T-Cell Receptor Transgenic Response to an Endogenous Polymorphic Autoantigen Determines Susceptibility to Diabetes

NOD and NOD.CB17-Prkdc^scid/J (NOD-scid) mice were originally obtained from The Jackson Laboratory (Bar Harbor, ME) or the breeding colony at the Barbara Davis Center for Childhood Diabetes (Denver, CO). BALB/c mice were originally obtained from The Jackson Laboratory. Mice were bred and maintained under specific pathogen-free conditions at the University of Colorado Health Sciences Center’s Center for Laboratory Animal Care (Denver, CO), the Scripps Rodent Breeding Colony (La Jolla, CA), and Southern Illinois University School of Medicine animal facility (Springfield, IL).

Rearranged TCR α- and β-chain genes were amplified from BDC-6.9 T-cell clone mRNA by RT-PCR using TaqDNA polymerase (Perkin-Elmer Life Sciences, Boston, MA) and the following primers: 5′-TGTGAGTCGACATGAAAACATACGCTCCTACATTATT-3′ (α-5′); 5′-AGTCAGTCGACGTCAGTCTTGCAGACCTC-3′ (α-3′); 5′-GTCGACACTAGTCACATGGGCTCCATTTTCCTCAGT-3′ (β-5′); and 5′-AGGAGGC CGACTTGGCCGGTGAAGAACGGCTCA-3′ (β-3′). PCR products were cloned directly into the pCRII vector (Invitrogen Life Technologies, Carlsbad, CA). Following sequence verification, the α- and β-genes were subcloned into separate mCD4e/p-sil expression vectors kindly provided by D. Littman (construct “i”) (25). Briefly, this vector contains the murine CD4 minimal enhancer (339 bp), minimal promoter (489 bp), exon I, part of intron 1, and 32 bp of exon 2. A key intronic silencer element is absent, making this promoter capable of directing expression in both CD4 and CD8 T-cell lineages, as has been previously demonstrated (26). The CD4 exonic sequences encode 5′ untranslated regions. Appropriate orientation of each gene was confirmed by sequencing 5′ and 3′ junctions. Linear NotI fragments containing each TCR transgene (α 5,740 bp and β 5,860 bp) were purified from unwanted vector sequences using the Qiagen Gel Extraction kit (Qiagen, Valencia, CA), mixed in equal molar ratios and injected into F2 (C57BL/6By × BALB/cBy) embryos. Tail DNA from offspring was screened by PCR using the above primers. In the event that decreased fertility might result from early diabetes onset on the NOD background, transgenic founder mice were initially crossed to BALB/c mice and then subsequently backcrossed to either NOD or NOD.C6 mice (see description below) for three to five generations to generate 6.9TCR/NOD or 6.9TCR/NOD.C6 mice, respectively.

To create NOD.C6(N/B) mice, BALB/c mice were backcrossed to NOD mice for eight generations. Offspring were screened using PCR of tail DNA with microsatellite primers for D6Mit60 (ResGen; Invitrogen). Only mice that maintained heterozygosity at D6Mit60 were selected to breed for the next generation. Intercrossing was initiated with the eighth backcross generation to create NOD.C6 mice (homozygous BALB/c or B/B at D6Mit60). Additional genotyping was performed to establish an interval of BALB/c genome between D6Mit52 and D6Mit57. Transgenic mice were backcrossed to NOD.C6(N/B) or NOD.C6(B/B) mice for at least five generations to create 6.9TCR/NOD.C6(N/N), 6.9TCR/NOD.C6(N/B), or 6.9TCR/NOD.C6(B/B) mice. To simplify nomenclature, NOD.C6(B/B) mice are referred to in some instances as NOD.C6.

Peripheral blood was collected or single cell suspensions of spleen, peripheral lymph nodes, and thymus were made and erythrocytes lysed using 0.85% ammonium chloride. Mononuclear cell suspensions were stained with a combination of anti-CD4-FITC (fluorescein isothiocyanate), anti-CD8-allophycocyanin, and anti-Vβ4-biotin or anti-CD3-biotin. Anti-Vβ4 and anti-CD3 binding were visualized with streptavidin-phycoerythrin. All antibodies were purchased from BD Pharmingen (San Diego, CA). Immunofluorescence was detected using a FACS (fluorescence-activated cell sorter) Calibur, and analysis was performed on cells with lymphoid light-scatter properties using CellQuest version 3.2 software (Becton Dickinson).

