The Effect of XPD Polymorphisms on Digestive Tract Cancers Risk: A Meta-Analysis

Background The Xeroderma pigmento-sum group D gene (XPD) plays a key role in nucleotide excision repair. Single nucleotide polymorphisms (SNP) located in its functional region may alter DNA repair capacity phenotype and cancer risk. Many studies have demonstrated that XPD polymorphisms are significantly associated with digestive tract cancers risk, but the results are inconsistent. We conducted a comprehensive meta-analysis to assess the association between XPD Lys751Gln polymorphism and digestive tract cancers risk. The digestive tract cancers that our study referred to, includes oral cancer, esophageal cancer, gastric cancer and colorectal cancer. Methods We searched PubMed and EmBase up to December 31, 2012 to identify eligible studies. A total of 37 case-control studies including 9027 cases and 16072 controls were involved in this meta-analysis. Statistical analyses were performed with Stata software (version 11.0, USA). Odds ratios (ORs) with 95% confidence intervals (CIs) were used to assess the strength of the association. Results The results showed that XPD Lys751Gln polymorphism was associated with the increased risk of digestive tract cancers (homozygote comparison (GlnGln vs. LysLys): OR = 1.12, 95% CI = 1.01–1.24, P = 0.029, P heterogeneity = 0.133). We found no statistical evidence for a significantly increased digestive tract cancers risk in the other genetic models. In the subgroup analysis, we also found the homozygote comparison increased the susceptibility of Asian population (OR = 1.28, 95% CI = 1.01–1.63, P = 0.045, P heterogeneity = 0.287). Stratified by cancer type and source of control, no significantly increased cancer risk was found in these subgroups. Additionally, risk estimates from hospital-based studies and esophageal studies were heterogeneous. Conclusions Our meta-analysis suggested that the XPD 751Gln/Gln genotype was a low-penetrate risk factor for developing digestive tract cancers, especially in Asian populations.


Introduction
Digestive tract cancers, especially gastric, esophageal and colorectal cancers, are a major global health problem. Globocan data in 2008 showed [1] that the standardized incidence of colorectal cancer, gastric cancer and esophageal cancer were located in 4th, 6th and 9th in all tumors, respectively. The standardized mortality rate of gastric cancer, coming after lung cancer and breast cancer, ranked the third place. Moreover, colorectal cancer and esophageal cancer also ranked top ten in cancer mortality rankings. The incidence of different cancer varies widely among different racial and ethnic groups which may be partly attributed to lifestyle and genetic background [2]. Exposure to environmental carcinogens can cause different types of DNA damage that subsequently lead to carcinogenesis of different tissues, if left unrepaired [3].
DNA repair mechanisms, such as the nucleotide excision repair (NER), base excision repair pathway (BER) and double-strand break pathway, are essential for maintaining genome integrity and preventing carcinogenesis. NER, the most versatile, well studied DNA repair mechanism in humans, is mainly responsible for repairing bulky DNA damage, such as DNA adducts caused by UV radiation, mutagenic chemicals, or chemotherapeutic drugs [4]. The major component of NER, xeroderma pigmentosum group D (XPD or ERCC2), mapped in chromosome 19q13.3, spans over 20 kb, contains 23exons and encodes the 761-amino acid Table 1. Characteristics of XPD polymorphisms Included in the Meta-analysis. protein. It has two functions: nucleotide excision repair and basal transcription as part of the transcription factor complex (TFIIH) [5]. Mutations on different sites in XPD gene can give rise to repair and transcription defects, and altered DNA repair capacity can render a higher risk of developing different types of cancer [5][6][7][8][9][10][11]. Several polymorphisms of XPD were identified, like Asp312Asn, Lys751Gln, Arg194Trp and Arg399Gln. The XPD polymorphic loci that has been of particular interest in molecular epidemiology studies is the Lys751Gln polymorphism (rs13181) in exon 23 [12]. The lysine to glutamine transition at position 751 in exon 23 may affect different protein interactions, diminish the activity of TFIIH complexes, and alter the genetic susceptibility to cancer [13]. Genetic variant in XPD Lys751Gln had been demonstrated to be associated with some cancers risk in different meta-analysis, such as esophageal cancer, gastric cancer, colorectal cancer, breast cancer, prostate cancer, lung cancer and bladder cancer [14][15][16][17][18][19][20][21][22][23]. However, due to an insufficient number of publications, they did not calculate pooled odds ratios (ORs) of digestive tract cancers comprehensively. In consideration of the extensive role of XPD in digestive tract cancers, we performed a meta-analysis of all 37 eligible case-control studies: oral cancer, esophageal cancer, gastric cancer http://www.sciencedirect.com/science/article/ pii/S0188440911000853 -bib10and colorectal cancer, to derive a more precise association of XPD Lys751Gln polymorphism and different types of digestive tract cancers risk.

