The major predisposing genetic component in type 1 diabetes maps to the major histocompatibility complex locus in both mice and humans. To verify the HLA class II association with disease pathogenesis, we adopted the transgenic approach. Expression of HLA-DQ8, the molecule showing the strongest association with human type 1 diabetes, in the diabetes-predisposing milieu of NOD mice in the absence of the endogenous class II molecule I-Ag7 did not render susceptibility to type 1 diabetes. To study if providing a local proinflammatory environment would lead to diabetes in these mice, Aβo.NOD.DQ8 were bred with C57BL/6 mice expressing tumor necrosis factor (TNF)-α in the β-cells of the islets of Langerhans. Surprisingly, although diabetes was evident in the F1 intercross expressing rat insulin promoter (RIP)-TNF, offspring lacking either endogenous or transgenic class II molecules developed accelerated diabetes with high frequency in both sexes. Moreover, expression of any functional class II molecule seemed to confer significant protection from diabetes in this model. Thus, neonatal expression of TNF-α in an islet-specific manner bypassed the requirement of CD4+ T-cells and resulted in diabetes that could be mediated by CD8+ T-cells. We also show for the first time that diabetes in NOD.RIP-TNF mice can occur independent of inheritance of NOD-derived idd1.

Type 1 diabetes is an organ-specific autoimmune disease resulting from the interaction of several genetic elements and yet undetermined environmental factors. It is characterized by a predominantly T-cell-mediated destruction of insulin-producing β-cells of the islets of Langerhans culminating in life-long insulin dependence. More than 18 genetic loci are implicated in the etiopathogenesis of type 1 diabetes, of which idd1 (in mice) or IDDM1 (in humans), which encompass the major histocompatibility complex (MHC) gene complex, has consistently shown the strongest association (1). A significant majority of patients with type 1 diabetes express certain HLA class II molecules, such as HLA-DQ8/DR4 or DQ2/DR3 (2). However, the precise roles played by these MHC class II molecules in the pathogenesis of type 1 diabetes still remain speculative. Promiscuous peptide presentation by the disease-associated class II molecules resulting in improper negative selection of autoreactive T-cells remains as the favored hypothesis (3). A significant influence of class I molecules is also suspected because CD8+ T-cells also play a crucial role in the disease pathogenesis (4).

Much of the understanding of the immunopathogenesis of several autoimmune diseases, including diabetes, has been feasible only because of the availability of suitable animal models (5). Of the two models of spontaneous type 1 diabetes, NOD mice and BB rats, the former is widely used where the disease occurs with increased frequency and disease onset is earlier in females (6). The strong structural (7) and functional (8) similarity between the NOD class II molecule, H2-Ag7, and its human counterpart, HLA-DQ8, suggest a common etiological role. Spontaneous loss of tolerance to GAD65, one of the major autoantigens implicated in type 1 diabetes, in HLA-DQ8 transgenic mice on a diabetes-resistant B10 background lends support to this hypothesis (9,10). Although diabetes was seldom seen in these mice, mild insulitis in a small percentage of islets was a consistent finding (9). Absence of disease per se in the face of demonstrable autoreactivity in HLA-DQ8 transgenic mice indicated the requirement of additional factors/stimuli. Islet-specific expression of the costimulatory molecule B7 happens to be one such factor/stimulus, which can drive the nonpathogenic autoreactive T-cells to the pathogenic pathway in these mice (11,12). This effect was HLA-DQ8 specific because rat insulin promoter (RIP)-B7 mice transgenic for HLA-DQ6 were diabetes free (11). Even though HLA-DQ8 exhibited such strong pro-diabetogenic properties, transgenic expression of DQ8 in the diabetes-predisposing milieu of NOD mice lacking endogenous class II (I-Ag7) did not confer susceptibility to type 1 diabetes (13). Nevertheless, NOD mice transgenic for DQ8 were susceptible to spontaneous disease in the presence of I-Ag7, indicating that DQ8 is permissible to diabetes (14).

