Polymorphisms of XRCC1 gene, alcohol consumption and colorectal cancer
Article first published online: 30 MAR 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 116, Issue 3, pages 428–432, 1 September 2005
How to Cite
Hong, Y.-C., Lee, K.-H., Kim, W.-C., Choi, S.-K., Woo, Z.-H., Shin, S.-K. and Kim, H. (2005), Polymorphisms of XRCC1 gene, alcohol consumption and colorectal cancer. Int. J. Cancer, 116: 428–432. doi: 10.1002/ijc.21019
- Issue published online: 10 JUN 2005
- Article first published online: 30 MAR 2005
- Manuscript Accepted: 12 JAN 2005
- Manuscript Received: 10 AUG 2004
- Inha University. Grant Number: INHA-22055-01
- Korea Science & Engineering Foundation. Grant Number: R01-2002-000-00554-0(2003)
- colorectal cancer;
To evaluate contribution of polymorphisms of the XRCC1 gene to the risk of colorectal cancer, we conducted a case-control study of 209 colorectal cancer cases and 209 age- and gender-matched controls in the Korean population. We tested the hypothesis by constructing allele combinations with known SNP. Allelic variants of the XRCC1 gene at codons 194, 280 and 399 were analyzed in lymphocyte DNA by PCR-RFLP. We observed an increased risk of colorectal cancer associated with the 399Gln allele. The odds ratio (OR) was 1.61 (95% confidence interval [CI] 1.09–2.39) for the 399Gln allele. When combined allele-specific OR were calculated after estimating frequencies, 3 common allele combinations were found to be associated with an increased risk of colorectal cancer. The OR for the 194Trp-280Arg-399Arg was 1.48 (95% CI = 1.06–2.07) using 194Arg-280Arg-399Arg as the reference. The OR for the 194Arg-280His-399Arg and the 194Arg-280Arg-399Gln were 1.78 (95% CI = 1.09–2.89) and 1.78 (95% CI = 1.23–2.59), respectively. Analysis after controlling for smoking, exercise and dietary habits indicated that alcohol consumption (≥80 g/week) is a significant risk factor of colorectal cancer (OR = 2.60, 95% CI = 1.46–4.62). An increased risk for colorectal cancer was identified in alcohol drinkers with the risky allele combinations. Our results suggest that polymorphisms in the XRCC1 genes may contribute to colorectal cancer susceptibility, and some evidence was obtained of a genetic modification for the relationship between alcohol intake and colorectal cancer. © 2005 Wiley-Liss, Inc.
Colorectal cancer is one of the most commonly diagnosed cancers in North America and Western Europe,1 and is the fourth most common cause of cancer in South Korea.2 Inherited deficiencies in DNA repair have been associated with an individual's susceptibility to cancer.3 Therefore, polymorphisms of DNA repair genes may increase the risk of colorectal cancer.
The human XRCC1 gene, one of the DNA repair genes, was identified because of its ability to restore DNA repair activity in a Chinese hamster ovary mutant cell line EM9.4 The XRCC1 protein is involved in base-excision repair, and interacts with DNA ligase III, DNA polymerase β, poly(ADP-ribose)polymerase (PARP), polynucleotide kinase and AP endonuclease I.5, 6, 7, 8 The base-excision repair pathway is designed to remove non-bulky base adducts, which are produced by methylation, oxidation or reduction by ionizing radiation or oxidative damage.9, 10 Three coding polymorphisms of the DNA repair gene XRCC1 (Arg194Trp, Arg280His and Arg399Gln) have been identified in man, suggesting altered efficiency due to amino acid substitutions.11, 12
Alcohol consumption has been associated with the production of reactive oxygen species, which are known to cause DNA lesions that can be removed by the DNA base-excision repair pathway.13 Meta-analyses of alcohol consumption in relation to colorectal cancer have reported a small or moderate additional risk among alcohol drinkers.14, 15 The association between alcohol consumption and colorectal cancer is not conclusive, as some reports have shown an increased risk and others no excess risk.16, 17, 18, 19, 20 Given the possible association between alcohol consumption and colorectal cancer, polymorphisms in the XRCC1 gene, which have the potential to alter the structure of DNA repair enzyme, could affect the risk of colorectal cancer in alcohol drinkers.
