Studies on Peripheral T Cell Deletion and Anergy Induced by Superantigens

To whom correspondence should be addressed:

1. Department of Pathology
   Room B-2
   Montreal, QC, Canada H3A 2B4
   
2. Department of Pathology
   McGill University
   

INTRODUCTION

Superantigens (superAgs) are products of bacteria and viruses that stimulate large subpopulations of T cells and cause a wide variety of clinical syndromes such as food poisoning, toxic shock syndrome, scalded skin syndrome, and scarlet fever (reviewed in 1). These syndromes result from T cell activation with massive release of cytokines. However, following an early phase of T cell proliferation, superAgs can induce specific forms of T cell tolerance, namely anergy and deletion. SuperAgs interact directly with major histocompatibility (MHC) class II molecules of the antigen-presenting cells (APC) and with specific Vß-segments of the T cell receptor (TcR) (reviewed in 1-3). Thus, depending on the frequency of the responding Vß subpopulation(s), 5 -30% of the entire T cell repertoire can be recognized by a given superAg. This differs from T cell responses to conventional antigens, where less than 0.1% of cells react.

SuperAgs are ideal tools for probing T cell tolerance phenomena. Functional tolerance can be easily assessed by measuring the strong in vitro mitogenic response induced by superAg. Their limited TcR specificity enables us to monitor the proportion of T cells that bear the relevant Vß-TcR by flow cytometric analysis (1-3).

The exogenous and the endogenous superAgs used in this study were, respectively, the Staphylococcus aureus enterotoxin B (SEB) and the product of the intrinsic mouse mammary tumor virus 7 (MMTV-7), also commonly referred to as Mls-1a. SEB stimulates T cells bearing Vß3, 7, 8, 11, and 17 segments, whereas the Mls-1a molecule triggers the activation of T cells bearing Vß6, 7, 8.1, and 9 segments. However, due to the relative frequencies of the Vß-expressing T cell populations in most murine strains, the observed effects of SEB are generally caused by the activation of Vß8+/CD4+ T cells, and Mls-1a by the Vß6+/CD4+ T cells. When injected intravenously in mice, superAgs induce anergy as well as partial peripheral deletion of CD4+ T cells bearing superAg-specific TcR-Vß segments (4-7). The current study is aimed at defining the characteristics and nature of T cell anergy and peripheral deletion.

METHODS

BALB/c mice (H-2d, Mls-1b, Mls-2a, Lyt-1.2, Lyt-2.2), DBA/2 mice (H-2d, Mls-1a, Mls-2a, Lyt-1.1, Lyt-2.1) were purchased from the Charles River Laboratory (St-Constant, QC, Canada), and C3H/HeJ mice (H-2k, Mls-1b, Mls-2a) from the Jackson Laboratory (Bar Harbor, ME, USA). BALB/c and DBA/2 mice were chosen because these strains are MHC identical but differ in Mls-1 expression. C3H/HeJ mice were used as a source of stimulator cells to generate mixed lymphocyte cultures (MLC) across an MHC barrier. Adult BALB/c mice (four to six weeks old) were thymectomized in our facility as described (8) and rested for seven days before the onset of treatments. For consistency, since male and female responses may differ in magnitude, all the experiments were performed in female mice.

Monoclonal antibodies and Reagents

Phycoerythrin (PE)-conjugated anti-CD4, PE-conjugated anti-CD8, anti-Lyt-1.1, anti-Lyt-2.1, and anti-Thy-1.2 were purchased from Cedarlane (Hornby, ON, Canada). The following monoclonal antibodies (MoAbs) were produced as culture supernatants and purified by protein G (Pharmacia, Montreal, QC, Canada) affinity chromatography: anti-Vß6 (44-22-1 hybridoma) (9), anti-Vß8.1-8.2 (KJ16 hybridoma) (10), anti-Vß14 (14.2 hybridoma) (11), and anti-TcR-aß (H57-579 hybridoma) (12). For flow cytometry analysis the purified MoAbs were directly conjugated with fluorescein isothiocyanate (FITC). Cyclosporin A (CsA) was a kind gift from Sandoz (Dorval, QC, Canada). Staphylococcal enterotoxin A (SEA) was purchased from Toxin Technology (Sarasota, FL) and SEB was purchased from Sigma Chemical Co. (St. Louis, MO, USA).

