Immunopathogenic role of TH1 cells in autoimmune diabetes: evidence from a T1 and T2 doubly transgenic non-obese diabetic mouse model

To improve the feasibility of in vivo monitoring of autoreactive T cells in the diabetogenic process, we generated T1 and T2 doubly transgenic non-obese diabetic (NOD) mice in which transgenic human CD90 (hCD90) is simultaneously expressed on IFN-gamma-producing cells or murine CD90.1 (mCD90.1) is expressed on IL-4-producing cells. These transgenic NOD mice develop diabetes with the same kinetics and incidence as wild type NOD mice, permitting the physiological characterization of CD4(+)hCD90(+) cells, which represent T(H)1 cells in lymphoid organs and at the site of insulitis. CD4(+)hCD90(+) cells had a higher capacity to secret IFN-gamma than CD4(+)hCD90(-) cells in an autoantigen-specific manner. Transgenic mice treated with GAD65 plasmid were protected from autoimmune diabetes, and had a lower number of CD4(+)hCD90(+) cells, confirming the pathogenic role of CD4(+)hCD90(+) cells in autoimmune diabetes. To further investigate the effect of IL-12 on the development of T(H)1 cells in autoimmune diabetes, we crossed these doubly transgenic mice to IL-12p35-deficient NOD mice. Despite severe disturbance of diabetes in p35(-/-) mice, the frequency of T(H)1 cells in these mice was slightly lower than in wild type mice. These data support the pathological role of IL-12 in autoimmune diabetes and suggest the existence an IL-12-independent pathway of T(H)1 development.     

An islet-homing NOD CD8+ cytotoxic T cell clone recognizes GAD65 and causes insulitis

T cells play a central role in the development of diabetes both in man and in the non-obese diabetic (NOD) mouse. Both the CD4(+) and CD8(+) subsets of T cells are required for the normal development of IDDM in NOD mice. Islet reactive CD4(+) T cells play a clear pathogenic role as evidenced from the isolation of diabetogenic CD4(+) T cell clones. CD8(+) T cells seem to be involved in the initiation of diabetes as lack of these cells leads to protection from diabetes. We have isolated a GAD(65) reactive, cytotoxic CD8(+) T cell clone R1 that produces large quantities of IFNgamma and accelerates the onset of insulitis. This clone proliferates and produces IFNgamma in response to GAD(65) presenting APCs and kills GAD(65) presenting targets. Furthermore, it expresses TNFalpha, CD25, CD28, CD44, CD45 and LFA1, but not CD95L This is the first example of a GAD(65)specific CD8(+) T cell clone that accelerates the onset of the insulitis, although it does not appear to accelerate the onset of diabetes.     

In vivo diabetogenic action of CD4+ T lymphocytes requires Fas expression and is independent of IL-1 and IL-18

CD4(+) T lymphocytes are required to induce spontaneous autoimmune diabetes in the NOD mouse. Since pancreatic β cells upregulate Fas expression upon exposure to pro-inflammatory cytokines, we studied whether the diabetogenic action of CD4(+) T lymphocytes depends on Fas expression on target cells. We assayed the diabetogenic capacity of NOD spleen CD4(+) T lymphocytes when adoptively transferred into a NOD mouse model combining: (i) Fas-deficiency, (ii) FasL-deficiency, and (iii) SCID mutation. We found that CD4(+) T lymphocytes require Fas expression in the recipients' target cells to induce diabetes. IL-1β has been described as a key cytokine involved in Fas upregulation on mouse β cells. We addressed whether CD4(+) T cells require IL-1β to induce diabetes. We also studied spontaneous diabetes onset in NOD/IL-1 converting enzyme-deficient mice, in NOD/IL-1β-deficient mice, and CD4(+) T-cell adoptively transferred diabetes into NOD/SCID IL-1β-deficient mice. Neither IL-1β nor IL-18 are required for either spontaneous or CD4(+) T-cell adoptively transferred diabetes. We conclude that CD4(+) T-cell-mediated β-cell damage in autoimmune diabetes depends on Fas expression, but not on IL-1β unveiling the existing redundancy regarding the cytokines involved in Fas upregulation on NOD β cells in vivo.     

