The Arg⁹⁷² Variant in Insulin Receptor Substrate-1 Is Associated With an Increased Risk of Secondary Failure to Sulfonylurea in Patients With Type 2 Diabetes

The study group consisted of 477 unrelated Caucasian type 2 diabetic patients who were consecutively recruited according to the following criteria: men and women with onset of diabetes after age 35 years, absence of ketonuria at diagnosis, and anti-GAD⁻ antibody. Type 2 diabetes was diagnosed according to the American Diabetes Association criteria (21). Patients were excluded if they had chronic gastrointestinal diseases associated with malabsorption, chronic pancreatitis, history of any malignant disease, history of alcohol or drug abuse, liver or kidney failure, clinical problems potentially causing hyperglycemia including infection, thyroid disease, surgery, and treatments able to modify glucose metabolism such as corticosteroids or estrogens. In addition to appropriate diet, all patients were treated with second- or third-generation sulfonylureas at >15 mg/day (i.e., for most patients glibenclamide 5 mg, three tablets/day) until their plasma glucose rose to >300 mg/dl or they developed hyperglycemic symptoms (typically thirst or polyuria). At this point, metformin was added up to 1,700 mg/day in most patients. Patients with secondary sulfonylurea failure were defined as those requiring insulin due to uncontrolled hyperglycemia (fasting plasma glucose >300 mg/dl) despite sulfonylurea-metformin combined therapy, appropriate diet (i.e., iso- or hypocaloric diet depending on presence of overweight), and absence of any conditions causing hyperglycemia. The median duration of therapy with oral agents before failure was 10 years. The ethical committee approved the protocol, and informed written consent was obtained from all participants. All the investigations were performed in accordance with the principles of the Declaration of Helsinki.

Fasting blood glucose, total and HDL cholesterol, and triglycerides were measured by enzymatic methods (Roche Diagnostics, Mannheim, Germany). Blood HbA_1c concentrations were analyzed by high-pressure liquid chromatography (Menarini Diagnostics, Italy) with coefficients of variation <4%.

Genomic DNA was isolated from peripheral blood according to standard procedures. The Gly972Arg of the human IRS-1 sequence was detected by digestion of PCR products with restriction enzyme BstNI as previously described (14,20,22). The Pro12Ala polymorphism in exon B of the peroxisome proliferator–activated receptor (PPAR)-γ gene was detected by digestion of PCR products with restriction enzyme BstUI as previously described (23).

Fisher’s exact test was used to test for differences in genotype frequencies. Categorical variables were compared by χ² test. Continuous data are shown as means ± SD. Because only one subject was found to be homozygous for the Arg⁹⁷² IRS-1 variant, she was combined with the heterozygous group for statistical analysis. Non-normally distributed parameters were logarithmically transformed to approximate a normal distribution. Differences between means were compared using unpaired Student’s t test. All differences were also tested after adjusting for sex, age, and BMI. The Hardy-Weinberg equilibrium between the two genotypes was evaluated by χ² test. The OR for a possible specific influence of the IRS-1 genotype on secondary failure to sulfonylurea was calculated using a logistic regression analysis while adjusting for age, sex, BMI, degree of glycemic control, age at diagnosis, duration of diabetes, and Pro12Ala polymorphism of the PPAR-γ2 gene. All tests were two-sided, and P < 0.05 was considered statistically significant. All analyses were performed using SPSS version 10.0 software program for Windows.

