Association between genetic variants in the XPG gene and gastric cancer risk in a Southern Chinese population

Xeroderma pigmentosum group G (XPG) recognizes and excises DNA damage on the 3′ side during the DNA repair process. Previous studies indicated that XPG gene polymorphisms may associate with gastric cancer susceptibility, but results were inconsistent. We evaluated the association of five potentially functional XPG polymorphisms (rs2094258 C>T, rs751402 C>T, rs2296147 T>C, rs1047768 T>C, and rs873601 G>A) with gastric cancer susceptibility in 1142 gastric cancer cases and 1173 controls. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated using logistic regression models. Overall, no significant association was detected between any of selected polymorphism and gastric cancer risk. However, we found that individuals carrying 3-4 risk genotypes were at significantly higher risk of gastric cancer than those with 0-2 risk genotypes (OR=1.32, 95% CI=1.04-1.68, P=0.021). The stratification analysis revealed that the cumulative effect of risk genotypes (3-4 vs. 0-2) on gastric cancer were more prominent among subgroups older than 58 years and men. In conclusion, our results indicated that none of the selected XPG polymorphism could significantly alter gastric cancer susceptibility alone. These polymorphisms might collectively confer increased gastric cancer susceptibility. These findings would be strengthened by larger prospective multicenter studies involving different ethnic populations.


INTRODUCTION
Gastric cancer is one of the most common cancers with high mortality, ranking as the fifth most common and the third deadliest cancer in the world [1]. Decreasing trends in gastric cancer incidence and mortality have been reported in most industrialized countries, whereas it is still prevalent in developing countries, predominantly in China [2]. Despite remarkable progress achieved in multimodal therapy strategies, the survival of gastric cancer remains poor with overall 5year survival rates hovering around 25% [3]. Helicobacter pylori (H. pylori) infection is a wellestablished risk factor for gastric cancer. However, some countries with a high H. pylori infection rate have disproportionately low gastric cancer incidence or mortality [4][5][6]. These observations suggested that rather than any single factor alone, the development of gastric cancer stem from a combination of multiple factors, such as H. pylori infection, nutritional deficiencies, a high salt or a low fiber diet, smoking, alcohol consumption, high body mass index [7,8], and genetic predisposition [9].
DNA repair system is responsible for maintaining the stability and integrity of human genomic DNA [10], and DNA repair genes may serve as potential biomarkers for cancer predication and prognosis [11]. Nucleotide excision repair (NER), one of the highly evolutionarily conserved pathway, can monitor and repair a variety of DNA damages [12,13]. Failure to repair DNA damages may lead to a number of human diseases including xeroderma pigmentosum (XP) [14]. Xeroderma pigmentosum group G (XPG) gene is one of eight key genes [XPA to XPG, and excision repair cross complementing group-1 (ERCC1)] in the NER pathway [15]. XPG can recognize and cut DNA lesion on the 3' side to ensure the proper repair of damaged DNA [16,17]. XPG also serves as a nonenzymatic scaffolding for subsequent 5' incision by the XPF/ERCC1 heterodimer during the NER process [18].
Thus far, a number of studies have reported the relationship between single nucleotide polymorphisms (SNPs) in the XPG gene and cancer risk, including lung cancer [19,20], gastric cancer [21][22][23][24], esophageal squamous cell carcinoma [25], colorectal cancer [26][27][28], and neuroblastoma [29]. However, only a few papers with small sample sizes are available regarding the role of XPG gene SNPs in gastric cancer carcinogenesis, and conclusions remain conflicting [21][22][23][24]. Therefore, we performed this study to precisely determine the association between five potentially functional SNPs (rs2094258 C>T, rs751402 C>T, rs2296147 T>C, rs1047768 T>C and rs873601G>A) in the XPG gene and gastric cancer susceptibility with a total of 1142 patients and 1173 cancer-free controls in a Southern Chinese population.

Population characteristics
The final analysis consisted of 1142 cases and 1173 healthy controls (Supplemental Table S1). There were 65.59% and 67.26% men in cases and controls (P=0.393), respectively. However, regarding age, smoking status, drinking status, and pack-years, there existed significant difference (P<0.0001) between the cases and controls. Thus, we further adjusted for these variables in the multivariate analyses. Of the gastric cancer patients, 240 (21.02%) cases were diagnosed with gastric cardia adenocarcinoma, while 902 (78.98%) cases were with non-gastric cardia adenocarcinoma. In term of stage, 140 (12.26%), 329 (28.81%), 456 (39.93%), and 217 (19.00%) cases were classified as TNM stage I, II, III, and IV, respectively, according to the 7 th Edition of the American Joint Committee on Cancer (AJCC) [30].

