The level of peptide-mhc complex determines the susceptibility to autoimmune diabetes: studies in HEL transgenic mice

Eur. J. Immunol. 2001. 31: 3453 3459 Autoimmune diabetes 3453 The level of peptide-mhc complex determines the susceptibility to autoimmune diabetes: studies in HEL transgenic mice Richard J. DiPaolo and Emil R. Unanue Department of Pathology and Immunology and Center for Immunology, Washington University School of Medicine, St. Louis, USA We report a mouse model for the spontaneous development of autoimmune diabetes: the 3A9 T cell receptor (TCR) transgenic mouse, which contains T cells that recognize the 52 61 family of hen egg-white lysozyme (HEL) peptides in the context of MHC class II I-A k molecules, was bred to the ILK3 mouse, that expresses HEL protein via the rat insulin promoter (RIP). Despite partial tolerance of 3A9 T cells in ILK3 mice, spontaneous diabetes developed in 64 % of 3A9xILK3 mice by 20 weeks of age. We provide evidence that APC from peripancreatic nodes have a large content of peptide-mhc complex and stimulate 3A9 T cells. We also report that cross presentation of HEL from beta cells to APC is 26-fold more efficient than presentation of soluble HEL. We previously reported on a biochemical margin of safety, based on the observation that activation of naive 3A9 T cells required 100-fold more peptide- MHC complexes than required for deletion of 3A9 thymocytes. We speculate that the high local density of autologous peptide-mhc complexes can be a determining factor that leads to the activation of autoreactive CD4 T cells and, consequently, to the development of autoimmunity. Key words: Antigen presentation / Diabetes / Autoimmunity / Cross-presentation / Tolerance Received 15/7/01 Accepted 17/9/01 1 Introduction Factors that lead to the initial activation of T cells that participate in autoimmune reactions are only partially understood. In an attempt to further our understanding of the events that lead to activation of self reactive T cells in autoimmune diabetes, several murine models have been developed in which beta cells of the islets of Langerhans are induced to express protein antigens which are then presented by either class I or class II MHC molecules [1 8]. Only a few systems have been developed in which mice transgenic for T cells reactive to a protein expressed transgenically in the pancreas have resulted in spontaneous diabetes [4, 5]. Studies have implicated roles for background genetics, the timing of antigen expression, and the frequency of autoreactive T cells in determining whether tolerance or autoimmunity develops [4, 6, 9 12]. Recent reports provide evidence that antigens expressed under the rat insulin promoter (RIP) can be presented in the lymph nodes draining the pancreas [10, Abbreviation: HEL: Hen egg-white lysozyme [I 22273] 13 15]. In some systems, cross-presentation within the lymph nodes draining the pancreas has been demonstrated to lead to tolerance of T cells [16 18]. Presentation of pancreatic antigens within the pancreatic lymph nodes has also been demonstrated to lead to activation of transgenic CD4 + T cells in other systems [14, 15]. It is not clear why antigen presentation with the pancreatic lymph nodes can lead to tolerance or activation, depending on the system. Factors may include: the level and timing of antigen expression, the affinity of the T cell receptor (TCR), the frequency of autoreactive T cells, and the level of peptide-mhc complexes presented within the lymph nodes. We report here on the development of spontaneous diabetes in mice bearing a TCR (3A9) to the chemically dominant epitope of hen egg lysozyme (HEL), when HEL is expressed in beta cells under the RIP. Although negative selection occurs in the thymus of 3A9xILK3 mice, it is incomplete, and autoreactive T cells enter the periphery and cause disease. We provide evidence that APC from peri-pancreatic nodes have a high number of peptide-mhc complex and stimulate proliferation of naive 3A9 T cells. We conclude that the level of peptide presentation in the pancreatic lymph nodes or pancreas is an important factor in the activation of self reactive T cells and the development of autoimmunity. WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001 0014-2980/01/1212-3453$17.50+.50/0