Regulatory T cells contribute to diabetes protection in lipopolysaccharide-treated non-obese diabetic mice

It is well established that viral, parasitic or bacterial infections can prevent type 1 diabetes (T1D) occurrence in non-obese diabetic (NOD) mice. On the other hand, defects in CD4(+) Regulatory T cell (Treg) numbers and/or function contribute to T1D aetiology in NOD mice and in humans. In this work, we formally tested whether the protective role of the bacterial product lipopolysaccharide (LPS) on diabetes incidence results from enhanced Treg activity. We first report that weekly administration of LPS to young prediabetic NOD mice, presenting or not insulitis at the time of treatment, afforded full protection from diabetes. Taking advantage from the high but incomplete penetrance of diabetes in NOD mice raised in specific pathogen free (SPF) conditions we compared untreated disease-free old animals with gender- and age-matched LPS-treated mice. Histological and flow cytometry analysis indicated that LPS treatment did not prevent islet infiltration or priming of diabetogenic T cells but increased Foxp3(+) and CD103(+) Treg frequency and numbers. By performing adoptive transfer experiments into alymphoid NOD/SCID recipients, we further demonstrated that CD25(+) cells from LPS-treated NOD mice, but not from naturally protected animals, maintained diabetogenic cells at check. Our study suggests that T cell regulation represents a cellular mechanism to explain the 'hygiene hypothesis' and reinforces the notion that immune activity consolidates dominant tolerance.     

Immunotargeting of insulin reactive CD8 T cells to prevent Diabetes

Insulin is one of the earliest targeted autoantigens in the immune destruction of insulin-producing beta cells by autoreactive CD4 and CD8 T cells in type 1 diabetes. In this study, we used Non-obese diabetic (NOD) transgenic T cells engineered to express MHC class I-insulin peptide complexes linked to a T cell activation component (InsCD3-ζ), to target insulin-reactive CD8 T cells. We showed that activated, but not naïve, InsCD3-ζ CD8 T cells killed diabetogenic insulin-reactive CD8 target cells in vitro, inducing antigen-specific cell death mediated via both the release of perforin and the Fas–Fas ligand pathway. In vivo, InsCD3-ζ CD8 T cells migrated to the pancreatic lymph nodes of NOD mice after adoptive transfer. Concomitant with this, infiltration of CD8 T cells was also reduced in the pancreatic islets. Finally, in vivo, we showed that diabetes induced by adoptive transfer of insulin-reactive T cells was reduced following injection of activated InsCD3-ζ CD8 T cells. Furthermore, young NOD mice injected with InsCD3-ζ CD8 T cells developed a lower incidence and delayed onset of diabetes. Thus, using this novel system we have demonstrated that InsCD3-ζ CD8 T cells can directly kill insulin-reactive CD8 T cells in vitro and by targeting insulin-specific CD8 T cells early in the course of disease alter the progression of spontaneous diabetes in vivo in NOD mice.     

CD154-dependent priming of diabetogenic CD4(+) T cells dissociated from activation of antigen-presenting cells

We followed the fate of K(d)- or I-A(g7)-restricted beta cell-autoreactive T cells in monoclonal TCR-transgenic NOD mice expressing or lacking CD154. 8.3-NOD.RAG-2(-/-)/CD154(-/-) mice, which bear autoreactive CD8(+) T cells, developed diabetes with the same incidence and tempo as 8.3-NOD.RAG-2(-/-)/CD154(+) mice. Recruitment of CD154(-/-) 8.3-CD8(+) CTL was accelerated by CD154(+)CD4(+) T cells, by expression of a B7.1 transgene in beta cells or by treatment of the mice with CpG-DNA or an agonistic anti-CD40 antibody. In contrast, the autoreactive CD4(+) T cells maturing in 4.1-NOD.RAG-2(-/-) mice lost their diabetogenic potential if they lacked CD154, even in the presence of CD154(+)CD4(+) T cells, B7.1 molecules on beta cells, CpG-DNA treatment, or systemic CD40 ligation. These results demonstrate the existence of a novel, CD154-dependent pathway of CD4(+) T cell activation that is independent of CD40-mediated activation of APCs.     

