The Anatomy of T-Cell Activation and Tolerance Anna Mondino*T, Alexander Khoruts*, and Marc K

Home , Clonal deletion, Germinal center, Immune tolerance, Naive T cell, Peripheral tolerance, T cell

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 2245-2252, March 1996 Review The anatomy of T-cell activation and tolerance Anna Mondino*t, Alexander Khoruts*, and Marc K. Jenkins Department of Microbiology and the Center for Immunology, University of Minnesota Medical School, 420 Delaware Street S.E, Minneapolis, MN 55455

ABSTRACT The mammalian im- In recent years, it has become clear that TCR is specific for a self peptide-class I mune system must specifically recognize a full understanding of immune tolerance MHC complex) T cell that will exit the and eliminate foreign invaders but refrain cannot be achieved with reductionist in thymus and seed the secondary lymphoid from damaging the host. This task is vitro approaches that separate the individ- tissues (3, 4). In contrast, cortical CD4+ accomplished in part by the production of ual lymphocyte from its in vivo environ- CD8+ thymocytes that express TCRs that a large number of T lymphocytes, each ment. The in vivo immune response is a have no avidity for self peptide-MHC bearing a different antigen receptor to well-organized process that involves mul- complexes do not survive and die by an match the enormous variety of antigens tiple interactions of lymphocytes with each apoptotic mechanism. Cortical epithelial present in the microbial world. However, other, with bone-marrow-derived antigen- cells are essential for the process of pos- because antigen receptor diversity is gen- presenting cells (APCs), as well as with itive selection because they display the self erated by a random mechanism, the im- nonlymphoid cells and their products. The peptide-MHC complexes that are recog- mune system must tolerate the function of anatomic features that are designed to op- nized by CD4+ CD8+ thymocytes and also T lymphocytes that by chance express a timize immune tolerance toward innocuous provide essential differentiation factors self-reactive antigen receptor. Therefore, self antigens and increase the efficiency of (5). Positively selected T cells continue during early development, T cells that are recognition and elimination of pathogens their journey toward the central region of specific for antigens expressed in the thy- are the subjects of this review. the thymus by first crossing the cortico- mus are physically deleted. The popula- medullary junction and then entering the tion of T cells that leaves the thymus and Thymic Tolerance medulla. The medulla contains the prod- seeds the secondary lymphoid organs ucts of positive selection-i.e., mature contains helpful cells that are specific for The thymus is the primary lymphoid organ CD4+ CD8- or CD4- CD8+ T cells, that antigens from microbes but also poten- where immature precursors develop into exit the thymus into the blood and seed tially dangerous T cells that are specific mature T cells (1). It is also the site where the secondary lymphoid tissues, where for innocuous extrathymic self antigens. immature T cells that express self-reactive they will be available to be activated by a The outcome of an encounter by a periph- T-cell antigen receptors (TCRs) are elim- foreign peptide-MHC complex. eral T cell with these two types ofantigens inated (2). The thymus is organized into The population of positively selected is to a great extent determined by the three physically distinct areas: the outer cortical T cells will include cells that have inability of naive T cells to enter nonlym- subcapsular zone, the cortex, and the in- a strong avidity for self peptide-MHC phoid tissues or to be productively acti- ner medulla (1). The subcapsular zone is a complexes that are expressed on thymic vated in the absence of inflammation. thin space that lies directly below the outer APCs. These T cells are potentially auto- capsule that encases the thymus. This area reactive and must be eliminated. Work Current evidence indicates that self- contains the most immature CD4- CD8- from many laboratories has shown that reactive lymphocytes can be inactivated at T-cell precursors that have recently mi- this occurs via physical deletion involving all stages of their development by a variety grated to the thymus from the bone mar- apoptosis (6). Experiments with TCR of mechanisms. Perhaps the best under- row. From the subcapsular zone, the im- transgenic mice that also express the an- stood is clonal deletion of lymphocytes at mature T cells migrate deeper into the tigen the transgenic T cells are specific for an early stage of development in the pri- thymus and enter the cortex, which is have shown that clonal deletion can occur mary lymphoid organs- e.g., T cells in the richly populated by a special type of epi- either early in the CD4+ CD8+ stage in thymus-if their antigen receptors have a thelial cell that expresses high levels of the cortex (7, 8) or later in the CD4+ high affinity for self antigens expressed class I and II major histocompatibility CD8+ stage in the medulla (9, 10). The there. However, this process alone cannot complex (MHC)-encoded molecules. In site where a given self-reactive T cell will account for tolerance that exists toward the cortex, the T-cell precursors begin to be deleted in the thymus is probably de- antigens that are not well expressed in the express both CD4 and CD8 and to rear- termined by the location of the cells that primary lymphoid organs. Examples of range their TCR a and 13 genes such that present the relevant self peptide-MHC such antigens are proteins that are ex- each cell will express a different rear- complex (6). In cases where T cells are pressed exclusively in the parenchymal ranged TCR-a(3 heterodimer. The speci- deleted early in the CD4+ CD8+ stage in tissues, proteins that are only expressed at ficity of the TCR expressed by a given the cortex, the relevant complex is prob- certain developmental stages of the or- CD4+ CD8+ thymocyte for a peptide- ably presented by cortical epithelial cells. ganism, and the infinite variety of other- MHC complex expressed on another cell Because blood-borne macromolecules wise innocuous environmental proteins (the APC) will determine the thymocyte's fate. present in the air and diet. Because pep- Evidence from fetal thymic organ Abbreviations: APC, antigen-presenting cell; tides derived from these proteins are pre- cultures suggests that a CD4+ CD8+ thy- HEL, hen egg lysozyme; IL, interleukin; MHC, sented to mature T cells outside of the mocyte that expresses a TCR with low but major histocompatibility complex; PALS, peri- thymus, peripheral tolerance cannot be measurable avidity for a self peptide- arterial lymphatic sheath; TCR, T-cell antigen explained by a developmentally regulated MHC complex will be "positively select- receptor; TNF, tumor necrosis factor; TGF-f3, to as in case of that it will transforming growth factor f; HEV, high en- susceptibility deletion, the ed," is, be allowed to differen- dothelial venule. immature T cells, but rather depends on tiate into a mature CD4+ CD8- (if its *A.M. and A.K. contributed equally to the factors that are extrinsic to the self- TCR is specific for a self peptide-class II preparation of this manuscript. reactive T cell. MHC complex) or a CD4- CD8+ (if its tTo whom reprint requests should be addressed. 