To the Editor: IA-2 is a major autoantigen in Type 1 diabetes mellitus [1]. Autoantibodies to IA-2 are detected in approximately 70% of newly diagnosed Type 1 diabetes patients and in less than 1% of the control subjects. Based on sequence analysis, IA-2 is an enzymatically inactive transmembrane protein tyrosine phosphatase and is located in secretory vesicles of neuroendocrine cells, particularly pancreatic islets and brain cells. IA-2 knockout studies in mice showed elevated glucose tolerance tests and depressed insulin release [2]. Islets isolated from IA-2−/− mice showed less insulin release as compared to islets from IA-2+/+ mice when the cultures were switched from basal (3.3 mmol/l) to high glucose (27.7 mmol/l) concentrations. These and other studies argue that IA-2 is involved in insulin secretion.

The non-obese diabetic (NOD) mouse is the most widely studied animal model for Type 1 diabetes [3]. Gene linkage analyses have identified at least 19 susceptibility loci (Idd1-Idd19). The murine IA-2 gene (Ptprn) is located on chromosome 1 (41 cM) and is a potential candidate gene for Idd5.2 [4]. The role of IA-2 in the NOD mouse model, however, has not been determined.

In the present study, IA-2+/− mice (129/Sv x C57BL/6) were backcrossed to NOD/LtJ mice (Jackson Laboratory, Bar Harbor, Me., USA) for eight generations. IA-2+/− NOD mice then were intercrossed to obtain IA-2−/− NOD mice. All protocols were approved by our Institutional Animal Care and Use Committees. The IA-2 genotypes were determined by PCR using specific primers for the targeted locus and for the wild-type locus [2]. IA-2 protein expression in brain, analysed by Western blot, showed the presence of IA-2 in IA-2+/+ NOD mice, but its absence in IA-2−/− NOD mice (Fig. 1). IA-2+/+ NOD and IA-2−/− NOD mice then were injected with cyclophosphamide, a known accelerator of diabetes in NOD mice [5]. Animals were followed twice a week and mice with two consecutive non-fasting blood glucose determinations with values above 250 mg/dl were scored as diabetic. As seen in Fig. 1, 44.4% of IA-2+/+ NOD and 36.4% of IA-2−/− NOD mice developed diabetes. Histologic examination of pancreatic islets from these animals showed insulitis. The number of IA-2−/− NOD mice in our colony is still small, but five mice between 20 and 40 weeks of age already have developed diabetes spontaneously. Recently, we succeeded in knocking out IA-2β, a protein closely related to IA-2, and its role in the development of diabetes in NOD mice is currently under investigation.

Fig. 1
figure 1

Development of diabetes in IA-2−/− NOD mice. Eight-week-old IA-2+/+ NOD (■, n=18) and IA-2−/− NOD (□, n=22) female mice were treated (arrows) with cyclophosphamide (250 mg/kg i.p.) on days 0 and 14 and the percent that developed diabetes was determined. Insert shows Western blot

Full size image

The knockout studies described here show that IA-2 is not required for the development of diabetes in NOD mice and earlier knockout studies showed that glutamic acid decarboxylase 65 (GAD65) also is not required for the development of diabetes in NOD mice [6]. Although there is some evidence in NOD mice that these antigens could be involved in a cell-mediated immune response [7], little if any autoantibodies are made to IA-2 or GAD65 in NOD mice [8]. In contrast to NOD mice, close to 90% of newly diagnosed patients with Type 1 diabetes have autoantibodies to IA-2 and/or GAD65 and these autoantibodies have become important predictive markers [9]. Thus, there are major differences between the NOD mouse model and the human disease. Although IA-2 is not involved in the diabetes of NOD mice, the extent to which the autoimmune response to IA-2 may contribute to the pathogenesis of Type 1 diabetes in humans has yet to be solved.