Introduction

Colorectal cancer (CRC) is the third most common gastrointestinal tract malignancy and the fourth leading cause of cancer mortality worldwide. Nowadays, CRC is becoming one of major challenge to public health in China because of lifestyle changes and the process of westernization [1]. Accumulating evidence supports that genetics play an important role in the etiology of CRC [2]. CRCs frequently occur in some hereditary cancer syndromes with germline mutations of DNA repair genes; however, genetic factors responsible for sporadic CRCs have not been largely determined.

DNA repair has an essential role in protecting the genome from damage by endogenous and environmental agents [3]. Further, repair of DNA damage is a complex process in maintaining genome integrity, which is conducted by a series of DNA repair pathways, consisting of more than 130 genes. Nucleotide excision repair (NER) is one of major pathway, which deal with the UV radiation or chemical agents induced DNA damage, is responsible for removal of a wide variety of structural unrelated bulky DNA adducts and helix-distorting lesions [4, 5]. Previous studies provided evidence that those polymorphisms of the NER pathway genes may alter the DNA repair capacity and thus could play an important role in carcinogenesis or response to chemotherapy [6–9].

Xeroderma pigmentosum group G (XPG) is one of the NER genes and responsible for 1,186 amino acid structure-specific endonuclease activity, thereby playing a key role in NER of helix-distorting DNA damage [5]. Previous studies investigated the role of common polymorphism Asp1104His (G-to-C transition in 13q33 exon15, rs17655) of XPG in the etiology of various cancers [10–13]. Only two studies demonstrated the association of Asp1104His polymorphism of XPG with CRC risk. Liu et al. observed that XPG Asp1104His (GC or CC genotype) had an increased risk for CRC [14]. However, another study did not found any association of Asp1104His with CRC risk and found no significant results either [15]. Taken together, the two results remain conflicting rather than conclusive and need further investigations for solving these discrepancies.

Xeroderma pigmentosum group C (XPC) is located at chromosome 3p25, which is also one of the eight core genes in the NER pathway, particular plays an important role in the early steps (damage recognition, open complex formation and reparation) of genome NER [5]. Recent studies have showed that polymorphisms of XPC gene may alter the DNA repair capacity (DRC) and modulate the susceptibility to cancer. The Ala499Val (C/T) in exon 9 of XPC gene has been previously identified in several tumors. However, very little information is available about the association of Ala499Val polymorphism with CRC risk. Only Wu et al. reported that individuals carrying the Ala499Val CT + TT genotypes showed a significantly decreased risk of rectal cancer in a small size [16].

The DNA repair gene WRN is located at chromosome 8p11-12. WRN protein is an important member of the RecQ helicase family, which is involved in multiple DNA repair pathways, protecting the genome from incorrect recombination during mitosis and maintaining its stability. Several studies have shown a relationship between WRN expression and malignancy and have indicated that epigenetic inactivation of WRN is of importance in carcinogenesis [17–20]. Since mutations in the WRN gene can accelerate aging, it has been reasoned that WRN polymorphisms may also be associated with age-related pathologies, including the cancer susceptibility. Cys1367Arg (rs1346044), one of the common SNPs in WRN gene, usually manifests T to C variation and leads to a protein variation of 1367 Cys to Arg. In previous studies, WRN Cys1367Arg had been found to be associated with many diseases, including chronic kidney disease, glioma, and familial breast cancer [21–23]. However, the association between the Cys1367Arg and CRC susceptibility and prognosis has not been explored yet.

Therefore, in the present study, we analyzed the common genetic polymorphisms in the DNA repair genes XPG, XPC, and WRN in a large sample case–control study from 890 CRCs and 910 healthy controls, and attempted to elucidate the association between these polymorphisms and CRC susceptibility, clinicopathological features, and clinical outcomes with oxaliplatin-based chemotherapy.

