Deficiencies in Gut NK Cell Number and Function Precede Diabetes Onset in BB Rats¹

The etiopathogenesis of autoimmune diabetes mellitus is still unclear, but both genetic and environmental factors appear to participate. Two important environmental factors–diet and infectious agents–may participate in this process through mediation of the gut immune system. The gut harbors distinct lymphoid populations that respond to alternative activation signals to control the physiologic inflammation induced by huge intralumenal Ag loads (1). These lymphoid populations participate in mucosal processes like oral tolerance and controlled (or physiologic) chronic inflammation (1). Defects in regulation of the intestinal immune system are associated with celiac disease (2), inflammatory bowel disease (3, 4, 5), and possibly autoimmune diabetes (6). Lack of tolerance to dietary gluten has clearly been implicated in the immunopathogenesis of celiac disease (2), and several studies have documented a correlation between type 1 diabetes and celiac disease in humans (7, 8).

The role of intraepithelial lymphocytes (IELs)³ in immune regulation at mucosal surfaces remains incompletely understood, and the phenotype and function of IELs with characteristics of NK cells have only recently been described (9, 10, 11). Intraepithelial NK (IENK) cells are deficient in humans with celiac disease (9), but there is little or no information concerning the normal number or function of such cells. In mice, IENK cells kill intestinal epithelial cells in vitro in a perforin-dependent manner, an activity that is increased by IL-15 (10). It has been speculated that IENK cells activated by IL-15 may participate in mucosal immunity by killing infected intestinal epithelial cells (10). In the rat, IENK cells comprise a significant fraction of the total IEL compartment, particularly in young animals. The percentage of IENK cells ranges from ∼50% of total IELs at 4 wk of age to ∼17% at 16 wk of age and older (11). We have shown that rat IENK cells are different from splenic NK cells in two respects (11). First, IENK cells express high levels of ADP-ribosyltransferase 2 (ART2)⁴ and CD25, which are nearly undetectable on splenic NK cells. Expression of ART2 on peripheral rat T cells is a marker of regulatory capability (12, 13), as is expression of CD25 on a population of peripheral mouse T cells that can prevent diabetes in nonobese diabetic mice (14). Second, unlike splenic NK cells that secrete little or no IL-4 and secrete IFN-γ only in response to stimulation (15), a substantial fraction of unstimulated rat IENK cells spontaneously secrete IL-4 and/or IFN-γ (11). Taken together, these data suggest that IENK cells may contribute to immune regulation at the level of the gut epithelium.

Diabetes-prone BB (BBDP) rats are severely T cell lymphopenic and spontaneously develop autoimmune diabetes (16). Several lines of evidence implicate the gut immune system in the pathogenesis of this disorder. For example, hydrolyzed casein-containing diets reduce the incidence of spontaneous diabetes whereas both gluten-rich diets and orally administered insulin appear to accelerate the disorder (17). Interestingly, increased permeability of the gut has been described in BBDP rats (18) as it has in children with type I diabetes (19). Diabetes-resistant BB (BBDR) rats are nonlymphopenic and never become spontaneously diabetic when housed in a viral Ab-free environment (16). However, BBDR rats do become diabetic in response to several immunomodulatory perturbants; these include infection with Kilham rat virus, depletion of ART2.1⁺ regulatory T cells, and/or immune system activation with polyinosinic-polycytidylic acid (poly(I:C)) (16), now known to be a ligand for Toll-like receptor 3 (20). How defects in the gut immune system may contribute to diabetes pathogenesis in BB rats remains poorly understood.

We now report that, in comparison with normal WF rats, the total number of IELs is deficient before diabetes onset in both BBDP rats and in BBDR rats treated to induce the disease. In both cases, the deficiency in IELs was associated with a marked reduction in 1) the number of IENK cells, 2) NK-sensitive YAC-1 cell cytotoxicity, and 3) the number of cells in the intraepithelial compartment that spontaneously secreted IL-4 and/or IFN-γ. The data suggest that defects in gut immunity may predispose to systemic autoimmunity and diabetes in the BB rat.

