PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of carcinLink to Publisher's site
 
Carcinogenesis. 2009 May; 30(5): 785–792.
Published online 2009 March 6. doi:  10.1093/carcin/bgp058
PMCID: PMC2675653

Genetic susceptibility to esophageal cancer: the role of the nucleotide excision repair pathway

Abstract

In this case–control study with 387 White esophageal patients and 462 White controls matched to cases by age and sex, we evaluated the associations between 13 potential functional polymorphisms in eight major nucleotide excision repair (NER) genes and esophageal cancer risk. In individual single nucleotide polymorphism analysis, after adjustment for multiple comparisons, the heterozygous GT genotype of the ERCC1 3′ untranslated region (UTR) was associated with an increased risk, whereas the homozygous variant genotype TT was associated with 60% reduction in risk with an odds ratio (OR) of 0.40 (95% confidence interval [CI] = 0.19–0.86). The heterozygous AG genotype of XPA 5′ UTR was at 2.11-fold increased risk (95% CI = 1.33–3.35) and the risk reached 3.10-fold (95% CI = 1.94–4.95) for the homozygous variant GG genotype. These associations were also significant when restricted the analyses in patients with esophageal adenocarcinoma. Further, the CT genotype of the RAD23B Ala249Val was associated with increased esophageal cancer risk (OR = 1.44; 95% CI = 1.05–1.97), whereas the poly-AT−/+ genotype of the XPC intron 9 conferred a decreased risk (OR = 0.71, 95% CI = 0.51–0.97). In joint analysis, individuals carrying 1 (OR = 2.64, 95% CI = 1.57–4.52) and ≥2 (OR = 2.74, 95% CI = 1.58–4.75) unfavorable genotypes exhibited significantly increased risk for esophageal cancer risk with significant dose-response trend (P for trend = 0.006). The pathway-based risk was more evident in ever smokers, overweight/obese individuals, men and ever drinkers. Our results support the hypothesis that increasing numbers of unfavorable genotypes in the NER predispose susceptible individuals to increased risk of esophageal cancer. These findings warrant further replications in different populations.

Introduction

Esophageal cancer remains one of the most deadly malignancies, with a 5-year survival rate of <20% (1). With ~16,470 new cases and 14,280 deaths expected in the USA in 2008, carcinoma of the esophagus/gastroesophageal junction is the seventh most common cause of cancer death in men in the USA (1). There are two main types of esophageal cancer: squamous cell carcinoma and adenocarcinoma. The known risk factors for esophageal cancer include tobacco consumption, alcohol consumption, obesity, history of Barrett's esophagus and esophageal reflux disease (27).

Cigarette smoking is a major risk factor for esophageal cancer. Tobacco smoke contains a variety of carcinogens that react with DNA to form bulky DNA adducts. Obesity is a known risk factor for esophageal adenocarcinoma. Obesity may cause or exacerbate gastroesophageal reflux and the reflux exposure has been shown to cause oxidative DNA damage and the formation of bulky DNA adducts, which requires nucleotide excision repair (NER) pathway. Other DNA repair pathways such as the base excision repair pathway and double-strand break pathway are also involved in the repair of oxidative DNA damage. The NER pathway plays important roles in the repair of bulky lesions, such as pyrimidine dimers, photo-products, larger chemical adducts and cross-links and in the maintenance of genomic stability (8,9). Deficiencies in the DNA repair capacity of cells are a predisposing factor for cancer (10). NER involves global genomic repair that removes DNA lesions throughout the genome, and transcription-coupled repair that acts specifically on DNA lesions in the transcribed strand of transcriptionally active genes (11,12). The NER process involves at least four steps: damage recognition by a complex of bound proteins, including XPC-RAD23B, XPA and RPA in global genomic repair or ERCC6 and Cockayne syndrome type A in transcription-coupled repair; unwinding of the DNA by the transcription factor IIH complex that includes XPD and XPB; removal of damaged single-stranded fragment by 5′ ERCC1–XPF complex and 3′ XPG endonucleases and synthesis by DNA polymerases (13,14). Many of the genes involved in NER are highly polymorphic and their roles in cancer susceptibility have been studied (1517). To date, few studies have been conducted to evaluate the association between genetic variants in NER pathway and esophageal cancer risk.

