Genes within the Idd5 and Idd9/11 Diabetes Susceptibility Loci Affect the Pathogenic Activity of B Cells in Nonobese Diabetic Mice¹

Type 1 diabetes (T1D)³ in humans and the NOD mouse model results from T cell-mediated autoimmune destruction of insulin-secreting pancreatic β cells. Although B cells are also among the earliest leukocytes to infiltrate the pancreatic islets of NOD mice (1, 2) and the specific autoantibodies they secrete are excellent diagnostic markers of future risk for T1D in both humans and NOD mice (3), their presence was believed to be secondary to T cell activation. However, an important role for B cells in T1D development was discovered when NOD mice made deficient in this lymphocyte population, either by introducing a functionally inactivated Igμ H chain gene (termed NOD.Igμ^null) or treatment with IgM-specific Abs, were strongly protected from disease (4, 5). One study has described a human T1D patient that was deficient in B cells (6). Although of interest, this report does not universally exclude a role for B cells in T1D development in humans because it is likely that different patient populations may have varying sets of contributory factors leading to disease. Hence, while their role may be bypassed in some cases, B cells may still significantly contribute to T1D in a large proportion of human patients.

Although the mechanism is not clear, maternal transmission of autoantibodies appears to have an early diabetogenic role in NOD mice (7). However, susceptibility of NOD.Igμ^null mice to T1D could only be restored by reconstitution with NOD B cells, and not by infusion of autoantibodies, revealing another pathogenic mechanism for this lymphocyte population (8). Successive studies revealed that NOD B cells also make an important contribution to T1D as a subset of APC with a preferential ability to expand autoreactive CD4⁺ T cell responses (8, 9, 10, 11). Such a role was attributed to their unique ability to specifically capture, through membrane-bound Ig molecules (or BCR), β cell autoantigens for subsequent MHC class II-mediated presentation (12, 13). An implication of this finding is that in addition to defects underlying the generation of autoreactive T cells, another essential component of T1D susceptibility may include disruptions in mechanisms that normally prevent the development or activation of B cells expressing self-reactive BCR. This was investigated by comparative analyses of transgenic stocks of NOD and nonautoimmune-prone C57BL/6 (B6) mice in which virtually all B cells express a BCR specific for hen egg lysozyme (IgHEL), and develop in an environment where this protein was absent or expressed as a nominal membrane-bound or soluble neo-self-Ag (sHEL) (14). Immature IgHEL B cells of NOD and B6 origin did not differ in their ability to be deleted upon high-avidity engagement of membrane-bound HEL molecules. However, immature IgHEL B cells in NOD mice were found to have a significantly lower ability than those of B6 origin to be deleted or anergized following low-avidity Ig engagement of sHEL (14).

Multiple susceptibility (Idd) genes located both within and outside particular MHC haplotypes interactively contribute to T1D in both humans and mice (15, 16). Given their important role in T1D development, it seemed likely that the pathogenic action of at least some Idd genes in NOD mice (and potentially humans) is to allow the development and/or functional activation of autoreactive B cells. Indeed, the mechanism by which the nonexpressed H2-Ea allele within the NOD H2^g7 MHC haplotype contributes to T1D development partly entails a B cell component (17). In this study, we investigated whether specific non-MHC Idd genes also regulate the diabetogenic potential of B cells in NOD mice. We found genes on proximal chromosome (Chr.) 1 (Idd5) and distal Chr. 4 (Idd9/11) both contribute to T1D by affecting B cells in NOD mice, potentially by regulating responsiveness to activation and tolerance induction stimuli.

Mice were housed at The Jackson Laboratory under specific pathogen-free conditions with free access to food (NIH31A/6% fat diet; Ralston Purina) and acidified water or under similar conditions at the Garvan Institute Animal Facility (Sydney, Australia). Female mice were used in all experiments. Derivation of NOD.Igμ^null (currently at N10; Ref. 5), nonobese-resistant (NOR)/Lt (18), NOD.Idd5^B10 (line R444 at N13; Ref. 19) and R467 at N >10 (20), NOD.Idd13^NOR (line D2Mit490-Mit144^NOR at N21; Ref. 21), NOD.Idd5^NOR and NOD.Idd9/11^NOR mouse lines (N5; Ref. 22) are previously described. Three additional backcrosses of NOD.Idd5^B10 line R444 to NOD/Lt were performed to produce subcongenic stocks containing B10-derived genomic regions from D1Mit74 to 303 and D1Mit249 to 132, which were then fixed to homozygosity. The R2s Idd5^B10 subcongenic strain was produced by additional backcrossing of the previously described R2 line (20).