SuperAg immunization

SuperAg immunizations were performed by tail vein injections of 107 Mls-1a+ splenocytes (RBC depleted), or 50-100 ug of a staphylococcal enterotoxin (either SEA or SEB) in a volume of 0.25 ml of PBS. The day of superAg injection was considered day 0 of our experiments.

Flow cytometry analysis

For two color flow cytometry analysis, lymph node (LN) cells (pooled cervical, axillary, para-aortic and mesenteric LN) were stained with either PE-anti-CD4 or PE-anti-CD8 and with one of the FITC-labelled-anti-Vß MoAbs mentioned above. One color flow cytometry analysis was performed with FITC-labelled MoAbs alone. LN cells (5x105) were incubated with PE-conjugated MoAb for 15 minutes at 4[[ring]]C, then washed three times with phosphate buffered saline (PBS) containing 1% fetal calf serum (FCS) (Gibco, Burlington, ON, Canada), and then incubated under the same conditions with the second FITC-labelled-anti-Vß MoAb. Propidium iodide (Sigma Chemical Co.) was added after the last wash. Dead cells were excluded based on propidium iodide staining and forward scatter. In each sample, 104 cells were analyzed on a FACScan (Becton and Dickenson, Mississauga, ON, Canada) flow cytometer, and results were plotted on a logarithmic scale. Statistical analyses were performed using the Student's t test. In some experiments spleen cells were analyzed. Before staining, spleen cells (RBC depleted) were depleted of B cells by two rounds of sequential J11.D2 MoAb and complement treatment, performed as previously described (13). Following depletion of dead cells on a lymphocyte-M gradient (Cedarlane), spleen cells were stained and analyzed as described above for LN cells.

In vitro proliferative assays

LN cells or purified T cells were cultured in 96-well flat bottom plates (Gibco) in quadruplets at 5x105 cells/well with either: i) for SEB superAg stimulation experiments, 106 syngeneic irradiated (2000 rads) splenocytes of untreated BALB/c mice and 10 ug/ml of SEB, or 1 ug/ml of SEA as a control; or ii) for Mls-1a superAg experiments, 106 irradiated (1000 rads) DBA/2 splenocytes or irradiated (1000 rads) allogeneic C3H/HeJ (H-2k, Mls-1b) splenocytes. Cells were cultured in RPMI-1640 (Gibco) containing 10% FCS, L-glutamine, 50 uM 2-mercaptoethanol, and penicillin/streptomycin. Cells were pulsed with [3H]-thymidine (ICN, Montreal, QC, Canada) on day 3 of culture and harvested on day 4, unless otherwise indicated. For experiments that assessed the effect of exogenous IL-2 or mixed lymphocyte culture (MLC) supernatant additions, rIL-2 (Cedarlane) and MLC supernatant (taken from cultures of BALB/c T cell responders with CBA/J splenocytes) were added in the culture well at the beginning of the proliferation assay. For experiments where responder T cells were mixed in the same culture well, the T cells were purified by passage through a Cellect T cell purification column (Biotex Laboratories Inc., Edmonton, Canada) before their addition to culture wells. "Day in culture" refers to the day that the cells were harvested. In all cases, the cells were harvested after 18 hours of incubation with [3H]-thymidine (1 uCi/well). In the case of stimulation with either SEA or SEB, [[Delta]]c.p.m. = (c.p.m. experimental) - (c.p.m. control [no superAg]). In the case of Mls-1a or MLC stimulation, [[Delta]]c.p.m. = (c.p.m. experimental) - (c.p.m. with syngeneic stimulator cells). Statistical analyses were performed using the Student's t test.

