X-Ray Repair Cross-Complementing Group 1 ( XRCC 1 ) Genetic Polymorphisms and Thyroid Carcinoma Risk : a Meta-Analysis

Thyroid cancer is of special concern in endocrinology because it accounts for more than 90% of all endocrine cancers and contributes to more than 50% of all deaths from endocrine cancers (Gilfillan, 2010; Aschebrook-Kilfoy et al., 2011). So far, exposure to ionizing radiation is the only well established risk factor for thyroid cancer, especially when it occurs in early stages of life (Papadopoulou et al., 2009). However, there are evidences that many gene polymorphisms including DNA repair genes influence on thyroid cancer susceptibility (Gudmundsson et al., 2012, Jendrzejewski et al., 2012). X-ray repair cross complementing group 1 (XRCC1), located on chromosome 19q13.2–13.3, with 33 kilobases in length, and encodes a scaffold protein involved in the repair of DNA single-strand break (SSB) formed by ionizing radiation and alkylation damage (Chou et al., 2008). XRCC1 exert its role by interacting with other repair proteins such as OGG1, poly (ADP ribose) polymerase (PARP), polynucleotide kinase, and proliferating cell nuclear antigen (PCNA) (Hoeijmakers,


Introduction
Thyroid cancer is of special concern in endocrinology because it accounts for more than 90% of all endocrine cancers and contributes to more than 50% of all deaths from endocrine cancers (Gilfillan, 2010;Aschebrook-Kilfoy et al., 2011). So far, exposure to ionizing radiation is the only well established risk factor for thyroid cancer, especially when it occurs in early stages of life (Papadopoulou et al., 2009). However, there are evidences that many gene polymorphisms including DNA repair genes influence on thyroid cancer susceptibility (Gudmundsson et al., 2012(Gudmundsson et al., , 2001Marsin et al., 2003). There have been more than 300 validated SNPs (single nucleotide polymorphism) in the XRCC1 gene, however, only three SNPs have been extensively studied including Arg194Trp, Arg280His and Arg399Gln. The three SNPs have been reported to be associated with many types of cancer, such as lung cancer (Dai et al., 2012), gastric cancer (Chen et al., 2012), breast cancer (Huang et al., 2009) and leukemia .
There are also some case-control studies conducted to explore the association of SNPs of XRCC1 and thyroid cancer risk (Zhu et al., 2004;Chiang et al., 2008;Siraj et al., 2008;Akulevich et al., 2009;Ho et al., 2009;Sigurdson et al., 2009;Fard-Esfahani et al., 2011;García-Quispes et al., 2011;Ryu et al., 2011). However, the results were not consistent. For example, Ryu's study (Ryu et al., 2011) have found that the XRCC1 Arg194Trp Arg/ Trp genotype was significantly associated with a decreased risk of papillary thyroid carcinoma compared to that of Arg/Arg genotype (OR with 95 CI; 0.550 [0.308-0.983]), however, no such association was observed in Esfahani's study (Fard-Esfahani et al., 2011); Quispes's study (García-Quispes et al., 2011) have found a positive association (OR = 1.58, 95% CI 1.05-2.46) for XRCC1 Arg280His, however, no such tendency was found in Chiang's study (Chiang et al., 2008); Akulevich's study (Akulevich et al., 2009) have found that XRCC1 Arg399Gln polymorphisms was associated with a decreased risk of PTC according to the multiplicative and dominant models of inheritance, however, Siraj's study (Siraj et al., 2008) found no such association. These inconsistent results failed to clarify the complicated relationship between XRCC1 polymorphism and thyroid cancer risk. To reliably explore the effect of XRCC1 variants (Arg399Gln, Arg280His, and Arg194Trp) on thyroid cancer, we conduct this meta-analysis including all of the evidence to date.

Search strategy
Eligible articles were retrieved by searching the PubMed bibliographical database (up to July 30, 2012) using the following combination of keywords: (XRCC1 OR X-ray repair cross-complementing gene 1) AND (thyroid) AND (polymorphism OR polymorphisms OR variants OR variant). In addition, we checked the references in the retrieved articles to avoid missing studies. There was no restriction on language in this search.