Similar to the situation where provision of local costimulation with transgenic B7 resulted in diabetes in mice expressing HLA-DQ8 (11), we envisaged that islet-specific expression of the pro-inflammatory cytokine tumor necrosis factor (TNF)-α would also facilitate the study of HLA-DQ8 association with type 1 diabetes. In this context, it is important to know that even though neonatal expression of TNF-α under the control of RIP resulted in extensive insulitis in diabetes-resistant B6 mice (15), it led to diabetes only in NOD mice (16), indicating that only the NOD genetic background is permissible for induction of diabetes by RIP-TNF (17). Therefore, in the present study, HLA-DQ8 transgenic mice on NOD background (Aβo.NOD.DQ8) were mated with the diabetes-resistant RIP-TNF mice on the B6 background.

Generation of HLA transgenic RIP-TNF mice.

C57BL/6 mice expressing TNF-α in the β-cells of the islets of Langerhans under the control of RIP have been described earlier (15). Aβo.NOD.DQ8 mice, the 10th generation backcross between Aβo.HLA-DQ8 transgenic mice and NOD mice, homozygous for 15 NOD idd alleles (Aβo.NOD.DQ8), has been described elsewhere (13). B6.RIP-TNF and Aβo.NOD.DQ8 were mated to generate F1 offspring. Littermates obtained by intercrossing these first generation offspring were typed and used in the study. Independent segregation of genes would result in offspring with several different genotypes (Table 1). Presence or absence of appropriate MHC class I/class II molecules were determined by PCR and by flowcytometry as described elsewhere (13). Presence of RIP-TNF was confirmed by PCR. Mice were bred and maintained at the Immunogenetics Mouse Colony at the Mayo Clinic (Rochester, MN).

Diabetes monitoring.

The 4- to 6-week-old mice were moved from the barrier facility to the conventional facility after weaning, and their glycemic status was monitored weekly using a Glucometer (Bayer, Pittsburgh, PA). Diabetes was diagnosed when two consecutive random blood glucose levels were >13.9 mmol/l.

Histopathology and immunohistochemistry.

Histopathology was performed as described earlier (13). For immunofluorescence staining, pancreata were embedded in an OCT compound (Tissue Tek; Sakura FineTek, Torrance, CA), immersed in chilled Isopentane, and frozen immediately in liquid nitrogen. Thin sections made from frozen tissue were first air-dried, fixed in cold acetone, and subsequently treated with 1% paraformaldehyde. Sections were subsequently incubated with rat anti-mouse primary antibodies, CD4 (GK1.5), CD8 (Lyt 2.2), CD19 (Pharmingen, San Diego, CA), and CD11b (Pharmingen). A rat isotype control was also included. In the next step, sections were incubated with goat anti-rat secondary antibody (Jackson Immunoresearch Lab, West Grove, PA) followed by rabbit anti-goat IgG/fluorescein isothiocyanate conjugate (ICN, Costa Mesa, CA). Sections were examined under a fluorescent microscope.

Statistical analysis.

The statistical analyses were performed using SAS software (Version 4.0.4, SAS Institute, Cary, NC).

NOD/B6/RIP-TNF mice develop diabetes, which is dependent on islet-specific expression of TNF.

Previous studies have shown that the F1 intercrosses of B6.RIP-TNF mice and NOD mice seldom developed diabetes (15,17). This indicated that the presence of TNF did not overcome the dominant protective effects of B6-derived diabetes-resistant alleles when inducing diabetes in NOD F1 intercrosses. However, in the present study, diabetes developed in NOD/B6.RIP-TNF F1 intercrosses. Mice belonging to either sex were equally susceptible. However, diabetes only developed in littermates expressing TNF (males, 58/112, 49.15%; females, 38/106, 35.85%; overall, 96/218, 42.2%). Littermates lacking RIP-TNF were diabetes free during the entire observation period (males, 0/67; females, 0/69), confirming the prior reports that the presence of NOD background per se does not result in diabetes in NOD × B6 crosses (18). The possibility that transgenically expressed TNF could be directly toxic to the β-cells is unlikely, because if this were the case, the incidence of diabetes would have been similar in all the groups and would have occurred even in the B6.RIP-TNF founder line.