3454 R. J. DiPaolo and E. R. Unanue Eur. J. Immunol. 2001. 31: 3453 3459 2 Results and discussion 2.1 Spontaneous diabetes in 3A9xILK3 mice ILK3 mice, which express a membrane form of HEL on beta cells of the islets of Langerhans, did not develop diabetes. Diabetes did develop when ILK3 mice were crossed to the 3A9 TCR transgenic mice, which contain T cells reactive to the 52 61 core peptide from HEL in the context of the class II MHC I-A k (Fig. 1). After 20 weeks of age, 64 % of 3A9xILK3 mice became hyperglycemic (27 of 44 mice). Histological examination of the pancreas from all 3A9xILK3 mice showed typical lymphocytic infiltration in and around the islets (data not shown). These results differ from a previous report in which 3A9 TCR transgenic mice on the H-2 k background were crossed to the ILK3 mice on the H-2 b background, resultinginanh-2 k/b mixed MHC expression [8]. It was previously reported that positive selection of 3A9 T cells was inefficient on the H-2 k/b mixed background in the absence of HEL [19]. Indeed, we have confirmed that in the H-2 k/b background there are very few 3A9 TCRexpressing cells ( X 1 %) in the thymus and periphery, and those that are in the periphery proliferate less to HEL protein. We have repeated the studies with the 3A9xILK3 mice in the H-2 k/b background and also observed a low incidence of diabetes (1 in 19 mice). We attribute the difference in results, at least in part, to the effect of the H-2 k/b mixed MHC expression on the development of 3A9 T cells. The conclusion drawn in the previous report [8] was that 3A9 T cells were partially tolerized by expression of HEL in the pancreas, and could cause insulitis but not diabetes. However, in the present studies using the H-2 k/k background, 3A9 T cells do cause diabetes. 2.2 Evaluation of thymocytes in 3A9xILK3 mice Thymi were examined from 3A9 single transgenic and 3A9xILK3 double transgenic mice. 3A9 TCR expression was determined by FACS analysis with a clonotypic antibody, 1G12, specific for the 3A9 TCR (Fig. 2). Compared to the single transgenic mice, thymi from 3A9xILK3 mice contained a 2-fold decrease in total number of cells, a 7-fold reduction in the number of CD4 single-positive cells, and a 12-fold reduction in the number of CD4 single positive cells expressing the 3A9 TCR (Fig. 3 A, B, C). We conclude that negative selection of 3A9 thymocytes in ILK3 mice is incomplete. We previously showed that negative selection of 3A9 T cells was a very sensitive process that took place on a low number of peptide-mhc complexes. In an in vitro thymocyte culture system, negative selection of 3A9 double-positive thymocytes occurs between 1 and 30 peptide-mhc complexes per APC [20]. Similarly, we showed that negative selection of 3A9 T cells was complete in ML-5 mice, another HEL transgenic line that expresses soluble HEL in the serum at Fig. 1. Spontaneous development of diabetes in 3A9xILK3 mice. Blood glucose levels were monitored in 3A9xILK3 mice between 5 20 weeks of age. Mice were considered diabetic after two consecutive blood glucose readings greater than 250 mg/dl. Fig. 2. CD4 vs. CD8 profiles and 3A9 TCR expression. CD4 vs. CD8 staining of thymocytes from 3A9 (A) and 3A9xILK3 (B). There was a reduction in the number of CD4 singlepositive cells in 3A9xILK3 mice, indicating that negative selection was occurring. Cells were also stained with a clonotypic antibody for the 3A9 TCR (1G12). Panels C and D show the expression of 3A9 TCR gated on CD4 singlepositive cells from 3A9 (C) and 3A9xILK3 (D) mice.

