relationship between xrcc1 polymorphisms and susceptibility to prostate cancer in men from han,...

Asian J Androl 2007; 9 (3): 331–338

.331.Tel: +86-21-5492-2824; Fax: +86-21-5492-2825; Shanghai, China

Relationship between XRCC1 polymorphisms and suscepti-bility to prostate cancer in men from Han, Southern China

Zheng Xu1, Li-Xin Hua1, Li-Xin Qian1, Jie Yang1, Xin-Ru Wang2, Wei Zhang1, Hong-Fei Wu1

1Department of Urology, First Affiliated Hospital of Nanjing Medical University, Nanjing 210029, China2School of Public Health, Nanjing Medical University, Nanjing 210029, China

Abstract

Aim: To investigate the association among XRCC1 polymorphisms, smoking, drinking and the risk of prostate cancer(PCa) in men from Han, Southern China. Methods: In a case-control study of 207 patients with PCa and 235 cancer-free controls, frequency-matched by age, we genotyped three XRCC1 polymorphisms (codons 194, 280 and 399)using the polymerase chain reaction-restriction fragment length polymorphism (PCR-RELP) method. Results: Amongthe three polymorphisms, we found that the XRCC1 Arg399Gln variant allele was associated with increased PCa risk(adjusted odd ratio [OR]: 1.67, 95% confident interval [CI]: 1.11–2.51), but the XRCC1 Arg194Trp variant allele hada 38% reduction in risk of PCa (adjusted OR: 0.62, 95% CI: 0.41–0.93). However, there was no significant risk ofPCa associated with Arg280His polymorphism. When we evaluated the three polymorphisms together, we found thatthe individuals with 194Arg/Arg wild-type genotype, Arg280His and Arg399Gln variant genotypes had a significantlyhigher risk of PCa (adjusted OR: 4.31; 95% CI: 1.24–14.99) than those with three wild-type genotypes. In addition,we found that Arg399Gln variant genotypes had a significant risk of PCa among heavy smokers (adjusted OR: 2.04;95% CI: 1.03–4.05). Conclusion: These results suggest that polymorphisms of XRCC1 appear to influence the riskof PCa and may modify risks attributable to environmental exposure. (Asian J Androl 2007 May; 9: 331–338)

Keywords: XRCC1; polymorphism; prostate cancer; genetic susceptibility; molecular epidemiology

.Original Article .

DOI: 10.1111/j.1745-7262.2007.00263.xwww.asiaandro.com

© 2007, Asian Journal of Andrology, Shanghai Institute of Materia Medica, Chinese Academy of Sciences. All rights reserved.

Correspondence to: Dr Hong-Fei Wu, Department of Urology,First Affiliated Hospital of Nanjing Medical University, 300 Guang-zhou Rd., Nanjing 210029, China.Tel: +86-25-8371-8836 ext. 6603 Fax: +86-25-8378-0079E-mail: [email protected] 2006-08-29 Accepted 2006-11-08

1 Introduction

Prostate cancer (PCa) is one of the most commonmalignancies in men in the Western world. Prostate isthe leading site for cancer incidences, accounting for 31%

of new cancer cases in men [1]. The incidence of PCavaries greatly with race and geography. The incidence inblack Americans is 60 times higher than that of the Hanpopulation in China, so the research on pathogenesis ofPCa from a genetic aspect is of important significance[2]. It is well known that in the carcinogenic process mul-tiple points at which genetically-determined host character-istics and/or environmental factors might influence indi-vidual susceptibility through affecting DNA-repair capac-ity and other cellular processes [3]. The mechanism ofPCa development is similar to that of other major tumors,which are dependent on the interactions of genetic factors

.332.

XRCC1 polymorphisms and risk of prostate cancer

http://www.asiaandro.com; [email protected]

and environmental agents. Living organisms suffer con-tinuous damage from diverse environmental agents and nor-mal cellular metabolism products. DNA repair is essentialin protecting the genome of cells from environmentalhazards, such as tobacco smoking. Reduced DNA repaircapacity (DRC) might result in a higher risk for many peoplein developing cancer [4]. Therefore, the ability to monitorand repair carcinogen-induced DNA damage is the deter-minant factor of host susceptibility to carcinogenesis.