Peptide p277 of HSP60 signals T cells: inhibition of inflammatory chemotaxis

Peptide p277 is a 24-amino acid fragment of the heat shock protein 60 molecule, first discovered to be an antigen for diabetogenic T-cell clones in non-obese diabetic (NOD) mice. Therapeutic vaccination with p277 can arrest the spontaneous diabetogenic process both in NOD mice and in humans associated with a T(h)1 to T(h)2 cytokine shift specific for the autoimmune T cells. We now report that p277 can directly signal human T cells via innate toll-like receptor (TLR)-2, leading to up-regulation of integrin-mediated adhesion to fibronectin, and inhibition of chemotaxis to the chemokine SDF-1alpha in vitro. Resting CD45RA(+) T cells responded to lower concentrations of p277 than resting CD45RO(+) T cells, but activation of CD45RO(+) T cells greatly increased their sensitivity to p277. Mouse T cells, but not macrophages, were also sensitive to the innate effects of peptide p277, and adoptive transfer of diabetes by splenic T cells from NOD mice could be inhibited by p277 treatment before transfer. Thus, T cells do respond innately to p277, and signaling by soluble p277 through TLR2 could contribute to the treatment of type 1 diabetes; p277 may stop the destruction of beta cells by signaling in concert both innate and adaptive receptors on T cells.     

Islet-infiltrating T cell clones from non-obese diabetic mice that promote or prevent accelerated onset diabetes

In humans and non-obese diabetic mice (NOD), insulin-dependent diabetes mellitus (IDDM) results from a spontaneous T cell-dependent autoimmune destruction of the insulin-producing pancreatic beta cells. Previous data suggest that a delicate balance between autoaggressive T cells and suppressor-type immune phenomena determine whether expression of autoimmunity is limited to insulitis or progresses to IDDM. To resolve the cellular basis of this intricate network of pathogenic CD4+ and CD8+ T cells and the role of T cells in suppressive immune phenomena. T cell clones were propagated directly from islets of NOD mice at the onset of insulitis. Insofar as insulitis, but not IDDM, is universal in NOD mice, we have screened for the in vivo effects of the islet-infiltrating T cell clones upon expression of IDDM, not insulitis. A CD4+ T cell clone, IS-3S7D, proliferates in response to islet antigen(s) and its transfer into prediabetic NOD mice promotes the rapid onset of IDDM. An interleukin 2 (IL 2)-dependent noncytolytic, V beta 11+ CD8+. T cell clones IS-2.15, prevents an accelerated onset diabetes in two distinct models. The present study, which documents the presence of CD4+ diabetogenic T cell clones and CD8+ T cell clones that dampen autoimmunity, gives tangible evidence that opposing autoimmune processes may determine whether an autoimmune-prone host develops frank disease.     

During the early prediabetic period in NOD mice,the pathogenic CD8(+) T-cell population comprises multiple antigenic specificities

In the NOD mouse model of type 1 diabetes, major histocompatibilitycomplex (MHC) class I-restricted CD8(+) T cells are essential for disease development. However, the extent of diversity of their antigenic specificities during early pathogenesis remains unclear. An insulin-derived peptide was recently identified as the epitope for the NOD-derived diabetogenic T-cell clone G9C8. To explore the possibility that the early pathogenic CD8(+) T-cell population comprises additional antigenic specificities, we employed the T-cell clones AI4 and NY8.3, both of which are pathogenic and represent specificities present in early insulitic lesions. The clones responded to distinct fractions of chromatographically separated class I MHC-bound peptides purified from NOD-derived NIT-1 beta cells, and neither clone recognized the insulin-derived peptide. NIT-1 cells represent an unlimited peptide source that will allow for the future isolation and sequencing of the novel multiple epitopes targeted early in the autoimmune response by pathogenic CD8(+) T cells.     

CD40 ligation releases immature dendritic cells from the control of regulatory CD4+CD25+ T cells

We report that disruption of CD154 in nonobese diabetic (NOD) mice abrogates the helper function of CD4+CD25- T cells without impairing the regulatory activity of CD4+CD25+ T cells. Whereas CD4+ T cells from NOD mice enhanced a diabetogenic CD8+ T cell response in monoclonal TCR-transgenic NOD mice, CD4+ T cells from NOD.CD154(-/-) mice actively suppressed it. Suppression was mediated by regulatory CD4+CD25+ T cells capable of inhibiting CD8+ T cell responses induced by peptide-pulsed dendritic cells (DCs), but not peptide/MHC monomers. It involved inhibition of DC maturation, did not occur in the presence of CD154+ T-helper cells, and could be inhibited by activation of DCs with LPS, CpG DNA, or an agonistic anti-CD40 mAb. Thus, in at least some genetic backgrounds, CD154-CD40 interactions and innate stimuli release immature DCs from suppression by CD4+CD25+ T cells.     