Diabetes in NOD/B6/RIP-TNF mice is independent of functional class II molecules.

o.NOD.DQ8 mice used for initial breeding lack endogenous class II molecules because the H2-Aβ gene has been inactivated by gene targeting (19). The functional I-E molecule is also not expressed in these mice because of a natural mutation in the H2-Eα promoter region, which precludes the expression of the I-Eα chain (20). Therefore, these mice express only the transgenic HLA-DQ8. On the other hand, B6.RIP-TNF mice do express functional H2-A molecules. The littermates generated by crossing Aβo.NOD.DQ8 and B6.RIP-TNF can therefore express different combinations of class II molecules, as shown in Table 1, or may not express any class II molecules at all. Interestingly, littermates completely lacking any functional class II molecules were indeed susceptible to diabetes (males, 14/17, 82.35%; females, 12/18, 66.67%; overall, 26/35, 74.28%). The onset of diabetes was earliest in this group of mice, and the disease progressed rapidly. Most importantly, the incidence of disease was highest in this group (P < 0.001, χ2 test; Fig. 1 and Table 2). However, diabetes in this group also depended on islet-specific expression of TNF because littermates lacking both class II and RIP-TNF were diabetes free (overall, 0/24; males, 0/7; females, 0/17). There were no significant differences in B220+ cells, macrophages, NK cells, or CD8+ T-cells in the secondary lymphoid organs between littermates of Aβo.NOD.RIP-TNF mice with and without RIP-TNF, indicating that transgenic expression of TNF in the islets does not alter the development of immune components (data not shown). Similarly, there was no appreciable difference in these above parameters between euglycemic and hyperglycemic Aβo.NOD.RIP-TNF mice (data not shown).

Expression of class II molecules downregulates diabetes.

Compared with class II-deficient littermates, the incidence of diabetes in mice that expressed any single functional class II molecules (Aβ−/−.DQ8+.NOD.RIP.TNF+, Aβ+/−.DQ8-.NOD.RIP.TNF+, and Aβ+/+.DQ8-.NOD.RIP.TNF+) was lower. In this group, the onset of disease was delayed and the disease progression was slower; however, the incidence of disease did not differ significantly. There was also no significant difference in the incidence of diabetes in the HLA-DQ8 mice that were homozygous or heterozygous for I-Ab. Interestingly, mice homozygous for H2-Ab and also expressing HLA-DQ8 (Aβ+/+.DQ8+.NOD.RIP.TNF+) had the lowest incidence of diabetes (P < 0.05 when compared with mice expressing any single class II molecule, and P < 0.001 when compared with mice devoid of all class II molecules; χ2 test; Fig. 1 and Table 2). This result indicated that the presence of functional class II molecules and therefore CD4+ T-cells downregulates disease incidence and severity in RIP-TNF mice. We also generated three lines of Aβ°.DQ8.RIP.TNF mice lacking CD8, β2-microglobulin, or the classic class I molecules Kb and Db. So far, the preliminary results indicate that these mice are diabetes free (data not shown). This result implies that CD8+ T-cells are most likely the effectors. We are currently generating these lines (CD8−/−, β2-microglobulin−/−, and KbDb−/−) on the class II-deficient background to prove a similar role for CD8+ T-cells in class II-deficient RIP-TNF mice.

Put together, these results indicate that 1) diabetes in NOD.RIP-TNF mice can occur even in the absence of functional class II molecules and CD4+ T-cells, and 2) expression of functional class II molecules offers significant protection from diabetes.

Histopathology of islets.

Infiltration of islets with mononuclear cells was evident as early as 3 weeks in RIP.TNF+ mice (Fig. 2), as shown in earlier studies (15,21), whereas the littermates lacking RIP-TNF did not show evidence of insulitis. The infiltrates were primarily composed of B-cells and macrophages. T-cells were also detected. No significant difference in the extent of insulitis or pattern of insulitis in different groups of mice based on the expression of class II molecules could be appreciated (data not shown). CD4+ T-cells were present in the class II-positive islets but were almost absent in mice lacking class II molecules, whereas CD8+ T-cells were present in islets from both groups (Fig. 3). B-cells and macrophages could be seen in islets from both groups; however, there seemed to be a difference in the distribution of B-cells and macrophages. Whereas the islets from mice expressing class II molecules had more macrophages, islets from mice lacking class II molecules had more B-cells (Fig. 3). Salivary glands, kidneys, and thyroid glands were devoid of any mononuclear cell infiltration (data not shown) as described previously (15).