2245 Downloaded by guest on September 28, 2021 2246 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996) have poor access to the cortex (11), cor- blown reactivity to extrathymic self anti- their activation. The spleen has a similar tical clonal deletion is probably most rel- gens is to be prevented. organization but is unique among second- evant for antigens that are produced by ary lymphoid organs because it lacks af- the cortical epithelial cells themselves- Lymphocyte Migration Through ferent lymphatic drainage and is thus ded- e.g., MHC molecules and housekeeping Secondary Lymphoid Tissues icated to process blood-borne antigens proteins-or are produced in large exclusively. The spleen is divided into the amounts by other cells in the cortex-e.g., After T cells leave the thymus, they recir- red pulp, which contains red blood cells thymic hormones. In several cases where T culate between the blood and lymph with and macrophages, and the white pulp, cells are deleted in the medulla, the rele- intermediate stops in secondary lymphoid which has T- and B-cell-rich areas like the vant antigen is preferentially produced by tissues (13) (lymph nodes, Peyer's patches, lymph nodes (Fig. 1B). A central artery bone marrow-derived medullary APCs (9) and the spleen) (1). The secondary lym- enters the spleen and branches into arte- or is preferentially picked up from the phoid tissues are designed to co-localize rioles that end in splenic sinuses present in serum, processed, and presented by med- antigen, APCs, and naive T and B cells. the red pulp. The arterioles dump their ullary APCs (10). This process is facili- Peripheral lymph nodes are strategically contents of blood cells and fluids into the tated by the concentration of interdigitat- located throughout the body to collect and sinuses, providing the major means by ing dendritic cells, the most potent induc- process antigens arriving via the lymphatic which blood-borne material enters the ers of clonal deletion (12), at the vasculature. Lymphatic vessels originate spleen. T cells and interdigitating den- corticomedullary junction (1). as closed-end capillaries that collect tis- dritic cells are concentrated along the sue-derived lymph fluid and lymphocytes. length of the arterioles to form the peri- The Peripheral Tolerance Paradox Not surprisingly, the richest networks of arterial lymphatic sheath (PALS) that lymphatic vessels are found in sites of constitutes the T-cell-rich zone of the Because clonal deletion of self-reactive T greatest foreign antigenic load-e.g., skin spleen and is analogous to the lymph node cells in the thymus is such an efficient and mucosal surfaces (1). Lymph and paracortex. Follicles rich in B cells and process, one might speculate that this is lymphocytes are carried into lymph nodes follicular dendritic cells are attached to by afferent lymphatic vessels, pass through the PALS at various points along its the only T-cell tolerance mechanism that the lymph node cortex and medulla and length, and both the follicles and PALS is required. However, not all self antigens exit via efferent lymphatic vessels (Fig. are surrounded by the marginal zone. The are expressed in, or gain access to, the 1A). This flow pattern ensures that lymph- marginal zone is a unique structure of the thymus. Self antigens that are not ex- borne antigens will percolate through the spleen and consists of a mixture of T and pressed in the thymus are by definition densely clustered lymphocytes and APCs B cells, including memory B cells. expressed in areas of the body that may be that make up the inner part of the lymph Secondary lymphoid tissues are not patrolled by mature T cells. Tolerance to node, optimizing the chances that effec- static structures; rather, lymphocytes are these types of antigens is particularly tive antigen presentation will occur. Mac- constantly coming in and going out. This problematic because, unlike immature T rophages line the subcapsular space in a is an essential feature of the immune cells that are programmed to die when position where they can phagocytose par- system, which must move its lymphocyte stimulated, peripheral T cells are designed ticles and produce proinflammatory cyto- repertoire through the secondary lym- to respond in a positive fashion. There- kines that, as detailed below, play a key phoid tissues where lymph- or blood- fore, the challenge of the immune system role in adaptive immunity. Interdigitating borne antigenic material will be captured is to retain the capacity to activate periph- dendritic cells, the most potent APCs at and presented. Naive-i.e., cells that have eral T cells that are specific for microbial activating naive T cells (14), are inter- never been stimulated-T and B cells use antigens but silence the functions of pe- spersed throughout the central paracortex molecules like L-selectin to exit the blood ripheral T cells that are specific for extra- that also contains mainly T cells plus a few and enter lymph nodes by binding to spe- thymic self antigens without being able to B cells. Most B cells are located in follicles cialized blood vessels called high endothe- delete these T cells in the thymus. It is in close association with follicular den- lial venules (HEVs) that are located in the therefore important to define the events dritic cells, although a few T cells are also paracortex (15). In contrast, naive lym- that result in effective antimicrobial T-cell present. A key feature of this arrangement phocytes enter the splenic white pulp via immunity if we are to understand the is the juxtaposition of T and B cells with an HEV- and L-selectin-independent critical steps that must be avoided if full- the dendritic cell types that are critical for mechanism, probably by leaving the red B 0D 01

TCell Resting B Cell 0O.- Germinal Center B Cell 0 Macrophage Follicular Dendritic Cell - Perartenal MarginalMarginaZoneZon > Lymphatic Sheath Interdigitating Dendritic Cell

FIG. 1. Schematic depiction of the structure of a lymph node (A) and an area of splenic white pulp (B). The white pulp is shown as a transverse cross-section through one branch of the central arteriole. Downloaded by guest on September 28, 2021 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996) 2247 pulp sinus and directly entering the mar- and process free antigen that enters the Surprisingly, the initial interaction be- ginal zone (1, 15, 16). Naive T cells are lymph node and spleen, many studies in- tween splenic antigen-specific B cells and deposited into the T-cell zones, whereas dicate that antigen-bearing, interdigitat- helper T cells occurs in the T-cell-rich naive B cells migrate to primary follicles. ing dendritic cells are essential for the PALS (25-27), a site where very few B Both cell types remain in a given second- initial activation of naive T cells (14). In cells reside. ary lymphoid tissue site for 10-20 h, and, addition, the immature dendritic cells that Some of the activated T (28, 29) and B if they do not encounter their cognate migrate from the tissue may be critical cells (30) then migrate from the T-cell antigen, they leave and recirculate to an- because these types of cells have been zone into follicles to initiate the germinal- other