Although candidate genes controlling autoimmune disease can now be identified, a major challenge that remains is defining the resulting cellular events mediated by each locus. In the current study we have used NOD-InsHA transgenic mice that express the influenza hemagglutinin (HA) as an islet Ag to compare the fate of HA-specific CD8+ T cells in diabetes susceptible NOD-InsHA mice with that observed in diabetes-resistant congenic mice having protective alleles at insulin-dependent diabetes (Idd) 3, Idd5.1, and Idd5.2 (Idd3/5 strain) or at Idd9.1, Idd9.2, and Idd9.3 (Idd9 strain). We demonstrate that protection from diabetes in each case is correlated with functional tolerance of endogenous islet-specific CD8+ T cells. However, by following the fate of naive, CFSE-labeled, islet Ag-specific CD8+ (HA-specific clone-4) or CD4+ (BDC2.5) T cells, we observed that tolerance is achieved differently in each protected strain. In Idd3/5 mice, tolerance occurs during the initial activation of islet Ag-specific CD8+ and CD4+ T cells in the pancreatic lymph nodes where CD25+ regulatory T cells (Tregs) effectively prevent their accumulation. In contrast, resistance alleles in Idd9 mice do not prevent the accumulation of islet Ag-specific CD8+ and CD4+ T cells in the pancreatic lymph nodes, indicating that tolerance occurs at a later checkpoint. These results underscore the variety of ways that autoimmunity can be prevented and identify the elimination of islet-specific CD8+ T cells as a common indicator of high-level protection.

Immune tolerance is established and maintained by multiple mechanisms that begin during thymic development and continue in the periphery. As a result, the development of autoimmune diseases, such as type 1 diabetes, which is caused by autoimmune destruction of the insulin-producing β cells in the pancreatic islets, are generally of complex etiology and mediated by multiple genetic and environmental factors (1, 2). Results from family studies in humans and analyses of congenic strains of the spontaneously diabetic NOD mouse have shown that in addition to the NOD MHC region, at least 15 regions of the NOD genome collectively contribute to disease susceptibility (2, 3). In congenic NOD mice, the genetic regions controlling diabetes, designated as insulin-dependent diabetes (Idd)5 loci, have been replaced by resistant alleles obtained from non-diabetic strains of mice such as C57BL, allowing for analysis of their role in the disease process (4). Although the presence of resistance alleles at a single Idd locus can produce a measurable decrease in diabetes (5, 6, 7, 8), particular combinations of Idd loci have been shown to provide nearly complete protection from diabetes (5, 6, 9). Less than 2% of Idd3/5 mice and only 4% of Idd9 mice develop diabetes by 7 mo of age (9). The immune molecules encoded by the prime candidate genes defined in Idd3/5 and Idd9 mice include IL-2 (Idd3; Refs.5 and 6), IL-21 (2), CTLA-4 (Idd5.1; Refs.8 and 10), NRAMP1 (Idd5.2; Ref.11), and 4-1BB (Idd9.3; Ref.12).

We showed previously that mice expressing the hemagglutinin of influenza virus (HA) in pancreatic β cells are tolerant to HA in both the CD4+ and CD8+ T cell compartments (13). The only HA-specific T cells that can be retrieved from B10.D2 or BALB/c-InsHA mice demonstrate low avidity for Ag and are unable to cause diabetes. In contrast to these results, we showed that NOD-InsHA mice that express influenza HA in their pancreatic β cells retain autoreactive CD8+ T cells that have high avidity for self-Ag and can cause diabetes upon transfer into prediabetic NOD-InsHA mice, demonstrating that there is a defect in the induction of peripheral CD8+ T tolerance in NOD mice (14).

The factors underlying the retention of these high avidity anti-self-specific CD8+ T cells are highly influential in the development of autoreactivity. We hypothesize that critical genetic susceptibility loci regulate the maintenance of self-specific CD8+ T cells. We tested this hypothesis by determining the impact of highly protective NOD susceptibility regions, Idd3/5 and Idd9, on the behavior of islet Ag-specific CD8+ T cells.

BALB/c and B10.D2 mice were purchased from the breeding colony of The Scripps Research Institute. NOD/MrkTac mice were purchased from Taconic Farms. B10.D2-InsHA+/− transgenic mice were generated and characterized as previously described (13) and bred onto the BALB/c background for >10 generations (BALB-InsHA+/−) or onto the NOD background for 13 generations (NOD-InsHA+/−) (14). The NOD Idd3/5 congenic strain (B6 at Idd3, B10 at Idd5.1 and Idd5.2) has been described previously (9). The Idd3 region present in the Idd3/5 strain is an ∼2.7 Mbp congenic interval derived from the B6 strain that is between but does not include the proximal and distal markers D3Nds55 (35.7 Mbp) and D3Nds76 (38.4 Mbp) on chromosome 3 (6). The Idd5 region present in the Idd3/5 strain is an ∼28 Mbp congenic interval derived from the B10 strain that is between but does not include the proximal and distal markers D1Mit478 (53.2 Mbp) and D1Mit134 (52.9 Mbp) on chromosome 1 (11). The NOD Idd9 congenic strain (B10 at Idd9.1, Idd9.2, and Idd9.3) used in this study (available from Taconic Farms as Line 905) has a disease frequency identical (L. S. Wicker, unpublished observations) to that described previously for the NOD.B10 Idd9R28 congenic strain (6, 9). The Idd9 interval is an ∼38 Mbp region derived from the B10 strain that is between but does not include the proximal and distal markers D4Mit200 (115.3 Mbp) and D4Mit59 (153.6 Mbp) on chromosome 4 (our unpublished observations). Idd3/5 and Idd9 congenic mice were backcrossed to NOD-InsHA+/− transgenic mice and are referred to as Idd3/5-InsHA and Idd9-InsHA, respectively. CD8+Thy1.1+ clone-4 TCR BALB/c were previously described (15) and backcrossed to NOD for ≥12 generations. Idd9+/+ homozygous clone-4 Thy1.1+ mice were generated by intercrossing (NOD clone-4+ Thy1.1+ × Idd9+/+)F1 mice. The generation of BDC 2.5/NOD TCR transgenic mice has been described (16). Recombinant vaccinia virus expressing the H-2Kd-restricted epitope IYSTVASSL, amino acid residue 518–526 (Vac-KdHA) was provided by J. R. Bennink and J. Yewdell (National Institutes of Health, Bethesda, MD). Hemagglutinin influenza virus A/PR/8/34 (H1N1) peptide (518IYSTVASSL526, restricted by H-2Kd) was synthesized as described (14).

