Genetic Polymorphisms Involved in Carcinogen Metabolism and DNA Repair and Lung Cancer Risk in a Japanese Population

Original Article

Genetic Polymorphisms Involved in Carcinogen Metabolism and DNA Repair and Lung Cancer Risk in a Japanese Population Chikako Kiyohara, PhD,* Takahiko Horiuchi, PhD,† Koichi Takayama, PhD,‡ and Yoichi Nakanishi, PhD‡

Background: Several components of overall lung carcinogenesis, carcinogen metabolic and DNA repair pathways may be involved in individual genetic susceptibility to lung cancer. Methods: We evaluated the role of cytochrome P450 (CYP) 1A1 rs4646903 and rs104894, glutathione S-transferase (GST) M1 and GSTT1 deletion polymorphisms, GSTP1 rs1695, x-ray repair, excision repair cross-complementing group 2 (ERCC2) rs13181, complementing defective in Chinese hamster 1 rs25487, and XRCC3 rs861539 in a casecontrol study comprising 462 lung cancer cases and 379 controls in a Japanese population. Unconditional logistic regression was used to assess the adjusted odds ratios (OR) and 95% confidence intervals (95% CI). Results: CYP1A1 rs4646903 (OR = 1.72, 95% CI = 1.25–2.38), rs1048943 (OR = 1.40, 95% CI = 1.02–1.92), the GSTM1 deletion polymorphism (OR = 1.38, 95% CI = 1.01–1.89), GSTP1 rs1695 (OR =1.48, 95% CI = 1.04–2.11), ERCC2 rs13181 (OR = 1.89, 95% CI = 1.28–2.78), and Chinese hamster 1 rs25487 (OR = 1.54, 95% CI = 1.12–2.13) were associated with lung cancer risk whereas the GSTT1 deletion polymorphism and XRCC3 rs861539 were not. A pertinent combination of multiple “at-risk” genotypes of CYP1A1 rs4646903, the GSTM1 deletion polymorphism and ERCC2 rs13181 was at a 5.94-fold (95% CI = 2.77–12.7) increased risk of lung cancer. Conclusions: A pertinent combination of multiple at-risk genotypes may detect a high-risk group. Further studies are warranted to verify our findings. Key Words: Carcinogen metabolism, DNA repair, Epidemiology, Genetic polymorphism, Lung cancer. (J Thorac Oncol. 2012;7: 954–962)

L

ung cancer is a major cause of cancer-related death in developed countries, and the overall survival rate is still

*Department of Preventive Medicine; †Department of Medicine and Biosystemic Science; and ‡Research Institute for Diseases of the Chest, Graduate School of Medical Sciences, Kyushu University, Japan. Supported by a Grant-in-Aid for Scientific Research (B) (2139019) from the Ministry of Education, Science, Sports and Culture, Japan. Disclosure: The authors declare no conflicts of interest. Address for Correspondence: Dr. Chikako Kiyohara, Department of Preventive Medicine, Graduate School of Medical Sciences, Kyushu University, Maidashi 3-1-1, Higashi-ku, Fukuoka 812–8582, Japan. E-mail: [email protected] Copyright © 2012 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/12/0706-0954

954

extremely poor. Although tobacco smoking is an established risk factor for lung cancer, approximately one in 10 smokers develops lung cancer in their lifetime, indicating an interindividual variation in susceptibility to tobacco smoke.1 Various DNA alterations can be caused by exposure to tobacco smoke and environmental and endogenous carcinogens. Cells with damaged DNA are usually destroyed through apoptosis; however, aberrant cells may escape normal growth control and acquired mutations may alter apoptosis, thereby allowing for the development of lung cancer.2 Carcinogenesis therefore requires multiple genetic changes, as can occur within the context of long-term, repeated exposure to carcinogenic compounds.3,4 Furthermore, individuals may have a unique combination of polymorphic traits that modify genetic susceptibility and response to environmental carcinogens. Such carcinogens must be metabolically activated to exert their deleterious effects by phase-I enzymes such as those encoded by the cytochrome P450 (CYP) supergene family, but this is counteracted by the ongoing detoxification of carcinogens by phase-II enzymes such as those encoded by the glutathione S-transferases (GST) supergene family.5 Therefore, DNA damage itself is a balance between activation and detoxification of carcinogens that involve phase-I and -II metabolic enzymes, many of which are polymorphic. The capacity to repair DNA damage induced by activated carcinogens is also a host factor that may influence lung cancer risk.6 Numerous epidemiological studies suggest that genetic polymorphisms involved in activation, detoxification, and DNA repair are associated with increased risk of developing lung cancer.7,8 Identifying metabolic and DNA-repair polymorphisms involved with exposure to carcinogens will help clarify the process by which lung cancer develops, and may thus indirectly lead to prevention. Candidate-susceptibility genes for lung cancer have been extensively studied, with most of the work focusing on mechanistically plausible polymorphisms in genes coding for enzymes involved in the activation, detoxification, and DNA repair of damage caused by tobacco smoke. Alterations in these pathways are hypothesized to affect an individual’s processing of tobacco carcinogens and therefore their risk of developing lung cancer. Thus it is expected that pertinent components of lung carcinogenesis may be involved in individual genetic susceptibility to tobacco-smoke exposure. As CYP1A1, GSTM1, GSTT1, GSTP1, excision repair cross-complementing group 2 (ERCC2), Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

x-ray repair complementing defective in Chinese hamster 1 (XRCC1), and XRCC3 enzymes are involved in the tobacco-smoke carcinogenesis process, the genes, which code for these enzymes, have been of considerable interest as candidate-susceptibility genes for lung cancer.5 As a functional polymorphism alters the expression of a gene or the structure/function of the gene product, a functional polymorphism may have some direct and/or immediate clinical relevance. CYP1A1 rs4646903 and rs1048943,9 GSTM1 and GSTT1 deletion polymorphisms, GSTP1 rs1695,10–12 ERCC2 rs13181,13,14 XRCC1 rs2548715,16 and XRCC3 rs86153917 are potentially functional. The aim of this study is to assess the possible association of genetic polymorphisms in genes coding metabolic activation (CYP1A1 rs4646903 and rs1048943), detoxification (GSTM1, GSTT1 deletion polymorphisms, and GSTP1 rs1695) and DNA-repair enzymes (ERCC2 rs13181 in the nucleotide excision repair (NER) pathway, XRCC1 rs25487 in the base excision repair (BER) pathway, and XRCC3 rs861539 in the double-strand break repair (DSBR) pathway in pertinent components of overall lung carcinogenesis.