secondary lymphoid tissue (17). As shown to take -up and process large center reaction. In the follicles, activated described below, the tropism of lympho- amounts of antigen, a property that is lost CD4+ T cells produce critical lymphocytes for secondary lymphoid tissues is in mature dendritic cells (14, 21). kines that promote B-cell proliferation lost following activation and replaced with The co-localization of dendritic cells and isotype switching and express CD40 a tropism for nonlymphoid tissues. and naive T cells in the paracortex and ligand that binds to CD40 on the B cells. PALS increases the chances that naive T Immunohistochemical analyses have lo- T-Cell Activation by will find calized CD40 ligand-expressing, lympho- Peripheral cells with the appropriate TCRs kine-producing T cells in close proximity Microbial Antigens an antigen-bearing APC. The interaction to antibody-forming B cells in the PALS of CD28 (22) expressed by the T cell and and follicles (31). CD40 signaling in B cells As diagrammed in Fig. 2A, introduction of B7-1 or B7-2 (22) expressed by the den- has been shown to be important for stim- foreign antigen dramatically alters the be- dritic cells plays an important role in the ulating B-cell proliferation (32, 33) and havior of those lymphocytes within sec- clonal expansion of naive T cells in the antibody production (32) and for enhanc- ondary lymphoid tissue that are specific paracortex. Paracortical dendritic cells ex- ing the T-cell costimulatory activity of B for the antigen in question. Foreign anti- press B7-2 molecules in situ (23), and cells (34). CD40 ligand has been recently gen that is deposited in tissue may travel immature tissue dendritic cells increase shown to deliver an activation signal to T passively to the nearest lymph node byway expression of B7 molecules during their cells that is important for their clonal of the draining afferent lymphatic vessels migration from tissue to the lymph node expansion (35, 36). The importance of the (1). Alternatively, immature dendritic (24). Once activated by antigen-bearing CD40-CD40 ligand interaction is demon- cells present in the tissue-e.g., Langer- dendritic cells, the specific T cells prolif- strated by the findings that primary and hans cells of the skin-may take up the erate in the paracortex and become com- secondary antibody responses and germi- antigen and, in response to inflammatory petent to receive further activation signals nal-center formation are blocked by anti- cytokines, migrate via the afferent lymph from antigen-bearing macrophages and B CD40 ligand antibody treatment in vivo to the node (18-20). Antigen that is cells (14). The latter population includes (37, 38) and that CD40 ligand-deficient present in the blood will be taken up by antigen-specific B cells that use their sur- mice (39) and humans (40) fail to produce red pulp macrophages or diffuse from the face immunoglobulin molecules to effi- IgG or form germinal centers. red pulp into the white pulp. Although B ciently internalize and process the antigen The proliferating B-cell blasts push cells and macrophages probably take up and present peptide-MHC complexes. nonactivated follicular B cells out of the

--.- Skin =..Ie .r-':

/ I a~ICeIII,

Non-Lymphoid, - Tissues / Folic

Peptide!' Antigen CD40 MHC B7 AMie o F o * ymphokines Proinflammatory Cytokines CD4CL TCR CD2B

FIG. 2. Diagrammatic representation of the events associated with the induction of T-cell immunity by injection of antigen in complete Freund's adjuvant (CFA) (A) or peripheral tolerance by soluble antigen alone (B). FDC, follicular dendritic cell. Downloaded by guest on September 28, 2021 2248 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996) way and begin to form the characteristic Developmentally Regulated Antigens node follicles. After day 5, the antigen- germinal center, composed of a dark zone specific T cells gradually disappeared containing actively proliferating B-cell If autoimmunity is to be prevented, the from the lymph nodes. Adoptively trans- blasts that are undergoing somatic hyper- aforementioned sequence of events must ferred mice primed in this way produced mutation of immunoglobulin heavy and be avoided when peripheral self antigens high levels of antigen-specific IgG and light chain genes and a light zone contain- are presented. This situation could occur developed a local delayed-type hypersen- ing nonmitotic B cells (41, 42). Antigen- when developmentally regulated self an- sitivity reaction when challenged with an- antibody complexes, produced in the T- tigens, for example proteins associated tigen, suggesting that humoral and cellular cell zone by B cells during their initial with pregnancy, lactation, puberty, and immunity had been induced (M.K.J., un- interaction with T cells, are bound by aging, appear in the blood for the first time published observation). In contrast, when follicular dendritic cells (43). These com- after mature T cells have seeded the sec- adoptively transferred mice were injected plexes serve as a source of antigen for ondary lymphoid organs. In these cases, it intravenously with soluble antigen alone, selection of the highest affinity B cells, is likely that new proteins are presented to the antigen-specific T cells proliferated presumably those cells that have acquired a mature T cell repertoire that has not had transiently in the paracortical regions of the most beneficial immunoglobulin so- the chance to purge itself ofT cells specific peripheral lymph nodes but never went matic mutations. It is thought that these B for peptide-MHC complexes derived into follicles, and most of the T cells cells capture antigen very efficiently and from these proteins. Thus, these peptide- rapidly disappeared. These animals did thus outcompete other antigen-binding B MHC complexes would be expected to be not produce antibody (M.K.J., unpub- cells for immunoglobulin- or CD40-medi- just as "foreign" to the immune system as lished observation). Given the depen- ated survival signals provided by follicular those derived from microbes. Clues about dence of germinal-center formation on dendritic cells and helper T cells. The ger- a possible explanation for this paradox CD4+ T cells (39, 40, 52), it is likely that minal-center reaction peaks about 2 weeks come from experiments performed by this failure is related to the inability of after initial antigen exposure and wanes Dresser (46, 47), which showed that the antigen-specific T cells to enter the folli- thereafter (26), due in part to the dying off immunogenicity of purified antigens was cles. When adoptively transferred mice ofmany ofthe antigen-specific T and B cells. determined by factors in addition to for- that were first injected with antigen alone However, by way of a poorly understood eignness. He showed that even foreign were rechallenged with antigen in adju- process, some of the antigen-specific T and antigens failed to induce immunity unless vant, the few remaining TCR transgenic T B cells survive to become memory cells that they were injected with an adjuvant (46). cells proliferated poorly in the draining are capable of mounting rapid and efficient Adjuvants are components of microbes, lymph nodes, suggesting that their activa- secondary responses (44). for example lipopolysaccharide or mu- tion was