Defective CD8 T Cell Peripheral Tolerance in Nonobese Diabetic Mice

Defective CD8ⴙ T Cell Peripheral Tolerance in Nonobese Diabetic Mice1 Huub T. C. Kreuwel,† Judith A. Biggs,† Ingrid M. Pilip,* Eric G. Pamer,* David Lo,† and Linda A. Sherman2† Nonobese diabetic (NOD) mice develop spontaneous autoimmune diabetes that involves participation of both CD4ⴙ and CD8ⴙ T cells. Previous studies have demonstrated spontaneous reactivity to self-Ags within the CD4ⴙ T cell compartment in this strain. Whether CD8ⴙ T cells in NOD mice achieve and maintain tolerance to self-Ags has not previously been evaluated. To investigate this issue, we have assessed the extent of tolerance to a model pancreatic Ag, the hemagglutinin (HA) molecule of influenza virus, that is transgenically expressed by pancreatic islet ␤ cells in InsHA mice. Previous studies have demonstrated that BALB/c and B10.D2 mice that express this transgene exhibit tolerance of HA and retain only low-avidity CD8ⴙ T cells specific for the dominant peptide epitope of HA. In this study, we present data that demonstrate a deficiency in peripheral tolerance within the CD8ⴙ T cell repertoire of NOD-InsHA mice. CD8ⴙ T cells can be obtained from NOD-InsHA mice that exhibit high avidity for HA, as measured by tetramer (KdHA) binding and dose titration analysis. Significantly, these autoreactive CD8ⴙ T cells can cause diabetes very rapidly upon adoptive transfer into NOD-InsHA recipient mice. The data presented demonstrate a retention in the repertoire of CD8ⴙ T cells with high avidity for islet Ags that could contribute to autoimmune diabetes in NOD mice. The Journal of Immunology, 2001, 167: 1112–1117.

I

nsulin-dependent diabetes mellitus (IDDM)3 occurs as the result of T cell-mediated destruction of the insulin-producing pancreatic ␤ cells and represents a failure on the part of the immune system to maintain self-tolerance. Susceptibility to diabetes is determined by both environmental and genetic factors. Genetic predisposition is polygenic, with as many as 15–20 genes involved in progression to type I diabetes in mice and humans (1). These observations suggest that autoimmune disease represents a situation in which numerous events conspire to override multiple checkpoints that normally maintain self-tolerance (2). One of the best rodent models of autoimmune diabetes is the nonobese diabetic mouse (NOD). These mice develop spontaneous diabetes that is mediated by both CD4⫹ and CD8⫹ T cells (3–9). The strongest genetic determinants of autoimmune diabetes in both mice and man is the MHC genotype. The presence of certain MHC class II molecules, HLA-DQ8 in humans and I-Ag7 in NOD mice, is strongly associated with disease susceptibility (10 –12). These molecules share the same nonaspartic acid substitution at residue 57 in the ␤-chain. As a consequence, the role of CD4⫹ T cells and the MHC class II molecules they recognize has been a major focus of research in diabetes. Several laboratories have demonstrated an *Sections of Infectious Diseases and Immunobiology, Yale University School of Medicine, New Haven, CT 06520; and †Department of Immunology, The Scripps Research Institute, La Jolla, CA 92037 Received for publication May 1, 2001. Accepted for publication May 7, 2001. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by National Institutes of Health Grants DK50824 and DK57644 (to L.A.S.). H.T.C.K. is a recipient of a Fellowship from the Juvenile Diabetes Foundation. 2 Address correspondence and reprint requests to Dr. Linda A. Sherman, Department of Immunology, The Scripps Research Institute, 10550 North Torrey Pines Road, IMM-15, La Jolla, CA 92037. E-mail address: [email protected] 3 Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; NOD, nonobese diabetic; HA, hemagglutinin; InsHA, transgenic mouse expressing the HA from influenza virus on pancreatic islet ␤ cells under the control of rat insulin promoter.

Copyright © 2001 by The American Association of Immunologists

enhanced proliferative response to self-Ags within the CD4⫹ T cell compartment (13–15). However, there is strong evidence that spontaneous autoimmune diabetes that occurs in NOD mice is mediated not only by CD4⫹ T cells, but also by CD8⫹ T cells. NOD mice bred to MHC class I-deficient mice (␤2m⫺/⫺) or treated with anti-CD8 Abs develop neither insulitis nor diabetes (16 –20). Also, it was found that proinsulin-specific CD8⫹ T cell populations from NOD mice can cause diabetes in the absence of CD4⫹ T cell help (21) and accumulate in the islets of young NOD mice (22). The current study was undertaken to determine the extent to which there may be a defect in tolerance to islet Ags within the CD8⫹ T cell repertoire of the NOD mouse. To address this question we have used a model pancreatic Ag, the hemagglutinin (HA) molecule of influenza virus, that is expressed on ␤ cells in the pancreatic islets under control of the rat insulin promoter (InsHA). When this molecule is expressed in BALB-InsHA or B10.D2-InsHA mice, tolerance is induced and maintained even after immunization with influenza virus (23–27). One hallmark of successful tolerance in this model is a fundamental alteration in the HAspecific repertoire such that all CD8⫹ T cells specific for HA exhibit low avidity for HA compared with the HA-specific response of conventional mice (23, 24, 27). We have previously reported that peripherally induced tolerance plays a major role in eliminating high-avidity CD8⫹ T cells specific for HA. HA-specific naive T cells encounter Ag that is cross-presented by professional APCs in the lymph nodes draining the pancreas. This results in their activation and is followed by their functional elimination (25, 26). We have investigated the degree of tolerance to HA within the CD8⫹ T cell compartment of NOD-InsHA mice. We show that under identical stimulatory conditions, high-avidity CD8⫹ T cells specific for this Ag can be retrieved from NOD-InsHA mice, but not from BALB-InsHA mice. Importantly, these HA-specific CD8⫹ T cells can cause diabetes very rapidly upon adoptive transfer into irradiated NOD-InsHA recipient mice. These data demonstrate a deficiency in self-tolerance within the CD8⫹ T cell repertoire of NOD mice such that CD8⫹ T cells with high affinity for an 0022-1767/01/$02.00