PATIENTS AND METHODS Study Subjects and Data Collection Lung cancer patients were recruited at Kyushu University Hospital (Research Institute for Diseases of the Chest, Kyushu University, Japan) and its collaborating hospitals. Eligible cases were newly diagnosed, and histologically confirmed primary lung cancer cases from 1996 to 2008 were selected. Histological types were categorized into four major types according to the International Classification of Diseases for Oncology (ICD-O), second edition: adenocarcinoma (8140, 8211, 8230–8231, 8250–8260, 8323, 8480–8490, 8550–8560, 8570–8572), squamous cell carcinoma (8050–8076), smallcell carcinoma (8040–8045) and large-cell carcinoma (8012– 8031, 8310). The participation rate among the cases was 100%. Potential controls (379) were selected from inpatients without a clinical history of any type of cancer, past or present, ischemic heart disease or chronic respiratory diseases, who were admitted to departments other than the departments of respiratory medicine in collaborating hospitals during the same period because hospital controls are more motivated and are more easily accessible for DNA samples. Controls were not, individually or in larger groups, matched to cases. Controls were approached by their attending physicians to be recruited as control subjects. None of the controls refused to participate in this study. A self-administered questionnaire was used to collect data on demographic and lifestyle factors such as age, years of education, smoking, alcohol consumption, environmental tobacco exposure from spouse, etc. All subjects were unrelated ethnic Japanese. The study protocol was approved by our institutional review board, and all participants provided written informed consent.

Genetic Analyses Genomic DNA was extracted from blood samples. Genotyping was conducted with blinding to case/control

Genetic Polymorphisms and DNA Repair and Lung Cancer

status. The CYP1A1 rs4646903 and rs1048943 polymorphisms were genotyped by the methods described by Hayashi et al.18 and Nakachi et al.,19 respectively. The GSTM1 and GSTT1 deletion polymorphisms were determined by the methods described by Zhong et al.20 and Chenevix-Trench et al.,21 respectively. The genotyping of ERCC2 rs13181, XRCC1 rs25487, and XRCC3 rs861539 was performed using the method described by Matullo et al.17 For quality control, both assays were repeated randomly on 5% of all samples, and the replicates were 100% concordant.

Statistical Analysis Comparisons of means and proportions were based on unpaired t test and χ2 test, respectively. The distribution of the CYP1A1 rs4646903, CYP1A1 rs1048943, GSTP1 rs1695, ERCC2 rs13181, XRCC1 rs25487 or XRCC3 rs861539 genotypes in controls was compared with that expected from Hardy-Weinberg equilibrium (HWE) by the χ2 (Pearson) test, and linkage disequilibrium (LD) was assessed with D’ and r2. HWE was not tested for GSTM1 and GSTT1 deletion polymorphisms because the genotyping methods we employed did not distinguish between non-null heterozygotes and non-null homozygotes. We hypothesized that lung cancer patients would carry more active alleles (CYP1A1), null activity alleles (GSTM1, GSTT1), and less active alleles (GSTP1, ERCC2, XRCC1, XRCC3) than the control subjects. Unconditional logistic regression was used to compute the odds ratios (ORs) and their 95% confidence intervals (CIs), with adjustments for several covariates (age, sex, smoking status, drinking status, and education) found to be associated with risk. Subjects were considered current-smokers if they had smoked or stopped smoking less than 1 year before either the date of diagnosis (lung cancer patients) or the date of completion of the questionnaires (controls). Nonsmokers were defined as those who had never smoked in their lifetime. Former-smokers were those who had stopped smoking 1 or more years before either the date of diagnosis of lung cancer (lung cancer patients) or the date of completion of the questionnaires (controls). Based on “Healthy Japan 21” (National Health Promotion in the 21st Century), heavy drinkers were defined as those who drank more than 60 g of alcohol per day.22 As “Healthy Japan 21” has emphasized drinking an appropriate volume of alcohol (20 g of alcohol per day), appropriate drinkers were defined as those who did not exceed 20 g of alcohol intake per day. The appropriate volume of alcohol use may have a protective effect on life expectancy and morbidity.23 Moderate drinkers were defined as those who drank more than 20 g alcohol per day but not exceeding 60 g per day. Unlike cigarette smoking, ingested alcohol is eliminated from the body by various metabolic mechanisms, and the alcohol elimination process begins almost immediately. Significant relationships between excessive drinking and lung cancer have been reported whereas appropriate drinking has not shown the same effects.24 In terms of alcohol consumption, the subjects were classified into the following two groups on the basis of their intake for at least 1 year: those who drink more than 20

Copyright © 2012 by the International Association for the Study of Lung Cancer

955

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Kiyohara et al.

g of alcohol per day (excessive drinkers) and those who drink less than 20 g of alcohol per day (appropriate drinkers). Genotype impact was assessed by a score test for each genotype as follows: 0, homozygous for the major allele; 1, heterozygous; and 2, homozygous for the minor allele. The cumulative “at-risk” genotype (at least one at-risk allele) impact was assessed by ordinal scores 0 (no at-risk genotypes), 1 (one at-risk genotype), 2 (two at-risk genotypes) and 3 (three at-risk genotypes). Gene–gene and gene–gene–gene interactions were statistically evaluated on the basis of the likelihood test, comparing the models with and without (multiplicative scale) terms for interaction. LD between the CYP1A1 rs4646903 and rs1048943 polymorphisms was calculated with HaploView software (version 4.2, Broad Institute, Cambridge, MA). 25 All statistical analyses were performed using the computer program STATA Version 10.1 (STATA Corporation, College Station, TX). All p values were two-sided, with those less than 0.05 considered statistically significant.

