X-Ray Repair Cross Complementing 4 (XRCC4) Genetic Single Nucleotide Polymorphisms and the Liver Toxicity of AFB1 in Hepatocellular Carcinoma

Our previous reports have shown that the genetic single-nucleotide polymorphisms (GSNPs) in the DNA repair gene X-ray repair cross complementing 4 (XRCC4) are involved in the carcinogenesis of hepatocellular carcinoma (HCC) induced by aflatoxin B1 (AFB1). However, the effects of GSNPs in the coding regions of XRCC4 on hepatic toxicity of AFB1 have been less investigated. We conducted a hospital-based clinic tissue samples with pathologically diagnosed HCC (n = 380) in a high AFB1 exposure area to explore the possible roles of GSNPs in the coding regions of XRCC4 in AFB1-induced HCC using liver toxicity assays. A total of 143 GSNPs were included in the present study and genotyped using the SNaPshot method, whereas the liver toxicity of AFB1 was evaluated using AFB1-DNA adducts in the tissues with HCC. In the clinicopathological samples with HCC, the average adduct amount is 2.27 (cid:1) 1.09 μ mol/mol DNA. Among 143 GSNPs of XRCC4, only rs1237462915, rs28383151, rs762419679, rs766287987, and rs3734091 significantly increased the levels of AFB1-DNA adducts. Furthermore, XRCC4 GSNPs (including rs28383151, rs766287987, and rs3734091) also increased cumulative hazard for patients with HCC. These results suggest that the liver toxicity of AFB1 may be modified by XRCC4 GSNPs.


Introduction
Aflatoxin B1 (AFB1) is an important type I chemical toxicant mainly produced by the toxigenic strains of Aspergillus flavus (A. flavus) and Aspergillus parasiticus (A. parasiticus) [1,2]. This carcinogen is often taken into human body via contaminating human foods such as nuts and cereals and displays its toxic effects, especially hepatic toxicity [1][2][3][4][5][6][7][8]. AFB1-induced hepatic effects consist of acute toxic damages (such as severe DNA damage, severe liver degeneration and necrosis, and the failure of hepatic function) and chronic cumulative damages (such as a series of cumulative DNA damage, slight hepatocellular degeneration and necrosis, chronic inflammation, liver cirrhosis, and liver cancer) [3][4][5]. Increasing evidence has shown that under the same exposure of AFB1, some individuals feature severe hepatic damage; others have no noticeable damage [9][10][11][12][13][14]. This suggests that different individuals have different responses to the toxic effects of AFB1 and genetic factors may play a central role in the AFB1-induced hepatic toxicity.
X-ray repair cross complementing 4 (XRCC4), an important DNA repair gene involved in nonhomologous end-joining (NHEJ) repair pathway, plays a scaffold function via stabilizing and localizing DNA repair enzymes LIG IV, Ku70/80 heterodimer, and the DNA-dependent protein kinase (DNA-PK) catalytic subunit (DNA-PKcs) in the ends of DNA double-stranded breaks (DSBs) during NHEJ [15,16]. In the past decades, growing reports have exhibited that the abnormal structures and functions of XRCC4 may alter the capacity of DNA repair and ultimately result in human diseases [17][18][19][20][21][22]. Several recent studies have also shown that the genetic alterations in the coding regions of XRCC4 can modify hepatocellular carcinoma (HCC) risk and prognosis [23][24][25][26][27]. However, the effects of this genetic alteration on the hepatic toxicity of AFB1 is unclear. Here, we conducted a clinical sample study exposure to explore whether the genetic single-nucleotide polymorphisms (GSNPs, a type of genetic alterations) in the coding regions of XRCC4 modified the effects of AFB1 on hepatic damage.