The clinical characteristics of the study subjects stratified according to therapy and genotype are reported in Table 1. As expected, patients with secondary failure to sulfonylurea showed significant differences in age (P < 0.001), duration of the disease (P < 0.0001), age of clinical diagnosis (P < 0.0001), fasting plasma glucose levels (P < 0.0001), and HbA_1c concentrations (P < 0.03), as compared with patients well controlled with oral therapy. Of the total patients, 53 (11.1%) were heterozygous for the Arg⁹⁷² IRS-1 variant, 1 (0.2%) was homozygous, and the remainder (88.7%) were homozygous for the wild-type allele. The genotype frequency was in the Hardy-Weinberg equilibrium. The genotype frequency of the Arg⁹⁷² IRS-1 variant was 8.7% among diabetic patients well controlled with oral therapy and 16.7% among patients with secondary failure to sulfonylurea. Thus, the Arg⁹⁷² IRS-1 variant was associated with secondary failure to sulfonylurea with carriers having a relative risk of 2.1 (95% CI 1.18–3.70, P = 0.01). Adjustment for age, sex, BMI, degree of metabolic control estimated as fasting plasma glucose and HbA_1c concentrations, age at diagnosis, and duration of diabetes in a logistic regression analysis with secondary failure to sulfonylurea as a dependent variable did not change this association (OR 2.2 [95% CI 1.16–4.46], P = 0.016). To further exclude the influence of duration of diabetes in our study population, we analyzed subgroups matched for sex, age, BMI, and duration of diabetes (50 men/33 women vs. 49 men/48 women, P = 0.19; 64.9 ± 10 vs. 67 ± 8 years of age, P = 0.09; 29.0 ± 5.5 vs. 29.5 ± 2.7 kg/m², P = 0.51; 20.3 ± 4.7 vs. 21.1 ± 5.1 years of diabetes duration, P = 0.51; respectively, for the diabetic patients well controlled with oral therapy and patients with secondary sulfonylurea failure). The genotype frequency of the Arg⁹⁷² IRS-1 variant was 7.2% among diabetic patients well controlled with oral therapy and 17.5% among patients with secondary failure to sulfonylurea. Thus, also in this subgroup analysis, which takes into account major confounding factors, the Arg⁹⁷² IRS-1 variant was associated with secondary failure to sulfonylurea with carriers having a relative risk of 2.7 (1.02–7.28, P = 0.045). There were no differences in clinical and metabolic characteristics between patients carrying the Arg⁹⁷² variant and homozygous wild-type patients within each of the two groups, i.e., patients well controlled with oral therapy and patients with secondary failure to sulfonylurea (Table 1). It has been reported that the Ala allele of the Pro12Ala polymorphism in the PPAR-γ2 gene is associated with a decreased risk of type 2 diabetes (24,25). These results have been confirmed in some, but not all, studies (26–30). PPAR-γ is the target of thiazolidinediones, a new class of insulin-sensitizing agents used clinically to treat type 2 diabetes. In vitro, the Pro12Ala change is associated with reduced transcriptional activity of PPAR-γ2 (24), thus raising the possibility that the Pro12Ala variant could influence the response to oral hypoglycemic agents. Although no patient in our study was treated with thiazolidinediones because these drugs were not available in Italy at the time of recruitment, we analyzed the prevalence of Pro12Ala variant in our cohort of patients. The genotype frequency of the Ala variant was 12.3% among diabetic patients well controlled with oral therapy and 17.6% among patients with secondary failure to sulfonylurea. Addition of the Pro12Ala polymorphism to the logistic regression analysis model, including as confounding factors age, sex, BMI, degree of metabolic control, age at diagnosis, and duration of diabetes, did not change the association of the Arg⁹⁷² IRS-1 variant with secondary failure to sulfonylurea (2.0 [1.38–3.86], P = 0.038). These results extend previous data showing that the Pro12Ala polymorphism in PPAR-γ does not affect the therapeutic response to pioglitazone (31) and indicate that this common variant is unlikely to be responsible for secondary failure to oral hypoglycemic agents.

Secondary sulfonylurea failure is a common evolution of type 2 diabetes. According to this view, 30% of newly diagnosed diabetic patients randomly assigned to receive sulfonylureas in the U.K. Prospective Diabetes Study had been switched to insulin therapy because initially effective sulfonylurea therapy failed within the first 6 years of therapy (2). It has been shown that a progressive deterioration of β-cell function is a better predictor of secondary sulfonylurea failure than other factors, including duration of diabetes, young age at diagnosis, BMI, insulin sensitivity, or glycemic control (5,6). However, genetic factors determining the progressive β-cell dysfunction that leads to sulfonylurea failure remain unsettled. Because previous studies have demonstrated that the common Arg⁹⁷² IRS-1 variant may affect insulin secretion to sulfonylurea (19,20), in the present study we investigated the relationship between the Arg⁹⁷² variant and the risk for secondary failure to sulfonylurea in an Italian cohort of type 2 diabetic patients. We found that the Arg⁹⁷² IRS-1 variant was associated with secondary failure to sulfonylurea with patients carrying this variant having a relative risk of 2.0 as compared with noncarriers also after adjustment for age, sex, BMI, metabolic control, age at diagnosis, duration of diabetes, and Pro12Ala polymorphism of the PPAR-γ2 gene.

