Nonobese Diabetic Strain Research Articles

Congenic mapping and functional analysis of a second component of the MHC-linked diabetogenic gene (Idd16)

By using a congenic non-obese diabetic (NOD) mouse strain that possesses a recombinant majorhistocompatibility complex (MHC) from a diabetes-resistant sister strain, the CTS mouse, we havepreviously mapped a second component of the MHC linked susceptibility gene (Idd16) to the <11.8-centiMorgan (cM) segment of chromosome 17 adjacent to, but distinct from class II A and E genes(Idd1). To further localise and characterise Idd16, CTS-derived non-diabetogenic genetic interval,identified in our previous study, was transferred onto the NOD background, and a new recombinantwith recombination breakpoint at 1.3-cM proximal to the previous congenic chromosome was obtainedamong 49 intercrosses of heterozygous mice. The frequency of type 1 diabetes in mice possessing thenew recombinant chromosome was lower than that in the parental NOD strain, as in the case of theprevious congenic strain. The tumor necrosis factor a gene (Tnf), a strong candidate gene for type 1diabetes, is located within the newly localized Idd16 interval. Although our previous study indicatedthat the sequences of exons and exon-intron junctions of Tnf in the NOD mouse were identical to thosein the control mouse, defective expression of TNFα may well contribute to Idd16 effect. We thereforedetermined plasma TNFα levels in NOD and CTS mice. TNFa levels under basal conditions and afterstimulation with interferon gamma (IFN-gamma) and lipopolysaccharide (LPS), however, werecomparable among NOD, CTS and control mice. These data suggest that Idd16 is located in the <10.5-cM segment of chromosome 17 adjacent to Idd1, and that defective expression of TNFa may not beresponsible for Idd16 effect, at least under the condition used in the present study.

New Zealand mixed mice: a genetic systemic lupus erythematosus model for assessing environmental effects.

Systemic lupus erythematosus (SLE) is a multiphenotypic autoimmune disease. The hallmark of SLE is the production of anti-double-stranded DNA autoantibodies and the deposition of immune complexes in target tissues such as the kidney, skin, and brain. Additional phenotypic traits are the presence of arthritis, anemia, central nervous system involvement, and a variety of autoantibodies. Females of childbearing age are particularly at risk. Recent genetic analysis of murine SLE shows that susceptibility is under complex polygenic control. It is also apparent that environmental factors contribute to the induction and exacerbation of SLE. We describe here the genotypic and phenotypic characterization of a group of recombinant inbred strains of SLE-prone mice that were derived from NZB and NZW progenitors, the parental strains of the classic female F1 hybrid lupus model. Recombination and reassortment of these ancestral genomes resulted in the NZM (New Zealand mixed) strains with strain-specific patterns of renal disease penetrance and other autoimmune traits such as Coombs positive anemia and neurologic deficits. Multiple susceptibility loci of the ancestral strains demonstrate that SLE is inherited as a threshold trait. Because some of these loci co-localize with the susceptibility loci of the insulin-dependent diabetes of nonobese diabetic strain, it is apparent that there are disease-specific as well as autoimmunity-promoting genes. It is proposed that the NZM strains, particularly those with reduced disease penetrance or partial genotypes, provide an improved genetic model for assessment of the effects of environmental agents on SLE and autoimmunity.

New concepts for the development of autoimmune exocrinopathy derived from studies with the NOD mouse model.

The non-obese diabetic (NOD) mouse is now recognized as an appropriate model to study autoimmune exocrinopathy prevalent in human Sjögren's syndrome patients. With increasing age, NOD mice undergo histopathological changes similar to human Sjögren's syndrome patients, but more importantly, exhibit the same clinical manifestation of declining exocrine tissue secretory function. Studies with the immunodeficient NOD-scid mouse have provided evidence for the temporal loss in the expression of several major salivary proteins and a decreased presence of acinar cells in salivary tissues. The diminished presence of acinar cells is accompanied by an increase in the enzymes associated with apoptosis in the absence of T- and B-lymphocytes. Despite these alterations, NOD-scid mice, unlike NOD mice, do not lose secretory function. Recent analyses of a second congenic NOD strain, the NOD.Igmnull, which lacks B-lymphocytes, indicate the histological presence of a T-cell infiltrate of the exocrine glands, increased caspace activity and induction of the biochemical alterations in protein expression observed in NOD and NOD-scid mice. NOD.Igmnull mice also do not lose secretory function, but can be manipulated to generate a reduced secretory response following the infusion of IgG fractions from autoimmune NOD mice or Sjögren's syndrome patients. These observations, in the absence of components of the adaptive arm of the immune system, have given rise to the concept that autoimmune exocrinopathy develops in two phases. The initial phase is lymphocyte independent and occurs as a consequence of an innate error in exocrine tissue homeostasis or differentiated function. The subsequent tissue specific immunological attack, generated in part by B-cell autoantibodies, is responsible for the loss of secretory function. Our preliminary observations in both NOD mice and Sjögren's syndrome patients is that antibody directed against the cell surface muscarinic/cholinergic receptors appears to play an important part in the onset of clinical disease.