NOD/Lt, B6, and C57BL/10 (B10) mice used for flow cytometric analysis and in vitro experiments were purchased from the Animal Resources Centre (Perth, Australia) and NOR/Lt mice were obtained from the Walter Eliza Hall Institute (Melbourne, Australia). NOD and B6 mice carrying the IgHEL and sHEL transgenes were previously described (14, 23, 24). IgHEL and sHEL transgenes were each introduced into NOD.Idd5^B10 (R444) and NOD.Idd9/11^NOR congenic mice by intercrossing each transgenic with congenic mice, followed by backcrosses to the congenic strains to fix the B10 or NOR-derived regions to homozygosity. NOD, B6, and congenic background stocks hemizygous for the IgHEL or sHEL transgenes were intercrossed to produce IgHEL/sHEL double-transgenic mice. All experimental procedures using mice were approved by The Jackson Laboratory’s Animal Care and Use Committee or the Garvan Institute Animal Ethics Committee.

Bone marrow (BM), blood, peritoneal cavity, lymph node (LN), or splenic leukocyte suspensions from the indicated mice were assessed for proportions of B cell subsets by multicolor flow cytometry using the FACSCalibur instrument with CellQuest acquisition software (BD Biosciences), and previously described methods and reagents (14). Analysis of flow cytometric data was performed using FlowJo software (Tree Star).

Cohorts of female NOD.Igμ^null mice were lethally irradiated in two 600R split doses from a ¹³⁷Cs source between 4 and 6 wk of age, and then reconstituted by i.v. injection with 5 × 10⁶ T cell-depleted syngeneic BM (SBM) cells admixed with 5 × 10⁶ purified splenic B cells from indicated strains. Splenic B cells were purified using the previously described (8) MACS-negative depletion system (Miltenyi Biotec) which routinely achieved >90% purity. Control chimeras consisted of NOD.Igμ^null female mice reconstituted with 5 × 10⁶ SBM cells only. BM/B cell recipients were then monitored weekly between 8 and 21 wk postreconstitution for T1D development (see below). Upon diabetes onset or at the end of incidence study (21 wk postreconstitution), spleens from recipient mice were assessed by flow cytometry to determine percentage reconstitution of B cells, CD4⁺ and CD8⁺ T cells using previously published methods and reagents (8).

T1D development was assessed by measuring glycosuric values with Ames Diastix (Diagnostics Division, Bayer). Values of >3 were considered indicative of T1D onset. Statistical differences in T1D onset between different experimental groups were tested using Kaplan-Meier life table analyses.

B cells were purified from pooled spleens or LN of indicated strains as described above. Proliferative responses of triplicate aliquots of 1 × 10⁵ B cells stimulated with 1 μg/ml AffiniPure goat anti-mouse IgM F(ab′)₂ (Jackson ImmunoResearch Laboratories) in the presence or absence of the CD40-specific mAb HM40-3 (BD Biosciences) at a concentration of 5 μg/ml were assessed as previously described (14). Control wells contained no stimulatory agents. Proliferative responses are presented as mean Δcpm ± SEM for triplicate wells of each sample.

