ORIGINAL ARTICLE DNA Repair Gene ERCC1 and ERCC2/XPD Polymorphisms and Risk of Squamous Cell Carcinoma of the Head and Neck Erich M. Sturgis, MD; Kristina R. Dahlstrom, BS; Margaret R. Spitz, MD, MPH; Qingyi Wei, MD, PhD Objective: To determine the effect of the ERCC1 C8092A polymorphism and the ERCC2/XPD G23591A polymorphism on the risk of squamous cell carcinoma of the head and neck (SCCHN). Design: A hospital-based case-control study. Subjects: A total of 330 newly diagnosed case subjects with SCCHN and 330 cancer-free control subjects matched on age (±5 years), sex, smoking status, and alcohol use. All subjects were non-hispanic whites. Methods: After informed consent was obtained, blood was drawn for genotyping. The ERCC1 C8092A polymorphism was typed by single-strand conformational polymorphism analysis. The ERCC2/XPD G23591A polymorphism was typed by polymerase chain reaction based restriction fragment length polymorphism analysis with the enzyme StyI. The 2 analysis was used to assess differences in genotype and allele frequencies. Multivariate logistic regression analysis was performed to estimate the risk of SCCHN for individuals having these genotypes after adjustment for age, sex, tobacco smoking, and alcohol use. Results: The DNA was available and genotyping was ultimately successful for 313 case subjects and 313 control subjects. The ERCC1 8092CC genotype and the ERCC2/XPD 23591A allele were associated with nonsignificantly increased risks of SCCHN: odds ratios, 1.15 (95% confidence interval [CI], 0.84-1.59) and 1.28 (95% CI, 0.93-1.76), respectively, whereas having both risk genotypes was associated with an even higher risk of SCCHN: odds ratio, 1.78 (95% CI, 0.99-3.17). When considering both polymorphisms, we found a significant allele dose effect (P=.04). Conclusions: These 2 polymorphisms may contribute to the risk of SCCHN, but larger studies are needed to confirm their role in SCCHN. Combining common DNA repair gene polymorphisms into models of genetic risk of SCCHN may improve risk estimates. Arch Otolaryngol Head Neck Surg. 2002;128:1084-1088 From the Departments of Head and Neck Surgery (Dr Sturgis) and Epidemiology (Drs Sturgis, Spitz, and Wei and Ms Dahlstrom), The University of Texas M. D. Anderson Cancer Center, Houston. ANNUAL ESTIMATES show that the number of current smokers in the United States has remained relatively stable at approximately 45 million over the past 3 decades, and the absolute incidence of newly diagnosed cancers has never exceeded 1.4 million. 1,2 These facts suggest that most smokers never develop cancer and that genetic differences probably influence individuals responses to environmental carcinogens and consequently their risk for cancer. Although the concept of genetic susceptibility may seem intuitive, it is extremely complex and involves multiple cellular systems regulated by hundreds of genes. It is unlikely that common genetic variants, most of which are single nucleotide polymorphisms (SNPs), will significantly influence risk without environmental exposures. Therefore, current efforts have focused on determining genotype frequencies in the population and which SNPs will be important to include in future, more complex, genetic models of cancer risk assessment. We have previously reported that the risk of squamous cell carcinoma of the head and neck (SCCHN) was associated with poor DNA repair phenotype in response to the classic tobacco carcinogen, benzo[a]pyrene diol epoxide. 3-6 The damage induced by it is repaired primarily by the nucleotide excision repair pathway. 7 In 1998, Shen et al 8 identified 31 SNPs of ERCC1, ERCC2/XPD, and ERCC4/XPF of the nucleotide excision repair pathway. Be- 1084
