmmi~nity2, 617-627. Livik, F., Wilsoii, A., MacDonald, H. R., and Schatz, D. G. (1997). Cup lineage-committed thymocytes can be rescued by the y6 T cell receptor (TCR) in the absence of the TCRP chain. Eiir. J. I ? 7 1 i n U i l O ~ .27, 2948-2958.
Liiffert. D., Ehlicli. A.. Miiller, LV..i t n d Kajewsky. K. ( 15196).Surrogate light chain expression is rcqnired to estal)lish iiiiiiiiiiioe;lot)iiliii hea~?.chain allrlic excliision diiring early R cell developnieirt. Zttttttiitiit!/ 4, 1 3 - 1-14. Ixwc*nthal. J. \V.. Howe, K . C . , Cercdig, R.. iintl hl;icDonald. 11. I70% in adult peripheral blood (Parker et nl., 1990). In contrast, cells expressing Vy9/ V a l are more frequent very early in development and continually decrease thereafter. More than differential Vy gene expression, the sequential appearance of particular Vy-Jy or VS-DS-JS combinations suggests that, as in the mouqe, human y6 T cells arise in a developmentally ordered fashion (Vietor and Koning, 1990). In addition, both clones and entire subsets of 76 T cells may peripherally expand driven by antigenic stimula1991,1992). tion (De Liberoet al., 1991;Parkeret al., 1990; Uyeinuraet d., Despite similarities in the development of peripheral y6 T-cell populations, enormous differences exist between species regarding the tissue distributions and overall sizes of y6 T-cell subsets (reviewed in Born et n l , 1994). For example, y6 T cells are comparable in numbers to a@T cells in sheep, cattle, pigs, and chickens, despite being rather infrequent in primates and rodents. The biological significance of these differences is not yet understood. If indeed y6 T-cell subsets are functionally specialized (see later), species may vary in their requirements for these functions, e.g., because of their life-styles (Hein and Mackay, 1991) or due to organismic structural properties. It may thus be helpful to consider y6 T-cell subsets within an organism as separate entities and to attempt to identify
and assess their individual functions as an alternative to comparing y6 and ap T cells in a more global way. 111. Specificity
A. IMPI,IC:ATIONS o~ y6 TCR STH~JCTLJHE
The y6 TCR is expressed as a heterodimeric cell surface molecule (Haas et nl., 1993; Raulet, 1989). y and 6 chains each consist of one variable and one constant domain. Cell surface expression of the heterodimer and signaling through the y6 TCR require association with tlie CD3 complex of transmembrane proteins. Thus, tlie overall structure of the y6 TCK complex resembles that of the ap TCR complex far more than that of iiniiiunoglobulins,and the similarity between the two types of TCRs initially suggested that ligands recognized by the two types of T cells may also be similar. A more detailed analysis of y6 TCR structure has revealed further similarities but also substantial differences with a0 TCRs. The genomic organization of the genes encoding TCR y and 6 chains has been described in detail elsewhere (Arden et , 1995; Chien et al., 198%; Clark et nl., 1996; Hayday et al., 1985; Iwasliima et d.,1988; Lefranc et nl., 1989; Lefranc and Rabbits, 1989; Lefranc and Rabbitts, 1985, 1990; Ratlibun et nl., 1988; Raulet, 1989; Zhang ct nl., 1994). Briefly, TCR y genes are clustered at a distinct lociis with considerable variations in genomic organization between the different species. There are only a few Vy genes, each of which typically combines through gene rearrangement only with its most proximal J segment. There are no Dy segments. The V, J, and C gene segments form small clusters. In mice, for example, there are four such clusters and seven Vy-J-C combinations possible within these clusters. Vy-J-C combinations between clusters do occur but are rare. Moreover, the TCR y repertoire is additionally limitcd as junctional variations in rearranged genes are not extensive, and in productive rearrangements involving Vy5 and VyG, are typically absent. These two invariant TCR y chains in mice each form heterodimeric receptors with the same, equally invariant, TCR 6 chain. This lack of diversity in a portion of the y6 TCRs iinplies that ligands recognized by the y6 T-cell populations carrylng these receptors are invariant as well. TCRS and a gene loci are interspersed such that productive TCR a gene rearrangements eliminate most or all gene segments of the TCR 6 cluster, wlierea? TCR 6 rearrangements per se do not prevent subsequent TCR a rearrangements (Cliien ut a/., 1987a,b; Porcelli et al., 1991). This same peculiar genomic organization persists in distant species (e.g., mice and humans), suggestive of evolutionary conservation and functional impor-
84
WI1,L.I BORN rt nl
tance. An obvious consequence of this arrangement is that y6 T cells are far less likely to express TCR a chains, and, inversely, that a@T cells are prevented from expressing TCR S. There are far fewer V6 genes than V a genes (e.g., approximately 10 times fewer in mice.) The V6 genes are interspersed with V a genes, and some are closely related to or even identical with V a genes. Assuming that y6 TCRs make contact with their ligands in a fashion similar to immunoglobulins or a/3 TCRs (Engel and Hedrick, 1988; Jorgensen et al., 1992), i.e., via the three CDR-equivalent loops present in both y and S TCR chains, yS TCRs containing V6's that are also Va's should have similar or identical specificities for two of the six predicted points of ligand contact. The similarities in CDRl and CDR2 equivalents could indicate a structural basis for a similar ligand bias of y6 T cells and a@T cells, e.g., favoring recognition of MHC class I/IIlike molecules. Among the CDR loops, CDRS is of particular interest. For immunoglobulins, X-ray crystallography has shown that CDRS loops of both heavy and light chains are typically involved in antigen contact (Engel and Hedrick, 1988;Jorgensen et al., 1992). In the case of a/3TCRs recognizing peptide/ MHC complexes, the CDRS loops tend to make contact with the bound peptides. Davis and Bjorkman (1988) estimated that the potential number of different y6 TCRs in mice actually exceeds that of a/3TCRs or immunoglobulins. This is primarily due to an enormous potential for junctional diversity within the TCR 6 gene locus. Assuming that CDR3 is also critical for antigen recognition in y6 T cells, Rock et nl. (1994) cornpared CDRS length distributions of the various types of antigen receptors in mice and humans. CDRS lengths were defined as the distance from the J regionencoded Gly-X-Gly motif to the nearest preceding V region-encoded Cys, minus four amino acids. Among transcripts of IgH, -L(Kand A) and TCR a, @, 7 , and 6, CDR3s of IgH and TCR 6 were indeed the longest and most variable in size. In contrast, CDRS size variations of TCR a and @ are much more constrained, with almost indentical CDR3 lengths. In Ig heavy chains, CDR3 loops are often extensive with a wide range of lengths, whereas CDR3 loops of light chains tend to be much shorter with little variations in length. It is unclear whether this difference has any particular significance, but the great variability of heavy chain junctions coincides with the ability of antibodies to recognize structurally diverse antigens. Much in contrast, CDR3 loops of TCR a and p chains are typically similar in lengths. This likely reflects the fact that the size of the ligands for a/3 T cells, which are small peptides nestled within a peptidebinding groove, is almost invariant. Moreover, it suggests that that CDRS loops of TCR a and /3 chains play similar roles, consistent with evidence that both can make contact with the bound peptide fragments (Fremont
Ih.IMlII\’OHEGUI,4TORY FIINCTIOVS OF y6 T CELLS
85
et nl., 1996). Chien and cohhorators (1996) have pointed out that tlie heterogeneity of CDR3 loops in y6 TCRs is more consistent with ligand recognition in a manner similar to that of iiniiiunoglobulins than of cr,P TCRs. This new concept is further supported by their observations of the distinctive ligand requirements of two MHC-reactive y6 T-cell clones (see later).
B. SELECTION OF y6 T CELI.S Most peripheral crp T cells have gone through a screening process that promotes the development of self-MHC-restricted ligand specificities, while preventing the survival of potentially autoaggressive clones. This balancing act is accomplished through positive and negative selection mechanisms that favor cells with relatively low but distinct affinities for self-MHC class 1/11 molecules, but having the potential of much higher affinities for self-MHC-antigen complexes (Robey and Fowlkes, 1994). Because the CDR3 regions of the a@TCR are primarily involved in ligand binding (Engel and Hedrick, 1988; Jorgensen et al., 1992), they are also the primary targets of selection. Nevertheless, the survival of any given cell may also depend on the TCR V genes or even on the accessory molecules it expresses. Because CDR3 sequences of y6 T-cell subsets tend to be nonrandom, it made sense to postulate that the peripheral repertoire of y6 TCRs is shaped by selection as well. However, the generation of particular junctional sequences could also be driven by genetic mechanisms and be entirely independent of TCR-ligand interactions. In this regard, the invariant junctional sequences of the murine VyS/V61 and Vy6NG1-positive subsets have been examined in some detail. Normal development of the Vy5’ subset requires a fetal thymus environment, consistent with the possibility of thymic selection (Ikuta et nl., 1990). Moreover, whereas productive rearrangements of Vy5, Vy6, and V61 genes in fetal thymocytes are essentially invariant,junctional diversity has been seen in nonproductive rearrangements that are not subject to selectional forces ( Itoliara and Tonegawa, 1990; Lafaille et nl., 1989). Finally, when thymic expression of the y6 TCR was prevented by modulation with anti-TCR mAbs, frequencies of productive rearrangements of tlie same genes with noncanonical sequences increased (Lafaille et d . ,1990).These findings are all consistent with thymic selection of the two invariant y6 T-cell subsets in mice. However, other findings indicate that thymic selection is not necessary in the generation of tlie invariant ~6 TCRs. In mice into which nonfunctional TCR y gene substrates were introduced as transgenes, canonical TCR Vy5-Jyl and Vy6-Jyl junctions were generated at high frequencies, even though there was no possibility of selection for certain surface-expressed
86
\VILLI BORN ef d
protein products (Asamow ct al., 1993). Similarly, in mice lacking TCR 6, Vy5-Jyl and Vy6-Jyl canonical junctions are generated in normal frequencies (Itohara et nl., 1993). More extensive experimentation with transgenic mice has emphasized the importance of short homology repeats near these gene junctions and a lack of terminal deoxynucleotidyl transferase (TdT) in the generation of the invariant y6 TCRs (Zhang et al., 1995). Specifically, transgenic recombination substrates revealed that di- and trinucleotide repeats in the coding regions and in P elements have strong effects on the site of recombination. In addition, forced expression of TdT at early developmental stages decreased the frequencies of canonical junctions while increasing the frequency of in-frame noncanonicaljunctions containing N nucleotides. These data seem to indicate that early in development, a directional mechanism of rearrangement, aided by the absence of TdT activity, gves rise to the invariant y6 TCRs in mice (Allison and Havran, 1991; Raulet et al., 1991). Bias in the expressed TCR repertoire has also been noticed among other inurine y6 T-cell subsets. For example, differences between mouse strains in the proportions of V&+ cells among splenic and intestinal yS populations have been reported (LefranGois et al., 1990; Sperling et al., 1992). Among intestinal y6 T-cell subsets, a dominant V&'" phenotype was found to be linked genetically to certain MHC class I1 alleles, expressed in the periphery but not in the thymus (LefranGois t>t al., 1990). In pulmonary y6 T cells of BALB/c mice (but not in BALB/b or C57BLlG mice), characteristic TCR junctional sequences (referred to as BID in rearrangements involving V65, and GxYS in rearrangements involving Vy4) are present in extraordinarily high frequencies (Sim, 1995; Sim and Augustin, 1991a,b).The same junctional sequences were present in mice carrying the nude mutation on a BALB background, suggesting a relatively thymus-independent mechanism. Moreover, although C57BW6 mice lack BID or GxYS sequences in the periphery, they are able to generate them because they could be found in the thymus. The presence of BID arid GxYS cannot be explained by an absence of TdT activity nor have genetic mechanisms been identified that could be responsible for these sequences. However, a correlation between the presence of an endogenous retrovirus and BID sequences was noted (Sim and Augustin, 1993). Therefore, BID and perhaps also GxYS may represent examples of extrathyinic junctional selection of y6 T cells. If y6 T cells are indeed selected, are they subject to similar mechanisms as those operational for a@T cells and, more specifically, are the selecting ligands similar? In mice lachng µglobulin (P,-M), CD8' c.0 T cells fail to develop, presumably because they are not positively selected in the absence of MHC class I expression (Zijlstra et al., 1990). However, the same mice showed normal distributions of y6 T cells in thymus, peripheral
lymphoid tissues, and intraepithelial locations (Koller et d., 1990; Zijlstra ct nl., 1990). Neither abnormalities in Vy or V6 gene usage were noted nor were y6 T cells dysfunctional when stimulated with anti-TCR mAbs. Similarly, normal nuiiibers of splenic y6 T cells were found in double mutant mice lacking both class I and class I1 MHC expression (Grusby et nl., 1993). Nevertheless, in inice expressing transgenic y6 TCRs with specificities for class Ib molecules encoded in the Tla region, PL-Mexpression was required for the normal development of transgene-expressing y6 T cells (Pereira ct n l , 1992; Wells et al., 1991, 1993). In the absence of P,-M, thymic transgene’ y6 T cells exhibited an immature phenotype (HSA’ ), and peripheral y6 T cells were reduced in numbers. Similarly, in mice expressing transgenic class Ib, changes in the peripheral y6 TCR repertoire have been noted. Thus, while the majority of y6 T cells in mice s e e m to develop normally in the absence of MHC class I/II-dependent selection, a minority requires these ligands for their development. This pattern fits well with the observation that in contrast to a0 T cells, specificities for allogeneic MHC among y6 T cells are rare (Bux et nl., 1985; O’Brien et nl., 1989). Nevertheless, the finding that some y6 T cells are selected in a MHC-dependent fashion probably indicates that y6 T cells as a whole are subject to selectional mechanisms, albeit for the most part not involving M HC-type molecules. Alternatively, ligand selection could be exceptional among 76 T cells. only occurring among those cells whose TCRs happen to have aP T cell-like specificities. y6 T cells that fit these ciiteria include a sinall population ofCD8‘ cells in inice and rats mediating tolerance to certain inhaled antigens and apparently endowed with MHCrestricted peptide antigen specificities (McMenainin ct nl , 1994).
C . y6 T CELLSARE STIMULATBL)H1’ M % N Y D I F F E R EMOLECULES ~T Because immunization with soluble antigens has largely failed to elicit antigen-specificy6 T cells in vivo (for possible exceptions, see the examples discussed under tolerance to ingested and inhaled antigens), y6 T cells liavc been screened in uitro with a variety of antigens in the hope of finding specificities by chance, withoiit prior sensitization in vivo This approach is reminiscent of early attempts to find antigen specificities of myeloma proteins, prior to the availability of hybiidomas. The screening of myeloma proteins in binding assays with large collections of various chemical compounds led to tlie discovery of niimerous hapten specificities ( Janeway and Travers, 1997). Similarly, random stimulation of y6 T cells, clones, and hybridomas with various antigens revealed numerous responses to hoth peptidic and nonpeptidic substances. However, the biological significance of these responses remains unclear at present.
Among partially defined antigens, heat-killed bacteria, bacterial extracts, mycobacterial purified protein derivative (PPD, a partially purified culture supernatant of mycobacteria such as M . tuberculosis H37Rv), low molecular weight protease-resistant components of mycobacterial extracts, and polyGT (pGT, a random copolymer of glutamic acid and tyrosine, molecular weight 20-50,000) were all found to stimulate murine and/or human yS T cell responses in vitro (Dembic and Vidovic, 1990; Holoshitz et al., 1989; Kabelitz et al., 1990; O’Brien et al., 1989; Panchamoorthy et al., 1991; Pfeffer et nl., 1990, 1992). With the exception of pGT, all have been reported to elicit polyclonal yet subset-specific reactivity, in this regard reminiscent of the superantigen responses of ap T cells (Herman et al., 1991). The authors’ unpublished data indicate that pGT also elicits polyclonal reactivity of yS T cells in vitro, in the absence of aj3 T cells and without requirement for in uivo priming. Subset specificity and, in some cases, TCR gene transfection indicate that these responses are indeed y6 TCR dependent, but it has remained unclear whether they involve direct binding interactions between the TCR molecules and the antigens. Molecularly defined soluble antigens have also been found to stimulate y6 T-cell responses in vitro. These include tetanus toxoid (Kozbor et al., 1989), mycobacterial 60-kDa heat shock protein (HSP-60) (Haregewoin et al., 1990; O’Brien et al., 1989) HSP-60-derived peptides (Born et al., 1990a; Fu et d . , 1994a), staphyloccocal enterotoxin A, (Rust et al., 1990) listeriolysin 0 (Guo et al., 1995), and lipopolysaccharides (LPS) (Reardon et al., 1995; Tsuji et al., 1996). More recently, nonpeptidic components, mostly derived from mycobacteria, have been isolated that were found to be stimulatory for human but not murine y6 T cells (Biirk et al., 1995; Constant et al., 1994; Schoel et al., 1994; Tanaka et al., 1994, 1995). Chemical characterization of these molecules has revealed that they are all of low molecular inass and contain phosphate groups. Otherwise, their structures are diverse, ranging from nucleotide derivatives to isoprenyls and sugars. As with the less defined antigens listed earlier, these substances all elicit polyclonal yet subset-specific and TCR-dependent responses of yS T cells. Nevertheless, the mechanisms underlying their stimulatory activities remain unresolved. Responses to phosphate-containing antigens have been studied most extensively. Some of these molecules elicit proliferative and cytokine responses at very low molar concentrations (Tanaka et al., 1995). Indirect stimulation of yS T cells can be excluded, as accessory cells are not strictly required to induce these responses. Stimulation of cell lines and clones occurs very rapidly, making it unlikely that the phosphate-antigens function by inducing the de not100expression of cellular ligands (Lang et nl., 1995). A requirement for cell contact has
been interpreted as evidence for antigen presentation (Morita ct al., 1995). However, professional antigen presenters are not required and there is no evidence for antigen processing. Although phosphate ligands could be antigens in the classical sense, it seems e q u J y possible that they function a s niolecular "adjuvants," enhancing preexisting TCR-clependent cellnlar interactions. The unique susceptibility of y6 T cells to phosphate antigens could be of considerable importance for the selective induction of y6 T cells in the course of an immune response to bacterial and perhaps other pathogens (Burk et al., 1997). There is ample evidence that y6 T cells can function as iminunoregiilators (see later), ancl the presence of phosphate antigens might help determine the flavor of their regulatory activities. However, in the absence of rodent responses to phosphate antigens, suitable model systems in which in vim effect\ could be tested remain to be identified.
D. MECHANISM OF LJGAND R I X O ( , N I T I OOF ~ MHC M ~ I . E ( ' U I , E - S P E ~ .y6 I F JT~ ,CEJ.LS In recent years, nnmerous y 6 T-cell clones have been isolated that respond specifically to classical and nonclassical MHC molecules (Bluestone et d., 1988: Bonnedle ct d., 198%; Houlden et d.,1989; Matis and Bluestone, 1991; Matis e t a / , 1987, 1989; Porcelli et al., 1989; Porcelli and Modlin, 1995; Kellahan et a / . , 1991). Most of these cells were selected by allo-antigen stimulation, but others were not, consistent with an inherent or selected bias for the recognition of MHC moleciiles among y6 T cells (Ito et d., 1990). For two MHC-reactive clones, the mechanism of ligand recognition has been analyzed in considerable detail. These were derived from BALB/c nidnu mice (H-2") stimulated with low-density B1O.BR (H2') spleen cells. One clone, LBKS, recognizes the mouse MHC class I1 inolecules I-EL'" (Matis et a l , 1989). LBK5 differs in V( D)J junctional sequences but not V, D, or J segments from another clone, LKDl , derived from a B 10.BR nioiise immunized with B10.D2 (H-2") cells arid specific for 1-A', suggesting that CDR3 determinants dictate antigen specificity for these clones (Rellahan et d., 1991). Schild and collaborators (1994) demonstrated in an elegant set of experiments that LBKS recognizes I-ELindependently of bound pepticles and that no conventional antigen processing pathway is required for the recognition of this molecule by y8 T cells. Even isolated I-EL protein immobilized on a plastic surface stimulated LBKS to an extent similar to that of cells expressing I-EL,thus eliminating the possibility that an additional molecule, such as a superantigen, could be involved in the stimulation. Moreover, epitope mapping using cells expressing mutated I-E molecules indicated that ap T cells ancl LBKS recognize different regions of
90
WILL1 BORU ct nl
I-Ek.Specifically, mutations in the a helices of I - E a and -P that affected a0 T-cell recognition did not alter LBKS stimulation, whereas a mutation at position 79 of I-Ea abolished the response of LBK5 while not interfering with the responses of any of the I-E-specific (YPT cells tested. In addition, the response of LBK5 was influenced by a polymorphic solvent-exposed residue at position 67 of I-EP and by the carbohydrate at a82 (Chien et al., 1996). Another clone termed G8, also derived from alloantigen-immunized BALB/c ndnu mice (Matis et al., 1987), recognizes both the products of the nonclassical class I genes T10 and T22 (Schild et al., 1994;Weintraub et al., 1994).Again, there was no indication of a requirement for conventional antigen-processing pathways in ligand recognition. G8 recognized T10/ T22 molecules expressed in Drosophila cells, which are considered to be incapable of any type of antigen processing or presentation, and it even responded to T10 expressed in Escherichia coli, after the recombinant molecules were immobilized on plastic. That peptide loading in fact is not important for the recognition of the MHC molecule by G8 was also suggested by the failure to elute peptides from the stimulatory T10 molecule and by the finding that T10 does not require peptides for cell surface expression (Chien et al., 1996). T22 is also recognized by another y6 T-cell clone, K N 6 (Ito et al., 1990). G8 and K N 6 express different y6 TCRs. KN6 was derived without alloantigen selection, suggesting that y6 T cells bearing this particular specificity occur more frequently. The T22 gene was also mutagenized, and transfectomas were tested for their ability to stimulate KN6 responses (Moriwaki et al., 1993). In this study, some mutations located at the floor of the predicted peptide-binding groove reduced KN6 reactivity. The response pattern was interpreted to suggest that bound peptides could play a role in ligand recognition, but the alternative possibility that the mutations altered T22 surface expression was not ruled out. In sum, the recognition of TL molecules by y6 T cells remains somewhat less well defined than that of MHC class I1 because possible bound ligands have not been ruled out as strictly and because aP T cells with specificities for TL are not as readily available for comparisons as those with specificities for MHC class I1 (Chien et al., 1996).
E. LICAND RECOGNITIONOF A HERPES SIMPLEX VIHUS-SPECIFIC y6 T-CELLCLONE Another murine y6 T-cell clone, Tg14.4, was found to respond to a herpes simplex virus type I transmexnbrane glycoprotein, gI ( Johnson et nl., 1992). Several observations indicated that gI is also recognized without processing or presentation (Sciammas et al., 1994): (1) anti-gI antibodies
1h.lMUNOKEC:~lLATORY FUNCTIONS OF 7 8 T CELLS
91
blocked the response of Tg14.4, (2) a mutated form of gI not expressed on the cell surface was not stimulatory, and (3) a form of gI expressed as a cell surface protein in the antigen-processing mutant RMA-s remained stirnulatoiy. The TCR of Tg14.4 is coinposed of rearranged V68 and Vy2 variable genes. Junctional sequences have been determined, but a inutational analysis defining requirenients on the TCR structure for the recognition of HSV gI has not yet been reported. Recognition of HSV and its product(s) by y6 T cells could be of considerable importance in host resistance (see later).
F. REQUIREMENTS FOR y6 T-CEI.LAC.TIVATION CARRY IMPI,ICATIONS FOR SPECIFICITY For the activation of a0 T cells through the TCR, multivalent ligands capa1)le of cross-linking the TCRs are required. Foreign peptide antigens are typically rendered polyvalent through their display on the surfice of antigen-presenting cells. Experimentally, a similar situation can be created using iiniiiobilized anti TCR antibodies. TCR cross-linking induces signaling through TCR-associated transmembrane molecules collectively referred to as the CD3 complex. Like a@ TCRs, y6 TCRs are associated with the CD3 complex and require it both for TCR surface expression and for signaling (Haas et nl., 1993). Moreover, like a0 T cells, y6 T cells are activated following TCR cross-linking. First, this has been shown with anti-TCR antibodies and later using antigenspecific 76 T-cell clones. Thus, using y6 T-cell clones recognizing MHC class I1 (I-EL),MHC class I (TlO/T22), and HSV g1 proteins, it was shown that soluble forins of these proteins are only stiinulatory when iininobilized (Chien et nl., 1996; Schild Pt al., 1994; Sciammas et nl., 1994). Interestingly, requirements for TCR cross-linking also seein to exist with substances found to elicit polyclonal y6 T cell responses in citro. Thus, stirnulatory bacterial extracts, randoni amino acid copolymers, and sinall peptides such as those derived from HSP-60 d l are inherently polyvalent, partially insoluble, or need to bc immobilized in order to elicit y6 Tcell responses (Deinbic and Vidovic, 1990; Fu et nl., 1994a, unpublished observations). Similarly, the inore recently discovered low molecular weight phosphate-containing compounds, which are capable of stimulating polyclonal responses of human y6 T cells but are soluble in free form, require cell contact, and thus probably a primitive form of presentation, in order to be stimulatory (Morita et nl., 1995). The requirement for TCR cross-linking in 76 T-cell activation seems to be an indication that y6 T-cell specificities are norinally trained on ligands expressed on the cell surface where they inherently have cross-linking properties, as opposed to soluble ligands. It is still not clear, however,
whether the ligands are primarily heterologous and complexed with autologous cell surfaces or whether they are autologous and perhaps indicators of activation, stress, or inflammation. There are also some indications that signal processing in y6 T cells differs from that in ap T cells. Thus, at least epidermal y6 T cells in mice may use FceRIy for signal transduction instead of CD35. Signaling has not yet been studies in great detail in 76 T cells. Possible implications of the presence of different signal transducers in the two types of T cells have been discussed elsewhere (Leclercq and Plum, 1996).
G. POTENTIAL USE OF sTCR CONSTRUCTS I N DETERMININC; y6 T-CELLLIGAND SPECIFICITIES Candidate ligands for y6 TCRs still need to be confirmed by measuring binding interactions. If yS TCR ligand binding is Ig like (Chien et al., 1996), there is hope that such studies will be facilitated by higher affinities than those of ap TCR ligand interactions, but this remains to be seen. For this type of experiment, soluble and perhaps polyvalent forms of y6 TCRs will be required, similar perhaps to the engineered multivalent ligands for a/3 TCRs (Altman et ul., 1996). Although y6 TCRs can be isolated directly from cells expressing them, conventional methods yield only sinall quantities of mostly denatured protein (Born et al., 1987). As demonstrated with crp TCRs, coinparatively large amounts can be generated using soluble TCR (sTCR) constructs (Fields et al., 1995). The first y6 sTCR reported was derived from a chimeric construct in which the extracellular domains of the mouse Vyl.1-Cy4 and Vy6.2-C6 TCR chains of y6 T-cell hybridoma T195/BW were fused to the hinge region, CH2 and CH3 domains of human IgGl heaLy chain, and transiently expressed in COS cells (Eilat et d., 1992). The chimeric proteins were produced intracellularly at rather high levels, the hvo protein chains formed disulfatelinked, glycosylated heterodimers, and correctly paired receptor chains were found secreted into the culture medium. In addition to confirming the identity of the chimeric secreted TCR yG-IgH heterodimer with Vyland TCR &specific antibodies, reactivity with an anticlonotypic inAb (F10/ 56) suggested that the fusion protein retained a conformation identical or at least similar to that of the native TCR. This chimeric TCR construct has been used in attempts to identifi. a putative autoantigen recognized by hybridoma T195/BW and similar V y l + cell lines (see later), thus far without success. In a CHO expression system, others have demonstrated efficient secretion of nonchimeric disulfidelinked human y6 TCR by introducing translational termination codons upstream from the sequences encoding TCR chain transmembrane regions (Davodeau et al., 1993a). Based on its reactivity with several anti-y and
-6 mAbs, tlie recoinbinant protein appeared to be folded correctly. It also proved to be immunogenic, dlowing the generation of mAbs capable of recognizing both soluble and ineiiibr;ine-l)oiind, native y6 TCRs. A high sensitivity of the interchain disulfide bridge to digestion with papaiii suggested that the sTCR C-terminal portions were in ii more extended configuration than tlie corresponding regions in iiiiinunoglobiilins. However, the soluble y6 heteroclimer remained stable after removal of tlie interchain disulfide link, suggestive of strong noncovalent forces capable of holding the two chains together. More recently, the V6 doniain of another huinan y6 TCR, derived from a clonc specific. for the HLA-A2 molecule (Palliard et nl., 1989), has been ciystallized (Lebedeva ct nl., 1996). Here, the V doniain of TCR 6 was expressed a s a secreted recombinant protein within the periplasinic space ofE. coli (Studieret al., 1990).It was then crystallized in a form suitable for X-ray diffraction analysis (orthorhoiiibic crystals, space group P21212 with unit cell diinensions (1 = 69.9, 11 = 49.0, c = 61.6 diffraction to beyond 2.3 resolution), but the final result of this analysis still awaits publication. Using a baculo\irus expression system, a soluble nonchiineric form of an iiivariaiit inurine y GTCR, initially identified in the tongue and reprocluctive tract but later alw found in several other tissues (Hayes et nl., 1996; Heybome ot nl., 1992; Itohara et d., 1990) and after bacterial infections (Mukasu ct al., 1997: Roark et nl., 199G),has also been generated (Roark, 1995). Reactivity with anti-TCR mAbs suggests that this soluble heterodiiner is also appropriately folded and therefore may be used to generate antibodies against this particular 76 TCR. The structural analysis of y6 TCRs still lags beliind that of ap TCRs, although experiences with ap T-cell-derived molecules likely will accelerate the characterization of y6 TCH heterodiiners. More iinportantly perhaps, as y6 T-cell functions are still poorly defined, crucial insights inay be gained from the structure of their antigen receptors.
A,
A
IV. Functions
A. DISTINCTIVE I N T E H A C T I O N S \VITll
Cl.,Yu, Q.-C., and Fuclis. E. 11993).Targeting expression of keratinocyte growth factor to keratinocytes elicits striking changes in epietlrelid differentiation in transgenic mice. EMBO J . 12, 97:3-986. Guo, Y.,Ziegkr, H. K.. SaHey. S. A,, Niesel, D. \%'.,\'aid!-a, S., and Klirnpel, G. R. (1995). Hninan T-cell recognition of Listcrin ttioiioc,!/togote.c: Recognition of listeriolysin 0 by TcRap' and TcRy6' T cells. Zrfcct. I n t i n u t i . 63, 2288-2294. Guy-setsof I)o\ine y6 T cells with unicpie ccll surface phrnotype and tissiie distribntion. Z t t i t i i i i r i d ~ ~ g92, ! ~ :340-:345. Maki. I,complex-linked specificity of y6 receptor-bearing T lyniphoe)tes. Notiire 330, 262-264. Matis, L. A,. Fry. A. M., Cron. It. Q., Cotterinan, M. M.. Dick. R. F., and Blnestone, J. A. (1989). Stnictiire and specificity of it c ' I1 alloreactive y6 T cell receptor heterodimer. Scimcr 245, 746-749. Mavadd;tt, N., Rohinson. B. W.. Rosc, A. I I . , Manning. I,. S.. and Garlepp, M. J. (19Ys3). An analysis of the rehtionship between ganinia delta T cell receptor \' gene usage and non-major Iristocornpatability coiiiplex-lcstrictetl cytotoxicity. Z t n t n u r i d . Ccll B i d . 71, 27-37.
136
WII.LI HORN rt rrl.
McMenamin, C., McKersey, M., Kiihnlein, P.. Hunig, T., and Holt, P. G. (1995). y6 T cells down-regulate primary IRE responses in rats to inhaled soluble protein antigens. J. Zmitiunol. 154, 4390-4394. McMenaniin, C., Oliver, J., Girn, B., Holt, B. J., Kees, U. R., Thomas, W. R., and Holt, P. G . (1991). Regulation of T-cell sensitization at epithelial surfaces in the respiratory tract: Suppression of IgE responses to inhaled antigens by CD3+ TcR a-//3-lymphocytes (putative ylS T cells). Zmvnunohg/ 74, 234-239. McMenaniin, C., Pimm, C., McKersey, M., and Holt, P. G. (1994). Regulation of IgE responses to inhaled antigen in mice by antigen-specific y6 T cells. Science 265, 18691871. Meeusen, E., Fox, A,, Brandon, M., and Lee, C.-S. (1993).Activation ofuterine intraepithelial y6 T cell receptor-positive lyinphocytes during pregnancy. Eur. J . Itnrnunol. 23, 11121117. Melanied, D., and Friedman, A. (1993). Direct evidence for anergy in T lymphocytes tolerized by oral administration orf ovalbumin. E m J. Imnarrnol. 23, 935-942. Mengel, J., Cardillo, F., Aroeira, L. S., Williams, O., Russo, M., and Vaz, N. M. (1995).Antiy6 T cell antibody blocks the induction and maintenance of oral tolerance to ovalbumin in mice. Imniunol. Lett 48, 97-102. Mincheva-Nilsson,L., Baranov,V., Yeung, M. M.-W., Hammarstrijm, S., and Hammarstriim, M.-L. (1994). Immunomorphologic studies of human decidua-associated lymphoid cells in normal early pregnancy. J . Immlinol. 152, 2020-2032. Mincheva-Nilsson,L., Hammarstriini, S., and Harnmarstrom, M.-L. (1992).Human decidual leukocytes from early pregnancy contain high numbers of yS+ cells and show selective down-regulation of alloreactivity.J. Inunrrnol. 149, 2203-221 1. Mixter, P. F., Camerini, V., Stone, B. J.. Miller, V. L., and Kronenberg, M. (1994). Mouse lymphocytes that express a y 6 T cell antigen receptor contribute to resistance to salmonella infection in ciuo. Itgect. Inirnzrn. 62, 4618-4621. Miyagawa, Y., Matsuoka, T., Baba, A., Nakaniura, T., Tsuno, T., Tamura, A,, Agematsu, K., Yabuhara, A., Uehara, Y., Kawai, H., and Miyasaka, K. (1992). Fetal liver T cell receptor y/6+ T cells as cytotoxic T lymphocytes specific for maternal alloantigens. J. Exp. Med. 176, 1-7. Mombaerts, P., Arnoldi, J., Russ, F., Tonegawa, S., and Kaufmann, S. H. E. (1993).Different roles of a/3 and y6 T cells in immunity against an intracellular bacterial pathogen. Nature 365, 53-56. Mombaerts, P., Clarke, A. R., Rudnicki, M. A,, Iacoinini, J., Itohara, S., Lafaille, J. J., Wang, L., Ichikawa, Y., Jaenisch, R., Hooper, M. L., and Tonegawa, S. (1992). Mutations in Tcell antigen receptor genes a! and /3 block thymocyte development at different stages. Nature 360, 225-231. Mombaerts, P., Mizoguchi, E., Ljunggren, H. G., Iacomini, J., Ishikawa, H., Wang, L., Grusby, M. J., Climcher, L. H., Winn, H. J., and Bhan, A. K. (1994).Peripheral lymphoid development and function in TCR mutant mice. Znt. Zinrntrno~.6, 1061-1070. Morita, C. T., Beckman, E. M., Bukowski, J. F., Tanaka, Y.. Band, H., Bloom, B. R., Colan, D. E., and Brenner, M. B. (1995).Direct presentation ofnonpeptide prenyl pyrophosphate antigens to human y6 T cells. Immunity 3, 495-507. Morita, C. T., Verma, S., Aparicio, P., Martinez-A, C., Spits, H., and Brenner, M. B. (1991). Functionally distinct subsets of human y / 6 T cells. Eur. J. Inmiunol. 21, 2999-3007. Moriwaki, S., Korn, B. S., Ichikawa, Y., Van Kaer, L., and Tonegawa, S. (1993). Amino acid substitutions in the floor of the putative antigen-binding site of H-2T22 affect recognition of a y6 T-cell receptor. Proc. Natl. Acnd. Sci. USA 90, 11396-11400.
Mowat, A . M. (1987). The regulation o f iiiiniiiiic responses to dietary protein antigcws. z l r t r l ~ l ~ r t ~?br/r!/ l/. 8, 933-98. ;uid Nomoto, K. (1995). Mrlkasa, A , . Hioroniatsri, K., Matsiizaki, (;., O’Brieii. R., B c i r i i ~\i’.. Bacterial infection of the testis leading to ;uitoaggrrssive i l i i i i i r i i i i h triggers apparently opposed responses of a/3 antl y6 T cc~lls.J . Z i r t r r t i ~ r i o / , 155, 2047-2056. Miikasa. A,, Lahn. M.. Pfluiri, E. K.. Borii. it’., antl R. L., 0. B. (1997). E d e n c e that the sanics y6 T cells respond driiiiig iiiti~ction-inducrd ;iid antoinimune inflainmation. J . Z t r i t r i i i r t o / . 159, 5787-,5794. Munk,M . E., Fazioli, R. A., Calich. \’, L. G.. and Kaiifimtin, S. H. E. (199s).P~mcoccit/ir,i[k,s bmsi/iotsis-stiiiriilated hiinian y / 6 T ccsllssupport a n t i l d y production by B cells. Infect. Z r m i i r t i . 63, 160% 1610. Munk. M. E.. Teixeira. H . (;., Kiiner, A . Fazioli, R. A., Gdicli, \’. L. G., and Kaufinann, S. H. E. (1994).Hriinan a/3 a i d yd T cells from iinexposeed individuals respond to protein , ~ i ~ i Ziif. o i ~ Z r /n nf~ ,u t.i d~.6, 171 7-1725 antigens of the yeast form of ~ ~ r ~ ~ ~ ~ ~hrt/silicmi.s. Nagler-Anderson. C., McNair, L. 4.,arid Cradock. A. (1992). Srlf-reactive, T cell receptory6’, Iymphoc\tes from the intestinal e p i t h e h i i ofwcwiling mice. J . Zrritriutid. 149,2315-
2322. Nakajiina, H., Moore, Y. I . , and Takiguchi. M . (1997).Fiinctional expression of the N K B 1 killer cell inhibiton. receptors on a y6 CTL clone. Zrritrirrriogeiictic:v 45, 226-228. Nakajiina, 1-1.. Tomiyaina, H.. and Takigiichi. M . (1995).Inhiliitioii of gamma delta T cell recognition by receptors for MHC class I niolecriles. J . Zttttrtutid 155, 4139-4142. Nishiniiira, 13.. Emoto, M., I-liromatsu, K.. Yani;inioto. S.. Matsuura, K., Gomi, H.. Iketla, T., Itoliara, S . , ant1 Yoshikai, Y. (1995). The role of y6 T cells in priniing macrophages to prodiicc, tiimor necrosis f’actor-a. Eiir-. /. I t r i t r t r r r i o / . 25, 1465-1468. Nishiniiira, H.,Hiroiiiatsii. K.. Kobayaslii. N., C.ral)steiii, K. €1.. P at o n , R., Siigainrira, K., Bhestone, J. A.. and Yoshikai, Y. (1996). IL-15 is a novvl growth fiictor for miirine y6 T cells induced by Snlrrtortc~/k~/infection J . Z r r i t r i r r r t o / . 156, 663-669. Nishiiiiiira, Y., Eto, M.. Maedii, T., Hiroinntsu, K.. K o b a ~ d i i ,N., Noitloto, K., Kong, Y.-Y., antl Nomoto, K. (1994).Inhibitiori of skin scnograft rejection by depleting T-cell reccytor a/3-heariiig cells withorit T-cell rweptor yd-bearing cells or natural killer cells by monoclonal mtil~otly.Ztrtniiiriology 83, 196-204. O’Brieii, R. I,., FII,Y.-X., Crarifill. R.. Dallas. A.. Reardon. C.. Lang, J.. Carding. S. R., Kiiho, R., and Born, M‘. (1992). €Ieat shock protein Hsp-60 reactive y6 cells: A large, divrrsified T Iyiiipliocyte siihset with highly focused specificity. Proc. Nod. Accd. Sci. USA 89, 4348-4:352. O’Brien, R. L.. Happ. M. P.. Dallas, A,. Pa1nic.r. E., Kril)o. R., and Born, tV. K. (1989). receptor yd by an antigen Stimiilatioii of a niajor siibset of lyniplioc derived from Myco1)acteriuiii tuberculo )to, K. (1990). Sequential Oliga, S., Yoshikai, Y., Takeda. Y.. Hiro ty during an i.p. infection qqw;iriiice of $6- and dp-l)earing T c with Listeria monocytogeiies. Eirr. J . Z Olive, C . ( 1997). Modillation of exprrimental all(.rgic c~ncephaloinyelitisin mice by imnirinization with peptide specific for the y6 T cell recvptor. I t t m i i i i o / . Cell Hiol. 75, 102-106. Orsini. D. L., Res. P. S., Van Laar, J. M., Mliller, L. M.. Soprano, A. E., Kooy, Y. M., Tak, P. P.. and Koning, F. (1993). A subset of V61+ T cells proliferates in response to EpsteinBarr \iriis-transfornie(l I3 ce11 hies in \itro. S m d . J . Z r n n i u r i o / . 38, Orsini, D. L., =in Gild, M., Kooy, Y. M.. Struyk, L., Klein. G., vim den Elsen. P., and Koning, F. (1994). Furictional and niolrcrih charactcarization of B-cell responsive V61’ gainina delta T cells. Rut-. J . f r n m r r t t o / . 24, 3199-3204.
138
CZ'I1.LI BORN c/ nl
Ota, Y., Kobata, T., Seki, M., Yagita, H., Shiniada. S., Hrlang, Y.-Y.. Takagaki, Y., and Okiiniiira, K. (1992). Extrathymic origin of VylN66 T cells in the skin. Ettr. J . Iintnnnol. 22,595-598. Palliard, X., Yssel, H,, Blanchard, D., Waitz, J. A,, deVries, J. E., and Spitz, H. (1989). Antigen specific and MHC nonrestricted cytotoxicity of T cell receptor aBi and y6+ human T cell clones isolated in IL-4. I. Ztnmtnol. 143, 452-457. Panchamoorthy, G., McLean, J., Modlin, R. L., Morita, C. T., Ishikawa, S., Brenner, M. B., and Band, H. (1991). A predominance of the T cell receptor of Vy2/V62 subset i n human inycobaceria-responsivc. T cells suggests germline gene encoded recognition. 1. Imnunol. 147, 3360-3369. Pao, W., Wen, L.. Smith, A. L., Culbranson-Judge. A,. Zheng, B., Kelsoe, G., MacLennan, 1. C. M., Owen, M. J., and Hayday, A. C. (1996).y6 T cell help of B cells is induced by repeated parasitic infection, in the absence of other T cells. Cztrr. Biol. 6, 1317-1325. Pardoll, D. M., Fowlkes, B. J., Bluestone, J. A,, Kniisbeek, A., Maloy, W. L., Cohgan, J. E., and Schwartz, R. €I. (1987). Differential expression of two distinct T-cell receptors during thymocyte development. Natnre 326, 79-81. Parker, C. M., Groh, V., Band, H., Porcelli, S. A., Morita, C., Fahhi, M., Class, D.. Strominger, J. L., and Brenner, M. B. (1990). Evidence for extrathymic changes in the T cell receptor y / 6 repertoire. I. E q i . Med. 171, 1597-1612. Parr, E. L., and McKenzie, I. F. C. (1979).Demonstration of Ia antigens on mouse intestinal epithelial cells by immunoferritin labeling. In~nunogenetics8, 499-508. Peng, S. L., Madaio, M. P., Hughes, 1). P. M., Crispe, I. N., Owen, M. J., Wen, L., Hayday, A. C., and Craft. J. (1996). Murine lupus in the absence of (YP T cells. J. I n i ~ i r ~ n d . 156, 4041-4049. Pereira, P., Zijlstra, M., McMaster, J., Loring, J. M., Jaenisch, R., and Tonegawa, S. (1992). Blockade of transgenic y6 T cell development in B2-microglobulindeficient mice. E M B O J. 11, 25-31. Peterinan, G. M., Spencer, C., Sperling. A. I., and Bluestone, J. A. (1993). Kole of y6 T cells in inurine collagen-induced arthritis. J. Immnnol. 151, 6546-6558. Pfeffer, K., Schoel, B., Gulle, H., Kaufinann, S. H. E., and Wagner, H. (1990). Primary responses to human T cells to mycobacteria: '4frequent set of y / 6 T cells are stimulated by protease-resistant ligands. Etrr. 1.Itnmnnol. 20, 1175-1 179. Pfeffer, K., Schoel, B., Plesnila, N., Lipford, G . B., Kromer, S., Deusch, K., and Wagner, H. (1992). A lectin-binding, protease-resistant mycobacterial ligand specifically activates 199' human y6 T cells. J . Zmmcnol. 148, 575-583. Porcelli, S., Brenner, M. B., and Band, H. (1993). Biology oftlie human y6T-cell receptor. I r n m m l . Reo. 120, 137-183. Porcelli, S., Brenner, M. S., Greenstrill, J. K., Balk, S. P., Terhorst, C., and Bleicher, P. (1989). Recognition of cluster of differentional antigens by human CD4-CD8- cytolytic T lymphocytes. Nature 341, 447-450. Porcelli, S. A., and Modlin, R. L. (1995).CD1 arid the expanding universe of T cell antigens. J. Iminunol. 155, 3709-3710. Ptak, W., and Askenase, P. W. (1992). y6 T cells assist ap T cells in adoptive transfer of contact sensitivity. J . Zm?nunol. 149, 3503-3508. Ptak, Mi., Szczepanik, M., Ramabhadran, R., arid Askenase, P. W. (1996). Immune or nonnal y6 T cells that assist a/3 T cells in elicitation of contact sensitivity preferentially me Vy5 and V64 variable regon gene segments. /. Imtnunol. 156, 976-986. Quaroni, A., and Isselhacher, K. J. (1985). Study of intestinal cell differentiation with monoclonal antibodies to intestinal cell surface components. Deo. Rial. 111, 267-279.
Quere, P., Cooper, hl. D., and Tliorhecke, G. J. (1990). Cliaractcrizatiori of' suppressor T cells for aiitibody protluctioii I)? cliickrn spleen cells. I. Antigen-induced suppressor cells are CTH', TcRI' (76) T cells. z t t L t t I f l t i O / ~ J @ /71, 517-522, Rajagopalan. S., Mao, C., and Ilatta, S. K. ( 1992). Patliogenic autoailtibody-inducing y/S T helpw cells from patients with lupiis IirphiitiT express tiiiiisrial T cell receptors. -Glitz. ~ ~ ~ l ~ f i f l f~ I t Ot / l . f t I f / f ~ ~ J62, ~ ~ ~344-350, f ~ l ~ ~ ~ . Hajagopalaii. S., Zordan, T.. Tsokos. G. C., antl DattLi. S . K. (1990). Pathogenic antiDN.4 autoantibocly-iii(1~iciiigT helper cell lines from patients with active Inpus nephritis: Isolation of CD4-8 T helper cell lines tliat cxprvss the yS T-cell antigen receptor. Proc. N ~ t l4cod. Sci. USA 87, 7020-7024. Rakasz, E., Sandor. M.. Hagrn, M.. aiid Lycli, R. G. (1996).Activation features of intraepithelid gminra delta T-cells of the muriiie vagina. Z t t t t t w i i d . Left. 54, 129-1:34. Hast, J . P.. Anderson. hl., Strong, S.J., h e r , (;,, Idtirraii, R. T., ant1 Litinan, G. W. (1997). a, p, y a i d S T cell antigen receptor geiies arose early i n vertebrate phylogeny. Zii~nirrnity 6, 1-11. Ratlibiiil, G.A.. Born. \V., Krizie!, LV. A , . uid Trickt~r,1'. LV. (1988).Diversity of the mouse T cell receptor Cyl gene: Striictiiral aiiiilysis in C57HIJKa. Zi,rtiitriiogr,tLr.fics 27, 121-126. Raulet, 11. H. (1989). The stnicturr, liinctioii. a i i d inolecular genetics o f tlie ylS T cell . Z t t m i m d . 7, 175-207, receptor. L h ~ r i t t Rco. Rarilet, D. 11. (1994). Ilow yS T cells d i e ii hing. Cirrr. B i d . 4, 246-248. Kaulet. D. H., Sprwer, D. M . . Hsiaiig, Y . - H . . C ~ o l c l i i i m ,J. P., Bix, M., Liao. N.-S., Zijlstra, M., Jaenisch, R., antl Correa, I. ( 1991).Control ofy6 T-cell developnient. Z t r i t t w t u d Rrti. 120, 185-204. Rcardon, C., Lefrmcois. L., Farr, A,. Kubo,It.. O'Brien. R.. and Born, M'. (1990). Exprcassion of' ylS T cell receptors on lynipliocytcs from thc Inctatiiig maminary gland. 1.E s p M o d 172, 1263-1266. Reardori, C. L., Heybome, K., Tsuji. M.. Zavala, F., Tigelaar, R. E., O'Brieii K. I>., and Born, W.K. (1995). Miirine epitlrrinal VyFjn'SI-TCR' T cells r q o n d to B cell lines untl liI)ol")lysacctiaridcs. I , Incest. L)wttwtol. 105, 58S-61S. Rella1i:iii. B. L., Bluestone, J. A.. Houlden. B. A . , Cotteriiian, M. M., mid Matis, L. A . (1991). Junctional secpences inHiience the specificity of ylS T cell receptors. /. E x p Med. 173, 503-506. Rliciiis. 1,. A,. Yoiuig, E. M . . Nordlrind, M . I,., Berning. 11. B.. a i d Nordlund, J. J. (19x7). Rapid inchiction of Thy-1 antigenic niarkers on krratinocytes and epiderinal iminuiie cells i n tlie skin of mice following topical tre;itment with coiiiiiioii preservatives used in topical medicatioris ailid in foods. /. Z t w r s f . Z A > t - t f r d o / . 89, 489-494. Itoark, C. E. (1995). A study on iiiuriue liver gaiiui1ddrlta T lymphocytes. Pl1.D. The ' Roark, C . E., Volliner. M . . Cairiphell, P., Borii, W.K.. and O'Biieii, R. L. (1996). Respo~isc~ /. Z t t ~ i t i i m d 156,2214-2220. . o f a yS-TCH irioiioinorphic subset during 1)acteii;iliiift~tioii. Roberts, A. D.. Ordway, D. J., mid Orine. I. M . (1993). Listeria rriot~ocytogc,tir.sinfcction i n 8 2 microglobulin-deficieiit inicc. Ztlji,ct. Z t t u t i u t i , 61, 111:3- 1116. Roberts, K., Y C J ~ [ J ~ I I IW. I ~ IM . . , Kehn, 1'. J.. a i i d Sllcvxh, E. M. (1991). The vitronectin receptor serves as an accessory niolecdr for the activation of a subset of ylS T cells. I. E x p . Met!. 173, 231-240, Koberts, S. J., Smith. A. L., West, '4.B., Wen, L., Fiiidy, R. C., Owen, M. J., and Hayday, A. C . (1996). T-cell L Y and ~ yS' deficient mice display ahioriiial but distinct phenohpes toward a natural, widespread infection o f the intestinal epithelium. Proc. N o t / . Actid. Sci. U S A 93, 1 1774- 11779, Robey, E., and Fowlkrs. B. J. (1994). Selective events in T cell development. h n i l . Hec. Z i t u t i i L t w l , 12, 675-706.
140
\VILLI UOHN ct
NI
Rock, E. P., Sibbald, P. R., Davis, M. M., and Chien, Y.-H. (1994).CDR3 length in antigenspecific immune receptors. J. E x p Med. 179, 323-328. Rothenberg, M. E., Weber, W. E., Longtine, J. A,, and Hafler, D. A. (1996).Cytotoxicgalrllua delta I lymphocytes associated with an Epstein-Barr virus-induced posttransplantation lyinplioproliferative disorder. Clirr. Ittittiunol. Inttiuntpztliol. 80, 1266- 1272. Rust, C. J. J., Verreck, F., Vietor, €I., and Koning, F. (1990). Specific recognition of staphylococcal eliterotoxin A by Iiumaii T cells hearing receptors with the Vy9 region. Nature 346, 572-574. Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A,, Eisen, H., and Tonegawa, S. (1984). Complete primaiy stiuctiire of a heterodimeric T-cell receptor deduced from cDNA sequences. Nature 309, 757-762. Saito. T., Hochstenbach, F., Marusic-Calesic, S., Kruisberk, A. M., Bremier, M., and Germain, R. N . (1988).Surface expression of only yS and/or a/3T cell receptor heterodimers by cells with four (a,P,y,S,)functional recrptor chains. J , Exp. Med 168, 1003- 1020. Sano, Y., Dudley, E., Carding, S., Lin, R.-H., IIayday, A. C., and Janeway, C. A. Jr., (1993).yS T-cell lines isolated from intestinal epithelium respond to a B-cell lymphoma. zfnttlunology 80, 388-394. Satoh, M., Seki. S., Hashimoto, J. W., Ogasawara, K.. Kobayashi, T., Kumagai, K., Matsuno, S., and Takeda, K. (1996). Cytotoxic gammadelta or alphabeta T cells with a natural killer c d l maker, CD56, induced from human peripheral blood lymphocytes by a combination of IL-12 and IL-2. J. Z t t m i t m d . 157, 3886-3892. Scharton, T. M., and Scott, P. (1993). Natural killer cells are a source of interferon y that drives differentiation of CD4+ T cell subsets and induces early resistance to Leishttimiia tnajor in mice. J. Exp. M e d 178, 567-577. Schild, H., Mavaddat, N., Litzenherger, C., Ehrich, E. W.. Davis, M. M., Bluestone, J. A,, Matis, L., Draper, R. K., and Chien, Y.-H. (1994).The nature of major histocompatibility complex recognition by yS T cells. Cell 76, 29-37. Sclioel, B., Sprenger, S., and Kairfinann, S. H. E. (1994).Phosphate is essential for stiniulation afVySVS2 T lymphocytesby inycohacterial low molecular weight ligand. Eur. J. Z t t m u t i d . 24, 1886-1892. Sciammas, R., Johnson, R. M., Sperling, A. I., Brady. W., Liiisley, P. S., Spear, P. G., Fitch, F. W., and Bluestone, 1. A. (1994). Unique antigen recognition by a herpesvirus-specific TCR-yS cell. J. Zn~rncmol.152, 5092-5397. Sciammas, R., Kodukula, P., Tang, Q., Hendricks, H. L., and Bluestone, J. A. (1997).T cell receptor-y/S cells protect mice from herpes simplex virus type 1-induced lethal encephalitis. /. Exp. Med. 185, 1969-1975. Shiohara, R., Moriya, N., Gotoh, C., Sajzawa, K., and Nagashima, M. (1988). 1,ocally administered monoclonal antibodies to lymphocyte function-associated antigen 1 and to L3T4 prevent cutaneous graft-vs-host disease. J. Ittiniund. 141, 2261-2267. Shiohara, T., Moriya, N., Gotoh, C., Hayakawa, J., Nagashima, M., Saizawa, K., and Ishikawa, 13. (1990). Loss of epidermal integrity by T cell-mediated attack induces long-term local resistance to suhsequent attack. I. I ~ i d ~ c t iof~ iresistance i correlates with increased in Thy-lt epidermal cell numbers. J . Exp. Mrd. 171, 1027-1041. Shiohara, T., Moriyi, N., Hayakawa, 1.. Itohara. S., and Ishikawa, H. (1996). Resistance to cutaneous graft-vs.-host disease is not induced in T cell receptor S gene-mutant mice. J . Exp. Met/. 183, 1483-1489. Shiohara, T., Nariinatsu, H., and Nagashima, M. (1987). Induction of cutaneous graft-vshost disease by d o - o r self-Ia-reactivehelper T cells in mice. Transplantation 43,692-698. Sim, G.-K. (1995). Intraepithelial lymphocytes arid the irnrnuiie system. Ado. Iintnunol. 58, 297-343.
1hlhlUYOREC;ULATORY FUNCTIONS OF y6 T C:El,l.b
141
Siin. G.-K., and Augustin. A. (1991a). Dominant expres,sion of the T cell receptor BALB invariant 6 (BID) chain in resident pnhnonaiy l>mphocytes is due to selection. Errt-, /. Z t t i t t i r r d . 21, 859-861. Sini, G.-K., and Augustin, A. (199lb). Extrathvnric positive selection of y6 T cells: \Jy,Jyl rearrangements with “CxYS” junctions. I . Itnttiirtio/. 146, 2349-2445. Sim, G.-K., and Augustin. A. (1993).The presence of a i i endogenons murine leukemia Liirus sequence correlates with the peripheral expansion of y 6 T cells bearing the BALB invariant delta (BID) T cell receptor 6.1. Exp. M L , ~178, . 1819-1824. Sim, G.-K., Rajaserkar, R., Dessing, M.. antl Angustin, A. (1994). Homing antl iri s i t r r diffi2rentiationof resident pulnionay lymphoc>tes. Int. Z t t ~ t ~ ~ r i r i6, o l . 1287-1295. Skeeii. M. J., antl Ziegler, H. K. (1993;i). Intluction of niiirine peritoneal y6 T cells ant1 their role in resistance to biicteiial infection. /. En-11. Med 178, 971-984. Skeen, M. J., and Ziegler, H. K. ( I 99311). Intercellular interactions and cytokine responsivcness of peritoneal d p and y/S T cells from Lis(c,t-ici-iiifected inice: Synerpstic effects of interleukin 1 and 7 on y/6 T cells. I, E.vp hlcrl. 178, 985-996. SGderstroni, K., Bucht, A , , Halapi, E., Gronberg, A,, Magnusson. I., and Kiessling, H. (1W6). Increased frequency of al)nornial ganinia delta T cells in blood of patients mith inflainniatoy bowel diseases. /. Ittitttrrtio!. 156, 2331-2339. Sowder, J. T.. C h i , C.-L. H., Ager, L. L.. C1i;in. M. M.. and Cooper, M. D. (1988). A largr subpopulation of avian T cells express ii Iiomologue oftlie nianinialian Ty/6 receptor. /. Exp. hlerl. 167, 315-322. Spaner, D.. Cohen, B. L., Miller, H. G . , arid Phillips, R. A. ( 1995). Antigen-presenting cells for naive transgenic y6 cells: Potent activation by activated cup T cells. I. Z r t ~ m t r d . 155, :3866-3876. Spaner, D.. Migita, K., Ochi, A,, Shannon, J., Miller. R. G.. PeleiI’d. P.. Tonegawa, S., and Phillips, R. A. (1993). y6 T cells differentiate into a fiinctional h i t nonproliferative state during ii normal irnniune response. Proc. Not/. Acnd. Sci. U S A 90, 8415-8419. Sperling, A. I., Cron, H. Q.. Decker, D. C., Stern, D. A,, and Bluestone, J. A. (1992). Peripheral T cell receptor y6 variable gene repertoire maps to the T cell receptor loci and is influenced by positive selection. /. Z i t i t t i r i t i o ! . 149, 3200-3207. Sperling, A. I., and Wortis, 11. H. (1989).CD4.. CD8- y/6 cells froin normal inice respond to a syigeneic B cell ljinphoma antl can induce its differentiation. Int. Irtttrwnol. 1, 434-442. Stenger, S., Mazzaccaro, R. J., Uyemrira, K..Cho, S., Banies, P. F.. Rosat, J.-P., Sette, A,, Br~nner,M. B.. Porcelli, S. A,, Blooin, B. R., and Modlin, R. L. (1997). Differential effects of‘ cytolytic T cell subsets on intracrllriliir infixction. S ’ C ~ P ~ I C C276, 1684-1687. Studicr, F. \\’.. Rosenberg, A. H., Diinn. J. J., and Dubendorff, J. \V. (1990). Use of T7 RNA plyinerase to direct the expression of cloned genes. hfethody E t m p u d 185, 80-89. Sun, Y,, Chrn, Z., C:hilng, S. W., Zeng, H., and Gorczyiski. R. M. (1997). TCR diversity in y6TCII’ hybridomas derived from niice givcn po1ta1vein donor-specificpreimmunization and skin allografts. Mol. MK/. 3, 89- 102. Sunaga, S., Maki, K., Koniagata. Y.. Miytzaki, J.-l., and I h t a , K. (1997). Developinentally ordered V-J recoinbination in inouse T cell receptor y locus is not perturbed by targeted deletion of the \7y4 gene. 1. Iinttiutio~.158, 422:3-4228. Suzriki. T., Hirornatsu, K., Antlo, Y., Okmioto, T.. Tomotla. Y.. antl Yoshikiu, Y. (199s). Regulatory role of y6 Tcells in uterine intraepithelial lyinphocytes in maternal antifetal iininune response. 1. Zmvmtiol. 154, 4376-4484. Szczepanik, M., Anderson, L. H., Ushio, I I . , Ptak, W., Owcw, M. J., Hayday, A. C.. and Askenase, P. \V. (1996). y6 T cells from tolerized ap T cell receptor (TCR)-deficient mice inliibit contact sensitivity-effectorT cells in vivo. and their interferon-y production in citro. J . Exp. Mcd. 184, 2129-2139.