During the normal course of diabetes in the NOD strain of mice, both CD4+ and CD8+ T-cells seem to be necessary for the development of diabetes (4). NOD mice deficient in MHC class I expression and thereby lacking CD8+ T-cells are protected from insulitis and type 1 diabetes (4). Similarly, NOD mice lacking the class II transactivator (CIITA−/−), which have drastically reduced surface class II (therefore lacking CD4+ T-cells), show periinsulitis but are nevertheless protected from diabetes (22). However, CD8+ T-cells seem to be important in the early phase of the disease (4). NOD.RIP-TNF mice (with intact NOD idd1) lacking class I molecules (β2 microglobulin−/−) were protected from diabetes (16). Susceptibility of RIP-TNF mice expressing class I molecules alone (without any class II) and the absence of disease in CD8 or β2 microglobulin null mice (data not shown) in the present study confirm the pathogenic role of CD8+ T-cells.

However, in an earlier study, abolishing the expression of class II molecules by mutating CIITA neither protected nor accelerated diabetes in NOD.RIP-TNF mice, implying that the class II molecules and CD4+ T-cells are not involved in disease pathogenesis in this model (22). Furthermore, residual class II expression and presence of CD4+ T-cells (albeit in reduced numbers) in CIITA−/− mice (23) might complicate the conclusion drawn from CIITA−/−.NOD.RIP-TNF mice. However, in the present study, the incidence of diabetes was significantly reduced in mice expressing any single competent class II molecule (H2-Ab or DQ8) and was reduced more so when both class II molecules were present, indicating that expression of functional class II molecules (and therefore presence of CD4+ T-cells) is protective. The reason(s) is not clear as to why expression of H2-Ab and DQ8 (as seen in the present study) protected RIP-TNF mice from diabetes, while H2-Ag7 had no such modulatory role in an earlier study (16). Nevertheless, several nonmutually exclusive reasons could be put forth. It could be due to the anti-diabetogenic effects of the genes closely associated with H2-Ab and the DQ8 transgene (positional effect), because both of these molecules are from a diabetes-resistant background. It could also be due to the intrinsic differences in class II molecules themselves. More importantly, we envisage a protective/regulatory role for CD4+CD25+ T-cells (T regulatory cells).

Several studies have shown that CD4+ T-cells expressing CD25, the IL-2R α-chain, have an important immunoregulatory role (24). CD4+CD25+ T regulatory cells have been shown to suppress the activity of both CD4+ and CD8+ T-cells. Because expression of class II molecules on the cortical epithelial cells seems to mediate the selection of CD4+CD25+ immunoregulatory T-cells (25), these cells should therefore be absent in class II-deficient mice, and, hence, Aβo.NOD.RIP-TNF mice have an increased incidence of disease. The role of CD4+CD25+ regulatory T-cells in spontaneous type 1 diabetes has been addressed earlier using the murine model of the disease (26). A recent human study has also shown that patients with type 1 diabetes also have low numbers of resting CD4+CD25+ T-cells (27). The recent demonstration by Green et al. (28) that CD4+CD25+ T-cells efficiently inhibit diabetes in RIP-B7.RIP-TNF double transgenic mice underscores the significance of these immunoregulatory T-cells in type 1 diabetes. However, a recent study has shown that absence of CD4+CD25+ T-cells per se is not sufficient to result in autoimmunity. Nevertheless, in an appropriate pro-inflammatory environment, deficiency of CD4+CD25+ T-cells can result in autoimmunity (29). The results of the present study correlate well with this hypothesis. Littermates lacking class II molecules (thereby CD4+CD25+ Treg cells) rarely developed type 1 diabetes in the absence of RIP-TNF, whereas they readily did so when TNF was expressed in the islets. In addition to this result, recent studies have shown that CD4+CD25 T-cells also play a regulatory role (3032). Thus, a low incidence of diabetes in class II-positive littermates could be due to several mutually nonexclusive mechanisms.

Even though neonatal expression of TNF-α in the islets of NOD mice can override the requirement for several of the NOD idd alleles, homozygosity for NOD alleles at the MHC locus (i.e., idd1) is mandatory for the development of accelerated diabetes in NOD.RIP-TNF mice (15,17). Interestingly, in the present study, diabetes was evident even in the F1 intercross of Aβo.NOD.DQ8 and B6.RIP-TNF mice (both carrying the B6-derived diabetes-resistant idd1 allele because the inactivated H2-Aβ gene was also made originally on a diabetes-resistant background), indicating that the presence of NOD-derived idd1 is not mandatory. NOD idd1 consists of Kd, Db, a nonexpressing H2-E, and H2-Ag7 (18). In addition to the clear-cut involvement of NOD class II, class I molecules also seem to be important in the disease pathogenesis (4). Curiously, most, if not all, of the class I-restricted CD8+ T-cells/clones identified thus far from NOD mice seem to be restricted to Kd (3335). However, as evident from the present study, expression of Kd is not always required, and diabetes can also occur in mice expressing Kb and Db.