A complex system of DNA repair enzymes has a vi-tal role in protecting the genome of cells from all kindsof carcinogenic exposure. The DNA repair enzymeXRCC1 is thought to be involved in base excision repair(BER) of oxidative DNA and single-strand breaks repair[5]. Human XRCC1 is mapped at human chromosome19q13.2–13.3, spans a genomic distance 33 kb, con-tains 17 exons, and transcripts a protein of 633 aminoacids (69.5 kDa). Although XRCC1 has no known en-zymatic activity, it can interact with several important re-pair proteins through its different domains, such as DNAligase Ⅲ at its breast cancer susceptibility gene C terminusⅡ (BRCT-II) domain, DNA polymerase β at its NH2terminus, poly (ADP-ribose) polymerase (PARP) 1 and 2at its BRCT-I domain, human AP endonuclease, poly-nucleotide kinase (PNK), human 8-oxoguanine DNAglycosylase (OGG1) and proliferating cell nuclear antigen(PCNA) at the central section of XRCC1 protein [6]. Threecommon polymorphisms that lead to amino acid substitu-tions in XRCC1 gene have been described in codon 194(exon 6, base C to T, amino acid Arg to Trp), codon 280(exon 9, base G to A, amino acid Arg to His) and codon399 (exon 10, base G to A, amino acid Arg to Gln) [7].

Although a large number of molecular epidemiologi-cal studies have been conducted to evaluate the role of thethree polymorphisms on cancer risk, the results are con-flicting rather than conclusive. In the present study, wegenotyped the three polymorphisms and evaluated the as-sociations between the polymorphisms and their haplotypeswith PCa risk in a hospital-based case-control study.

2 Materials and methods

2.1 Study subjectsThe study subjects consisted of 207 cases with newly

diagnosed PCa and 235 cancer-free male controls re-cruited from the First Affiliated Hospital of Nanjing Medi-cal University (Nanjing, China) between September 2003and April 2006. All cases were patients diagnosed with

PCa through biopsy of puncture or operation. Serologi-cal (prostate specific antigen [PSA], prostatic acidphosphatase), physical and other auxiliary examinationswere conducted on all controls to exclude the possibilityof PCa, and any control was excluded from the study ifhe ever had an abnormal PSA test (i.e., ≥ 4ng/dL), or heever had an abnormal digital rectal examination, or hehad any previous cancer diagnosis. The mean age of thecases was 71.2 vs. 69.7 years in the controls. All sub-jects were ethnic Han Chinese, permanently residing inJiangsu or Anhui Province, China. Each subject wasinformed about the aims and requirements of the study,and informed consent for participation was obtained inaccordance with institutional guidance at Nanjing Medi-cal University. A structured questionnaire was adminis-tered by interviewers to collect information on demo-graphic data and lifestyle characteristics. In our research,smoking more than five cigarettes per day for more than5 years was defined as a ‘smoking habit’. Taking smokeinto the lung when smoking was defined as “deepsmoking”, while only taking smoke into the mouth wasdefined as “superficial smoking”. Pack-years of smoking(cigarettes per day/20) × (years with smoked) were cal-culated to indicate the cumulative smoking dose. “Drink-ing habit” was defined as drinking at least three times perweek for more than 10 years. “Family history of cancer”was defined as any cancer in first-degree relatives (parents,siblings or children). After interview, approximately 5 mLof venous blood sample was collected from each subject.

2.2 GenotypingGenomic DNA was isolated and purified from anti-

coagulated blood by the traditional phenol/chloroformextraction and ethanol precipitation, dissolved in TEbuffer (pH = 7.4) and stored at –20ºC for genotyping.

The XRCC1 Arg194Trp, Arg280His and Arg399Glnpolymorphisms were determined using the polymerasechain reaction-restriction fragment length polymorphism(PCR-RFLP) method. Three separate PCR assays wereused to detect the polymorphisms in exon 6, exon 9 andexon 10 of XRCC1 using primers of (a) XRCC1 194F,5'-GCCCCGTCCCAGGTA-3'; and XRCC1 194R 5'-AGCCCCAAGACCCTTTCACT-3' and (b) XRCC1 280F,5'-CCAGTGGTG CTAACCTAATC-3'; and XRCC1 280R5'-CACTCAGCACCAGTACCACA-3' and (c) XRCC1399F, 5 '-TTGTGCTTTCTCTGTGTCCA-3'; andXRCC1 399R 5'-TCCTCCAGCCTTTACTGATA-3'. PCRconditions were 95ºC for 5 min, followed by 37 cycles of