Eur. J. Immunol. 2001. 31: 3453 3459 Autoimmune diabetes 3455 Fig. 3. Average cell numbers from the analysis of six individual 3A9 mice and nine individual 3A9xILK3 mice. Indicated are the total number of thymocytes (A) and spleen cells (D) and from each the number of CD4 + cells (B, E), and CD4 + 1G12 + positive cells (C, F). 10 20 ng/ml [21]. We have directly quantitated the amount of 48 62 peptide presented by thymic APC in ML-5 mice: from 132 mice thymi we detected 3.4 pmol of 48 63 peptides which is the order of 250 complexes per 1 10 6 MHC class II I-A k molecules [22]. In the ILK3 mice, HEL was detectable in the serum at levels between 1 2 ng/ml, which is about tenfold less than in the serum of ML-5 mice. We expect that the level of presentation of HEL by thymic APC to be also much lower than in ML-5, below the threshold required to delete all of the 3A9 T cells. Thus attempts to directly quantitate the content of 48 63 in thymic APC are not feasible. In separate experiments ILK3 +/ mice were immunized with 10 nmoles of HEL in CFA, and the number of HELreactive T cells was quantitated by limiting dilution analysis. The frequency of HEL-reactive T cells was reduced from 1 in 12,000 lymph node cells in ILK3 / mice to 1 in 41,000 in ILK3 +/ mice, indicating that partial tolerance was also seen in the normal T cell repertoire. Together these data indicate that the level of presentation in the thymus of ILK3 mice is at a low density, which allows for partial deletion of autoreactive T cells. 2.3 Evaluation of peripheral T cells in 3A9xILK3 mice We next examined the spleens of 3A9 and 3A9xILK3 mice. Compared to 3A9 single-transgenic mice, 3A9xILK3 mice contained similar numbers of total splenocytes, but the number of CD4 + cells expressing the 3A9 TCR were reduced by eightfold (Fig. 3, E, F). Fig. 4. Proliferation of 3A9 TCR-positive T cells from either 3A9 (open circles) or 3A9xILK3 (closed triangles) mice to HEL protein. Shown is the proliferation of 2 10 4 CD4 + cells expressing the 3A9 TCR (as determined by FACS analysis with a clonotypic antibody) in a total of 5 10 5 splenocytes per well. Results from two other experiments confirmed that T cells from 3A9xILK3 mice had an average of 50 % reduction in the amount of proliferation compared to T cells from 3A9 mice. Using the clonotypic antibody for the 3A9 TCR (1G12), we normalized the number of 3A9 TCR-positive cells and tested their ability to proliferate to HEL protein in vitro. The T cells from 3A9xILK3 mice maintained the ability to proliferate to HEL in vitro, but the level of proliferation was consistently lower than T cells from 3A9 singletransgenic mice (Fig. 4). 3A9 T cells from 3A9xILK3 mice had slightly lower levels of TCR and CD62L (data not shown). It is not clear whether the reduction in proliferation was caused by partial unresponsiveness due to a previous encounter with antigen in the periphery, or as a result of their selection in the thymus. 2.4 HEL presentation in peri-pancreatic lymph nodes In trying to determine the events that lead to diabetes in this model, we examined the level of HEL presentation in various lymph nodes of ILK3 mice by testing the proliferation of naive 3A9 T cells. In the present study, we isolated cells from the brachial, inguinal, and peripancreatic lymph nodes from ILK3 mice and tested their ability ex vivo to stimulate proliferation of naive 3A9 T cells. Fig. 5 shows that only APC from the peripancreatic lymph nodes of ILK3 mice contained enough complexes to stimulate proliferation of naive 3A9 T cells. The level of presentation by cells from the peripancreatic lymph nodes equaled the 50 % maximal levels of proliferation for these naive T cells stimulated by the lymph node cells pulsed with HEL, which is between 0.001 and 0.01 mm HEL (Fig. 5 inset). We previously determined that 300 1000 peptide-mhc complexes are required for half maximal proliferation of naive 3A9 T cells [20]. (The previous estimate was determined by pulsing APC with various amounts of peptide and directly