Eight-week-old mice were injected i.p. with 107 PFU of recombinant vaccinia-KdHA (Vac-KdHA). Three weeks later, mice were sacrificed and responder splenocytes were cultured as described (14). Effector CTL were harvested 6 days after stimulation and relative cytotoxic activity was determined as described (14).

CD25+ T cells were depleted in vivo by injecting 300 μg of PC61 mAb (endotoxin status <0.5 endotoxin units/ml) i.p. 3 days before injection of clone-4 cells. Recipient mice treated with purified rat IgG (Jackson ImmunoResearch Laboratories) served as isotype controls. After 3 days, at which time mice were injected with CFSE-labeled CD8+Thy1.1+ clone-4 cells, the degree of depletion was determined by staining PBLs of treated mice with FITC-conjugated anti-CD25 Abs (7D4) and analyzing with a FACScan and CellQuest software (BD Biosciences). This resulted in >85% CD4+CD25+ T cell depletion in PC-61-treated mice.

Naive CD8+Thy1.1+ clone-4 TCR cells were isolated from the inguinal, axillary, cervical, and mandibular lymph nodes of NOD clone-4 TCR mice (6–8 wk of age) as described (17). CD4+Vβ4+ T cells from BDC 2.5 mice and CD4+ T cells from either NOD-InsHA or Idd 3/5-InsHA were isolated as described (17). CFSE labeling was performed as described (18). Eight- to 10-wk-old female recipient mice were injected i.v. in the tail vein with either 5 × 106 purified CD8+Thy1.1+ clone-4 TCR or BDC 2.5 cells in a volume of 200 μl of HBSS on day 0. In some experiments, CD25+ T cells were removed during CD4+ T cell isolation and 15 × 106 of CD4+ or CD4+CD25 T cells were adoptively transferred i.v. in the tail vein of NOD-InsHA and Idd3/5-InsHA recipients 3 days prior to injection of 5 × 106 purified CFSE-labeled CD8+Thy1.1+ clone-4 TCR.

Four days after clone-4 CFSE-labeled transfer, recipient mice were sacrificed and pancreatic lymph nodes (PancLN) and a mixture of other lymph nodes including inguinal, axillary, cervical, and mandibular were excised and processed separately to obtain single cell suspensions. NOD clone-4 cells were stained with Thy1.1-PE and CD8-PerCP, BDC 2.5 cells were stained with CD4-PerCP and Vβ4-biotin followed by streptavidin-PE. All mAbs were obtained from BD Biosciences. Cells were analyzed with a FACS HTS (BD Biosciences) and FlowJo software (Tree Star)

Four days after injections of 5 × 106 purified CD8+Thy1.1+ clone-4 cells, islets were isolated using a modified previously published protocol (19). Briefly, the PancLN were removed and the remaining pancreas tissue was carefully cut into small pieces using fine scissors and digested in 1 mg/ml collagenase P (Roche Diagnostics) and 1 μg/ml DNase I (Roche Diagnostics) by shaking (300 rpm) at 37°C for 20 min. The islets were then purified by centrifugation on Ficoll gradient (Histopaque-1077; Sigma-Aldrich). Thereafter, to obtain single cells, islets were dissociated using a nonenzymatic dissociation solution (Sigma-Aldrich).

The FTY720 analog AAL-(R), which inhibits lymphocyte egress from lymph nodes and thymus (20, 21, 22, 23), was obtained from Dr. V. Brinkmann (Novartis, Basel, Switzerland). The drug was dissolved in water and injected i.p. (3 mg/kg) into 2-wk-old NOD-InsHA or 7-wk-old BALB/c-InsHA recipients. NOD-InsHA mice were treated continuously with FTY720 every other day until the time of analysis. The mice received 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 NOD T cells injected i.p. at 6 wk of age and were sacrificed for analysis 4 days later. At that time, the pancreata were excised, fixed overnight in 10% (v/v) formalin solution (Sigma-Aldrich), and processed for paraffin embedding. Paraffin-embedded tissue was cut by using a microtome, and sections were placed onto saline-coated Superfrost slides for processing (Fisher Scientific). Tissue sections were deparaffinized in xylene and rehydrated in graded ethanol to distilled water. Sections were stained with eosin (Sigma-Aldrich) and counterstained with Mayer’s hematoxylin and analyzed for the presence of infiltrates. Pooled peripheral or PancLN were analyzed by flow cytometry as described above. BALB/c-InsHA mice were treated continuously with FTY720 for 7 days and received 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells on day 3 after the initiation of treatment. Recipient mice were sacrificed 4 days later for analysis of peripheral and PancLN.

All statistical analysis was performed with a Student t test. Values of p > 0.05 were considered insignificant.

Clone-4 TCR mice express a TCR specific for a Kd-restricted epitope of the influenza HA obtained originally from an HA-specific CTL clone derived from a B10.D2 mouse immunized with the A/PR8/34 virus strain of influenza (15). CD8+ T cells were purified from NOD, BALB/c, or B10.D2 Thy1.1 clone 4 TCR mice, labeled with CFSE, and adoptively transferred into syngeneic InsHA recipients. As reported previously, naive clone-4 cells first encounter HA Ag that is cross-presented in the PancLN of InsHA recipients (24). On the BALB/c or B10.D2 background, clone-4 cells undergo an abortive activation that results in deletion (Fig. 1,A). In contrast, we found the NOD clone-4 T cells proliferated vigorously and there was a 2- to 3-fold increase in the percentage of clone-4 cells that accumulated as dividing cells (Fig. 1,A). As a result, many more dividing cells were recovered from NOD-InsHA mice (17,447 ± 3,718) as compared with BALB/c-InsHA or B10.D2-InsHA (2,717 ± 435, and 4,888 ± 1,261, p < 0.001 and p = 0.004). Such proliferation is Ag specific, and does not occur in the PancLN of HA-negative NOD mice (Fig. 1 B).

FIGURE 1.

