CD8+ T cells play an important role in the initiation of insulitis and in the destructive stage leading to insulin-dependent diabetes mellitus. A string of recent studies has led to the identification of numerous HLA-A2-restricted epitopes derived from pancreatic β cell Ags. It is hoped that assays detecting responses of patient PBMC to such epitopes might be instrumental for early diagnosis of β cell-directed autoimmunity and for monitoring trials of immunointervention. However, it remains unclear whether the results of assays studying PBMC reflect responses of islet-infiltrating lymphocytes, and to what extent they correlate with disease risk and/or activity. We have used female and male humanized NOD mice expressing HLA-A2 in addition to murine MHC class I molecules to study spontaneous responses of islet-infiltrating blood, spleen, and lymph node lymphocytes of various age groups to a panel of 16 epitopes. Twelve of these are restricted by HLA-A2, have previously been shown to be recognized by patient CTL, and have identical sequences in human and murine autoantigens. Using an IFN-γ ELISPOT assay, we find highly similar hierarchies of epitope immunodominance in the different T cell compartments, including peripheral blood and pancreatic islets. Moreover, we demonstrate that most of the epitopes eliciting dominant responses in humans display similar status in the mouse model. These results emphasize the potential of humanized mice as tools for studying spontaneous autoimmune CTL responses, and they provide a strong rationale for the development and use of assays monitoring responses of CD8+ PBMC in human type 1 diabetes.

Insulin-dependent diabetes mellitus (IDDM)3 is characterized by the progressive destruction of pancreatic β cells by autoreactive, Ag-specific T lymphocytes (1). Both CD4+ Th cells and CD8+ CTL play important roles in the autoimmune process (2). The causal implication of CTL in IDDM is suggested by numerous observations (3). CD8+ cells are dominant among islet-infiltrating lymphocytes and are required for efficient diabetes induction upon adoptive lymphocyte transfer in the model of spontaneously diabetic NOD mice (4). Moreover, results obtained with β2-microglobulin-deficient mice and cell populations demonstrate that CTL are involved both in initiation of β cell-directed autoimmunity and in ultimate destruction of β cells preceding manifest disease (5, 6, 7). The isolation of CD8+ T cell clones capable of transferring disease or precipitating it when expressed in transgenic NOD mice provides further evidence for the pathogenic role of CD8+ cells in IDDM (8, 9, 10).

Considering the cited evidence, there is strong interest in developing methods suitable for specific and sensitive detection of CTL in human peripheral blood. It is hoped that such tests can facilitate early detection of β cell-directed autoimmunity as well as help assess the risk of ultimately developing IDDM, either alone or by complementing the risk assessment afforded by islet Ag autoantibodies (11). Moreover, they might be useful for selection of candidate individuals for immunointervention trials and for monitoring such trials. Finally, monitoring of autoreactive CD8+ T cells might provide guidance for preventive intervention since these cells can themselves be targets for effective immunotherapy administered in the prediabetic stage (12).

After a prolonged period of slow progress, in part due to the lack of sufficient epitope mapping data (reviewed in Ref. 13), the last 2 years have seen substantial advances toward the goal of developing suitable CD8+ T cell tests for use in humans. As a result of the efforts of several laboratories, including ours, a large number of epitopes derived from important β cell Ags including preproinsulin (PPI), 65-kDa glutamic acid decarboxylase (GAD65), IA-2, islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP), glial fibrillary acidic protein, and islet amyloid polypeptide have been identified (14, 15, 16, 17, 18, 19, 20, 21). Most of these are presented by the frequent HLA class I allele HLA-A*0201, produced by natural Ag processing in cells expressing the source protein, and are recognized by one or several patients with recent-onset IDDM. We have undertaken the so far most extensive study in HLA-A*0201 patients, using a large panel of epitopes and an IFN-γ ELISPOT assay, and shown that IDDM patients can be distinguished from controls with a sensitivity and specificity both >80%, even when a limited panel of only five epitopes is used (14, 16). Thus, satisfactory methods may now be available to detect HLA-A2-restricted islet-specific CD8+ T cells in human peripheral blood, and such methods might soon be extended to additional HLA class I alleles, using efficient strategies such as epitope mapping by DNA immunization of HLA class I humanized mice (14, 22).

Considering that broad-scale monitoring of individuals at risk of IDDM for autoreactive CD8+ T cells may be at hand, it becomes essential to define what precise information can be derived from the detection of islet Ag-recognizing CTL in peripheral blood. One important notion frequently brought forward is that lymphocytes circulating in the blood might not be representative of islet-infiltrating cells, with respect to specificity, and kinetics or amplitude of responses (23). Another open question is to what extent IFN-γ secreting peripheral blood CTLs provide information about the risk of developing IDDM, and possibly the imminence of disease, as opposed to merely highlighting an ongoing autoimmune cellular response (24). Although these and other questions are difficult or at least very time-consuming to answer in humans, humanized spontaneously diabetic mice may provide an experimental system in which answers with high relevance for humans can be obtained rapidly (25, 26). Expression of HLA-A2 in NOD mice accelerates disease onset, suggesting that HLA-A2 mediates pathogenic CD8+ T cell responses in these mice (27). This conclusion is corroborated by the observation that NOD mice expressing HLA-A2 as the sole MHC class I molecule still develop spontaneous disease (26). Herein we took advantage of HLA-A2 humanized NOD mice (27) and of our panel of HLA-A2 restricted islet Ag epitopes (14, 16) to compare HLA-A2- and Kd-restricted spontaneous CTL responses of CD8+ lymphocytes residing in the β cell target organ to the responses of circulating lymphocytes.

Previously described NOD mice expressing a single-chain HLA-A2 molecule (HHD) composed of (N-terminal to C-terminal) β2-microglobulin, the α1 and α2 domains of HLA-A2, and the α3 domain of Ld were produced by direct transgene injection of NOD zygotes by Dr. D. Serreze (The Jackson Laboratory), as described (20), and bred in our specific pathogen-free mouse facility. In our animal facility, diabetes incidence for this strain, at 70% for females and 35% for males by week 35, is equivalent to standard NOD mice. However, with earliest disease onset from week 11 for females and week 14 for males, disease onset is accelerated by ∼2 wk in NOD-HHD relative to standard NOD mice. Nonautoimmune HHD mice, produced by injection of the same transgene into C57BL/6 × SJL oocytes and backcrossing to H-2Db and β2-microglobulin knockout C57BL/6 mice (28), were a generous gift from Dr. F. Lemonnier (Institut Pasteur, Paris). BALB/c mice and nonobese-resistant (NOR) mice were purchased from Charles River Laboratories, and NOD mice were raised in our animal facility. Before removal of organs, mice were euthanasized by i.p. injection of a dilute solution of 2,2,2-tribromoethanol in tert-amylalcohol (Sigma-Aldrich). The project was approved by the regional animal care and use review board.