RESULTS

Effects of dose and repeated superAg injections on superAg reactive T cells

SuperAgs are potent inducers of anergy as demonstrated in Fig. 1(A, B). In vitro proliferative assays indicate a five- to seven-fold reduction in reactivity to the superAg used to immunize the mice (p < 0.001). The extent of the decrease in proliferative responses to superAg in anergic T cells reached a maximum at a dose of 105 Mls-1a+ cells (Fig. 1A) and 25 ug of SEB (Fig. 1B), in Mls-1a-injected and SEB-injected mice, respectively. The extent of the partial deletion of superAg-reactive T cells in these mice also reached a plateau at the same doses. Both anergy and partial deletion are specific to the superAg-reactive T cell populations. The non-Mls-1a-reactive T cells (Vß14+/CD4+ T cells) showed no significant variation in their relative frequencies in Mls-1a-injected BALB/c (Mls-1b; H-2d) mice. The in vitro proliferative responses to irradiated allogeneic C3H/HeJ (Mls-1b; H-2k) spleen cells (MLC stimulator) remained comparable to PBS-injected BALB/c mice (Fig. 1A). Likewise, non-SEB-reactive T cells (Vß6+/CD4+ T cells) in SEB-injected mice showed no reduction in their numbers (not shown), but their relative frequencies were slightly increased. This effect is likely due to the loss of SEB-reactive T cells (Fig. 1B). The in vitro responses to SEA (non Vß8-TcR specific) were not affected in these mice. Multiple injections of Mls-1a+-cells did not augment the deletion, nor did they induce a further reduction of the in vitro proliferative responses to Mls-1a (Fig. 2A). Thus, the latter observation shows that the partial deletion of Vß6+/CD4+ T cells cannot be enhanced by reexposing these cells to Mls-1a molecules. The remnant of in vitro reactivity by Mls-1a-anergic T cells could be attributed to Mls-1a-reactive T cells which, despite the multiple injections of Mls-1a+-cells, did not encounter the superAg in vivo (or at least not in a way that would normally induce anergy in these cells). Alternatively, the in vitro responses observed could be mounted against conventional antigens on Mls-1a cells that are not "anergizing" due to either small antigen concentrations or to other reasons currently unknown to us. In contrast, repeated injections of SEB enhanced the deletion of Vß8+/CD4+ T cells, reducing their relative frequencies from 13.4% (SEB injected 1 time) to 7.9% (SEB injected 5 times) (Fig. 2B). This enhanced reduction in reactive T cells was accompanied by a further slight decrease of the in vitro reactivity to SEB. The discrepancies between Mls-1a- and SEB-reactive T cell populations in their sensitivity to multiple injections may reflect a difference in the bioavailability of the two superAgs. Mls-1a is an endogenous protein expressed by the injected cells and is likely to persist in the periphery of the mice as long as the cells survive. SEB, on the other hand, is an exogenous protein which makes it more susceptible to rapid elimination. However, presently we can only speculate on the reasons for this difference.

It should be noted that superAg-induced anergy is not limited to a single mouse strain since anergy was successfully induced in various murine strains (4-5,14). However, in different strains, the degree of peripheral deletion obtained can vary from 0% to about 75% of superAg-specific T cells.

In vivo kinetic studies of anergy and peripheral deletion induction in sham and thymectomized mice

Injections of superAgs into thymectomized (Tx) mice were used to determine if anergy and deletion were peripheral phenomena. As shown in Fig. 3 (A and C), the induction of anergy in sham and Tx mice had the same kinetic profile. These results demonstrate that anergy induction does not require the presence of a thymus. In vitro proliferative assays showed that the low levels of reactivity to the superAg used for immunization were maintained for almost 100 days after immunization (p < 0.01). Subsequently, superAg reactivity increased slightly in superAg-treated sham mice, possibly due to the generation of new superAg-specific T cells (see below). However, it should be noted that the level of reactivity observed by day 224 post-immunization in sham mice remained low when compared to non-immunized mice (21% and 33% of normal proliferative responses for SEB and Mls-1a, respectively) (p < 0.01). We also observed, as have others (6), that for SEB, the partial elimination of Vß8+/CD4+ is preceded by an expansion of these cells on day 3 post-immunization. It is followed by a rapid decline that stabilizes by day 7 post-immunization (Fig. 3B). In Mls-1a-injected BALB/c mice, no expansion was found in Vß6+/CD4+ T cells prior to their deletion (Fig. 3D). The latter observation varies with different mouse strains. Webb and colleagues found that B10.BR (Mls-1b, H-2k) mice injected with AKR/J (Mls-1a, H-2k) splenocytes had a strong expansion of Vß6+/CD4+ T cells in the first few days after immunization (14). This was subsequently followed by a deletion to about 25% of the initial levels by day 22 post-immunization. We also detected an approximately 2% increase at day 3 post-immunization (from 14.2 +/- 0.2 in PBS-injected vs. 16.3 +/- 0.1 in Mls-1a-injected mice) before observing a deletion in Vß6+/CD4+ T cells in CBA/CaJ (Mls-1b/Mls-2b, H-2k) mice injected with CBA/J (Mls-1a/Mls-2a, H-2k) splenocytes (unpublished data). However, as we report here for BALB/c (Mls-1b, H-2d) mice injected with DBA/2 (Mls-1a, H-2d) splenocytes, no such expansion before deletion could be observed in C3H/HeJ (Mls-1b, H-2k) mice injected with C3H.CE (Mls-1a, H-2k) splenocytes (4). The reasons for these discrepancies are not known. In both SEB- and Mls-1a-injected sham mice, a slow increase of superAg-specific T cell numbers was observed starting on day 30 post-immunization, which corresponds to the increase of in vitro reactivity (Fig. 3). Since no increase in reactivity to superAg was observed in these mice, this suggests that de novo T cells are responsible for the increase of in vitro superAg reactivity.