How TNF-α may cause/accelerate diabetes has been an area of active research. Green et al. (16,36) have shown that the local expression of TNF-α under the control of RIP resulted in early apoptosis of β-cells. The recent demonstration that TNF-receptor is indeed expressed in β-cells and is critical for the development of diabetes (37) supports the notion that TNF-mediated apoptosis of β-cells is a crucial event not only in the RIP-TNF model but also during the natural course of the disease, even in NOD mice (38). Apoptosis of β-cell recruits dendritic cells (DCs), which may then cross prime CD8+ T-cells independent of the CD40-CD40L pathway (15). A recent study based on a CD8+ T-cell-dependent diabetes model has also shown that in situ β-cell death can effectively prime CD8+ T-cells (39). In situ demonstration of DCs by immunohistochemical staining of affected islets underscores the importance of DCs in the pathogenesis (17).

Thus, in conclusion we have shown that accelerated diabetes in NOD.RIP-TNF mice is independent of the NOD allele at idd1 and is independent of the expression of functional class II molecules. Expression of class II is rather inhibitory to the development of diabetes. Expression of TNF in the pancreas thus seems to bypass the requirement of class II molecules for antigen presentation and might lead to a direct effector pathway in the context of class I and CD8 T-cells independent of CD4+ T-cells. This model can be useful to dissect roles of diabetogenic genes and class I/CD8 in the disease pathogenesis and to study therapeutic/preventive modalities.

FIG. 1.

Cumulative incidence of diabetes. Mice belonging to different groups were monitored weekly to assess their glycemic status. Diabetes was diagnosed when two consecutive random blood glucose levels were >13.9 mmol/l. Figures in parentheses indicate total number/group.

FIG. 1.

Cumulative incidence of diabetes. Mice belonging to different groups were monitored weekly to assess their glycemic status. Diabetes was diagnosed when two consecutive random blood glucose levels were >13.9 mmol/l. Figures in parentheses indicate total number/group.

Close modal
FIG. 2.

Histopathology of islets. Pancreata were collected from mice at indicated time points and fixed in buffered formalin. Thin sections were stained with hematoxylin/eosin and evaluated microscopically. A and D: RIP-TNF+ mice at 3 weeks; B and E: RIP-TNF+ mice at 24 weeks; C and F: RIP-TNF mice at 24 weeks. Magnification in AC, ×20, and in DF, ×60.

FIG. 2.

Histopathology of islets. Pancreata were collected from mice at indicated time points and fixed in buffered formalin. Thin sections were stained with hematoxylin/eosin and evaluated microscopically. A and D: RIP-TNF+ mice at 3 weeks; B and E: RIP-TNF+ mice at 24 weeks; C and F: RIP-TNF mice at 24 weeks. Magnification in AC, ×20, and in DF, ×60.

Close modal
FIG. 3.

Immunofluorescence. Cryosections of pancreas from Aβ−/−.NOD.DQ8+.RIP-TNF mice (AF) or Aβ−/−.DQ8.NOD.RIP-TNF mice (GL) were stained either with hematoxylin and eosin (A and G) or with CD4 (B and H), CD8 (C and I), CD19 (D and J), CD11b (E and K), or a control antibody (F and L), as described in research design and methods. Magnification in A–F, ×100, and in G–L, ×200.

FIG. 3.

Immunofluorescence. Cryosections of pancreas from Aβ−/−.NOD.DQ8+.RIP-TNF mice (AF) or Aβ−/−.DQ8.NOD.RIP-TNF mice (GL) were stained either with hematoxylin and eosin (A and G) or with CD4 (B and H), CD8 (C and I), CD19 (D and J), CD11b (E and K), or a control antibody (F and L), as described in research design and methods. Magnification in A–F, ×100, and in G–L, ×200.