Impaired IL-4 production by CD8+ T cells in NOD mice is related to a defect of c-Maf binding to the IL-4 promoter

CD8(+) T cells play an important role in the induction of the autoimmune response in non-obese diabetic (NOD) mice. Here we describe abnormalities in the control of cytokine production by NOD CD8(+) T cells. NOD CD8(+) T cells had an increased propensity to produce IFN-gamma upon TCR activation, in both adult and 2-week-old mice. NOD CD8(+) T cells had a reduced capacity to produce IL-4 in type 2 conditions compared to CD8(+) T cells from the diabetes-resistant strains BALB/c and C57BL/6. Both GATA-3 and c-Maf, two positive transactivators for IL-4 gene expression, were expressed in type 2 conditions at comparable levels in NOD CD8(+) T cells. The GATA-3 was functional since normal levels of IL-5 were produced and the IL-4 promoter was hyperacetylated in NOD CD8(+) T cells. In contrast, c-Maf failed to bind to its responsive element as determined by chromatin immunoprecipitation (ChIP) assay. These results suggest that NOD CD8(+) T cells possess an increased propensity to produce IFN-gamma and impaired c-Maf-dependent DNA binding activities in vivo that lead to reduced IL-4 production following TCR activation. These defects may facilitate the development of the autoimmune response by inducing an overall type 1-biased immune response in NOD mice.     

Disruption of the homeostatic balance between autoaggressive (CD4+CD40+) and regulatory (CD4+CD25+FoxP3+) T cells promotes diabetes

Although regulatory T cells (Tregs) are well described, identifying autoaggressive effector T cells has proven more difficult. However, we identified CD4loCD40+ (Th40) cells as being necessary and sufficient for diabetes in the NOD mouse model. Importantly, these cells are present in pancreata of prediabetic and diabetic NOD mice, and Th40 cells but not CD4+CD40(-) T cells transfer progressive insulitis and diabetes to NOD.scid recipients. Nonobese-resistant (NOR) mice have the identical T cell developmental background as NOD mice, yet they are diabetes-resistant. The seminal issue is how NOR mice remain tolerant to diabetogenic self-antigens. We show here that autoaggressive T cells develop in NOR mice and are confined to the Th40 subset. However, NOR mice maintain Treg numbers equivalent to their Th40 numbers. NOD mice have statistically equal numbers of CD4+CD25+forkhead box P3+intrinsic Tregs compared with NOR or nonautoimmune BALB/c mice, and NOD Tregs are equally as suppressive as NOR Tregs. A critical difference is that NOD mice develop expanded numbers of Th40 cells. We suggest that a determinant factor for autoimmunity includes the Th40:Treg ratio. Mechanistically, NOD Th40 cells have low susceptibility to Fas-induced cell death and unlike cells from NOR and BALB/c mice, have predominantly low Fas expression. CD40 engagement of Th40 cells induces Fas expression but further confers resistance to Fas-mediated cell death in NOD mice. A second fundamental difference is that NOD Th40 cells undergo much more rapid homeostatic expansion than Th40 cells from NOR mice.     

Targeted expression of the anti-apoptotic gene CrmA to NOD pancreatic islets protects from autoimmune diabetes

The activation of apoptosis is a critical mechanism by which pancreatic beta cells are destroyed in type 1 diabetes (T1DM). Strategies aimed at interfering with the apoptotic pathways could therefore be of potential therapeutic value. To this end, we generated NOD transgenic mice with targeted expression of the anti-apoptotic gene Cytokine response modifier A (CrmA) to pancreatic beta cells using the rat insulin promoter and the reverse tetracycline transactivator to express CrmA in a temporally controlled manner. Two lines of transgenic mice were studied whose expression of CrmA occurred only after feeding doxycycline food. Islet expression of CrmA partially protected pancreatic beta cells from the cytokine-mediated cytotoxicity in vitro and reduced modestly the spontaneous development of diabetes in NOD mice in vivo. In addition, beta cells from NOD CrmA mice were significantly protected from the destruction by diabetogenic T cells after adoptive transfer. More strikingly, NODCrmA mice were significantly resistant to the diabetogenic activity of a potent insulin-specific CD8 T-cell clone. Since these adoptive transfer models mainly represent the effector phase rather than the initiation phase of autoimmune diabetes, our data suggest that the latter is more sensitive to CrmA protection. We conclude that anti-apoptotic genes such as CrmA might be potential candidates to enhance islet graft survival in T1DM.     

