AIM: To determine whether -1195 A→G and/or -765 G→C polymorphisms in Cyclooxygenase-2 (COX-2) may have a risk modifying effect on the development of esophageal carcinoma in a Dutch Caucasian population.

METHODS: Two study groups were recruited, 252 patients with esophageal carcinoma and 240 healthy controls, matched for race, age, gender and recruiting area. DNA was isolated from whole blood and used for genotyping. PCR products were digested with restriction enzymes and products were analyzed by agarose gel electrophoresis. Odds ratios (OR) and 95% confidence intervals (CI) were estimated.

RESULTS: The distribution of the -1195 A→G polymorphism was significantly different in esophageal cancer patients compared to controls. The -1195 GG genotype resulted in a higher risk of developing esophageal adenocarcinoma (OR = 3.85, 95% CI: 1.45-10.3) compared with the -1195 AA genotype as a reference. The -765 G→C genotype distribution was not different between the two groups. The GG/GG haplotype was present more often in esophageal adenocarcinoma patients than in controls (OR = 3.45, 95% CI: 1.24-9.58; with AG/AG as a reference). The same trends were observed in patients with squamous cell carcinomas, however, the results did not reach statistical significance.

CONCLUSION: Presence of the COX-2 -1195 GG genotype and of the GG/GG haplotype may result in a higher risk of developing esophageal carcinoma.

During the last few decades, the incidence of esophageal carcinoma has sharply increased in Western-lifestyle countries. Two main types of esophageal carcinoma exist, adenocarcinoma and squamous cell carcinoma. The main difference between adenocarcinoma and squamous cell carcinoma is the cell type from which the tumor originates; glandular or squamous epithelial cells, respectively.

Adenocarcinoma of the esophagus predominantly occurs in Western societies. There is a strong and probably causal relation between gastro-esophageal reflux and the development of esophageal adenocarcinoma[1]. Gastro-esophageal reflux may cause damage to the esophageal tissue due to the high concentrations of acid and bile salts, which may induce metaplasia and cell proliferation, thereby increasing the risk of mutations. This can lead to Barrett’s esophagus with high grade dysplasia and ultimately to adenocarcinoma of the esophagus[12].

In contrast to adenocarcinoma, squamous cell carcinoma of the esophagus is thought to be caused predominantly by specific lifestyle or environmental factors such as heavy smoking in combination with alcohol use, chewing of tobacco or consumption of spicy foods and hot beverages[3]. In certain developing countries such as China, India or Iran, squamous cell carcinoma of the esophagus is very common, probably due to particular lifestyle habits[3]. As a result, damage to esophageal tissue may occur and tissue renewal may increase. This increased cell proliferation can lead to mutations, dysplasia and carcinoma.

Cell proliferation may play a key role in tumor genesis and cyclooxygenases (COXs) are important regulatory enzymes in this process. COXs are enzymes that catalyze the conversion of free arachidonic acid into prostaglandin H₂, which is the precursor of prostaglandins, prostacyclin and thromboxanes. These regulatory compounds play a role in many biological processes such as cell proliferation, angiogenesis, immune function and inflammation, which are all crucial in the development and progression of neoplasms[4]. The human COX family consists of three members, COX-1-3[45]. COX-1 is found in most tissues and plays a role in homeostasis of many physiologic processes. COX-3 is an alternative splice product of COX-1 and is believed to be involved in the regulation of pain and fever. COX-2 is probably very important in the development and progression of neoplasms. COX-2 is an inducible enzyme whose expression can be induced by pro-inflammatory and mitogenic stimuli like cytokines and growth factors. COX-2 plays an important role in the development of otherwise healthy tissue into metaplastic and dysplastic tissue, as well as in the development and progression of a tumor, by taking part in the regulation of cell proliferation, cell transformation, tumor growth, metastasis and invasion. COX-2 is often found over-expressed in gastrointestinal tumors, including those of the esophagus[6–10]. Tumors which exhibit a high level of COX-2 seem to be more aggressive[6] and patients bearing those tumors showed a significantly reduced survival[10]. In addition, when COX-2 expression in laboratory animals was suppressed with medication, fewer animals developed esophageal adenocarcinoma[11]. Therefore, the role of COX-2 in the development of normal or metaplastic tissue into neoplasms seems evident.