Peptides were purchased from Schafer-N and were generally >80% pure, as confirmed by chromatographic and mass spectrometric analysis. Concentrated stocks (10 mM) were prepared in DMSO and stored as single-use aliquots of 2 μl at −80°C. A pool of four HLA-A2 restricted viral peptides was used as negative control: influenza matrix protein 158–66 (GILGFVFTL), CMV pp65495–503 (NLVPMVATV), EBV BMLF1280-88, and HIV gag77–85 (SLYNTVATL).

Blood was collected from anesthetized mice by cardiac puncture using a heparinized syringe and diluted in an equal volume of RPMI 1640 medium with 10% FCS (Invitrogen). Cells were pelleted by centrifugation at 500 × g for 10 min at room temperature, followed by lysis of RBC by incubation in 5 ml of 150 mM NH4Cl, 1 mM KHCO3, and 0.1 mM EDTA for 5 min. Finally, white blood cells were pelleted, resuspended in 2 ml of complete RPMI 1640 medium, and used for ELISPOT assays.

Spleens or lymph nodes were placed in 5 ml of HBSS in a sterile petri dish and dissociated by forcing through a 40-μm cell strainer (BD Biosciences), pelleted, and washed once in HBSS. Then RBC were lysed as described above, and the remaining cells were washed twice with HBSS and resuspended in complete RPMI 1640 medium. For control experiments (shown in Tables II and III), 2.5 × 104 lymphocytes from spleen, lymph nodes, or peripheral blood were incubated for 5 days with 50 syngeneic irradiated (3500 rad) islets and 50 U/ml recombinant human IL-2 (R&D Systems) and then tested in ELISPOT assays using irradiated splenocytes as APCs, as described below in detail.

Table II.

CD8+ T cells from nonautoimmune mice do not recognize β cell autoantigensa

EpitopeHHD Female (10 wk)NOR Female (10 wk)
Spleen+b IL-2LN+b IL-2Blood+b IL-2IsletsSpleen+b IL-2LN+b IL-2Blood+b IL-2Islets
P1 3.4 19 8.3 65 35 74 618 17 15 60 51 45 61 662 
P3 3.4 11 12 60 65 74 618 12 16 36 78 20 77 314 
P4 8.7 20 62 70 100 633 18 20 57 58 63 63 845 
P6 1.7 17 15 79 22 73 507 16 10 63 64 74 71 312 
G4 15 13 70 77 87 729 16 12 64 55 77 74 712 
I3 4.7 21 10 68 65 81 475 15 15 58 58 61 60 653 
IG1 13 23 43 92 65 665 20 24 55 64 87 71 803 
Anti-CD3 1,017 460 3,831 2,385 5,488 3,955 45,221 1,362 1,293 2,826 2,545 3,044 3,261 26,637 
Negative control 7.3 17 27 54 42 42 570 11 13 53 51 63 58 285 
Control + 3 SD 19 33 66 124 102 102 1,407 29 30 129 108 151 127 894 
Control + 4 SD 23 38 79 147 122 122 1,686 35 35 154 127 180 150 1,097 
Control + 5 SD 27 44 92 171 142 143 1,965 40 41 180 146 210 173 1,300 
EpitopeHHD Female (10 wk)NOR Female (10 wk)
Spleen+b IL-2LN+b IL-2Blood+b IL-2IsletsSpleen+b IL-2LN+b IL-2Blood+b IL-2Islets
P1 3.4 19 8.3 65 35 74 618 17 15 60 51 45 61 662 
P3 3.4 11 12 60 65 74 618 12 16 36 78 20 77 314 
P4 8.7 20 62 70 100 633 18 20 57 58 63 63 845 
P6 1.7 17 15 79 22 73 507 16 10 63 64 74 71 312 
G4 15 13 70 77 87 729 16 12 64 55 77 74 712 
I3 4.7 21 10 68 65 81 475 15 15 58 58 61 60 653 
IG1 13 23 43 92 65 665 20 24 55 64 87 71 803 
Anti-CD3 1,017 460 3,831 2,385 5,488 3,955 45,221 1,362 1,293 2,826 2,545 3,044 3,261 26,637 
Negative control 7.3 17 27 54 42 42 570 11 13 53 51 63 58 285 
Control + 3 SD 19 33 66 124 102 102 1,407 29 30 129 108 151 127 894 
Control + 4 SD 23 38 79 147 122 122 1,686 35 35 154 127 180 150 1,097 
Control + 5 SD 27 44 92 171 142 143 1,965 40 41 180 146 210 173 1,300 
a

The IFN-γ spot numbers for the negative control (pool of viral peptides) are shown as raw data; all other data are given as differentials (spot numbers for test peptide − spot numbers for negative control); differentials ≤ 0 are indicated as 0. LN, Lymph node. The shading indicates where data are 3 (light grey), 4 (intermediate grey), and 5 (dark grey) SD above negative controls.

b

Cells from the same source as in the left adjacent column were preincubated for 5 days with islets and IL-2 before ELISPOT analysis.

Table III.