impaired. A very similar set of During the course of the primary im- ramyl dipeptide, that induce inflammation observations was reported by Kyburz et al. mune response, activated T cells lose L- (48). Dresser (47) and others (49) also (53) using CD8+ TCR transgenic T cells selectin that is involved in lymph node showed that injection of foreign proteins specific for a viral peptide-class I MHC entry and increase expression of adhesion without an adjuvant induced a state of complex. Therefore, in these situations, molecules on their surface-e.g., LFA-1 long-lasting immunological tolerance such even though the antigen (chicken ovalbu- and VLA-4-that permit extravasation that subsequent injection of the same pro- min or viral protein) was completely for- into nonlymphoid tissues (44). The traf- tein together with an adjuvant failed to eign to the host, administration of the ficking of activated lymphocytes into tis- induce immunity. These results suggested antigen under noninflammatory condi- sues is governed by interactions with vas- that: (i) adjuvants create an inflammatory tions caused an abortive type of T-cell cular endothelium in inflamed tissues in vivo environment that is conducive to activation that resulted in tolerance in- (45). Inflammatory mediators and cyto- adaptive immunity; and (ii) in the absence stead of immunity. kines-e.g., interleukin 1 (IL-1) and tu- of this environment, antigen exposure re- A scenario by which adjuvants shift the mor necrosis factor a (TNF-a)-pro- sults in tolerance. According to this model T-cell response toward immunity (Fig. duced by macrophages and parenchymal (50, 51), developmentally regulated self 2A) instead of peripheral tolerance (Fig. cells in response to microbial molecules, proteins would not cause autoimmunity 2B) is shown in Fig. 2. As noted above, the induce the expression of selectins, integrin because they would be presented to T cells CD28-B7 and CD40 ligand-CD40 inter- ligands, and immobilized chemokines on in the absence of inflammation, whereas actions play key roles in the productive nearby vascular endothelial cells. There- microbial antigens would induce immu- immune response, and, thus, adjuvants fore, previously activated T cells that ex- nity because they would always be pre- may work by causing the induction of B7 press higher levels of LFA-1 and VLA-4 sented to T cells in the presence of in- on APCs or CD40 ligand on the respond- are preferentially recruited into inflamed flammation caused by components of ing T cells. Microbial products have been tissues, although in cases of chronic in- their cell walls or membranes. shown to induce B7 and other costimula- flammation, naive T cells can also enter To study this situation in more detail, tory molecules on normally B7-negative because the local vascular endothelium our laboratory directly tracked the fate of resting B cells (54), and inflammatory takes on the properties of HEV (15). The antigen-specific T cells following in vivo cytokines produced in response to micro- net effect of this process is to allow anti- exposure to a soluble antigen alone or to bial products cause immature dendritic gen-specific lymphocytes to migrate to the that same antigen emulsified in Freund's cells to migrate from the inflamed site to actual sites of inflammation where the adjuvant (29). A low but detectable num- the regional lymph nodes and express high invading pathogen has entered the body. ber of TCR transgenic T cells were trans- levels of B7 (18-20, 24). Therefore, in the In summary, T-cell immunity to micro- ferred into normal syngeneic recipients, absence of inflammation, fewer costimu- bial antigens depends on inflammation at and the number and location of the trans- latory APCs would be predicted to be the site of entry; co-localization of micro- ferred T cells was followed by immuno- present in the draining lymph nodes. Sev- bial antigen, APCs, and naive T cells in histology or flow cytometry by using an eral lines of evidence indicate that this secondary lymphoid tissues; and clonal anti-clonotypic antibody that recognizes situation favors T-cell tolerance. Preven- expansion and survival of antigen-specific only T cells that express the transgenic tion of CD28-B7 interactions in vivo has T cells, followed by their migration into TCR. Following antigen injection in ad- been shown to inhibit T-cell immunity follicles and then into nonlymphoid tis- juvant, antigen-specific T cells prolifer- (55-59) and result in antigen-specific T- sues. As outlined below, peripheral tolerated extensively in the paracortical region cell tolerance in several (56-58), but not ance can operate at any of these steps de- of the draining lymph nodes and after all (55), instances. Injection of B7- pending on the nature of the self antigen. several days appeared in draining lymph negative resting B cells induces T-cell tol- Downloaded by guest on September 28, 2021 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996) 2249 erance in vivo (60, 61), although it has question (H-Y-expressing male cells) can- germinal center) and dies or becomes been argued (61) that this is related to a not be readily cleared. Functional unre- unresponsive. The lengths to which the tolerogenic program that occurs in naive T sponsiveness related to TCR and CD8 immune system has gone to exclude po- cells whenever they are stimulated by an modulation may also be involved in im- tentially self-reactive lymphocytes from APC other than a dendritic cell and not by munological tolerance to the fetus. Preg- follicles suggests that the follicle provides the poor costimulatory function of resting nancy-associated loss of TCR and CD8 an essential microenvironment for the B cells. However, the ability of B cells to has been described for CD8+ TCR trans- generation of memory T and B cells. The induce T-cell tolerance is greatly en- genic T cells in pregnant mice that carry a actual mechanism by which T or B cells hanced if recipient mice are also treated fetus expressing the class I MHC molecule are excluded from follicles is unclear, how- with anti-CD40 ligand antibody (62) or if for which the transgenic TCR is specific ever, although ineffective competition for the B cells are CD40-deficient (63). One (68). Tolerance does not persist in the follicular niches and failure of the produc- explanation for this is that inhibition of mother postpartum, suggesting that the tion of a chemoattractant by follicular CD40 signaling prevents the induction of tolerogens are fetal cells that migrate into cells are possibilities. B7-dependent T-cell costimulatory activ- the mother. Although it is possible that In summary, the bulk of the evidence to ity in the B cells (34). An alternative tolerance in these cases is caused by anti- date suggests that T-cell tolerance specific explanation, however, is that CD40 ligand gen presentation in a non-inflammatory for developmentally regulated antigens delivers a critical activation signal to the T environment as diagrammed in Fig. 2B, it (antigens that appear after a mature T-cell cell that prevents death or induction of is equally possible that chronic stimulation repertoire has formed) is induced in sec- functional unresponsiveness (35, 36). In- by antigens that cannot be easily cleared ondary lymphoid tissues as a consequence dependent of the