Conserved T cell receptor α-chain induces insulin autoantibodies

Abstract

A fundamental question is what are the molecular determinants that lead to spontaneous preferential targeting of specific autoantigens in autoimmune diseases, such as the insulin B:9–23 peptide sequence in type 1 diabetes. Anti-insulin B:9–23 T cell clones isolated from prediabetic NOD islets have a conserved Vα-segment/Jα-segment, but no conservation of the α-chain N region and no conservation of the Vβ-chain. Here, we show that the conserved T cell receptor α-chain generates insulin autoantibodies when transgenically or retrogenically introduced into mice without its corresponding Vβ. We suggest that a major part of the mystery as to why islet autoimmunity develops relates to recognition of a primary insulin peptide by a conserved α chain T cell receptor.

Multiple autoantigens are the targets of T cell-mediated islet autoimmunity of the nonobese diabetic (NOD) mouse, including uncharacterized autoantigens (1). In terms of gene knockouts, only gene knockouts that alter the expression of insulin (e.g., insulin 1 vs. insulin 2 genes) or change the sequence of the insulin molecule dramatically influence progression to diabetes (2–8) of NOD mice. In addition Kay and coworkers recently reported that the immune response to islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP) depends on an immune response to proinsulin (9).

Wegmann and coworkers reported that the majority of CD4 T cells infiltrating NOD islets react to insulin; >90% of these recognize insulin B chain 9–23 peptide amino acids (insulin B:9–23) and these T cells were able to accelerate the development of diabetes on transfer to nondiabetic mice (10). The T cell receptor (TCR) α- and β-chains derived from one of these clones (BDC 12-4.1) as a transgenes induces diabetes (11). Wen and coworkers have isolated a B:9–23-reactive T cell clone (2H6) whose TCR protects from diabetes both as a cell line and as a TCR transgenic (12). In addition, a CD8 T cell clone recognizing insulin B:15–23 is diabetogenic (13).

T cell receptors are composed of an α- and a β-chain, each of which are produced by a process combining multiple gene segments. For the α-chain there is “random” selection of ≈1/100 Vα-sequences and 1/50 Jα-sequences with a variable intervening N region (between Vα and Jα), which are combined with a nonvariable constant (Cα) sequence.

Of note, the T cell receptors of CD4 anti-B:9–23 clones exhibited a dominant conserved motif, with a conserved Vα/Jα-chain sequence, but multiple N region amino acids (e.g., E, A, GAN, ASG, S) and no conservation of the Vβ-chain (14, 15). In addition, the anti-B:9–23 protective 2H6 clone had the Jα-sequence identical to Wegmann clones, but a different Vα-sequence (16). Finally, antigen-presenting cells with both I-A^d and I-A^g7 (which share the same I-A α-chain) are able to present the B:9–23 peptide to the anti-B:9–23 clones, and both BALB/c (I-A^d) and NOD mice (I-A^g7) respond to immunization with the B:9–23 peptide with generation of high levels of insulin autoantibodies that cannot be absorbed by the immunizing peptide (and thus represent insulin autoantibodies rather than simply peptide autoantibodies) (17). Immunization of NOD mice with the B:9–23 peptide in adjuvant despite inducing insulin autoantibodies prevents diabetes. The observations above suggest an unusual recognition by a potentially low-stringency TCR dependent primarily on Vα- and Jα-sequences for recognition of a key islet autoantigen.

To directly test the hypothesis that such an α-chain motif is sufficient to engender anti-insulin autoimmunity, we created a series of transgenic mice expressing different combinations of the BDC 12-4.1 TCR α- and β-chains.

Results

Mice with Only the Conserved TCR α-Chain Express Insulin Autoantibodies.

To study the influence of the conserved Vα-chain in the absence of other α-chains we combined α-chain constant region mutation (Cα−/−) with our anti-B:9–23 TCR α-chain transgene (coding for a complete α-chain TCR). Mice with the Cα-mutation (Cα−/−) lack the constant region α-gene segment and thus cannot produce endogenous α-chains. Thus, the mice we produced have only an anti-B:9–23 α-chain but can normally produce diverse TCR β-chains.

