DNA Repair Gene XRCC3 Codon 241 Polymorphism, Its Interaction with Smoking and XRCC1 Polymorphisms, and Bladder Cancer Risk

DNA repair plays a key role in carcinogenesis through the removal of DNA damage induced by endogenous and environmental sources. We are interested in polymorphic genes involved in the repair of smoking-induced DNA damage and in their association with bladder cancer risk. The main risk factor for bladder cancer is cigarette smoking, which can increase risk up to four times (1, 2).

Cigarette smoke is a rich source of chemical carcinogens and ROS⁴. Chemical carcinogens form bulky adducts on DNA, whereas ROS can induce base damage, SSBs, and DSBs (3). Bulky adducts are repaired through the nucleotide excision repair pathway (4). ROS-induced base damage and SSBs are repaired through the base excision repair pathway (5), whereas DSBs can be repaired by either HRR or nonhomologous end joining (6). We have previously reported an association between bladder cancer risk and allelic variations in the XRCC1 gene, which is involved in BER (7).

In this study, we focused on the XRCC3 gene, which is required for efficient repair of DSBs through the HRR pathway (8), for repair of DNA cross-linking (9), and for chromosomal segregation (10). During HRR, the XRCC3 protein interacts with the Rad51 protein, enabling Rad51 protein multimers to assemble at the site of damage (9, 11). Furthermore, Rad51 has also been found to colocalize with the XRCC1 protein after base damage, suggesting coordination between XRCC1-dependent SSB repair and recombination events during DNA replication (12).

A common polymorphism in the XRCC3 gene in codon 241 results in a Thr to Met substitution (13). Using a bladder cancer case-control study, we analyzed a possible association between the XRCC3 codon 241 polymorphism and bladder cancer and a possible interaction with cigarette smoking as a measure of ROS exposure. We also tested for an XRCC1-XRCC3 interaction and a possible effect modification of this interaction on the effect of cigarette smoking.

Bladder cancer patients (n = 235) and control individuals (n = 213) were enrolled from the urology clinics at Duke University Medical Center and the University of North Carolina Hospitals, as described previously (2, 7). Briefly, cases had histologically confirmed transitional cell carcinoma. Controls were urology patients without a history of cancer from the same clinics, the most frequent diagnoses being benign prostate hypertrophy, impotence, and urinary incontinence. Controls were frequency matched to cases based on ethnicity (black, white), sex, and age at interview (10-year interval). After giving written informed consent, all individuals were administered a questionnaire that detailed their smoking exposure history and provided blood samples collected under protocols approved by the institutional review boards of each participating institution.

DNA was extracted from peripheral blood lymphocytes by standard methods. The DNA segment surrounding codon 241 was amplified by PCR using 50 ng of DNA in a final volume of 15 μl containing 1× PCR buffer, 0.2 mm dNTPs, 2 mm MgCl₂, 0.8 μm forward primer (at 17,724 bp: 5′-TTGGGGCCTCTTTGAGA-3′), 0.8 μm reverse primer (at 18,258 bp: 5′-AACGGCTGAGGGTCTTCT-3′), and 0.5 units of AmpliTaq Gold (Perkin-Elmer, Foster City, CA). PCR reactions were carried out in a Perkin-Elmer 9700 thermocycler with an initial denaturation step of 8 min at 94°C, followed by 30 cycles at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min. PCR products were digested with NlaIII restriction enzyme (New England Biolabs, Beverly, MA), resolved in 3% NuSieve 3:1 agarose gels (FMC Bioproducts, Rockland, ME), and stained with ethidium bromide. NlaIII recognizes an invariant restriction site around bp 17,963, which serves as an internal control for complete enzyme digestion. The XRCC3 codon 241 Met-allele creates a NlaIII restriction site at bp 18,067, a site that is not present in the codon 241 Thr-allele.