The Journal of Immunology

1113

islet Ag are retained within the repertoire. This may help explain why NOD mice develop diabetes.

counter (VWR Scientific, San Diego, CA). Relative cytotoxic activity (percentage) was calculated as: 100 ⫻ (sample release ⫺ spontaneous release)/ (maximum release ⫺ spontaneous release).

Materials and Methods

Flow cytometry

Mice

CTLs were stained 4 days after stimulation, purified through a Ficoll-Paque gradient (Pharmacia Biotech, Uppsala, Sweden), and then washed in FACS buffer (HBSS containing 1% (w/v) BSA; Sigma) and 0.02% (w/v) sodium azide (Sigma). Cells were incubated on ice with PE-conjugated KdHA tetramer and FITC-conjugated anti-CD8 (BD PharMingen, La Jolla, CA) for 1 h. Tetramers were produced as previously described, except that KdHA peptide was used (24). Cells were washed three times with FACS buffer. Propidium iodide (Sigma) was added after the final wash at 1 mg/ml to exclude dead cells in all experiments. Cells were analyzed with a FACScan and CellQuest software (BD Biosciences)

BALB/c and NOD/Shi were purchased from the breeding colony of The Scripps Research Institute (La Jolla, CA). InsHA-transgenic mice were generated and characterized as previously described (23) and were bred onto the BALB/c background for at least 10 generations (BALB-InsHA) or onto the NOD background for 13 generations (NOD-InsHA). All mice were bred and maintained under specific pathogen-free conditions in The Scripps Research Institute vivarium. All experimental procedures were conducted according to the guidelines laid out in the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Cell lines The SV40-transformed H-2d cell line B10.D2 was maintained in RPMI 1640 medium (Life Technologies, Gaithersburg, MD) supplemented with 10% (v/v) heat-inactivated FCS (Gemini Biological Products, Calabasas, CA), 2 mM glutamine (Life Technologies), 5 ⫻ 10⫺5 M 2-ME (Sigma, St. Louis, MO), and 50 mg/ml gentamicin (Gemini Bio-Products). Cells were cultured in a humidified incubator at 37°C with 5% (v/v) CO2 and were used as targets in CTL assays. The Bio.oz cell line was originally obtained from B. Knowles (University of Pennsylvania, Philadelphia, PA).

Peptide HA influenza virus A/PR/8/34 (H1N1) peptide (518IYSTVASSL526, restricted by H-2Kd) was synthesized by the core facility of The Scripps Research Institute using an Applied Biosystems model 430A synthesizer (Foster City, CA). Purity was ⬎85%, as determined by mass spectrometry and reversed-phase HPLC analysis on a Vydac C18 column (Vydac, Hesperia, CA).

Virus Influenza virus A/PR/8/34 H1N1 (PR8) was grown in the allantoic cavity of 10- to 11-day-old hen’s eggs. Upon isolation, the allantoic fluid was titrated for HA using chicken RBC and stored at ⫺70°C.

Production of effector CTL and adoptive transfers Eight-week-old mice were injected i.p. with 1200 HA units of influenza virus A/PR/8 in the form of allantoic fluid. Three weeks later, mice were sacrificed, and responder splenocytes were seeded into 24-well tissue culture plates at 7 ⫻ 106 cells/well in 1 ml complete RPMI 1640. APC splenocytes were irradiated (3000 rad) and then pulsed for 1 h with 5 nM KdHA peptide in a humidified incubator at 37°C with 5% (v/v) CO2. After washing three times in complete RPMI 1640 to remove unbound peptide, 1 ml containing 6 ⫻ 106 cells was added to the responder splenocytes and cultured in a humidified incubator at 37°C with 5% (v/v) CO2. For adoptive transfers, effector CTLs were injected i.v. into sublethally irradiated (600 rad) 4-wk-old BALB-InsHA or NOD-InsHA.

CD4⫹ T cell depletion CD4⫹ T cells were depleted in vitro by incubating splenocytes with antiCD4 Ab (RL172) supernatant on ice for 1 h. Cells were then centrifuged, the supernatant was discarded, and cells were resuspended in Low-Tox rabbit complement (Accurate Chemical and Scientific, Westbury, NY), and incubated for an additional hour at 37°C. Cells were then washed three times in complete RPMI 1640 and used for primary cultures. Depletion was complete as checked by staining cells with FITC-conjugated anti-CD4 Abs and analyzing with a FACScan and CellQuest software (BD Biosciences, Mountain View, CA).

