An Updated Meta-analysis Between the Association of XRCC 1 Arg 399 Gln Polymorphism and Hepatocellular Carcinoma Risk

Background: Various studies have evaluated the relationship between X-ray repair cross-complementing group 1 (XRCC1) Arg399Gln polymorphism and hepatocellular carcinoma (HCC) risk, but the conclusions have been inconsistent and underpowered. The purpose of this updated meta-analysis was to examine whether XRCC1 Arg399Gln polymorphism confers susceptibility to HCC. Methods: Eligible studies extracted from PubMed, Embase, Cochrane Library, VIP (chinese) and CNKI (chinese) up to November 2013 were included in


Introduction
Hepatocellular carcinoma (HCC) is the fifth most frequent diagnosed cancer in men and the seventh in women respectively, and the third most common cause of cancer-related death worldwide exceeded only by lung cancer and stomach cancer (Ferlay et al., 2010;Jemal et al., 2011).With a dramatic increase in incidence and mortality, HCC has become a global health challenge and has aroused growing public concern.It is accepted that HCC is a complex and multi-factorial disease, and its carcinogenesis still remains elusive (Sato et al., 2011;Forner et al., 2012).Risk factors enhance a person's chance of getting diseases.Some factors including chronic hepatitis B or C, obesity, diabetes, excessive alcohol consumption, pre-existing liver cirrhosis together with exposure to aflatoxin B1 are main known risk factors for HCC (Gomaa et al., 2008;Hagymasi et al., 2008;Caldwell et al., 2009;Forner et al., 2012).Besides, genetic factors have been reported to influence host's susceptibility and may play a vital role in the progression of HCC (Farazi et al., 2006;Sato et al., 2011;El-Serag 2011).

An updated meta-analysis between the association of XRCC1 Arg399Gln polymorphism and hepatocellular carcinoma risk
Xiao-Lian Zhang, Yu Lu, Shi Yang, Qi-Liu Peng, Jian Wang, Li Xie, Yan Deng, Yu He, Tai-Jie Li, Xue Qin*, Shan Li* Recently, attention has focused on genetic variations in DNA repair pathways, as unrepaired DNA damage can lead to unregulated cell growth and even cancer (Hoeijmakers, 2001).The base excision repair (BER) pathway is one of the four major DNA repair pathways for the processing of small lesions caused by alkylation and oxidation damage (Almeida et al., 2007).The X-ray repair cross-complementing group 1 (XRCC1) is one of the key DNA repair proteins involved in BER and singlestrand breaks (SSBs) repair through interacting with DNA ligaseⅢ and the complexes with DNA polymerase and poly (ADP-ribose) polymerase (PARP) (Masson et al., 1998;Vidal et al., 2001).Human XRCC1 gene spans 33 kb on chromosome 19q13.2-13.3 and composes of 17 exons (Mei et al., 2013;Li et al., 2013).The Arg399Gln polymorphism (rs25487) is G/A substitution at position 28152 on exon 10, which could alter XRCC1 function, diminish repair kinetics, and influence susceptibility to adverse health effect, such as cancer.To date, XRCC1 Arg399Gln polymorphism has been extensively explored the association with HCC risk (Long et al., 2004;Han et al., 2004;Chen et al., 2005;Kirk et al., 2005;Long et al., 2006;Borentain et al., 2007;Ren et al., 2008;Su 2008;Wu 2009;Kiran et al., 2009;Jia et al., 2010;Zeng et al., 2010;Pan et al., 2011;Tang et al., 2011;He et al., 2012;Han et al., 2012;Guo et al., 2012;Jung et al., 2012;Bose et al., 2013;Gulnaz et al., 2013;Mohana Devi et al., 2013).However, the sample sizes of these previous studies were limited and the molecular epidemiological studies into HCC risk are contradictory instead of conclusive.In addition, previous meta-analysis investigating this issue also generated conflicting results (Zhang et al., 2010;Li et al., 2013;Wu et al., 2013).In order to obtain a more accurate assessment of this relationship under different genetic models, an updated meta-analysis containing a total of 21 published studies was performed, which may provide more comprehensive evidence for the relationship of XRCC1 Arg399Gln variants with HCC risk.