New concepts for the development of autoimmune exocrinopathy derived from studies with the NOD mouse model

The non-obese diabetic (NOD) mouse is now recognized as an appropriate model to study autoimmune exocrinopathy prevalent in human Sjögren's syndrome patients. With increasing age, NOD mice undergo histopathological changes similar to human Sjögren's syndrome patients, but more importantly, exhibit the same clinical manifestation of declining exocrine tissue secretory function. Studies with the immunodeficient NOD-scid mouse have provided evidence for the temporal loss in the expression of several major salivary proteins and a decreased presence of acinar cells in salivary tissues. The diminished presence of acinar cells is accompanied by an increase in the enzymes associated with apoptosis in the absence of T- and B-lymphocytes. Despite these alterations, NOD-scid mice, unlike NOD mice, do not lose secretory function. Recent analyses of a second congenic NOD strain, the NOD.Igmnull, which lacks B-lymphocytes, indicate the histological presence of a Tell infiltrate of the exocrine glands, increased caspace activity and induction of the biochemical alterations in protein expression observed in NOD and NOD-scid mice. NOD.Igmnull mice also do not lose secretory function, but can be manipulated to generate a reduced secretory response following the infusion of IgG fractions from autoimmune NOD mice or Sjögren's syndrome patients. These observations, in the absence of components of the adaptive arm of the immune system, have given rise to the concept that autoimmune exocrinopathy develops in two phases. The initial phase is lymphocyte independent and occurs as a consequence of an innate error in exocrine tissue homeostasis or differentiated function. The subsequent tissue specific immunological attack, generated in part by B-cell autoantibodies, is responsible for the loss of secretor function. Our preliminary observations in both NOD mice and Sjögren's syndrome patients is that antibody directed against the cell surface muscarinic/cholinergic receptors appears to play an important part in the onset of clinical disease.

Glutamate decarboxylase and GABA in pancreatic islets: lessons from knock-out mice.

The GABA-synthesizing enzyme glutamic acid decarboxylase (GAD) is expressed in pancreatic beta-cells and GABA has been suggested to play a role in islet cell development and function. Mouse beta-cells predominantly express the larger isoform of the enzyme, GAD67, and very low levels of the second isoform, GAD65. Yet GAD65 has been shown to be a target of very early autoimmune T-cell responses associated with beta-cell destruction in the non-obese diabetic (NOD) mouse model of Type 1 diabetes. Mice deficient in GAD67, GAD65 or both were used to assess whether GABA is important for islet cell development, and whether GAD65 is required for initiation of insulitis and progression to Type 1 diabetes in the mouse. Lack of either GAD65 or GAD67 did not effect the development of islet cells and the general morphology of islets. When GAD65-/-(129/Sv) mice were backcrossed into the NOD strain for four generations, GAD65-deficient mice developed insulitis similar to GAD65+/+ mice. Furthermore, at the low penetrance of diabetes in this backcross, GAD65-deficient mice developed disease at the same rate and incidence as wildtype mice. The results suggest that GABA generated by either GAD65 or GAD67 is not critically involved in islet formation and that GAD65 expression is not an absolute requirement for development of autoimmune diabetes in the NOD mouse.

Evidence for the presence of insulin-dependent diabetes-associated alleles on the distal part of mouse chromosome 6.