mechanism, the ability results in "exhaustion" of the T cells (69). of an abortive form of T-cell activation of B-cell antigen presentation in the ab- As noted above, antigen-specific CD4+ that occurs when an inflammatory envi- sence of dendritic-cell antigen presenta- T cells did not enter follicles following a ronment is absent. tion to induce T-cell unresponsiveness tolerogenic injection of antigen in the may be an important tolerance mecha- absence of an adjuvant (29). A similar type Nonlymphoid Antigens nism for self antigens expressed only by B of follicular exclusion has been shown for cells-e.g., peptide-MHC complexes de- tolerized antigen-specific B cells. Good- A second class of antigens that peripheral rived from variable components of immu- now et al. (70) have produced double T cell tolerance must deal with are pro- noglobulins (61). transgenic mice that express hen egg ly- teins expressed exclusively in nonlym- Adjuvants may also facilitate T-cell im- sozyme (HEL)-specific immunoglobulin phoid tissues, for example proteins ex- munity by stimulating cells of the innate on all of their B cells and also express a pressed only in the pancreas, liver, or immune system to produce inflammatory soluble form of HEL. HEL-specific T cells kidney. Because these proteins are not cytokines that promote T-cell survival. are tolerant in these mice and, thus, the expressed in the thymus, there is no op- Injection of superantigens that directly transgenic B cells are constantly exposed portunity to clonally delete T cells specific crosslink the TCR-I3 chain on T cells and to HEL in the absence of T-cell help. The for peptide-MHC complexes derived class II MHC molecules on APCs induces transgenic B cells are not deleted but from them. transient T-cell proliferation followed by become functionally unresponsive be- The first line of protection for these unresponsiveness and death (64), as in the cause of a proximal block in secreted types of self antigens is related to the fact case of peptide antigen injection in the immunoglobulin signaling (71). When that naive T cells are relatively confined to absence of adjuvant (29, 53). Marrack and transferred into HEL, single transgenic blood and secondary lymphoid organs and coworkers (65) have shown that the T cells mice that possess a diverse B-cell reper- cannot efficiently enter uninflamed non- that express the appropriate TCR-13 toire, the unresponsive B cells from the lymphoid tissues (13). Therefore, nonlym- chains survive much longer if lipopolysac- double transgenic mice accumulated in phoid antigens are tolerated, in part be- charide is injected along with the super- the T-cell zones, did not enter follicles, cause naive T cells do not get access to the antigen. TNF-a appears to be partially and died shortly thereafter (72). Because tissues that express them. This could ex- responsible for this effect. In addition, the this phenomenon was not observed in the plain the results of Ohashi et al. (76), who disappearance of T cells that is induced by intact, double transgenic mice, the authors studied double transgenic mice expressing the injection of antigen in the absence of concluded that the follicular exclusion of lymphocytic choriomeningitis virus glyco- adjuvant does not occur if the T cells lack the transferred unresponsive B cells was protein (LCMV-GP) in the B3 islet cells of Fas (66), implicating Fas-Fas ligand inter- due to failed competition with normal B the pancreas and also a TCR specific for actions in this peripheral deletion pathway. cells. These investigators (73) and others LCMV-GP/H-2Db on all CD8+ T cells. Tolerance also results when CD8+ T (74, 75) also demonstrated that activated Despite the fact that most of the T cells in cells are suddenly exposed to their cog- germinal-center B cells that are suddenly these animals were specific for LCMV-GP nate antigen in the periphery. Rocha and exposed to a cross-reactive antigen for peptide-H-2Db, they did not develop di- von Boehmer (67) showed that when which there is no T-cell help leave the abetes. Tolerance was not due to deletion CD8+ T cells from female donors express- germinal center and enter the T-cell zone or T-cell anergy since the functional re- ing a transgenic TCR specific for the male but are unable to reenter the germinal sponse of transgenic T cells, when mea- H-Y antigen were transferred into male center and die. This is probably an impor- sured in vitro, was the same as that of T recipients, the T cells proliferated tran- tant tolerance mechanism that eliminates cells from TCR transgenic mice that did siently, and then most of the cells died B cells that acquire immunoglobulin so- not express LCMV-GP. When the doubly leaving behind a population that was func- matic mutations that confer reactivity for transgenic animals were infected with tionally unresponsive due to loss of sur- a self antigen present in the lymphoid LCMV, the pancreas was infiltrated with face TCR and CD8 molecule expression tissue. Therefore, in three different tolero- TCR transgenic T cells and the f3 cells (66). These surface molecule changes genic situations-activation of naive T cells were destroyed. These results are consis- were not noted in the aforementioned in the absence of inflammation, stimula- tent with the idea that the naive TCR cases where peripheral tolerance was in- tion of naive T-cell zone B cells in the transgenic T cells did not enter the pan- duced by the injection ofsoluble antigen in absence of T-cell help, or confrontation of creas and thus ignored the antigen- the absence of an adjuvant (29, 53). Loss activated germinal center B cells with a bearing R cells. However, activation of the of TCR and accessory molecules may be a cross-reactive antigen in the absence of T T cells in secondary lymphoid tissue by the consequence of chronic stimulation be- cell help-the specific lymphocyte popu- virus infection would be expected to cause cause in the H-Y system the antigen in lation is excluded from the follicle (or the expression of adhesion receptors re- Downloaded by guest on September 28, 2021 2250 Review: Mondino et aL Proc. Natl. Acad. Sci. USA 93 (1996) quired for extravasation, thus allowing the CD4+ naive T-cell specific for a self pep- taining immune privilege (94). For T cells to enter the pancreas and destroy tide-class II MHC complex were to wan- example, Sertoli cells (95) and ocular tis- the , cells. der into a nonlymphoid tissue, the T cell sues (96) constitutively express Fas ligand Although the results mentioned above would not be activated because these tis- and can kill previously activated T cells suggest that it is limited, some degree of sues normally do not express class II MHC that normally express Fas. Thus, Bellgrau surveillance of most nonlymphoid tissues molecules. Evidence for a lack of consti- et al. (95) showed that, while allogeneic by naive T cells probably does occur. A tutive self antigen-class II MHC presen- testis tissue transplanted under the kidney peripheral tolerance mechanism would be tation is provided by the finding that trans- capsule survived for weeks without any required to deal with this, especially for genic B7 expression alone in pancreatic (3 evidence of rejection or lymphocyte infil- CD8+ T cells that recognize class I MHC cells (80-82) or keratinocytes (83, 