Cyclooxygenase-2 and Inducible Nitric Oxide Synthase Gene Polymorphisms and Risk of Reflux Esophagitis, Barrett's Esophagus, and Esophageal Adenocarcinoma

The incidence of esophageal adenocarcinoma has increased in recent years, and Barrett's esophagus is a recognized risk factor. Gastroesophageal reflux of acid and/or bile is linked to these conditions and to reflux esophagitis. Inflammatory disorders can lead to carcinogenesis through activation of “prosurvival genes,” including cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS). Increased expression of these enzymes has been found in esophageal adenocarcinoma, Barrett's esophagus, and reflux esophagitis. Polymorphic variants in COX-2 and iNOS genes may be modifiers of risk of these conditions. In a population-based case-control study, we examined associations of the COX-2 8473 T>C and iNOS Ser⁶⁰⁸ Leu (C>T) polymorphisms with risk of esophageal adenocarcinoma, Barrett's esophagus, and reflux esophagitis. Genomic DNA was extracted from blood samples collected from cases of esophageal adenocarcinoma (n = 210), Barrett's esophagus (n = 212), and reflux esophagitis (n = 230) and normal population controls frequency matched for age and sex (n = 248). Polymorphisms were genotyped using TaqMan allelic discrimination assays. Odds ratios and 95% confidence intervals were obtained from logistic regression models adjusted for potential confounding factors. The presence of at least one COX-2 8473 C allele was associated with a significantly increased risk of esophageal adenocarcinoma (adjusted odds ratio, 1.58; 95% confidence interval, 1.04-2.40). There was no significant association between this polymorphism and risk of Barrett's esophagus or reflux esophagitis or between the iNOS Ser⁶⁰⁸ Leu polymorphism and risk of these esophageal conditions. Our study suggests that the COX-2 8473 C allele is a potential genetic marker for susceptibility to esophageal adenocarcinoma. (Cancer Epidemiol Biomarkers Prev 2008;17(3):727–31)

The incidence of esophageal adenocarcinoma has been increasing markedly over the last three decades (1). Barrett's esophagus increases the risk of esophageal adenocarcinoma by 30- to 125-fold (2).

Traditionally, reflux esophagitis has been seen as an early development on the reflux esophagitis-Barrett's esophagus-esophageal adenocarcinoma spectrum (3). However, recent evidence has suggested that reflux esophagitis and Barrett's esophagus may be distinct clinical entities (4). Although the underlying stimulus in these three conditions is gastroesophageal reflux of acid and/or bile, it is likely that various genetic and environmental factors interact to determine an individual's response to refluxate in the esophagus.

Inflammatory disorders, which can be caused by chemical agents (e.g., gastric refluxate and cigarette smoke), are recognized risk factors for carcinogenesis. Chronic inflammation leads to activation of “prosurvival genes,” including cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS; ref. 5). Increased expression of both these enzymes has been found in esophageal adenocarcinoma (6-8), Barrett's esophagus (6, 9), and reflux esophagitis (10). It has been shown that downstream mediators on the COX-2 and iNOS pathways lead to tumor promotion by stimulation of proliferation, promotion of angiogenesis, and inhibition of apoptosis (11, 12).

Various single nucleotide polymorphisms have been identified in the COX-2 and iNOS genes.

The COX-2 T>C polymorphism at position 8473 in the 3′ untranslated region is in an area shown to contain multiple regulatory elements, which alter COX-2 expression in a murine model (13). The iNOS Ser⁶⁰⁸ Leu allele (C>T polymorphism) leads to an amino acid alteration in an important regulatory domain of the enzyme (14). The minor allele frequency of these polymorphisms in Caucasian populations has been estimated at 35% and 13%, respectively.⁸

Because of their putative functional roles, these polymorphisms have been studied in association with various cancers but not in esophageal adenocarcinoma or its precursor lesions Barrett's esophagus and reflux esophagitis. We examined the association between COX-2 8473 T>C and iNOS Ser⁶⁰⁸ Leu polymorphisms and esophageal adenocarcinoma, Barrett's esophagus, and reflux esophagitis in a population-based case-control study.

The study methods have been described in detail elsewhere (15). Briefly, the Factors Influencing the Barrett's Adenocarcinoma Relationship study collected data and samples between March 2002 and July 2005. The study recruited four groups of Caucasian subjects: (a) patients with esophageal adenocarcinoma, (b) patients with long segment Barrett's esophagus, (c) patients with reflux esophagitis, and (d) normal population controls. Subjects were recruited throughout the island of Ireland with the exception of reflux esophagitis patients who were recruited only in Northern Ireland.