Type 1 diabetes (IDDM) is a complex disorder with multifactorial and polygenic etiology. A genome-wide screen performed in a BC1 cohort of a cross between the nonobese diabetic (NOD) mouse with the diabetes-resistant feral strain PWK detected a major locus contributing to diabetes development on the distal part of chromosome 6. Unlike the majority of other Idd loci identified in intraspecific crosses, susceptibility is associated with the presence of the PWK allele. Genetic linkage analysis of congenic lines segregating PWK chromosome 6 segments in a NOD background confirmed the presence of the Idd locus within this region. The genetic interval defined by analysis of congenic animals showed a peak of significant linkage (P = 0.0005) centered on an approximately 9-cM region lying between D6Mit11 and D6Mit25 genetic markers within distal mouse chromosome 6. [Genetic markers polymorphic between the NOD and PWK strains are available as a supplement at http://www.genome.org]

Development of an I-Ag7-expressing Antigen-presenting Cell Line: Intrinsic Molecular Defect in Compact I-Ag7Dimer Generation

Insulin-dependent diabetes mellitus (IDDM) results from chronic, T-cell dependent, autoimmune destruction of the insulin-producing β-cells in the Langerhans’ islets of the pancreas. Non-obese diabetic (NOD) mice spontaneously develop IDDM that resembles human type I diabetes. The susceptibility to diabetes in the NOD strain is a complex polygenic trait that determines a phenotype of immune alterations. The unique MHC class II molecule expressed by NOD mice (I-Ag7) plays a major role in the development of disease. Recently, it has been reported that I-Ag7molecules generate a lower proportion of compact αβheterodimers, compared to other haplotypes. However, it is not clear whether this reflects an intrinsic defect of this molecule to bind peptide stably or is the result of abnormal processing and/or peptide loading into the I-Ag7molecule. Our aim was to develop and characterize a suitable antigen-presenting cell (APC) that expressed I-Ag7in the context of a non-diabetes-prone antigen processing and presentation machinery. Here, we report the generation of a mouse DAP.3 fibroblast cell line (DAP.3Ag7) that constitutively expresses high levels of I-Ag7. Using DAP.3 cells transfected with I-Ag7or I-Ak, we show that the expression of compact dimers in the same cell type is proportionally less for I-Ag7molecules than for I-Akmolecules, implying an intrinsic defect of the I-Ag7molecule as the cause for the low generation of compact dimers. However, DAP.3Ag7cells are able to process and present antigen, as indicated by I-Ag7-dependent IL-2 production by a GAD67-specific NDO T-cell hybridoma after stimulation with GAD and live, but not fixed, DAP.3Ag7cells. The IL-2 response to GAD when presented by DAP.3Ag7was significantly higher than the response to GAD presented by NOD splenocytes. Based on these data, we conclude that the low generations of compact dimers is an intrinsic feature of I-Ag7molecules and not affected by other genes in the NOD background. The DAP.3Ag7cell line should be a valuable tool with which to dissect the role of the I-Ag7molecule in antigen presentation and T-cell activation in NOD mice, which clearly contributes to the development of IDDM.

Congenic mapping of the insulin-dependent diabetes (Idd) gene, Idd10, localizes two genes mediating the Idd10 effect and eliminates the candidate Fcgr1.

The development of autoimmune diabetes in the nonobese diabetic (NOD) mouse is under the control of multiple insulin-dependent diabetes (Idd) genes. The Idd3 gene, originally defined as a broad peak of linkage on mouse chromosome 3, was subsequently identified as two genes, Idd3 and Idd10, separated by at least 20 cM. The resistance alleles of Idd3 and Idd10 individually confer only partial protection from diabetes but, in combination, result in profound resistance to disease due to an epistatic genetic interaction. In this study, we used newly developed congenic strains to further localize Idd10. Surprisingly, we found that Idd10 itself comprises at least two linked loci: Idd10 and the newly designated Idd17. Idd17 was localized to a 1.1-cM region between D3Mit26 and D3Mit40, proximal to Fcgr1, a candidate gene encoding the high affinity Fc receptor for IgG. Idd10 was localized to a 10-cM region between D3Mit213 and D3Mit106, distal to Fcgr1. Thus, Fcgr1 was excluded as a candidate for either Idd10 or Idd17, despite the fact that the NOD strain expresses a mutant form of the receptor. Interestingly, although Idd10 and Idd17 participate in a genetic interaction with each other, Idd10 but not Idd17 participates in the genetic interaction with Idd3. Our study on chromosome 3 begins to reveal the extent of the polygenic nature of autoimmune diabetes, and demonstrates that the use of congenic strains is an effective mapping strategy, even in the dissection of multiple, linked genes with subtle effects.