20 with large-cell carcinoma, data not shown). As controls were not selected to match lung cancer patients on age and sex, there were significant differences in age (p < 0.001) and sex ratios (p < 0.001) between lung cancer patients and controls. Compared with control subjects, the lung cancer patients were more likely to report a history of smoking (p < 0.001) and a history of excessive drinking (p < 0.001). Naturally, pack-years of smoking was higher among cases than controls (p < 0.001). Lung cancer patients were less likely to be educated compared to the control subjects (p < 0.001). Table 2 shows associations between metabolic polymorphisms and lung cancer risk. Genotype distributions of CYP1A1 rs4646903, CYP1A1 rs1048943 and GSTP1 rs1695 were consistent with HWE among controls. Among controls, the frequency of the CC genotype of CYP1A1 rs4646903 was 11.4% whereas that of the Val/Val genotype of CYP1A1 TABLE 2.  Association Between Metabolic Polymorphisms and Risk of Lung Cancer

RESULTS The distributions of selected characteristics among subjects are summarized in Table 1. Our analysis included 462 lung cancer patients (242 with adenocarcinoma, 131 with squamous cell carcinoma, 69 with small-cell carcinoma, and TABLE 1.  Selected Characteristics of Lung Cancer Cases and Controls Characteristics

Cases (n = 462)

Controls (n = 379)

p

Age (yr), mean (95% CI) Sex, n (%)   Male   Female Smoking status, n (%)   Current smoker   Former smoker   Never smoker Pack-years, mean (95% CI) Drinking status, n (%)   Appropriate drinker*   Excessive drinker† Exposure to environmental tobacco smoke among nonsmokers, n (%)   Positive   Negative Education, mean   (95% CI)

66.1 (65.2–66.9)

55.9 (54.6–57.2)

<0.001

287 (62.1) 175 (37.9)

283 (74.7) 96 (25.3)

198 (42.9) 111 (24.0) 153 (33.1) 36.4 (33.3–39.5)

129 (34.0) 41 (10.8) 209 (55.2) 16.0 (13.8–18.2)

<0.001 <0.001

284 (61.5)

175 (46.2)

<0.001

178 (38.5)

204 (53.8)

284 (61.5)

175 (46.2)

99 (64.7) 54 (35.3) 14.0 (13.8–14.2)

135 (64.6) 74 (35.4) 14.9 (14.6–15.1)

95% CI, 95% confidence interval * Subjects who drink less than 20 g of alcohol per day. †    Subjects who drink more than 20 g of alcohol per day.

956

<0.001

Polymorphism CYP1A1 rs4646903 TT (ancestral†) TC CC

Cases n (%)

Controls n (%)

148 (32.0) 218 (47.2) 96 (20.8) p‡<0.0001

169 (44.6) 167 (44.1) 43 (11.4) § p = 0.857

TC + CC vs. TT CYP1A1 rs1048943 Ile/Ile (AA, ancestral†) Ile//Val (AG) Val/Val (GG)

<0.001

GSTT1 deletion Non-null Null GSTP1 rs1695 Ile/Ile (AA) Ile//Val (AG) Val/Val (GG,   ancestral†)

0.982 <0.001

Ile/Val + Val/Val vs. Ile/Ile

1.0 (reference) 1.50 (1.06–2.11) 2.63 (1.61–4.28) ptrend < 0.0001 1.72 (1.25–2.38)

p

0.021 < 0.001 0.001

248 (53.7)

233 (61.5)

1.0 (reference)

160 (34.6) 54 (11.7) p‡ = 0.004

125 (33.0) 21 (5.54) § p = 0.436

1.18 (0.84–1.66) 2.86 (1.54–5.32) ptrend = 0.003 1.40 (1.02–1.92)

0.040

Ile/Val + Val/Val vs. Ile/Ile GSTM1 deletion Non-null Null

Adjusted* Odds Ratio (95% Confidence Interval)

0.338 0.001

194 (42.0) 268 (58.0) p‡ = 0.008

194 (51.2) 185 (48.8)

1.0 (reference) 1.38 (1.01–1.89)

0.044

245 (53.0) 217 (47.0) p‡ = 0.284

215 (56.7) 164 (43.3)

1.0 (reference) 1.17 (0.86–1.61)

0.322

323 (69.9) 120 (26.0) 19 (4.11)

289 (76.3) 84 (22.2) 6 (1.58)

1.0 (reference) 1.37 (0.94–1.97) 3.22 (1.12–9.30)

0.098 0.031

p‡ = 0.032

p§ = 0.971

ptrend = 0.012 1.48 (1.04–2.11)

0.031

* Adjusted for age, sex, education, smoking status, and drinking status. †    Defined by National Center for Biotechnology Information SNP database. ‡    p for χ2 test. §    p for Hardy-Weinberg equilibrium test among controls.

Copyright © 2012 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Genetic Polymorphisms and DNA Repair and Lung Cancer

TABLE 3.  Association Between the Polymorphisms Involved in DNA Repair and Risk of Lung Cancer Polymorphism ERCC2 rs13181 Lys/Lys (AA) Lys/Gln (AC) Gln/Gln (CC, ancestral)

Cases n (%)

Controls n (%)

Adjusted* Odds Ratio (95% Confidence Interval)

332 (71.9) 111 (24.0) 19 (4.11) ‡ p = 0.003

308 (81.3) 65 (17.2) 6 (1.58) § p = 0.240

1.0 (reference) 1.61 (1.07–2.42) 5.61 (1.99–15.8) ptrend < 0.0001 1.89 (1.28–2.78)

Lys/Gln + Gln/Gln vs. Lys/Lys XRCC1 rs25487 Arg/Arg (GG, ancestral†) Arg/Gln (GA) Gln/Gln (AA)

243 (52.6) 171 (37.0) 48 (10.4) ‡ p < 0.0001

242 (63.9) 121 (31.9) 16 (4.22) § p = 0.859

Arg/Gln + Gln/Gln vs. Arg/Arg XRCC3 rs861539 Thr/Thr (CC, ancestral†) Thr/Met (CT) Met/Met (TT)

352 (76.2) 97 (21.0) 13 (2.81) ‡ p = 0.626

295 (77.8) 77 (20.3) 7 (1.85) § p = 0.455

Thr/Met + Met/Met vs. Thr/Thr

1.0 (reference) 1.38 (0.98–1.94) 2.59 (1.36–4.95) ptrend = 0.002 1.54 (1.12–2.13)

1.0 (reference) 0.95 (0.65–1.39) 1.34 (0.47–3.78) ptrend = 0.929 0.98 (0.68–1.42)

p

0.022 0.001 0.001

0.067 0.004 0.008

0.778 0.584 0.915

* Adjusted for age, sex, education, smoking status, and drinking status. †    Defined by National Center for Biotechnology Information SNP database. ‡    p for χ2 test. §    p for Hardy-Weinberg equilibrium test among controls.