Inclusion and exclusion criteria
For an article to be included in this meta-analysis, it must accord to the following criteria: 1) case-control studies evaluating the association between XRCC1 polymorphisms and thyroid cancer risk; 2) sufficient published data available to calculate odds ratio (OR) with 95% confidence interval (CI). Those not designed as casecontrol studies, reviews, and those provided no controls or no usable data were excluded.

Data extraction
A predesigned data extraction table was used to extract the data by two independent reviewers. Disagreement was resolved by discussion. The following information was extracted from each included article: journal name, first author, year of publication, area and ethnicity, sample size, design type, inclusion and exclusion criteria, the number of genotypes in both cancer cases and controls, and the results of the studies.

Statistical analysis
Statistical analysis was conducted using STATA statistical software (version 11; Stata Corporation, College Station,Texas). Two-sided P values less than 0.05 were considered statistically significant. The strength of XRCC1 SNPs in Arg194Trp, Arg280His, and Arg399Gln and thyroid cancer risk was assessed by odds ratios (ORs) with the corresponding 95%CI for each study. The OR and its 95% CI in each comparison were assessed for the genotypes: 1) aa versus AA (a was for the minor allele and A was for the major allele); 2) Aa versus AA; 3) Dominant genetic model ( aa+Aa versus AA); 4) Recessive genetic model ( aa versus Aa+AA). The chi-squared (χ 2 ) based Q-statistic test was used to assess heterogeneity across studies, p < 0.10 was considered statistically significant    (Egger et al., 1997). In the control populations, Hardy-Weinberg equilibrium (HWE) was tested. In addition, subgroup analysis for ethnicity (Asian and Caucasian Mixed population), design type (HCC(hospital based casecontrol study) and PCC(population based case-control study)) and sample size ( smaller (total sample<400) and larger (total sample≥400)) was conducted, and influence analysis was performed by omitting each study to find potential outliers. Two authors performed the statistical analysis independently and got the same results.

Literature selection and study characteristics
Twelve articles were retrieved from PubMed, three of which were excluded after detailed assessment (one was case report, one was not about thyroid cancer and one was review). Finally, nine studies met the inclusion criteria (1620 cases and 3557 controls). Three of these were conducted in an Asian population (499 cases and 674 controls), six in a Caucasian population (1121 cases and 2883 controls). There were 9 studies with a total of 1620 casses and 3557 controls for Arg399Gln polymorphism, 6 studies with a total of 932 cases and 2270 controls for Arg194Trp polymorphism, and 7 studies with a total of 1432 cases and 3356 controls for Arg280His polymorphism. Genotype distributions in the controls of all studies were in agreement with HWE. The detailed characteristics of the studies are shown in Table 1.

Sensitivity analysis
Influence analysis was conducted to assess the influence of each individual trial on the pooled ORs by sequential omission of individual studies. The results suggested that no individual trial significantly affected the pooled ORs (Figure 1 a,b).

Publication bias
For each of the three SNPs, publication bias was examined by funnel plots qualitatively and estimated by Begg's and Egger's tests quantitatively. Taken the Arg399Gln polymorphism for example, the shapes of the funnel plot did not indicate any evidence of obvious asymmetry in dominant genetic model (Figure 2). Moreover, the p values from the Begg's test (p=0.88) and Egger's test (p=0.84) were all greater than 0.05, indicating no publication bias.

