Absence of autoantibodies against glutamate decarboxylase (GAD) in the non-obese diabetic (NOD) mouse and low expression of the enzyme in mouse islets

Type 1 diabetes pathogenesis and the role of inhibitory receptors in islet tolerance

Type 1 diabetes (T1D) affects over a million Americans, and disease incidence is on the rise. Despite decades of research, there is still no cure for this disease. Exciting beta cell replacement strategies are being developed, but in order for such approaches to work, targeted immunotherapies must be designed. To selectively halt the autoimmune response, researchers must first understand how this response is regulated and which tolerance checkpoints fail during T1D development. Herein, we discuss the current understanding of T1D pathogenesis in humans, genetic and environmental risk factors, presumed roles of CD4+ and CD8+ T cells as well as B cells, and implicated autoantigens. We also highlight studies in non-obese diabetic mice that have demonstrated the requirement for CD4+ and CD8+ T cells and B cells in driving T1D pathology. We present an overview of central and peripheral tolerance mechanisms and comment on existing controversies in the field regarding central tolerance. Finally, we discuss T cell- and B cell-intrinsic tolerance mechanisms, with an emphasis on the roles of inhibitory receptors in maintaining islet tolerance in humans and in diabetes-prone mice, and strategies employed to date to harness inhibitory receptor signaling to prevent or reverse T1D.

Discovery of native autoantigens via antigen surrogate technology: application to type 1 diabetes

A fundamental goal in understanding the mechanisms of autoimmune disease is the characterization of autoantigens that are targeted by autoreactive antibodies and T cells. Unfortunately, the identification of autoantigens is a difficult problem. We have begun to explore a novel route to the discovery of autoantibody/autoantigen pairs that involves comparative screening of combinatorial libraries of unnatural, synthetic molecules for compounds that bind antibodies present at much higher levels in the serum of individuals with a given autoimmune disease than in the serum of control individuals. We have shown that this approach can yield "antigen surrogates" capable of capturing disease-specific autoantibodies from serum. In this report, we demonstrate that the synthetic antigen surrogates can be used to affinity purify the autoantibodies from serum and that these antibodies can then be used to identify their cognate autoantigen in an appropriate tissue lysate. Specifically, we report the discovery of a peptoid able to bind autoantibodies present in about one-third of nonobese diabetic (NOD) mice. The peptoid-binding autoantibodies were highly enriched through peptoid affinity chromatography and employed to probe mouse pancreatic and brain lysates. This resulted in identification of murine GAD65 as the native autoantigen. GAD65 is a known humoral autoantigen in human type 1 diabetes mellitus (T1DM), but its existence in mice had been controversial. This study demonstrates the potential of this chemical approach for the unbiased identification of autoantigen/autoantibody complexes.

Compartmentalization of GABA synthesis by GAD67 differs between pancreatic beta cells and neurons

The inhibitory neurotransmitter GABA is synthesized by the enzyme glutamic acid decarboxylase (GAD) in neurons and in pancreatic β-cells in islets of Langerhans where it functions as a paracrine and autocrine signaling molecule regulating the function of islet endocrine cells. The localization of the two non-allelic isoforms GAD65 and GAD67 to vesicular membranes is important for rapid delivery and accumulation of GABA for regulated secretion. While the membrane anchoring and trafficking of GAD65 are mediated by intrinsic hydrophobic modifications, GAD67 remains hydrophilic, and yet is targeted to vesicular membrane pathways and synaptic clusters in neurons by both a GAD65-dependent and a distinct GAD65-independent mechanism. Herein we have investigated the membrane association and targeting of GAD67 and GAD65 in monolayer cultures of primary rat, human, and mouse islets and in insulinoma cells. GAD65 is primarily detected in Golgi membranes and in peripheral vesicles distinct from insulin vesicles in β-cells. In the absence of GAD65, GAD67 is in contrast primarily cytosolic in β-cells; its co-expression with GAD65 is necessary for targeting to Golgi membranes and vesicular compartments. Thus, the GAD65-independent mechanism for targeting of GAD67 to synaptic vesicles in neurons is not functional in islet β-cells. Therefore, only GAD65:GAD65 homodimers and GAD67:GAD65 heterodimers, but not the GAD67:GAD67 homodimer gain access to vesicular compartments in β-cells to facilitate rapid accumulation of newly synthesized GABA for regulated secretion and fine tuning of GABA-signaling in islets of Langerhans.

Regulation of GAD65 expression by SMAR1 and p53 upon Streptozotocin treatment

Background

GAD65 (Glutamic acid decarboxylase 65 KDa isoform) is one of the most important auto-antigens involved in Type 1 diabetes induction. Although it serves as one of the first injury markers of β-islets, the mechanisms governing GAD65 expression remain poorly understood. Since the regulation of GAD65 is crucial for the proper functioning of insulin secreting cells, we investigated the stress induced regulation of GAD65 transcription.