In this study, we applied a pathway-based multigenic approach, which has been shown to augment the effects of individual polymorphisms and enhance cancer risk assessment (15), to investigate the individual and joint effects of 13 potential functional polymorphisms in eight major NER genes (ERCC1, XPD, XPA, XPG, XPC, RAD23B CCNH and ERCC6). We also used classification and regression tree (CART) analysis (18) to explore high order gene–gene interactions in modulating esophageal cancer risk. We hypothesized that common sequence variants of the NER pathway genes predispose susceptible individuals to increased risk of esophageal cancer.

Materials and methods

Study subjects

Esophageal cancer cases were recruited from the University of Texas M. D. Anderson Cancer Center in an ongoing study beginning in October 2004. Cases were identified by reviewing daily clinic schedule of all patients coming to the Department of Gastrointestinal Medical Oncology for clinic visits. One of the eligibility criteria is that all patients are diagnosed within 1 year. There was no age, sex or stage restriction on recruitment.

Healthy controls were identified from a pool of control subjects recruited in ongoing case–control studies (19). Briefly, controls were recruited in collaboration with the Kelsey-Seybold Clinic, the largest multispecialty medical organization in Houston, TX, that provides care to over 400 000 patients at 18 clinic locations. Potential controls were identified by reviewing short survey forms distributed to individuals visiting the clinic for health checkups or for addressing health concerns. The short survey forms elicit information on interest to participate in the study and basic demographic characteristics for matching. Potential control subjects were subsequently contacted by telephone to confirm their willingness to participate and screen for eligibility, and an appointment was scheduled at a Kelsey-Seybold Clinic site convenient to the participant. To be eligible, controls must have no prior history of cancer (except non-melanoma skin cancer). On the day of the interview, the controls visited the clinic specifically for the purpose of participating in this study but not for any treatment purposes. Control subjects were frequency matched to the cases by age (±5 years), gender and ethnicity. The response rates for cases and controls were 91.4 and 76.7%, respectively.

For both cases and controls, after written informed consent was obtained, trained M. D. Anderson staff interviewers administered a 45 min risk factor questionnaire to study participants. Data were collected on demographic characteristics, occupational history, tobacco use history, medical history and family history of cancer. At the end of the interview, a 40 ml blood sample was obtained from each participant and delivered to laboratory for molecular analyses. This study was approved by the institutional review boards of M. D. Anderson Cancer Center and Kelsey-Seybold Clinics.

An individual who had never smoked or had smoked <100 cigarettes in his or her lifetime was defined as a never smoker. An individual who had smoked at least 100 cigarettes in his or her lifetime but had quit >12 months before diagnosis (for cases) or before the interview (for controls) was coded as a former smoker. Current smokers were those who were currently smoking or had quit <12 months before diagnosis (for cases) or before the interview (for controls). Ever smokers included both former smokers and current smokers. Drinkers were defined as those who drank alcoholic beverages at least once a week for 1 year or more.

A total of 387 White patients and 462 White controls were included in this analysis. Of the 387 cases, 348 (89.92%) were classified having adenocarcinoma, whereas 39 (10.08%) had squamous cell carcinoma.

Genotyping

Genotyping procedures for NER single nucleotide polymorphisms (SNPs) were described previously (15). Genomic DNA was isolated from peripheral blood lymphocytes using the QIAamp DNA blood maxi kit (Qiagen, Valencia, CA). Except for XPA A23G, XPD Asp321Asn and XPC-poly-AT (PAT) polymorphism, which were genotyped using polymerase chain reaction–restriction fragment length polymorphism. Genotyping was performed using the Taqman method with a 7900 HT sequence detector system (Applied Biosystems, Foster City, CA) for the following NER SNPs: ERCC1 3′ untranslated region (UTR); XPD Lys751Gln; XPG Asp1104His; XPC Lys939Gln; XPC Ala499Val; RAD23B Ala249Val; CCNH Val270Ala; ERCC6 Met1097Val and ERCC6 Arg1230Pro. As a start, we used a candidate gene approach to examine the association in 13 potentially functional SNPs of eight major NER genes. Potentially functional SNPs were defined as non-synonymous SNPs with minor allele frequency >5%, SNPs located at 5′ UTR and splice site. Functional significance of these SNPs has been reported in previous studies (2023). In our laboratory, strict quality control procedures are implemented to ensure high genotyping accuracy. Positive controls and negative controls were included in each plate. A total of 5% of the samples were randomly selected and run in duplicates to ensure accuracy of the genotyping. Laboratory staff was blind of case–control status of the samples.