BBDR/Wor and BBDP/Wor rats were obtained from BRM Incorporated (Worcester, MA). ACI, BN, Copenhagen, DA, F344, LEW, Sprague Dawley, and WF rats were obtained from Harlan Sprague Dawley (Indianapolis, IN). Animals of either sex were used in all experiments. All animals were certified to be serologically free of Sendai virus, pneumonia virus of mice, sialodacryoadenitis virus, rat corona virus, Kilham’s rat virus, H1 (Toolan’s virus), mouse poliovirus (GD7), Reo-3, Mycoplasma pulmonis, lymphocytic choriomeningitis virus, mouse adenovirus, Hantaan virus, and Encephalitozoon cuniculi. All animals were housed in a viral Ab-free facility until used and maintained in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Research Council, National Academy of Sciences, 1996) and the guidelines of the Institutional Animal Care and Use Committee of the University of Massachusetts Medical School.

Fluorochrome- or biotin-conjugated mAbs directed against γδTCR (clone V65), CD3 (clone G4.18), NKR-P1A (clone 10/78), and RT7.2 (clone HIS41) were obtained from BD PharMingen (San Diego, CA). Tissue culture supernatant containing the NDS58 rat anti-rat RT7.1 (IgG2b) was obtained from Serotec (Oxford, U.K.). Hybridoma cells secreting the DS4.23 rat anti-rat ART2.1 (IgG2b) and the isotype control B21-2 rat anti-mouse I-A^b,d (IgG2b) are maintained in our laboratory. The B21-2 hybridoma was obtained from American Type Culture Collection (Manassas, VA). mAbs prepared from culture supernatant of these hybridomas were prepared for use in flow cytometry as described (21). For measurement of IL-4-secreting cells by ELISPOT assay, paired mAbs (purified OX-81 and biotinylated B11-3) directed against IL-4 were obtained from BD PharMingen. For measurement of IFN-γ by ELISPOT, the capture mAb (clone DB-1) was obtained from BioSource International (Camarillo, CA) and the detection rabbit anti-rat IFN-γ polyclonal Ab was obtained from Torrey Pines Biolabs (La Jolla, CA) and biotinylated as described (22). Isotype control mouse IgG1, IgG2a, IgG2b, and IgG3, secondary Ab (anti-rat IgG2b), and alkaline phosphatase-conjugated streptavidin were purchased from BD PharMingen.

BBDR rats 25–28 days of age were treated with i.p. injections of DS4.23 anti-ART2.1 mAb (50 μg five times weekly) plus poly(I:C) (5 μg/100 g body weight three times weekly) as described (23, 24); BBDR rats treated in this way uniformly develop autoimmune diabetes (23, 24). Rats in the present study were treated for only 1 wk, less than the minimum time required for diabetes onset (25), and then killed for experimental analysis.

IELs were prepared from rat small intestine by Percoll density gradient centrifugation as described (26), with minor modifications. Briefly, rats were killed in an atmosphere of 100% CO₂, and the entire small intestine was immediately removed en bloc and gently flushed with cold RPMI 1640 medium (Life Technologies, Grand Island, NY) to remove lumenal contents. The intestine was then incubated in ice-cold RPMI 1640 medium for 60–120 min to delaminate epithelial cells and IELs from the basement membrane. Detached cells were recovered by flushing the intestine two times with a total of 50 ml of HEPES-buffered HBSS medium containing 1 mM DTT (Sigma-Aldrich) and 10% Fetalclone I (a defined neonatal bovine serum; HyClone Laboratories, Logan, UT) at 37°C. IELs were then purified by Percoll density gradient centrifugation, and viable cells exhibiting lymphoid morphology were quantified by the method of trypan blue using a hemocytometer.