Identification of eligible studies
Using PubMed, we identified all published case-control studies which investigated the association between the XPD Lys751Gln polymorphism and digestive tract cancers risk using a retrieving query formulation ''(XPD or ERCC2) polymorphisms AND (colorectal cancer OR gastric cancer OR esophageal cancer OR oral cancer)''.The digestive tract cancers in this article refer to oral cancer, esophageal cancer, gastric cancer and colorectal cancer. We also searched references in published articles and reviews on this topic in PubMed. Eligible studies had to meet the following criteria: (a) only case-control designs were considered, (b) The study explored the correlation between different types of digestive tract cancers and XPD Lys751Gln polymorphism. Major exclusion criteria were (a) no control population, (b) no available genotype frequency. (c) Genotypic distribution of the controls was not in agreement with Hardy-Weinberg equilibrium (HWE). (d) Duplication of the previous publications, the largest or most recent publication was selected.

Data Extraction
Information was carefully extracted from all eligible publications independently by two authors according to the inclusion criteria listed above. If the two pieces of typed data were different, a third investigator would be asked to check and to make sure all data were right. The following information was extracted from each study: first author, year of publication, country of study population, ethnicity, source of controls, number of cases and controls with different genotypes and HWE (Table 1).

Statistical Analysis
We assessed the departure from the Hardy-Weinberg equilibrium for the control group in each study using Pearson's goodnessof-fit x 2 test with 1 degree of freedom. Heterogeneity among studies was checked by the random-effects model (the Der Simonian and Laird method) if there was significant heterogeneity [24]. A P value of more than the nominal level of 0.05 for the Q Table 1. Cont.

Meta-analysis
Table2 lists the main results of the meta-analysis for XPD Lys751Gln: having the Gln/Gln genotype is a risk factor for digestive tract cancers: GlnGln vs. LysLys: OR = 1.12, 95% CI = 1.01-1.24, P = 0.029, P heterogeneity = 0.133. I 2 = 20.9% (Figure 2). We did not find any significant association between the other genetic models and digestive tract cancers. The results of stratified analysis by cancer type, source of controls and ethnicity were shown in table 2. The Gln/Gln vs. Lys/Lys genotype had an elevated risk in Asian population (OR = 1.28, 95% CI = 1.01-1.63, P = 0.045, P heterogeneity = 0.287, I 2 = 14.2%; Figure 3). High heterogeneity was found in esophageal cancer and hospital-based studies, so the results can not be pooled together. In addition, the results did not suggest any association between XPD Lys751Gln polymorphism and digestive cancers susceptibility for all genetic models in European individuals or in population-based studies overall.

Sensitivity analysis
In the sensitivity analysis, when each particular study had been removed meta-analyses were conducted repeatedly. The corresponding pooled ORs were not qualitatively altered with or without this study. As shown in Figure 4, the most influencing single study on the overall pooled OR estimates seemed to be the one conducted by Mariana et al, which had a relatively large sample size. However, after the removal of the study, the result of the meta-analysis did not been influenced significantly: Gln/Gln vs. Lys/Lys: OR = 1.17, 95% CI: 1.05-1.30, indicating high stability of our results.

Heterogeneity analysis
There was moderate heterogeneity among these studies in GlnGln+GlnLys vs.LysLys comparisons and Gln/Gln vs. Lys/Lys comparisons, but not in the other genetic models. We explored the source of heterogeneity for dominant model by cancer type, ethnicity, source of control, and found that esophageal cancer and hospital-based studies contributed to substantial heterogeneity (Table3). One reason may be that hospital-based studies had relatively small samples and were more prone to random error and false positive or negative results. Furthermore, it is very likely that

Publication Bias
Begg's rank correlation method and Egger's weighted regression method were used to assess publication bias. There was no evidence of publication bias in XPD Lys751Gln (Begg's test P = 0.284, Egger's test P = 0.324, t = 1.00, 95% CI = 0.41-1.21). We present funnel plot for ORs of Gln/Gln versus Lys/Lys ( Figure 5).