Statistical analyses

Pearson's χ2 test for categorical variables and the Student's t-test for continuous variables were performed to analyze the differences in distribution between cases and controls. Multivariate logistic regression analysis was conducted to estimate odds ratios (ORs) along with 95% confidence intervals (CIs), while adjusting for confounding variables. A likelihood ratio test was used to test for interactions among variables. This test compares the likelihood of a full model including the interaction term with a reduced model without the interaction term. A trend test was performed to test for a linear trend in the ORs. All statistical analyses were two sided. All analyses were performed using the Intercooled Stata 10.0 statistical software package (Stata Co., College Station, TX). To account for multiple comparisons, we used the false discovery rate (FDR) function implemented in the R software (Version 2.5) to estimate the FDR based on the Benjamini–Hochberg method (24). We calculated the FDR-adjusted P-values at 5% level to assess whether the resulting P-values were still significant after adjusting for multiple comparisons.

Haplotype and diplotype frequencies were analyzed using the HelixTree Genetics Analysis Software (Golden Helix, Bozeman, MT). Haplotypes were inferred using the expectation–maximization algorithm implemented in the HelixTree software. The adjusted ORs and 95% CIs for each haplotype were calculated using multivariate logistic regression using the most abundant haplotype as the reference group.

To explore gene–gene interaction in modulating esophageal cancer risk, we applied a recursive partitioning technique (18). The recursive partitioning was derived from the methodology of CART. In CART, a tree-based model is created by recursive partitioning the data and allows identify effect modifications between variables that are less visible by traditional logistic regression. The algorithm splits the study samples into a number of homogenous subgroups based on risk factors. The final model is a tree structure with terminal nodes defining a range of risk subgroups. CART analysis was performed using the RPART package in the R software (version 2.5).

Results

A total of 387 confirmed White esophageal cancer cases and 462 healthy White controls were analyzed in this study (Supplementary Table I is available at Carcinogenesis Online). By study design, there were no significant differences between cases and controls in terms of age (P = 0.76) and sex (P = 0.24). However, the case population had higher percentage of current smokers, whereas the control population had higher percentage of never smokers (P < 0.0001) (Supplementary Table I is available at Carcinogenesis Online). Among smokers, cases also reported to have higher pack year of smoking (35 versus 20, P < 0.001), longer duration of smoking (32 versus 22, P < 0.001) and higher number of cigarettes per day (26.2 versus 22.6, P = 0.01) than the controls. Cases exhibited higher body mass index (BMI) than controls (29.11 versus 28.41, P = 0.05). There were no significant differences between the distribution of drinking status between cases and controls (P = 0.29) (Supplementary Table I is available at Carcinogenesis Online).

In the analysis of individual SNPs, compared with the XPA 5′ UTR AA genotype, the heterozygous AG genotype was associated with a significantly increased risk with an OR of 2.11 (95% CI = 1.33–3.35), and the risk further increased to 3.10 (95% CI = 1.94–4.95) for the homozygous variant GG genotype (Table I). The association remained significant after FDR adjustment at 5% level. The ERCC1 3′ UTR variant T allele was associated with reduced risk in the recessive model (OR = 0.35, 95% CI = 0.16–0.75). In contrast, compared with the wild-type CC genotype of the RAD23B Ala249Val, the CT and TT genotype combined exhibited a significantly increased risk of esophageal cancer in the dominant model (OR = 1.35, 95% CI = 1.00–1.83). Further, the PAT−/+ genotype of the XPC intron 9 was associated with significantly decreased risk (OR = 0.71, 95% CI = 0.51–0.97). Other individual SNPs did not show significant association with the overall esophageal cancer risk.

Table I.
The association between single SNPs in NER pathway and esophageal cancer risk

Because adenocarcinoma is the major histology type, we restricted the above analyses in the adenocarcinoma cases only. Consistent with the above analysis that included all cases, the G allele of the XPA 5′ UTR was associated with significantly increased risk of esophageal adenocarcinoma (Table I). Moreover, consistent with the analysis that included all cases, the variant T allele of the ERCC1 3′ UTR was associated with significantly reduced risk (OR = 0.39, 95% CI = 0.18–0.83) in a recessive model (Table I). These associations remained significant after FDR adjustment for multiple comparisons at 5% level (Table I). Because the above-restricted analyses on adenocarcinoma cases yielded similar results as on all cases, subsequent analyses were performed using all cases and ORs were estimated for overall esophageal cancer risk.