Study population
This was a hospital-based molecular epidemiological study conducted in high AFB1 exposure area, Guangxi Zhuang Region, China. All participants were newly diagnosed HCC cases and recruited from the Affiliated Hospitals of Youjiang Medical University for Nationalities (located at Bose region, a major AFB1 exposure area) between January 2010 and January 2013 inclusively. The inclusive criteria of cases consisted of (a) cases with ultimately histopathologically confirmed HCC; (b) cases without any evidence of hepatitis virus infection; (c) cases with the history of AFB1 exposure which was defined according to positive history of peripheral serum AFB1-albumin adducts [5,24]; and (d) cases with available tumor tissue samples and clinicopathological data.
According to the criteria, a total of 380 cases with HCC were recruited in this study during the period. With informed consent, the tissue samples with HCC for all patients and clinicopathological data were collected. Additionally, survival follow-up information was also collected through cases themselves or their family contact. In this study, the last follow-up date was set on January 31, 2019. The protocol for clinical samples was approved by Youjiang Medical University for Nationalities Medical Ethics Committee.

The evaluation of AFB1-related hepatic toxicity
Hepatic toxicity of AFB1 was evaluated using AFB1-DNA adducts in the tissue samples with HCC, and the amounts of AFB1-DNA adducts were tested by the previously described enzyme-linked immunosorbent assay (ELISA).

GSNP selection
All GSNPs of XRCC4 gene were first screened from the SNPdatabase (http:// asia.ensembl.org/Homo_sapiens/Gene/Variation_Gene/ Table?db=core;g=  ENSG00000152422;r=5:83077498-83353787). According to the data from SNPdatabase, a total of 143 GSNPs can result in missense mutations and the change of amino acids in XRCC4 protein, and thus they were ultimately selected for final analyses.

Genotypic analyses
Genomic DNA in all tumor tissue samples with HCC was standard phenolchloroform extraction binding with proteinase K. The GSNPs of XRCC4 were genotyped using SNaPshot method (Applied Biosystems [ABI], Foster City, CA) as previously described [28]. For quality control, all laboratory personnel were blind to the status of every sample with hepatocarcinoma, and controls were also included in each analysis.

Statistical analysis
The test for genotypic distribution of XRCC4 GSNPs among HCC cases featuring different AFB1-DNA adducts was accomplished using student t-test or one-way analysis of variance (ANOVA) test. Multiple tests were adjusted using a Bonferroni correction, and the threshold for GSNP screening was defined as α = 3.53 Â 10 À4 . Kaplan-Meier survival model with log-rank test and Cox regression model (the selection of significant varies based on forward-step method with likelihood ratio test) was used to analyze the association between XRCC4 GSNPs and HCC outcomes. Cumulative hazard value for the effects of XRCC4 GSNPs on the hepatic toxicity for AFB1 and corresponding 95% confidence interval (CI) were calculated using hazard ratio (HR) from significant multivariate Cox regression model (including all significant variates). All statistical analyses were performed with SPSS statistical package (Version 18, SPSS Institute, Chicago, IL, USA).

The characteristics of subjects
All subjects suffered from hepatic carcinoma, and Table 1 summarized their characteristics. The mean age of all participants was 50.74 AE 11.55 years, and more than 70% of them are male. For these cancer patients, 70.3% (267/380) and 26.3% (100/380) cases featured TNM II and III stages of tumor, and they also had an average AFB1 exposure value of 2.27 AE 1.09 μmol/mol DNA.

