Abstract
Two major risk factors for bladder cancer are smoking and occupational exposure to chemicals. The XPC protein is crucial in the recognition and initiation of the nucleotide excision repair pathway which repairs the DNA adducts formed by carcinogens found in cigarette smoke and chemicals. Polymorphisms in the XPC gene have been shown to influence an individual's DNA repair capacity, and hence, increase that individual's susceptibility to cancer. We undertook a case-control study of 547 bladder cancer cases and 579 cancer-free controls to investigate the association between 22 XPC polymorphisms and bladder cancer susceptibility, and investigated gene-environment interactions. We showed that the nonsynonymous polymorphism Ala499Val was in strong linkage disequilibrium with two polymorphisms in the 3′-untranslated region (Ex15-184 and Ex15-177) with Lewontin's D′ ≥ 0.99 and r2 ≥ 0.82. Individuals homozygous for the minor allele of Ala499Val, Ex15-184, or Ex15-177 had an increased risk of bladder cancer compared with those homozygous for the common allele [adjusted odds ratio (95% confidence interval), 1.65 (1.05-2.59), 1.82 (1.12-2.97), and 1.82 (1.12-2.96), respectively]. The associations were somewhat stronger for smokers and those occupationally exposed to chemicals, although tests for gene-environment interactions were not significant. (Cancer Epidemiol Biomarkers Prev 2006;15(12):2537–41)
Introduction
Carcinogens in tobacco smoke and industrial chemicals are two major risk factors for transitional cell carcinoma (TCC). The bulky DNA adducts formed by these carcinogens are repaired by nucleotide excision repair. Reduced DNA repair capacity (DRC) is associated with an increased cancer risk (1). Inherited polymorphisms in DNA repair genes might influence an individual's DRC (2) by altering amino acid sequences, mRNA splicing, and transcription efficiency.
The xeroderma pigmentosum complementation group C (XPC) protein is involved in the recognition and initiation of the global genome repair pathway of nucleotide excision repair. XPC binds to HR23B to form the stable XPC-HR23B complex, which recognizes and binds to damaged DNA (3). XPC protein deficiency causes the clinical disorder xeroderma pigmentosum, characterized by extreme sensitivity to sunlight and an increased risk of skin cancer (4). Similarly, XPC knockout (XPC−/−) mice are susceptible to UV-induced skin tumors (5), but may also develop acetylaminofluorene-induced lung and liver tumors (6).
To date, only five XPC polymorphisms have been studied in bladder cancer, i.e., Arg492His, Ala499Val, Lys939Gln, intervening sequence (IVS) 11-6, and poly(AT) (intron 9; refs. 7-10). Although three studies, including the larger studies of Wu et al. and Garcia-Closas et al. showed no association with bladder cancer risk (8-10), Sanyal et al. (7) found an increased risk for cases with the Lys939Gln homozygous variant. Also, variants of Lys939Gln, IVS11-6, and poly(AT) have been associated with increased risk of various other cancers (11-13). The IVS11-6 (A to C transition) at intron 11 causes abnormal mRNA splicing with skipping of exon 12 and hence reduced DRC (14). The poly(AT) (insertion/deletion) polymorphism in intron 9 was associated with reduced DRC in a normal population (15). The Lys939Gln (A to C transition) was found to influence irradiation-specific DNA repair in normal human lymphocytes (16). We hypothesized that other XPC polymorphisms may also influence DRC and hence bladder cancer risk. Therefore, we undertook a large study in TCC bladder cases and controls of 22 potentially functional XPC polymorphisms.
Materials and Methods
Study Subjects
The design of this case-control study has been described in detail previously (8). Briefly, 547 cases of bladder TCC were recruited from August 2002 to April 2004 at our institution in Leeds, United Kingdom. Five hundred and seventy-nine cancer-free controls were recruited from the community from 1997 to 2000 (n = 227; ref. 17) and from the hospital otolaryngology and ophthalmology departments from 2002 to 2004 (n = 352) in the same region. Each subject donated a blood sample and completed a structured health questionnaire regarding smoking, occupation, and family history. Occupational exposure was defined as participating in occupations involving rubber/plastics industries, laboratories, printing, paints, dyes, or diesel fumes. Histopathologic information regarding tumor stage and grade were obtained from hospital records.