As more than one SNP of XPD, ERCC6 and XPC were included in the study, we assessed the association between haplotypes of these three genes and esophageal cancer risk (Table II). No association was observed for XPD haplotypes and esophageal cancer risk. For ERCC6, only three haplotypes were observed in the study population (AG, AC and GG) and none of the haplotypes was associated with esophageal cancer risk. For XPC, compared with the most abundant ‘C–A’ haplotype, the haplotype ‘C–C’ was associated with 3.37-fold increased risk (95% CI = 1.81–6.27) and the association remained significant after FDR adjustment for multiple comparisons at 5% level (Table II).

Table II.
XPD, ERCC6, XPC haplotypes and esophageal cancer risk

To test the hypothesis that multiple SNPs in the NER pathway may have a cumulative effect on esophageal cancer risk, we estimated the combined effect of these SNPs and then stratified the analyses by host characteristics (Table III). The unfavorable genotypes were those showing significant association with esophageal cancer risk in the above single SNP analysis after FDR adjustment for multiple comparisons at a significance level of 0.05, specifically, the wild-type and heterozygous genotypes of ERCC1 3′ UTR (GG and GT), the GG and AG genotypes of the XPA 5′ UTR and the variant genotypes of RAD23B Ala249Val (TT and CT). Compared with the reference group, individuals carrying zero unfavorable genotype, individuals carrying one (OR = 2.64, 95% CI = 1.57–4.52) and two or more (OR = 2.74, 95% CI = 1.58–4.75) unfavorable genotypes exhibited significantly increased risk for esophageal cancer with significant dose-response trend (P for trend = 0.006) (Table III).

Table III.
NER pathway and esophageal cancer risk

Stratified analyses showed that this significant dose-response trend was apparent in ever smokers, subjects with BMI ≥25 (overweight and/or obese), men and drinkers (Table III). For example, among ever smokers, compared with reference group (zero unfavorable genotype), those with one and two or more unfavorable genotypes had ORs of 2.78 (95% CI = 1.50–5.14) and 3.08 (95% CI = 1.61–5.88), respectively (P for trend = 0.005); among overweight and/or obese subjects, subjects carrying one and two or more unfavorable genotypes exhibited increased risks of 3.52-fold (95% CI = 1.15–2.05) and 3.82-fold (95% CI = 1.82–7.98), respectively (P for tend < 0.001). Similar trends were observed among men and among ever drinkers (Table III). However, the interaction was only significant with drinking status (P for interaction = 0.03) (Table III). To see if removing of squamous cases could affect results, we also performed the joint pathway analysis and stratified analysis in patients with adenocarcinoma only. We observed similar significant associations in ever smokers, overweight and obese subjects, males and ever drinkers (results not shown).

We then applied CART analysis to explore high order gene–gene interactions. Figure 1 depicted the tree structure generated using the CART analysis. The final tree structure contained six terminal nodes, representing a range of low versus high-risk subgroups as defined by the different combination of genotypes of NER SNPs. The first split was XPA 5′ UTR, separating individuals into GG and AG+AA genotypes. Subsequent splits were XPC Ala499Val, ERCC6 Arg1230Pro, XPG Asp1104His and ERCC1 3′ UTR. To calculate ORs as defined by the six terminal nodes, we chose terminal node 1 (individuals carrying the AA or AG genotypes of the XPA and individuals with heterozygous or wild-type of the XPC Ala499Val) as the reference group. The ORs of terminal nodes ranged from 1.13 (terminal node 3) to 2.93 (terminal node 5) (Table IV). The highest risk group was individuals in terminal node 5 [homozygous GG of the XPA, wild-type of ERCC6 Arg1230Pro, wild-type of XPG Asp1104His and at least one variant C allele of the ERCC1 3′ UTR (OR = 2.93; 95% CI = 1.67–5.13)] (Figure 1 and Table IV).

Table IV.
CART terminal node and esophageal cancer risk
Fig. 1.
CART analysis of NER pathway and esophageal cancer risk. ORs and 95% CIs (in parenthesis) are presented underneath each terminal node box. Please refer to Table II for name of the SNPs. For each SNP, ‘WW’ represents wild type, ‘WM’ ...

Discussion

In this case–control study, we studied the effects of genetic polymorphisms in genes involved in the NER pathway as predisposition factor for esophageal cancer. Our results suggest that genetic variations in several NER genes showed significant association with esophageal cancer risk. Similar results were obtained when the analyses were restricted to patients with esophageal adenocarcinoma. Moreover, we took a pathway-based polygenic approach to assess the combined effects of multiple SNPs involved in the NER pathway. Our results suggest that esophageal cancer risk increased with the increasing number of unfavorable genotypes in the NER pathway. In stratified analyses, we showed that the effects of NER pathway genes on esophageal cancer risk might be modified by smoking status, obesity, gender and drinking status.