XRCC4 GSNPs increased AFB1-DNA adducts
A total of 143 GSNPs in the coding regions of XRCC4 gene were selected in our final analyses, and Table 2 showed the genotypic distribution of all GSNPs. To evaluate the effects of these potential GSNPs on AFB1-DNA adducts, the role of each GSNP in the coding regions of XRCC4 gene was tested using Student t-test or ANOVA test with the adjustment of multiple test. Among these GSNPs, only rs1237462915 (cat#SNP016, at codon 38), rs28383151 (cat#SNP026, at codon 56), rs762419679 (cat#SNP069, at codon 127), rs766287987 (cat#SNP112, at codon 203), and rs3734091 (cat#SNP138, at codon 247) significantly affected the levels of AFB1-DNA adducts in the tumor tissues with HCC. The adduct amounts of their wild genotypes (defined as XX genotype) were 2.15 AE 0.97 μmol/mol DNA, 2.07 AE 0.99 μmol/mol DNA, 2.12 AE 0.86 μmol/mol DNA, 2.11 AE 0.89 μmol/mol DNA, and 2.09 AE 0.97 μmol/mol DNA, respectively. For their mutant heterozygotic genotypes (defined as XY genotype), the amounts of AFB-DNA adduct were from 2.64 to 4.33 μmol/mol DNA, whereas the adduct levels were from 3.04 to 5.78 for the mutant homozygotic genotypes (defined as YY genotype) ( Table 2).

XRCC4 GSNPs modified the AFB1-related HCC prognosis
Because the poor prognosis of patients with HCC has been associated with the toxicity of AFB1, we followed up the survival information of all patients and explored whether positive GSNPs of XRCC4 modified HCC outcomes, including overall survival (OS) and disease recurrence-free survival (RFS) (Figures 1 and 2). Results from Kaplan-Meier survival model (based on the cumulative risk models) and Cox regression model analyses showed that compared with their wild e SD is not determined. * SD for genotype yy is not determined and P-value is used for genotypes xx and xy. Table 2.
The association between SNPs in the coding region of XRCC4 and AFB1-DNA adducts in tissues with hepatocellular carcinoma. Cumulative hazard function was plotted by Kaplan-Meier methodology, and P value was calculated with twosided log-rank tests. The relative hazard ratio (HR) values for genotypes were calculated using multivariable Cox regression models (with all significant variables) based on forward-step method with likelihood ratio test.

Figure 2.
XRCC4 GSNPs significantly correlating with the disease recurrence-free survival (RFS) of hepatocellular carcinoma (HCC). Cumulative hazard function was plotted by Kaplan-Meier methodology, and P value was calculated with two-sided log-rank tests. The relative hazard ratio (HR) values for genotypes were calculated using multivariable Cox regression models (with all significant variables) based on forward-step method with likelihood ratio test.