Search of eligible studies
Eligible studies about XRCC1 Arg399Gln polymorphism and HCC risk were identified by systematic searches of PubMed, Embase, Cochrane Library, VIP (chinese) and CNKI (chinese) that contained all of the records published up to November 2013, using the following key words in both English and Chinese: XRCC1 or X-ray repair cross-complementing group 1, HCC or hepatocellular carcinoma and polymorphism.Although the search was initially performed without restrictions of language, the final analysis we only allowed the full-text articles published in English and Chinese.Additional eligible records were checked by a manual search of the references in the retrieved studies.Review articles were also inspected to find other relevant publications.

Selection criteria
The studies had to satisfy all the included criteria: (a) assessed the association between XRCC1 Arg399Gln polymorphism and HCC risk; (b) a case-control or cohort study; (c) studied on human beings; (d) provided sufficient data to calculate an odds ratio (OR) and a 95% confidence interval (CI); (e) published in English or Chinese language; (f) if more than one article reported on the same or overlapping data, only the study with the largest sample size was included.Family-based design study, meta-analysis, letters, case reports, reviews and editorials were excluded.

Data extraction
Based on the inclusion criteria, literature searches and identification of eligible articles were carried out by two independent reviewers (Xiaolian Zhang and Yu Lu).Then, two separate investigators extracted data from all eligible studies and the result was reviewed by a third reviewer (Xue Qin).The following data was extracted from each study: the first author's name, publication year, country, ethnicity, genotyping method, source of controls, number of cases and controls, genotypes frequency and Hardy-Weinberg equilibrium (HWE) of controls.

Statistical analysis
HWE was calculated for control groups of each study using the goodness-of-fit (χ 2 or Fisher's exact test), and p<0.05 was considered representative of deviation from HWE.The strength of association between the XRCC1 Arg399Gln polymorphism and HCC susceptibility was evaluated by odds ratio (OR) together with their 95 % confidence interval (CI) under the allele model (A vs G), the homozygous model (AA vs GG), the heterozygous model (AG vs GG), the dominant model (AA+AG vs GG) and the recessive model (AA vs GG+AG).The statistical significance of the pooled OR was determined by the Z test, and p<0.05 was considered statistically significant.A Q-test and the I 2 test were performed to assess statistical between-study heterogeneity assumption (Higgins et al., 2002;2003).If the result of the Q-test was p<0.10 or I 2 >50%, indicating there was heterogeneity among studies, the pooled OR estimate of the each study was calculated by the a random-effects (the DerSimonian and Laird method) model (DerSimonian et al., 1986), otherwise fixed-effects model (the Mantel-Haenszel method) was applied (Mantel et al., 1959).Sensitivity analysis was performed to assess the stability of the results by sequential omission of individual studies.Potential publication bias was diagnosed by Egger's linear regression test (p<0.05was considered representative of statistically significant publication bias) (Egger et al., 1997) and visual observation of Begger's funnel plot.All analyses were performed by STATA version 12.0 (Stata Corporation LP, College Station, Texas, USA).All the tests were two-sided, p<0.05 was considered statistically significant.

Characteristics of studies
Based on the search criteria, 78 records were identified during the initial search.When review the title and abstract, only 26 full-text studies were preliminarily identified for further detailed examination (Figure 1).Under the inclusion criteria, five of these articles were excluded: four were overlapped subjects (Yu et al., 2003 Yang et al., 2004;Long et al., 2005;Li et al., 2012), one was family-based design study (Ding et al., 2012).At last, a total of 21 studies with 4,170 cases and 5,030 controls were included in the final meta-analysis.The eligible studies were published from 2004 to 2013.Of the 21 studies, there were 19 studies for Asians, one study for Africans and one study for Caucasians, respectively.13 studies were English and eight were Chinese literatures.The distribution of genotypes for XRCC1 Arg399Gln polymorphism in the controls were in consistent with HWE except the six studies (Han et al., 2004;Long et al., 2006;Su 2008;He et al., 2012;Guo et al., 2012;Bose et al., 2013).The main characteristics of the 21 case-control studies were summarized in Table 1.

Sensitivity analysis
Sensitivity analysis was conducted to evaluate the influence of the individual studies on the pooled OR by sequential omission of each eligible study.The corresponding pooled OR were not changed when any single study was removed, indicating that the statistical results did not suggest significant effects, revealing the stability and credibility of the results.