Increased proliferation of NOD clone-4 cells in the PancLN of NOD-InsHA mice as compared with BALB/c-InsHA and B10.D2-InsHA. A, A total of 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 BALB/c or B10.D2 cells were injected into three to four syngeneic InsHA hosts; B, 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 NOD cells were injected into NOD or 8- and 4-wk-old NOD-InsHA recipient mice. Four days later, cells were pooled from peripheral or pancreatic lymph nodes and analyzed by FACS. Histograms depict the amount of CSFE in gated CD8+Thy1.1+ cells. The percentage of proliferating cells (one or more divisions) is indicated in each histogram. Data are representative of 1 of 10 experiments each including three to four mice per group and each with similar results.

FIGURE 1.

Increased proliferation of NOD clone-4 cells in the PancLN of NOD-InsHA mice as compared with BALB/c-InsHA and B10.D2-InsHA. A, A total of 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 BALB/c or B10.D2 cells were injected into three to four syngeneic InsHA hosts; B, 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 NOD cells were injected into NOD or 8- and 4-wk-old NOD-InsHA recipient mice. Four days later, cells were pooled from peripheral or pancreatic lymph nodes and analyzed by FACS. Histograms depict the amount of CSFE in gated CD8+Thy1.1+ cells. The percentage of proliferating cells (one or more divisions) is indicated in each histogram. Data are representative of 1 of 10 experiments each including three to four mice per group and each with similar results.

Close modal

Idd3/5 mice resist the occurrence of diabetes, insulitis, and the development of insulin autoantibodies as compared with NOD mice (9), suggesting they have re-established functional tolerance. To determine whether genetic susceptibility loci affect the parameter of CD8+ T cell tolerance, NOD-InsHA and Idd3/5-InsHA mice were each immunized with a vaccinia recombinant virus expressing the HA epitope recognized by Kd-restricted CD8+ T cells (vac-KdHA), and the HA-specific CTL response was compared. Unlike NOD-InsHA mice, the Idd3/5-InsHA mice were found to be tolerant to HA, as they displayed a poor HA-specific CD8+ T cell response that is similar to that of non-diabetes prone strains, such as BALB/c-InsHA and B10.D2-InsHA (Fig. 2 A).

FIGURE 2.

HA-specific CD8+ T cell reactivity is observed in NOD-InsHA recipients, but not in BALB/c-InsHA, B10.D2-InsHA, Idd3/5-InsHA, or Idd9-InsHA hosts. A, Splenocytes from vaccinia-KdHA-primed BALB-InsHA, B10.D2-InsHA, NOD-InsHA, Idd3/5-InsHA, Idd9-InsHA, or (B) BALB/c, B10.D2, NOD, Idd3/5, and Idd9 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 μM Kd HA peptide. Background lysis on target cells without Ag was below 10% at the highest E:T ratio. Data represent one of three experiments with similar results and depicts the mean-specific lysis of eight mice (A) and four mice (B) per group.

FIGURE 2.

HA-specific CD8+ T cell reactivity is observed in NOD-InsHA recipients, but not in BALB/c-InsHA, B10.D2-InsHA, Idd3/5-InsHA, or Idd9-InsHA hosts. A, Splenocytes from vaccinia-KdHA-primed BALB-InsHA, B10.D2-InsHA, NOD-InsHA, Idd3/5-InsHA, Idd9-InsHA, or (B) BALB/c, B10.D2, NOD, Idd3/5, and Idd9 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 μM Kd HA peptide. Background lysis on target cells without Ag was below 10% at the highest E:T ratio. Data represent one of three experiments with similar results and depicts the mean-specific lysis of eight mice (A) and four mice (B) per group.

Close modal

Idd9 mice are also highly protected from diabetes, but unlike Idd3/5, Idd9 mice develop insulitis and insulin autoantibodies (6, 9). To assess the influence of this genetic region on CD8+ T tolerance, Idd9-InsHA mice were immunized with vac-KdHA. Again, it was found that CD8+ T tolerance to HA was restored (Fig. 2,A). Therefore, in contrast to conventional NOD-InsHA mice, mice expressing protective Idd3/5 or Idd9 regions achieve normal CD8+ T tolerance to the HA islet Ag. HA-negative BALB/c, B10.D2, NOD, and Idd3/5 or Idd9 littermates were all found to respond vigorously to HA (Fig. 2 B).

Based on our observation that the lack of CD8+ T tolerance in NOD-InsHA mice correlated with an accumulation of HA-specific clone-4 cells in the PancLN, we next assessed the accumulation of NOD clone-4 cells in the Idd congenic lines. As compared with results obtained with NOD-InsHA recipients, clone-4 cells exhibited reduced accumulation in PancLN of Idd3/5-InsHA congenic mice (Fig. 3). Thus, the expression of protective alleles at Idd3, Idd5.1, and Idd5.2 by the recipients blocked the accumulation of HA-specific CD8+ T cells. Importantly, activated clone-4 cells were present in the islets of NOD-InsHA, but few cells were found in the islets of Idd3/5-InsHA recipients, consistent with the observed lack of accumulation of activated clone-4 cells in the PancLN of Idd3/5-InsHA recipients (Fig. 3).

FIGURE 3.

Islet-specific CD8+ T cells accumulate in the PancLN and migrate to the islets of NOD-InsHA and Idd9-InsHA mice but not in the Idd3/5-InsHA. A, A total of 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected into NOD-InsHA, Idd3/5-InsHA, and Idd9-InsHA hosts. Four days after transfer, recipient mice were sacrificed and lymphocytes from peripheral lymph nodes, PancLN, and pancreas were isolated and analyzed by FACS as described in Fig. 1. The percentage of proliferating cells (one or more divisions) is indicated in each histogram and are representative of the results obtained from eight independent experiments, each containing three to four mice per group. B, The total numbers of clone-4 cells in PancLN is depicted and represent the mean results obtained from four independent experiments. C, The absolute numbers of divided clone-4 cells was calculated by multiplying the total numbers of clone-4 cells in the PancLNs by the percentage of dividing cells. The error bars denote the variation corresponding to one SD.

FIGURE 3.

Islet-specific CD8+ T cells accumulate in the PancLN and migrate to the islets of NOD-InsHA and Idd9-InsHA mice but not in the Idd3/5-InsHA. A, A total of 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected into NOD-InsHA, Idd3/5-InsHA, and Idd9-InsHA hosts. Four days after transfer, recipient mice were sacrificed and lymphocytes from peripheral lymph nodes, PancLN, and pancreas were isolated and analyzed by FACS as described in Fig. 1. The percentage of proliferating cells (one or more divisions) is indicated in each histogram and are representative of the results obtained from eight independent experiments, each containing three to four mice per group. B, The total numbers of clone-4 cells in PancLN is depicted and represent the mean results obtained from four independent experiments. C, The absolute numbers of divided clone-4 cells was calculated by multiplying the total numbers of clone-4 cells in the PancLNs by the percentage of dividing cells. The error bars denote the variation corresponding to one SD.

