Abstract
Adenocarcinoma (ADC) is the most frequent histological type of lung cancer and comprises the majority of lung cancers in non‐smokers. Thus, genetic factors responsible for ADC susceptibility need to be determined to establish efficient ways of preventing the disease. The OGG1 gene, encoding a glycosylase for 8‐hydroxyguanine, an oxidatively damaged promutagenic base, has the polymorphism Ser326Cys, and OGG1‐326Cys protein was indicated to have a lower ability to prevent mutagenesis than the OGG1‐326Ser protein. Case‐control studies to date suggest that the OGG1‐326Cys allele is associated with a higher risk for several types of cancers, including overall lung cancer. However, the contribution of this polymorphism to lung ADC risk is unclear. In the present study, the OGG1‐Ser326Cys polymorphism was assessed for association with lung ADC risk using a case‐control study of a Japanese population consisting of 1097 cases and 394 controls. Odds ratios (OR) of the 326Cys allele carriers increased in a dose‐dependent manner with allele number (P for the trend test = 0.04). The OR of homozygotes for the 326Cys allele was increased significantly when homozygotes for the 326Ser allele were used as a reference (OR = 1.5, 95% confidence interval [CI] = 1.0–2.1, P = 0.04). Furthermore, the overall OR for lung ADC of the Cys/Cys homozygotes out of a total of 1925 ADC patients and 3449 controls from six case‐control studies reported up to the present were 1.43 (95% CI = 1.11–1.84, P = 0.0045). These results indicate that OGG1‐326Cys functions as a risk allele for lung ADC development. (Cancer Sci 2006; 97: 724–728)
Abbreviations:
- 8OHG
8‐hydroxyguanine
- ADC
adenocarcinoma
- AP
apurinic/apyrimidinic
- CI
confidence interval
- EGFR
epidermal growth factor receptor
- NCCH
National Cancer Center Hospital
- NNGH
National Nishi‐gunma Hospital
- SCC
small cell carcinoma
- SNP
single nucleotide polymorphism
- SQC
squamous cell carcinoma
- OR
odds ratio
Lung cancer is the leading cause of cancer‐related death in the world.( 1 ) ADC, SQC and SCC are three major histological types of lung cancer. It is known that the development of SQC and SCC is strongly associated with smoking, whereas that of ADC is less associated compared with the other two types, indicating that carcinogenic processes are different between ADC and SQC and SCC.( 1 ) Recently, ADC has become the most frequent histological type of lung cancer, and it comprises the majority of lung cancers in non‐smokers.( 2 ) Thus, genetic factors, as well as environmental factors, responsible for the susceptibility of ADC in each individual need to be determined to establish efficient ways of preventing the disease. The OGG1 gene encodes a protein with DNA glycosylase and AP lyase activities that remove 8OHG, an oxidatively damaged promutagenic base, from double‐stranded DNA.( 3 , 4 ) We previously found a SNP at codon 326 associated with the amino acid change Ser326Cys.( 5 ) The lower ability of OGG1‐326Cys protein than OGG1‐326Ser protein to prevent mutagenesis by 8OHG has been shown by both in vitro ( 5 , 6 ) and in vivo ( 7 ) studies. It was also demonstrated that homozygotes for the OGG1‐326Cys allele have a lower capacity to repair 8OHG than others.( 8 ) Frequencies of the OGG1‐326Cys allele are different among populations, and the frequency is higher in Asians (i.e. >40%) than in Caucasians (i.e. <20%).( 9 ) As reviewed by Weiss et al.( 10 ) several case‐control (association) studies have been undertaken on several populations, including Caucasians and Asians, and the OGG1‐326Cys allele encoding the less active OGG1 protein was reported to be associated with the risk of a variety of human cancers such as esophageal, prostate, orolaryngeal and nasopharyngeal cancers. Case‐control studies have also been undertaken on the association of this SNP with lung cancer risk.( 9 , 11 , 12 , 13 , 14 , 15 , 16 ) A recent meta‐analysis showed that the overall OR of the homozygotes for the OGG1‐326Cys allele against those for the OGG1‐326Ser allele was 1.24 (95% CI = 1.01–1.53),( 17 ) suggesting that the SNP plays a role in susceptibility to overall lung cancer. However, association of the SNP with risk of each histological type of lung cancer, such as ADC, is still unclear. In our previous study of a Japanese population,( 9 ) the OGG1‐326Cys allele was associated with an increased risk for lung ADC, although the association was not statistically significant. In a study of another Japanese population, such an association was not observed.( 12 ) In four other studies of Caucasians and Hawaiians, the OGG1‐326Cys allele was associated with increased risks for lung ADC, and in three of them, the associations were statistically significant.( 11 , 13 , 14 , 15 ) Thus, in the present study, the association of the OGG1‐Ser326Cys SNP with lung ADC risk was further investigated by conducting a case‐control study of a Japanese population consisting of 1097 lung ADC cases and 394 controls, which were enrolled in two hospitals. We also calculated the overall OR for lung ADC of the homozygotes for the OGG1‐326Cys allele in 1925 ADC patients and 3449 controls analyzed in six case‐control studies published up to the present, including the present study.