Associations between XPG gene polymorphisms and gastric cancer risk
The genotype frequencies of cases and controls for the five XPG SNPs and their associations with gastric cancer risk were summarized in Table 1. Observed genotype frequency distributions of all SNPs among the control subjects were in agreement with Hardy-Weinberg equilibrium (HWE). In the single factor analysis, no significant associations were observed between any of all the five polymorphisms and gastric cancer risk before and after adjusting for age, gender, pack-years, smoking and drinking status. We then determined the risk genotypes for each SNP based on its association with gastric cancer susceptibility. If a genotype of a SNP was associated with increase gastric cancer risk [odds ratio (OR)>1], the genotype was considered as a risk genotype, even if the association was not significant. When we combined the five polymorphisms, we observed that carriers of 3-4 risk genotypes had a significantly increased gastric cancer risk by 32%, when compared to carriers of 0-2 risk genotypes [OR=1.32, 95% confidence interval (CI)=1.04-1.68, P=0.021]. However, this association were weakened and became borderline significant (adjusted OR=1.29, 95% CI=0.99-1.69, P=0.062) after adjustment for age, gender, pack-years, smoking and drinking status.

Stratification analysis
In the stratified analysis by age, gender, smoking status, pack-year, drinking status, tumor sites and TNM stage, we further evaluated the effects of all the five SNPs and provided the results for rs751402 C>T, rs873601 G>A polymorphisms. The effects of combined risk genotypes on gastric cancer risk were also shown. We failed to AGING (Albany NY) find any significant association with gastric cancer risk for any studied variants among subgroups (Table 2). When the risk genotypes were combined, the significant associations with 3-4 risk genotypes were observed in individuals older than 58 years (adjusted OR=1.90, 95% CI=1.06-3.41, P=0.030) and men (adjusted OR=1.50, 95% CI=1.07-2.11, P=0.019), when compared to 0-2 risk genotypes.

Haplotype analysis
The frequency of inferred haplotypes of XPG gene based on observed genotypes and their association with the risk of gastric cancer were shown in Table 3. None of the haplotype was associated with gastric cancer risk significantly.

DISCUSSION
In the present study, we investigated the impact of five potentially functional XPG SNPs on gastric cancer risk in a Chinese Han population from South China. Our analysis indicated that none of these SNPs could individually influence the gastric cancer susceptibility. However, the individuals carrying 3-4 risk genotypes had a significantly increased gastric cancer risk, especially among those older than 58 years and men. To the best of our knowledge, this is the largest study to investigate the association of these five XPG polymorphisms with the gastric cancer risk by far.
XPG is an indispensable component of the NER pathway, which is responsible for the cleavage of DNA on the 3' side of lesion and also recruit PCNA to the damage sites for the subsequent gap-filling DNA synthesis in mammals [31]. It is reported that XPG also participates in other cellular processes, such as transcription-coupled DNA repair and RNA polymerase II transcription [32,33].
Recently, several studies have been carried out to explore the role of XPG polymorphisms in gastric cancer susceptibility; however, inconsistent results have been reported. We previously evaluated the association between XPG (rs2094258 C>T, rs2296147 T>C and rs873601 G>A) and gastric cancer risk in an Eastern Chinese population with 1125 cases and 1196 controls. We found that the rs873601 G>A polymorphism (located in the 3' UTR) was significantly associated with an increased gastric cancer risk [22]. We also demonstrated that rs873601 A allele was significantly associated with reduced mRNA expression level of