We conducted a hospital-based case-control study of colorectal cancer in Incheon, South Korea, where the age-adjusted incidence rates were 27.0 among men and 17.5 among women per 100,000 person-years during the period 1997–2000.2 To elucidate the role of the XRCC1 gene polymorphisms in colorectal cancer development, we carried out genotype association analyses. We also hypothesized that the effect of alcohol intake on colorectal cancer is modulated by polymorphisms of the XRCC1 gene. To our knowledge, no previous study has examined the effect of XRCC1 gene polymorphisms on the association between alcohol intake and colorectal cancer. The Arg194Trp, Arg289His and Arg399Gln polymorphisms of the XRCC1 gene, and their allele combinations were analyzed to evaluate genetic susceptibility to colorectal cancer, and the possible modification effect on the relationship between alcohol intake and colorectal cancer risk.
Material and methods
Study population and sample collection
We recruited 209 subjects newly diagnosed with colorectal cancer at the outpatient clinics of general surgery between 2001–2003 at the Inha University Hospital, Incheon, South Korea. They consisted of 145 patients diagnosed with rectal cancer and 64 patients diagnosed with colon cancer (that included ascending, sigmoid and descending colon). All subjects voluntarily participated, completed a self-administered questionnaire and provided peripheral blood.
Controls were selected by 1:1 matching for age (±3 years) and gender after initial random sampling from the Incheon Health Examinee Cohort (15,540 participants), which consisted of subjects who voluntarily visited the health-screening clinic at the same hospital.
A questionnaire administered to the subjects included questions on alcohol consumption, dietary patterns, smoking history and exercise frequency. Self-reported alcohol use was evaluated as a categorical variable (non-users, <1 time/week, 1–2 times/week, 3–4 times/week or daily) and the amount of alcohol consumed was recorded with the type of alcoholic beverage. Considering the alcohol consumption amount and distribution among these study subjects, we chose the 80 g/week as a cutoff point where 27.8% of the subjects reported to consume 80–160 g/week or more. Dietary habit and smoking status were classified as categorical variables. Information on exercise was obtained as frequency for a week.
Our study was approved by the Institutional Review Board of Inha University Hospital and written informed consent was obtained from all participants.
Genomic DNA was extracted from peripheral blood lymphocytes using a QIAamp DNA Blood Mini Kit. We determined the genotypes of the XRCC1 genes by PCR-RFLP.21, 22 All of the PCR reactions were carried out in a total reaction volume of 20 μl containing 50 ng of DNA, 1 U Taq polymerase in 1× PCR buffer, 1.5 mM MgCl2, 250 μM dNTPs and 1 μM of each primer. Thermal cycling conditions consisted of an initial denaturation step at 94°C for 5 min, followed by 30 cycles of 94°C for 1 min, annealing (at 62°C for Arg194Trp and Arg399Gln, and 68°C for Arg280His) for 1 min and at 72°C for 1 min, followed by a final extension step at 72°C for 7 min. The PCR products were digested with MspI for Arg194Trp and Arg399Gln, and with RsaI for Arg280His, and analyzed in a 3 % Metaphor gel (BMA, Walkersville, ME).
Univariate analyses were carried out to calculate the crude odds ratios (OR) and their 95% confidence intervals (CI), of the associations between colorectal cancer and the investigated genotypes and haplotypes, singly or in combination. Odds ratio and 95% CI, adjusted for smoking, alcohol consumption, exercise frequency and dietary habits, were then estimated for the various genotypes or combined alleles using logistic regression analysis. Combined allele analyses were carried out using the computer program PHASE (http://www.stat.washington.edu/stephens/phase.html) to construct haplotype structures and to estimate their frequencies. Statistical analyses were carried out with the SAS (version 6.12) statistical package. All tests of statistical significance were 2-sided.
The mean ages of the colorectal cancer patients and the controls were 59.5 (SD = 12.0) and 59.0 (SD = 11.8) years, respectively. The proportion of men and women were the same in the patient and control groups (i.e., 116 men and 93 women).
Table I presents the distribution of the alleles for the Arg194Trp, Arg280His and Arg399Gln polymorphisms of the XRCC1 gene. Genotype distributions at each locus were consistent with the Hardy-Weinberg equilibrium for the controls (p = 0.629 for Arg194Trp, p = 0.991 for Arg280His and p = 0.954 for Arg399Gln). The allele frequencies of 2 polymorphisms, Arg194Trp and Arg280His, were not statistically different in the patient and control groups (p > 0.05). In contrast, colorectal cancer risk showed a significant association with the 399Gln allele (p = 0.047). Table II shows that the adjusted OR for heterozygotes or homozygotes for the 399Gln allele was 2.00 (95% CI = 1.15–3.47) using 399Arg homozygotes as the reference group.