Islet cells were used as a source of antigen to stimulate T-cell proliferation and cytokine production. Islets were harvested from whole pancreas by collagenase digestion as previously described (27,28).

To measure proliferative responses, mesenteric and pancreatic lymph node cells were harvested and pooled from 9- to 10-week-old 6.9TCR/NOD or nontransgenic age-matched NOD mice. Lymph node cells (5 × 10⁵ cells/well) were cultured in duplicate in the presence or absence of islet cells from NOD (1–2 × 10⁴ cells/well) or BALB/c (1.5 × 10⁴ cells/well) mice. All cultures were maintained in complete medium, which was prepared by supplementing Dulbecco’s modified Eagle’s culture medium (Life Technologies, Gaithersburg, MD) with 44 mmol/l sodium bicarbonate, 0.55 mmol/l l-arginine, 0.27 mmol/l l-asparagine, 1.5 mmol/l l-glutamine, 1 mmol/l sodium pyruvate, 50 mg/l gentamicin sulfate, 50 μmol/l 2-ME, 10 mmol/l HEPES, and 10% FCS. Cultures were incubated at 37°C with 10% CO₂. After 60 h, cultures were pulsed with [³H]thymidine (1 μCi/well), and 12 h later incorporation was determined using a Tomtec Harvester 96 Mach III (Tomtec, Hamden, CT) and a Wallac 1450 MicroBeta liquid scintillation counter (Perkin-Elmer Life Sciences).

To measure interferon (IFN)-γ production, spleen and lymph node mononuclear cells were harvested and pooled from a 12-week-old 2.5TCR/NOD mouse or from 8-week-old 6.9TCR/NOD or nontransgenic NOD littermate mice. Cells (5 × 10⁵/well) were cultured in the presence or absence of islet cells (10,000/well) from NOD.C6 or BALB/c mice. Triplicate samples were cultured in complete medium for 72 h, and then cell-free supernatants were assayed for IFN-γ by enzyme-linked immunosorbent assay using purified and biotinylated antibody pairs according to an established BD Pharmingen protocol.

Splenocytes were harvested from 6.9TCR/NOD, 6.9TCR/NOD.C6, or appropriate nontransgenic littermate mice. Mononuclear cells were prepared using gradient centrifugation with Lympholyte M (Cedarlane Laboratories Limited, Hornby, ON, Canada). Flow cytometry was used to determine the percentage of CD4 cells present. Cell concentrations were normalized based on CD4 cell number before injection of adult (intravenous; 2 × 10⁶ CD4 equivalents) or neonatal (intraperitoneal; 2.4 × 10⁶ CD4 equivalents) NOD-scid recipients. Alternatively, before injection of neonatal NOD-scid recipients, CD4 cells were purified from the spleens of 6.9TCR/NOD.C6 mice using mouse CD4 (L3T4) MicroBeads, MS+ Separation Columns, and the MiniMACS magnet system from Miltenyl Biotec (Auburn, CA). Purity of the CD4+ population was always >95%, as determined by flow cytometric analysis. Diabetes onset was monitored as described below.

Blood glucose measurements were made using OneTouch Ultra (Life Scan, Milpitas, CA) or Precision QID (MediSense, Waltham, MA) glucose meters. Mice were considered diabetic when blood glucose levels were >15 mmol/l (270 mg/dl). As a less-invasive screening method, urine glucose was measured using Diastix in some cases (Miles Laboratories, Elkhart, IN). Hyperglycemia was always verified for mice with >2% glucosuria. Urine and/or blood glucose levels were measured daily, weekly, or every 2 weeks as appropriate for the expected disease time course. Diabetes onset was graphed using Kaplan-Meier nonparametric survival curves, and statistical significance was determined by the Mantel-Cox log-rank test with StatView software.

Pancreas tissue from 6.9TCR/NOD, 6.9TCR/NOD.C6, or appropriate nontransgenic littermate mice was paraffin embedded, sectioned, and stained as previously described (29,30) with hematoxylin and eosin to determine mononuclear cell infiltration and/or with aldehyde-fuchsin to determine the extent of β-cell granulation. Sections from each mouse (containing on average 12 islets per section) were microscopically examined, and each islet was scored for the level of infiltration or β-cell granulation. Photographs were taken at ×200 (original magnification).