Patients and methods

Subjects

This study included 890 patients with CRC who were admitted to the General Hospital of CNPC in Jilin between 2006 and 2012. Approximately, 89 % of contacted patients consented to enrollment in the study. Finally, 890 patients were included in the study. The principal clinical characteristics were obtained from the interviewer-administered health risk questionnaires and medical records. Tumor differentiation or pathological grade of these CRC patients was performed according to the World Health Organization criteria and DUKE’s criteria, respectively. The patients received several chemotherapy regimens, including oxaliplatin-based chemotherapy (Folfox regimen: oxaliplatin, leucovorin plus 5-FU; Xelox regimen: oxaliplatin plus capecitabine), LV5–FU2 (leucovorin plus 5-FU), or FU alone (fluorouracil).

We also included 910 unrelated age- and gender–matched healthy controls. The population had no known medical illness or hereditary disorders and was not taking any medications. Before its commencement, this study was approved by the Research Ethics Committee of the First Affiliated Hospital of Dalian Medical University, and informed consent was obtained from each participant.

Genotyping assay

Genomic DNA was isolated from a leukocyte cell pellet of each blood sample, using the TaKaRa DNA Blood Mini Kit (TaKaRa Biotechnology (Dalian) Co., Ltd, Dalian, China). The XPG, XPC, and WRN Cys1367Arg genetic polymorphisms were genotyped by PCR–RFLP assay. PCR was performed using 50∼100 ng of genomic DNA, 300 nM of each primer, 200 nM dNTPs, and 0.5 U Taq polymerase in PCR buffer (TaKaRa Biotechnology (Dalian) Co., Ltd, Dalian, China) in a total volume of 25 μl. For XPG Arg1104His, the 271-base pair (bp) PCR products were digested with Hsp92II overnight at 37 °C. The T allele was uncut, and the polymorphic allele (C) produced a 227-bp and 44-bp fragment. The PCR products of XPC Ala499Val were digested with SacII (Promega Corporation, U.S.A.) at 37 °C overnight. The T allele was uncut, and the C allele had a SacII restriction site, 2 bands were generated (131-bp and 21-bp fragment), while the T allele had a single fragment with a size of 152-bp. For WRN Cys1367Arg, the 193-bp PCR products were digested with HSP92II (Promega Corporation, U.S.A.) at 37 °C overnight. The C allele was cut into 118-bp, 53-bp, and the T allele was cut into 90-bp and 53-bp bands. Samples were coded for case–control status, and at least 10 % of the samples were randomly selected and subjected to repeat analysis as quality control for verification of genotyping procedures. Two researchers independently performed RFLP and reviewed all genotyping results.

Statistical analysis

The SPSS software package version 16.0 (SPSS Inc, Chicago, USA) was used for statistical analyses. The association between the distributions of demographic, epidemiologic, and clinical parameters was assessed by using the chi-square (Pearson χ 2 test) or Fisher exact test. The associations between genotypes and CRC risk were calculated by odd ratios (OR) and 95 % confidence interval (CI). The population genetic analysis program SNPAlyze 2.2 (Dynacom Co. Ltd, Yokohama, Japan) was used for linkage disequilibrium (LD) analysis, and the Hardy–Weinberg equilibrium test. The disease-free survival (DFS) was defined as the time between the diagnosis and an occurrence of relapse, death, or last known follow-up. The overall survival (OS) was defined as the time between the diagnosis and death, or last known follow-up. The DFS or OS curves were plotted by using the Kaplan–Meier method, and the statistical differences in survival among groups were compared by log-rank test. The independent prognostic values of different polymorphism in XPG, XPC, and WRN gene and their association with survival time of CRC patients were analyzed by univariate or multivariate Cox hazards regression model. The Cox hazard ratio and its 95 % CI were obtained accordingly. All statistical significance was set at P < 0.05, and all tests were two-sided.