Transitional type-1 (T1) B cells were enriched from pooled RBC-depleted spleens of indicated mice by performing a streptavidin-microbead (Miltenyi Biotec) magnetic depletion of T cells, monocytes/granulocytes, and T2/follicular (FO) B cells stained with biotinylated mAbs specific for CD3 (145-2C11), Mac-1 (M1/70), and CD23 (B3B4), respectively (BD Biosciences). Recovered cells were subsequently stained with mAbs specific for B220 (RA3-6B2), CD21/35 (7G6), and CD23 (B3B4) conjugated to allophycocyanin, FITC, and PE (BD Biosciences), respectively, in addition to treatment with propidium iodide (PI; Sigma-Aldrich) to exclude dead cells. FACSAria flow cytometer (BD Biosciences) was used to sort viable T1 B cells, defined as B220⁺, CD21/35⁻, CD23⁻, and PI⁻, to >99% purity. To perform the deletion assays, triplicate aliquots of 1 × 10⁵ cells were seeded into 96-well plates in 200 μl of complete RPMI 1640 medium with varying concentrations of AffiniPure goat anti-mouse IgM F(ab′)₂. Cell were incubated for 24 h at 37°C in a 5% CO₂ incubator and then harvested and stained with Annexin V-FITC (BD Biosciences) and PI in the presence of annexin-friendly buffer (10 mM HEPES (pH 7.4), 140 nM NaCl, 2% dialyzed FBS, 0.02% NaN₃). Data are presented as the percentage of T1 B cells that remained viable (annexin V⁻, PI⁻) in each stimulated culture compared with nonstimulated control cultures.

We hypothesized that some non-MHC Idd genes functionally contribute to T1D in NOD mice by allowing the development of autoreactive B cells. The closely NOD-related, but T1D-resistant, NOR strain proved to be a valuable resource for testing this hypothesis. NOR shares a large proportion of its genome (∼88%) with NOD mice, including the H2^g7 MHC haplotype (25), with the balance derived from the C57BLKS/J (BKS) strain, which itself is of a mixed B6 and DBA/2-like origin (25). BKS-derived genomic regions on Chr. 1, 2, and 4 are mainly responsible for the strong T1D resistance of NOR mice (21, 22, 25). We hypothesized that some subset of these NOR non-MHC Idd-resistance loci may elicit T1D protection by diminishing development of pathogenic B cells. This possibility was tested by determining whether repopulation with B cells from NOD vs NOR donors differentially abrogated the normal T1D resistance of NOD.Igμ^null recipients. Lethally irradiated NOD.Igμ^null mice were reconstituted with SBM admixed with purified NOD or NOR B cells. We previously found this approach overcomes the difficulty that unmanipulated NOD.Igμ^null mice are not tolerant of directly infused B cells, as evidenced by their ability to rapidly reject them through a CD8⁺ T cell response (8). Over a 21-wk observation period, T1D developed at a significantly lower frequency in NOD.Igμ^null mice reconstituted with SBM plus B cells from NOR than NOD donors (30 vs 69.6%, respectively; Fig. 1 A). As discussed later, this was not due to differential repopulation by NOD vs NOR B cells. These results indicated certain non-MHC Idd genes in NOD mice are involved in facilitating the function or development of diabetogenic B cells.

The strongest T1D protective effect in NOR mice is encoded by genes located within two closely linked loci on distal Chr. 2, designated Idd13a and Idd13b (21, 25). Therefore, we compared T1D development in NOD.Igμ^null mice reconstituted with SBM and B cells from standard NOD donors or a stock congenic for a segment of NOR Chr. 2 encompassing both the Idd13a and Idd13b regions (NOD.Idd13^NOR) (21). Over 21 wk, both groups of recipients developed a similar high incidence of T1D (Fig. 1 B). Thus, while Idd13 region genes strongly contribute to T1D resistance in NOR mice, they do not function at the B cell level.

Genes on Chr. 1 and 4, which respectively reside within portions of the previously mapped Idd5 and Idd9/11 loci, also make significant contributions to T1D resistance in NOR mice (22, 25). Thus, we tested whether either of these loci dampen the diabetogenic capacity of NOR B cells. Lethally irradiated NOD.Igμ^null mice were reconstituted with SBM admixed with B cells from NOD stocks congenic for NOR-derived Idd5 or Idd9/11-resistance regions (termed NOD.Idd5^NOR and NOD.Idd9/11^NOR, respectively). Over a 21-wk follow-up period, T1D developed at a significantly lower frequency in NOD.Igμ^null mice repopulated with B cells from NOD.Idd9/11^NOR (31%) than NOD (60%) donors (Fig. 1,B). It should be noted that compared with those from NOD donors, the ability of B cells from NOD.Idd9/11^NOR and NOR mice to support T1D development were equally diminished. These results indicated allelic variants of a gene(s) within the Idd9/11 region on Chr. 4 contribute to T1D susceptibility or resistance by affecting the pathogenicity of B cells. In contrast, NOD.Idd5^NOR B cells elicited a rate of T1D development in NOD.Igμ^null recipients that was slightly lower, but not significantly different, than those from NOD donors (50 vs 60%; Fig. 1 C).