In designing a new diabetes model we took advantage of the previously well-characterized T-cell clone, BDC-6.9, which was derived from a newly diabetic NOD mouse. The BDC-6.9 clone is I-A^g7 restricted, specific for an unknown islet β-cell antigen, and diabetogenic upon injection into NOD or NOD-scid mice (29–31). Unlike the specificities of other NOD-derived T-cell clones, the BDC-6.9 clone is selectively responsive to islet antigen from NOD and NOD-related mouse strains (32). For example, BDC-6.9 cells do not proliferate or secrete cytokines when presented with islet cell antigens from BALB/c mice. Because we sought to exploit this discrimination, we first generated transgenic mice (6.9TCR) using rearranged TCR genes from the BDC-6.9 T-cell clone. Using PCR to screen for transgene integration, 7 founder mice were identified from 38 offspring: 1 integrated only 6.9 α-, 1 integrated only 6.9 β-, and 7 integrated both α- and β-transgenes (data not shown). Only founders that integrated genes for both chains were bred for further analysis.

Transgene β-chain expression was detected by flow cytometry and was used in addition to PCR to identify transgenic mice (Fig. 1). Intact expression of the BDC-6.9 TCR expression could not be demonstrated by flow cytometry since neither a clonotypic antibody nor an α-chain-specific antibody (Vα13) exists. However, offspring from two lines (9729 and 9812) demonstrated consistent germline transmission of α- and β-TCR chains (data not shown) and possessed peripheral T-cells exhibiting a high degree of expression of the transgene (Fig. 1) (Table 1). The CD4-to-CD8 ratios in peripheral blood lymphocytes were typically more than 4, and >85% of CD4 and 95% of CD8 T-cells expressed the transgene Vβ chain (Fig. 1) (Table 1). As studies progressed, no difference was observed between these lines, and therefore data from the 9729 line are shown unless otherwise indicated. A sample of Vα expression in the transgenic repertoire was obtained by examining peripheral CD4+ T-cells by flow cytometry with two commercially available Vα antibodies (Vα2 and 8.3). Significantly fewer Vα2 (2.4%)- and Vα8.3 (0.7%)-positive cells were found in transgenic compared with nontransgenic mice (Vα2 [9.8%] and Vα8.3 [4.8%]).

In the absence of BDC-6.9 TCR α-chain expression data, we sought to confirm the functional expression of the BDC-6.9 TCR by taking advantage of the established selective specificity of the BDC-6.9 T-cell clone (32). T-cells from 6.9TCR/NOD transgenic mice proliferated when provided with islet cell antigen from NOD, but not BALB/c mice (Fig. 2). In contrast, T-cells from age-matched nontransgenic littermates did not proliferate significantly to islet antigens from either source, reflecting the diverse repertoire of young NOD mice. Furthermore, 6.9TCR/NOD splenocytes (as few as 5 × 10⁵) transferred disease when injected into adult NOD-scid recipients (data not shown). Both in vitro and in vivo peripheral 6.9TCR/NOD T-cell responses were consistent with a repertoire composed primarily of T-cells expressing the TCR of BDC-6.9.

To determine whether disease incidence was affected by expression of the BDC-6.9 TCR, blood glucose levels of a cohort of 6.9TCR/NOD mice (from backcross generations 3–5) were compared with nontransgenic littermates. Disease onset was significantly accelerated not only among female 6.9TCR/NOD mice but also among transgenic males (Fig. 3A). More than 60% of all transgenic mice from backcrosses 3–5 were hyperglycemic by 12 weeks of age, whereas nontransgenic littermates remained euglycemic. The kinetics of disease onset were extreme, with hyperglycemia being detected as early as 4 weeks of age. At 3–5 weeks of age, NOD mice typically show few if any signs of impending disease, whereas at this time point, islet tissue from 6.9TCR/NOD mice showed profound inflammation (Fig. 3B). Islets were extensively infiltrated, and islet β-cell degranulation was significant (Fig. 3C and D). These results indicate that autoimmune diabetes was well established in young 6.9TCR/NOD mice. Disease was clearly accelerated, and the sex bias typically observed among NOD mice was almost nonexistent.

Given that both the BDC-6.9 T-cell clone and 6.9TCR/NOD T-cells exhibited strain-specific antigen responsiveness, we sought to control autoantigen expression in vivo. Previous mapping studies localized the gene encoding (or regulating) the β-cell autoantigen to a region on chromosome 6 (Fig. 4A). In these studies, microsatellite markers D6Mit58 and D6Mit60 showed 100% correlation with inheritance of the strain-specific phenotype (32). That is, the BDC-6.9 clone proliferated in response to islet antigen from mice with at least one NOD allele (defined at D6Mit58/60), but not if both alleles were of BALB/c origin.