84) tration, allogeneic testis tissue derived molecules that are expressed on most non- does not induce CD4+ T-cell-mediated from Fas ligand-deficient mutant gld mice lymphoid tissues. Evidence for peripheral autoimmunity. It is therefore likely that was promptly rejected. Similarly, injection tolerance in this situation comes from ignorance is the major mechanism of tol- of herpes simplex virus type 1 into the studies by Arnold and coworkers (77-79) erance for class II MHC-restricted, tissue- anterior chamber of the eye resulted in who have constructed doubly transgenic specific peptides in normal individuals. initial inflammatory infiltration of the eye mice expressing TCR specific for class I However, class II MHC molecules are in normal and gld mice (96). However, MHC Kb molecules on all CD8+ T cells upregulated on many nonlymphoid co- while extensive apoptosis in infiltrating and also Kb in the liver but not the thymus. stimulation-deficient tissues during in- cells and little tissue damage were subse- The absence of Kb expression in the thy- flammation, and this could serve as a quently seen in normal mice, the course in mus was shown by lack of its detection by means of terminating chronic T-cell re- gld mice was characterized by expansion of immunohistology and reverse transcrip- sponses. The finding that constitutive ex- the inflammatory infiltrate and significant tase-PCR. The Kb-reactive TCR trans- pression of B7-1 by keratinocytes greatly tissue damage. Soluble mediators may genic cells had reduced levels of TCR and increases and prolongs cutaneous de- also contribute to the immunosuppressive CD8 expression and failed to respond to layed-type hypersensitivity reactions to environment found within immune- Kb in vivo or in vitro. Therefore, tolerance exogenously applied contact sensitizers is privileged sites. For instance, ocular fluids was probably induced as naive TCR trans- consistent with this possibility (83, 84). contain high concentrations oftransforming genic T cells recognized Kb in the liver. growth factor (3 (TGF-(3), as well as certain Although it is possible that tolerance Antigens Expressed in Privileged Sites immunosuppressive neuropeptides (97). could have been induced in regional Similar immunosuppressive microenviron- lymph nodes, this seems unlikely since it Some organs-e.g., brain, gonads, and ments are also present in the brain, testis, would have required that the T cells rec- eye-are considered immune-privileged and placenta (97). Preferential skewing of ognized passively acquired Kb on the sur- sites because foreign antigens that are the immune response toward humoral face of a lymph node APC. Naive T-cell placed within them do not provoke an rather than cell-mediated immunity to an- surveillance may be higher in tissues like immune response and often tolerance is tigens placed into immune-privileged sites, the liver that have a rich blood supply and induced instead. These organs maintain perhaps by favoring T-cell differentiation where blood flow is reduced in sinusoids even stronger barriers to routine entry of toward the less destructive Th2 phenotype, providing a better opportunity for naive T lymphocytes than other nonlymphoid tis- may be another possible mechanism of cells to leave the circulation. sues. Such restriction may be required to avoiding tissue damage (97, 98). One explanation for why self antigen- avoid disruption of normal physiologic MHC presentation by nonlymphoid tis- function of these sites by the destructive Mucosal Antigens sues induces tolerance instead of autoim- force of the immune response. One exam- munity is that most nonlymphoid tissues ple is the blood-brain barrier, where tight The mucosal surfaces are exposed to the do not express costimulatory molecules junctions between endothelial cells of the greatest number of foreign antigens. How- like B7 that are required to activate naive brain vasculature prevent the access of ever, while effective immunity usually de- T cells and, perhaps, prevent tolerance. lymphocytes to the central nervous system velops against microbial pathogens invad- This idea has been validated by the results (15, 85). The importance of the blood- ing through the mucosa, immune toler- of Flavell and coworkers (80, 81) using brain barrier as a self-tolerance mecha- ance is maintained against innocuous transgenic mice that express I-E class II nism is illustrated by experimental models antigens that are components of diet and MHC molecules alone, B7-1 alone, or I-E of autoimmunity involving the central ner- air. Some discrimination between poten- plus B7-1 on pancreatic (3 cells. Coexpres- vous system. One example, experimental tially pathogenic intestinal contents and sion of both I-E and B7-1, but not I-E or autoimmune encephalomyelitis, can be in- food is provided by the physical barriers B7-1 alone, on pancreatic (3 cells resulted duced by CD4+ T cells specific for certain created by the intestinal epithelium. In- in T-cell-mediated destruction of the pan- components of myelin (86, 87). However, tegrity of adhesion among intestinal epi- creas instead of tolerance. Furthermore, this can only be achieved in certain strains thelial cells and the layer of mucus coating transplanted I-E+, B7-1+ islets were re- of mice that are notable for increased the epithelium are critical in protecting jected by normal mice but not by the vascular sensitivity within the central ner- the intestine from bacterial penetration transgenic mice that express only I-E in vous system (87, 88). Even in this context (99-101). However, a barrier that is even the pancreas, demonstrating that the T disease often requires the use of pertussis more specific against immunogenic lumi- cells in the latter mice were truly tolerant toxin that is known to further disrupt the nal contents is created by secreted IgA to I-E and not simply ignorant of it. This blood-brain barrier (89-91). molecules. Production of IgA is depen- important set of results shows that non- Immune ignorance may also be an im- dent on absorption of antigens into Pey- lymphoid tissues that express class II portant mechanism of tolerance for other er's patches, lymph node-like structures MHC molecules are capable of inducing immune-privileged sites. Thus, extensive that are scattered throughout the intes- CD4+ T-cell tolerance, whereas nonlym- tight junctions also exist among Sertoli tine. Specialized intestinal epithelial cells, phoid tissues that express both class II cells lining the seminiferous tubules in the called M cells, efficiently translocate bac- MHC and costimulatory molecules induce testis, as well as among the retinal pigment teria and macromolecules into Peyer's T-cell immunity. It should be noted, how- epithelial cells in the eye (92, 93). How- patches, where particles with adjuvant ever, that constitutive expression of class ever, it is now clear that active immuno- properties are likely to lead to a produc- II MHC molecules is not the normal sit- regulatory mechanisms employed by res- tive immune response, including IgA pro- uation. In a normal individual, even if a ident tissue cells are also critical for main- duction (102). In contrast, soluble food Downloaded by guest on September 28, 2021 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996) 2251