Close modal

Surprisingly, when we followed the fate of NOD clone-4 cells adoptively transferred into Idd9-InsHA mice, we found expression of protective alleles of Idd9 by the recipient did not prevent the accumulation of clone-4 cells (Fig. 3). Indeed, the level of accumulation of activated NOD clone-4 cells in Idd9-InsHA recipients was comparable to that seen in NOD-InsHA recipients in both the percentage of proliferating cells and absolute numbers (Fig. 3).

We found no obvious differences in the expression levels of several key activation markers (CD44, CD62L, and CD49d) by the activated clone-4 cells of NOD-InsHA, Idd9-InsHA, and Idd3/5-InsHA recipients (data not shown).

In the results presented above the protective Idd9 loci were expressed only by the host, and not by the NOD clone-4 cells. To determine whether endogenous expression of the protective Idd9 allele by the clone-4 CD8+ T would promote their deletion, we produced NOD clone-4 Thy 1.1 mice that are homozygous for the expression of Idd9. Upon adoptive transfer, Idd9 clone-4 CFSE-labeled cells were found to accumulate in the PancLN of both NOD-InsHA and Idd9-InsHA recipient mice (Fig. 4). This result indicates that expression of Idd9 by CD8+ T cells does not affect their accumulation in the PancLN.

FIGURE 4.

Idd9 clone 4 islet-specific CD8+ T cells accumulate in the PancLN of NOD-InsHA and Idd9-InsHA mice. A total of 5 × 106 syngeneic purified CSFE-labeled CD8+Thy1.1+Idd9+/+ clone-4 cells were injected into NOD-InsHA and Idd9-InsHA mice. Four days later, cells from PancLN were pooled and analyzed by FACS as described in Fig. 1.

FIGURE 4.

Idd9 clone 4 islet-specific CD8+ T cells accumulate in the PancLN of NOD-InsHA and Idd9-InsHA mice. A total of 5 × 106 syngeneic purified CSFE-labeled CD8+Thy1.1+Idd9+/+ clone-4 cells were injected into NOD-InsHA and Idd9-InsHA mice. Four days later, cells from PancLN were pooled and analyzed by FACS as described in Fig. 1.

Close modal

To determine whether the presence of the Idd resistance intervals alter expression levels of the HA transgene in NOD congenic β cells, purified islet cells from NOD-InsHA, Idd3/5-InsHA, and Idd9-InsHA were stained with the mouse 37/38 anti-HA mAb followed by FITC-conjugated goat anti-mouse mAb. Levels of HA expression by islets cells from NOD-InsHA and Idd-InsHA congenic mice were found to be highly similar to each other, and similar to the level in BALB-InsHA mice (data not shown).

To assess whether increased clone-4 accumulation would be due to increased availability of cross-presented HA Ag through islet destruction (25, 26), we examined the amount of proliferation/accumulation of clone-4 T cells in the PancLN of NOD-InsHA mice experiencing minimal insulitis. First, we used the sphingosine-1-phosphate receptor agonist, FTY720, which efficiently blocks egress of T cells from lymph nodes (20). Continuous treatment of NOD mice with FTY720 prevents both islet infiltration and the occurrence of diabetes (27, 28). Of significance to our studies, despite its drastic effect on T cell circulation and migration, FTY720 does not prevent T cell activation in secondary lymphoid tissue (22, 29).

Islet Ag-specific T cells first become activated in the PancLN of NOD mice at 3 wk of age (30). Therefore, to prevent migration of activated T cells to the islets, treatment with FTY720 was initiated when the NOD-InsHA mice were 2 wk of age. After continuous treatment for 4 wk, mice received CFSE-labeled clone-4 cells and were sacrificed 4 days later to assess the extent of lymphocyte infiltration in the islets, and the accumulation of clone-4 cells in the PancLN. The islets of recipients treated with FTY720 showed very little infiltration. Only 1 of 22 islets (4.5%) examined showed mild peri-insulitis. In contrast, 26 of the 37 islets (70%) examined from the untreated NOD-InsHA mice showed clear evidence of either insulitis (51%) or peri-insulitis (19%), indicating that the presence of the drug successfully reduced islet infiltration. Despite this, the percentage of activated clone-4 cells accumulating in the PancLN of FTY720 treated NOD-InsHA recipients was still 2.5-fold greater than the number in FTY720 treated BALB-InsHA recipients (Fig. 5,A), indicating that insulitis was not required to cause the enhanced accumulation of clone-4 cells seen in NOD-InsHA recipients. It should be noted that the percentage of proliferating clone-4 cells was reduced in the PancLN of both NOD-InsHA and BALB-InsHA recipients as a result of treatment with FTY720 (compare Fig. 5,A with Fig. 1). This did not reflect a decrease in the numbers of activated clone-4 cells, but rather a large increase in numbers of naive clone-4 cells accumulating in lymph nodes as a result of the drug treatment.

FIGURE 5.

Increased clonal expansion of NOD clone-4 cells in the PancLN of NOD-InsHA mice can occur in the absence of pancreatic T cell infiltrates. A total of 5 × 106 syngeneic purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected into (A) mice treated with FTY720 as described in Materials and Methods and (B) adults and 4-wk-old NOD mice. Four days after transfer, cells from PancLN from three to four mice were isolated, pooled, and analyzed by FACS as described in Fig. 1.

FIGURE 5.

Increased clonal expansion of NOD clone-4 cells in the PancLN of NOD-InsHA mice can occur in the absence of pancreatic T cell infiltrates. A total of 5 × 106 syngeneic purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected into (A) mice treated with FTY720 as described in Materials and Methods and (B) adults and 4-wk-old NOD mice. Four days after transfer, cells from PancLN from three to four mice were isolated, pooled, and analyzed by FACS as described in Fig. 1.

Close modal

As a second approach, clone-4 cells were transferred into 4-wk-old NOD-InsHA recipients, an age at which insulitis is minimal (31). The young mice exhibited levels of proliferation similar to that observed in adult NOD-InsHA mice (Fig. 5 B). Taken together, these results indicate that islet destruction is not responsible for enhanced accumulation of clone-4 cells in NOD-InsHA recipients.