The backcrosses either lacked endogenous constant α-genes (Cα−/−) or were able to produce endogenous TCR α-chains (Cα+/−) and had either the individual α- or β-BDC 12-4.1 anti B:9–23 TCR transgenes or both the 12-4.1 α- and β-chain transgenes (Table 1). We evaluated the 12-4.1 TCR α-chain expression level of splenocytes by using 12-4.1 TCR α-specific primers by real-time PCR. As expected, the 12-4.1 TCR α-gene expression was 250–300 times higher than NOD mice whether on the Cα−/− background or Cα+/− (12-4.1 TCR α+β-Cα−/− mice: 256 times, 12-4.1 TCR α+β−Cα+/− mice: 291 times). The backcrossed mice were evaluated for the production of insulin autoantibodies, insulitis and diabetes. As summarized in Table 1 and shown in Fig. 1, mice with the anti-B:9–23 TCR α-chain (A–D) produced insulin autoantibodies, regardless of the presence or absence of the anti-B:9–23 TCR β-chain, whereas mice bearing the 12-4.1 TCR β-chain transgene but not α-chain transgene (Fig. 1H) did not produce insulin autoantibodies (12-4.1 TCR α+β−Cα−/− vs. 12-4.1 TCR α−β+Cα+/−: P < 0.001, 12-4.1 TCR α+β+Cα−/− vs. 12-4.1 TCR α−β+Cα+/−: P < 0.001, 12-4.1 TCR α+β−Cα+/− vs. 12-4.1 TCR α−β+Cα+/−: P < 0.01, 12-4.1 TCR α+β+Cα+/− vs. 12-4.1 TCR α−β+Cα+/−: P > 0.05, Mann–Whitney t tests, peak insulin autoantibodies). The insulin autoantibodies of mice with the TCR α transgene appeared between 4 and 8 weeks of age. The levels of autoantibodies peaked and then usually disappeared by ≈20 weeks of age, similar to the time course of insulin autoantibodies in wild-type NOD or NOR mice (18). Seventy-four percent of the mice with both the α- and β-chain transgenes (14/19, Cα−/−), and thus the complete BDC 12-4.1 anti-B:9–23 TCR expressed insulin autoantibodies. This was slightly (not significantly P = 0.25) more than the 54% of mice positive for insulin autoantibodies with just the BDC 12-4.1 α-chain T cell receptor transgene (20/37, Cα−/−) where only the BDC 12-4.1 α-chain is available (Cα−/−) and the mouse utilizes endogenous TCR β-chains. If the mouse can also generate endogenous α-chains (Cα+/−) in addition to having the BDC 12-4.1 α-chain transgene there is no enhancement (43%, 7/16) of expression of insulin autoantibodies (Table 1). In the absence of the α-chain transgene there is little or no expression of insulin autoantibodies. In contrast to the α-chain transgene, the β-chain transgene by itself does not induce insulin autoantibodies. Thus, in these early backcross mice the BDC 12-4.1 α-chain transgene is both necessary and sufficient for the generation of insulin autoantibodies.

Table 1.

Insulin autoantibodies, insulitis, and diabetes in backcross mice relative to α and/or β anti-B:9–23 BDC 12-4.1 T cell receptor transgenes and ability to produce endogenous TCR α-chains (Cα)

12-4.1 anti B:9–23 TCR chains	Cα knockout	Expression of IAA	% islets with insulitis	Development of diabetes
α+β−	Cα−	20/3754%	Peri-islet 4.4%	0/370%
α+β+	Cα−	14/1974%	Intraislet 60%	1/195%
α+β−	Cα+	7/1643%	Intraislet 44%	1/166%
α+β+	Cα+	6/1638%	Intraislet 30%	0/160%
α−β−	Cα−	0/170%	No 0%	0/170%
α−β+	Cα−	0/1010%	No 0%	0/100%
α−β−	Cα+	1/1010%	Intraislet 57%	1/1010%
α−β+	Cα+	0/140%	No 0%	0/140%

IAA, insulin autoantibodies.

Fig. 1.