Recently, several functional Single Nucleotide Polymorphisms in the COX-2 gene have been discovered which may contribute to the variance in inter-individual COX-2 expression. The -1195 A→G substitution in the COX-2 promotor was found to be associated with a lower expression of COX-2 in a Chinese population[12].

Another SNP, -765 G→C was first described in a UK population[13]. This polymorphism was shown to result in a lower promoter activity, which could subsequently lead to a lower expression of COX-2.

The purpose of this study is to determine the possible modulating effect of the COX-2 polymorphisms -1195 A→G and -765 G→C on the risk for developing esophageal cancer in a Dutch Caucasian population.

Patients with esophageal carcinoma and controls were matched for race, age, gender and recruiting area. Table 1 shows the characteristics of the patients and controls. The COX-2 genotype distributions in patients and controls are summarized in Table 2. The polymorphisms tested here were distributed according to the Hardy-Weinberg criteria, P-values in patients and controls were 0.98 and 0.47 for the -765 G→C polymorphism and 0.21 and 0.24 for the -1195 A→G polymorphism, respectively.

No significant differences in the distribution of the -765 G→C polymorphism between patients with esophageal carcinoma and controls were observed (P = 0.80; χ² test). However, a significant difference in the distribution of the -1195 A→G polymorphism between patients and controls was observed (P = 0.02; chi square test). The -1195 G/-1195 G genotype was present more often in patients with esophageal carcinoma (whole group) as compared to the -1195 A/-1195 A genotype in controls (OR 3.57, 95% CI 1.39-9.13, P = 0.005). When analyzed according to the type of tumor, ORs were 3.85 (95% CI 1.45-10.3) for patients with adenocarcinoma and 3.29 (0.95-11.4) for patients with squamous cell carcinoma (Table 2). Allele frequencies in all patients with esophageal cancer (-1195 A vs -1195 G) also differed significantly from those in controls (P = 0.02).

When comparing the squamous cell carcinoma group (n = 70) with the adenocarcinoma group (n = 174), there were no significant differences with respect to the -1195 genotype distribution: -1195 AA, 55.7% vs 57.5%; -1195 AG, 37.1% vs 33.9% and -1195 GG, 7.2% vs 8.6% (P = 0.97). For the -765 genotypes no differences in distribution between the squamous cell carcinoma and adenocarcinoma groups were found: -765 GG, 69.5% vs 68.7%; -765 GC, 27.1% vs 28.2% and -765 CC, 3.4% vs 3.1% (P = 0.95).

Table 3 shows the results of a comparison of the distribution of the COX-2 -765 and -1195 haplotypes, according to the type of tumor. Only one significant difference was found, the GG/GG haplotype was present more often in the esophageal adenocarcinoma group than in the control group (OR = 3.45, 95% CI = 1.24-9.58). However, the number of individuals present in these subgroups was very small (n = 13 vs n = 6, respectively). The same trend was observed in the squamous cell carcinoma group, however, statistical significance was not reached.