proteins can directly pass from the gut anisms limit the activation of naive T cells & Steinman, R. M. (1994) J. Exp. Med. lumen into the bloodstream (103-105) to secondary lymphoid tissues that drain 180, 1849-1860. and thus would be likely to induce T-cell sites of inflammation, and control the 24. Larsen, C. P., Ritchie, S. C., Pearson, tolerance as a result of noninflammatory effector functions of activated T cells that T. C., Linsley, P. S. & Lowry, R. P. antigen presentation, as shown in Fig. 2B. enter nonlymphoid tissues. The fact that (1992) J. Exp. Med. 176, 1215-1220. Support for this comes from the finding these limitations are only overcome during 25. Gray, D. (1988) Immunology 65, 73-79. that oral administration of soluble foreign 26. Jacob, J., Kassir, R. & Kelsoe, G. (1991) microbial infections prevents the T-cell sys- J. Exp. Med. 173, 1165-1175. protein antigens in the absence of an tem from mounting full-blown autoimmune 27. Liu, Y. J., Zhang, J., Lane, P. J. L., Chan, adjuvant results in systemic T-cell toler- responses to peripheral self antigens. E. Y. T. & MacLennan, I. C. M. (1991) ance, whereas feeding foreign antigen plus Eur. J. Immunol. 21, 2951-2962. a mucosal adjuvant, such as cholera toxin, We thank Drs. Yoji Shimizu and Dan Muel- 28. Fuller, K. A., Kanagawa, 0. & Nahm, results in T-cell immunity and IgA pro- ler for their critical reviews of the manuscript. M. H. (1993) J. Immunol. 151, 4505- duction (106-108). Another reason for 4512. normal tolerance of food antigens may be 1. Picker, L. J. & Siegelman, M. H. (1993) 29. Kearney, E. R., Pape, K. A., Loh, D. Y. due to the local suppressive microenviron- in Fundamental Immunology, ed. Paul, & Jenkins, M. K. (1994) Immunity 1, ment in the gut. Immunosuppressive mol- W. E. (Raven, New York), 3rd Ed., pp. 327-339. ecules such as TGF-,3 and IL-10 are made 145-197. 30. Jacob, J. & Kelsoe, G. (1992)J. Exp. Med. abundantly in the intestine. In addition to 2. Kappler, J. W., Roehm, N. & Marrack, P. 176, 679-687. resident parenchymal cells, another (1987) Cell 49, 273-280. 31. Van den Eertweght, A. J. M., Noelle, 3. Hogquist, K. A., Gavin, M. A. & Bevan, R. J., Roy, M., Shepherd, D. M., Aruffo, source of these molecules may be regula- M. J. (1993)J. Exp. Med. 177, 1469-1473. A., Ledbetter, J. A., Boersma, W. J. A. & tory T cells that are preferentially acti- 4. Ashton-Rickardt, P. G., Van Kaer, L., Claassen, E. (1993) J. Exp. Med. 178, vated by fed antigens (109, 110). By what- Schumacher, T. N. M., Ploegh, H. L. & 1555-1565. ever mechanism, TGF-/ and IL-10 clearly Tonegawa, S. (1993) Cell 73, 1041-1049. 32. Noelle, R. J., Roy, M., Shepherd, D. M., play a role in tolerance to mucosal anti- 5. Anderson, G., Jenkinson, E. J., Moore, Stamenkovic, I., Ledbetter, J. A. & gens because TGF-,B1- and IL-10-defi- N. C. & Owen, J. J. T. (1993) Nature Aruffo, A. (1992) Proc. Natl. Acad. Sci. cient animals develop inflammatory (London) 362, 70-73. USA 89, 6550-6554. bowel disease (111-113). Interestingly, 6. Sprent, J. & Webb, S. R. (1995) Curr. 33. Liu, Y. J., Joshua, D. E., Williams, G. T., mice lacking TGF-,31 also develop sys- Opin. Immunol. 7, 196-205. Smith, C. A., Gordon, J. A. & MacLen- temic autoimmunity, suggesting an impor- 7. Kisielow, P., Bluthmann, H., Staerz, nan, I. C. M. (1989) Nature (London) tant systemic role of this cytokine in reg- U. D., Steinmetz, M. & von Boehmer, H. 342, 929-931. (1988) Nature (London) 333, 742-746. 34. Ranheim, E. A. & Kipps, T. J. (1993) J. ulating the immune response. 8. Bogen, B., Dembic, Z. & Weiss, S. (1993) Exp. Med. 177, 925-935. EMBO J. 12, 357-363. 35. Grewal, I. S., Xu, J. & Flavell, R. A. Summary 9. Surh, C. D. & Sprent, J. (1994) Nature (1995) Nature (London) 378, 617-620. (London) 372, 100-103. 36. van Essen, D., Kikutani, H. & Gray, D. Clonal deletion via apoptotic cell death 10. Zal, T., Volkmann, A. & Stockinger, B. (1995) Nature (London) 378, 620-623. appears to be the intrinsic response of (1994) J. Exp. Med. 180, 2089-2099. 37. Foy, T. M., Shepherd, D. M., Durie, immature thymocytes that receive high 11. Nieuwenhuis, P., Stet, R. J. M., We- F. H., Aruffo, A., Ledbetter, J. A. & No- levels of TCR signaling. The structure of genaar, J. P. A., Wubbena, A. S., Kam- elle, R. J. (1993) J. Exp. Med. 178, 1567- the thymus is designed to maximize expo- pinga, J. & Karrenbeld, A. (1988) Immu- 1575. sure to thymic self peptide-MHC com- nol. Today 9, 372-375. 38. Foy, T. M., Laman, J. D., Ledbetter, plexes by passing immature thymocytes 12. Matzinger, P. & Guerder, S. (1989) Na- J. A., Aruffo, A., Claassen, E. & Noelle, through a gauntlet of cortical ture (London) 338, 74-76. R. J. (1994) J. Exp. Med. 180, 157-163. epithelial 13. Mackay, C. R., Marston, W. L. & Dud- 39. Xu, J., Foy, T. M., Laman, J., Elliott, cells and corticomedullary dendritic cells. ler, L. (1990) J. Exp. Med. 171, 801-817. E. A., Dunn, J. J., Waldschmidt, T. J., In contrast, peripheral tolerance, which by 14. Steinman, R. M. (1991) Annu. Rev. Im- Elsemore, J., Noelle, R. J. & Flavell, definition must act on mature T cells that munol. 9, 271-296. R. A. (1994) Immunity 1, 423-431. are specific for extrathymic self antigens, 15. Girard, J. P. & Springer, T. A. (1995) 40. Korthauer, U., Graf, D., Mages, H. W., depends on factors other than the avidity Immunol. Today 16, 449-457. Briere, F., Munoreedevi, P., Malcolm, S., of the TCR for self peptide-MHC com- 16. Arbores, M. L., Ord, D. C., Ley, K., Rat- Ugazio, A. G., Notarangelo, L. D., plexes. Many of these factors consist of ech, H., Maynard-Curry, C., Otten, G., Levinski, R. J. & Kroczek, R. A. (1993) anatomic barriers that limit the migration Capon, D. J. & Tedder, T. F. (1994) Im- Nature (London) 361, 539-543. and activation of self-reactive T cells. The munity 1, 247-260. 41. Berek, C., Berger, A. & Apel, M. (1991) homing receptors expressed by naive cells 17. Smith, M. E. & Ford, W. L. (1983) Im- Cell 67, 1121-1129. restrict T cells to munology 49, 83-94. 42. MacLennan, I. C. M. (1994) Annu. Rev. secondary lymphoid 18. Macatonia, S. E., Knight, S. C., Edwards, Immunol. 12, 117-139. tissues, limiting the detection of tissue- A. J., Griffiths, S. & Fryer, P. (1987) J. 43. Szakal, A. K., Kosco, M. H. & Tew, J. G. specific self antigens expressed in nonlym- Exp. Med. 166, 1654-1667. (1989) Annu. Rev. Immunol. 7, 91-111. phoid tissues. In addition, class II MHC 19. Larsen, C. P., Steinman, R. M., Witmer- 44. Sprent, J. (1994) Cell 76, 315-322. and costimulatory molecules are not ex- Pack, M. D., Hankins, D. F., Morris, P. J. 45. Imhof, B. A. & Dunon, D. (1995) Adv. pressed by non-lymphoid tissues, prevent- & Austyn, J. M. (1990) J. Exp. Med. 172, Immunol. 58, 345-416. ing productive activation of self-reactive T 1483-1494. 46. Dresser, D. W. (1961) Nature (London) cells that do gain access. T cells that 20. Cumberbatch, M. & Kimber, I. (1995) 191, 1169-1171. recognize developmentally regulated self Immunology 84, 31-55. 47. Dresser, D. W. (1962) Immunology 5, antigens in secondary lymphoid organs do 21. Sallusto, F., Cella, M., Danieli, C. & 378-388. so in a noninflammatory environment Lanzavecchia, A. (1995)J. Exp. Med. 182, 48. Warren, H. S., Vogel, F. R. & Chedid, that does not allow the T cells to enter 389-400. L. A. (1986) Annu. Rev. Immunol. 4, follicles or become memory cells. 22. June, C. H., Bluestone, J. A., Nadler, 369-388. Finally, L. M. & Thompson, C. B. (1994) Immu- 49. Chiller, J. M., Habicht, G. S. & Weigle, immune-privileged tissues that are highly nol. Today 15, 321-331. W. 0. (1971) Science 171, 813-815. susceptible to immune-mediated destruc- 23. Inaba, K., Witmer-Pack, M., Inaba, M., 50. Janeway, C. A., Jr. (1989) Cold Spring tion protect themselves from T-cell attack Hathcock, K. S., Sakuta, H., Azuma, M., Harbor Symp. Quant. Biol. 54, 1-13. with immunosuppressive cytokines and Yagita, H., Okumura, K., Linsley, P. S., 51. Matzinger, P. (1994) Annu. Rev. Immu- surface molecules. Together, these mech- Ikehara, S., Muramatsu, S., Hodes, R. J. nol. 12, 991-1045. Downloaded by guest on September 28, 2021 2252 Review: Mondino et al. Proc. Natl. Acad. Sci. USA 93 (1996)