Close modal
TABLE 1

List and genotypes of mice used in the present study

DQ8TNF
−/− − 
+/+ 
+/+ − 
+/− 
+/− − 
−/− 
−/− − − 
+/+ − 
+/+ − − 
+/− − 
+/− − − 
−/− − 
DQ8TNF
−/− − 
+/+ 
+/+ − 
+/− 
+/− − 
−/− 
−/− − − 
+/+ − 
+/+ − − 
+/− − 
+/− − − 
−/− − 

B6.RIP-TNF mice were mated with Aβ°.NOD.DQ8 (N10). The first generation offspring was intercrossed for two to three generations and was used in the study. Presence or absence of appropriate class I, class II, and RIP-TNF was determined by PCR and/or flowcytometry.

TABLE 2

Incidence of diabetes

DQ8TNFMalesFemalesTotal
−/− − 14/17 (82.35) 12/18 (66.67) 26/35 (74.28) 
+/+ 0/3 2/10 (20) 2/13 (15.38) 
+/+ − 5/10 (50) 2/6 (33.33) 7/16 (43.75) 
+/− 9/11 (81.81) 2/12 (16.67) 11/23 (47.83) 
+/− − 9/23 (39.13) 9/31 (29.03) 18/54 (33.33) 
−/− 21/48 (43.48) 11/29 (37.93) 32/77 (41.56) 
−/− − − 0/10 0/7 0/17 
+/+ − 0/3 0/4 0/7 
+/+ − − 0/7 0/11 0/18 
+/− − 0/7 0/9 0/16 
+/− − − 0/7 0/4 0/11 
−/− − 0/33 0/34 0/67 
DQ8TNFMalesFemalesTotal
−/− − 14/17 (82.35) 12/18 (66.67) 26/35 (74.28) 
+/+ 0/3 2/10 (20) 2/13 (15.38) 
+/+ − 5/10 (50) 2/6 (33.33) 7/16 (43.75) 
+/− 9/11 (81.81) 2/12 (16.67) 11/23 (47.83) 
+/− − 9/23 (39.13) 9/31 (29.03) 18/54 (33.33) 
−/− 21/48 (43.48) 11/29 (37.93) 32/77 (41.56) 
−/− − − 0/10 0/7 0/17 
+/+ − 0/3 0/4 0/7 
+/+ − − 0/7 0/11 0/18 
+/− − 0/7 0/9 0/16 
+/− − − 0/7 0/4 0/11 
−/− − 0/33 0/34 0/67 

Data in parentheses are %. Mice belonging to different groups were monitored weekly to assess their glycemic status. Diabetes was diagnosed when two consecutive random blood glucose levels were >13.9 mmol/l.

Generation of HLA transgenic mice was supported by National Institutes of Health (NIH) Grant AI14764. Financial support from the American Diabetes Association is also acknowledged. R.A.F. is an investigator at the Howard Hughes Medical Institute. Work by R.A.F. was supported by grants from NIH (ROI-DK51665 and P01-DK53015) and the American Diabetes Association.

We thank Julie Hanson and her crew for excellent animal husbandry and Michelle Smart for characterizing the transgenic mice. We also thank Tim Plummer, Karen Lien, and J.L. Platt, MD, for help in immunostaining and fluorescence microscopy.