Esophageal adenocarcinoma cases (ages ≤85 years) were patients with a histologic confirmation of adenocarcinoma within the esophagus. Cases from Northern Ireland were identified from electronic pathology records from all pathology laboratories within the province. Republic of Ireland cases were identified from the main hospitals involved in the diagnosis and treatment of esophageal cancer.

Barrett's esophagus patients were eligible for inclusion if ≥3 cm of typical Barrett's mucosa were seen at endoscopy and the presence of specialized intestinal metaplasia was confirmed by histologic examination of biopsy specimens. Patients with dysplasia on histologic examination were not included.

Reflux esophagitis patients were included if there was macroscopically visible erosive esophagitis at endoscopy. Erosive esophagitis was defined as mucosal breaks or erosions within the esophagus (grades 2-4 in the Savary Miller/Hetzel-Dent classification or grades B, C, or D in the Los Angeles classification were included).

Eligible control subjects were adults without a history of esophageal or other gastrointestinal cancer, a known diagnosis of Barrett's esophagus or reflux esophagitis. Controls were frequency matched by sex and 5-year age band to the distribution of esophageal adenocarcinoma patients and selected at random from the General Practice Master Index.

All subjects underwent a structured computerized interview with trained interviewers. The questions covered demographic characteristics, occurrence of gastroesophageal reflux symptoms, smoking, alcohol intake, dietary history, medication use, occupational history, and anthropometric measures. The reference time point used in the analyses in this study for body mass index, smoking, reflux symptoms, and alcohol use was >5 years before the interview.

DNA was extracted from venous blood samples using the Puregene DNA purification kit (Gentra Systems). The COX-2 8473 T>C (rs5275) and iNOS Ser⁶⁰⁸ Leu (rs2275218) polymorphisms were genotyped using predesigned and validated TaqMan allelic discrimination assays (Applied Biosystems). Reactions were done using the Applied Biosystems Prism 7900HT sequence detection system, with an annealing temperature of 60°C.

Samples were processed without knowledge of their case-control status. Positive and negative control samples were included in each reaction. Genotyping was repeated for a random selection of 10% of samples, with 100% concordance for all assays. Any samples labeled as “undetermined” by the sequence detection system software were excluded from the analysis. One subject was “undetermined” after genotyping for each assay.

Deviation from Hardy-Weinberg equilibrium among cases and controls was tested for using the goodness-of-fit χ² test. Genotype frequencies among cases of esophageal adenocarcinoma, Barrett's esophagus, and reflux esophagitis were compared with controls using the χ² test for association while adjusting for potential confounding factors by logistic regression.

Significant differences were observed between the study groups after analysis of baseline characteristics (Table 1). We therefore identified age at interview, sex, body mass index 5 years before interview (kg/m²), years of full-time education, smoking status (never, ex-smoker, current smoker), and alcohol intake (g/d) as potential confounding factors. Trend tests were conducted by assigning the ordinal values 1, 2, and 3 to homozygous wild-type, heterozygotes, and homozygous variant genotypes, respectively, and by adding these scores as a continuous variable in the logistic regression model. Statistical significance was set at P < 0.05. All analyses were done using SPSS for Windows version 13.0 (SPSS).

The Factors Influencing the Barrett's Adenocarcinoma Relationship study was approved by the Research Ethics Committee of Queen's University Belfast, the Clinical Research Ethics Committee of Cork Teaching Hospitals, and the Research Ethics Committee Board of St. James's Hospital Dublin.

A total of 227 esophageal adenocarcinoma patients, 224 Barrett's esophagus patients, 230 reflux esophagitis patients, and 260 population controls were recruited as part of the Factors Influencing the Barrett's Adenocarcinoma Relationship study. The demographic characteristics of each group of subjects are presented in Table 1. The participation rate of eligible, alive esophageal adenocarcinoma patients was 74.2% and the overall esophageal adenocarcinoma response rate was 63.9%. The participation rates of Barrett's esophagus and reflux esophagitis patients and control subjects were 82.4%, 68.7%, and 41.8%, respectively. DNA samples were available for analysis from 210 (92.5%) esophageal adenocarcinoma patients, 212 (94.6%) Barrett's esophagus patients, 230 (100%) reflux esophagitis patients, and 248 (95.4%) controls.