CD8+ T cells from NOD-HHD mice recognize β cell autoantigensa

EpitopeHHD-NOD MaleHHD-NOD FemaleNOD Female
SpleenLNBloodIsletsSpleenIL-2LNIL-2BloodIL-2IsletsSpleenLNBlood
P1 30 112 106 1,319 36 48 139 169 329 623 3,057 14 32 
P2 20 79 72 899 25 54 71 154 228 473 1,457 59 85 
P4 103 1,049 675 1,445 243 136 1,351 1,190 2,080 1,947 4,778 143 1,352 1,044 
P5 38 112 99 1,382 30 46 124 175 277 319 1,707 ND ND ND 
P6 27 108 89 941 50 47 145 178 349 510 2,982 36 62 
G4 27 87 88 889 42 40 113 204 319 458 2,582 23 20 
I3 21 107 95 826 40 54 131 202 368 484 2,757 45 67 
IG1 124 1,072 741 1,372 184 156 1,341 1,355 2,553 2,211 4,919 153 1,501 1,032 
Anti-CD3 552 4,553 5,674 25,933 1,451 1,751 6,472 8,206 7,516 8,379 52,703 937 19,450 12,620 
Negative control 11 53 53 279 12 15 66 71 147 132 843 86 125 
Control + 3 SD 31 129 120 1,005 32 36 136 122 328 291 2,160 10 303 329 
Control + 4 SD 37 155 142 1,246 38 43 160 139 388 344 2,603 13 375 397 
Control + 5 SD 44 181 165 1,487 45 50 183 156 449 397 3,043 15 447 465 
EpitopeHHD-NOD MaleHHD-NOD FemaleNOD Female
SpleenLNBloodIsletsSpleenIL-2LNIL-2BloodIL-2IsletsSpleenLNBlood
P1 30 112 106 1,319 36 48 139 169 329 623 3,057 14 32 
P2 20 79 72 899 25 54 71 154 228 473 1,457 59 85 
P4 103 1,049 675 1,445 243 136 1,351 1,190 2,080 1,947 4,778 143 1,352 1,044 
P5 38 112 99 1,382 30 46 124 175 277 319 1,707 ND ND ND 
P6 27 108 89 941 50 47 145 178 349 510 2,982 36 62 
G4 27 87 88 889 42 40 113 204 319 458 2,582 23 20 
I3 21 107 95 826 40 54 131 202 368 484 2,757 45 67 
IG1 124 1,072 741 1,372 184 156 1,341 1,355 2,553 2,211 4,919 153 1,501 1,032 
Anti-CD3 552 4,553 5,674 25,933 1,451 1,751 6,472 8,206 7,516 8,379 52,703 937 19,450 12,620 
Negative control 11 53 53 279 12 15 66 71 147 132 843 86 125 
Control + 3 SD 31 129 120 1,005 32 36 136 122 328 291 2,160 10 303 329 
Control + 4 SD 37 155 142 1,246 38 43 160 139 388 344 2,603 13 375 397 
Control + 5 SD 44 181 165 1,487 45 50 183 156 449 397 3,043 15 447 465 
a

Data presentation is identical to Table II. All mice were 10 wk old. LN, Lymph node.

After placing a clamp at the junction of the common bile duct and the duodenum of an anesthetized mouse, the bile duct was cannulated with a 30-gauge needle, and 2 ml of a solution (0.33 mg/ml in HBSS) of purified collagenase (Liberase, Roche Applied Science) was injected. The pancreas was removed and digested for 10 min at 37°C. Digestion was stopped by the addition of 30 ml cold HBSS supplemented with 10% FCS, followed by centrifugation at 300 × g at 4°C for 5 min. The material was suspended in 30 ml HBSS with FCS, poured through a tissue sieve (Cellector 40-mesh screen, Bellco Glass), washed with cold HBSS buffer, and suspended in 10 ml Histopaque-1077 (Sigma-Aldrich). The cell suspension was overlaid with 10 ml HBSS and centrifuged for 10 min at 800 × g. The interface was collected and diluted in complete RPMI 1640 medium, and islets were hand-picked under a microscope. Islets were suspended in complete RPMI 1640 medium with 50 U/ml IL-2, seeded into 24-well plates at 50 islets and 1 ml/well, and placed for 5 days in a 37°C incubator. Cells were then collected and digested for 5 min at 37°C in 0.1 ml of 0.5 mg/ml trypsin with 0.22 mg/ml EDTA (PAA Laboratories). Digestion was stopped with complete RPMI 1640 medium, and cells were washed twice and finally suspended in 2 ml of complete RPMI 1640 medium for use in ELISPOT assays.

Ninety-six-well polyvinylidene difluoride plates (Millipore) were coated overnight with an anti-IFN-γ Ab (U-CyTech Biosciences) according to the instructions of the manufacturer. Plates were washed five times in PBS and blocked for 1 h at 37°C with 200 μl/well of RPMI 1640 medium with 10% FCS. Lymphocytes were then added in a total volume of 200 μl/well of RPMI 1640 medium supplemented with 10% FCS and 1 U/ml IL-2. Splenocytes were cultured at 2 × 105/well without additional APCs, while the same amount of irradiated (3500 rad) splenocytes was used as APCs in ELISPOT assays examining other cell populations. Cultured islet-infiltrating lymphocytes, and lymphocytes from spleen, blood, or lymph nodes preincubated with IL-2 and islets, were added to APCs at 104 cells/well, lymph node cells (nonpreincubated) at 3 × 104 cells/well, and peripheral blood lymphocytes (nonpreincubated) at 2 × 104 cells/well. mAb 145 2C11 recognizing CD3 was used as positive control at a concentration of 1 μg/ml. Peptides were added at 7 μM and incubated with cells at 37°C for 40 h. Cells were then removed, plates were washed, and a secondary biotinylated anti-murine IFN-γ mAb (U-CyTech Biosciences) was diluted in PBS with 10% FCS was added. Following incubation for 2 h at room temperaure, alkaline phosphatase-conjugated ExtrAvidin (Sigma-Aldrich) diluted in PBS with 0.5% FCS was added for a 1-h incubation at room temperature. After removal of the bottom, plates were washed and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate was added (Sigma-Aldrich). Ten minutes later the reaction was stopped by washing with water. The plates were air-dried for 1 h before spot counting using an AID reader (Autoimmun Diagnostika). All data are means of sixtuplicate wells and expressed as spot-forming cells (SFC) per 106 responder cells. A receiver-operator characteristics analysis was performed to define the cut-off for positive responses. The cut-off was set at the average basal reactivity + 3 SD, as this was the value giving the best sensitivity (i.e., number of positive responses in autoimmune HHD-NOD mice) and specificity (i.e., number of negative responses in nonautoimmune HHD and BALB/c mice). ELISPOT assays for blood, spleen, and regional lymph nodes were performed separately on lymphocytes isolated from four or five mice. For islet-infiltrating lymphocytes, four assays each with pooled lymphocytes from six mice were performed.

Data were evaluated using Prism 4 software (GraphPad Software). To compare immunodominance ranking in two groups, we used the Mann-Whitney U test, and for comparison of more than two groups, the Kruskal-Wallis test. All data were corrected for multiple comparisons.

We asked whether HLA-A2-restricted epitopes derived from islet Ags and recognized by HLA-A2+ patients were also recognized during spontaneous development of diabetes in HLA-A2-expressing NOD mice. Among the 21 HLA-A2-restricted epitopes derived from PPI, GAD65, IA-2, and IGRP recently examined by us (14, 16), 12 were completely conserved between the human and murine islet Ag sequences (Table I). An additional HLA-A2-restricted epitope, proinsulin A1–10, differing between human and murine PPI by a single conservative substitution in a non-HLA anchor position, was also tested. Given that NOD-HHD mice also express murine MHC class I molecules, we included the dominant H-2Kd-restricted epitopes IGRP206-14 and PPI-B15–23, as well as a heteroclitic analog of the former epitope, NRP-V7, in our epitope panel (10, 29). Tetramers formed with NRP-V7 stain a larger number of CTL than do tetramers with the native IGRP epitope, and they may be capable of predicting disease in female NOD mice (24). Consistent with a previous report (20), we found that HHD-NOD mice express lower levels of cell-surface HLA-A2 molecules than do human PBL; expression was also lower than in control HHD mice (28) produced on a C57BL/6 background (Fig. 1).

