Idd9.1 Locus Controls the Suppressive Activity of FoxP3⁺CD4⁺CD25⁺ Regulatory T-Cells

8.3-NOD and 8.3-NOD.Rag2^−/− mice have been described (15). NOD.B10 Idd9R28 (line 1104)–, NOD.B10 Idd9R11 (line 1105)–, and NOD.B10 Idd9.3 (line 1106)–congenic mice were developed as described previously (5). Development of the NOD.B10 Idd9.1 (line 1565) and NOD.B10 Idd9.2 (line 1566) strains and the fine-mapping of Idd9.1 and Idd9.2 will be the subjects of future reports. 8.3-NOD.B10 Idd9R28, 8.3-NOD.B10 Idd9R11, 8.3-NOD.B10 Idd9.1, 8.3-NOD.B10 Idd9.2, and 8.3-NOD.B10 Idd9.3 mice were established by crossing 8.3-NOD mice with NOD.B10 Idd9R28 (line 1104), NOD.B10 Idd9R11 (line 1105), NOD.B10 Idd9.1 (line 1565), NOD.B10 Idd9.2 (line 1566), and NOD.B10 Idd9.3 (line 1106) mice, respectively. The line number associated with each strain refers to the unique line designation for strains bred at Taconic (Germantown, NY). Lines 1565, 1566, and 1106 are available from Taconic through the Emerging Models Program. Lines 1104 and 1105 are no longer extant. Figure 1 A shows a map of the congenic strains used in this study in comparison to Idd9-congenic strains described by others (9,16,17). Lines 1565, 1566, and 1106 contain the B10 Idd9.1, Idd9.2, and Idd9.3 regions alone, respectively. The markers that define the boundaries of the lines described here are given in supplementary Table 1 (available in an online appendix at http://diabetes.diabetesjournals.org/content/early/2009/10/14/db09-0648/suppl/DC1), along with the primer sequences of novel markers developed for this study. supplementary Fig. 1 lists the single nucleotide polymorphisms (SNPs) distinguishing B10, B6, and NOD in the Idd9.1 region defined by line 1565. All genes containing polymorphisms are noted. Also noted are the regions of overlap of the B10 Idd9 congenic line with R201 of Ueno et al. (9) and the two Brodnicki et al. (16,17) B6 Idd9/Idd11–congenic lines defining the smallest Idd11 interval published to date. Information on the congenic strains is also available at http://www.t1dbase.org/page/DrawStrains (select the strain of interest). Note that the B10 Idd9 fragment in line 1565 overlaps the distal 30% of the B10 Idd9.1 fragment carried by the R201 strain from Ueno et al., and that Brodnicki et al.'s Idd11-congenic strains were produced using B6 donor mice. The 8.3-NOD.Igμ^−/−, 8.3-NOD.B10 Idd9.1/Igμ^−/−, and 8.3-NOD.B10 Idd9.1/Rag2^−/− mice were established by crossing 8.3-NOD and 8.3-NOD.B10 Idd9.1 with NOD.Igμ^−/− and NOD.Rag2^−/− mice, respectively, followed by backcrossing 8.3-TCR⁺, mutant heterozygous F1 mice to 8.3-NOD.B10 Idd9.1 mice and intercrossing 8.3-TCR⁺, mutant heterozygous B10 Idd9.1 homozygous littermates. All mice were housed in specific pathogen-free conditions.

Peptides (NRP-A7 and TUM) were purchased from Mimotopes (Clayton, Victoria, Australia). All monoclonal antibodies (mAbs) were purchased from PharMingen (San Diego, CA), unless indicated otherwise. Anti-mouse glucocorticoid-induced tumor necrosis factor receptor (GITR), anti-mouse/rat FoxP3, and anti–folic receptor 4 (FR4) were from eBiosciences (San Diego, CA). Streptavidin-peridinin-chlorophyll-protein complex was from Becton Dickinson (San Jose, CA).