Asian J Androl 2007; 9 (3): 331–338


95ºC for 40 s, 60ºC (for codon 194) or 55ºC (for codon280 and codon 399) for 40 s, 72ºC for 60 s and a finalelongation step at 72ºC for 10 min. Following PCR, 10 µLaliquot were removed and subjected to restriction diges-tion separately with Pvu II (for codon 194), Rsa I (forcodon 280) and Msp I (for codon 399) (New EnglandBiolabs, Ipswich, MA, USA). The Pvu II restricted prod-ucts of XRCC1 codon194 Arg/Arg, Arg/Trp and Trp/Trp genotypes had band sizes of 491, 491/294/197 and294/197 bp, respectively. The Rsa I restricted productsof XRCC1 codon 280 Arg/Arg, Arg/His and His/His geno-types had band sizes of 145/56, 201/145/56 and 201 bp,respectively. The Msp I restricted products of XRCC1codon 399 Arg/Arg, Arg/Gln and Gln/Gln genotypes hadband sizes of 375/240, 615/375/240 and 615 bp,respectively.

2.3 Statistical analysisAll differences in select demographic variables, pack-

years of smoking, smoking method, alcohol use, familyhistory of cancer, frequencies of the XRCC1 genotypesbetween the case and control groups were evaluated byusing χ2-test. Unconditional univariate and multivariatelogistic regression analyses were performed to obtain thecrude and adjusted odds ratios (OR) for the risk of PCa

and their 95% confidence intervals (CI). When we se-lected variables for the multivariate analysis, any variablewhose univariate test has a P-value < 0.2 was a candi-date for the multivariable model along with all variablesof known clinical importance. Once the variables hadbeen identified, we began with a model containing all ofthe selected variables. The multivariate adjustment in-cluded age, pack-years of smoking, alcohol use and fam-ily history of cancer. Considering the potential interac-tion among the three polymorphisms on the risk for PCa,the associations between the combined genotypes of thethree polymorphisms risk of PCa was evaluated. Thegenotypes were further stratified by subgroups of age,pack-years of smoking, smoking method, alcohol useand family history of cancer. All tests of statistical sig-nificance were two-sided. All statistical analyses wereperformed with Statistical Analysis System software(version 9.1.3e; SAS Institute, Cary, NC, USA).

3 Results

3.1 Characteristics of the study populationTable 1 shows the demographic characteristics of

the case and control groups. The cases and controlsappeared to be well-matched in age (grouping by 70 years

Table 1. Distribution of selected variables and XRCC1 alleles in prostate cancer (PCa) cases and controls. a Two-sided χ2-test.

Variables Case (n = 207) Control (n = 235)

Pa

n % n % Age (years) ≤70 97 46.9 129 54.9 0.092 >70 110 53.1 106 45.1 Pack-years of smoking 0 66 31.9 122 51.9 <20 25 12.1 46 19.6 < 0.001 ≥20 116 56.0 67 28.5 Smoking method Never smoking 66 31.9 122 51.9 Superficial smoking 53 25.6 58 24.7 < 0.001 Deep smoking 88 42.5 55 23.4 Drinking status Never 153 73.9 193 82.1 0.037 Ever 54 26.1 42 17.9 Family history of cancer No 159 76.8 216 91.9 < 0.001 Yes 48 23.2 19 8.1

.334.



old, P = 0.092) and gender (all male). However, therewere more heavy smokers (56.0%) and deep smokers(42.5%) among the cases than the controls (28.5% and23.4%, respectively), and these differences were statisti-cally significant (P < 0.001). In addition, cases had moreoften been alcohol user (26.1% vs. 17.9%, P = 0.037),and more often had a first-degree relative with cancer(23.2% vs. 8.1%, P < 0.001).