Studies using NOD mice harboring various overlapping B10-derived congenic regions have revealed Idd5-mediated T1D susceptibility results from a combination of at least three separate genes, designated Idd5.1, Idd5.2, and Idd5.3 (Fig. 2) (20) (L. S. Wicker, unpublished observations). As the NOR Idd5 region, which is of B6 origin, only overlaps with the previously described Idd5.2 region (Fig. 2), we wanted to determine whether the full complement of resistance alleles at the Idd5 locus could affect the diabetogenic activity of B cells. Irradiated NOD.Igμ^null mice were reconstituted with SBM and B cells from the NOD.Idd5^B10 R444 congenic line, which contains B10-derived alleles at Idd5.1, Idd5.2, and Idd5.3 (20). Unlike those from NOD.Idd5^NOR donors, NOD.Idd5^B10 B cells elicited a significantly lower frequency of T1D in NOD.Igμ^null recipients than those of standard NOD origin (27.7 vs 60.3%, respectively, Fig. 1,C). NOD.Igμ^null mice were also reconstituted with SBM and B cells from one of the four NOD.Idd5^B10 subcongenic lines illustrated in Fig. 2. B cells from none of the subcongenic lines, containing different combinations of B10-derived Idd5.1, Idd5.2, and Idd5.3 variants, could completely replicate the protective effect of the originally tested NOD.Idd5^B10 R444 line. B cells derived from the subcongenic line containing B10-derived genome between D1Mit249 and D1Mit132, which encompasses the Idd5.1, Idd5.2, and Idd5.3 genes according to recently described boundaries, came closest to the protection offered by R444 B cells (46.4 vs 27.7%, respectively; Fig. 2). The discrepancy in protection offered by B cells from these two Idd5 congenic lines suggests this effect could result from the presence of an additional polymorphic gene in the region of Chr. 1 where the two lines differ (i.e., between D1Mit74 and D1Mit249). However, the major conclusion from these collective results is that multiple Idd5 region genes interactively determine the extent of diabetogenic B cell development.

It was possible that quantitative differences in ability to reconstitute NOD.Igμ^null recipients correlated with the capacity of the various tested B cell populations to induce T1D. This was assessed by comparing B cell levels in diabetic and nondiabetic mice within each group (i.e., those reconstituted with one type of B cell) and also comparing reconstitution between different groups that had varying overall susceptibility to disease (Table I). In the majority of cases, we found no significant differences in B cell reconstitution between diabetic vs nondiabetic mice within a particular group. In the two cases, a difference was noted (NOD.Idd13^NOR and NOD.Idd5^B10 donor groups), B cell reconstitution levels were lower in diabetic than nondiabetic mice. Between groups, B cell reconstitution levels varied, but did not correlate with T1D incidence (r² = 0.0048).

B cell development in NOD mice has been reported to differ from that in nonautoimmune prone strains (14, 26, 27). Differences include a notable deficiency in immature T1 B cells and enlarged populations of transitional T2/premarginal zone and marginal zone (MZ) subpopulations. This altered state of B cell development may have important implications for T1D development because similar differences are observed in other autoimmune prone strains (28, 29, 30). Hence, we investigated whether differences in the levels of certain splenic B cell subpopulations of the various mouse strains used as donors (all 6-wk-old females) may underlie their variable diabetogenic capacity. Consistent with past reports, NOD mice had lower numbers of T1 B cells and higher numbers of pre-MZ/T2 and MZ subpopulations compared with nonautoimmune prone C57BL/10 mice (Fig. 3 and Table II). We were unable to differentiate between the T2 and T3 B cell subpopulations described by Allman et al. (31) because NOD mice appear to be deficient in expression of the AA4.1 marker (P. A. Silveira, unpublished observation and Ref. 32). Despite their decreased diabetogenic activity, NOR, NOD.Idd9/11^NOR, and NOD.Idd5^B10 splenic B cells demonstrated similar T1, T2/pre-MZ, FO, and MZ profiles as those from NOD mice (Fig. 3 and Table II). We also examined various other lymphoid organs and compartments, including BM, blood, peritoneal cavity, and pancreatic LN to determine whether genes within the Idd5^B10 and Idd9/11^NOR congenic regions could affect the development or migration of B cells at these sites. No such changes were found (data not shown). Together, these results indicate T1D-resistance alleles within the Idd5^B10 and Idd9/11^NOR congenic regions do not decrease B cell pathogenicity by altering their subset distribution patterns from that seen in nonautoimmune prone NOD mice.