2.5 10 3456 R. J. DiPaolo and E. R. Unanue Eur. J. Immunol. 2001. 31: 3453 3459 Fig. 5. Proliferation by naive 3A9 T cells when co-cultured 1 10 6 cells from peri-pancreatic lymph nodes, brachial lymph nodes, and inguinal lymph nodes from either ILK3 +/ mice or ILK3 / negative controls. The inset shows the proliferation of 3A9 T cells to 1 10 6 lymph node cells pulsed with various amounts of HEL protein. Only cells isolated from peri-pancreatic lymph nodes of ILK3 mice presented enough HEL to lead to proliferation of naive 3A9 T cells. The presentation of HEL within the peri-pancreatic lymph nodes reached the 50 % maximum proliferation for 3A9 T cells, equal to the level of presentation of 0.003? M of soluble HEL protein. directly measuring the number of peptide-mhc complexes with a monoclonal antibody specific for the complex [23]). This level of presentation is achieved with the lymph nodes draining the pancreas, and is not reached in the spleen or more distant lymph nodes. 2.5 Cross-presentation of HEL from beta cells The high level of presentation of HEL in the lymph nodes draining the pancreas could be the result of HEL shedding from the surface of the beta cells and draining into the lymph nodes, or HEL that is transported by APC from the pancreas to the draining nodes. We tested the ability of HEL to be cross-presented by APC in vitro, using proliferation of naive 3A9 T cells as a readout for presentation. Single beta cell suspensions were made from pancreata of ILK3 +/ mice, and added to APCs and naive 3A9 transgenic T cells. Fig. 6 shows that beta cells from ILK3 mice were cross-presented very efficiently, leading to the proliferation of naive 3A9 T cells. In fact, as few as 1,000 beta cells led to levels of presentation comparable to about 700 pm exogenous HEL protein. Identical results were observed when islet cells were isolated from allogenic mice (data not shown). Separately, we directly quantitated the amount of HEL in beta cells: 1.7 10 5 beta cells yielded 900 fmol of HEL. Fig. 6. Proliferation of 3A9 T cells to HEL (A) and to islets from ILK3 +/ mice (B). Single-cell suspensions of beta cells were isolated from the pancreas ILK3 mice, and added in the indicated number with 1 10 6 splenocytes from 3A9 transgenic mice. Note that as few as 1,000 beta cells donated the equivalent of 700 pm HEL protein. Thus, when 1,000 beta cells were introduced into the culture they carried 5.3 fmol of HEL (26.5 pm): to achieve similar levels of stimulation by 3A9 T cells required 140 fmol of exogenous HEL (or 700 pm). Thus, the HEL associated with membranes of beta cells was 26-fold more efficient at being presented than when HEL was given as a soluble protein. (Assuming that there were 5 APC in the culture expressing the 300 peptide-mhc complexes, we calculate that for exogenous HEL, 1 peptide-mhc complex was formed from every 1,100 HEL molecules, while for the membrane HEL on beta cells, the number was 1 from every 42 HEL molecules, a remarkable efficiency for presentation). Although these data are not direct evidence for how the high levels of presentation are achieved in vivo, these are some of the first data that quantitatively measure the amount of antigen that can be cross-presented byapcfrombetacells. 2.6 Comments We have termed the difference between the levels of peptide-mhc complexes in the thymus required for deletion of double-positive thymocytes, and the levels required for activation of mature naive T cells, as the biochemical margin of safety [20]. For 3A9 transgenic T cells, thymic deletion occurs between 1 10 complexes per APC, while activation requires 100 1,000 complexes [20]. In this ILK3 system the level of presentation in the thymus partially deleted 3A9 T cells, indicating that the 48 62 peptide was presented at very low abundance, probably on the order of single digit complexes per APC. Normally, a 3A9 T cell that escapes negative selection would not be activated in the periphery because of the requirement for 100-fold higher levels of presentation in