The associations between polymorphisms in NER genes and esophageal cancer risk have been studied in a few studies. Tse et al. (25) studied four SNPs in NER pathway in their associations with esophageal adenocarcinoma: XPD Lys751Gln; XPD Asp312Asn; ERCC1 8092 C/A and ERCC1 118C/T. They reported that the variant alleles of the XPD Lys751Gln, ERCC1 8092 C/A and ERCC1 118C/T were each individually associated with increased risk of esophageal adenocarcinoma. We found a borderline significant association for the variant allele of the XPD Lys751Gln in a recessive model. Similar to our study, they did not find an association for the XPD Asp312Asn SNP. Our study did not include the two ERCC1 SNPs in their study. In another study of the two XPD SNPs and a XRCC1 SNP by the same group, Liu et al. (26) reported significantly increased risk of adenocarcinoma associated with the variant allele of the XPD Lys751Gln but not XPD Asp312Asn. In a Chinese population, Guo et al. (27) reported that the G allele of the XPA 5′ UTR SNP was associated with significantly decreased risk of esophageal squamous carcinoma and that the association was even stronger among subjects who had no family history of upper gastrointestinal cancers; however, no associations were found for the XPC intron 9 SNP, the XPC Ala499Val SNP or the XPC Lys939Gln SNP (27).

Except for the association with esophageal cancer, polymorphisms in NER genes have been extensively studied for their associations with other cancers, such as lung cancer, bladder cancer, head and neck cancer, breast cancer and oral premalignant lesion (1517,28). In the individual SNP analysis of the current study, our result showed a significantly increased esophageal cancer risk associated with the XPA 5′ UTR G allele. Previous studies have generated inconsistent results regarding the association between the XPA 5′ UTR polymorphism with cancer susceptibility. The G allele was associated with reduced lung cancer risk (21,29,30), whereas the homozygous GG genotype was associated with significantly increased risk for oral premalignant lesions (28).

We found that individuals carrying at least one copy of the variant T allele of the RAD23B Ala249Val were associated with increased esophageal cancer risk. Shen et al. (31) reported increased lung cancer risk associated with the variant T allele of the RAD23B Ala249Val, while results from other studies are inconsistent (28,32). In our study, we also found that compared with the wild-type GG genotype, the heterozygous GT genotype of the ERCC1 3′ UTR was associated with an increased risk, whereas the homozygous variant genotype TT conferred a decreased risk of esophageal cancer. The protective effect of the TT genotype was even stronger in a recessive model. There are no reports on the association between this SNP and esophageal cancer risk.

Taken together, the inconsistent results generated from the individual SNPs may be due to different etiological factors in different cancers. It should be noted that the results derived from single-SNP analysis tend to have high FDRs because multiple hypotheses are tested simultaneously and the probability of type I error rates increases with the number of tests. In this study, we adopted the FDR adjustment for multiple comparisons. The significant findings of XPA 5′ UTR, ERCC1 3′ UTR and RAD23B Ala249Val were all adjusted for multiple comparisons. Nevertheless, the associations in esophageal cancer etiology warrant further validation in independent populations.

As to whether the unfavorable genotypes identified have potential functional impact, we previously reported more efficient DNA repair capacity associated with the G allele of the XPA 5′ UTR SNP (21,22), supporting the findings of Guo et al. (27) that the G allele of the XPA 5′ UTR SNP was associated with significantly decreased risk of esophageal cancer and also consistent with the decreased risk of lung cancer found in other studies (21,29,30). To our knowledge, there are no reports on the correlations with DNA repair phenotype for the ERCC1 3′ UTR SNP and RAD23B Ala249Val SNP.