3.2 Genotype distributions of XRCC1 polymorphismsamong cases and controls

The distribution of XRCC1 genotypes in case pa-tients and control subjects and their associations withrisk of PCa are presented in Table 2. The genotype fre-quencies of these three polymorphisms among the con-trols were all in agreement with the Hardy-Weinberg equi-librium (χ2-test: P = 0.214 for Arg194Trp, 0.524 forArg280His and 0.680 for Arg399Gln). The frequencyof XRCC1 194Arg/Arg genotype was 49.8% in casesand 39.1% in controls, suggesting that 194Arg/Arg geno-type might be a risk genotype. Logistic regression analysisrevealed that the 194Arg/Trp heterozygote had a signifi-cantly decreased risk (adjusted OR: 0.62; 95% CI: 0.40–0.95) and the 194Trp/Trp homozygote had a non-sig-nificantly decreased risk (adjusted OR: 0.62; 95% CI:

0.31–1.25), compared with 194Arg/Arg homozygouswild-type. When we combined the variant genotypes(Arg/Trp+ Trp/Trp) assuming a co-dominant allele effect,the combined variant genotypes had a 38% reduction inrisk of PCa (adjusted OR: 0.62; 95% CI: 0.41–0.93).However, the Arg399Gln variant allele was associated withincreased PCa risk. The data indicated that individuals with399Gln allele (Arg/Gln+Gln/Gln) were more prevalent inthe cases (47.9%) compared to the controls (34.9%). Thecombined variant genotypes were at a 1.67-fold greaterrisk of PCa than homozygous wild-type genotype, andthe adjusted ORs for 399Arg/Gln and 399Gln/Gln geno-types were 1.61 (95% CI: 1.05–2.47) and 2.09 (95% CI:0.85–5.14), respectively, as shown in Table 2. However,the XRCC1 Arg280His polymorphism was not associ-ated with a significantly increased risk of PCa.

3.3 Association between the combined genotypes ofXRCC1 polymorphisms and risk of PCa

Combined genotype analyses of the three XRCC1polymorphisms are shown in Table 3. Assessment ofintra-gene interactions in the logistic regression modelsof the three XRCC1 polymorphisms revealed statisticallysignificant interactions (P < 0.001). Compared with wild-type for all three genotypes, Arg399Gln variant alone was

Table 2. XRCC1 variant genotypes and prostate cancer (PCa) risk. aAdjusted for age, pack-years of smoking, alcohol use and family historyof cancer. bUsed as reference group.

Genotype Cases (n = 207) Controls (n = 235) Crude OR Adjusted ORa

No. % No. % (95% CI) (95% CI)XRCC1 Arg194Trp

Arg/Arg 103 49.8 92 39.1 1 (Reference)b 1 (Reference)b

Arg/Trp 84 40.5 117 49.8 0.64 (0.43–0.95) 0.62 (0.40–0.95)Trp/Trp 20 9.7 26 11.1 0.69 (0.36–1.31) 0.62 (0.31–1.25)

Arg/Trp +Trp/Trp 104 50.2 143 60.9 0.65 (0.45–0.95) 0.62 (0.41–0.93)XRCC1Arg280His


Arg/His 40 19.3 39 16.6 1.2 (0.74–1.95) 1.16 (0.69–1.9)His/His 2 1.0 3 1.3 0.78 (0.13–4.72) 0.64 (0.10–4.35)

Arg/His+His/His 42 20.3 42 17.9 1.17 (0.73–1.88) 1.12 (0.67–1.87)XRCC1Arg399Gln


Arg/Gln 85 41.1 72 30.6 1.67 (1.12–2.49) 1.61 (1.05–2.47)Gln/Gln 14 6.8 10 4.3 1.98 (0.85–4.63) 2.09 (0.85–5.14)

Arg/Gln+Gln/Gln 99 47.9 82 34.9 1.71 (1.17–2.51) 1.67 (1.11–2.51)

Asian J Androl 2007; 9 (3): 331–338


associated with a statistically significant increased risk of PCa(adjusted OR: 3.19; 95% CI: 1.48–6.88). Moreover, we foundthat the individuals with 194Arg/Arg wild-type genotype,Arg280His and Arg399Gln variant genotypes had a signifi-cantly higher risk of PCa (adjusted OR: 4.31; 95% CI: 1.24–14.99) than those with three wild-type genotypes.

3.4 Stratified analyses of the association between XRCC1polymorphisms and risk of PCa

The dichotomized genotypes (the XRCC1 194Arg/Trp+Trp/Trp versus Arg/Arg genotypes, the XRCC1280Arg/His+His/His vs. Arg/Arg genotypes and the XRCC1399Arg/Gln+Gln/Gln versus Arg/Arg genotypes) werefurther examined for subgroups of the variables listed inTable 1. The stratified, adjusted ORs are presented inTable 4. Among individuals no more than 70 years old,the OR for XRCC1 Arg194Trp combined variant geno-types was 0.55 (95% CI: 0.30–1.00). Among non-smokers,the OR for XRCC1 Arg194Trp combined variant genotypeswas 0.48 (95% CI: 0.24–0.95). Among heavy-smokers(Pack-years of smoking ≥ 20), the OR for XRCC1Arg399Gln combined variant genotypes was 2.04 (95%CI: 1.03–4.05). Among non-drinkers and subjects with-out a family history of cancer, the ORs for XRCC1Arg399Gln combined variant genotypes were 1.74 (95%CI: 1.07–2.83) and 2.02 (95% CI: 1.26–3.25), respectively.

