Selected Polymorphisms of Base Excision Repair Genes and Pancreatic Cancer Risk in Japanese

Background Although several reports have described a possible association between DNA repair genes and pancreatic cancer (PC) in smokers, this association has not been fully evaluated in an Asian population. We assessed the impact of genetic polymorphisms in the base excision repair (BER) pathway on PC risk among Japanese. Methods This case-control study compared the frequency of 5 single-nucleotide polymorphisms (SNPs) of BER genes, namely rs1052133 in OGG1, rs1799782 and rs25487 in XRCC1, rs1130409 in APE1, and rs1136410 in PARP1. SNPs were investigated using the TaqMan assay in 185 PC cases and 1465 controls. Associations of PC risk with genetic polymorphisms and gene–environment interaction were examined with an unconditional logistic regression model. Exposure to risk factors was assessed from the results of a self-administered questionnaire. We also performed haplotype-based analysis. Results We observed that the minor allele of rs25487 in XRCC1 was significantly associated with PC risk in the per-allele model (odds ratio = 1.29, CI = 1.01–1.65; trend P = 0.043). Haplotype analysis of XRCC1 also showed a statistically significant association with PC risk. No statistically significant interaction between XRCC1 polymorphisms and smoking status was seen. Conclusions Our findings suggest that XRCC1 polymorphisms affect PC risk in Japanese.


INTRODUCTION
][3][4][5] Because early detection and curative treatment of PC are very difficult, the 5-year survival rate is only 5.5%. 6This suggests that epidemiologic approaches to identifying PC high-risk groups have an important role in decreasing the number of PC deaths.
[9][10] A positive association between PC risk and a family history of PC has also been reported, suggesting the possible involvement of genetic factors in PC incidence. 5,10,112][23] With regard to the association between PC risk and DNA repair genes, genetic polymorphisms in the BER pathway are the most extensively investigated in molecular epidemiologic studies. 14However, most studies of the association between PC risk and BER gene polymorphisms have been conducted in white populations.
We conducted a case-control study in a Japanese population to evaluate the impact of genetic polymorphisms in BER pathways on PC risk among Asians.The study focused on 4 key genes in this repair pathway: X-ray repair crosscomplementing group 1 (XRCC1), apurinic/apyrimidinic endonuclease (APE1), 8-oxoguanine DNA glycosylase (OGG1), and poly(ADP-ribose) polymerase 1 (PARP1).

Study subjects
The case subjects were 185 PC patients with no prior history of cancer who were diagnosed at Aichi Cancer Center Hospital (ACCH), Nagoya, Japan, between January 2001 and November 2005.The control subjects comprised 1465 randomly selected non-cancer ACCH outpatients, during the same period, who had no history of any cancer.All subjects were enrolled in the Hospital-based Epidemiological Research Program II at ACCH (HERPACC-II) at the time of their first visit to ACCH.The framework of HERPACC-II has been described elsewhere. 24,25Briefly, all first-visit ACCH outpatients aged 20 to 79 years were asked to complete a self-administered questionnaire on their lifestyle before development of the presenting symptoms.Outpatients were also asked to provide a 7-mL blood sample.Approximately 95% of eligible subjects completed the questionnaire, and 50% provided blood samples.All data were loaded into the HERPACC database, which is periodically synchronized with the hospital cancer registry system.Approximately 35% of subjects were diagnosed with cancer within a year of the first visit.In the present study, we defined the case population as patients who received a diagnosis of PC within 1 year of the first visit, ie, we used the window period from first visit to final diagnosis of PC rather than prospectively identifying cases.A total of 75.7% of PC cases were histologically confirmed, among which 92.1% were diagnosed as having ductal adenocarcinoma.Our previous study showed that the lifestyle patterns of firstvisit outpatients corresponded with those of individuals randomly selected from Nagoya's general population, which confirms the external validity of the study. 26The present study was approved by the Ethics Committee of Aichi Cancer Center, and informed consent was obtained at first visit from all participants.

Assessment of exposure
Exposure to potential PC risk factors was assessed from responses to the self-administered questionnaire, which was completed before diagnosis during the first visit to ACCH and reviewed by trained interviewers.Subjects were specifically questioned about their lifestyle before the onset of presenting symptoms.Daily alcohol consumption in grams was determined by summing the amount of pure alcohol in the average daily consumption of Japanese sake (rice wine), shochu (distilled spirit), beer, wine, and whiskey.Cumulative smoking exposure was measured in pack-years (PYs), ie, the product of the average number of packs per day and the number of years of smoking.Height and body weight at baseline and weight at age 20 years were self-reported.Current body mass index (BMI) and BMI at age 20 were calculated by dividing the weight in kilograms by the height in meters squared.Family history of PC was considered positive when at least 1 parent or sibling had a history of PC.