Discussion
In this study, we investigated the association between the GSNPs in the coding regions of XRCC4 gene and the toxic effects of AFB1 on the liver. We found that five XRCC4 GSNPs, including rs1237462915 (at codon 38), rs28383151 (at codon 56), rs762419679 (at codon 127), rs766287987 (at codon 203), and rs3734091 (at codon 247), significantly increased the amount of AFB1-DNA adducts in tissues with HCC (2.07-2.15 μmol/mol DNA for XY genotypes and 2.64-4.33 μmol/mol DNA for YY genotypes, respectively) and progressed the cumulative hazard of AFB1 hepatic toxicity.
AFB1 acts as a type of human chemical toxicant, and the toxic effects of this toxicant are characterized by organophilism (mainly causing hepatic damage), genic toxicity (mainly inducing DNA damages such as hotspot mutation at codon 249 of TP53 gene, AFB1-DNA adduct formation, and so on), and carcinogenicity (mainly resulting in HCC) [6][7][8]. Among the hepatic toxicity of AFB1, the formation of AFB1-DNA adducts in hepatic cells is a key step during the metabolism of this toxicant [9][10][11][12][13][14]. Evidence from molecular epidemiological studies and clinical studies has proved that the levels of AFB1-DNA adducts in the hepatic tissues are positively associated with the levels and time of AFB1 exposure [3,24,[26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44]. This is indicative of AFB1-DNA adduct acting as the biomarker for AFB1's toxic capacity in the liver. In this study, AFB1-DNA adduct in the tumor tissues with HCC was used to evaluate hepatic toxicity related to AFB1, mainly because normal liver tissue samples cannot be obtained. Our results exhibited HCC tumor samples from high AFB1 exposure areas have an average adduct amount of 2.27 AE 1.09 μmol/mol DNA. Supporting our findings, several studies from high AFB1 exposure areas Nanning and Tiandong, China, have also shown the similar level of DNA adducts [4,5,26,27,37,39,45]. Taken together, the amount of AFB1-DNA adducts should be able to reflect the hepatic toxic potential of AFB1.
XRCC4, a key gene in the V(D)J recombination repair pathway, is located at 5q14.2 and consists of 13 exons (PubMed). Normally, XRCC4 is mainly expressed in genital meatus, alimentary tract, and lymphoid tissue; however, its expression will noticeably increase in other tissues such as the skin and liver under the condition of in vitro and in vivo injuries. This gene's encoding protein plays a vital role in both NHEJ and the completion of V(D)J recombination via acting as a scaffold protein for DNA ligase IV and DNA-PK in the repair of DNA DSBs [15,19]. Mutations in XRCC4, including GSNPs and other non-GSNPs variants, can cause endocrine dysfunction, microcephaly, short stature, and diseases [16,21]. With the development of human Geno projects, more than 1000 GSNPs are identified. Among these GSNPs, we focused on genetic alterations in the coding regions of XRCC4, mainly because they will result in missense mutations and ultimately cause the structure damage and function deficiency of XRCC4 protein. Molecular epidemiological studies have displayed that the GSNPs in the XRCC4 genes can increase DNA repair capacity and increase the risk of some tumors such as lung cancer, colon cancer, HCC, and so on [21,[46][47][48][49][50][51]. Evidence from in vitro and in vivo studies has also proved that XRCC4 GSNPs increase the amount of DNA damage and induce more gene mutations [23,24,26,27]. In our study, we tested the genotypic distributions of all known GSNPs in the coding region of XRCC4 in liver tumor tissues. Five positive GSNPs were identified, and they result in the change of amino acid D to Y at codon 38 for rs1237462915, A to T at codon 56 for rs28383151, I to T at codon 127 for rs762419679, Q to H at codon 203 for rs766287987, and A to S at codon 247 for rs3734091, respectively. Although evidence that several other GSNPs, including rs761695470, rs758779099, rs144653114, rs1277864722, and rs777195630, increased the amounts of AFB1-DNA adducts was not statistically significant according to our defined threshold value, their effects should not be neglected because small-size samples may underestimate values.
Because the toxic effects of AFB1 also modify the prognosis of patients with HCC [26,27,33,52,53], we accomplished patients' survival analyses on the basis of the cumulative risk models and found only rs28383151, rs766287987, and rs3734091 polymorphisms shortened HCC cases' OS and RFS. Supporting our findings, several previous reports have proved that XRCC4 GSNPs can alter the levels of XRCC4 mRNA and protein expression and dysregulation of XRCC4 expression increasing the amount of AFB1-DNA adducts and mutative risk of TP53 gene [23,24,26,27].
To conclude, this study is the first report investigating the modified function of XRCC4 GSNPs on AFB1's hepatic toxicity. Our findings suggest that the GSNPs in the coding regions of XRCC4 gene, like rs1237462915, rs28383151, rs762419679, rs766287987, and rs3734091, may alter the DNA repair capacity of DNA damage induced by AFB1. If these individuals with mutant genotypes of these GSNPs decrease their exposure to AFB1, they will be free from toxic effects of AFB1 on hepatic damage. Several limitations should be focused for our study. First, relatively small-size samples may underestimate the effects of XRCC4 GSNPs on AFB1 hepatic toxicity. Second, the hospital-based design may result in selective bias. Third, we only accomplished the cumulative risk analyses but not the cumulative survival analyses. Finally, we did not finish functional and mechanical analyses. Thus, XRCC4 GSNPs may be valuable biomarkers for predicting the toxic effects of AFB1 on the liver once the present findings were proved by larger samples and toxic function analyses.