1
Tisch R, McDevitt H: Insulin-dependent diabetes mellitus.
Cell
85
:
291
–297,
1996
2
Field LL: Genetic linkage and association studies of type I diabetes: challenges and rewards.
Diabetologia
45
:
21
–35,
2002
3
Rosmalen JG, van Ewijk W, Leenen PJ: T-cell education in autoimmune diabetes: teachers and students.
Trends Immunol
23
:
40
–46,
2002
4
Wong FS, Janeway CA Jr: The role of CD4 vs. CD8 T cells in IDDM.
J Autoimmun
13
:
290
–295,
1999
5
Taneja V, David CS: HLA class II transgenic mice as models of human diseases.
Immunol Rev
169
:
67
–79,
1999
6
Cheta D: Animal models of type I (insulin-dependent) diabetes mellitus.
J Pediatr Endocrinol Metab
11
:
11
–19,
1998
7
Lee KH, Wucherpfennig KW, Wiley DC: Structure of a human insulin peptide-HLA-DQ8 complex and susceptibility to type 1 diabetes.
Nat Immunol
2
:
501
–507,
2001
8
Wen L, Wong FS, Sherwin R, Mora C: Human DQ8 can substitute for murine I-A(g7) in the selection of diabetogenic T cells restricted to I-A(g71).
J Immunol
168
:
3635
–3640,
2002
9
Abraham RS, Kudva YC, Wilson SB, Strominger JL, David CS: Co-expression of HLA DR3 and DQ8 results in the development of spontaneous insulitis and loss of tolerance to GAD65 in transgenic mice.
Diabetes
49
:
548
–554,
2000
10
Abraham RS, Wen L, Marietta EV, David CS: Type 1 diabetes-predisposing MHC alleles influence the selection of glutamic acid decarboxylase (GAD) 65-specific T cells in a transgenic model.
J Immunol
166
:
1370
–1379,
2001
11
Wen L, Wong FS, Tang J, Chen NY, Altieri M, David C, Flavell R, Sherwin R: In vivo evidence for the contribution of human histocompatibility leukocyte antigen (HLA)-DQ molecules to the development of diabetes.
J Exp Med
191
:
97
–104,
2000
12
Wen L, Chen NY, Tang J, Sherwin R, Wong FS: The regulatory role of DR4 in a spontaneous diabetes DQ8 transgenic model.
J Clin Invest
107
:
871
–880,
2001
13
Kudva YC, Rajagopalan G, Raju R, Abraham RS, Smart M, Hansen J, David CS: Modulation of insulitis and type 1 diabetes by transgenic HLA-DR3 and DQ8 in NOD mice lacking endogenous MHC class II.
Hum Immunol
63
:
987
–999,
2002
14
Raju R, Munn SR, Majoribanks C, David CS: Islet cell autoimmunity in NOD mice transgenic for HLA-DQ8 and lacking I-Ag7.
Transplant Proc
30
:
561
,
1998
15
Picarella DE, Kratz A, Li CB, Ruddle NH, Flavell RA: Transgenic tumor necrosis factor (TNF)-alpha production in pancreatic islets leads to insulitis, not diabetes: distinct patterns of inflammation in TNF-alpha and TNF-beta transgenic mice.
J Immunol
150
:
4136
–4150,
1993
16
Green EA, Wong FS, Eshima K, Mora C, Flavell RA: Neonatal tumor necrosis factor alpha promotes diabetes in nonobese diabetic mice by CD154-independent antigen presentation to CD8(+) T cells.
J Exp Med
191
:
225
–238,
2000
17
Green EA, Flavell RA: Tumor necrosis factor-alpha and the progression of diabetes in non-obese diabetic mice.
Immunol Rev
169
:
11
–22,
1999
18
Wicker LS, Todd JA, Peterson LB: Genetic control of autoimmune diabetes in the NOD mouse.
Annu Rev Immunol
13
:
179
–200,
1995
19
Cosgrove D, Gray D, Dierich A, Kaufman J, Lemeur M, Benoist C, Mathis D: Mice lacking MHC class II molecules.
Cell
66
:
1051
–1066,
1991
20
Mathis DJ, Benoist C, Williams VE 2nd, Kanter M, McDevitt HO: Several mechanisms can account for defective E alpha gene expression in different mouse haplotypes.
Proc Natl Acad Sci U S A
80
:
273
–277,
1983
21
Higuchi Y, Herrera P, Muniesa P, Huarte J, Belin D, Ohashi P, Aichele P, Orci L, Vassalli JD, Vassalli P: Expression of a tumor necrosis factor alpha transgene in murine pancreatic beta cells results in severe and permanent insulitis without evolution towards diabetes.
J Exp Med
176
:
1719
–1731,
1992
22
Mora C, Wong FS, Chang CH, Flavell RA: Pancreatic infiltration but not diabetes occurs in the relative absence of MHC class II-restricted CD4 T cells: studies using NOD/CIITA-deficient mice.