Mature B cells in NOD mice reportedly respond more robustly to several types of stimuli compared with those from nonautoimmune prone strains such as B6 and BALB/c (33, 34). This hyperresponsiveness may constitute a mechanism that enhances the ability of NOD B cells to contribute to T1D. Hence, we assessed whether Idd9/11 or Idd5 loci regulate B cell responsiveness. The ability of B cells from NOD, NOD.Idd9/11^NOR, NOD.Idd5^B10, and B10 mice to proliferate upon IgM cross-linking with or without CD40 costimulation (mimicking Ag encounter with or without T cell help) was compared (Fig. 4,A). To ensure any differential responses were not be due to subset variations, we used B cells from pooled inguinal, axillary, cervical, and mesenteric LN which consist primarily of the FO subset. We deliberately did not take pancreatic LN as this is believed to be the primary site of lymphoid activation preceding T1D (35, 36), and thus may differ in levels of activated B cells between each strain depending on their disease susceptibility. As previously observed, NOD B cells proliferated significantly more than those from B10 mice under both stimulation conditions. NOD.Idd5^B10 B cells showed a similar degree of proliferation as those from standard NOD mice after anti-IgM treatment with or without CD40 stimulation, indicating genes in the congenic region do not affect this phenotype. Surprisingly, NOD.Idd9/11^NOR B cells proliferated significantly more than those from NOD mice in response to both sets of stimuli. Thus, we tested whether allelic variants within the Idd9/11^NOR region allowed similar superresponsiveness by NOR B cells. To our surprise, the responsiveness of NOR B cells to both forms of stimulation was not only greater than those from NOD, but also the NOD.Idd9/11^NOR strain (Fig. 4 B). Therefore, while a Idd9/11 region gene(s) contributes to the ability of NOR B cells to respond more vigorously to proliferation stimuli than those from NOD mice, other factors also control this phenotype.

In nonautoimmune prone mice, immature B cells in the BM or at the T1 developmental stage in the spleen are susceptible to apoptosis induced by Ag engagement of the BCR. We previously showed both of these subpopulations in NOD mice have a higher threshold for deletion upon BCR cross-linking than those from nonautoimmune prone B6 mice (14). This aberrant phenotype could contribute to the increased numbers of mature self-reactive diabetogenic B cells in NOD mice. We therefore investigated whether genes within Idd5 or Idd9/11 regions contribute to the higher deletion resistance of immature B cells in NOD mice. Given they are the easier population to obtain at levels sufficient to allow experimental evaluation, we compared the sensitivity of purified splenic T1 B cells from wild-type NOD, NOD.Idd5^B10, NOD.Idd9/11^NOR, and B6 mice to apoptotic death by culturing them for 20 h with various concentrations of anti-IgM F(ab′)₂ (Fig. 5). Consistent with our previous observation (14), more NOD than B6 T1 B cells remained viable after BCR stimulation. After BCR stimulation, both NOD.Idd5^B10 and NOD.Idd9/11^NOR T1 B cells survived to a similar extent as those from standard NOD mice. This finding indicated that genes within the Idd5 or Idd9/11 region do not contribute to the decreased diabetogenic capacity of B cells by restoring their ability to be deleted at an immature stage of development upon BCR engagement.