Mesenteric lymph node (MLN) and pancreatic lymph node (PLN) DCs were purified from collagenase-digested MLNs and PLNs using anti-CD11c mAb-coated magnetic beads (Miltenyi Biotec). Purified DCs were pulsed with NRP-A7 or TUM peptides (1 μmol/l) for 1 h at 37°C and then used in proliferation and cytokine secretion assays using 8.3-CD8⁺ T-cells as responders.

Diabetes was monitored by measuring urine glucose with Diastix (Bayer, ON, Canada); animals were considered diabetic after two readings ≥3+. Differences between diabetes survival curves were compared with the Kaplan-Meier log-rank test using Prism software (Graphpad). Insulitis scores were measured on 14 hematoxylin-eosin–stained pancreas sections 150 μm apart, using the following criteria: 0, intact islet; 1, peri-insulitis; 2, up to 25% of the islet infiltrated; 3, 25–50% infiltration; 4, >50% infiltration.

Splenic CD8⁺ T-cells (2 × 10⁴/well) were incubated with NRP-A7/TUM peptide-pulsed (0.0001–1 μmol/l) antigen-presenting cells (APCs; 10⁵ irradiated splenocytes/well) for 2 or 3 days (for cytokine and proliferation assays, respectively) at 37°C in 5% CO₂. Cytokines in the supernatants were measured by enzyme-linked immunosorbent assay. The 3-day cultures were pulsed with 1 μCi [³H] thymidine during the last 18 h and harvested. The regulatory activity of CD4⁺CD25⁺ T-cells was measured by adding 0.125–1 × 10⁴ CD4⁺CD25⁺ T-cells preactivated with anti-CD3 mAb and recombinant (r)IL-2 to CD8⁺ cell–APC cocultures for 48 h (interferon γ [IFN-γ] content measurements) or 72 h (proliferative activity, assessed by [³H] incorporation during the last 18 h of culture).

Splenic CD8⁺ T-cells purified from 8.3-NOD and 8.3-NOD.B10 Idd9.1 mice using anti-CD8–coated microbeads (Miltenyi Biotec) (2 × 10⁴ cells/well) were stimulated with NRP-A7–pulsed irradiated NOD splenocytes (10⁵ cells/well) for 3 days and expanded in 0.5 units/ml of rIL-2 (Takeda, Osaka, Japan) for 7–10 days. Cytolytic activity of cytotoxic T-lymphocytes (CTLs) was measured using NRP-A7– or TUM-pulsed (1 μmol/l) RMA-SK^d cells.

Splenic CD8⁺ T-cells were purified using iMAG CD8 beads (BD Bioscience) following the manufacturer's protocols, labeled with carboxyfluorescein succinimidyl ester (CFSE; 2.5 μmol/l), and injected intravenously (10⁷ CD8⁺ T-cells) into 9- to 11-week-old hosts (NOD or NOD.B10 Idd9.1). Hosts were killed 6 days later and their PLNs and MLNs examined for presence of CD8⁺ CFSE⁺ cells. To purify CD4⁺CD25⁺ cells, lymph node and/or splenic cells were enriched for CD4⁺ cells, incubated with anti–CD25-phycoerythrin (PE), and separated using anti-PE mAb-coated beads (Miltenyi Biotec). The purity was >85% for CD4⁺CD25⁺ cells. The mice were injected intravenously with 2 × 10⁵ cells and monitored for diabetes for 15 weeks.

Data were compared using log-rank, Mann-Whitney U, or two-way ANOVA tests.

To determine whether the B10-derived Idd9 region that protects from spontaneous type 1 diabetes can also protect from disease accelerated by the presence of a diabetogenic TCR, we introgressed two copies of the B10 Idd9.1-9.3 and Idd9.2-Idd9.3 intervals (Fig. 1 A) into transgenic NOD mice expressing a diabetogenic T-cell receptor, 8.3 (15). This TCR is representative of a large fraction of islet-associated CD8⁺ T-cells in NOD mice that use highly homologous TCRα chains (15,18,19) and recognize the mimotopes NRP-A7 and NRP-V7 in the context of the major histocompatibility complex (MHC) molecule H-2K^d (20). These T-cells are already a significant component of the earliest NOD islet CD8⁺ infiltrates (19,–21), are pathogenic (15,18), target a peptide from islet-specific glucose 6 phosphatase catalytic subunit–related protein (IGRP_206–214, similar to NRP-A7) (22), and are unusually frequent in the periphery (>1/200 CD8⁺ T-cells) (23).