Effect of MHC Transgene Expression on Spontaneous Insulin Autoantibody Class Switch in Nonobese Diabetic Mice

A study of spontaneous anti-insulin autoantibodies in nonobese diabetic (NOD) mice revealed that when first detected, the antibodies are immunoglobulin M (IgM), but by age 10 weeks, immunoglobulin G (IgG) autoantibodies have appeared in many of these animals. When NOD strains, partially or completely protected from IDDM by the insertion of transgenes in the class II region, were compared, it was found that the switch to IgG autoantibodies was inhibited and the autoantibodies remained IgM indefinitely. We speculate that the switch to IgG may be a marker of events leading to IDDM in NOD mice and an indication that T-cell help has been generated for responses to beta-cell antigens. Such help not only directs the development of IgG autoantibodies, but more importantly, allows the emergence of potentially pathogenic T-cell clones that are capable of infiltrating the pancreas and mediating beta-cell damage.

Association of an androgen-responsive T cell phenotype with murine diabetes and Idd2.

T cells are involved in the induction and suppression of autoimmune diabetes in nonobese diabetic (NOD) mice. Because the incidence of diabetes is 13-fold greater in NOD/Smrf females, we searched for T cell phenotypes that showed sexual dimorphism and associated with diabetes in backcross segregants. The percentage of CD4+PBL was higher in NOD/Smrf females than males, was intermediate in [NOD X NON] F1 mice and approximated a 1:1 distribution in F1 mice backcrossed to either NOD or NON parental strains, suggesting primary control of the phenotype by an incompletely dominant gene, but not excluding additional effects by other genes. We term this primary gene Tlf(T lymphocyte frequency) because it also influenced the percentage of CD8+ T cells, although to lesser extent and independently from the MHC previously shown to lower the CD8+ T cell fraction in NON mice. Tlf segregated with diabetes in BC1 females, suggesting linkage with at least one diabetic locus. Genotyping of markers for Idd1, Idd2, and Idd3/10 revealed that Tlf mapped with Idd2 on chromosome 9. Dihydrotestosterone simultaneously lowered CD4+ PBL levels and prevented diabetes in NOD females while, in vitro, it had a differential effect on Con A elicited cytokines, increasing IL-2 22% and decreasing IL-4 39% (p < 0.0001). Thus the Tlf phenotype in NOD females, like diabetes, can be modulated by androgens.

Anchored polymerase chain reaction based analysis of the V beta repertoire in the non-obese diabetic (NOD) mouse.

We have performed extensive analyses of T cell receptor V beta usage in the thymus, the spleen and the infiltrated islets of preclinical non-obese diabetic (NOD) mice. A semiquantitative anchored polymerase chain reaction (An-PCR) protocol has been developed for this purpose. The validity of the method has been first assessed by antibody staining with a panel of anti-V beta monoclonal antibodies (mAb). The results obtained by An-PCR are accurate, reproducible, and in good agreement with cell surface protein staining. A strict comparison between thymus and spleen repertoires reveals no major V beta-specific deletion except the already reported V beta 3 deletion due to Mtv-3. Certain V beta such as V beta 15, 18, 20 are found with a low frequency in the spleen, but the fact that they are also scarce in the thymus probably reflects a poor availability of these genetic elements during beta chain rearrangement rather than negative selection. Other V beta, such as V beta 2, V beta 12 and V beta 14 are significantly more abundant in the spleen than in the thymus. This finding was confirmed by mAb staining for V beta 2 and V beta 14. The expansion asymmetrically affects the CD4+ subset and can be traced back to the mature, single-positive thymocyte subset, suggesting an intrathymic positive selection event. V beta repertoires in infiltrated islets of 13- and 18-week-old, non-diabetic mice are polymorphic. Practically all the V beta found in the peripheral lymphoid tissues are present in the islets, in similar proportions. The major exception is V beta 12, one of the V beta which is subject to expansion during intrathymic differentiation and which is further augmented in the islets, both at 13 and 18 weeks. This increase probably reflects further peripheral amplification of the V beta 12-bearing subset due to encounter with the same ligand as in the thymus or with a cross-reactive motif. Finally, the nucleotide sequencing of all the V beta segments in usage in the NOD strain confirms the absence of allelic polymorphism of V beta-coding regions.

I-E+ nonobese diabetic mice develop insulitis and diabetes.