Materials and Methods
Case‐control study
All of the 1097 cases and 394 controls were Japanese. The first sample consisted of 1031 ADC cases and 253 controls, which were enrolled in the NCCH from 1999 to 2004, and the second sample consisted of 66 ADC cases and 141 controls, which were enrolled in the National Nishi‐gunma Hospital (NNGH) from 1999 to 2002. All ADC cases, from whom informed consent and blood samples were obtained, were included consecutively in this study without any particular exclusion criteria. All of the cases were diagnosed as ADC by cytological or histological examinations according to the WHO classification.( 18 ) Controls were patients of NCCH or NNGH, and they were selected with a criterion of no history of cancer. This study was approved by the institutional review boards of NCCH and NNGH. Smoking history was obtained via interview using a questionnaire. Smoking habit was expressed in pack‐years, which was defined as the number of cigarette packs smoked daily multiplied by years of smoking, both in current smokers and former smokers. Smokers were defined as those who had smoked regularly for 12 months or longer at any time in their life, whereas non‐smokers were defined as those who had not. There were no individuals who had smoked regularly for less than 12 months. From each individual, a 10 or 20 mL whole‐blood sample was obtained. Genomic DNA was isolated from the samples, and 10 ng DNA was subjected to genotyping for the OGG1‐Ser326Cys SNP by pyrosequencing using the PSQ96 system (Pyrosequencing AB, Uppsala, Sweden) as described previously.( 19 ) A recent analysis of Japanese individuals chosen randomly from five areas in Japan showed that the fraction of homozygotes for the OGG1‐326Cys allele comprised approximately 20% of Japanese people.( 20 ) Therefore, when we hypothesize an OR of homozygotes for the OGG1‐326Cys allele as being 1.5, the statistical power to detect association in the present case‐control study was 0.95.
Statistical analysis
Differences in the allele and genotype distributions between the cases and controls were evaluated using the χ2‐test. Hardy–Weinberg tests were carried out using the TFPGA software (http://bioweb.usu.edu/mpmbio/). The strength of association of OGG1 genotypes with ADC risk was measured as crude OR and OR adjusted for sex, age (49 years, 50–59 years, 60–69 years, 70 years) and smoking dosage (pack‐years: 0, 1–49, 50) with 95% CI by unconditional logistic regression analysis.( 21 ) Homogeneity test for the OR of the Cys/Cys homozygotes was undertaken according to a method described previously( 21 ) to examine whether or not OR were common among five previous studies and the present study, regardless of their genotype frequencies based on adjusted OR with 95% CI in the reports.( 9 , 11 , 12 , 14 , 15 ) As homogeneity of the OR was not rejected, the overall OR of the Cys/Cys homozygotes in the subjects of these six studies was calculated based on adjusted OR with 95% CI in the reports( 9 , 11 , 12 , 14 , 15 ) by the inverse‐variance method.( 21 ) All of the statistical analyses were carried out using SAS version 9 software (SAS Institute, Cary, NC, USA).
Results and Discussion
We conducted a case‐control study consisting of 1097 cases and 394 controls (Table 1). The cases and controls were enrolled in two hospitals, NCCH and NNGH; 1031 cases and 253 controls from NCCH plus 66 cases and 141 controls from NNGH. In the NCCH sample, smokers were more frequent in the cases than in the controls. In the NNGH sample, the frequency of smokers was similar between the cases and controls, due to the fact that the controls of NNGH included patients with smoking‐associated lung diseases, such as chronic obstructive pulmonary disease, as reported previously.( 22 )
Table 1.