142
\VILLI BORN
“1
al
Takada, H., Matsuzaki, G., €Iiromatsu, K., and Nornoto, K. (1994). Analysis of the role of natural killer cells in Listerict triotiocytogrne.~infection: Relation between natural killer cc.lls and T-cell receptor y6 T cells in the host defence niechanism at the early stage of infection. Zititt~u~~ology 82, 106-112. Takagaki, Y., DeCloux, A,, Bonneville, M., and Tonegawa, S . (198%). Diversity of y6 Tcell receptors on inurine intestinal intraepitlielial lymphocytes. Notiire 339, 712-714. Takagaki, Y., Nakanishi, N.. Ishida, I., Kanagawa, O., and Tonegawa, S. (198913). T cell receptor-y and -6 genes preferentially utilized by adiilt thymocytes for the surface expression. J . Im?tiutiol. 142, 2112-2121. Tanaka, Y., Morita, C. T., Tanaka, Y., Nieves, E.. Brenner, M. B., and Bloom, B. H. ( 19Yq5). Natural and synthetic non-peptide antigens recognized by human y6 T cells. Notrirc 375, 155-158. Tanaka, Y., Sano, S., Nieves, E., De Libero, G., Rosa, D., Modlin, H. L., Brenner, M. B., Bloom, B. R., and Morita, C. T. (1994). Nonpeptide ligands for human y6 T cells. Yroc. N ( ~ t lAcccd. . sci. USA 91, 8175-8179. Tigelaar, R. E., and Lewis, J. M. (1994).I n “Basic Mechanisms of Physiologic and Abr:rrant Lyniplioproliferation in the Skin” (W. C . Lambert, B. Giannotti, and W. van Vloten, eds.), pp, 39-55. Plenum Press, New York. Tripp, C. S., Wolf, S. F., and Unanue, E. R. (1993). Interleukin 12 and tumor necrosis factor (Y are costirnulators of interferon y production by natural killer cells in severe combined irnir~uiiotleficiei~cy inice with listeriosis, and iiiterleukin 10 is a physiologic antagonist. Proc. Notl. Acrrd. Sci. USA 90, 3725-3729. Tschachler, E., Scliuler, G., Hutterer, J., Leibl. H., Wolff, K., and Stingl, G. (1983). Expression of Thy-I antigen by inurine epiderinal T cells. J . Ztiuest. Denn. 81, 282-285. Tsuji, M., Eyster, C., O’Brien, R. L., Born, M’. K., Bapna, M., Reichel, M., Nussenmeig, R. S., andzavala, F. (1996).Phenotypic and functional properties of murine y8Tcell clones derived from malaria iriimunized (YOT cell-deficient mice. Znt. Zmmunol. 8, 359-366. Tsuji, M., Mombaerts, P., Lefrancois, L., Nussenmeig, R. S., Zavala, F., and Tonegawa, S. (1994). y6 T cells contribute to immunity against the liver stages of malaria in (YOTcell-deficient mice. Proc. Nntl. Acod. Sci. USA 91, 345-349. Tsuzuki, T., Yoshikai, Y., Ito, M., Mori, N., Ohbayashi, M., and Asai, J. (1994). Kinetics of intestinal intraepithelial lymphocytes during acute graft-versus-host disease in mice. Eur. /. Imnnunol. 24, 709-715. Unanue, E. R. (1997).Inter-relationship among macrophages, natural killer cells and neutrophils in early stages of Listerin resistance. Ciirr. Opin. Imtnunol. 9, 35-43. Us:iini, J., Hiromatsu, K., Matsumoto, Y., Maeda, K., Inagaki, H., Suzuki, T., and Yoshikai, Y.(1995).A protective role of y6 T cells in primary infection with L i s t d o t~umxytogenes in autoirnniune non-obese diabetic mice. Ittinircnology 86, 199-205. Uyeinura, K., Deans, R. J., Band, H., Ohintan, J., Pancliainoorthy, G., Morita, C., Rea, T. H., and Modlin, R. L. (1991). Evidence for clonal selection of y/S T-cells in response to a human pathogen. J. E x p . Mcrl. 174, 683-692. Uyemura, K., Klotz, J., Pirinez, C., Ohmen, J., Wang, X.-H., Ho, C., Hoffman, W. L., and Modlin, R. L. (1992). Microanatomic clonality of y6 T cells in human leischinaniasis lesions. J , hi?tumd. 148, 1205-121 1. Vaessen, L. M. H., chiwehand, A. J., Baan, C. C., Jutte, N . H. P. M., Balk, A. 1%. M. M., Claas, F. H. J., and Weimar, W. (1991). Phenotypic and functional analysis of T cell receptor 78-bearing cells isolated from human heart allografts.]. h ~ mt 4 n d .147,846-850. Vaessen, L. M. B., Schipper, F., Knoop, C., Claas, F. H. J., and Weimar, W. (1996). Inverted V61N62 ratio within the T cell receptor (TCR)-yG T cell population in peripheral blood of heart transplant recipients. C . in Exp. Ztrttnunol. 103, 119-124.
IblMUNOHEGU1,ATORY F U N C T I O N S OF yS T CELLS
143
Vandekerckliovt.. R . A. E., Datenia, C.,Ktriniiig, F., Goulniy. E., Persijii, G. G.. \’:in R o o d , J. I., Claas, F. 11. J.. and Dc Vries. J. E. (1990). Analysis o f the donor-specific cytotoxic T Iyiiiphocyte repertoire in a patient with a long term suni\ing allograft: Frequen specificity, and phenot.ype of tlonar-reacti\,e T cell rweptor (TCK)-aP’ ;ind TCR-yG cio1nes..I r t t 1 7 t l l ~ ~ l ~144, )/. iw-12~4. \‘an der Ileytle, H. C . , Elloso, M. M., Chaiig, W.-L.. Kaplan, M., Manning, D. D., and \Veitlanz. CV. 1’. ( 1995). y6 T cells fiiirctioii in cell-inetliated iinmunity to acutcx bloodstage Pkistiioditrrti dtabafidi nr/at?ii m;ilaria. /. Z t t i r i i u i i d . 154, 3985-3990. van tier Heyde, €1. C.. Elloso, M . M., Roopenian. D. C:.. Maiming, D. D.. and Weidariz, \Y. P. (199:3a), Expansion of the C113 ,CIX yS T cell sul)sct in the spleens of niicc tirlrillg rloI~-ietililiiliood-stage I11;tianLI.El4,l 1. r ~ t t ~ r ~ l l 23, r l o ~1846-1850. . \7an tier Heyde, H . C., Manning, 1). D., and Weitlanz, \V.P. (199:3b). Role of CD4’ T cells in the expansion of the CD4-, CDS- yG T cell siibset i i i the spleens of mice during blood-stage inalaria. /. Z r n t i i u t i o / . 151, 6311-6317. Vassalli, P. (1992). The pathophysiolo~of tiii~iornwrosis factors. Antiti. Hrc. Z i t i t i i u i i d . 10, 411-452. Vitlovic, D., and Dembic, Z.(1991). @I-I restricted g:iinma d d t a T cc4ls can help B cells. Curr. Top. Alicrobiol. I r t ~ r t i i r r t o l , 173, 239-244. Vietor, I*., and Koniug, F. (1990). y6 T-cell receptor rcapertoire iii huinaii peripherd blood a i d thyiirris. Z t t i i r t ~ ~ n ~ ~ r 31, i i ~ t340-346. ic~. Vincent, M. S., Roessiler, K . , Lynch. D., \Vilsoii, D.. Cooper. S. M., Tschopp, J., Sigal, L. H., and Budd, R. C. (199G). Apoptosis of Fas-high CD4* spovial T cells by horreliareactive Fas-ligand (high) y6 T cells iii Lyirie d i n t i s . /. bkp. ‘!fed 184, 2109-2117. W e i n t ~ ,11. L., Frietlman, A , . Miller, A , , Kliorrry, S. J., Al-Salhagh. A,, Saritos, L.. Sayegh, %I., Nnssenhlatt. R. B., Trentham, D. E., ; i d Hafler, D . A. (1994). Oral tolerance: Iininiunologic iriechanisms and treatment of:iniirial and human organ-specific autoimmune diseases by oral administration of autoantigens. At lt iu H m ZrnrtitirioZ. 12, 809-837. Weintraul). B., Eckinaiin. L., Okairwto, S.. Hense, M., llediick, S. M., and Fierer, J. (1997). Role of c.P and y6 T cells in tlir host rrsponse to Sdrtioricllri infection as demonstrated iri T-cell-rcceptor-deficient mice or defined Zty genotypes. Infect. Irnitiuii. 65,2306-2312. Weintraub, B. C., Jackson. M. R., and Hetlrick, S. M. (1994). y6 T cells can recognize nonclassical MHC i n the absence of corrveiitional ailtigenic peptides. 1.Zminuno/. 153, 305 1-3058. Wells. F. B., Gahm, S.-J., Hedrick, S. M.. Bluestone. J. A , , Deiit, A,, and Matis, L. A. (1991). Reyuirenwnt for positive selection of yG reccptol--be~iriiigT cells. Science 253, 903-905. W d l s . F. B., Tatsuiini, Y.. Bluestone, J . A , , Hrdrick, S. M., Allison, J. P.. and Matis, L. A. (1993). Plienotypic and fimctional an;ilysis of positive selection in the y/S T cell lineage. 1.E.xp. Mid. 177, 1061-1070. Welsh. R. M., Lin, M.-Y.. Lohtnan, H. I,.. \’arga. S. M.. Zaronainski. C. C., and Selin, L. K. (199i). a11 and y6 T-cell networks a r i d their r o k s in natural resistance to b i r d infc~tions.Itnrrrrttiol. Rcc. 159, 79-93. Wen, I,.. Pao, W., \Vong, F . S., Prng, Q.?Craft, J., Zlieng. B., Kelsoe, G.. Dianda, L., Owen, M. J., and Hayday, A. C. (1996). Cerminal center forination, iminunogkhulin class s\vitchiirg. and autoai1til)ocly protlriction diiveil by “inon a@” T cells. /. Esp. hlrd
183,2271-22&2. Wcn. L., Roberts. S. J., Viriey, J. L.. CYong, F. S., Mallick, C., Findly. R. C . , Peng, Q., Craft, J. E., Owen, M. J.. and Hayday. A . C. (1994). Iinmunoglobulin synthesis arid generalized aritoiiiriiiiiniv in inice congenitally deficient in Cup( +) T cells. Nature 369, 654-658.
144
WILL1 B O R N rf ol.
Werner, S., Peters, K. G., Longaker, M. T., Pace, F. F., Banda, M. J., and Williams, L. T. (1992). Large induction of keratinocyte growth factor expression in the dermis during wound healing. Proc. Natl. Acnd. Sci. USA 89, 6896-6900. Werner, S., Smola, H., Liao, X., Longaker, M. T., Krieg, T., Hofschneider, P. H., and Williams, L. T. (1994).The function of KGF in inorphogenesis of epithelium and reepithelialization of wounds. Science 266, 819-822. Wildner, G., Hiinig, T., and Thurau, S. R. (1996). Orally induced, peptide-specific y/S TCR' cells suppress experimental autoimmune uveitis. Eur. /. Imrriunol. 26,2140-2148. Wilhelm, M., and Tony, H.-P. (1994). An in vitro model for the expansion of Vy962 T lymphocytes during devebpment. S m n d 1.bmnutlo~.40, 521-528. Winoto, A,, and Baltimore, D. (1989a). cup lineage-specific expression of a T cell receptor gene by nearby silencers. Cell 59, 649-655. Winoto, A,, and Baltimore, D. (1989b).Separate lineages of T cells expressing the afi and y6 receptors. Nature 338, 430-432. Wright, A., Lee, J. E., Link, M. P., Smith, S. D., Carroll, W., Levy, R., Clayberger, C., and Krensky, A. M. (1989). Cytotoxic T lymphocytes specific for self tumor immunoglobulin express T cell receptor 6 chain. J. Exp. Med. 169, 1557-1564. Yeh, C.-J. G., Bulmer, J. N., Hsi, B.-L., Tian, W.-T., Rittershaus, C., and Ip, S. H. (1990). Monoclonal antibodies to T cell receptor ylS complex react with human endometrial glandular epithelium. Placenta 11, 253-261. Yokota, K., Ariizumi, K., Kitajima, T., Bergstresser, P. R., Street, N. E., and Takashima. A. (1996). Cytokine-mediated communication between dendritic epidermal T cells and Langerhans cells: In oitro studies using cell lines. /. ZniinuiioL 157, 1529-1537. Yoshimoto, T., Bendelac, A,, Watson, C., Hu-Li, J,, and Paul, W. E. (1995). Role of NK1.I T cells in a TH2response and in immunoglobulin E production. Science 270, 1845-1847. Zhang, X.-M., Tonnelle, C.. Lefranc, M.-P., and Huck, S. (1994). T cell receptor y cDNA in human fetal liver mtl thymus: Variable regions of y chains are restricted to Vyl or V9, due to the absence of splicing of the V10 and V11 leader intron. Eur. /. Iniinunol. 24, 571-578. Zhang, Y., Cado, D., Asaniow, D. M., Komori, T., Alt, F. W., Raulet, D. H., and Allison, J. P. (1995). The role of short homology reports and TdT in generation of the invariant yS antigen receptor repertoire in the fetal thymus. Iintnunity 3, 439-447. Zijlstra, M., Bix, M., Simister, N . E., Loring, J. M., Raulet, D. H., and Jaenisch, R. (1990). fie-microglobulin deficient mice lack CD4-8+ cytolytiic T cells. Nature 344, 742-746. +
This article was accepted for publication on March 24, 1998
STATs as Mediators of Cytokine-Induced Responses TIMOTHY HOEY' AND MICHAEL J. GRUSBYt 'rulorik, Inc., Souh Son Fmncisco, Colifornio 94080 ond beparfment of immunology ond Infectious Diseases, Homrd School of Public Heolh, Deparfment of Medicine, Horvord Medico1 School, Boston, Mossochusetis 02 1 15
I. Introduction
Cytokines are a family of secreted proteins that have important roles in regulating the growth and differentiation of multiple cell types. Although it has long been appreciated that the interaction of cytohnes with their specific cell surfilce receptors results in the induction of new gene transcription, it is only very recently tliat tlie signaling pathways leading from the cytokine receptor to the nucleus have been elucidated. Two novel families of proteins, Jaks ( Janus fhmily tyrosine kinases) and STATs (signal transducers and activators of transcription), have been identified and shown to be important mediators of cytokine-induced signaling. Although originally characterized using interferon (IFN) as a model cytokine, it is now clear that the Jak-STAT signaling pathway is critically important for mediating the biologic effects of a nuniber of different cytokines. Several excellent reviews have lieen published on various aspects of the Jak-STAT signaling pathway (Ihle ct d., 1995; Schindler and Darnell, 199Fj; Ihle, 1996;Darnell, 1997; O'Shea, 1997; Leonard and O'Shea, 1998). The aim of this review is to focus on the biology and biochemistry of the STAT proteins. It reviews how STAT structure is related to function, considers what STAT-deficient mice have told us about the biology of STATs, and discusses new advances in our understanding of how STAT function is regulated. II. The STAT Gene Family
STAT proteins were first identified as components of a DNA-binchng coniplex induced rapidly in response to IFN stimulation. IFN-a induced the activation of a multiprotein coinplex composed of a 11$3-kDaand either a 91- or a 84-kDa protein together with a 48-kDa DNA-binding protein (Fu et a/., 1990). This coinplex was termed interferon-stimulated gene factor-3 ( ISGF-3), and subsequent work demonstrated that tlie 113-kDa protein and tlie 91/84-kDa proteins (which arise from differential splicing of tlie saine gene, see later) became phosphorylated on tyrosine following IFN-a stimulation (Schindler et ol., 1992b). Interestingly, IFN-.)Istimulation led to tlie activation of a different protein complex composed only of the 91-kDa protein and termed ganiina-activated factor (GAF) (Sliuai et 115
( qnnzlat C I W J tn \~.&I!IK Pr?v 211 nglit, 01 r t p r d u c t ~ 111 w nit lnnn re\enc=d OIlBi Zii(dY9 $ 3 0 IJIJ
146
TIMOTIIY IIOEY AND MICHAEL J GRUSBY
al., 1992). The subsequent cloning of the genes encoding these proteins (Fu et al., 1992; Schindler et al., 1992a) led to the eventual identification of the 91-kDa protein (activated by both IFN-a and IFN-7) as Statl and the 113-kDa protein (activated only by IFN-a) as Stat2. Since then, five additional members of this gene family have been cloned, including Stat3 (Akiraet nl., 1994;Zhonget al., 1994), Stat4 (Yamamotoet al., 1994; Zhong et al., 1994), Stats Sa and Sb (Wakao et nl., 1994; Hou et nl., 1995; Liu et nl., 1995; Mui et al., 1995) and Stat6 (Hou et al., 1994; Quelle et al., 1995). It has been pointed out that additional members of the STAT family have been sought through the use of polymerase chain reaction and homologybased screening methods to no avail (Ihle, 1996),suggesting that the seven genes identified to date probably represent the entire gene family. STAT proteins appear to be part of an evolutionarily conserved signal transduction pathway as homologues have been identified in Drosophila (Yan Pt al., 1996) and more recently in Dictyostelium (Kawata et al., 1997). In the mouse, Stat genes are found tightly linked in three clusters (Copeland et al., 1995). Statl and Stat4 are colocalized on chromosome 1, Stat2 and Stat6 are found on chromosome 10, and Stat3, Stada, and StatSb reside on chromosome 11. These observations have led to the suggestion that the family arose through the tandem duplication of an ancestral gene, followed by duplication events of the linked genes and their subsequent dispersion to other chromosomes. As the STAT gene identified in Drosophila is most closely related to StatS, it is possible that this STAT represents the ancestral gene that was then duplicated to yield the Stat3 gene. The high degree of sequence conservation between Stat5a and StatSb suggests that they arose via a more recent gene duplication event. 111. Structural and Functional Domains in STAT Proteins
A. TIIESH2 DOMAIN A N D R~~CEFTOH BINDING The functional domains that have been identified in STAT proteins are shown in Fig. 1. One striking feature of the STAT Family of transcription factors is the presence of a Src homology 2 (SH2) domain. This domain was first identified on receptor tyrosine kinases and functions to bind specific tyrosine-phospliorylated protein sequences (Koch et al., 1991). The STAT SH2 domains, however, represent a subfamily of SH2 domains. For example, this region is approximately 75% identical among STATs 1, 3,4, and 5 , whereas the SH2 domains of Statl and Stat6, the most distantly related members of the STAT family, are 35% identical. In contrast, the conservation between the STAT and the src SH2 domain is much lower. The STAT SH2 domains are only 1Ei-20% identical to src, and the alignment between the STATs and src includes many gaps and nonconservative
STAT\
147
Frc:. 1. Functional doniains of STAT protfsins. Several frinctioiial doniains of STAT molecdrs ha\^ been mapped. These include tliv N-tcrmiiial interaction chnain requircd for tc,tramrrization, a centrd DNA-binding domaiii. ii conserved SH2 tloniain, a conscned tyrosine residue that liecomes phosplio~latetl(111 ;wtivatioii. and ii C-tcrminal transcription activation domain.
substitutions ( Fu, 1992). The highest dcgree of consemition is surrounding the pliospliotyrosine-binding pocket around residue 600 near tlie N-terminal region of the SH2 domain. Several lines of evidence indicate that the SH2 domain-pliospliopeptide interaction provides the specificity for recruitment of the different STAT proteins to various cytokine receI&m. For example, exchange of the Stat1 and Stat2 SH2 domains resulted in an alteration of tlie receptor specificity for the chimeric proteins (Heim et d., 1995). In a reciprocal manner, it was sliown that insertion of the Stat3 docking site sequence YXXQ (see Table I ) derived from a 1 3 0 to the ervthropoietin (EPO) receptor resulted in recruitment of stat3 rather than Stat5 (Stall1 et nl., 1995). In several cases, synthetic peptides d e r i \ ~ from ~ l a cytolune receptor have been found to bind to STAT proteins iti vitro. For example, a phospliorylated peptide derived from the IFN-y receptor intracellular domain was shown to bind STAT1 nionoiners (Creenlund et al., 1994). A useful assay for STAT-peptide interaction is based on the fact that receptorderived peptides can disnipt STAT diiners, leading to an inhibition of DNA-binding activity (Hou et nl., 1994).As mentioned earlier, each STAT SH2 domain appears to have unique peptide-binding specificity. Sequences that can be bound by tlie different STAT proteins, and the receptors from which they are derived, are listed in Table I. Based on studies with src, tlie sequences C-terminal to the phospliotyrosine residue are critical for binding specificity (Songyang et al., 1993). In general, residues at the 1 and + 3 position relative to the phosphotyrosine appear to govern STAT binding to the receptor. Taken together, these data are consistent with the idea that receptor-binding specificity is mediated by the SH2 domain of the STAT protein and the sequence immediately downstream of the phosphotyrosine in the receptor. However, other regions of the STAT protein may also play a role in recruitment to the receptor. The best evidence for this is the observation that deletion of the N-terminal region
+
148
TIMOTHY IlOEY A N D MICHAEL J. G H U S H Y
STAT
Receptor
Sequence
Position
Stat1 state Stat3
IFN-YR IFN-CYH gp130
pYDKPH pYL’FFR pYRHQV pYFKQN pYLPQT pYQPQA pYLPSN pYLSLQ pY LVLD pYKAFS pYKPFQ
440 466 767 814 905 96 1 800 5 10 343 578 606
Stat4 Stat5 Stat6
LIF-R IL-1BR IL-2R Epo-R IL-4n
of Stat2 led to decreased phosphorylation in response to IFN-a, suggesting an inability of the STAT molecule to bind to the receptor (Qureshi et al., 1996). It is not known, however, to what extent this is the case for other STAT Iuroteins as well. For Stat6 at least. the N-terminal domain is not required for tyrosine phosphorylation in response to IL-4 (Mikita et al., 1996). Subsequent to its recruitment to the cytokine receptor, the STAT protein is phosphorylated by Jak kinases on a tyrosine residue located downstream of the SH2 domain around residue 700. Based on in vitro kinase assays and transfection experiments, there does not appear to be much, if any, specificity in the phosphorylation of STAT proteins by the Jaks (reviewed in Ihle, 1996). Furthermore, selective protein interactions between STAT proteins and Jak kinases have not been reported. Thus, the specificity of STAT activation is determined by interaction with the cytokine receptor instead of the Jak kinase.
B. DIMERIZATION Subsequent to their tyrosine phosphorylation, STAT proteins leave the receptor and form dimers through a reciprocal interaction between the SH2 domain of one molecule and the phosphotyrosine of the other. The binding affinities between STAT SH2 domains and receptor-derived peptides are, in general, much greater than those for the interaction of the STAT SH2 domain and the STAT phosphotyrosine peptide and, similar to other SH2 domain-phosphotyrosine peptide interactions, the association and dissociation rates are very fast (Greenlund et al., 1995). An important parameter in STAT dimerization is that the strength of binding in the dimer is the square of the individual SH2-phosphotyrosine peptide affinities.
STATs
149
Therefore, following tyrosine phosphorylation and rapid dissociation from the receptor, STAT proteins have a strong preference for dimerization rather than reassociatioil with the cytokine receptor. It is not known to what extent other regions of the STAT protein contribute to diinerization. Interestingly though, the region from approximately 250 to 310 contains a heptad repeat of hydrophobic residues likely to form an a-helix and similar to the leucine zipper motif involved in the diinerization of other DNA-binding proteins, such as those of the 13-Zip and B-HLH classes (Jones, 1990).
C. N~JCI,EAR LOCALIZATION The transport of proteins from the cytoplasm to the nucleus usually requires a protein sequence termed a nuclear localization signal (NLS). The NLS has not been deterniined for any STAT protein. The activities of Ran, a sniall GTPase that plays a general role in nuclear import, and NPI-1, a component of the nuclear pore-targeting complex, appear to be required for the nuclear transport of Statl after IFN-.)I signaling (Sehrnoto et nl., 1996, 1997). Moreover, experiments with mutant forins of Statl have indicated that dimerization is required for nuclear accuinulation of the protein, a s mutation of Y701 in STAT1 eliminated STAT diinerization as well as nuclear transport (Sekimoto et d ,1996). It has been reported, however, that this same mutant Statl protein is nearly a s effective as wildtype Statl in restoring caspase expression in Statl-deficient cells (Kumar et al., 1997). This observation leaves open the intriguing possibility that low levels of STAT inonomers may be able to niake their way to the nucleus and act as transcriptional regulators through protein-protein rather than protein-DNA complexes. D. DNA BINDING The STAT DNA-binding domain has been identified as a region in the central part of the inolecule between ainino acids 300 and 500. The experimental evidence for this was provided by analyzing the DNA-binding specificity of Stat1 and Stat6 chiineric proteins (Schindler et ut., 1995). Using a similar strategy with Statl and Stat3 hybrids, the DNA-binding specificity region was found to reside between residues 400 and 500 (Horvath ct nl., 199-5).This sequence is not related to any previously characterized DNA-binding stnictural motif. Three classes of DNA-binding sites have been described for the STAT family. The first is the interferon-stimulated regulatory clement ( ISRE) that inediates transcriptional regulation in response to IFN-a. This site binds ISGF-3, consisting of a Statl/Stat2 heterodiiner plus p48 ( F u et cd., 1990). All three proteins in the complex appear to contact DNA directly
150
TIMOTHY HOET AND MI(:HAEL
J.
GHUSRY
(Qureshi et al., 1995). The second class of cis-acting sites was originally defined as IFN-?-activated sequence (GAS) elements. These sites bind homodimers of Statl (Shuai et al., 1993). Subsequently, it was shown that Statl, Stat3, Stat4, and Stat5 can all recognize the same site. This binding site is a palindrome with the sequence TTCNNNGAA ( N 3 site). The highest affinity sites contain CRG ( R = C or G) as the central three nucleotides (Horvath et nl., 1995). IL-4-regulated gene transcription is mediated by a closely related sequence, TTCNNNNGAA (N4 site), that differs in the spacing between the inverted repeats. This binding site corresponds to the optimal recognition site for Stat6 (Schindler et al., 1995). Thus, for STAT hornoQmers, only Stat6 clearly has a distinct specificity and can selectively recognize the IL-4 response element. Given that most STAT proteins bind to a similar DNA sequence, the question arises of how specificity in transcriptional regulation is achieved. Analysis of STAT-binding sites in various promoters has indicated that these sites are often clustered together (Guyer et nl., 1995; John et nl., 1996; Xu et al., 1996). The proximity of the sites allows dimer-dimer interactions that enable STAT proteins to selectively recognize sites that are diverged from the optimal consensus sequence and that are individually lower in affinity (Vinkemeier et al., 1996; Xu et al., 1996). Cooperative binding interactions are mediated by the association of adjacently bound diiners through the N-terminal region of the proteins (Vinkemeier et al., 1996; Xu et d., 1996). Analysis of the crystal structure of the N-terminal domain of Stat4 indicates that this region is composed of eight helices that form a hook-like structure, which appears to mediate dimer-dimer interactions (Vinkemeier et al., 1998). This domain is highly conserved among the STAT family, being between 25 and 50% identical among the six STATs. The N-terminal domain is not essential for dimerization or for binding to single high-affinity site, however (Xu et al., 1996). Interestingly, the sequence of Stat6 in this region is the most diverged among the STAT family. Because Stat6 has the unique ability to recognize the N4 consensus sequence, this protein apparently does not use the mechanism of cooperative binding to adjacent sites to achieve selectivity.
E. TRANSCRIPTIONAL ACTIVATION Subsequent to DNA binding, the STAT proteins function to activate or, in some cases, repress gene transcription. The STAT transcriptional activation domain was first identified as the C-terminal domain in Statl and Stat3 (Wen et al., 1995; Bromberg et al., 1996). The corresponding region also functions in transcription activation in the other STATs (Mikita et al., 1996; Moriggl et al., 1996; Qureshi et al., 1996). Interestingly, this region of the protein is deleted in the STATlP form, which is derived
STAT\
151
from an alternatively spliced form of the Stat 1 transcript. Tlie expression of two naturally occnrring isoforms of STAT proteins that either have (a form) or lack ( p forin) tlie C-terminal domain has also been found for 1996), StatFj (Wang et nl., StatS (Schaefer rt nl., 1995; Caldenlioven clt d., 19961, and Stat6 (Patel et a/., 1998). Generally, the /3 forms of the STAT nioleciiles have been shown to act as dominant negative inhibitors of their a forin counterparts. Statl and Stat3 proteins can also lie serine phosphorylated in their C-terminal domains, and this modification appears to positivily regulate tile potency of the activation domain (Wen ct a/., 1995; Zhang et d., 1995). At present, however, tlie precise niechanisms by which STAT proteins mediate transcriptional activation remain to be elucidated. In addition to the protein interactions between STAT dimers described earlier, interactions of STAT proteins with other transcription factors also occur and, in inany cases, are both critical for transcriptional activation and promoter selectivity. Tlie most well-studied case is the ISGF-3 complex, consisting of StatllStat2 and p48. The domain of Statl that interacts with p48 has been identified as the region between amino acids lS0 and 2SO (Horvath et nl., 1996). Although a StatUStat2 heterodimer alone can hind to a palindrornic N 3 sequence characteristic of the GAS sites and direct IFN-a-induced transcription (Li ct d., 1996), interaction with p48 alters the DNA-binding specificity of the complex such that it now binds to the ISRE site. p48 is a member of the IRF fhmily of transcription factors (Harada et nl., 1989).These proteins have been implicated in many aspects of iinrnune function, including cytokme signaling arid cellular proliferation. Thus. it is tempting to speculate that other members of the IRF gene family may interact with STAT proteins in a manner similar to p48. In iddition to p48, STAT proteins have been shown to interact functionally with several other classes of transcription factors. For example, Statl and Stat2 have been shown to associate with the transcriptional coactivator p300/CBP (Bhattachaiya et d . ,1996; Zliang vt a / ., 1996a).Two interaction sites on Statl, comprising either the N- or the C-terminal domains, were defined for its interaction with p3OO/CBP, although a transcriptional activation fiinction for the N-terminal domain had not previously been reported. In tlie case of State, the C-terniinal domain w a s found to be the site for p300/CBP interaction. Interestingly, the Stat2 activation domain is unique among tlie STAT family in that it is rich in acidic amino acids. I n contrast to its interaction with CREB or jun proteins, the binding of p300/CBP to Statl or Stat2 does not appear' to be dependent on phosphorylation of the STAT protein. Finally, studies with pSOO/CBP have indicated that these proteins have histone acetylase activity (Bannister and Kouzarides, 1996; Ogryzko et ol., 1996). Acetylated histones have been associated with
1.52
TIMOTIIY HOEY AND blI(:I-IAEL J. CRUSBY
active genes and an “open” chromatin configuration (reviewed in Turner, 1993).Thus, it may be that this chromatin remodeling activity allows access to the promoter for other transcription factors and is the result of synergistic interactions between STAT proteins and other transcriptional regulators. An interaction between c-jun and Stilt3 was discovered through a twohybrid screen to look for proteins that can bind the N-terminal domain of c-jun (Shaefer et nl., 1995). Stat3 can stimulate expression of the IL-6 response element in the a2-macroglobulin promoter by worhng together with c-jun. Interestingly, Sliaefer et $. (1995) isolated an alternatively spliced version of Stat3 that lacks the C-terminal transcription activation domain. This /3 form of Stat3 was more effective than the full length when tested in combination with c-jun, and it may be that the short form of Stat3 is specialized for c-jun interaction. Stat5 was found to interact functionally with glucocorticoid and progesterone receptors (Stocklin et nl., 1996).These proteins mediate signaling in response to prolactin and steroid hormones and are involved in activation of the &casein promoter. Similarly, Stat3 binds to, and enhances, the transcriptional activity of the glucocorticoid receptor (Zliang et nl., 1997). Finally, IL-4 response elements are bound by Stat6 and c/EBP proteins (Hou et d , 1994; Delphin and Stavnezer, 1995).Activation of transcription in response to IL-4 from these composite elements requires the transcription activation domain of both proteins, and they appear to interact on DNA to stabilize each other’s binding (Mikita et nl., 1996).A theme that has emerged from these studies is that STAT proteins are often involved in combinatorid interactions with other transcriptional activators. Indeed, in several cases it appears that STAT proteins may not be able to maximally activate gene transcription independently of these interactions. For example, Statl synergistically activates the ICAM-1 promoter in coinbination with Spl; mutation of the Spl site eliminates IFN-?-inducible transcription without affecting Statl binding (Look et al., 1995). IV. STAT-Deficient Mice
With the exception of Stat2, each of the STAT genes has now been functionally inactivated in mice by gene targeting (Table 11). Perhaps the most important conclusion that can be drawn from the collective phenotypes of the STAT-deficient mice is that, in viva, STAT activation seems to be biologically critical for only a select group of cytokines. This is in marked contrast to the in uitro observations demonstrating that almost every known cytokine can activate one or more STAT proteins. For example, a number of ligands other than the IFNs, including IL-6 (Lamer et nl., 1993), IL-10 (Finblooin and Winestock, 1995), and epidermal growth
STAT\
153
factor (Fu and Zhang, 1993), had been sliown, in citro, to lead to the activation of Statl on ligand binding to their specific cell surface receptors. Despite these prior observations, no defects in the in uico responsiveness of Statl-deficient mice to these cytokines and hormones were noted. Siinilarly, leptiri was shown to activate Stat6 in uitro (Chilardi et d., 1996),yet Stat6deficient mice are not obese. Thus, even if these ligands do activate thc Jak-STAT pathway in zjico, then clearly they must be able to activate other biologically relevant signaling pathways as well.
A. STAT~-DICFICIENT MI<E The Statl gene was the first member ofthe STAT family to be disriiptetl d.,1996). Statl-deficient aninials are developmentally nornial and are Iiom at thc expected Mendelian frequencies. Perliaps not unexpectedly, the predominant phenotype of Stat 1deficient mice is the complete lack of responsiveness to both IFN-a and IFN-7. As a resiilt, the transcriptional activation of IFN-inducible genes such 21s IRF-1 a i d MHC class I and class I1 molecules is coiripletely impaired in the mutant mice. The functional consequence of this lack of IFN responsiveness is that Statl-deficient animals are highly susceptible to infections with viruses such as MEW and VSV and with microbial pathogens such as Listeria ~nonoc!itogc.n~~. What was perhaps most surprising from the analysis of this first STAT knockout was the demonstration of the plivsiologic specificity of the JakSTAT signaling pathway. As discussed earlier, the analysis of thcw animals demonstrated that while Statl is critical for inediating biologic responses to IFNs, activation of this transcription factor is clearly not necessary for the responses to a number of other ligaiids. Whetlier the activation of Statl in vitro hy ligands other than IFNs is simply an artifact of nonphysiologic in mice (Diirbin ~t nl., 1996; Meraz ct
154
TIMOTIIY HOEY AN11 MICIIAEL J GRUSBI
stimulation or a minor component of the biologic response to these stimuli remains to be determined.
B. STAT3-DEFIC:IENT MICE Stat3 deficiency leads to embryonic lethality between 6.5 and 7.5 days of gestation (Takeda et al., 1997). Although it is not clear why Stat3deficient embryos fail to complete development, it is interesting to note that a number of different cytokines and growth factors important in growth and differentiation (LIF, EGF, CNTF, G-CSF, etc.) activate Stat3. In this regard, the period of embryonic death observed in Stat3-deficient mice is earlier than that seen in mice deficient in receptors for LIF or CNTF or even the common signal transduction receptor gp130, suggesting that Stat3 deficiency may represent the combined effect of a loss in the receptor signaling of two or more critical growth factors.
c. STAT4-DEFICIENT MICE 1L-12 is the only known ligand capable of activating Stat4, and thus not suprisingly, Stat4-deficient mice show impaired IL- 12-induced responses (Kaplan et al., 1996b; Thierfelder et al., 1996). Thus, IL-12-induced increases in IFN-.), production, cellular proliferation, and natural killer cell cytotoxicity are all abrogated in the absence of Stat4. Moreover, the differentiation of IFN-.), producing T h l cells is markedly impaired in Stat4deficient inice. One group (Kaplan et al., 1996b) noted that this was accompanied by the enhanced development of Th2 cells, consistent with the phenotype of IL-12R-deficient mice (Magram et al., 1996). Interestingly, Stat4 has a very restricted tissue-specific distribution of expression, with high levels expressed in a developmentally restricted manner in the testis (Herrada and Wolgemuth, 1997). Nevertheless, Stat4 deficiency has no discernible effects on spermatogenesis and Stat4-deficient male mice are fertile. D. STAT5-DEFICIENT MICE Perhaps the most surprising finding from the Stat5a and Stat5b knockout mice is their distinct phenotypes, demonstrating that, despite their sequence similarity, these two STAT proteins have distinct functions. Stat5adeficient mice show impaired lobuloaveolar mammopoiesis and females fail to lactate after parturition, consistent with an inability to respond to prolactin (Liu et al., 1997). Although expressed with a similar pattern during mammary gland development, Stat5b clearly cannot compensate for StatFja in maininary gland development. In contrast, Stat5b is required for the sexual dimorphism of body growth rates and growth hormone pulseregulated liver gene expression (Udy et d., 1997).Once again, these results
highlight the degree of biologic specificity in Jak-STAT signaling that has been uncovered in STAT-deficient mice.
E. STATG-DEFKTENT MKT Stat6-deficient mice have been indepentlently generated by three groups, all of which reported similar phenotypes (Kaplan et nl., 1996a; Shimoda et d., 1996; Takeda et nl., 1996b). As mentioned earlier, Stat6 was originally purified and cloned based on its abili? to be activated in response to IL-4 (Hou et a/., 1994), and StatG-deficient mice are niarkedlv impaired in their ability to respond to this cytokine. For example, the expression of genes known to be regulated by IL-4, such a s CD23, IL-4Ra, and MHC class I1 molecules, is not increased following the IL-4 stimulation of Stat6deficient lymphocytes. I n addition, Stat6-deficient mice fiail to generate an IgE response, following either infection with the nematode Nippostrotigylr~sbmsiliensis or immunization with the polyclonal stiinulus anti-IgD. Finally, the differentiation of Th2 cells by Stat6-deficient lymphocytes was f h i d to be niarkedly impaired. Interestingly, however, not all IL4-induced responses are abrogated in Stat6-deficient lymphocytes. For example, IL-4 could rescue Stat6-deficient T cells from undergoing apoptosis (Vella et al., 1997; Kaplan et al., 1998) and probably mediates these effects through an alternate signaling pathway involving the activation of a/., 1996). insulin receptor substrate (Zamorano t The IL-4Ra chain is a shared coinpoilent of both IL-4 and IL-13 receptor complexes and. like IL-4, IL-13 is capable of activating Stat6. Not surprisingly, many IL-13-induced responses are also abrogated i'n Stat6-deficient 1996a) mice, including IL-13-driven TI12 differentiation (Kaplan et d., and IL-13-mediated increases in the expression of MHC class I1 antigens and the production of nitric oxide in macropliages (Takeda ut al., 1996a). V. STAT Function in Cellular Proliferation and Disease
The role of STAT proteins in regulating cellular proliferation has been somewhat controversial, although several lines of evidence now suggest that the Jak-STAT signaling pathway is indeed important for this process. Experinients in which cytokine receptors harboring niutations that prevent the docking of STAT proteins often gave conflicting results, in some cases implicating STAT proteins as being important (Danien et al., 1995) and in other cases dispensable (Qiielle a]., 1996) for a proliferative response. In contrast, the expression of a dominant negative Stat5 protein led to impaired IL-3-induced proliferation (Mui et nl., 1996). Perhaps more iinportantly, lympliocytes from inice deficient in Stat4 (Kaplan et nl., 199613; Thierfelder et al., 1996), S t a t 5 (Nakajima et al , l997), and Stat6 (Kaplan c7t
156
TIMOTHY IIOEY A N D MICHAEL J GRUSAY
et nl., 1996a; Sliimoda et al., 1996; Takeda et d., 1996b) show impaired proliferative responses to IL-12, IL-2, and IL-4, respectively. Interestingly, the failure of IL-2 to induce expression of the high-affinity IL-2Ra chain was shown to be responsible for the proliferative defect seen in Stat5adeficient lymphocytes (Nakajiina et nl., 1997). In contrast, failure of IL-4 to upregulate the expression of the IL-4Ra chain was found not to be responsible for the proliferative defect of State-deficient lymphocytes. In thi5 case, Stat6 was shown to regulate tlie expression of ~ 2 7 ~ ' pthe ' , major cdk inhibitor expressed in lymphocytes (Kaplan et al., 1998). Thus, STAT proteins appear to act at multiple levels to regulate cellular proliferation. Perhaps tlie most compelling evidence that STAT proteins are involved in cellular proliferation is tlie observations of constitutive activation of tlie Jak-STAT signaling pathway following the transformation of cells with certain vinises such as HTLV-1 (Migone et nl., 1995) or oncogenes such as v-Src (Yu et nl., 1995), v-Abl (Danial et al., 1995), or v-Eyk (Zong et d , 1996). Moreover, cell lines from a number of patients with lymphoinas and leukemias were sliown to have constitutively activated Jak-STAT activity (Zhang et nl., 1996b; Cliai et al., 1997; Takemoto et al., 1997). Finally, STATs also appear to be important in mediating the antiproliferative effects of IFN-.)I, and do so by regulating the induction of the cdk inhibitor p21""' (Chin et nl., 1996). Interestingly, this activity of Stat1 may be part of the mechanism responsible for thanatophoric dysplasia type I1 dwarfism, where mutations in fibroblast growth factor receptor 3 lead to constitutive tyrosine kniase activity, activation of Statl, and increased expression of p21CLAF' in cartilage from patients with this disease (Su et nl., 1997). VI. Regulation of STAT Function
As outlined earlier, STAT proteins play an important role as regulators of gene transcription in a variety of cellular processes. Thus, it is perhaps not suprising that a number of different mechanisms appear to exist to negatively regulate STAT function. One of these, the generation of alternatively spliced forms that lack a region in the C-terminal domain and thus act as dominant negatives, has already been discussed, although it is at present unclear how the relative levels of the various spliced forms of the STAT molecule are regulated in any given cell. Evidence also shows that the action of an a s yet unidentified pliosphatase results in the dephosphorylation of the critical tyrosine residue required for STAT dimerization and DNA binding (Haspel et nl., 1996; Shuai et nl., 1996). Finally, it has been demonstrated that the degradation of STAT proteins through either the
STAT,
157
1997) or proteosome pathway (Kim and Maniatis, 1996; Rayanade ct d., the action of caspases (King and Goodburn, 1998) can also occur. A novel mechanism for regulating STAT activity has emerged with the identification of two new classes of STAT inhibitors. The first is the CIS/ SOCS/JAB/SSI class of cvtokine-indncible STAT inhibitors that appear to be aide to inhibit STAT activation 1iy binding to either the pliosphorylated Jak kinase or the phosphorylated cytokirie receptor (Yoshimuraet aZ., 1995; Endo ct nl., 1997; Naka ct al., 1997; Starr el nl., 1997). Some 20 proteins containing the so-called SOCS box have now been identified (Hilton et al., 1998), and it will now tie of great interest in determining the exact ~nechanismby which these proteins inhibit specific Jak-STAT signaling events. A second class of proteins, the protein inhibitor of activated STAT (PIAS),has also been identified (Chung ct al., 1997).Two members, PIASZ and PIAS3, have been characterized and shown to interact specifically with Stat1 and Stat3, respectively. Tliur, the PIAS molecules appear to represent specific STAT inhibitors, ;tiid thcre will undoubtedly be great interest in exploring their potential a s tlierapeutic targets. Finally, there are now several exainpks of STAT proteins competing with other transcription factors for overlapping DNA-binding sites. Thus, the Ets fknily protein Elf-1 represses IL-2Ru gene transcription by binding to an Ets-binding site that overlaps with a StatS-binding site (Lecine ct d., 1996). In addition, BCL-6 appears to compete with Stat6 binding to several target genes as BCL-6-deficient mice exhibit augiiiented IL-41997). Finally, Stat6 induced Th2 responses (Dent ct a / . , 1997; Ye et d., can antagonize the binding of other transcription factors, such as NF-KB 1997). suggesting one niedianisin in the E-selectin gene (Bennett et d., by which STAT proteins can act a s transcriptional repressors. VII. Summary and Perspective
In inany respects, it is remarkable the degree to which we understand the biology and biochemistry of the STAT proteins given their recent discovery. Nevertheless, there are still inany important issues related to Jak-STAT signaling that remain to be addrerssed. First, additional structural information of the individual domains in the STAT molecule will be very informative. For example, the STAT SH2 doinuins represent a distinct subfamily relative to the src-related SH2 domains, and therefore it is likely the STAT SH2 domains will have wine unique structural features. Because STAT DNA-binding domains do not show any obvious sequence siinilarity to other DNA-binding motifs, this structure will be the first of its class. Perhaps the most interesting aspect of the structure is that it will allow us
158
TIMOTHY IIOEY AND MICIIAEI,
1
GRUSRY
to gain insight into the interaction of the different domains and how these interactions change during the transition from monoiners to dirners. In addition to structural information, there are inany questions regarding the functions of the STAT family that remain poorly understood. One of the hallmarks of cytokine signaling is the interplay between different signaling pathways. Further investigation of the mechanisms of activation, and inactivation, of STAT proteins will be critical in obtaining a more detailed understanding these processes. In particular, the inechanisrns of transcriptional activation by STAT proteins and their synergistic interactions with other transcription factors will be very important to pursue. Finally, the analysis of STAT-deficient mice has revealed that these proteins are essential for various cytokine signaling pathways during the iininune response. Therefore, we can also look forward to the development of antagonists of STAT function that may prove to be useful in therapeutically modulating immune responses.
REFERENCES Akira, S., Nishio, Y., Inoue, M., Warig, X-J., Wri, S., Matsusaka, T., Yoshida, K., Sudo, T., Nanito, M., iind Kishimoto, T. (1994). Cell 77, 63-71. Baiinister, A. J., and Kouzarides, T. (1996). Il’nttcre 384, 641-643. Bennett, B. L., Cruz, R., Lacson, R. G., a i i c l Manning, A. M. (1997). J , B i d . Chon. 272, 10212-10219. Bhattachalya, S., Eckner, R., Grossinan, S., Oldread, E., Arauy, Z., D’Andrea, A,, and Livingston, D. M. (1996). Naturc, 383, 344-:347. Broinberg, J. F., Morvatli, C. M., Wen, Z.,Srhreiber, R. D., and Damell, J. E. (1996). Proc. N d Accid. Sci. USA 93, 7673-7678. Caldenhoven. E., van Dijk, T. B., Solari, R., Arinstrong, J., Raaijiiiakers, J. A. M., Lainniers, J. J.. Koenderman, L., and de Groot, R. P. (1996). /. Biol. Cherri. 271, 13221-13227. C h i , S. K.. Nichols, G. L., arid Rothmaii, P. ( 1997).1.Imnirrttol. 159, 4720-4728. C h i , Y. E., Kitagawa, M., Su, W.-C. S., You. Z.-H., Iwamoto, Y., and Fu, X.-Y. (1996). Science 272, 71 9-722. Chiuig, C. D., Liao, J., Liu, B., Rao. X.. Jay, l’., Berta. P., and Shuai. K . (1997). Science 278, 1803-1805. Copeland, N. G., Gilbert, D. J.. Schinder, C., Zhorig, Z., Wen, Z., Darnell, J. E., Jr,, Mui, A. L.-F., Miyajima, A., Qiielle, F. W., Ilile, J. N., and Jenkins, N . A. (1995). Genomic:r. 29, 225-228. Damen, J. E., Wakao, H., Miyajinia, A., Krosl, J., Humphries, R. K.. Cutler, R. L., arid Krystal, G.(1995). EMBO 1. 14, 5557-5568. Danial, N. N.. Peruis, A., and Rothinan, P. B. (1995).Science 269, 1875-1877. Damell, J. E., Jr. (1997). Science 277, 1630- 1635. Delphin, S., and Stavnezer, J. (199s)./. E x p Med. 181, 181-192. Dent, A. L., Shaffer, A. L., Yu, X., Allinan, D., arid Staudt, L. M. (1997). Scirrice 276, 589-592. Durbin, J. E., Hackenmiller, H., Sinion, M. C . , and Levy, D. E. (1996). Cell 84, 443-450. Endo, T. A,, Masuliara, M., Yokouchi, M., Suzuki, R., Sakamoto, H., Mitsui, K., Matsumoto, ,4., Taniinura, S., Ohtsubo, M.. Misawa, H., Miyazaki, T., Leonor, N., Taniguchi, T., Fujita, T., Kanakura, Y., Komiya, S., and Yosliimura, A. (1997). Nature 387, 921-924.
STAT.;
Finblooni, D. S.. ant1 Winestock, K. D. (1995).J. I i i i t t w i i o l . 155, 1079-1090, Fu, X.-Y. (1992). Cell 70, 323-335. Fu. X-Y.,Kessler, D. S., \'eals, S. A.. I,e\y, 11. E.. i i ~ d11:irneIl. J. E., Jr. (1990). Proc. Nod. Acntl. Sei. USA 87, 8555-8559. Fu, X-Y., Schinder. C . , lmprota. T., Ael)ersold, K., and Dal.nell, J. E., Jr. (1992). Proc. Not!. Acad. Sci. USA 89, 7840-7843. Fu, X.-Y., and Zliarig, J.-J. (1993).Cell 74, 1135-1145. Cliilardi, N., Zieglrr. S., Wiestiier, A., Stoffel. R.. Hriin, M . H., and Skoda. R. C. (1996). Proc. N d .Actid Sci. U S A 93, 6231-62:35. Greerilund, A. C., Farrar, M . A.. \'iviaiio, B. L., arid Schreiber, R. D. (1994). EMBO /. 13, 1591-1600. Greenluntl, A. C . , Morales, M . O . , Viviaiio, 13. L.. Yan, Y.,Krowleski, J., and Schreiber, R. D. (1995). Z r i i i i i i r i i i t / j 2, 677-687. Guyer. N. B., Sevcms, C . W., Wong. P.. Fengliali, C. A , , antl Wright, T. M . (1995). J . ~ l l i l l i l L l f ~ J155, ~. 3472-3480. Harada, H., Fujita, T.. Miyamoto, M., Kiinuni, Y.,Mariiyima, M . , Furia, A , , Miyata, T.. and Taiiiguchi. T. (1989). Cell 58, 729-739. IIaspel, R. L.. Salditt-Georgieff, M., and Ilarnell, J. E.. Jr. (1996).EhfBOJ. 15, 6262-6268. Heim, M . H., Kerr, I. M., Stark. G. R., antl Darnell, J. E.. Jr. (1995).Scierice 267, 1347-1349, Herrada, G., and Wolgemiith. D. J. (1997).J , Cell. Sci. 110, 1543-1553. Hilton, D. J.. Richardson, R. T.. Alcaxaiitier, \V. S., \'iney, E. M.. Willson, T. A , , Sprogg, N.S., Stair, R., Nicholson. S. E., Metcalf, I)., aiid Nicola, N . A. (1998).Proc. Ntztl. Acad. Sci. USA 95, 114-119. Howath, C. M., Stark, C. R., Kerr, 1. M., and D;iniell, J. E., Jr. (1996). Mol. Cell. B i ~ i l . 13, 3951-3963. Horvath, C. M . , U'en. Z.,and Ilarnell, J. E., Jr. (199.5).Geiics Deo. 9, 984-994. Hou, J.. Schinder, U., Herizel, W.J., Ilo, T. %., Brsseiir, M.,and McKnight, S. L. (1994). Scietice 265, 1701-1706. Hou. J., Schiiidler, U.. Henzel, W. J.. M'ong, S. C., and McKnight, S. L. (1995).Zttimuriily 2, :321-:329. Ihle, J. N. (1996). Cell 84, 3:31-:334. Ihle, J. N.. Witliuln. B. A,, Quelle, F. \V.,Yarnainoto. K., and Silvennoinen, 0. (1995). Aiiiifr, Hec. Zniriiunol. 13, :369-398. John, S.,Robbins, C. M . , and Leonard. M'. J. (1996). EMBO J 15, 5627-5635. Jones, N . (1990). C d l 61, 9-11. Kaplaii, M. €I.,Daniel, C., Schintller, U.. antl Grnshy, M . J. (1998).Mol. Cell. B i d . 18, 19962003. Kaplan. M . H., Schiridlrr, U., Sriiilty, S. T., a r i d Gmshy, M . J. (l996a).Znmrrnity 4,313-:319. Kaplan, bl. H., Sun, Y.-L., Hory. T., antl Grusly, M . J. (1996b). Ntitrrm 382, 174-177. Kawatn. T., Shevclic~iiko.A,, Fukrizawa, M.. Jerinyn. K. A., Totty, N. F., Zliukovskaya, N. \',, Sterling, A. E., Mann, M., and i\'illiuiis, J. G. (1997). Cell 89, 909-916. Kiln, T. K.. and Maniatis, T. (1996). Sciriice 273, 1717-1719. King, P., a i d Goodbum, S. (1998).J. B i d . Clioii. 273, 8699-8704. Koch, C. A , , Anderson, D., Moran, M . F., Ellis, C,, arid Pawson, T. (1991). Sciericc 252, 668-674. Kumar, A , . Coniriiane. M., Flickingrr, T. \V,, H o n d i , C . M., and Stark, G. R. (1997). ScieticP 278, 1630-1632. Lamer, A. C.. Ilavid, M . , Feldman, G. M . , Igarashi, K.-I., Hackett. R. H., Webb, D. S. A . , Sweitzc,r, S. M., Pc~tricoin,E. F., 111. and Finbloom, D. S. (1993). Scieiice 261, 1730-1733.
160
TIMOTHY IIOEY AN11 MICHAEL J. GRUSBY
Lecine, P., Algarte, M., Rameil, P., Beadling, C., Buchler, P., Nabholz, M., and Imbert, J. (1996).Mol. Cell. Biol. 16, 6829-6840. Leonard, W. J., and O’Shea, J. J. (1998). Atinti. Reo. Zrnrnutd 16, 293-322. Li, X., Leung, S., Qureshi, S., Damell, J. E., and Stark, G. R. (1996). I. Biol. Chem. 271, 5790-5794. Lin, X., Robinson, G. W., Gouilleux, F., Groner, B., and Hennighausen, L. (1995). Proc. Natl. Actid. Sci. USA 92, 8831-8835. Liu, X., Robinson, G . W., Wanger, K.-U., Garrett. L., Wynshaw-Boris, A,, and Henighamen, L. (1997). Gcnrs Dev. 11, 179-186. Look, D. C., Pelletier, M. R., Tidwell, R. M., Roswit, W. T., and Holtznian, M. J. (1995). 1. B i d . Chem. 270, 30264-30267. Magrarn, J., Connaughton, S. E., Warner, R. R., Carvajd, D. M., Wu, C.-Y., Ferrante, J., Stewart, C., Sarmiento, U., Faherty, D. A,, and Gately, M. K. (1996).Zmnitcnity 4,471-481. Meraz, M. A,, White, J. M., Sheehan, K. C. F., Bach, E. A,, Rodig, S. J., Dighe, A. S., Kaplan, D. H., Riley, J. K., Greenhind, A. C., Campbell, D., Carver-Moore, K., DuBois, R. N., Clark, R., Agnet, M.. and Schreiber, R . D. (1996). Cell 84, 431-442. Migone, T.-S., Lin, J.-X., Cereseto, A,, Mulloy, J. C., O’Shea, J. J., Franchini, G., and Leonard, W. J. (1995). Science 269, 79-81. Mikita, T., Campbell, D., Wu, P., Williamson, K., and Schintller, U. (1996).Mol. Cell. B i d . 16,581 1-5820. Moriggl, R., Gouilleux-Gruart, V., Jahne, R., Berchtold, S., Gartman, C., Liu, X., Hennighansen, L., Sotiropoulos, A,, Groner, B., and Gouillex, F. (1996).Mol. Cell. Biol. 16,56915700. Mui, A. L.-F., Wakao, I-I., Kinoshita, T., Kitamnra, T., and Miyajima, A. (1996).EMBO J . 15, 242552433. Mni, A. L.-F., Wakao, H., O’Farrell, A,-M., Harada, N., and Miyajirna, A. (1995). E M B O J . 14, 1166-1175. Naka, T., Narazaki, M., Hirata, M., Matsumoto, T., Minamoto, S., Aono, A., Nishirnoto, N., Kajita, T., Taga, T., Yoshizaki, K., A h a , S., and Kishirnoto, T. (1997). Nature 387, 924-929. Nakajiina, H., Liu, X.-W., Wynshaw-Boris, A., Rosenthal, L. A., Imada, K., Finbloom, D. S., Hennighausen, L., and Leonard, W. J. (1997). Znirrwnity 7, 691-701. Ogryzko, V. V., Schlitz, R. L., Russanova, V. Howard, B. H., and Nakatani, Y. (1996). Cell 87,953-959. O’Shea, J. J. (1997). Zntinuriity 7, 1-11. Patel, B. K. R., Pierce, J. H., and LaRochelle. W. J. (1988). Proc. Natl. Acad. Sci. USA 95, 172-177. Quelle, F. W., Shinioda, K., Thierfelder, W., Fischer, C., Kim, A,, Ruben, S. M., Cleveland, J. L., Pierce, J. H., Keegan, A. D., Nelins, K., Paul, W. E., and Ihle, J. N. (1995). Mol. Cell. Biol. 15, 3336-3343. Quelle, F. W., Wang, D., Nosaka, T., Thierfelder, W. E., Stravopodis, D., Weinstein, Y., and Ihle, J. N. (1996). Mol. Cell. B i d . 16, 1622-1631. Qureshi, S. A., Leung, S., Kerr, I. M., Stark, G. R., and Darnell, J. E., Jr. (1996).Mol. Cell. B i d . 16, 288-293. Qureshi, S. A., Salditt-Georgieff, M., and Darnell, J. E., Jr. (1995). Proc. Nntl. Acad. Sci. USA 92, 3829-3833. Rayanade, R. J., Patel, K., Ndubnis, M., Sharnia, S., Omura, S., Etlinger, J. D., Pine, R., and Sehgal, P. B. (1997).J. Biol. Chem. 272, 4659-4662. Schindler, C., and Damell, J. E., Jr. (1995). Armu. Reo. Biochem. 64, 621-651.