Genotyping
We selected XPC polymorphisms based on potential function rather than on patterns of linkage disequilibrium. Hence, all the XPC polymorphisms with allele frequencies of >1.0% located in the promoter region (up to 1,000 bases upstream of the gene), 5′ and 3′ untranslated region (UTR), and exons were selected. Only those polymorphisms in the intronic regions with known or potential splicing effects (within 100 bases up and downstream of an exon) were selected (15% of all intronic single nucleotide polymorphisms of >1% allele frequency). The 22 polymorphisms chosen for this study are listed in Table 1.
The location, dbSNP reference number, genotype distribution, minor allele frequency and ORs for the 22 XPC polymorphisms
Polymorphisms* . | Genotypes . | Controls (%) . | Cases (%) . | aOR (95% CI) . | P . | Redundant polymorphisms† . |
---|---|---|---|---|---|---|
Ex1-434 (G > A), rs3731054, promoter | G/G | 517 (92.3) | 500 (94.3) | 1.00 | Nil | |
G/C | 43 (7.7) | 30 (5.7) | 0.78 (0.48-1.27) | 0.31 | ||
C/C | 0 (0) | 0 (0) | NC | NC | ||
C allele frequency | 0.04 | 0.03 | ||||
Ex1-356 (G > A), rs3731055, promoter | G/G | 549 (98.6) | 515 (98.7) | 1.00 | Nil | |
G/A | 8 (1.4) | 7 (1.3) | 1.09 (0.39-3.07) | 0.87 | ||
A/A | 0 (0) | 0 (0) | NC | NC | ||
A allele frequency | 0.01 | 0.01 | ||||
Ex1-12 (G > C), rs2607775, promoter | G/G | 133 (24.8) | 138 (27.1) | 1.00 | IVS11-37 (C > G), rs2470353 | |
G/C | 262 (48.9) | 247 (48.4) | 0.95 (0.71-1.29) | 0.76 | ||
C/C | 141 (26.3) | 125 (24.5) | 0.91 (0.64-1.28) | 0.58 | ||
C allele frequency | 0.51 | 0.49 | ||||
Ex1-7 (G > A), rs3731058, promoter | G/G | 559 (99.8) | 536 (99.6) | 1.00 | Asp454Asp (T > A), rs3731128, exon 8 | |
G/A | 0 (0) | 2 (0.4) | NC | |||
A/A | 1 (0.2) | 0 (0) | NC | |||
G/A + A/A | 1 (0.2) | 2 (0.4) | 2.15 (0.18-25.7) | 0.54 | ||
A allele frequency | 0.002 | 0.002 | ||||
Leu16Val (C > G), rs1870134, exon 1 | C/C | 523 (98.5) | 501 (98.8) | 1.00 | Nil | |
C/G | 8 (1.5) | 6 (1.2) | 0.95 (0.32-2.81) | 0.93 | ||
G/G | 0 (0) | 0 (0) | NC | NC | ||
G allele frequency | 0.008 | 0.006 | ||||
IVS6-22 (G > A), rs3731125, intron 6 | A/A | 484 (86.3) | 462 (86.5) | 1.00 | Nil | |
A/G | 75 (13.4) | 70 (13.1) | 1.01 (0.70-1.44) | 0.97 | ||
G/G | 2 (0.4) | 2 (0.4) | 0.80 (0.11-5.91) | 0.83 | ||
G allele frequency | 0.07 | 0.07 | ||||
Arg492His (G > A), rs2227999, exon 8 | G/G | 507 (89.4) | 473 (88.4) | 1.00 | Nil | |
G/A | 59 (10.4) | 60 (11.2) | 1.04 (0.70-1.54) | 0.84 | ||
A/A | 1 (0.2) | 2 (0.4) | 1.60 (0.14-18.6) | 0.71 | ||
G/A + A/A | 60 (10.6) | 62 (11.6) | 1.05 (0.71-1.55) | 0.80 | ||
A allele frequency | 0.05 | 0.06 | ||||