Effects of in vitro addition of T cell growth factors to anergic T cells

Anergic T cells produce small amounts of IL-2 when stimulated with their specific superAg (4-6). In vitro T cell proliferative response to exogenous rIL-2 and MLC supernatant was assessed to determine if the anergic state was caused by a lack of growth factor production (Fig. 4). Addition of rIL-2 to Mls-1a-anergized T cells (Fig. 4A) or to SEB-anergized T cells (Fig. 4C) had no significant effect on their proliferative responses to the superAg to which they were tolerant. A comparable lack of response was obtained by the addition of MLC supernatant for both Mls-1a- and SEB-anergized T cells (Fig. 4B and 4D, respectively). Thus, anergic T cells fail to react to their specific superAg by at least two means: first, the blockage of IL-2 production, and second, the lack of responsiveness to T cell growth factors.

Anergic T cells are not suppressed by immunoregulatory T cells

To insure that unresponsive T cells were not under the influence of antigen-specific suppressor T cells, we added syngeneic Mls-1a-anergized T cells to normal T cells of BALB/c origin (Mls-1b, H-2d). We challenged these cells in vitro with Mls-1a+ cells (DBA/2: Mls-1a, H-2d), or with Mls-1b allogeneic spleen cells with disparate MHC (C3H/HeJ: Mls-1b, H-2k). In order to appreciate the reduction of the proliferative response due to the replacement of normal responder T cells by weakly-responding T cells (anergic T cells), we used, as a control, a mixture of normal responder T cells with T cells of DBA/2 (Mls-1a, H-2d) mice which do not respond to the syngeneic stimulating cells. DBA/2 mice possess only a small number of potentially Mls-1a-reactive T cells, and have the same MHC as the normal responder T cells (BALB/c: Mls-1b, H-2d). Therefore, they do not generate a significant MLC response to the BALB/c T cells (<3000 c.p.m.). In a typical antigen-specific suppression assay, a suppression of >80% can be observed with less than 10% of suppressive cells in culture (15,16). As shown in Fig. 5, the proliferative responses of all the normal:anergic responder T cell mixtures were maintained over 50% of the normal proliferative response. In samples with higher concentrations of anergic T cells (i.e., 75% of total T cells), the proliferative responses were on average 55% of the normal proliferative response. Notably, the proliferative responses of the normal:anergic responder T cell ratios were comparable to those of the normal:control (DBA/2 T cells) mixtures. This argues strongly against a suppressive effect. The higher proliferative responses of the normal:anergic ratios of 1:2 and 1:3 over the normal:control of the same ratios are likely due to the contribution of non-anergized T cells of the anergic population to the normal response. The allogeneic responses did not vary significantly in all groups and ratios studied (data not shown). Based on these results, it seems unlikely that the superAg model of anergy induction is caused by immunoregulatory T cells.