Spleens and mesenteric lymph nodes were removed from rats killed in an atmosphere of 100% CO₂. Single cell suspensions were prepared by gentle extrusion through stainless steel sieves into medium consisting of cold RPMI 1640 medium supplemented with 5% Fetalclone I as described (26, 27). Spleen cell preparations were washed in medium and erythrocytes were lysed in a hypotonic buffer (0.15 M NH₄Cl, 1.0 mM KHCO₃, 0.1 mM Na₂EDTA, pH 7.4).

Cells for bone marrow transplantation were prepared from 6-wk-old donor rats as described (28). Bone marrow recipients 6 wk of age were treated with 7 Gy of gamma radiation (∼0.1 Gy/min) using a ¹³⁷Cs source (Gammacell 40; Atomic Energy of Canada, Ottawa, Ontario, Canada) and immediately injected with 25 × 10⁶ bone marrow cells via a lateral tail vein.

One, two, and three-color flow cytometric analyses were performed as described (11). Cells were analyzed using a FACScan instrument (BD Biosciences, Sunnyvale, CA). Analyses were gated on lymphoid cells, which were identified by forward and side light-scatter profiles. A minimum of 25,000 events was analyzed for each sample.

Cytotoxic activity of lymphoid populations was measured by ⁵¹Cr-release microcytotoxicity assay as described (11), with minor modifications. Briefly, YAC-1 target cells in growth phase were labeled with ⁵¹Cr as sodium chromate (150 μCi/10⁶ cells; New England Nuclear, Boston, MA), and 1.0 × 10^4 51Cr-labeled cells were added to each well of a 96-well microtiter plate. Unfractionated IELs were added at E:T cell ratios ranging from 10:1 to 50:1, and the plates were incubated for 4 h at 37°C in a humidified atmosphere of 95% air-5% CO₂. All assays were performed in triplicate and averaged. The calculated IENK to target cell ratio was determined by multiplying the IEL to target cell ratio for each data point by the percentage of NKR-P1A⁺CD3⁻ IELs present in an aliquot of the sample as determined by flow cytometry.

Total releasable radioactivity (“cpm_maximal”) was determined by incubating an aliquot of ⁵¹Cr-labeled target cells with 1% Triton X-100. After incubation, cells were pelleted by centrifugation, and 100-μl aliquots of supernatant were transferred to a separate microtiter plate containing 100 μl of Optiphase Supermix β-scintillation fluid (Wallac, Gaithersburg, MD) and counted (“cpm_test”) using a 1450 Microbeta Trilux instrument (Wallac). In all assays reported here, release of ⁵¹Cr in the absence of effector cells (“cpm_spontaneous”) was <15% of total ⁵¹Cr release. Specific cytotoxicity was calculated as a percentage using the raw cpm and the formula: specific lysis (%) = ((cpm_test − cpm_spontaneous)/(cpm_maximal − cpm_spontaneous)) × 100%.

Spontaneous production of IL-4 and IFN-γ by rat IELs was measured by ELISPOT assay as described (11). The optimum number of IELs per well was determined in preliminary experiments and ranged from 2.5 to 20 × 10⁴. The number of discrete spots visible in each well was counted using a light microscope at 2× power. Each cytokine assay was performed using four to six wells, with all wells but one assayed using the cytokine-specific developing Ab. The remaining well was assayed using the irrelevant biotinylated control Ab. For analysis, the number of spots observed in the control well for each assay was subtracted from the number of spots counted in each of the three to five test wells. The maximum number of spots observed in any control well was always ≤3. The adjusted number of spots present in each of the test wells was then averaged.

Parametric data are presented as arithmetic means ± 1 SD except in Fig. 3 where the SEM is used. Comparisons of three or more means used one-way ANOVA and the Scheffe or least squared difference procedure for a posteriori contrasts (29). In cases where variances were inhomogeneous, data were log transformed before analysis.