Statistical analysis
All statistical analyses were performed using Stata version 10 (Stata Corp., College Station, TX, USA).A P-value less than 0.05 was considered statistically significant.Differences in characteristics between cases and controls were assessed using the chi-square test.Odds ratios (ORs) and 95% CIs were estimated using an unconditional logistic regression model adjusted for potential confounders.Potential confounders considered in multivariate analysis were age, sex, PYs of smoking (<5, ≥5 but <20, ≥20 but <40, or ≥40), drinking habit (nondrinker, <23, ≥23 but <46, or ≥46 g/day), current BMI (<18.5, ≥18.5 but <22.5, ≥22.5 but <25, ≥25 but <27.5, or ≥27.5 kg/m 2 ), BMI at age 20 (<18.5, ≥18.5 but <22.5, ≥22.5 but <25, ≥25 but <27.5, or ≥27.5 kg/m 2 ), history of diabetes mellitus (yes or no), and family history of PC (yes or no).Interactions between environmental factors and genotypes were assessed by using likelihood ratio tests within logistic regression models, with and without interaction terms.In haplotype analysis, we used haplotype-effects logistic regression for case-control data. 30This haplotype analysis uses phased and unphased SNP genotype data to estimate haplotype effects and haplotype-environment interactions in a case-control study. 30,31It fits haplotype-effects logistic regression by using the retrospective likelihood method in a special case of a rare disease and a single candidate gene in the Hardy-Weinberg Equilibrium (HWE), under the assumption of gene-environment independence. 30,31Accordance with the HWE was assessed by the chi-square test.

RESULTS
The background characteristics of the subjects are shown in Table 1.Men accounted for 68.7% of case subjects and 74.9% of controls.As compared with the control group, the case group had a significantly higher prevalence of heavy smokers (P = 0.005), a higher prevalence of a history of diabetes mellitus (P < 0.001), a lower current BMI (P = 0.021), and a higher BMI at age 20 (P = 0.018).
We analyzed potential interactions of rs1799782 and rs25487 in XRCC1 and rs1130409 in APE1 with known risk factors for PC, such as smoking, alcohol consumption, overweight, diabetes mellitus, and family history of PC.Because of the low frequency of minor homozygous subjects, these subjects were combined with heterozygous subjects in this analysis.The exposure variables used were current and former BMI (BMI <25 or ≥25 kg/m 2 ), alcohol consumption (<23 or ≥23 g ethanol/day), smoking (<5 or ≥5 PYs), heavy smoking (<40 or ≥40 PYs), history of diabetes mellitus (yes or no), and family history of PC (yes or no).There were no statistically significant interactions (eTable 1).We also analyzed interactions of rs1799782 and rs25487 in XRCC1 and rs1130409 in APE1 with smoking duration and intensity and detected no statistically significant interactions (eTable 2).
Table 3 shows the results of haplotype analysis for XRCC1.The R-squared (R 2 ) value between rs1799782 and rs25487 was 0.15.Haplotype CA in XRCC1 was associated with a statistically significant increase in PC risk (OR = 1.32,CI = 1.01-1.71;P = 0.042) as compared with the most common haplotype, CG, in XRCC1.As shown in Table 4, we defined haplotype CA in XRCC1 as a risk haplotype and analyzed the potential interactions between risk haplotype and known PC risk factors (smoking, alcohol consumption, overweight, diabetes mellitus, and family history of PC).The results showed significant interactions between a family history of PC and the risk haplotype in XRCC1 (interaction P = 0.020).