cause polymorphisms of the nucleotide excision repair genes may be associated with differences in DNA repair capacity, 9 we hypothesized that genetic variants in the nucleotide excision repair pathway may influence risk of SCCHN. In the present study, we examine an SNP of the 3 untranslated region (UTR) of ERCC1 that has been reported to be linked to adult-onset glioma, 10 and an SNP of exon 10 of ERCC2/XPD that has been linked to altered DNA repair capacity. 9 SUBJECTS, MATERIALS, AND METHODS STUDY SUBJECTS Between May 1995 and March 2001, patients with incident (newly diagnosed) SCCHN were recruited from patients seen in the Head and Neck Center at our institution to participate in an ongoing molecular epidemiologic study. After providing informed consent, all participating patients agreed to donate 30 ml of blood for biomarker testing and to complete a detailed questionnaire eliciting demographic, exposure, and family history information. Cancer-free control subjects were selected from a pool of healthy controls, identified from enrollees in a local managed-care organization, to participate in ongoing hospital-based case-control studies. The control subjects were matched to the case subjects on age (±5 years), sex, smoking status, and alcohol use. To eliminate the possibility of racial confounders, only non-hispanic whites were included in this study. Smokers were defined as those who had smoked more than 100 cigarettes in their lifetimes. Drinkers were defined as those who had drunk alcoholic beverages at least once a week for more than 1 year. GENOTYPING We used leukocyte cell pellets obtained from the buffy coat by centrifugation of 1 ml of whole blood for DNA extraction and performed polymerase chain reaction (PCR) amplification as previously described. 9,10 We used PCR assays to amplify the 3 UTR of ERCC1 and exon 10 of ERCC2/XPD, which contain the polymorphisms of interest. The primers used were 5 - TGAGCCAATTCAGCCACT-3 and 5 -TAGTTCCTCAGTTTC- CCG-3, which generate a 255 base pair (bp) fragment for the 3 UTR of ERCC1, and 5 -CTGTTGGTGGGTGCCCGTATC- TGTTGGTCT-3 and 5 -TAATATCGGGGCTCACCCTGCAG- CACTTCCT-3, which generate a 751-bp fragment for exon 10 of ERCC2/XPD. As previously described, 10 we used singlestrand conformational polymorphism assay to type the ERCC1 3 UTR polymorphism. The restriction enzyme StyI(NewEngland Biolabs, Beverly, Mass) was used to distinguish the 23591 polymorphism of exon 10 of ERCC2/XPD in which the gain of a StyI restriction site occurs in the polymorphic allele. The wildtype allele has a single StyI restriction site resulting in 2 bands (507 and 244 bp), and the polymorphic allele has 2 StyI restriction sites and, therefore, has 3 bands (474, 244, and 33 bp, not visible). The PCR product was digested with 10 U of StyI, in the 10 buffer supplied with the restriction enzyme and 2% bovine serum albumin at 37 C for 16 hours. The digestion products were separated on a 2% NuSieve 3:1 agarose gel (FMC Bio- Products, Rockland, Me) and photographed with Polaroid film (Cambridge, Mass). STATISTICAL ANALYSIS Table 1. Frequency Distribution Analysis of Demographic and Risk Factors* Variable Cases We first performed univariate analysis to calculate the frequency of each allele and genotype. By tabulation, we also examined the concordance between the genotype frequencies of 2 polymorphisms. We compared the observed genotype frequencies with those calculated from the Hardy-Weinberg equation (p2+2pq+q 2 =1, where p is the frequency of the wildtype allele and q is 1 p). We calculated the odds ratios (ORs) and their 