52. Cosgrove, D., Gray, D., Dierich, A., 71. Cooke, M. P., Heat, A. W., Shokat, 92. Greenwood, J., Howes, R. & Lightman, Kaufman, J., Lemeur, M., Benoist, C. & K. M., Zeng, Y., Finkelman, F. D., Lins- S. (1994) Lab. Invest. 70, 39-52. Mathis, D. (1991) Cell 66, 1051-1066. ley, P. S., Howard, M. & Goodnow, C. C. 93. Dynn, M. (1988) in Cell and Tissue Biol- 53. Kyburz, D., Aichele, P., Speiser, D. E., (1994) J. Exp. Med. 179, 425-438. ogy: A Textbook of Histology, ed. Weiss, Hengartner, H., Zinkernagel, R. M. & 72. Cyster, J., Hartley, S. & Goodnow, C. C. L. (Urban & Schwarzenberg, Baltimore), Pircher, H. (1993) Eur. J. Immunol. 23, (1994) Nature (London) 371, 389-395. pp. 931-972. 1956-1962. 73. Shokat, K. M. & Goodnow, C. C. (1995) 94. Streilein, J. W. (1995) Science 270,1158- 54. Liu, Y. & Janeway, C. A., Jr. (1991) Int. Nature (London) 375, 334-338. 1159. Immunol. 3, 323-332. 74. Pulendran, B., Kannourakis, G., Nouri, 95. Bellgrau, D., Gold, D., Selawry, H., 55. Linsley, P. S., Wallace, P. M., Johnson, S., Smith, K. G. C. & Nossal, G. J. V. Moore, J., Franzusoff, A. & Duke, R. C. J., Gibson, M. G., Greene, J. L., Ledbet- (1995) Nature (London) 375, 331-334. (1995) Nature (London) 377, 630-632. ter, J. A., Singh, C. & Tepper, M. A. 75. Han, S., Zheng, B., Dal Porto, J. & 96. Griffith, T. S., Brunner, T., Fletcher, (1992) Science 257, 792-795. Kelsoe, G. (1995)J. Exp. Med. 182, 1635- S. M., Green, D. R. & Ferguson, T. A. 56. Lenshow, D. J., Zeng, Y., Thistlewaite, 1644. (1995) Science 270, 1189-1192. J. R., Montag, A., Brady, V., Gibson, 76. Ohashi, P. S., Oehen, S., Buerki, K., 97. Ksander, B. R. & Streilein, J. W. (1994) M. I., Linsley, P. S. & Bluestone, J. A. Pircher, H., Ohashi, C. T., Odermatt, B., Chem. Immunol. 58, 117-145. (1992) Science 257, 789-791. Mallissen, B., Zinkernagel, R. M. & 98. Wilbands, G. A. & Streilein, J. W. (1990) 57. Lin, H., Bolling, S. F., Linsley, P. S., Wei, Hengartner, H. (1991) Cell 65, 305-317. Immunology 71, 383-389. R.-Q., Gordon, D., Thompson, C. B. & 77. Shonrich, G., Momburg, F., Malissen, 99. Beachey, E. H. (1981) J. Infect. Dis. 143, Turka, L.A. (1993) J. Exp. Med. 178, M., Schmitt-Verhulst, A.-M., Malissen, 325-345. 1801-1806. B., Hammerling, G. J. & Arnold, B. 100. Mantle, M., Basaraba, L., Peacock, S. C. 58. Wallace, P. M., Rodgers, J. N., Leytze, (1992) Int. Immunol. 4, 581-590. & Gall, D. G. (1989) Infect. Immun. 57, G. M., Johnson, J. S. & Linsley, P. S. 78. Hammerling, G. J., Schonrich, G., Fer- 3292-3299. (1995) J. Immunol. 154, 5885-5985. ber, I. & Arnold, B. (1991) Immunol. 101. Hermiston, M. L. & Gordon, J. I. (1995) 59. Kearney, E. R., Walunas, T. L., Karr, Rev. 122, 47-67. Science 270, 1203-1207. R. W., Morton, P. A., Loh, D. Y., Blue- 79. Arnold, B., Schonrich, G. & Hammer- 102. Keren, D. F. (1992) Sem. Immunol. 4, stone, J. A. & Jenkins, M. K. (1995) J. ling, G. J. (1993) Immunol. Today 14, 217-226. Immunol. 155, 1032-1036. 12-14. 103. Cornell, R., Walker, W. A. & Issel- 60. Enyon, E. E. & Parker, D. C. (1992) J. 80. Guerder, S., Picarella, D. E., Linsley, bacher, K. J. (1971) Lab. Invest. 25, 42- Exp. Med. 175, 131-138. P. S. & Flavell, R. A. (1994) Proc. Natl. 48. 61. Fuchs, E. J. & Matzinger, P. (1992) Sci- Acad. Sci. USA 91, 5138-5142. 104. Warshaw, A. L., Walker, W. A., Cornell, ence 258, 1156-1158. 81. Guerder, S., Meyerhoff, J. & Flavell, R. R. & Isselbacher, K. J. (1971) Lab. In- 62. Parker, D. C., Greiner, D. L., Phillips, (1994) Immunity 1, 155-166. vest. 25, 675-684. N. E., Appel, M. C., Steele, A. W., Durie, 82. Harlan, D. M., Hengartner, H., Huang, 105. Bruce, M. G. & Ferguson, A. (1986) Im- F. H., Noelle, R. J., Mordes, J. P. & M. L., Kang, Y.-H., Abe, R., Moreadith, munology 59, 295-300. Rossini, A. A. (1995) Proc. Natl. Acad. R. W., Pircher, H., Gray, G. S., Ohashi, 106. Wilson, A. D., Bailey, M., Williams, Sci. USA 92, 9560-9564. P. S., Freeman, G. J., Nadler, L. M., N.A. & Stokes, C. R. (1991) Eur. J. 63. Buhlmann, J., Foy, T. M., Aruffo, A., June, C. H. & Aichele, P. (1994) Proc. Immunol. 21, 2333-2339. Crassal, K. M., Ledbetter, J. A., Green, Natl. Acad. Sci. USA 91, 3137-3141. 107. Hornquist, E. & Lycke, N. (1993) Eur. J. W. R., Xu, J. C., Shultz, L. D., Roope- 83. Williams, I. R., Ort, R. J. & Kupper, T. S. Immunol. 23, 2136-2143. sian, D., Flavell, R. A., Last, L., Noelle, (1994) Proc. Natl. Acad. Sci. USA 91, 108. Xu-Amano, J., Kiyono, H., Jackson, R. J. & Durie, F. H. (1995) Immunity 2, 12780-12784. R. J., Staats, H. F., Fujihashi, K., Bur- 645-653. 84. Nasir, A., Ferbel, B., Salminen, W., rows, P. D., Elson, C. O., Pillai, S. & 64. Webb, S., Morris, C. & Sprent, J. (1990) Barth, R. K. & Gaspari, A. (1994)J. Clin. McGhee, J. R. (1993) J. Exp. Med. 178, Cell 63, 1249-1256. Invest. 94, 892-898. 1309-1320. 65. Vella, A. T., McCormack, J. E., Linsley, 85. Schlosshauer, B. (1993) Bioessays 15, 109. Chen, Y., Kuchroo, V. K., Inobe, J., P. S., Kappler, J. W. & Marrack, P. 341-346. Hafler, D. A. & Weiner, H. (1994) Sci- (1995) Immunity 2, 261-270. 86. Zamvil, S. S. & Steinman, L. (1990) ence 265, 1237-1240. 66. Singer, G. C. & Abbas, A. (1994) Immu- Annu. Rev. Immunol. 8, 579-621. 110. Powrie, F. (1995) Immunity 3, 171-174. nity 1, 365-371. 87. Tuohy, V. K., Sobel, R. A. & Lees, M. B. 111. Shull, M. M., Ormsby, I., Kier, A. B., 67. Rocha, B. & von Boehmer, H. (1991) (1988)J. Immunol. 140, 1868-1873. Pawlowski, S., Diebold, R. J., Yin, M., Science 251, 1225-1228. 88. Yong, T., Bebo, Jr., B. F., Sapatino, Allen, R., Sidman, C., Proetzel, G., Cal- 68. Tafuri, A., Alferink, J., Moller, P., Ham- B. V., Welsch, C. J., Orr, E. L. & Linthi- vin, D., Annunziata, N. & Doetschman, merling, G. J. & Arnold, B. (1995) Sci- cum, D. S. (1994) J. Neurotrauma 11, T. (1992) Nature (London) 359, 693-699. ence 270, 630-633. 161-171. 112. Kulkarni, A. B., Huh, C.-G., Becher, D., 69. Zinkernagel, R. M., Moskophidis, D., 89. Bernard, C. C. A. & Carnegie, P. R. Geiser, A., Lyght, M., Flanders, K. C., Kundig, T., Oehen, S., Pircher, H. P. & (1975) J. Immunol. 114, 1537-1540. Roberts, A. B., Sporn, M. B., Ward, J. M. Hengartner, H. (1993) Immunol. Rev. 90. Linthicum, D. S. & Frelinger, J. (1982)J. & Karlsson, S. (1993) Proc. Natl. Acad. 131, 199-223. Exp. Med. 155, 31-40. Sci. USA 90, 770-774. 70. Goodnow, C. C., Crosbie, J., Jorgensen, 91. Linthicum, D. S., Munoz, J. J. & Blas- 113. Kuhn, R., Lohler, J., Rennick, D., Ra- H., Brink, R. A. & Basten, A. (1989) kett, A. (1982) Cell. Immunol. 73, 299- jewski, K. & Muller, W. (1993) Cell 75, Nature (London) 342, 385-391. 310. 263-274. Downloaded by guest on September 28, 2021