Discussion
DNA repair mechanisms play essential roles in maintaining the genomic stability which is constantly challenged by endogenous (reactive oxygen species) and exogenous agents (ionizing radiation) (Ming et al., 2012). XRCC1 plays an important role in the DNA repair pathway because it could specifically interact with nicked and gapped DNA, rapidly and transiently responds to DNA damage in cells, thus may serve as a strand-break sensor (Mani et al., 2004). In addition, XRCC1 could interact with many proteins known to be involved in BER andSSBR, so it has been proposed that XRCC1 may function as a scaffold protein able to coordinate and facilitate the steps of various DNA repair pathways (Mani et al., 2007). It is widely accepted that alterations in XRCC1 may play important roles in the process associated with the etiology of cancers because of the alteration of base excision repair functions (Monaco et al., 2007). The functional significance of XRCC1 Arg194Trp polymorphism is due to the location in an evolutionarily conserved region, and the occurrence of chromosomal breaks is largely increased among cases with the Arg/Arg genotype (Vodicka et al., 2007). The Arg280His is located in the PCNA-binding region of XRCC1, and could potentially alter the structure   DOI:http://dx.doi.org/10.7314/APJCP.2012.13.12.6385 XRCC1 Genetic Polymorphisms and Thyroid Carcinoma Risk: a Meta-Analysis of XRCC1 and its ability to interact with apurinic/ apyrimidinic endonuclease (Yan et al., 2009). Th e XRCC1 Arg399Gln polymorphism is located within a relatively non-conserved region between conserved residues of the BRCA1 COOH terminus domain, and may associated with higher sister chromatid exchange frequency and prolonged cell-cycle delay in response to ionizing radiation (Hu et al., 2001;Matullo et al., 2006).
Our meta-analysis included six studies with a total of 932 cases and 2270 controls for Arg194Trp polymorphism, seven studies with a total of 1432 cases and 3356 controls for Arg280His polymorphism and nine studies with a total of 1620 cases and 3557 controls for Arg399Gln polymorphism. To our knowledge, this is the first metaanalysis evaluated the association between XRCC1 polymorphisms and thyroid cancer risk. Unfortunately, we failed to observe any association between XRCC1 polymorphisms and thyroid cancer risk in the overall analysis based on all of the included studies. However, when we performed subgroup analyses by ethnicity, design type and sample size, we found that: 1) elevated risk in aa vs AA analysis (OR=2.03, 95%CI= 1.24-3.31, p=0.01) and recessive genetic model analysis (OR=1.93, 95%CI= 1.20-3.08, p=0.01) in the larger sample size trials for XRCC1 Arg194Trp polymorphism; 2) decreased thyroid cancer risk in subgroup analysis based on ethnicity in Aa vs AA analysis (OR=0.84, 95%CI= 0.72-0.98, p=0.03) and in a dominant genetic model (OR=0.84, 95%CI= 0.72-0.97, p=0.02) in Caucasian population for XRCC1 Arg399Gln polymorphism; 3) decreased thyroid cancer risk in subgroup analysis based on design type in Aa vs AA analysis (OR=0.72, 95% CI= 0.54-0.97, p=0.03) among the PCC trials for the Arg399Gln polymorphism.
In this meta-analysis, we did not observe any association of XRCC1 Arg194Trp polymorphism, Arg280His polymorphism and Arg399Gln polymorphism with thyroid cancer risk in the overall analysis, it may be due to the shortness of available data and the large heterogeneities in the studies. Therefore subgroup analysis based on ethnicity, design type and sample size was conducted to avoid potential bias. We did find a increased thyroid cancer risk associated with the Arg194Trp polymorphism in the larger sample size trials, it may be due to the high statistical power in the larger sample size trials. Previous study have found that the frequency distribution of Arg399Gln allele significantly varied in different ethnicities (Xing et al., 2002;Tumer et al., 2010), so it was essential to conduct a subgroup analysis based on ethnicities. In this subgroup analysis, we also found a decreased thyroid cancer risk associated with the Arg399Gln polymorphism among the Caucasian population. Although our results are suggestive, there are still some limitations in our meta-analysis. First, heterogeneity among the studies, resulting from different design type, ethnicity and sample size or some other factors, may influence the results of the analysis. Although we have conducted subgroup analysis based on these issues, the results may be also biased and the statistical power may be reduced by doing so, thus the results may be interpreted with caution due to a small number of studies. Second, environmental and lifestyle factors may alter the associations between gene polymorphisms and cancer risk. However, the relationship between XRCC1 gene polymorphism and thyroid cancer risk was analyzed without consideration of these interactions because of the lack of sufficient original data, which should be further studied. Third, publication bias may have occurred though the funnel plot did not show it since negative findings were likely to be unreported.
In conclusion, this comprehensive meta-analysis has evaluated all published data currently available on XRCC1 polymorphisms and thyroid cancer risk. This metaanalysis suggest that XRCC1 Arg399Gln polymorphism may be associated with decreased thyroid cancer risk among Caucasian population and the XRCC1 Arg194Trp may be associated with a tendency of increased thyroid cancer risk in the two larger sample size trials . We did not observe any association of XRCC1 Arg280His polymorphism with thyroid cancer. However, our results may be biased because of the relatively small number of objects, therefore, further larger studies should be conducted to validate the conclusion.