Eur. J. Immunol. 2001. 31: 3453 3459 Autoimmune diabetes 3457 the periphery. Activation would require a relatively high level of peptide-mhc complexes displayed by the APC or a lowering of the threshold for T cell activation due to inflammation or other factors. In fact, several systems have shown that autoimmunity can be induced when the T cell threshold for activation is lowered, by providing signals to either the T cells or APC [2, 3, 24 30]. The system described here is unique in that the self reactive T cells are partially tolerized, yet they cause spontaneous diabetes without the need for an external stimulus to lower the threshold for activation, or to induce inflammation. Our conclusion is that the low level of presentation in the thymus allows for the escape of self reactive T cells, and the high local level of peptide presentation in the lymph node draining the pancreas allows for the activation of naive 3A9 T cells. This is an example of the biochemical margin of safety being broken, and the result is the development of spontaneous diabetes. We suggest that there are four conditions in ILK3 mice that contribute to the breaking of the biochemical margin of safety in this system. First is the limited availability of HEL for presentation in the thymus, which results in the escape of autoreactive cells into the periphery. But these cells will require 100-fold higher levels of peptide-mhc presentation to become activated in the periphery. Second is the high local concentration of the immunogen in the pancreas, i.e. calculated to be 3 10 6 HEL molecules per beta cell. Third is a strong chemically dominant peptide derived from HEL, which is one of the most effective peptides selected by any class II molecule [31, 32]. Indeed, the 52 61 family of peptides is highly selected by I-A k based on having highly favorable constellation of anchor amino acid [32 34]. The fourth is the highly efficient manner in which HEL is transferred from beta cells for presentation to 3A9 T cells. Our estimates indicate that 26-fold less HEL is required to achieve the number of complexes required to stimulate 50 % activation of naive 3A9 T cells when HEL is presented associated with beta cells compared to when it is offered as a soluble protein. These last three factors lead to the high levels of presentation of the 48 62 peptide within the pancreatic lymph nodes that result in breaking of the biochemical margin of safety, which allows for the activation of 3A9 T cells and the development of diabetes. Of course, other considerationsmayplayamajorroleintheinitiationofautoimmunity such as the number of effector cells, which in our case, with a transgenic TCR, is not limiting, or regulatory cells that may modulate the process. The lack of a requirement for an external signal, such as a viral infection, or immunization suggests that the level of class II-presented peptides can be sufficient to initiate the process of activation of both APC and CD4 T cells. Perhaps for this process to happen spontaneously, presentation of a high density of peptide-mhc complexes are required due to the lower levels of costimulatory molecules on resting APC. This level of presentation is probably achieved by limited number of peptides that are expressed in high concentrations in local tissues and are favored for presentation on a particular MHC. In addition to the other factors that have been implicated in determining whether self-reactive T cells are activated and participate in autoimmunity, these data suggest that the level of peptide presentation is one important parameter in the process of initiating autoimmune reactions. 