Ala499Val (C > T), rs2228000, exon 8 | C/C | 317 (56.1) | 279 (51.9) | 1.00 | Ex15-699 (C > G), rs2229090, 3′UTR | |
C/T | 210 (37.2) | 202 (37.6) | 1.12 (0.86-1.45) | 0.40 | ||
T/T | 38 (6.7) | 57 (10.6) | 1.65 (1.05-2.59) | 0.03 | ||
T allele frequency | 0.25 | 0.29 | ||||
Arg687Arg (G > A), rs3731151, exon 11 | G/G | 301 (53.8) | 303 (57.2) | 1.00 | IVS3+73 (G > T), rs3731081, intron 3; IVS6-70 (A>C), rs3731128, intron 6; IVS14-51 (G > A), rs3731175, intron 14 | |
G/A | 216 (38.6) | 200 (37.7) | 0.92 (0.71-1.18) | 0.50 | ||
A/A | 42 (7.5) | 27 (5.1) | 0.60 (0.35-1.01) | 0.054 | ||
A allele frequency | 0.27 | 0.24 | ||||
Poly(AT) (− > +), not listed, intron 9 | −/− | 204 (35.4) | 215 (39.5) | 1.00 | IVS11-6 (C > A), rs2279017, intron 11; IVS14-40 (G>A), rs2733532, intron 14; Lys939Gln (A > C), rs2228001, exon 15 | |
−/+ | 288 (49.9) | 242 (44.5) | 0.82 (0.63-1.07) | 0.15 | ||
+/+ | 85 (14.7) | 87 (16.0) | 0.99 (0.69-1.42) | 0.95 | ||
+ allele frequency | 0.397 | 0.382 | ||||
Ex15-184 (T > A), rs2470352, 3′UTR | T/T | 346 (61.8) | 285 (54.8) | 1.00 | Ex15-177 (A > G), rs2470458, 3′UTR | |
T/A | 183 (32.7) | 186 (35.8) | 1.23 (0.95-1.61) | 0.12 | ||
A/A | 31 (5.5) | 49 (9.4) | 1.82 (1.12-2.97) | 0.02 | ||
A allele frequency | 0.22 | 0.27 | ||||
Ex15-111 (C > G), rs1126547, 3′UTR | C/C | 444 (78.9) | 415 (77.7) | 1.00 | Nil | |
C/G | 112 (19.9) | 112 (21.0) | 1.06 (0.78-1.43) | 0.72 | ||
G/G | 7 (1.2) | 7 (1.3) | 1.09 (0.37-3.25) | 0.87 | ||
G allele frequency | 0.11 | 0.12 |
Polymorphisms* . | Genotypes . | Controls (%) . | Cases (%) . | aOR (95% CI) . | P . | Redundant polymorphisms† . |
---|---|---|---|---|---|---|
Ex1-434 (G > A), rs3731054, promoter | G/G | 517 (92.3) | 500 (94.3) | 1.00 | Nil | |
G/C | 43 (7.7) | 30 (5.7) | 0.78 (0.48-1.27) | 0.31 | ||
C/C | 0 (0) | 0 (0) | NC | NC | ||
C allele frequency | 0.04 | 0.03 | ||||
Ex1-356 (G > A), rs3731055, promoter | G/G | 549 (98.6) | 515 (98.7) | 1.00 | Nil | |
G/A | 8 (1.4) | 7 (1.3) | 1.09 (0.39-3.07) | 0.87 | ||
A/A | 0 (0) | 0 (0) | NC | NC | ||
A allele frequency | 0.01 | 0.01 | ||||
Ex1-12 (G > C), rs2607775, promoter | G/G | 133 (24.8) | 138 (27.1) | 1.00 | IVS11-37 (C > G), rs2470353 | |
G/C | 262 (48.9) | 247 (48.4) | 0.95 (0.71-1.29) | 0.76 | ||
C/C | 141 (26.3) | 125 (24.5) | 0.91 (0.64-1.28) | 0.58 | ||
C allele frequency | 0.51 | 0.49 | ||||
Ex1-7 (G > A), rs3731058, promoter | G/G | 559 (99.8) | 536 (99.6) | 1.00 | Asp454Asp (T > A), rs3731128, exon 8 | |
G/A | 0 (0) | 2 (0.4) | NC | |||
A/A | 1 (0.2) | 0 (0) | NC | |||
G/A + A/A | 1 (0.2) | 2 (0.4) | 2.15 (0.18-25.7) | 0.54 | ||
A allele frequency | 0.002 | 0.002 | ||||