We next determined whether islet Ag-specific CD4+ T cells, like islet Ag-specific CD8+ T cells, would also be found to accumulate in the PancLN of NOD and Idd9, but not in Idd3/5, recipients. To this end, we examined the response of CFSE-labeled CD4+ T cells from NOD BDC2.5 mice that express a class II-restricted, islet Ag-specific TCR (16). Upon adoptive transfer, these CD4+ T cells were found to accumulate in the PancLN of both NOD and Idd9 recipients (53 ± 2%, and 48 ± 4%, respectively), but not in Idd3/5 mice (23 ± 8%, Fig. 6). Thus, both CD4+ and CD8+ T cells specific for islet Ags accumulated in NOD and Idd9 recipients, but not in Idd3/5.

FIGURE 6.

Proliferation of transferred BDC2.5 cells in the PancLN of NOD Idd9 mice but not in NOD Idd3/5 mice. A total of 5 × 106 purified CSFE-labeled CD4+Vβ4+ BDC 2.5 cells were injected into three to four NOD, Idd3/5 and Idd9 hosts (8 wk of age). Histograms depict the amount of CSFE, in gated CD4+Vβ4+ cells. Data correspond to three to four mice per group and are representative of one of three experiments with similar results.

FIGURE 6.

Proliferation of transferred BDC2.5 cells in the PancLN of NOD Idd9 mice but not in NOD Idd3/5 mice. A total of 5 × 106 purified CSFE-labeled CD4+Vβ4+ BDC 2.5 cells were injected into three to four NOD, Idd3/5 and Idd9 hosts (8 wk of age). Histograms depict the amount of CSFE, in gated CD4+Vβ4+ cells. Data correspond to three to four mice per group and are representative of one of three experiments with similar results.

Close modal

Several reports have demonstrated an important role for CD4+CD25+ regulatory T cells (Tregs) in maintaining tolerance in the periphery (32, 33). To assess the role of CD4+CD25+ Tregs in preventing the accumulation of clone-4 cells in the PancLN of NOD-InsHA mice and Idd3/5-InsHA, anti-CD25 mAb was given to mice to deplete Tregs 3 days before the transfer of clone-4 cells. In contrast to NOD-InsHA mice, such depletion enhanced the accumulation of clone-4 cells in the PancLN of Idd3/5-InsHA recipients (53 ± 10% vs 27 ± 3%, p = 0.004, Fig. 7,A) restoring it to the level seen in NOD-InsHA recipients. This finding raised an interesting question: are Tregs also required to promote CD8+ T tolerance in other non-diabetes prone strains of mice, such as B10.D2 and BALB/c? To address this issue, we examined the proliferation of CFSE-labeled B10.D2 clone-4 cells in B10.D2-InsHA recipients that received prior treatment with anti-CD25. Elimination of CD25+ cells did not alter the amount of accumulation, suggesting CD25+ Tregs are not normally required to promote deletion of islet-specific CD8+ T cells in conventional strains of mice (21 ± 4% vs 24 ± 14%, p = 0.773, Fig. 7 A). Similar results were obtained on the BALB/c background (data not shown).

FIGURE 7.

CD4+CD25+ Treg depletion in Idd3/5-InsHA mice increases the accumulation of islet-reactive CD8+ T cells. A, A total of 300 μg of PC61 (anti-CD25) mAb were injected into NOD-InsHA and Idd3/5-InsHA or B10.D2-InsHA mice. Three days later, mice received 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells of NOD or B10.D2 origin. B, A total of 15 × 106 purified CD4+ T cells from either Idd3/5-InsHA or NOD-InsHA mice and 15 × 106 purified CD4+CD25 from Idd3/5-InsHAmice were injected i.v. into Idd3/5-InsHA recipient mice and 3 days later 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected. Four days after transfer, cells from PancLN from three to four mice were isolated and analyzed by FACS as described in Fig. 1. The data are representative of one of three experiments with similar results including each time four mice per group.

FIGURE 7.

CD4+CD25+ Treg depletion in Idd3/5-InsHA mice increases the accumulation of islet-reactive CD8+ T cells. A, A total of 300 μg of PC61 (anti-CD25) mAb were injected into NOD-InsHA and Idd3/5-InsHA or B10.D2-InsHA mice. Three days later, mice received 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells of NOD or B10.D2 origin. B, A total of 15 × 106 purified CD4+ T cells from either Idd3/5-InsHA or NOD-InsHA mice and 15 × 106 purified CD4+CD25 from Idd3/5-InsHAmice were injected i.v. into Idd3/5-InsHA recipient mice and 3 days later 5 × 106 purified CSFE-labeled CD8+Thy1.1+ clone-4 cells were injected. Four days after transfer, cells from PancLN from three to four mice were isolated and analyzed by FACS as described in Fig. 1. The data are representative of one of three experiments with similar results including each time four mice per group.

Close modal

These results suggested that Idd3/5 mice contain an endogenous population of autoreactive CD4+ T cells that were kept in check by the activity of Tregs so they could no longer promote the accumulation of newly activated islet Ag-specific T cells. To directly assess the presence of such a population, we first purified CD4+ cells from Idd3/5-InsHA mice and tested their ability to promote accumulation of clone-4 cells when coinjected into Idd3/5-InsHA recipients. As anticipated, no such accumulation was evident (Fig. 7,B). However, if the CD4+ population was depleted of CD25+ cells before coinjection with the clone-4 cells, a significant enhancement in the level of accumulation of clone-4 cell was observed (71 ± 5% vs 34 ± 6%, p = 0.0013, Fig. 6). Thus, removal of CD25+ cells from total Idd3/5-InsHA CD4+ T cells revealed the presence of an endogenous CD4+ population that could promote the accumulation of clone-4 cells. Of interest, we could also promote the accumulation of clone-4 cells in Idd3/5-InsHA recipients by coinjection with total, purified CD4+ T cells from NOD-InsHA donors (62 ± 1% vs 34 ± 6%, p = 0.001, Fig. 7 B).

We demonstrate that two nonoverlapping combinations of resistance alleles, Idd3/5 and Idd9, each of which provides potent protection from diabetes, repair the defect in CD8+ T cell tolerance in NOD-InsHA mice, as assessed by the reduced response to HA subsequent to infection with a virus that contains the dominant Kd-HA peptide. Furthermore, this is accomplished by distinct mechanisms in each protected line, which further promotes the hypothesis that restoration of CD8+ T cell tolerance may be critical to achieve a high level of protection from diabetes.