The development of type I diabetes in the nonobese diabetic (NOD) mouse is under the control of multiple genes, one or more of which is linked to the major histocompatibility complex (MHC). The MHC class II region has been implicated in disease development, with expression of an I-E transgene in NOD mice shown to provide protection from insulitis and diabetes. To examine the effect of expressing an I-E+ or I-E- non-NOD MHC on the NOD background, three I-E+ and three I-E- NOD MHC congenic strains (NOD.H-2i5, NOD.H-2k, and NOD.H-2h2, and NOD.H-2h4, NOD.H-2i7, and NOD.H-2b, respectively) were developed. Of these strains, both I-E+ NOD.H-2h2 and I-E- NOD.H-2h4 mice developed insulitis, but not diabetes. The remaining four congenic strains were free of insulitis and diabetes. These results indicate that in the absence of the NOD MHC, diabetes fails to develop. Each NOD MHC congenic strain was crossed with the NOD strain to produce I-E+ and I-E- F1 mice; these mice thus expressed one dose of the NOD MHC and one dose of a non-NOD MHC on the NOD background. While a single dose of a non-NOD MHC provided a large degree of disease protection to all of the F1 strains, a proportion of I-E+ and I-E- F1 mice aged 5-12 mo developed insulitis and cyclophosphamide-induced diabetes. When I-E+ F1 mice were aged 9-17 mo, spontaneous diabetes developed as well. These data are the first to demonstrate that I-E+ NOD mice develop diabetes, indicating that expression of I-E in NOD mice is not in itself sufficient to prevent insulitis or diabetes. In fact, I-E- F1 strains were no more protected from diabetes than I-E+ F1 strains, suggesting that other non-NOD MHC-linked genes are important in protection from disease. Finally, transfer of NOD bone marrow into irradiated I-E+ F1 recipients resulted in high incidences of diabetes, indicating that expression of non-NOD MHC products in the thymus, in the absence of expression in bone marrow-derived cells, is not sufficient to provide protection from diabetes.

RFLP analysis of the MHC class III region defines unique haplotypes for the non-obese diabetic, cataract Shionogi and the non-obese non-diabetic mouse strains

The non-obese diabetic (NOD) mouse strain which spontaneously develops diabetes is a model for human Type 1 (insulin-dependent) diabetes mellitus. At least one of several genes controlling diabetes in the NOD mouse has been mapped to the MHC. Although previous experiments have implicated the MHC class II genes in the development of the disease, the existence of other MHC linked susceptibility genes has not been ruled out. In order to identify these susceptibility genes we have further characterized the MHC haplotype of the NOD mouse and two non-diabetic sister strains, the non-obese non-diabetic (NON) and cataract Shionogi (CTS). We have examined the mouse MHC class III region for the presence of homologous genes to 17 newly isolated human MHC class III region genes (G1, G2, G4, G6, G7a/valyl-tRNA synthetase, HSP70, G8, G9, G10, G12, G13, G14, G15, G16, G17 and G18). We detect unique hybridizing DNA fragments for 16 of the 17 genes in six inbred mouse strains (NOD, NON, CTS, B10, BALB/c and CBA/J) indicating that this part of the H-2 region is similar to the human MHC class III region. Using a panel of restriction enzymes we have defined RFLPs for 6 (G2, G6, HSP70, G12, G16, G18) of the 16 cross-hybridizing probes. The RFLPs demonstrate that NOD, NON and CTS mouse strains each have a distinct MHC haplotype in the MHC class III region.

Autoimmune syndromes in major histocompatibility complex (MHC) congenic strains of nonobese diabetic (NOD) mice. The NOD MHC is dominant for insulitis and cyclophosphamide-induced diabetes.