Results

The baseline characteristics and clinical features of 890 CRC patients and 910 controls are summarized in Table 1. There were no significant differences in the distributions of gender and age between CRC patients and controls (P = 0.646 and P = 0.625, respectively). The mean age at diagnosis (SD) was 58.2(10.3) and at the range of 23∼79 years old. All the cases included in this study, 57.2 % of CRC patients were male, and 84.6 % of them were in the status of ever smoking. Among 890 patients, the vast majority patients were in grade 2 (G2, moderate, 82.4 %), while only 8.7 % of them were in grade 1 (G1, well). Most of CRC patients (80.7 %) received oxaliplatin-based chemotherapy, whereas 11.3 % of them underwent fluorouracil chemotherapy.

Table 1 Clinical variables of the CRC patients in the study (N = 890)
Full size table

We carried out a standard allelic association analysis and genotype distribution in XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms between CRC patients and controls, outlined in Table 2. The observed genotype frequency among individuals in control group fit well to Hardy–Weinberg equilibrium (P = 0.498 for XPG Arg1104His, P = 0.067 for XPC Ala499Val, P = 0.630 for WRN Cys1367Arg). A significant increased in the frequency of CC genotype in XPC Ala499Val polymorphism was observed [CC vs TT, P < 0.001; adjusted OR (95 % CI) = 2.194 (1.598–3.011)]. Moreover, there was a significant difference in the CT/CC genotypes or C allele of the XPC Ala499Val between CRC patients and controls and appeared to be associated with an 1.212-, 1.294-fold increased risk of CRC [P = 0.049, adjusted OR (95 % CI) = 1.212 (0.966–1.476); P < 0.001, adjusted OR (95 % CI) = 1.294 (1.132–1.479), respectively] in recessive model. For WRN Cys1367Arg polymorphism, the heterozygous variant CT genotype or homozygous variant TT genotype showed a significantly increased risk of CRC in comparison to the CC genotype [CT vs CC, P < 0.001; adjusted OR (95 % CI) = 3.371 (1.752–6.487); TT vs CC, P < 0.001; adjusted OR (95 % CI) = 3.131 (1.663–5.895)]. Additionally, compared with the corresponding homozygous CC genotypes, the variant genotypes CT/TT of WRN Cys1367Arg had a significant increased risk of developing CRC after adjusted for age and gender in the recessive model [P < 0.001; adjusted OR (95 % CI) = 3.183(1.694–5.981)]. No significant difference was detected for CRC risk and WRN Cys1367Arg polymorphism in the dominant model. However, there were no significant associations observed between XPG Arg1104His polymorphism and the risk of CRC (Table 2).

Table 2 Frequency distribution of XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms and the associations with CRC risk
Full size table

To further test our hypothesis that multiple SNPs in the DNA repair genes may have a joint effect on CRC risk, we estimated the combined effect based on two different SNPs. In this study, XPG Arg1104His C allele, XPC Ala499Val C allele, and WRN Cys1367Arg T allele were considered as favorable high-risk alleles. Table 3 showed the risk of CRC increased significantly with the number of putative risk genotypes for the combined XPG Arg1104His and XPC Ala499Val [two risk genotypes: P = 0.016; adjusted OR (95 % CI) = 1.523 (1.079–2.512); >2 risk genotype: P = 0.005, adjusted OR (95 % CI) = 1.718 (1.176–2.512), respectively] and adjusted OR = 1.380, 95 % CI = 1.046–1.821, P = 0.022 for combined XPC Ala499Val and WRN Cys1367Arg. However, when XPG Arg1104His and WRN Cys1367Arg were considered together, no apparent increase in CRC risk was observed as the number of risk genotypes increased, as illustrated in Table 3.

Table 3 The combined effect analysis between DNA repair genes polymorphisms and CRC susceptibility
Full size table

We further investigated the correlation of XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms with clinical variables of 890 CRC patients (Table 4). We only observed that the distribution frequency of the XPG Arg1104His polymorphism was associated with tumor differentiation (grade 1, 2, 3) of patients. The frequency (32.5 %) of XPG Arg1104His GG genotype in the patients with grade 3 was significantly higher than that (16.9 %) in patients with grade 1 (P = 0.043) (Table 4). However, no significant correlation of genotype distributions of XPC Ala499Val, and WRN Cys1367Arg polymorphisms were found with the clinicopathological characteristics.