95% confidence intervals (CIs) for the genotypes by logistic regression analysis with adjustment for age, sex, smoking status, and alcohol use. All of the statistical analyses were performed with Statistical Analysis System software (Version 6; SAS Institute Inc, Cary, NC). RESULTS Controls P Value Age, y 50 99 (31.6) 98 (31.3) 51-63 110 (35.2) 112 (35.8).99 63 104 (33.2) 103 (32.9) Sex Male 234 (74.8) 229 (73.2) Female 79 (25.2) 84 (26.8).65 Smoking status Smoker 235 (75.1) 230 (73.5) Nonsmoker 78 (24.9) 83 (26.5).65 Alcohol status Drinker 249 (79.5) 247 (78.9) Nondrinker 64 (20.5) 66 (21.1).84 Site Oral cavity 109 (34.8)...... Pharynx 145 (46.3)...... Larynx 59 (18.9)...... Two-sided 2 test. Includes 126 patients with oropharyngeal cancer and 19 patients with hypopharyngeal cancer. Initially, we identified 330 patients with SCCHN and 330 cancer-free control subjects. Of these, DNA was unavailable or PCR was unsuccessful in 17 patients and 17 control participants. Consequently, the ultimate sample size was 313 case subjects and 313 control subjects. The subjects were well matched on age, sex, smoking status, and alcohol use (Table 1). All subjects were non-hispanic whites. All cases were incident squamous cell carcinomas of the oral cavity, oropharynx, hypopharynx, or larynx (Table 1). Because only 19 patients with hypopharyngeal cancer were recruited, they were grouped with oropharyngeal patients for subgroup analysis. Genotype distributions are summarized in Table 2. The ERCC1 8092 genotype distribution for case and control subjects were in Hardy-Weinberg equilibrium (P=.88 and P=.54, respectively). The variant ERCC1 8092A allele was less frequent in the case subjects (0.230) than in the control subjects (0.248), and the homozygous wildtype ERCC1 8092CC genotype was more common in the case group (58.5%) than in the control group (55.0%), suggesting the ERCC1 A allele has a protective effect. The ERCC2/XPD 23591 genotype distribution in case and control subjects were in Hardy-Weinberg equilibrium (P=.09 1085
Table 2. ERCC1 8092 and ERCC2/XPD 23591 Genotype/Allele Frequencies and Concordance* ERCC2/XPD G23591A Genotype Cases Controls ERCC1 C8092A Genotype GG GA AA All GG GA AA All CC 95 (30.4) 80 (25.6) 8 (2.6) 183 (58.5) 103 (32.9) 57 (18.2) 12 (3.8)* 172 (55.0) CA 26 (8.3) 78 (24.9) 12 (3.8) 116 (37.0) 38 (12.1) 69 (22.0) 20 (6.4) 127 (40.6) AA 2 (0.6) 7 (2.2) 5 (1.6) 14 (4.5) 1 (0.3) 9 (2.9) 4 (1.3) 14 (4.5) All 123 (39.3) 165 (52.7) 25 (8.0) 313 (100) 142 (45.4) 135 (43.1) 36 (11.5) 313 (100) *All data are number (percentage) of subjects. ERCC2/XPD 23591G allele frequency, 0.657/0.669 (cases/controls); ERCC2/XPD 23591A allele frequency, 0.343/0.331 (cases/controls). Test for Hardy-Weinberg equilibrium in the control subjects: P =.94. ERCC1 8092C allele frequency, 0.770/0.752 (cases/controls); ERCC1 8092A allele frequency; 0.230/0.248 (cases/controls). Test for Hardy-Weinberg equilibrium in the control subjects: P=.54. Table 3. ERCC1 8092 and ERCC2/XPD 23591 Risk/Combination Genotype Frequencies and Risk Estimates* Genotype Cases Controls Odds Ratio (95% Confidence Interval) Crude Adjusted ERCC1 8092 AA/AC 130 (41.5) 141 (45.1) 1.00 1.00 CC 183 (58.5) 172 (54.9) 1.15 (0.84-1.58) 1.15 (0.84-1.59) ERCC2/XPD 23591 GG 123 (39.3) 142 (45.4) 1.00 1.00 GA/AA 190 (60.7) 171 (54.6) 1.28 (0.93-1.76) 1.28 (0.93-1.76) Genotype combinations Neither ERCC1 8092CC nor ERCC2/XPD 23591 GA/AA 28 (9.0) 39 (12.5) 1.00 1.00 Either ERCC1 8092CC or ERCC2/XPD 23591 GA/AA 197 (62.9) 205 (65.5) 1.34 (0.79-2.26) 1.34 (0.79-2.27) Both ERCC1 8092CC and ERCC2/XPD 23591 GA/AA 88 (28.1) 69 (22.0) 1.78 (1.00-3.17) 1.78 (0.99-3.17) Adjusted for age, sex, smoking status, and alcohol use. Trend test for the allele dose effect: P =.04. and P=.94, respectively). The variant ERCC2/XPD 23591A allele was more frequent in the case subjects (0.343) than in the control subjects (0.331), and the homozygous wildtype ERCC2/XPD 23591GG genotype was less common in the case group (39.3%) than in the control group (45.4%), suggesting that the ERCC2/XPD A allele increases risk. Risk estimates are summarized in Table 3. The ERCC1 8092CC genotype was associated with a borderline increased risk of SCCHN: adjusted OR, 1.15 (95% CI, 0.84-1.59). The ERCC2/XPD 23591A allele was also associated with a borderline increased risk of SCCHN: adjusted OR, 1.28 (95% CI, 0.93-1.76). Furthermore, having both of these risk genotypes (ie, both ERCC1 8092CC and ERCC2/XPD 23591 GA/AA) was associated with a significantly increased risk of SCCHN, and having only 1 risk genotype was associated with a borderline increased risk, suggesting an allele dose effect (trend test, P=.04). Subgroup analyses of the combined effect of ERCC1 8092CC and ERCC2/XPD 23591 GA/AA risk genotypes are summarized in Table 4. In smokers and drinkers, the risk estimates approached significance: adjusted OR for smokers, 1.46 (95% CI 0.95-2.25) and for drinkers, 1.40 (95% CI; 0.94-2.10). Furthermore, the combined risk genotype was associated with a significantly increased risk for pharyngeal cancer: adjusted OR, 1.59 (95% CI, 1.02-2.49). COMMENT In this study, we assessed the risk of SCCHN associated with ERCC1 8092 and ERCC2/XPD 23591 genotypes in a hospital-based, case-control analysis of 626 non- Hispanic white subjects closely matched on age, sex, smoking status, and alcohol use. Our findings were consistent with the prior report by Chen et al 10 of an association between ERCC1 8092CC and adult-onset glioma. The frequency of the variant ERCC1 8092A allele in our control subjects (0.248) was similar to the control group of Chen and colleagues (0.270). However, in that study, a significant risk associated with the ERCC1 8092CC genotype was found only in a glioma histologic subgroup of 28 oligoastrocytomas. 10 There are no other case-control data of tumor risk associated with the ERCC1 8092 genotype. Our findings are also consistent with an association between the ERCC2/XPD 23591A allele and decreases in DNA repair capacity reported in a casecontrol study of lung cancer risk. 9 Although Lunn et al 11 reported that the ERCC2/XPD 23591A allele is associated with better DNA repair as measured by a cytogenetic assay of chromatid aberrations induced by irra- 1086
Table 4. Stratification Analysis of Combination Risk Genotype Frequencies, Odds Ratios, and 95% Confidence Intervals* Subjects With Both ERCC1 8092CC and ERCC2/XPD GA/AA Genotypes Odds Ratio (95% Confidence Interval) Variable Cases Controls Crude Adjusted Age, y 50 24 (24.2) 23 (23.5) 1.04 (0.54-2.01) 1.08 (0.54-2.13) 51-63 32 (29.1) 22 (19.6) 1.68 (0.90-3.13) 1.67 (0.89-3.14) 63 32 (30.8) 24 (23.3) 1.46 (0.79-2.71) 1.40 (0.75-2.61) Sex Male 65 (27.8) 50 (21.8) 1.38 (0.90-2.10) 1.39 (0.90-2.13) Female 23 (29.1) 19 (22.6) 1.41 (0.69-2.84) 1.41 (0.69-2.85) Smoking status Smoker 66 (28.1) 49 (21.3) 1.44 (0.94-2.21) 1.46 (0.95-2.25) Nonsmoker 22 (28.2) 20 (24.1) 1.24 (0.61-2.50) 1.24 (0.61-2.51) Alcohol status Drinker 74 (29.7) 57 (23.1) 1.41 (0.94-2.11) 1.40 (0.94-2.10) Nondrinker 14 (21.9) 12 (18.2) 1.26 (0.53-2.98) 1.26 (0.52-3.02) Site Oral cavity 28 (25.7) 69 (22.0) 1.22 (0.74-2.03) 1.21 (0.72-2.02) Pharynx 45 (31.0) 69 (22.0) 1.59 (1.02-2.48) 1.59 (1.02-2.49) Larynx 15 (25.4) 69 (22.0) 1.21 (0.63-2.30) 1.22 (0.63-2.34) Adjusted