In the current study, BALB/c mice were backcrossed for eight generations with NOD/LtJ mice and then intercrossed to generate NOD.C6 congenic mice. Microsatellite analysis of chromosome 6 allowed development of NOD mice that possessed NOD [NOD.C6(N/N)], BALB/c [NOD.C6(B/B)], or both [NOD.C6(N/B)] loci as defined by D6Mit60. Additional microsatellite analysis of the eighth backcross generation established that the BALB/c loci extended from D6Mit52 telomeric to D6Mit57 (data not shown).

Because type 1 diabetes is a complex polygenic disorder and both susceptibility and resistance loci have been identified (17,18,21,23), it was possible that BALB/c susceptibility and/or resistance genes might be located within or adjacent to the D6Mit52-57 interval. The presence of such elements could affect diabetes onset in our new model independent of TCR expression. To test this, we followed the onset of diabetes among male and female nontransgenic NOD.C6 mice. Onset of hyperglycemia was nearly identical among NOD.C6 mice of all three genotypes [NOD.C6(BB), (N/B), or (N/N)] (Fig. 5E and F). Onset of hyperglycemia among NOD.C6 females followed the typical time course for NOD mice, where ∼80% become diabetic at 3–7 months (33). Together, these data suggested that even if BALB/c disease-resistance genes remained within the NOD.C6 genome, they had little influence on diabetes onset among nontransgenic mice.

To verify the deficiency in autoantigen, islet tissue from NOD.C6 mice was used to stimulate T-cells from either 6.9TCR/NOD or 2.5TCR/NOD mice. For comparison, islet tissue from BALB/c mice was also tested. Islets obtained from both BALB/c and NOD.C6 mice elicited robust IFN-γ secretion from 2.5TCR/NOD cells (Fig. 4B). This is consistent with the high frequency of BDC-2.5 TCR-expressing T-cells in 2.5TCR/NOD mice and the known reactivity of the parental BDC-2.5 T-cell clone, neither of which can discriminate between NOD and BALB/c islet antigen (32). In contrast, 6.9TCR/NOD cells produced little or no IFN-γ in response to either islet source. Even though islet tissue from NOD.C6 mice maintained expression of at least one naturally occurring islet autoantigen (recognized by 2.5TCR/NOD cells), expression of the BDC-6.9-specific antigen was not detectable.

Once established, eighth-generation NOD.C6 or NOD.C6(N/B) mice were bred to 6.9TCR/NOD mice to generate transgenic mice that possessed NOD [6.9TCR/NOD.C6(N/N)], BALB/c (6.9TCR/NOD.C6), or both [6.9TCR/NOD.C6(N/B)] loci at D6Mit60. As with the 6.9TCR/NOD transgenics, nearly all peripheral T-cells from the 6.9TCR/NOD.C6 mice expressed Vβ4 (Fig. 5A and B). Expression was independent of chromosome 6 genotype, demonstrating that 6.9 T-cells successfully completed thymic positive selection even in the absence of endogenous autoantigen (Fig. 5A and B) (Tables 2 and 3). However, the level of Vβ4 expression was lower in peripheral CD4 cells from transgenic mice compared with that of controls (Fig. 5B). Reduction in TCR levels does not appear to result from peripheral downregulation of receptor expression since similar levels are found on single positive CD4 thymocytes of transgenic mice. Rather, this may reflect the inherent level of expression required for 6.9TCR transgenic CD4 cells to proceed successfully through thymic education, avoiding negative selection. Given the major histocompatibility complex class II restricted nature of the 6.9TCR, it was not surprising that thymic profiles of transgenic mice exhibited strong skewing to the CD4 lineage. However, the presence of double-negative CD3-positive thymocytes was unexpectedly increased in all 6.9TCR transgenic mice [Fig. 5A: percentage double negative (mean ± SD) for nontransgenic NOD.C6(N/N), 3.0 ± 0.3; for 6.9TCR/NOD.C6(N/N), 28.0 ± 11.0; and for 6.9TCR/NOD.C6(B/B), 28 ± 5.1%]. This likely results from early TCR expression during the double-negative stage of thymocyte development (data not shown), as has been previously described (26) with other transgenes using the same promoter. In some, but not all, 6.9TCR-Tg/NOD.C6(N/N) mice, a reduction in the percentage of double-positive thymocytes was observed compared with that of nontransgenic or 6.9TCR-Tg/NOD.C6(B/B) mice (Fig. 5A). This may be secondary to diabetes-induced elevations in glucocorticoids (34,35) because loss of double-positive thymocytes occurred in mice that were hyperglycemic at the time of analysis. Analysis of other 6.9TCR-Tg/NOD.C6(N/N) mice that were not yet hyperglycemic yielded profiles similar to those of 6.9TCR-Tg/NOD.C6(B/B) mice (Table 3). Since increased glucocorticoid levels are associated with hyperglycemia and can result in increased thymocyte apoptosis (36,37), the loss of double-positive thymocytes in some 6.9TCR-Tg/NOD.C6(N/N) mice may be glucocorticoid induced.