Serum anti-insulin autoantibody levels. (A–D) Mice with the 12-4.1 α-chain transgene. (E–H) Mice lacking the 12-4.1 α-chain. (A) BDC 12-4.1 TCR α+β−, Cα−/−, n = 37. (B) BDC 12-4.1 TCR α+β+, Cα−/−, n = 19. (C) BDC 12–4.1 TCR α+β−, Cα+/−, n = 16. (D) BDC 12-4.1 TCR α+β+, Cα+/−, n = 16. (E) BDC 12-4.1 TCR α−β−, Cα−/−, n = 17. (F) BDC 12-4.1 TCR α−β+, Cα−/−, n = 10. (G) BDC 12-4.1 TCR α−β−, Cα+/−, n = 10. (H) BDC 12-4.1 TCR α−β+, Cα+/−, n = 14. Lines connect results for individual mice. Upper-right number is the IAA positive mouse number out of the total mouse number. mIAA, micro insulin autoantibodies.

Anti-B:9–23 TCR β-Chain Allelic Exclusion Was Efficient.

With the constant region α-chain knockout (Cα−/−) we could guarantee that mice had only the introduced TCR α-chain of the transgene and no endogenous α-chains. Expression of a T cell receptor β-chain suppresses expression of a second β-chain (allelic exclusion). To confirm efficient β-chain allelic exclusion we analyzed expression of TCR β-chains by flow cytometry. Fig. 2 illustrates expression for the transgenic β-chain of BDC 12-4.1, which is a Vβ2 with or without the Cα knockout (Cα−/− or Cα+/−). As expected, a high percentage of CD4 T cells expressed Vβ2 (94.5 ± 3.34% Cα−/−, 99.2 ± 0.33% Cα+/−). To analyze suppression of expression of an endogenous β-chain in the presence of the BDC 12-4.1 β-chain we quantitated T cells expressing Vβ8.2. As expected, Vβ8.2 was suppressed relative to nontransgenic mice (Fig. 2 B and D). For mice with only the single transgenic α-chain (Cα−/− or Cα+/−) and able to produce endogenous β-chains 7–10% of CD4 T cells were Vβ2 (versus 95% with the Vβ2 transgene) and 16–20% Vβ8.2 (Fig. 2 A and C). These latter percentages are similar to nontransgenic NOD mice (data not shown). Despite the high expression of the Vβ2 transgene, mice with only the Vβ2 BDC 12-4.1 transgene did not express insulin autoantibodies—a marked difference from the α-chain transgene. In addition, the β-chain transgene by itself for mice that are able to generate endogenous α and β T cell receptors completely blocked expression of insulitis relative to mice lacking the transgene (e.g., Table 1, transgene TCR α−, β−, and Cα+/− 57% with insulitis vs. TCR α−, β+, and Cα+/− 0%, P < 0.01, insulitis score).

Fig. 2.

Flow cytometry of peripheral blood mononuclear cells with FITC-conjugated Vβ 2 and APC-conjugated CD4 antibodies (Right) and FITC-conjugated Vβ 8.1–8.2 and APC-conjugated CD4 (Left). One of three different representative flow cytometry analyses was shown. Upper-right quadrant number shown represents percentage Vβ (2 or 8.1–8.2) of CD3-gated cells.

Splenocytes from Mice Having Only BDC 12-4.1 α-Chain Transgene, Without the BDC 12-4.1 β-Chain Transgene Secreted IFN-γ in Response to Insulin B:9–23 Peptide.

Similar to nontransgenic NOD mice we could not demonstrate enzyme-linked immunospot (ELISPOT) responses to insulin peptide B:9–23 in nonimmunized 12-4.1 α-chain transgenic NOD mice (Cα−/−). After immunization with B:9–23 peptide both wild-type NOD mice and mice with only the 12-4.1 α chain (Cα−/−) responded with IFN-γ-secreting cells by ELISPOT analysis. As shown in Fig. 3, the 12-4.1 TCR α-chain only (Cα−/−) transgenic splenocytes produced IFN-γ when stimulated with the B:9–23 peptide. In contrast splenocytes from mice with only the 12-4.1 TCR β-chain transgene (Cα+/−) and endogenous α-chain did not produce IFN-γ (P < 0.01). The number of IFN-γ-producing cells from the mice with only the 12-4.1 TCR α-chain transgene was not different than wild-type NOD mice. As expected, no IFN-γ production was detected in mice with only the 12-4.1 TCR α-chain (Cα−/−) transgene stimulated with the control tetanus toxin peptide (Fig. 3 Right).