Table I.

HLA-A2- and H-2Kd-restricted autoantigenic epitopes tested

CodeAutoantigen/PositionSequenceRestriction% Patients RespondingaReference
P1 PPI-B5–14 HLCGSHLVEAb A2 13c Unpublished 
P2 PPI-B10–18 HLVEALYLVd A2 19 16  
P3 PPI-B18–27 VCGERGFFYTd A2 40 16  
P4 PPI-B15–23 LYLVCGERG Kd NA 10  
P5 PPI-A1–10 GIVEQCCTSIe A2 16  
P6 PPI-A12–20 SLYQLENYCd A2 33.3 16  
G1 GAD110–118 FLQDVMNIL A2 14  
 (GAD67)f ––LEVD––    
G2 GAD114–123 VMNILLQYVV A2 50 35  
 (GAD67VD–––N––R    
G3 GAD141–149 LLQEYNWEL A2 14  
 (GAD67––EGMEGFN    
G4 GAD536–545 RMMEYGTTMV A2 25 14  
 (GAD67L–––S–––––    
I1 IA-2790–798 TIADFWQMV A2 14  
I2 IA-2805–813 VIVMLTPLV A2 42 14  
I3 IA-2830–839 SLYHVYEVNL A2 14  
I4 IA-2962–970 ALTAVAEEV A2 14  
IG1 IGRP206–214 VYLKTNVFL Kd NA 36 
IG2 (Analog) KYNKANVFL Kd NA 36 
CodeAutoantigen/PositionSequenceRestriction% Patients RespondingaReference
P1 PPI-B5–14 HLCGSHLVEAb A2 13c Unpublished 
P2 PPI-B10–18 HLVEALYLVd A2 19 16  
P3 PPI-B18–27 VCGERGFFYTd A2 40 16  
P4 PPI-B15–23 LYLVCGERG Kd NA 10  
P5 PPI-A1–10 GIVEQCCTSIe A2 16  
P6 PPI-A12–20 SLYQLENYCd A2 33.3 16  
G1 GAD110–118 FLQDVMNIL A2 14  
 (GAD67)f ––LEVD––    
G2 GAD114–123 VMNILLQYVV A2 50 35  
 (GAD67VD–––N––R    
G3 GAD141–149 LLQEYNWEL A2 14  
 (GAD67––EGMEGFN    
G4 GAD536–545 RMMEYGTTMV A2 25 14  
 (GAD67L–––S–––––    
I1 IA-2790–798 TIADFWQMV A2 14  
I2 IA-2805–813 VIVMLTPLV A2 42 14  
I3 IA-2830–839 SLYHVYEVNL A2 14  
I4 IA-2962–970 ALTAVAEEV A2 14  
IG1 IGRP206–214 VYLKTNVFL Kd NA 36 
IG2 (Analog) KYNKANVFL Kd NA 36 
a

Percentages are taken from the publications cited in the right-hand column. NA, Not applicable.

b

Epitope identical between human PI and murine PI-2 only (substitution in murine PI-1).

c

Epitope recognized by 2 of 13 HLA-A2+ patients with recent-onset type 1 diabetes in our studies.

d

Epitope with identical sequence in human PI and murine PI-1 and PI-2.

e

Epitope with conservative substitution (D for E) in position 4 of murine PI-1 and PI-2.

f

Homologous sequences in murine GAD67; dashes indicate identity.

FIGURE 1.

Expression of HLA-A2 by HHD-NOD mice. Splenocytes from NOD-HHD or control HHD and NOD mice, and human A2+ peripheral blood lymphocytes were stained with fluorescent mAb BB7.2 specific for HLA-A2 and analyzed for fluorescence intensity on a FACSCalibur flow cytometer.

FIGURE 1.

Expression of HLA-A2 by HHD-NOD mice. Splenocytes from NOD-HHD or control HHD and NOD mice, and human A2+ peripheral blood lymphocytes were stained with fluorescent mAb BB7.2 specific for HLA-A2 and analyzed for fluorescence intensity on a FACSCalibur flow cytometer.

Close modal

To study spontaneous CD8+ T cell responses to the selected epitopes, we used our IFN-γ ELISPOT assay recently developed to examine responses of PBMC from patients with recent-onset IDDM (16). However, because only very limited numbers of lymphocytes can be obtained from murine blood, lymph nodes, and islets, we modified the assay by adding an excess of splenic APC for Ag presentation to these cell populations. Moreover, to obtain a sufficient number of islet-infiltrating lymphocytes, we adopted a protocol in which islets are cultured for 5 days in the presence of IL-2 before lymphocytes are recovered for examination (20, 30). Given that syngeneic splenocytes might themselves respond to islet Ag peptides, we used irradiation to suppress IFN-γ secretion by APC in response to stimulation by autoantigenic peptides. Fig. 2 shows that irradiation reduced IFN-γ secretion of splenocytes in response to both HLA-A2- and Kd-restricted β cell epitopes to background levels, while the response to polyclonal stimulation by an Ab to CD3 was partially preserved. Irradiation also abolished IFN-γ secretion of islet-infiltrating lymphocytes in response to autoantigens (not shown). We therefore established a standard protocol in which peptides were presented to lymph node, blood, and islet lymphocytes by splenic APC from the same animal.

FIGURE 2.

Irradiation abolishes IFN-γ secretion in response to islet Ag epitopes. Spleen cells, irradiated or not with 3500 rad, were seeded at 2 × 105 cells/well in 96-well plates and incubated for 40 h with 7 μM peptides, or an Ab against CD3 used 1 μg/ml, as indicated. Solvent DMSO (DM) was used as negative control. The number of IFN-γ secreting cells was then measured by ELISPOT, as described in Materials and Methods.

FIGURE 2.

Irradiation abolishes IFN-γ secretion in response to islet Ag epitopes. Spleen cells, irradiated or not with 3500 rad, were seeded at 2 × 105 cells/well in 96-well plates and incubated for 40 h with 7 μM peptides, or an Ab against CD3 used 1 μg/ml, as indicated. Solvent DMSO (DM) was used as negative control. The number of IFN-γ secreting cells was then measured by ELISPOT, as described in Materials and Methods.