We first quantified the number of IELs recoverable from 4- to 6-wk-old nondiabetic WF, BBDR, and BBDP rats. The numbers recovered from unmanipulated BBDR rats and WF rats were statistically similar (Table I, groups 1 and 3, p = NS). The number of IELs recovered from BBDP rats was significantly less (Table I, group 2, p < 0.001); the same was true of the number of IELs recovered from BBDR rats that had been treated with anti-ART2.1 mAb plus poly(I:C) for 1 wk (Table I, group 4, p < 0.001). The numbers of IELs recovered from BBDP and treated BBDR rats were statistically similar (p = NS). One week represents approximately half the time required to induce autoimmune diabetes in BBDR animals (23, 24). None of the animals were diabetic, and differences in numbers of IELs among WF, BBDR, and BBDP rats were therefore not a consequence of diabetes per se.

The rat DS4.23 anti-rat ART2.1 mAb depletes most peripheral ART2.1⁺ T cells (24), and nearly 100% of IELs in the rat are ART2⁺ (26). However, as shown in Fig. 1, most residual IELs in BBDR rats treated with anti-ART2.1 mAb plus poly(I:C) for 1 wk continued to express high levels of ART2.1. By incubating freshly isolated IELs from treated BBDR rats with the secondary FITC-conjugated anti-rat mAb alone, we observed that some anti-ART2.1 mAb from the treatment protocol was residually bound to the surface of IELs recovered from these animals (Fig. 1,D). Consistent with previous reports (30), 1 wk of treatment appeared to deplete peripheral (lymph node) ART2.1⁺ lymphocytes completely; there was no evidence that residually bound Ab was masking the Ag on these cells (Fig. 1, A and C).

It has been reported that the number of intestinal γδ T lymphocytes is extremely low in BBDP rats (31). We next tested whether a selective deficiency in γδ T cells fully accounts for the reduced numbers of IELs observed in prediabetic BB rats. As shown in Table II, we confirmed that both the number and percentage of intraepithelial γδ T cells in BBDP rats (group 2) are decreased in comparison to WF rats (group 1) and untreated BBDR rats (group 3). Interestingly, the absolute number of γδ-TCR⁺ IELs in BBDR rats did not change following 1 wk of anti-ART2.1 mAb plus poly(I:C) treatment (group 4), and the percentage of these cells was actually increased in treated animals. These data suggest that a selective deficiency in an IEL subpopulation other than intraepithelial γδ T cells might better account for the reduced total IEL numbers seen in prediabetic BB rats.

NKR-P1A⁺CD3⁻ IENK cells are particularly interesting because these cells comprise a major fraction of the IEL compartment in normal WF rat gut and may have an immunoregulatory function (11). Consistent with previous observations (11, 26), ∼50% of IELs in 4- to 6-wk-old WF rats expressed the NKR-P1A⁺CD3⁻ IENK phenotype (Table II). Percentages of NKR-P1A⁺CD3⁻ IENK cells in five other rat strains with no known susceptibility to autoimmunity (F344, DA, Sprague Dawley, ACI, and Copenhagen rats) were 53.5 ± 2.5%, 45.1 ± 3.4%, 31.0 ± 2.5%, 28.4 ± 11.2%, and 22.0 ± 4.6%, respectively. In contrast, only ∼12% of IELs in age-matched BBDP rats expressed the IENK phenotype (p < 0.02 vs nonautoimmune prone strains), and the absolute number of recoverable IENK cells in these animals was only ∼8% of that observed in WF rats (Table II).

Both the percentage and number of IELs expressing the IENK phenotype in age-matched, untreated, nondiabetic BBDR rats were less than half that observed in WF controls (Table II). Unexpectedly, both the percentage and number of IENK cells in the IEL compartment of BBDR rats treated with anti-ART2.1 mAb plus poly(I:C) for 1 wk were reduced >90% in comparison to untreated control BBDR rats (Tables II and III).