STAT\
161
Schindler, C., FII.X.-Y., Improta, T., Aebersold, R., mid Ilarnell, J. E., Jr. (1992a).Proc. N ~ 1 1 Actrd. . Sci. l J S A 89, 7836-7839. Scliiiidler, C., Shuai, K., Prezioso, 1‘.R., and Darnrll, J. E., Jr. (19921)).Scicvicr 257,809-813. Schindlrr. U.,R’II.P.. Rothe, M., Brasseur. M., iciitl McKniglit, S. L. (1YY5). Ztrttnrinif!y 2, 689-697. Sekiiiioto. T., Iiiiaiiioto, N., Nakajiina, K.. Hiraiio, T., iuitl Yoneida, Y. (1997). EMBO J. 16, 7067-7077, Sekiiiioto, T., Nakajinia, K.. Trac1iilxiii:i. T., liiriiiw, T., a d Yoneda, Y. ( 1996). I. B i d . C l u w . 271, 31017-:31020. Shaefer. T. S., Sanders, L. K.. and Nithaiis. D. (19951. Pntc. NafI. Aca4/. Sci. USA 92, 90979101. Shimoda. K., van Deursen, J.. Sarigster. M. I’.,Sariiwar, S. R., Carson, H. T., Tripp, R. A,, Chii, C . , Quelle. F. W., Nosaka, T.. Vigiidi. D. A . A,, I>ohert).,P. C., Grosveld, C . , Paul, \l’ E.,.antl IIilv, J. N . (1996). Notrirc 380, 6:30-63:3. Shuai. K..Liao. J.. ;ant1 Song. M . M . (1Y96). Mo/. CON.Biol. 16, 4932-4941. Shuai, K., Schindler, C., Presiozo. 1’. K., and Ilarncll. J. E., Jr. (1992).Science 258, 18081812. Shiiai. K..Star-k, C . l i . ~Kerr, I. A., ;ind Dtiniell, J. E., Jr. (1993). Science 261, 1744-1746, Sonpmg, Z.,S h o e h i , S. E.. Chaudhiiri. M., Gish. G., Paiwson, T.. liaser. W.G., King, F.. Roberts. T., Ratnofshy. S.. Leclileitlrr. R. J., Ned, H . G., Birge, R. B.. Fajardo, J. E., Chou,M. M.. Hanafusa. H.. Schaffliausen, B., and Cantle), L. C. (1993).Cell 72, 767-778. Stahl. N., Farruggella, T. J , , Uorilton, T. C.. Zhoiig, %., Damell. J. E.. antl Yancopoulos, C:. D. (1995). Scicttrx. 267, 1:349-135:3. Starr, R., Willson, T. A., Vine). E. M., Murray, L. J . I,., Rayiier, 1. R., Jenkins, B. J., Gonda, T. 1.. Alexander, 11;. S., Metcalf. D.. Nicola, N . A,, and Hilton, D. J. (1997). Notrire 387, 917-921. Stocklin, E., Wissler. M., Coilleiis, F.,i n i d Gronei-. B. (1996). Nnfrcr(7 383, 726-728. Su, 11’.-C. S.. Kitagawa. M.. Xiw. N., XY, B., Gm)f’;tlo,S.. Clio. J.. Deng, C., Horton, \V. A , . ant1 FII. X.-Y. (1997). Nature 386, 288-292. Takedii, K., K;iinanaka, M.. Tanaka, T., Kisliimoto. T., and Akira, S. (199Ga).J . Ztwnitnol. 157, 322O-m2. Takeda. K..Nogriclii, K., Slii. U’., Tmaka. T., Matsunioto. M., Yoshitla, N.. Kishiiiioto, T.. and Akira, S. (1995). Proc. Not[. Accid Sci. LISA 94, 3801-3804. Takeda. K., Taiiakn, T.. Shi, \%’.-Matsrinioto, M., Minaini. M., Kasliiwamiira, S., Nakanishi, K., Yoshitla, N.. Kishiinoto. T., a i d Akira, S. (lCJ96h). Notrirc. 380, 625-630. Takemoto, S., Miilloy, J . C., Cereseto, A , . Migone. T.-S.. l’atel, B. K. H.. Matsuoka, M.. Yamaguchi, K., Takatsuki, K.. Kmiiihira. S.. \t’liit(~, J. D., Leonard, \V. 1..M’aldinanii, T., and Franchirii. G. (1997). Proc. .\’of/. .4cvtl. Sci. U S A 94, 1:3897-13902. Thierfdtler, \V. E.. van Deurseii. J. M., l’ainamoto. K,, Tripp, H. A,, Sarawar, S.K.. Carson, R. T.. Sangster, M. Y., \’iguali, D. A. A , . I>olre%. P. C.. Grosveld, G.C.. and Ihle, J. N. (1996). Nntrirc 382, 171-174. Turner, B, M. (199:3). Cell 75, ,543, Udy, (;. B., Towers. K. P., Snell. R. C., \b’ilkins. R. J., Park, S.-II., Rain, P. A,. M’a?cman, D. J.. and Dave?, H . CV. ( 1907). Yroc. NotI. Actill. Sci. U S A 94, 7239-7244. \’ella, A,, Tcague, T. K..Ihle, J., Kapplrr. I., and Marrack. I‘. (1997)./. Eq,. Med. 186, 32S-:3:3O \’inkciiic-ier, U..rand autoimmune disease similar to the human disese systemic lupus erythematosus (SLE). The mutations l p + g (allelic to lpr) and gld (generalized lymphoproliferative disease) cause very similar diseases. In all three cases, aberrant T cells accumulate; these T cells are Thy-1 positive, CD4 and CD8 negative (“double-negative” cells), polyclonal (Y and j3 T-cell receptor positive, and B220 positive. In lpr mice, insertion of a retroviral early transposable element (ETn) into intron 2 of the CD95 gene causes a splicing defect, premature termination, and a greatly reduced expression of CD95 mRNA. In 1pTLX mice a point mutation (isoleucine to asparagine) in the intracellular “death domain” of CD95 abolishes transinision of the apoptotic signal. In gld mice a point mutation in the C terminus of CD95L impairs its ability to interact successfully with CD95 to cause apoptosis. Thus, a failure of apoptosis accounts for the complex immune disorder in lpr and gld mutant mice (for reviews, see Krammer et nl., 1994a,b; Nagata and Golstein, 1995). In humans a similar disease with a dysfunction of the CD95/CD95L system has been reported (Fisher ct al., 1995: Rieux-Laucat et al., 1995). Children with this “autoimrnune lymphoproliferative syndrome” (ALPS) show massive, nonmalignant lymphadenopathy, an altered and enlarged T-cell population, and a massive autoimmune disorder. Many of these children show a crippling mutation in the death domain (DD) but are heterozygous for this defect. Because the parents are not sick, whereas the children are, a secondary as yet unknown defect must exist that is responsible for the appearance of the symptoms. The pathology of Ipr and glrl mutant mice is characterized by the accumulation of aberrant T cells and autoantibody production by B cells. Autoantibody production is the direct result of a defect in the B cells themselves. Transplantation experiments in which T cells could “help” normal or lpr mutant B cells led to the exclusive production of lpr autoantibodles. Data imply that the lpr defect affects B cells directly. Breeding experiments showed that the development of the complete lpr pathology is strongly dependent on the genetic background of the lpr mutation. In some cases, autoimmunity can occur in the absence of lymphadenopathy, which reinforces the view that autoantibody production is not a consequence of excess
168
PETER F I . KRAMMER
aberrant helper T cells. Thus, autoinirnunity and lymphadenopathy are distinct aspects of the lpr pathology, due to B-cell and T-cell defects, respectively (Krammer et nl., 1994a,b). The finding that Ipr and gld mutations are defects in the CD95/CD95L system has helped greatly in determining the physiological role of CD95mediated apoptosis in the immune system. In the normal lymphoid system, apoptosis occurs in primary lymphoid organs, such as the bone marrow, liver, and thymus. There it is used to eliminate useless precursor cells with nonrearranged or aberrantly rearranged nonfunctional antigen receptors. In addition, apoptosis is essential for the deletion of autoreactive T cells in the thymus. This mechanism, therefore, is the basis of central selftolerance. In lymph nodes and the spleen, apoptosis causes deletion in T and B cells. Peripheral deletion by apoptosis is a second line where the immune system establishes self-tolerance and where it downregulates an excessive immune response. Only lymphocytes that survive this process may determine immunological memory. The mechanism of peripheral deletion was previously unknown and has now become elucidated, at least in part. Data show that the CD95/CD95L system contributes substantially to the elimination of peripheral lymphocytes. V. Role of the CD95/CD951 System in Deletion of Peripheral T Cells
To investigate the role of CD95/CD95L in TCR-mediated apoptosis, malignant Jurkat T cells as a model were triggered via the TCR in the presence of blocking anti-CD95 antibodies. These reagents prevented antiCD3-induced apoptosis. Dexamethasone-induced apoptosis, however, was not prevented. The same findings were obtained with activated human Tcell clones and activated peripheral human CD4 positive T cells. Taken together, these results and those by others in vitro and in vivo suggest that CD95/CD95L are involved specificallyin TCR-triggered apoptosis. T-cell apoptosis occurs as “fratricide” by interaction of the membranebound receptor with the membrane-bound ligand on neighboring T cells that kill each other. TCR-triggered CD95-mediated apoptosis is also found in single Jurkat T cells. A single TCR-activated T cell in the absence of costimulation may autonomously decide to die by apoptosis, employing, at least in part, the CD95 pathway. These results suggest a minimal model in which TCR-induced death in activated T cells involves CD95/CD95Lmediated suicide. Collectively, TCR-induced CD95-mediated apoptosis may occur in several forms: fratricide, paracrine death, and autocrine suicide (Dhein et al., 1995). The CD95/CD95L system, however, is not the only death system that plays a role in the deletion of peripheral T
cells. Data from Zheng et nl. (1995) suggest that late after triggering of the TCK in vitro TNF-RII and TNF dominate over the CD95/CD95L system. Elucidation of the TClUCD95RNF-R death mechanism has shed new light on peripheral T-cell tolerance by deletion, on suppression of the immune response, and on the development of memory in the surviving T lymphocyte pool. It can be predicted that as these phenomena are studied in greater depth, other less well-5tudied death system may coinpleinent the ones already invoked in these processes. In addition, it is conceivable, and in fact recent data would point in that direction, that the CD9S/CD95L system has an essential role for B-cell deletion similar to the one described for T cells. Here, however, B-cell deletion may be brought about by CD95L positive T cells that kill CD9S positive B cells (Rothstein et nl., 1995). This review has already described some of the basic features associated with CD95KD95L in t1i-e immune system, but the CD95/CD9FjL system also has an important role in another organ: the liver. VI. Role of the CD95/CD95L System in liver Homeostasis
The liver belongs to the many nonlymphoid tissues that express CD95. CD95 is expressed in the developing (French, 1996) and the mature (Leithhser, 1993) liver. Primary hepatocytes from mice (Ni, 1994) and humans (Galle, 1995) were demonstrated to be sensitive toward CD95mediated apoptosis in uitro. A physiological role of CD9ij in maintaining liver homeostasis has been suggested ;is mice deficient in CD95 develop increased cellularity and substantial liver hyperplasia (Adachi, 1995). CD9SKD95L might have a less prominent role in apoptotic processes in other organs. This role may not be revealed easily in CD95 mutant or knockout mice due to adaptive differentiation and may show better in conditional organ-specific knockout mice. VII. Signal Transduction of CD95-Mediated Apoptosis
Signal transduction of CD9-5has only been clarified recently, at least in part. Thus, oligoinerization of CD95 was shown to be a prerequisite for the transduction of the apoptotic signal. Several subclasses of the agonistic apoptosis-inducing monoclonal antibody. anti-APO-1, have been used to induce the CD9Fj death signal. The best death-inducing isotype, clearly, was the subtype IgG 3. This IgC subtype has the property of self-aggregationvia its Fc part. Other IgG subclasses of the antibody showed a differential cytotoxic activity in the declining order IgG 3 >> IgG 1> IgG 2a > IgG 21). IgG 2b did not show any activity anymore. Because all antibodies carried the same combining site specificity, the reason for this differential
170
PETER M. KKAMMER
apoptotic activity had to be the different activity of the different Fc parts. IgG 2b anti-APO-1 was inactive because it only dimerized CD95 receptors. With additional cross-linking activity the IgG 2b antibody became as active as the self-oligomerizing IgG 3 antibody. Also, apoptosis was not induced with a F(ab’)?anti-APO-1 but with F(ab’)., anti-APO-1 cross-linked with F(ab’)z anti-Ig light chain. Thus, CD95 dimers do not induce apoptosis, whereas oligomerized CD95 receptors show this activity (Dhein et al., 1992). The most probable structure to transmit an apoptotic signal is a CD95 trimer, as this structure corresponds to the predicted trimer structure of members of the TNF-R superfamily by X-ray crystallography (e.g., LTa! complexed with TNF-R1; Banner et al., 1993). VIII. The Death Domain
The intracellular part of CD95 does not possess any consensus sequence that would have predicted the use of a known signaling pathway. A deletion of 15 amino acids of the C terminus of CD95 was shown to increase CD95-mediated apoptosis. Further deletions inhibited the CD95 signal completely. When the sequence of the intracellular part of CD95 was compared with the one of TNF-RI, a homology region of 68 amino acids could be defined. Moreover, using deletion and point mutagenesis, Tartaglia et al. (1993) defined a region of TNF-RI that was essential for the cytotoxicity mediated by the receptor. This stretch was 80 amino acids long, comprises the domain described by Itoh and Nagata (1993), and was later called the death domain. The DD also contains the valine residue (ValzJ8)that is mutated in IpfK mice and which abolishes the signaling of apoptosis. Only several years later were additional DD-containing receptors isolated (DR3, DR4, and DR5). Also, further DD proteins were found, two of which bind to CD95: FADD (MORT 1; Chinnaiyan et al., 1995; Boldin et al., 1995) and RIP (Stanger et al., 1995). These molecules were all cloned in the yeast two-hybrid system with the cytoplasmic part of CD95 used as bait. The yeast two-hybrid system selects for low-affinity interactions. Therefore, some of these molecules still have to prove their importance in an in vivo system. IX. CD95 Associated Signaling Molecules
Almost all published signaling molecules found to bind to the CD95 receptor have been found by the yeast two-hybrid system with the cytoplasmic part of CD95 as bait. These proteins are shown in Table 111. The two DD proteins FADD and RIP bind to the CD95 D D directly. Overexpression of FADD and RIP causes apoptosis. All other molecules
171
Signaling Molecule
Reference
FADD (MORT11
Cliiriiiniym ct nl (1995), Boldlrl cf a1 (199s)
RIP DAXX FAFl UBC-FAP (UBC9) FAP-1 Sentnn
Stmger et (zI (1995) Ynng et nl (1997b) C l l U P t al (1995) Wnght ct nl (1996) L i t o rt (11 (1995) Okurd cf a1 (1996)
(see Table 111) do not possess a DD. Except for FAP-1, binding of all proteins to CD95 is decreased or abrogated by tlie lpeg mutation. Overexpression of DAXX (Yang et al., 1997), FAF 1, and UBC-FAP increases CD95-niediated apoptosis, whereas a blocking effect has been described for FAP-1 and Sentrin. Except for the DD at the C terminus, no other known domain has been identified for FADD. Interestingly, expression of the DD of FADD alone did not cause suicide of tlie cells whereas expression of the N terminus did. This is tlie reason why this part of tlie niolecule is called the “death effector domain” (DED). In the yeast two-hybrid system, RIP showed affinity to CD95 and to a lesser degree to TNF-€31. RIP has a domain homologous to the protein kinases and its DD alone i5 sufficient to kill a cell. DAXX was cloned recently. It increases apoptosis and activates Jim N-terminal kinases ( J N K ) . Its C-terminal DD binds to the DD of CD95 and TNF-RI. When this part of DAXX is expressed alone it inhibits both apoptosis and JNK activation. UBC-FAP is the human hornologue of “ubiquitin-conjugating enzyme 9”(UBC9) in yeast S. cemisiae. UBC9 controls the cell cycle at G2 to M and regulates the degradation of cyclins. A negative regulatory role has been suggested for tlie C terminus of CD95 because the deletion of the last 15 amino acids of CD95 increases the sensitivity toward CD95-mediated apoptosis (Itoh et nl., 1993). This region of CD95 interacts with FAP-1 (for Fas-associated phosphatase l ) ,a protein phosphatase that is capable of inhibiting CD95-mediated apoptosis. Recent reports, however, could not show the interaction of FAP-1 with CD95 in mice (Cuppen et al., 1997).In addition, deletion of the 15 carboxyterminal amino acids of mouse CD95 did not show any effect on CD95mediated apoptosis. Another molecule capable of inhibiting CD95mediated apoptosis is sentrin, which, like RIP and DAXX, also binds to
172
PETER H . KRAMMER
TNF-RI in the yeast two-hybrid system. FAP-1 showed additional homology to ubiquitin, NEDDS, and Sint3, a S. cerevisiae protein. X. Other Signaling Molecules Invoked in CD95 Signaling
Several other partly classical signaling pathways have been described that are thought to play a role in generating the CD95 death signal. Cerainide is cleaved from membrane sphingolipids by sphingomelinases (SMases). Sphingolipids are molecules that form a central part of the cell membrane. Several authors have reported that for CD95-mediated apoptosis the activities of the acid SMase (Kolesnick et al., 1994; Cifone et al., 1994), as well as of the neutral SMase (Tepper et al., 1995), are essential. In addition, tyrosine phosphorylation was described as a regulatory modification on the triggering of many receptors. Thus, the activity of protein tyrosine kinases (PTK) was also found to be essential for CD95mediated apoptosis. Furthermore, protein kinase C (PKC) activity may play a role in modulating the cytotoxic signal. Thus, PKC inhibitors such as H7 and HA 1004 sensitized cells toward CD95-mediated apoptosis, whereas activation with the phorbol ester PMA led to the development of resistance (Copeland et al., 1994). This listing should not create the impression that all molecules mentioned are of equal importance. The direct role of some of them in CD95 signaling needs to be confirmed in the in uivo situation (Table IV). XI. Proteins of the Bcl-2 Family
Bcl-2 was first discovered in the development of human follicular lymphoma (Tsujimoto et al., 1984). A t( 14 : 18) translocation of the bcl-2 gene in this tumor led to a deregulation of its expression because it became controlled by the enhancer of the Ig heavy chain gene (Graninger et al., TABLE IV OTIIER SIC:NAI.INC: MOIXCULESINVOKEDI N CD95 SIGNAI.INC: Signaling Molecule
Bcl-2 family PKC Cerainide
HCP PTK Caspases
Activity Controversial Blocker Activator/stimnlator Activator/stimulator Activator/stirnulator Activator/stimulator
1987). It soon became clear that overexpression of bcl-2 was not sufficient for the transformation of cells. In fact, other genetic alterations were also necessary (Vaux et al., 1988; Reed et al., 1988; Cook et al., 1985; Metcalf et al., 1987). To demonstrate this, elegant experiments using c-myc and bcl-2 transgenic mice were suitable whose offspring showed an increased tumor incidence (Strasser et al., 1990). Although this has been contested lately (HSUet al., 1997), Bcl-2 may heterodiinerize with other proteins in physiological systems. These inolecules interact by domains (BH1 and BH2) that have a high degree of homology (Yin et al., 1994). Together, these molecules form the Bcl-2 family (Sedlack et al., 1995). Interaction with Bax is probably the most important function of Bcl-2 (Oltvai et al., 1993, Yin et nl., 1994; Hanada et al., 1995). Bax occurs in several splice variants; Bax-a, Bax-P, and two forms of Bax-y. Only the function of Bax-a is known. Overexpression of Bax-a! in cells leads to apoptosis, and protection by the simultaneous expression of Bcl-2 is relieved (Oltwai et al., 1993). Thus, Bax-a expression is also found in tissues that show increased apoptosis (Krajewski et al., 1994). Another member of the Bcl-2 family is Bcl-x (Boise et al., 1993). Bcl-x occurs in two splice variants, which act differentially on apoptosis. The long form, Bcl-xL,and Bcl-2 both block apoptosis of cells, e.g., deprived of growth hormones. The short variant of Bcl-x, Bcl-x,, antagonizes the effect of Bcl-2. In different tissues expression of the two variants correlated with the life expectancy of the cells of these tissues (Boise et d., 1993). Bcl-x, and Bcl-2 inay exert their negative effects on different apoptotic pathways. Bcl-x, blocks apoptosis in P-lymphoblastoid cells induced by immunosuppressive agents whereas Bcl2 does not seein to protect at all (Gottschalk et al., 1994; Choi et al., 1995; Cuende et al., 1993). Also, mechanisms of in oivo resistance of T-cell lymphomas do not seein to involve Bcl-2 (Debatin and Krammer, 1995; Boise et al., 1995a). Additional molecules belonging to the Bcl-2 family are Bad, Bik, and Bak-1. Bad is a modulator of the function of Bax (Yang et al., 1995); Bik associates with the adenovirusprotein E1B 19K, Bcl-2, and Bcl-xL(Boyd et al., 1995); and Bak-1 acts like Bad and increases apoptosis (Chittenden et al., 1995; Farrow et al., 1995; Kiefer et al., 1995). In mammals, the A1 (Lin et al., 1993) and Mcl-1 genes (Kozopas et al., 1993) show homology to bcl-2. The protein BHRF was isolated froin the Epstein-Barr virus and tlie protein LMW5-H1 froin tlie African swine fever virus (Neilan et nl., 1993). Both proteins share homology with Bcl-2. BHFR blocks the apoptosis of infected cells whereas no further function is known for LMW5-H1. Another protein, BAG-1, which does not show sequence homology to the Bcl-2 family, interacts with Bcl-2, however, and increases the antiapoptotic effect of Bcl-2 when cotransfected with Bcl-2 (Takayama et al., 1995).
174
PETER 13. KRAMMER
Expression of Bcl-2 in cells protects against different apoptosis-inducing stimuli such as irradiation, DNA damage, glucocorticoids, sodium axide, Ca’+ influx, heat shock, and oxygen radicals (Vaux et al., 1988; McDonell et al., 1989; Strasser et al., 1991b; Sentman et al., 1991; Miyashita and Reed, 1992; Reed, 1994). In contrast to the earlier examples, there are cases, however, in which apoptosis cannot be blocked by Bcl-2. Such a case, for example, is the apoptosis involved in the negative selection of thymocytes (Sentman et al., 1991; Strasser et al., 1991a),cell death induced by TNF (Vanhaesebroeck et al., 1993), or lysis of target cells by cytotoxic T cells (Vauxet al., 1992a).The effect of Bcl-2 on CD95-mediated apoptosis is unclear. The described effects of Bcl-2 vary from complete (Jaattela et al., 199s) or partial inhibition (Itoh et al., 1993; Martin et al., 199Sa; Enari et al., 1995a) to the observation that bcl-2 does not protect cells from CD95-mediated apoptosis at all (Strasser et al., 1995; Memon et al., 1995). These controversial data have been resolved by experiments described by Scaffidi et al. (1998) and will be discussed further after a more complete discussion of the CD95-signaling pathway. XII. The Death-Inducing Signaling Complex (DISC)
The three-dimensional structure of the CD95 DD has been determined by nuclear magnetic resonance spectroscopy. It consists of six antiparallel, amphipathic a helices arranged in a novel fold, which is likely to be important for binding intracellular signaling molecules (Huanget al., 1996). Using a different method than the yeast two-hybrid system, a complex of proteins was identified that associated only with stimulated CD95 (Kischkel et al., 1995).Treatment of CD95-positive cells with the agonistic mAb anti-APO-1 and subsequent immunoprecipitation of CD95(APO-l/Fas) with protein A-Sepharose resulted in the identification of four ytotoxicitydependent APO-1-associated proteins (CAP1-4). The CAP2-4 proteins could be revealed on two-dimensional IEF/SDS gels within seconds after CD95 triggering. Together with CD95 these proteins formed a complex called the death-inducing signaling complex (DISC). Using a specific rabbit antiserum7CAPi and CAP2 couldbe identified as two different serinephosphorylated species of FADD and demonstrated that FADD bound to CD95 in a stimulation-dependent fashion in tjivo. While the binding of FADD to CD95 is stimulation dependent the phosphorylation of FADD is not. Therefore, the role of FADD phosphorylation still remains a mystery. It is not excluded, however, that it is important to give the entire DISC a structural conformation in order to function properly. When FADD-DN (the C-terminal DD-containing part) was stably transfected into cells the DISC also formed. However. FADD-DN was recruited
CD9551APO- l/F.is)-ME UIATEU APOPTOSIS
175
to CD95 instead of the endogenous FADD. Analysis on two-dimensional gels revealed that CAP3 and CAP4 were not part of the DISC anymore ). proteins were therefore prime candidates (Chinnaiyan et al., 1 9 9 6 ~ These for the signaling molecules. Using nanoelectrospray tandem mass spectrometry, sequence information of CAP3 and CAP4 was obtained that led to the retrieval of a full-length clone from a cDNA data base that contained all sequenced peptides (Muzio et al., 1996). This protein contained two DED at its N terminus and showed tlie typical domain structure of an ICE-like protease at its C terminus. It was therefore termed FLICE (for FADD-like ICE). The same molecule was also found by two other groups and was named MACH and MchS (Boldin et nl., 1996; Fernandes-Alnemri et nl., 1996). FLICE belongs to cysteine proteases, which are now called caspases (Alnemri et nl., 1996),and got the name caspase-8. Thus, identification of caspase-8 and its location in the DISC on CD95 triggering connected two levels in the CD95 apoptosis pathway; the CD95 receptor level with tlie intracellular level of the apoptosis executioners, the caspases. The finding that caspase-8 was identified as part of the in civo CD95 DISC suggested that its activation occurred at tlie DISC level. It has been shown that tlie entire cytoplasmic caspase-8 is converted into active caspase-8 subunits at the DISC (Medema et nl., 1997a). After stimulation, FADD and caspase-8 are recruited to CD95 within seconds after receptor engagement. Although this has not yet been shown, direct binding of caspase-8 may cause a structural change in tlie molecule that results in autoproteolytic activation on which the active subunits p10 and p18 are released into the cytoplasm. Part of caspase-8 prodomain stays bound to the DISC. Anti-caspase-8 monoclonal antibodies have been generated that show that from tlie eight published caspase-8 isoforms, only two, caspase8/a and 8/b, were expressed predominantly on the protein level in 13 different cell lines tested (Scaffidi et al., 1997). Both isoforms are recruited to the DISC and are processed with similar kinetics. Recoinbinant caspase8 lacking tlie prodomain has been reported to cleave caspase-8 in uitro, suggesting autocatalytic cleavage of caspase-8 at tlie DISC and an amplification step with caspase-8 at the top of a caspase cascade (Srinivasula et al., 1996a,b; Muzio et al., 1997). However, Medema et al. (1997a) used the DISC and could not confirm this observation. Recombinant caspase8 lacking the prodomain might therefore display a different substrate specificity compared to full-length caspase-8 in ciwo. Most investigators believe that active caspase-8 cleaves various cellular death substrates, including other caspases, such as caspase-3, thus initiating the execution of apoptosis. The caspase-8 gene has been mapped to chromosome 2q33 to 35 (Kisclikel et nl., 1998; Rasper et al., 1998). Furthermore, caspase-8 has been shown to be cleaved and activated by granzyine B in perforin killing
176
PETER 11. KRAMMER
by cytotoxic T lymphocytes. Thus, capsase-8 is the primary, although not the only, initiator of a protease cascade in this type of T killer cell activity (Medema et al., 1997b). XIII. Downstream Caspases in CD95 Death Receptor Signaling
Caspase-8 belongs to a growing family of cysteine proteases (Ahemri et al., 1996). Caspases are crucial for the execution of apoptosis, as is CED-3 during nematode development (Yuan et al., 1993). So far, 11 known caspases can be subdivided into three families based on sequence homology. The ICE-like protease family includes ICE (caspase-1) (Thornberry et al., 1992; Ceretti et al., 1992), WICH-2ACE-relII (caspase-4) (Faucheu et al., 1995; Munday et al., 1995; Kamens et al., 1995), TY/ICEre1111 (caspase-5) (Munday et al., 1995; Faucheu et al., 1996), and ICH3 (caspase-11) (Wanget al., 1996).The CED-3-like family includes CPP3W YAMNapopain (caspase-3) ( Fernandes-Alnemri et al., 1994; Tewari et al., 1995b; Nicholson et al., 1995), Mch2 (caspase-6) ( Fernandes-Alnemri et al., 1995a), Mch3/ICE-LAP3/CMH-l (caspase-7) ( Fernandes-Alnemri et al., 1995b; Lippke et al., 1996; Duan et al., 1996a), caspase-8 (FLICE/ MACH/Mch5) (Muzio et al., 1996; Boldin et al., 1996; Fernandes-Alnemri et al., 1996), MchG/ICE-LAPG (caspase-9) (Duan et al., 1996b; Srinivasula et al., 1996b),and Mch4/FLICE2 (caspase-10) (Fernandes-Alnemri et al., 1996; Vincenz and Dixit, 1997). The third subfamily consists of Nedd2/ ICH-1 (caspase-2) only (Wang et al., 1994; Kumar et al., 1994) (Fig. 1). Caspases are synthesized as proenzymes that are activated by proteolytic cleavage. The active enzyme is a heterotetrameric complex of two large subunits containing the active site and two small subunits, as deduced from the crystal structure of caspase-1 and caspase-3 (Wilson et al., 1994; Walker et al., 1994; Mitt1 et al., 1997). Activation of caspases has been reported for a variety of apoptotic stimuli, including signal transduction through the “death receptors.” Thus, crmA, a viral inhibitor of caspase-1 and other caspases, blocks apoptosis induced by these receptors (Tewari et al., 1995b; Enari et al., 1995b; Los et al., 1995; Miura et al., 1995). As discussed earlier, either caspase-8 or a similar caspase, such as caspase-10, may be involved in a similar fashion as in CD95 signaling in the signal transduction of the other death receptors. Overexpression of caspases in mammalian or insect cells induces apoptosis, but only caspase-3, -6, -7, and -8 were shown to be activated in vivo on triggering of the death receptors (Medema et al., 1997a; Scaffidi et al., 1997; Orth et al., 1996b; Duan et al., 1996a,b; Schlegel et al., 1996; Faleiro et al., 1997). In vitro caspase-8 was shown to cleave caspase-3, -4, -7, -9, and -10 directly. Caspase-2 and -6 were cleaved indirectly by other caspases
177 Caspase-5 (ICEreplIl, TY)
ICE-
subfamily
Caspase-4 (TX, ICH-2, ICE,pII) Caspase-1 (ICE) Caspase-7 (Mch3, ICE-LAPS, CMH-1) Caspase-3 (CPP32, Yama, apopain) Caspase-6 (Mch2) ^^
I,
,”* _,_,__’
Caspase-8 (FLICE, MACH, Mch5) Caspase-10 (Mch4)
I”-.{-
CED-3sub f am I I y
.yLI.-LIII..--x
Caspase-2 (ICH-1) Caspase-9 (ICE-LAPG, Mch6)
Fic:. 1 . The caspasr f’aniilyof ICE/CED-:3-hoiiiologouscystein proteases. Synonyms are given in parentheses. Phylogenetic relationships were generated by PILEUP algorithms of the “Wisconsin COG sequence analysis package.” K n o w i caspase family itieml)ers arc Caeiiorlubrlitis & ~ O J J . S CED-5 and 10 enzynies of hiimaii origin. Similar niolecules have also Ixen identified in other species (e.g., iiiouse, guinea pig. antl Uro.soplda rnclariogcister, not shown). Based on the sequences of ICE and CED-3, the proteins can he divided into hvo subgroups. C. rlqyms CED-3 is closely reliited to huiiian caspise-3. References for caspase-1: Thoriiberr>i et al. (1992); Ceretti ct 01. (1992): caspase-2: Wang et al. (1994); Kuinar et (11. (1994); caspase-3: Fernantles-Alnemri ct (11. (1994); Tewari et al., 199511; Nicholson ~i 01. (1995): caspase-3: Fauclieii ef al. (1995); Munday et 01. (1995); Kamens at rrl. (1995);caspase-5: Munday at a!. (1995); Faucheu c>t a!. (1996);caspase-6: FernandesAlneinri c’t rrl. (lW5a); caspase-7: Feriiaiitles-Alneiiiri et rrl. (1995h); Lippke et nl. (1996); Duan r f 01. (1996a):caspase-8: Muzio ct 01. (l99G):Boldin ct 01. (1996): Fernandes-Aliiemri et (11. (1996); caspise-9: Duan et nl. (199Fb): Srinivasula rt rrl. (1996h); antl caspase-10: Fernandez-Alneniri ei ril. (l99G); ~’iticoiizaiid Dhit (1997).
and activated by caspase-8 in cellular extracts (Muzio et al., 1997). Therefore, caspase-8 is able to start a cascade of caspases. The order of caspases in this cascade is not yet clear. Orth et al. (199611)place caspase-6 upstream of caspase-3 and -7, whereas it has also been demonstrated that activated caspase-3 can cleave and activate caspase-6, -7, and -9 ( Fernandes-Alnemri et al., 1995a,b; Fernandes-Alnemri et al., 1996; Srinivasula et al., 1996b). There is no report so far, however, that demonstrates that a single caspase is crucial for apoptosis signaling by death receptors. Caspase-1 was proposed to be a key molecule in CD95-mediated apoptosis, and thymocytes
178
PETER €1. KRAMMER
from caspase-I-’- mice were reported to be resistant to CD95-induced cell death (Kuida et al., 1995). However, others did not find an impairment of apoptosis in c a s p a s e - P mice (Li et al., 1995b; Smith et al., 1997) or failed to demonstrate activation of caspase-1 on CD95 triggering (Muzio et al., 1997).Therefore, it is unlikely that caspase-1 plays a role in apoptosis signaling through death receptors. Alternatively, another caspase-l-like caspase can substitute for its function in different cellular contexts. Similarly, mice deficient of caspase-3 showed an alteration of brain development and thymocytes were not affected at all (Kuida et al., 1996). However, the activity of caspase-3 might be redundant with other caspases. Therefore, the knockout technology may not always be suitable in investigating the role of a single caspase in apoptosis signaling. Different splice variants of numerous caspases have been reported (Wang et al., 1994; Alnemri et al., 1995; Fernandes-Alnemri et al., 1996; Fernandes-Alneniri et al., 1995; Boldin et al., 1996; Wang et al., 1996c; Vincenz and Dixit, 1997) and shown to function either as activators or as inhibitors of caspase activation. Some ofthem may also represent nonfunctional proteases. Most of the reported splice variants were described on the mRNA level, and the number of isoforms expressed as proteins is clearly more limited (Scaffidi et al., 1997). The activity of caspases characterizes the execution phase of apoptosis. Therefore, the search for substrates cleaved by caspases during apoptosis should provide insight into the more downstream events in apoptosis signaling. Several of these so-called “death substrates” are known so far. First, caspases can cleave and activate other caspases as described earlier. Second, molecules involved in DNA repair, such as poly(ADP-ribose) polymerase (PARP) or the catalytic subunit of the DNA-dependent protein kinase (DNA-PK),were described to be cleaved and thereby inactivated by downstream caspase-3-like caspases (Lazebnik et nl., 1994; Gu et al., 1995a; Casciola-Rosen et al., 1995; Song et al., 1996b).Third, the ribonucleoproteins U1-70kDa, C1 and C2, components responsible for the splicing reaction of precursor mRNA, are inactivated by cleavage through caspases (Casciola-Rosenet al., 1994; Tewari et al., 1995a;Waterhouse et al., 1996). Therefore, caspases may inactivate cellular processes that prevent apoptosis from proceeding. Another set of “death substrates” contains molecules that are involved in the regulation of other signaling processes. One example is the sterol regulatory element-binding proteins SKEBP-1 and SREBP-2 that are cleaved by caspase-3 or -7 (Wang et al., 1995; Pai et al., 1996). The delta isoforin of protein kinase C (PKCG) is also a target for caspases during CD95- or TNF-R-mediated apoptosis. Cleavage activates PKCS, which is associated with chromatin condensation and nuclear fragmentation
(Ghayur et al., 1996; Emoto et al., 1995). Another example is a GDP dissociation inhibitor for the Ras-related Rho family GTPases, D4-GDI cleaved and inactivated during CD95-mediated apoptosis (Na et al., 1996: Danley et al., 1996). This leads to defective Rho GTPase regulation. The PISTLRE kinase, a member of a superfamily of cdc2-like kinases, is also cleaved by caspases during CD95-induced apoptosis. This is accompanied by increased activity of this lanase and may lead to apoptosis (Bunnell et al., 1990; Lahti et al., 1995; Beyaert ct ~ l . 1997). , Cytosolic phospholipase A? (PLA?)was also demonstrated to be a target of caspases during TNFinduced apoptosis, leading to its activation. In addition, the inhibition of PLA? was shown to partially inhibit TNF-induced apoptosis (Wissing et al., 1997). Therefore, the cleavage of these signaling molecules may be involved in the downstream execution of apoptosis. However, the mechanism by which these signaling nioleciiles can transduce the apoptotic signal remains elusive. Another set of death substrates are structural proteins of the cell. Thus, larnin cleavage by caspase-6 may occur during TNF-R- and CD9Fi-induced apoptosis (Orth et al., 1996a; Neamati et al., 1995;Zhivotovsky et al., 1997). Similarly, a-fodrin is cleaved during apoptosis induced by CD95 or by TNF-R1 triggering (Cryns et a/., 1996; Martin et al., 1995a). However, the caspase inhibitor DEVD protects cells from CD95-induced apoptosis but does not prevent fodrin proteolysis. This indicates that cleavage of this protein can be uncoupled from apoptotic cell death (Cryns et al., 1996). Even actin was found to be cleaved proteolytically by caspase-3 on triggering with different apoptotic stimuli. However, contradicting reports exist on the cleavage of actin in oivo during CD95-mediated apoptosis (Mashinia et a!., 1995; Kayalar et al., 1996: Chen et al., 1996: Song et al., 1997; Mashiina et al., 1997). Finally, Gas% a component of the microfilament system, was reported to be cleaved on induction of apoptosis (Brancolini et nl., 1995).Therefore, cleavage of these structural proteins may account for some of the massive morphological changes, such as lnelnbrane blebbing, nuclear fragmentation, and the formation of apoptotic bodies during apoptosis. Some oncoproteins, such as HI) and MDM-2, were found to be cleaved and inactivated by caspases during apoptosis (Bing et al., 1996; Janicke et al., 1996; Ehrhardt et al., 1997).The functional relevance of cleaving these substrates remains unclear. Cleavage and activation of the 112l-activated lilnase PAK2 during CD95and TNF-mediated apoptosis have been reported. PAK2 was also reported to activate the Jun kinase pathway. PAK2 may therefore link the caspases and J N K activation during apoptosis signaling. Interestingly, when the activity of PAK2 is blocked by a dominant negative mutant, the form.tal., 1996; Coustan-Smith et d . , 1996; Krajewski et d.,1995). p53 was found to lie involved in transcriptional expression of the proapoptotic BX gene following DNA damage. Thus, soine forms of cliernotherapy-induced apoptosis seem to involve upregulation of H~LXexpression mediated by p53 (Miyashita et 01, 1994; Miyashita and Heed, 1995). A major role for the CD95 system in drug-induced apoptosis has been suggested. Cytotoxic drugs commonly used in tlie chemotherapy of leukemias strongly induce CD95L exprejsioii in CD95 positive tumor cells (Hata et ul., 199Fj; Miiller et d., 1997; Friesen ct d., 1996). CD9SL is expressed in a membrane form or secreted by the tumor cells exposed to the drug. Binding of CD95L to CD95 then triggers the apoptosis cascade in chemosensitive cells. This scenaiio is malogous to activatioii-induced death in activated T cc4s following T-cell receptor stimulation (Dhein et al., 1995; Stahnke et nl., 1997).Triggering of the CD95/CD95L system by anticancer drugs was originally discovered in leukemia cells but was also found to be
186
PETEH 11. KKAMMEH
involved in drug-induced apoptosis in other chemosensitive tumors such as hepatoblastoma, neuroblastoma and brain tumors (Hata et al., 1995; Muller et al., 1997; Micheau et al., 1997; Fulda et al., 1997a). In some tumors, treatment with cytotoxic drugs upregulates CD95L and CD95. CD9g upregulation seems to depend on functional, wild-type p53 (Munker et al., 1995; Muller et al., 1997; Fulda et al., 1997a). Regardless of the primary target of the cytotoxic drug in the tumor cell, data would suggest that activation of the key effector molecules for apoptosis, such as the ones in the caspase cascade, is crucial for the antitumor effect, and inhibition of caspase activation in tumor cells is associated with resistance against anticancer drugs (Los et al., 1997; Kaufmann et al., 1993; Bellosillo et al., 1997). It is not clear whether this applies to all drugs, e.g., apoptosis induced by glucocorticoids may use a different effector system (Geley et al., 1997). In cases of drug treatment in which CD95/CD95L is used, it is assumed that the described (see earlier discussion) CD95 signaling pathways are initiated. As described in these situations, caspase-8 plays a crucial role. The role of caspase-8 may even be essential when CD95/CD95L is not used. This is exemplified by treatment of neuroblastomas with Betulinic acid. Here, caspase-8 cleavage is clearly observed (Fulda et al., 1997b). Initially, drugs may affect diverse cellular functions such as cellular metabolism, DNA, or the mitotic apparatus. Cellular damage is sensed by p53 and/or leads to activation of a cellular stress program (Herr et al., 1997). In the following phase a relatively uniform apoptosis program of the cell is activated. Here, CD95LJCD95 interaction represents one of the important trigger mechanisms for the apoptosis cascade leading to protease activation. However, CD95/CD95L represents only a part of an amplifier machinery, as drug-induced production of other apoptosisinducing ligands (TNFa and TRAIL) has also been found (Muller et al., 1997). Alterations in each phase of drug-induced apoptosis may lead to drug resistance. Thus, in addition to established mechanisms of drug resistance, such as increased drug efflux by membrane pumps, the fiilure to activate the apoptosis program represents another mode of drug resistance in tumor cells (Friesen et al., 1997; Knipping et al., 1995). In this respect, a favorable treatment outcome is shown for patients with CD9Fj positive AML undergoing chemotherapy. This situation was compared to CD95 negative AML. Data from this study demonstrate that CD95 expression in tumors in relationship to therapy may be clinically relevant and worth investigating (Min et al., 1996). Thus, the CD95/CD95L system is important for effective chemotherapy. It also participates in reciprocal interactions of tumor cells with cells of the immune system (Debatin, 1997). Thus, CD95 positive tumor cells may become targets for killer T cells,
NK cells, or LAK cells (Micheau et nl., 1997;Yoshihiro et a / . , 1997).Factors that upregulate CD95 expression in tumor cells, such a s cytotoxic drugs or cytohnes, may therefore be involved in the induction of tumor regression by immune cells if these cells resist the attack by CD9SL, which may be produced by the tumor cell itself. These data demonstrate that the treatment of tumor cells with various chemotherapeutic drugs causes an upregulation of the CD95 death system. As more tumor types are checked, it is conceivable that other death systems, such as DR3, TNF-K 1/11, and TRAIL, may be used in a similar fashion as the CD9Fj system. However, the basic features and their impact on tumor treatment will be the same and may cause a change in paradigm in drug treatment and in the evaluation of drug sensitivity and resistance. In a successful antitumor therapy the CD95 signaling cascade, including ICE-like proteases, is switched toward sensitivity. Considerations of drug resistance in the future, therefore, will also have to incorporate testing the functionality of the CD95 and possibly other death systems. It is clear that this has far-reaching therapeutic consequences. The previously described decisive advances in the explanation of the mechanism of chemotherapy may pro\& an explanation of why chemotherapy may have a positive and a negative side. On the one hand, it helps to eliminate the tumor. On the other hand, it may provide the tumor with a lethal weapon, boost immune evasion, and make the tumor an immune privileged site. The lethal weapon is CD95L. CD9FjL can be made by tumors constitutively or its expression might be induced, e.g., a s detailed earlier, by drug treatment of the tumor. CD95L may have a deleterious effect on the immune system, particularly on the activated attacker antitumor T lymphocytes, provided they are in a CD95 apoptosis-sensitive state. The experimental background of this situation is outlined in different in vitro and in oivo model systems in several papers that may change our view of tumor/host interactions (O'Connell et nl., 1996; Hahne et nl., 1996; Strand et nl., 1996). The previous experiments and their interpret,nt'ions have been criticized, and growth of CD9SL positive tumors was observed to be blocked by infiltrating neutrophils and granulocytes (Seino et al., 1997; Arai et nl., 1997).In any case, data point toward a situation in which (1) attacker T lymphocytes, e.g., activated T (killer) cells, express two killing systems: the perforidgranzyme and the CD95/CD95L system; and (2) tumor cells express CD95L and may even have become resistant against its hlling activity (e.g., by downregulating CD95) such as in most hepatomas. Previously, the prevailing view was that the T cells recognize the tumor, do not receive sufficient costiinulatory signals, and are therefore not activated properly. This view now adds a different turn. It says that the tumor is not passive but has the weapon to fight the immune system
188
PETER H. KR-\MMER
(“strike back’) and cause the depletion of attacker lymphocytes and immune suppression. Thus, both T cells and tumor cells are equipped with a set of almost equal weapons. Several parameters may determine who wins the encounter between T cells and tumor cells. The size of the tumor-reactive T-cell pool, the quantity of CD95 and CD95L, and sensitivity and resistance toward CD95 signaling may all be important. In the worst case the tumor has become apoptosis resistant, e.g., by downregulating CD95 or by switching the signaling pathway to insensitive, expresses CD95L, and encounters activated T cells in a CD95 apoptosis-sensitive state. The net result of such an interaction could be T-cell depletion, imiiiunosuppression, and immune escape. The aim of future therapy is to cause tumor rejection by temporarily inducing T-cell resistance toward apoptosis and maintaining sensitivity in the tumor (Krammer et al., 1997, 1998). XVIII. The CD95 Death System in AIDS
In AIDS the number of T cells, e.g., in the peripheral blood, productively infected with HIV is relatively low (in the range of one in several thousand). This implies that T-cell depletion in this disease may also affect noninfected CD4 positive T cells. Depletion of T cells may occur by induction of apoptosis. This assumption raises several questions, mainly: (1) what is the (main) mechanism of apoptosis of noninfected CD4 positive T cells and (2) how is sensitivity toward apoptosis induced in noninfected CD4 positive T cells? The hypothesis that regulatory viral gene products (e.g., HIV-1 Tat) made by HIV-infected cells may penetrate HIV-noninfected cells and render these cells hypersensitive to TCR-induced CD95-mediated apoptosis was tested. HIV Tat might induce a prooxidative state in the affected cells, increase CD95L expression, and facilitate TCR-triggered CD9fj-mediated suicide. Further sensitization of the CD4 positive T cells might result from the binding of HIV a 1 2 0 to the CD4 cell surface receptor and from the cross-linking of bound gpl20 by gp120 antibodies in the patients’ sera. This hypothesis was tested in an in vitro model system using Jurkat T cells or peripheral blood lymphocytes triggered via the TCR and incubated with synthetic Tat or with Tat, gpl20, and anti-gpl20, respectively. TCR-triggered apoptosis was accelerated by the combination of these reagents. Significantly, most but not all of the apoptosis seen under these conditions could be blocked by F(ab’)zanti-CD95 and soluble CD95-Fc receptor decoys (Westendorp et al., 1995; BBuinler et al., 1996; for a review, see Kraminer et al., 1994a,b).These data reinforce the hypothesis that the depletion of noninfected CD4 positive T cells in AIDS might involve TCR-triggered CD95-mediated suicide sensitized by HIV Tat and
a 1 2 0 and anti-gpl20 antibodies in patients' sera. These findings warrant testing this hypothesis in animal model systems and in patients and, if correct, open new exciting possibilities of AIDS therapy by stabilizing the pool of CD4 positive T cells. To test further whether the CD95 system might play a role in AIDS, CD95 expression on isolated peripheral 1)lood lymphocytes was investigated. Although CD95 expression was highly variable, it was increased significantly in CD4 positive and CD8 positive T cells from HIV-1 positive children in comparison to HIV-1 negative controls (BBumler ct d . ,1996). Thus, CD95 expression on T cells ofthe CD4 positive and the CD8 positive subset is increased in HIV infection. The mechanism of this increase niay be directly related to HIV infection or stimulation ljy HIV products or may be caused by a general stiniulation of T cells in this disease. It is interesting that the increase of CD95 expression is observed in both CD4 positive and CD8 positive T cells from HIV- 1 positive children. Assuming a role of CD95 receptor and CD95L in apoptotic depletion of CD4 positive T cells in AIDS requires, thcrefore, that sensitization of the CD95 pathway is mainly directed at the CD4 positive subset of T cells. As suggested earlier, sensitization may involve stimulation of CD4 on these cells by gpl2O and anti-gpl2O antibodies. In addition, other viral gene products such as Tat may have an additional sensitizing effect. The author's data, therefore, provide an interesting potential link between CD95mediated apoptosis and T-cell depletion in AIDS. In view of the discovery of new death receptors, it is conceivable that they are also involved in CD4 positive T-cell depletion in AIDS in a similar fashion as the CD95 system. In addition to CD95-mediated apoptosis, a novel type of apoptosis induced via CXCR4 (fusiidLESTR) and CD4 in CD4 positive T cells as detailed later was found (Bemdt c>td . ,1998).As described earlier, apoptosis can be induced in CD4 positive T lyinphocytes by HIV-1 a 1 2 0 and anti-gpl20 antibodies. Because a 1 2 0 binds to T cells via CD4 and the chernokine receptor CXCR4 ( fiisidLESTR), the role of these molecules was investigated in gpl2O-induced apoptosis. A novel and rapid type of apoptosis induced by both cell surface receptors, 0 4 and CXCR4, in T cell lines, human PBL, and CD4KXCR4 transfectants was found. Significantly, apoptosis was observed exclusively in CD4 positive but not in CD8 positive T cells (Table V). Induction of the new type of cell death did not involve the signaling cascades normally initiated by CD4, CXCR4, and some of the known death receptors (p56''!', G , a , and caspase-8, respectively). The potency of this phenomenon and its specificity for CD4 positive T cells would suggest that it might play a significant role in T helper cell depletion in AIDS. On the basis of these data, the use of antichemokine
190
PETEK I I KHAMMER
TABLE V COMIJAHISON OF NOVELAPOPTOTICCELLDEATH MEDIATEDBY CD4 CD%-MEDIATED APopTosis CD95-Mediated Apoptosis
DNA degradation Kinetics (onset) Changes in FSC/SSC" Loss of niemhrane asylllllletry
Cliroinatin condensation
A !bn! 7-AAD positivity' Inhibition by zVAD-fink" Caspase-3 cleavage Caspase-8 cleavage PARP' cleavage
AND
CXCR4 WITH
CD4- or CXCR4-Mediated Apoptosis
+ + + + + + + + + +
-12 hr
-15 min
+ + + + + -
-
Foward scatter/sideward scatter. Loss of initocliondrial inembranc potcritial. ' Staining of apoptotic cells with the fluorescelit c l y ~7-airiino-actirioiiiycinD " Peptide inhibitor of caspasw ' Poly (ADP ribose) polyimrrasr. " 'I
receptor antibodies meant to prevent HIV-1 infection might be dangerous. The use of tlie natural ligand of CXCR4, SDF-la, or its derivatives, however, could be considered for therapy, as it inhibits infection as well as CXCR4-mediated apoptosis. Studying the apoptotic signaling cascade triggered by CD4 and CXCR4 might prove to be useful in tlie identification of therapeutic strategies aimed at intervening with the progressive loss of CD4 positive T cells in HIV-1-infected individuals. XIX. Further Considerations on the Role of Apoptosis in the Clinic
Apoptosis has come a long way from the description of its morphological features to a partial niolecular understanding of its signaling pathways and its physiological and pathological consequences. Although our understanding of the different routes to deaths is far from complete, the first glimpses at the different signaling pathways make it possible to predict important areas of research in the future. Apoptosis adds a new chapter in the understanding of the pathogenesis of many diseases. Generally, there are diseases with too little apoptosis and diseases with too much apoptosis. As described earlier, cancer could be looked on as a disease with too little apoptosis where the net increase of the tumor burden is the sum of an
increased growth rate and a decreased apoptotic rate. However, AIDS involves too much apoptosis in cells of the lymphoid and nonlymphoid compartment, particularly in CD4 positive T cells. These two diseases are only two examples in which apoptosis and its regulation have gone wrong. The molecular understanding procided by the elucidation of the signaling pathways of apoptosis may make it possible, in the future, to reestablish the normal level of apoptosis. With respect to the CD95 system, which may serve as a paradigm for other death systems, apoptosis could be regulated at several decisive points:
1. the receptor (structure, glycosylation, density, expression as soluble versus membrane-bound receptor), 2. the CD95-mediated pathways (DISC formation, caspase cascade, induction of apoptosis via the mitochondria1 pathway, alteration of the death substrates, stimulation of the apoptotic pathway, inhibition of the apoptotic pathway at the receptor level by, e.g., FLIPS, or further downstream by Bcl-2 or family members or by other inhibitory or stimulating proteins), 3. CD95L (structure, glycosylation, density, expression as soluble ligand versus membrane bound ligand), and 4. regulation of induction versus prevention of apoptosis in different cellular systems. It is important to increase our understanding of the molecular events that determine these particular steps and to develop methods of targeting apoptosis regulatory moleciiles to specific cells of hfferent tissues. Using such methods, the modulation of apoptosis will gain its place in the therapeutic tools used for the treatment of diseases. Alternatively, therapeutic windows for apoptosis modifiers have to be found that only affect diseased cells and leave normal cells intact.
ACKNOWLEDGMENTS I thank all my previous and present collaborators and the previous and present nrcmhers of my group. particularly M. E. Peter, F. C. Kischkel. and C. Scaffidi for their input and discussion, C. Scaffidi for reading of the manuscript, and H . Sauter for expert secretarial assistance. This work was funded by Deutsche Krehshilfe Dr. Mildred Scheel Stiftung: Gerinan-Israeli Cooperation in Cancer Research: AIDS grant German Federal Health Agency; Tumor Centre Heidell)erg/Maiinheiin: BMBF Forderschwerpunkt “Clinicalbiomedical research; AIDS F’erbund Heidelberg; Wilhelm-Sander Stiftung; and SFB Transplanation. I apologize to all my colleagues who have done excellent work in the field and whose papers have not been quoted comprehensively. It was not possible to be encyclopedic in this exponentially growing field.
192
PETER H. KRAMMER
REFERENCES Adachi, M., Suematsu, S., Kondo, T., Ogasawara, J., Tanaka, T., Yoshida, N., and Nagata, S. (1995). Targeted mutation in the Fas gene causes hyperplasia in peripheral lymphoid organs and liver. Nut. Genet. 11, 294-300. Alderson, M. R., Tough, T. W., Davis Smith, T., Braddy, S., Falk, B., Schooley, K. A,, Goodwin, R. G., Smith, C. A,, Ramsdell, F., and Lynch, D. H. (1995).Fas ligand mediates activation-induced cell death in human T lymphocytes.]. Exp. Med. 181, 71-77. Alnernri, E. S., Livingston, D. J., Nicholson, D. W., Salvesen, G., Thornberry, N. A,, Wong, W. W., and Yuan, J. (1996). Human ICEICED-3 protease nomenclature. Cell 87, 171. Anderson, D. M., Maraskovsky, E., Billingsley, W. L., Dougall, W. C., Tometsko, M. E., Roux, E. R., Teepe, M. C., DuBose, R. F., Cosnian, D., and Galibert, L. (1997). A homologue of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell function. Nature 390, 175-179. Anel, A,, Simon, A. K., Auphan, N., Buferne, M., Boyer, C., Gohein, P., and SchmittVerhulst, A. M. (1995). Two signalling pathways can lead to Fas ligand expression in CD8+ cytotoxic T lymphocyte clones. Eur. J. Zmmunol. 25, 3381-3387. Arai, H., Gordon, D., Nabel, E. G., and Nabel, G. J. (1997). Gene transfer of Fas ligand induces tumor regression in oioo. Proc. Natl. Acad. Sci. USA 94, 13862-13867. Baens, M., Chaffanet, M., Cassiman, J. J . , van den Berghe, H., and Marynen, P. (1993). Construction and evaluation of a hncDNA library of human 12p transcribed sequences derived from a somatic cell hybrid. Genomics 16, 214-218. Baumler, C., Herr, I., Bohler, T., Krammer, P. H., and Debatin, K.-M. (1996). Increased ex vivo expression of the CD95 ligand in T cells of HIV infected children. Blood 88,17411748. Bajorath, J., and Aruffo, A. (1997). Prediction of the three-dimensional structure of the human Fas receptor by comparative molecular modelling.]. Comput. AidedMol. Des. l l , 3-8. Banner, D. W., D’Arcy, A,, Janes, W.. Gentz, H., Schoenfeld, H. J., Broger, C., Loetscher, H., and Lesslauer, W. (1993).Crystal structure of the soluble human 55 kd TNF receptorhuman TNF beta complex: Implications for TNF receptor activation. Cell 73, 431-445. Bellosillo, B., Dalmau, M., Colomer, D., and Cil, J. (1997). Involvement of CED3/ICE proteases in the apoptosis of B-chronic lyinphocytic leukemia. Blood 89, 3378-3384. Bemdt, C., Mopps, B., Angerrniilller, S., Gierschik, P., and Krammer, P. H. (1998).CXCR4 and CD4 mediate a novel type of apoptosis in CD4 positive T cells. Submitted for publication. Bertin, J., Armstrong, R. C., Ottilie, S., Martin, D. A., Wang, Y., Banks, S., Wang, G. H., Senkevich, T. G., Alnemri, E. S., Moss, B., Lenardo, M. J., Tomaselli, K. J., and Cohen, J. I. (1997). Death effector domain-containing herpesvirus and poxvirus proteins inhibit both Fas- and TNFR1-induced apoptosis. Proc. Natl. Acad. Sci. USA 94, 1172-1176. Beyaert, R., Kidd, V. J., Cornelis, S., Van de Craen, M., Denecker, G., Lahti, J. M., Gururajan, R., Vandenabeele, P., and Fiers, W. (1997). Cleavage of PITSLRE kinases by ICEICASP-1 and CPP3WCASP-3 during apoptosis induced by tumor necrosis factor. ]. B i d . Chem. 272, 11694-11697. Bing, A., and Dou, Q. P. (1996). Cleavage of retinoblastoina protein during apoptosis: An interleukin 1 beta-converting enzyme-likeprotease as candidate. Cancer Res. 56,438-442. Black, R. A., Rauch, C. T., Kozlosky, C. J., Peschon, J. J., Slack, J. L., Wolfson, M. F., Castner, B. J., Stocking, K. L., Reddy, P., Srinivasan, S., Nelson, N., Boiani, N., Schooley, K. A., Gerhart, M., Davis, R., Fitzner, J. N., Johnson, R. S., Paxton, R. J., March, C. J., and Cerretti, D. P. (1997). A metalloproteinase disintegrin that releases tumour-necrosis factor-alpha from cells. Nature 385, 729-733. Bodnier, J. L., Bums, K., Schneider, P., Hofniann, K., Steiner, V., Thome, M., Bornand, T., Hahne, M., Schroter, M., Becker, K., Wilson, A., French, L. E., Browning, J. L.,
MacDonald, H. R., and Tschopp, J. (1997).TRAMP, a novel apoptosis-mediating receptor with sequence homology to tumor nrcrosis fictor receptor 1 antl Fas(Apo-l/Cl)95). I t t i t n u n i t ! / 6, 79-88. Boise, L. H., Gonzakz Garcia. M.. Postrnia. C. E.. Ding, L., Lindsten, T., Turka, L. A,, Mao, X., Nunez. G.. and Thompson. C;. B. (1993).Ucl-x, a bcl-2-related gene that functions as ii doininant regulator of apoptotic ccll dcath. Cvll 74, 597-608. Boise. L. H., Gottschalk. A. R.. Quintans. J., and Thompson, C. B. (1995a). Bcl-2 and Bcl2-rc-latedproteins in apoptosis regnlation. Curr. T o p Microbial. Zttirtiu?~ol. 200, 107-121. Boisr. L. H., Minn, A. J., Noel. P. J.. Junr, C. f l . . Accavitti, M. A,, Lindsten, T., and Thonipson, C:. B. (199%).CD28 costininlation can promote T cell survival by enhancing the expression of Bcl-X,,.Z t t t t i i t m i f y 3, 87-98. Boldin, M. P.. Goncharov, T. M., Goltscxv, Y. V.. d Wallach, D. (1996). Involvement of MACH, a novel MORTUFADD-interacting protcase. in Fas/APO-1- and TNF receptorindnced cell death. Cell 85, 803-815. Boldin, M . P., Varfolomeev. E. E., Pancrr. Z., Mett, I. L., Canionis, J. H., and Wallach, D. (1995). A novel protein that interacts with the death tloniain of Fas/APO-1 contains a sequence motif related to the death tlomain. J . B i d . Chon. 270, 7795-7798. Boyd. J. M.. Gallo, G . J., Elangovan, B.. Ilonghton, A. B., Malstrom, S., Avery, B. J., Ebb, R. C., Subramanian, T., Chittenden, T., Lritz, R. 1.. rt 01. (199s). Bik, a novel deathinducing protein shares a distinct seqnence motif with Bcl-2 family proteins and interacts with viral arid cehilar suni\~alproiiiotingproteins. 0 t m ) g e n e 11, 1921-1928. Brancolini, C., Benedetti. M., antl Schneich~r,C. (1995). Microfilament reorganization during apoptosis:The role ofGas2. a possible snl)strate for ICE-like proteases. EMBOJ. 14,51795190. Brojatscli, J., Naughton, J., Rolls. M. M . , Zingler, K., ant1 Young, J. A. (1996). CAR1, a TNFR-related protein, is a cellular receptor for cytopatliic avian leukosis-sarcomaviruses and inrdiates apoptosis. Ctdl 87, 845-855. Broonie, 11. E., Dargan, C. M..Krajewski, S., and Reed, J. C. (1995). Expression ofBcl-2, Bcl-x. antl Bax after T cell activation antl IL-2 withtlrawa1.J. I t t i n i t m d 155, 231 1-2317. Urowiing, J. L., Ngani E. K. A,. Lawton, P., DvMarinis, J.. Tizard, R., Chow, E. P., Hession, C.. O'Brine Greco, B., Folry, S. F., and \+';ire, C. F. (1993). Lyniphotoxin beta, a novel member of tlie TNF family that forms il lieteronieric complex with lymphotoxin on the cell surface. C d l 72, 847-8Fj6. Uninner. T., Mogil. R. J., LaFace, D., Yoo, N . J., hlahboubi, A,, Echeverii, F., Martin, S. J., Force, W. I{., Lynch, D. H., Ware, C . F., c't al. (199s). Cell-autonomons Fas (CD9S)/ Fas-ligand interaction inecliates activation-intlucetlapoptosis in T-cell hybridoinas. Nottire 373,44 1-444. B~i~iriell, B. A,, Heath, L. S., Adams, D. E., Lahti, J. M., and Kidd, \', J. (1990). Increased expression of a 58-kDa protein kinase leads to changes in the CHO cell cycle. Proc. N d . A c d . Sci. USA 87, 7467-7471. Camerini, D., Walz. G., Lornen, W.A,, Borst, J., and Seed, B. (1991). The T cell activation antigen C D E is a inember of the nerve gro\vth f;ictor/tumor necrosis fiictor receptor gene f"ami1y./. I t t i t n u n i d , 147, 3165-3169. Casciola Rosen. L. A., Anhalt, G. J., ant1 Rosen, A. (199ij). DNA-dependent protein kinase is one of a subset of autoantigens specifically cleaved early during apoptosis.J . E x p . Med. 182, 1625-1634. Casciola Rosen, L. A,, Miller, D. K., Anhalt, G.J., and Rosen, A. (1994). Specific cleavage of the 70-kDa protein component ofthe U1 sinall nuclear ribonucleoprotein is a characteristic biochernical featnre of apoptotic cell death. J , B i d . Ckem. 269, 30757-30760. Cerretti, D. P., Kozlosky, C. J.. Mosley. B., Nelson, N., Van Ness, K., Greenstreet. T. A,, March. C. J., Kronlieiin, S. R., Dmck, T., Carinizzaro, L. A,, et (11. (1992). Molecular cloning of the interleukin-1 beta converting enzyme. Scimcr 256, 97-100.