Close modal

To establish the specificity of the assay, we first tested lymphocytes from nonautoimmune mice for their response to islet Ags (Table II). All samples were tested in sixtuplicate, and responses exceeding the background spot number by at least 3 SD were considered positive. Spleen, lymph node, and islet-infiltrating blood lymphocytes from male (not shown) and female HHD mice, from female BALB/c mice expressing H-2Kd (not shown), and from NOR mice sharing MHC class I and class II alleles with NOD mice did not secrete IFN-γ in response to four viral epitopes with high HLA-A2 binding affinity, whereas stimulation by an Ab against CD3 induced vigorous IFN-γ secretion by all cell populations. Islet Ag-derived epitopes, both HLA-A2 and Kd restricted, did not elicit IFN-γ responses above background levels in HHD, BALB/c, and NOR mice, suggesting that mice not prone to autoimmunity do not harbor detectable β cell-specific CTL in blood, spleen, islets, and pancreatic lymph nodes (Table II). Preincubation of lymphocytes from spleen, lymph nodes, or blood with IL-2 and irradiated islets also did not result in detectable IFN-γ responses against islet autoantigens, demonstrating that nonautoimmune mice did not harbor CD8+ precursors that could readily be activated and/or expanded upon incubation with Ag and growth factors.

Next we used the described assay for a systematic analysis of islet Ag-specific CTL responses in HHD-NOD mice. In addition to comparing responses of islet-infiltrating cells to those of spleen, lymph node, and blood lymphocytes, we examined responses over time. Our analysis included 5-wk-old mice presumably close to the onset of autoimmunity, 10- and 15-wk-old mice with full-blown autoimmunity in which “epitope spreading” presumably has taken place (13, 31), and diabetic mice. We also compared responses between female and male mice, with the following reasoning. If IFN-γ secretion is indicative of the risk of diabetes, young (i.e., potentially prediabetic) female and male mice should produce different ELISPOT results, because female mice are twice as likely to develop diabetes as are male mice. Conversely, an assay associated with disease should produce similar results in both genders when diabetic mice are analyzed.

For comparison with data obtained with nonautoimmune mice (Table II), Table III presents selected assay results for 10-wk-old HHD-NOD mice. Although male HHD-NOD displayed levels of IFN-γ secretion in the absence of specific Ag and upon anti-CD3 stimulation that resembled control HHD mice, both background and anti-CD3-stimulated IFN-γ secretion was somewhat increased in female HHD-NOD mice. In striking contrast to nonautoimmune mice, HHD-NOD mice displayed significant responses to the entire epitope panel tested. Although both female and male mice showed broad responses, responses had a higher magnitude in female mice. Control NOD mice not expressing HLA-A2 displayed strong responses against the Kd-restricted epitopes but not against five A2-restricted epitopes, confirming that the latter were presented exclusively by HLA-A2 (Table III).

Fig. 3 shows a comprehensive view of the entire data set obtained with HHD-NOD mice and allows for several general conclusions. First, islet-infiltrating lymphocytes show higher background IFN-γ secretion than do lymphocytes from other sources. Second, for all autoantigenic peptides, the number of spots varies between the different organs, ranging from 100 to 5000 for the most immunodominant peptides. The highest proportion of cells recognizing islet Ags is found among islet-infiltrating lymphocytes. This was not due to preincubation of islet-infiltrating lymphocytes with islet Ag and IL-2, because preincubation of HHD-NOD lymphocytes from spleen, lymph node, or blood under the same conditions did not increase ELISPOT responses significantly (Table III). Autoantigen-specific peripheral blood lymphocytes rank second in frequency, followed by lymph node cells, while spleens contain very few such cells.

FIGURE 3.

Global view of IFN-γ secretion in response to 16 autoantigenic peptides. The figure shows the means of four or five mice per time point, cell compartment, and gender. SD are not shown to avoid excessive visual clutter (see Fig. 4 for a subset of data with SD). All data are background-subtracted, that is, data represent the number of SFC for the test peptide minus the number of SFC obtained with a pool of HLA-A2-restricted viral peptides. Scales are adjusted for each organ but are identical for both genders. The colored flags above the bars indicate where data are 3 (yellow flag), 4 (orange flag), or 5 (red flag) SD above negative controls.

FIGURE 3.

Global view of IFN-γ secretion in response to 16 autoantigenic peptides. The figure shows the means of four or five mice per time point, cell compartment, and gender. SD are not shown to avoid excessive visual clutter (see Fig. 4 for a subset of data with SD). All data are background-subtracted, that is, data represent the number of SFC for the test peptide minus the number of SFC obtained with a pool of HLA-A2-restricted viral peptides. Scales are adjusted for each organ but are identical for both genders. The colored flags above the bars indicate where data are 3 (yellow flag), 4 (orange flag), or 5 (red flag) SD above negative controls.

Close modal

Among the 15 epitopes studied (IG2 = NRP-V7, a variant of IG1, is not counted here), the two restricted by H-2Kd elicited the strongest IFN-γ responses. This was true for all time points and for all organs, although the relative dominance of the Kd-restricted responses varied between organs (see below). However, differing from observations in standard NOD mice (24, 30, 32), we found equivalent responses to PPI-B15–23 and IGRP206-14, again at all time points and in all organs. All tested A2-restricted peptides were also recognized by IFN-γ-secreting cells, albeit to a somewhat different extent. There was a clear hierarchy, with peptides P1, P3, P6, G4, and I3 inducing consistently higher spot numbers than the other A2-restricted epitopes (see below).

An analysis of IFN-γ responses in different organs reveals that the immunodominance (i.e., the ratio between specific spot number for compared epitopes) of the two Kd-restricted epitopes is much less pronounced among islet-infiltrating lymphocytes than among the other lymphocyte populations. However, the epitope hierarchy is identical for all four studied compartments, as confirmed by statistical analysis of ELISPOT data. When only HLA-A2-restricted epitopes are considered, epitope ranking in every single lymphocyte compartment correlates highly (corrected p < 0.0038 for all comparisons) with ranking in every other compartment of female or male mice. The same applies to the complete epitope set, with the exception of the correlation between male and female splenocytes (corrected p = 0.072). This highly significant correlation applies also when data for individual time points are analyzed. For example, for female mice, islet and blood hierarchy ranking correlate with p < 0.0038 at 5 wk, 10 wk, and diabetes onset, and with p = 0.049 at 15 wk; for male mice, hierarchies correlate with p < 0.0038 at all time points. When the same analysis is performed on A2-restricted subdominant epitopes only, the same high level of correlation is observed (corrected p < 0.0038 for all comparisons, male and female mice).