Cytotoxicity assay Target cells were prepared by incubating B10.D2 cells at 37°C with 200 ␮Ci sodium 51Cr (NEN, Boston, MA) for 1 h in the presence or the absence of various concentrations of KdHA peptide as indicated. Target cells were washed four times, resuspended in complete RPMI 1640, and seeded into 96-well plates at 1 ⫻ 104 cells/well in 100 ␮l. Effector CTL were harvested 6 days after stimulation and seeded into duplicate wells containing the target cells at various E:T cell ratios, making a final volume of 200 ␮l. Plates were incubated at 37°C in a humidified incubator with 5% (v/v) CO2 for 5 h. Plates were centrifuged, and 100 ␮l supernatant was removed from each well to assess isotope release using an ICN Isomedic gamma radiation

Analysis of blood glucose level The glucose concentration in blood obtained from the retro-orbital plexus of mice was measured using the Accu-ChekIII (Roche, La Jolla, CA). Animals were considered diabetic if blood glucose levels were ⬎300 mg/dl.

Results

HA-specific CD8⫹ T cells persist in NOD-InsHA Previous studies have revealed that BALB-InsHA and B10.D2InsHA mice are profoundly tolerant to HA and do not develop diabetes even after infection with influenza virus (23). To determine whether CD8⫹ T cells in NOD mice undergo tolerance to a model pancreatic Ag, we produced NOD mice expressing the same InsHA transgene (NOD-InsHA). NOD-InsHA females develop spontaneous diabetes at a similar rate and incidence as normal NOD females (70% incidence by 28 wk of age) (28). Also, the level of HA expression on pancreatic ␤ cells is similar for NODInsHA and BALB-InsHA mice, as measured by flow cytometry using anti-HA Abs (data not shown). To assess the HA-specific CD8⫹ T cell repertoire in NOD-InsHA mice, recipient animals were immunized with influenza virus A/PR/8/34 H1N1 (PR8) that contained the same HA molecule as that used by the transgene. Three weeks after priming, spleen cells from BALB/c or NOD mice were restimulated in vitro with a concentration of KdHA peptide that has previously been shown to be optimal for stimulating high-avidity HA-specific CD8⫹ T cells (27). As previously shown, CD8⫹ T cells from BALB/c, but not BALB-InsHA, mice respond vigorously to HA (Fig. 1A). In contrast to BALB-InsHA, a good response to HA was detectable in NOD-InsHA mice (Fig. 1B). However, the response by NOD-InsHA mice was clearly weaker than that of nontransgenic NOD responders, suggesting that at least partial tolerance to HA had occurred as a result of expression of the InsHA transgene. It is also important to note that not every adult NOD-InsHA animal responded to HA, suggesting variability in the degree of tolerance induction. One explanation for the increased HA reactivity observed for NOD-InsHA CD8⫹ T cells compared with BALB-InsHA could be the help of autoreactive NOD-InsHA CD4⫹ T cells present during primary in vitro cultures. It has been shown that NOD CD4⫹ T cells can proliferate spontaneously toward self-peptides in vitro (13–15). Increased levels of cytokines could be responsible for the observed increased CD8⫹ T cell proliferation and cytotoxicity in response to HA. To determine whether CD4⫹ T cell help was necessary for the enhanced HA-specific CD8⫹ T cell response by NOD-InsHA, influenza virus-primed splenocytes from BALB/c and BALB-InsHA or NOD and NOD-InsHA were depleted of CD4⫹ T cells using anti-CD4 Abs and complement before in vitro culture. No decrease in CD8⫹ T cell cytotoxicity was observed if CD4⫹ T cells were depleted before establishment of the primary cultures (data not shown).

1114

FIGURE 1. HA-specific CD8⫹ T cells persist in NOD-InsHA. Splenocytes from influenza virus primed BALB/c (A, f), BALB-InsHA (A, 䡺), NOD (B, F), NOD-InsHA (B, E), BALB/c-InsHA neonates (C, Œ), and NOD-InsHA neonates (D, ‚) were cultured with homologous, irradiated (3000 rad) APCs pulsed with 5 nM KdHA peptide. After 6 days, specific lysis was tested in a 5-h chromium release assay on B10.D2 targets (H-2Kd positive) pulsed with 25 ␮M KdHA peptide. Background lysis on target cells without Ag was ⬍16% at the highest E:T cell ratio. Data are representative of three independent experiments.

Previous studies of the mechanism of tolerance to HA in BALBInsHA mice indicated that tolerance does not occur until several weeks after birth. This correlated with the ability of naive HAspecific CD8⫹ T cells to first detect HA Ag within the draining lymph nodes of the pancreas (25). Consistent with these previous results, both BALB-InsHA and NOD-InsHA neonates respond vigorously to HA (Fig. 1, C and D). These results strongly suggest that the peripheral deletion of HA-specific CD8⫹ T cells that normally occurs in BALB-InsHA and B10.D2-InsHA over time (25), does not occur in NOD-InsHA mice. HA-specific CD8⫹ T cells in NOD-InsHA mice have high avidity toward self-Ags Previous studies from this laboratory have demonstrated that the most significant difference between HA-specific CD8⫹ T cells from BALB-InsHA or B10.D2-InsHA mice and nontransgenic mice is a reduction in T cell avidity for HA (24, 27). This reduction is attributable to elimination of T cells that express TCRs with high affinity for HA. The residual low-avidity T cells are unable to cause diabetes (27). The observed high level of cytotoxicity by HA-specific CTLs from NOD-InsHA mice could be attributable to the presence of either a large number of CD8⫹ T cells with low avidity for HA or a small number of HA-specific cells with high avidity for HA. To distinguish between these alternative possibilities, the avidity of the HA-specific CD8⫹ T cells from NODInsHA was compared with that of HA-specific CD8 T cells from NOD mice on the basis of their lytic activity on targets pulsed with varying concentrations of KdHA peptide (Fig. 2A). Both NOD and NOD-InsHA responded with comparable avidities, although it took more NOD-InsHA CTLs to obtain comparable lysis (3- to 10-fold). This suggested that the HA-specific CD8⫹ T cells from