When clone-4 T cells were used to assess the fate of islet-specific CD8+ T cells as they first encounter Ag, we observed that both NOD-InsHA and Idd9-InsHA mice accumulated large numbers of proliferating CFSE-labeled clone-4 T cells in their PancLN whereas Idd3/5-InsHA (Fig. 3), B10.D2-InsHA, and BALB/c-InsHA mice exhibited 3-fold less accumulation (Figs. 1,A and 3). In contrast to the results reported by King et al. (34), we found that proliferation of clone-4 cells in NOD mice was dependent on the presence of HA Ag. This difference may be due to the fact that they examined proliferation of CD8+ T cells in older diabetic mice.

As observed for clone-4 CD8+ T cells, both NOD and Idd9 mice accumulated greater numbers of CD4+ BDC2.5 cells in their PancLN than did Idd3/5 recipients (Fig. 6). This was somewhat surprising as it was reported previously that there is no significant difference in proliferation of BDC2.5 cells in NOD and non-diabetes prone C57BL/6 H2g7 congenic hosts (35). Considering that the protective Idd3/5 genetic regions are the same in C57BL mice and Idd3/5 mice, this suggests that other genetic regions on the C57BL background may modify the level of activation/accumulation of the BDC2.5 cells. Indeed, BDC2.5 cells are far more aggressive in causing diabetes on the C57BL/6 background than in NOD mice, suggesting the presence of protective genetic regions on the NOD background not found in C57BL mice (36).

Due to the fact that the recipient strains in which large numbers of clone-4 and BDC2.5 cells accumulate, NOD and Idd9, both produce islet Ag-specific Abs and exhibit islet infiltrates (9), we considered the possibility that enhanced proliferation of clone-4 cells may result from increased availability of islet Ags through β cell destruction by infiltrating CD4+ and CD8+ T cells (25). However, neither of the two strategies that were used to decrease insulitis succeeded in reducing the accumulation of activated CD8+ T cells in adult or young NOD-InsHA (Fig. 5) and also in young congenic Idd9-InsHA (data not shown), indicating that insulitis and islet destruction is not required for the enhanced accumulation of clone-4 cells. Furthermore, it was reported recently that elimination of T cell-mediated β cell destruction does not reduce the levels of cross-presented β cell autoantigen in the PancLN of NOD mice (37). These data strongly argue that enhanced accumulation in the PancLN of NOD mice is not the result of islet damage by T cells. Thus, there is a fundamental deficiency in peripheral tolerance of CD8+ and CD4+ T cells in NOD mice that occurs at the site where they first encounter Ag. Preliminary results suggest that enhanced delivery of costimulatory signals by dendritic cells may underlie the accumulation of clone-4 cells, as provision of anti-B7.1 and anti B7.2 mAbs markedly reduced the level of accumulation of clone-4 cells in the PancLN of NOD-InsHA mice (X. Martinez and H. T. C. Kreuwel, unpublished results).

A major clue to the mechanism responsible for the reduced accumulation of islet Ag-specific CD4+ and CD8+ T cells in Idd3/5 recipients was obtained when we found that the removal of CD25+ cells from Idd3/5-InsHA mice resulted in a profound enhancement in the accumulation of transferred clone-4 cells in the PancLN. It is of interest that such tolerance could also be overcome by providing purified CD4+ T cells obtained from NOD-InsHA mice (Fig. 7 B). This suggests that the Tregs in the Idd3/5 mice are unable to suppress the activity of the large number of transferred autoreactive CD4+ T cells. Consistent with this possibility, we found that although cotransfer of purified CD4+ T cells from Idd3/5-InsHA mice and clone-4 cells into Idd3/5-InsHA hosts did not increase the accumulation of clone-4 cells, if the CD4+ T cell population was first depleted of CD25+ cells, a significant increase in the accumulation of clone-4 cells was observed. This suggests a balance between Tregs and a population of autoreactive CD4+ T cells is responsible for maintaining tolerance in Idd3/5 mice. This concept is compatible with recent observations made by Bluestone and colleagues (38, 39) showing that diabetes in the NOD mouse results from a disruption in the balance between autoreactivity and immune regulation in the periphery. In addition, we have observed that CD4+ T cells from Idd3/5-InsHA donors are unable to suppress the activation of clone-4 T cells in NOD-InsHA recipients, thus suggesting the balance in NOD mice is tipped in favor of activation of islet Ag-specific T cells (X. Martinez and L. A. Sherman, unpublished observation).

The observation that Tregs were instrumental in maintaining tolerance during the initial activation of islet-specific T cells in the PancLN of Idd3/5 mice raised the important question of whether Tregs served a similar role in promoting peripheral tolerance in conventional, non-diabetes prone strains of mice. However, removal of CD25+ T cells did not increase the accumulation of clone-4 cells in the PancLN of B10.D2-InsHA (Fig. 7 A) or BALB/c-InsHA (not shown) recipients. Therefore, non-diabetes prone mice do not require the presence of Tregs in their PancLN to promote CD8+ T tolerance. Surprisingly, removal of CD25+ T cells in NOD-InsHA had no effect on the accumulation of clone-4 cells suggesting that Tregs cells are either not fully functional or not able to control the activation of clone-4 in NOD mice. The percentage of CD4+CD25+ in the PancLN of NOD-InsHA (10.4%) and Idd3/5-InsHA (12.4%) did not differ significantly. Taken together these results suggest that the underlying basis for autoimmunity has not been eliminated in Idd3/5-InsHA mice and that self-reactive T cells are continuously held in check by Tregs. This may be due to the fact that Idd3/5 mice still harbor diabetes-promoting alleles of many other Idd genes. Nevertheless, Idd3/5 mice have achieved an effective way to suppress the disease at an early stage of T cell activation through the development of Tregs. Additional insights into the mechanisms by which tolerance is restored in Idd3/5 mice should be discernible in future studies using mice expressing either Idd3 or Idd5 protective loci.