DISCUSSION

SuperAgs stimulate a large number of mature T cells bearing specific Vß-segments of the TcR. They achieve this effect by binding to both TcR Vß-segments and MHC class II molecules of APCs. This property has proven extremely useful for studying immunological phenomena, especially T cell tolerance (reviewed in 17). SuperAgs provided the means for obtaining conclusive evidence for thymic clonal deletion of autoreactive T cells and enabled the demonstration of peripheral clonal anergy and deletion (4-6,18). In this study, we examined the peripheral forms of tolerance induced by superAgs. We investigated the effects of dose and injection regimens, the in vivo kinetics, and the effects of T cell growth factors on anergic T cells. We have also investigated the possible contribution of specific suppression mechanisms for maintaining the anergic condition.

We found that augmenting the Mls-1a immunizing dose (in mice given only one dose) had no effect on either anergy or deletion. This was true even for Mls-1a-injected mice which received 1000 times the number of donor cells required to reach the maximum degree of anergy and deletion. Unfortunately, due to its toxic effects, SEB could not be injected at doses that exceeded 100 ug. Nevertheless, a dose of 100 ug of SEB was still four times higher than the minimal dose needed for optimal anergy and deletion induction. Although multiple superAg injections had no effect on peripheral anergy and deletion in Mls-1a-injected mice, in SEB-injected mice increasing the number of injections produced a concomitant decrease in SEB-reactive T cell levels and in vitro proliferative responses to SEB. The reason for the discrepancies between Mls-1a and SEB is unknown. Several factors may explain the differences observed, namely: the rates of superAg elimination, the differences in superAg tissue distribution and differences in the types of APC involved.

SuperAg injections in Tx mice revealed that the superAg-induced anergy and partial deletion were peripheral events. Tx and sham mice had similar kinetic profiles for both anergy and deletion. Long-term studies in these mice showed that anergy lasted at least 112 days for SEB-immunized Tx mice and for more than 224 days in Mls-1a-immunized Tx mice. In vitro reactivity to the superAg used for immunization in sham mice increased concomitantly with superAg-reactive T cell numbers. However, even at day 224 post-immunization the proliferative responses were still low when compared to PBS-injected mice: 21% and 33% of normal responses for SEB- and Mls-1a-injected sham mice, respectively. These findings show that anergy is a long-lasting phenomenon. Following the initial phase of T cell deletion, there was no evidence of further loss of anergic T cells. It should be noted that these cells were also non-responsive to direct cross-linking of their TcR with anti-Vß-TcR MoAbs in the presence of splenocytes, even if exogenous rIL-2 was added (our unpublished data and 5). Thus, it seems unlikely that anergic cells respond to other antigens once "anergized".

The unresponsive state of anergic T cells is not simply due to a limitation of IL-2 production by these cells (4-6). It has been found that the addition of rIL-2 directly to culture wells could not reverse the unresponsive state of anergic T cells (4,6). MLC supernatant also had no effect on the proliferative response of these cells. The inability of anergic T cells to respond to exogenous IL-2 could be due to the lack of high-affinity IL-2 receptor expression or to an impaired signal transduction induced by these receptors. However, since the IL-2 receptors were shown to be expressed on anergic T cells of both Mls-1a- and SEB-immunized mice (4,5), it would appear that the IL-2 responsiveness defect is more likely associated with signal transduction events.

A role for suppressor T cells was suggested by some authors to explain the lack of responsiveness in anergic cells. Using a model where lethally irradiated (DBA/2xB10.D2)F1 mice grafted with B10.D2 cells developed severe graft-versus-host reaction (GVHR) to non-H-2 antigen (19), Bruley-Rosset and colleagues observed that the mortality rate could be markedly reduced by donor pre-immunization against Mls-1a+ cells (20). This effect was attributed to Mls-1a generated double-negative suppressor T cells. To determine if suppressor T cells were active in our models, we mixed purified T cells of PBS-injected control mice with purified T cells of Mls-1a-immunized mice in various proportions and challenged them with Mls-1a+ cells. The results showed no evidence of suppression in these cell-mixing experiments. It was important to confirm the validity of the superAg-induced tolerance as an anergy model since it is the only alternative to the T cell anergy observed in TcR transgenic mice. Moreover, the superAg-induced anergy model has the particular advantage of providing an internal control in the populations of T cells bearing non-reactive TcR-Vß segments.