Immature IgHEL-expressing B cells in B6 mice that do not undergo deletion following low-avidity engagement with the neo-self-Ag sHEL are instead rendered functionally anergic (37). Such anergic B cells are normally short-lived and remain unresponsive (37), even when provided with BCR and CD40 activating signals (14). In contrast, identical transgenes on the NOD background resulted in B cells being induced into only a weak anergic state, that was readily reversed by BCR and CD40 stimulation (14). Hence, we investigated whether genes within the Idd5 or Idd9/11 regions contribute to impaired B cell anergy in NOD mice. B6, NOD, NOD.Idd5^B10, and NOD.Idd9/11^NOR genetic background mice hemizygous for the IgHEL and sHEL transgenes were intercrossed. Our previous study had shown that in B6 background mice, total numbers of HEL-specific B cells in the spleen were decreased by ∼30% when they developed in the presence vs the absence of their cognate Ag expressed as a soluble self-protein (14). In contrast, the presence of sHEL as a soluble self-Ag did not influence the number of HEL-specific B cells found in the spleens of NOD background mice (14). Congenic introduction of homozygous Idd5^B10 or Idd9/11^NOR-resistance loci to NOD background mice did not result in an enhanced ability of sHEL expressed as a soluble self-Ag to numerically limit the development of IgHEL-expressing B cells (data not shown). Nevertheless, compared with those maturing in the absence of soluble self-Ag, IgHEL-expressing B cells exposed to sHEL during their development in each of the B6, NOD, NOD.Idd5^B10, and NOD.Idd9/11^NOR genetic background stocks down-regulated IgM on their surface, were restricted from entering the MZ and secreted little specific Ab spontaneously (data not shown).

Our previous work (14), as well as the results described above, indicated that in NOD genetic background mice a state of anergy is induced in IgHEL-expressing B cells maturing in the presence of soluble self-Ag. However, unlike the case in B6 background controls, the anergic state of IgHEL-expressing B cells in NOD mice that had matured in the presence of soluble self-Ag is readily reversed by stimulation through the Ig and CD40 receptors (14). Thus, we determined whether congenic introduction of the Idd5^B10 or Idd9/11^NOR-resistance loci to NOD background mice resulted in a restored ability of IgHEL-expressing B cells maturing in the presence of soluble self-Ag to be anergized in a nonreversible fashion. B cells were purified from the spleens of IgHEL single-transgenic and IgHEL/sHEL double-transgenic progeny from NOD, B6, NOD.Idd5^B10, and NOD.Idd9^NOR genetic background mice, and stimulated with IgM and CD40 cross-linking Abs (Fig. 6). As expected, IgHEL B cells from B6 control mice maturing in the presence sHEL were properly anergized as evidenced by lower levels of anti-IgM and CD40 stimulated proliferation than those that developed in the absence of cognate soluble self-Ag (Fig. 6,A). Also as previously observed, IgHEL-expressing B cells from NOD mice maturing in the presence of sHEL were not properly anergized because they proliferated equivalently in response to anti-IgM and CD40 stimulation as those that did not encounter soluble self-Ag during their development (Fig. 6). Significantly, in contrast to NOD, but similar to B6 background mice, IgHEL-expressing B cells from both the NOD.Idd5^B10 and NOD.Idd9/11^NOR stocks did anergize properly when maturing in the presence of soluble self-Ag (Fig. 6, B and C). Restoration of proper anergy induction was irrespective of the baseline level of B cell responsiveness in the NOD.Idd5^B10 and NOD.Idd9/11^NOR stocks. This was demonstrated by the finding that while IgHEL-expressing B cells from NOD.Idd9/11^NOR mice that matured in the absence of sHEL were more responsive to anti-IgM and CD40 stimulation than those of standard NOD origin, when developing in the presence of cognate soluble self-Ag they were still functionally repressed. These collective results indicate one mechanism whereby genes within the Idd5 and Idd9/11 regions contribute to T1D development in NOD mice is by inducing defects in the ability of B cells that recognize soluble self-Ags to be properly anergized.