Fig. 3.

Dose–response to B:9–23 peptide, expressed as IFN-γ spots per million by ELISPOT assay of splenocytes. We tested three mice with only BDC 12-4.1 TCR α (Cα−/−), two mice with only BDC 12-4.1 TCR β (Cα +/−), and four wild-type NOD mice. Graphs represent typical results from one assay showing the number of IFN-γ-producing cells in response to a different dose of B:9–23 peptide and tetanus toxin (TT) peptide. All mice were immunized with B:9–23(100 μg per mouse) in complete Freund's adjuvant (CFA) by s.c. route 2 weeks before the assay. Splenocytes were stimulated by different concentrations of B:9–23 peptide (Left) or TT peptide (Right). Each line represents an individual mouse.

Mice Carrying Both BDC 12–4.1 TCR α- and β-Chain (Cα−/−) Develop Insulitis, Whereas Mice with Only BDC 12–4.1 TCR α-Chain Develop Only Peri-Islet Insulitis.

In contrast to production of insulin autoantibodies and IFN-γ, we found insulitis predominantly in mice with both the α- and β-chain transgenes. In mice with only the α-chain and not the β-chain transgene, <5% of total islets (10/315) displayed peri-islet insulitis and <2% of total islets (4/315) displayed intraislet insulitis (Fig. 4A), whereas mice with both the α- and β-chains had high levels of insulitis (22/69; with intraislet insulitis) (Fig. 4D) (P < 0.05, insulitis score) similar to NOD mice. We found insulitis for the single α-chain transgene (Cα−/−) only in mice >25 weeks of age, whereas we found insulitis in mice with both the α- and β-chains (Cα−/−) <10 weeks of age (data not shown). All mice that were Cα+/− and thus could generate multiple different Cα T cell receptors had insulitis (Fig. 4 B–D and E), except for mice with the TCR β-chain transgene without the TCR α-chain transgene (Fig. 4F). This BDC 12-4.1 β-chain transgene thus suppresses insulitis unless paired with its original BDC 12-4.1 α-chain. The α- and β-chain transgenic mice (Cα−/−) did not develop sialitis consistent with tissue specificity of targeting insulin (data not shown). In these early backcross mice, only 1 of 19 mice with both the α- and β-chain BDC 12-4.1 TCR (Cα−/−) progressed to diabetes and none of the BDC 12-4.1 α chain only TCR transgenic (Cα−/−) mice developed diabetes (Table 1).

Fig. 4.

Histology of transgenic mice showing islets stained for glucagon. We evaluated 15 mice (315 total islets, age range: 5–30 weeks of age) with only the 12-4.1 α-chain (Cα−/−); eight mice (89 total islets, age range: 8–29 weeks of age) with the 12-4.1 α-chain (Cα+/−); five mice (103 total islets, age range: 16–29 weeks of age) without both 12-4.1 α- and β-chain (Cα+/−); five mice (69 total islets, age range: 8–23 weeks of age) with both 12-4.1 α- and β-chain (Cα−/−); six mice (74 total islets, age range: 7–25 weeks of age) with both α- and β-chain (Cα+/−); seven mice (139 total islets, age range: 9–29 weeks of age) with 12-4.1 β-chain (Cα+/−). (A) BDC 12-4.1 TCR α+β−, Cα−/−, evaluated at 23 weeks. (B) BDC 12-4.1 TCR α+β−, Cα+/−, evaluated at 20 weeks. (C) BDC 12-4.1 TCR α−β−, Cα+/−, evaluated at 27 weeks. (D) BDC 12-4.1 TCR α+β+, Cα−/−, evaluated at 23 weeks. (E) BDC 12-4.1 TCR α+β+, Cα+/−, evaluated at 13 weeks. (F) BDC 12-4.1 TCR α-β+, Cα+/−, evaluated at 29 weeks.

Inhibition of Diabetes Development in Adoptive Transfer by BDC12-4.1 TCR α-Transgenic Cells.