Fig. 4 shows data for the two immunodominant Kd-restricted and for the two A2-restricted epitopes eliciting the strongest responses in such a manner that the evolution of responses over time is visualized. This representation highlights another difference between islet-infiltrating lymphocytes and lymphocytes from other sites. Although IFN-γ responses are low at 5 wk of age among lymph node, spleen, and male blood cells, islet-infiltrating cells already show high-level responses at this time. Interestingly, the same is true for blood lymphocytes from female but not from male mice, suggesting that early appearance of islet Ag-specific T cells in peripheral blood may correlate with increased disease risk. Most responses reach peak or plateau levels at 10 wk and remain at a similar level through overt manifestation of diabetes.

FIGURE 4.

Evolution of representative IFN-γ responses over time. The figure shows a subset of the data in Fig. 2 plotted over time. Data are means ± SD.

FIGURE 4.

Evolution of representative IFN-γ responses over time. The figure shows a subset of the data in Fig. 2 plotted over time. Data are means ± SD.

Close modal

Male and female mice also display a number of suggestive other differences. Both Ag-specific and, to a lesser extent, background IFN-γ secretion are of greater magnitude in female mice. As a result, specific responses to subdominant epitopes emerge more clearly in female than in male mice. For example, applying the criterion of responses exceeding background + 5 SD at three or more tested time points and in two or more organs for identification of subdominant A2-restricted epitopes, female mice show responses to five epitopes (P1, P3, P6, I3, G4), while male mice respond to a single epitope (P5). Additionally, this list suggests a subtle but consistent difference with respect to recognition of subdominant epitopes by the two genders, because epitope P5 is recognized less efficiently by female mice. However, both of the observed differences between genders are also observed for diabetic mice, so they cannot be interpreted as evidence for distinct disease risk.

It is also of interest to compare recognition of epitopes between IDDM patients and humanized HHD-NOD mice. Using a cut-off of recognition by ≥25% of patients, epitopes P3, P6, G2, G4, and I2 are immunodominant in patients (Table I). This compares to subdominant status of epitopes P1, P3, P6, I3, and G4 in the humanized mice. Thus, not only the complete epitope panel recognized by patients is autoantigenic in the mice, but also three of five epitopes dominant in patients have similar status in the mice. Nevertheless, there may be some shifts in immunodominance in the humanized mice, as exemplified by epitope I2, which is recognized by 42% of patients but is not subdominant in mice.

Herein we have undertaken a comprehensive analysis of spontaneous CD8+ T cell responses to islet cell Ags in a humanized mouse. The broad range and the specificity of the observed responses emphasize the potential of such mice as models for studying autoimmune responses restricted by HLA class I molecules. HLA-A2 transgenic NOD mice devoid of murine MHC class I molecules have been used to map autoantigenic CTL epitopes derived from β cell Ags in two studies, one published before and one after submission of this paper (20, 33). These studies resulted in identification of three epitopes derived from murine IGRP (20) and three others from PPI (33); however, only one of these epitopes had a sequence conserved in humans. Different from these studies, we focus on a large number of HLA-A2-restricted epitopes previously tested for recognition by patient CTL and with sequences conserved in mice, and we report observations that allow a number of initial general conclusions to be drawn concerning the use of humanized mice in IDDM.

One parameter indicative of the potential of humanized mice for studying human IDDM is the specificity of autoimmune CTL response. Considering the lower expression levels of the single-chain HHD molecule relative to native MHC class I molecules (20, 28), the immunodominance of the Kd-restricted responses against IGRP206-14 and PPI-B15–23 was not surprising. Due to the presence of the Kd molecule, the HHD-NOD mouse used by us might be interpreted as a model for a setting in which HLA-A2 presents subdominant epitopes in an individual carrying other alleles that restrict dominant CTL responses. Although it is difficult to compare the frequencies of CTL recognizing Kd-restricted epitopes with published data frequently obtained using direct tetramer analysis of islet-infiltrating cells (10, 24, 32), our data are consistent with a recent report by Wong and associates (30). These authors analyzed IL-2-expanded islet-infiltrating lymphocytes by IFN-γ ELISPOT, and they reported a frequency of 0.4% for cells recognizing NRP-V7, in agreement with our finding for 10-wk-old female mice (Fig. 3).

Although these considerations suggest that murine MHC class I molecules functioned normally in the presence of HLA-A2, it was conversely conceivable that the former reduced CTL stimulation by the latter through competition for APCs or cytokines, or by directly competing for binding of HLA-A2 ligands. However, several observations suggest that this was not the case. First, the complete absence of responses to HLA-A2-restricted epitopes in standard NOD mice shows that the latter cannot be presented by H-2Kd or H-2Db molecules. Moreover, the frequencies observed in this study of CTL recognizing subdominant HLA-A2-restricted epitopes (∼1 in 300) is very similar to that reported by Jarchum et al. (33) for dominant CTL epitopes in HHD-NOD mice lacking murine MHC class I molecules (1 in 160 to 1 in 500, assuming that 50% of islet-infiltrating cells are CD8+). Thus, the relatively weak IFN-γ responses against HLA-A2-restricted epitopes are not due to the presence of murine MHC class I molecules. Note that most IFN-γ responses against the same epitopes in IDDM patients are even weaker, ranging from 10 to 140 spots (vs ∼300 spots in HHD-NOD mice) per 1 × 106 PBL (16).

An important observation in this study is the identical epitope hierarchy in all studied organs. This includes islets despite the fact that the relative immunodominance of the Kd-restricted epitopes is attenuated in this organ. The latter finding may be due to less efficient export from islets of CTL-recognizing subdominant epitopes, or to poorer survival of such CTL in a nonislet environment; alternatively, CTL-recognizing subdominant epitopes may proliferate more vigorously in vitro in the presence of IL-2 than do CTL-recognizing dominant epitopes. Conservation of epitope hierarchy between islets and peripheral blood provides a strong rationale to studies that examine human PBL as markers of the autoimmune responses taking place in pancreatic islets, and such conservation suggests that the specificity of the latter is accurately reflected in circulating lymphocytes. Such studies are further encouraged by our finding that autoreactive CTL are more abundant among PBL than among splenocytes or lymph node T cells.