The most important finding of this study is a significant trend of increased esophageal cancer risk with increasing numbers of unfavorable genotypes in the NER pathway. Consistent with this observation, in the study of Tse et al. (25), when four NER SNPs were combined in a pathway analysis, an increased risk of esophageal adenocarcinoma was observed with the number of variant alleles in a dose-response manner. Similar gene-dosage effect has been observed in other cancer types recently (15,33). As cancer is a complex and multifactorial disease that often occurs through interplay between multiple genetic and environmental factors, and it is seldom that polymorphisms in individual genes would have substantial effects on the overall risk. The rationale behind the gene–cancer risk association is that genetic variants may result in alterations in phenotypes (i.e. DNA repair capacity). In a recent study to evaluate the correlation between NER SNPs and mutagen sensitivity, an indirect measure of DNA repair capacity as quantified by the chromatid breaks induced by a mutagen challenge in short-term cultures of peripheral blood lymphocytes, we showed that the correlation between mutagen sensitivity and any individual NER SNP was modest at most and lacked a consistent pattern; however, when genotypes were combined in a pathway analysis and the effects were assessed across a panel of NER genes, we observed a consistent increasing trend in mutagen sensitivity (decreased DNA repair capacity) with increasing number of risk alleles in the NER pathway (22). In another study, Matullo et al. (34) found a dose-response relationship between the number of adverse alleles in DNA repair gene SNPs and the level of DNA adducts in peripheral blood cells, suggesting a progressive decrease in DNA repair capacity as the number of adverse alleles in DNA repair genes increases. Taken together, evidence suggested that increased number of risk alleles and/or unfavorable genotypes in NER pathway might result in decreased DNA repair capacity and thus increased cancer risk.

In stratified analysis, we found that the increased risk of esophageal cancer associated with NER genetic variants was evident in ever smokers, in overweight/obese subjects, in men and in drinkers. The stronger association in smokers is biologically possible. Cigarette smoke contains a variety of carcinogens that form bulky DNA adducts that require NER, and deficiency in NER may cause accumulation of unrepaired DNA adducts and consequently genetic instability and cancer. This finding is well supported by several other studies, which have reported similar interactions between DNA repair capacity and dose or duration of smoking (15,35).

The association between obesity and risk of esophageal adenocarcinoma has been reported in previous studies and results are generally consistent with an increased risk associated with increased BMI (2,5,6). In a recent systematic review and meta-analysis of 14 studies (two cohort and 12 case–control), it was found that a high BMI (>25) conferred an increased risk of esophageal adenocarcinoma in males (OR = 2.2, 95% CI = 1.7–2.7) and females (OR = 2.0, 95% CI = 1.4–2.9) (6). Obesity is known to cause or exacerbate gastroesophageal reflux and the reflux exposure in the form of acid and bile has been shown to cause loss of the p53 tumor suppressor protein, as well as inducing non-specific DNA damage in esophageal cells, including some forms of oxidative DNA damage and the formation of bulky DNA adducts, which requires NER repair pathway (3638). Our data suggest that decreased DNA repair capacity, as reflected by the carrying of high number of unfavorable genotypes in NER pathway, may place overweight and/or obese individuals at higher risk of esophageal cancer.

Our data also suggested that the increased risk of esophageal cancer associated with NER genetic variants was more evident in men than in women. It should be noted that our study population comprised mainly of men so that the small sample size of women reduced the statistical power to detect an effect, if exists. Also, the proportion of smokers was higher in men than in women and men tend to be heavier smokers than women. Therefore, the gender differences observed in this study may be partly attributed to differences in smoke exposure between gender groups. Further, it is suggested that obesity-related gastroesophageal reflux might be stronger among subjects with abdominal patterns of fat deposition, which is more commonly seen in men.

Finally, we found that compared with never drinkers, ever drinkers in our study population exhibited stronger association. Alcohol may increase esophageal adenocarcinoma risk by promoting gastroesophageal reflux, leading to the formation of bulky DNA adducts that require NER. Thus, compared with never drinkers, ever drinkers with suboptimal DNA repair capacity may be at an increased risk due to the accumulation of DNA adducts and genetic instability. However, due to the moderate sample size of our study population, the interaction between drinking and NER pathway should be interpreted with caution.

To explore the high-order interactions among the NER SNPs, we applied the CART analysis to define high versus low-risk subgroups. With this analytic approach, subgroups of individuals with a range of esophageal cancer risk profile were identified based on combinations of NER genotypes. The terminal nodes as defined by specific SNP combinations reflect risk subgroups resulted from the interactions between individual SNPs. Thus, the grouping of specific genotypes in the nodes of CART analysis may not be consistent with the results from single SNP analysis (Table I), which did not take into the account of the interactions with other SNPs. Further, because of the moderate sample size of this study, the number of subjects becomes small in terminal nodes and therefore these results should be interpreted with caution.