Although also highly protected from diabetes, Idd9 mice produce insulin-specific autoantibodies and have islet infiltrates (9, 12) Idd9 mice accumulate islet Ag-specific CD4+ T (BDC2.5) and CD8+ T cells (clone-4) in their PancLN. Therefore, it was quite surprising to find that Idd9-InsHA mice have restored CD8+ T cell tolerance to HA (Fig. 2,A). We next considered the possibility that to achieve tolerance, the transferred T cells must express protective Idd9. However, we found that Idd9 clone-4 cells also accumulated within the PancLN of Idd9-InsHA recipients (Fig. 4). Thus, the expression of protective Idd9 genes does not affect the initial activation of islet Ag-specific CD8+ T cells, but rather, it affects the ability of the activated CD8+ T cells to either develop effector function and/or to survive after leaving the PancLN. Future studies will address the mechanism of CD8+ T cell tolerance in these mice.

In conclusion, our data demonstrates that islet-reactive CD8+ T cells are vulnerable to tolerance induction at multiple stages in the periphery, and implies that multigenic cooperativity is required to perpetuate a sustained, destructive islet attack. This provides a variety of independent targets with potential usefulness for modulating tolerance. Our results also have important implications for strategies to predict diabetes susceptibility in humans, as they suggest the hypothesis that a common feature of individuals highly protected from diabetes may be a deficiency in islet Ag-specific CD8+ T cells. This may preclude the occurrence of disease, even in individuals that harbor autoreactive CD4+ T cells.

We thank all members of the Sherman laboratory for discussions, Kristi Marquardt and Judith Biggs for excellent technical assistance, Shelly Gassert for excellent secretarial assistance, and the Flow Cytometry Core Facility at The Scripps Research Institute.

The authors have no financial conflict of interest.

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 grants from the National Institutes of Health (DK57644, DK50824, and T32-AI07606 (to L.A.S.); AI055509-01 (to H.R.)), the Juvenile Diabetes Research Foundation (JDRF) and the Wellcome Trust (to L.S.W. and K.H.). X.M. is a recipient of a Fellowship from the Swiss National Science Foundation (PBGEB-103070 and PA00A-105081). H.T.C.K. is a recipient of a Fellowship from the JDRF. The availability of NOD congenic mice through the Taconic Farms Emerging Models Program has been supported by grants from the Merck Genome Research Institute, National Institute of Allergy and Infectious Diseases, and the JDRF.

5

Abbreviations used in this paper: Idd, insulin-dependent diabetes; HA, influenza hemagglutinin; PancLN, pancreatic lymph node; Treg, regulatory T cell.