Our TCR α-transgenic mice produced insulin autoantibodies, but similarly to mice immunized with the insulin B:9–23 peptide with induction of insulin autoantibodies (19), we hypothesized that they might not develop diabetes because of the presence of a regulatory response. To determine whether this transgenic-specific CD4 T cell response has a protective role in diabetes pathogenesis, α-chain-only transgenic spleen cells were transferred together with spleen cells from diabetic NOD females. As shown in Fig. 5, splenocytes from diabetic NOD females transfer diabetes into 100% of NOD scid recipients within 60 days. The cotransfer of splenocytes from the mice with only the TCR α-chain with diabetogenic NOD splenocytes significantly delayed transfer of diabetes compared with transfer of only the diabetic NOD splenocytes. By contrast, nontransgenic NOD splenocytes did not significantly delay development of diabetes.

Fig. 5.

Adoptive transfer of diabetes. Diabetic splenocytes (1.5 × 10⁷ per recipient) from wild-type NOD mice were transferred into NOD scid mice (8–10 weeks of age) alone (filled circles, n = 6) or together with a double number of splenocytes from the mice with only BDC 12-4.1 TCR α-chain (Cα−/−) (30–52 weeks of age) (filled squares, n = 8), or young NOD mice (8–10 weeks of age) (filled triangles, n = 10).

Retrogenic Mice Expressing a Different (BDC12–4.4) Conserved TCR α-Chain with Different N Region Amino Acids also Develop Insulin Autoantibodies and Diabetes.

The anti-B:9–23 conserved α-chains share Vα- and Jα-sequences, but junctional N region amino acids are diverse (15). To investigate whether a different conserved α-chain is also sufficient to induce anti-insulin autoimmunity, we created NOD mice retrogenic for the BDC 12-4.4 α-chain, which has the exactly same sequences as the 12-4.1 α-chain except the N region (N region: alanine for the 12-4.4 vs. GAN for the 12-4.1). As reported previously, the 12-4.4 T cell clone was able to induce diabetes in young NOD and NOD scid mice (20). Recipient NOD scid mice were transplanted with Cα−/− bone marrow cells carrying the 12-4.4 α-chain gene, thus T cells of the 12-4.4 α-chain retrogenic mice express only the 12-4.4 α-chain coupled with various β-chains from the endogenous β-chain genes. As shown in Fig. 6, three of six 12-4.4 α-chain-only retrogenic mice developed high levels of insulin autoantibodies, as approximately half of mice transgenic for only the 12-4.1 α-chain did (Fig. 1 A). Strikingly, one 12-4.4 α-chain retrogenic mouse also developed diabetes by 8 weeks after bone marrow transplant. Of note, the 12-4.4 retrogenic mice were generated on “full” NOD background, whereas the 12-4.1 transgenic mice analyzed were backcrossed onto NOD mice only three to four times.

Fig. 6.

Serum anti-insulin autoantibody levels of 12-4.4 α-chain retrogenic NOD mice. Three of six NOD scid mice retrogenic for the 12-4.4 α-chain in Cα−/− bone marrow cells developed insulin autoantibodies. Lines connect results for individual mice. mIAA, micro insulin autoantibodies (data for two mIAA negative mice overlap).

Discussion

Our prior studies indicate that insulin peptide B:9–23 is a key target of the immune pathogenesis of NOD diabetes, despite evidence for additional targets (21–23). The current study indicates that a conserved Vα/Jα-chain targeting the B:9–23 peptide is sufficient to drive the expression of high levels of insulin autoantibodies, even when only this TCR α-chain is expressed (Cα knockout mice per retrogenic mice). It is important to note that expression of the Vα-conserved transgene is greatly increased in both Cα+/− and Cα−/− mice potentially contributing to induction of autoimmunity with the single α-chain transgene. In man there is evidence that insulin B:9–23 (20), insulin A:1–15 (21), as well as other sequences are targets of autoimmunity, and Kent and coworkers found that insulin A:1–15 reacting T cell clones derived from pancreatic lymph nodes used conserved T cell receptor chains, but there are currently no data in terms of T cell receptor utilization for T cells invading islets.