Results

The present study shows that SMAR1 regulates GAD65 expression at the transcription level. Using a novel protein-DNA pull-down assay, we show that SMAR1 binding is very specific to GAD65 promoter but not to the other isoform, GAD67. We show that Streptozotocin (STZ) mediated DNA damage leads to upregulation of SMAR1 and p53 expression, resulting in elevated levels of GAD65, in both cell lines as well as mouse β-islets. SMAR1 and p53 act synergistically to up-regulate GAD65 expression upon STZ treatment.

Conclusion

We propose a novel mechanism of GAD65 regulation by synergistic activities of SMAR1 and p53.

Modulation of diabetes in NOD mice by GAD65-specific monoclonal antibodies is epitope specific and accompanied by anti-idiotypic antibodies

Type 1 diabetes is caused by the autoimmune destruction of pancreatic beta cells. Here we show that administration of a human monoclonal antibody (b96.11) specific to the 65-kDa isoform of glutamate decarboxylase (GAD65) to prediabetic non-obese diabetic (NOD) mice significantly delays the onset of autoimmune diabetes. We found this effect to be epitope-specific, as only b96.11 showed this therapeutic property, while a GAD65-specific human monoclonal control antibody (b78) derived from the same patient, but specific to a different determinant of GAD65, had no significant effect on the progression of disease. Administration of b96.11 or b78 to NOD mice was accompanied by the generation of anti-idiotypic antibodies. Importantly, the induced anti-idiotypic antibodies were specific for the immunizing antibody and blocked the binding of GAD65 by the respective antibody. These findings suggest a potential role for the internal image of the GAD65 determinant recognized by b96.11 in the anti-idiotypic antibody, supporting an immunomodulatory role for GAD65-specific autoantibodies, as originally postulated by Jerne.

Antibodies to ICA512/IA-2 in rodent models of IDDM

Antibodies to ICA512/IA-2 are a well established marker of human IDDM and can be detected prior to and soon after the onset of insulin dependency. The non-obese diabetic (NOD) mouse and the diabetes-prone BB rat develop spontaneous diabetes as a consequence of T-cell mediated autoimmune destruction of islet beta-cells, but the occurrence of autoantibodies is controversial. We tested sera from NOD mice and BB-rats for anti-ICA512 by a radioimmunoprecipitation assay (RIP). In sequential serum samples from 20 NOD mice, of which 15 developed diabetes, low levels of anti-ICA512 were demonstrable. Anti-ICA512 appeared close to the onset of hyperglycaemia and was usually transient. Non-diabetic NOD mice also produced anti-ICA512, but at a later age and at lower levels than the diabetic NOD mice. In a cross-sectional analysis of sera from BB rats, low levels of anti-ICA512 were present in 11/20 (55%) of non-diabetic-diabetes prone (DP) BB rats, 0/4 (0%) of diabetic DP BB rats, and 1/6 (17%) of diabetes-resistant BB rats. Anti-ICA512 was not detected in rats of other strains, including three Sprague-Dawley rats with streptozotocin-induced diabetes. In both NOD mice and BB rats the anti-ICA512 reactivity was directed to the cytoplasmic domain of the protein. The transient appearance of anti-ICA512 close to the onset of diabetes in NOD mice and the loss of these antibodies after diabetes onset is consistent with the occurrence of anti-ICA512 in human IDDM. Thus in both human IDDM and rodent models, anti-ICA512 is a marker of the impending onset of diabetes and disappears after diabetes onset.

Double-label immunofluorescence study of glutamic acid decarboxylase in the fetal and adult ovine pancreas by light and confocal microscopy: evidence for predominant beta-cell coexpression

Glutamic acid decarboxylase (GAD) is present in the central nervous system and in several nonneuronal tissues including the pancreatic islets. There are two isoforms with molecular weights of 65 kDa (GAD65) and 67 kDa (GAD67). The cellular specificity of the two molecular forms of GAD and their levels within the mammalian islets may be species-dependent, being coexpressed in both beta and in non-beta cells. We have examined the ovine pancreas, from the adult and fetal stages of late gestation, for the expression of GAD65 within the islet cells by double-label immunofluorescence light and confocal microscopy. In the adult tissue, GAD65 was colocalized in a majority of the beta cells (> 95%), with only a few glucagon and somatostatin cells (< 5%) showing immunolocalization. During the fetal stages GAD65 also showed a similar predominant beta-cell coexpression. The enzyme was also detected in a few fetal glucagon (< 5%) but not somatostatin cells. In the degenerating large fetal islets, GAD65 was also observed in the majority of the residual beta cells. These results demonstrate that in the ovine pancreas GAD65 is expressed during fetal development and is predominantly beta-cell-restricted. This pattern of expression is maintained during adult life. However, the physiological role of pancreatic GAD and/or its biosynthetic product, gamma-aminobutyric acid, in islet function in the sheep and in other ruminants remains unclear.

Islet cell antibodies: variable immunostaining of pancreatic islet cells and carcinoid tissue

OBJECTIVES

Islet cell antibodies (ICA) in sera of patients with autoimmune diabetes mellitus generally stain not only the insulin-producing beta cells but also the non-beta cells of the islets of Langerhans. The antibodies have been reported to react also with the chromaffin cells of carcinoid tissue. In the present study, we examined in detail the reactivity of 10 ICA-positive sera of patients with new onset insulin-dependent diabetes mellitus (IDDM) and two sera of patients with stiff-man syndrome.