1
Atkinson, M. A., G. S. Eisenbarth.
2001
. Type 1 diabetes: new perspectives on disease pathogenesis and treatment.
Lancet
358
:
221
-229.
2
Todd, J. A., L. S. Wicker.
2001
. Genetic protection from the inflammatory disease type 1 diabetes in humans and animal models.
Immunity
15
:
387
-395.
3
Onengut-Gumuscu, S., P. Concannon.
2002
. Mapping genes for autoimmunity in humans: type 1 diabetes as a model.
Immunol. Rev.
190
:
182
-194.
4
Hamilton-Williams, E. E., D. V. Serreze, B. Charlton, E. A. Johnson, M. P. Marron, A. Mullbacher, R. M. Slattery.
2001
. Transgenic rescue implicates β2-microglobulin as a diabetes susceptibility gene in nonobese diabetic (NOD) mice.
Proc. Natl. Acad. Sci. USA
98
:
11533
-11538.
5
Podolin, P. L., M. B. Wilusz, R. M. Cubbon, U. Pajvani, C. J. Lord, J. A. Todd, L. B. Peterson, L. S. Wicker, P. A. Lyons.
2000
. Differential glycosylation of interleukin 2, the molecular basis for the NOD Idd3 type 1 diabetes gene?.
Cytokine
12
:
477
-482.
6
Lyons, P. A., N. Armitage, F. Argentina, P. Denny, N. J. Hill, C. J. Lord, M. B. Wilusz, L. B. Peterson, L. S. Wicker, J. A. Todd.
2000
. Congenic mapping of the type 1 diabetes locus, Idd3, to a 780-kb region of mouse chromosome 3: identification of a candidate segment of ancestral DNA by haplotype mapping.
Genome Res.
10
:
446
-453.
7
Greve, B., L. Vijayakrishnan, A. Kubal, R. A. Sobel, L. B. Peterson, L. S. Wicker, V. K. Kuchroo.
2004
. The diabetes susceptibility locus Idd5.1 on mouse chromosome 1 regulates ICOS expression and modulates murine experimental autoimmune encephalomyelitis.
J. Immunol.
173
:
157
-163.
8
Vijayakrishnan, L., J. M. Slavik, Z. Illes, R. J. Greenwald, D. Rainbow, B. Greve, L. B. Peterson, D. A. Hafler, G. J. Freeman, A. H. Sharpe, et al
2004
. An autoimmune disease-associated CTLA-4 splice variant lacking the B7 binding domain signals negatively in T cells.
Immunity
20
:
563
-575.
9
Robles, D. T., G. S. Eisenbarth, N. J. Dailey, L. B. Peterson, L. S. Wicker.
2003
. Insulin autoantibodies are associated with islet inflammation but not always related to diabetes progression in NOD congenic mice.
Diabetes
52
:
882
-886.
10
Ueda, H., J. M. Howson, L. Esposito, J. Heward, H. Snook, G. Chamberlain, D. B. Rainbow, K. M. Hunter, A. N. Smith, G. Di Genova, et al
2003
. Association of the T-cell regulatory gene CTLA4 with susceptibility to autoimmune disease.
Nature
423
:
506
-511.
11
Wicker, L. S., G. Chamberlain, K. Hunter, D. Rainbow, S. Howlett, P. Tiffen, J. Clark, A. Gonzalez-Munoz, A. M. Cumiskey, R. L. Rosa, et al
2004
. Fine mapping, gene content, comparative sequencing, and expression analyses support Ctla4 and Nramp1 as candidates for Idd5.1 and Idd5.2 in the nonobese diabetic mouse.
J. Immunol.
173
:
164
-173.
12
Lyons, P. A., W. W. Hancock, P. Denny, C. J. Lord, N. J. Hill, N. Armitage, T. Siegmund, J. A. Todd, M. S. Phillips, J. F. Hess, et al
2000
. The NOD Idd9 genetic interval influences the pathogenicity of insulitis and contains molecular variants of Cd30, Tnfr2, and Cd137.
Immunity
13
:
107
-115.
13
Lo, D., J. Freedman, J. Hesse, R. D. Palmiter, R. L. Brinster, 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
-1022.
14
Kreuwel, H. T., J. A. Biggs, I. M. Pilip, E. G. Pamer, D. Lo, L. A. Sherman.
2001
. Defective CD8+ T cell peripheral tolerance in nonobese diabetic mice.
J. Immunol.
167
:
1112
-1117.
15
Morgan, D. J., R. Liblau, B. Scott, S. Fleck, H. O. McDevitt, N. Sarvetnick, D. Lo, L. A. Sherman.
1996
. CD8+ T cell-mediated spontaneous diabetes in neonatal mice.
J. Immunol.
157
:
978
-983.
16
Katz, J. D., B. Wang, K. Haskins, C. Benoist, D. Mathis.
1993
. Following a diabetogenic T cell from genesis through pathogenesis.
Cell
74
:
1089
-1100.
17
Lyman, M. A., S. Aung, J. A. Biggs, L. A. Sherman.
2004
. A spontaneously arising pancreatic tumor does not promote the differentiation of naive CD8+ T lymphocytes into effector CTL.
J. Immunol.
172
:
6558
-6567.
18
Hernandez, J., S. Aung, K. Marquardt, L. A. Sherman.
2002
. Uncoupling of proliferative potential and gain of effector function by CD8+ T cells responding to self-antigens.
J. Exp. Med.
196
:
323
-333.
19
Guo, Z., D. Mital, J. Shen, A. S. Chong, Y. Tian, P. Foster, H. Sankary, L. McChesney, S. C. Jensik, J. W. Williams.
1998
. Immunosuppression preventing concordant xenogeneic islet graft rejection is not sufficient to prevent recurrence of autoimmune diabetes in nonobese diabetic mice.
Transplantation
65
:
1310
-1314.
20
Rosen, H., C. Alfonso, C. D. Surh, M. G. McHeyzer-Williams.
2003
. Rapid induction of medullary thymocyte phenotypic maturation and egress inhibition by nanomolar sphingosine 1-phosphate receptor agonist.
Proc. Natl. Acad. Sci. USA
100
:
10907
-10912.
21
Rosen, H., G. Sanna, C. Alfonso.
2003
. Egress: a receptor-regulated step in lymphocyte trafficking.
Immunol. Rev.
195
:
160
-177.
22
Xie, J. H., N. Nomura, S. L. Koprak, E. J. Quackenbush, M. J. Forrest, H. Rosen.
2003
. Sphingosine-1-phosphate receptor agonism impairs the efficiency of the local immune response by altering trafficking of naive and antigen-activated CD4+ T cells.
J. Immunol.
170
:
3662
-3670.
23
Mandala, S., R. Hajdu, J. Bergstrom, E. Quackenbush, J. Xie, J. Milligan, R. Thornton, G. J. Shei, D. Card, C. Keohane, et al
2002
. Alteration of lymphocyte trafficking by sphingosine-1-phosphate receptor agonists.
Science
296
:
346
-349.
24
Morgan, D. J., H. T. Kreuwel, 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
-727.
25
Kurts, C., J. F. Miller, R. M. Subramaniam, F. R. Carbone, W. R. Heath.
1998
. Major histocompatibility complex class I-restricted cross-presentation is biased towards high dose antigens and those released during cellular destruction.
J. Exp. Med.
188
:
409
-414.
26
Zhang, Y., B. O’Brien, J. Trudeau, R. Tan, P. Santamaria, J. P. Dutz.
2002
. In situ β cell death promotes priming of diabetogenic CD8 T lymphocytes.
J. Immunol.
168
:
1466
-1472.
27
Maki, T., R. Gottschalk, A. P. Monaco.
2002
. Prevention of autoimmune diabetes by FTY720 in nonobese diabetic mice.
Transplantation
74
:
1684
-1686.
28
Yang, Z., M. Chen, L. B. Fialkow, J. D. Ellett, R. Wu, V. Brinkmann, J. L. Nadler, K. R. Lynch.
2003
. The immune modulator FYT720 prevents autoimmune diabetes in nonobese diabetic mice small star, filled.
Clin. Immunol.
107
:
30
-35.
29
Pinschewer, D. D., A. F. Ochsenbein, B. Odermatt, V. Brinkmann, H. Hengartner, R. M. Zinkernagel.
2000
. FTY720 immunosuppression impairs effector T cell peripheral homing without affecting induction, expansion, and memory.
J. Immunol.
164
:
5761
-5770.
30
Hoglund, P., J. Mintern, C. Waltzinger, W. Heath, C. Benoist, D. Mathis.
1999
. Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes.
J. Exp. Med.
189
:
331
-339.
31
Delovitch, T. L., B. Singh.
1997
. The nonobese diabetic mouse as a model of autoimmune diabetes: immune dysregulation gets the NOD.
Immunity
7
:
727
-738.
32
Shevach, E. M..
2002
. CD4+ CD25+ suppressor T cells: more questions than answers.
Nat. Rev. Immunol.
2
:
389
-400.
33
Sakaguchi, S..
2000
. Regulatory T cells: key controllers of immunologic self-tolerance.
Cell
101
:
455
-458.
34
King, C., A. Ilic, K. Koelsch, N. Sarvetnick.
2004
. Homeostatic expansion of T cells during immune insufficiency generates autoimmunity.
Cell
117
:
265
-277.
35
Turley, S., L. Poirot, M. Hattori, C. Benoist, D. Mathis.
2003
. Physiological β cell death triggers priming of self-reactive T cells by dendritic cells in a type-1 diabetes model.
J. Exp. Med.
198
:
1527
-1537.
36
Wakeland, E. K., K. Liu, R. R. Graham, T. W. Behrens.
2001
. Delineating the genetic basis of systemic lupus erythematosus.
Immunity
15
:
397
-408.
37
Yamanouchi, J., J. Verdaguer, B. Han, A. Amrani, P. Serra, P. Santamaria.
2003
. Cross-priming of diabetogenic T cells dissociated from CTL-induced shedding of β cell autoantigens.
J. Immunol.
171
:
6900
-6909.
38
Belghith, M., J. A. Bluestone, S. Barriot, J. Megret, J. F. Bach, L. Chatenoud.
2003
. TGF-β-dependent mechanisms mediate restoration of self-tolerance induced by antibodies to CD3 in overt autoimmune diabetes.
Nat. Med.
9
:
1202
-1208.
39
Salomon, B., D. J. Lenschow, L. Rhee, N. Ashourian, B. Singh, A. Sharpe, J. A. Bluestone.
2000
. B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes.
Immunity
12
:
431
-440.