The Gly⁹⁷²→Arg IRS-1 Variant Is Associated With Type 1 Diabetes in Continental Italy

Genetic variance in the insulin receptor substrate-1 (IRS)-1 is thought to play a key role in the insulin resistance that characterizes type 2 diabetes (1–6). Transfection studies have demonstrated that the most common IRS-1 variant, Arg⁹⁷², which involves a Gly 224 Arg substitution at codon 972, impairs insulin signaling via the phosphatidylinositol-3 (PI3)-kinase pathway (7), and in some (but not all) studies, this variant has been found with an increased frequency among type 2 diabetic patients (1–6).

Interestingly, carriers of the Arg⁹⁷² substitution have been found to have lower fasting insulin and C-peptide levels than noncarriers (8,9), suggesting that this IRS-1 variant might also play a role in the secretory capacity of the β-cells themselves. Indeed, impaired insulin secretion has also been observed in rat insulinoma (RIN) cells overexpressing the Arg⁹⁷² IRS-1 polymorphism (10) in human islets naturally carrying the variant (11), and even in normal glucose-tolerant subjects with the Arg972 variant (9). These observations raise the intriguing hypothesis that genetic defect in the IRS-1/PI3 kinase pathway might also be involved in the inadequate insulin secretion that characterizes type 2 diabetes.

More recent studies suggest that the Arg⁹⁷² IRS-1 variant also plays a role in β-cell survival (12). We have found that human Arg⁹⁷² islets contain a significantly higher number of apoptotic cells than their wild-type counterparts, and they are also resistant to the antiapoptotic effects of insulin (12). It has been speculated that apoptosis plays a crucial role in the autoimmune destruction of β-cells characterizing type 1 diabetes (13). An increase in apoptosis might have pathological consequences in diabetes-prone individuals, who have an autoreactive T-cell repertoire that can be mobilized by the exposed β-cell antigens. These considerations prompted us to examine the possible relations between the Arg⁹⁷² IRS-1 variant and type 1 diabetes.

We evaluated the frequency of the IRS-1 Arg⁹⁷² variant in 307 unrelated type 1 diabetic patients (age at onset: 0–14 years) and 243 unrelated control subjects from a single region in central Italy. Control subjects had no history of type 2 diabetes, type 1 diabetes, or other autoimmune diseases. The results of this case-control analysis are shown in Table 1. Arg⁹⁷² was significantly more common in type 1 diabetic patients than in control subjects (P = 0.0008), conferring an OR of 2.5. The frequency of IRS-1 Arg⁹⁷² among patients with HLA genotypes representing moderate-to-high risk for type 1 diabetes was not significantly different from that of the low-risk genotype subgroup of patients. The Arg972 IRS-1 variant was associated with a positive predictive value (PPV) for onset of type 1 diabetes, among children 15 years of age, of 0.32%, and it increased the PPV of the high-risk genotype DR3/DR4-DQB1*0302 by 3% (from 4 to 7%). Transmission disequilibrium test (TDT) of 140 type 1 diabetes simplex families from the same area revealed that the Arg⁹⁷² IRS-1 variant was transmitted from heterozygous parents to affected probands at a frequency of 70.2% (P < 0.02), whereas the rate of transmission to unaffected siblings was not significantly different from the expected 50% (Table 2).

This surprisingly strong association between the Arg⁹⁷² IRS-1 variant and type 1 diabetes is at variance with previous reports (14,15), including that of Cox et al. (16), which involved genome screening of 767 type 1 diabetes multiplex families. In fact, the latter study revealed no linkage between the disease and region 2q36, where the IRS-1 gene has been mapped. It is important to recall, however, that at the present time single genes with limited effects are more likely to be identified by population-based case-control studies and linkage disequilibrium analysis within families, which were used in the present study. The fact that the Caucasian population studied by Cox et al. was ethnically mixed, whereas our analysis was confined exclusively to continental Italians, may also have contributed to the discrepancy between our findings.

The Arg⁹⁷² variant also displayed marginal correlation (P = 0.051) with fasting insulin levels (reflected by C-peptide concentrations) at the time of diagnosis. These data were available for only 168 of the 307 patients, and it is possible that a stronger correlation might have emerged if this variable had been evaluated in a larger population. Nevertheless, our finding is compatible with previous observations indicating that the Arg⁹⁷² variant negatively influences the secretory function of the β-cell (9–12).

The significance of a defect of this type in type 1 diabetes is difficult to determine. Whereas type 2 diabetes is clearly characterized by gradual loss of insulin secretion, the insulin deficit in type 1 diabetes is the result of more or less massive and rapid autoimmune destruction of the β-cell mass (1,13). However, a recent report indicates that in the early stages of the disease, certain type 1 diabetic patients show some degree of β-cell dysfunction, manifested primarily by hyperglycemia after glucose loading due to a defective insulin secretion (17). Interestingly, a similar impairment in insulin secretion has also been observed in the IRS-1 null mice, both in vivo and in vitro (18). Our findings confirm the growing body of evidence reflecting the negative effects of IRS-1 Arg972 on the ability of pancreatic β-cells to secrete insulin. This impairment might be related to the increased apoptosis and defective maturation of insulin granules, which we have recently documented in human pancreatic islets with the Arg972 IRS-1 variant (11,12). In fact, some investigators have suggested that increased β-cell apoptosis might function as a trigger for the autoimmune response in type 1 diabetes (13). Insulin has potent antiapoptotic effects that may depend on activation of the PI3 kinse/Akt pathway, and it is reasonable to assume that any significant reduction in insulin secretion could lead to an increase in “physiological” β-cell death. Alternatively, accumulation of immature insulin granules (as a result of the defect mentioned above) could function as an independent stimulus of apoptosis by provoking endoplasmic reticular stress, as recently observed (19). Regardless of it’s cause(s), however, a substantial increase in the rate of β-cell apoptosis might exceed the capacities of the macrophage clearance system, and, in a genetically predisposed individual, exposure of β-cell antigens released from the remaining cells might be sufficient to trigger an autoimmune reaction (13).

The hypotheses discussed above are highly speculative, and the association we have found between IRS-1 Arg972 is obviously a preliminary finding that will have to be verified in other populations. The Arg972 variant of the IRS-1 seems to play a complex role in the pathogenesis of type 2 diabetes, affecting both peripheral insulin sensitivity and the functional capacities of the pancreatic β-cells themselves. In light of our findings, it is possible to speculate that the same mechanisms—in the presence of a genetically determined predisposition—might also result in or contribute to the clinical manifestations of type 1 diabetes.