This study demonstrated the development of autoreactive B cells, which serve as a critical subset of APC in T cell-mediated autoimmune T1D in NOD mice, is at least partly a consequence of pathogenic functions mediated by genes within the Idd5 and Idd9/11 disease susceptibility loci on Chr. 1 and 4, respectively. Specifically, mechanisms normally enabling B cells to be rendered functionally anergic following low-avidity Ig engagement of soluble self-Ag are impaired in NOD mice by the activity of genes within the Idd5 and Idd9/11 loci. Our experiments show clear differences in these tolerance-induction mechanisms using highly purified B cells from standard NOD mice and stocks congenic for Idd5 or Idd9/11-resistance variants. Hence, we are confident at least some polymorphic genes within the Idd5 and Idd9/11 loci act intrinsically through B cells to allow or prevent the development of functional autoreactive clonotypes. The association of these Idd susceptibility loci with the development of pathogenic B cells in NOD mice provides further affirmation of the important role this lymphocyte population plays in T1D development. It was also noted that while B cells of NOR, NOD.Idd5^B10, and NOD.Idd9/11^NOR origin supported an overall significantly lower rate of T1D development in NOD.Igμ^null recipients than those from NOD donors, in all these groups, the initial kinetics of disease onset were quite similar. We currently do not have an explanation for this phenomenon.

Enlarged numbers of splenic MZ B cells have been reported in NOD mice (14, 27, 38), which is of interest because this population is characterized by an ability to mount rapid responses to blood borne Ags and an enhanced capacity to act as APC for naive CD4 T cells (39). Furthermore, a link between MZ B cells and T1D development has been proposed due to evidence showing that: 1) insulin-specific B cells have an enhanced capacity to enter the MZ compartment (40); 2) depletion of MZ B cells protects NOD mice from disease (27); and 3) linkage analysis of an F₂ cross between NOD and B6 mice indicated a significant association between the Idd11 susceptibility locus and the enlarged MZ compartment (38). Despite this evidence, we found resistance alleles within Idd5^B10 or Idd9/11^NOR congenic regions did not decrease the diabetogenic capacity of NOD B cells by diminishing their entry into MZ compartment. Furthermore, our findings are consistent with a recent report showing NOD mice carrying a B6-derived Idd11 congenic region continue to have an enlarged MZ B cell compartment (41). The discrepancy in results between linkage analyses and congenic mice casts some doubt over the ability of Idd11 to regulate MZ B cell development. Nevertheless, our congenic stock data cannot rule out the possibility that the regulation of MZ B cells may require interactions between genes in the Idd9/11 region and elsewhere.

NOD B cells respond more robustly to several types of stimuli compared with those from the nonautoimmune prone B6 and BALB/c strains (34). Underlying NOD B cell hyperresponsiveness may be dysregulations of signaling pathways such as the heightened activation potential of the cardinal proinflammatory transcription factor NF-κB, and also increased expression of costimulatory molecules such as CD40, CD72, CD80, and CD86 (11, 33, 42). These molecular differences may lower the activation threshold for NOD B cells and enhance their capacity to interpret exposure to self-Ag in a proinflammatory context, thereby contributing further to the breakdown of tolerance in this strain. Although our results are consistent with the hyperresponsiveness of NOD B cells compared with those from nonautoimmune prone B10 and B6 mice, resistance alleles within the Idd5^B10 or Idd9/11^NOR regions did not dampen this phenotype. In contrast, NOR and NOD.Idd9/11^NOR B cells both proliferated more vigorously in response to activation stimuli than those of NOD origin. This indicates the superresponsiveness of NOR B cells is controlled by a gene(s) within the Idd9/11 locus. Strong responsiveness following Ig engagement is essential for activation-induced cell death (AICD), a tolerance mechanism that normally leads to the elimination of autoimmune B cells receiving chronic stimulation from self-Ags (43). Despite being hyperresponsive, NOD B cells have been previously reported to be more resistant to AICD compared with those from nonautoimmune prone B6 and BALB/c mice (34), perhaps contributing to enhanced development of self-reactive clones. Interestingly, based on analyses of an otherwise NOD genetic background stock, the ability of a congenically introduced NOR-derived gene(s) in the Idd9/11 region to enhance Ig-mediated signaling responses did not also result in an increased capacity of immature B cells to undergo AICD upon engagement of a self-Ag. The shared resistance of NOD and NOR B cells to AICD was also previously observed by Hussain et al. (34). This may be due to the fact that while differing at the Idd9/11 locus, NOD and NOR mice share many other T1D susceptibility genes which could contribute to the impaired ability of B cells from both strains to undergo AICD. However, while not affecting susceptibility to AICD, our data indicate that the ability of the NOR-derived Idd9/11-resistance locus to allow for enhanced Ig receptor-mediated responsiveness is likely to increase the ability of autoreactive B cells to be anergized following engagement of self-Ag.