Although the conservation of only Vα-sequences and the functional effect of the α-chain transgene by itself suggest a central role of the α-chain-conserved sequences, we find that the transgenic β-chain influences phenotype. By itself the 12-4.1 β-chain dramatically decreases insulitis [0% insulitis 12-4.1 TCR α−β+Cα+/− versus 57% 12-4.1 TCR α−β−Cα+/−, P < 0.01 (insulitis score)]. The current study demonstrates that the complete TCR α- and TCR β-chains of the anti-B:9–23 peptide not only induce insulin autoantibodies but also intraislet insulitis. Mice with either the 12-4.1 α-chain alone or the 12-4.1 β-chain alone did not have significant insulitis. Thus, the 12-4.1 α- and the 12-4.1 β-chains both influence development of insulitis and diabetes. Transfer experiments indicated inhibition of disease transfer by splenocytes from α-chain-only transgenics consistent with induction of regulation. Immunization with insulin peptide B:9–23 of NOD mice induces a similar phenotype with induction of high titers of insulin autoantibodies and diabetes prevention (19, 24). The 12-4.1 α-chain may also regulate islet autoimmunity when coupled with different β-chains.

The specificity of the usual α/β T cell receptor of mice is created by a combinatorial process, with choice of ≈1/100 Vα* 1/50 Jα* 1/25 Vβ* 1/2 Dβ* 1/12 Jβ* and α and β N region and β nDn amino acid modifications, giving >10¹⁵ potential combinations (25). In contrast if only a specific Vα and Jα combination is primarily required for targeting a peptide such as insulin B:9–23, the precursor frequency of autoreactive T cells might be greatly increased. We hypothesize that such a low-stringency T cell receptor defined by conservation of primarily germ line Vα and Jα chain that interacts with a crucial insulin peptide of the NOD, is a potential molecular mechanism underlying the frequent targeting of islet cells and development of anti-insulin autoimmunity of the NOD mouse. Indeed, our results indicate that two different conserved α-chains having the same germ line Vα and Jα but varying the junctional region resulted in the development of insulin autoimmunity. Such genome-encoded specific molecular targeting may be a general mechanism contributing to spontaneous organ-specific autoimmune disorders and other T cell phenotypes. Although the results with two different conserved α-chains are consistent with the hypothesis of a central role of this conserved sequence for NOD diabetes, they do not prove that only this sequence is essential and further study of addition T cell receptor chains and models is essential. The T cell receptor of NKT cells with its conserved α-chain (Vα14 Jα18 (mice) or Vα24 Jα18 (humans)) reacting with glycolipids (26, 27) is an example of receptor specificity generated primarily by specific genome encoded alpha chain segments.

To address the broader question of relevance to multiple autoimmune diseases, knowledge of primary peptide targets and T cell receptor conservation for multiple disorders is required. Given initial TCR-biased targeting of self-antigens as demonstrated in NOD mice, abnormalities of regulation would be needed for disease expression. It is of interest that the majority of children with foxp3 mutations and absence of foxp3 regulatory cells develop type 1 diabetes as neonates and young children (28). Given the remarkable associations of specific HLA genotypes with type 1A diabetes across population we hypothesize that there will be primary autoantigenic peptides, similar to the NOD mouse, facilitating the frequent targeting of human islet β cells and a concomitant conserved T cell receptor utilization for given diabetes associated HLA genotypes.

Materials and Methods

Mice.

The BDC 12-4.1 TCR transgenic mice were generated as described in ref. 11. For the current study, the BDC 12-4.1 T cell receptor transgenic mice on the FVB background (α #32 transgene, β [num] 82 transgene) mice were crossed and backcrossed onto NOD Cα knockout mice obtained from The Jackson Laboratory (strain 004444) to create mice that were then backcrossed to the NOD. No more than four backcrosses on the NOD (Cα−/−) background occurred for the mice described in this article. The lack of Vα2⁺ cells by flow cytometry confirmed that the TCR-Cα knockout mice lacked α-chain expression in the absence of the BDC 12-4.1 T cell receptor. The presence of TCR-α and TCR-β transgenes was determined at weaning by testing genomic DNA isolated from tail snips as described in ref. 11. Retrogenic mice were generated as described in ref. 29. Cα−/− bone marrow cells transduced with retroviral vectors carrying the 12-4.4 α-chain gene were infused into irradiated NOD scid mice. Mice were bred and housed under specific pathogen-free conditions at the University of Colorado Health Sciences Center for Comparative Medicine (CCM) in Aurora, CO. with an approved protocol from the University of Colorado Health Sciences Center Animal Care and Use Committee.