The preferential ability of B lymphocytes to act as diabetogenic APC in NOD mice depends on expression of self-antigen-specific immunoglobulin receptors

B lymphocytes partially contribute to autoimmune type 1 diabetes (T1D) as a subset of APC with a preferential ability to trigger pathogenic CD4 T cells. We hypothesized that this resulted from the unique ability of B lymphocytes to take up pancreatic beta cell proteins through Ig mediated capture. T1D was significantly delayed, but not prevented, in a NOD stock in which the B lymphocyte Ig repertoire was strongly restricted because of the allelic exclusion induced by transgenic Ig molecules specific for the disease irrelevant hen egg lysozyme (HEL) protein (NOD.IgHEL mice). However, introducing the Ig(mu)null mutation to eliminate the small residual numbers of non-transgenic B lymphocytes in the NOD.IgHEL stock strongly suppressed T1D to the same low levels that characterize B lymphocyte deficient NOD.Ig(mu)null mice. In contrast to standard NOD mice, both the NOD.IgHEL.Ig(mu)null and NOD.Ig(mu)null stocks were unable to generate T cell responses against the candidate diabetes autoantigen, glutamic acid decarboxylase. These results indicate that Ig-mediated capture of beta cell autoantigens accounts for why B lymphocytes have a greater capacity than other APC subtypes to trigger diabetogenic T cells. Hence, defects in B lymphocyte, as well as T lymphocyte, tolerance induction mechanisms may contribute to T1D in NOD mice.     

IFN-γ receptor deficiency prevents diabetes induction by diabetogenic CD4(+) , but not CD8(+) , T cells

IFN-γ is generally believed to be important in the autoimmune pathogenesis of type 1 diabetes (T1D). However, the development of spontaneous β-cell autoimmunity is unaffected in NOD mice lacking expression of IFN-γ or the IFN-γ receptor (IFNγR), bringing into question the role IFN-γ has in T1D. In the current study, an adoptive transfer model was employed to define the contribution of IFN-γ in CD4(+) versus CD8(+) T cell-mediated β-cell autoimmunity. NOD.scid mice lacking expression of the IFNγR β chain (NOD.scid.IFNγRB(null) ) developed diabetes following transfer of β cell-specific CD8(+) T cells alone. In contrast, β cell-specific CD4(+) T cells alone failed to induce diabetes despite significant infiltration of the islets in NOD.scid.IFNγRB(null) recipients. The lack of pathogenicity of CD4(+) T-cell effectors was due to the resistance of IFNγR-deficient β cells to inflammatory cytokine-induced cell death. On the other hand, CD4(+) T cells indirectly promoted β-cell destruction by providing help to CD8(+) T cells in NOD.scid.IFNγRB(null) recipients. These results demonstrate that IFN-γR may play a key role in CD4(+) T cell-mediated β-cell destruction.     

Immune modulation induced by tuberculosis DNA vaccine protects non-obese diabetic mice from diabetes progression

We have described previously the prophylactic and therapeutic effect of a DNA vaccine encoding the Mycobacterium leprae 65 kDa heat shock protein (DNA-HSP65) in experimental murine tuberculosis. However, the high homology of this protein to the corresponding mammalian 60 kDa heat shock protein (Hsp60), together with the CpG motifs in the plasmid vector, could trigger or exacerbate the development of autoimmune diseases. The non-obese diabetic (NOD) mouse develops insulin-dependent diabetes mellitus (IDDM) spontaneously as a consequence of an autoimmune process that leads to destruction of the insulin-producing beta cells of the pancreas. IDDM is characterized by increased T helper 1 (Th1) cell responses toward several autoantigens, including Hsp60, glutamic acid decarboxylase and insulin. In the present study, we evaluated the potential of DNA-HSP65 injection to modulate diabetes in NOD mice. Our results show that DNA-HSP65 or DNA empty vector had no diabetogenic effect and actually protected NOD mice against the development of severe diabetes. However, this effect was more pronounced in DNA-HSP65-injected mice. The protective effect of DNA-HSP65 injection was associated with a clear shift in the cellular infiltration pattern in the pancreas. This change included reduction of CD4(+) and CD8(+) T cells infiltration, appearance of CD25(+) cells influx and an increased staining for interleukin (IL)-10 in the islets. These results show that DNA-HSP65 can protect NOD mice against diabetes and can therefore be considered in the development of new immunotherapeutic strategies.     