CD8⫹ T CELL TOLERANCE IN NOD MICE

FIGURE 2. HA-specific CD8⫹ T cells in NOD-InsHA mice have high avidity toward self-Ags, as measured by dose titration analysis and KdHA tetramer binding. A, Splenocytes from influenza virus-primed NOD (F) and NOD-InsHA (E) were cultured with homologous, irradiated (3000 rad) APCs pulsed with 5 nM KdHA peptide. After 6 days, specific lysis was tested in a 5-h chromium release assay on B10.D2 targets (H-2Kd positive) pulsed with the indicated concentration of KdHA peptide. Background lysis on target cells without Ag was ⬍15%. Data are for two animals per group and are representative of three independent experiments. Results are plotted at comparable lysis. B, Cells from secondary cultures of influenza virus-primed NOD and NOD-InsHA were purified through a Ficoll-Paque gradient and stained with anti-CD8-FITC Abs vs KdHA tetramer-PE Abs 4 days after stimulation with homologous, irradiated (3000 rad) APCs pulsed with 5 nM KdHA peptide. Dead cells were excluded using propidium iodide.

NOD and NOD-InsHA mice differed quantitatively, but were comparable in their avidity for HA. Another measure of T cell avidity is the amount of MHC-peptide tetramers bound by individual CD8⫹ T cells (27, 29). Using this technique, we have previously shown a reduced level of KdHA tetramer binding by CD8⫹ T cells derived from tolerant BALBInsHA or B10.D2-InsHA animals compared with CD8⫹ T cells from mice that do not express HA (27). Therefore, to further assess whether there was a difference in the avidity for HA between CD8⫹ T cells from NOD and NOD-InsHA mice, CD8⫹ T cells stimulated with HA were compared on the basis of their ability to bind KdHA tetramers (Fig. 2B). HA-specific CTLs from NODInsHA bound tetramer as well as those from NOD CTLs (mean fluorescence intensity, 134 vs 142), although there were fewer HAspecific CTLs present in the NOD-InsHA cultures (11 vs 40%). Taken together, these data indicate that NOD-InsHA mice exhibit the presence of reduced numbers of HA-specific cells with high avidity, rather than large numbers of CD8⫹ T cells with low avidity, for HA. HA-specific CD8⫹ T cell reactivity in unprimed NOD-InsHA mice, but not NOD These results demonstrate a significant HA response in NOD-InsHA mice that had been primed in vivo with influenza virus. To

The Journal of Immunology determine whether HA-specific CD8⫹ T cells could also be detected in unprimed NOD-InsHA mice, KdHA tetramers were used in an attempt to detect HA-specific CD8⫹ T cells in the spleen, lymph nodes, and pancreas of NOD-InsHA mice. This failed, presumably due to the low frequency of these cells in vivo. To increase the numbers of such cells, splenocytes from unprimed NOD-InsHA were stimulated in vitro with KdHA peptide-pulsed APCs. After three stimulations, HA reactivity was measured in a CTL assay using target cells pulsed with KdHA peptide (Fig. 3). By this method we could detect HA-specific CD8⫹ T cells from two of the seven NOD-InsHA tested. In contrast, none of the cultures from NOD mice developed HA-specific CTL. Adoptive transfer of HA-reactive CTLs into irradiated hosts Another measure of the degree to which InsHA mice are tolerant to HA is the fact that the HA-specific CD8⫹ T cells from B10.D2 or BALB/c, but not B10.D2-InsHA or BALB/c-InsHA mice, can cause diabetes upon adoptive transfer into InsHA mice (27) (Table I). To determine whether the HA-specific CD8⫹ T cells retrieved from NOD-InsHA mice could lead to ␤ cell destruction and diabetes, HA-reactive CTLs obtained from NOD-InsHA were adoptively transferred into irradiated NOD and NOD-InsHA mice (Table I). As shown in Table I, HA-reactive CTL lines from NODInsHA were able to cause diabetes very rapidly in irradiated NODInsHA recipient mice. No diabetes was observed if the recipients did not express HA in the pancreas, indicating that ␤ cell destruction was Ag specific.