We used standard methods for 2 × k contingency tables, including Fisher’s exact test, as appropriate, to analyze categorical variables without adjustment for covariates. We checked among controls for differences between the observed genotypic frequencies and those expected under the Hardy-Weinberg law using estimates of the disequilibrium coefficient (14). When adjusting for age, sex, or ethnicity, and when examining interactions between the polymorphism and smoking, we used standard logistic regression methods (15). We tested for interaction on a multiplicative scale. Given the small number of blacks in our study, data from blacks are included with whites in combined OR estimates after including ethnicity as a covariate. For gene-environment interaction analyses, we used categorized versions of pack-years as measures of smoking exposure and years of smoking as a continuous variable as described previously (2). We examined the combined effects of the continuous smoking variable and the XRCC3 genotype and the combined effects of smoking with the XRCC3 and XRCC1 genotypes using an approach that we have used previously (2, 7). This approach fits a series of logistic regression models and compares them, testing relevant hypotheses with likelihood ratio tests. Within a model, differences between slopes represent genotype-exposure interactions, and differences between intercepts represent genotype effects among nonsmokers. All tests were two-tailed. All analyses were done using the statistical package Egret for Windows (Cytel Software Corporation, Cambridge, MA).

Demographic and smoking information on this study have been previously reported (2, 7). Estimates and tests reported below are for whites and blacks combined, adjusted for age, sex, and ethnicity.

The frequency of the codon 241 Met allele among cases and controls were 0.39 and 0.34, respectively, for whites and 0.26 and 0.12 for blacks. Among white or black controls, there were no significant differences between the observed genotypic frequencies and those expected under the Hardy-Weinberg law (P = 0.78 for whites and P = 0.61 for blacks). The OR for bladder cancer for those individuals with one copy of the codon 241 Met allele compared with those with none was 1.2 (95% CI: 0.8–1.9), and the corresponding OR for those with two copies of the Met allele was 1.5 (95% CI: 0.8–2.7; Table 1). For the remaining analysis, we combined subjects who carried one or two copies of the Met allele (OR = 1.3; 95% CI = 0.9–1.9).

Among never smokers, we observed little association between subjects with the codon 241 Thr/Met or Met/Met genotypes and bladder cancer (Table 1). Among moderate smokers (1–35 pack-years), subjects with the codon 241 Thr/Met or Met/Met genotypes had a slightly lower OR than those homozygous for the Thr allele, with considerable overlap in the 95% CI for the two estimates (Table 1), giving a gene effect within this smoking strata of 0.7 (95% CI: 0.4–1.4). However, among the high-level smoking category (>35 pack-years), the codon 241 Thr/Met or Met/Met genotypes were associated with a higher bladder cancer risk than that of subjects that smoked the same amount but who carried the Thr/Thr genotype (Table 1), giving a gene effect within this smoking strata of 2.1 (adjusted 95% CI: 1.1–4.2). Tests of interaction that compared the gene effect in each of the smoking categories to that among never smokers were not statistically significant (Table 1). A test for XRCC3-smoking interaction using years of smoking as a continuous variable detected no interaction (χ²_1df = 1.1; P = 0.29).

We examined whether the XRCC3 polymorphism had differential effects depending on XRCC1 genotype using data we had previously obtained on XRCC1 polymorphisms in these subjects (7). Our earlier data on XRCC1 supported an inverse association between the codon 194 Arg/Trp genotype and bladder cancer (no subjects carried the Trp/Trp genotype) and a separate inverse association between the codon 399 Gln/Gln genotype and cancer risk. For this analysis, we used as a reference the genotype group with the largest sample size, which was the highest risk group, i.e., subjects with the XRCC1 codon 194 Arg/Arg and the XRCC3 codon 241 Thr/Met or Met/Met genotypes (Table 2). Our results suggest that the inverse association between the codon 241 Thr/Thr genotype and cancer risk might be stronger among subjects who also carry the XRCC1 codon 194 Arg/Trp genotype than among those with the XRCC1 codon 194 Arg/Arg genotype, however, the test for interaction did not reach statistical significance (P = 0.09). The combined presence of the XRCC1 codon 399 and XRCC3 codon 241 polymorphisms showed a similar pattern, but the test for interaction was far from statistical significance (P = 0.61).