rs1048943 was 5.54%. The frequencies of the null genotypes of GSTM1 and GSTT1, and the Val/Val genotype of GSTP1 rs1695 among controls were 48.8%, 43.3%, and 1.58%, respectively. After adjustment for age, sex, education, smoking status, and drinking status, the minor homozygotes of the CYP1A1 rs4646903 (OR = 2.63, 95% CI = 1.61–4.28, p < 0.001), CYP1A1 rs1048943 (OR = 2.86, 95% CI = 1.54–5.32, p = 0.001) and GSTP1 rs1695 (OR = 3.22, 95% CI = 1.12– 9.30, p = 0.031) polymorphisms were significantly associated with an increased lung cancer risk. We excluded pack-years (the number of packs of cigarettes smoked/day multiplied by years of smoking) from the logistic models because of high correlation (pack-year increases with advancing age, correlation coefficient = 0.52, p <0.0001) with age (avoiding the problem of potential collinearity). The null genotype of the GSTM1 deletion polymorphism was at a 1.38-fold (95% CI = 1.01–1.89, p = 0.044) increased risk of lung cancer whereas the null genotype of the GSTT1 deletion polymorphism was not. Table 3 shows associations between the polymorphisms involved in DNA repair and lung cancer risk. The frequencies of the minor genotypes of ERCC2 rs13181, XRCC1 rs25487 and XRCC3 rs861539 among controls were 1.58%, 4.22% and 1.85%, respectively. ERCC2 rs13181, XRCC1 rs25487, and XRCC3 rs861539 polymorphisms did not deviate from HWE in controls. An increased risk of lung cancer was seen for subjects carrying either minor homozygote of ERCC2 rs13181 (OR = 5.61, 95% CI = 1.99–15.8, p = 0.001) or XRCC1 rs25487 (OR = 2.59, 95% CI = 1.36–4.95, p = 0.004) compared with

subjects carrying the corresponding major homozygote. The minor homozygous genotype of XRCC3 rs861539 was not associated with lung cancer risk. On the basis of our hypothesis, we designated the allele that is presumed to increase the risk of lung cancer as the at-risk allele. To achieve adequate statistical power, subjects with at least one at-risk allele (termed at-risk genotype) were bundled in one group for subsequent analysis. Because the CYP1A1 rs4646903 polymorphism was at LD with the CYP1A1 rs1048943 polymorphism (D’ = 0.82, r2 = 0.31, data not shown), rs1048943 was excluded from the subsequent analysis. Table 4 shows associations between polymorphisms involved in metabolic (the CYP1A1 rs4646903 and GSTM1 deletion polymorphisms) pathways and three DNA-repair pathways (NER, BER and DSBR) and lung cancer risk. As for ERCC2 rs13181 (NER)-containing pertinent combination, multiple at-risk genotypes dose dependently increased lung cancer risk (OR = 1.67, 95% CI =1.03–2.72 for one at-risk genotype; OR = 1.67, 95% CI =1.03–2.72 for two “at-risk” genotypes; OR = 5.94, 95% CI = 2.77–12.7; p < 0.001). In the remaining two pertinent combinations, increasing numbers of at-risk genotypes also increased lung cancer risk in a dose dependent manner (ptrend < 0.0001 for rs25487; ptrend = 0.002 for rs861539). The impact of the combination of multiple at-risk genotypes was strongest in the combination including ERCC2 rs13181 and weakest in the combination including XRCC3 rs861539. No threeway gene–gene–gene interactions were significant (data not shown).

Copyright © 2012 by the International Association for the Study of Lung Cancer

957

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Kiyohara et al.

TABLE 4.  Combinations of CYP1A1 rs464903, GSTM1 Deletion Polymorphism and DNA-Repair Polymorphisms and Risk of Lung Cancer Number of “at-risk” genotypes Phase I

Phase II

DNA Repair

Cases/ Controls

Adjusted* OR (95% CI)

rs4646903 0 1 0 0 1 1 0 1

GSTM1 0 0 1 0 1 0 1 1

rs13181 0 0 0 1 0 1 1 1

44/72 100/81 67/68 15/15 121/87 35/26 22/14 58/16

1.0 (reference) 1.87 (1.09–3.20) 1.40 (0.79–2.48) 1.96 (0.73–5.27) 2.16 (1.27–3.66) 2.22 (1.07–4.62) 2.43 (1.00–5.87) 5.94 (2.77–12.7)

0.023 0.246 0.182 0.004 0.033 0.049 <0.001

rs4646903 0 1 0 0 1 1 0 1

GSTM1 0 0 1 0 1 0 1 1

rs25487 0 0 0 1 0 1 1 1

29/54 68/62 55/54 30/33 91/72 67/45 34/28 88/31

1.0 (reference) 1.99 (1.04–3.78) 1.48 (0.76–2.89) 1.60 (0.74–3.46) 2.22 (1.19–4.14) 2.29 (1.17–4.50) 2.16 (1.00–4.67) 4.55 (2.29–9.05)

0.037 0.246 0.228 0.013 0.016 0.050 <0.001

rs4646903 0 1 0 0 1 1 0 1

GSTM1 0 0 1 0 1 0 1 1

rs861539 0 0 0 1 0 1 1 1

49/67 100/83 77/59 10/20 126/86 35/24 12/23 53/17

1.0 (reference) 1.58 (0.92–2.73) 1.50 (0.84–2.66) 0.68 (0.27–1.72) 1.95 (1.14–3.31) 1.57 (0.77–3.21) 0.68 (0.28–1.64) 3.41 (1.62–7.19)

0.099 0.171 0.410 0.014 0.214 0.391 0.001

Adjusted* OR (95% CI)

p

} } } } } }

p

1.0 (reference) 1.67 (1.03–2.72)

0.038

2.20 (1.34–3.61)

0.002

5.94 (2.77–12.7) ptrend < 0.001

<0.001

1.0 (reference) 1.71 (0.97–3.03)

0.066

2.23 (1.26–3.95)

0.006

4.55 (2.29–9.05) ptrend < 0.001

<0.001

1.0 (reference) 1.43 (0.88–2.34)

0.150

1.64 (1.00–2.71)

0.050

3.41 (1.62–7.19) ptrend = 0.002

0.001

* Adjusted for age, sex, education, smoking status, and drinking status.

Table 5 presents the associations between polymorphisms involved in metabolic (the CYP1A1 rs4646903 and GSTP1 rs1695 polymorphisms) and three DNA-repair pathways and lung cancer risk. In three pertinent combinations, the risk was more than fourfold in subjects with three at-risk genotypes, compared to those with no at-risk genotypes. A dose-dependent relationship between lung cancer risk and the number of at-risk genotypes was also observed in all pertinent combinations. The combination including ERCC2 rs13181 produced the highest OR (OR = 5.76, 95% CI = 1.69–15.8). Three-way gene–gene–gene interactions did not reach statistical significance (data not shown). We further examined two-way gene–gene interactions either within one pathway or between two pathways. Significant interactions were found between CYP1A1 rs4646903 and XRCC3 rs861539 (p for interaction = 0.030), GSTT1 deletion polymorphism and ERCC2 rs13181 (p for interaction = 0.008), and GSTP1 rs1695 and XRCC3 rs861539 (p for interaction = 0.001). All other two-way

958

gene–gene interactions were not statistically significant (data not shown).