|Alleles||Patients (%)||Controls (%)||p-value1|
|XRCC1 Codon 194|
|Arg||274 (65.6)||284 (67.9)|
|Trp||144 (34.4)||52 (32.1)||0.463|
|XRCC1 Codon 280|
|Arg||369 (88.3)||380 (80.4)|
|His||49 (11.7)||38 (19.6)||0.213|
|XRCC1 Codon 399|
|Arg||312 (74.6)||336 (80.4)|
|Gln||106 (25.4)||82 (19.6)||0.047|
|XRCC1 genotypes||Patients (%)||Controls (%)||OR (95% CI)||Adjusted OR1 (95% CI)|
|XRCC1 Codon 194|
|Arg/Arg||90 (43.0)||101 (48.3)||1||1|
|Arg/Trp||94 (45.0)||82 (39.2)||1.29 (0.85–1.94)||1.32 (0.74–2.36)|
|Trp/Trp||25 (12.0)||26 (12.5)||1.08 (0.58–2.00)||0.96 (0.40–2.33)|
|Arg/Trp or Trp/Trp||119 (57.0)||108 (51.7)||1.24 (0.84–1.82)||1.23 (0.72–2.11)|
|XRCC1 Codon 280|
|Arg/Arg||161 (77.0)||173 (82.8)||1||1|
|Arg/His||47 (22.5)||34 (16.2)||1.49 (0.91–2.43)||2.00 (0.94–4.26)|
|His/His||1 (0.5)||2 (1.0)||0.54 (0.05–5.98)||1.64 (0.10–27.58)|
|Arg/His or His/His||48 (23.0)||36 (17.2)||1.43 (0.88–2.32)||1.98 (0.95–4.14)|
|XRCC1 Codon 399|
|Arg/Arg||112 (53.6)||136 (65.1)||1||1|
|Arg/Gln||88 (42.1)||64 (30.6)||1.67 (1.11–2.51)||2.18 (1.23–3.88)|
|Gln/Gln||9 (4.3)||9 (4.3)||1.21 (0.47–3.16)||1.03 (0.31–3.67)|
|Arg/Gln or Gln/Gln||97 (46.4)||73 (34.9)||1.61 (1.09–2.39)||2.00 (1.15–3.47)|
The Phase program predicted 4 different allele combinations in 836 chromosomes. The estimated frequencies of the combined alleles in the cases and controls are shown in Table III. Frequencies of the 4 common combined alleles were significantly different in the patient and control groups (p = 0.007). The OR for 194Trp-280Arg-399Arg was 1.48 (95% CI = 1.06–2.07) using 194Arg-280Arg-399Arg as the reference. The OR for 194Arg-280His-399Arg and 194Arg-280Arg-399Gln were 1.78 (95% CI = 1.09–2.89) and 1.78 (95% CI = 1.23–2.59), respectively. The adjusted OR for 194Trp-280Arg-399Arg and 194Arg-280His-399Arg did not reach statistical significance. The other 4 possible allele combinations, namely, 194Trp-280His-399Arg, 194Trp-280Arg-399Gln, 194Arg-280His-399Gln, and 194Trp-280His-399Gln, were predicted to be absent, suggesting that the 3 SNP are in linkage disequilibrium.
|Combined alleles1||Patients (%)||Controls (%)||Crude OR (95% CI)||Adjusted OR2 (95% CI)|
|194Arg-280Arg-399Arg||119 (28.5)||164 (39.2)||1||1|
|194Trp-280Arg-399Arg||144 (34.4)||134 (32.1)||1.48 (1.06–2.07)||1.05 (0.69–1.60)|
|194Arg-280His-399Arg||49 (11.7)||38 (9.1)||1.78 (1.09–2.89)||1.64 (0.83–3.26)|
|194Arg-280Arg-399Gln||106 (25.4)||82 (19.6)||1.78 (1.23–2.59)||1.59 (1.01–2.49)|
Alcohol was found to be associated significantly with colorectal cancer in the logistic regression model controlling for genotypes (p = 0.001), whereas smoking, exercise and dietary habits were not associated significantly with colorectal cancer (p > 0.05).