As with 6.9TCR/NOD mice, we observed an acceleration of disease onset among both male and female 6.9TCR/NOD.C6(N/N) mice (from backcross generations 5–8) compared with that of nontransgenic littermates (Fig. 5C and D). Hyperglycemia was detected as early as 21 days of life. Disease onset was also significantly accelerated among transgenic mice possessing only one NOD allele [6.9TCR/NOD.C6(N/B)]. However, disease penetrance before 12 weeks appeared to be reduced compared with that of 6.9TCR/NOD mice (compare Fig. 3A and Fig. 5C and D). This difference notwithstanding, diabetes still progressed with an overall accelerated time course for 6.9TCR/NOD.C6(N/N) and 6.9TCR/NOD.C6(N/B) mice.

Most importantly, and in contrast to transgenic NOD.C6 mice containing NOD alleles within the C6 locus, it was found that with 6.9TCR/NOD.C6 mice, in which the C6 locus is homozygous for BALB/c, no animals became hyperglycemic during the course of this study. This was true for both males and females over a period of up to 349 days. Despite the complete lack of hyperglycemia among 6.9TCR/NOD.C6 mice, islet infiltration was still evident (Fig. 6A). In young (8- to 13-week-old) mice, the extent of infiltration (Fig. 6B ) and degranulation among 6.9TCR/NOD.C6 mice was relatively low compared with their 6.9TCR/NOD counterparts, but was still detectable. Whether this islet inflammation is due to a small population of nontransgenic islet-reactive T-cells or to an unrelated defect in immune function is currently unclear and awaits analysis of 6.9TCR/NOD.C6 mice on a scid or Rag^null background.

Because peri-insulitis without frank diabetes occurs in 6.9TCR/NOD.C6 mice, it was possible that an abnormality in the transgenic T-cell pool existed, rendering them no longer diabetogenic. However, either purified CD4 lymph node T-cells or unseparated splenocytes from nondiabetic 6.9TCR/NOD.C6 mice were capable of transferring disease to NOD-scid recipients (Fig. 7A–C). The transfers with purified CD4 T-cell populations are especially compelling since these experiments indicate that transfer of disease by the transgenic T-cells does not require the presence of CD8 T-cells, recapitulating the behavior in vivo of the diabetogenic parent BDC-6.9 clone. The kinetics of disease transfer were similar to those observed with the transfer of splenocytes from diabetic nontransgenic littermates. Thus, despite the lack of diabetes in 6.9TCR/NOD.C6 mice, CD4 transgenic T-cells retained their diabetogenic function when placed in an antigen-competent host.

Generation of 6.9TCR transgenic mice has allowed the development of a flexible model of autoimmune diabetes whereby controlled expression of a naturally occurring autoantigen effectively determines disease development. As is observed with the BDC-6.9 T-cell clone, T-cells from the 6.9TCR transgenic mouse respond only to islet cells from NOD mice and not to BALB/c islet cells. This feature of restricted islet cell antigen specificity of the BDC-6.9 clone allowed for the expression of the TCR in a NOD mouse lacking the autoantigen for these T-cells. The NOD.C6 congenic mouse is BALB/c in the chromosome 6 interval containing the BDC-6.9 antigen locus, and the expression of the BDC-6.9 TCR in these mice leads to a transgenic animal that is protected against developing diabetes. In the 6.9TCR mouse on the NOD.C6 background, we did not observe diabetes in animals for up to 1 year. Whereas transgenic mice with one or two NOD alleles on chromosome 6 (∼67 cM) develop diabetes with accelerated kinetics, hyperglycemia does not occur if both alleles are of BALB/c origin. Importantly, splenocytes from 6.9TCR/NOD.C6 mice remain functionally capable of responding to cognate antigen as demonstrated by disease development following transfer into NOD-scid recipients. Therefore, absence of hyperglycemia in 6.9TCR/NOD.C6 mice most likely reflects a change in islet cell autoantigen expression.

