Introduction

Gastric cancer is a major health problem worldwide due to its prevalence, poor prognosis and limited treatment options. It is well established that both environmental and genetic factors involved in the development of gastric cancer [1, 2].

DNA double-strand breaks (DSBs) occur spontaneously during the cell cycle and are induced by a variety of exogenous agents. DSBs are major lesions that destroy the integrity of the DNA molecule. To combat this potentially lethal damage, two related repair pathways, namely homologous recombination (HR) and non-homologous DNA end joining (NHEJ), have been evolved. The HR is a template guided and error-free pathway. In most cases, the NHEJ pathway results in the loss of a few nucleotides at the broken ends, making this pathway error-prone. Unrepaired DSBs may result in genetic instability and ultimately may enhance the rate of cancer development. It have been reported that NHEJ deficiencies can lead to increased genomic instability and cause increased tumorigenesis [3–8]. Ku70 and Ku80 form a Ku complex which is a heterodimeric DNA binding complex involved in the repair of DSBs as a member of the NHEJ pathway. Ku80 is encoded by the X-ray repair cross-complementing group 5 gene (XRCC5; OMIM: 194364) [9]. It has been reported that mouse cells deficient for Ku80 display a marked increase in chromosomal aberrations, including breakage, translocations, and aneuploidy [4]. It has been well shown that the X-ray repair cross-complementing group 4 gene (XRCC4; OMIM: 194363) plays important role in repair of DSBs [10]. Deficiency of XRCC4 in primary murine cells causes growth defects, premature senescence, IR sensitivity, and inability to support V(D)J recombination have been reported [11].

Several genetic polymorphisms in the XRCC4 and XRCC5 have been reported in human. An Insertion/Deletion (I/D) of a 30 bp sequence in intron 3 of the XRCC4 gene (rs28360071) has been reported. Associations between the I/D polymorphism of XRCC4 and susceptibility to several types of cancers, including childhood leukemia, prostate, oral and colorectal cancers were reported [12–15]. A variable number of tandem repeats of a 21 bp (VNTR) polymorphism at the promoter region of the XRCC5 gene have been reported [9]. This polymorphism has four alleles: 3R, 2R, 1R and 0R [16]. It is established that the mRNA level of XRCC5 decreased as function of number of tandem repeats at the promoter region of the gene [17]. Considering that environmental factors induced DNA damage and on the other hand, functional polymorphisms in the DNA repair genes may alter the cellular DNA repair capacity; therefore, study of association between genetic variations in DNA repair genes and risk of gastric cancer is potentially important. It should be noted that the over expression of the XRCC5 in several types of cancers (such as gastric and colorectal cancers) has been reported [18–20]. Take together, it may be hypothesized that the 3R and 2R (High repeated) alleles compared to 0R and 1R (Low repeated) alleles alter the risk of development of cancers. There is no so much data on associations between polymorphisms of the XRCC5 and susceptibility to cancers [21–26]. However, there is no study investigating the association between VNTR polymorphism at the promoter region of the XRCC5 and I/D polymorphism of XRCC4 and susceptibility to gastric cancer. Therefore, the present case-control study was carried out.

Materials and Methods

Subjects

The present case-control study was consisted of 159 (56 females, 103 males) with pathologically confirmed primary gastric cancer that were recruited from chemotherapy department of Nemazi hospital in Shiraz (Fars province, south of Iran). Age and gender frequency-matched controls were randomly selected from the healthy blood donors. A total of 242 (75 females, 167 males) healthy controls was included in the study. Exclusion criteria for controls included any previous cancer history and diagnosed psychiatric diseases. The mean age (SD; Min-Max) of the patients and the controls were 57.3 (12.9; 24–85) and 56.7 (9.8; 31–82) years, respectively. There was no significant difference between cases and controls for age of subjects (P > 0.05). The Iranian population is one of the most heterogeneous populations [27, 28]. Therefore, we selected our patients and controls from Persian (Caucasian) Muslims living in Fars province (southern Iran). Information on family history of cancer was collected for all participants via in-person interviews. A person with at least one first-degree relative with diagnosed cancer is considered to have a positive family history. Informed consent was obtained from each subject before the study. Ethical approval for the current study was obtained from Shiraz University institutional review board.

DNA Extraction and Genotyping Analysis

Blood samples were collected from the subjects. Genomic DNA was extracted from whole blood samples. Genotypic analysis for the I/D polymorphism of the XRCC4 was determined by PCR assay, as described previously [12]. PCR amplification was performed using following primers: forward primer 5’-TCC TGT TAC CAT TTC AGT GTT AT-3’ and reverse primer 5’-CAC CTG TGT TCA ATT CCA GCT T-3’. The genotypic analysis for the VNTR polymorphism of XRCC5 was determined by PCR assay, as described previously [29]. PCR amplification was performed using following primers: forward primer 5’-AGG CGG CTC AAA CAC CAC AC-3’ and reverse primer 5’-CAA GCG GCA GAT AGC GGA AAG-3’. The VNTR alleles divided into two groups: L (means low repeated alleles; 0 and 1 repeat) and H (means high repeated alleles; 2 and 3 repeats) alleles. For quality control, 15 % of randomly selected samples were repeated to verify genotyping results and 100 % concordance was found.

Statistical Analysis

A Chi-square test was performed for the XRCC5 VNTR and XRCC4 I/D polymorphisms to determine if the control group demonstrated Hardy-Weinberg equilibrium. The associations between the genotypes of the study polymorphisms and gastric cancer risk were assessed by calculating odds ratios (ORs) and 95 % confidence intervals (CIs). In the analysis, the persons with LL and II genotypes assumed as reference groups, for VNTR XRCC5 and I/D XRCC4 polymorphisms, respectively.