Publication Bias
Begg's funnel plot and Egger's test were performed to assess the publication bias of included studies in all comparison models.Begg's funnel plot is relatively straightforward to observe whether the publication bias is present, and Egger's test was used to provide statistical evidence of symmetries of the plots.The shape of the funnel plots showed no obvious asymmetry (Figure 3) and then result of Egger's test did not show statistical evidence for bias (AG vs GG: p=0.870).

Discussion
XRCC1 gene is a key DNA repair gene involved in BER, which play an important role in the stability and integrity of the genome and the pathogenesis and development of human cancers (Poehlmann et al., 2010).Three common single nucleotide polymorphisms (SNP) in the XRCC1 gene, including Arg399Gln (rs25487), Arg280His (rs25489) and Arg194Trp (rs1799782) are extensively studied in many studies and produced nonconservative changes (Shen et al., 1998).These mutations that can alter XRCC1 function may contribute to the risk of cancers.
Although chronic hepatitis B or C, obesity, diabetes, excessive alcohol consumption, pre-existing liver cirrhosis and exposure to aflatoxin B1 have been identified as significant risk factors, there is limited understanding on the molecular mechanisms HCC (Gomaa et al., 2008;Hagymasi et al., 2008;Caldwell et al., 2009;Forner et al., 2012).It is accepted that the carcinogenesis of HCC is a multistep process, and multiple factors including environmental and genetic factors are involved in this complex process (Sato et al., 2011).Many epidemiological studies investigating the association between XRCC1 Arg399Gln polymorphism and HCC risk have provided inconsistent results.Two previous meta-analysis (Zhang et al., 2010;Wu et al., 2013) both conducted with seven studies showed that the XRCC1 Arg399Gln polymorphism might not be risk factors for HCC.But another meta-analysis including 13 literatures (Li et al., 2013) obtained conflicting results.Previous meta-analysis did not contain all appropriate studies, and lack of stratified analysis by ethnicity, which may lead to a deviation to final result.Therefore, an updated meta-analysis including 21 studies with 4, 170 cases and 5, 030 controls was performed to comprehensively assess the relationship between XRCC1 Arg399Gln polymorphism and HCC risk.Subgroup analysis was categorized by ethnicity.Our meta-analysis statistical data showed that XRCC1 Arg399Gln genotypes were associated with an increased risk of HCC, especially among the Asians.However, a reduced risk was detected among the Caucasians and no association was found among the Africans.This indicated that different populations living in different environment and genetic backgrounds may influence the association between XRCC1 Arg399Gln polymorphism and HCC risk.However, the conclusion should be interpreted with caution, because the study which focused on Caucasians and Africans both had one included study with a small sample size.Our results of this meta-analysis also may cause by chance because studies with a small sample size may be underpowered or may have generated a fluctuated risk assessment, so that further studies need to be performed to improve the statistical power.
A comprehensive analysis was performed to show the association between XRCC1 Arg399Gln polymorphism and HCC risk, but some limitations remain.Firstly, only two published studies included in this meta-analysis focused on Caucasians and Africans.Secondly, the sources of heterogeneity that existed among the studies were not addressed.Thirdly, this meta-analysis was based on unadjusted data, whereas a more precise analysis stratified by gender, age, smoking status, and environmental factors could be conducted if individual data were available.Finally, the controls were not uniformly defined.Most of the controls were chosen from healthy populations, but some were HBV or HCV positive, inpatient, and outpatient without HCC.Therefore, non-differential misclassification bias was possibly existed because these studies which including the controls may have different risks to develop HCC.
In conclusion, our meta-analysis of 21 case-control studies demonstrated that there was an increased risk between XRCC1 Arg399Gln polymorphism and HCC risk, especially in Asians, and a reduced risk in Caucasians.Due to limitations showed above in this analysis, it is necessary that well-designed and more-detailed studies with larger populations are needed to further evaluate the associations.Moreover, gene-gene and gene-environment interactions should be taken into consideration in future analysis.

Figure
Figure 3. Begg's Funnel Plot of XRCC1 Arg399Gln Polymorphism and HCC Risk for Publication Bias in the Heterozygous Model (AG vs GG).

Table 2 . Meta-Analysis of the Association between XRCC1 Arg399Gln and HCC
OR, odds ratio; CI, confidence interval; F, fixed effects model; R, random effects model of HCC under the allele model (A vs G: OR=0.