Adoptive transfer of islet antigen-autoreactive T cell clones to transgenic NOD.Ea(d)mice induces diabetes indicating a lack of I-E mediated protection against activated effector T cells

Transgenic insertion of the MHC class II Ea(d)gene in NOD mice restores I-E expression and prevents T-cell-mediated autoimmune diabetes (IDDM). The specific molecular and cellular mechanisms responsible for the diabetes resistance of transgenic NOD.Ea(d)mice remain unclear. We adoptively transferred islet antigen-specific T cell clones into NOD and transgenic NOD.Ea(d)mice to evaluate the level of protection provided by I-E expression against activated effector T cells. We have found that neither neonatal or 3-5-week-old I-E-expressing NOD.Ea(d)mice can completely inhibit the diabetogenic activities of activated islet antigen-specific T cell clones. These data indicate that Ealpha protein expression in NOD antigen presenting cells (APC) does not reduce islet autoantigen presentation in the context of I-A(g7)below the threshold required for stimulation of effector/memory diabetogenic T cells. Our results suggest that the mechanism of Ealpha protein-mediated diabetes resistance in NOD mice may be "antigen ignorance," in which the quantity of islet autoantigens presented in the context of I-A(g7)by APC is reduced below the threshold required to activate nai;ve islet antigen-specific T cells.     

Development of autoreactive diabetogenic T cells in the thymus of NOD mice

Type 1 diabetes results from destruction of pancreatic beta cells by beta cell-specific autoreactive T cells in the nonobese diabetic (NOD) mouse. Defects in thymic negative selection are thought to result in failure to delete potential beta cell-reactive T cells, contributing to the development of autoimmune diabetes. We investigated this possibility by comparing the deletion profile of double-positive (DP) thymocytes in NOD mice with diabetes-resistant strains of mice after anti-CD3 Ab treatment to trigger the TCR-mediated signaling pathway. We found that immature NOD CD4+CD8+ DP thymocytes have a lower activation threshold than C57BL/6 and Balb/c thymocytes. This was confirmed by showing that NOD DP thymocytes have a higher level of ERK and JNK phosphorylation. The low activation threshold of immature thymocytes resulted in rapid deletion of strongly activated immature DP thymocytes by negative selection, whereas weakly activated immature thymocytes differentiated more efficiently into CD69+CD3high DP thymocytes by positive selection. SP thymocytes, particularly CD4-CD8+ T cells that were efficiently generated from activated DP thymocytes, could induce severe insulitis and diabetes in NOD.scid mice. We conclude that the development of autoreactive diabetogenic T cells results from inordinate positive selection due to the low activation threshold of DP thymocytes in NOD mice.     

CD8+ T lymphocytes specific for glutamic acid decarboxylase 90-98 epitope mediate diabetes in NOD SCID mouse

During the past decade, glutamic acid decarboxylase (GAD) has been considered a crucial beta-cell autoantigen involved in type 1 diabetes in the NOD mouse and human. Recently, the etiological role of GAD has remained controversy. In the NOD mouse, some previous studies argued in favor of a regulatory role for GAD-specific CD4+ T cells, and no diabetogenic CD8+ T cells specific for GAD have been identified so far, discrediting the importance of GAD in beta-cell injury. Here, we identified, in the NOD model, a relevant GAD CD8+ T cell epitope (GAD(90-98)) using immunization with a plasmid encoding GAD, a protocol relying on in vivo processing of peptides from the autoantigenic protein. In pancreatic lymph nodes of na?ve female NOD mice, CD8+ T lymphocytes recognizing GAD(90-98) peptide were detected during the initial phase of invasive insulitis (between 4 and 8 weeks of age), suggesting an important role for these cells in the first stage of the disease. GAD(90-98) specific CD8+ lymphocytes lysed efficiently islet cells in vitro and transferred diabetes into NOD(SCID) mice (100%). Finally, diabetes was accelerated greatly in 3-week-old female NOD mice injected i.p. with GAD(90-98), strengthening the role of GAD-specific CTLs in diabetes pathogenesis.