Table IV shows that alcohol intake (80 g or more a week) is a significant risk factor of colorectal cancer (OR = 2.60, 95% CI = 1.46–4.62) after controlling for smoking, exercise and dietary habits. The referent combined allele, 194Arg-280Arg-399Arg and the other 3 allele combinations, 194Trp-280Arg-399Arg, 194Arg-280His-399Arg, and 194Arg-280Arg-399Gln, were analyzed to compare effects on the association between alcohol intake and colorectal cancer. Although colorectal cancer risk resulting from alcohol intake was 1.62 (95% CI = 0.80–3.26) for the 194Arg-280Arg-399Arg, an increased risk for colorectal cancer was observed for alcohol drinkers with the other 3 combined alleles.
|Amount of alcohol intake||Patients (%)||Controls (%)||OR (95% CI)||p-value|
|<80 g/wk||145 (69.4)||157 (75.1)|
|≥80 g/wk||64 (30.6)||52 (24.9)||2.60 (1.46–4.62)||0.001|
|<80 g/wk||86 (73.8)||121 (72.3)|
|≥80 g/wk||33 (26.2)||43 (27.7)||1.62 (0.80–3.26)||0.178|
|<80 g/wk||95 (66.0)||101 (75.4)|
|≥80 g/wk||49 (34.0)||33 (24.6)||3.30 (1.60–6.80)||0.001|
|<80 g/wk||35 (71.4)||31 (81.6)|
|≥80 g/wk||14 (28.6)||7 (18.4)||7.19 (1.31–39.68)||0.024|
|<80 g/wk||74 (69.8)||61 (74.4)|
|≥80 g/wk||32 (30.2)||21 (25.6)||2.97 (1.26–7.03)||0.013|
The results of our study demonstrate that those who carry 399Gln allele have a higher risk of colorectal cancer than those carrying 399Arg homozygotes, and that there is a significant association between an increased risk of colorectal cancer and the combined allele 194Arg-280Arg-399Gln. Our data also suggest that alcohol intake is a significant risk factor of colorectal cancer. In particular, we observed an increased colorectal cancer risk associated with alcohol consumption in combined alleles 194Trp-280Arg-399Arg, 194Trp-280His-399Arg and 194Arg-280Arg-399Gln. This finding suggests different susceptibilities in alcohol drinkers to colorectal cancer depending on the XRCC1 genotype.
XRCC1 is believed to form complexes with DNA ligase III, via a breast cancer COOH-terminus (BRCT) domain at its COOH terminus, and with DNA polymerase β at its NH2 terminus, to repair gaps left during base excision repair.23 Shen et al.11 reported 3 coding polymorphisms, i.e., Arg194Trp, Arg280His and Arg399Gln, in the XRCC1 gene.
The Arg194Trp and the Arg280His amino acid substitutions reside in the linker region that separate the DNA polymerase β domain of XRCC1 from the PARP-interacting domain. The Arg399Gln change occurs in the COOH-terminal of the PARP-interacting domain, and within a relatively nonconserved region between conserved residues of the BRCT domain.11 Even though these SNP have not been proposed to cause a complete loss of XRCC1 function, they were postulated to have an effect on the DNA-protein complex assembly and ultimately DNA repair function.3, 24, 25
Our findings suggest that the Arg399Gln polymorphism is associated with the risk of colorectal cancer. Individuals with the 399Gln allele were found to have 2 times greater risk of colorectal cancer than those with the homozygous 399Arg genotype. The SNP at codons 194 and 280 were not found to be significantly associated with colorectal cancer in our study.
The 399Gln allele of the XRCC1 gene has been reported to be associated with higher levels of aflatoxin B1-DNA adducts and glycophorin NN mutations in placental DNA, suggesting that the Arg399Gln polymorphism may result in deficient DNA repair.21 It was also reported to be associated with reduced repair of NNK-induced genetic damage in human lymphocytes26 and with the frequency of P53 mutations in oral squamous cell carcinomas.27 In a colorectal cancer study among Egyptians, the 399Gln allele was associated with an earlier age of onset.28 In contrast, an experimental study reported that the Arg399Gln polymorphism does not affect the DNA repair function, suggesting that it has little or no biological impact on cancer susceptibility.29
Studies on the Arg194Trp polymorphism have produced inconsistent results with respect to the risk of various cancers, even though most published studies have reported a reduced risk of cancer for the 194Trp allele.30 The inconsistent results of association studies between the polymorphisms and cancer risk may be due to linkage disequilibrium with a determinant variant of the XRCC1 gene.31 Interestingly, Hsieh et al.27 reported linkage disequilibrium between Arg194Trp and Arg399Gln, by showing that all carrying the 194Trp allele also carried the 399Arg allele. We also observed a close linkage among 3 SNP, specifically, 194Trp was found to be linked to 280Arg and 399Arg. The previously reported cancer risk associated with the 194Trp allele, therefore, could be attributed to the polymorphisms Arg280His or Arg399Gln.