Discussion We have previously shown that expression of HA in the pancreas is sufficient to achieve T cell tolerance in InsHA mice (23). InsHA mice that were thymectomized, irradiated, and then reconstituted

FIGURE 3. HA-specific CD8⫹ T cell reactivity in unprimed NOD-InsHA, but not NOD mice. Splenocytes from unprimed NOD and NODInsHA (8 wk of age) were stimulated weekly with homologous, irradiated (3000 rad) APCs pulsed with 5 nM KdHA peptide for 3 consecutive weeks. Six days after the last restimulation, specific lysis was tested in a 5-h chromium release assay on B10.D2 targets pulsed with 25 ␮M KdHA peptide. Data are the results of two independent experiments. Background lysis was ⬍20% at the highest E:T cell ratio.

1115 Table I. Adoptive transfer of HA-reactive CTLs into irradiated hostsa Adoptive Transfer

BALB3BALB-InsHA BALB-InsHA3BALB-InsHA NOD3NOD-InsHA NOD3NOD NOD-InsHA3NOD-InsHA NOD-InsHA3NOD PBS3NOD-InsHA

Diabetic/Total

8/8 0/8 7/7 0/8 8/8 0/7 0/6

a Groups of three to four mice were irradiated (700 rad) and injected i.v. with HA-reactive CTLs 4 days after stimulation in vitro. Animals were considered diabetic if blood glucose levels were above 300 mg/dl. Animals became diabetic within 10 days after transfer. Results are the total of two independently performed experiments with three to four mice per group.

with bone marrow and thymus from nontransgenic littermates demonstrated tolerance to HA. Further evidence of the important role of peripheral tolerance for this Ag is the fact that InsHA mice do not acquire tolerance until several weeks after birth (25). The requirements of the peripheral tolerance mechanism were elucidated in studies in which naive T cells from an HA-specific TCRtransgenic line, Clone-4 TCR mice, were adoptively transferred into InsHA mice (25, 26). This mechanism of peripheral tolerance requires that the self-Ag be cross-presented by professional APCs in the lymph nodes that drain the pancreas. There, HA-specific CD8⫹ T cells become activated, divide, and are eliminated. These results are similar to those of Heath and coworkers (30, 31), who first demonstrated this type of peripheral CD8⫹ T cell tolerance to a different model pancreatic Ag, OVA. The purpose of this study was to determine whether the CD8⫹ T cell compartment in NOD mice is responsive to self-Ags. Previous data have demonstrated reactivity to certain self-Ags within the CD4 compartment, such as glutamic acid decarboxylase, carboxypeptidase, insulin, 60-kDa heat shock protein, and peripherin (13–15). It was further shown that expression of an unusual MHC class II I-Ag7 is necessary for autoproliferation to occur, but is not sufficient, since CD4⫹ T cells from I-Ag7-positive B10 mice do not exhibit autoproliferation nor do CD4⫹ T cells from NOD mice that lack I-Ag7 molecules (32, 33). This suggests that multiple NOD genes are required for CD4⫹ T cell autoproliferation to occur. However, the progression toward autoimmune diabetes requires not only CD4⫹ T cells, but also CD8⫹ T cells (16 –20). It was therefore of interest to investigate whether CD8⫹ T cells exhibit autoreactivity to islet Ags. In this study, we present data demonstrating a defect in the peripheral deletion of CD8⫹ T cells specific for a model pancreatic self-Ag expressed in NOD mice. This defect results in the presence of autoreactive HA-specific CD8⫹ T cells that have high avidity toward this self-Ag and are very aggressive in causing autoimmune diabetes upon adoptive transfer into irradiated recipient mice. Such CD8⫹ T cells are normally eliminated from adult InsHA mice, as they are not found in either BALB-InsHA or B10.D2-InsHA mice (23, 24, 27). Although a defect in the induction of peripheral tolerance was found in NOD-InsHA mice, clearly some tolerance did occur, since the CTL response toward HA was significantly lower in some NOD-InsHA mice compared with that in NOD mice. It is tempting to speculate that such variability in CTL responses may parallel the significant variability that occurs in the age of onset of diabetes in NOD animals. The frequency of HA-reactive CD8⫹ T cells in unprimed NODInsHA is low, and several rounds of in vitro stimulation were required for their detection (Fig. 3). Although HA-specific CD8⫹ T cells are also present in unprimed NOD mice, no HA-specific