3 Materials and methods 3.1 Mice The 3A9 mice used in this study were initially obtained from Dr. Mark Davis (Stanford University School of Medicine, Stanford, CA) [35]. ILK3 mice that express HEL under RIP were obtained from Dr. Christopher Goodnow [8]. These mice were back crossed to B10.BR and maintained in the Washington University mouse facility. 3A9 mice were screened by FACS analysis of peripheral blood lymphocytes with FITC conjugated anti-cd4 (GK1.5), and biotinylated anti-v g 8 (MR5-2). ILK3 mice were screened by PCR [8]. 3A9 mice were bred to ILK3 +/ mice and the resulting doubletransgenic mice were used in these studies. The 3A9xILK3 mice expressing MHC class II I-A k/b were generated by breeding the 3A9 (I-A k/k ) mouse to the ILK3 mice on the (I-A b/b ), and analyzing the resulting F1 mice. 3.2 Antibodies Anti-CD4 FITC (GK1.5), and biotinylated anti-v g 8 (MR5-2) were purchased from Pharmingen. Anti-CD8 quantum red was purchased from Sigma (Saint Louis, MO). The 1G12 antibody, which detects the 3A9 TCR, was produced by our laboratory and tested as a biotinylated purified antibody [21]. 3.3 Diabetes study Blood glucose levels were measured in 3A9xILK3 mice 1 2 times weekly between the ages of 5 and 20 weeks. Mice were considered diabetic after two consecutive blood glucose readings above 250 mg/dl. 3.4 Examination of lymphocytes Thymi, spleens and lymph nodes were removed and mechanically disrupted into single-cell suspension in DMEM with 10 % FCS. For 3A9 and 3A9xILK3 proliferation assays, 2 10 4 3A9 T cells (as determined by FACS analysis with the

3458 R. J. DiPaolo and E. R. Unanue Eur. J. Immunol. 2001. 31: 3453 3459 1G12 clonotypic antibody) in a total of 5 10 5 splenocytes were added in a 96-well U-bottom plate with the indicated amount of HEL protein. To maintain a constant cell number, 3A9 splenocytes were diluted into B10.BR splenocytes when normalizing the number of 3A9 TCR + cells. HEL protein was obtained from Sigma Chemical Co., St. Louis, MO. In assays for HEL presentation by lymph node APC, 1 10 6 cells from peri-pancreatic lymph nodes, brachial lymph nodes, or inguinal lymph nodes were added per well to 1 10 6 spleen and lymph node cells from 3A9 transgenic mice in 96-well U-bottom plates. For cross-presentation assays, single-cell suspensions of beta cells were prepared from pancreata of ILK3 +/ mice. The experiments were also done with beta cells from allogenic mice from the H-2 g7 background, with identical results. The indicated number of beta cells were added to 96-well U-bottom plates in the presence of 1 10 6 splenocytes from 3A9 trangenic mice. In proliferation assays, cells were cultured for 72 h and pulsed with 0.5? Ci/well of [ 3 H]thymidine during the last 12 18 h of culture. 3.5 HEL quantitation in beta cells 1.7 10 5 beta cells were lysed in 40 mm MEGA8/MEGA9 detergent (Sigma) in PBS with enzyme inhibitors (1 mm PMSF, 10 mm iodoacetamide, and 20? g/ml leupeptin). The suspensions were centrifuged at 8,000 rpm for 30 min at 4 C. To estimate the amount of HEL, a competitive ELISA was performed with a known standard of HEL protein, along with several dilutions of the beta cell extract. The anti-hel monoclonal antibody was F10.6.6, used at 0.05? g/ml. In this assay, a 50 % inhibition was obtained using 0.05 pmol of HEL protein. Acknowledgements: We thank Katherine Frederick for breeding and monitoring the blood glucose of the mice used in these studies, Carlos Velazquez for assistance in quantitatingthehelexpressedinbetacells,anddr.osamikanagawaandthomascirritoforhelpfulcommentsonthismanuscript. This study was supported by grants from the National Institutes of Health and the Kilo Diabetes and Vascular Research Foundation. References 1 Hanahan, D., Heritable formation of pancreatic beta-cell tumours in transgenic mice expressing recombinant insulin/simian virus 40 oncogenes. Nature 1985. 