DISCUSSION The polymorphisms involved in carcinogen metabolism and DNA repair such as CYP1A1 rs4646903, CYP1A1 rs1048943, GSTM1 deletion, GSTT1 deletion, GSTP1 rs1695, ERCC2 rs13181, XRCC1 rs25487, and XRCC3 rs861539 were determined in a total of 841 Japanese subjects (462 lung cancer cases and 379 controls). The CC (minor) genotype is most common among Asians (14%) and least common among whites (1.2%) with African-Americans (5.9%) intermediate between these groups.26 The prevalence of the CC genotype in this control population (11.4%) was similar to that reported in other Japanese populations (8.0%,27 10.4%,28 10.6%19 and 12.9%29). Thus, the minor homozygous genotype in Asians was about ten times more than that in whites. The CC genotype of CYP1A1 rs4646903 was at a 2.63-fold increased risk

Copyright © 2012 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Genetic Polymorphisms and DNA Repair and Lung Cancer

TABLE 5.  Combinations of CYP1A1 rs464903, GSTP1 rs1695 and DNA Repair Polymorphisms and Risk of Lung Cancer Number of “at-risk” genotypes Phase II

DNA Repair

Cases/ Controls

Adjusted* OR (95% CI)

p

rs4646903 0 1 0 0 1 1 0 1

rs1695 0 0 1 0 1 0 1 1

rs13181 0 0 0 1 0 1 1 1

89/107 149/123 22/33 21/24 72/45 64/35 16/5 29/7

1.0 (reference) 1.49 (0.97– 2.28) 0.90 (0.46–1.78) 1.11 (0.52–2.41) 2.08 (1.21–3.56) 2.46 (1.37–4.39) 5.85 (1.69–20.2) 5.76 (1.69–15.8)

0.066 0.770 0.782 0.008 0.002 0.005 0.001

rs4646903 0 1 0 0 1 1 0 1

rs1695 0 0 1 0 1 0 1 1

rs25487 0 0 0 1 0 1 1 1

57/86 109/102 27/22 53/45 50/32 104/56 11/16 51/20

1.0 (reference) 1.78 (1.09–2.91) 1.77 (0.85–3.70) 1.77 (0.97–3.22) 2.52 (1.32–4.80) 2.62 (1.53–4.47) 1.62 (0.63–4.16) 4.08 (2.02–8.23)

0.022 0.130 0.064 0.005 <0.001 0.312 <0.001

rs4646903 0 1 0 0 1 1 0 1

rs1695 0 0 1 0 1 0 1 1

rs861539 0 0 0 1 0 1 1 1

95/94 172/124 31/32 15/37 54/45 41/34 7/6 47/7

1.0 (reference) 1.46 (0.95–2.23) 1.10 (0.57–2.12) 0.43 (0.21–0.90) 1.46 (0.83–2.57) 1.16 (0.63–2.12) 1.30 (0.38–4.41) 4.99 (2.04–12.2)

0.083 0.774 0.026 0.193 0.638 0.672 <0.001

Phase I

Adjusted* OR (95% CI)

} } } } } }

p

1.0 (reference) 1.33 (0.89–1.98)

0.162

2.41 (1.54–3.77)

<0.001

5.76 (1.69–15.8) ptrend < 0.001

0.001

1.0 (reference) 1.78 (1.13–2.79)

0.013

2.46 (1.52–3.96)

<0.001

4.08 (2.02–8.23) ptrend < 0.001

<0.001

1.0 (reference) 1.19 (0.80–1.78)

0.388

1.31 (0.82–2.09)

0.252

4.99 (2.04–12.2) ptrend = 0.002

<0.001

* Adjusted for age, sex, education, smoking status, and drinking status. OR, odds ratio; CI, confidence interval.

of lung cancer. Therefore, this association was frequently reproduced in Asian populations but not in non-Asian populations because of the low prevalence of the C allele (low statistical power).30 Substrates for and inducers of CYP1A1 include polycyclic aromatic hydrocarbons such as benzo(a)pyrene (BP) in tobacco smoke. A genotypephenotype association study of CYP1A1 demonstrated that the C allele of CYP1A1 rs4646903 is associated with high CYP1A1 inducibility.9 Therefore, it is biologically plausible that the CC genotype associated with high enzyme activity was related to an increased risk of lung cancer. Incidentally, CYP1A1 rs1048943 (Ile462Val) linked to CYP1A1 activity,9 and was also shown to be associated with lung cancer in a Japanese population.31,32 As shown in our study, this polymorphism is closely linked to CYP1A1 rs4646903.31,33 The phenotypic absence of GSTM1 and GSTT1 activity is a result of the homozygous deletion of these genes.10,11 The GSTP1 rs1695 single nucleotide polymorphism (SNP)

(Ile105Val) is an A (Ile) to G (Val) transition that confers reduced catalytic activity.12 The homozygous deletion of the GSTM1 gene was shown to occur in approximately 50% of the population of various ethnic origins whereas the frequency of the GSTT1 null genotype in white populations is 30% or less; the frequency of the null genotype in Asian populations is similar to that of the GSTM1 null genotype.34 Individuals with the Val/Val genotype of GSTP1 rs1695 are most common among African-Americans (19%) and least common among Japanese (0–3.1%) with whites (6.5–11.7%) intermediate between these groups.35 The frequencies of the GST genotypes were similar to the reported figures in Japanese populations. In this study, the GSTM1 null genotype and the Val/ Val genotype of GSTP1 rs1695 were significantly associated with an increased risk of lung cancer but the GSTT1 null genotype was not. The GSTM1 deletion polymorphism has most widely been evaluated for the association with lung cancer risk and a significant association was found in both whites and Asians in a meta- and pooled analyses.36 However, the

Copyright © 2012 by the International Association for the Study of Lung Cancer

959

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Kiyohara et al.