Lung adenocarcinoma cases and controls used for the present case‐control study
| Variable | NCCH | NNGH | Both NCCH and NNGH | |||
|---|---|---|---|---|---|---|
| Controls | Cases | Controls | Cases | Controls | Cases | |
| Total | 253 | 1031 | 141 | 66 | 394 | 1097 |
| Age (mean ± SD; years) | 64 ± 11 | 60 ± 10 | 65 ± 14 | 67 ± 10 | 64 ± 12 | 60 ± 10 |
| Sex † | ||||||
| Male | 144 (57) | 616 (60) | 100 (71) | 41 (62) | 244 (62) | 657 (60) |
| Female | 109 (43) | 415 (40) | 41 (29) | 25 (38) | 150 (38) | 440 (40) |
| Smoking habit † | ||||||
| Non‐smoker | 134 (53) | 431 (42) | 46 (33) | 24 (36) | 180 (46) | 455 (42) |
| Smoker | 117 (46) | 598 (58) | 91 (64) | 42 (64) | 208 (53) | 640 (58) |
| Unknown | 2 (1) | 2 (0) | 4 (3) | 0 (0) | 6 (2) | 2 (0) |
| Smoking (mean ± SD; pack‐years) | 17 ± 28 | 24 ± 29 | 30 ± 36 | 28 ± 27 | 22 ± 32 | 25 ± 29 |
Numbers in parentheses refer to percentages. NCCH, National Cancer Center Hospital; NNGH, National Nishi‐gunma Hospital.
Genotypes for the OGG1‐Ser326Cys SNP in all of the cases and controls were determined using the pyrosequencing method. Distribution of the 326Ser and 326Cys alleles was in Hardy–Weinberg equilibrium both in the cases and controls (P < 0.05). The relative risks of the genotypes were calculated as crude and adjusted OR (Table 2). Heterozygotes and homozygotes for the 326Cys allele showed increased crude and adjusted OR in both hospital samples, when homozygotes for the 326Ser allele were used as a reference. When the subjects from the two hospitals were combined, increases in OR of homozygotes for the 326Cys allele were statistically significant. The P‐value for the trend test was also statistically significant. Thus, it was suggested that OGG1‐326Cys functions as a risk allele for ADC development in a dose‐dependent manner, and homozygotes for the 326Cys allele, which are predicted to have the lowest mutation suppression ability, have a significantly increased risk for lung ADC.
Table 2.
Association of the OGG1 genotype with lung adenocarcinoma risk
| Hospital | Genotype | No. controls (%) | No. cases (%) | Crude OR (95% CI, P) | Adjusted OR † (95% CI, P) | P for trend test † |
|---|---|---|---|---|---|---|
| NCCH | Ser/Ser | 79 (31) | 267 (26) | Reference | Reference | 0.09 |
| Ser/Cys | 126 (50) | 514 (50) | 1.2 (0.9–1.7, 0.2) | 1.2 (0.9–1.7, 0.3) | ||
| Cys/Cys | 48 (19) | 250 (24) | 1.5 (1.0–2.3, 0.03) | 1.4 (0.9–2.2, 0.09) | ||
| NNGH | Ser/Ser | 44 (31) | 18 (27) | Reference | Reference | 0.6 |
| Ser/Cys | 64 (45) | 30 (45) | 1.2 (0.6–2.3, 0.7) | 1.2 (0.6–2.4, 0.7) | ||
| Cys/Cys | 33 (23) | 18 (27) | 1.3 (0.6–3.0, 0.5) | 1.3 (0.6–2.9, 0.6) | ||
| Both | Ser/Ser | 123 (31) | 285 (26) | Reference | Reference | 0.04 |
| Ser/Cys | 190 (48) | 544 (50) | 1.2 (0.9–1.6, 0.1) | 1.2 (0.9–1.6, 0.3) | ||
| Cys/Cys | 81 (21) | 268 (24) | 1.4 (1.0–2.0, 0.03) | 1.5 (1.0–2.1, 0.04) |
Adjusted for sex, age and smoking dosage (National Cancer Center Hospital [NCCH] and National Nishi‐gunma Hospital [NNGH]), or for sex, age, smoking dosage and hospital (both). CI, confidence interval; OR, odds ratio.
Tobacco smoke is known to cause oxidative damage to genomic DNA.( 23 ) Therefore, it can be assumed that individuals with reduced OGG1 activity represent a group that is more prone to acquiring gene mutations and ADC by smoking. Thus, we assessed the effect of smoking on the contribution of the OGG1‐Ser326Cys SNP to ADC risk. Namely, OR in non‐smokers and smokers were calculated separately as shown in Table 3. Increases in OR of OGG1‐326Cys allele carriers were similar between non‐smokers and smokers with a P‐value for modification by smoking of 0.27. Thus, the effect of the OGG1‐Ser326Cys SNP on lung ADC risk was not modified by smoking in this study population. The present results suggest that 8‐OHG generated by factors other than tobacco smoke induce mutations in epithelial cells of the peripheral lung where lung ADC develops, and that the capacity of repairing 8‐OHG by OGG1 protein is responsible for lung ADC susceptibility of each individual.