194
PETER H . KRAMMEH
ChaudhaI);, P. M., Eby, M., Jasmin, A,, Bookwaker, A., Murray, J., and Hood, L. (1997). Death receptor 5 , a new member of the TNFR family, and DR4 induce FADD-dependent apoptosis and activate the NF-kappaB pathway. Zrnmunity 7, 821-830. Chen, Z., Naito, M., Mashiina, T., and Tsuruo, T. (1996). Activation of actin-cleavable interleukin lbeta-converting enzyme (ICE) Family protease CPP-32 during chemotlierapeutic agent-induced apoptosis in ovarian carcinoma cells. Cancer Res. 56, 5224-5229. Cheng, J., Zhou, T., Liu, C., Shapiro, J. P., Braiier, M. J., Kiefer, M. C., Barr, P. J., and Mountz. J. D. (1994). Protection froin Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 263, 1759-1762. Chinnaiyan, A. M., O’Rourke, K., Tewari, M., m d Dixit, V. M. (1995). FADD, a novel death doinain-containing protein, interacts with the death domain of Fas and initiates apoptosis. Cell 81, 505-512. Chinnaiyan, A. M.. O’Rourke, K.. Yu, G. L., Lyons, R. H., Garg, M., Duan, D. R., Xing, L., Gentz, R., Ni, J., and Dixit, V. M . (1996a). Signal transduction by DR3, a death domain-containing receptor related to TNFR- 1 and CD95. Science 274, 990-992. Chinnaiyan, A. M., Orth, K., O’Rourke, K., D i m , H., Poirier, G. G., and Dixit, V. M. (19961~). Molecular ordering of the cell death pathway: Bcl-2 and Bcl-xL function upstream of the CED-Mike apoptotic proteases. J. B i d . Cheiii. 271, 4573-4576. Chinnaiyan, A. M., Tepper, C. G . , Seldin, M. F., O’Rourke, K., Kischkel, F. C., Hellbardt, S., Krammer. P. H., Peter, M. E., and Dixit, V. M. (1996~). FADD/MORTl is a coininon mediator of CD9S ( FadAPO-1) and tumor necrosis factor receptor-induced apoptosis. J. B i d . Cherri. 271, 4961-4965. Chinnaiyan, A. M., O’Rourke, K., Lane, B. R., and Dixit, V. M. (1997).Interaction of CED4with CED-3 and CED-9: A molecular framework for cell death. Science 275,1122-1126. Chittenden, T., Harrington, E. A,. O’Connor, R.. Fleinington, C., Lutz, R. J., Evan, G. I., and Guild, B. C. (1995). Induction of apoptosis by the BcI-2 hoinologue Bak. Nature 374, 733-736. Chiu, V. K., Walsh, C. M., Liu, C. C., Reed, J. C., and Clark, W. R. (1995). Bcl-2 blocks degranulation but not Fas-based cell-mediated cytotoxicity./. Ztiztnunol. 154,2023-2032. Choi, M. S., Boise, L. H., Gottschalk, A. R., Quintans, J., Thompson, C. B., and Klaus, G. G. (1995).The role of Bcl-Xl,in CD4O-mediated rescue froin anti-mu-induced apoptosis in WEHI-231 B lymphoma cells. Eur. J . Ittiniund 25, 1352-1357. Chu, K., Niu, X., and Williams, L. T. (1995). A Fas-associated protein factor, FAF1, potentiates Fas-mediated apoptosis. Proc. Not/. Acad. Sci. USA 92, 11894-1 1898. Cifone, M. G., De Maria, R., Roncaioli, P., Rippo. M. R., Azuma, M., Lanier, L. L., Santoni, A., and Testi, R. (1994).Apoptotic signaling through CD9g (Fas/Apo-1) activates an acidic sphingoniyelinase.J. E x p Metl. 180, 1547-1552. Cook, W. D., Metcalf, D., Nicola, N. A,, Burgess. A. W., and Walker, F. (1985). Malignant transformation of a growth factor-dependent inyeloid cell line by Abelson virus without evidence of an autocrine mechanism. Cell 41, 677-683. Copeland, K. F., Haaksma, A. G., Goudsmit, J., Krammer, P. H., and Heeney, J. L. (1994). Inhibition of apoptosis in T cells expressing human T cell leukemia virus type I Tax. AIDS Res. Hton. Retrooir. 10, 1259-1268. Coustan-Smith, E., Kitanaka, A,, Oui, C.-H., McNinch, L., Evans, W. E., Rairnondi, S. C., Brehm, F. G., Arico, M., Campano, D. (1996). Clinical relevance of Bcl-2 overexpression in clinical childhood acute lymphoblastic leukemia. B l o o d 87, 1140- 1146. Crowe, P. D., VanArsdale, T. L., Walter, B. N., Ware, C. F., Hession, C., Ehrenfels, B., Browning, J. L., Din, W. S., Coodwin, R. G., and Smith, C. A. (1994). A lymphotoxinbeta-specific receptor. Science 264, 707-710. Cryns, V. L., Bergeron, L., Zhu, H., Li, H., and Yuan, J. (1996). Specific cleavage of alpha-fodrin during Fas- and tumor necrosis factor-induced apoptosis is mediated by an
interleiikin-lheta-coiiverting enzynie/Ced-3 protease distinct from the poly(ADP-iibose) polymerase protease. J . B i d C/trrn. 271, 31277-:31282. Cuende. E., Ales Martinez. J. E.. Ding, L., Gonzalcz Garcia, M., Martinez. C.,and Niinez. G. (1993). Programmed cell death by bcl-2-tlependent and intlependent niechanisins in B Iymphoma cells. EhlBO J . 12, 1555-1560. Wieringa, B.. and Hentlriks, W.(1997). N o evidence for involvement of mouse protein-tyrosine pIiosPhatase-BAS-like Fas-associated phosphatase-1 in Fasmediated apoptosis. 1.B i d . c h i . 48, :3021Fj-:30220. Danley, D. E., Chiiang. T. H.. and Bokoch. G. M. (1996). Defective Rho GTPase regulation by IL-1 Iwta-converting enzvn~e-mediatetlcleavage of D4 GDP dissociation inhibitor. I. ~ I t I t J l f ~ J I O157, ~. 500-50:3. Dbaiho, G. S. ( 1997). Regulation of the stress response by cerainide. Bioclwn. Soc. Trans 25, 557-561. Debatin, K . M.. and Krammer, P. 11. (1'395). Resistance to APO-1 (CD95)induced apoptosis in T-ALL is determined by a BCL-2 independent anti-apoptotic program. Leukcinin 9, 815-8". Debatin, K.-M. (1997). Cytotoxic drugs. progrmiined cell death, and the irninrine system: Defining new roles in an old play. J . Nntl. Caticrr Zti.rt. 89, 750-751. Ilegli-Esposti, M . A.. Sinolak. 1'. J., \Valczak. H., Waiigh, J., Huang, C. P., DuBose, R. F.. Goodwin, R. G.. and Smith, C. .4.(1997a). Cloning and characterization of T R A L R 3 , a novel member of the emerging TRAIL receptor fainily. J . Eq,. Merl. 186, 1165- 1170. Degli-Esposti, M. A,, Dongali, W. C., Sinolak. P., LVaugIi, J. Y.,Smith, C. A,, and Goodwin, R. G. (199711).The novel receptor TRAIL-H4 induces NF-kappaB and protects against TRAIL-metliated apoptosis. yet retains an incomplete death dornain. Zmniutiity 7, 813-820. Demhic, Z., Loetscher. H., Gubler, U.. Pan. Y. C . . Lahm, H. It'., Gentz, R., Brockhaus, M . . and Lesslauer, W.(1990). T\vo huinan TNF receptors have similar extracelliilar, hut distinct intracellular. domain sequences. C!ytokittc, 2, 2 3 - 2 3 ? , Dhein, J.. Daniel, P. T., Trauth, €3. C., Oehm, A,. Moller, P.. and Kraininer, P. H. (1992). Induction of apoptosis by monoclonal antihody anti-APO-1 class switch variants is dependent on cross-linking of APO-1 cell surface antigens. J . ltwttrmd. 149, 3166-3173. Dhein, J., Walczak, H., B#umler, C., Debatin, K. M., and Krammer. P. H. (1995).Aiitocrine T-cell siiicide mediated b!. APO-l( Fas/CD95). Nmtrrre 373, 4.38-441. Dive, C., Evans, C. A,, and Whetton, A. D. (1992). Intliiction of apoptosis-new targets for cancer chcmotherapy. C a w e r B i d . 3, 417-427. Dole. M. G., Jasty, R., Cooper, M. J.. Thonipson, C. R., Nunez, G., and Castle, V. P. (1995). Bcl-xL,is expressed i n neuroblastoma cells and modrilates cheinotlierapy-indiIced apoptosis. Cntuxr H t x 55, 2576-2582. Duan, H.. Cliinnaiyan, A. M., Hudson, P. L., Wing, J. P., He, W.W., and Dixit, V. M . (1996a). ICE-LAP3, a novel niainnialian Iioinologiir~of the Caetior~trrhrlitiselegarts cell death protein Ced-3 is activated during Fas- ant1 tumor necrosis factor-induced apoptosis. J. Biol. C l i ~ i 271, . 1621-1625. Duan, H.. Orth. K . , Chinnaiyin, A. M.. Poirier. G. G.,Froelich, C. J.. Ile, W.W., and Ilixit. V. M. (l996h). ICE-LAP6, a novel nirin1)er of the ICE/Ced-3 gene family, is activated Iiy the cjtotoxic T cell protease granzymc B. J , B i d . C h e m 271, 16720-16724. Durkop, H., Latza, U.. I l i i n i i n c l , M., Eitelbach, F., Seed. B., a r i d Stein, H. (1992). Molecular cloning and expression of a new ineinlwr of thc iiwvc growth factor receptor family that is characteristic for IIodgkin's disease. Cell 68, 421 -427. Eck, M . J., and Sprang, S. R. (1989).The structure of tumor necrosis factor-alpha at 2.6 A resolution: Iinplications for receptor binding. J , B i d . Chetm 264, 17595- 17605.
196
PETER 11 KHkMMER
Eck, M. J., Ultsch, M., Rinderknecht, E., de Vos, A. M., and Sprang, S. R. (1992). The structure of human lymphotoxin (tumor necrosis factor-beta) at 1.9-A resolution. J. Biol. Chem. 267, 2119-2122. Emery, J. G., McDonnell, P., Burke, M. B., Deen, K. C., Lyn, S., Silverman, C . , Dul, E., Appelbaum, E. R., Eichman, C., DiPrinzio, R., et al. (1998). Osteoprotegerin is a receptor for the cytotoxic ligand TRAIL. J. Biol. Chem. 273, 14363-14367. Emoto, Y., Manome, Y., Meinhardt, G., Kisaki, H., Kharbanda, S., Robertson, M., Ghayur, T.. Wong, W. W., Kamen, K., Weichselbaum. R., et al. (1995). Proteolytic activation of protein kinase C delta by an ICE-like protease in apoptotic cells. EMBOJ. 14,6148-6156. Enari, M., Hase, A,, and Nagata, S. (1995a). Apoptosis by a cytosohc extract from Fasactivated cells. E M B O J. 14, 5201-5208. Enari, M., Hug, H., and Nagata, S. (1995b). Involvement of an ICE-like protease in Fasmediated apoptosis. Nature 375, 78-81. Enari, M., Sakahira, H., Yokoyama, H., Okawa, K., Iwamatsu, A,, and Nagata, S. (1998). A caspase-activated Dnase that degrades DNA during apoptosis, and its inhibitor ICAD. Nature 391, 43-50. Erhardt, P., Toniaselli, K. J., and Cooper, G. M. (1997).Identification ofthe MDM2 oncoprotein as a substrate for CPP32-like apoptotic proteases. J , Biol. Chem. 272, 15049-15052. Erickson, S. L., de Sauvage, F. J., Kikly, K., Carver Moore, K., Pitts Meek, S., Gillett, N., Sheehan, K. C., Schreiber, K. D., Goeddel, D. V., and Moore, M. W. (1994). Decreased sensitivity to tumour-necrosis factor but normal T-cell development in TNF receptor-2deficient mice. Nature 372, 560-563. Faleiro, L., Kobayashi, R., Feamhead, H . , and Lazebnik, Y. (1997). Multiple species of CPP32 and Mch2 are the major active caspases present in apoptotic cells. EMBO J. 16, 2271-2281. Farber, S., Diamond, L. K., Mercer, H. D., et al. (1948). Temporary remissions in acute leukemia in children produced by folic acid antagonist 4-aminopteoylglutamicacid (aminopterin). N. Engl. J , Med. 28, 787. Farrow, S. N., White, J. H., Martinou, I., Raven, T., Pun, K. T., Grinhani, C . J., Martinou. J. C., and Brown, R. (1995). Cloning of a bcl-2 homologue by interaction with adenovirus E1B. Nature 374, 731-733. Faucheu, C., Blanchet, A. M., Collard Dutilleul, V., Lalanne, J. L., and Diu Hercend, A. (1996).Identification of a cysteine protease closely related to interleukin-1 beta-converting enzyme. Eor. 1.Biochein. 236, 207-213. Faucheu, C., Diu, A., Chan, A. W., Blanchet, A. M . , Miossec, C., Herve, F., Collard Dutilleul, V., Gu, Y., Aldape, R. A,, Lippke, J. A,, et al. (1995). A novel human protease similar to the interleukin-1 beta converting enzyme induces apoptosis in transfected cells. EMBO J. 14, 1914-1922. Feniandes-Alnemri, T., Armstrong, H. C . , Krebs, J., Srinivasula, S. M., Wang, L., Bullrich, F., Fritz, L. C., Trapani, J. A,, Tomaselli, K. J., Litwack, G., and Alnemri, E. S. (1996). In vitro activation of CPP32 and Mch3 by Mch4, a novel human apoptotic cysteine protease containing two FADD-like domains. Proc. Natl. Acad. Sci. USA 93,7464-7469. Fernandes-Alnemri, T., Litwack, G., and Alnemri, E. S. (1994). CPP32, a novel human apoptotic protein with homology to Cnenorhahdiitis elegans cell death proteiii Ced-3 and mammalian interleukin-1 beta-converting enzyme. J. Biol. Chein. 269, 30761-30764. Feniandes-Alnemri, T., Litwack, G., and Alnemri, E. S. (1995a). Mch2, a new member of the apoptotic Ced-3/ICE cysteine protease gene family. Cancer Re.s. 55, 2737-2742. Femandes-Alnemri, T., Takahashi, A,, Armstrong, R., Krebs, J., Fritz, L., Tomaselli, K. J., Wang, L., Yu, Z., Croce, C. M., Salveson, G . , et al., (1995b). Mch3, a novel human apoptotic cysteine protease highly related to CPP32. Cancer Res. 55, 6045-6052. Fisher, D. E. (1994).Apoptosis in cancer therapy: Crossing the threshold. Cell 78,539-542.
Fisher, G. N.,Rosenherg, F. J., Straus. S. E., Dale, J. K., Middleton, L. A,, Lin, A. Y.. Strober, W., Lenardo, M. J.. and Prick, J. M. (1995). Dominant interfering Fas gene mutations impair apoptosis in a hriman autoininiunc lyiirplroproliferative syndrome. Cell 81, 935-946. French, L. E., Ilahne, M.. Viard. I., Rdlgnil)er, C;., Zanone, R., Becker, K., Muller, C., and Tschopp, J. (1996).Fas and Fas ligand i n embryos and adult mice: Ligand expression in several irniniine-piivilege(l tissues and coexpression in adult tissues characterized by apoptotic cell turnover. I. Cull. Biol. 133, 335-34:3. Friesen, C.. Herr, I., Kraninirr, P. IJ., and Ilehatin, K.-M. (1996). Involvenrent of the CD95(APO-l/Fas) receptor/ligand system in dnig indnced apoptosis in leiikemia cells. Notiire Mcrl. 2, 574-Fj77. Friesen. C.. Fulda, S., and Debatin, K . - M . (1997). Deficient activation of the CD95 (APO-l/Fas) system i n dnig-resistant cells. Lc.trkettiin 11, 1833-1841. Fnlda, S.. Los, M..Friesen, C., Scaffidi, C;. A,, Mier, W., Benrdict, M., Nufiez, G., Kramrner, P. H . , Peter, M. E., and Debatin, K.-M. (1998). Betulinic acid triggers CD95(APO-l/ Fas) and p53-independent apoptosis via activation of caspases in neuroectoderinal turnors. Concur Re.s. 57, 4956-4964. Fulda, S., Sieverts, H., Friescii, C.. Herr. I.. and Debatin, K.-M. (1997).The CD95 (APO-1/ Fas) system mrdiates dnig indiiced apoptosis in neuroblastoma cells. Coricer Rus. 57, 3823-3829. G a k , P. R., I-Iofinann, W. J., b’alczak. I{., Schdler, H . , Otto, G., Streininel, W., Kraminer, P. I I . , and Riinkel, L. (1995).Involvement ofthe CD95 (APO-1/Fas) receptor and ligand in liver damage. J, Exp. hled. 182, 1223-1230. Geley, S., Hartmann, B. L., Kapelari, K., Egk, A., Villnnger. A,. Heidacher, D. Greil, R., Auer, B., and Kofler, R. (1997). The interleukin 1 beta-converting enzyme inhibitor CrmA prevents APO-l/Fas- hut not glucocorticoid-iiIduced poly(ADP-ribose) polymerase cleavage and apoptosis in 1pphol)lastic leiikemiu cells. FEBS Lett. 402, 36-40. Ghayur, T., Banerjee. S., Hngunin, M., Butler. D., Herzog, L., Carter, A,, Quintal, L., Sekut, L., Talanian, R.. Paskind, M., Wong. W.. Kainen, R., Tracey, D., and Allen, 11. (1997).Caspase-1 processes IFN-gamnra-indiicing factor and regulates LPS-induced IFNgamma production. Nuttrrc 386, 619-623. Ghaynr, T., Hngunin, M., Talanian, R. \’.> Ratnofsky, S.,Quinlan, C., Emoto, Y., Pandey, P., Datta, R.. Huang, Y., Kharhanda, S., Allen, Ii., Kamen, R., Wong, W., and Knfe, D. (1996). Proteolytic activation of protein kinase C delta by an ICE/CED 3-like protease induces clraracteristics of apoptosis. J. EXJJ.Med 184, 2399-2404. Godfrey, W. R., Faponi, F. F.. Harara. M. A,. Bnck, D., and Engleman, E. G. (1994). Identification of a Iiuinan OX-40 ligantl, a costiiniilator of C D 4 f T cells with homology to tumor necrosis factor. J . EX)). Merl. 180, 757-762. Goltsev, Y. V., Kovalenko, A. V.,Arnold, E., Varfolomeev, E. E., Brodianskii, V. M., and Wallach, D. ( 1997). CASH, a novel caspase lroinologue with death effector domains. 1.Biol. Cherti. 272, 19641-19644. Goodwin, R. G.. Alderson, M . R.. Smith, C. A,, Armitage, R. J., VandenBos, T.. JerLy, R., Tough, T. W., Schoenborn. M. A,, Davis Smith, T.. Hennen, K . , ut (11. (1993a). Molecular and biological characterization of a ligand for CD27 defines a new Family of cytokines with IionioloFn, to tninor nwrosis factor. Cell 73, 447-456. Goodwin. R. G., Din, W.S., Davis Smith. T., Anderson, D. M., Girnpel. S. D., Sato, T. A,. Maliszewsla, C. R., Brannan, C. I.. Copeland. N. G., Jenkins, N. A., et ul. (1993b). Molecular cloning of a ligand for the indncible T cell gene 4-1BB: A member of an emerging family of cytokines with homology to tumor necrosis factor. Eur. J . Intmnnol. 23, 2631-2641.
198
PETER H . KRAMMER
Gorczyca, W., Bigman, K., Mittelman, A,, Ahtned, T., Gong, J., Melamed, M. R., and Darzynkiewicz, Z. (1993). Induction of DNA strand breaks associated with apoptosis during treatment of leukemias. Leukemia 7, 659-670. Gottschalk, A. R., Boise, L. H., Thompson, C. B., and Quintans, J. (1994). Identification of iinmunosuppressaiit-induced apoptosis in ii inurine B-cell line and its prevention by bcl-x but not bcl-2. Proc. Natl. Acad. Sci. IJS4 91, 7350-7354. Graf, D., Korthauer, U.. Mages, H. W., Senger, G., and Kroczek, R. A. (1992). Cloning of TRAP, a ligand for CD40 on human T cells. Eur. 1. Imtiuno!. 22, 3391-3194. Graninger, W. B., Seto, M., Boutain, B., Goldman, P., and Korsnieyer, S. J. (1987).Expression of Bcl-2 and Bcl-2-Ig fusion transcripts in normal and neoplastic cells. I. Clin. Inoest. 80, 1512-1515. Gray, P. W., Agganval, B. B., Benton, C. V., Bringman, T. S., Henzel, W. J.. Jarrett, J. A,, Leung, D. W., Moffat, B., Ng, P., Svedersky, L. P., et al. (1984). Cloning and expression of cDNA for human lymphotoxin, a lymphokine with tumour necrosis activity. Nature 312, 721-724. Grell, M., Douni, E., Wajant, H., Lohden, M., Clauss, M., Maxeiner, B., Georgopoulos, S., Lesslauer, W., Kollias, G., Pfizenmaier, K . , et al. (1995). The transmembrane form of tumor necrosis factor is the prime activating ligand of the 80 kDa tumor necrosis factor receptor. Cell 83, 793-802. Gu, Y., Kuida, K., Tsutsui, H., Ku, G., Hsiao, K., Fleming, M. A,, Hayashi, N., Higashino, K., Okaniura, H., Nakanishi, K., Kurimoto, M., Tanimoto, T., Flavell. R. A,, Sato, V., Hardmg, M. W., Livingston, D. J., and Su, M. S. (1997). Activation of interferon-gamma inducing factor mediated by interlerikin-lbeta converting enzyme. Science 275,206-209. Gu, Y., Sarnecki. C., Aldape, R. A,, Livingston, D. J., and Su, M. S. (1995). Cleavage of poly(ADP-ribose)polymerase by interleukin-1 beta converting enzyme and its honiologs TX and Nedd-2. /. Biol. Chem. 270, 18715-18718. Hahne, M., Himoldi, D., Schroter, M., Romero, P., Schreier, M., French, L. E., Schneider, P., Bomand, T., Fontana, A,, Lienard, D., Cerottini, J., and Tschopp, J. (1996). Melanoma cell expression of Fas(Apo-l/CD95) ligand: Irnplications for tumor immune escape. Science 274, 1363-1366. Han, D. K. M., Chaudhary, P. M., Wrigth, M. E., Friedman, C., Trask, B. J., Riedel, R. T., Baskin. D. G., Schwartz, S. M., and Hood, L. (1997). MRIT, a novel death-effector domain-containing protein, interacts with caspases and Bcl-X,, and initiates cell death. Cell Bid. 94, 11333-11338. Hanada, M., Aime Sempe, C., Sato, T., and Reed, J. C. (1995). Structure-function analysis of Bcl-2 protein. Identification of conserved domains important for hornodimerization with Bcl-2 and heterodirnerization with Baw. 1. Biol. Chem. 270, 11962-11969. Hata, EI., Matsuzaki, I%., Takeya, M., Yoshida, M., Sonoki, T., Nagasaki, A,, Kuribayashi, N., Kawano, F., and Takatsuki, K. (1995). Expression of Fas/APO-l(CD95) and apoptosis in tumor cells from patients with plasma cell disorders. Blood 86, 1939-1945. Hermine, O., Haioun, C., Lepage, E., d’Agay, M.-F., Briere, J., Lavignac, C., Fillet, G., Salles, G., Marolleau, J.-P., Diebold, J., Reyes, F., and Gaulard, P. (1996). Prognostic significance of Bcl-2 protein expression in aggressive non-Hodgkin’s lymphoma. Blood 87, 265-272. Herr, I., Bohler, T., Wilhelm, D., Angel, P., and Debatin, K.-M. (1997). Activation of the CD95 (APO-l/Fas) signaling by ceramide mediates cancer therapy induced apoptosis. EMBO J. 16, 6200-6208. Hsu, H., Solovyev, I., Colombero, A,, Elliott, R., Kelley, M., and Boyle, W. J. (1997).ATAR, a novel tumor necrosis factor receptor family member, signals through TRAF2 and TRAF5. /. Biol. Chem. 272, 13471-13474.
Hn, S., Vincenz. C.,Uuller. M., and Disit, \’. M . (1H97a). A novel family of Lira1 death effector domain-containing molecules that inhibit both CI195- and tumor iiecrosis ftactor receptor-1-indiicrtl upoptosis. J . Biol. Clwtti. 272, 9621-9624, I I u , S.,Vincenz. C . , Ni, J.. Centz, R., airtl Ilixit. \’, M . (19971,). I-FLICE, a novel inliilitor of tnmor ncwosis factor rccvptor-1- a i d CD9S-inchiced apoptosis. J . Biol. CIirttL. 272, 172S55-17257. IIiiang. B., Eberstadt, M.. Olejniczak, E. T.. Me~itlows.R. I!. ~ i n t lFesik, S . W.(1996).NMR strnctnre and inntugenesis ofthe Fas (Al’O-I/C1195)cleath tloinain. N(ztrrrc, 384,638-641. Inoliara, N., Koseki, T., Hu, Y..C h i , S..and Niinez, G. (1997). CLARP. a death effector doiliain-containing protein interacts with caspase-8 and regiilates apoptosis. Crll Biol. 94, 10717-10722. lrinler. M., Tlionirs. M., Hahnc>, M., Schnc~itlrr,P., Hofiiiaiin, K., Striner, V.,Bodnicr. J. I,,, Schroter, M., Biinis, K., Mattmarin, C., liiinoltli, D., French, L. E., and Tschopp, J. (1997). Inhihition of death receptor signals hy cellular FLIP. Nottire 388, 190- 19!5. Itoh, N.. Yonehara, S., Ishii. A.. Yoneliara. M., Mizushinin, S.,Sanieshinia, M., Ilase, A , , Seto, Y., and Nagata, S. (1991). The poIy~q)ticleencodetl by the cDNA for hiinian cell surface antigen Fas can nietliate apoptosis. Ccll 66, 233-243. Itoh. N..Tsiijimoto. Y., and Nagata, S. (199:3).Effbct of bcl-2 on Fas antigen-mediated cell death. 1. Z t t i i t ~ r t t i o l .151, 621-627. Jiiiittrlii, M.,Benedict, M., Tewari, M.. Sliayman, J. A,, ant1 Dixit, V. M. (1995). Bcl-x antl Bcl-2 inhibit T N F and Fas-induced apoptosis antl activation of pliospholipase A2 in breast c;ircinoin;i cells. Otirogcwc. 10, 2297-2:305. Janicke, R. U., L%’alker,P. A,, Lin. X Y.,and Porter. A . G. (1996). Specific cleavage of the retinoldastoma protein by an ICE-like protcasc in apoptosis. EA4BO J . 15, 6969-6978. Johnson, D.. Lanahan.A , . Biick, C. R., Schgal. A,, Morgan. C., Mercer, E., Bothwell, M.. and Chao, M. (1986).Expression and striictrire ofthe hnin;in N G F receptor. Cell 47,545-554. Jones. E. Y., Stuart. D. 1.. and Walker, N. P. (1992). (;r).stal stnictnre of TNF. Zirmrcnol. Ser. 56, 93- 127. Ju, S. T.. Panka. D. J., Cni, H., Ettinger, R., el Kliatil), M.. Sherr, D. H., Stanger, B. Z., and Marshak Rothstein. A. ( 1995).Fas(C;DYS)/FasLinteractions required for prograinmed cell deatli after T-cell acti\.ation. Notrrrc 373, 444-448. Kamens, J., Paskintl, M., Hugiinin, M., Talanian, 11. V., Allen, 11.. Banach, D., Buinp, N.. Hackett, M . . Johnston, C. G.,Li. P., ~t al. (199Ti). Itlentification and characterization of I C I 1-2, a novel member of the interleiikin-l I)et;t-coiiverting enzyme family of cysteine proteases. J . Biol. C/iem. 270, 15250- 15256. Kayagaki. N.. Kawasaki. A,. Ehuta, T.. Ohinoto. H.. Ikeda, S.. inoue, S., Yoshino, K.. Okiimnra, K.,and Yagita. H. ( 1995). Metalloprotrinasc-mediated release of human Fas ligand. J . E q i . h f d 182, 1777-1783. Ka)ial;ir. C.. Ord, T., Testa, M . P., Zhong. L. T.. antl Bredesen, D. E . (1996). Cleavage of actin by interlcnkin 1 hrto-converting enqxie to reverse DNase I inhibition. Proc. N d l . Aiwl. Sci. l G A 93, 2234-2238. Kaufniann. S. H., Desnoyers, S., Ottaviiilo. Y.. Davidson, N . E., and Poirier. G. G. (1993). Specific proteolytic cleavage of pol? ( ADP-rit)ose) plyinerase: An early marker of chemotherapy-indiiced apoptosis. CattcrJt-Rex. 53, 3976-33985. Kerr, J. F., Wyllie, A. H., and Cnrrie, A. It. (1972).Apoptosis: A hasic bio1ogic;tl phenomenon with wide-ranging iniplications i n tissnc. kinetics. Br. J . Cottcrr 26, 239-257. Kiefer. M. C., Braner. M. J., Powers. \’. C . , \Vu, J . J., Uniansky, S. R., Toinei. L. D., and Barr. P. J. (1995). Modulation of apoptosis by the widely distril)uted Bcl-2 Iioinologue Bak. Notirrc, 374, 736-739. Kischkel, F. C., IIcllhardt, S.. Belirniann. I., Grrmer, M., Pawlita. M., Krammer. P. 11.. and Peter, M , E. (1995).Cytotoxicity-drpentlent APO-1 ( Fas/CD95)-associated proteins
200
PETER ti. KKAMMEH
form a death-inducing signaling complex (DISC) with the receptor. EMBOJ.14,55795588. Kischkel, F. C., Kioschis, P., Weitz, S., Poustka, A., Lichter, P., and Kraminer, P. H. (1998). Localization of the human and inurine caspase-8 (FLICE/MACH/MchS) genes by FISH. Submitted for publication. Kitson, J., Raven, T., Jiang, Y. P., Goeddel, D. V., Giles, K. M., Pun, K. T., Grinham, C. J., Brown, R., and Farrow, S. N. (1996). A death-domain-containing receptor that mediates apoptosis. Nature 384, 372-375. Klas, C., Debatin, K.-M., Jonker, R. R., and Krammer, P. H. (1993). Activation interferes with the apoptotic pathway in mature human T cells. Znt. Znirnunol. 5(6), 625-630. Knipping, E., Debatin, K.-M., Heilig, B., Eder, A.. and Krammer, P. H. (1995).Identification of soluble APO-1 in supernatants of human B and T cell lines and increased serum levels in B and T cell leukemias. Blood 85, 1562-1569. Kolesnick, R. N., Haiinovitz Friedman, A,, and Fuks, Z. (1994). The sphingoinyelin signal transduction pathway mediates apoptosis for tumor necrosis factor, Fas, and ionizing rahation. Biochem. Cell Biol. 72, 471-474. Kozopas, K. M., Yang, T., Bnchan, H. L., Zhou, P., and Craig, R. W. (1993). MCL1, a gene expressed in programmed myeloid cell differentiation, has sequence similarity to BCLS. Proc. Natl. Acad. Sci. USA 90, 3516-3520. Krajewski, S., Blomqvist, C., Franssila, K., Krajewska, M., Wasenius, V.-M., Niskanen, E., Nordling, S., and Reed, J. C. (1995). Reduced expression of proapoptotic gene Bax is associated with poor response rates to combination chemotherapy and shorter survival in women with metastatic breast adenocarcinoma. Cancer Res. 55, 4471-4478. Krajewski, S., Krajewska, M., Shabaik, A,, Miyashita, T., Wang, H. G., and Reed, J. C. (1994). Immunohistochemical determination of in vivo distribution of Bax, a dominant inhibitor of Bcl-2. Am. J Pathol. 145, 1323-1336. Krammer, P. H. (1997).The tumor strikes back: New data on expression of the CD95(APOW a s ) receptodigand system may cause paradigm changes in our view on drug treatment and tumor immunology. Cell Death Diff 4, 362-364. Krammer, P. H., Behrmann I., Daniel, P., Dhein J., and Debatin K.-M. (19944. Regulation of apoptosis in the immune system. Curr. +in. Zmmnunol. 6, 279-289. Krammer, P. H., Dhein, J., Walczak, H., Behrmann, I., Mariani, S., Matiba, B., Fath, M., Daniel, P. T., Knipping, E., Westendorp, M. 0.. Stricker, K., Baumler, C., Hellbardt, S., Germer, M., Peter, M. E., Debatin, K.-M. (1994b). The role of APO-1 mediated apoptosis in the immune system. Ininwrnol. Rev. 142, 175-191. Krammer, P. H., Galle, P. R., Mdler, P., and Debatin, K.-M. (1998). CDSS(APO-l/Fas)mediated apoptosis in normal and malignant liver-, colon- and hematopoietic cells. I n “Advances in Cancer Research” (G. F. Vande Wonde and G. Klein, eds.), vol. 75, pp. 252-273. Academic Press. Kuida, K., Lippke, J. A,, Ku, G., Harding, M. W., Livingston, D. J., Su, M. S., and Flavell, R. A. (1995). Altered cytokine export and apoptosis in mice deficient in interleukin-1 beta converting enzyme. Science 267, 2000-2003. Kuida, K., Zheng, T. S., Na, S., Kuan, C., Yang, D., Karasuyama, H., Rakic, P., and Flavell, R. A. (1996). Decreased apoptosis in the brain and premature lethality in CPP32- deficient mice. Nature 384, 368-372. Kumar, S., Kinoshita, M., Noda, M., Copeland, N . G., and Jenkins, N. A. (1994). Induction of apoptosis by the mouse Nedd2 gene, which encodes a protein similar to the product of the Cuenorhubditis elegans cell death gene ced-3 and the mammalian IL-1 betaconverting enzyme. Genes. Dev. 8, 1613-1626.
Kwon, B. S., Tan. K. B., Ni. J., Lee, K. O., Kim, K. K.. Kim, Y.J., \Vang, S., Gentz. R., Y u , C:. L., Harrop, J., Lyn. S. D.. Silverinan, ( ; . ~Porter, T. G.,Triineh, A,, and Yonng, P. 13. (1997).A newly itlentifietl ineml)er of the tiimor necrosis factor receptor superfamily &it11a \vide tissue distril)ntion and involvcwic.nt in Iyinphoc>-te;ictivation. /. Biol. Chcrn. 272, 14272- 14276. Lacry, D. L.. Tininis, E.. Tan, €I. L.,Kelley, M. J., Ihiistan, C . R.. Bnrgess, T., Elliott, R., Colonibero, A.. Elliott, G . , Scrilly, S.. ct cr/. (1998).Osteoprotegerin ligand is a cytokine that regulatc,s osteoclast differentiation and activation. Cull 93, 165-176. Lahti. J. M.. Xi;ing, J.. Heath, I,. S.. C a n p n a , D.. and Kitld, V. J. (1995).PITSLRE protein kinase acti\it\. is associatd with apoptosis. 1 ! 4 0 / . C:c// Biol. 15, 1-1 1. LeithYnser, F.. Uhein, J., Mechtershciiiier. C . , Koretz, K., Bruderlein, S., Ilenne, C., Schniidt. A., Debatin, K. M., Krammrr. P. H,, and Miiller, P. (199.3). Constitutive and intlncetl expression of APO- 1, a ne\v nlrmhrr of the nerve growth factorltumor necrosis Factor receptor superfamily. in normal and neoplastic cells. Lab. Itwest. 69, 415-429. Lenardo, M. J. (1991). Interleukin-2 progrmis inoiisv alpha beta T lymphocytes for apoptosis. Nattirc. 353, 858-861. Levine. A. J. (1997).1353, the cellular gatekeeper for growth antl tli\.ision. Crl/ 88, 323-331. Li, P.. Allen, II., Benrrjee, S.,Franklin, S., Herzog. L.. Johnston. C., McDowt~ll,J., Paskintl. M., Rotlinan. L., Salfeld. J., et i d , (1995). Mice deficient in IL-1 beta-converting enzyme are defective i n production of inaturc. IL-I beta and resistant to ( d o t o x i c shock. C P / / 80,401-411. Liclrtrr, P., \Valczak. II., Ll’eitz, S. Belirnrann. I., and Krammer. P. €1. (1992).The hiinlan APO-I antigen maps to 1Oq23. ii region which is syntenic with iiioiise cliroinosoine 19. Ccnornits 14, 179-180. Lin, E. Y.,Orlofkhy, A,, Berger, M . S.. and Pnstowsky. M. B. (1993). Characterization of Al, ii novel 1iernoI)oietic-spc.cific early-response gene with sequence similarity to 1x1-2. /. r l l l r r l l ~ r t o / .151, 1979-1988. Lippke, J. A,, Gn, Y.. Samecki, C., Caroii, P. R., and SII. M. S . (1996). Identification and characterization of CPP3WMch2 homolog 1, a iiovcl cysteine protease siinilar to CPP32. /. Biol. Chcttt. 271, 1825-1828. Liu, X..Zon, II.. Slaughter. C.. and \\’ang, X. (1997). DFF. a heterodiineric protein that fnnctions downstream of caspse-3 to trigger D N A fi.agnicWation during apoptosis. Cell 89, 17.5-184. Loetscher, H., Pan, Y. C.. Lahin, 14. Lj’., Ckntz, It., Brockliaris. M.. Tahiichi, H.. and Lesslauer, \V, ( 1990).Moleciilar cloning and vspression of the human 55 kd tririior necrosis factor receptor. Crd 61, 351 -359. Los. M.. Herr, I . , Friescn, C.. Fnltla, S.. Scliiilze-Ostlioff, K.. antl Dehatin. K.-M. (1997). Crossresistance of CD9.5- and tlmg-induced apoptosis as a consequence of deficient activation of caspises ( IC:E/Cetl-:3 proteases). Blood 90, 31 18-3129. Los. M., Van de Craen, M., Penning, L. C., Schenk. I I . , Westendorp, M.. Baeurrle, P. A,. Driige, LV.. Kramint~r,P. I.I., Fivrs. \’\:.> antl Schiilze Osthoff, K. (1995). Requirement of iin ICElCED-.3 protease for Fas/APO-I -mcdiatcd apoptosis. Notiire 375, 81-83. Lowe. S.W’.,Kirley, H. E., Jacks, T.. and IIonsnian, D. E. (1993).p53-tlcpendent apoptosis nrodiilates the cytotoxicity of anticancrr agrnts. Ci.11 74, 957-967. Mallett, S.,Fossnin, S.. and Barclay, A . N . i 1990). Characterization of the MRC OX40 antigen of activated CD4 positive T lyinph?-tes: A molecule related to nerve growth factor receptor. EAlBO /. 9, 1063- 1068. Mariani, S. M.. Matilxi, B., Biiiimler. C., and Kraininer, P. H. (1995). Regulation of cell surface APO-I/Fas (CD95) ligand expressioll l)y inrtalloproteases. Eiir. /, I t i i r r i i m o / . 25, 2303-2307.
202
I’ETEH M . KHAMMEH
Mariani, S. M., and Kraminer, P. H. (1998). Ilifferential regulation of TRAlL/APO2L and CD95L in transformed cells of the T and B lymphocyte lineage. Eur. J. I i n t ~ i u d 28, 973-982. Marsters, S. A,, Sheridan, J. P.. Doiiahue, C. J., Pitti, R. M., Gray, C. L., Goddard, A. D., Bailer. K. D., and Ashkenazi, A. (1996). Apo-3, a new member of the tumor necrosis factor receptor family, contains a death domain and activates apoptosis and NF-kappa B. Curr. Bid. 6, 1669-1676. Marsters, S. A,, Sheridan, J. P., Pitti, R. M., Huang, A. D., Skubatch, A. D., Baldwin, A. D., Yuan, A. D., Gurney, A. D., and Goddard, A. D. (1997). A novel receptor for APOZmRAIL contailis a truncated death domain. Cirrr. B i d 7, 1003-1006. Martin, S. J., O’Brien, G. A,, Nishioka, W. K., McGahon, A. J., Mahboubi, A,, Saido, T. C., and Green, D. R. (1995a).Proteolysis of fodiin (non-erythroidspectrin)chringapoptosis. J. B i d . Cheni. 270, 6425-6428. Martin, S. J., Reutelingsperger, C. P., McCahon, A. J., Rader, J. A,, van Schie, R. C., LaFace, D. M., and Green, D. R. (1995b). Early redistribution of plasma membrane phosphatidylserine is a general feature of apoptosis regardless of the initiating stimulus: Inhibition by overexpression of Bcl-2 and Abl. J. Exp. Med. 182, 1545-1SC56. Masliima, T., Naito, M., Fujita, N., Noguchi, K., and Tsuruo, T. (1995). Identification of actin as a substrate of ICE and an ICE-like protease and involvement of an ICE-like protease but not ICE in VP-16-induced U937 apoptosis. Biocherti. Biophys. Re.9. Commun. 217, 1185-1192. Mashima, T., Naito, M., Noguchi, K., Miller, D. K., Nicholson, D. W., and Tsunio, T. (1997).Actin cleavage by CPP-3Yapopain duriiig the development of apoptosis. Oncogene 14, 1007-1012. McDonnell, T. J., Deane, N., Platt, F. M., Nunez, G., Jaeger, U., McKeam, J. P., and Korsrneyer. S. J. (1989). bcl-2-iminiinoglobulin transgenic mice demonstrate extended B cell survival and follicular lymphoproliferation. Cell 57, 79-88. Medema, J. P., Scaffidi, C., Kischkel, F. C., Shevchenko, A,, Mann, M., Krammer, P. H., and Peter, M. E. (19974. FLICE is activated by association with the CD95 death-inducing signaling complex (DISC). E M B O J. 16, 2794-2804. Medenia, J. P., Toes, R. E. M., Scaffidi, C., Zheng, T. S.,Flavell, R. A,, Melief, C. J. M., Peter, M. E., Offringa, R., and Krammer, P. H. (1997b). FLICE activation by granzyme B during CTL mediated apoptosis. Eur. J. Itnmunol. 27, 3492-3498. Medema, J , P., Scaffidi,C., Krammer, P. H., and Peter, M. E. (1998).Bcl-xl,acts downstream of FLICE/caspase-8, J. B i d . Chem. 273, 3388-3393. Memon, S. A,, Moreno, M. B., Petrak, D., and Zacharchuk, C. M. (1995). Bcl-2 blocks glucocorticoid- but not Fas- or activation-induced apoptosis in a T cell hybridonia. J. Ininiunol. 155, 4644-4652. Metcalf, D., Roberts, T. M., Cherington, V., and Dunn, A. R. (1987).The in vitro behavior of hemopoietic cells transformed by polyoma middle T antigen parallels that of primary human inyeloid leukemic cells. EMBO J. 6, 3703-3709. Micheau, 0.. Solary, E., Hammann, A,, Martin, F., and Dimanche-Boitrel, M. T. (1997). Sensitization of cancer cells treated with cytotoxic dnigs to Fas-mediated cytotoxicity. J. Natl. Cancer Inst. 89, 783-789. Min, Y. H., Lee, S.,Lee, J. W., Chong, S. Y., Mahn, J. S., and KO, Y. W. (1996). Expression of Fas antigen in acute inyeloid leukemia is associated with therapeutic response to chemotherapy. Br. J. Haemnutol. 93, 928-930. Miossec, C., Dutilleul, V., Fassy, F., and Diu-Hercend, A. (1997). Evideiice for CPP32 activation in the absence of apoptosis during T lymphocyte stimulation. J. Biol. Chem. 272, 13459-13462.
Mittl, P. R.. Di Marco, S.. Krelis, J. F., Hai. X., Karanewsky, 11. S., Priestle, J. P.. Tomaselli, K. J.. and Chitter, M. C. (1997). Structure of recoinbinant hiiman CPP:32 in complex with the tetrapeptide acctyl-Asp-Val-.41~i-Aspflrioronirthyl ketone. J . Biol. CIwn. 272, (3539-6547. Miura, M., Friedlander, R. M., and Yuan, J. ( 1995).Tunior necrosis factor-indriced apoptosis is niediated by a CrniA-sensitive cell death pathway. Proc. Natl. Acad. Sci. USA 92, 83188322. Miyashita, T., and Reed, J. C. ( 1992). bcl-2 gene transf&r incrcwes relative resistance of S49. 1 and LVEH17.2 lyniphoid cells to cell death antl DNA fragmentatioil indiiced hy glucocorticoids a i i d rnultiple clieiiiotlierapeutic. drugs. Cancer Res. 52, 5407-541 1 , Mipashita, T., and Reed, J , C;. (1993). Bcl-2 oncoprotein blocks chemotherapy-induced apoptosis in ii human leukemia cell liiic.. Blootl 81, 151-157. Miyashita, T., Krajewski, S., Krajc,wska. M.. LYang. H. G., Lin, H. K., Lieberinann, D. A,. Hoffman. B., and Heed, J. C. (1994). Tiiirior s iip 1mwo r p53 is a regulator of Bcl-2 and ba); gene expression in vitro antl in vivo. Oticogctie 9, 1799-1805. Miyashita. T., and Reed, J. C. (1995). Tunior suppressor p5C is a tlirect transcriptional activator of the huniaii hi^^ geiir. C d l 80, 293-289. M o k r . P., Koretz, K., Lritliiinsrr, F.. Briiderlein, S., Heiiiie, C., Quentmeier, A , , and Krammer, P. 11. (1994). Expressions of APO-1 (CD95). a member of the NGF/TNF receptor siiperfamily, in norinal and iieoplastic coloii epithelium. Int. /. Cnriccr 57, 1-7. Mongkolsapaya, J., C o y e r , A. E., Xu, X N., Morris, G,, McMichael, A. J., Bell, J. I., antl Screaton. G. R. (1998).Lymphocyte iiihibitor of TRAIL (TNF-related apoptosis-indiicing ligaiitl): A new receptor protecting lyinphocytes from the death ligand TRAIL./. I u i i t i u r w l . 160, :3-6. Montgomery, R. I., Warner, M . S., Luin, 8 . J., iind Spear, P. C. (1996). Herpes simplex virus-1 entiy into cells inetliutetl by ii novel member of the TNF/NGF receptor family. Cell 87, 427-436. Muss, M. L,, Jinq S. !!., M i h > kl. E.. BiirkIi;irt%\V.> ( h r t e r > tl. I , . . Chen. W.J.. ., Price. I., Chien, K. R.. and Evans, R. M. (1994). RXRa mutant inice esta6lisli a genetic hasis for vitamin A signaling in heart niorpliogenesis. Gmes Dec 8, 1007- 1018. Siido. T., Nisliikawa, S., Ohno. N., Akiyania. N.. Euriakoshi, M.. Yosliida, H., and Nishikawwa, S. I. (1993). Expression and fiinctioii of the interlrukin 7 receptor in murine lymphocytes. Proc. N d Acmrl. Sci. USA 90, 9125-9129. Tacliiliana, K., Nakajima, T., Sato, A,, Igarashi, K., Shida, H., lizasa, H., Yoshida, N., Yosliie, O., Kishimoto, T.. antl Nagasawa, T. (1997). CXCR4lfiisin is not a species-specific 1,arricr in miirine cells for HIV-1 entiy. J. E r p M d . 185, 1865-1870. Tan;ibe, S., Heesen, M., Yoshizawa, I., Berman, M. A,, Luo, Y., Bleul, C . C., Springer, T. A,, Okiitla. K,,Gvrard. N., and l h r f , M. E. (1997). Functional expression of the
228
TAKASHI NACASAWA et a1
CXC-cheniokine receptor-4fusin on mouse niicroglial cells and astrocytes. I. Zrntntmol. 159, 905-911. Tashiro, K., Tada, H., Heilker, R., Shirozu, M., Nakano, T., and Honjo, T. (1993). A cloning strategy for secreted protein and type I mernbrane proteins. Science 261, 600-603. voii Freeden-Jeffiy, U., Vieira, P., Lucian, L. A,, McNeil, T., Burdach, S. E. G., arid Murray, R. (1995). Lymphopenia in interleukin-7 gene-deleted mice identifies IL-7 as a nonredundant cytokine. I. Exp. Med. 181, 1519-1526. This article was accepted for publication on March 24, 1998.
h l ) \ h Y ( E \ I Y I M M l I Y O L O C . ~ \ O L 71
T Lymphocyte Tolerance: From Thymic Deletion to Peripheral Control Mechanisms BRlGlTTA STOCKINGER Division of Molecular Immunology, National Institute for Medical Research, The Ridgeway, Mill Hill, London, NW7 IAA United Kingdom
1. Introduction
Our understanding of tolerance induction in lymphocyte subpopulations has been exponentiallyadvanced hy the generation of T- and B-cell receptor transgenic mice. Initial contradictions, which seemed to appear as different transgenic models were published, have given way to overall consensus, at least as far as central tolerance induction in the thymus is concerned. It is now possible to dwriminate between clonal elimination and other ways of rendering lymphocytes unresponsive. The debate of whether tolerance is due to suicide or murder has been decided in fhvor of the former. Despite the assumption that we understand most of the phenomena resulting in central tolerance induction in thymus and bone marrow, a new area of uncertainty lies in our understanding of how tolerance can be induced and maintained in the periphery. This review concentrates on central and peripheral tolerance in the T-cell population only, given that this series has dealt comprehensively with self-tolerance in the B-cell population (Goodnow rt al., 1995). Table 1 lists T-cell receptor (TCR) transgenic mouse strains published to date. II. Central Tolerance Induction in the Thymus
A. CLONAL DELETION AS T I I E MAJORPHIN( IPLE OF TOLERANCE INDUCTION IN THE TIIIMUS
The first direct demonstration for the elimination of potentially autoreactive cells in the thymus came from studies of the T-cell repertoire to endogenous superantigens (Kappler et al., 1987; MacDonald et al., 1988). T cells bearing Vp segments that conferred binding to endogenous mouse inainmary tumor viruses (Mtv)were absent from mice carrying these superantigens in their genome. Analysis of thymic subpopulations established that there was deletion of T cells with Vp segments conferring superantigen reactivity at the immature CD4+8+stage. Shortly thereafter the first T-cell receptor transgenic inoiise strain appeared (Kisielow et al , 1988) bearing T-cell receptors specific for an epitope on the male antigen 229
C opynght 0 l9YY In Academic P r m hll nglit\ 01 rqxmliictwn I I ) d m form n w nrtl “65 27ifd99 $30 00
230 TABLE I T-CELLRECEPTOR TRANSCXNIC MOUSESTRAINS TCR Transgenic Strain 2 c (Sha et al., 1988) H-Y (Kisielow et nl., 1988) C6 (Donek et al., 1996) F5 (Mamalaki et al., 1993) KB5.C20 (Schonrich et al., 1991) BM3 (Spoonas et a/., 1994) OT-1 (Hogquist et al., 1994) 8.3 TCR (Verdaguer et al., 1997) K TCR (Geiger et nl., 1992) F3 (Morahan, et al., 1991) N15 (Ghendler et al., 1997) Clone-4 TCR (Morgan et a/., 1996) P14 (Pircher et al., 1989)
Specificity H2-Ld + QL9 peptide (Sykulev et al., 1994) H-Y peptide from Smcy (Markiewicz et al., 1998) H-Y peptide from Smcy (Scott et al., (1995) InHuenza NP 366-374 H2-K"
MHC Restriction
Clonotypic ab, Rag-/-, scid, or TCRa-1-
H2h class I
1B2 clonotypic mAh
H2-D"
Rag2-1T3.70 chotypic 111.4b
H2-Kk
Clonotypic mAh Va8, Vpll
H2-D" H 2'
Ragl -1Desire clonotypic niAb Ragl -198 clonotypic mAb Ragl-lVa2 mAh
H2' Ovalbuinin 257-264
HZ-K"
Undefined islet ag
H2-K"
SV40 T
h2-k1
H2-Kh VSV nuclear N52-59
H ~ - K ~
Va8 mAb Rag2-I-
InHueiiza hemaglutinin LCMV gp 32-42
H~-K"
DO.ll.10 (Murphy et d., 1990) AD10 (Kaye et al., 1989)
Ovalbntnin 323-339
H2-A'
Pigeon cytochrome c 88-104
H~-E'
2B4 (Berg et al., 1989)
Moth or pigeon cytochronie c Undefined islet ag
H2-E' H2-AI"
scid TCR a-I-
Influenza HA 111-119
H2-E"
Rag2-ITCRa -16.5 clonotypic mAb
IgG2a" 435-451 MBP 1-11 MBP 1-9
H2-A' H2-A" H-2A"
BDC.2.5 (Katz et d, 1993) 14.3d (Kirberg et al., 1994) B5 (Granucci et al., 1996) (Goverman et al., 1993) (Liu et al., 1995)
H2-Db
V a 2 , Vp8.1 mAh Rag2-IKJI-26 clonotypic tnAb
An ti-Vp3 Anti-Val1 Rag2-1-
231
T LYMPHOCYTE TOLERANCE
TABLE I (continued) TCR Trancgenic Strain
Specificity
MHC Restriction
~______
~
Clone 19 (1,afaille et a/ , 1994) Tag TCRl (Forstcxr ct (11, 1995) .3A9 (Ho ct a1 , 1994) 5C C7 (Seder et a l , 1992)
A18 (Zal c’t 0 1 , 1994) A1 (Zelenika et a / , 1998) 4B2A1 (Bogen et a/., 1993) Smarta (Tourne ct a[., 1997) TEa (Grubin P t d.,1997) SEP (Klein e t a ! . , 1998)
DEP (Klein c>t nl , 1998)
TCR-LACK (McSorley et al., 1997) T2.5-5 (Scott pf d., 1994) 4.1 TCR (Schmidt et d . , 1997) ”
Clonotypic ah, Rag-/-, scitl. or TCRa -1-
MRP 1-9
Rag1 -1-
SV40 large T 362-384
H2-A‘
HEI, 46-61 Pigeon cytoclrroine c 88-104 Mouse C5 106-121 HY peptide rniknown A”5 91-101
9H5 clonotypic mAb Rag1-13A9 anti-a antiserum Ragl -1-
H2-Ek 112-Eh 112-E“
LCMV pp 61-80
H2-Ah
Ragl -1Rag1-1scid GBI 13 clonotypic mAb Va2, Vp8.3 mAb
Ea52-68 C-reactive protein 80-94 C-reactive protein 89-101 Lei.dirrurnia mc!jor LACK 158-173 IIA 126-138 Undefined islet ag
112-Ah H2-Ah
Va2, Vp6 inAb
€12-A”
Vall.2 inAb
H2-A”
Va8, Vp4 mAb
Rag2-1-
Reagents identifying T-cell receptor in FACS analysis
H-Y. Again, deletion of T cells specific for H-Y was apparent in immature thymocytes at the double positive CD4’8’ stage in male mice. However, the time point of deletion was not identical for these systems. Generally, superantigen reactive T cells were deleted only at the late CD4+8+stage before transition to single positive mature cells, whereas deletion in CD8 T-cell receptor transgenic mice seemed to affect thymocytes at an earlier stage, leading to profound reduction of the double positive subpopulation and highly reduced thymus cellnlarity. However, CD4 T-cell receptor transgenic strains with physiological ligands showed a phenotype of “late” deletion as seen in superantigen-driven deletion (Douek et al., 1996; Zal et al., 1994: Zelenika et al., 1998).Although the timing of tolerance induction may depend on the localization of antigen (Pircher et al., 1989; Schneider ct al., 1992)or antigen-presenting cells (APC), large numbers of TCR transgenic strains, both of CD8 and CD4 lineage, provided evidence of a
232
BXIGITTA STOCKINGEX
heterogeneity in the timing of deletion, which is best explained by the overall avidity of T-cell receptor/antigen/major histocompatibility complex (MHC) interactions (Oehen et al., 1996). Transgenic strains vary in the time of onset and levels of T-cell receptor expression in different thymic subpopulations and in their coreceptor expression and receptor affinity. In addition there are differences in concentration and localization of cognate antigen and its access to different APC populations. Interactions that exceed a threshold of avidity tend to result in deletion, and an increase in one of the parameters that define the overall avidity can tip the balance. This is illustrated in various transgenic models; e.g., overexpression of CD8 coreceptor was shown to change a positive selection signal to a negative selection signal (Lee et al., 1992; Robey et al., 1991),expression of different levels of transgenic V p l l resulted in different degrees of deletion in the presence of H2-E (Homer et al., 1993), and changes in the density of alloantigen influenced the deletion pattern for thymocytes with a class I alloreactive receptor (Auphan et al., 1991). Deletion of antigen-specific thymocytes apparently requires a lower threshold of antigen stimulation than activation of mature T cells (Pircher et al., 1991; Vasquez et al., 1994). Thymic deletion can be induced in T-cell receptor transgenic mice through the injection of large amounts of cognate peptide. In these cases deletion affects the bulk of immature double positive thymocytes, resulting in a drastic reduction of thymus cellularity. However, a potential complication of this procedure is the concomitant activation of mature single positive thymocytes as well as peripheral T cells, resulting in the release of cytokines such as tumor necrosis factor (TNF) (Mamdaki et al., 1992; Martin and Bevan, 1997; Tarazona et al., 1998). It is currently unclear whether the nonspecific component of deletion is a consequence of direct TNF action or whether TNF acts on stromal cells, which in turn induce an apoptotic process in thymocytes (Lerner et al., 1996). Viral or bacterial infections can lead to transient thymic involution due to the deletion of immature double positive thymocytes, emphasizing that cytokine production in the thymus or periphery following strong T-cell activation can contribute to T-cell deletion in the thymus in a nonspecific manner (Wang et al., 1994; Bonyhadi et nl., 1993; Godfraind et al., 1995).