impact of the GSTM1 null genotype was somewhat stronger in Asians (OR = 1.33, 95% CI = 1.06–1.67) than whites (OR = 1.10, 95% CI = 1.01–1.19). In contrast, the deletion GSTT1 polymorphism has not been suggested to be associated with lung cancer in other meta- and pooled analyses.37 In a pooled analysis, Asians who carried at least one Val allele of GSTP1 rs1695 compared with those with the Ile/Ile genotype were shown to be at increased risk of lung cancer (OR = 1.35, 95% CI = 1.07–1.70).38 The GSTM1 enzyme catalyzes the detoxification of genotoxins including aromatic hydrocarbon epoxides such as BP 7,8-diol 9,10-epoxide (BPDE) and products of oxidative stress such as DNA hydroperoxides.39,40 Similarly, the GSTP1 enzyme can utilize a variety of potential carcinogens, including cigarette-smoke–derived chemicals such as BPDE and acrolein.41 The GSTT1 enzyme utilizes potential carcinogens including constituents of cigarette smoke such as alkyl halides.11 Cigarette smoke is causally related to BPDEDNA adduct formation.42,43 BP is metabolically activated to BPDE by CYP1A1. Polycyclic aromatic hydrocarbons-DNA adducts, especially the BPDE-DNA adduct, have been linked to an increased risk of lung cancer.44,45 It is biologically plausible that the genotypes involved in the accumulation of BPDE (null genotype of GSTM1 or the Val/Val genotype of GSTP1 rs1695) may increase lung cancer risk. At least four pathways of DNA repair, such as BER, NER, mismatch repair, and DSBR, operate on specific types of damaged DNA.46 Many studies have investigated the associations of lung cancer with the genetic polymorphisms involved in the BER (repair of small DNA lesions, such as oxidized or reduced bases, nonbulky adducts, and DNA single-strand break),47 NER (repair of bulky DNA adducts, crosslinks and oxidative damages),47 and DSBR pathways in different populations. Among the known functional genetic polymorphisms of the DNA-repair genes, ERCC2 rs13181, XRCC1 rs25487, and XRCC3 rs86539 have been studied most commonly. The Gln/Gln genotype of ERCC2 rs13181 was common in Europe and North America (10%–15%) whereas this genotype was uncommon in Asians (about 1%).48 The frequency of the Gln/ Gln genotype in our study was comparable to that in Asian studies. Among genetic polymorphisms in the NER pathway, an increased risk of lung cancer patients carrying the Gln/Gln genotype of ERCC2 rs13181 (Lys751Gln) was found in our meta-analysis (OR = 1.30, 95% CI = 1.14–1.49).49 The Gln/ Gln genotype of ERCC2 rs13181 was reported to be associated with lower DNA-repair capacity.13,14 XRCC1 rs25487 is a more common polymorphism than rs1799782 (Arg194Trp) and rs25489 (Arg280His), and there was no major variation by ethnicity.50 The frequency of the Gln/Gln genotype in our study fell within the reported figures (3.3%– 9.4% among Asians).50 Among genetic polymorphisms in the BER pathway, the Gln/Gln genotype of the XRCC1 rs25487 (Arg399Gln) polymorphism was associated with an increased risk of lung cancer in our other meta-analysis (OR = 1.34, 95% CI = 1.16–1.54).51 A significant association between the Gln/Gln genotype of XRCC1 rs23487 and decreased DNA capacity has been reported.15,16 According to the haplotype map (HapMap) SNP database,52 the T allele frequency of rs8615139 is most common among whites (42.9%) and least common among Han Chinese (5.8%); Japanese (11%) have intermediate

960

frequencies. The frequency of the T allele in our study (14%) was similar to that in HapMap SNP database. As the Met allele of XRCC3 rs861539 (Thr241Met) was significantly associated with higher DNA adduct levels, it may be associated with low DNA-repair capacity.17 The association between this polymorphism and lung cancer risk remains inconsistent, however. As the BER and NER pathways play a pivotal role in the removal of DNA damage induced by cigarette smoking, it is plausible that ERCC2 rs13181 and XRCC1 rs25487 but not XRCC3 rs81539 were associated with lung cancer risk. To the best of our knowledge, the association between lung cancer risk and the pertinent combination of multiple atrisk genotypes has not been explored. Individuals who carry three at-risk genotypes in the pathways of metabolic and DNA repair had a greater risk of developing lung cancer (Tables 4 and 5). In particular, the pertinent combination of multiple at-risk genotypes of CYP1A1 rs4646903, GSTM1 deletion polymorphism, and ERCC2 rs13181 was at a 5.94-fold increased risk of lung cancer. This finding is biologically plausible because CYP1A1, GSTM1, and ERCC2 enzymes are relevant to BPDEDNA adduct formation/removal. Similarly, the pertinent combination of multiple at-risk genotypes of CYP1A1 rs4646903, GSTP1 rs1695 and ERCC2 rs13181 was at a 5.76-fold increased risk of lung cancer. However, the observed OR was slightly higher than the expected one (5.94 > 4.49 = 1.72*1.38*1.89 for GSTM1-containing and 5.76 > 4.81 = 1.72*1.48*1.89 for GSTP1-containing). This was just as valid for the remaining combinations. This result suggests the absence of gene–gene–gene interactions. Although the OR for the combination of CYP1A1 rs4646903, GSTP1 rs1695, and XRCC3 rs25487 (4.99) was twofold higher than the expected OR (2.49 = 1.72*1.48*1.54), there is no evidence of interaction. Although the aim of this study is to assess the possible association between lung cancer risk and the pertinent combination of multiple at-risk genotypes in three pathways, we further examined two-way gene–gene interactions. Most epidemiological studies on lung cancer have considered main effects of single-nucleotide polymorphisms or gene-environment interactions and rarely gene–gene interactions. Several studies have suggested a possible interaction between GSTM1 deletion polymorphism and CYP1A1 rs4646903 or rs1048943.53–55 Although we could not detect such interactions, the combinations of CYP1A1 rs4646903 and XRCC3 rs861539, GSTT1 deletion polymorphism and ERCC2 rs13181, and GSTP1 rs1695 and XRCC3 rs861539 showed a significant interaction. Variations in lifestyle factors may complicate the validation of the polymorphism– polymorphism (gene–gene) interactions. Another reason for difficulty in validating the gene–gene interactions may be the differences in study design and study-sample characteristics used. Further studies, which will need to be substantially larger than the association studies published to date, are needed to validate our findings. A limitation of our study population was the sample size. We cannot exclude chance as an explanation of the significant association observed. However, as the minor allele frequency of CYP1A1 rs4646903 is relatively higher in Japanese than in other ethnic populations,26 the Japanese are a suitable population for study of the role of either CYP1A1 rs4646903 alone or with other polymorphisms in human carcinogenesis.