Despite the increasing attention devoted to superAgs in recent years, we remain unsure how they induce T cell anergy in mice, especially in view of their potent T cell mitogenic properties. SuperAgs are presented by MHC class II molecules without antigen processing by APCs (21,22). This process enables these molecules to stimulate T cells directly by simultaneously binding the TcR and class II molecules (23). Thus, T cells can be triggered by any class II-expressing cells, including some that may not be competent APCs. Consequently, T cells may not receive the proper costimulatory signals required for complete activation. Such lack of costimulatory signals in CD4+/Th1 clones has been reported to induce an anergic state in these cells (reviewed in 24). The costimulatory signal needed for full activation of these clones was found to be transmitted by the T cell's CD28 receptor after ligation with the BB-1/B7 molecule expressed on the surface of competent APCs (25,26). Since the weak expression of the BB-1/B7 molecule has been associated with the poor stimulatory property of resting B cells (25), it is conceivable that the superAg-induction of anergy is caused by resting B cell presentation of superAg to reactive T cells.

Our study demonstrates that following i.v. injection of superAgs, anergy and partial deletion are thymus-independent events. This state of anergy is long-lasting and cannot be reversed by the addition of T cell growth factors. We also ruled out the contribution of suppressor cells in the maintenance of anergy.

The biological significance of superAg-induced T cell anergy and deletion is unclear. This induced state of unresponsiveness may protect recipients from subsequent injury by these molecules if re-exposed. This effect may be relevant in diseases where shock or other serious conditions are induced by superAgs (1). A better understanding of peripheral T cell anergy and deletion could lead to new approaches in clinical immune tolerance. Finding ways to improve peripheral tolerance could have a significant impact on transplantation medicine. We have obtained some promising results in this matter by potentiating peripheral deletion with the immunosuppressant cyclosporin A (7). Furthermore, investigation of autoimmune diseases like juvenile diabetes (type 1) and rheumatoid arthritis may reveal activation of T cells by superAgs as a pathogenic mechanism. A recent study by Conrad and colleagues is the first one to suggest a direct link between superAg activation and the induction of juvenile diabetes (27). Therefore, treatments aimed at inducing tolerance in superAg-reactive T cells could be of benefit in autoimmune diseases. Our study reveals that such a state of tolerance can be induced and is very long-lasting.

ACKNOWLEDGEMENTS

This study was funded by the MRC of Canada, the Canadian Diabetes Association, and the Juvenile Diabetes Foundation International. LEV is the recipient of an MRC of Canada Studentship. The authors gratefully acknowledge H. Ste-Croix for her excellent technical assistance.

REFERENCES

1. Kotzin BL. Superantigens and their role in disease. Hospital Practice 1994;29:59-70.

2. Hodes RJ, Abe R. T cell recognition of Mls-like superantigens: analysis of TcR requirements, superantigenic ligands, and signal transduction. Seminars in Immunology 1992;4:319-327.

3. Johnson HM, Russell JK, Pontzer CH. Staphylococcal enterotoxin microbial superantigens. FASEB Journal 1991;5:2706-2712.

4. Rammensee HG, Kroschewski R, Frangoulis B. Clonal anergy induced in mature Vß6+ T lymphocytes on immunizing Mls-1b mice with Mls-1a expressing cells. Nature 1989;339:541-544.

5. Rellahan BL, Jones LA, Kruisbeek AM, Fry AM, Matis LA. In vivo induction of anergy in peripheral Vß8+ T cells by staphylococcal enterotoxin B. Journal of Experimental Medicine 1990;172:1091-1100.

6. Kawabe Y, Ochi A. Selective Anergy of Vß8+, CD4+ T cells in staphylococcus enterotoxin B-primed mice. Journal of Experimental Medicine 1990;172:1065-1070.

7. Vanier LE, Prud'homme GJ. Cyclosporin A markedly enhances superantigen-induced peripheral T cell deletion and inhibits anergy. Journal of Experimental Medicine 1992;176:37-46.

8. Herbert WJ, Kristensen F. Laboratory animal techniques for immunology. In: BD Weir, editor. Applications of Immunological Methods in Biomedical Sciences. Boston: Blackwell Scientific Publications, 1986:133.1-133.36.

9. Payne J, Huber BT, Cannon NA, et al. Two monoclonal rat antibodies with specificity for the b-chain variable region Vß6 of the murine T-cell receptor. Proceedings of the National Academy of Sciences of the United States of America 1988;85:7695-7698.