B. WHICHAPC PRESENT SELF-ANTIGEN FOR NEGATIVE SELECTION? Based on initial studies with bone marrow chimeras, positive and negative selection of thymocytes was ascribed to distinct cell types. Cortical epithelial cells were shown to effect positive selection, whereas bone marrowderived hematopoietic cells were responsible for tolerance induction (Lo and Sprent, 1986; Sprent et al., 1975; Bevan, 1977; Zinkernagel et al.,
1979; Marrack et al., 1988; Bix and Raulet, 1992; Cosgrove et al., 1992; Brocker et al., 1997; Markowitz et al., 1993). However, the involvement of epithelium and especially cortical epithelium remained controversial. Although some reports denied the contribution of cortical epithelium to deletion (Laufer et ul., 1996), others report at least partial effects, such as anergy induction, attributable to epithelium (Hugo et al., 1994; Gao and Sprent, 1990; Roberts et al., 1990; Ramsdell et nl., 1989; Tanaka et al., 1993). The situation was coinplicated further by the choice of antigens and the functional readoiits that were used to assess tolerance. Although thyinocytes with specificity for major liistocoinpatibility antigens appeared to ignore the presence of allo-MHC molecules on thymic epithelium (Ready et d ,1984; von Boehmer and Schubiger, 1984), it became clear subsequently that cortical epithelium and peripheral APC can present distinct sets of self-peptides on allo-MHC (Marrack et al., 1993). Thus, a functional readout of T-cell tolerance on peripheral APC may miss components of epitheliuin-induced tolerance. This was illustrated directly in experiments showing that thymic epithelium could tolerize in a tissuespecific manner (Bonomo and Matzinger. 1993). Furthermore, in contrast with cortical epithelium, medullaiy epithelium may be quite efficient in tolerance induction (Hoffniann et al., 1992; Burkly et al., 1993; Klein et al., 1998) so that results obtained in radiation bone inarrow chimeras do not solely address the tolerance-inducing potential of cortical epithelium. Different experimental systems come to seemingly contradictory conclusions even when addressing tolerance induction to similar antigens. While expression of H2-E in medullary epithelimn was sufficient in one case to induce tolerance in VPFj and V P l l CD4 T cells reactive to superantigen, grafting of H2-E positive branchial clefts, which are free of hematopoietic cells, f d e d to tolerize these T cells, although they induced tolerance to I-E itself (Salaiin c>tal., 1992). Furthermore, the potential to tolerize may depend on antigen presentation pathways available to different thvmic APC. For instance, CD4 T cells specific for pigeon cytochrome c were tolerized whether the antigen was expressed as a membrane protein or targeted to mitochondria after synthesis in the cytoplasm (Oehen et ul., 1996). However, expression of P-gal as a nuclear protein only resulted in tolerance induction when it was expressed by medullary epitheliuin, but not bone marrow-derived cells (Oukka et al., 1996). The tolerance-inducing potential of all thymic M HC class I1 expressing cells for a defined self-antigen in a TCR transgenic system was investigated. This was done in reaggregation cultures with fetal thymocytes from C5 T-cell receptor transgenic mice and purified thymic APC isolated from mice that naturally express the self-antigen CS in their blood circulation
234
B R I C; ITTA STOC KI N C, E R
(Volkmannet al., 1997).Because these experiments relied on endogenously processed self-antigen in APC subpopulations purified from thymus of adult CS+donors, they directly addressed the inherent capacity of different thymic APC to internalize, process, and present for T-cell deletion. In this system, cortical epithelial cells and medullary epithelial cells proved as efficient as dendritic cells in deletion; in fact, the only APC population that was unable to tolerize the T cells were macrophages. However, the levels of endogenous antigen in C5 expressing mice are high (about SO pg/ml in blood) so that the comparative efficiency of cortical epithelial cells for negative selection may not have been put to a stringent test. Taken together, the available collection of data suggests that thymic epithelial cells from cortex and medulla can be involved in negative selection, but they may not be as efficient as dendritic cells in low avidity situations (limiting amounts of antigen, inefficient internalization of exogenous antigen, reduced levels or affinity of T-cell receptors). The ability of thymic macrophages to induce negative selection has been put into question before. They were defective in clonal deletion of H2-E reactive T cells (Miyazaka et al., 1993) and their abject failure to tolerize CS specific T cells was notable. Low levels of MHC class I1 and adhesion molecules such as ICAM-1 (see later) presumably contributed to this phenotype. It remains to be tested whether macrophages are able to induce negative selection for MHC class I-restricted T cells.
C. ROLEOF OTHERCELLSURFACE RECEPTORSI N THYMIC NEGATIVESELECTION Given the importance of costimulatory signals in the activation of mature T cells, it appears likely that negative selection requires more than just a signal through the T-cell receptor. The most obvious candidate costimulatory molecules-B7.1 and B7.2 and their interaction with CD28-do not seem as obligatory for thymic deletion as they are for the induction of responses in peripheral mature T cells (Shahinian et al., 1993; Walunas et al., 1996; Tan et al., 1992). CD28 knockout mice do not show impaired negative selection, and the contribution of B7 molecules to negative selection in various in vitro or in vivo models of negative selection remained ambiguous (Punt et al., 1994; Jones et al., 1993; Amsen and Kruisbeek, 1996). This could partly be due to problems with the assays used for negative selection; only one study addressed physiological negative selection in the presence of self-antigen in viuo. In addition, a likely explanation is that there are more costimulatory molecules operative in negative selection. This is supported by the finding that cortical epithelial cells that do not express B7 molecules are efficient APC for induction of negative selection in C5-specific thymocytes (Volkmann et al., 1997). Additional
T LYMPHOCYTE TOI,EHAN(:E
235
candidates are ICAM-1 (Carlowet al., 1992; Pircher et al., 1993;Kishimoto et nl., 1996), CD40 (Foy et nl., 1995), and CD30 (Amakawa et al., 1996). The role of Fas (CDYS), another member of the TNFR family, and TNF in thymic negative selection is ambiguous. Mice deficient in Fas and Fas ligand have normal T-cell development in the thymus and show apparently unimpaired negative selection of superantigen reactive thymocytes, although one study describes a modulating effect of Fas on the early stages of negative selection (Castro et al,, 1996). In addition, Fas appears to play a role in negative selection of “semimature” T cells in the thymus, a population that has undergone positive selection and downregulated one of the coreceptors, but has not acquired full maturity to exit the medulla yet (Kishimoto and Sprent, 1997). In the majority of studies, negative selection of antigen-specific transgenic thymocytes or of superantigen reactive T cells was reported to occur in the absence of Fas or Fas ligand and blocking of TNF did not interfere with negative selection either (Sidman et al., 1992; Parijs et al., Abbas, 1996; Sytwu et nl., 1996; Adachi et al., 1996). Although most of the negative selection experiments were carried out by injection of peptide, with the potential problems of specificity mentioned earlier, one study used HY TCR transgenic mice arid deletion of transgenic thymocytes by endogenously expressed ligand in male mice was found to proceed normally in Fas ligand or Fas-deficient HY TCR transgenic mice (Schwartz, 1997). Thus, whereas evidence that both Fas and TNF play an important role in the deletion of peripheral T cells following activation (activation induced cell death) is unequivocal, available data do not indicate a major role in thymic negative selection. The regulation of negative selection in the thymus is also influenced by thymus-derived steroid hormones. Nur77, a ligand-dependent transcription factor that belongs to the steroid receptor superfamily, is strongly involved in thymic negative selection. Expression of a dominant negative version of this molecule in transgenic mice interferes with the deletion of self-specific thymocytes (Zhou et al., 1996: Calnan et al., 1995). Interestingly, despite defective clonal deletion in the thymus, peripheral selfspecific T cells that are present in increased numbers in such mice were found to be functionally anergic and prone to increased Fas-mediated deletion. What is the quantitative impact of thymic negative selection on the normal T-cell repertoire? In normal mice, expression of endogenous superantigens has been shown to markedly decrease the total number of reactive T cells reaching maturity (Simpson et al., 1993).In bone marrow chimeras expressing MHC molecules on radioresistant thymic epithelial cells, but not on bone marrow-derived cells, the number of mature thymocytes was increased twofold, indicating that about half of the positively selected
236
BRICITTA STOCKINGER
thymocytes may normally be lost through negative selection (van Meemijk et al., 1997). Although this is a substantial number, it is nevertheless marginal if compared with the extent of thymocyte death due to “neglect” (Surh and Sprent, 1994). D. SIGNALING FOR NEGATIVE SELECTION An important question is what dictates whether contact with antigen will result in deletion or in activation of a T cell. Are tliymocytes inherently more “deletable” than mature T cells? Burnet states in his clonal selection theory that the immune system is geared to distinguish between self and non-self. This, however, should not be interpreted to mean that a T cell “knows” what is a self-antigen to which it is supposed to be tolerant and what is a foreign antigen to which it should develop an immune response. Clearly the developmental stage of a T cell is important for that distinction. While mature peripheral T cells can tie deleted (whether contact is with “self” or “foreign” antigen), this process is preceded by at least partial activation if not full-blown differentiation to effector cell status (see Section 11,2). Developing thymocytes, however, receive signals through engagement of T-cell receptor and coreceptor with antigedMHC complexes before developing the capacity to mount an immune response. Curiously, however, signals through the T-cell receptor can have diametrically opposed outcomes in tlie thymus, leading to either positive or negative selection. The current consensus is that the avidity of such signals determines the fate of developing thyniocytes. Thus, any antigen encountered in the thymus whose interaction with T-cell receptors triggers a signal above a certain threshold would initiate deletion. Although it was not possible to directly establish clonal deletion as the underlying mechanism inducing tolerance in the classical cases of neonatally induced tolerance to alloantigens (Billingham et al., 1953; Owen, 1945; Anderson et al., 1951), several subsequent examples of experimental tolerance induction strongly suggested, or even directly demonstrated, this mechanism (Qin et al., 1989; MacDonald et al., 1988; Ciliak and Lelimann-Grube, 1978). It should be pointed out that this interpretation of the neonatal tolerance experiments does not imply that tlie neonatal period per se is more conducive to tolerance induction rather than induction of immunity. Even within the thymus, susceptibility to deletion extends to a certain window of T-cell differentiation only. Once thymocytes have reached the fully mature single positive state, they respond to anti-TCR stimulation in a manner analogous to peripheral T cells (Stockinger et al., 1996; Volkmann et al., 1997; Ramsdell et al., 1991; Nikolic Zugic and Bevan, 1990; Dyall and Nikolic Zugic, 1995; Barthlott et al., 1997). It is likely, but remains to be formally proven,
'r
LYMPHOWTTE TOLERANCE
237
that the signaling machinery in a mature thymocyte resembles that of a mature peripheral T cell, rather than that of a double positive thymocyte. Insights into TCR signaling pathways have emerged through analysis of nonreceptor protein tyrosine hnases expressed in lymphocytes, such as p59"", p56Ick,and ZAP-70, all of which are required for normal T-cell '~ development and signaling. While the presence of ZAP-70 and ~ 5 6 ' are obligatory for positive selection and negative selection in the thymus and essential for signaling in mature T cells (Negishi et al., 1995; Hashimoto et nl., 1996), differences einerged at distal points in the signaling cascade, indicating that T-cell receptors can deliver different types of signals by variably recruiting distinct, downstream signal transduction pathways (Alberola Ila et nl., 1997). Among the large number of molecules that pass on protein kinase-derived signals are small GTP-bindng proteins, notably p2lrAS, which participates in signaling from a wide variety of protein tyrosine kinases. Expression of a dominant negative form of p21ras was shown to block activation of the IL-2 gene in T-cell lines, indicating its important role in relaJing TCR-mediated signals. Transgenic expression of dominant negative p21" profoundly inhibited thymocyte positive selection, but did not interfere with negative selection, suggesting that TCR-derived signals resulting in negative selection may not use a ras-mediated pathway (Swan et al., 1995). Similarly, dominant negative MAP kinase MEKl affected positive, but not negative, selection (Alberola-Ila et al., 1995). The Rho family GTP exchange factor Vav (Turner et d ,1997), which is critical for proliferation and IL-2 production of mature T cells and essential for positive selection in the thymus, only plays a partial role in negative selection. Vav mutants show draniatically reduced calcium fluxes, suggesting that a defect in this pathway accounts for an absence of positive selection and reduced efficiency of negative selection. Blocking of calcium fluxes with the intracelMar calcium chelator has previously been shown to either abolish negative selection (Vasquez et ul., 1994) or at least to block negative selection to weak, but not strong, ligands (Kane and Hedrick, 1996). Thus, in some situations, interfering with a particular signaling component might affect positive and negative selection differentially because of quantitative differences in the strength of the signals required for these two processes. However, a mere quantitative effect has been excluded, at least for the involvement of the ras signaling pathway, as complete inhibition by simultaneous expression of dominant negative ras and MEKl still only affected positive, but not negative, selection ( Alberola-Ila et al., 1995). E. ESCAPEFROM THYMIC; DELETION While deletion of self-specificthyrnocytes might be considered the major component in the induction of immunologiical tolerance, ample evidence
238
BHIGITTA STOCKINGER
shows that thymocytes can escape from deletion, even if self-antigen is present in the thymus. Escape from thymic deletion is possible for thymocytes that have decreased levels of either coreceptor (Teh et al., 1989) or T-cell receptor (Schonrich et al., 1991; Mamalaki et al., 1995), which may result in decreased avidity and therefore signals that are not sufficient to induce apoptosis. However, it is possible that an apparent phenotype of receptor downregulation is due to the expression of a second endogenously rearranged T-cell receptor in addition to the transgenic T-cell receptor. Such endogenous rearrangements, leading to the expression of additional T-cell receptor a chains, are prominent in most TCR transgenic mouse strains not backcrossed to Rag-/- and can influence the degree of negative selection. Expression of a second T-cell receptor in addition to a selfantigen-specific TCR has been shown to allow escape of T cells from thymic deletion despite the presence of abundant levels of self-antigen (Zal et al., 1996). This potential drawback does not apply to the mice described by Mamalah et al., (1995). In double transgenic Ragl-/- mice expressing an NPspecific CD8 T-cell receptor and endogenous NP peptide, a proportion of CD8 cells appears to be deleted in the thymus, but there are CD8 cells in the periphery. These have reduced levels of receptor and coreceptor and characteristically upregulated levels of CD44, a marker attributed to memory or activated cells (Stout and Suttles, 1992; Budd et al., 1987). Likewise, double transgenic mice that express the antigen influenza hemagglutinin (HA) under control of the K light chain promoter and CD4 T cells specific for HA have antigen nonresponsive CD4 cells in the periphery with increased levels of CD44 (Lanoue et al., 1997). Interestingly, in all reported cases of escape from thymic deletion, despite the presence of self-antigen in the thymus, the escapees in the periphery were functionally inert. In the case of HA and NP-specific T cells their state of unresponsiveness must have been imprinted in the thymus, as already the mature single positive thymocytes abnormally expressed the activation markers found on the peripheral T cells. In other cases (e.g., Zal et al., 1996), dual TCR expressing T cells escaped from thymic deletion in the presence of abundant levels of self-antigen because their levels of the self-antigenspecific (C5) TCR were so low that they remained ignorant of its presence. There is no systematic analysis concerning the life span of peripheral escapees. If there is any correlation with the B-cell system one could assume that functionally unresponsive peripheral T cells probably have a short half-life.
F. CAVEAT. Is OURPERCEPTION OF CLONAL DELETION IN THE THYMUS AN ARTIFACTOF TRANSGENIC MODELS? In the B-cell system an intriguing mechanism preventing wholesale loss of B cells through negative selection has been described, termed receptor
T LYMPHOCYTE TOLEHANCE
239
editing. The original papers based on two different B-cell receptor transgenic mouse models stated that editing of light chains at the pre-B-cell stage allows survival of a proportion of B cells that otherwise would have expressed a self-specific receptor and be deleted (Gay et al., 1993; Tiegs et al., 1993). Pelanda et al. (1997) illustrate that B cells with a self-specific receptor that had been targeted into the endogenous locus by homologous recoinbination are not subject to deletion, but undergo very efficient receptor editing. In conventional, noiitargeted B-cell receptor transgenic mice the extent of clonal deletion might have been overestimated, as the transgenes cannot be deleted by Ig gene rearrangements. If this situation is applied to T-cell selection in the thymus, it is immediately obvious that the occurrence of T cells carrying receptors with other TCR a chains, due to incomplete allelic exclusion at the TCR a locus, are far more prominent in mice that express the self-antigen recognized by the transgenic TCR. This is usually interpreted to be due to the accumulation of such cells in the face of massive deletion of autoreactive cells. No TCR transgenic mouse exists so far that was created by homologous recombination into the correct locus, and the number of T cells with endogenously rearranged receptors varies widely in different TCR transgenic models. This begs the question of whether the assumption of thymic deletion, similar to the B-cell situation, may be an overestimate. Given the fact of massive loss of T cells through Failed positive selection, it might be preferable to avoid deletion following self-antigen recognition by an editing mechanism, such as additional rearrangements of TCR a chains, which would allow a second chance to create an innocuous T-cell receptor. This would be a way to avoid “holes in the repertoire” created by eliminating a wide range of cross-reactive receptor specificities. The existence of autoimmune diseases prompts the question of whether this constitutes failure of thymic negative selection or breakdown of control mechanisms in the periphery. It stands to reason that it must be impossible for every self-antigen in the body to have access to the thymus for tolerance induction. T cells with specificity for autoantigens can be found in the repertoire of healthy humans (Ota et nl., 1990; Pette et al., 1990; Salvetti et nl., 1991; Sommer et al., 1991).Surprisinglythough, a nuinber of presumed tissue-specific proteins have been found to be expressed in the thymus. A number of protein structures considered specific components of the central nervous system are contained in the thymus (Mathisen et al., 1993; Pribyl et al., 1996: Fritz and Zhao, 1996). Transgenic mice constructed with regulatory elements from the insulin gene can express transgenes both in pancreas and in thymus. This was initially thought to reflect the consequence of the promiscuous expression of hybrid insulin-promoter transgenes, but in the meantime thymic expression was demonstrated for a nuinber of pancreas-specific genes, including endogenous, nontransgenic
240
BRIGI'ITA STOCKINGER
elements. (Jolicoeur et nl., 1994; Smith et al., 1997; Antonia et al., 1995). It is impossible to decide whether autoimmunity can arise as a consequence of changes in levels or onset of thymic expression of tissue-specific antigens and whether the target structures recognized reflect thymic expression. In this context the success of experimental tolerance induction affecting dlabetes, by injection of a single protein, such as GAd or insulin (reviewed by Wegmann, 1996), is certainly perplexing, Taken together, it must be assumed that T cells carrying receptors specific for autoantigens are present in the periphery. Whether those T cells ever make contact with self-antigen and what mechanisms exist to prevent their activation and progression to autoimmune disease are the questions facing us under the topic of peripheral tolerance. 111. Peripheral Tolerance
A. T-CELLIGNORANCE, T-CELLRECIRCULATION There must be many self-antigens expressed in peripheral tissues that never access the thymus and therefore cannot be expected to have induced deletion of their specific T cells, yet autoimmune disease is a relatively rare phenomenon. This means that there are effective ways of avoidlng the activation of potentially self-reactive T cells. Conceptually the simplest of possibilities is to avoid contact of T cells with their self-antigen on peripheral tissue. This is termed T-cell ignorance and at its basis is the particular pattern of lymphocyte recirculation, which differs crucially for naive T cells compared with effector and memory T cells (Mackay, 1993). Naive T cells recirculate through secondary lymphoid tissue, but do not access extralymphoid sites such as skin, joints, or pancreas. In contrast, once T cells are activated, they will recirculate through lymphoid and extralymphoid sites (reviewed in Butcher and Picker, 1996). A striking example of T-cell ignorance was described by Ohashi et al. (1991) in a transgenic system. Transgenic mice expressing lyinphochorionieningitis (LCMV) viral glycoprotein in pancreas p islet cells were crossed with mice carrying a transgenic T-cell receptor specific for a peptide within the LCMV glycoprotein. The transgenic T cells remained unresponsive (tolerant) and phenotypically naive unless they were infected with LCMV virus. Infection abolished unresponsiveness and resulted in CD8 T-cell-mediated diabetes. However, T-cell ignorance was not always the outcome when antigen was expressed in the pancreas and the differences may be due to the type of T cell (CD8 or CD4) and the quality of its TCR, as well as its affinity and the amount of antigen expressed (Lo et al., 1992; Morgan et al., 1996: Degermann et al., 1994; Scott et al., 1994; Morahan et a/.),
T LYMPHOCYTE: TOLERANCE
24 1
Some organs, such as the eye, brain, and testis, are considered immuneprivileged sites because antigens expressed in them do not induce immune responses. These organs maintain even stronger barriers to the entry of recirculating lymphocytes than other nonlymphoid sites. Apart from immunological ignorance, mechanisms invoked for maintaining tolerance to such “sequestered” antigens include a switch to liumoral rather than cellmediated immunity (see Section II,D) and the presence of soluble mediators with suppressive characteristics, e.g., TGF-P (Ksander and Streilein, 1994). In cases where ignorance prevails it might be explained with differential migration patterns, as naive T cells are not expected to get into contact with their self-antigen if it is not present in lymphoid organs. However, a number of observations cloud this simplistic view of peripheral tolerance. First, there seems to be large-scale traffic of virgin T cells through extralymphoid sites in early ontogeny (demonstrated in the sheep fetus; Kimpton ct al., 1995), which would indicate that there is contact between naive T cells and tissue-specificantigens early in development. Thymic emigrants were shown to migrate to tlie skin within a few hours after exiting the thymus and it was suggested that these cells might be more susceptible to tolerance induction than circulating peripheral T cells (Washington et al., 1995). Furthermore, it was shown in a transgenic model that neonatal, but not adult, T cells had access to alloantigen expressed exclusively on keratinocytes (Alferink et nl., 1998). The accessibility of peripheral sites appears to be due to a developmental switch in expression of the lymphocyte homing receptor and adhesion molecules, favoring MAdCAM as addressin and a4 p 7 as homing receptors in fetal and early neonatal life (Mebius et nl., 1996). If this is the case one might expect contact of potentially self-reactive T cells and self-antigen on peripheral tissues, at least during a short period in life. Recognition of antigen on nonprofessional APC. such as tissue cells, is unlikely to induce a productive immune response due to the lack of signal 2 (a costiinulatory signal). Such a recognition event may have multiple potential outcomes, such as ignorance, anergy, or partial activation, resulting in differentiation to a regulatory subset. The various mechanisms invoked for induction and maintenance of peripheral tolerance will be discussed in detail. Even more perplexing is the role of dendritic cells in the activation of T cells with specificity for self-antigens expressed on peripheral tissues. To what extent do dendritic cells acquire such antigens, do they present them to T cells, and where? Polly Matzinger, in her “danger” hypothesis, makes tlie case that a productive immune response against any antigen, be it intrimic or foreign to an organism, requires a danger signal that activates APC, notably dendritic cells. The absence of such signals (which
242
BRIGI’ITA STOCKINGER
constitute bacterial or viral products, injury, necrotic death, inflammatory cytokines, etc.) in a quiescent immune system would preclude induction of a T-cell response and thus guarantee self-tolerance in the periphery (Matzinger, 1994). Because the exact nature of the danger signals resulting in activation and migration of dendritic cells are very difficult to define, it is not known if there is a degree of “constitutive” traffic of dendritic cells carrying peripheral self-antigen into draining lymph nodes. A report by Kurts et al. (199713) suggests that dendritic cells (or bone marrowderived APC at least) carry ovalbumin expressed in the plasma membrane of pancreas p cells into the draining lymph nodes. There is apparently no indication of an insult to the pancreas resulting from overexpression of this foreign protein and no inflammation, suggesting constitutive migration of dendritic cells transporting this antigen into lymphoid organs. If this was routinely the case and if dendritic cells are constitutively able to offer costimulatory signals, one might expect a much larger potential for the induction of autoimmune recognition events. More recently, such cross presentation for the activation of CD8 T cells by dendritic cells has been demonstrated to occur as the result of apoptosis, but not necrosis, of cells expressing the antigen in question (Albert et al., 1998). Curiously, no similar clear-cut data exist with respect of antigen pick up by dendritic cells for the activation of CD4 T cells, although one might have thought that it should be easier for exogenous antigens to reach the conventional class I1 presentation pathway rather than the class I presentation pathway. Although the jury is still out to decide whether overexpression of ectopic proteins in various organs will cause disturbance of normal cell physiology, resulting perhaps in increased cell turnover and release of cellular constituents, another point of consideration is the status of maturity of the antigenpresenting dendritic cells. Resting tissue dendritic cells do not constitutively express high levels of MHC class I1 and B7, although these molecules can be upregulated very rapidly. Is it conceivable that constitutively migrating dendritic cells fail to get signals inducing their maturation and that such dendritic cells are tolerogenic rather than immunogenic? It will be very difficult to devise experiments to test this ponit as it is currently impossible to separate dendritic cell migration and maturation or even isolate dendritic cells avoiding their activation.
B. PERIPHERAL DELETION, EXHAUSTION Studies of immune responses to endogenous and exogenous superantigens provided compelling evidence that contact of mature T cells with antigen can lead to their deletion (Webb et al., 1990; Jones et al., 1990; Critchfield et al., 1994; Kawabe and Ochi, 1991).The notion that prolonged exposure of lymphocytes to antigen can lead to “exhaustion” has been
T LYMPIIOCYTE TOLERANCE
243
around much earlier (Byers and Sercarz, 1968; Sterzl and Silverstein, 1967), but in the age of transgenic mice it became possible to study the fate of individual T cells. Deletion of mature T cells as a consequence of excessive stirnulation or antigen persistence is now widely demonstrated not just for superantigens, but also for conventional antigens (MacDonald et al., 1991; Moskophidis et al., 1993; Rocha and von Boehmer, 1991; Kyburz et al., 1993; Bogan, 1996). In contrast to the deletion of iininature thymocytes, mature T cells undergo a period of activation and even expansion before they are deleted, although this may have limited functional consequences (Renno et al., 1995; Forster and Lieheram, 1996; Kurts et al., 1997b; Ehl et al., 1998). Clearly, high antigen dose and chronic stimulation favor elimination of both CD4 and CD8 T cells. The silencing of T cells following persistent engagement of their antigen receptors in the periphery may represent a continuous process, ranging from activation to unresponsiveness to deletion, depending on signal strength and exposure time. Induction of peripheral deletion can be prevented or even reversed in the presence of proinflamniatory substances, such as lipopolysaccharide or poly : IC, or infections with VSV, Listeria, vaccinia virus, or Nipostmngylzis bmsiliensis (Ehl et al., 1998; Chiller and Weigle, 1973; Rocken et al., 1992; Vella et al., 1995), emphasizing the role of additional stimuli other than antigen for the generation of productive immune responses ( Janeway et al., 1996; Matzinger, 1994). One mechanism of deletion is terminal differentiation to short-lived effector cells that may die by apoptosis due to lyinphokine depletion. The latter mechanism could be prominent for CD8 T cells. The absence of CD4 help (Kirberg et al., 1993; Guerder and Matzinger, 1992; K~irtset a?.,1997a)promotes tolerance induction by deletion in the CD8 population, and conversely CD4 T cells can prolong the survival of CD8 T cells, perhaps by supplying cytokines, such as IL-2. Both Fas- and TNF-mediated cell death can contribute to the deletion of effector T cells (Sytwu et aZ., 1996: Crispe, 1994; Parijs et nl., 1996). In addition, a feedback regulatory mechanism that controls the intensity of immune responses is involved in the susceptibility of mature T cells to programmed cell death. The latter operates when T cells are exposed to large concentrations of antigen under fully activating conditions and results from the exposure to IL-2 followed by cell cycle progression (Lenardo et a1 , 1995). The physiological significance of the latter mechanism in self-tolerance is unclear, but it has been proposed as a strategy for high dose antigen therapy of autoimmune encephalomyelitis (Critchfield et a1 , 1994). C. ANERCVA N D PARTIAL SIGNALING: I N T CELLS The activation of T cells requires at least two signals: a signal through the T-cell receptor and a second signal presumed to involve a costimulatory
244
RHI(;ITTA STOCEINGEH
molecule(s) on an antigen-presenting cell (Bretscher and Cohn, 1970; Lafferty et nl., 1983; Mueller et al., 1989). The antigenic signal alone cannot activate T cells and has been shown to result in unresponsiveness, termed anergy (Mueller et al., 1989; Schwartz, 1990; Boussiotis et al., 1994b). Classical T-cell anergy has been defined in vitro using Thl T-cell clones (Jenhns and Schwartz, 1987; Schwartz, 1997); anergic cells were reported to have a block in p21 ras activation (Fields et nl., 1996), decreased activation of several MAP kinases (Li et al., 1996), reduced c-Fos and JunB protein induction (Mondino et al.. 1996), and fail to phosphorylate the AP-1 protein required for IL-2 gene activation (Kang et nl., 1992). Anergy induced in vitro can be reversed by IL-2 and by signaling through the common y chain of the IL-2, 4,and 7 receptors and does not lead to cell death (Boussiotis et al., 19944. While the molecular mechanisms leading to anergy in T cell clones are quite well defined, the situation is far more complex in v i m Anergy is used as a generic term for nonresponsiveness and the underlying reasons and mechanisms for the failure to respond can be manyfold. Anergy in vivo is well documented in the B-cell system; transgenic B cells developing in the presence of soluble self-antigen (HEL) are functionally unresponsive due to a biochemical block in the tyrosine kinase-signaling cascade initiated by IgM and IgD receptors. This inhibits the accumulation of phosphotyrosine on the receptor-associated CD79a and p chains and on the collaborating syk tyrosine kinase (Cooke et al., 1994). Anergic cells were shown to have a very short half-life and failed to enter follicles when competing with normal B cells (Cyster et al., 1994): thus B-cell anergy can be viewed as a slow form of deletion. Induction of tolerance in mature peripheral T cells has a long-standing history. In early studies, tolerance was induced by injection of high or low doses of soluble, deaggregated proteins without any adjuvants (Dresser, 1961;Chiller et al., 1971; Mitchison, 1964; Romball and Weigle, 1993), but the success of this strategy was variable and not extendable to every protein. The underlying mechanisms remained undefined, but could be explained by more recent evidence that antigens presented to T cells in the absence of inflammation (as caused by adjuvants) tend to be tolerogenic (Janeway et al., 1996; Matzinger, 1994; Kearney et d.,1994; Pape et a!., 1997; Kyburz et a/., 1993). Additional mechanisms resulting in high dose tolerance were discussed in Section I1,B. The CTLA-4 molecule, which is induced in T cells on activation, has been invoked in the control ofperipherd tolerance (Chambers et al., 1996). CTLAWB7 interactions regulate T-cell expansion, and CTLA-4 deficient mice have a profound lymphoproliferative disorder (Waterhouse et nl., 1996). It is clear that the uncontained proliferation in CTLA-4 deficient mice is attributable to CD4 T cells (Chambers et al., 1997), suggesting
T LYMPHOCYTE TOI.ERANCE
245
that this molecule is particularly important in the homeostatic control of CD4 T-cell responses. CTLA-4 ligation results in the inhibition of IL-2 production and consequent arrest in cell cycle progression (Kruminel and Allison, 1996). CTLA-4 could also compete for CD28/B7 interactions and thus deliver negative signals, resulting in T-cell mergy (Thompson and Allison, 1997). Induction of anergy following priming with peptide in incomplete Freunds adjuvant was prevented in the presence of antiCTLA-4 antibodies (Perez et d.,1997), suggesting that some form of costimulation is involved in the induction of anergy. The presence of a CTLA-4-specific antibody before the onset of insulitis in mice with a diabetogenic transgenic T-cell receptor precipitated the onset of diabetes dramatically, whereas antibody injection at later stages had no effect S, engagement of CTLA-4 at different (Luhder et al., 1998).T ~ L Idifferential time points in ongoing T-cell responses can have distinct consequences for antigen recognition. One cell type that is frequently associated with the induction of anergy in T cells are B cells. They are shown to be dispensable for the initial priming of naive T cells in viuo (Lassila ct al., 1988; Ronchese and Hausmann, 1993), although some aspects of T-cell priming seem to be impaired in mice lacking B cells (Kurt-Jones et nl., 1988; Ron and Sprent, 1987). Moreover, presentation of antigen by B cells in uivo does not seem a neutral event, but instead induces tolerance (Eynon and Parker, 1992; Fuchs and Matzinger, 1992). It is noteworthy that these experiments all involved antigen nonspecific B cells that either internalized antigen via nonspecific endocytosis or expressed it intracellularly (H-Y antigen in inale B cells). This may make a crucial difference for the functional activity of B cells. First, nonspecific antigen uptake is very inefficient in B cells unless they are activated (Chesnut et a)., 1982), whereas antigen-specific B cells internalize large amounts of antigen via Ig specific endocytosis (Lanzavecchia, 1990). Not only does this guarantee increased occupancy of MHC class I1 molecules with a specific peptide, but cross-linking of Ig receptors (Finkehnan et a1 , 1992) and antigen internalization via Ig receptors also results in the upregulation of B7 inoleciiles (Ho et al., 1994). B cells that do not signal effectively tlirough their antigen receptors fail to activate T cells for CD40 ligand upregulation, IL-2 secretion, and proliferation. However, they may still affect partial activation, resulting in upregulation of activation markers such as CD69 and CD44 (Ho et al., 1994; Croft ct al., 1997). This phenotype, as mentioned earlier, seems to be a hallmark of anergic B cells. Taken together, one may have to view with caution a general assumption that B-cell antigen presentation to naive T cells in uivo will cause anergy, as long as this has not been tested with antigen-specific B cells. B cells that encounter antigen that they can internalize via their
246
BRIGITTA STOCKINGER
Ig receptors may be as capable of inducing productive T-cell responses as dendritic cells. However, high doses of protein antigens or peptides injected in vivo are likely to be targeted to B cells not expressing cognate Ig receptors, given the high frequncy of B cells compared with antigenspecific B cells or with dendritic cells, and would not result in effective B7 upregulation. Under these circumstances it is conceivable that CTLA-4 competes with CD28 for limited B7 molecules so that full activation of T cells is stalled through the curtailment of IL-2 production. The life span of anergized peripheral T cells within the same host (following thymectomy) has not been studied systematically. Instead, the fate of adoptively transferred T cells that were either activated or anergized in the adoptive host has been determined (Rocha et al., 1995). These experiments were performed with donor T cells from Rag-/- transgenic mice, thus avoiding the complication of endogenous, additional T-cell receptors contributing to the survival of such cells (Tanchot and Rocha, 1997). MHC class I-restricted, H-Y-specific T cells are anergized when transferred into host mice that contain the male antigen on every cell in the body. Anergized cells persist for long time periods, provided the host thymus does not generate new T cells. In contrast, they disappear gradually in mice that export T cells into the periphery. This indicates that anergic T cells may have a reduced life span, not necessarily because of intrinsic properties predisposing them to apoptosis, but rather because, like anergic B cells, they lose out in the competition for space in lymphoid organs (Mondino et al., 1996). Is anergy reversible? Anergy induced in T-cell clones in vitro is reversible with IL-2 (Beverley et al., 1992),but this is not usually the case for anergic T cells in vivo (Rocha and von Boehmer, 1991; Lanoue et al., 1997). It is widely believed that a state of T-cell anergy is dependent on the continuous presence of antigen and is reversible upon removal of antigen. This was usually tested by adoptive transfer of previously anergized cells into an antigen-free environment (Rocha et al., 1993; Bachmann et al., 1994; Alferink et al., 1995).This approach can hardly be considered physiological, but an analysis by Tafuri et al. (1995) provided evidence for a transient antigen-specific state of tolerance to paternal alloantigens during pregnancy. Numbers of T cells carrying receptors specific for the paternal alloantigen decreased in thyinectomized pregnant mice and reverted to normal levels after delivery. It was not possible, however, to attribute the changes in specific T cells to a particular mechanism such as anergy, receptor modulation, or deletion because of the large nonspecific fluctuations of spleen cell numbers typical during pregnancy. Ectopic expression of antigen on tissue cells devoid of costimulatory molecules has been used as a means to induce antigen-specific unrespon-
T L Y hl P I 1O( :ITE TO L.EH A NC E
247
siveness. Numerous models of organ-specific TCR transgenic inice were developed whose phenotypes with respect to tolerance induction vary substantially (Adams et al., 1987; Boliine and Pilstrom, 1991; Miller ct al., 1989; Murphy et a/., 1989; Loet c d . , 1991,Cruerderet al., 1994; Hiii-ninerling et nl., 1991).Crosses of mice expressing antigens in a tissue-specific manner with TCR transgenic inice have made it possible to analyze phenotypic and functional effects on specific T cells. Evidence for antigen encounter by peripheral T cells usually is the expression of activation markers such as CD44 or CD69. Graded degrees of peripheral T-cell unresponsiveness, accompanied by T-cell receptor downregulation, were observed in TCR transgenic mic; with tolerance to K” expressed in the liver, glial cells, or slun keratinocytes (Schonrich et (11, 1991, 1992; Ferber et al., 1994). Similarly, the pattern of antigen expression in the thymus and periphery for HEL-specific transgenic T cells resulted in a spectrum of T-cell unresponsiveness, ranging from elimination in the thymus to impaired function of peripheral T cells (Akkaraju Pt al., 1997).In other cases, T cells appeared ignorant of the presence of self-antigen, notably in the case of inyelin basic protein (Goverman et al., 1993; Lafaille ct al., 1994; Liu et d.,199Fj), which might have been due to low avidity of the TCR and/or inefficient presentation of the self-peptide. Apart from the properties of the different T-cell receptors, a likely factor in tlie outcome of antigen recognition by peripheral T cells is the nature and size of the organ expressing self-antigen and the amount of self-antigen present. Th~is,it was suggested that organ vasculature may influence the availability of antigens to the immune system. Antigens expressed on hepatocytes may induce tolerance more easily than those on skin or brain because access-for circulating T cells is easier. In addition, tlie difference in inass of the thyroid coinpared with tlie pancreas was cited as a likely factor for determining the inore pronounced tolerizing effect of the thyroid in this system (Akkaraju et al., 1997). Yet another factor influencing peripheral tolerance induction could be the onset of expression of self-antigens. In a transgenic model of pancreasspecific expression of SV40 T antigen, T cells were tolerant when SV40 T was expressed in the fetal stage. If the antigen was expressed postiiatally only, T cells remained active and caused spontaneoiis autoimmunity (Adams et al., 1987; Geiger ct al., 1992). This might be explained by the different migration patterns of recent thymic emigrants compared with naive, peripheral T cells inentiolied earlier (Section 11,A). It should be stressed, however, that in the case of overexpression of SV40 T or of MHC class I molecules (Miller et al., 1989),the expression per se caused dysplasia and cell death in the pancreas, which would be sufficient to provide a source of self-antigen that could be presented by professional APC such as dendritic cells. Therefore, any potential tolerizing effect of peripheral
248
BHI(;I‘ITA STOCKINGER
tissue cells could have been overwhelmed by the dissemination of selfantigen. Developmentally regulated self-antigens under normal conditions may fail to induce immune responses in the absence of inflammatory stimuli.
D. IMMUNE DEVIATION The term “anergic” suggests functional disability, yet over a number of years anergic T cells have been attributed reactivities that are not compatible with complete inertia. An interesting observation was downregulation of Thl characteristics and a shift to Th2 development in Tho clones treated with “anergizing” stimuli (Gajewski et al., 1994), which suggested that anergized T cells, rather than becoming functionally inert, may instead switch their response potential to a different functional phenotype. A direction of research for which the term “immune deviation” (Asherson and Stone, 1965) was used described a shift from harmful, destructive immune responses to a response mode that did not cause pathology (Rocken and Shevach, 1996).In nonobese diabetic (NOD) mice that spontaneously develop diabetes, the presence of Th1 cells in the islets correlated with disease, whereas the presence of cells producing Th2 cytokines conferred resistance to disease (Liblau et nl., 1995). In a transgenic diabetes model, aniinals with a certain genetic background did not develop autoimmune disease, although their pancreas tissue was infiltrated with activated T cells. This phenotype was attributed to a preferential induction of Th2 rather than Th1 subsets (Scott et al., 1994). Immune deviation is the active mechanism preventing immune responses in the eye (Streilein et al., 1997). In some autoimmune models it was possible to block the induction of disease by the transfer of T cells or thymocytes expressing TI12 cytokines (Powrie and Mason, 1990; Saoudi et al., 1993: Powrie et al., 1994), suggesting immune regulatory properties of the injected cells. In other cases, a regulatory role by Th2 cells was not observed. Thus, it was shown that whereas Th2-like responses correlated with protection from diabetes, these cells could not prevent disease induction by Thl cells (Katz et al., 1995). This illustrates that the view of harmful Th 1 responses being tolerized by a regulatory mechanism involving immune deviation may be an oversimplification.
E. REGULATORY T CELLS After a decade of intensive research into suppressor T cells and suppressor factors, these cells disappeared from center stage following the discovery that the I-J subregion, which supposedly encoded their restriction element (Dorf and Benacerraf, 1985), was not present in the H-2 locus (Kronenberg et nl., 1983). The following interesting statement from a
review on antigen-specific siippressor factors illustrates tlie feelings of a large section of the immunological coininunity: “ F ~ t(Ireas c ofiiizniuriologic. research havc endured srich ctridcnt criticism or oigcnrlered siich faintheartecl ~ y i p o rCIS t the vtiidy of antigeii-Ypw$c \iipprecsion $the i i i i n i i i n c response. . . 1n a very real &iw, tlioce w ~ i pc$)niie(~ o iiiric~ioftlie early work in the j e l d hear responddity Fir the ozitccist statiiy of suppression. With the increciying number of soliiblc; iiier1intor.r c i n d cavcarles qf iiitcr(icting T cells, icliicIi poliidaterl rcciezcc of tlic srihj,jcct in the 198Os, the concept of antigen-s pecijic riipprmsion c i n d sripprewor -factors .siiiiphj hcctimc too complicated and w m disinisscd us artifiict” (O’Hara, 1995). Nevertheless, the notion that immune regulation is the underlying caiise for sonie phenotypes of tolerance sunived. I n contrast to so-called passive mechanisms of tolerance induction, such a s deletion or anergy, immune regulation can be considered an active or doniiiiant form of immunological tolerance. As such, immune regulation offers far wider potential to interfere with tlie autoiminune activation of T cells, which explains the large nuinber of ex-perimental approaches that are uwd to investigate its mechanisms. A number of these approaches ‘ire reviewed in Iiiiinunological Reviews 149 (1996), and will not be discussed in much detail here as the subject area is beyond the scope of this review. One experimental system in wliich the influence of regulatory T cells is invoked is autoimmunity induced by neonatal thymectoiny or generally by the induction of lyinphopenia. This treatment induces a multitude of organspecific autoimmune diseases that can be cured by the transfer of T cells or thyinocytes (reviewed in Gleeson c’t ( I / . , 1996): It is argued that there may be a subset of thymocytes that tire programmed for the control of self-reactive T cells (Saoudi Lit al., 1996: Modigliani et nl , 1996; Le Douarin et al., 199G), but details of their specificity and way of differentiation remain unclear. Irrespective of the problem if and how regulatory T cells are educated in th e t h yin 11s, the re are n u 111e rou s ways for reg11latory in echanisin s invoked in the periphery. Active suppression by regiilatory T cells is a mechanism invoked in oral tolerance (reviewed in M’einer, 1997) and this involves a number of inhibitory cytokines, such as IL-4, IL-10, and TGF-P so that the mode of action is via secreted nonantigen-specific “suppressor factors.” Tolerance to transplantation antigens can be induced if the source of antigen is given at the same time as antibodies to either CD4 or CD8 (Cobbold et d., 1996).This type of tolerancc is “infectioiis,” i.e., it can be transferred by T cells. The primary effector of rejection and the regulatory population both belong to the CD4 T-cell subset, and it appears as if sonie kind of “infectious anergy” induction is underlying the regulation. More recent experimental data may provide a link between cytokine-mediated
250
BRlGlTTA STOCKINGER
suppression and anergy. In addition, they give insight into the problem of the antigen specificity of regulatory T cells, as even in TCR transgenic models where it should be possible to identify T cells with a given antigenic specificity, it often appears that T-cell regulation is found in Rag+ TCR transgenic mice, but not in their Rag-/- counterparts (Forster and Lieberam, 1996: Alferink et al., 1998). This could be taken as evidence that T cells with other specificities than the transgenic receptor generated by endogenous TCR rearrangements are involved in regulation. Alternatively, it could simply indicate that active regulation is difficult to detect in Rag-/- transgenic mice due to the large numbers of monospecific T cells. A Rag-/- TCR transgenic model has been described where T cells specific for hemagglutinin (HA) escape from thymic deletion, but become anergized when crossed with mice expressing HA under control of the immunoglobulin-rc promoter. Anergized HA-specific T cells had impaired ability to cause diabetes following adoptive transfer into mice expressing HA on pancreas islet cells, whereas HA-specific T cells (isolated from mice that did not express endogenous HA) caused diabetes rapidly. Impairment of autoreactive potential was linked with the secretion of large amounts of IL-10 by anergic T cells (Buer et ul., 1998). This cytokine has a longstanding reputation for the downregulation of immune responses and could therefore qualify as a “suppressor factor.” IL-10 was reported to interfere with the costimulatory activity of APC (Ding et al., 1993), the dendritic cell-induced interferon-y production by T cells (Macatonia et al., 1993), or IL-2 production and IL-2R a chain upregulation (de Waal Malefyt et al., 1993; Groux et al., 1996). IL-10 secreting ovalbumin-specific T-cell clones prevented autoimmune inflanirnatory bowel disease when coinjected with the disease-initiating CD4+CD45RBh’subset (Groux et al., 1997). Future experiments, perhaps taking advantage of the fact that antigenspecific T cells can be identified in transgenic systems, should allow clarification of whether particular modes of antigen stimulation, such as stimulation by APC lacking costimulatory molecules or partial triggering of T-cell receptors resulting in altered signaling patterns, can cause a regulatory phenotype within cells of a given specificity. The intriguing possibility is that such regulatory cells may act on other members of the same clone that have not been exposed to the “tolerizing” stimulus themselves and switch their normal response pattern.
ACKNOWLEDGMENTS I thank Liz Simpson, Rose Zamoyska, and Dimitris Kioussis for critical comments on the manuscript. Among their corrections, they all pointed out a few references that were missing. I suspect that the reference list, however long it is, inevitably left out some citations
T LYMHIOCYTE TOLEHANCE
25 1
others might liave considered important. I have cited rebiews wherever applicable so that the reader can obtain original references or further details from those. If there are omissions in the reference list, I can only stress that they have not been intentional. I am grateful to the members of my laboratory for their patience with nie during the final weeks ofwriting.
REFE HENC:ES Adachi, M., Sueniatsu, S., Suda, T., Watanahe, D., Fukuyima, H., Ogasawara, J., Tanaka, T., Yoshida, N., and Nagata, S. (1996). Enhanced and accelerated lyinplioproliferation in Fas-null mice. Proc. Nntl. Acad Sci. USA 93, 2131-2136. Adam, T. E., Alpert, S.. and Hanahan, D. (1987). Non-tolerance and autoantibodies to a transgenic self antigen expressed in pancreatic beta cells. Nottire 325, 223-228. Akkaraju, S., Ho, W. Y., Leong, D., Canaan. K., Davis, M. M., and Goodnow, C . C. (1997). A range of CD4 T cell tolerance: Partial inactivation to organ-specific antigen allows nondestructive thyroiditis or insulitis. Zittrnurtity 7, 255-271. Alberola, Ila, J., Takaki, S., Kerner, J. D., and Perlmutter, R. M. (1997). Differential signaling by lymphocyte antigen receptors. AIIW. Rez;. Z i t i i i t u n d . 15, 125-154. Alberola, Ila, J., Forbush, K. A , , Seger. R., Krebs, E. G.. and Perlmutter, R. M. (1995). Selective requirement for MAP kinase activation in thyinocyte differentiation. Nature 335, 620-623. Albert, M. L., Sauter, B.. and Bhardwaj, N. (1998). Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs N ~ t i r r e392, 86-89. Alferink, J., Schittek, B., Schonrich, G., Hammerling, G. J., and Arnold, B. (1995). Long life spin of tolerant T cells and the role of antigen in maintenance of peripheral tolerance. Int. Zitmu1101 7, 331-336. Alferink, J., Tafuri, A,, Klevenz, A,. Arnold, B., and Hiininierling, G. J. (1998). Tolerance induction in mature T lymphocytes. Nooorti.~Fouitdotiott Syinposiuiti 215, 191- 196. Amakawa, R., Hakem, A,, Kundig, T. M.. Matsnyiina, T., Siniard, J. J., Tiinnis, E., Wakehani, A,, Mittnlecker, H. W., Griesser. II., Takimoto, H., Schniits, R., Shahinian, A,, Ohashi, P., Penninger, J. M., and Mak, T. W. (1996). Impaired negative selection of T cells in Hodgkin’s disease antigen CD30-deficient mice. Cell 84, 551-562. Anisen, D.. and Kruisbeek, A. M. (1996). CD28-I37 interactions function to costiinulate clonal deletion of double-positive thymocytes. Ziit. Z i t t i t ~ u i i d 8, 1927-1936. , E., Lainpkin, G. H., and Meda\yar. P. B. (1951). The use of Anderson, D.. B i l h ~ h a m R. skin grafting to distinguish between nionozygotic and dizygotic twins in cattle. Heredity 5.379-397. Antonia, S. J.. Geiger, T.. Miller, J., and Flavell, K . A . (199.5). Meclianisrns of iiniriune tolerance induction through the thymic exprcssion of a peripheral tissue-specific protein. znt. ZIlt?tiU?tO~.7, -715-725. Asherson. G. L., and Stone, S. I I . (1965). Selective and specific inhibition of 24 hour skin reactions i n the guinea-pig. Immune deviation: Description of the phenomenon and the effect of splenectoiny. Ziti~ttuirology9, 205-217. Auphan, N., Schonricli, G., Malissen, M., Barad, M . , Hlinmerling, G. J., Arnold, B., Malissen, B., and Schmitt-verliulst, A.-M. (1991j . Influence of antigen density on the degree of clonal deletion in T cell receptor transgenic mice. I i t t . Zntitirtnol. 4, 541. Bachmann, M . F.. Rohrer, U . H., Steinhoff, U., Burh. K., Skuntz. S., Arnheiter, H., Hengartner, H.. and Zinkernagel, R. M. (1994). T helper cell unresponsiveness: Kapid induction in ;~ntigen-transgenicand reversion in non-transgenic mice. Eur. /. Zmttunol. 24, 2966-2973. Barthlott, T., Kuliler, H., Pirclier, H., and Eichmann, K. (1997). Differentiation of CD4(high) CD8(low) coreceptor-skewed thymocytes into mature CD8 single-positive cells independent of MHC: class I rpcognition. Eur. /. Iitirttunol. 27, 2024-2032.
252
13HIC:I'ITA STOCKINllsby superantigens. However, antigen ~)iiidin~recognition occurs at a site that is probably conserved in all Vy2 chains- and close to the highly variiible CDR3 region of the TCR responsible for fine antigen specificity (Rocket nl., 1994; Schild et nl., 1994). Because of this oligoclonal stimulation, the frequency of myco1)acteiiastimulated y6 T cells is at least in the same or even higher order of magnitude ;is tlie frequency of antigen-specific c.p T cells stiiiiulated clonally by bacteria. Evidence shows that growth factors such as IL-2 and IL- 15 produced by activated CD4+ T cells are required for the proliferation of human 76 T cells. Thus, although stimulation of y6 T cells with phosplioligands induces tlie activation markers CD2Fj and CD69, they fail to proliferate (Wesch et nl., 1997). This is further corroborated by findings that Vy2 T cells froin HIV-infected individuals do not respond to mycobacterial antigens due to their deficiency of antigen-specific CD4+ Thl cells ( W t d et nl., 1996). Importantly, because IL-10 production strongly inhibits this response, this mechanism of controlling y6 T-cell proliferation can only occur following the generation of TI11 cytokines (Pechholtl et nl., 1994). Siinilar to human y6 T cells, murine y6 T cells, which do not recognize phospholigands but peptides, expand to an oligoclonal population of y6 T cells expressing the Vyl chain 011 stirnulation hth HSP6O (O’Brien rt (11.. 1989; Fu ~t a / , , 1994; Belles et ( I / . , 1996). Data from many groups have proposcd a role for y6 T cells in inurine listeriosis, a s these cells accuinulate and secrete IFNy in response to Listerici infection (Hiromatsu ct d . ,1992; Skeen and Ziegler, 1993). It has been suggested that this activation of y6
314
U L I W H E. SCHAIRLE cf
(I[
T cells requires the iiiacropliajie-produced cytokines IL-12 and IL-1 (Skeen and Ziegler, 1995) and is also dependent on CD45 (Fujise et al., 1997). Using a comparison between 76 and a@TCR KO mice, the authors’ group found that either a@ or y6 T cells were sufficient for early protection against Listeria infection. However, treatment of (YOTCR KO mice with anti-yS inAb exacerbated the disease, suggesting that compensation inechanisms occur between these two T-cell populations (Moinbaerts et al., 1993). This is also illustrated in mycobacterial infections. Infection of mice iv with low doses of M. tuberciilosis revealed increased bacterial growth in 6 TCR gene disrupted inice between diiys 15 and 30, although by day 120 postinfection both mutant and wild-type mice were able to adequately contain the infection. In contrast, higher doses that were not lethal in control mice rapidly proved fatal for the y6 T-cell deficient KO mice (Ladel et d ,1995a). However, D’Souza and colleagues (1997), using a range of doses of M . tuberculosis of high virulence administered by aerosol, found equivalent survival rates between wild-type and KO mice. A possible scenario to explain the apparent differences in these results is that in the high-dose iv infection of 6 TCR KO mice, the compensatory activation of ap T cells occurs too late to control this overwhelming infection and consequently the mice die. In the case of the lower iv infection and the aerosol infection, the organism do not overwhelm the animals prior to the activation of ap T cells, which then control the infection. During the course of these experiments, it was revealed that the 6 TCR knockout mice exhibited an increased influx of neutrophils into the mycobacterial granuloina in contrast to the priinarily lymphocytic infiltrate in wild-type inice (D’Souza et al., 1997), which may also account for the death of animals when higher doses are administered iv. Based on this and the observation that Listeria-infected S TCR KO mice develop liver abscesses rather than granulomas (Mombaerts et al., 1993), it has been proposed that yST cells primarily play a regulatory role in infections with intracellular bacteria by limiting the inflaiiimatory response that leads to tissue damage. Supporting this notion is the observation that splenic y6 T cells produce IL-10 during the course of a Listeria infection that coincides with maximum IFNy production and a decrease in inflammation and tissue damage, suggesting an important immunoregulatoiy role for these cells in controlling Th1 cell responses (Hsieh et al.. 1996). Consistent with their regulatory function, yS T cells have been implicated in the control of IFNy production by N K cells in response to Listeria infection (Ladel et al., 1996).
2. Cells Controlled by Group I CDl Molecules As discussed in Section V, CD1 molecules are a group of nonpolymorphic MHC-reIated polypeptides with the unique ability to present glycolipid
antigens to unconventional T cells. T cells that respond to CD 1 presented glycolipids express the arj3 TCH and are either CD4- CD8 or CD8' (reviewed in Porcelli et (11, 1996) and are capable of antigen-specific proliferation and IFNy secretion. Furthermore, these cells are c'ipable of lysing antigen-presenting cells infected with live niycobacteria but not uninfected CD1' cells (Stenger et nl., 1997).Although both double negative CD4-CD8 and CD8' CD1-restricted cells lysed infected APCs, they did so via different niechanistns. DN cells affected cell death via the Fas-FasL interaction, whereas CD8' cells einployed perforin- and granzymecontaining cytotoxic granules to mediate tlie lytic effect (Stenger et nl., 1997). Importantly, these elegant experiments also revealed that only lysis of infected macrophages by CD8' CD1-dependent T cells reduced the viability of the intracellular mycobacteria. Thus tliese experiments suggest that T cells reactive with microl)ial glycolipids have ;I dual role in the response to an invading pathogen: an initnunoreplatory role inducing the elimination of infected APC via apoptosis and the granule-dependent killing of bacteria.
3. Group 2 CD1-Coiitrolled T Ccll~-NKl T Cells The murine group I1 CD1 molecules control the development of N K l ' a@ T cells (Bendelac et nl , 1997).Tlie absolute requirement for CD1 in the development of these cells has been confirmed following the generation of CD1.l gene disruption mice that are devoid of NK1 T cells (Chen et nl., 1997; Mendiratta et al., 199'7) These specialized T cells develop mainly in the thymus and display a defined tissue distribution, accounting for the inajority of liver T cells, 20-30% of bone marrow T cells, 10-20% of mature thymocytes, and 0.5-1% of splenocytes, and are also found in the intestine within the lamina propria and the Peyers patches (Vicari and Zlotnik, 1996). Interestingly, the recognition of CD1.l by T cells is highly dependent on tlie cell type in which this molecule is expressed (Park c>t al., 1998) Despite the control by an MHC class I-related molecule, NK1 T cells are not CD8', but rather CD4' or DN. Many experiments have been performed to elucidate a function for these cells in tlie regulation of the immune response in general and specifically whether these cells play a role in tlie host response to infections. What is now undisputed is that these cells rapidly produce IL-4 when stimulated in viuo with anti-CD3 ~ of producing t)otll (Yosliiinoto; i d Paul, 1994)and that they are a l capable IL-4 and IFNy in tjitro when stimulated with anti-CD3 or CD1 (Arase et al., 1993; Chen and Paul, 199'7). It was originally proposed that these T cells produced the IL-4 required to promote the development of coiiventional CD4' TI12 cells (Yosliiinotoet nl., 199,5).However it is now established that nuce devoid of these cells sucli as & i n KO and CD1 KO mice
316
ULRICH E . SCH.41BLE et nl.
are capable of mounting functional Th2 responses, despite the diminished early IL-4 burst (Brown et nl., 1996; Smiley el al., 1997). I n the context of intracellular bacterial infections, experiments using L. iwnocqtogenes and M . bovis BCG have revealed that NK1 T cells in the target organ for Listeria follobing iv infection) rapidly lose this liver T~-il-secretingproperty following infection, and this modulation is mediated via IL-12 (Emoto et al., 1995). Infection of mice with Propionibncter i r i i n acnes confirmed these results and additionally implicated IL-18 in the effect (Matsui et al., 1997). Furthermore, when beige mice were infected with S. entericn serovar Cholerasuis, a suppression of conventional NK cell function was observed and a concomitant expansion of TL-4producing NK1 T cells in the peritoneal cavity of the mutant mice was noted (Enomoto et al., 1997). These mice are unable to mount a successful TI11 response and are consequently more susceptible to the infection, but these effects can be abrogated by the administration of a neutralizing antiIL-4 mAb. Together, these results suggest that although other sources of IL-4 may induce Th2 function, early IL-4 production by NK1 T cells as it default consequence of microbial infection can modulate the outcome of the specific immune response. More recent experinients suggest that during infection NK1 T cells acquire IFNy-secreting properties and lose their ability to produce IL-4. This shift in cytokine production is paralleled by alterations in the density of the cell (Emoto et al., submitted). Thus NK1 T cells may support, rather than counteract, protective immunity to intracellular bacteria.