315: 115 122. 2 Ohashi, P. S., Oehen, S., Buerki, K., Pircher, H., Ohashi, C. T., Odermatt, B., Malissen, B., Zinkernagel, R. M. and Hengartner, H., Ablation of tolerance and induction of diabetes by virus infection in viral antigen transgenic mice.cell 1991. 65: 305 317. 3 Oldstone,M.B.,Nerenberg,M.,Southern,P.,Price,J.and Lewicki, H., Virus infection triggers insulin-dependent diabetes mellitus in a transgenic model: role of anti-self (virus) immune response. Cell 1991. 65: 319 331. 4 Scott, B., Liblau, R., Degermann, S., Marconi, L. A., Ogata, L., Caton, A. J., McDevitt, H. O. and Lo, D., A role for non-mhc genetic polymorphism in susceptibility to spontaneous autoimmunity. Immunity 1994. 1: 73 83. 5 Degermann, S., Reilly, C., Scott, B., Ogata, L., von, B. H. and Lo, D., On the various manifestations of spontaneous autoimmune diabetes in rodent models. Eur. J. Immunol. 1994. 24: 3155 3160. 6 Forster, I., Hirose, R., Arbeit, J. M., Clausen, B. E. and Hanahan, D., Limited capacity for tolerization of CD4 + T cells specific for a pancreatic beta cell neo antigen. Immunity 1995. 2: 573 585. 7 Morgan, D. J., Liblau, R., Scott, B., Fleck, S., McDevitt, H. O., Sarvetnick, N., Lo, D. and Sherman, L. A., CD8(+) T cellmediated spontaneous diabetes in neonatal mice. J. Immunol. 1996. 157: 978 983. 8 Akkaraju, S., Ho, W. Y., Leong, D., Canaan, K., Davis, M. M. and Goodnow, C. C., A range of CD4 T cell tolerance: partial inactivation to organ-specific antigen allows nondestructive thyroiditis or insulities. Immunity 1997. 7: 255 271. 9 Lo, D., Freedman, J., Hesse, S., Palmiter, R. D., Brinster, R. L. and Sherman, L. A., Peripheral tolerance to an islet cell-specific hemagglutinin transgene affects both CD4 + and CD8 + T cells. Eur. J. Immunol. 1992. 22: 1013 1022. 10 Forster, I. and Lieberam, I., Peripheral tolerance of CD4 T cells following local activation in adolescent mice. Eur. J. Immunol. 1996. 26: 3194 3202. 11 Kurts,C.,Sutherland,R.M.,Davey,G.,Li,M.,Lew,A.M.,Blanas, E., Carbone, F. R., Miller, J. F. A. P. and Heath, W. R., CD8 T cell ignorance or tolerance to islet antigens depends on antigen dose. Proc. Natl. Acad. Sci. USA 1999. 96: 12703 12707. 12 Morgan, D. J., Kurts, C., Kreuwel, H. T., Holst, K. L., Heath, W. R. and Sherman, L. A., Ontogeny of T cell tolerance to peripherally expressed antigens. Proc. Natl. Acad. Sci. USA 1999. 96: 3854 3858. 13 Kurts, C., Heath, W. R., Carbone, F. R., Allison, J., Miller, J. F. and Kosaka, H., Constitutive class I-restricted exogenous presentation of self antigens in vivo. J. Exp. Med. 1996. 184: 923 930. 14 Hoglund, P., Mintern, J., Waltzinger, C., Heath, W., Benoist, C. and Mathis, D., Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes. J. Exp. Med. 1999. 189: 331 340. 15 Sarukhan, A., Lechner, O. and von Boehmer, H., Autoimmune insulitis and diabetes in the absence of antigen-specific contact between T cells and islet beta-cells. Eur. J. Immunol. 1999. 29: 3410 3416. 16 Kurts, C., Kosaka, H., Carbone, F. R., Miller, J. F. and Health, W. R., Class I-restricted cross-presentation of exogenous selfantigens leads to deletion of autoreactive CD8(+) T cells. J. Exp. Med. 1997. 186: 239 245. 17 Heath, W. R., Kurts, C., Miller, J. F. A. P., and Carbone, F. R., Cross-tolerance: A pathway for inducing tolerance to peripheral tissue antigens. J. Exp. Med. 1998. 187: 1549 1553. 18 Adler,A.J.,Marsh,D.W.,Yochum,G.S.,Guzzo,J.L.,Nigam, A., Nelson, W. G., and Pardoll, D. M., CD4 + T cell tolerance to parenchymal self-antigens requires presentation by bone marrow-derived antigen-presenting cells. J. Exp. Med. 1998. 187: 1555 1564.