DISCUSSION
In this case-control study, we found that the genetic polymorphism rs25487 in XRCC1 was associated with increased risk of PC in a Japanese population.However, we did not detect any statistically significant interactions with smoking.
The BER pathway has a primary role in the repair of oxidative base lesions such as 8-hydroxyguanine, formamidopyrimidines, and 5-hydroxyuracil. 21Oxidative damage to DNA may lead to mutations that activate oncogenes or inactivate tumor suppressor genes and may eventually increase the probability of genetic alterations developing into neoplastic events. 21Sequence variants in BER genes are thought to modulate DNA repair capacity and are consequently suspected of being associated with altered cancer risk. 20With regard to the association between PC risk and DNA repair pathway, genetic polymorphisms in the BER pathway are the most widely studied in Western epidemiologic studies. 14To our knowledge, however, this is the first study to detect an association between XRCC1 polymorphisms and the development of PC in an Asian population.
Two common variants of XRCC1 are rs1799782 (in which T substitutes for C) and rs25487 (in which A substitutes for   G), which lead to amino acid substitutions Arg194Trp and Arg399Gln, respectively. 32In particular, the minor allele (A) for rs25487, namely the 399Gln allele, was found to be associated with a higher frequency of glycophorin mutant, elevated DNA adduct levels, higher baseline sister chromatid exchange frequency, and increased sensitivity to ionizing radiation, all of which might be due to reduced BER function. 18In the present study, the minor allele of rs25487 in XRCC1 was associated with elevated PC risk.Thus, our finding is consistent with the potential mechanisms described above.
8][19][20][21] A synergistic effect between these gene polymorphisms and tobacco smoking in relation to PC risk has been suspected. 14xidative DNA damage produced by smoking is expected to be repaired by the BER system, which includes the XRCC1 gene. 33In this study, we observed a significant main effect of rs25487 in XRCC1 but found no statistically significant interaction between XRCC1 loci and smoking.Although the reason for this difference is unclear, the results may have been affected by the heterogeneity of study populations, differences in ethnicity, the small sample size, and potential confounders.On the other hand, the minor allele of rs25487 in XRCC1 was associated with increased risk of PC among light but not heavy smokers.These findings are consistent with past reports, and possible explanations for the limited effect among heavy smokers include enhanced apoptosis at the time of cell division from heavy smoking and induction of DNA repair capacity in response to DNA damage. 20,21n addition, we found that haplotype CG in XRCC1 had a higher impact in those with a family history of PC than in those without such a history.To our knowledge, this phenomenon has not been reported previously and might be explained by family history acting as a surrogate for interaction with known/unknown genetic susceptibility factors.We did not observe any interaction with the ABO genotype, which has been reported to have a significant association with PC risk (data not shown). 34Given our small study population, the finding of an interaction with family history may have been due to chance.Further study is required to duplicate our findings in a larger cohort.
Our study has several methodological issues that warrant discussion.First, the control population was selected from non-cancer patients at ACCH.It is reasonable to assume that this was the same population from which the case subjects were derived, which would bolster the internal validity of our study.Second, with regard to external validity, we previously showed that individuals selected randomly from our control population were similar to the general population of Nagoya City in terms of the exposure of interest. 26owever, only 50% of patients agreed to the collection of a blood sample, and the characteristics of these participants might differ from those of the general population.Third, case-control studies have an intrinsic information bias.However, the HERPACC system is less prone to this bias than are typical hospital-based studies, as the data for all patients were collected before diagnosis.In addition, the results from our self-administered questionnaire may be inaccurate.However, any such misclassification would be nondifferential and would likely underestimate the causal association. 35Fourth, with regard to SNP analysis, because 2 SNPs (rs1052133 in OGG1 and rs1136410 in PARP1) were not in accordance with HWE in our study, further studies that include these factors are required.Fifth, residual confounding by known and unknown risk factors might be present; given the small number of cases, our findings require replication in a larger study.Finally, our study was limited to a Japanese population, and the results cannot necessarily be extrapolated to other populations.
In summary, this case-control study showed a significant association between genetic polymorphisms in XRCC1 and PC risk in a Japanese population.Further investigation of our findings in larger populations are warranted, and the biological mechanisms responsible for the association should be fully elucidated.

Table 1 .
Comparison of selected characteristics of pancreatic cancer (PC) patients and non-cancer controls a P-values calculated by chi-square test.b BMI: body mass index.

Table 2 .
Distribution of cases and controls, and odds ratios for pancreatic cancer associated with selected BER gene polymorphisms Unconditional logistic regression model adjusted for age, sex, current BMI, BMI at age 20, smoking status, drinking habit, history of diabetes mellitus, and family history of pancreatic cancer.

Table 3 .
Haplotype frequencies of XRCC1, and odds ratios for pancreatic cancer associated with XRCC1 haplotype Abbreviation: OR, odds ratio.a SNP 1 is rs1799782; SNP 2 is rs25487.b Haplotype-effects logistic regression for case-control data was used.Multivariable adjustment by age, sex, current BMI, BMI at age 20, smoking status, drinking habit, diabetes mellitus, and family history of PC.