DESIGN

The sera were analysed by immunofluorescence and by immunoperoxidase staining of human islets as well as by immunoprecipitations using 35S-methionine labelled rat islet lysates. In addition, immunofluorescence analyses of carcinoid tissues were carried out.

RESULTS

Eight of the 10 IDDM-positive sera reacted with all islet endocrine cells, whereas two sera showed staining restricted to the beta cells, as did the two sera of the patients with stiff-man syndrome. All beta cell 'selective' sera, but only 6 of 8 'whole' islet positive sera, immunoprecipitated the 64.kDa glutamic acid decarboxylase (GAD) antigen. The staining of carcinoid tissue was variable and did not correlate with the 'whole' or 'selective' staining pattern of islets.

CONCLUSION

The data underline a heterogeneity of ICA and indicate the presence of a separate, non-GAD antigen in islet cells. It is possible that in future studies, a resolution of ICA titres with respect to different types of islet cellular reactivity might provide insights into the pathogenesis of IDDM and improve the prognostic implications of antibody determinations.

Pancreatic expression of antigens for islet cell antibodies in non-obese diabetic mice

Diabetes-prone NOD mice of both sexes and at different ages were compared to control mice with regard to the level of pancreatic expression of certain autoantigens: antigens for islet cell antibodies (ICA antigens) and glutamic acid decarboxylase (GAD) 67 kDa. ICA antigens were compared by immunofluorescence using serial dilutions of ICA positive human sera so that differences of fluorescence intensity were due only to differences in amounts of antigen. Pancreatic GAD67 mRNAs were compared by polymerase chain reaction followed by Southern hybridization with 32P-probes and densitometry of autoradiographic bands. GAD67 product and gamma-aminobutyric acid (GABA) were compared by immunoperoxidase staining. As compared to BALB/c, C57BL6, Swiss, or F1 mice, NOD mice displayed higher ICA antigen levels (P < 0.01) both before and after insulitis onset (at 7 days, 15 days, 1 month, 2 months). ICA antigens were scarcely detectable by the first day of life, and increased with age from 7 days to 2 months (P < 0.01; n = 10 for each strain and at each age). Both before and after insulitis onset (4 days, 7 days, 15 days, 1 month, 2 months), amounts of GAD67 mRNAs were higher (P < 0.01) in NOD mice than in BALB/c mice (n = 8 for each age in each strain). This was already noted in foetuses on Day 18 of gestation (n = 8). After birth, amounts of GAD67 mRNAs increased up to 1 month (P < 0.04) and then decreased in older mice. The staining intensity of pancreatic sections with antisera against either GAD67 or GABA was higher (P < 0.04) in islets from NOD mice than in those from control mice. Whatever the age, no significant difference was noted between female and male NOD mice with regard to ICA antigens or GAD67. The expression of ICA antigens and GAD67 was intermediate in NOD x BALB/c F1 mice when compared to parental strains. We conclude that whatever the age, NOD mice strongly express ICA antigens and GAD67. This peculiarity was detectable very early, in embryos for GAD67 but after birth for ICA antigens. The timing of antigen expression may underlie the development of diabetes. The antigen overexpression might affect early completion of self-tolerance and, during later life, might also contribute to amplification of the anti-beta cell autoimmune response due to the existence of more targets for effector mechanisms.

Brain-reactive autoantibodies in BB/d rats do not recognize glutamic acid decarboxylase

The BB rat spontaneously develops insulin-dependent diabetes mellitus (IDDM) similar to that in humans. The most practical markers of beta cell autoimmunity are circulating antibodies to islet cell components. In particular autoantibodies to the enzyme glutamic acid decarboxylase (GAD) are a common feature of IDDM development in humans. This study aims at investigating the prevalence and levels of autoantibodies in BB rats to antigens in a semipurified, GAD-enriched preparation from rat brain. Eighteen diabetes-prone BB/d rats (10 male and eight female) were tail bled weekly from age 28 days to 113 days and antibodies detected on the rat brain preparation by ELISA. Antibody levels were expressed as arbitrary units relative to a standard positive serum. Individual rats varied in the time and order of antibody appearance and IDDM onset, with the earliest occurrence being 42 days and 69 days, respectively. In some rats antibody production was maintained but declined in others. By 113 days 85% of diabetic rats had at some time been positive for autoantibodies to brain components, compared with 25% of non-diabetics (P = 0.09 by Fisher's exact test). Immunoabsorption studies using recombinant rat GAD-65 or recombinant human GAD-67 failed to inhibit the binding of BB rat sera to the original rat brain preparation. A capture ELISA using GAD-6 MoAb to capture GAD-65 from rat brain preparation or from a preparation of recombinant rat GAD-65, failed to detect anti-GAD antibodies in BB rats. Immunofluorescent staining of tissue sections showed the autoantibodies to be brain-specific, but having distinct staining patterns to the anti-GAD antibodies of Stiff Man Syndrome serum. In conclusion, BB rats possess autoantibodies reactive with rat brain antigens which may be associated with IDDM. However, these are not directed against GAD.