8.3-NOD.B10 Idd9R28 mice, carrying the complete B10 Idd9 interval that affords >95% protection from spontaneous type 1 diabetes (5), displayed a significantly reduced frequency of diabetes compared with their noncongenic 8.3-NOD counterparts (∼40 vs. 80%) (Fig. 1,B). The shorter B10 Idd9.2-Idd9.3 (Idd9R11) interval, which provides 60% protection from spontaneous type 1 diabetes (5), was also antidiabetogenic, but to a much lesser extent (Fig. 1,B). To ascertain whether the higher antidiabetogenic effect of Idd9 was due to additive/synergistic effects of loci contained in different subregions, we developed 8.3-NOD mice congenic for each of the individual Idd9 intervals (Idd9.1, Idd9.2, and Idd9.3); each of the individual Idd9 congenic regions affords non–TCR-transgenic NOD mice protection from spontaneous type 1 diabetes as shown in Fig. 1,C. The fine-mapping of the Idd9.1 and Idd9.2 regions is in progress and will be the subject of separate reports (D.B.R. et al. and Hamilton-Williams et al., manuscripts in preparation). The Idd9.3 region is 1.2 Mb with only 15 genes, including the prime candidate Tnfrsf9, which encodes 4–1bb (also known as CD137) (24,–26). As shown in Fig. 1 D, the incidence of diabetes in 8.3-NOD.B10 Idd9.1 mice was virtually identical to that seen in 8.3-NOD.B10 Idd9R28 mice, suggesting a minor contribution of the B10 Idd9.2 and Idd9.3 alleles on 8.3-CD8⁺ T-cell–enhanced type 1 diabetes. In agreement with these observations, the diabetes survival curves corresponding to 8.3-NOD.B10 Idd9.2 and 8.3-NOD.B10 Idd9.3 mice were remarkably similar to those seen in 8.3-NOD.B10 Idd9R11 mice. This is in contrast to what occurs in nontransgenic NOD mice, where Idd9.2 and Idd9.3 augment the antidiabetogenic effects of each other and of Idd9.1 (5). Because the Idd9.1 interval is as protective in 8.3-NOD mice as it is in non–TCR-transgenic NOD mice (5), these data suggest that expression of the diabetogenic TCR overwhelms the antidiabetogenic properties of Idd9.2 and Idd9.3, but not the effects of Idd9.1.

Because the antidiabetogenic effect of the B10 Idd9.1 region in 8.3-NOD.B10 Idd9.1 mice was associated with slower progression of insulitis (Fig. 2,A), we asked whether this region's antidiabetogenic activity was mediated by effects on the developmental biology of 8.3-CD8⁺ T-cells. Cytofluorometric studies indicated that the thymi and spleens of both types of mice contained similar absolute and relative numbers of double-positive CD4⁺CD8⁺ thymocytes and/or single-positive CD4⁺ or CD8⁺ T-cells, respectively (Fig. 2,B and data not shown). In addition, the transgenic CD8⁺ T-cells of both types of mice expressed comparable levels of the transgenic Vβ8.1⁺ TCR and CD8 (Fig. 2,C), and bound NRP-V7/K^d and IGRP_206–214/K^d tetramers (recognized with high and intermediate avidity, respectively) with similar frequency (Fig. 2,D) and mean fluorescence intensity (Fig. 2 E).

Functional in vitro assays revealed that the splenic CD8⁺ T-cells of both types of mice proliferated equally well and secreted similar levels of IFN-γ and IL-2 in response to NRP-A7 peptide stimulation over a range of concentrations (Fig. 3,A). Likewise, in vitro–differentiated 8.3-CD8⁺ T-cells from both strains killed peptide-pulsed target cells with similar efficiency (Fig. 3 B).