Leu16Val (C > G), rs1870134, exon 1 | C/C | 523 (98.5) | 501 (98.8) | 1.00 | Nil | |
C/G | 8 (1.5) | 6 (1.2) | 0.95 (0.32-2.81) | 0.93 | ||
G/G | 0 (0) | 0 (0) | NC | NC | ||
G allele frequency | 0.008 | 0.006 | ||||
IVS6-22 (G > A), rs3731125, intron 6 | A/A | 484 (86.3) | 462 (86.5) | 1.00 | Nil | |
A/G | 75 (13.4) | 70 (13.1) | 1.01 (0.70-1.44) | 0.97 | ||
G/G | 2 (0.4) | 2 (0.4) | 0.80 (0.11-5.91) | 0.83 | ||
G allele frequency | 0.07 | 0.07 | ||||
Arg492His (G > A), rs2227999, exon 8 | G/G | 507 (89.4) | 473 (88.4) | 1.00 | Nil | |
G/A | 59 (10.4) | 60 (11.2) | 1.04 (0.70-1.54) | 0.84 | ||
A/A | 1 (0.2) | 2 (0.4) | 1.60 (0.14-18.6) | 0.71 | ||
G/A + A/A | 60 (10.6) | 62 (11.6) | 1.05 (0.71-1.55) | 0.80 | ||
A allele frequency | 0.05 | 0.06 | ||||
Ala499Val (C > T), rs2228000, exon 8 | C/C | 317 (56.1) | 279 (51.9) | 1.00 | Ex15-699 (C > G), rs2229090, 3′UTR | |
C/T | 210 (37.2) | 202 (37.6) | 1.12 (0.86-1.45) | 0.40 | ||
T/T | 38 (6.7) | 57 (10.6) | 1.65 (1.05-2.59) | 0.03 | ||
T allele frequency | 0.25 | 0.29 | ||||
Arg687Arg (G > A), rs3731151, exon 11 | G/G | 301 (53.8) | 303 (57.2) | 1.00 | IVS3+73 (G > T), rs3731081, intron 3; IVS6-70 (A>C), rs3731128, intron 6; IVS14-51 (G > A), rs3731175, intron 14 | |
G/A | 216 (38.6) | 200 (37.7) | 0.92 (0.71-1.18) | 0.50 | ||
A/A | 42 (7.5) | 27 (5.1) | 0.60 (0.35-1.01) | 0.054 | ||
A allele frequency | 0.27 | 0.24 | ||||
Poly(AT) (− > +), not listed, intron 9 | −/− | 204 (35.4) | 215 (39.5) | 1.00 | IVS11-6 (C > A), rs2279017, intron 11; IVS14-40 (G>A), rs2733532, intron 14; Lys939Gln (A > C), rs2228001, exon 15 | |
−/+ | 288 (49.9) | 242 (44.5) | 0.82 (0.63-1.07) | 0.15 | ||
+/+ | 85 (14.7) | 87 (16.0) | 0.99 (0.69-1.42) | 0.95 | ||
+ allele frequency | 0.397 | 0.382 | ||||
Ex15-184 (T > A), rs2470352, 3′UTR | T/T | 346 (61.8) | 285 (54.8) | 1.00 | Ex15-177 (A > G), rs2470458, 3′UTR | |
T/A | 183 (32.7) | 186 (35.8) | 1.23 (0.95-1.61) | 0.12 | ||
A/A | 31 (5.5) | 49 (9.4) | 1.82 (1.12-2.97) | 0.02 | ||
A allele frequency | 0.22 | 0.27 | ||||
Ex15-111 (C > G), rs1126547, 3′UTR | C/C | 444 (78.9) | 415 (77.7) | 1.00 | Nil | |
C/G | 112 (19.9) | 112 (21.0) | 1.06 (0.78-1.43) | 0.72 | ||
G/G | 7 (1.2) | 7 (1.3) | 1.09 (0.37-3.25) | 0.87 | ||
G allele frequency | 0.11 | 0.12 |
NOTE: aOR, odds ratios adjusted for subjects' gender, age, smoking, occupational exposure, and family history of bladder cancer. ORs in boldface represent significant results with P ≤ 0.05.
Abbreviations: NC, not calculated; −, wild-type allele; +, variant allele with insertion of 83 bases of A and T.
For each polymorphism, its name, dbSNP reference number, and XPC gene location is stated.