The relatively conserved hierarchy of immunodominance between humans and HHD-NOD mice is another important observation. Three of the five epitopes dominant in humans are subdominant in the mice, and three of the epitopes subdominant in the mice are dominant in humans. Moreover, weak responses to epitope G2, recognized by 50% of A2+ patients, is probably related to different expression patterns of the two GAD isoforms in humans and mice. It is well known that murine β cells express almost exclusively GAD67 while human β cells express GAD65 (34). Among the four epitopes derived from GAD65 tested in this study, three, including epitope G2, differ in four or more positions from murine GAD67, while epitope G4, the only subdominant GAD epitope in the mouse, shows only two substitutions that, in addition, do not affect HLA anchor positions (Table I). Therefore, we hypothesize that GAD-specific CTL are primed in the mouse mainly by autoimmunization against GAD67, and that our ELISPOT test detected mainly CTL primed against GAD67 and cross-reacting with a GAD65 peptide relatively conserved between the two isotypes. This leaves only I2 as an epitope derived from an Ag expressed in β cells and dominant in humans but replaced as a subdominant epitope in the mouse by I3, presumably due to well-documented differences in Ag processing and/or the T cell repertoire (35). The largely conserved epitope hierarchy between humans and HHD-NOD mice has several implications. First, NOD mice carrying HLA class I molecules, and expressing murine class I molecules or not, should allow for identification of CTL epitopes recognized during spontaneous IDDM development in humans. Second, such mice should be models in which strategies for immunomodulation or prevention of IDDM by targeting of specific HLA class I-restricted epitopes can directly be tested.

Finally, the data presented in this study provide some insight concerning the relationship between detection of autoreactive CTL and the risk of developing IDDM, or the presence of the disease. First, the complete absence of CTL responses, both Kd- and A2-restricted, in control mice demonstrates that, in the HHD-NOD model, detection of IFN-γ-producing CTL clearly indicates an abnormal autoimmune response. Moreover, the absence of responses in NOR mice demonstrates that the observed IFN-γ responses are not due to the mere presence of I-Ag7. Second, our data are compatible with the proposition of Trudeau and associates that a number of autoreactive CD8+ T cells above a certain threshold, rather than the mere presence of such cells, announces IDDM (24). Being twice as likely than male mice to develop disease, female mice consistently harbored higher numbers of islet Ag-specific cells than did male mice. However, interpretation of this phenomenon is complicated by the fact that this difference applies also to diabetic mice. Third, the subdominant status of epitope P5 in male but not in female mice may indicate that subtly altered epitope hierarchies may distinguish individuals developing disease from those remaining disease-free. Fourth, early (here at 5 wk) detection of autoreactive CTL in peripheral blood may indicate increased disease risk. However, in humans, defining any of the parameters indicating high risk—threshold values, epitope profiles, early time windows—will be a challenge.

We are grateful to D. Serreze (The Jackson Laboratory, Bar Harbor, ME) for the gift of the HHD-NOD strain, and to F. Lemonnier (Pasteur Institute, Paris) for providing transgenic HHD mice for control experiments.

The authors have no financial conflicts 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 Grant 05-PCOD-036 of the Agence Nationale de Recherche and by a grant from the European Foundation for the Study of Diabetes.

3

Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; PPI, preproinsulin; GAD, glutamic acid decarboxylase; IGRP, islet-specific glucose-6-phosphatase catalytic subunit-related protein; SFC, spot-forming cells.

