Minute defects in the expression of MHC E molecules lead to impaired protection from autoimmunity in NOD mice.

The E complex of the major histocompatibility complex (MHC) can prevent the spontaneous development of diabetes in nonobese diabetic (NOD) mice transgenic for the Ea gene. None of three promoter-mutated Ea constructs with Ea expression directed to different subsets of immunocompetent cells exerts full protection in NOD mice. The promoter-mutated constructs are all capable of mediating intrathymic elimination of I-E-restricted T cells. Thus, thymic negative selection is not responsible for the protective effect but a more complex effect is likely. Here we show that combinations of two or three different mutated Ea constructs do not protect against intra-islet insulitis either. We also show that spleen cells from protected animals are sufficient to protect NOD mice in adoptive transfer experiments. The only detectable expression defects in splenic cells or cells influencing the repertoire of splenic cells are in the B-cell compartment. Furthermore, in three construct combinations, the differences to wild-type expression are extremely small. Thus, we conclude that even minute disturbances of the E expression pattern might reduce the protection of NOD mice from insulitis.

Monokine-producing cells predominate in the recruitment phase of NOD insulitis while cells producing Th1-type cytokines characterize the effector phase.

Cells infiltrating the Langerhans' islets of prediabetic NOD females were isolated from 6 weeks to 6 months of age. These cells were assayed at a single-cell level for production of eight different cytokines by intracellular immunofluorescent staining. Quiescent in vivo preactivated cells were detected by in vitro stimulation with PMA and ionomycin for 4 h. The cell recruitment phase, between 6 and 12 weeks of age, is predominated by production of the monokines IL-1alpha, IL-6, and TNF After stimulation IFN-gamma and occasional IL-10 and GM-CSF producing cells could also be observed. This cytokine pattern occurs simultaneously with increasing insulitis, and we suggest that these cytokines are important in attracting inflammatory cells to the islets and maintaining the inflammatory state. A high frequency of endocrine cells producing IL-6 during this period may denote a stress response caused by initial beta-cell destruction due to cytokines released by the inflammatory cells. During the effector phase, between 4 and 6 months, there is a characteristic Th1 cytokine profile with lymphocytes producing IL-2, IFN-gamma and TNF, supposedly TNF-beta. No IL-4 production could be detected and IL-10 was very rarely found, indicating the absence of a Th2 response. Our findings show that the effector phase in NOD insulitis is a Th1 rather than a Th2-mediated event. We also demonstrate that cytokines that may cause initial tissue destruction are produced during the recruitment of inflammatory cells.

Alleviation of insulitis in NOD mice is associated with expression of transgenic MHC E molecules on primary antigen-presenting cells.

Major histocompatibility complex (MHC) class II genes are important in the pathogenesis of insulin-dependent diabetes mellitus (IDDM) both in the mouse and in man. The non-obese diabetic (NOD) mouse, which is a good model for human IDDM, has a particular MHC class II with an A complex consisting of A alpha d and the unique A beta g7 chain, as well as an absent E molecule due to a deletion in the Ea promoter region. Transgenic insertion of a functional Ea gene protects against insulitis and diabetes, but when the transgene expression is restricted to certain compartments of the immune system by deleting parts of the promoter region, the protection against insulitis is disrupted. We have analysed three promoter-mutated lines where one lacks expression on B cells and has a reduced expression on approximately 1/3 of the dendritic cells and macrophages (Sma), one lacks thymic cortical expression and has a slightly reduced B-cell expression (delta X), and one lacks expression in the thymic medulla, on macrophages, dendritic cells and about half of the B cells (delta Y). None of these lines is protected against insulitis, but Sma and delta X display a reduced intensity of insulitis, with an average of 10-15% of the islets infiltrated in each mouse, while delta Y resembles non-transgenic mice with 30-35% infiltrated islets. Bone-marrow chimeras between Sma and delta Y mice demonstrate that peripheral cells of Sma origin reduce insulitis significantly when developed in the delta Y host, while insulitis is enhanced when delta Y bone marrow is given to Sma mice. This shows that E expression on the primary antigen-presenting macrophages and dendritic cells is of crucial importance to the alleviation of insulitis.