10. Haskins K, Hannum C, White J, et al. The antigen-specific, major histocompatibility complex restricted receptor on T cells. IV. An antibody to a receptor allotype. Journal of Experimental Medicine 1984;160:452-471.

11. Liao N-S, Maltzman J, Raulet DH. Positive selection determines T cell receptor Vß14 gene usage by CD8+ T cells. Journal of Experimental Medicine 1989;170:135-143.

12. Kubo RT, Born W, Kappler JW, Marrack P, Pigeon M. Characterization of a monoclonal antibody which detects all murine aß T cell receptors. Journal of Immunology 1989;142:2736-2742.

13. Prud'homme GJ, Bocarro DC, Luke ECH. Clonal deletion and autoreactivity in extrathymic CD4-CD8- double negative T cell receptor-a/ß T cells. Journal of Immunology 1991;147:3314-3318.

14. Webb S, Morris C, Sprent J. Extrathymic tolerance of mature T cells: clonal elimination as a consequence of immunity. Cell 1990;63:1249-1256.

15. Adorini L, Harvey MA, Miller A, Sercarz EE. Fine Specificity of regulatory T cells. II. Suppressor and helper T cells are induced by different regions of hen egg-white lysozyme in a genetically nonresponder mouse strain. Journal of Experimental Medicine 1979;150:293-306.

16. Germain RN, Benacerraf B. A single major pathway of T-lymphocyte interactions in antigen-specific immune suppression. Scandinavian Journal of Immunology 1981;13:1-10.

17. Herman A, Kappler JW, Marrack P, Pullen A. Superantigens: mechanism of T cell stimulation and role in immune responses. Annual Review of Immunology 1991;9:745-772.

18. Kappler JW, Staerz UD, White J, Marrack PC. Self-tolerance eliminates T cells specific for Mls-modified products of the major histocompatibility complex. Nature 1988;332:35-40.

19. Halle-Pannenko O, Pritchard LL, Motta R, Mathé G. Non-H-2 antigens can induce high GVH mortality in adult recipients of normal cells. Biomedicine 1978;29:253-261.

20. Bruley-Rosset M, Miconnet I, Canon C, Halle-Pannenko O. Mls-1a generated suppressor cells. I.Suppression is mediated by double-negative CD3+CD5+CD4-CD8- a/ß T cell receptor-bearing cells. Journal of Immunology 1990;145:4046-4052.

21. Mollick JA, Cook RG, Rich RR. Class II MHC molecules are specific receptors for staphylococcus enterotoxin A. Science 1989;244:817-820.

22. Fraser JD. High-affinity binding of staphylococcal enterotoxins A and B to HLA-DR. Nature 1989;339:221-223.

23. Fleischer B, Schrezenmeier, Conradt P. T lymphocyte activation by Staphylococcal Enterotoxins: role of class II molecules and T cell surface structures. Cellular Immunology 1989;120:92-99.

24. Mueller DL, Jenkins MK, Schwartz RH. Clonal expansion versus functional clonal inactivation: a costimulatory signalling pathway determines the outcome of T cell antigen receptor occupancy. Annual Review of Immunology 1989;7:445-480.

25. Jenkins MK, Taylor PS, Norton SD, Urdhal KB. CD28 delivers a costimulatory signal involved in antigen-specific IL-2 production by human T cells. Journal of Immunology 1991;147:2461-2466.

26. Linsley PS, Brady W, Grosmaire L, Aruffo A, Damle NK, Ledbetter JA. Binding of the B cell activation antigen B7 to CD28 costimulates T cell proliferation and IL-2 mRNA accumulation. Journal of Experimental Medicine 1991;173:721-730.

27. Conrad B, Weidmann E, Trucco G, et al. Evidence for superantigen involvement in insulin-dependent diabetes mellitus aetiology. Nature 1994;371:351-355.

Laurent Vanier received a Ph.D. from McGill University (Montreal, Quebec) in 1995. He is currently a second-year medical student at the University of Montreal (Montreal, Quebec). His work on peripheral T cell anergy was conducted at the Department of Pathology and Centre for Clinical Immunobiology and Transplantation at McGill University. His current research interests are in immunological tolerance. This project was funded by the Canadian Diabetes Association and the Medical Research Council of Canada.