No Association between the I105V Polymorphism of the Glutathione S-Transferase P1 Gene (GSTP1) and Prostate Cancer Risk: A Prospective Study¹

GSTs³conjugate glutathione to harmful substances facilitating their removal. Allelic variants of the GSTP1 gene have been identified, one isoleucine (GSTP1-1/I-105) and one valine (GSTP1-1/V-105) at codon 105. The allele frequency varies between ethnic groups. The isozymes have activities that differ by substrate. GSTs noncovalently bind and sequester steroid hormones. Smaller epidemiologic studies have reported an association between the GSTP1-1/Val-105 allele and higher risk of CaP.

We sought to test the hypothesis that men homozygous for GSTP1-1/Val-105 would be at higher risk for developing CaP than men who are GSTP1-1/Ile-105 homozygotes or heterozygotes and to determine whether smoking alters any such association. We also hypothesized that circulating levels of androgens would be higher in men homozygous for GSTP1-1/Val-105.

Blood was obtained in 1982 from 14,916 men enrolled in The Physicians’Health Study and men were followed until 1995. Study characteristics including blood collection and processing, stage classification,clinical follow-up and documentation of prostate cancer were described previously.

Genomic DNA was isolated from whole blood using a Qiamp DNA extraction kit (Qiagen, Chatsworth, CA) with investigators blinded to status. Samples were diluted to 20 ng/μl and stored at −20°C. Twenty ng of DNA were added to the PCR reaction mixture: 0.3 mm primers P105F and P105R, 1.5 mm MgCl₂, and 1.5 units of Qiagen Taq in a final volume of 22 μl. Cycling conditions were 94°C for 2 min, 35 cycles of 94°C for 30 s, 60°C for 1 min, 72°C for 30 s, and 72°C for 8 s in a Perkin-Elmer GeneAmp PCR System 9600 or an MJ Research PTC-200 (MJ Research, Waltham, MA). Digestion was with 2.5 units of BsmA1 (New England Biolabs,Beverly, MA), separated on a 2% agarose gel stained with ethidium bromide. The genotypes were successfully determined for 590 cases and 803 controls.

Plasma hormone concentrations were measured as described by Gann et al. (5).

We used the χ² test to evaluate whether the distribution of GSTP1 alleles or genotypes varied among cases and controls. We calculated ORs as an estimate of relative risk and 95%CIs from logistic regression models. To increase power in the analysis of genotype by stage and grade, we included all controls while controlling for the matching factors of baseline age (5-year intervals)and smoking status. We used stratified analysis to evaluate whether CaP risk associated with genotype varied by smoking status and assessed interaction by use of a cross-product term. From a previous analysis,plasma steroid hormone concentrations were available for 389 controls. We calculated mean levels of testosterone, dihydrotestosterone, and androstanediol glucuronide by genotype and assessed whether hormone levels varied by genotype using the Kruskal-Wallis test.

The valine frequencies did not differ between CaP cases and controls (Table 1). Neither val/val nor ile/val genotype was associated with CaP overall or with high or low stage/grade disease(Table 1). No difference (P-interaction = 0.4) was found in the OR for developing CaP by GSTP1 genotype between 761 ever-smokers (ile/ile referent, ile/val OR,0.84; 95% CI, 0.62–1.13; val/val OR, 0.80; 95% CI,0.47–1.36) and 632 never-smokers (ile/ile referent; ile/val OR, 0.91; 95% CI, 0.65–1.28; val/valOR, 0.87; 95% CI, 0.52–1.48). We saw no important differences by stage/grade of CaP when considering smoking status either. Levels of testosterone, dihydrotestosterone and androstanediol glucuronide did not differ by GSTP1 genotype (all P ≥0.08).

We had 80% power to detect as statistically significant an OR of 1.4 when comparing the val/val and ile/valgenotypes to the ile/ile genotype.

The population examined in this study was predominantly United States Caucasian. Allele and genotype frequencies vary at this locus between ethnic groups; the impact it may have on CaP occurrence in other groups requires further study. Smoking status was assigned at baseline in 1982. Heavy, recent smoking may be a risk factor for developing highly aggressive CaP.

In this large, nested case-control study, we found no evidence of a differential risk for CaP among primarily Caucasian, United States men possessing the isoleucine or valine variants of codon 105 of GSTP1 in any genotypic combination. Our large sample size resulted in narrow CIs and allowed us to examine high and low stage/grade CaP while maintaining adequate statistical power. The upper CI limit of 1.22 excludes a large risk associated with the val/val genotype. Selection or survival bias was not an issue because of the prospective design and the high follow-up response rate.

We thank Stephanie Parker, Rachel Meyer, and Kathryn Starzyk for administrative support and management of the plasma samples and Xiaoyang Liu for assistance with computer programming.