for age, sex, smoking status, and alcohol use within each subgroup. diation, they assessed only 29 subjects, making genotyping subgroup analysis particularly subject to chance findings. In fact, they reported an ERCC2/XPD 23591 variant A allele frequency of 0.420, which is much higher than the 0.331 we found in our 313 control subjects. Furthermore, ERCC2/XPD is a component of the nucleotide excision repair pathway, which is responsible for removal of tobacco-induced adducts and UV-induced dimers but not the repair of chromatid breaks induced by irradiation. In a study of 96 patients with lung cancer and 94 cancer-free controls, Butkiewicz et al 12 also suggested the ERCC2/XPD 23591A allele may have a potential protective effect, but a case-control analysis of adult-onset glioma found no association with the ERCC2/XPD 23591 genotype. 13 In comparison, the study by Spitz et al 9 of more than 450 subjects found that in the case subjects the ERCC2/XPD 23591A allele was associated with suboptimal DNA repair function as measured by the wellestablished host cell reactivation assay. 9 Such suboptimal DNA repair is associated with increased risk of lung cancer 14 and SCCHN. 3 These discrepancies in results between studies are obviously due to differences in sample sizes and the assays used. However, these results also suggest that these individual genotypes probably have only a modest effect on cancer risk, if any, and that more studies with larger samples are needed to clarify the role of these genotypes in cancer risk. Future genetic models will probably include multiple genotypes to significantly and reliably predict cancer risk. Furthermore, efforts to determine the effect of such polymorphisms on repair function must be encouraged. It is unlikely that such common genetic variants have a major effect on cancer risk independent of environmental exposures. The increased risk in those exposed to tobacco and alcohol suggests a potential geneenvironment effect. In other words, these genotypes may put individuals at increased risk only if they are also exposed to a carcinogen. Some epidemiological data suggest that tobacco and alcohol exposure may have a greater effect on pharyngeal cancer risk than on oral cavity or laryngeal cancer risk. 15 This may explain why we found these risk genotypes to be most common in the pharyngeal cancer subgroup. Regardless, these findings in subgroup analyses are preliminary, may be due to chance, and must be confirmed in larger studies. Accepted for publication February 13, 2002. This study is supported in part by start-up funds from The University of Texas M. D. Anderson Cancer Center (Dr Sturgis) and research grants CA 55769 (Dr Spitz), CA 70334, ES 11740 (Dr Wei), and CA 16672 (to M. D. Anderson Cancer Center) from the National Institutes of Health, National Cancer Institute, Bethesda, Md. This work was presented in part at the annual meeting of the American Head and Neck Society, Palm Desert, Calif, May 14-16, 2001. We thank Margaret Lung, RN, for assistance with recruiting patients; Maureen Goode, PhD, and Chris Yeager, BA, BS, for scientific editing; and Deanna Thomas, BS, for manuscript preparation. Corresponding author: Erich M. Sturgis, MD, Department of Head and Neck Surgery, Box 441, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030-4009 (e-mail: esturgis @mdanderson.org). REFERENCES 1. Giovino GA, Schooley MW, Zhu BP, et al. Surveillance for selected tobacco-use behaviors United States, 1900-1994. Mor Mortal Wkly Rep CDC Surveill Summ. 1994;43(suppl S-3):1-43. 1087
2. Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancer statistics, 2001. CA Cancer J Clin. 2001;51:15-36. 3. Cheng L, Eicher SA, Guo Z, Hong WK, Spitz MR, Wei Q. Reduced DNA repair capacity in head and neck cancer patients. Cancer Epidemiol Biomarkers Prev. 1998;7:465-468. 4. Wang LE, Sturgis EM, Eicher SA, Spitz MR, Hong WK, Wei Q. Mutagen sensitivity to benzo[a]pyrene diol epoxide and the risk of squamous cell carcinoma of the head and neck. Clin Cancer Res. 1998;4:1773-1778. 5. Li D, Firozi PF, Chang P, et al. In vitro BPDE-induced DNA adducts in peripheral lymphocytes as a risk factor for squamous cell carcinoma of the head and neck. Int J Cancer. 2001;93:436-440. 6. Cheng L, Sturgis EM, Eicher SA, Spitz MR, Wei Q. Nucleotide excision repair gene expression and the risk for squamous cell carcinoma of the head and neck. Cancer. 2002;94:393-397. 7. Hoeijmakers JHJ. Genome maintenance mechanisms for preventing cancer. Nature. 2001;411:366-374. 8. Shen MR, Jones IM, Mohrenweiser H. Nonconservative amino acid substitution variants exist at polymorphic frequency in DNA repair genes in healthy humans. Cancer Res. 1998;58:604-608. 9. 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. 10. Chen P, Wiencke J, Aldape K, et al. Association of an ERCC1 polymorphism with adult-onset glioma. Cancer Epidemiol Biomarkers Prev. 2000;9:843-847. 11. Lunn RM, Helzlsouer KJ, Parsad R, et al. XPD polymorphisms: effects on DNA repair proficiency. Carcinogenesis. 2000;21:551-555. 12. Butkiewicz D, Rusin M, Enewold L, Shield PG, Chorazy M, Harris CC. Genetic polymorphisms in DNA repair genes and risk of lung cancer. Carcinogenesis. 2001;22:593-597. 13. Caggana M, Kilgallen J, Conroy JM, et al. Associations between ERCC2 polymorphisms and gliomas. Cancer Epidemiol Biomarkers Prev. 2001;10:355-360. 14. Wei Q, Cheng L, Amos CI, et al. Repair of tobacco carcinogen-induced DNA adducts and lung cancer risk: a molecular epidemiologic study. J Natl Cancer Inst. 2000;92:1764-1772. 15. Franceschi S, Talamini R, Barra S, et al. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx, and esophagus in northern Italy. Cancer Res. 1990;50:6502-6507. Call for Photographs ARCHIVES OF OTOLARYNGOLOGY HEAD & NECK SURGERY Covers W ith the January 2001 issue, the ARCHIVES OF OTOLARYNGOLOGY introduced nonmedical photographs as cover art for the journal. We are bombarded with medical and technical information every minute of every day and this is our way of offering you, our readers, a moment to reflect, smile, breathe a little more deeply, maybe even escape for just a second and relax a bit. Do you have a scenic photograph you have taken that you think would make a great cover shot? We d love to see it! Submissions should be from our readers, reviewers, authors, or anyone affiliated with the journal, and MUST be formatted horizontally. They can be black and white or color and at least 3.5 5 in but no larger than 8 10 in. If you wish to submit a digital photograph, please call our office at (404) 778-2322 for guidelines. Due to legal concerns, no recognizable people should appear in the picture, and please include details about where the picture was taken, how you happened to be there, and anything else you think is interesting about the image. We need the photographer s complete name, highest academic degree, city and state of residence, and a statement explaining how he or she is affiliated with the journal. Send submissions to ARCHIVES OF OTOLARYNGOLOGY, 1440 Clifton Rd NE, Suite 400, Atlanta, GA 30322. If you would like your photo returned, please enclose a self-addressed, stamped envelope. Cover photos will be chosen at the discretion of the ARCHIVES editorial staff. Michael M. E. Johns, MD Editor 1088