Similar results were obtained in vivo. CFSE-labeled 8.3-CD8⁺ T-cells from 8.3-NOD.B10 Idd9.1 mice proliferated in the pancreatic (but not mesenteric) lymph nodes (PLNs and MLNs, respectively) of wild-type NOD hosts as efficiently as those derived from 8.3-NOD donors (Fig. 3 C).

Taken together, these observations indicated that the B10 Idd9.1 allele does not impair the development or function of IGRP_206–214-reactive CD8⁺ T-cells.

Whereas CFSE-labeled 8.3-CD8⁺ T-cells proliferated equally well in vivo regardless of the Idd9.1 genotype of the donor mice, CFSE-labeled 8.3-CD8⁺ T-cells from both 8.3-NOD and 8.3-NOD.B10 Idd9.1 mice proliferated significantly less in the PLNs of NOD.B10 Idd9.1 hosts than in the PLNs of NOD hosts (Fig. 3,C and D), without any obvious effects on the relative or absolute numbers of recruited/retained 8.3-CD8⁺ T-cells (Fig. 3,E). This suggested that B10 Idd9.1 somehow impairs the cross-priming of 8.3-CD8⁺ T-cells by endogenous autoantigen-loaded dendritic cells. This was not due to interstrain differences in autoantigen loading of DCs in vivo but rather to impaired PLN DC function because whereas ex vivo, NRP-A7–pulsed DCs purified from the MLNs of NOD and NOD.B10 Idd9.1 mice could stimulate naive 8.3-CD8⁺ T-cells with similar efficiency, NRP-A7–pulsed DCs purified from the PLNs of NOD.B10 Idd9.1 mice could not do so as efficiently as NRP-A7–pulsed DCs purified from the PLNs of NOD mice (Fig. 3 F).

To ascertain whether the CD8⁺ T-cell–extrinsic effects of B10 Idd9.1 on 8.3-CD8⁺ T-cell cross-priming in vivo were associated with B10 Idd9.1's antidiabetogenic activity, we introduced a recombination activating gene (Rag2) deficiency into 8.3-NOD and 8.3-NOD.B10 Idd9.1 mice. Although penetrance of type 1 diabetes in 8.3-NOD.Rag2^−/− mice is significantly lower than in 8.3-NOD mice (owing to lack of CD4⁺ T-cell help) (15), the incidence of diabetes in 8.3-NOD.Rag2^−/− and 8.3-NOD.B10 Idd9.1/Rag2^−/− mice was essentially the same (Fig. 4,A), indicating that the RAG-2 deficiency had completely abrogated the antidiabetogenic effects of the B10 Idd9.1 allele. Because these mice export only 8.3-CD8⁺ T-cells (but not CD4⁺ T-cells or B-cells) to the peripheral lymphoid organs (15), we reasoned that the protective activity of B10 Idd9.1 was mediated by endogenous T-cells and/or B-cells. To distinguish between these two possibilities, we introduced an immunoglobulin μ heavy chain gene (Igμ) deficiency into 8.3-NOD and 8.3-NOD.B10 Idd9.1 mice and followed both types of B-cell–deficient mice for development of diabetes. B-cell–deficient 8.3-NOD.Igμ^−/− mice developed diabetes essentially like B-cell–competent 8.3-NOD mice, indicating that type 1 diabetes development in the 8.3-NOD model, unlike the case in nontransgenic NOD mice, is B-cell independent (Fig. 4 B). In contrast, 8.3-NOD.B10 Idd9.1/Igμ^−/− mice remained protected from type 1 diabetes despite lacking B-cells, indicating that the protective effects of the B10 Idd9.1 region are B-cell independent and mediated by endogenous T-cells.