We next sought to determine whether the dramatic reduction in the absolute number of NKR-P1A⁺CD3⁻ IENK cells observed in BBDR rats following 1 wk of treatment with anti-ART2.1 mAb plus poly(I:C) was restricted to this population or occurred in other lymphocyte populations as well. As shown in Table II, the number of intraepithelial γδ T lymphocytes is similar in untreated and treated BBDR rats. As shown in Table III, the numbers of splenic NKR-P1A⁺CD3⁻ NK cells in untreated and treated BBDR rats are also similar to each other (p = NS). In addition, the number of NKR-P1A⁻CD3⁺ intraepithelial T lymphocytes in BBDR rats was unchanged by 1 wk of anti-ART2.1 mAb plus poly(I:C) treatment, and NKR-P1A⁻CD3⁺ splenic T cells were reduced by ∼40% (Table III).

NKR-P1A⁺CD3⁺ NK-T cells comprise a relatively small fraction of the total IEL population in rats (11). Consistent with this observation, the percentage of intraepithelial NK T cells in 4- to 6-wk-old untreated animals was 5.2 ± 0.7% in WF rats (n = 4), 2.6 ± 0.8% in BBDR rats (n = 12, p < 0.001), and 1.4 ± 0.2% in BBDP rats (n = 4, p < 0.001 vs both WF and BBDR). These differences in percentage were paralleled by differences in the absolute number of NKT cells present in the IEL preparations, but these numbers were all very low, ranging from 330 × 10³ cells in WF rats to 210 × 10³ in BBDR rats to 30 × 10³ in BBDP rats.

We next tested the hypothesis that the residual population of IENK cells in prediabetic BB rats would exhibit reduced functionality. To do so, we measured specific lysis of NK-sensitive YAC-1 cells incubated with unfractionated IELs freshly isolated from 6-wk-old WF, BBDR, and BBDP rats. Consistent with previous reports (11), IELs from WF rats exhibited efficient YAC-1 cell killing (Fig. 2 A). YAC-1 specific lysis was consistently less in IELs from BBDP rats, untreated BBDR rats, and BBDR rats treated for 1 wk with anti-ART2.1 mAb plus poly(I:C).

Essentially all YAC-1 cell killing activity in the IEL compartment of rats is associated with the IENK cell subpopulation (11). Using flow cytometry, we determined that NKR-P1A⁺CD3⁻ IENK cells comprised 36% of IELs in WF rats, 11% in BBDP rats, 14% in untreated BBDR rats, and 2% in treated BBDR rats. With this information, we reanalyzed the data in Fig. 2,A using calculated IENK:target cell ratios based on these percentages (see Materials and Methods). The analysis (Fig. 2 B) suggests that, on a per cell basis, the YAC-1 cell killing activity of IENK cells from WF rats is high, that of BBDP rat IENK cells is low, and that of IENK cells from untreated and treated BBDR rats is intermediate.

Unstimulated rat IELs spontaneously produce IL-4 and/or IFN-γ, an activity entirely attributable to the NKR-P1A⁺CD3⁻ IENK cell subpopulation (11). We next tested the hypothesis that the residual population of IENK cells in prediabetic BB rats would exhibit reduced cytokine production. To do so, we performed ELISPOT analyses. As shown in Fig. 3 (upper panel), we confirmed that unstimulated IELs isolated from 4- to 6-wk-old WF rats spontaneously produce both IL-4 and/or IFN-γ. In contrast, very few IELs from age-matched BBDP rats or untreated BBDR rats produced either cytokine. Treatment of BBDR rats with anti-ART2.1 mAb plus poly(I:C) further reduced the number of IL-4 and/or IFN-γ-producing IELs in these animals.

Because these data were obtained from total IEL populations, we also measured the percentage of IENK cells present in the IEL populations tested (Fig. 3, lower panel). Inspection of the two panels of the figure shows that, as would be expected from previous reports (11), the number of cells producing IL-4 and/or IFN-γ correlated with the percentage of IENK cells present.