The impaired ability of immature NOD B cells to undergo deletion upon Ig cross-linking is likely to result in increased numbers of self-reactive clonotypes that mature and serve as APC for diabetogenic CD4 T cells. However, our studies showed that allelic variants causing diabetes resistance within the Idd5 or Idd9/11 regions were both unable to correct the deletion resistance of immature B cells in NOD mice. The inability of Idd5 genes to control deletion of immature B cells was particularly surprising given previous reports have mapped a locus to this region in NOD mice that causes apoptosis resistance in total lymphocytes, mature T cells, and thymocytes (44, 45, 46). Although this difference may be indicative of distinct genetic control of apoptosis resistance in T and B cells of NOD mice, we cannot discount that genes within Idd5 may still be able to regulate the deletion of B cells upon interaction with other Idd genes. This may also be the case for T cells because the impaired deletion of thymocytes in NOD mice has been demonstrated to be the product of several Idd genes in addition to Idd5 (47, 48). Nevertheless, our data indicate that both the Idd5^B10 or Idd9/11^NOR regions independently decrease the development of diabetogenic B cells through the regulation of some mechanism other than the induction of deletion.

Anergy induction serves as another important mechanism for the maintenance of B cell tolerance to soluble Ags (49), such as the T1D-relevant autoantigen insulin (50). Using the IgHEL/sHEL transgenic model, we previously showed that unlike those from B6 controls, NOD B cells developing in the presence of cognate soluble self-Ag are maintained in only a very weak anergic state that can be readily reversed by BCR and CD40 stimulation (14). The aberrant ability of self-reactive B cells in NOD mice to overcome anergy after BCR and CD40 stimulation would seem relevant to their increased capacity to act as diabetogenic APC, because this defect should allow them to rapidly expand after initially encountering their Ag in the presence of CD40L positive β cell autoreactive CD4 T cells. A recent study by Acaveda-Suarez et al. (40) addressed whether similar anergy defects give rise to B cells specific for pancreatic β cell Ags in NOD mice. This was achieved by introducing an insulin-specific Ig transgene into NOD and B6 mice. In contrast to results in IgHEL/sHEL-transgenic mice, anergy induction was equivalent in NOD and B6 B cells expressing this particular insulin-reactive Ig transgene. However, this result was inconsistent with the fact that insulin-specific autoantibodies are detected in standard NOD but not B6 mice (51), implying nonanergic B cells specific for this β cell autoantigen are generated in the former strain. One possible reason for this contradiction might be the insulin-reactive B cells which remain functional in standard NOD mice bypass tolerance induction because the Ig they have a lower functional avidity for Ag than the transgenic variant studied by Acaveda-Suarez et al. (40). Our current study strongly implicates defective induction of B cell anergy as an important contributory factor to T1D development in NOD mice. This conclusion is supported by the finding that compared with those from standard NOD mice, B cells from NOD stocks congenic for the Idd9/11^NOR or Idd5^B10-resistance regions had a decreased ability to support T1D development in NOD.Igμ^null recipients, and also showed an enhanced ability to be anergized when maturing in the presence of cognate soluble self-Ag.

In future studies, B cell phenotypes controlled by the Idd5 and Idd9/11 loci will be used to aid in susceptibility gene discovery by both a candidate approach and microarray analyses. Understanding the underlying molecular and genetic aspects leading to the development of autoreactive B cells contributing to T1D may not only advance our understanding of disease pathogenesis, but also illuminate potential new targets for therapeutic intervention in humans.

The authors have no financial conflict of interest.