Prompted by these findings and the observation that the protective effect of the B10 Idd9.1 interval of line 1565 was not associated with increased development of thymic iNKT cells (9) (supplementary Fig. 2), we investigated whether B10 Idd9.1 might enhance the development and/or function of FoxP3⁺CD4⁺CD25⁺ regulatory T-cells (Tregs). We compared the relative size of the FoxP3⁺CD4⁺ T-cell subset in thymuses, spleens, MLNs, and PLNs of 8.3-NOD and 8.3-NOD.B10 Idd9.1 mice. The lymphoid organs of 8.3-NOD.B10 Idd9.1 mice contained higher percentages (albeit not absolute numbers) of FoxP3⁺CD4⁺ T-cells than those of 8.3-NOD mice, although the differences were relatively small and statistically significant only for thymuses and MLNs (Fig. 5,A and B, and supplementary Fig. 3A). These differences were not a peculiarity of TCR-transgenic mice, because the thymuses and MLNs of nontransgenic NOD.B10 Idd9.1 mice also contained significantly higher percentages of these T-cells compared with NOD mice (Fig. 5,C). As expected, there were no differences in the percentages of FoxP3⁺ T-cells within the CD4⁺CD25⁺ T-cell subsets (Fig. 5,A and C). These CD4⁺CD25⁺ T-cells had all the hallmarks of CD4⁺CD25⁺ Tregs described in other strains: they expressed GITR, FR4, CD134, and transforming growth factor β (TGF-β; Fig. 5,D), and produced IFN-γ, IL-4, and IL-10 when stimulated with plate-bound anti-CD3 mAb in the presence, but not absence, of rhIL-2 (Fig. 5,E). The CD4⁺CD25⁺ T-cells of transgenic and nontransgenic NOD and NOD.B10 Idd9.1 mice expressed similar levels of all these markers, including CD25 and FoxP3 (Fig. 5 F and data not shown).

We next asked whether these B10 Idd9.1-associated differences in the relative size of the FoxP3⁺CD4⁺CD25⁺ Treg subset were accompanied by differences in suppressive activity. We first compared the ability of CD4⁺CD25⁺ (and CD4⁺CD25⁻) T-cells purified from NOD.B10 Idd9.1 and NOD mice to suppress the activation of naive 8.3-CD8⁺ T-cells from 8.3-NOD mice in response to NRP-A7 peptide–pulsed APCs. The CD4⁺CD25⁺ T-cells purified from 8.3-NOD.B10 Idd9.1 or NOD.B10 Idd9.1 mice had significantly higher regulatory activity than the CD4⁺CD25⁺ T-cells purified from 8.3-NOD or NOD mice, respectively, at different CD4⁺CD25⁺/CD8⁺ T-cell ratios (Fig. 6,A and B, and supplementary Fig. 3B). The CD4⁺CD25⁻ T-cells of both types of mice were not suppressive in these assays (data not shown), as described previously (27). These strain differences in Treg function were also true in vivo. Whereas 8.3-NOD.Rag2^−/− hosts transfused with CD4⁺CD25⁺ T-cells from wild-type NOD donors developed spontaneous diabetes essentially like nonmanipulated 8.3-NOD.Rag2^−/− mice (15), 8.3-NOD.Rag2^−/− hosts transfused with CD4⁺CD25⁺ T-cells from NOD.B10 Idd9.1 mice remained diabetes free throughout the follow-up period (Fig. 6 C).

Taken together, these data are consistent with the idea that diabetes resistance afforded by the B10 Idd9.1 interval is mediated, at least in part, by enhanced development and function of CD4⁺CD25⁺ regulatory T-cells.

Analyses of Idd-congenic NOD mice, in which type 1 diabetes–associated chromosomal regions are replaced by the corresponding regions found in nondiabetes-prone strains of mice, such as B6 and B10, have demonstrated the existence of more than 20 non–MHC-linked type 1 diabetes–associated regions/loci (1,28). The nature of the genes that are responsible for these associations has been resolved for only a few Idd regions, including Idd16-β2m (29), Idd3-Il2 (14), Idd5.1-Ctla4 (30,31), and Idd5.2-Nramp1 (30,32,33). Furthermore, with the exception of Idd3 (14), which controls Treg development and function, the mechanisms by which most of these loci contribute to the diabetogenic process remain unclear.