1
Bach, J. F..
1995
. Insulin-dependent diabetes mellitus as a β-cell targeted disease of immunoregulation.
J. Autoimmun.
8
:
439
-463.
2
Roep, B. O..
2003
. The role of T-cells in the pathogenesis of type 1 diabetes: from cause to cure.
Diabetologia
46
:
305
-321.
3
Liblau, R. S., F. S. Wong, L. T. Mars, P. Santamaria.
2002
. Autoreactive CD8 T cells in organ-specific autoimmunity: emerging targets for therapeutic intervention.
Immunity
17
:
1
-6.
4
Bendelac, A., C. Carnaud, C. Boitard, 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
-832.
5
de Jersey, J., S. L. Snelgrove, S. E. Palmer, S. A. Teteris, A. Mullbacher, J. F. Miller, R. M. Slattery.
2007
. β cells cannot directly prime diabetogenic CD8 T cells in nonobese diabetic mice.
Proc. Natl. Acad. Sci. USA
104
:
1295
-1300.
6
Serreze, D. V., E. H. Leiter, G. J. Christianson, D. Greiner, D. C. Roopenian.
1994
. Major histocompatibility complex class I-deficient NOD-B2mnull mice are diabetes and insulitis resistant.
Diabetes
43
:
505
-509.
7
Wicker, L. S., E. H. Leiter, J. A. Todd, R. J. Renjilian, E. Peterson, P. A. Fischer, P. L. Podolin, M. Zijlstra, R. Jaenisch, L. B. Peterson.
1994
. β2-microglobulin-deficient NOD mice do not develop insulitis or diabetes.
Diabetes
43
:
500
-504.
8
Anderson, B., B. J. Park, J. Verdaguer, A. Amrani, P. Santamaria.
1999
. Prevalent CD8+ T cell response against one peptide/MHC complex in autoimmune diabetes.
Proc. Natl. Acad. Sci. USA
96
:
9311
-9316.
9
Graser, R. T., T. P. DiLorenzo, F. Wang, G. J. Christianson, H. D. Chapman, D. C. Roopenian, S. G. Nathenson, D. V. Serreze.
2000
. Identification of a CD8 T cell that can independently mediate autoimmune diabetes development in the complete absence of CD4 T cell helper functions.
J. Immunol.
164
:
3913
-3918.
10
Wong, F. S., J. Karttunen, C. Dumont, L. Wen, I. Visintin, I. M. Pilip, N. Shastri, E. G. Pamer, C. A. Janeway, Jr.
1999
. Identification of an MHC class I-restricted autoantigen in type 1 diabetes by screening an organ-specific cDNA library.
Nat. Med.
5
:
1026
-1031.
11
van Endert, P., Y. Hassainya, V. Lindo, J. M. Bach, P. Blancou, F. Lemonnier, R. Mallone.
2006
. HLA class I epitope discovery in type 1 diabetes.
Ann. NY Acad. Sci.
1079
:
190
-197.
12
Han, B., P. Serra, A. Amrani, J. Yamanouchi, A. F. Maree, L. Edelstein-Keshet, P. Santamaria.
2005
. Prevention of diabetes by manipulation of anti-IGRP autoimmunity: high efficiency of a low-affinity peptide.
Nat. Med.
11
:
645
-652.
13
Di Lorenzo, T. P., M. Peakman, B. O. Roep.
2007
. Translational mini-review series on type 1 diabetes: systematic analysis of T cell epitopes in autoimmune diabetes.
Clin. Exp. Immunol.
148
:
1
-16.
14
Blancou, P., R. Mallone, E. Martinuzzi, S. Severe, S. Pogu, G. Novelli, G. Bruno, B. Charbonnel, M. Dolz, L. Chaillous, et al
2007
. Immunization of HLA class I transgenic mice identifies autoantigenic epitopes eliciting dominant responses in type 1 diabetes patients.
J. Immunol.
178
:
7458
-7466.
15
Hassainya, Y., F. Garcia-Pons, R. Kratzer, V. Lindo, F. Greer, F. A. Lemonnier, G. Niedermann, P. M. van Endert.
2005
. Identification of naturally processed HLA-A2-restricted proinsulin epitopes by reverse immunology.
Diabetes
54
:
2053
-2059.
16
Mallone, R., E. Martinuzzi, P. Blancou, G. Novelli, G. Afonso, M. Dolz, G. Bruno, L. Chaillous, L. Chatenoud, J. M. Bach, P. van Endert.
2007
. CD8+ T-cell responses identify β-cell autoimmunity in human type 1 diabetes.
Diabetes
56
:
613
-621.
17
Ouyang, Q., N. E. Standifer, H. Qin, P. Gottlieb, C. B. Verchere, G. T. Nepom, R. Tan, C. Panagiotopoulos.
2006
. Recognition of HLA class I-restricted β-cell epitopes in type 1 diabetes.
Diabetes
55
:
3068
-3074.
18
Pinkse, G. G., O. H. Tysma, C. A. Bergen, M. G. Kester, F. Ossendorp, P. A. van Veelen, B. Keymeulen, D. Pipeleers, J. W. Drijfhout, B. O. Roep.
2005
. Autoreactive CD8 T cells associated with β cell destruction in type 1 diabetes.
Proc. Natl. Acad. Sci. USA
102
:
18425
-18430.
19
Standifer, N. E., Q. Ouyang, C. Panagiotopoulos, C. B. Verchere, R. Tan, C. J. Greenbaum, C. Pihoker, G. T. Nepom.
2006
. Identification of novel HLA-A*0201-restricted epitopes in recent-onset type 1 diabetic subjects and antibody-positive relatives.
Diabetes
55
:
3061
-3067.
20
Takaki, T., M. P. Marron, C. E. Mathews, S. T. Guttmann, R. Bottino, M. Trucco, T. P. DiLorenzo, D. V. Serreze.
2006
. HLA-A*0201-restricted T cells from humanized NOD mice recognize autoantigens of potential clinical relevance to type 1 diabetes.
J. Immunol.
176
:
3257
-3265.
21
Toma, A., S. Haddouk, J. P. Briand, L. Camoin, H. Gahery, F. Connan, D. Dubois-Laforgue, S. Caillat-Zucman, J. G. Guillet, J. C. Carel, et al
2005
. Recognition of a subregion of human proinsulin by class I-restricted T cells in type 1 diabetic patients.
Proc. Natl. Acad. Sci. USA
102
:
10581
-10586.
22
Gauvrit, A., M. Debailleul, A. T. Vu, P. Sai, J. M. Bach.
2004
. DNA vaccination encoding glutamic acid decarboxylase can enhance insulitis and diabetes in correlation with a specific Th2/3 CD4 T cell response in non-obese diabetic mice.
Clin. Exp. Immunol.
137
:
253
-262.
23
Walter, U., P. Santamaria.
2005
. CD8+ T cells in autoimmunity.
Curr. Opin. Immunol.
17
:
624
-631.
24
Trudeau, J. D., C. Kelly-Smith, C. B. Verchere, J. F. Elliott, J. P. Dutz, D. T. Finegood, P. Santamaria, R. Tan.
2003
. Prediction of spontaneous autoimmune diabetes in NOD mice by quantification of autoreactive T cells in peripheral blood.
J. Clin. Invest.
111
:
217
-223.
25
Serreze, D. V., Y. G. Chen.
2005
. Of mice and men: use of animal models to identify possible interventions for the prevention of autoimmune type 1 diabetes in humans.
Trends Immunol.
26
:
603
-607.
26
Serreze, D. V., M. P. Marron, T. P. Dilorenzo.
2007
. “Humanized” HLA transgenic NOD mice to identify pancreatic β cell autoantigens of potential clinical relevance to type 1 diabetes.
Ann. NY Acad. Sci.
1103
:
103
-111.
27
Marron, M. P., R. T. Graser, H. D. Chapman, D. V. Serreze.
2002
. Functional evidence for the mediation of diabetogenic T cell responses by HLA-A2.1 MHC class I molecules through transgenic expression in NOD mice.
Proc. Natl. Acad. Sci. USA
99
:
13753
-13758.
28
Pascolo, S., N. Bervas, J. M. Ure, A. G. Smith, F. A. Lemonnier, B. Perarnau.
1997
. HLA-A2.1-restricted education and cytolytic activity of CD8+ T lymphocytes from β2 microglobulin (β2m) HLA-A2.1 monochain transgenic H-2Db β2m double knockout mice.
J. Exp. Med.
185
:
2043
-2051.
29
Amrani, A., J. Verdaguer, P. Serra, S. Tafuro, R. Tan, P. Santamaria.
2000
. Progression of autoimmune diabetes driven by avidity maturation of a T-cell population.
Nature
406
:
739
-742.
30
Wong, C. P., L. Li, J. A. Frelinger, R. Tisch.
2006
. Early autoimmune destruction of islet grafts is associated with a restricted repertoire of IGRP-specific CD8+ T cells in diabetic nonobese diabetic mice.
J. Immunol.
176
:
1637
-1644.
31
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.
32
Lieberman, S. M., T. Takaki, B. Han, P. Santamaria, D. V. Serreze, T. P. DiLorenzo.
2004
. Individual nonobese diabetic mice exhibit unique patterns of CD8+ T cell reactivity to three islet antigens, including the newly identified widely expressed dystrophia myotonica kinase.
J. Immunol.
173
:
6727
-6734.
33
Jarchum, I., J. C. Baker, T. Yamada, T. Takaki, M. P. Marron, D. V. Serreze, T. P. DiLorenzo.
2007
. In vivo cytotoxicity of insulin-specific CD8+ T-cells in HLA-A*0201 transgenic NOD mice.
Diabetes
56
:
2551
-2560.
34
Pleau, J. M., M. Throsby, A. Esling, M. Dardenne.
1997
. Ontogeny of glutamic acid decarboxylase gene expression in the mouse pancreas.
Biochem. Biophys. Res. Commun.
233
:
227
-230.
35
Yewdell, J. W., S. M. Haeryfar.
2005
. Understanding presentation of viral antigens to CD8+ T cells in vivo: the key to rational vaccine design.
Annu. Rev. Immunol.
23
:
651
-682.