Table 4 Association between XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms and clinicopathological features in CRC patients
Full size table

We further evaluated the correlation of XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms with the prognosis of CRC patients with oxaliplatin-based chemotherapy (N = 718).

We found that XPG Arg1104His polymorphism have a significant impact on DFS (log-rank test: P < 0.001; Fig. 1, A1) but not OS for CRC patients. The estimated median DFS time for XPG Arg1104His CG or CC genotype carriers were longer [136 months (127–142), 85 months (75–95), respectively] compared with the GG genotype carriers [44 months (40–48)]. The multivariate Cox regression analysis also established XPG Arg1104His as an independent prognostic factor [CG genotype, adjusted HR (95 % CI) = 0.163 (0.107–0.248), P < 0.001; CC genotype, adjusted HR (95 % CI) = 0.333 (0.235–0.470), P < 0.001; CG/CC genotype, adjusted HR (95 % CI) = 0.333 (0.235–0.470)]. Furthermore, the XPC Ala499Val polymorphism was associated with a longer DFS (log-rank test, P = 0.018; Fig. 1 B1). The estimated median PFS was 83 months for patients with CT genotype (95 % CI = 64–101) and 59 months for patients with TT genotype (95 % CI = 51–77). Moreover, the multivariate COX regression analysis also verified that XPC Ala499Val CT genotypes showed a better survival [CC genotype, adjusted HR (95 % CI) = 0.691 (0.528–0.904), P = 0.007; CT/CC genotype, adjusted HR (95 % CI) = 0.602 (0.389–0.934), P = 0.024], outlined in Table 5.

Fig. 1
figure 1

Kaplan–Meier survival curves illustrating the DFS and OS in CRC patients with XPG Arg1104His, XPC Ala499Val, and WRN Cys1367Arg polymorphisms after postoperative oxaliplatin-based chemotherapy. A1, B1, C1 disease-free survival (DFS). A2, B2, C2 overall survival (OS)

Full size image
Table 5 Cox regression analysis of potential factors for DFS and OS in CRC patients after oxaliplatin-based chemotherapy
Full size table

In addition, no correlation of WRN Cys1367Arg polymorphism was found with DFS (log-rank test, P = 0.437; Fig. 1 C1) along with a tendency toward longer OS (log-rank test, P = 0.059; Fig. 1, C2) for patients with postoperative oxaliplatin-based chemotherapy.

Discussion

DNA repair has an essential role in protecting the genome from damage by endogenous and environmental agents. Multiple genetic alterations contribute to CRC susceptibility. Further, increasing studies provided evidence that epigenetic modifications such as germline polymorphisms in DNA repair pathway genes can either change protein coding or alter the levels of transcription or translation, thereby reducing DNA repair capacity and inducing genetic instability or carcinogenesis [24]. In present molecular epidemiological case–control study, we sought to identify genetic polymorphisms of DNA repair genes XPG, XPC, and WRN that confer individual susceptibility to CRC and as predictors of response of patients to chemotherapy and survival.

Our results obtained by analyzing 890 CRC patients and 910 controls, and a significant association between XPC Ala499Val polymorphism and CRC risk were observed. To date, there have been only three studies exploring relationship between the XPC Ala499Val variation and CRC risk [16, 25, 26].Similarly, Wu et al. found that Ala499Val variant individuals were at an increased risk for developing rectal cancer in a Chinese population, but in relative small samples. In contrast, the TT genotype carriers showed case-cohort studies, no association was found between Ala499Val genotypes and CRC risk. Previously, the associations between Ala499Val and risk of other variety of cancers have been studied, such as nasopharyngeal cancer, lung cancer, bladder cancer, and squamous cell cancer [27–30]. However, those findings were also inconsistent. For example, the individuals with heterozygous CT genotype or homozygous variant TT genotype had a lower risk for oral premalignant lesions cancer and gastric cancer [31, 32]. In contrast, the TT genotype showed an increased risk of nasopharyngeal cancer, lung cancer, bladder cancer, and squamous cell cancer. Thus, it might be stated that the role of XPC Ala499Val polymorphisms in cancer risk may vary with different diseases or race/ethnicity, and sample size, a possibility that warrants further investigation. Further functional assays for the expressional influence of the polymorphism will be needed.