Positive family history of cancer in first-degree relatives is one of the strongest risk factor [1, 2, 30–32]. Therefore, the participants were stratified by their family history of cancers (negative and positive) and the data were reanalyzed. For these analyses, persons who have negative family history and the reference genotypes (as mentioned above) assumed as reference groups.

Data on family history in the control subjects were missed for some participants. In order to study the potential effect of the family history on gastric cancer risk as well as the risk associated with genotypes of the study polymorphisms, the “sensitivity analysis” was used. For this analysis we tested two assumptions for the missing data of the family history in the control group: all the missed data had negative family history; and alternatively, 25 % of them had positive (and 75 % had negative) family history.

Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS Inc., Chicago, IL, USA) (version 11.5). A probability of P < 0.05 was considered statistically significant. All statistical tests are two-sided.

Results and Discussion

Detailed genotype distributions among cases and controls are summarized in Table 1. The genotypic frequencies of the VNTR XRCC5 (χ2 = 4.90, df = 6, P = 0.559) and I/D XRCC4 (χ2 = 0.032, df = 1, P = 0.857) polymorphisms did not show significant deviation from Hardy-Weinberg equilibrium in control subjects.

Table 1 Genotypic prevalence of the VNTR polymorphism at promoter region of the XRCC5 among gastric cancer patients and healthy control subjects
Full size table

Table 2 shows the association between the XRCC4 polymorphism and susceptibility to gastric cancer. For I/D polymorphism of XRCC4, although there was no significant association between the ID genotype and the risk of gastric cancer (OR = 1.34, 95 % CI: 0.84–2.15, P = 0.215), the DD genotype increased the risk of gastric cancer (OR = 1.85, 95 % CI: 1.04–3.29, P = 0.034) (Table 2). The risk of gastric cancer increased as a function of number of the D allele (χ2 = 4.54, P = 0.033). Table 2, also shows the association between the VNTR XRCC5 polymorphism and susceptibility to gastric cancer. Neither the LH (OR = 1.03, 95 % CI: 0.66–1.62, P = 0.882) nor the HH (OR = 0.90, 95 % CI: 0.50–1.61, P = 0.735) genotypes altered the risk of gastric cancer in comparison with the LL genotype (Table 2).

Table 2 Associations between the I/D XRCC4 and VNTR XRCC5 polymorphisms and risk of gastric cancer
Full size table

Several studies revealed that positive family history of cancer in first-degree relatives is one of the strongest risk factor for cancers, including gastric cancer [30–32]. The prevalence of positive family history among controls and cases were 14.0 and 28.4 %, respectively. Therefore there was significant association between family history and risk of gastric cancer (OR = 2.43, 95 % CI: 1.41–4.20, P = 0.001). We further analyzed to see if the family history of cancer influenced the association of the XRCC4 I/D polymorphism and risk of gastric cancer. After stratification of the subjects according to their family history of cancer, either the ID (OR = 3.19, 95 % CI: 1.35–7.50, P = 0.008) or the DD genotypes (OR = 4.62, 95 % CI: 1.63–13.0, P = 0.004) among positive family history persons, increased the risk of gastric cancer compared with the reference group. It should be noted that there was no association between the genotypes of I/D polymorphism and risk of gastric cancer among negative family history persons (Table 2).

For the VNTR XRCC5 polymorphism, and after stratification of the subjects according to their family history of cancer, the LH + HH genotypes among positive family history persons, increased the risk of gastric cancer compared with the reference group (OR = 2.88, 95 % CI: 1.34–6.18, P = 0.006). It should be noted that there was no association between the genotypes of VNTR polymorphism and risk of gastric cancer among negative family history persons (Table 2).

The results of the present study have some limitations. Data on the family history were missing for 64 participants in control group. Using “sensitive analysis” it is possible to estimate the potential effect of this variable on the study by assuming various degrees of maldistribution of the variable in the control group and seeing how it would affect the results. As mentioned in “Statistical analysis” section, we tested two distributions for the missing data among controls. Table 3 shows the statistical analysis under our two assumptions. Therefore the present case-control study suggests that the study polymorphisms of the XRCC4 and XRCC5 genes might be risk factors for gastric cancer development among persons with positive family for cancer. Also we found that these associations are not false findings due to maldistribution of missing data.

Table 3 Association between genetic polymorphisms of I/D XRCC4 and VNTR XRCC5 and risk of gastric cancer under two assumptions for missing data in control group
Full size table

The similar association between other polymorphism of XRCC5 (C74468A) and alter gastric and esophageal cancers were reported [23]. Dong et al. [23] found that in subjects with a familial history of gastric cancer, the C allele of XRCC5 C74468A seemed to be a protective factor for the incidence. A similar trend was found in the case of esophageal cancer [23].

The influence of 2R and 3R alleles on suppressed expression of the XRCC5 [17] might confer risk to DNA repair leading to genomic instability and gastric cancer. Hence, the VNTR polymorphism in the promoter region of XRCC5 gene could serve as an important prognostic marker in gastric cancer development among persons who have positive family history.

It is also important to keep a number of limitations in mind in considering the present findings. It should be mentioned that the small sample size and thus the limited statistical power to detect differences between the case and control groups, especially when multiple testing was done, is the main limitation of our study. In future, our findings should be confirmed by large-scale studies.