Copyright © 2012 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Case-control studies tend to be susceptible to selection bias, particularly in the control group. Controls should represent the source population from which the cases were drawn. When using hospital-based cases, it may not be possible to define the population from which the cases were drawn. Hospital controls may be more appropriate because the study population can be defined as potential hospital users.56 Generally, the reported participation rates were slightly higher in hospitalbased case-control studies than in population-based case-control studies.57 Participation rates of cases and controls were very high in this study. High participation rate may reduce the possibility of selection bias.57 However, as the possibility of selection biases could not be completely excluded in case-control studies, our findings should be interpreted with caution. The current sample size may prevent any conclusive inference. Replication of findings from additional studies with larger sample sizes is very important before any causal inference can be drawn. In conclusion, the best strategy at present may not be a simple combination of multiple at-risk genotypes but a pertinent combination of multiple at-risk genotypes that are involved in the overall lung carcinogenesis pathway. Although lung-cancer–associated polymorphisms should be studied for a pertinent combination of multiple at-risk genotypes, no such studies have been completed to date. The pertinent combination of multiple at-risk genotypes of CYP1A1 rs4646903, GSTM1 deletion polymorphism, and ERCC2 rs13181 was at high risk of lung cancer. This result is biologically plausible because CYP1A1, GSTM1, and ERCC2 enzymes are relevant to BPDE-DNA adduct formation/removal. However, as at-risk genotype combinations are rare in the general population, the public-health implications of our finding for primary prevention actions should be carefully considered. Larger studies are needed to validate the risk associated with these rare genotypes that identify at-risk genotypes of metabolic and DNA repair as potential contributors to lung cancer risk. REFERENCES 1. Kellermann G, Jett JR, Luyten-Kellermann M, Moses HL, Fontana RS. Variation of microsomal mixed function oxidase(s) and human lung cancer. Cancer 1980;45:1438–1442. 2. Carson DA, Ribeiro JM. Apoptosis and disease. Lancet 1993;341: 1251–1254. 3. Stewart BW, Kleihues P. Mechanisms of tumor development. In World Cancer Report. Lyon: IARC Press; 2003:83–126. 4. Stewart BW, Kleihues P. Human cancers by organ site. In World Cancer Report. Lyon: IAR Press; 2003:181–270. 5. Kiyohara C, Yoshimasu K, Takayama K, Nakanishi Y. Lung cancer susceptibility: are we on our way to identifying a high-risk group? Future Oncol 2007;3:617–627. 6. Berwick M, Vineis P. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J Natl Cancer Inst 2000;92:874–897. 7. Kiyohara C, Otsu A, Shirakawa T, Fukuda S, Hopkin JM. Genetic polymorphisms and lung cancer susceptibility: a review. Lung Cancer 2002;37:241–256. 8. Kiyohara C, Yoshimasu K, Shirakawa T, Hopkin JM. Genetic polymorphisms and environmental risk of lung cancer: a review. Rev Environ Health 2004;19:15–38. 9. Kiyohara C, Hirohata T, Inutsuka S. The relationship between aryl hydrocarbon hydroxylase and polymorphisms of the CYP1A1 gene. Jpn J Cancer Res 1996;87:18–24. 10. Seidegård J, De Pierre JW, Birberg W, Pilotti A, Pero RW. Characterization of soluble glutathione transferase activity in resting mononuclear

Genetic Polymorphisms and DNA Repair and Lung Cancer

leukocytes from human blood. Biochem Pharmacol 1984;33:3053–3058. 11. Pemble S, Schroeder KR, Spencer SR, et al. Human glutathione S-transferase theta (GSTT1): cDNA cloning and the characterization of a genetic polymorphism. Biochem J 1994;300 (Pt 1):271–276. 12. Yuille M, Condie A, Hudson C, et al. Relationship between glutathione S-transferase M1, T1, and P1 polymorphisms and chronic lymphocytic leukemia. Blood 2002;99:4216–4218. 13. Spitz MR, Wu X, Wang Y, et al. Modulation of nucleotide excision repair capacity by XPD polymorphisms in lung cancer patients. Cancer Res 2001;61:1354–1357. 14. Au WW, Salama SA, Sierra-Torres CH. Functional characterization of polymorphisms in DNA repair genes using cytogenetic challenge assays. Environ Health Perspect 2003;111:1843–1850. 15. Au WW, Navasumrit P, Ruchirawat M. Use of biomarkers to characterize functions of polymorphic DNA repair genotypes. Int J Hyg Environ Health 2004;207:301–313. 16. Lunn RM, Langlois RG, Hsieh LL, Thompson CL, Bell DA. XRCC1 polymorphisms: effects on aflatoxin B1-DNA adducts and glycophorin A variant frequency. Cancer Res 1999;59:2557–2561. 17. Matullo G, Palli D, Peluso M, et al. XRCC1, XRCC3, XPD gene polymorphisms, smoking and (32)P-DNA adducts in a sample of healthy subjects. Carcinogenesis 2001;22:1437–1445. 18. Hayashi SI, Watanabe J, Nakachi K, Kawajiri K. PCR detection of an A/G polymorphism within exon 7 of the CYP1A1 gene. Nucleic Acids Res 1991;19:4797. 19. Nakachi K, Imai K, Hayashi S, Watanabe J, Kawajiri K. Genetic susceptibility to squamous cell carcinoma of the lung in relation to cigarette smoking dose. Cancer Res 1991;51:5177–5180. 20. Zhong S, Howie AF, Ketterer B, et al. Glutathione S-transferase mu locus: use of genotyping and phenotyping assays to assess association with lung cancer susceptibility. Carcinogenesis 1991;12:1533–1537. 21. Chenevix-Trench G, Young J, Coggan M, Board P. Glutathione S-transferase M1 and T1 polymorphisms: susceptibility to colon cancer and age of onset. Carcinogenesis 1995;16:1655–1657. 22. Healthy Japan 21. Available at: http://www.kenkounippon21.gr.jp/ kenkounippon21/about/kakuron/index.html. Accessed October 7, 2011. 23. Holman CD, English DR, Bower C, Kurinczuk JJ. NHMRC recommendations on abstinence from alcohol in pregnancy. National Health and Medical Research Council. Med J Aust 1996;164:699. 24. Benedetti A, Parent ME, Siemiatycki J. Lifetime consumption of alcoholic beverages and risk of 13 types of cancer in men: results from a casecontrol study in Montreal. Cancer Detect Prev 2009;32:352–362. 25. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005;21:263–265. 26. Garte S, Gaspari L, Alexandrie AK, et al. Metabolic gene polymorphism frequencies in control populations. Cancer Epidemiol Biomarkers Prev 2001;10:1239–1248. 27. Tanimoto K, Hayashi S, Yoshiga K, Ichikawa T. Polymorphisms of the CYP1A1 and GSTM1 gene involved in oral squamous cell carcinoma in association with a cigarette dose. Oral Oncol 1999;35:191–196. 28. Sato M, Sato T, Izumo T, Amagasa T. Genetic polymorphism of drugmetabolizing enzymes and susceptibility to oral cancer. Carcinogenesis 1999;20:1927–1931. 29. Le Marchand L, Donlon T, Kolonel LN, Henderson BE, Wilkens LR. Estrogen metabolism-related genes and breast cancer risk: the multiethnic cohort study. Cancer Epidemiol Biomarkers Prev 2005;14:1998–2003. 30. Lee KM, Kang D, Clapper ML, et al. CYP1A1, GSTM1, and GSTT1 polymorphisms, smoking, and lung cancer risk in a pooled analysis among Asian populations. Cancer Epidemiol Biomarkers Prev 2008;17:1120–1126. 31. Hayashi S, Watanabe J, Nakachi K, Kawajiri K. Genetic linkage of lung cancer-associated MspI polymorphisms with amino acid replacement in the heme binding region of the human cytochrome P450IA1 gene. J Biochem 1991;110:407–411. 32. Sugimura H, Wakai K, Genka K, et al. Association of Ile462Val (Exon 7) polymorphism of cytochrome P450 IA1 with lung cancer in the Asian population: further evidence from a case-control study in Okinawa. Cancer Epidemiol Biomarkers Prev 1998;7:413–417. 33. Tefre T, Ryberg D, Haugen A, et al. Human CYP1A1 (cytochrome P(1)450) gene: lack of association between the Msp I restriction fragment length polymorphism and incidence of lung cancer in a Norwegian population. Pharmacogenetics 1991;1:20–25.