This study has several limitations. As a hospital-based case–control study, this study may be subject to potential selection bias. However, as our study was testing a genotype-driven hypothesis rather than an environment-driven hypothesis, selection bias is of less concern. Further, as all case–control studies, recall bias resulted from differential responses to environmental exposures between cases and controls may exist. However, recall bias should not affect genotype data. Thus, we anticipated no substantial influence of recall bias on the genetic–disease associations. Further, since the questionnaires on smoking habits were not directly validated in the study population using biomarkers, misclassification of exposure to cigarette smoke may exist, especially when smokers intend to report less smoking if they know smoking causes the disease. However, the bias due to this possible underestimation of smoking intensity should not be substantial because the only smoking variable used in the stratified analysis is smoking status (never versus ever), but not smoking intensity or smoking duration. The selection of SNPs in this study is still limited and is based on prior knowledge of potential functional significance of SNPs that have been related to cancer risk. A more comprehensive tagging SNP-based approach and a haplotype block analysis would better confirm the association and provide more complete information about the associations of NER genes and esophageal cancer risk.

In conclusion, we studied the effects of genetic variants in NER pathway on the susceptibility of esophageal cancer using a pathway-based multigenic approach. Our results support that common sequence variants of the NER pathway genes predispose susceptible individuals to increased risk of esophageal cancer and that the association may be modified by smoking status, obesity, gender and alcohol drinking. Due to the moderate sample size of the current study, these findings warrant further replications in different populations.

Supplementary material

Supplementary Table I can be found at http://carcin.oxfordjournals.org/

Funding

National Cancer Institute (R01 CA111922); Dallas, Park, Cantu and Smith Families; Rivercreet Foundation; University of Texas M. D. Anderson Cancer Center.

Supplementary Material

[Supplementary Data]

Acknowledgments

Conflict of Interest Statement: None declared.

Glossary

Abbreviations

BMI
body mass index
CART
classification and regression tree
CI
confidence interval
FDR
false discovery rate
NER
nucleotide excision repair
OR
odds ratio
SNP
single nucleotide polymorphism
PAT
poly-AT
UTR
untranslated region