AGING (Albany NY)
XPG gene. These three polymorphisms were also genotyped in 337 gastric cancer cases and 347 controls by Yang and coworkers [24]. Intriguingly, they found that the rs2296147 T>C polymorphism was associated with a decreased gastric cancer risk, while the rs2094258 C>T polymorphism was associated with an increased gastric cancer risk [24]. In a study by Duan et al. [21], composed of 400 gastric cancer cases and 400 healthy controls, both rs751402 C>T and rs2296147 T>C polymorphisms were shown to significantly increase gastric cancer risk. Recently, Chen et al. [23] explored the association of rs2094258 C>T, rs751402 C>T, rs2296147 T>C and rs873601 G>A polymorphisms with gastric cancer susceptibility in 692 cases and 771 healthy controls. However, only XPG rs873601 G>A polymorphism appeared to be associated with the risk of gastric cancer. This controversy regarding the association might be partly due to ethnic and demographic differences, or insufficient statistical power caused by small sample size.
With this in mind, we conducted the current study with 1142 cases and 1173 controls. We found no significant association between variant genotypes of XPG polymorphisms and gastric cancer risk. However, the individuals carrying 3-4 risk genotypes were at significantly increased gastric cancer risk, especially for individuals older than 58 years and men. Overall, the negative results might be partially ascribed to the mild effect of each variant. In addition, the moderate sample size in this study might not be large enough to detect relatively weak association. Besides, complex interactions between environmental and genetic factors should be taken into account while measuring the true associations of XPG gene polymorphisms with gastric cancer.
Despite that this is the largest study to extensively analyze the association of five potentially functional XPG polymorphisms with gastric cancer in a Southern Chinese population, there still exists some limitations. First, frequency matching between cases and controls in this research were only performed on gender, but not on age, smoking and drinking status. We used multivariate logistic regression analysis to minimize the impact of these confounding factors, to some extent. Second, gastric cancer is a heterogeneous disease which might be influenced by other related factors such as H. pylori infection, diet, occupational exposure, and environmental factors. Since such information on participants was missing, the results should be explained with caution. Third, due to the hospital-based case-control design, our study was inevitably suffered from the selection bias. Moreover, the conclusions drawn from subjects residing in South China may not well represent other Chinese populations in the different regions. Fourth, only five potentially functional SNPs were included in this study. As a result, SNPs from the coding and the intron regions that may also be related to gastric cancer risk could be omitted. Finally, we only investigated the association between XPG gene polymorphisms and gastric cancer risk. Genetic variations in other genes (e.g. KDM5A, DNAH7 [34], PLCE1 [35], PSCA [36,37], PRKAA1 [38], MUC1 [39]) reported to be specifically associated with gastric cancer initiation and progression were not investigated in the current study.
In conclusion, we found that none of the XPG rs2094258 C>T, rs751402 C>T, rs2296147 T>C, rs1047768 T>C and rs873601 G>A polymorphisms was associated with gastric cancer susceptibility. However, cumulative effects of risk genotypes (3-4) on the risk of gastric cancer were observed. Further well-designed, prospective studies with large-scale multicenter investigations involving different ethnicities are required to verify our conclusions.

Study subjects
The study protocol was approved by the institutional review board of Sun Yat-sen University Cancer Center. All participants of this study signed individual informed consent. This study consisted of 1142 patients and 1173 healthy controls as we describe previously [40]. All subjects were unrelated ethnic Han Chinese population from Southern China, mainly from Guangdong, Guangxi, and Hainan province. In general, the response rate of cases and controls was more than 85%.

SNP selection and genotyping
Five potentially functional SNPs in the XPG gene were selected for this study as we described previously [ 28,29]. Briefly, we searched the potentially functional candidate SNPs located in the 5'-flanking region, exon, 5' UTR, and 3' UTR, which might affect transcription activity and the microRNA binding site activity. As predicted by SNPinfo software (http://snpinfo.niehs.nih.gov/snpinfo/snpfunc.htm), five SNPs (rs2094258 C>T, rs751402 C>T, rs2296147 T>C, rs1047768 T>C and rs873601 G>A) were potentially functional (Supplemental Table 2). All these SNPs have a minor allele frequency no less than 5% for Chinese Han subjects. There is no significant linkage disequilibrium (LD) (R 2 <0.8) among these SNPs. DNA samples were genotyped by the Taqman real-time PCR method as we described previously [22,41].

Statistical analysis
Goodness-of-fit χ 2 test was used to check whether genotype frequency distribution of each polymorphism in controls were in accordance with HWE. We compared the differences in demographic variables as well as genotype frequencies between cases and controls by using the two-sided χ 2 test. ORs and 95% CIs were used to estimate the effect of SNPs and haplotypes on gastric cancer risk. Adjusted ORs were calculated by unconditional multivariate logistic regression analysis, with adjustment for age, gender, pack-years, smoking and drinking status. We determined the risk genotypes for each SNP based on its association with gastric cancer susceptibility. If a genotype of a SNP was shown to increase gastric cancer risk (OR>1), the genotype was regarded as a risk genotype. For example, as to the rs2094258 C>T polymorphism, ORs of 1.02 (heterozygous model) and 1.17 (homozygous model) indicated that the T allele carriers (CT/TT) may have an increased risk when compared to those with CC genotypes (Table 1). Thus, the CC wild-type genotype carriers was define as 0, while the CT or TT genotype carriers was defined as 1. We then divided subjects into two groups based the number of risk genotypes. Carriers of 3-4 risk genotypes represented those carrying 3-4 risk genotypes of the five SNPs, while 0-2 risk genotypes represented those carrying 0-2 risk genotypes. All statistical analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC). A P value of <0.05 was considered as statistically significant.