The development of autoimmune diabetes in the nonobese diabetic (NOD) mouse is controlled by multiple genes. At least one diabetogenic gene is linked to the major histocompatibility complex (MHC) of the NOD and is most likely represented by the two genes encoding the alpha and beta chains of the unique NOD class II molecule. Three other diabetogenic loci have recently been identified in the NOD mouse and are located on chromosomes 1, 3, and 11. In addition to the autoimmune diabetes which is caused by destruction of the insulin-producing beta cells in the pancreas, other manifestations of autoimmunity are seen in the NOD mouse. These include mononuclear cell inflammation of the submandibular and lacrimal glands, as well as the presence of circulating autoantibodies. To determine the effect of the non-MHC diabetogenic genes on the development of autoimmunity, we constructed the NOD.B10-H-2b (NOD.H-2b) strain, which possesses the non-MHC diabetogenic genes from the NOD mouse, but derives its MHC from the C57BL/10 (B10) strain. The NOD.H-2b strain does not develop insulitis, cyclophosphamide-induced diabetes, or spontaneous diabetes. It does, however, develop extensive lymphocytic infiltrates in the pancreas and the submandibular glands that are primarily composed of Thy 1.2+ T cells and B220+ B cells. In addition, autoantibodies are present in NOD.H-2b mice which recognize the "polar antigen" on the insulin-secreting rat tumor line RINm38. These observations demonstrate that the non-MHC genes in the NOD strain, in the absence of the NOD MHC, significantly contribute to the development of autoimmunity. The contribution of a single dose of the NOD MHC to autoimmunity was assessed with a (NOD x NOD.H-2b)F1 cross. Although only approximately 3% of F1 females developed spontaneous diabetes, approximately 50% of both female and male F1 mice developed insulitis, and 25% of females and 17% of males became diabetic after treatment with cyclophosphamide. These data demonstrate that the MHC-linked diabetogenic genes of the NOD mouse are dominant with decreasing levels of penetrance for the following phenotypes: insulitis greater than cyclophosphamide-induced diabetes greater than spontaneous diabetes.

Pancreatic islet ganglioside expression in nonobese diabetic mice: comparison with C57BL/10 mice and changes after autoimmune beta-cell destruction.

Recent observations have shown that the presumed target antigen of cytoplasmic islet cell antibodies (ICA) has properties of a monosialo-ganglioside migrating between GM2 and GM1 standards (GM2-1) and that ICA binding is higher in nonobese diabetic (NOD) than in C57BL/10SnJ mouse pancreatic frozen sections. This study aimed to characterize the ganglioside expression in NOD mouse islets in comparison with the control C57BL/10SnJ strain, taking into account possible sex differences, variations with age, and changes after autoimmune beta-cell destruction. Thus, acidic glycolipid composition was analyzed 1) in isolated islets from 11-week-old female and male NOD mice and age-matched female and male C57BL/10SnJ mice, and 2) in whole pancreas of both NOD and control mouse strains at different ages (4, 8, and 18 weeks) and of female NOD mice before and after diabetes onset. The acidic glycolipid GM2-1 is expressed in isolated female NOD islets, male NOD islets, and C57BL/10SnJ mouse islets, but quantitative analysis showed an increased amount of GM2-1 in NOD vs. C57BL/10 islets. GM3 is a ganglioside fraction expressed in female and male NOD mice and not in the C57BL/10 strain, whereas GD3 characterizes the C57BL/10 strain islets. GM2-1 is the sole ganglioside fraction in the whole pancreas to clearly decrease with age in the NOD mouse, and diabetes onset in this strain is associated with a significant decrease in the expression of this component as well as of GM3, whereas other pancreatic ganglioside (GD3, GD1a, and GT1b) levels did not significantly decrease; no age-related ganglioside change was observed in the C57BL/10SnJ mouse. Interestingly, the observed increased ICA binding in NOD islets is paralleled by the increased expression of GM2-1 islet ganglioside, and beta-cell destruction in NOD mice is associated with a significant decrease in the amount of this ganglioside in the pancreas.

Prevention of autoimmune insulitis by expression of I-E molecules in NOD mice.

The NOD (non-obese diabetic) mouse spontaneously develops insulin-dependent diabetes mellitus (IDDM) characterized by autoimmune insulitis, involving lymphocytic infiltration around and into the islets followed by pancreatic beta (beta) cell destruction, similar to human IDDM. Genetic analysis in breeding studies between NOD and C57BL/6 mice has demonstrated that two recessive genes on independent chromosomes contribute to the development of insulitis. One of the two recessive diabetogenic genes was found to be linked to the major histocompatibility complex (MHC). This is of interest, because the NOD strain has a unique class II MHC: it does not express I-E molecules as no messenger RNA for the alpha-chain of I-E is visible in Northern blot analysis; I-A molecules are not detected with any available monoclonal antibodies or by allo-reactive or autoreactive T-cell clones, although their expression is demonstrated with a conventional antiserum to Ia antigens. To examine whether the unusual expression of class II MHC molecules may be responsible for the development of autoimmune insulitis, we attempted to express I-E molecules in NOD mice selectively, without introducing other genes on chromosome 17 by using I-E-expressing C57BL/6 (B6(E alpha d)) transgenic mice. We report here that the expression of I-E molecules in NOD mice can prevent the development of autoimmune insulitis.