To examine if the combined presence of the XRCC1 codon 194 and XRCC3 polymorphisms could modify the effect of cigarette smoking on bladder cancer risk, we used years of smoking as a predictor of bladder cancer risk in the logistic regression model, and we fitted separate regression lines for each of the four combined XRCC1-XRCC3 genotypes (fitted lines in Fig. 1,A). This four-line model fit our data significantly better than a model with a single regression line for all four genotype combinations (χ²_6df = 12.9; P = 0.04). The fitted regression lines suggested that individuals who carry the XRCC1 codon 194 Trp/Arg genotype and the XRCC3 codon 241 Thr/Thr genotype (protective genotype) have a different smoking dose response from any of the other three combined genotypes (risk genotypes). Therefore, we examined a data-suggested hypothesis. The three risk genotypes did not appear to differ in their smoking dose response (χ²_4df = 2.1; P = 0.72), thus, we fitted a common smoking dose-response line for the combined at-risk genotypes (Fig. 1 B). This combined smoking dose response appeared quite distinct from that of the single protective genotype (χ²_2df = 10.8; P = 0.004). In particular, the slopes of the two lines appeared to differ, suggesting a possible gene-gene-smoking three way interaction (χ²_1df = 3.0; P = 0.08).

Two case-control studies have found a significant positive association between the XRCC3 codon 241 Met allele and melanoma (16) and bladder cancer (17), whereas a lung cancer study found no evidence of an association among Caucasians and a nonsignificant positive association among African-Americans (18).

Our analysis of the XRCC3 gene alone, ignoring smoking and the XRCC1 genotype, provides little evidence of a positive association between the XRCC3 codon 241 Met variant allele and bladder cancer. Among heavy smokers, subjects with the Met allele had roughly twice the risk of those who had comparable smoking histories but who did not carry the Met allele; however, the test for interaction with smoking exposure did not reach statistical significance. In contrast, Matullo et al. (17) reported that the positive association between the Met allele and cancer risk seemed stronger among ex-smokers and nonsmokers than among current smokers. However, like our study, they found little evidence for gene-smoking interaction.

The XRCC1 and XRCC3 combined analyses suggested that the observed small difference in risk between different XRCC3 genotypes may be more relevant among individuals with the XRCC1 codon 194 Arg/Trp genotypes, however, the test for interaction did not reach statistical significance (P = 0.09). The combined XRCC1 codon 194 Arg/Trp and XRCC3 codon 241 Thr/Thr genotypes had a stronger inverse association with bladder cancer among lower-dose smokers than higher-dose smokers, although again the test for interaction did not reach statistical significance (P = 0.08). Genotype risks converged at high-smoking dose, which differs from what was observed with XRCC3 alone, where genotype risks differed most among heavier smokers. We should note that the size of our study has very limited power for assessing three-way interactions; therefore, our findings must be interpreted with caution and need to be validated in larger studies. As yet, no studies have directly examined the function of the XRCC3 polymorphism. However, a statistically significant association was described between the codon 241 Met variant and higher frequency of kinetochore-positive micronuclei among healthy cigarette smokers.⁵ Given that cigarette smoking induces micronucleus formation (19), the reported association suggests that this allele may have impaired function. An XRCC1-XRCC3-smoking interaction seems plausible. Both proteins play important roles in the repair of strand breaks. The XRCC1 protein can detect and repair SSBs (20), which may convert to DSBs after DNA replication (21). The XRCC3 protein plays a role in the repair of DSBs through the HRR by its interaction with the Rad51 protein (8, 9, 11). Furthermore, XRCC1 has been found to colocalize with Rad51 after DNA damage (12). Evidence for such an interaction between a base excision repair pathway and a recombination repair gene has been found in studies in the yeast Schizosaccharomyces pombe (22). Our findings highlight the importance of integrating the information from multiple genes in relevant pathways to better identify people at risk from environmental exposures.

We thank Lyle Lansdell and Patty Blanton for data handling, sample retrieval, and management; Drs. James Mohler, David Paulson, and Cary Robertson for patient enrollment; Sue Edelstein for graphic design; and Drs. Carla H. van Gils and Jane C. Schroeder for critical review of this manuscript.