Sturgis EM, Dahlstrom KR, Spitz MR, Wei Q. DNA Repair Gene ERCC1 and ERCC2/XPD Polymorphisms and Risk of Squamous Cell Carcinoma of the Head and Neck. Arch Otolaryngol Head Neck Surg. 2002;128(9):1084–1088. doi:10.1001/archotol.128.9.1084
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).
A hospital-based case-control study.
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.
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 χ2 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.
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).
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.
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 al8 identified 31 SNPs of ERCC1, ERCC2/XPD, and ERCC4/XPF of the nucleotide excision repair pathway. Because 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
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.
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′-TAGTTCCTCAGTTTCCCG-3′, which generate a 255–base pair (bp) fragment for the 3′ UTR of ERCC1, and 5′-CTGTTGGTGGGTGCCCGTATCTGTTGGTCT-3′ and 5′-TAATATCGGGGCTCACCCTGCAGCACTTCCT-3′, which generate a 751-bp fragment for exon 10 of ERCC2/XPD. As previously described,10 we used single-strand conformational polymorphism assay to type the ERCC1 3′ UTR polymorphism. The restriction enzyme StyI (New England 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 wild-type 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 BioProducts, Rockland, Me) and photographed with Polaroid film (Cambridge, Mass).
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 + q2 = 1, where p is the frequency of the wild-type 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).
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 wild-type 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 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 wild-type 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).
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 al10 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 case-control study of lung cancer risk.9 Although Lunn et al11 reported that the ERCC2/XPD 23591A allele is associated with better DNA repair as measured by a cytogenetic assay of chromatid aberrations induced by γ irradiation, 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 γ irradiation.
In a study of 96 patients with lung cancer and 94 cancer-free controls, Butkiewicz et al12 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 al9 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 well-established host cell reactivation assay.9 Such suboptimal DNA repair is associated with increased risk of lung cancer14 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 gene-environment 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: email@example.com).