To determine whether the apparent deficiencies in IENK cell numbers and cytokine production are unique to BB rats, we studied four other rat strains. IELs were isolated from 4- to 6-wk-old F344, BN, DA, and LEW rats and analyzed for both the percentage of IENK cells present and the number of cytokine-producing cells (Fig. 3). WF, F344, and DA rats were similar with respect to the percentage of IENK cells (45–54%) and the relatively high number of IELs that spontaneously produced IL-4 (47–119 spots/10⁵ IELs) and/or IFN-γ (33–94 spots/10⁵ IELs). Both the percentage of IENK cells and the number of cytokine-producing IELs in these three strains were much greater than in any of the BB rat groups. Fig. 3 also shows that BN and LEW rats were similar to each other with respect to the number of IELs producing IL-4 and/or IFN-γ. These numbers were much lower than in WF, F344, and DA rats and comparable to the numbers observed in BB rats. The percentage of LEW IELs that were IENK cells (18%) was similar to that in untreated BBDR rats (17%), whereas the percentage of BN IELs that were IENK cells (33%) was higher, but still not as high as that observed in WF, F344, or DA rats.

We next sought to determine whether reduced IENK cell number and function in BBDP rats results from defects in hemopoietic progenitors or defects in the microenvironment of the host. To do so, we performed reciprocal bone marrow transplantations between 6-wk-old BBDP and WF rats and analyzed the animals 6 wk later. It is already known that transplantation of WF bone marrow prevents diabetes in BBDP rats (32). The RT7 allotypic marker was used to distinguish donor-origin cells from residual recipient cells. We have previously reported that the percentage of NKR-P1A⁺ cells present in the WF rat IEL compartment declines with age, decreasing from ∼50% in 4-wk-old animals to ∼17% in animals 16 wk of age and older (11). In the present study, we observed that WF bone marrow appeared to generate an IEL compartment in BBDP rats similar to that which would be expected (11) for a 12-wk old WF rat with respect to both the number of IELs and the percentage of IENK cells (Table IV, group 5). The number and percentage were comparable to those observed in both irradiated WF rats transplanted with syngeneic WF bone marrow (group 2) and in unmanipulated 12-wk-old control WF rats (group 1). In contrast, transplantation of BBDP bone marrow into irradiated WF recipients failed to generate a normal IEL compartment of BB rat origin (group 6). Although the number of IELs recovered was high, most of these IELs were of WF origin (RT7.2⁺), and the percentage of IENK cells was similar to that in control BBDP rats (Table IV, groups 3 and 4).

These data document major deficiencies of intraepithelial NK cell number and function before the onset of diabetes in spontaneously diabetic BBDP rats and in BBDR rats that are treated to induce the disease. A less pronounced but nonetheless significant reduction in IENK cell number and function was also observed in unmanipulated BBDR rats but not in WF rats. Differences in IENK cell number and function among BBDP, BBDR, and WF rats were not a consequence of diabetes per se because none of the rats were diabetic when studied.

In the BBDP rat, autoimmune diabetes occurs spontaneously but the animals are T cell lymphopenic due to a mutation in the Ian4L1 gene (33, 34). Our data reveal that the lymphopenia observed in the periphery of BBDP rats extends to the IEL compartment in general and to the IENK subpopulation in particular. Our analysis of bone marrow chimeras demonstrates that the BBDP rat’s deficiency in IELs and IENK cells, like its deficiency in peripheral T cells (35), is intrinsic to hemopoietic stem cells.