Eur. J. Immunol. 2001. 31: 3453 3459 Autoimmune diabetes 3459 19 Tourne, S., Kouskoff, V., Ho, W., Davis, M., M., Benoist, C. and Mathis, D., Inhibition of thymocyte positive selection by natural MHC: peptide ligands. Eur. J. Immunol. 1999. 29: 394 402. 20 Peterson, D. A., DiPaolo, R. J., Kanagawa, O. and Unanue, E. R., Cutting edge: Negative selection of immature thymocytes by a few peptide-mhc complexes: Differential sensitivity of immature and mature T cells. J. Immunol. 1999. 162: 3117 3120. 21 Peterson, D. A., DiPaolo, R. J., Kanagawa, O. and Unanue, E. R., Quantitative analysis of the T cell repertoire that escapes negative selection. Immunity 1999. 11: 453 462. 22 Velazquez, C., DiPaolo, R. and Unanue, E. R., Quantitation of lysozyme peptides bound to class II MHC molecules indicates very large differences in levels of presentation. J. Immunol. 2001. 2001: 5488 5494. 23 Dadaglio,G.,Nelson,C.A.,Deck,M.B.,Petzold,S.J.and Unanue, E. R., Characterization and quantitation of peptide- MHC complexes produced from hen egg lysozyme using a monoclonal antibody. Immunity 1997. 6: 727 738. 24 von Herrath, M. G., Dockter, J. and Oldstone, M. B., How virus induces a rapid or slow onset insulin-dependent diabetes mellitus in a transgenic model. Immunity 1994. 1: 231 242. 25 von Herrath, M. G., Guerder, S., Lewicki, H., Flavell, R. A. and Oldston, M. B., Coexpression of B7-1 and viral ( self ) transgenes in pancreatic beta cells can break peripheral ignorance and lead to spontaneous autoimmune diabetes. Immunity 1995. 3: 727 738. 26 von Herrath, M. G., Allison, J., Miller, J. F. and Oldstone, M. B., Focal expression of interleukin-2 does not break unresponsiveness to self (viral) antigen expressed in beta cells but enhances development of autoimmune disease (diabetes) after initiation of an anti-self immune response. J. Clin. Invest. 1995. 95: 477 485. 27 Harlan, D. M., Hengartner, H., Huang, M. L., Kang, Y. H., Abe, R., Moreadith, R. W., Pircher, H., Gray, G. S., Ohashi, P. S., Freeman, G. J. and et al., Mice expressing both B7-1 and viral glycoprotein on pancreatic beta cells along with glycoproteinspecific transgenic T cells develop diabetes due to a breakdown of T-lymphocyte unresponsiveness. Proc. Natl. Acad. Sci. USA 1994. 91: 3137 3141. 28 Harlan,D.M.,Barnett,M.A.,Abe,R.,Pechhold,K.,Patterson, N.B.,Gray,G.S.andJune,C.H.,Very-low-dose streptozotocin induces diabetes in insulin promoter-mb7-1 transgenic mice. Diabetes 1995. 44: 816 823. 29 Green, E. A., Eynon, E. E. and Flavell, R. A., Local expression of TNFalpha in neonatal NOD mice promotes diabetes by enhancing presentation of islet antigens. Immunity 1998. 9: 733 743. 30 Allison, J., Stephens, L. A., Kay, T. W., Kurts, C., Heath, W. R., Miller, J. F. and Krummel, M. F, The threshold for autoimmune T cell killing is influenced by B7-1. Eur. J. Immunol. 1998. 28: 949 960. 31 Nelson, C. A., Roof, R. W., McCourt, D. W. and Unanue, E. R., Identification of the naturally processed form of hen egg white lysozyme bound to the murine major histocompatibility complex class II molecule I-Ak. Proc. Natl. Acad. Sci. USA 1992. 89: 7380 7383. 32 Nelson, C. A., Viner, N. J., Young, S. P., Petzold, S. J. and Unanue,E.R,A negatively charged anchor residue promotes high affinity binding to the MHC class II molecule I-Ak. J. Immunol. 1996. 157: 755 762. 33 Fremont, D. H., Monnaie, D., Nelson, C. A., Hendrickson, W. A. and Unanue, E. R., Crystal structure of I-Ak in complex with a dominant epitope of lysozyme. Immunity 1998. 8: 305 317. 34 Latek, R. R., Petzold, S. and Unanue, E. R., Hindering auxiliary anchors are potent modulators of peptide binding and selection by I-Ak class II molecules. Proc.Natl.Acad.Sci.USA2000. 97: 11460 11465. 35 Ho,W.Y.,Cooke,M.P.,Goodnow,C.C.andDavis,M.M., Resting and anergic B cells are defective in CD28-dependent costimulation of naive CD4 + T cells. J. Exp. Med. 1994. 179: 1539 1549. Correspondence: Emil R. Unanue, Center for Immunology and the Department of Pathology, Washington University School of Medicine, St. Louis, MO 63110, USA Fax: 314-362-4096 e-mail: unanue pathology.wustl.edu