Table 3.
Effect of OGG1 genotypes on lung adenocarcinoma risk according to smoking history
| Genotype | Smoking habit | No. controls (%) | No. cases (%) | Crude OR (95% CI, P) | Adjusted OR † , ‡ (95% CI, P) |
|---|---|---|---|---|---|
| Ser/Ser | Non‐smoker | 55 (31) | 105 (23) | Reference | Reference |
| Smoker | 65 (31) | 178 (28) | Reference | Reference | |
| Ser/Cys | Non‐smoker | 87 (48) | 245 (54) | 1.5 (1.0–2.2, 0.06) | 1.3 (0.8–2.0, 0.2) |
| Smoker | 101 (49) | 299 (47) | 1.1 (0.8–1.6, 0.7) | 1.1 (0.7–1.7, 0.7) | |
| Cys/Cys | Non‐smoker | 38 (21) | 105 (23) | 1.5 (0.9–2.4, 0.1) | 1.3 (0.8–2.3, 0.3) |
| Smoker | 42 (20) | 163 (26) | 1.4 (0.9–2.2, 0.1) | 1.6 (1.0–2.6, 0.08) |
Adjusted for gender, age, smoking dosage, and hospital;
P for interaction of smoking and OGG1 genotypes (trend) = 0.27. CI, confidence interval; OR, odds ratio.
In the present case‐control study, homozygotes for the OGG1‐326Cys allele were associated significantly with lung ADC risk. However, previous case‐control studies on other populations did not necessarily show such associations.( 9 , 11 , 12 , 13 , 14 , 15 ) The frequency of the OGG1‐Cys326 allele is largely different among populations with different ethnicities.( 9 , 11 , 12 , 13 , 14 , 15 ) The allele frequencies in controls for the two previous case‐control studies of Japanese subjects and for the present study were different;( 9 , 12 ) therefore, the frequency might also be different among Japanese populations. Otherwise, such a difference might be a deviation due to small numbers of subjects.( 9 , 12 ) Thus, we calculated the overall risk of the Cys/Cys homozygotes for lung ADC by combining subjects of the present study with those of previous case‐control studies. Five studies were chosen based on the criteria that adjusted OR of Cys/Cys homozygotes against Ser/Ser homozygotes, as well as the number of subjects for each genotype, were available in the literature( 9 , 11 , 12 , 14 , 15 ) (Table 4). Homogeneity test for the OR of the Cys/Cys homozygotes was undertaken to examine whether or not the OR are common among five previous studies and the present study, regardless of the frequency of the Cys/Cys homozygotes. Homogeneity of OR in these six studies was not rejected (P = 0.29), therefore homogeneity of OR in these studies was supported. Thus, the overall OR for the six studies was calculated as 1.43 (95% CI = 1.11–1.84, P = 0.0045), indicating that homozygotes for the OGG1‐326Cys allele were significantly associated with lung ADC risk. The overall OR for the three studies on Japanese subjects was calculated as 1.27 (95% CI = 0.94–1.72, P = 0.12). Although the OR was not statistically significant, an increased risk in homozygotes of the OGG1‐326Cys allele was also suggested in the Japanese population.
Table 4.
Odds ratios (OR) of individuals with the OGG1‐326 Cys/Cys genotype against those with the OGG1‐326 Ser/Ser genotype for lung adenocarcinoma
| Ethnity of subjects | Country | No. subjects | Frequency of the Cys/Cys genotype in controls | Adjusted OR | 95% CI | Reference | |
|---|---|---|---|---|---|---|---|
| Cases | Controls | ||||||
| Asian | Japan | 78 | 197 | 0.14 | 1.34 | 0.53–3.39 | 9 |
| Asian | Japan | 138 | 240 | 0.23 | 0.81 | 0.44–1.52 | 12 |
| Asian | Japan | 1097 | 394 | 0.21 | 1.47 | 1.02–2.13 | This study |
| Caucasian | Germany | 50 | 105 | 0.019 | 1.84 | 0.41–14.4 | 11 |
| Caucasian | USA | 63 | 350 | 0.023 | 4.20 | 1.10–14.8 | 14 |
| Caucasian | Eastern Europe | 499 | 2163 | 0.036 | 1.66 | 1.04–2.66 | 15 |
| Overall | 1925 | 3449 | 1.43 | 1.11–1.84* | |||
| Overall | Japan | 1313 | 831 | 1.27 | 0.94–1.72** | ||
P = 0.0045;
P = 0.12. CI, confidence interval.