Insulin Autoantibody (IAA) Assay.

IAA expression was measured with the micro IAA assay (30) and a value of 0.01 or greater is considered positive.

Real-Time Quantitative PCR.

cDNAs from the mice spleen were used as templates in SYBER green real-time PCR assays on an ABI 7700 Sequence Detection System (Applied Biosystems). The primers used in this study were: 12-4.1 TCR alpha F (5′-GGCGAGCAGGTGGAGCAG-3′), 12-4.1 TCR α R (5′-GGAGTCACAGTTAAGAGAGATCC-3′), TCR Cα F (5′-GCTGTGTACCAGTTAAAAGATCCT-3′), TCR Cα R (5′-CTTGGAATCCATAGCTTTCATGTC-3′). Sample data were analyzed according to the comparative cycle threshold method and were normalized by TCR Cα expression in each sample.

Flow Cytometry.

Flow cytometry was performed by using the FACSCalibur (BD Biosciences). Peripheral blood mononuclear cells (PBMCs) were isolated by using Lympholyte-Mammal from Cedarlane Laboratories. All fluorescein-labeled antibodies were purchased from BD PharMingen A minimum of 10,000 cells were analyzed by flow cytometry. Expression of the β-chain in the founder was confirmed by flow cytometry of PBMCs stained with a FITC-conjugated Vβ2 antibody (for 12-4.1) and a FITC-conjugated Vβ 8.1–8.2 antibody.

Enzyme-Linked Immunospot Assay.

IFN-γ ELISPOT (enzyme-linked immunosorbent spot analysis) kits (BD Biosciences) were used for splenocyte analysis. Spleen cells were prepared by homogenization in red blood cell lysis buffer and cultured at 5 × 10⁵ cells per well in triplicate. Cells were cultured in Click's medium (Sigma-Aldrich) and 0.5% heat-inactivated mouse serum with 10 mM Hepes, pH 7.4 (GIBCO) and 8 mM l-glutamine (GIBCO) for 48 h at 37°C at 5% CO₂. Peptides used for stimulation were high-performance liquid chromatography purified (>98%) and dissolved in sterile lipopolysaccharide-free saline at a neutral pH (Genemed Synthesis Inc.). 3-Amino-9-ethylcarbazole substrate and Chromogen solution (BD Biosciences) were used as the detection method, and wells were counted and analyzed by using the Immunospot reader and software version 3.

Histology.

The pancreata obtained from the mice were fixed in 10% formalin, then embedded in paraffin. Paraffin-embedded tissue sections were stained with polyclonal guinea pig antiglucagon antibodies (Linco Research), followed by incubation with a peroxidase-labeled anti-guinea-pig IgG antibody (Kirkegaard & Perry Laboratories). A minimum of 10 islets from each mouse were microscopically observed for the presence of insulitis, and the levels of insulitis were scored according to the following criteria for analysis: 0, no lymphocyte infiltration; 1, islets with peri-islet insulitis; 2, <25% of the islet area infiltrated; 3, >25% infiltrated or small retracted islets. An Olympus BX51 microscope connected to a high-resolution digital camera (Pixera) and the software Image-Pro Plus (Media Cybernetics) was used.

Adoptive Transfer Experiments.

NOD scid mice (8–10 weeks of age) were used as recipients in adoptive transfer experiments. Diabetic splenocytes from various donors were injected i.p. (1.5 × 10⁷ per mouse) alone or together with BDC12-4.1 TCR α transgenic spleen cells (30–52 weeks of age) or young NOD (8–10 weeks of age) spleen cells. The mice were monitored for blood glucose biweekly after adoptive transfer, and the mice developed diabetes were killed immediately.

Diabetes.

Glucose was measured weekly with the FreeStyle blood glucose monitoring system (TheraSense), and the mice were considered diabetic after two consecutive blood glucose values of >250 mg/dl

Statistics.

The peaks of IAA, flow cytometry of Vβ expression, IFN-γ production, and insulitis score were analyzed with Mann–Whitney t tests. Survival curves were analyzed with the log-rank test. Statistical tests used PRISM software (Graphpad). P values <0.05 were considered statistically significant.