1116 CD8⫹ T cells were detected in cultures from NOD mice. One explanation for this could be that some HA-specific CD8⫹ T cells in NOD-InsHA mice are endogenously primed by pancreatic HA and that this facilitates their ability to respond in vitro to stimulation with HA peptide. Although the present study showed a defect in peripheral tolerance in the NOD mouse, we cannot exclude the possibility that there are additional defects that may impair thymic tolerance toward self-Ags in NOD mice. However, if this is the case, it is unlikely that the basis for any such defect in tolerance within the CD8 compartment is the same as the linear avidity model, first proposed by Fathman and coworkers (34). According to this model, the instability of the I-Ag7 molecule is believed to contribute to defective negative selection, which, in turn, is attributed to the autoproliferation in the CD4 compartment. However, the restriction element in the HA model (Kd) is the same in BALB/c and B10.D2 mice and therefore is unlikely to contribute to defective negative selection in the CD8 compartment. The increased reactivity to HA in NOD-InsHA animals within the CD8⫹ T cell compartment may be due to a number of different factors. It is possible that the presence of CD4⫹ T cell help prevents the peripheral deletion of autoreactive CD8⫹ T cells (35– 38). In a different transgenic model of CD8⫹ T cell tolerance it was shown that autoreactive CD4⫹ T cells could prevent peripheral deletion of CD8⫹ T cells induced by cross-presentation of selfAgs. This resulted in autoimmunity, rather than tolerance (38). Since NOD animals contain many autoreactive CD4⫹ T cells, this could prevent the normal induction of CD8⫹ T cell tolerance in the periphery. We are currently testing this hypothesis. Another possibility that could explain the defective peripheral tolerance of autoreactive CD8⫹ T cells is a general defect in T cell apoptosis and function. It has been shown that both mature NOD T lymphocytes as well as immature thymocytes can resist a variety of apoptosis-inducing stimuli, such as cyclophosphamide, dexamethasone, IL-2 deprivation, and activation-induced cell death (39 – 48). This defect in programmed cell death could contribute to defective thymic and peripheral deletion of autoreactive T cells in NOD mice. A third possibility could be a defect in APC function. Recent data suggest a crucial role for professional APCs in the induction and maintenance of T cell tolerance by cross-presentation of selfAgs (30, 31). In both NOD mice and diabetic patients, a defect in maturation and function was found in monocyte-derived dendritic cells (49 –52). It is possible that this defect in maturation leads to inefficient cross-presentation of self-Ag and therefore inefficient presentation of self-Ags to CD8⫹ T cells. We are currently investigating this possibility. In conclusion, we demonstrate a defect in the induction of CD8⫹ T cell peripheral tolerance in NOD mice, leading to the presence of high-avidity, islet-specific autoreactive CD8⫹ T cells that may contribute to autoimmune diabetes. In future experiments we will use CD8⫹ T cells from a TCR transgenic line specific for HA to elucidate the mechanisms responsible for such a tolerance defect. As there is no a priori reason to expect that tolerance to conventional islet Ags differs from tolerance toward HA, it is likely that NOD mice are defective in peripheral tolerance within their CD8 compartment toward other islet Ags as well. Indeed, the recent demonstration of high-avidity, ␤ cell-specific CD8⫹ T cells in NOD mice may reflect such a defect (22, 53).

References 1. Tisch, R., and H. McDevitt. 1996. Insulin-dependent diabetes mellitus. Cell 85: 291. 2. Sinha, A. A., M. T. Lopez, and H. McDevitt. 1990. Autoimmune diseases: the failure of self tolerance. Science 248:1380.