Discussion
XPD plays a crucial role in NER, which is significant in the elimination of certain DNA cross-links, ultraviolet (UV) photolesions, and bulky chemical adducts. The XPD protein possesses both single-strand DNA-dependant ATP ase and 59-39 DNA helicase activities, which is essential for NER pathway and transcription [66]. Genetic variation in XPD may contribute to impaired DNA repair capacity and increased cancer risk. The Lys to Gln change at position 751 of XPD results in complete changes about the charge configuration of the amino acid, which affects the interactions of XPD protein and its helicase activator [67]. To date, a number of epidemiological studies have been conducted to evaluate the role of Lys751Gln polymorphism on several cancer risks, but the results remain controversial. As far as we know, several previous meta-analyses on XPD Lys751Gln polymorphism and cancers risk have been performed, such as gastric cancer, colorectal cancer, esophageal cancer, breast cancer and bladder cancer [14][15][16][17][18][19][20][21][22][23]. But to date, there is no meta-analysis on the association between digestive tract cancers risk and XPD Lys751Gln polymorphism. In order to derive a more precise estimation of relationship, we performed this meta-analysis of 37 studies, including 9027 cases and 16072 controls.
Through analyzing genotypes from the 37 eligible studies, we found the Gln/Gln genotype carries might be at potential risk to digestive tract cancers. The Lys to Gln variation on position 751 of XPD resulted in complete changes about the electronic configuration of the amino acid, which affected the interactions of XPD protein and its helicase activator [68]. Digestive tract cancers represent a homogenous group of malignancies in some ways. Different primary sites of digestive tract cancers have some shared risk factors. For example, except for smoking and alcohol consumption, eating rough, spicy, hot and non-digestible food is likely to damage the digestive tract tissue. In addition, H.Pylori infection is a major cause of gastric cancer, while nitrites derived from red meat and processed meat is a key risk factor for esophageal cancer and colorectal cancer. Such risk factors and their tissue specificity raise the possibility that the XPD polymorphism may be associated with digestive tract cancers risk. The functional XPD Lys751Gln polymorphism resulting in decreased activity of XPD protein may increase risk of digestive tract cancers on the basis of damage tissue.  In stratified analysis by cancer type, we found that all genetic models did not appear to have an effect on the risks of esophageal, gastric, colorectal and oral cancers. This was different from Ling Yuan's and Wu XB's studies [69,70]. However Bo Chen et al. [71] detected that Gln/Gln genotype carriers might have an increased risk of gastric cancer in the Helico-bacter pylori (H.pylori)-positive population, but not in the Helico-bacter pylori (H. pylori)-negative population. One possible explanation is that the modulation of digestive tract cancers risk may depend not only on a single gene/ single nucleotide polymorphism, but also on a joint effect of multiple polymorphisms within different genes or pathways, or on close interaction between polymorphisms and environmental factor. The other is that Helicobacter pylori infection is one of the clear etiologies of gastric cancer and maybe there is some relationship between helicobacter pylori and the polymorphic loci. In the subgroup of ethnicity, we found significant association  between XPD Gln/Gln polymorphism and increased risks of digestive tract cancers in Asians but not in European. We think ethnic differences and diverse live environment may partly explain the phenomenon. Furthermore, we believed differences in diet, such as food structure and cooking way, were the main cause of this result. In addition, it was also likely that the observed ethnic differences may be due to chance because studies with small sample size may have insufficient statistical power to detect a slight effect or may have generated a fluctuated risk estimate [72]. In summary, this meta-analysis indicated that XPD Lys751Gln polymorphism, individuals carrying the variant homozygote Gln/ Gln may increase the susceptibility of digestive tract cancers. And, significant associations were detected among Asians population. It should be noted explicitly: first, the effective sample size is much smaller for the Gln/Gln vs. Lys/Lys analyses than the other genetic models and therefore it is more prone to random error and false positive results; second, the results for GlnGln vs. GlyLys+ LysLys, while not statistically significant (OR 1.09, 95% CI = 0.99-1.20, P = 0.072, P heterogeneity = 0.385), strengthen our conclusion about which genetic model is most appropriate. Largescale case-control and population-based association studies are warranted to validate the risk identified in the current metaanalysis and investigate the potential gene-gene and geneenvironment interactions on digestive tract cancers risk.

Supporting Information
Checklist S1 (DOC) Author Contributions Figure 5. Begg's funnel plot for publication bias test (Homozygote comparison). Each point represents a separate study for the indicated association. doi:10.1371/journal.pone.0096301.g005