Relatively few studies have been conducted on the effects of the Arg280His polymorphism.30 A case-control study of lung cancer reported that the 280His allele carried a higher risk than the homozygous wild-type genotype, and a possible interaction with alcohol consumption.25
We used a statistical method to infer phase at linked loci from genotypes and construct allele combinations using the software Phase v 2.0. The algorithm of the program starts with some initial haplotype reconstruction, repeatedly choosing an individual at random, and estimates the individual's haplotypes under the assumption that all the other haplotypes are correctly reconstructed.32 Even though the statistical method to construct haplotypes does not provide information of phase structure exactly, it was reported to be sufficiently accurate for the reconstruction of haplotypes, particularly for genomic regions with high linkage disequilibrium and less ambiguity.32, 33 Because the region studied was relatively small, covering approximately 1.9 kb for the 3 polymorphisms of XRCC1 gene, and the 4 possible allele combinations were estimated to be absent in the analysis, there is likely to be strong linkage disequilibrium in the region. We confirmed the linkage disequilibrium by showing that D′ values were 0.997 or more among the 3 SNP.34 For the ambiguity problem, we evaluated accuracy in estimating combined allele frequencies indirectly by comparing the Phase algorithm with the expectation–maximization algorithm.35, 36, 37 When we checked individual probabilities of combined allele estimation, the inference results between the 2 algorithms were virtually identical and the expected posterior probabilities were very close to one (≥0.999) for all subjects. These observations indicate that the reconstruction of allele combinations was sufficiently accurate in this analysis.
Alcohol consumption is associated with the production of free radical intermediates such as, hydroxyethyl free radicals and reactive oxygen species.13, 38 These intermediates may generate DNA base lesions, which should be removed by the base excision repair pathway. XRCC1 may be involved in the base excision repair of alcohol-induced DNA damage, therefore, those with a reduced repair capacity due to XRCC1 polymorphisms could be more susceptible to the development of colorectal cancer.25
We found that alcohol consumption (80 g or more a week) is a risk factor of colorectal cancer irrespective of XRCC1 genotypes. Similarly, a meta-analysis of 22 studies on alcohol and colorectal cancer risk reported an increased risk for the dose of alcohol consumption of 25 g/day.15 Furthermore, the risk associated with alcohol consumption was modulated by the XRCC1 allele combinations in our study. Because DNA repair plays a critical role in protecting against DNA damage conferred by alcohol intake, a reduced DNA repair capacity would constitute a significant risk factor for alcohol-associated cancers. Alcohol consumption per se may interfere with the repair of damaged DNA.13 Therefore, alcohol intake and XRCC1 polymorphisms may inhibit DNA repair in concert and thus promote colorectal cancer development.
These findings should be considered in light of a number of limitations. Recall bias is unavoidable because we assessed the effects of alcohol consumption based on a self-administered questionnaire among cancer patients at outpatient clinics and among controls at a health examination center. Moreover, we cannot exclude the possibility of misclassifications of disease status, because the controls were not rigorously examined for the absence of colorectal cancer. Our study also involved a relatively small number of subjects, thus limiting its statistical power. The wide confidence interval for OR of the association between alcohol intake and colorectal cancer after subgrouping according to allele combinations for the evaluation of the genetic modification was probably due to the small sample size. Future studies with large numbers of subjects are required to confirm the relationship. Because we evaluated only 3 commonly studied SNP in the XRCC1 gene, other SNP in linkage disequilibrium with the studied SNP may also be necessary to be examined to refine the allele combination structures.
Despite these limitations, our epidemiologic approach has several strengths. The problem of confounding was reduced by choosing the age- and gender-matched controls from the Incheon Health Examinee Cohort. We evaluated the influence of polymorphisms in the XRCC1 gene, not only based on SNP, but also on combined alleles. Moreover, the allele combination approach yielded a more informative relationship between the XRCC1 polymorphisms and colorectal cancer.
In conclusion, our study demonstrates that polymorphisms of the XRCC1 genes, particularly based on allele combinations of the 3 SNP, Arg194Trp, Arg280His and Arg399Gln, contribute to colorectal cancer susceptibility. We also found possibility for a gene–environmental interaction between alcohol intake and the XRCC1 genotype. Although our study does not conclusively identify the polymorphisms of the XRCC1 gene that contribute to the development of colorectal cancer, we suggest that association studies between allele combinations and the risk of cancer development would be useful for determining genetic susceptibility.
H. Kim was partially supported by the Basic Research Program of the Korea Science & Engineering Foundation (No. R01-2002-000-00554-0(2003).
- 1Cancer incidence in five continents. vol. VII. IARC Scientific Publication No. 143. Lyon: IARC, 1997., , , , .
- 2Cancer incidence in Incheon, South Korea, during 1997-2000: results from Incheon Cancer Registry. 25th Annual Scientific Congress and Meeting of the International Association of Cancer Registries. Lyon: IARC, 2003., , , .