The distribution and frequency of genotypes for cases and controls are shown in Table 2. There were no significant departures from Hardy-Weinberg equilibrium for either of the polymorphisms in any of the subject groups. For controls, the frequency of polymorphic variants was 0.32 (COX-2 8473 C) and 0.18 (iNOS Ser⁶⁰⁸ Leu). These were similar to frequencies reported in Caucasians in the SNP500 Database.⁹

There was no statistically significant association between the iNOS Ser⁶⁰⁸ Leu polymorphism and risk of esophageal adenocarcinoma, Barrett's esophagus, or reflux esophagitis either in the unadjusted analysis or after adjustment for a number of other potential confounding factors (see Table 2). In addition, there was no statistically significant association between the COX-2 8473 T>C polymorphism and risk of Barrett's esophagus or reflux esophagitis. However, the presence of at least one COX-2 8473 C allele was associated with a significantly increased risk of esophageal adenocarcinoma both before and after adjustment for potential confounding factors (see Table 2). The adjusted odds ratio (OR) of COX-2 8473 TC + CC versus TT was 1.58 [95% confidence interval (95% CI), 1.04-2.40; P = 0.03]. In addition, a significant additive effect across this genotype was shown (P_trend = 0.04).

To the best of our knowledge, this is the only case-control study to date to explore the association between COX-2 and iNOS polymorphisms and risk of esophageal adenocarcinoma, Barrett's esophagus, or reflux esophagitis. Our study results indicate that the presence of the COX-2 8473 C allele may predispose an individual to esophageal adenocarcinoma. However, this polymorphism was not associated with risk of its precursor conditions Barrett's esophagus or reflux esophagitis. Furthermore, the iNOS Ser⁶⁰⁸ Leu polymorphism did not seem to modulate risk for any of the studied esophageal conditions.

The strengths of our study include the population-based design, the relatively large sample size, and the efforts made to minimize misclassification bias by characterizing phenotypes clearly. Detailed interview data allowed us to control for a range of potential confounding factors, and DNA was available for analysis in almost all study subjects (>90%). Both polymorphisms were in Hardy-Weinberg equilibrium and had sufficient statistical power (>80%) to detect an OR of >2.0. We carefully selected two polymorphisms to evaluate by a review of published data, thus reducing the likelihood of type I error.

Previous studies in Caucasians investigating association between COX-2 8473 T>C and various cancers have produced conflicting results. Increased risks of breast (16), lung (17), and basal cell (18) carcinoma have been described. However, other studies in breast (19), prostate (20), and colorectal (21) cancers have reported null findings. There is no direct evidence in humans that this polymorphism, located in the 3′ untranslated region of exon 10 in the COX-2 gene, has a functional role. However, variation in this region has been shown to alter COX-2 expression in a murine model (13). Increased expression of the COX-2 enzyme in esophageal adenocarcinoma has been well shown in previous studies (6-8). We therefore hypothesize that the expression of the COX-2 8473 C allele, which was associated with an increased risk of esophageal adenocarcinoma, may be higher than that of the “wild-type” 8473 T allele. However, further in vitro analysis of the genetic regulation of COX-2 expression will be necessary before a conclusion on the functionality of the COX-2 8473 T>C polymorphism can be drawn.

Only two studies to date have investigated an association between the iNOS Ser⁶⁰⁸ Leu polymorphism and cancer both in Asian populations. From China, Shen et al. reported no significant association with gastric cancer (22), and similar null findings were reported in a Korean cervical cancer study (23). These results may in part be explained by the relatively low variant allele frequency in these populations (15.2% and 13.8%, respectively; refs. 22, 23), limiting study power. This may also be a factor in our study, with a variant allele frequency of 18% in controls.

A further limitation of our study was that misclassification bias may have arisen because asymptomatic control subjects did not have endoscopy done to exclude Barrett's esophagus or reflux esophagitis. However, given that the population prevalence of Barrett's esophagus has been estimated at 1.2% and reflux esophagitis at 5.7% (24, 25), this is unlikely to have had a major effect on our findings. In addition, not all patients with reflux esophagitis diagnosed at endoscopy had biopsies taken to definitively exclude Barrett's esophagus, which may have led to further misclassification bias. However, the prevalence of histologic Barrett's esophagus in subjects with endoscopically diagnosed reflux esophagitis has been estimated at 2.6% in a similar Caucasian population (25). If applied to the 230 reflux esophagitis subjects in our study, only 6 subjects with Barrett's esophagus may have been misclassified.

Given that the sample size of our study is relatively small compared with those in more common cancers, such as prostate (20) and breast (19), chance findings cannot be excluded and we did not have adequate power to detect a modest excess risk associated with the iNOS Ser⁶⁰⁸ Leu polymorphism (OR < 2.0) or to examine the role of possible gene-environment interactions.

In addition, the low participation rate in the control group may have introduced selection bias, which is a common problem in all case-control studies. However, controls in our study were selected at random from the population via a General Practice database.

In conclusion, the COX-2 8473 C allele may be a risk factor for esophageal adenocarcinoma, but further larger population-based studies are required to confirm this finding.