Copyright © 2012 by the International Association for the Study of Lung Cancer

961

Journal of Thoracic Oncology • Volume 7, Number 6, June 2012

Kiyohara et al.

34. Kiyohara C. Genetic polymorphism of enzymes involved in xenobiotic metabolism and the risk of colorectal cancer. J Epidemiol 2000;10:349–360. 35. Yamamura K, Kiyohara C, Nakanishi Y, Takayama K, Hara N. [Lung cancer risk and genetic polymorphism at the glutathione S-transferase P1 locus in male Japanese]. Fukuoka Igaku Zasshi 2000;91:203–206. 36. Benhamou S, Lee WJ, Alexandrie AK, et al. Meta- and pooled analyses of the effects of glutathione S-transferase M1 polymorphisms and smoking on lung cancer risk. Carcinogenesis 2002;23:1343–1350. 37. Raimondi S, Paracchini V, Autrup H, et al. Meta- and pooled analysis of GSTT1 and lung cancer: a HuGE-GSEC review. Am J Epidemiol 2006;164:1027–1042. 38. Cote ML, Chen W, Smith DW, et al. Meta- and pooled analysis of GSTP1 polymorphism and lung cancer: a HuGE-GSEC review. Am J Epidemiol 2009;169:802–814. 39. Smith G, Stanley LA, Sim E, Strange RC, Wolf CR. Metabolic polymorphisms and cancer susceptibility. Cancer Surv 1995;25:27–65. 40. Heagerty AH, Fitzgerald D, Smith A, et al. Glutathione S-transferase GSTM1 phenotypes and protection against cutaneous tumours. Lancet 1994;343:266–268. 41. Hayes JD Pulford DJ. The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit Rev Biochem Mol Biol 1995;30:445–600. 42. Mooney LA, Santella RM, Covey L, et al. Decline of DNA damage and other biomarkers in peripheral blood following smoking cessation. Cancer Epidemiol Biomarkers Prev 1995;4:627–634. 43. Mollerup S, Berge G, Baera R, et al. Sex differences in risk of lung cancer: Expression of genes in the PAH bioactivation pathway in relation to smoking and bulky DNA adducts. Int J Cancer 2006;119:741–744. 44. Tang D, Santella RM, Blackwood AM, et al. A molecular epidemiological case-control study of lung cancer. Cancer Epidemiol Biomarkers Prev 1995;4:341–346. 45. Li D, Wang M, Cheng L, Spitz MR, Hittelman WN, Wei Q. In vitro induction of benzo(a)pyrene diol epoxide-DNA adducts in peripheral

962

lymphocytes as a susceptibility marker for human lung cancer. Cancer Res 1996;56:3638–3641. 46. Hoeijmakers JH. Genome maintenance mechanisms for preventing cancer. Nature 2001;411:366–374. 47. Wood RD, Mitchell M, Lindahl T. Human DNA repair genes, 2005. Mutat Res 2005;577:275–283. 48. Benhamou S, Sarasin A. ERCC2 /XPD gene polymorphisms and lung cancer: a HuGE review. Am J Epidemiol 2005;161:1–14. 49. Kiyohara C, Yoshimasu K. Genetic polymorphisms in the nucleotide excision repair pathway and lung cancer risk: a meta-analysis. Int J Med Sci 2007;4:59–71. 50. Hung RJ, Hall J, Brennan P, Boffetta P. Genetic polymorphisms in the base excision repair pathway and cancer risk: a HuGE review. Am J Epidemiol 2005;162:925–942. 51. Kiyohara C, Takayama K, Nakanishi Y. Association of genetic polymorphisms in the base excision repair pathway with lung cancer risk: a meta-analysis. Lung Cancer 2006;54:267–283. 52. NCBI dbSNP database. Available at: http://www.ncbi.nlm.nih.gov/ projects/SNP/snp_ref.cgi?rs=861539. Accessed October 7, 2011. 53. Nakachi K, Imai K, Hayashi S, Kawajiri K. Polymorphisms of the CYP1A1 and glutathione S-transferase genes associated with susceptibility to lung cancer in relation to cigarette dose in a Japanese population. Cancer Res 1993;53:2994–2999. 54. Vineis P, Veglia F, Anttila S, et al. CYP1A1, GSTM1 and GSTT1 polymorphisms and lung cancer: a pooled analysis of gene-gene interactions. Biomarkers 2004;9:298–305. 55. Belogubova EV, Ulibina YM, Suvorova IK, et al. Combined CYP1A1/ GSTM1 at-risk genotypes are overrepresented in squamous cell lung carcinoma patients but underrepresented in elderly tumor-free subjects. J Cancer Res Clin Oncol 2006;132:327–331. 56. dos santos Silva I. Case-control studies. In Cancer Epidemiology: Principles and Methods. Lyon: IARC Press, 1999:189–212. 57. Morton LM, Cahill J, Hartge P. Reporting participation in epidemiologic studies: a survey of practice. Am J Epidemiol 2006;163:197–203.

Copyright © 2012 by the International Association for the Study of Lung Cancer