References

1. Jemal A, et al. Cancer statistics, 2008. CA Cancer J. Clin. 2008;58:71–96. [PubMed]
2. Vaughan TL, et al. Obesity, alcohol, and tobacco as risk factors for cancers of the esophagus and gastric cardia: adenocarcinoma versus squamous cell carcinoma. Cancer Epidemiol. Biomarkers Prev. 1995;4:85–92. [PubMed]
3. Lagergren J, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N. Engl. J. Med. 1999;340:825–831. [PubMed]
4. Lagergren J, et al. The role of tobacco, snuff and alcohol use in the aetiology of cancer of the oesophagus and gastric cardia. Int. J. Cancer. 2000;85:340–346. [PubMed]
5. Lindblad M, et al. Body mass, tobacco and alcohol and risk of esophageal, gastric cardia, and gastric non-cardia adenocarcinoma among men and women in a nested case-control study. Cancer Causes Control. 2005;16:285–294. [PubMed]
6. Kubo A, et al. Body mass index and adenocarcinomas of the esophagus or gastric cardia: a systematic review and meta-analysis. Cancer Epidemiol. Biomarkers Prev. 2006;15:872–878. [PubMed]
7. Freedman ND, et al. A prospective study of tobacco, alcohol, and the risk of esophageal and gastric cancer subtypes. Am. J. Epidemiol. 2007;165:1424–1433. [PubMed]
8. Hiyama T, et al. Genetic polymorphisms and esophageal cancer risk. Int. J. Cancer. 2007;121:1643–1658. [PubMed]
9. Hoeijmakers JH. Genome maintenance mechanisms for preventing cancer. Nature. 2001;411:366–374. [PubMed]
10. Berwick M, et al. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J. Natl Cancer Inst. 2000;92:874–897. [PubMed]
11. Volker M, et al. Sequential assembly of the nucleotide excision repair factors in vivo. Mol. Cell. 2001;8:213–224. [PubMed]
12. Van Der Wees C, et al. Nucleotide excision repair in differentiated cells. Mutat. Res. 2007;614:16–23. [PubMed]
13. Friedberg EC. How nucleotide excision repair protects against cancer. Nat. Rev. Cancer. 2001;1:22–33. [PubMed]
14. Christmann M, et al. Mechanisms of human DNA repair: an update. Toxicology. 2003;193:3–34. [PubMed]
15. Wu X, et al. Bladder cancer predisposition: a multigenic approach to DNA-repair and cell-cycle-control genes. Am. J. Hum. Genet. 2006;78:464–479. [PubMed]
16. Michiels S, et al. Polymorphism discovery in 62 DNA repair genes and haplotype associations with risks for lung and head and neck cancers. Carcinogenesis. 2007;28:1731–1739. [PubMed]
17. Crew KD, et al. Polymorphisms in nucleotide excision repair genes, polycyclic aromatic hydrocarbon-DNA adducts, and breast cancer risk. Cancer Epidemiol. Biomarkers Prev. 2007;16:2033–2041. [PubMed]
18. Zhang H, et al. Recursive Partitioning in the Health Sciences. New York, NY: Springer; 1999.
19. Hudmon KS, et al. Identifying and recruiting healthy control subjects from a managed care organization: a methodology for molecular epidemiological case-control studies of cancer. Cancer Epidemiol. Biomarkers Prev. 1997;6:565–571. [PubMed]
20. Spitz MR, et al. Modulation of nucleotide excision repair capacity by XPD polymorphisms in lung cancer patients. Cancer Res. 2001;61:1354–1357. [PubMed]
21. Wu X, et al. XPA polymorphism associated with reduced lung cancer risk and a modulating effect on nucleotide excision repair capacity. Carcinogenesis. 2003;24:505–509. [PubMed]
22. Lin J, et al. Mutagen sensitivity and genetic variants in nucleotide excision repair pathway: genotype-phenotype correlation. Cancer Epidemiol. Biomarkers Prev. 2007;16:2065–2071. [PubMed]
23. Zhu Y, et al. Modulation of DNA damage/DNA repair capacity by XPC polymorphisms. DNA Repair (Amst.) 2008;7:141–148. [PMC free article] [PubMed]
24. Benjamini Y, et al. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. Royal Stat. Soc. Ser. 1995;B57:289–300.
25. Tse D, et al. Polymorphisms of the NER pathway genes, ERCC1 and XPD are associated with esophageal adenocarcinoma risk. Cancer Causes Control. 2008;19:1077–1083. [PMC free article] [PubMed]
26. Liu G, et al. XRC.C1 and XPD polymorphisms and esophageal adenocarcinoma risk. Carcinogenesis. 2007;28:1254–1258. [PubMed]
27. Guo W, et al. Polymorphisms of the DNA repair gene xeroderma pigmentosum groups A and C and risk of esophageal squamous cell carcinoma in a population of high incidence region of North China. J. Cancer Res. Clin. Oncol. 2008;134:263–270. [PubMed]
28. Wang Y, et al. Nucleotide excision repair pathway genes and oral premalignant lesions. Clin. Cancer Res. 2007;13:3753–3758. [PubMed]
29. Kiyohara C, et al. Genetic polymorphisms in the nucleotide excision pathway and lung cancer risk: a meta-analysis. Int. J. Med. Sci. 2007;4:59–71. [PMC free article] [PubMed]
30. Park JY, et al. Polymorphisms of the DNA repair gene xeroderma pigmentosum group A and risk of primary lung cancer. Cancer Epidemiol. Biomarkers Prev. 2002;11:993–997. [PubMed]
31. Shen M, et al. Polymorphisms in the DNA nucleotide excision repair genes and lung cancer risk in Xuan Wei, China. Int. J. Cancer. 2005;116:768–773. [PubMed]
32. Huang WY, et al. Excision repair gene polymorphisms and risk of advanced colorectal adenoma: XPC polymorphisms modify smoking-related risk. Cancer Epidemiol. Biomarkers Prev. 2006;15:306–311. [PubMed]
33. Zheng SL, et al. Cumulative association of five genetic variants with prostate cancer. N. Engl. J. Med. 2008;358:910–919. [PubMed]
34. Matullo G, et al. Combination of DNA repair gene single nucleotide polymorphisms and increased levels of DNA adducts in a population-based study. Cancer Epidemiol. Biomarkers Prev. 2003;12:674–677. [PubMed]
35. Gorlova OY, et al. Genetic susceptibility for lung cancer: interactions with gender and smoking history and impact on early detection policies. Hum. Hered. 2003;56:139–145. [PubMed]
36. Olliver JR, et al. Risk factors, DNA damage, and disease progression in Barrett's esophagus. Cancer Epidemiol. Biomarkers Prev. 2005;14:620–625. [PubMed]
37. Jolly AJ, et al. Acid and bile salts induce DNA damage in human oesophageal cell lines. Mutagenesis. 2004;19:319–324. [PubMed]
38. Sihvo EI, et al. Oxidative stress has a role in malignant transformation in Barrett's oesophagus. Int. J. Cancer. 2002;102:551–555. [PubMed]

Articles from Carcinogenesis are provided here courtesy of Oxford University Press