(a
4. M H C Class lb-Restricted CD8+ T Cells
Experimental infections of mice with listeriae have revealed a subset of CD8' CTL that are restricted by the nonclassical MHC class Ib gene product, H2-M3, and these T cells form a significant component of the total MHC class I-restricted T-cell response (Bouwer et al., 1997; Lenz and Bevan, 1997). As described in Section V, these CD8+T cells recognize N-f-met-containing peptides (Pamer et al., 1992).What role these T cells actually play in protection against intracellular bacteria remains to be seen, although cloned non-M HC-restricted T cells specific for L. rnonocytogenes are protective in a inurine model of infection (Kaufmann et al., 1988). Furthermore, evidence showing that these cells are present and responsive in conventionally housed, nonspecific pathogen-free mice suggests either that the epitopes that are recognized are cross reactive or that priming requirements are not as stringent for these nonclassical MHC molecules (Lenz and Bevan, 1997). Together with the fact that fonnylated proteins are uncommon in the mammalian host and that the CD8+ T-cell responses generated occur across a variety of MHC haplotypes (Gulden et al., 1996),
it has been proposed that stimulation of these cells could form part of a vaccination stratea. It should be noted, however, that no homologous T cells have been identified in humans. C. COSTIMUIATION So far this review has considered the interaction of the MHC-antigen complex arid the appropriate cytokine environment as being critical in activating the specific T-cell response to intracellular pathogens. In addition, cognate receptor interactions such as CD40KD40L and B7/CD28/ CTLA4 are important to ensure that the correct T-cell response is generated. These molecules provide a second signal following engagement of the TCR in the two-step activation process necessary to activate naive T cells, with the failure to receive this second signal resulting in anergy (Jenkins, 1994). CD40L (CD154) is preferentially expressed on T cells, and KO inice with a CDlS4 gene disruption are severely impaired in their primary Tcell responses to protein antigens, as well as in their ability to activate macrophages to produce a variety of products such as TNFa, IL-1, and NO involved in antimicrobial defense (Grewal et nl., 1995). Importantly, the interaction between CD40 and CD154 during antigen presentation to T cells results in tlie secretion of IL-12, which is critical for the generat’1011 of a Th1 response (Grewal and Flavell, 1996). Experimental infection of KO inice with the intracellular protozoan L. mujor revealed that mice deficient either in CD40 or its ligand develop a more severe form of the disease, which correlated with the inability of the macrophages from these mice to produce NO and TNF and to respond to IFNy (Campbell et nl., 1996; Kamanaka ct nl., 1996). I n contrast, preliminary infection studies with L. monocytogenes revealed equivalent growth of this organism in both wild-type arid CD 154-deficient mice witliiii 48 hr and no difference in tlie clearance of a sublethal challenge dose. These findings suggest that both the early T-cell independent control mechanisms and the development of T-cell-mediated immunity to listeriae develop norinally in CD 154-deficient mice (Grewal et al., 1997). One explanation for the discrepancy in these results is that listeriae and other intracellular bacteria are inore potent IL12 inducers, via coinponents such as LTA, and therefore CD40KD154 interactions are less critical. Consistent with this, experiments using CD1-54 KO mice infected with M . ttiberculosis revealed no difference in either survival times or bacterial loads in KO mice as coinpared to CS7BLJ6 wild-type mice (Campos-Net0 cf (11, 1998). As M . triberculosis can exert direct effects on rnacrophages, inducing IL-12, TNF, and NO production (see Section III), this points to an activation of macrophages in a T-
cell-independent manner, thus bypassing the requirement for the CD40/ CD154 interaction. The interaction of CD28/CTLA4 (CDl52) with B7-1 (CD80) or B7-2 (CD86) on APC is a necessary requirement for T cells to produce the autocrine growth factor, IL-2. However, those T cells that utilize IL-4 as a growtli factor appeared not to depend on costirnulation, which has led to the hypothesis that Thl and Th2 cell responses require differential second signals, i.e., the T-cell differentiation toward IL-4-producing Th2 cells is less dependent on B7 than the induction of IFNy-secreting Th1 cells (reviewed in Gause ct al., 1997). Further indication of this differential activation requirement was provided using CD28-deficient mice that show reduced T helper cell activity and decreased Ig class switching (Shahinian et al., 1993). In this case, however, it appeared to be the IL-4-producing cells exhibiting a higher dependency on the presence of CD28. The current hypothesis remains that signaling via members of the B7 family is required to maintain T-cell effector functions whereas the CD40/CD154 interactions appear to primarily play a role in T-cell induction. With the interactions between these costimulatory molecules being critical in initiating and maintaining an effective T-cell response, any modulation of their expression by an intracellular pathogen may provide a means of subverting the host iininune response and enhancing its chances of survival. To date, most data on the modulation of costimulatory molecules by pathogens have been provided by Leishnlania, which induce macrophages to downregulate surface expression of B7-1 following infection both in oitro and in vivo (Kaye et al., 1994; Kaye, 1995). More recent reports have also demonstrated a leishmania1 protein, LeIF, which upregulates B7 expression on inonocytes and macrophages (Probst et al., 1997). A similar situation is seen following infection of macrophages in oitro with live S. t y p l i i n i i i r i t m , where surhce expression of B-7 was downregulated following infection with live bacteria, but was increased if the macrophages were pulsed with dead bacteria (Gupta et at , 1996).Heat-killed B. abortus, which is currently being considered as a vaccine candidate, induces a strong Tlil cell response, partially mediated by its ability to upregulate the expression of B7-1 and B7-2 on liuman inonocytes (Zaitseva ct nl., 1996).Together, these data suggest that high levels of costimulatory molecules are necessary for the initiation of successful Th1 responses, which lead to macrophage activation. D. REMJLATIONOF T H E IMMUNE RESPONSEHY CYTOKlNE5 1 Mncrophnge Activatioii A critical step in the resolution of infection by an intracellular bacterium is the elimination of the organism by the antibacterial properties of the
IKTHACELLIJIAR HACTEHIA A N D THE I M M U N E SYSTEM
319
activated macrophage. The central cytokine required to initiate these events is IFNy, which can synergize with macropliage-produced TNFa to mediate many niicrobicidal mechanisms, such as ROI and NO production discussed in Section I11 (reviewed in Unanue, 1997; MacMicking et al., 1997). The absolute requirement for IFNy in resolving intracellular infections was first deinonstrated by antibody-mediated neutralization of this cytokine in vivo and was subsequently confirmed using KO mice with a deletion of either the lFNy receptor a chain or the IFNy gene itself. All of these inice were highly susceptible to infections with listeriae, mycobacteria, salmonellae, and yersiniae (Buchmeier and Schreiber, 1985; Cooper et ul., 1993; Flynri et a l , 1993; Autenrieth et al., 1994; Hess et al., 199613). Similarly, patients with a inlitation in the receptor for IFNy suffer from an increased susceptibility to mycobacterial infections due to a defect in their ability to activate rnacrophages in response to this cytokme ( Jouanguy et aE., 1996; Newport et ul., 1996). In addition to the requirement for IFNy in the induction of NO, macrophage activation by this cytokme can modulate other cellular functions. Following infection of macrophages, inycobactena reside in intracellular compartments that fail to acidify due to the paucity of the v-H'ATPase (Sturgill-Koszyckiet al., 1994).Activation of infected macrophages with IFNy leads to the accumulation of the vH' ATPase within the inycobacterial phagosome, its subsequent acidification, and segregation froin iron delivery via Tf (Schaible et al., 1998). As a consequence, the antitnycobacterial capacity of IFN 7-activated macrophage5 is enhanced greatly ( F l e d and Kaufniann, 1988; Chan et al., 1992; Schaible ct a l , 1998). In contrast, in hutnans, IFNy fails to consistently induce the antiniycobacterial effects of macrophages in vitro. However, the addition of 1,25-dihydroxyvitaniin D1, the biological active metabolite of vitainin ]I1,maximizes the tiiberculostatic potential of the activated 1987). Moreover, an alternamacrophage (Crowle et al., 1987; Rook et d., tive niicrobicidal mechanism ha5 been described that occurs in the absence of IFNy in which exogenous ATP mediates lysis of M . bovis BCG-infected human macrophages via P2Z receptors, resulting in death of the mycobacteria (Lainmas ct al., 1997). 2 Gr[mulonui Formation A granuloim is an organized lesion formed by infiltrating T cells and inononuclear cells at the site of infection. Forination of granulomas is a characteristic feature of inany infections caused by intracellular bacteria and they are often critical in restricting bacterial replication and confining pathogens to discrete foci. In listeriosis, for example, sterilizing immunity can be accompanied by granuloma formation, although the granulomas formed are often incomplete and are not essential for the eradication of
listeriae (Mielke et nl., 1989). In contrast, a successful imrnune response against mycobacteria is generally accompanied by tuberculoid granulomas. In both cases, similar mechanisms of cell recruitment occur, which are highly orchestrated events regulated largely by cytokines produced by T cells. Infection models with listeriae, brucellae, M . bovis BCG, and M . aviurn revealed that CD4+ T cells are critical for granuloma formation (Mielke, 1991; Lade1 et al., 199Sb; Hgnsch et al., 1996). Furthermore, the use of y6 T-cell KO inice suggests that these cells also exert considerable influence on granuloma formation. A deficiency in this T-cell subset resulted in abscess formation rather than the organized lesions generally observed in listerial infections (Mombaerts et al., 1993) and neutrophildominated inflammation following aerosol infection with M . tuberctilosis (D'Souza et al., 1997). However, in both cases antibacterial resistance was virtually unaffected by the altered tissue response. Early studies by Kindler and colleagues (1989) demonstrated the absolute requirement for TNF in the development of bactericidal granulomas in M . boois BCG infection, which has subsequently been confirmed using p55 TNF receptor KO mice or the soluble TNF receptor (Senaldi et al., 1996). However, KO mice infected with M . tuberculosis revealed that the number of granulomas formed was equivalent to that observed in wildtype mice, yet necrosis was only noted in TNF-competent mice, suggesting that this cytokine plays a role in immunopathology as well as bacterial containment (Flynn et al., 1995). It is now appreciated that other cytokines, particularly IFNy (Mielke et al., 1992) and IL-10, also play a critical role in granuloma formation (Wynn et d . , 1997). Data from a number of laboratories have revealed that the most important step in this process is the activation of nonspecifically immigrating CD4' T cells to produce IFNy, TNF, and IL-2, leading to the activation of macrophages and, together with the appropriate chemokine production (e.g., MCP-l), their attraction to the infection site (Mielke et al., 1997). Extravasation is then mediated via specific adhesion molecules (e.g., LFA-1, ICAM-l), which are upregulated during this process. However, although ICAM-1 KO mice failed to form granulomas following aerosol infection with M. tuberculosis, this inability did not affect the outcome of infection (Johnson et al., 1998). Consistent with this are observations that both ICAM-1 and P-selectin KO mice also control infections with L. monocytogenes as well as M . bovis BCG with equal efficiency to wild-type mice (Steinhoff et al., 1998).Taken together, these results indicate that although granuloma formation may be the desired result of protective immunity, especially against mycobacterial infections, the failure of this process need not adversely influence the outcome of infection in every case.
I N T I ~ A ( ~ ~ I . I . l I L . BACTERIA 41~ .AN11 TI{E IhlMliNE SYSTEM
32 1
response^ As inentioiicd previously, intracellular bacteria do not ally induce a Th2 response followinginfection. However, concomitant with tlie induction of a Thl response to eradicate the liactenn is the risk of inflammatory mediated tissue damage following an uncontrolled iinmune response. Consequently, counterregiilatory cytokines are necessaiy to downregulate the production of inflammatory cytokines such :is IFNy and TNF. One of the most important cytokines in this regulatory process is IL-10, which is produced hy a variety of cell Q p s , including Th2 cells, B cells, and macrophages (Moore ct a / . , 1993). Its major role s e e m to be to inhibit the prodiiction of IL-12 (D'Andrea ct d . , 1993) and to antagonize the effects of IFNy on macrophage activation. For instance, IL-10 induces internalization of MHC class I1 molecules from the surface of APC (Koppelnian et ml., 1997). Thus its regulation plays a critical role in infections where macrophages are the host cells for pathogens. This has been confirmed in experiments where transgenic mice that overproduce IL-10 failed to clear invcobacterial infections, despite the abundmt production of IFNy by T cells, with IL- 10 overriding the antimycobacteiial signals received by the macrophages (Murray r t d., 1997). The iniportance of the regulatory role of IL-10 is emphasized by experiments in which IL-10 KO mice infected with T g m d i i succumb to enhanced liver pathology and necrosis due to incrcwsed production of both IL-12 and IFNy (Gazzinelli et d . , 1996). Interestinglv, it has been shown that IL-12 can induce T cells to produce IL-10, altliough not as efficiently as it mediates IFNy production. Thus the possibility exists that IL-12 can regulate its own production via IL-10 (Meyaard et ol., 1996) and therefore provide a rnechanisrn to downregulate tissue destruction following pathogen-induced iininune responses. 3. Copitrol of Itzapproprioto
4 Efcct
oti
Init,iiine
Lytti),hn),oei~i~
In addition to their biological activities on effector cells of tlie immune system, such as T cells, macrophages, and N K cells, several cytokines including IL-12 and IL-4 can, in combination with other factors, enhance the surviviil and growth of earl?, lieinatopoietic stem cells (reviewed in Chehinii and Trinchieri, 1994; Sonoda, 1994). IL-12 appears to have a dual role in hematopoiesis, in that it enhances the proliferation of inyeloid and B-cell precursors and yet, if NK cells are present, can also inhibit hematopoletic colony forination as a result of the IL-12-induced production of IFNy and TNFa (Cheliimi and Trinchieri, 1994). Sindarly, IL-4 also exhibits diverse effects on hematopoiesis as it can act on committed as well as earl>,progenitors enhancing granulopoiesis but inhibiting mono-
322
ULRICII E. SCHAIBLE et 01.
poiesis from CD34+ bone marrow precursor cells (Snoeck et al., 1996). To date there is little information on what effect, if any, infectious agents may have on these processes. However, the authors have described for the first time the rapid induction of IL-4, but not IL-12, in murine bone marrow cells by M. tuberculosis, M. bovis BCG, and one of their major cell wall components, mannose-capped LAM (Collins et al., 1998). Experiments are currently underway to identify the mechanism of this induction and its possible implications on hematopoiesis and the outcome of infection. Overexpression of IL-4 within the bone marrow in transgenic mouse models has revealed abnormal T-cell development (Tepper et al., 1990) and also an increase in eosinophils with enhanced phagocytic capacity (Sullivan et al., 1992).This suggests that alterations in cytokine concentrations within the bone marrow induced by microbial components may influence the growth and maturation o f cells involved in host defense. VIII. Host Genetics Influencing the Outcome of Infection
The impact of the genetic makeup ofthe host has been well illustrated by a study on an epidemic o f tuberculosis in an isolated population of Yanomami Indians of the Amazon rain forest. This population had no contact with tuberculosis prior to the early 1960s. This study showed an extraordinarily high incidence of active disease, correlating with elevated levels of M. tuberculosis-specificIgG4 antibodies and almost no skin test responses, even among BCG-vaccinated individuals (Sousa et al., 1997a). This emphasizes the importance of host genetic composition on resistance to tuberculosis and, by implication, the evolutionary pressure that microbial pathogens exert on the selection of the human race. It maybe speculated that such high susceptibility reflects the genetic situation at the beginning of a newly introduced pathogen causing epidemic disease as exemplified by tuberculosis in 18th century Europe and more recently by the worldwide spread of HIV.
A. NRAMP For many years it has been appreciated that resistance against intracellular pathogens is, in part, genetically controlled and that inherited factors can modulate the course of disease in infections with a broad clinical spectrum such as leprosy and tuberculosis. In particular these genetic traits can control the progression from infection to disease, which divides individuals into resistant and susceptible groups. Studies in the murine system have revealed a single, dominant autosomal gene on chromosome 1, originally termed Zty/Lsh/Bcg, which controls innate resistance to the unrelated intracellular pathogens, L. major, M . bovis BCG, and S. typhimurium (Blackwell et al., 1994; Zwilling and Hilburger, 1994). Following
IKTHACELLULAH BACTERIA AND TllE IMMIJNE SYSTEM
323
positional cloning and gene targeting, the Nrampl (natural resistanceassociated macrophage protein) gene was identified and shown to be allelic to Bcg (Vidal et al., 1995). This gene belongs to a family of genes that also includes Nramp2 located on mouse chromosome 15. Nrainpl encodes for a hydrophobic, membrane-associated protein expressed exclusively in phagocytic cells. In congenic mouse strains, a single glycine to aspartic acid substitution at position 169 of the fourth transmembrane domain is associated with the Bcg-susceptible (Bcg’) phenotype (Vidal et al., 1996). Conflicting reports exist as to what functional effects this mutation changes, with original data claiming that the amino acid substitution resulted in decreased expression of the molecule (Vidal et al., 1996),whereas more recent data suggest that there is equivalent expression in bone marrow macrophages from BALBlc mice (Bcg’) and CBA mice (BCG‘),indicating that the protein is present but nonfunctional (Atkinson et al., 1997). The exact function of the Nramp1 protein is as yet not fully elucidated, but the use of gene deletion mouse mutants and transfected cell lines have conclusively illustrated the pleiotropic effects on macrophage function that this gene controls. These include the production of ROI, NO, and TNFa from macrophages as well as the regulation of MHC class I1 expression and antigen presentation (Arias et al., 1997; Lang et al., 1997). Using an antibody generated against the C-terminal35 amino acids of murine Nrampl, confocal microscopy has localized this protein to late endosomesAysosornesas well as latex bead phagosomes, which would place the protein in close proximity to an intracellular pathogen (Atkinson et al., 1997; Gruenheid et al., 1997). Moreover, phagosomes containing live M . bovis BCG within macrophages expressing functional Nrampl fuse with V-H+ATPasecarrylng vesicles and subsequently acidify, in contrast to their mutant counterparts (Hackam et al., 1998). Furthermore, a high degree of sequence similarity to the yeast protein SMF-1 has indicated that Nrampl may function as a divalent cation transporter (Blackwell et nl., 1995). Mice with microcytic anemia have been shown to have a mutation in the closely related Nrainp2 gene, which is considered to encode for an iron transporter (Fleming et al., 1997). As many bacteria have an obligate requirement for iron for their intracellular survival (Wooldridge and Williams, 1993), this provides an explanation for an essential role of Nrainpl in resistance and opens up an area for potential therapeutic intervention. The human equivalents NRAA4Pl and NRAMP2 have been identified on chromosomes 2q and 12q, respectively. The gene for human NRAMPl has been cloned and sequenced, and polymorphisms and sequence variants have been described and used in linkage studies for susceptibility to tuberculosis and leprosy (Liu et al., 1995). Early evidence suggested that there was n o evidence for an LshlltylBcg gene hoinologue influencing suscepti-
bility to leprosy (Shaw et cd., 1993). However, a more recent large case control study from The Gambia, West Africa, comparing more than 400 tuberculosis patients with healthy controls, revealed that individuals heterozygous for two N R A M P l alleles were overrepresented among the tuberculosis cases (Belhny et al., 1998). In the murine system, data by Medina and North (1996, 1998; Medina et al., 1996), studying a number of inbred and recombinant mouse strains infected with virulent M. tuberculosis both iv and by aerosol, indicate that Nrampl polymorphisms have no influence on the outcome of infection. In this study, the fact that mouse strains homozygous for the Nrampl susceptibility allele were more resistant than those expressing resistant alleles seems more than coincidental, although as yet this finding cannot be explained. At least it can be concluded that genes other than Nranipl strongly influence susceptibility to experimental tuberculosis infection in mice (Medina and North, 1998). This is in line with data demonstrating that resistance to M . tuberculosis in mice only becomes evident after the onset of specific immunity, i.e., in the lung 30 days postinfection (Nadeau et al., 1995).
B. MHC Segregation analyses in human populations have indicated HLA linkage with various disease states. In terms of intracellular bacteria, most studies have analyzed linkage in leprosy and tuberculosis. Blackwell and colleagues ( 1997) analyzed the irnmunogenetics of both leishmanial and mycobacterial infections. This family study revealed that there was linkage to and allelic association of HLA molecules with leprosy, but not tuberculosis, which is consistent with what was found in a study population in India (Ghosal et nl., 1996). However, a population in Cambodia showed association of an HLA-DQ allele with clinical tuberculosis (Goldfield et al., 1998). Similarly, Rajalingam and colleagues ( 1997) identified an association of HLA-DR2 with mycobacterial disease. It was suggested that these specialized HLA alleles may preferentially present inycobacterially derived peptides. Supporting this notion is the observation that DR17 can only bind its specific peptide at neutral pH rather than at the acidic pH found in most peptide loading compartments (Geluk et al., 1997), which is the pH that the mycobacterial pliagosome maintains (Sturgill-Koszyclaet al., 1994).Studies on experimental tuberculosis in mice provided contradictory results. One study failed to show strong correlation with certain MHC genes, although some minor influence was observed (Medina and North, 1998), whereas in a different set of experiments, H-2k associated with resistance and H2” and H-2“ with susceptibility to tuberculosis (Apt et al., 1993). Although most linkage analyses have been performed for infections with a broad spectrum of clinical disease, there are reports of HLA segregation
in trachoma caused by C tracliowztis. In a study population in Oman, HLA-DR16 was associated with blinding trachoma whereas HLA-DR53 correlated uitli resistance (White ct d . ,1997).With more data now becoining available froin worldwide studies it is clear that HLA associations differ between population groups, making it difficult, or even impossible, to define a universal association for resistance and susceptibility, and underlining the influence of natural selection under the pressure of infection on M HC polymorphism. Within the MHC locus there are genes encoding for TNF and its receptors as well as genes for molecules of the antigen-processing machinery (proteasome subunits and TAP1,2), and these genes have also heen examined for polvmorphisms affecting disease states. In addition to HLA polyinorphisms,-there are indications that the polyinorphism of TAP molecules in rats modifies the spectrum of antigenic peptides presented by MHC class Ia m o l t d e s (Powis, 1997). Although inurine and human TAP molecules show only limited polyinoi-pliisni (Lobigs and Mullbacher, 1993; Yewdell cf a/ , 1993),correlation of certain TAP alleles with disease severity has been claimed in human tuberculosis and leprosy. The frequency of TAPZ-A/F’was increased in patients with pulmonary tuberculosis and that of TAP2-B9 in patients with tuberculoid leprosy (Rajalingam et nl., 1997). Despite the reported associations with M HC, an analysis in The Gambia of the relative contribution of MHC and non-MHC genes in the response to foreign antigens, including those derived from M . tzihcwulosis, indicated that genes outside the MHC lociis were more influential (Jepson et nl , 1997). In a different study, several cases of familial disseminated M . aviuin infections have been explained by a defect in antigen processing that diniinishes IFNy production. APCs from these patients can present influenza hemagglutinin peptides to T cells but are unable to process the native protein for presentation (D’Souza ct 01 , 1996).
R C. O T ~ I EGENES Resistance to sonie intracellular pathogens in the inurine models of infection are not influenced by tither N r m i p or MHC genes. A/J inice are an inbred strain that are niore susceptible to infections with both listeiiae and legionellae. A defect in the phagocyte inflammatory response caused by a deficiency in the C5 complement component was shown to be the major reason for the high susceptibility of these mice to infection with listeriae. Congeiiic inice sufficient for C5 were restored in their ability to control infection (Gervais et d., 1989). The natural resistance or susceptibility of inbred strains of mice to infections with L. pneumoplzitn is controlled by a single, dominant gene on chromosome 13, designated Lgii1. The phenotypic result of expression of this gene is the presence or
326
ULRICH E. SCHAIBLE rt al.
absence of intracellular replication of the bacteria inside host macrophages (Dietrich et al., 1995). Currently, a high resolution linkage map is being prepared of this region to accurately pinpoint and eventually clone the gene (Beckers et al., 1997). In an elegant set of experiments conducted by Murphy and colleagues, another locus has been identified that may influence host resistance to intracellular infections by controlling the T-cell response that is made. The genetic background of mice is an important determinant of IL-12 responsiveness and can influence the onset and severity of disease. CD4' T cells from a DO1l.10 ap TCR transgenic mouse were transferred into either BALB/c or B10.D2 mice, which both made equivalent levels of IL12. Analysis of the T-cell populations revealed that those transferred into the BALB/c mouse made increased IL-4 and decreased IFN-y, whereas the converse was true for the B10.D2 strain (Gorham et al., 1996). As both mice are congenic at the H-2 locus (H-2Dd),non-H-2 genetic loci must be controlling this effect. It has been shown more recently that the downregulation of the IL-12 receptor P chain in the BALB/c mouse but not the B10.D2 mouse strain is involved in this phenomenon (Gorham et al., 1997). A single dominant gene locus termed Tpml (T-cell phenotype modifier) controlling IL-12 responsiveness has been mapped on mouse chromosome 11 and on human chromosome 5 (Gorham et al., 1996). This region of the genome contains a dense cluster of genes of immunological importance, including IL-4, IL-5, and IL-13, as well as T-cell signalling molecules, and currently the molecular identity of Tpml remains to be determined. Consistent with the location of a resistance locus on chromosome 11, serial backcrosses of resistant B10.D2 mice onto the susceptible BALB/c background reveaIed several candidate loci, conferring resistance to L. major infection. However, despite one of these being located on chromosome 11, no single locus was required and the authors postulate that a variety of combinations of these loci may interact, resulting in resistance (Beebe et al., 1997). IX. Immune Intervention Strategies
Detailed knowledge about the biology of infections with intracellular bacteria and the immune response elicited not only allows us to dissect the complex cross talk between pathogen and host, it is also of basic importance for developing rational preventive and therapeutic intervention strategies. Although vaccination is the most successful prophylactic measure against infectious diseases in general, most vaccines currently in use are directed against viruses or extracellular bacteria. The success of these vaccines in principle rests on antibodies that either prevent pathogen
invasion o r neutralize toxins or virulence factors. To date, only hvo vaccines arc in use against intrncellular Iiwcteria-in both cases with only liinited success. The tuberculosis vaccine BCG is an ntteniiated strain derived from tlie virulent M . Iiouis strain, the agent of tuberculosis in cattle and occasionally in human Since its first iisc in 1921. BCC has been administered with lo\v side e ects :3 billion times \vorltl\vide, hut its protective efficacy is highly variable, ranging from 0 to 80% against puliiionaiy tiiberculosis (Ginsberg, 1998). Generally, the protecti6n achieved by BCG is 1iettc.r against niiliary and meningeal tuberculosis in children (46- loo%), but is inefficient against adult tuberculosis, which is usually caused by the reactivation of dormant hi. trihcrm1o.vi.s(Coltlitz rt d . , 1994; Fine, 1995). Increasing incidences of tulwrculosis in developing coiintries and a recent upswing in case numbers in the United States and eastern Europe since the mici-198Os are mainly caused Iiv tIrr HI\’ epidemic, by insufficient compliance wit11 tIie long duration of cIi(.inotlierapy, ancl by globa~uriiariization and pauperization. Moreover, Ad. triherculosis strains liave emerged that are resistant against most of thc cIieiiiotlierapeutic drugs c u r r e d y iwdal)le. This serious 1ie:iltli situation mikes the tlevelopment of an effective vaccine against tuberculosis a high priorit), in adjunct to iinprovetl chemotherapy. The other vaccine ciirrently employed against an iiitracellular bacterium is tlie attenuated S. enfcrictr serovar TfjphiTy21 strain. Becaiise of its low and sIiort-livecI protection, it is only aciministerec1 to travelers to areas where a high risk of coiitl-actingtyplioid exists. There is not only a need for more effective, inexpensive,and easy to apply vaccines against tuberculosis and typhoid fever, h i t also against otherworldwide liealth burdens caused by intricelhilar bacteria such as C. tiaclu,t,zriti.~,which represents the predoininant ciiiise of congenital blindness in developing countries. Rational development of a new generation of vaccines h a s to take into account the unique character of the immiine response that is effective against the respective pathogen. Accordingly, the requirements for such abaccine inc:lutle (i) induction of the “right” T-cell populations arid cytokine patterns that are rapidly mobiliztd and, at the same tinre, mediate protective and long-lasting (nienioiv) immunity; (ii) activation of the “right” effector mechanisins to prevent infection or to eliminate the infectioiis agent; (iii) exclusive specificity for tlie responsiible pathogen to avoid the risk of autoaggression through cross-reactive antigens; (iv) coverage of antigens that are expressed by the hacteria in the host and by various isolates or strains; and (v) iiiiiiiunogenicityiiiogeiricity fbr all MHC haplotypes within 1111111a11 pop111iit’1011s. First of all, a \roccine against an intracchlar bacterium should induce the appropriate combination of specific CD4’ and Cll8’ T cells that produce IFNy to subsequently activate macrophages to eliminate the bac-
328
UI.RICII E SCII4IULF: uf NI
teiia. Vaccines directed against some bacteria have to activate MHC class I-restricted CTL more profoundly than others. This is of particular importance for bacteria that parasitize nonprofessional APC that do not express MHC class I1 molecules or only at low densities. Additional strategies include antigens that stimulate unconventional T cells, such as glycolipids for CD1 reactive T cells and phospholigands for y 8 T cells. Unconventional T cells respond faster than conventional T cells and can therefore promote the generation of Th 1 cells. Glycolipids have the additional advantages (i) that they induce CTL, although they are derived from endo-/phagosomes, and (ii) that CD1 is a nonpolymorphic molecule. However, the overall contribution of unconventional T cells to the efficacy of vaccineinduced protection against intracellular bacteria remains to be determined. To rationally design an effective vaccine, the following aspects are important: i. the choice of antigens and their display and physicochemical nature (intracellular, surface associated or secreted proteins, glycolipids, phospholigands); ii. vaccine types (whole proteins or peptides, whole live attenuated bacteria, recombinant bacteria, naked DNA); and iii. formulations and ways of delivery (higMow doses, adjuvants, immunomodulatory cytokines, immunostimulatory DNA sequences).
Currently, various vaccine strategies against the intracellular bacteria discussed here are being tested with a focus on vaccines against tuberculosis and typhoid fever. In addition, some of these vaccines are also examined as carriers for heterologous antigens. A. IDENTIFICATION OF PROTECTIVE ANTIGENS
During recent years, an array of studies has been published describing proteins from various species of intracellular bacteria and their immunological relevance. Antigens that are released or actively secreted by live bacteria are recognized by the immune system early during infection when the number of degraded bacteria is low, whereas somatic proteins are only recognized after bacterial death and degradation (Kaufmann and Andersen, 1998). Therefore, most studies have concentrated on the identification of secreted proteins as candidate antigens for protection. The superior effectiveness of secreted vs somatic antigens for inducing protective immune responses has been demonstrated most convincingly in studies using either recombinant salmonellae or recombinant listeriae expressing defined antigens (Hess et al., 1996a, 1998; Shen et al., 1998). For M . tuberculosis, more than 200 proteins were found to be released into the culture supernatant, including antigens of potential relevance to protection (Andersen et
nl., 1991; Andersen, 1994; Sonnenberg and Belisle, 1997; Kaufinann and Andersen, 1998). Several studies found that mycobacterial proteins released into the culture supernatants can induce protective T cells (Hubhard ct d . ,1992; Pal and Honvitz, 1992; Andersen, 1994; Roberts et nl., 1995). Vaccination with secreted proteins (including the Fn-binding 30-kDa protein of the Ag85 complex) from M . triliercrilosis protected guinea pigs against aerosol challenge with tuliercle bacilli (Honvitz et nl., 1995). Moreover, Ag8rj and a 29-kDa antigen (CFP29) induce IFNy-secreting T cells in mice (Andersen et al., 1995; Rosenkrands ct nl., 1998). ESAT-6, a secreted 6-ltDa protein that is expressed exclusively in M . tiiherczilosis but not in hl. b i x i y BCG, induced T cells in mice tliat expressed a phenotype potentially associated with protection (Andersen ct d.,1995; Brandt et'crl., 1996; Harboe ef d., 1996). However, it appears that ESAT-6 given in different forins of vaccinations fails to induce satisfactory protection. Other mycobacterial antigens released into tlie culture supernatant, such as the proline-rich coiiiplex (Roinain cf nl., 1993), the chaperones a-crystallin and GroES (Young and Garbe, 1991;Verbon ct nl., 1992), SOD (Andersen et a / . , 1991; Andersen, 1997), MPTS1 (Wiker ct d . , 1992), MPT63, and MPT64 (Haslov et d . ,1995) still await further analysis with regard to their protective potential. Still, these proteins encoinpass the most promising vaccine candidates as dead mycohacteria are insufficient to induce a protective immune response (Orine, 1988). It should be noted that none of the antigens tested proved to be superior over BCG and that the final outcome of such studies will most probahly show that not all secreted proteins are protective. Enormous effects have been put into tlie analysis of the total protein spectniin of various bacteria. A number of groups are working to establish the proteomes of intracellular bacteria, such as M . ttdw-culasis, M . bovis BCG (Sonnenberg and Belisle, 1997; Urquhart et d., 199'7; Welding11 et d . ,1998; Jungblut ct a/., in preparation), S. trplzimurizrni (Qi et d . ,1996; O'Connor ct d.,1997; Burns-Keliher ut d.,1998),Brcicdn ouis, B. melitcrrsis (Teixeira-Gomes et al., 1997a,b),and C. trnclioimtis (Bini et d., 1996). These efforts benefit from, and compleinent greatly, ongoing genome sequence iinalyses of various bacterial pathogens, including M . tnherctiZo.sis, which has been announced to be contiguous (Cole et NZ., 1998). Apart from identification of protein patterns exl&ssed under various conditions, proteoine analyses will also facilitate the determination of posttranslational modifications such as acylation, glycosylation, and phosphorylation of the respective proteins.
B. ATTENUATEDA N D RECOMBINANT VAC:CINE STRAINS Use of live-attenuated bacteria as vaccines has the benefit that the organisms replicate in the host for at least sonie time and therefore gener-
330
U L R I C H E SCIIAIBLE rl
(I/
ally induce long-lasting specific immunity. Moreover, they resemble the fully virulent pathogen to a certain extent and thus may induce a similar iinmune response. Such vaccines currently in use encompass those against tuberculosis, M. bovis BCG, and typhoid fever, S. enterica serovar Typhi Ty21. As mentioned before, M. bovis BCG has been used worldwide as vaccine against tuberculosis and leprosy, although with variable and often low efficacy. Several hypotheses have been brought forward to explain the variable outcome of M. bovis BCG vaccination. These include (i) exposure and sensitization through environmental mycobacteria, which vary greatly between geographical regions (Stanford, 1991; Fine, 1995); (ii) lack of important genedantigens in BCG that are expressed exclusively in virulent mycobacteria (Harboe et d ,1996; Mahairas et al., 1996); and (iii) insufficient induction of CD8+ T cells by A4. bovis BCG, thus excluding an important iinmune mechanism for protection against tuberculosis (Muller et d., 1987; DeLibero et d.,1988; Flynn et al., 1992; Bonato et al., 1998). The latter obstacle may be overcome by the generation of recombinant M. bovis BCG expressing the listerial LLO, which facilitates targeting of BCG antigens into the MHC class I pathway (Hess et al., 1998). This construct will be improved further by including genes encoding M . tuberculosis-specificproteins, with the ultimate goal of targeting protective antigens to both MHC class I and I1 pathways. The abundance of glycolipids with inflammatory properties in the rnycobacterial cell wall and the high proportion of CpG motifs in the mycobacterial DNA most probably further promote a Thl type response. Recombinant vaccine strains such as reconibinant M. bovis BCG expressing heterologous antigens have already been used in vaccine studies against other infections: recombinant M. bovis BCG expressing the surface cysteine protease gp63 of L. nzxicana or the outer surface protein A of Bovelia burgdu$eri can protect against cha1lenge with the respective pathogen (Connell et al., 1993; Stover et al., 1993; Langermann et al., 1994). Another approach aims to identify virulence and survival factors for the generation of KO mutants by genetic deletion. These attenuated strains fail to thrive in the host but retain their immunogenic potential and may be considered safe vaccine strains. Such KO bacteria potentially encompass those lacking metabolically essential genes, i.e., auxotrophic mutants, which die in the host due to a lack of essential nutrients. Using transposon mutagenesis, auxotrophic M. bovis BCC mutants have been generated that are attenuated but still able to induce an immune response comparable to wild-type BCG (McAdam et al., 1995; Guleria et al., 1996). These vaccine strains may provide a safe way to immunize immunodeficient patients, which may develop generalized BCG infection on vaccination. Self-limiting vaccine strains such as the suicide L. nuinucytogenes strains
could proiide a solution to this obstacle because they survive long enough to retain their high iininunogenicity but ultimately die after escape into the cytoplasm (Dietrich ct nl., 1998). Auxotrophic S. trjpliiniuriimi AroAmutants, wliicli were generated by transposon deletion inutagenesis and have a defect in aromatic biosynthesis, inducc protective immunity against salmonella infection in mice (Hosieth and Stocker, 1981; Newton et al., 1989; Horniaeche et 0 1 , 1991; Harrison et a/ , 1997).Attenuated recombinant salmonellae have been genetically engineered to actively secrete various antigens by virtue of the E coli secretion apparatus. These recoinbinant salmonellae have been shown to deliver foreign proteins into the MHC class I and 11 pathway (Hess et nl., 1996a, 1997). Virtually every heterologous antigen could be introduced into this system, probahly also in combination with other antigens. Similarly, recombinant listeriae can also be used as vaccine vehicles (Jensen et (11, 1997; Shen et al., 1998). The recoinbinant salmonellae or recombinant listeriae vaccine carriers have tlie fiirther advantage that they can be administered orally. A different and probably more promising approach is the genetic deletion of certain virulence factor genes, which will diminish the intracellular sui-vival rate of the respective bacteria. By employing transposons tagged with a unique DNA sequence, genes of S. typhiniuriuui were identified that contribute to survival of the salinonellae in mice (Hensel et nl., 1995).This experimental system is currently also applied to other intracelliilar pathogens. However, this approach may be hindered by tlie fact that the capacity to survive intracellularly is based on many factors/genes. Based on the current knowledge of the cytohnes important for controlling intracellular bacteria, i.e., IL-12, GMCSF, and IFNy, recombinant vaccine strai 11s have been constructed that express these cytokines alone or together with defined antigens. Thus, recombinant M . h i s BCG strains expressing IL-2, IFNy, and GMCSF werc found to be superior to wildtype BCG in inducing protection in mice (Murray et d., 1996). Similarly, attenuated recombinant salmonellae exliressnig MIF, TNFa, and IFNy were used siiccessfully to direct the inimune response against L major in susceptible BALB/c mice toward a Th1 type T-cell response (Xu et d., 19981)).Although this approach implies that cvtokines secreted from intracellular bacteria are trafficked to the responsive cells without being degraded, some bacteria may also survive outside of their host cells to release sufficient concentrations of cytokines into the extracellular milieu. C.
S U B U N I T VACCINE5 A N D ADJU\’ANTS
The benefits and drawbacks of‘ purified native or recombinant proteins as vaccine antigens are obvious: These vaccines are highly specific and bear a low risk of side effects; however, when given alone, their immuno-
332
ULRICII E SCHAIBLE r f
n/
stiinulatory and protective efficacy is generally weak and short-lived. For example, after immunization of mice with a mixture of culture supernatant proteins from M . tztbercrrlosis in adjuvant, at best a protection level comparable to BCG vaccinations was achieved. However, BCG-induced protection was long lasting whereas protection induced by the subunit vaccine faded after 5 months (Roberts et al., 1995). Therefore, subunit vaccine formulations depend strongly on inimunostimulatory adjuvants that siniultaneously create the milieu required for the stimulation of protective T cells and generate a reservoir for slow antigen release at the site of application to facilitate long-lasting protection. Alum, the only adjuvant approved so far for use in humans, preferentially induces a Th2 response whereas adjuvants stimulating Thl cells are required for vaccination against intracellular bacteria. Currently, most of these adjuvants are hampered by the fact that their capacity to stimulate a Thl response is accompanied by severe inflammation resulting in tissue destruction. Bacterial glycolipids and oligonucleotides (see later) may be considered “natural” adjuvants due to their proinflammatory properties. As described in Section 111, LPS, LAM, and other bacterial glycolipids are strong inducers of cytokines triggering a Th1 response. Dissecting the adjuvant-associated stimuli that induce inflammation from those that elicit Thl cell responses would strongly improve subunit vaccine strategies. A number of adjuvants have been developed with reduced inflammatory potential but retaining the capacity to induce antigen-specific Thl responses. These include liposoines, microspheres, squalene, and dimethyl deoctadecyl ammonium bromide. Moreover, some adjuvants are endowed with the capacity to introduce antigens into the MHC class I pathway to trigger CD8’ T cells. Of these, iinmuriostiinulatory coinplexes (ISCOM) represent the most advanced adjuvants (Takahashi et nl., 1990; Heeg et nl., 1991; Morein et nl., 1996). Moreover, cytokines, i.e., IL-12, IFNy, or GMCSF, contained in the vaccine formulation can direct the immune reaction toward a Th1 type response (Bermudez and Kaplan, 1995). Peptide vaccines comprising defined protective epitopes while bearing the lowest risk of cross-reactivity because of their unique specificity have the disadvantage of being the least immunogenic. Due to the different peptide motifs defined by the various HLA haplotypes in the human population, only a vaccine encompassing a mixture of antigenic peptides will cover the repertoire of T-cell epitopes present in a heterologous popillation o f vaccinees. Such peptide mixtures are chemically linked during synthesis to form linear or branched multipeptide complexes. To induce an immune response, peptides or complexes thereof are linked to carrier proteins and strong adjuvants have to be incorporated. As a corollary, it is not surprising that up to now only one peptide-based vaccine has been
proven to induce sufficient protection, nainely that against foot and mouth diseuse in cattle (Ada, 1990). The finding that T cells elicited by synthetic peptides can recognize peptide-piilsed APC but not APC loaded with tlie complete protein, which present naturally processed peptides, adds to the arguments against peptide vaccines (Viner et crl , 1996). A different approach is Ixised on immunization with a combination consisting of antigenic peptides and certain HSP inolecules. Vaccination with HSPSO plus peptides has been sliown to mduce protection against challenge with LCMV (Ciupitu ct 01, 1998) and, siinilarly, application of HSP’iO and gp96 plus peptides protected mice against tumor growth (Udono and Srivastava, 1993; Suto, 1995).Although the underlying inolecular mechanisms are not fully understood, it is assumed that HSP preferentially bind peptides for M HC class I processing and subsequently introducc thein into tlie MHC class I processing pathway. Similarly, fusion proteins comprising HSP70 plus HIV1324 or an antigenic ovalb;imin-clerived peptide (SIINFEKL) induced specific iinniune responses including CTL to tlie respective antigens (Suzue and Young, 1996; Suzue et nl., 1997). The feasability of this approach to vaccination against intracellular tiacteria needs further exploration. D. N A ~ E D D N A The latest achievement in vaccine developinent is the direct application of naked plasmid DNA encoding the geneis) of tlie respective antigen(s) (Tang et nl., 1992; Tighe ct 01, 1998). Plasinid DNA is delivered by direct intramuscular injection, by use of a balistic device to “slioot” DNA-coated gold particles into the skin (gene gun), or by means of bacterial carriers (see later) Gene expression is usually under control of a strong viral promoter, siicli as tlie CMV promoter, wliich allows expression of encoded genes by the transfected inamnialian cells, most probably muscle cells. Integration of a secretion signal into tlie plasmid allows export of the antigen by the cell. Subsequently, the iuntigen can stimulate B cells and can also Le taken up by infiltrating nionocytes and dendritic cells, which will then stimulate T helper cells. This vaccine type is safe and easy to produce, store, and administer. So far, DNA vaccination has been einployed successfully in various experiinental infkction models such as influenza virus (Ulmer et al., 1993), Mycoplnsitici p i h u m i s (Barry et a1 , 1995), L. niiljor (Xu and Liew, 1995), an‘d B.hrirgtlorferi (Zliong et d . ,1996; Luke ct nl., 1997). In tlie case of M . pubtionis, expmsion libraries and sublibraries were used for immunization (Barry et nl , 1995). To screen a whole set of antigens at once represents an interesting approacli that is currently applied to other bacteria such as mycobacteria. To date, three studies have demonstrated that DNA vaccination with mycobacterial genes induces a protective iinmune re-
sponse against M ticbercttlosis in mice comparable or only slightly lower as compared with BCG. The antigens used were HSP6O and the 36-kDa proline-rich antigen from M . leprae (Tascon ct d., 1996), the secreted M . tiiberczdosis protein Ag8SA (Huygen ct al., 1996), and the 38-kDa glycolipoprotein (Zhu et al., 1997). In these studies, protection correlated with the induction of high levels of IFNy producing cells and CTL. Moreover, one report showed that a DNA vaccine based on tlie porin gene 0mpC of S. typliimuriuin induces a humoral immune response in mice (Lopez-Macias et d., 199s; reviewed in Sti-Lignell ct al., 1997). Following DNA vaccination, the inflammatory reaction induced by distinct bacterial DNA sequences, such as the CpG-motif, with immunostiniulatory potential promotes tlie infiltration of nionocytes. Similarly, inclusion of genes encoding cytokines and costirnulatory molecules into the vaccine plasmid can promote the appropriate irninune response as shown in the influenza system:The capacity of a plasmid containing the influenza nucleoprotein to stimulate CD8’ T cells wiis improved successfully by the addition of genes encoding IL-12 and GMCSF and the costiinulatory molecules B7-1 and B7-2 (Iwasaki ct al., 1997). Knowledge about the iinmimoinodulatory effect of certain DNA sequences came from early studies trying to identify the molecular entities of M . houis BCG, which are responsible for nonspecific tumor rejection after treatment of tuinor-bearing mice with BCG (Kataoka et al., 1992). The isolated component was identified as BCG-derived DNA, which appeared to be a strong inducer of N K cells and interferons a, b, and y (Sliiniadaet al., 1986; Mashiba et nl., 1988;Yainainoto et nl., 1988; Pisetsky, 1996). The responsible DNA sequences were found to contain at least one palindrornic stretch of the motif 5’-purine-purine-CG-pyrimidinepyriinidine-3’ such as GACGTC, GCXGCC, AGCGCT, and AACGTT (Tokunaga ct nl., 1992; Yamamoto ct d., 1992, 1994). Although the high G/C content in inycobacteria favors abundance of CpG motifs in M . houis BCG, this phenonienon is obviously not restricted to M . bouis BCG-derived DNA, a s DNA from E. coli also induces IFNy via IL-12 and TNFa (Halpern et nl., 1996).Therefore, certain properties of bacterial DNA must be distinct from mainrnalian DNA. The CpG motif occurs more frequently in bacterial (1/16 bp) than in inaininaliaii DNA (1/SO bp), with cytosins being less inethylated in bacteria (<SO%) than in mammals (70-90%) (Pisetsky, 1996;Tighe et d., 1998).The immunogenicity of a DNA vaccine was reduced significantly by methylating its CpG motif (Klinman et al., 1996). CpG containing immunostirnulatory DNA sequences (ISS) have been shown to be strong inducers of IL-6, IFNy, IL-12, and IL-18 and subsequent Th1 and CTL responses i15 well as high antibody titers dominated by the IgG2a isotype (Klinman c>tnl., 1996; Sato et NI., 1996; Lipford
nl., 1997: Koman rt d., 1997).This is most probably due to the infiltration of potent .4PC siich as dendritic cells and niacrophages to the site of intriidernial injection. Moreover, it lias been ~ O L I that I ~ ~ CpC: DNA sustained IL-12 production, thus promoting nonspecific resistance to L. tjwnocgtogcnes infection (Krieg ct nl., su1)mitted). The Th 1-dominated effect seen after intradernial DNA iiijectioii is quite opposite to the outcome of vaccine adininistratioii using a ixilistic device, wliicli ratlier leacis to a TMlike response (Feltquiite ct NI., 1997). It has been proposed that the 100fold lower amount of DNA iised in the gold particle delivery system compared to direct DNA injection l e d s to a dilution of the ininiunostimiilatory DNA sequences, thiis reducing any inflammatory reaction that results in TI11 cell activation (Tighe et a[.,1998). It has been reported that bacteriiil \rectors can be used as transport veliiclcs for recombinant DNA. Tliis vxcinatioii stratea opens the possibility of oral immunixatioii and will therefore additionally activate inucosal immune responses. So far, attenuated recombinant S. jexrieri ( Sizeinore ct d . , 199511, recoinbinant S. t z j p h i t t u i r i r i i j i (Darji rt nl., 1997), and suicide recoinbinant L. I,iotiocglojiciies ( Dietrich r f nl., 1998) liave been used in these qqxwiches. Taken together, a vaccine based on genes of iminunodoininai~tbacterial antigens probably coupled to cytokine ancVor costiniulaton~molecules inducing Th1 responses represent a promising candidate to control infections with intraettllular bacteria. Althongh there were only a few reports that upon DNA vaccination the foreign DNA has been integrated into the host genome (Doerfler et d., 1997; Dietrich et d., 1998), future studies on DNA vaccination have to take these safety aspects into account. Nevertheless, DNA vaccination represents a potent approach to battle infections with intracelliilar bacteria. ef
336
UI.Hl(:H E. SCHAIBLE rt ol.
REFERENCES Ahhas, A. K., Murphy, K. M., arid Shrr, A. (1996). Functional diversity of helper T lynphocytes. Nature 383, 787-793. Abou-Zeid, C., Hatliff, T. L., Wiker, H. G., Harboe, M., Bennedsen, J., and Rook, C . A. W. (1988).Characterization of fibronectin-binding antigens released by Mycobncterium tuberculosis and Mycohcteriurti both BCG. Infect. I m m z i r i . 56,3046-3051. Abu Kwaik, Y. (1998). Induced expression of the Legionellrr pneumoplrila gene encoding a 20-kilodalton protein during intracellular infection. Infect. Inmiun. 66,203-212. Abu Kwaik, Y., Eisenstein, B. I., and Engleberg, N. C. (1993). Phenotypic modulation by Legionella ~~ntrc~nopliila npoii infection of macrophages. Iifect. Ittiniun. 61,1320-1329. Ada, G.L. ( 1990). The iinmiinological principles of vaccination. Lancet 335,523-526. Agganval,A.,Kuinar, S.,Jaffe, H., Hone, D., Gross, M., and Sadoff,J. (1990).Oral Salmonella: Malaria circumsporozoite recombinants induce specific CD8' cytotoxic T cells. J. Exp. Med 172, 1083-1090. Ahearn, J. M., and Fearon, D. T. (1989).Structure and function ofthe complement receptors, CR1 (CD35) and CR2 (CD21). Adu. Irtimurio~.46, 183-219. Albert. M. L., Sauter. B., and Bharclwaj. N . (1998). Dendritic cells acquire mtigen from apoptotic cells and induce class I restricted CTLs. Nature 392,86-89. Altlrich, C. J.. DeCloux, A,, Woods, A. S., Cotter, H. J., Soloski, M. J., and Forinan, J. (1994). Identification of a Tap-dependent leader peptide recognized by alloreactive T cells specific for a class Ib antigen. Cell 79, 649-658. Alpuche-Aranda,C. M., Berthiauine, E. P., Mock. B., Swanson,J. A,, and Miller, S. I. (1995). Spacious phagosonie formation within niouse macrophages correlates with Salmonella serotype pathogenicity and host susceptibility. Itfect. Iminun. 63,4456-4462. Alpuche-Aranda, C. M,, Racoosin, E. L., Swanson. J. A,, and Miller, S. I. (1994). Salmonella stimulate macrophage macropinocytosis and persist within spacious phagosonies. 1.Esp. Med. 179,601-608. Alpuche Aranda, C. M., Swanson, J. A,, Loomis, W. P., and Miller, S. 1. (1992). Salinnnielh typhimuriuni activates virulence gene transcription within acidifed macrophage phagosomes. Proc. Nat. A c Sci. ~ USA 89, 10079-10083. Alhire, F.,Durandy, A., Emile, J.-F., Lamhametli, S., Le deist, F., Laminas, D., Dysdale, P., Jouanguy, E., Doffinger, R., Bernaudin, F., Jeppsson, O., Segal, A. W., Fischer, A,. Kuniararatne. D.. and Casanova. J.-L. ( I 998). IL-12 receptor deficiency selectively impairs rnycobacterial irnmunity Sciciice 280, 143.5-1438. Alvarez-Dominguez, C.,Barbieri, A. M., Beron, W., Wandinger-Ness, A,, and Stahl, P. D. (1996). Phagocytosed live Listeria nionocytogenes influences Rab5-regulated in vitro phagosoine-endosoine fusion. J. B i d . Ch.em. 271, 13834-13843. Alvarez-Dominguez,C.,Roberts, R., and Stahl, P. D. (1997). Internalized Listeria nronocytogenes modulates intracellular trafficking and delays maturation of the phagosome. J. Cell Sci. 110,731-743. Andersen, P. (1994). Effective vaccination of mice against Mycobucterizun tuberculosis infection with a soluble mixture of secreted mycobacterial proteins. Infect. Immrn. 62, 2536-2544. Andersen, P. (1997). Host responses and antigens involved in protective iminunity to Mycobacterium tuhercuZosis. Scand. /. Irnniunol. 45, 115-131. Andersen, P., Andersen, A. B., Sorensen, A. L., and Nagai, S. (1995). Recall of long-lived irnmunity to Mycobaderiuni tzibercldosis infection in mice. 1.Irnnirrnol. 154,3359-3372. Andersen, P., Askgaard. D., Ljungqvist, L, Bennedsen, J., and Heron, I. (1991). Proteins released from Myc(~bncteriumtubercubsis during growth. Infect. bnrnunol. 59, 19051910.
INTHACELLUIAH RA(:TF:HI.A AND THE IMhllJR'E SYSTEhl
337
Aiitoine. J. C.. J ~ ~ i i i i i i i C., e , I,ang, T., Prina, E., de Cliastellier, C., and Frchel, C . (1991). Localizatioii of major 1iistocoinpatil)ilitvcoiiiples class II nrolecules i n phagolvsosomes of niiirine nimup1iage.s infrctetl \tit11 I , ( + d t t r m i o ~ l t t t m l t t f ' J I , Y i . S .It!/&. I t t t r r t i r t i . 59, 764-775 Appellierg, R. Castro, A. G.,Pedrosa, J.. ;urd Miiiopiio. P. (1994). Role of interleiikin-6 iri tht. induction of protecti\v T cells thiiing iirycolxictrrial infections in mice. Ir~triiirtioZogy 82, 361-364. Appelberg. R., and Orme, 1. M. (1993).Effcxtor iiic&misnis involved in cytokine-rrictliated IYl..,. L ri- :iostiisis i ) f nl!/")~(ic.tc.r-iirri~m : i i t t t t iiilrc.~ioiisin iiniiiinc~niiacrrqihiigv~.Irtivizrtt&)gy 80, 352-359. Apt. A. S., Adienko, V. G.. Nikonenko. B. \'.. Kniiiinik, I. B.. Moroz, A. M.,iiiitl Skanienr, E. (1993).Distinct II-2coniplex control of mortality, iirrd iniinuiie responses to tul)crciilosis iiifrction in \irgin arid BCG-v;icciiiatrtf mice. Clitt. Esp. f r r i t r i i r t i ~ ) l .94, 3322-329, Arasr, H., Arue. N., Nakagawa, K.. Good. R. A , . and Oiroe, K. (1993). NKl.l'CD4'CD8thyniocytes with specific^ lynipliokine secretion. Err,.. /, Z r r t t r i i t t i o l . 23, 307-310. Alias, M.. Rojas, M.. Zahaleta. J., Rodiigiiez. J. I., Paris, S. C.. Barrera, L. F., and Giircia, L. F. (1997). 1nliil)ition of virulent 12lyeijbNc,tc.r~rtititic/wrcrtZo.si.rby Bcg(r) and Bcg(s) macrophages correlates with nitric oxide productioii. /. Ir!fecf.Di.v. 176, 1552-1.558. Armstrong. J. A,, and D'Arcy Hart, P. D. (1971). Response of cultured niacrophages to Mycobcrc.trt-iirrtItitbcr~dosis,with ohenations on f i i k i o i i of lysosomes with phagosomes. 1. Esp. Afcrl. 134, 713-740. Arnistrong, J. A,, a n d Hart, P. D. (197.5). Phagosoilre-lvsosoiri~ interactions in cultured macrophages infected with virulent tril)ercle bacilli: Reversal ofthe usuiil nonfiision pattern aiitl olxeiv;itioiis OH hacterial s u n i d /. Exp. hftv1. 142, 1- 16. Arnold, It.. and Konig. W, (199'3). Iiitrrlrukiir-8 release from huiriaii iiriitropliils after phugocytosis of Listc,ricr rnotic)cylogfric,s and Y(Jr.sitiioeritc,rocoliticci. /. M i d . Microbial. 47, 55-62. Athison, P. C.. Blackwell. J. M., ant1 Bartoii, C. f1. (1997. Nrampl locus encvcles a 65 kDa iiiterferon-y-iiitlricihle proteiii i i i niurinc niacrophages. Bioe/wni. 1. 325, 779-786. Augustin, A,. Kubo, R. T., and Sini, G. K. (1989).Resident l~iilmona~lyinphoc~Zes cxpressiiig the yl6 T cell receptor. Natirt-c, 240, T39-241. Autenrieth, I. B., Berr, M., Bohii, E., Kaufinaiin, S. El. E., and Heesemann, J. (1994). Iniinune responses to Yersinirr oitcrocolitictr in siisceptible BALB/c and resistant CSSBW 6 mice: An essential role for ganiina interf(.ron. Itfict. I t t i t t i i t t i , 62, 2590-2599. Bachmann, M. F., Lutz, M. B., Laytoii, C. T., llarris, S. J.. Fc>hr,T., Rescigno, M., and Ricciardi-C'istagiioli. P. ( 1996). Dendritic cc~llsprocess exogenous viral proteins and c i m s like particles for class I preserrtation to CDH' c),totodc T lymphoc\tes. Eirr. J. IrttrriittioZ. 26, 2595-2600. Balaji. K. N..and Boom, \V, H . (1998). Processing of h.l!yco/xrcfen'rrtn tuhrrcirkxis bacilli hy human iiionocytes for CD4' a/3 and y6 T cells: Role of particdate antigen. Irfert. 1 ~ 1 1 ~ t i i ~ t 66, i . 98-106. Balaji, K. N., Schw;irider. S . K., Rich, E. A., and Boonr, W. 11. (1995).Alvcolar iiiacrophages a s accrssory cells for hrinian y6 T cells activated hy A.i!/cobncteritrtri tnbcrcdosis. /. Z t r i t t i i t r t d . 154, 5959-5968. Halk. S. P., Burke, S.. Polischuk, J. E., F r a n k M. E.. Yaiig, L., Porcelli. S.. Colgan, S. P., and Bluiiiberg. R. S. [ 1994). Beta ~-microglohiilin-independentMHC class 111 molecnle expressed by Iinnian intestinal epithrlium. Scic~itc.~ 265, 259-262. Balk, S. P., Ehert, E. C.. Hliiinenthal. R. I.,., McDermott, F. V., Wucherpfennig. K. W., Landau. S. B., arid Hlumberg, R. S. (1991). Oligoclonal expansion and C D l recognition by Iiiiiiiiiii intestinal intrac~pithelialIyiiiphoc)tes. Scirtlce 253, 1411-141 5.