Regarding XPG Arg1104His polymorphism, we found no association between the remaining variants of XPG and CRC risk. Up to now, only two reports have identified the correlation between CRC risk and the investigated Arg1104His genotypes. Nevertheless, consistent with our finding, Gil et al. [15] found no association between XPG Arg1104His and CRC susceptibility. In another case-cohort study, an increased CRC risk among individuals with the variant allele of Arg1104His was observed [14]. Taken together, these data suggests that further work is needed to extend these findings by carrying out extended haplotype analyses or related genes and to replicate the observations in other studies.

To the best of our knowledge, this is the first study to examine the polymorphism WRN Cys1367Arg and CRC risk and prognosis in a relatively large sample of Chinese population. Previous studies examined the role of WRN Cys1367Arg polymorphism in modulating cancer susceptibility, except CRC. Our findings suggest that WRN Cys1367Arg heterozygous variant or homozygous variant carriers were at a significantly increased risk of CRC, which was consistent with previous study on breast cancer [23] and chronic kidney disease (CKD) [21]. Wirtenberger et al. revealed a significant association of the WRN Cys1367Arg polymorphism with high risk of familial breast cancer on German familial breast cancer patients [23]. Moreover, Yoshida et al. found that T→C (Cys1367Arg) polymorphism of WRN was significantly associated with the prevalence of chronic kidney disease (CKD) in a Japanese population [21]. While inconsistent with our results, Nakayama et al. found that WRN Cys1367Arg SNP was to be a protective factor of bone and soft tissue sarcomas (BSTSs) in Japan [18]. In another Brazilian population-based case control study, found that WRN Cys1367Arg SNP was not involved either in susceptibility to developing gliomas or in patient survival [22].

Although increasing studies investigated the relevance between the polymorphisms of DNA repair pathway genes and CRC risk or prognosis, the roles of these genes in carcinogenesis remain complex and need to be further elucidated. Oxaliplatin (the third-generation platinum derivative compound) is one of the effective chemotherapy for CRC, and the genetic polymorphisms involved in the DNA repair pathway plays an important role on changing the efficacy of chemotherapy. In the present pharmacogenetic study, we further assessed the association between XPG, XPC, and WRN polymorphisms and prognosis of CRC patients receiving oxaliplatin-based chemotherapy. Our results suggest that XPG Arg1104His variant carriers had a longer DFS in comparison to the CC genotype carriers in the CRC patients receiving oxaliplatin-based chemotherapy. Similarly, Liu et al. found a significantly longer progression-free survival (PFS) after oxaliplatin-based chemotherapy in patients with Asp1104His polymorphism. Moreover, the multivariate COX regression analysis also verified that carrying XPC Ala499Val CT genotype has a significant impact on prognosis. Taken together, our study provides significant information on prognostic value of XPG and XPC for customized chemotherapy to improve oxaliplatin treatment efficacy. Nevertheless, this study was performed only in Chinese population, and need to be replicated in other race/ethnicities or geographic areas.

In conclusion, this case–control study indicates that genetic polymorphisms of XPC Ala499Val, WRN Cys1367Arg were alone and combined significantly associated with the CRC risk. Furthermore, our study identified that XPC Ala499Val, XPG Arg1104His genetic variation had a significant impact on survival in CRC patients with oxaliplatin-based chemotherapy, which indicate that these genetic polymorphisms could be used as candidate molecular prognostic and predictive markers of CRC patients toward individualizing treatment strategies.