Considering the OR for lung ADC and for overall lung cancer (i.e. 1.24),( 17 ) the contribution of the OGG1‐Ser326Cys SNP to lung cancer, including ADC, seems small. Considering the diversity of the OGG1‐Cys326 allele frequency among populations as described above, the associations of the OGG1‐Ser326Cys SNP would be masked or over‐represented by deviations among the populations. Thus, further case‐control studies on subjects that are chosen carefully to represent cases and controls in the same population, such as a nested case‐control study designated in a large‐scale cohort study, will be effective to validate the association of the OGG1‐Ser326Cys SNP with lung ADC risk. In addition, recent studies have revealed that genetic and epigenetic alterations accumulating in cancer cells are different among histological subtypes of lung ADC, as represented by more frequent EGFR gene mutations in bronchioloalveolar carcinomas than in others, suggesting the differences in pathogenic pathways among ADC subtypes.( 24 , 25 , 26 ) Therefore, it is possible that the OGG1‐Ser326Cys SNP is preferentially associated with the risk of an ADC subtype. Thus, comparative association studies among ADC subtypes should be an important issue and are being investigated in our laboratory.
Acknowledgments
This work was supported in part by Grants‐in‐Aid from the Ministry of Health, Labor and Welfare for research on human genome and tissue engineering and for cancer research (16S‐1). We thank Ms Rumie Sasaki‐Matsumura and Sachiyo Mimaki for technical assistance. The continuous encouragement by Dr Mitsumasa Okada and Dr Nobue Ito of the Graduate School of Science, Toho University is gratefully acknowledged. N. Y. was an awardee of a Research Resident Fellowship from the Foundation for Promotion of Cancer Research in Japan during the study.
References
- 1. Parkin M, Je T, Boffetta P, Samet J, Shields PG, Caporaso NE. Lung cancer epidemiology and etiology. In: Harris CC, ed. World Health Organization Classification of Tumors: Pathology and Genetics, Tumours of Lung, Pleura, Thymus and Heart. Lyon: IARC Press, 2004: 31–4. [Google Scholar]
- 2. Colvy T, Noguch IM, Henschk EC et al. Adenocarcinoma. In: Harris CC, ed. World Health Organization Classification of Tumors: Pathology and Genetics, Tumours of Lung, Pleura, Thymus and Heart. Lyon: IARC Press, 2004: 35–44. [Google Scholar]
- 3. Boiteux S, Radicella JP. The human OGG1 gene: structure, functions, and its implication in the process of carcinogenesis. Arch Biochem Biophys 2000; 377: 1–8. [DOI] [PubMed] [Google Scholar]
- 4. Shinmura K, Yokota J. The OGG1 gene encodes a repair enzyme for oxidatively damaged DNA and is involved in human carcinogenesis. Antioxid Redox Signal 2001; 3: 597–609. [DOI] [PubMed] [Google Scholar]
- 5. Kohno T, Shinmura K, Tosaka M et al. Genetic polymorphisms and alternative splicing of the hOGG1 gene, that is involved in the repair of 8‐hydroxyguanine in damaged DNA. Oncogene 1998; 16: 3219–25. [DOI] [PubMed] [Google Scholar]
- 6. Dherin C, Radicella JP, Dizdaroglu M, Boiteux S. Excision of oxidatively damaged DNA bases by the human alpha‐hOgg1 protein and the polymorphic alpha‐hOgg1 (Ser326Cys) protein which is frequently found in human populations. Nucl Acids Res 1999; 27: 4001–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yamane A, Kohno T, Ito K et al. Differential ability of polymorphic OGG1 proteins to suppress mutagenesis induced by 8‐hydroxyguanine in human cells in vivo . Carcinogenesis 2004; 25: 1689–94. [DOI] [PubMed] [Google Scholar]
- 8. Lee AJ, Hodges NJ, Chipman JK. Interindividual variability in response to sodium dichromate‐induced oxidative DNA damage: role of the Ser326Cys polymorphism in the DNA‐repair protein of 8‐oxo‐7,8‐dihydro‐2′‐deoxyguanosine DNA glycosylase 1. Cancer Epidemiol Biomarkers Prev 2005; 14: 497–505. [DOI] [PubMed] [Google Scholar]