Redundant polymorphisms in strong linkage disequilibrium (both D′ and r2 ≥ 0.85). Data is similar and are therefore not presented.
Details of the genotyping using TaqMan techniques have been described previously (8). The primer and probe sets used are listed in Supplementary Table S1. As a quality control measure, 5% of the samples (n = 57) were re-genotyped with 100% concordance.
Statistical Analysis
All data were analyzed as previously described (8) using the Stata software version 8 (College Station, Statacorp, TX). We had 90%, 78%, and 55% power to detect an odds ratio (OR) of 1.50 for carriage of the minor allele with frequencies at 0.20, 0.10, and 0.05, respectively (two-sided, P = 0.05). Gene-environment interactions were assessed by stratification of subjects based on smoking status and occupational exposure, and P values were calculated based on the likelihood-ratio test. Haplotypes for the gene were estimated based on the expectation maximization algorithm using the program SIMHAP (http://www.genepi.com.au/projects/simhap/). A P = 0.05 for any test or model was considered to be statistically significant.
Results
Study Subjects
The demographic characteristics of the cases and controls have been described previously (8). The participation rate for cases and controls were 99% and 80%, respectively. The majority of subjects were Caucasian (98.6%) with no difference in mean age (cases, 72.8 years; controls, 71.9 years), although cases were more likely to be male (70.9% versus 65.5%, P = 0.05). Smoking prevalence was higher in cases (78% versus 66%, P < 0.001) as was occupational exposure (27.4% versus 16.8%, P < 0.001), and family history of bladder cancer (4.8% versus 2.2%, P = 0.02).
Genotyping
Genotyping was successful in 96.2% of samples (range, 91.7-99.6% for individual polymorphisms). The control genotype distributions were all in Hardy-Weinberg equilibrium. Although there were no significant differences among coding polymorphisms in genotype distributions between the two control groups, differences were found in two intronic polymorphisms (IVS11-37 and IVS14-51) at the 5% level (minor allele frequency, 0.54 versus 0.45; P = 0.01 and 0.29 versus P = 0.25 and 0.04, respectively). However, this was consistent with random variation, given the number of polymorphisms examined, so the two control groups were combined to increase our power to detect an association between the polymorphisms and bladder cancer risk.
The 22 XPC polymorphisms exhibited strong linkage disequilibrium (Supplementary Table S2). We have previously shown the strong linkage disequilibrium between XPC poly(AT), IVS11-6, and Lys939Gln (8). In addition, we also found strong linkage disequilibrium between Ala499Val and two polymorphisms in the 3′UTR (Ex15-184 and Ex15-177; D′ ≥ 0.99 and r2 ≥ 0.82). From Supplementary Table S2, it can be seen that there would be very little loss of information if only 12 of the 22 polymorphisms were considered.
ORs were calculated using logistic regression. Individuals homozygous for the variant allele of Ala499Val, Ex15-184, or Ex15-177 had an increased risk of bladder cancer [adjusted OR (95% CI), 1.65 (1.05-2.59), 1.82 (1.12-2.97), and 1.82 (1.12-2.96), respectively; Table 1]. Individuals homozygous for the variant allele of Arg687Arg or IVS14-51 had a borderline significant decreased risk of bladder cancer compared with those carrying the wild-type genotype [adjusted OR (95% CI), 0.60 (0.35-1.01), and 0.59 (0.35-0.99), respectively]. None of the remaining XPC polymorphisms were associated with bladder cancer risk.
Gene-Environment Interactions
To examine gene-environment interactions, individuals were stratified by smoking status (nonsmokers and combined ex- and current smokers) and occupational exposure (exposure and no exposure), and the association analyses were repeated (Table 2). Our data suggested possible gene-environment interactions between Ala499Val and both smoking and occupational exposure. Smokers homozygous for the variant allele 499Val had an approximately 3-fold increased risk of bladder cancer compared with nonsmokers carrying the wild-type genotype 499Ala/Ala. Similarly, individuals homozygous for the variant allele 499Val with a history of occupational exposure had an approximately 4-fold increased risk of bladder cancer compared with those carrying the wild-type genotype with no occupational exposure. However, the likelihood ratio tests of interaction were not significant (P = 0.64 and 0.57 for smoking and occupational exposure, respectively), consistent with a multiplicative effect for genotype and environment exposure on bladder cancer risk. Results were similar for Ex15-184 and Ex15-177 (data not shown). None of the remaining XPC polymorphisms showed significant gene-environment interactions.