CD8⫹ T CELL TOLERANCE IN NOD MICE 3. Hanafusa, T., S. Sugihara, H. Fujino-Kurrihara, J. I. Miyagawa, A. Miyazaki, T. Yoshioka, K. Yamada, H. Nakajima, H. Asakawa, N. Kono, et al. 1988. Induction of insulitis by adoptive transfer with L3T4⫹ and Lyt2⫹ T -lymphocytes in T-lymphocyte depleted NOD mice. Diabetes 37:204. 4. Miller, B. J., M. C. Appel, J. J. O’Neil, and L. S. Wicker. 1988. Both the Lyt-2⫹ and L3T4⫹ T cell subsets are required for the transfer of diabetes in nonobese diabetic mice. J. Immunol. 140:52. 5. Bendelac, A., C. Carnaud, C. Boitard, and J. F. Bach. 1987. Syngeneic transfer of autoimmune diabetes from diabetic NOD mice to healthy neonates: requirement for both L3T4⫹ and Lyt-2⫹ T cells. J. Exp. Med. 166:823. 6. O’Reilly, L. A., P. R. Hutchings, P. R. Crocker, E. Simpson, T. Lund, D. Kioussis, F. Takei, J. Baird, and A. Cooke. 1991. Characterization of pancreatic islet cell infiltrates in NOD mice: effect of cell transfer and transgene expression. Eur. J. Immunol. 21:1171. 7. Reich, E. P., R. S. Sherwin, O. Kanagawa, and C. A. Janeway, Jr. 1989. An explanation for the protective effect of the MHC class II I-E molecule in murine diabetes. Nature 341:326. 8. Shimizu, J., O. Kanagawa, and E. R. Unanue. 1993. Presentation of ␤-cell antigens to CD4⫹ and CD8⫹ T cells of nonobese diabetic mice. J. Immunol. 151: 1723. 9. Nagata, M., P. Santamaria, T. Kawamura, T. Utsugi, and J. W. Yoon. 1994. Evidence for the role of CD8⫹ cytotoxic T cells in the destruction of pancreatic ␤ cells in nonobese diabetic mice. J. Immunol. 152:2042. 10. Acha-Orbea, H., and H. McDevitt. 1987. The first external domain of the nonobese diabetic mouse class II I-A ␤ chain is unique. Proc. Natl. Acad. Sci. USA 84:2435. 11. Todd, J. A., J. I. Bell, and H. McDevitt. 1987. HLA-DQ ␤ gene contributes to susceptibility and resistance to insulin-dependent diabetes mellitus. Nature 329: 599. 12. Steinman, L. 1995. Escape from “horror autotoxicus”: pathogenesis and treatment of autoimmune disease. Cell 80:7. 13. Kaufman, D. L., M. Clare-Salzler. J. Tian, T. Forsthuber, G. S. P. Ting, P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, and P. V. Lehmann. 1993. Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes. Nature 366:69. 14. Tisch, R., X. D. Yang, S. M. Singer, R. S. Liblau, L. Fugger, and H. McDevitt. 1993. Immune response to glutamic acid decarboxylase correlates with insulitis in non-obese diabetic mice. Nature 366:72. 15. Ridgway, W. M., M. Fasso, A. Lanctot, C. Garvey, and C. G. Fathman. 1996. Breaking self-tolerance in nonobese diabetic mice. J. Exp. Med. 183:1657. 16. Katz, J., C. Benoist, and D. Mathis. 1993. Major histocompatibility complex class I molecules are required for the generation of insulitis in non-obese diabetic mice. Eur. J. Immunol. 23:3358. 17. Wicker, L. S., E. H. Leiter, J. A. Todd, R. J. Renjilian, E. Peterson, P. A. Fischer, P. L. Podolin, M. Zijlstra, R. Jaenisch, and L. B. Peterson. 1994. ␤2-microglobulin-deficient NOD mice do not develop insulitis or diabetes. Diabetes 43:500. 18. Serreze, D. V., E. H. Leiter, G. J. Christianson, D. Greiner, and D. C. Roopenian. 1994. Major histocompatibility complex class I-deficient NOD-␤2mnull mice are diabetes and insulitis resistant. Diabetes 43:505. 19. Sumida, T., M. Furukawa, A. Sakamoto, T. Namekawa, T. Maeda, M. Zijlstra, I. Iwamoto, T. Koike, S, Yoshida, H. Tomioka, et al. 1994. Prevention of insulitis and diabetes in ␤2-microglobulin-deficient non-obese diabetic mice. Int. Immunol. 6:1445. 20. Wang, B., C. Gonzalez, C. Benoist, and D. Mathis. 1996. The role of CD8⫹ T cells in initiation of insulin-dependent diabetes mellitus. Eur. J. Immunol. 26: 1762. 21. Wong, S. F., I. Visintin, L. Wen, R. A. Flavell, and C. A. Janeway, Jr. 1996. CD8 T cell clones from young nonobese diabetic (NOD) islets can transfer rapid onset of diabetes in NOD mice in the absence of CD4 cells. J. Exp. Med. 183:67. 22. Wong, S. F., J. Karttunen, D. Dumont, L. Wen, I.Visintin, I. M. Pilip, N. Shastri, E. G. Pamer, and C. A. Janeway, Jr. 1999. Identification of an MHC class I-restricted autoantigen in type I diabetes by screening an organ-specific cDNA library. Nat. Med. 5:1026. 23. Lo, D., J. Freedman, S. Hesse, R. D. Palmiter, R. L. Brinster, and L. A. Sherman. 1992. Peripheral tolerance to an islet cell-specific hemagglutinin transgene affects both CD4⫹ and CD8⫹ T cells. Eur. J. Immunol. 22:1013. 24. Morgan, D. J., H. T. C. Kreuwel, S. Fleck, H. I. Levitsky, D. M. Pardoll, and L. A. Sherman. 1998. Activation of low avidity CTL specific for a self epitope results in tumor rejection but not autoimmunity. J. Immunol. 160:643. 25. Morgan, D. J., C. Kurts, H. T. C. Kreuwel, K. Holst., W. R. Heath, and L. A. Sherman. 1999. Ontogeny of T cell tolerance to peripherally expressed antigens. Proc. Natl. Acad. Sci. USA 96:3854. 26. Morgan, D. J., H. T. C. Kreuwel, and L. A. Sherman. 1999. Antigen concentration and precursor frequency determine the rate of CD8⫹ T cell tolerance to peripherally expressed antigens. J. Immunol. 163:723. 27. Nugent, C. T., D. J. Morgan, J. A. Biggs, A. Ko, I. M. Pilip, E. G. Pamer, and L. A. Sherman. 2000. Characterization of CD8⫹ T lymphocytes which persist after peripheral tolerance to a self antigen expressed in the pancreas. J. Immunol. 164:191. 28. Degermann, S., C. Reilly, B. Scott, L. Ogata, H. von Boehmer, and D. Lo. 1994. On the various manifestations of spontaneous autoimmune diabetes in rodent models. Eur. J. Immunol. 24:3155. 29. Busch, D. H., I. M. Pilip, S. Vijh, and E. G. Pamer. 1998. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 8:353. 30. Kurts, C., W. R. Heath, F. R. Carbone, J. Allison, J. A. F. P. Miller, and H. Kosaka. 1996. Constitutive class I-restricted exogenous presentation of self antigens in vivo. J. Exp. Med. 184:923.

