Insulin Resistance–Related Gene Polymorphisms and Risk of Prostate Cancer

Insulin resistance and hyperinsulinemia have been hypothesized to increase the risk of prostate cancer (1-3). Insulin has mitogenic and antiapoptotic activity and may exert these properties directly on prostate epithelial cells (4). In addition, hyperinsulinemia may affect prostate cancer risk by increasing levels of free (bioactive) insulin-like growth factor-I (IGF-I) or testosterone (3). In previous investigations, prostate cancer risk was positively related to serum insulin levels (2) and also to the metabolic syndrome (5, 6).

Several genetic variants within the insulin signaling pathway have been associated with insulin resistance and hyperinsulinemia (7-10). Polymorphisms of the insulin (INS) and insulin receptor substrate-1 (IRS1) genes have also been associated with prostate cancer (11-13), but data are inconsistent. In addition, no studies have evaluated prostate cancer risk in relation to genetic variation of the insulin receptor (INSR)—a pivotal component of the insulin signaling pathway. We therefore examined the association of prostate cancer with common variants of the INS, INSR, IRS1, and IRS2 genes in a large cohort of Caucasian men.

The study sample comprised 1,053 case/control pairs from the Alpha-Tocopherol, Beta-Carotene (ATBC) Cancer Prevention Study, a cohort of 29,133 men aged 50 to 69 years residing in southwestern Finland who smoked at least 5 cigarettes/day and gave informed consent (14, 15). The study was approved by the institutional review boards of the National Public Health Institute of Finland and the National Cancer Institute.

Cases were individuals with incident prostate cancer (International Classification of Diseases 9, code 185) diagnosed by April 30, 2003, and identified through the Finnish Cancer Registry, which provides almost 100% case coverage (16). DNA was successfully extracted from whole blood for 980 cases and 876 controls. For cases identified through April 1999, medical records were reviewed centrally by two oncologists for staging. Prostate cancer stage was available for 592 of the cases with successful DNA extraction; 408 cases (69%) were stages 0 to 2, and 184 cases (31%) were stages 3 and 4 (17). Controls were subjects alive at the time of case diagnosis and were matched to the cases on age (±5 years), treatment assignment, and date of baseline serum blood draw (±30 days).

Single nucleotide polymorphisms (SNP) were selected using the public databases dbSNP⁷

and SNP-500⁸ and via a literature search on INS, INSR, IRS1, and IRS2. Criteria for inclusion were a minor allele frequency >5% in Caucasian individuals and potential functionality, e.g., SNPs in exons, exon/intron boundaries, putative regulatory regions, or association with insulin resistance or related outcomes in previous studies (7). Polymorphic loci identified in this manner were verified in a panel of 102 individuals (SNP-500 population; ref. 18). This led to the selection of one SNP in INS, five SNPs in INSR, three SNPs in IRS1, and one SNP in IRS2 (Table 2). Following this initial selection, we examined INSR more comprehensively by resequencing every 5 to 10-kb region around a SNP across the INSR gene region. This resulted in the identification of an additional 34 SNPs with minor allele frequencies exceeding 5% in Caucasians. Using the approach developed by Clayton et al. (htSNP2 software⁹), we identified 11 haplotype-tagging SNPs in INSR (including the original five) which predicted the 39 common SNPs among the SNP-500 Caucasian population with high probability (R_H² = 0.90).

Genotyping was done at the Core Genotyping Facility of the Division of Cancer Epidemiology and Genetics, National Cancer Institute, using TaqMan (Applied Biosystems).¹⁰

For validation purposes, TaqMan assays were initially applied to the 102 SNP-500 individuals (18) with sequence information; TaqMan results were 100% concordant. To assess quality control, duplicate masked specimens for 120 control samples were genotyped with 100% concordance. Departures from Hardy-Weinberg equilibrium were tested for each SNP using genotype distributions among the control participants.

To assess the association of genotypes and haplotypes with prostate cancer risk, we used unconditional logistic regression, with the most common genotype/haplotype serving as reference. Tests of trend were calculated using ordinal values of 1, 2, and 3 assigned to genotypes in order of homozygous for common allele, heterozygous, and homozygous for the rare allele, respectively. Haplotype blocks were specified among controls according to the method of Gabriel et al. (19), and haplotype frequencies were estimated using the expectation-maximization (EM) algorithm (20). We found little phase ambiguity in the reconstruction of haplotypes. All analyses were adjusted for age and treatment assignment. We also conducted conditional logistic regression models but found no substantive differences from unconditional models. Separate analyses were conducted for advanced cancers (stages 3 and 4).

Cases were slightly older and more likely to have a positive family history of prostate cancer than controls but were similar with respect to body mass index (BMI), physical activity, and smoking history (Table 1).

All genotypes were distributed in accordance with the Hardy-Weinberg equilibrium except for INSR Ex3+131C>T (P = 0.02). We found no association for any of the genotypes of INS, INSR, IRS1, and IRS2 with risk of prostate cancer (Table 2). Haplotype analysis for the INSR and IRS1 genes did not yield statistically significant findings (data not shown). In analyses restricted to advanced cases (stages 3 and 4), we found that carriers of the C allele at the INSR IVS7-126C>T locus had a 34% reduced risk of prostate cancer [OR, 0.69; 95% confidence interval (95% CI) 0.50, 0.96; P = 0.03] compared with men homozygous for the T allele. In a secondary analysis, the reduction in advanced disease risk was primarily among men with a BMI > 25 kg/m² (OR, 0.47; 95% CI, 0.31, 0.72; P = 0.0005).

Overall, our findings suggest that there is little association between the polymorphisms of the INS, INSR, IRS1, and IRS2 genes studied here and the risk of prostate cancer. The statistical power was sufficient (>0.80) to detect an odds ratio (OR) of ≥1.5 for all SNPs under the assumption of a dominant mode of inheritance. Given that our study included a large, well-defined sample of prostate cancer cases and controls, it is unlikely that our null findings are due to chance, although we cannot exclude the possibility of undetected weaker associations (OR < 1.5).

One possible exception to our null findings is a lower risk of advanced prostate cancer among carriers of the C allele at the INSR IVS7-126C>T locus (P = 0.03). However, using either the false discovery rate (21) or the Bonferroni correction method and assuming two or more statistical tests, the P value for the association would need to have been 0.025 or lower to be considered statistically significant.

Insulin resistance and compensatory hyperinsulinemia have been hypothesized to promote prostate carcinogenesis through either the direct promitotic/antiapoptotic properties of insulin or via alterations in the IGF and/or sex hormone pathways. In prior studies, the INS IVS1-6T/T variant has been found to predict insulin levels and type II diabetes (7, 22) and also prostate cancer (11). However, our study and other studies did not confirm the latter association (12, 23). One investigation reported an association between the IRS1 G972R polymorphism and prostate cancer (12), but a subsequent study was null (23). The IRS1 IVS1+12245 SNP, which is in moderate linkage disequilibrium with IRS1 G972R,¹¹

was not associated with prostate cancer risk in this population.

In conclusion, this large case-control study found little evidence for an association between allelic variants in insulin resistance–related genes and risk of prostate cancer.