Gene-environment interactions between Ala499Val, smoking, and occupational exposure
Environmental exposure . | Genotypes . | Control (n) . | Cases (n) . | Exposure-specific aOR . | Cumulative aOR . | P for interaction* . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Smoking status | ||||||||||||
Nonsmoker | C/C | 99 | 59 | 1.00 | 0.64 | |||||||
C/T | 79 | 52 | 1.08 (0.67-1.75) | |||||||||
T/T | 12 | 8 | 1.15 (0.44-3.02) | |||||||||
Ex- and current smoker | C/C | 218 | 220 | 1.00 | 1.60 (1.09-2.36) | |||||||
C/T | 131 | 150 | 1.13 (0.83-1.52) | 1.80 (1.20-2.72) | ||||||||
T/T | 26 | 49 | 1.82 (1.09-3.06) | 2.93 (1.09-3.06) | ||||||||
Occupational exposure | ||||||||||||
No | C/C | 265 | 210 | 1.00 | 0.57 | |||||||
C/T | 173 | 143 | 1.06 (0.80-1.42) | |||||||||
T/T | 33 | 39 | 1.49 (0.90-2.46) | |||||||||
Yes | C/C | 52 | 69 | 1.00 | 1.65 (1.08-2.51) | |||||||
C/T | 37 | 59 | 1.16 (0.67-2.03) | 2.22 (1.39-3.54) | ||||||||
T/T | 5 | 18 | 3.04 (1.04-8.91) | 54.17 (1.51-11.5) |
Environmental exposure . | Genotypes . | Control (n) . | Cases (n) . | Exposure-specific aOR . | Cumulative aOR . | P for interaction* . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Smoking status | ||||||||||||
Nonsmoker | C/C | 99 | 59 | 1.00 | 0.64 | |||||||
C/T | 79 | 52 | 1.08 (0.67-1.75) | |||||||||
T/T | 12 | 8 | 1.15 (0.44-3.02) | |||||||||
Ex- and current smoker | C/C | 218 | 220 | 1.00 | 1.60 (1.09-2.36) | |||||||
C/T | 131 | 150 | 1.13 (0.83-1.52) | 1.80 (1.20-2.72) | ||||||||
T/T | 26 | 49 | 1.82 (1.09-3.06) | 2.93 (1.09-3.06) | ||||||||
Occupational exposure | ||||||||||||
No | C/C | 265 | 210 | 1.00 | 0.57 | |||||||
C/T | 173 | 143 | 1.06 (0.80-1.42) | |||||||||
T/T | 33 | 39 | 1.49 (0.90-2.46) | |||||||||
Yes | C/C | 52 | 69 | 1.00 | 1.65 (1.08-2.51) | |||||||
C/T | 37 | 59 | 1.16 (0.67-2.03) | 2.22 (1.39-3.54) | ||||||||
T/T | 5 | 18 | 3.04 (1.04-8.91) | 54.17 (1.51-11.5) |
NOTE: aOR, ORs adjusted for subjects' gender, age, smoking, occupational exposure, and family history of bladder cancer.
P value of the likelihood ratio test for gene-environment interaction.
Haplotype Analysis
Because there was strong linkage disequilibrium among XPC polymorphisms, genotypes for one single nucleotide polymorphism would be expected to generate similar information for other polymorphisms. We chose a simple model in which we selected the four coding polymorphisms with amino acid substitutions, i.e., Leu16Val, Ala492His, Ala499Val, and Lys939Gln, to construct XPC haplotypes. Four common haplotypes accounted for 99% of all cases and controls (Supplementary Table S3). The more frequent XPC haplotype carrying the Ala499Val variant allele (CGTA) was associated with an increased risk of bladder cancer compared with the XPC haplotype carrying all wild-type alleles [crude OR (95% CI), 1.25 (1.00-1.57); P = 0.05].