J Immunol
162
:
4576
–4588,
1999
23
Williams GS, Malin M, Vremec D, Chang CH, Boyd R, Benoist C, Mathis D: Mice lacking the transcription factor CIITA: a second look.
Int Immunol
10
:
1957
–1967,
1998
24
Shevach EM, McHugh RS, Piccirillo CA, Thornton AM: Control of T-cell activation by CD4+ CD25+ suppressor T cells.
Immunol Rev
182
:
58
–67,
2001
25
Bensinger SJ, Bandeira A, Jordan MS, Caton AJ, Laufer TM: Major histocompatibility complex class II-positive cortical epithelium mediates the selection of CD4(+)25(+) immunoregulatory T cells.
J Exp Med
194
:
427
–438,
2001
26
Salomon B, Lenschow DJ, Rhee L, Ashourian N, Singh B, Sharpe A, Bluestone JA: B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes.
Immunity
12
:
431
–440,
2000
27
Kukreja A, Cost G, Marker J, Zhang C, Sun Z, Lin-Su K, Ten S, Sanz M, Exley M, Wilson B, Porcelli S, Maclaren N: Multiple immuno-regulatory defects in type-1 diabetes.
J Clin Invest
109
:
131
–140,
2002
28
Green EA, Choi Y, Flavell RA: Pancreatic lymph node-derived CD4(+)CD25(+) Treg cells: highly potent regulators of diabetes that require TRANCE-RANK signals.
Immunity
16
:
183
–191,
2002
29
McHugh RS, Shevach EM: Depletion of CD4+CD25+ regulatory T cells is necessary, but not sufficient, for induction of organ-specific autoimmune disease.
J Immunol
168
:
5979
–5983,
2002
30
Tarbell KV, Lee M, Ranheim E, Chao CC, Sanna M, Kim SK, Dickie P, Teyton L, Davis M, McDevitt H: CD4(+) T cells from glutamic acid decarboxylase (GAD)65-specific T cell receptor transgenic mice are not diabetogenic and can delay diabetes transfer.
J Exp Med
196
:
481
–492,
2002
31
Graca L, Thompson S, Lin CY, Adams E, Cobbold SP, Waldmann H: Both CD4(+)CD25(+) and CD4(+)CD25(−) regulatory cells mediate dominant transplantation tolerance.
J Immunol
168
:
5558
–5565,
2002
32
Wolf M, Schimpl A, Hunig T: Control of T cell hyperactivation in IL-2-deficient mice by CD4(+)CD25(−) and CD4(+)CD25(+) T cells: evidence for two distinct regulatory mechanisms.
Eur J Immunol
31
:
1637
–1645,
2001
33
Anderson B, Park BJ, Verdaguer J, Amrani A, Santamaria P: Prevalent CD8(+) T cell response against one peptide/MHC complex in autoimmune diabetes.
Proc Natl Acad Sci U S A
96
:
9311
–9316,
1999
34
Wong FS, Visintin I, Wen L, Flavell RA, Janeway CA Jr: CD8 T cell clones from young nonobese diabetic (NOD) islets can transfer rapid onset of diabetes in NOD mice in the absence of CD4 cells.
J Exp Med
183
:
67
–76,
1996
35
Quinn A, McInerney MF, Sercarz EE: MHC class I-restricted determinants on the glutamic acid decarboxylase 65 molecule induce spontaneous CTL activity.
J Immunol
167
:
1748
–1757,
2001
36
Green EA, Eynon EE, Flavell RA: Local expression of TNFalpha in neonatal NOD mice promotes diabetes by enhancing presentation of islet antigens.
Immunity
9
:
733
–743,
1998
37
Pakala SV, Chivetta M, Kelly CB, Katz JD: In autoimmune diabetes the transition from benign to pernicious insulitis requires an islet cell response to tumor necrosis factor alpha.
J Exp Med
189
:
1053
–1062,
1999
38
Mathis D, Vence L, Benoist C: Beta-cell death during progression to diabetes.
Nature
414
:
792
–798,
2001
39
Zhang Y, O’Brien B, Trudeau J, Tan R, Santamaria P, Dutz JP: In situ beta cell death promotes priming of diabetogenic CD8 T lymphocytes.
J Immunol
168
:
1466
–1472,
2002

Address correspondence and reprint requests to Dr. Chella S. David, Department of Immunology, Mayo Clinic, 200 First St., SW, Rochester, MN 55905. E-mail: davic4@mayo.edu.

Received for publication 28 June 2002 and accepted in revised form 5 November 2002.

CIITA, class II transactivator; MHC, major histocompatibility complex; RIP, rat insulin promoter; TNF, tumor necrosis factor.