The Journal of Immunology 31. Kurts, C., H. Kosaka, F. R. Carbone, J.F.A.P. Miller, and W. R. Heath. 1997. Class I-restricted cross-presentation of exogenous self-antigens leads to deletion of autoreactive CD8⫹ T cells. J. Exp. Med. 186:239. 32. Wicker, L. S., B. J. Miller., L. Coker, S. E. McNally, S. Scott, Y. Mullen, and M. C. Appel. 1987. Genetic control of diabetes and insulitis in the nonobese diabetic (NOD) mouse. J. Exp. Med. 165:1639. 33. Ridgway, W. M., H. Ito, M. Fasso, C. Yu, and C. G. Fathman. 1998. Analysis of the role of variation of major histocompatibility complex class II expression on nonobese diabetic (NOD) peripheral T cell response. J. Exp. Med. 188:2267. 34. Ridgway, W. M., and C. G. Fathman. 1999. A new look at MHC and autoimmune disease. Science 284:749. 35. Rees, M. A., A. S. Rosenberg, T. I. Munitz, and Singer, A. 1990. In vivo induction of antigen-specific transplantation tolerance to Qa1a by exposure to alloantigen in the absence of T-cell help. Proc. Natl. Acad. Sci. USA 87:2765. 36. Guerder, S., and P. Matzinger. 1992. A fail-safe mechanism for maintaining self-tolerance. J. Exp. Med. 176:553. 37. Kirberg, J., L. Bruno, and H. von Boehmer. 1993. CD4⫹8⫺ help prevents rapid deletion of CD8⫹ cells after a transient response to antigen. Eur. J. Immunol. 23:1963. 38. Kurts, C., F. R. Carbone, M. Barnden, E. Blanas, J. Allison, W. R. Heath, and J. F. A. P. Miller. 1997. CD4⫹ T cell help impairs CD8⫹ T cell deletion induced by cross-presentation of self-antigens and favors autoimmunity. J. Exp. Med. 186:2057. 39. Serreze, D. V., and E. H. Leiter. 1988. Defective activation of T suppressor cell function in nonobese diabetic mice: potential relation to cytokine deficiencies. J. Immunol. 140:3801. 40. Leijon, K., B. Hammarstro¨m, and D. Holmberg. 1994. Non-obese diabetic (NOD) mice display enhanced immune responses and prolonged survival of lymphoid cells. Int. Immunol. 6:339. 41. Zipris, D., A. H. Lazarus, A. R. Crow, M. Hadzija, and T. L. Delovitch. 1991. Defective thymic T cell activation by concanavalin A and anti-CD3 in autoimmune nonobese diabetic mice: evidence for thymic T cell anergy that correlates with the onset of insulitis. J. Immunol. 146:3763. 42. Zipris, D., A. R. Crow, and T. L. Delovitch. 1991. Altered thymic and peripheral T-lymphocyte repertoire preceding onset of diabetes in NOD mice. Diabetes 40:429. 43. Rapaport, M., J. A. Jaramillo, D. Zipris, A. H. Lazarus, D. V. Serreze, E. H. Leiter, P, Cyopick, J. S. Danska, and T. L. Delovitch. 1993. Interleukin 4

1117

44.

45.

46.

47.

48.

49.

50. 51. 52.

53.

reverses T cell proliferative unresponsiveness and prevents the onset of diabetes in nonobese diabetic mice. J. Exp. Med. 178:87. Garchon, H., J. J. Luan, L. Eloy, P. Be´dossa, and J. F. Bach. 1994. Genetic analysis of immune dysfunction in non-obese diabetic (NOD) mice: mapping of a susceptibility locus close to the Bcl-2 gene correlates with increased resistance of NOD T cells to apoptosis induction. Eur. J. Immunol. 24:380. Colucci, F., M. L. Bergman, C. Penha-Goncalves, C. M. Cilio, and D. Holmberg. 1997. Apoptosis resistance of nonobese diabetic peripheral lymphocytes linked to the Idd5 diabetes susceptibility region. Proc. Natl. Acad. Sci. USA 94:8670. Arreaza, G. A., M. J. Cameron, A. Jaramillo, B. M. Gill, D. Hardy, K. B. Laupland, M. J. Rapoport, P. Zucker, S. Chakrabarti, S. W. Chensue, et al. 1997. Neonatal activation of CD28 signaling overcomes T cell anergy and prevents autoimmune diabetes by an IL-4-dependent mechanism. J. Clin. Invest. 100:2243. Salojin, K., J. Zhang, M. Cameron, B. Gill, G. Arreaza, A. Ochi, and T. L. Delovitch. 1997. Impaired plasma membrane targeting of Grb2-murine son of sevenless (mSOS) complex and differential activation of the Fyn-T cell receptor (TCR)-␨-Cbl pathway mediate T cell hyporesponsiveness in autoimmune nonobese diabetic mice. J. Exp. Med. 186:887. Lamhamedi-Cherradi, S. E., J. J. Luan, L. Eloy, G. Fluteau, J. F. Bach, and H. J. Garchon. 1998. Resistance of T cells to apoptosis in autoimmune diabetic (NOD) mice is increased early in life and is associated with dysregulated expression of Bcl-x. Diabetologia 41:178. Serreze, D. V., H. R. Gaskins, and E. H. Leiter. 1993. Defects in the differentiation and function of antigen presenting cells in NOD/Lt mice. J. Immunol. 150: 2534. Langmuir, P. B., M. M. Bridgett, A. L. M. Bothwell, and I. N. Crispe. 1993. Bone marrow abnormalities in the non-obese diabetic mouse. Int. Immunol. 5:169. Jansen, A. M., M. van Hagen, and H. A. Drexhage. 1995. Defective maturation and function of antigen-presenting cells in type 1 diabetes. Lancet 345:491. Takahashi, K., M. C. Honeyman, and L. C. Harrison. 1998. Impaired yield, phenotype, and function of monocyte-derived dendritic cells in humans at risk for insulin-dependent diabetes. J. Immunol. 161:2629. Amrani, A., J. Verdaguer, P. Serra, S. Tafuro, R. Tan, and P. Santamaria. 2000. Progression of autoimmune diabetes driven by avidity maturation of a T-cell population. Nature 406:739.