However, lymphopenia is not a characteristic of human autoimmune diabetes, calling into question the suitability of the BBDP rat as a model for this disease. For this reason, we performed parallel analyses in the BBDR rat, which does not share the lymphopenia mutation of the BBDP rat but remains susceptible to the induction of diabetes as a function of other shared genetic loci (25, 36, 37). We document that these animals have normal numbers of IELs, but even before treatment to induce diabetes, they appear to be deficient in IENK cell number and function. YAC-1 cytotoxic activity, IL-4 production, and IFN-γ production in BBDR rat IELs are lower than in normal WF rat IELs, although higher than that in BBDP rat IELs. Based on our previous analyses of the IEL compartment of the rat (11), it is likely that these deficiencies are largely, if not entirely, due to deficiencies in the IENK subpopulation of IELs. Treatment to induce diabetes in BBDR rats using the well-characterized combination of anti-ART2.1 mAb plus poly(I:C) to perturb the immune system (16) was associated with further reductions in IENK cell number, cytotoxic activity, and cytokine production. Interestingly, the treatment protocol appeared to reduce NK cells in the IEL compartment but not the spleen. These data are all consistent with the hypothesis that abnormal IENK cell activity may be one of the underlying immunological defects in BB rats.

The constellation of NK-like phenotype and function, together with an unusual cytokine profile–constitutive expression of IL-4 and IFN-γ in the local microenvironment of the gut–indicate that intraepithelial NK cells are specialized cells that may subserve gut-specific functions. Given the present data, it is plausible to speculate that ART2⁺CD25⁺ rat IENK cells could play a role in the processes that regulate both the “physiologic inflammation” that is characteristic of the gut generally (1) and the genetic predisposition to autoimmunity in the BB rat specifically. Consistent with this hypothesis, we have also documented the existence of relatively deficient IENK cell number and cytokine production in two unrelated rat strains, LEW and BN, which are also susceptible to autoimmunity (BN, susceptible to HgCl₂-induced nephritis (38), and LEW, susceptible to experimental allergic encephalomyelitis (39)). These deficiencies were not observed in three strains, WF, F344, and DA, which have little or no predisposition to autoimmunity. The specific genetic and/or environmental factors that contribute to the deficiency of IENK cell number and function in rats susceptible to autoimmunity are unknown.

BBDP rats are reportedly deficient in intraepithelial γδ TCR⁺ lymphocytes (31), but the relationship of this deficiency to diabetes pathogenesis is unclear. We have confirmed the earlier observation in BBDP rats, but additional data from the BBDR rat suggest that a deficiency in the γδ TCR⁺ IEL subpopulation is not associated with diabetes expression. BBDR rats treated to induce diabetes have similar absolute numbers of γδ TCR⁺ IELs as do untreated animals, suggesting that the prediabetic state in BBDR rats is not characterized by reduced numbers of γδ TCR⁺ IELs. That the percentage but not the absolute number of IELs expressing the γδ TCR in treated BBDR rats is increased likely results from the selective loss of IENK but not other IEL populations.

Based on the data presented here and the well-recognized polygenic nature of type 1 diabetes in both humans (40) and BB rats (25, 36), we hypothesize that diabetes in the BB rat is a multistep process. The process begins with a genetically determined bone marrow-derived defect in regulatory gut IENK cells. The ensuing defect in immunoregulation at the level of the intestinal mucosa leads to abnormalities in gut-draining mesenteric lymph node cells, an inference supported by preliminary studies in which we have documented that cells capable of the adoptive transfer of autoimmune diabetes in BBDR rats appear in mesenteric lymph nodes before their appearance elsewhere in the periphery (R. Bortell, unpublished observations). Finally, in animals like the BB rat, in which additional susceptibility genes are present, the abnormal mesenteric lymph node populations proceed to induce insulitis and β cell destruction. This hypothesis, if proven correct, could help explain the linkage of environmental perturbation in the gut with abnormal peripheral immune responses against β cells in BB rats (16), nonobese diabetic mice (41), and humans (42) with autoimmune diabetes.

We thank Lisa Burzenski, Linda Leehy, Kelly Lake, and Michael Bates for technical assistance.