The current study aimed to elucidate mechanisms underlying the resistance to type 1 diabetes afforded by B10 Idd9 (5), to help pave the way toward eventual identification of the responsible gene. We compared the fate and diabetogenicity of a type 1 diabetes–relevant autoreactive CD8⁺ T-cell population in B10 Idd9–congenic versus noncongenic TCR-transgenic NOD mice. We found that this reductionist system, which has enabled us to solve Idd3 (14), is not informative to dissect mechanisms underlying Idd9.2- and Idd9.3-linked suppression, presumably because the large size of the transgenic CD8⁺ T-cell population born by these mice largely overwhelms the corresponding antidiabetogenic mechanisms. This, however, was not the case for Idd9.1. Our data show that Idd9.1 regulates the suppressive activity and antidiabetogenic potential of FoxP3⁺CD4⁺CD25⁺ Tregs. Tregs arising in B10 Idd9.1–congenic NOD mice were significantly more suppressive and antidiabetogenic than those arising in NOD mice. Because FoxP3⁺CD4⁺ T-cells suppress diabetogenic T-cell responses, at least in part, by inhibiting DCs in the pancreatic draining lymph nodes (27), these observations are consistent with the suboptimal ability of autoantigen-loaded (in vivo) and peptide-pulsed (ex vivo) PLN DCs from NOD.B10 Idd9.1 mice to support 8.3-CD8⁺ T-cell activation, both in vitro and in vivo; selective suppression of PLN DCs in these mice may result from cognate suppression of autoantigen-loaded DCs, which predominantly (if not exclusively) reside in the PLNs, by local autoregulatory CD4⁺CD25⁺ Tregs. Accordingly, we propose that B10 Idd9.1–“enhanced” Tregs contribute to B10 Idd9.1–linked type 1 diabetes inhibition by suppressing the activation of CD8⁺ (and possibly other) effector T-cells in the pancreatic draining lymph nodes, without affecting their developmental biology or intrinsic diabetogenic potential. These data might explain the observations of Chamberlain et al. in mice expressing a rat insulin promoter-driven tumor necrosis factor α transgene, in which B10 Idd9.1 somehow suppressed the ability of CD8⁺ T-cells to respond to islet-infiltrating APCs (34). The results are also compatible with the DC phenotype described by O'Keeffe et al. (12), with the caveat that this congenic strain, unlike the one described here, was produced using B6, as opposed to B10, donor mice and that B10 and B6 DNAs differ substantially in the distal region of chromosome 4 (supplementary Fig. 1). These observations, however, do not exclude, and are compatible with, possible effects of Idd9.1 polymorphisms on other lymphocyte populations. For example, it has been reported that a B10 chromosome 4 region partially overlapping the B10 Idd9.1 fragment studied here restores impaired iNKT cell development and an iNKT cell–related DC phenotype described in NOD mice (9). Our congenic line 1565, however, bears a B10 chromosome 4 fragment that overlaps only the distal 30% of the B10 chromosome 4 fragment present in Ueno et al.'s R201 line. Furthermore, line 1565 does not display Ueno et al.'s iNKT cell phenotype, suggesting that this phenotype is encoded in a gene that is not in the overlapping region. Also, our work does not exclude possible effects of genes within Idd9.1, Idd9.2, and/or Idd9.3 on peripheral tolerance of autoreactive T-cell clones other than those expressing the 8.3-TCR (35), such as for example CD8⁺ T-cells recognizing IGRP_206–214 with higher avidity (36).