Bancroft, G. J.. Schreiher, H. D., antl Uiiiinue, E. H. (1991). Natural innniiiiity: A T-cellindepeiident pathway of rnacrophage activation, defined in the scitl iiioiisc*. Z r t i i r i i r r i o l . RCL. 124, 5-24. Baorto, D. M., Gao, Z., Malaviya, R.. Dustin, M. L., van der Mcnve, A., Lublin, D. M., and Abraham, S . N . (1997). Survival of FimH-expressing enterobacteria in macrophages relies oii glycolipid traffic. Nature 389, 636-639. Barker, L. P., George, K. M., Falkow, S.. :ind Sinall, P. L. (1997). Differential trafficking of live aiitl d c d M!/c&zctsrirrrrt nwrinurri organisms in inacrophagrs. 1nfc.c.i. Z r r u r i r t > i . 65, 1497-1504. Barnes, P. F., Foiig, S . J., Brennan, P. J., Twomry, P., Mazumtler, A., and Modlin, R. L. ( 1990). 1,ocal production of tumor nccrosis fietor and IFN-y in tuberculous pleuritis. 1.~ I t l ~ t l ~ 145, ~ ~ I ~149J ~ . 154. Barnewall, R. E., Hikihisa, Y., arid Lee, E. [ I. (1997). Elirlicliia chnfJerisis inclusions are early cndosomes which selectively xcuninlate transferrin receptor. Irfect. I r i i r r i i t r i . 65, 1455-1461. Barly, M . A,, Lai, W.C., antl Johnston, S. A. (1995).Protection against mycoplasma intection using expression-library iiiiiiiunization. Nntrrrt, 377, 632-63-5. Barzu, S.. Benjelloiiii-Touinii, Z., Phalipon, A,, Sansonetti, P., and Parsot, C. (1997).Fuiictional analysis of the Shigdln Jurwri IpaC invasin by insertional niutagenesis. Itfict. f r r i t r i u u . 65, 1599- 160lj. Bauer, A,, Huttinger, R., Staffler, G., Hatismann, C., Schmidt. W., Majtlic, 0.. Knapp, DJ., antl Stockingrr, H. ( 1997). Analysis of the rrquirement for beta 2-inicroglobulin for expression and formation of huniaii C13 1 antigens. Brr. J . Z m t m t i o l . 27, 1366- 1373. Beaiiregartl. K . E., Lee, K. D., Collier, H. J.. and Swanson, J. A. (1997).pl-l-dependent perforation of macrophage phagosomes by hsteriolysin 0 from Listerin nionocytugcricx J. Erp. M i d . 186, 1159-1163. Beckers. M . C , Ernst, E., Diez. E., Morisscttca, C., Gewais, F., Hunter, K., Housman, D., Yoshida, S.. Skaniene, E., a i d Gros, P. (19971.High-resolution finkage map of nioiise chroinosome 13 in the vicinity of the host resistance locus Lgtil. Gerioniic,s 39, 254-263. Beckman, E. M., Melian, A. M., Behar, S. M., Sieling, P. A,. -Chatterjee, D., Matsmnoto, R., Rosat, J.-P., Modlin, R. L., and Porcelli, S. A. (1996). CDlc restricts responses of niycobacteria-specific T cells: Evidence for antigen presentation by a second member of the human CD1 family. J. Irritnirnol. 157, 280:3. Beckman, E. M., Porcelli, S. A., Morita, C. T., Behar, S. M., Furlong. S. T., and Breniier, M. B. (1994). Kecogiiition of a lipid antigen by CD1-restricted &3+ T cells. Natrcru 372,691-694. Beebe, A. M., Mauze, S., Schork, N. J., and Coffinan, R. L. (1997). Serial backcross mapping of multiple loci associated with resistance to Leishrrutnici riuqor in mice. frrimmiii!/ 6, 55-557. Belisle, J. T., Vissa, V. D., Sievert, T., Takayama. K., Brennan, P. J.. and Besra, G. S. (1'397). Role of the major antigen of Mycohncteriwri hibPrcdcwis in cell wall biogenesis. Science 276, 1420-1422. Bellainy, R., Ruweude, C., Corrah, T., McAdain, K. P. W. J., Whittle, 14. C., and Hill, A. V. S. (1998).Variations in the NRAhlPl geiic and susceptibility to tul)erculosis in west africans. N . Etigl. J. Med. 338, 640-644. Belles, C., Kiih1, A. K., Donoghue, A. J., Sano, Y., O'Brien, H. L., Born, W.. Bottoinly. K., and Carding, S. R. (1996). Bias in the yS T cell respoiise to Listerin nionoc~togenes: Vy%.3+ cells are a major coinporrent of the 76 T cell response to Listerin rrumoc!yfo~eties. J . I n i n i w i d . 156, 4280-4289.
Bellonr, M., Iezzi. (;., Ro\we, P.. Galati, C., Rochetti. A , , Pia Protti. M . , Davoust. J., Rngiirli, (:., ; i i d hlaiifretli, A. A . (1997). Processing of riigiilfid apoptotic I d e s ~ieltls T cdl cyitopes. J . I t t t t t t ~ ~ t t o /159, . 5591-53399. Bericlelac, A.. Lairtz. O., Qiiiiiih!,. M. E., Yewtlrll. J. \f7.. Bennink, J. R.. aird 13rutkiewvic.z. 1%. R. (1993). CD1 recognition hy nioiise N K I T Iyiiphocytes. Scimcc 268, 86:3-865. Heiidc~lac.A,, Rivera, M. N.. Park. S. 13.. and Koark. J. 13. (1997). Moiise C:Dl-spc~cific NKI T cells: Development. specificih. iind finictioii. Attttu. H(T. Z t t t t t i u t t o l . 15, 535-562. Hermiidez. L. E.. and K q l a n . C.(19953. Reconiliinant c!-tokiiies for controlling iiiycolxicterial infections. Trenrls L\ficrobiol. 3, 22-27. Berniudez, I,. E.. Parker. A , , urid Gootlman. J. 11. ( 1997). GrowtIi \iitlrin macrophagc.s incre:scs tlie (4ficiwcy of nf!t'."bo'.lf,riurti oriri t t i i n invading otlier niacropliages by a coiiipIcwciit recrptor-intlepentleiit patliwa!. It+c.t. Z t t t t t r t i t i . 65, 191~-1925. Bvrna, A., Kopf, M., Kul1)acki. R., Ll'eich. N.. Kdiler, C;.. and (:iittierez-Haiiios. J. C. (1994). Interleiihin-6is reqiiircd in \ivo for tht, rrgul;ition of'stein cells and conimittecl progcwitors of the Ixieniopoietic system. Z t t i t t i t i t t i t ! / 1, 72.5-7:31. Beron. W.,i~lvarrz-Doiiiingii(.z. C.. hluyorga. I,., and Stahl. P. D. (1995). Menilmne trafficking h n g the phagocytic piitlway. Tretrdy Crll Riol. 5, 100-104. Beutler, H.. and C;el.ami, A. (1987). Caclic,ctin: More thiur a tumor necrosis factor. .V. Eirgl. /. M d . 316, 379-3385. Bini, L.. Saiicliez-CainpilltJ, M., Santiicci. I\., Magi. €3.. Marzocclii, €3.. Comaiidiicci. M.,C:hristiansen, (;.> Birkelund, S., Crveiiini. H., \'rc>tori, E., Ratti, G.. and Palliiii. V. ( 1 996). Mapping of C h h q d i c i frclc/rotticitis proteins hy iiiiiiiol,ilinr-pc~l~tc~lanii(!f hvotlirntwsioiial electroplioresis: Spot itlentification hy N-terminal sequencing and iininiinolhtting. E/cc.fn)~,/ion,sis 17, 1&5- 190. Blackwell, J. M..Barton, C . H., \l%ite, J. K., Roucli, T. I., S11;iw. M. A., U'hiteheatl. S. H., Mock, B. A.. Searlc, S.. M'illians, H., ant1 Bakcsr, A. hl. (1994). Genetic regulation of leislimanial and inycolxicteiiiil infections: The Zd.drlZt!/lHcggene stoiv continiirs. Z t t t f r i r m d . Lf~lt.43,9:)- 107. Blackwell, J. M.. Barton, C:. H.. LC'liite. J . K.,S(wle. S.,Baker, A. M., \Villiams, H., and Sliaiv. M. A. (1995). Genoinic organization m t l sequeiice of the hiiinan NRAMP gene: Itlrntificntion antl niapping ol' a proinotrr region I)oIy~~i~)~phism. MoL. M c d 1, 194-2045. Blackwell, J. M..Black. C.F., Pcacock, C.,Miller, E. N., Sil)thorpe, D., Gnananandha, D.. S1i;iw. J. J.. Sihvira, F.. Liiis-Lainson. %.. Kanmr, 17.. Collins. A., and Shaw, M. A. (1997). Iminunogriietics of leisliinaiiial and iiiycohactvrid infections: The Belein Family Study. Plril. Tt-oti,T.Roy, Soc. Lottdori SPn H. Riol. Sci. 352, 1:331-1345. Blaiik, C., Fiichs. 11.. RappersI,erger, I;..Rollinghoff, M.. arid Moll. H . (1993).Parasitisin of epidrrnial 1,angerhans cells i n esperiniental ciit;ineous Icishmaiiiasis with Leislirtimici ttiujor. /. Zrtfi!ct. Dis. 167, 318-425. Blocker, A , . Severin, F. F., Biirkliardt, J . K.-Bingliam, J . B., Yu, H., Olivo, J. C., Scliroer, T. A.. II!wi;in, A. A , , and (:iiffiths. C;. [ 1997). blolecular requireincvits for bi-directional nioveincnt of phagosomes aloiig niicrotul)iiles. /. Ccdl Riol. 137, 11:3-129. Blockrr. A , , Sweriii, F. F.. Ilaberniaiiii. A , , H y i a n , A. A,, C;riffiths. G., antl Burkliardt, J. K. ( 1 996). Microtu~)ule-~issoeiate(1. I,rotein-tlependrnt I)inding of phiigosonies to rnicrotuhiilcs. J . Hiol. C/it,ttt. 271, 3803-381 1. B o n i a i i , H. C;, (1996).I'eptitle xtti1iotic.s: Holy or hcrcstic grails of innntt~iniiiiiiniv. S c m d J . Z t t i r t i u t u ) / . 43,47#5-482. Bonato. \'. I,.. Lima, \'. M., Tuscon, R. E.. Lowrie. I). H., i t r i d Silva, C. L. (1998).Identificatioii antl cli:iructl.riz;ition o f protecti\,c T cells in hspfij DNA-vaccinated and il~!/cohmctc,riritt~ tri/~c~n.rrlo.ri.s-inf~~cted niicc. Zrifict. Z j t i t t t r r t t . 66, 16% 175. +
340
ULHICH E S(:IlAII3LE
Pf
o/
Bonecchi, R., Bianchi, G., Bordignon, P. P., D'Ambrosio, D. A,, Lang, R., Borsatti. A., Sozzani, S., Allavena, P., Gray, P. A,, Mantovani, A., and Sinigaglia, F. (1998). Differential expression of chernokine receptors and chernotactic responsiveness of type 1 T helper cells (Thls) and Th2s. /. Ex/'. Men. 187, 129- 134. Boring, L., Gosling, J., Chensue, S. W., Kunkel, S. L., Farese, R. V., Jr., Broxmeyer, H. E., and Charo, I. F. (1997). Iinpaired inonocyte migration and reduced type 1 (Thl) cytokine responses in C-C cheinokine receptor 2 knockout mice.]. Clin. Znoest. 100,ZSSZ2561. Borremaiis, M., de Wit, L.. Volckaert, G., Ooms, J., De Bruyn, J,, Huygen, K., van Vooren. J,-P,,Stelandre, M., Verhofstadt, R., and Content, J. (1989). Cloning, sequence detennination and expression of a 32-kilodalton protein gene of Mycobncteriuin tuberculosis. Infect. Itntnuri. 57, 3123-3130. Bouwer, H. G., Seainan, M. S., Forrnan. J., and Hinrichs, D. J. (1997). MHC class Ihrestricted cells contribute to antilisterial immunity: Evidence for Qa-lh as a key restricting eleinent for Listeriri-specificCTLs. ]. Ininiuiiol. 159, 2795-2801. Brachet, V., Raposo, G., Aniigorena, S., and Mellman, I. (1997). Ii chain controls the transport of major histocoinpatihility complex class I1 molecules to and from lysosomes. /. Cell Biol. 137, 51-65 Brandt, L., Oettinger, T., Holm, A,, Andersen, A. B., and Andersen, P. (1996). Key epitopes on the ESAT-6 antigen recognized in mice during the recall of protective immunity to Mycobacteriuiii t u b e r a h i s . 1. Immunol. 157, 3527-3533. Brossay, L., Jullien, D., Cardell, S., Sydora, B. C., Burdin, N., Modlin, R. L., and Kronenberg, M. (1997). Mouse CD1 is mainly expressed on heinopoietic-derived cells. I. Ininiunol. 159, 1216-1224. Brossay, L., Tangri, S., Bix, M., Cardell, S., Locksley, R., and Kronenherg, M. (1998). Mouse CD1 autoreactive T cells have diverse patterns of reactivity to CD1' targets. /. Zinmunol. 160, 3681-3688. Broucpi, P., and Raoult, D. (1993). Proteinase K-sensitive and filterable phagosomelysosome fusion inhibiting factor in Afipici felis. Microb. Prithogen. 15, 187-195. Brown, D. R., Fowell, D. J., Corry, D. B., \Vynn, T. A,, Moskowitz, N . H., Cheever, A. W., Locksley, R. M., and Reiner. S. L. (1996). Beta 2-microglobulin-dependent NKI. 1' Tcells are not essential for Thelpercell2 innmine responses.]. Exp. Med. 184,1295-1304. Brown, E. J. (1991). Complement receptors and phagocytosis. Curr. @in. Immzinol. 3, 76-83. Brutkiewicz, R. R., Bennink, J. R., Yewdell, J. W., and Bendelac, A. (1995). TAP-independent, beta 2-microglobulin-dependent surface expression of functional mouse CD1.1. J . Exp. Med 182, 1913-1919. Buchmeier, N.. Bossie, S., Chen, C., Fang, F. C., Guiney, D. G., and Libby, S. J. (1997). SlyA. a transcriptional regulator of Salnumelln t!yp!iimtiriziin, is required for resistance to oxidative stress and is expressed in the intracellular environment of macrophages. Ir$ect. Zintnun. 65, 3725-3730. Buchmeier, N. A,, and Heffron, F. (1990). Induction of Salmonella stress proteins upon infection of inacrophages. Science 248, 730-732. Buchmeier, N. A,, and Heffron, F. (1991). Inhibition of inacrophage phagosoine-lysosome fusion by Srilnu~nellatyphimurium. Infect. lintnun. 59, 2232-2238. Buchmeier, N. A,, and Schreiber, R. D. (1985). Requirement of endogenous interferon-y production for resolution of Listerin numocytogenes infection. Proc. Nut. Acnd. Sci. USA 82, 7404-7408. Brims-Keliher, L., Nickerson, C. A,, Morrow, B., and Curtiss, R., 111 (1998). Cell-specific proteins synthesized by Snlnwnelln typhimurium. Infect. Zttiinun. 66, 856-861.
B ~ ~ s c D. h , H. , Hoiiwer. H. G., Hinrichs, I)..and Panier. E. C . (1997. A nonamer peptide derived from Listwio itiotioclltc)jiotcs irietalloprotease is presented to cytoljic T Iymplioc)tes. Znfict. Ztttitiuti. 65, Biisch. D. I-I.. Pilip, I. M., el l'mner. E. G. (1098).Coortlinate regulation of coinples T cell populations responding to hacterial infection. Ziiirtiuitity 8, Hyrd, T. F.. and Horwitz, M. A. (1989).Iiitrrferoii-y-~icti\~itetl human monocytes downregulate transferrin receptors and inhiliit the- intracc~ll~t1;tr multiplication of Lcgioiiello ptir, E. S.. :111(lS\V;~IISOII. X I , S, (ISJSJ5). I,gttl, it g ~ i tlr~tt e ~ I v ~ ( ~ I - I I Is~~I ~I (s~c\~ ~ ) t i l )to i l iI,(~giotr(~//(i t! /)rr~~~~ttio/)/ri/(i. 11);11)\to I I I O I I \ C , c l r r o ~ i ~ o s o1:3. l ~ ~Cc,trc~ttric~~ t~ 26, 4-1:3--1.50. l)i~~:uler, 11, C,, ll(~cl it! J . Ittittirttiol. 158, 5.781 -5.58:3. illfi,ctioll \\it11 1,istc~t-irltlrottoc~!/logc~t~c~v. 1)oerflc.r. \\'.. Sclr~~l)l)t~rt, H.. l l ( ~ I l ( 11.. ~ , Kam1lrc.r. (:.. Ililgrr-E\-~rshc.i~~i. K.. KIIoI)I;LII~II. l.1.. ;ultl l l c % ~ r11. ~ ~(1997). ~ s , Illt(-gr:ttio~rol'li)r('ig~~ I I S h ;111(l its c011~(~(~11(~11c(~5 ill I I I ; ~ I I I I I I ; L ~ ~ ; ~ I I s! stc21irs.I'tr~tr(lvAiol(~(~lrtrol. 15, 297L:iOl I)oIrert!. T. XI.. ant1 Sller. A. i 1997). Ilc~k~cts i ~ cc~ll-lrrrtli;ttvtI r i~irlrr~~l~it!al'kxct clrrolric. I ~ l t ~ i o innate. t resista~lcc,ol'~rricc> to J l ! / c ~ o l ) t r c ~ i ~ ~ t - o~.iritti i t ~ t t i iirli~ctio~r.]. Itttttrrttrol. 158, 48"-1% I . L ~ ~ I I I ; IE.. I I I\I\.' r l ~ l i ~ ~J,. i ( lRoli(l(,, . 11.. l'istor, S., 11:trtl. 51.. (;oc-l)(,l.\\... I~(~i~rrt~istt~r-\\';~cl~t(~r, kl.. ;LIICI(:l~;tkral~ort!~. T, ( lS)LI?). :\ I I O \ (,I l);tctc>ri;tl\-irr~Ic~ticc~ gc.~r(~ i l i I,i~t(,t-i(~ rtrotto(~!//og(~~r(~~\ r ( ~ ( ~ ~ ~ti)r i r (11ost ' ( l c~11~~ricro(il~lri('~rt i~~tc.ritctio~~ \\it1111o11101oq to t l ~ l>roli~rc>-ricl~ r r(3gio11 of' \-i~ic~~liir. EJIHO 1. 11, 1981- 1990. Do\\?lil~g,j. I;.. l';is111;1. 11.. \\'rigIrt. J . R.. T\\.i$g, 11. L. 11,. ; I I I ~k1arti11,\\'. 1. 11. (199.5). S~~rlactalrt protc.ilr A pro~rrotc.~ attachmrlit of' Jl!/c~o/)n(~/c,t-i~itti ftr/)c,t-c~rl/o\i.v to ;t1\(~1l:u~riacro~)lragcs tl~lri~rg i~rliactio~l \\-it11lr111ili111i ~ ~ ~ i r i ~ ~ ~ ~ o d\irlls. c . f i e1'1-oc.. i ( ~ ~S(I/. ~ c \ I\(.(id.Sci. rS:\ 92, 4848-48,52. Drxkc~,J. R., \I.cll)stpr. I'., (;a~r~l)ic,r, j. (;.. ;urtl hlrllrr~all.I. i 1997). I)rli\.en 01'13 cc.ll rcxcc.pto~-intel-~rulizrtla11tigc.11to c~lrt1osolnt.s ;urtl cl;~sr11 \c~sicl(.s.J . F:.vl). Jl(,t/. 186, 1.799-1:306. L)'Sot~xa, (;,. (;ool~vr.,,\. XI., F r ; t ~ ~ k 11. , ,I., X1~1~~~1cc;lro. 11. 1.. l3loo111.13. 11.. ;111(l0r111c.. I. XI. (1997). AII a ~ r t i - i ~ r l l a ~ r ~ ~ i rrole ; ~ t ofill~ ? -y6 ?' I!.~r~~>lroc\.te i l l ;~cc~~~irc*tl ilrrlrr~~liit\to J l ! j ( ~ ~ ~ / ) ~ i ( ~ t ( , t - t/~/)(,t-(,ri/o.~i.~. irittr J . Ittitt~t~rr(~/. 158, 1217-1 22 I. D'So~lz;t.S..Lr\ill. XI., Faith. ;\., Ysst~I,11.. 13c.1111c.tt. 13.. I,;tkr, K. A,. R r o \ v ~ ~I . . N.. alicl L:tlril). J . 11. (ID$J(i).Ilef&cti\.cx ;ultigc>~rp~-occ.\silrg;i\soc~i;~tc~cl \\.it11 lb~r~ilial tlissc~trril~atrcl 103, :35-:39. 111)col):~ct(~rio~i5, Cli~r.F:.v/). Zttttt~~i~roI. 1111lrn,K. I\.., aiitl Xl;t\firltl. 1;. 13. (199". I)c.li\c.~? of lig;u~tlsli.o~rrsorting c~1rtloso1i1c.s to of sortiirg erltlosolr~es.1. C d l Biol. 117, :301:310. lato c-nclosonlr~so c c ~ ~ I)!r s ~iiat~~ratiorr 111111n,\\'. A. (1994). ;\~~tol)lr;tp a~itlrel;ttecl ~~~c~clr;ulis~rrs of ~ ~ s o s o r r r c ~ - ~ ~ r c cp~r io; t~ct ~c ti~~ ~ dc.gfittlutio~~. Tt-(,trt/.sCcll Uiol. 4, 139-11:3. ~ I I I I I I I11. ~ ' \Y., . 11cs1rick. 11.. (;r('(xlll)(~rg,J.. Kric'gt'r, XI.. 1111c1 Joili(sr, ti. I\. (lU!J1). 1'11(' t\l>c2I ~ i i a c r o ~ ~ l r sca\c2irgc.r agt~ r c ~ c ~ p t oI~intls r to gfi~~~~-positi\.c. I)actc~ri;ta~itlrctcogni~ca\ lipotc~iclroici~citl.1'1-oc.. .\-of. i\c.nrl. S(,i. 1'SA 91, 1S ( i 3 1867. Elllvrs. XI. I{. \\. .. :t11(1 I);lrf?. 11. (I~\\TY,II Jl!~~~(~/)~i~~/(~t-ir~tti i/i/)('t.(,/l/o.\i.\ 6, :328-:3:3,5. :urd Irost cplls: Are I I ~ ! col):tctcxri;~lsltg;lrs thy kc,\ ;' Zt-(,rrt/v Jlic~tr)l~iol, Elrrt, S,. Slrilol~.k1, U,. ~ ~ I I L ~J.,I IClloi. , \ I . , ( ; I I I I L ~ ) L I ~S~., ~S ; I ~ ~ I(;.. ; IXi(,, I I , Q..; I I I ~11iIc,!.. I,. \I7. (1997). :\ no\.rl a~~tio\itla~rt gc5lrc2I'rolrr .\l!/c~o/~c~c~tc.t-i~i~ti trihc'r-(.rrlo~i~ 1. t.xp. .\I(,(/. 186, 1885-1896. Elom;~u,0.. K;tl~g;ts.1.1.. Salrll)c.rg. (;.. T I I I I ~ ~ ; ~J., I I S~~>~I I I. I I I I K I (. ,' ILii~kkil. I. A , , Tlrc~slvl'l'. 1.. Kraal, (:.. ;II~(IT ~ ? . g p a s o ~K.~ .(1995). (;lollil~gof' ;L ~ro\c,lI),~ctc~ria-l)i~rcli~rg reccJptor str~rct~lr;tll\. rc%l;~tc~l to sc.avrl1gc.r rrcc2ptors alrtl c,\;prcas\c~li l l s111)sc~t of' ~rr;~c~.o~)l~agc.\. C,'c,ll 80, 6():3-609. ~ I II- I,i.~tc,~-ici r~ro~roc.~/25, :3:321-:3:325. /ogr.trcJ.u.Errr 1. It~rt~rrttrol.
Engering, A. J.. Cella, M.. Flnitsina. D., Brockhaus, M.. Hoefslnit, E. C., Lanzavecchia. A,. and Pieters, J. (1997). The inannose receptor fiinctions as a high capacity and broad specificity antigen receptor in human dendritic cells. Eur. /. Irnmunol. 27, 2417-2425. English, B. I(.. Patrick, C. C., Orlicek, S. L., McCordic, R., and Shenep. J. L. (1996). Lipoteichoic acid from viridans streptococci induces the production of trnnor necrosis t"ictor and nitric oxide by innrine macropliages. /. Znjkct. Dis. 174, 1348-1351. Enonioto. A,. Nishimiira, II., and Yoshikai, Y. (1997). Predominant appearance of NK1.l' T cells prodircing IL-4 may be involved in thc. increased susceptibility of mice with the beige mutation tlrning Salmonella infection. J . Z t t m w t d 158, 2268-2275. Epstein, J., Eichl,aiuii, Q., Sherrif, S., and Ezekowitz, R. A . (1996).The collectins in innate iniinunity. Crrrr. O p h Zmmund 8, 29-35. Ernst, J. D. (1998). Macrophage receptors for I\lycohnctcririm tuhurcukosi.~.Zt$ect. Ztnrfum 66, 1277-1281. Ernst, W. A,, Maher, J., Cho, S.,Niazi, K. R., Chatterjee, D., Moody. D. H., Besra, G. S., Watanalie, Y., Jensen, P. E., Porcelli, S., Kr~inenberg,M., and Modlin, R. L. (1998). Molecnlar interaction of C D l b with lipoglycan antigens. Irnrrurnity 8, 331-340. Exley, M., Garcia, J., Balk, S. P., and Porcelli, S. (1997).Requirements for CDld recognition by hinnan invariant Va24'CD4-CD8- T cells. J. Exp. Med. 186, 109-120. Ezekowitz, R. A,, Sastry, K., Bailly, P., and W;irner, A. (1990).Molecular characterization of the human macrophage inannose receptor: Demonstration of multiple carbohydrate recognition-likedomains and phagocytosis of yeasts in Cos-1 ccl1s.J. Exp. Med. 172,17851794. Fang, F. C. (1997). Perspectives scries: Host/pathogen interactions. Mechanisms of nitric oxide-related antimicrobial activity. /. Clin. Innest. 99, 2818-2825. Fazal, N. (1997). The role of reactive oxygen species (HOS) in the effector mechanisms of n . B i d . Znt. 43, 399-408. hunian ~intiinycoliacterialiminunity. B i o c h ~ ~Mol. Fearon, D. T., and Locksley, R. M. (1996). The instnictive role of innate ininiunity in the 272, 50-53. acquired immune response. Sci cr~c~ Feltqnate. D.M., Heaney, S.,Webster. R. G., arid Robinson, H. L. (1997).Different Thelper cell types and antibody isotypes generated hy saline and gene gun DNA immunization. /. ~ t f 2 ? f L l t f l ~ J158, l. 2278-2284. Ferniri, C., Knight. A. M., Watts. C., and Picters, J. (1997). Distinct intracellular compartments involved in invariant chain degradation and antigenic peptide loatling of major histocompatiliility complex (MHC) class I1 niolecules. /. Cell Biol. 139, 1433-1446. Fine, P. E. (1995). Variation in protection by BCG: Implications of and for Iieterologous iniinunity. Lnricet 346, 1339-1315. Fineschi, B., ;mtl Miller, J. (1997). Entlosoitial proteases and antigen processing. Trenrls BiucJmi. Sci. 22, 377-382. Finlay. €3. B., and Cossart, P. (1997). Exploitation of mammalian host cell fnnctions by bacterial pathogens. Science 276, 718-725. Finlay, B. B., and Falkow, S. (1997). Coininon themes i n niicrobial pathogenicity revisited. Microbial. &lo/.B i d . Reo. 61, 136-169. Fischer Lindahl, K. (1997). On naming 132 haplotypes: Functional significance of MHC class Ib alleles. I ~ , i , , z u t i ~ ~ ~46, e ~53-62. e~~c.~ Fleming, M. D.. Trenor, C. C., 111, SII, M. A,, Foernzler, D., Beier, D. R., Dietrich, W. F., and Antlrews, N. C. (1997). Microcytic anaemia mice have a mutation in Nramp2, a candidate iron transporter gene. Nnture Genet. 16, 383-386. Flcsch, I. E., Hess, J. I f . . Huang, S.. Aguet, M., Rothe, J.. Bluethniann, I i . , and Kaufirianii, S. H. E. (1995).Early interleukin 12 production by macrophages in response to mycobacte-
rial infection clepeiitls 0 1 1 iritc4i.ron-y and tuiiiorriwrosis f'ktor a , / .Exp. Jlcd 181, 1615-
1621. Flescli, I. E., and Kaiil'inann, S. 1 I . E. (1988). Attempts to characterize tlic mechmisms iiivolveti in iiiycobacterial gro\vth inliibi tion b\-y-iiiterferon-ucti\,ntcd Imir marrow niiicrophages. Ztfict. Z t t i t t i i i t ~ ,56, 1464-1469. Flesch, 1. E., and Kaiifniaiirr, S. H. E. (1990;~). Stiniiilation of antibacterial macrophage acti\,ities l)y B-ccll stiiniilatory f k t o r 2 (interleukin-6). Ztfict. Z i t i t t u t t t . 58, 269-271. Flescli, 1. E., \l'aiidersee, A,, and Karifninnri, S. H. E. (2997). IL-4 secretioii by CD4' NKI ' T cells indiicc~sinonoc!te clieiiioattractairt proteiii-1 in early listvriosis. /. Z i t u i u i t i o / .
159, 7-10. Flesclr, I. E. A,, und Kaiifiiiaiiii, S. I I . E. (l990b).Activation oftulierciilostatic microphage functions liy iiiterfhii-y, interleukill 4 and tninor necrosis fiictor. Zrfect. Z t ~ i t t i i i t i . 58,
267.5-2677. Flescli, 1. F,. A,, aiid Kaufinann, S. 11. E. (1991).'vlcchanisins involved in iiiycobactcsl-ial growth inhibitiori Iiy y-iiitcrli.ron-~icti\-atrtl l ~ o i i c . inarro\v macrophages: I>,Saliiioiic,lla arid other stimuli dirpct macropinoc>tosis ofliacteria. Natri rt' 364,639-642. Freliel, C., tle Cliastellier, C,, I,ang, T., and Rastogi, N. (1986). Evidence for inhibition of fusion .' of Iysosoinal and prelysosotiial coiiipartineiits \\it11 phagosoiiies in macrophages infected wit11 pathogenic ,\Z!/cobacterinrti ocziLiti. Zt,fecf. Z t i ~ t t i r ~ n52, . 252-262. Friitli, U., Solioz, I., and Louis, J . A. (1993). L i t i o t i i o ttiujor interfercs \tit11 antigen prcsentatioii by irifectcd niacrophages. /. Z t t i t t i i i t i t d . 150, 1857-1861. FII, Y.-X., Roark, C . E., Kelly, K.. Drcvets, I]., Caiiiphell, P., O'Hrien, R., and Horn, M'. (1994). Immune protection aiid control of inHainmaton tissue necrosis l)y y6 T cc.lls.
/. Z n i ~ t ~ i ~ i r 153, o/. 3101-31 15.
Fujise, S., Kisliiliafii, K., Lee, K. Y.,Matsii7aki, G., and Nonioto, K. (1997). Normal micropliagct functions, but iiripaircd induction of y6 T cc~lls,at tlie site of bacterial infection in CD45 exon 6-deficierit micc,. Erir, 1.Z t t i t t t r L i i o l . 27, 2549-2556.
348
U L H I C I I E SCII.4IULE et
NI
Fujiwara, N. (1997).Distrihiition of antigenic glycohpids among Mycubacterizrm tuberc\lthsis strains and their contribution to virulence. Ktrkknkti 72, 193-205. Gallin, J. E., Pace, J., and Hayman, M. J. (1992). Involvement of the epidermal growth factor receptor in the invasion of cultiired inanimalian cells by S U ~ T I Utyphinmritori. ) ~ ~ N(itrtrc 357, 588-589. Gain, T.. and Weiss, J . ( 1997). Antimicrobial peptides of phagocytes and epithelia. Semiti. Hc,ttuitol. 34, 34:3-354. Garcia-del Portillo, F. (1996). Interaction of Salnionella with lysosomes of eukaryotic cells. Micruhiolugia 12, 259-266. Garcia-del Portillo. F.. and Finlay, B. B. (1995). Targeting of S a h u r t e h t y p h i ~ t i ~ i r i r ~ t t i to vesicles containing lysosoriial iiiembrane glycoproteins b p i sses compartmeiits with maiiiiose 6-phosphate receptors. 1. Cell B i d . 129, 81-97. Garcia, V. E., Sieling, P. A,, Gong, J., Barnes, P. F., Uyeinura, K.. Tanaka. Y., Bloom, B. R., Morita, C. T., ; i d Modliri, R . L. (1997). Single-cell cytokine andysis of$ T cell responses to iionpeptide mycobacterial antigens. 1. I t r i m i n u l , 159, 1328-1335. Gircia Vescovi, E., Soiiciiii, F., arid Groisman, E. A. (1996).Mg2' as an extracellular signal: environiiiental regulation of Salmonella virulence. Cell 84, 165-174. Guuse, W. C . , Ifalvorson, M. J., Lu, P., Greenwald. R., Linsley, P., Urban, J. F., and Finkelman, F. D. (1997). The fiinction of costimulatory inolecules and the development of IL-4-pro~l11cing T cells. Ittimzctcol. Torlay 18, 115-120. Gaynor, C. D., McCorinack. F. X., Voelker, D. R., McGowan, S. E., and Schlesinger, L. S. ( 1995).Pii1inoiiar-ysurfactant protein A mediates enhanced phagocytosis of Mycubacteririrtr tribcr~:rilosisby a direct interaction with human macrophages 1. Irtzmunol. 155, 533433-so5 1. c;uziiie11i, .. ' R. T., Wysocka, M.. Hieny, S., Scharton-Kersteri, T., Cheever, A,, Kuhii, R., Miiller, W., Trinchieri, G., and Slier, A. (1996). 111 the absence of endogenous IL-10, mice acutely infected with Toxojhwtri gotiriii succumb to a lethal immune response dependelit on CD4' T cells and accompanied by overproduction of IL-12, IFN-7 and TNF-a. J. l t m w m ~157, . 798-805. Geliran, S. J., Yamamoto, Y.. Newton, C., Toinioka, M., Widen, R., Klein, T., arid Friedman, H. (1995).LPS inhibits the intracellular growth of Legioiiella pneunwphiln in thioglycollate elicited muriiie peritoneal macrophages by iron-dependent, tryptophan-independent, ouygen-iridependentand arginine-indepc,ndent mechanisms.]. Leukocyte Biol. 57,80-87. Grluk, A.. van Meijgaarden, K. E., de Vries, R. R., Sette, A,, and Ottenhoff, T. 14. (1997). A DRl7-restricted T cell epitope from a secreted Mycobmterizrm tuberculosis antigen only binds to DR17 molecules at neutral 1'1%. Eur. J. Iriiitzunol. 27, 842-847. Gerckm, J., Pryjiiia, J . , Emst, M., :uid Flad, H. D. (1994). Defective antigen presentation by Mycubactet-iurtr ttrberciilosis-infected monocytes. Infect. b n t t i t i t i . 62, 3472-3478. Gervais, F., Desforges, C., and Skamme, E. ( 1989).The C5-sufficient A/]congenic mouse strain. Inflammato~response and resistancc: to Listeria ttwnocytogenes. 1. Ininiunol. 142, 2057-2060. Gllosal, A. C., Maiidal, T. K., Chowdhury, T. K., Mukherjee, K., and Mukherjee, D. (1996). IlLA phenotype frequencies in pulmonary tuberculosis. 1.Indian Med. Associut. 94, 328-330. Ghosh, R. N., Gelman, D. L., anti Mawfield, F. R. (1994). Quantification of low density lipoprotein and trarisferrin endocytic sorting in HEPp2 cells using confocal microscopy. J . CcN Sci. 107, 217772189, Ginsberg, A. M. ( 1998).Tu1)erculosisvaccine development. In The Jordan Report: Accelerated Development of Vaccines, pp. 75-79. Division of Microbiology and Infectious Diseases. N IH .
INTRACE L.1 .UI.AH H A C I E H I A A N D 'I-} 1 E I M M UiiE SYSTEM
349
Gobin, J., and Ilonvitz. M. A. (1996). Exocheliiis of Mycobcrctc,n'rr,,r tubercrtlosis reiiiove iron from Iiuinan iron-bincling proteins arid tlnnatr iron to mycol)actiiisin the ,%f.t~il)e~rcdo.si.s cell wall. I. Exp. A f c d 183, 1527-lS32. Cobin, J.. Moore. C. 11.. R r ~ w e ,J. B., Jr., \Vong, 11. K., Gi1,son. R. W..a d Honvitz, M . A. ( 1995). Iron acquisition 1)y A f ! i c , o ~ / r c ~ c , t ~frhc.rrri/osis: ftrti isolation and cliaracterization of a family of iron-binding exoclieliiis. Proc. Nut. Acmd. Sci. U S A 92, 5389-5193. Goldberg, A. Id., and Rock, K. L. (1992). Proteol proteasonws and antigen presentation. Nnturc, 357, 375-:379. Goklfield. A. E., Drlgado. J. C., Tliiin, S., Bozon, M. V., Ugliaro, A. M . . Turbay, D., Cohen. C., and Yunis, E. J. (1998). Association of im IILA-DQ allele with clinical tulwrculosis. /. h i , hfed. Associnf. 279, 226-2228. Goltlstein, J. L., Brown, M . S., Antkrson. R. G.-Rrissell. D. \i;,, and Schneider, \V. J. (1985). Recc.ptor-niediatet1 rndocytosis: Coiicrpts eiiirrging from tlie LDL receptor system. Aritm H m . Cell Riol. 1, 1-39, Gordon. A. €1 , D'Arcy Hart. I'., and Young, M . R. (19280). Ainiiionia inhibits phagosoirielysosoine fusion in niacropliages. NrrfitrrJ 286, 79-80, G o r h m . J. D . Ciiler, M . I,.. and Muipliy, K. M . (1997). Geiictic control of interleuhinJ. AfOl. h f d . 75, 502-51 I . Irriplications for disease patliogeiie f h h a n i , J. D., Guler, M. I,., Stccw, I{. (ttjis ;ind for its ahsence in A~!/cohortPrit~tt~ h i s . BCC;. Infict. Ittwmtt. 64,
16-22. Harding, C. V., Hoof, R. W., Allen, P. M., and Unanue, E. R. (1991). Effects of pH and polysaccliariclc,s on pepticlc binding to class 11 iiiajor Iiistoconipiitibilitycornplex inolecules. Proc. Not. Acnrl. Sci. USA 88, 2740-2744.
Hardiiig. C:. 1' , ;ind Song, K.( 1994). Pliagocytic processing of exogenous particulate antigens by niacrophages for prrsentotion by class I MHC niolrciiles.]. Z t t i t t w t m l . 153,49255493333, Harrison, J. A, \'illarreal-Rainos. B., Xlastroeni, P.. Ileiiiarco de Horniaeche, K., ant1 Morinaeche. C . E. (1997). C;orrehtcls of protection induced by live Aro- Snlttrottelln f!yp/iittiifriwti vaccines in the niurine t)phoitl iiiodel. Z r r t / t t i t i i o h g ! / 90, 618-625. Hnrrison. P. M., and Arosio, P. (1Y96). The ferritiris: Molecular properties, iron storage . 1275, 1631-20:3. fiinction ;in11crlliilar rrgulation. Biocltitrc. f i i o ~ j / r ! y ~.-\ctc/ ILii-tli, G., Clcinens, D. L.. ; i d Honvitz,. M.A. ( 1994).(;lutan>ine synthetase ofhfgc:oboctcrittrtc frrbrtu.tr(rais: Extracellii!ar release. antl c1i;iracterization oi'its enzyinatic activiiy. Proc. Mot. Atatl. Sci. U S A 91, 9:342-9346. Kirt!., J . T.. and Bevan. M. J , (1992). CD8' T cells sprcific for u single nonanier epitope of Iistrv-in ttiottoc.!/togcne,sLire protecti\.c. in vivo. J . E.Y/I..!fee/. 175, lfj-31-1538. Hart\.. J . T., antl Bevan. M. J . (1996).CD8T cell recognition ot'niacrophages and hepatoeytes rcwilts i n ininiunitv to Listetirr tiiotioc'!ytogr,,Ics I t f w t . Z t t i t t i t t t i . 64, 3632-336340, Hart\.. J. T., L ~ n z L. . L.. and Bevan, M. J. ( 19963).l'rinian, and secondary immune responses to Z,isfctic/ ttiotiocgtog[~ti[,.s.Ctrrr. Opiti. Itiitiiutiol. 8, 5%-5:30. Hasan, Z., Scltl~u,C . , K i i l i n , L., L e f h i t s . I., Yoiiiig, I),, Thoks. J., and Pirters, J. (1997). 1sol:ttion antl clianictrrizatiori of tlie niycohacteriul phagosonie: Segregation froin tlit. endosoitia~l~sosoi~i~il pathway. Mol. ,!ficwbiol. 24, 545-55:3. Haslov, K., Antlersen. A,, Nagai, S . , Gottschau, A , . Sormsen, T., and Andersen, P. (1995). Guinea pig celliilar inimnne responses to proteins sccreted by hfycobact~~riiitrt tidwrcihsis Zt!fr.ct. Z t t t t t i i u t . 63, 804-810. Havcll, E. A,, ;tnd Sehgal, P. B. (1991).Tiiiiior necrosis thctor-indrpendent IL-6 production thiring m i i r i n c Iisteriosis. J . I t r u r i r o i o l . 146, 756-761. Hayishi, T., Catanztro, A , , ~ n dI k . S. P. ( 1997). Apoptosis of human inonocytes and inacroplinge< by Mycobnrfcrirr t t i uoircttr sonic;itc,. Ztfict. Zttcttucti, 65, 5262-5271. Heeg. K . . Knon. M'.. and Wagner, II. (1991 ). Vaccinatioii of cliiss I major liist[icoinpatihili~ compleu ( !vI HC-restricted niurine CDX' qtotoxic T lynipliocytes towards soluble antigems: Ininiiiiiostiniulatiiig-ovalbuminconiplrws enter the c l ~ Is MHC-restricted antigcw patliway antl allo\v sensitization against the i i i i i i i i i r i c ) c l o i n i i i ~ i n t peptitle. Eur. J . I t t i t t i i i r i e d , 21, lC52:3-t527. IIeinzen, K. A., Scidmore, tvl. A,. Rocky, I). 11.. and IIackstadt, T. (1996). Differential interactioii \kith rndocytic and exocytic. pttli~vaysdistiiigiiisli parasitophorous vaciioles of Cosicllo but-rietii and Chkrtryrlio trocliotiwtis. Ztfict. Z j t i t t i u t i . 64, 796-809. Hensc.1, M., SlLea. J. E., Gleeson, C., Joiit's, M. D.. Ddton, E., and Holden, D. W. (1995). Siniiiltaneous identification of l)actc,rial \irnleiicr grnes by negative selection. Science 269, 400-403. Hess, J., Dirtrich. G . , C:entsclicv, I.. Miko, D., Goel)el. M'., and Kaufniann, S. H. E. (19%). Protcwtion against niririne listeriosis by air attcwuated recombinant Snbtiotir&r t ! / , d i i t t c t i t i ~ i t t i\.;icciiiestrain that secretes thr, iiatiirally somatic antigen superoxide clisniutasci. Zrficf. Z t r t t t i i c t t . 65, 128&1292. Hess, J.. Ckwtscliev, I., Miko, D., Welzel. M., Latlel, C., Goebel, W., arid Kaufiriaiin, S. H. E. ( lW6a). Siiperior efficacy ot'srcretetl over soiiiatic antigen display in recombinant S c d t t t o t t d l o \wcinr, induced protcvtion against listrriosis. Proc. N o t . Actrd. Sci. USA 93, 1458- 1463:3. Hess, J . , Latlel, C.. Miko. D.. and Kaufinanii, S. 1-1. E. (I996b).Sn/trionc7h f ! / p / i i t t u r r i u r t r aroAinfection in gene targeted immunodeficient mice: Maijor I-oleof CD4 ' TCR-a/3 cells and IFN-y i n 1)acterial clearance indepeiidriit of intracc-lliilar 1ocS;ition.J. Z t n t t m r i o L . 156,3321 :3:326.
Hess, J., Miko. D., Catic, A,, Lehmensiek, V.. Knssell, D. G., and Kanfmann, S. 11. E. (1998). Mycobocteritmt boois bacille Calmette-Guerin strains secreting Iisteriolysin of Listerio ttu)iiocytogetie.s. Proc. N o t . Acod Sci. USA 95, 5299-5304. Hewitt, E. W., Trenmann, A,, Morrice, N., Tatnell. P. J., Kay, J., and Watts, C. (1997).Natural processing sites for human cathepsin E and cxthepsin D in tetanus toxin: Iinp~icationsfor T cell epitope generation. /. I t i i t u u t w l . 159, 4693-4699. EIickey, E. K., and Cianciotto, N. P. (1997). A n iron- and fur-repressed Legionello ptiet’rtiop l i i k r gene that promotes intracellular infection and encodes a protein with similarity to the E.schcrichio ndi aerobiictin synthetases. Zrgect. Znin~un.65, 133- 143. Iliroinatsn, K., Yoshikai, Y., Matsuzaki, G., Ohga, S., Muratnori, K., Matsulrloto, K., Bluestone, J. A,, and Nomoto, K. (1992). A protective role of y6 T cells in primary infection with Listeria itumoc!/togetws in mice. J. Exp. Med. 175, 49-56. Hirsch, E., Itikura, V. M., Paul. S. M., and Hirsh, D. (1996). Functions of interleukin 1 receptor antagonist in gene knockout and overproducing mice. Proc. Not. Amd. Sci. USA 93, 11008-11013. Holniskov, U., Lawson, P.. Teisner, B., Tornoe. I., Willis, A. C., Morgan, C., Koch, C.. and Reid, K. B. (1997). Isolation and cliaracterization of a new niember of the scavenger receptor snperfamily, glycoprotein-340 (gp-340), a s a lung siirfactant protein-D binding niolecnle. J. Biol. Cheni. 272, 13743- 13749. Homutlr, M., Valentin-Weigand. P.. Rolide, M., and Gerlacli, G. F. (1998). Identification and characterization of a novel extracellular ferric reductase from M!ycobncteriuni pnratrcDrrculosis. Zrfect. Initnun. 66, 710-716. Hoppe, H. C., de Wet, B. J., Cywes, C., Daffe, M., and Elders, M. R. (1997). Identification of ~~I~os~~liatidylinositol mannoside BS H inycobxterial adhesin mediating both direct and opsoiiic binding to nonphagocytic mammalim cells. Zrfect. Z i t i t n u n . 65, 3896-3905. Hormaeche, C. E., Joysey, H. S . , Desilva, L., Izliur, M., and Stocker, B. A. (1991). Immunity conferred by Aro- Saltnonella live vaccines. Microb. Puthogen. 10, 149-158. Horuitz, M. A. (1983). The Legionnaires’ disease bacterium legion elk^ p n e u n q h i h ) inhibits pliagosoiiie-lysosoine fusion in Iinman tnonocytes.J. Exp. Alerl. 158,2108-21 26. IIotwitz, M. A. (1984). Phagocytosis of the Legionnaires’ disease bacterium (Legionelln ptieutiiophila) occurs by a novel niechanisin: engulfinent within a pseudopod coil. Cell 36, 27-33, Honvitz, M. A., Lee, B. W.. Dillon, B.J.. and Harth, G. (1995). Protective immunity against tuberculosis induced by vaccination with major extracellular proteins of A.iycob&riutn titbcrcdosis. Proc. Nut. Acad. Sci. USA 92, 1.530-1534. IIosieth, S. K., and Stocker, B. A. (1981). Aromatic-dependent Soltilotiella typhittiurinni are notivirulent and effective as live vaccines. Notiire 291, 238-239. IIsieh, B., Schrenzel. M. D., Mulvania, T.. Lepper, H. D., DiMolfetto-Landon, L., and Ferrick, D. A. (1996). In vivo cytokine production in inurine listeriosis: Evidence for itnmunoregulation by y6‘ T cells. J. Z t t i t w n d . 156, 232-237. Hsieh, C. S., Macatonia, S. E., Tripp, C. S., Wolf, S. F., O’Garra, A,, and Murphy, K. M. (1993). Development of TH1 CD4+ T cells through IL-12 produced by Listeria-induced macrophages. Science 260, 547-549. Hubbard, R. D.. Floty, C. M., and Collins, F. M. (1992). Immunization of mice with tnycobacterial culture filtrate proteins. Clin. Exp. Zniniuriol. 87, 94-98. Huller, L. A., and Peters, M. E. (1994). Mapping sinall GTP-binding proteins on high resolution two ditnensional gels by a coinbination of GTP binding and liibelling with in situ periodate oxidised GTP. Electrophoresis 15, 283-288. Huffman, G. R., Nataraj, C., and Kiirlander, R. J. (1996). Novel pathway for class I-restricted presentation of an exogenous, non peptide Listeria itiotiocytogenes antigen to CD8’ T cells. FASEB J. 10, A1309.
Ilunter. C. A.. Chizxonite. K..and Remiiigtoii, 1. S. (199.5). IL-1 /3 is required for IL-12 to induce production of IFN-y by NK cells: A role for IL-I /3 in the T cell-irid(~pendelit meclianisni o f resistance against intr;tcellul;ir pathogens. J . ~ l ? l t t l u t t ( J ~155, . 4:347-4:354. Hnnter. f:. A.. Tinims, J.. Pisacane, P.. Menon. S.. G i i , G.,Wdker, W..Astc-Anie.zaga, M . , Chizzonitr. R.. 13azan. J. F.. and Kastelein. R. A . (19971. C o n p i i s o n of the effects of interleukin- la. interleukin- 18 antl iiiterferoii-gaiiiina-intlticing factor on the protllwtion of iriterferon-jianiina by natural killer cells. Errr. J . Z t t i t r t t ~ t t ~ ~27, / . 2787-2792. 1Iuygr11,K.. Content. J.. Denis. 0 . .Montgonieiy, D. I,.. Yawman. A. M.. Deck, H. R., Deil'itt, C. M., Ornic,, I. M., Baldwin. S.. 13'Soiiza, C:.~D r w w t , A , , Lozrs, E., Vantlenl~usscl~e. P., Van Vooreii, J. P., Liu. M. A., and Uliiirr, J. B. (19%). Iiiiinuiiogc,nicitv :ind protcctive c4'ficacy of ii tuberciilosis DNA \wciiic.. Notirrc Mcd. 2 , 893-898. Igietsemr. J. IJ.,Perry. L. L., Ananaba. (;. A , . Uiiri, I . M., Ojior, 0 . 0..Kuinar, S. N., and Caldwrll. H . D. (1998). Clilaiiiydial infi,ctictn i i i inducil)le nitric oxide syithuse knockout inice. It!ff.c.i. Z t u t m t t i . 66, 128% 1286. Iniani, F., and Solosh. M . J. (1991). Heat sliock proteins ciin wgulatc expressioii of thr T1;i region-mcoded c l m Ib niolecule Qa-1, PIYJC. h'rrf.:\cvr/. Sci. C'SA 88, 10475-10479. Ireton, K . , ~ n i t Cossart. l P. (1997). IIost-puthogen iiitcwctions tltiiing ent? iincl actin-hascd nro\zwent of Lisfvrin tttotioc,!/togrric..v.Atttiri Rc~t:.Crttrt. 31, 1 13- 138. Isherg, IX' T cells reactive with ,Ilycnhncfrtirrtti ttrbc,rcrr/osi.~-infectetl antigc.ii presentiiig cells. J. Esp. A I d 187, 1633- 1640. Lin, J., antl Ficlit. T. A. (1995). Protein synthesis i n Bt7ccdk/ d ~ ~ r t iiidriced t~s during macrophage infection. 1tdk.t. 1 t t t t w t t . 63, 1409-1414. Liii, Y., Zhang M.. antl Barnes, P. F. ( 1998). Cheinokine production by a Iiiiman ulveolar epitlielial cell line i n respoiiw to ~ ~ I / ~ . ~ i / ~ ~ i ctrrbc~t-cirlo.sis. f ~ ~ t ~ i t t t i ih$ect. I t n t n i r n . 66, 1121 1126. Liiidahl. K. F.. Byers, 11.E., Dabhi. V. M.. Hovik. R.. Jones, E. P., Smith, C:. P., Wang, C . R.. Xiao, H., ant1 Yoshino. M. (1997).H2-M3. ii frill-senicr class I b histocoiiipati1)iIihi ~llltigell.L4rLtf11. &I;. f t t l t t t f l t l ( J / , 15, 851-879. 11 conipleues genenitcd Lindiier, R.. and Unaniie, E. R. (1996).Distinct antigen MllC b?. separiitc processing pathways. EAfBO J . 15, 6810-6920. Lipford, G. R.. Spiwasser. T., Bailer. M . , Zimrnermaiin. S., Koch, E. S., IIeeg, K.. iind L2'agnc.r. 11. (1997).Iinmriiiostiiniilatol-). DNA: srciiieiic(~-depericl[~nt prodiiction of potentially hariiifiil or useful qtokines. Etrr. J . Z t n t u r r t i d . 27, 3420-3426. Liu, J., Fiijiwara, T. M., Briii, N. T., Sanchez, F. O., C c h r , M., Paradis, A. J., Frappier. D., Skamenc., E., Gros, P.. Morgan. K.. et a[. (1995). I(1entification of polymorplrisnIs antl secpencr variiints in tlir Irurii;in Iioiiiologuc of the nioiise natural i-esistaircr,-associiitcd macrophage protein gene. Ant. /. Htrttt. GcttcJt.56, 845-853. Liu. Z., Simpson, R. J., and Cheers, C . ( 1994). Kole of IL-6 in activation of T cells for J . Z t t ~ t t w t o / . 152, 5375-53380. acquired celliilar resistance to Listerio ttiotioc!/togett~,~. Lobigs, M . , and Mullbacher. I\. (1993).The M I IC encotkt1 peptide transporters for class I MI IC: are lixnctioiiiilly not polyniorphic. P n x . Nnt. Auitl. Sei. USA 90, 2676-2680. Imckslry. R. M., antl Scott, P. (3991). Helper T-cell subscts in iiioiisc Ic~ishmaiiiasis:Induction, expansion and effi'ctor function. Z t l i t t u r t i d . Totkq 12, A58-61. I,oetscher, P.. Ugiiccioni, M.. 13ordoli, L., B;iggiolini, M., Moser, B.. Chizzolini, C., and Daym-. J. M. (1998).CCR.5 is cliaractc,ristic of Thl Iymphocytes. Nntrrrc 391, 344-345. Liihning. M.. Stroehmanii. A,. Coylc. A. J.. Grogoa, J. L., Lin, S.. Guttierez-Ramos, J.-C., I,e\inson. D.. Radbnich, A,, and Kiinirxlt. 1'.(1998).TlISTB is prelbntially expressed on iiiurinr TI12 cells independent of interleukin 4. interleiikin 5, and interleukin 10, antl inipoitant h r Th2 effector function. Proc. &'of/. Aced Sci. USA 95, 6930-6935. L,opcz-Macias, C., 1,opez-llrriiandez. M. A.. Gonzalez. C. R.. Isilxisi. A,. and OrtizNiwarrete. 1' (1995). lndiiction of iinti1)odies against Snhioticlln t y p k i OiiipC poiin by naked DNA ininirinization. Attti. N . 1..Arcid. Sci. 772, 285-288. Lri, 13.. Rutledqe. B. J., Gu. L.. Fiorilla. L.. Lukacs. N. b'.,K ~ n k e l .S. Id., North. R., C;er;ird, C.. d Rollins, B. J. ( I Y Y X ) . Abiiorinalities in inoiiocyte recruitment and cytokine exprcwion in nionocyte c1ieiiioattract;uit protein-1 deficient mice. j . Esp. Mcd. 187, 601-608.
358
ULRICH I.: S(:HAIBLE et 01.