4 Discussion

The XRCC1 protein has no known catalytic activitybut serves to orchestrate BER through its role as a cen-

tral scaffolding protein for DNA ligase Ⅲ, DNA poly-merase β, and PARP, and also through its function inrecognizing and binding to single-strand breaks [8]. TheArg399Gln polymorphism is located in the region of theBRCT-Ⅰ interaction domain of XRCC1 with poly(ADP-ribose) polymerase, the presence of the variant 399Glnallele has been shown to be associated with measurablereduced DRC as assessed by the persistence of DNAadducts [9,10], tumor-suppressor gene P53 mutations[11], increased red blood cell glycophorin A [9], elevatedlevels of sister chromatid exchanges [10] and prolongedcell-cycle delay [12]. Both the XRCC1 Arg194Trp andArg280His variants occur in the newly identified PCNAbinding region [13], but few studies have examined theinfluence of the 194Trp allele on the function of theXRCC1 protein, and there are relatively fewer studiesconducted to examine the association between Arg280Hisvariant and cancer risk. To the best of our knowledge, itis the first study that has investigated the association ofthese three polymorphisms and the risk of PCa in a Chi-nese population.

In this study, we investigated the associations ofthree polymorphisms of the DNA repair gene XRCC1with the risk of PCa in a southern Chinese population.When we evaluated each polymorphism separately, wefound that the Arg194Trp variant genotypes had a lowerrisk of PCa than the wild-type genotype, but theArg399Gln variant genotypes were associated with a sig-nificantly increased risk of PCa compared with the wild-type genotypes. However, the Arg280His polymorphismhad no main effect on PCa risk. When we analyzed

Table 3. Association between the combined genotypes of XRCC1 polymorphisms and risk of prostate cancer (PCa). aAdjusted for age,pack-years of smoking, alcohol use and family history of cancer. bUsed as reference group. WT, homozygous wild type; VT, heterozygousand homozygous variant. Genotype Case (n = 207) Control (n = 235) Crude OR Adjusted ORa

Codon 194 280 399 n % n % (95% CI) (95% CI)WT WT WT 18 8.7 33 14.0 1.00 (Reference)b 1.00 (Reference)b

VT WT WT 62 30.0 85 36.2 1.34 (0.69–2.59) 1.40 (0.68–2.84)WT VT WT 18 8.7 20 8.5 1.65 (0.70–3.89) 1.87 (0.75–4.66)WT WT VT 53 25.6 34 14.5 2.86 (1.39–5.86) 3.19 (1.48–6.88)VT VT WT 10 4.8 15 6.4 1.22 (0.46–3.27) 1.12 (0.38–3.28)WT VT VT 14 6.8 5 2.1 5.13 (1.59–16.57) 4.31 (1.24–14.99)VT WT VT 32 15.5 41 17.4 1.43 (0.69–2.99) 1.36 (0.62–3.00)VT VT VT 0 – 2 0.9 – –

P < 0.001

.336.



these three polymorphisms together, we found that indi-viduals with Arg399Gln variant genotype, Arg194Trp andArg280His wild-type genotypes were associated with astatistically significant increased risk of PCa compared withwild-type for all three genotypes (adjusted OR: 3.19;95% CI: 1.48–6.88). Moreover, we found that the indi-viduals with 194Arg/Arg wild-type genotype, Arg280Hisand Arg399Gln variant genotypes had a much higher risk

of PCa (adjusted OR: 4.31; 95% CI: 1.24–14.99). Suchfindings suggest that polymorphisms in XRCC1 maycontribute to the risk and therefore play a role in theetiology of PCa.