Association with Muscle-Invasive Disease
The relationship between XPC Ala499Val genotypes and the tumor grade and stage were examined. The median follow-up for bladder cancer cases was 28 months (range, 1-37 months). There was no significant relationship between Ala499Val genotypes and tumor grade (P = 0.98). However, cases carrying the Ala499Val variant allele were more likely to have muscle-invasive disease (either at presentation or due to progression from superficial disease; P = 0.02; Supplementary Table S4). Results were similar for Ex15-184 and Ex15-177 (P = 0.04 and 0.03, respectively). Those with intermediate stage (progression from superficial disease) contributed most to this difference, suggesting that this may be a chance finding.
Discussion
To our knowledge, this is the most comprehensive investigation of the association between XPC polymorphisms and bladder TCC risk. The homozygote variants of Ala499Val and two 3′UTR polymorphisms (Ex15-148 and Ex15-177) were associated with increased bladder TCC risk and the homozygote variants of Arg687Arg and IVS14-51 were associated with borderline reduced bladder TCC risk. Our findings were consistent with a case-control study of Ala499Val in lung cancer (18), but since we undertook this study, the 499Val variant was found not to be associated with bladder cancer in the large Caucasian studies of Wu et al. (9) and Garcia-Closas et al. (10). Neither study investigated the two 3′UTR polymorphisms. The 499Val minor allele frequency in our study (25%) was similar to the frequencies in Wu et al. (24%) and Garcia-Closas et al.'s studies (26%). Our discrepant findings may be explained by false positive or false negative results, or the different populations studied. To date, no functional studies have been undertaken on these polymorphisms. However, SIFT (http://blocks.fhcrcorg/sift//SIFT.html) and Polyphen (http://tux.embl-heidelberg.de/ramensky/) predict no effect on protein function or structure for the Ala499Val substitution. Therefore, one of the 3′UTR polymorphisms may be a functional variant, through its effect on mRNA transcription and stability, in which case, different patterns of linkage disequilibrium may obscure the association in the other studies.
This study was potentially limited by selection bias because both community and hospital controls were included. However, although there were some differences in the genotype distributions of a few intronic polymorphisms significant at the 5% level, we believe it unlikely that the use of these two control groups influenced the results. Recall bias for smoking, occupational exposure, and family history cannot be excluded. Our results must be interpreted cautiously bearing in mind multiple testing as 22 XPC polymorphisms were selected. However, these tests are highly correlated (equivalent to at most 12 independent tests).
Stratified analysis showed that the Ala499Val variant was only associated with bladder TCC risk among ex- and current smokers and those occupationally exposed (Table 2). The likelihood ratio test showed no significant evidence of gene-environment interaction, although this may be due to lack of statistical power. Neither Wu et al. (9) nor Garcia-Closas et al. (10) evaluated the interaction between the Ala499Val variant and environmental exposures (smoking/occupational exposure). However, Wu et al. (9) found a significant interaction between ever smoking and the combined variant alleles (n = 13) in the nucleotide excision repair pathway.
In subgroup analysis of cases only, patients with bladder cancer carrying the variant allele for Ala499Val or the two 3′UTR polymorphisms were more likely to have muscle-invasive disease at presentation or to develop muscle-invasive disease compared with those carrying the wild-type allele (Fisher's exact, P = 0.02). Gu et al. (19) found no effect of the Ala499Val genotype on the risk of superficial tumor progression but patient numbers were small (35 of 288 patients progressed). Muscle-invasive tumors have more genetic abnormalities than superficial tumors (20). We hypothesize that the variant genotypes are associated with an increased frequency of muscle-invasive disease due to reduced DRC causing increased tumor aggressiveness due to increased mutations in oncogenes and tumor suppressor genes, although the association is tentative and requires confirmation.
In conclusion, in our study, the XPC polymorphisms Ala499Val, Ex15-184, and Ex15-177 were associated with an increased risk of bladder cancer and these associations were stronger for smokers or occupationally exposed individuals. As negative results were recently found for Ala499Val in two larger series (9, 10), additional studies are required to confirm or refute the findings of this study.
Grant support: Yorkshire Cancer Research and Cancer Research UK.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Epidemiology Biomarkers and Prevention Online (http://cebp.aacrjournals.org/).
Acknowledgments
We thank Jo Robinson for help with blood collection; Dr. Juliette Randerson-Moor, Dr. Mark Harland, and Michael Churchman for assistance with genotyping; and Mr. G. Kelly and Mr. T. Dabbs for allowing us to approach their patients for use as control subjects.