It has been reported that patients with apparent secondary failure to sulfonylurea are positive for anti-GAD antibody, and they tend to develop insulin deficiency requiring insulin treatment in a short period (4,32). Interestingly, we have recently reported that the frequency of the Arg⁹⁷² IRS-1 variant is significantly increased in patients with type 1 diabetes compared with nondiabetic individuals, conferring an OR of 2.5 (22). However, the possibility that patients with type 1 diabetes or latent autoimmune diabetes in adults might account for the increased prevalence of the Arg⁹⁷² IRS-1 variant among patients with secondary failure to sulfonylurea was excluded a priori by recruiting exclusively anti-GAD⁻ patients.

There is evidence that the type of sulfonylurea may influence the development of secondary failure (3). Unfortunately, our data do not allow comparisons between the different types of sulfonylurea due to the small subgroups taking different sulfonylureas. Further longitudinal studies comparing the different types of sulfonylurea are needed to address this issue.

Sulfonylureas stimulate insulin release from pancreatic β-cells through the interaction with a specific plasma membrane receptor (SUR1) coupled to the ATP-sensitive K⁺ channel, leading to its closure. This causes depolarization of the β-cell membrane, leading to opening of voltage-sensitive Ca²⁺ channels, Ca²⁺ influx, and a rise in intracellular Ca²⁺, which stimulates exocytosis of insulin-containing granules. Several experimental studies support the notion that IRS-1 plays an important role in insulin secretion by regulating intracellular calcium homeostasis (8–11). Overexpression of IRS-1 in βTC6-F7 or RIN 1046-38 insulinoma cell lines resulted in an increased fractional insulin secretion in response to glucose or sulfonylurea (8,19). This event was associated with an increase in cytosolic Ca²⁺ concentrations due to inhibition of Ca²⁺ uptake by the endoplasmic reticulum through the activation of PI 3-kinase associated with IRS-1 (8,10). Further investigation revealed that the increase in cytosolic Ca²⁺ was the result of inhibition of the sarcoendoplasmic reticulum Ca²⁺-ATPase, a calcium pump that plays a major role in the sequestration of Ca²⁺ into the endoplasmic reticulum lumen (11). In isolated IRS-1^−/− knockout β-cells, insulin exocytosis and cytosolic Ca²⁺ concentrations were decreased, further supporting the role of IRS-1 in regulating insulin secretion and intracellular Ca²⁺ homeostasis (9). Taken together, these results are consistent with the possibility that the Arg⁹⁷² IRS-1 variant might impair insulin secretion stimulated by sulfonylureas by affecting intracellular calcium homeostasis, thus contributing to increase the risk to develop secondary failure to these oral agents. In vitro studies have suggested that sulfonylureas may induce apoptosis of pancreatic β-cell (33). Interestingly, we have recently demonstrated that human pancreatic β-cells isolated from donors carrying the Arg⁹⁷² IRS-1 variant exhibit an increased apoptosis compared with their wild-type counterparts, and they are also resistant to the antiapoptotic effects of insulin (34). These data raise the possibility that another potential mechanism, whereby the Arg⁹⁷² IRS-1 variant could contribute to the risk for secondary failure, might be related to increased susceptibility to apoptosis of pancreatic β-cells chronically exposed to sulfonylureas.

In conclusion, the present data demonstrate that the common Arg⁹⁷² IRS-1 variant is associated with increased risk for secondary failure to sulfonylurea in patients with type 2 diabetes. Experimental studies support the notion that this IRS-1 variant may influence the variability in patients’ response to sulfonylureas, thus representing a potential example of pharmacogenetics in type 2 diabetes.

We are grateful to Emanuela Laratta, Cristina Ricasoli, and Pierluigi De Nicolais for their expert technical assistance in genotyping. This study was supported in part by grants from European Community “EuroDiabetesGene” QLG1-CT-1999-00674 (to G.S.), Progetto di Ricerca Finalizzata-Ministero della Sanità (to G.S.), PRIN-COFIN 2002 (2002065949-002), and PRIN-COFIN 2003 (2003067733-02)-Ministero dell’Università e Ricerca Scientifica e Tecnologica (to G.S.).