The -1195 GG genotype was present more often in patients with esophageal carcinoma than in controls. This is in contrast to the findings of Zhang et al[12] who identified the -1195 AA genotype as a risk factor for esophageal carcinoma. It is commonly reported that COX-2 expression is higher in cancerous tissue, because high COX-2 expression contributes to and sustains inflammatory and pre-cancerous processes[46]. Zhang et al[12] also concluded that COX-2 mRNA expression in -1195 AA genotypes was much higher than the mRNA expression in tissues of patients with the -1195 GG genotype. Our findings now suggest that the COX-2 -1195 polymorphism has the opposite effect on esophageal carcinoma risk in Caucasians, as compared to Chinese patients. However, two limitations must be noted: firstly, we did not measure whether the COX-2 mRNA expression in -1195 AA genotypes was highest in our group of Caucasian patients, similar to the findings of Zhang et al[12] in Chinese patients. Secondly, there is a difference between our study population and that of Zhang et al[12]; the majority of our patients had adenocarcinoma (69%) and the minority suffered from squamous cell carcinoma (28%), whereas the Chinese patients in the study by Zhang et al[12] all had squamous cell carcinoma. In China, esophageal squamous cell carcinoma is significantly more common than adenocarcinoma, as it is mainly caused by lifestyle factors such as drinking hot beverages and eating spicy foods, whereas adenocarcinoma is associated with acid reflux as a result of the Western lifestyle[1]. In our patient group, we found no differences in the distribution of both COX-2 polymorphisms between patients with adenocarcinoma and squamous cell carcinoma, which suggests that the differences found when compared to the results of Zhang et al[12] could be assigned merely to racial differences rather than to differences in the type of tumor.

Another indication that racial differences in the study populations may explain the apparent contradictory results is obtained by comparing the distribution of the COX-2 polymorphisms in the Chinese and Dutch control populations. The genotype frequencies found in our Dutch controls for the -765 G→C and -1195 A→G polymorphisms were: 66.5% GG, 30.9% GC, 2.9% CC and 64.2% AA, 33.3% GA, 2.5% GG, respectively. Zhang et al[12] in a Chinese population reported genotype frequencies of 95.7% GG, 4.3% GC, 0% CC and 24.1% AA, 53.4% GA and 22.5% GG, respectively. Tan et al in Chinese controls more recently reported approximately the same genotype frequencies as Zhang et al: 95.2% GG, 4.8% GC, 0% CC and 23.7% AA, 53.2% GA and 23.1% GG, respectively[15].

On the other hand, our control group data on the COX-2 -765 genotype were in good agreement with other European control data recently reported from Denmark, being 73.2%, 24.8% and 2.0% for -765 GG, GC and CC genotypes, respectively[16]. In addition, the COX-2 polymorphism data in our patients are very similar to the recently reported COX-2 -765 and -1195 genotype distributions in Dutch esophageal adenocarcinoma patients by Moons et al[17], except for the -1195 GG genotype, which was present in 8.0% of our patients vs only 2.0% in the patients in the study by Moons et al[17].

The distribution of the -765 genotypes in the control group was not found to be significantly different when compared to the esophageal carcinoma group, whereas Moons et al[17] reported a significantly different -765 CC genotype distribution between a Dutch esophageal carcinoma group (n = 140) and a Barrett’s esophagus (n = 255) or reflux esophagitis (n = 240) patient group. It should be noted, however, that the number of -765 CC genotype individuals in these patient groups was very low, being seven, four and zero individuals, respectively[17]. Two main reasons for the difference in results between the two Dutch studies are as follows: firstly, our study was performed on a larger patient population than the study by Moons et al[17] (252 vs 140 patients), and secondly in our study, similar to the study by Zhang et al[12], a comparison between patients with esophageal cancer and healthy controls was made, in contrast to the study by Moons et al[17] where patients with Barrett’s esophagus or reflux esophagitis, both of which are at risk for esophagus carcinoma, were used for comparison.

Analyzing the COX-2 haplotypes showed that the GG/GG haplotype was present more often in the esophageal carcinoma group, which again is not in accordance with the results of Zhang et al[12] and Moons et al[17], who both found that the CA containing haplotypes carried the highest risk. Since the results of Zhang et al[12] and Moons et al[17] on different types of tumors (squamous cell carcinoma vs adenocarcinoma, respectively) are very similar, and more or less contradict our results, it was of interest to compare the haplotype distribution between our patients with squamous cell carcinoma vs adenocarcinoma. However, no significant differences were found.

In conclusion, the presence of the COX-2 -1195 GG genotype and of the GG/GG haplotype may result in a higher risk of developing esophageal adenocarcinoma and possibly also squamous cell carcinoma.