Lu, J. (1997). Collectins: Collectors of microorganisms for the innate immune system. Bioessuys 19, 509-518. Lucey. D. R., Clerici, M., and Shearer, G. M. (1996).Type 1 and type 2 cytokine dysregulation in human infectioris, neoplastic, and inflanimatory diseases. Clirr. Microbid Her;. 9, 532-562. Luke, C. J., Carner, K., Liang, X., and Barlmur, A. G. (1997). An OspA-based DNA vaccine protects mice against infection with Borrelin h i t r g d o r j k J. It@. Dis. 175, 91-97. Macatonia, S. E.. Hosken, N . A,, Litton. M., j'ieira, P., Hsieh, C . S.. Culpepper, J. A., Wysocka, M., Trinchieri, G., Murphy, K. M., and O'Garra, A. (1995). Dendritic cells produce 11,-12 and direct the development of Tlil cells from naive CD4' T cells. J. Imtntmol. 154, 5071-5079. MacMicking, J. D., Nathan, C., Hom, G . , Chartrain, N . , Fletcher, D. S., Trumbauer, M., Stevens, K., Xie, Q. W., Sokol, K., Hutchinson, N., et id. (1995). Altered responses to bacterial infection and endotoxic shock in mice lacking inducible nitric oxide syithase. Cell 81, 641-650. MacMicking, J. D., North, R. J.. Lacourse, R., Mudgett, J. S., Shah. S. K., and Nathan, C. F. (1997).Identification of nitric oxide synthase as a protective locus against tuberculosis. Proc. Nnt. Acnd. Sci. USA 94, 5243-5248. Magram, J., Connaughton, S. E., Warlier, R. R., Carvajal, D. M., Wu, C. Y., Ferrante, J., Stewart, C., Sarmiento, U., Faherty, D. A,, and Gately, M. K. (1996).IL-12-deficient mice are defective in IFNy production and type 1 cytokine responses. Imintiriity 4, 471-481. Mahairas, G. G., Sabo, P. J., Hickey, M. K., Singh, D. C., and Stover, C. K. (199G).Molecular analysis of genetic differences between A4ycobncteriim bouis BCG and virulent A4. boois. J. Bacteriol. 178, 1274-1282. Maher, J. K., and Kronenberg, M. (1997).The role of CD1 molecules in immune responses to infection. Cum. @in. 'Znmunol. 9, 456-461. Majeed, M., Ernst, J. D., Magnusson, K. E., Kihlstrom. E., and Stendahl, 0.(1994). Selective translocation of annexins during intracellular redistribiition of Chlrmiyrh imclionmtis in HeLa and McCoy cells. Infect. I n m u t i . 62, 126-134. Malaviya, R., Twesten, N . J., Ross, E. A,, Abraham, S. N., and Pfeifer, J. D. (1996). Mast cells process bacterial antigens through a phagocytic route for class I MHC presentation to T cells. J. I n i t n i t n o l . 156, 1490-1496. Martinez-Kinader, B.. Lipford, G. B., Wagner, II., and Heeg, K. (1995). Sensitization of M I-IC class I-restricted T cells to exogenous proteins: Evidence for an alternative I-restricted antigen presentation pathway. Iminz~iiology86, 287-295. Mashiba, H., Matsunaga, K., Tomoda, H., Furusawa, M., Jinii, S., and Tokunaga, T. (1988). In vitro augmentation of natiiral killer activity of peripheral blood cells from cancer patients by a DNA fraction from M!/cobucterizrm bouis BCG. Japan.J . Med. Sci. Biol. 41, 197-202. Matsui, K., Yoshimoto, T., Tsutsui, H., Hyodo, Y., Ilayashi, N., Hiroishi, K., Kawada, N., Okamura, H., Nakanishi, K., and Higashino, K. (1997).Pt-[~?ii~iiibncterirrm ncnes treatment diminishes CD4' NK1.1' T cells hut induces type I T cells in the liver by induction of IL-12 and IL-18 prodiiction from Kupffer cells. J. In~munol. 159, 97- 106. Matsumoto, H., Suzuki, K., Tsuyuguchi, K., Tanaka, E., Amitani, R., Maeda, 4.,Yamamoto, K., Sasada, M., and Kuze, F. (1997). Interleukin-2 gene expression in human monocytederived rnacrophages stiniulated with Mycolmcteriuni bouis BCG: Cytokine regillation and effect of NK cells. Infect. Iminun. 65, 4405-4410. Maurelli, A. T., Baudry, B., d'Hauteville, H., Hale, T. L., and Sansonetti, P. J. (1985). Cloning of plasmid DNA sequences involved in invasion of HeLa cells by Shigella.flexneri. Infect. Z t n i w i i . 49, 164-171.
accaro, R. J.. Grtldr, M., Jensen, E. R., \.tin Sainten. I I. M.,Ploegh. Ei. L., Rock, 13looin. 13. R. ( 1996). Major Iiistocoinp~itil)ilityclass I presentation of soluble antigen fiicilitatetl hy dllleo/inrto-iirttt firl~c~tr~irlo.sis infection. h e . A\’rit. Aceid. sci. c’.$i\ 93, 117fi6-11;91. McAtlain. R. A,. \Veisl)rod, T. K.. Martin, J.. Scrideri, J. I).. Brown, A. M.. Ckillo, J. I)., H l o o n i , 13. It., and Jncobs. \I. It-2' T cells. Itfict. I i i i i ) i t i i i . 55, 20:37-2041. Muntler, M., M d o . M . , Eicliniann, K.. iuid Motlolell. M . (1998). Murine inacrophages secrete interferon gainina upon coml)ined stiindation with interleukin (IL)-12 and IL18: A novel p at h n y . of aiitocrine macrophage activation. 1.E q i . Alctl. 187, 2103-2108. Murray. P. J., Aldo\ini, A.. and Yoring, R. A. ( 1996). Manipulation aiitl potrntiation of aiitiinycohac.teria1 iiiimiiiiity usiiig recoin1,inant b a d e Calniette-Guerin strains that secrcte cytokines. Proc. Nut. Acrid. Sri. I'SA 93, 934-9:39. Murray. P. J. h'uig, L., Oniifryk. C.. Teppvr, R . I., ant1 Young, R. A. (1997). T cellderivrd IL- 10 antagonizes macrophage function in niycobacterial infection. ]. I t n i t i t t r i o / . 158, 3 15-32 1. N d en u . J. J.. 4rl~ickle,L. I)., d Sknmene, E. (1995).Genetic dissection of inflammatory res~'oI1"'s. 1. z t l f l n t i l l i l , 45, "-33. Nataraj, C.. Brown. M . L., Poston. R. M., S h w w , S. M., Rich, R. R., Lindahl, K. F., aiid Kiirlander, R. J . (1996). H2-MSwt-restricted, Zistorio riionocytoecnrs-specific CD 8 T cells recognize ii novel, Iiytlrophobic. proteasc,-resistant, pciiodate-sensitive antigen. Int. Z I 1 1 I t L I i 1 1 0 1 . 8, 367-378. Nathan. C. F., and Hihiis, J. B. Jr. (1991). Role of nitric osidcl s>mthesis in macropliage antiniicrobial activity. Curr. Opiti. I r t i i j i u i i o l . 3, 65-70. Nathan, C. (1997). Intlriciblr nitric osidr syithasc: \$'hat differrnce does it make? . 1.C h i . /r1r;t,st. 100, 2417-2423. Ndson, C . A , . \'idavsly, I., \'iner, N . J.. Gross, M. L.. and Uiianue. E. R. (1997). Aininocomplex class terminal triiiiming of peptides for presentation on major 1iistocoiiip~~til)ility 11 Inolecllles. Pt-oc. Not. Acod. sci. C'SA 94, 628-633. Nelmiiuccwo. R. R.. €Iensclieii-Ediiian. A. H., Br~rgess,W. I I . , and Tciiner, A. J. (1997). cDiia clonins and priin;iry stnictiire analysis of ClqR(P).thr human Clq/Mbl/Spu receptor that inetliates enhanced pliagocytosis iii \itro. Ittiinunity 6, 119-129, Newport, M . J., Hiidcy, C:. M.. Hiiston. S.,Hawylo\vicz, C;. M., Oostra, B. A., Williamson. R., and I,ecin, M. (1996). A iiiutation iii the interft.roii-y-receptor gene and srisceptibilit\ to mycolxicterial infection. N . Engl. ] Mcd. 335, 1941 -1949. Newton. S. M.. Jacob, C. 0..and Stockctr, B. A . (1989). Immunc. response to cholera tosin epitope iiisertetl i n Salnionolla flagellin. Seicricc 244, 70-72. Nicholson. S., Boiiccini-Almeitla, M., Lapa e Silvo, J. R., Nathan, C., Xie, Q. W., Munifortl. R., Weidner. J. R., Calavcy, J.. Geng. J., Borchat. N.. c,t a / . (1996).Indncible nitric oxidr, sylithase in pulnionary alvcdar niacrophages from patients with tiiberciilosis. J . E.yp. &fed. 183, 2293-2302. Nigoii. J., Gilleron. M., Calirizac, €3.. Horlncry, J. I)., Herold. M.. Thnmlier, M., a~id Puzo, G. (1997). The phosphatidyl-inyo-inositol anchor of' the lipoarabinomannaiis froin M!/cobncfcn'roiihoui.~bacillus Calniette G w r i n : Hetcrogeneity. structure, and role in the L . 23094-23103. regii1;ition 6 cytokine secretion. /. Biol. C ~ ~ W272,
362
LJLRICH E. SCHAIBLE et nl.
Norbury, C. C., Chambers, B. J., Prescott. A. R.; Ljunggren, H. G., and Watts, C. (1997). Constitutive macropinocytosis allows TAP-dependent inajor histocompatibility coniplex class I presentation of exogenous soluble antigen by bone marrow-derived dendritic cells. Eur. J. Ziuniimol. 27, 280-288. Novick, P., and Zerial, M. (1997).The diversity of Rab proteins in vesicle transport. Cum. $in. Cell Biol. 9, 496-504. Nozaki, Y., Hasegawa, Y., Ichiyama, S., Nakashima, I., and Shimokata, K. (1997).Mechanism of nitric oxide-dependent killing of Mycobrtcterinm bouis BCG in human alveolar rnacrophages. Infect. Zmniiin. 65, 3644-3647. O'Brien, L., Roberts, B., and Andrew, P. W. (1996). In vitro interaction of Mycobacteriurn tuberculosis and macrophages: Activation of anti-mycobacterial activity of macrophages and mechanisms of anti-mycobacterial iinmunity. Cum. Top. Miorbiol. Znmunol, 215,
97- 101. O'Brien, R. L., Ilapp, M. P., Dallas, A,, Palmer, E., Kubo, R., and Born, W. K. (1989). Stimulation of a major subset of lymphocytes expressing T cell receptor y6 by an antigen derived from M!lcobacterium tuberdosis. Cell 57, 667-674. O'Connor, C. D., Farris, M., Fowler, R., and Qi. S. Y. (1997).The proteome of Salmonella enterica serovar T!yphiniurium: Current progress on its determination and some applications. Elec,-trophore,sis 18, 1483- 1490. Oh, Y. K., Alpuche-Aranda, C., Berthiaunie, E., Jinks, T., Miller, S. I., and Swanson, J. A. ( 1996). Rapid and complete fusion of macrophage lysosomes with phagosomes containing Salnwnellu typhiniurium. Infect. Ininwn. 64, 3877-3883. Oh, Y. K., and Straubinger, R. M. (1996).Intrarellular fate of Mycobacterium aoium Use of dual-label spectrofluorometry to investicate - the influence of bacterial viability and opsonization on phagosomal pH and phagosome-lysosome interaction. h f e c t . Znimun.
64, 319-325. Ohga. S., Ikeuchi, K., Kadoya, R., Okada, K., Miyazaki, C., Suita, S., and Ueda, K. (1997). Intrapulmonary Mycobacterium auinm infection as the first manifestation of chronic granuloinatous disease. J. Znfect. 34, 147-150. Ojcius, D. M., Dedba, Y. B., Kanellopoulos, J. M., Hawkins, R. A,, Kelly, K. A,, Rank, R. G., and Dantlyvarsat, A. (1998). Internalization of Chlamydia by dendritic cells and stimulation of Chlamydia-specificT cells. J . Zmnwtiol. 160, 1297-1303. Ojcius, D. M., Hellio, R., and Dautry-Varsat,A. (1997).Distribution of endosonial, lysosonial, and major histocompatibility complex markers in a monocytic cell line infected with C h h i y d i a psittaci. Infect. Zmnizrn. 65, 2437-2442. Ojcius, D. M., Thibon, M., Mounier, C., and Dautry-Varsat, A. (199s). pH and calcium dependence of heinolysis due to Rickettsia protoazekii: Comparison with phospholipase activity. Infect. Zmmurc. 63, 3069-3072. Oheira, S. C., and Splitter, G. A. (1995). CD8' type 1 CD44'"CD45 RB'" T lymphocytes control intracellular Bnicekz abortus infection as demonstrated in major histocompatibility complex class I- and class 11-deficient mice. Eur. /. Znununol. 25, 2551-2.557. Orme, I. M. (1988).Induction of nonspecific acqtiired resistance and delayed-type hypersensitivity, but not specific acquired resistance in mice inoculated with killed tnycobacterial vaccines. Infect. I n m i i n . 56, 3310-3312. Ortmann, B., Copeman, J., Lehner, P.. Sadasivan, B., Herberg, J. A,, Grandea, A. G., Riddell, S. R., Tanipe, R., Spies, T., Trowsdnle, J., and Cresswell, P. (1997).A critical role for tapasin in the assembly and function of multimeric MHC Class I-TAP complexes. Science 277, 1306-1309. Pal, P. C . ,and Honvitz, M. A. (1992).Immunization with extracellular proteins of Mycobacteriurri tuberculosis induces cell-mediated immune responses and substantial protective iinmunity in a guinea pig model of pulmonary tuberculosis. Infect. Zrnniun. 60,4781-4792.
IKTRt\CEI,LUL.AR BACI'ERIA 4 N l ) THE: I h l M U N F : SYSTEM
363
Pariier, E. C., Wang, C.-R.,Flaherty, L., Fischer Lintlahl, K., arid Bevan, M .J. (1992). I1-2M3 presents ii Listcv-in rru,tioc!ilogenc.I. peptide to c)$otoxic T lympl~ocytes.Cell 70, 215-223. Pamer, E. G.. Sijts, A. J., Villanucva, M. S., Biiscli, D. H., antl Vijh, S. (1997). M I I C class I antigen processing of Listcvia r ~ i [ , r t [ ) ~ , ~ / f ~proteins: ) ~ c ~ ~ t eIinplications .y for doininant and suljdomiiiaiit CTL responses. Z t t t t n t m d . Heti. 158, 129-136. Pancholi, P., Mirza, A., Bhardwaj, N . , and Steininan, R. M. (l993a). Sequestration froin iinmuiie CD4' T cells of mycobactr.ria growing i n Iiuinan macrophages. Science 260, 984-986. Panclroli, P., klirza, A., Schauf. I'., Steininon, R. M., a i i d Hhartlwaj, N. (1993b).Presentation s : of transfer froin infected microof inycobacterial antigens by h u n i m drndritic ~ ~ 4 Lack phages. Ir4ec-t. 1rtmi11ri. 61, 5326-5332. Parham, P. (1!494). The rise and fall of great s I genes. Sc,rtiiti. Z t ~ i t t i i ~ t ~6, o ~373-382. . Park, S. H., Roark, J. H., and Bendelac, A. (1998). Tissiw-specific recognition of iiiouse CD1 moleciiles. /, Zntrriuriol. 160, 3128-3134. Parker, C . M.,Groh, V..Band, ti.,Porcelli, S. A., Morita, C.. Fahbi. M., Glass, D., Strominger, J. L., and Brenner. kl. B. (1990). Evidcmce for extrathymic changes i n the T cell receptor ylS repertoive. /. Exp. Med. 171, 1597-1612. Pasparkis, M., Alexopoulou, L.. Douni, E.. and Kollias. G. (1996). Tumour necrosis factors in iinmune regulation: Everything that is interesting is . . . new! Cytokine Growth Fmtor Rec. 7, 223--229. Pasula, R., Downing, J. F., Wright. J. R., Kacliel, D. L., Davis, T.E.. Jr., and Martin, W.J., I1 (1997). Surfactant protein A (SP-A) mediates attachment of Myc~bactericirri tuhcrcdosi.y to niurine alveolar nwrophages. Atti. /. H e ~ pCdI. A4d B i d . 17, 209-21 7. Paul, \Y. E. (1997).Interleukin 4: Signalling incd~anisinsaiid control ofT cell differentiation. Cibn Fouiid. Syrtip. 204, 208-216: discrission 216-219. Paylie, N . R., ;ind Honvitz, M. A. (1987). Pliago:ocy$osis of L. pnfwrtu)phih is mediated hy human moiioc)te coinplement receptors. /. Exp. Met/. 166, 1077-1389. Pearson, A. M. ( 1996). Scavenger reports in iiinnte immunity. Ciirr. Opirt. Zttititwio/. 8,
20-28. Pechhold, K.. LVesch, D., Schondelmaier, S., and Kabelitz. D. (1994). Primary activation of Vy 9-expressing y6 T cells by M!phcicten'rttti riihr,r-crik)sis:Requirement for Th1 -type CD4 T cell lielp arid inhibition by IL-10. J , IfJittiurio~.152, 4984-4992. Peeteririans. I V . E., Raats, C. J., van Furtli. K., and Langernians, J. A. (1995). Mycobacterial S4-kilodalton heat shock protein induces tninor ncLcrosis fwtor (Y and interleukin 6, reactive nitrogen intermediates, and toxoplasniostatic activity in inurine peritoneal macropliages. Irlfect. Z t r t r r i i i i i . 63, 3454-3458. Pena, S.V.,antl Kreiisb, A . M. (1997).Granulysin, a new huinan cljtolytic aranrile-associated protein witli possible iii~olvementin cell mediated cytotoxicity. Seniin. I i r i r n t i r t d . 9, 1li-125. Peterson, P. K.. Gekkrr, G.. HI], S., Sheiig, U'. S., Anderson. \V. R., Ulevitch, R.J., Tobias, and Cliao, C . C. (1W5).CD14 receptor-mediated P. S., Gustafson, K. V.. Mohtor, T. W., uptake of noiiopsonized Al!i'."bncte,rirrirt ftilwrctiIo,vi,Eby human microglia. Zrlfect. Ztrirtmrt. 63, 1598-1692. Pfeffer, K., Matsiiyiuiia, T., Kundig, T. M., IVakeliani. A . , Kishihara, K., Shahinian, A., Wieginann, K.. Ohashi, P. S., Krolike, M . , arid Mak, T. \V.(1990). Mice deficient for thc 55 kD tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. infection. Cell 73, 457-467. Pfeffer, K., Sckioel, B., Gulle, H., Kaufiiiann, S. H. E., and Wagner, H . (1990). Primary responses of human T cells to inycobactcria: A frequent set of y/S T cells are stimulated by protease-resistant ligands. Eur. /. Z t n r t i u r i d . 20, 1175-1179.
364
ULRICH E SClIAlBLE f't nl
Pfeifer, J. D., Wick, M. J.. Roberts, R. L., Findlay, K., Normark, S. J., and Harding, C. V. (1993). Phagocytic processing of bacterial antigens for class I MHC presentation to T cells. Nature 361, 359-362. Pierre, P., Denzin, L. K., Hamniond, C., Drake, J. R., Amigorena, S., Cresswell, P., and Mellnian, I. (1996). HLA-DM is localized to conventional and unconventional MHC class 11-containing eiidocytic compartments. lrrirnunity 4, 229-239. Pieters, J. (1997). MHC class I1 compartments: specialized organelles of' the endocytic pathway in antigen presenting cells. Biol. Chern. 378, 751-758. Pisetsky, D. S. (1996). Immune activation by bacterial DNA: A new genetic code. Irnmunity 5, 303-310. Pitt, A,, Mayorga, L. S., Stalil, P. D., and Schwaitz, A. L. (1992). Alterations in the protein composition of maturing phagosomes. 1.Clin. Invest. 90, 1978-1983. Plum, G.,andClark-Curtiss, J. E. (1994).Induction ofMycobncteriirinaoiurri gene expression following phagocytosis by human macrophages. Infect. Irnmnun. 62, 476-483. Polotsky, V. Y., Belisle, J. T., Miknsova, K., Ezekowitz, R. A., and Joiner, K. A. (1997). Interaction of human mannose-binding proteiii with Myrobocteriuni auiurn. J . Infect. Dis. 175, 1159-1168. Pope, M., and Kotlarski, I. (1994). Detection of SalrtwrielZf~-specificL3T4' and Lyt-2+ T cells which can proliferate in vitro and mediate delayed-type hypersensiti\ity reactivity. Irnniunolog~8 1, 183- 191. Pope, M., Kotlarski, I., and Doherty, K. (1994). Induction of Lyt-2+ cytotoxic T lymphocytes following primary and secondary Salmonello infection. Ittimunology 81, 177-182. Porcelli, S., Brenner, M. B., Greenstein, J. L., Balk, S. P., Terhorst, C., and Bleicher, P. A. (1989). Recognition of clnster of differentiation 1 antigens by human CD4-CD8cytolytic T lymphocytes. Nature 341, 447-450. Porcelli, S., Morita, C. T.. and Brenner, M. B. (1992).C D l b restricts the response of hnnian CD4-8- T lymphocytes to a microbial antigen. Nature 360, 593-597. Porcelli, S. A. (1995). The CD1 family: A third lineage of aritigen-presenting molecules. Adu. Initnuriol. 59, 1-98, Porcelli, S. A,, Morita, C. T., and Modlin, R. L. (1996). T-cell recognition of non-peptide antigens. curr. Opin. 1mriund. 8, 510-516. Porcelli, S. A,, Segelke, B. W., Sngita, M., Wilson, I. A,, and Brenner, M. B. (1998). The ED1 family of lipid antigen presenting molecules. Irnrrwnol. T o d q 19, 362-368. Porter, E. M., van Dam, E., Valore, E. V., and Ganz, T. (1997).Broad-spectrum antimicrobial activity of human intestinal deferisin 5. Znfict. Zmrrwn. 65, 2396-2401. Powvis, S. J. (1997). Major histocompatibility complex class I molecules interact with both subunits of the transporter associated with antigen processing, TAP1 and TAP2. Eur. J. Inirnuriol. 27, 2744-2747. Prete, G. F. D., Carli, M. D., Alrnerigogna, F., Daniel, C. K., Delioe, M. M., Zancuoghi, G., Pizzolo, G., and Romagnani, S. (1995). Preferential expression of CD30 by human CD4+ T cells producing Th2-type cytokines. FASEB J. 9, 81-86. Prigozy, T. I., Sieling, P. A,, Clemens, D., Stewart, P. L., Behar, S. M., Porcelli, S. A,, Brenner, M. B., Modlin, R. L., and Kronenberg, M . (1997). The mannose receptor delivers lipoglycan antigens to endosoines for presentation to T cells by C D l b molecules. ImftllLYLity 6, 187-197. Prina, E., Antoine, J. C., Wiederanders, B., and Kirschke, H. (1990). Localization arid activity of various lysosonial proteases in Leishnlania a,rla;nnensis-infected macrophages. Infect. lrnniun. 58, 1730-1737. Prina, E., Lang, T., Glaichenhaus, N., and Antoine, J. C. (1996).Presentation ofthe protective parasite antigen LACK by Leishmania-infected macrophages. J. Initrtunol. 156, 43184323.
IN‘THA(XLL1J L A H BA( :TE RIA Ah’ 1) TI 1 Pi: I hl M U N E SYSI’E hf
36.5
Probst, I’., Skeib, Y. A., Steeves. M., Gcwassi. A,, Grabstein, K. H., and Reed, S. G. (1997). A Leishiiiania protein that modiilates iiitcrleukin ( ILL 12, IL- 10 a d tuiiior necrosis factoralpha production and expression of BT- 1 in h u m a i i moiioc)rte-tleriveclantigen-prescnting cells. Etcr J . Z t t m ~ r t t d27, 2634-2642. Pryjma, J., Baran, J.. Ernst, M.. Woloszyn. M., and Flad, H. D. (1994). Altered antigenpresenting cxpacity o f h u m m nionocytcbs after pliagocpsis of bacteria. It+ct. I t t i t t i t i t i . 62, 1961-1967. Pugin, J., Heumann, 1. D., Toniasz. A.. Kravchenko. V. \’,, Akainatsii, Y., Nishijiina, M., Clauser, M. P., Tobias, P. S., antl Ulecitch. K. J. (1994).CD14 is a pattern recognition receptor. I i r i t i i t i n ity 1, 509-5 16. Qi, S. Y., Moir. A., and O’Connor. C. (19%). Proteonir of Sultttonelln typliinitirinm SL1344: Itlentification of novel ahinidant cell envelope proteins and assignment to a twodiinensional reference map. J. Bnctcriol. 178, 5032-50338. Rabinowitz, S., Horstniann, H., Gordon. S., and Criffitlis, G. (1992).Iiniiiunocytoclieniic~il characterization of the endocytic and phagolysosoinal conipartinents in peritoneal ~nacrophages. J . Crll Biol. 116, 95-112. Raetz, C.. Ulevitch, R.. Wright, S., Siblry, C.. Ding, A,, arid Nathan, C. (1991). Cramnegative endotoxin: ,411 estraortlinq lipid with profound effects on eukaryotic signal transduction. FASEB J. 5, 2652-2660. Iiafic.-Kolpin. M., Essrnberg. K. C.. and Wyckoft: 1. H., I11 ( 1996). Identification and comparison of iiiacl.opliage-induce~1proteins antl proteins indnced under various stress conditions iii Bnrctlla ob&ttis. 1ifet.f. I t t t t t t i i i i . 64, 5274-5383. R;ijalingam. R.. Singal, D. P., arid Melira. N.K. (1997).Transporter associated with antigeni n g (TAP)genes and snsceptibilit\ to tuberculoid leprosy antl pullnonary tu1)erculosis. TissirP Arctigetis 49, 168-172. Ramhnkkana. A,, Salzrr, J. L., Yiirchcnco. P. D., antl Triomanen. E. I. (1997). Neural targeting of M~jcobcrtteritiittIelirav nietliated by the G dornain of the laininin-a2 chain. Cell 88, 811-821. Rathnian, M., Barker, L. P., imd Fdko\v, S. (1997). The unique trafficking pattern of SaltttottcUa t ~ ~ ~ ~ l t i ~ ~ i r c r i i ~ t t i - c ~phagosomes i i t ~ i i i i i i i g in niiirine niacrophages is independent of the inec1i;iiiism of bacterial entry. Infict. I t i t n i t t i t . 65, 1475-1485. Rathinan. M., Sjaastad, M. D.. ;md Falko\v, S. (1996).Acidification of phagosomes containing Smltttoi~c/lnt ! / p h i t t c r r r i u t ~ ti n niurine nincrophagt~s.Infict. I t w i t t i n . 64, 2765-2773. Kaulston, J. E. (1997). Response of Ch/nttiydin frcrcliotttntis serovar E to iron restriction i n vitro and evidence for iron-regiilated clilamydial proteins. ltfect. Z r t m ~ i i n .65, 4539-4547. Heilly, T. J., Baron. G. S., Nano, F. E., and Kiihlensclimidt, M. S. (1996).Characterization and sequencing of I I respirator). 1)urst-inhi1,iting acid phosphatase from Frutici.sella fuEareiwi.s. /. B i d Clienc. 271, 10973-109833. Reiinanri, J.. and Kaufiiiann. S. H. E. (1997). Altwnativc, antigen processing pathways for MMC-restricted epitopc presentation i n anti-infective immlmity. Ctrm. Opin h i t t 1 1 L ? l d . 9,462-469. Reiner, S. L., d Sedrr. R. A. (199s). T helper cell diffrrentiatioii i n iinininie response. Cirrr. Opiti. ~ m i c i r i i d7, :360-366. Reis e Sousa, C.. and Gerinain, R. N. (1995). Major liistocompatibili~complex class I presentation of peptitles derived froni soluhle exogenous antigen by a subset of cells engaged in phagocytosis. /. E x p hled. 182, 841-851. R e p t . J. M., Ikrthet. F. X., and Gicquel, B. (1995).The urease locus of M!/cobacterium trrberctikxsis md its utilization for the demonstration of allelic exchange in Mycohucteriz~nt boois bacillus Calmette-Guerin. Proc. N a t . Acnrl. Sci. USA 92, 8768-8772.
Reyrut, J. M., Lopez-Rainirez, G.,Ofrrdo, C., Gicquel, B., and Winter, N. (1996). Urease activity does not contribute dramatically to persistence of Mycolmcteririm boois bacillus Calniette-(~iierin.Irfmtrct.I t n t t i r ~ r i .64, 3934-3936. Rlaoatles. E. R., aud Ormr. 1. M. (1997). Siisceptibility of a panel of viruhit strains of Mycobac-fc~ri:r?n triberrrilosis to reactive nitrogen interinediates. Infect. I r n t t i r m . 65, 11891195. Rhoades, E. R., Coopcr, A. M., and Orine. I. M. (1995). Chemohne response in mice infected with Mycobocteriritti tribrrcrilosis. Infect. Z t t m r m 63, 3871-3877. Kicdel. 13. D., antl Kaufiiiann. S. H. E. (1997).Chenaokine secretion by huniaii polyinorphonuclear granulocytes after stiinulation with Bf!/cobnctcrinn~trrhercrilu niiinnan. Irtfiect. Z m t t i r i t i . 65, 4620-4623. Kiese. R. J., Wolf, P. R., Bromine, D., Natkin, L.. K., Villadangos, J. A,, Ploegh. I-I. L.. and Chapman, H. A. (1996). Essential role for cathepsin S ill M H C class 11-associatedinvariant chain processing and peptide loading. Ztnttirmity 4, 357-366. Roach, T. I., Barton, C. H.,Chatterjee, D., and Blackwell, J. M. (1993). Macrophage activation: li~~o~~rabinoiriaaira~~~a froin avinilent and virulent strains of Myt:obncteritnntuberciilosis differentially induces the rarly genes c-fos, KC, JE. and turnor necrosis factor-a. /. Irmnunol. 150, 1886-1896. h a r k , J. II., Park, S. H., Jayawardena, J., Kavita, U., Shannon, M., and Bendelac, A. (1998). CD1.1 expression by iiiotise iiiitigeii-preseiatingcells and niarginal zone B cells./. I t r i t t m r m l , 160, 3121-3127. Roberts, A. D., Sonnenberg, M. G., Ordway, D. J., F u r y , S . K., Brennan, P. J., Belisle, J. T., and Ornie, I. M. (1995). Characteristics of protective iininirnity engendered by vaccination of mice with purified culture filtrate protein antigens of Mycobmct~v-itcm frrbr~rculosis.I ~ r ~ t n r m ) / o85, g ~ 502-508. Rol)inson, D., Shihuya. K., Mui, A., Zonin, F., Murphy, E., Sam, T., Hartley, S. B., Mrnon, S., Kastelein, K., Bazan, F., and O’C;arra, A. (1997). IGIF does not drive TI11 development but sjnergizes with IL-12 for interferon-y production antl activates IRAK and NFKB. Z 7 ~ l t t ~ f i t i i t !7, / 571-581. Rock, E. P., Sibl)altl, P. H., Davis, M. M., and Cliien, Y.-H. (1994).CDH3 length in antigeiispecific immune receptors. /. Exp. Met/. 179, 323-328. Rockey. D. D., and Rosquist, J. L. (1994).Protein antigens of Clilutnydio psittnci present in infected cells Iiut not detected in the infectious elementary body. Ztfect. [triiririn. 62, 106-112. Rogers, H. W., Caller).. M. P.. Deck, B., and Unantie, E. R . (1996). Listericl t t z ~ r ~ [ ~ c y ~ [ ) ~ ~ , ~ i e . ~ iridiices apoptosis of infrcted liepatocytes. /. l i t i i n i ~ ~ i o 156, l. 679-684. R o j ~ M., , Barrera, L. F., Puzo, G., and Garcia, L. F. (1997). Differential induction of apoptosis liy virulent Mycobactr~riritti ttihercrilosis in resistant and susceptible inurine macrophagw Hole of nitric oxide and mycobacterial products. /. Itnrnzinul. 159, 13521361. Romagnani. S., Parronchi, P., D’Elios, M. M.. Rornagnani, P., Annunziato, F., Piccinni, M. P., Manetti, R., Sanipognaro, S.,Mavilia, C., De Cadi, M., Maggi, E., and Del Prete, C.. F. (1997). An update on human TI11 arid TI12 cells. Znt. Arch. Allergy Z m t ~ i u t i o l . 113, 153-156. Romain, F., Augier. J., Pesclier, P., antl Marchal, G. (1993). Isolation of a proline-rich ~nycol~acterial protein eliciting delayed-tpe hpersensitivity reactions only in guinea pigs iininunized with living inycobacteria. Proc. Not. Acnd. Sci. USA 90, 53224326. Roman, M., Martin-Orozco, E., Goodinan, J. S., Nguyen, M. D., Sato, Y.. Konaghy, A,, Kornbluth, H. S., Richman, D. D., Carson, D. A., and Raz, E. (1997).I ~ n ~ n ~ ~ n o s t i m ~ ~ l a t o r y DNA sequences function as T helper-1-promoting adjuvants. Nature Med. 3, 849-854.
ISTKACELLULAR BACTERIA AND THE I M M U N E SYSTEM
367
Rook, G. A. W., Taverne, J., Leveton, C., and Steele, J. (1987). The role of y-interferon, \.itamin D metabolites and tumor necrosis factor in the ~athogenesisof tuberculosis. I~n~nunology 62, 229-234. Rosat, J.-P., Grant, E. P., Beckman, E . M., Dascher, C. C., Sieling, P. A,, Frederique, D., Modlin, R., Porcilli, S. A,, Furlong, S. T., and Brenner, M. B. (1998). TCR a/3+CD8+ T cells recognize microbial phospholipid antigens in the context of C D 1 presenting molecules. J. Exp. Med. (in press). Rosenkrands, J., Rasmussen, P. R., Carnio, M., Jacobsen, S., Theisen, M., and Andersen, P. (1998). Identification and characterization of a 29-kilodalton protein from Mycohactet-ium tuher~~ilosis culture filtrate recognized by mouse memory effector cells. Infect. Iiizmun. 66, 2728-2735. Rosenshine, I., Ruschkowski, S., Foubister, V., and Finlay, B. B. (1994). Snbnonella typhiiiwTit~ininvasion of epithelial cells: Role of induced host cell tyrosine protein phosphorylatiori. Infect. Iinnlun. 62, 4969-4974. Rothe, J., Lesslauer, W., Lotscher, H., Lang, Y., Koebel, P., Kontgen, F., Althage, A,, Zinkernagel, R., Steinmetz, M., and Bluethmann, 11. (1993). Mice lacking the tumour necrosis Factor receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection by Listeria ~nonocytogenes.Nature 364, 798-802. Rothman, J. E. (1994). Mechanisms of intracellular protein transport. Nature 372, 55-63. Russell, D. G. (1998). What does "inhibition of phagosome/lysosome fusion" really mean? Trends Microbiol. 6, 212-214. Russell, D . G., Dant, J., and Sturgill-Koszycki, S. (1996). Mycohacteriz~maoiuin- and Mycobacterium tuberculo.sis-containingvacuoles are dynamic, fusion-competent vesicles that 156, are accessible to glycosphingolipids from the host cell plasmalemma. J. Itr~n~unol. 4764-4773. Russell, D. G., Sturgill-Koszvcki, S., VanIIeyningen, T., Collins, H. L., and Schaible, U. E. (1997).Why intracellular parasitism need not b e a degrading experience for Mycobacterilim. Plzil. Trans. Roy. Soc. Lonrlon Ser. B Biol. Sci. 352, 1303-1310. Russell, D . G., Xu, S., and Chakraborty, P. (1992). Intracellular trafficking and the parasitopl~orousvacuole of Leiskinania imxicana-infected macrophages. J. Cell Sci. 103, 11931210. Sallusto, F., Cella, M., Danieli, C., and Lanzavecchia, A. (1995). Dendritic cells use macropinocytosis and the mannose receptor to concentrate macromolecules in the major histocompatibility complex class I1 compartment: Downregulation by cjtokines and bacterial products. ]. Exp. Med. 182, 389-400. Sallusto, F., Mackay, C. R., and Lanzavecchia, A. (1997). Selective expression of the eotaxin receptor CCR3 by human T helper 2 cells. Science 277, 2005-2007. Salzinan, A. L., Eaves-Pyles, T., Linn, S. C., Denenberg, A. G., and Szabo, C. (1998). Bacterial induction of inducible nitric oxide synthase in cultured human intestinal epithelial cells. Gastroerlterology 114, 93-102. Samsom, J. N., Langermans, J. A,, Savelkoul, H. F., and van Furth, R. (1995). Tumour necrosis factor, but not interferon-y, is essential for acquired resistance to Listeria imnocytogenes during a secondary infection in mice. Irnmunology 86, 256-262. Sanceau, J., Falcoff, R., Beranger, F., Carter, D. B., and Wietzerbin, J. (1990). Secretion of interleulan-6 (IL-6) by human monocytes stimulated by muramyl dipeptide and tumour necrosis factor a. lrrlmunology 69, 52-56. Sander, B., Skansen-Saphir, U., Damm, O., Hakansson, L., Andersson, J., and Andersson, U. (1995). Sequential production of T h l and Th2 cytokines in response to live bacillus Calmette-Guerin. Immunology 86, 512-518.
368
ULRICH E. SCHAIBLE rt n /
Sastre, L., Kishimoto, T. K., Gee, C., Roberts, T., and Springer, T. A. (1986). The mouse leukocyte adhesion proteins Mac-1 and LFA-1: Studies on mRNA translation and protein glycosylation with emphasis on Mac-1. J . Zmiriunol. 137, 1060-1065. Sato, Y., Roman, M., Tighe, H., Lee, D., Corr, M., Nguyen, M. D., Silverman, G. J., Lotz, M., Carson, D. A,, and Raz, E. (1996). Imniunostiniulatory DNA sequences necessary fbr effective intradermal gene immunization. Science 273, 352-354. Schaible, U. E., Sturgill-Koszycki,S., Schlesinger, P. H., and Russell, D. G. (1998).Cytokine activation leads to acidification and increases maturation of Mycobacteriuin aoirrmcontaining phagosomes in murine macrophages. J. Zmmunol. 160, 1290-1296. Schild, H., Mavaddat, N., Litzenberger, C., Ehrich, E. W., Davis, M. M., Bluestone, J. A., Matis, L., Draper, R. K., and Chien, Y.-H. (1994).The nature of major histocompatibility complex recognition by yl6 T cells. Cell 76, 29-37. Schirmbeck, R., Bohm, W., Melber, K., and Reiinann, J. (1995). Processing of exogenous heat-aggregated (denatured) and particulate (native) hepatitis B surface antigen for class I-restricted epitope presentation. J. Zmmirr~ol.155, 4676-4684. Schirinbeck, R., and Reimann, J. (1994). Peptide transporter-independent, stress proteinmediated endosomal processing of endogenous protein antigens for major histocompatibility complex class I presentation. Eur. J. Ztr~i7tund.24, 1478-1486. Schirmbeck, R., and Reimann, J. (1996). “Empty” Ld molecules capture peptides from endocytosed hepatitis B surFace antigen particles for major histocompatibility complex class I-restricted presentation. Eur. J , Znimunol. 26, 2812-2822. Schlesinger, L. S. (1993). Macrophage phagocposis of virulent but not attenuated strains of Mycobacterii~rntuberculosis is mediated by mannose receptors in addition to complement receptors. J . bnmunul. 150, 2920-2930. Schlesinger, L. S., Bellinger-Kawahara. C. G., Payne, N. R., and Honvitz, M. A. (1990). Phagocytosisof Mycobucterium tuberculosis is mediated by human monocyte complement receptors and complement component C3. J . Immirncd. 144, 2771-2780. Schlesinger, L. S., and Homitz, M. A. (1991). Phenolic glycolipid-1 of Mycobacteriuni l q r a e binds complement component C3 in serum and mediates phagocytosis by human monocytes. J. Exp. Men. 174, 1031-1038. Schlesiriger, L. S., Hull, S. R.. and Kaufinan, T. M. (1994).Binding ofthe tenninal mannosyl units of lipoarabinomannan from a virulent strain of Myc(ibncterizcni tuberculosis to human inacrophages. J. Zrtitnun~d.152, 4070-4079. Schlesinger, L. S., Kaufinann, T. M., Iyer, S., Hull, S. R., and Marchiando, L. K. (1996). Differences in mannose receptor-mediated uptake of lipoarabinomannan from virulent and attenuated strains of Mycobacterium tuhercrilosis by human macrophages.]. Znmiunol. 157,4568-4575. Schoel, B., and Kaufinann, S. H. E. (1998).Influence of mycobacterial virulence and culture condition on y6 T cells. Microb. Patliogen. 24, 197-201. Schoel, B., Sprenger, S., and Kaufmann, S. H. E. (1994).Phosphate is essential for stimulation of Vy9VS2 T lymphocytes by mycobacterial low molecular weight ligand. Eur.J . Immunol. 24, 1886-1892. Schorey, J. S., Carroll, M. C., and Brown, E. J. (1997). A macrophage invasion mechanism of pathogenic mycobacteria. Science 277, 109 1-1093. Scidmore, M. A., Fischer, E. H., and Hackstadt. T. (1996). Sphingolipids and glycoproteins are differentially trafficked to the Chlamydia truchorruitis inclusion. J. Cell Biol. 134, 363-374. Seder, R. A., and Paul, W. E. (1994). Acquisition of lympholane-producing phenotype by CD4+ T cells. Annu. Reu. Z n ~ t ~ n12, ~ l 635-673. .
INTRACELLULAR RA(:TERIA A N D TllE IMhlUNE SYSTEM
369
Senaldi, G.. Yin, S., Sliaklee. C. L., Piguet. P. F., Mak, T. W., and Ulich, T. R. (1996). C o n ~ t t c ~ ~ o ~ f rpclruronr i u t t ~ and M!ycoboctrrium boois bacillus Calmette-G~ierin-i~i~I~~~~ed graiiiilom;i formation is inhibited in TNF receptor I (TNF-RI) knockout Inice and ly treatment with soluble TNF-RI. /. Z t n r ~ ~ i m d157, . 5022-5026, S h a h a n , A., Pfeffer, K., Lee, K. P., Kiindig. T. M.. Kishihara, K., Wakeham, A,. Kawai, K., Ohashi P. S., Tliompson, C . B., and Mak, T. W. (l99:3).Differential Tcell costinilllatory recjuiremeiits in CD28-deficient micc. Scicricc 261, 609-61 2. Shaw, M . A.. Atkinson, S., Dockrell, H., Hnssain. R., Lins-Lainson. Z., Shaw, J.. R u m s , F., Silveira, F.. Melid, S. Q., Kaukal), F., ct 01. (1993). An HFLP map for 2(]33-q37from multicase mycobacterial and leislinianial disease families: No evidence for an Lsl/Ity/ Hcg gene homolope influeiicingsusceptibilityto leprosy. Ann. Htort. Genet. 57,251-271, J. M., Rodgers, J. R.. and Rich, H. R . (1994). Antigen presentation by major histuconipitihility complex class I-B niolecules. Annu. Rec. Irnmunol. 12, 839-880. Shen, H.. Miller. J. F., Fan. X., Kolwyck, D., Ahnied. R., and Ha*, J. T. (1998).Coinpartmentalization of bacterial antigens: Differential effects on priniing of CD8 T cells arid protective irnmunity. Cell 92, 535-545. Shen, Z.. Remikoff. G . , Dranoff, G . , and Rock. K. L. (1997). Cloned dendritic cells can present exogenous antigens on both MHC class I and class 11 molecules. J , I I I I J I ~ I ~ ~ I I ~ , 158, 2723--2730. Shiniada, S . , Yano, 0..and Toknn;iga, T. (19x6). In vivo anginentation of natural killer cell activity with a deoxyribonucleicacid fraction of' BCG. Jopcin. J . Cnticer Res. 77, 808-816. Sibley, L. D., and Krahenbuhl, J. L. ( 1987).Al!/colmferiirtn lcpro~:-burdenedniacropliages are refractory to activation by y-interferon. I?rfccl. Z r t i r r i u t i . 55, 446-450. Sibley, L. D., and Krahenbrilil, J. L. (1988). Defective activation of granulonia macrophages from Alycoboctcritim kproc-infectrd nude inice. 1, Lrukocyte B i d . 43, 60-66. Sibley, L. D. Hunter, S. W.,Brennan, P. J.. and Kralimhiilil, J. L. (1988). Mycobactend 1il'"ar~ibinoiiianiiaii inhibits y-interf(~ron-inecliatedactivation of nixxopliages. Itrfect. Zitittitttl. 56, 1232-1236. Sibky, L. D., Weidner, E., and Kraheiil)iihl, J. I>.(1985). Phagosonie acidification blocked by intracellular Toroplastnci goridii. Notiire 315, 416-4 19. Sieling, P. A.. Chatterjee, D.. Porcelli, S . A,, Prigozy, T. I., Miuzaccaro, R. J., Soriano, T., Bloom, 13. 13., Brenrier, M. B., Kronenl)erg, M., Brennan, P. J., and Modliii, R . L. (1995). CDI -restricted T cell recognition of microbial lipoglycan antigens. Scirwe 269, 227-230. Sijts, A. J., and Pamer. E. G. (1997).Enliancetl intracellular dissociation of major histocompatibility ccrmplex class I-associated peptides: A mechanism for optimizing the spectrum of cell surface-presented cjotoxic T Iyinphocyte epitopes. /. Exp. Med. 185, 1403-141 1. Sijts, A. J.. Vikuiueva. M. S., mid Panier, E. G. (1996a). CTL epitope generation is tightly linked to cellular proteolysis ofa Listc.rin tiiorioc,llfogu,rcsantigen. I. Zmniunol. 156, 14971503. Sijts, A. J. A. M., Neisig, A,, Nref'jes, J. J., and Piinier, E. C;. (l99Gb).Two Listeria nionocytogenes CTL epitopes are processed from the saincb antigen with different efficiencies. /. f r r i t t ~ i i r t d156, 685-692. Sijts, A. J., Pilip, I., and Painer, E. G. (1997). The Listerin tiioriocytogenr,s-secretedp60 prott+i is an N-end rule sulxtrate in the cytosol of infected cells: Implications for major histocompatibility complex class I antigen processing of bacteriiil proteins. /. Biol. Clicni. 272, 19261- 19268. Silva, C. L., Silva, M. F., Pietro. R. C., and Lowrie, D. B. (1994). Protection against tuberculosis by passive transfer with T-cell clones recognizing mycobacterial heat-shock protein 65. Ztrittit4nohg!/ 83, 341-346.
370
ULHICH E. SCIlAIHLE 6't
N/
Silverinan, D. J., Santucci, L. A,, Meyers, N., i d Sekeyova, Z. (1992).Penetration of host cells by Rickettsia rickettsii appears to be mediated by a phospliolipase of rickettsia1 origin. Infect. Ztrittiiiri. 60, 2733-2740. Sinai, A. P., and Joiner, K. A. (1997).Safe haven: The cell hiology of nonfiisogenicpathogen vacuoles. Artnri. Ben. Micmbiol. 51, 415-462. Sizemore, D. R., Branstroin, A. A,, antl Sadoff, J. C. (1995). Attenuated Shigella as a DNA delivery vehicle for DNA-mediated immunization. Science 270, 299-302. Skeen, M. J., and Ziegler, H. K. (1993). Induction of inurine peritoneal y6 T cells and their role in resistance to bacterial infection. /. Exp. Med 178, 971-984. Skeen, M. J., and Ziegler, H. K. (1995). Activation of yS T cells for production of IFN-y is mediated by bacteria via niacropliage-derivecl cytokines IL-1 antl IL-12. J. Z i ~ ~ m i r t t o l . 154,5832-5841. Smiley, S. T., Kciplan, M. H., and Grusby, M. J. (1997). Iinmunoglobulin E production in the alisence of interleukin-4-secreting CD I-dependent cells. Science 275, 977-979. Snoeck, H. W., Lenjou, M., Nys, C., Lardon, F., Peetermans, M. E., Van Bockstaele, D. R., Moulijn. A,, Haenen, L., and Bernentan. Z.N. (199G).Interleukin 4 and interferoiiy costiniulate the expansion of early human myeloid colony-fomiing cells: Proposal of a motlel for the regulation of myelopoiesis by interleukin 4 and interferon-y and its integration with the regulation of the iminune response. Leukcvnin 10, 117- 122. Song, R., and Harding, C. V. (1996). Roles of proteasonies, transporter for antigen presentation (TAP), and beta 2-microglobrilin in the processing of bacterial or particulate antigens via an alternate class I MHC processing pathway. 1. Irtitriirnol. 156, 4182-4190. Sonnenberg, M. G., antl Belisle, J. T. (1997). Definition of Mycobacteritrni tiihercir1osi.s culture filtrate proteins by two-dimensionalpolyaerylaniide gel electrophoresis. N-terminal amino acid sequencing, and electrospray inass spectrometry. Ztlfect. I n i r n u n . 65, 45154524. Sonoda. Y. (1994). Interleuldn-4--a dual regulatory factor in hematopoiesis. Leukemia .!$Ulp'hOtlKl 14, 231 -240. Sousa, A. O., Salem, J. I., Lee, F. K., Vercosa, M . C., Cruaiid, P., Bloom, B. R . , Lagrange, P. H., and David, H. L. (1997a).An epidemic of tuberculosis with a high rate of tuberculin aiiergy among a population pre\ioiisly unexpostd to tuberculosis, the Yanomanii Indians an Amazon. P n x Nat. Acntl. Sci. USA 94, 13227713232, Sousa, C . R., Hieny, S., Scharton-Kerstrn, T.,Jankovic, D.. C h e s t , H., Gerniain, R. N., and Sher, A. (199%). In vivo inicroliial stimulation induces rapid CD40 ligand-iiidependeiit production of interleukin 12 by dendritic cells and their redistrihiition to T cell areas. J. Exp. Med. 186, 1819-1829, Stahl, P. D. (1992). The inannose receptor and other macrophage lectins. Cirrr. Opirt. ltnntrittol. 4, 49-52. Stahl, P. D., and Ezekowitz, R. A. B. (1998).The nianiio~ereceptor is a pattern receptor iiivolved in host defense. Ciirr. Opin. Zrntritinol. 10, 50-55. Stanford, J. L. (1991). Improving on BCG. Apttiis 99, 103-113. Starnbach, M. N.. Bevan, M. J., and Lampe, M. F. (1994).Protective cytotoxic T lymphocytes are indirced during inurine infection with Chkmydia tracimrutis. 1.ftrtrratnol. 153,51835189. Stein, M. A., Leung, K. Y., Zwick, M., Garcia-del Portillo, F.,and Finlay, B. B. (1996). Identification of a Salmonella virulence gene required for formation of filamentous structures containing lysosonial membrane glycoproteinswithin epithelial cells. Mol. Microbiol. 20, 151-164. Steinhoff, U., Kleinm, U., Greiner, M., Bordasch, K., and Kaufinann, S. H. E. (1998). Altered intestinal in~niunesystem but rrormal antibacterial resistance in the abselice of P-selectin and ICAM-1. /. Intttitirtol. 160, 61 12-6120.
Steininan. R 41., and Swanson, J. ( 1995). The rn(loc!!tic actiijih of dendritic cells. /. Exp. ilkd 182, 283-288, Steiiger. S., Mazzacc;iro, K. J.. CJycniiira. K..Clio. S., Barnes, P. F.. Kosat. J. P., Sette, A., Brcnner, M. B.. Porcrlli, S. A.. Bloom, H. R., and Modlin. K. L. (1997). Differential effects of c:ytol>tic T cell siibscBts o i i intracrlliilar infection. S'ciotccj 276, 1684-1687. of C D l on antigen Strnger. S., V i u i . K. H., m t l M o d h i . R. L. (1998). 1~0~11-rrgiilatiori tiihrrcrik).ui.s./. ~ t t t t t i ~ t m (in l . prcss)). presenting cc.lls by infection witti bf!/col~crc.tcririttt Stokes. R. IV., IIaidl, I . D.. Jefferirs. \V. A.. ;mtl Sprert. D. P. (1993). Mycobacteriamacrophage interactions: Macropliagc. phenot)Fe cleterniines tlie nonopsonic binding of ~l!/c~obnctct-iritti Irrbcwdo.ui.uto iriurinc. macrophagrs. /. I t t i t t i l i n o l . 151, 7067-7076. Stover. C. K.,Bansul. C;. P.. Hailson. M . S.. Biirlein. J. E., Palasq-nski. S. R., Youilg, J. F., Koenig. S.. Y o ~ n ~ D. g . B., Satlzienr, A., ant1 Barlioiir. A. G . (1993). Protective immunity eIicitetI by rcconihinaiit IiaciIIe Caliiiettc.-C:ueiiii ( BCG) r~pressingonter siirface protein A (OspA) lipoprotein: A c:iiididate I .yiw disease vuccine. /. E q ) . Mod. 178, 197-209. Striignell, R. A., Drew. D.. Mercitca, J.. DiN;italtl, S.. Firez, N., Dunstan, S. J.. Sirn~nons, C:. P.. and l'adolas, J. (1997).D N A \ x r i n c s for Iiacteiial infections. Z t t i t ) t r ~ t ~ oCcll l. Riol. 75, :364-369. Sturgill-Koszycki. S., Iladtlk P. I,., a i d Russell. D. G. (1997). Tlicl interaction betwreii niycobacteriuni antl the macrophage anal>.zetl l)y t\~~o-dirnensioiialpolyiic~lamitle gel t~IectropIioresis.E~t~ctro~~~ror~.sis IS, 25~8-2,565. Sturgill-Kosz!cki, S.. Scliailde, U. E.. antl Russell, D. C. (1996). Mycobactcrium-containing phagosoincs are accessihle to early c~ndosoirit~s and reflect a transitional statc in normal phagosoiiie biogenrsis. E M B O /. 15, 6960-6968. Sturgill-Koszycki, S.. Schlesinger. P. II . . (liakrahoit\, P.. Hadtlis, P. L., Collins, H. L., Fok, A. K.. Allen. H. I]., Cluck. S. L., lleusrr, J.. and Kiisscll, D. C. (1994). 1,ack of acidification i n Mycolwteriuin phagosonies p r o c h w l hy ~~xc11isioii of the vesicular protoii-ATPase. Scictrcc. 263, 678-681. Sugita, M.. and l3renner, M . B. ( 1 ~ ~ 5Association ). of the invaliant chain with major Iiistocompiitil,ility coiriple.; class Imoleciilrs directs trafficking to endocytic compartments. 1.H i i d C/rc.t)r.270, 1443-1448. Siigita, M., J w k n i a i i , R. M., van Donsc,l;iar, E.. Beliar. S. M., Rogers. H. A.. Peters, P. J., Brenner, b1. B., and PorccIIi, S. A. ( I 996). ~!!top~asiiiictail-depcwient IocaIization of CD 1 b anti~eii-preseiitinginoleciiles to M IICs. ScicJwc 273, 349-:352. Siigita, M . , Pr)rcelli, S. A,. iintl Brtwner, M . B. (1997). Assentbly and reteiition of C D l b Iiea\y cliuins in the endoplasniic reticiiliini. /. Z t t r t t i i ~ t t o / . 159, 2358-2365. S i i h m . L. b"., Bober. I,. A., Grace, M . J., Braun, S., Macosko. 11. D.. Pa)yandi, F., Sivo, C . C., and Nariila, S. K. (1992). Thts in vivo biological activip of inurine IL-4 and the role of entli)gcwous cytokiiitx /. Cdl. Biocltotr. Sirpp/. 16c, 10:3-109. Suzue, K., antl l'onng, H. A . (1996). Acljuviint-free lisp 70 fusion protein system elicits huinoral u i t l celliilar immiinc responses to HI\'-1 1724. /. Zttrtt~irncd 156, 873-879. Suzue, K.. Zhou, X.. Eism. H. N . , a i i d Yoling, R. A . (1997).Heat shock fnsion proteins as veliicles for antigcan tlelive? into the iiiajor Iiistocompatil~iIi~ complex closs I presentation pathway. Proc. Not. Acod. Sci. U S A 94, 13136-13151. Suzuki. H.. Kiirihara. Y.. Takeyi. M.. Kainiida. N.. Kataoka. M.,Jishage, K., Ueda, 0.. Sakaguclii. I I . . Higaslii. T., Suzuki, T., Takashinia, Y..Kawabe. Y., Cyiishi, IIontla, M., Kiirihara, H., Ahuratani, H., Doi, T., Matsunioto. A., Axuma, Toyoda, Y.. Itakura, El.. Yaznki. Y., IIoiinchi, S., Takahashi, K., Kruijt, J. K., vaii Berkel, T. J. C.. Steinbreclwr. U. P., Isliihashi. S., Maeda, N., Gordon, S., and Kotl;tin;i, T. (1997). A role for inicrophage scavengrr receptors in atherosclerosis and siisceptibility to infection. Nrrtiirr, 386. 292-296.
372
ULRICH E. SCMAIBLE el al
Svensson, M., Stockinger, B.. and Wick, M. J. (1997). Bone inarrow derived dendritic cells can process bacteria for MHC-I and MHC-I1 presentation to T cells. J. Immuriol. 158,4229-4236. Swanson, J. A,, and Baer, S. C. (1995).Phagocytosis by zippers and triggers. Trends Cell Biol. 5 , 89-92. Swanson, M. S., and Isberg, R. R. (1995).Association of Legionella pflelinwphlf~with the macrophage endoplasmic reticulum. Infect. Iinnmn. 63, 3609-3620. Swanson, M. S., and Isberg, R. R. (1996). Analysis of the intracellular fate of Legiorieh prieiirrwphila mutants. Atin. N . Y. Acud. Sci. 797, 8-18. Swier, K., and Miller, J. (1995). Efficient internalization of MHC class 11-invariant chain complexes is not sufficient for invariant chain proteolysis and class I1 antigen presentation. J. Imnunol. 155, 630-643. Szabo, S. J., Dighe, A. S., Guliler, U., and Miirphy, K. M. (1997). Regulation of the interleukin (IL)-12R beta 2 subunit expression in developing T helper 1 (Thl) and Th2 cells. I. E x p . Men'. 185, 817-824. Szday, G., Hess, J., and Kaufniann, S. H. E. (1994).Presentation of Listeria monocytogenes antigens by major histocompatibility class 1 inolecules to CD8 cytotoxic T lymphocytes independent of listeriolysin secretion and virulence. Eur. J. Imn~ztnol.24, 1471-1477. Szalay, G . , Ladel, C. H., and Kaufniann. S. H. E. (1995).Stimulation of protective CD8' T lyinphocytes by vaccination with nonliving bacteria. Proc. Nat. Acad Sci. USA 92, 12389- 12392. Sztein, M. B.. Tanner, M. K., Polotsky, Y., Orenstein, J. M., and Levine, M. M. (199Fj). Cytotoxic T lymphocytes after oral imrnrinization with attenuated vaccine strains of Saltnonella t y p h in humans. J. Imrmrnol. 155, 3987-3993. Takahashi, H.. Takesliita, T., Morein, B., Putney, S.,Germain, R. N., and Berzofsky, J. A. ( 1990). Induction of CD8' cytotoxicT cells by immunization with purified HJV-1 envelope protein in ISCOMs. Nature 344, 873-875. Takai, Y., Wong, G. G., Clark, S. C., Burakoff, S. J., and Herrmann, S. H. (1988). B cell stirnulatory factor-2 is involved in the differentiation of cytotoxic T lymphocytes./. Itnrrzrrnol. 140, 508-512. Takeda, K., Tsutsui, H., Yoshimoto, T., Adachi, O., Yoshida, N., Kishimoto. T., Okamura, H., Nakanishi, K., and Akira, S. (1998). Defective N K cell activity and Th1 response in IL-18 deficient mice. Iinrnrrnity 8, -383-390. Tan, M . C., Mommaas, A. C., Drijfhout. J . W., Jordens, R.. Ondenvater, J. J., Venvoerd, D., Mulder, A. A., van der Heiden, A. N., Scheidegger, D., Oomen, L. C., Ottenlioff, T. H., Tulp, A., Neefjes, J. J., and Koning, F. (1997). Mannose receptormediated uptake of antigens strongly enhances HLA class 11-restricted antigen presentation by cultured dendritic cells. Eur. 1. Zrttiniinol. 27, 2426-2435. Tanaka, Y., Morita, C. T., Nieves, E., Brenner, M. B., and Bloom, B. R. (1995). Natural and synthetic non-peptide antigens recognized by human y6T cells. Nature 375,155-158. Tanaka, Y., Sano, S., Nieves, E.. DeLibero, G . , Rosa, D., Modlin, R. L., Brenner, M. B., Bloom, B. R., and Morita, C. T. (1994).Nonpeptide ligands for human yS T cells. Proc. Nat. Acacl. Sci. USA 91, 8175-8179. Tang, D. C., DeVit, M., and Johnston, S. A. (1992). Genetic immunization is a simple method for eliciting an iinrnune response. Nature 356, 152-154. Taraska, T., Ward, D. M., Ajioka, R. S., Wyrick, P. B., and Davis-Kaplan, S. R. (199G).The late chlamydial inclusion membrane is not derived from the endocytic pathway and is relatively deficient in host proteins. Infect. lmmun. 64, 3713-3727. Tascon, R. E., Colston, M. J., Ragno, S., Stavropoulos, E., Gregory, D., and Lowrie, D. B. (1996). Vaccination against tnherculosis by DNA injection. Nature Med. 2, 888-892.
IhTRAC'ELL.IILAH RA(,TEKIA AYII THE I M M U Y E \\STEM
373
Taylor, M. E:.. and Drickainer, K. (199:3).Structural requirements for high affinity binding of complex ligands by the macrophage mannose receptor. 1. Biol. Chent. 268, 399-404. Teitell, M., 1Iolcoinbe. H. I. J O H N E. VOIAN.AKIS, NICOLEM. T H I E L E N S , STIIAN.AM V. L. NARAYANA, VI~K)NIQU Hossi, E \NU YLIANYUAN Xu 3x5
386
CONTENTS OF RECENT VOLUMES
Accessibility Control of V(D)J Recombination: Interleukin-18: A Novel Cytokine That Augments Both Inirate and Acquired Lessons from Gene Targeting Iinninnity WILLIAM M. HEMPEL,ISABELLE LEDUC, HARUKI OKAMURA, HIHOKCI TSUTSUI, NOELLEMATHIEU, RAJKAMAL TRIPATHI, SIIIN-ICHINO KASHIWAMUHA, TOMOHIRO A N D PIERRE FEHIiIEH YOSHIMOTO,A N D KENJINAKANISHI Interactions between the Iininunc System and Gene Therapy Vectors: Bidirectional CD4’ T-cell Induction and Effector Regulation of Response and Expression Functions: A Comparison of Immunity J O N A T H A N S. RHOMBEHC;,LISADEBHUYNE, against Soluble Antigens and Vird Infections A N D L I H l i l QIN ANNETTEOXENIUS, ROLF M. ZINKERNACEL, A N D HANSHENCARTNEH Major Histocompatibility Complex Genes Influence Individual Odors and Mating Preferences DUSTIN PENNA N D WAYNE Pons Olfactory Receptor Gene Regulation ANDREWCHESS
INDEX
Volume 70 Biology of the Interleukin-2 Receptor BRADH. NELSON A N D DENNIS M WILLERFOIIU Interleukin-12: A Cytokine at the Interface of Inflammation and Iininiinity G i o m o THIN(:IIIERI Recent Progress on the Regulation of Apoptosis by Bcl-2 F m i l y Members ANDYJ. MINK, RACHEL E. SWAIN, AVERIL MA,A N D CRAIG B. THOMP60N
Ciirrent Views in Intracellular Transport: Insights from Studies in Iinmunology VICTORW. HSU AN11 PETERJ. PETERS Phylogenetic Emergence and Molecular Evolution of the Iminunoglohiilin Family J O H N J. MAHCHALONIS, SAMUEL F. SCHLIITER, RALPHM. BERNSTEIN, SHANXIANC SHEN,A N D ALLENR. EDMUNDSON Current Insights into the “Antiphospholipid’ Syndrome: Clinical, Immunological, and Molecular Aspects A. KANDIAH,ANDRE] SALI, DAVID Y O N G H U A SHENG, EDWARD 1. VICTORIA, DAVID M. MARQUIS, STEPHEN M. COUTTS, A N D STEVEN A. KHILIS
INDEX
This Page Intentionally Left Blank