It is intriguing that the effects of the B10 Idd9.1 allele on Treg function and type 1 diabetes susceptibility observed in this study are remarkably similar to those we have described recently for B6 Idd3, where Treg development and function are enhanced by increased transcriptional activity of Il2 (14). Accordingly, it is tempting to suspect that regulatory lymphocytes, including the FoxP3⁺CD4⁺CD25⁺ subset, may be common targets of type 1 diabetes–associated chromosomal regions. This hypothesis is compatible with what we know about the type 1 diabetes gene associations in humans. In addition to the IL2-IL21 region on human chromosome 4q27 (2,4,37), and IL2RA on chromosome 10p15.1, which encodes CD25, the high-affinity receptor for IL-2 expressed by activated T-cells, and FoxP3⁺CD4⁺CD25⁺ Tregs (37,,–40), several other candidate human type 1 diabetes–associated loci play a role in Treg development and function. For example, the type 1 diabetes susceptibility gene CTLA4 on chromosome 2q33.2 encodes a negative regulator of T-cell–mediated immunity (31,41,42). The susceptibility allele generates reduced levels of an alternatively spliced transcript that encodes the soluble form of CTLA-4, which has been correlated with variations in the peripheral frequency of CD4⁺CD25⁺ Tregs, which constitutively express CTLA-4 (43). PTPN22 on chromosome 1p13.2, associated with susceptibility to a number of autoimmune diseases, including type 1 diabetes (40, 44,–46), encodes lymphoid tyrosine-phosphatase (LYP), another negative regulator of T-cell activation. A nonsynonymous polymorphism (R620W) abolishes the binding of Csk, a negative regulatory kinase, to LYP. Consequently, the R620W LYP variant results in a gain of function that presumably enables LYP to more effectively suppress T-cell signaling compared with the non–type 1 diabetes–associated variants. It has been suggested that PTPN22 polymorphisms contribute to autoimmune disease susceptibility by impairing negative selection of autoreactive thymocytes and by decreasing the number and function of CD4⁺CD25⁺ Tregs (47,48). Another locus possibly implicated in autoimmune disease susceptibility by modulating T-cell signaling is PTPN2 on chromosome 18p11.21, a nonreceptor type 2 tyrosine phosphatase that is a negative regulator of T-cell signaling (2).

The fine-mapping of Idd9.1 has not progressed sufficiently to productively comment on candidate genes present in the region. The B10-derived congenic interval present in the Idd9.1 strain used in this study (line 1565) has 329 annotated genes in this 18.5-Mb segment. The Idd11 region as defined by B6-derived congenic strains (16) is an ∼8-Mb region that is wholly contained within the Idd9.1 interval defined by the congenic strain used in this study. However, B6 and B10 are not identical by descent throughout the overlapping region (http://phenome.jax.org); from SNP analyses, B6 and B10 differ at 32.3% of the SNPs within the 8-Mb Idd9.1/Idd11 interval (supplementary Fig. 1). Notably, within this 8 Mb, there are small regions that appear to be identical-by-descent for B6 and B10, including ∼200 kb surrounding Lck, a previously highlighted candidate gene in the Idd9.1 region (16), where the ancestral haplotype shared by the B6 and B10 strains differs from that present in NOD mice (supplementary Fig. 1). Further fine-mapping and gene expression studies are required to define the genes that are responsible for Idd9.1- and Idd11-mediated type 1 diabetes protection and to determine whether the protective alleles are indeed shared by the B6 and B10 strains in this region.

In conclusion, we have shown that the B10 Idd9.1 locus affords diabetes resistance, at least in part, by promoting the function of CD4⁺CD25⁺ regulatory T-cells without altering the developmental biology of effector CD8⁺ T-cells.

This work was supported by grants from the Canadian Institutes of Health Research (to P.S.). J.Y. was supported by fellowships from the Canadian Diabetes Association and the Juvenile Diabetes Research Foundation (JDRF). P.S. is a Scientist of the Alberta Heritage Foundation for Medical Research (AHFMR). The JMDRC is supported by the Diabetes Association (Foothills). L.S.W. is supported by grants from the JDRF and the Wellcome Trust (WT) and is a JDRF/WT Principal Research Fellow. Support for this work was also provided by National Institute of Allergy and Infectious Disease (NIAID) Grant AI 070351 from the National Institutes of Health. The Cambridge Institute for Medical Research (CIMR) is in receipt of a Wellcome Trust Strategic Award (079895). The availability of NOD congenic mice through the Taconic Farms Emerging Models Program has been supported by grants from the Merck Genome Research Institute, NIAID, and the JDRF.

No potential conflicts of interest relevant to this article were reported.

We thank S. Bou, S. Thiessen, M. Deuma-Myers, H. Metselaar, M. DeCrom, and T. Irvine for technical assistance with mouse breeding, genotyping, and animal care; L. Kennedy and L. Robertson for flow cytometry; C. Gwozd for histology; Y. Yang for aGalCer/Cd1 tetramers; and S. Tsai, A. Shameli, and Y. Yang for helpful discussions.