T cells are unresponsive to transgenic MHC class II molecules expressed on pancreatic beta cells.

It has been proposed that the autoimmune attack on the pancreatic beta cells leading to insulin-dependent diabetes mellitus can be caused by the expression of MHC class II molecules on the beta cells. Transgenic mice expressing normal levels of allogeneic MHC class II Ak on the beta-cell surface (IP-Ak) do not develop either insulitis or diabetes, yet these mice are not tolerant to Ak when expressed on normal antigen-presenting cells. The authors have stimulated T cells from IP-Ak mice in vitro with Ak-expressing beta cells. Mice were also primed in vivo in order to facilitate the antiallogeneic response. The authors found that neither IP-Ak positive nor IP-Ak negative mice were able to respond to Ak-expressing beta cells, and that in vivo priming does not overcome this inability. They suggest that beta cells do not act as antigen-presenting cells, probably due to inability of delivering costimulatory signals. This strengthens the notion that MHC class II expression per se is not sufficient to induce an autoimmune attack on the beta cells.

Demonstration of a TH1 cytokine profile in the late phase of NOD insulitis.

Cells infiltrating the Langerhans' islets of prediabetic NOD females were isolated from 6 weeks to 6 months of age. These cells were assayed at a single-cell level for production of eight different cytokines by intracellular immunofluorescent staining. By in vitro stimulation with PMA and ionomycin for 4 hours the method is enhanced also to detect in vivo preactivated cells. During the early phase of insulitis from 6 to 12 weeks of age, mainly the monokines IL-1 alpha, IL-6, and TNF were detected. After stimulation, also IFN-gamma and low numbers of IL-10 and GM-CSF producing cells could be observed, but no IL-2 or IL-4 was seen. This cytokine pattern correlates with an increasing insulitis, and we suggest that these cytokines are important in attracting inflammatory cells to the islets, and may cause initial beta-cell destruction. During a later phase, between 4 and 6 months, there is a characteristic TH1 cytokine profile with production of IL-2 and IFN-gamma occurring after stimulation, as well as lymphocytes producing TNF, supposedly TNF-beta. During this period IL-10 was very rarely observed, and no IL-4 production could be found throughout the study. This indicates the absence of a TH2 cytokine profile in this lesion. In addition IL-6 production occurs in high frequencies at all ages, also in endocrine islet cells. We interpret this as a stress response caused by the inflammatory lesion. Our findings show that the effector phase in NOD insulitis is TH1 rather than TH2 mediated. We also demonstrate that cytokines, that may cause initial tissue destruction, are produced during the recruitment of inflammatory cells.

No evidence for TCR V beta repertoire changes influencing disease protection in E-transgenic NOD mice.

In order to study whether positive selection of T cells plays any role in the MHC-dependent protection from diabetes in the non-obese-diabetic (NOD) mouse, the T cell V beta repertoire has been studied in NOD mice and in NOD mice either transgenic for the wildtype MHC class II E alpha gene, or for delta Y, a promotor-mutagenized E alpha gene with a restricted tissue expression. The E alpha transgenic line is protected from both insulitis and diabetes. The delta Y transgenic line is neither protected from insulitis nor from diabetes, although it can perform both positive and negative E-mediated selection in the thymus. The V beta repertoire was studied in the pancreatic lymph nodes as these drain the area which is the target for the autoimmune attack. We see no evidence for E alpha TCR V beta repertoire differing from both nontransgenic NOD mice and delta Y mice despite its striking difference in susceptibility to autoimmunity. We conclude that none of the differences in the TCR V beta repertoire of E alpha-transgenic NOD mice hitherto observed are likely to explain the protective effect of E molecule expression in NOD mice.