- 9. Sugimura H, Kohno T, Wakai K et al. hOGG1 Ser326Cys polymorphism and lung cancer susceptibility. Cancer Epidemiol Biomarkers Prev 1999; 8: 669–74. [PubMed] [Google Scholar]
- 10. Weiss JM, Goode EL, Ladiges WC, Ulrich CM. Polymorphic variation in hOGG1 and risk of cancer: a review of the functional and epidemiologic literature. Mol Carcinog 2005; 42: 127–41. [DOI] [PubMed] [Google Scholar]
- 11. Wikman H, Risch A, Klimek F et al. hOGG1 polymorphism and loss of heterozygosity (LOH): significance for lung cancer susceptibility in a caucasian population. Int J Cancer 2000; 88: 932–7. [DOI] [PubMed] [Google Scholar]
- 12. Ito H, Hamajima N, Takezaki T et al. A limited association of OGG1 Ser326Cys polymorphism for adenocarcinoma of the lung. J Epidemiol 2002; 12: 258–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Le Marchand L, Donlon T, Lum‐Jones A, Seifried A, Wilkens LR. Association of the hOGG1 Ser326Cys polymorphism with lung cancer risk. Cancer Epidemiol Biomarkers Prev 2002; 11: 409–12. [PubMed] [Google Scholar]
- 14. Park J, Chen L, Tockman MS, Elahi A, Lazarus P. The human 8‐oxoguanine DNA N‐glycosylase 1 (hOGG1) DNA repair enzyme and its association with lung cancer risk. Pharmacogenetics 2004; 14: 103–9. [DOI] [PubMed] [Google Scholar]
- 15. Hung RJ, Brennan P, Canzian F et al. Large‐scale investigation of base excision repair genetic polymorphisms and lung cancer risk in a multicenter study. J Natl Cancer Inst 2005; 97: 567–76. [DOI] [PubMed] [Google Scholar]
- 16. Zienolddiny S, Campa D, Lind H et al. Polymorphisms of DNA repair genes and risk of non‐small cell lung cancer. Carcinogenesis 2005. [DOI] [PubMed]
- 17. 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: 925–42. [DOI] [PubMed] [Google Scholar]
- 18. Brambilla E, Travis WD, Colby TV, Corrin B, Shimosato Y. The new World Health Organization classification of lung tumours. Eur Respir J 2001; 18: 1059–68. [DOI] [PubMed] [Google Scholar]
- 19. Sakiyama T, Kohno T, Mimaki S et al. Association of amino acid substitution polymorphisms in DNA repair genes TP53, POLI, REV1 and LIG4 with lung cancer risk. Int J Cancer 2005; 114: 730–7. [DOI] [PubMed] [Google Scholar]
- 20. Yoshimura K, Hanaoka T, Ohnami S et al. Allele frequencies of single nucleotide polymorphisms (SNPs) in 40 candidate genes for gene–environment studies on cancer: data from population‐based Japanese random samples. J Hum Genet 2003; 48: 654–8. [DOI] [PubMed] [Google Scholar]
- 21. Breslow NE, Day NE. Statistical methods in cancer research, volume I – the analysis of case‐control studies. IARC Sci Publ 1980: 5–338. [PubMed]
- 22. Sunaga N, Kohno T, Yanagitani N et al. Contribution of the NQO1 and GSTT1 polymorphisms to lung adenocarcinoma susceptibility. Cancer Epidemiol Biomarkers Prev 2002; 11: 730–8. [PubMed] [Google Scholar]
- 23. Wynder EL, Hoffmann D. Smoking and lung cancer: scientific challenges and opportunities. Cancer Res 1994; 54: 5284–95. [PubMed] [Google Scholar]
- 24. Shigematsu H, Gazdar AF. Somatic mutations of epidermal growth factor receptor signaling pathway in lung cancers. Int J Cancer 2006; 118: 257–62. [DOI] [PubMed] [Google Scholar]
- 25. Matsumoto S, Iwakawa R, Kohno T et al. Frequent EGFR mutations in noninvasive bronchioloalveolar carcinoma. Int J Cancer 2006; 118: 2498–504. [DOI] [PubMed] [Google Scholar]
- 26. Toyooka S, Tokumo M, Shigematsu H et al. Mutational and epigenetic evidence for independent pathways for lung adenocarcinomas arising in smokers and never smokers. Cancer Res 2006; 66: 1371–5. [DOI] [PubMed] [Google Scholar]