The 6.9TCR/NOD mouse expresses TCR genes from BDC-6.9, a highly diabetogenic clone from our panel of Th1 cell clones (29). Unlike the situation with the TCR transgenic mouse made from the BDC-2.5 clone from our panel, in which diabetes incidence is quite low (13), the expression of the BDC-6.9 TCR on the NOD background results in a mouse exhibiting a markedly accelerated disease phenotype, with 60% of male and female transgenic animals developing hyperglycemia before 12 weeks of age. In this respect, the 6.9TCR/NOD mouse more closely resembles the 4.1-NOD TCR transgenic mouse described by Schmidt et al. (38), in which the T-cell repertoire was also derived from a pathogenic CD4+ T-cell clone.

Although autoreactive T-cells appear in the periphery of normal mice (and humans), their activation is controlled and autoimmunity is prevented by peripheral tolerance mechanisms that include regulatory cells (15,39,40). In some cases, regulatory T-cells are effective even in the presence of a nearly monospecific autoreactive T-cell population (13,41,42), and, for example, DX5⁺ cells limit the diabetogenic effect of T-cells in 2.5TCR/NOD mice (13). Our results would suggest that regulatory T-cells may not be effective at preventing disease in 6.9TCR/NOD mice, or, alternatively, these cells may be absent or severely reduced in number. The reason(s) for the striking difference in disease outcome between 6.9TCR/NOD (or 4.1-NOD) and 2.5TCR/NOD is not clear. One possibility is that autoantigen-specific T-cells may differ in affinity for antigen, trafficking capability, or other unknown characteristics, making them more or less affected by regulatory T-cells. Alternatively, differences in transgene expression and TCRα allelic exclusion may result in thymic development, which allows production of regulatory cells in the 2.5TCR mouse, but not the 6.9TCR mouse. Studies aimed at discriminating between these possibilities await analysis of transgenic NOD-scid or NOD-Rag^null mice.

Both susceptibility (Idd6 and Idd20) and resistance (Idd19) loci have been identified on chromosome 6 in crosses with Mus spretus, feral PWK, and the more closely related C3H mouse strains (43–45). However, recent mapping studies (32,45) show that none of these loci overlap with the locus controlling the BDC-6.9 autoantigen. Inheritance of BALB/c-derived susceptibility or resistance genes would be expected to change the incidence and/or kinetics of hyperglycemia. Although we cannot formally rule out simultaneous inheritance of susceptibility and resistance genes in the NOD.C6 mice, disease onset among nontransgenic NOD.C6 mice was nearly identical to that of NOD mice, strongly suggesting that NOD.C6 mice did not inherit BALB/c disease-modifying genes on chromosome 6.

In current NOD mouse model systems of autoimmune diabetes, determining whether an immune phenotype associated with disease development, such as increased activation-induced expression of type 1 cytokines (46), is caused by a genetic predisposition to disease or results from the disease process itself can be difficult. For example, these two possibilities cannot adequately be deciphered with current model systems where subclinical disease may be present (e.g., male NOD or young female NOD) or T-cells are absent (e.g., NOD-scid). The 6.9TCR/NOD.C6 congenic mouse, which possesses the genetic predisposition of the NOD strain but does not become hyperglycemic, may provide a new approach to separating disease development and genetic susceptibility. In other TCR transgenic mice produced with TCR from autoreactive T-cell clones, such as the 2.5TCR mouse, it cannot be determined at what point the T-cells are “naïve” for their self-antigen (47). However, in the 6.9TCR/NOD.C6 mouse in which the islet autoantigen is absent, T-cells arising in this mouse are truly naïve. Unchallenged mice will provide an opportunity to evaluate naïve T-cell function both in vitro and following in vivo immunization with peptide or cellular antigen. Therefore, 6.9TCR/NOD.C6 mice may be helpful in future studies aimed at defining complex susceptibility traits that might influence basal immune function in NOD mice.

This work was supported by National Institutes of Health Grant R01 DK 50561 to K.H., by a Juvenile Diabetes Research Foundation International (JDRF) Postdoctoral Fellowship Award to C.M.D., by a JDRF Advanced Postdoctoral Fellowship, by Southern Illinois University School of Medicine Central Research Committee and Excellence in Academic Medicine awards, and by an Illinois Health Improvement Foundation award to M.E.P.

The authors thank Anna Travelstead for flow cytometry data acquisition and the TSRI transgenic facility for microinjection and implantation of embryos.