A lot of researchers have studied the associationsbetween polymorphisms of the DNA repair gene XRCC1and cancer risk. Stern et al. [14] reported a case-con-trol study of the association between XRCC1 genotypes

Table 4. Stratified analyses of the association between XRCC1 polymorphisms and risk of prostate cancer (PCa). aAdjusted for age, pack-years of smoking, alcohol use and family history of cancer. WT, homozygous wild type; VT, heterozygous and homozygous variant. Arg194Trp Arg280His Arg399Gln

WT-Arg VT-Trp WT-Arg VT-His WT-Arg VT-GlnOR OR (95% CI)a OR OR (95% CI)a OR OR (95% CI)a

Cases/controls Cases/controls Cases/controls Cases/controls Cases/controls Cases/controls Age ≤70 1 0.55 (0.30–1.00) 1 0.88 (0.41–1.89) 1 1.46 (0.81–2.63)

46/42 51/87 79/106 18/23 53/87 44/42 >70 1 0.89 (0.46–1.70) 1 1.40 (0.62–3.12) 1 1.75 (0.91–3.36) 57/50 53/56 86/87 24/19 55/66 55/40 Pack-years

of smoking0 1 0.48 (0.24–0.95) 1 0.99 (0.42–2.33) 1 1.24 (0.61–2.49)

37/46 29/76 52/99 14/23 36/80 30/42<20 1 0.54 (0.16–1.78) 1 0.45 (0.08–2.69) 1 1.18 (0.35–3.93)

12/18 13/28 22/40 3/6 16/29 9/17

≥20 1 1.12 (0.56–2.26) 1 1.22 (0.53–2.80) 1 2.04 (1.03–4.05)

54/28 62/39 91/54 25/13 56/44 60/23 Smoking method Never 1 0.48 (0.24–0.95) 1 0.99 (0.42–2.33) 1 1.24 (0.61–2.49) 37/46 29/76 52/99 14/23 36/80 30/42

Superficial 1 0.84 (0.35–2.03) 1 0.97 (0.31–3.05) 1 2.28 (0.91–5.70)23/20 30/38 45/48 8/10 29/42 24/16

Deep 1 1.06 (0.46–2.44) 1 1.22 (0.45–3.36) 1 1.33 (0.61–2.91)43/26 45/29 68/46 20/9 43/31 45/24

Alcohol use Never 1 0.69 (0.42–1.13) 1 1.15 (0.61–2.17) 1 1.74 (1.07–2.83)

81/78 72/115 122/160 31/33 75/124 78/69 Ever 1 0.76 (0.28–2.06) 1 0.76 (0.24–2.35) 1 0.93 (0.34–2.54)

22/14 32/28 43/33 11/9 33/29 21/13 Family of history cancer No 1 0.79(0.49–1.27) 1 1.26 (0.69–2.28) 1 2.02 (1.26–3.25)

80/86 79/130 127/178 32/38 78/145 81/71 Yes 1 0.32(0.01–1.26) 1 0.57 (0.13–2.64) 1 0.32 (0.09–1.08)

23/6 25/13 38/15 10/4 30/8 18/11

Asian J Androl 2007; 9 (3): 331–338


and bladder cancer in a white US population. They foundsome evidence of a protective effect for subjects thatcarried at least one copy of the codon 194 variant allelerelative to those homozygous for the common allele(adjusted OR: 0.59; 95% CI: 0.3–1.0). Ritchey et al.[15] reported that XRCC1 399Gln/Gln genotype was as-sociated with an increased risk of PCa (OR: 2.18; 95%CI; 0.99–4.81). Chen et al. [16] also reported that a sig-nificantly increased risk of PCa was observed in white menwith the XRCC1 399Gln allele (OR: 1.6, 95% CI; 1.1–2.4).Our results from the Chinese population reported hereare consistent with previous findings that the XRCC1194Arg/Arg, 399Arg/Gln and Gln/Gln genotypes are theat-risk genotypes [14–16]. There are few studies inves-tigating the Arg280His polymorphism and cancer risk todate, and most reported a non-significant risk of cancerassociation with 280His allele [14, 17]. Therefore, thesedata strongly support our molecular epidemiologic find-ings that functional polymorphisms influencing the ac-tivity of XRCC1 are associated with an increased risk ofcancer.

Genetic polymorphisms often vary between ethnicgroups. For example, the frequencies of XRCC1 194Arg/Arg, Arg/Trp and Trp/Trp in our 235 southern China con-trol group were 39.1%, 49.8% and 11.1%, respectively,compared with 83%, 17%, and 0%, respectively, of 197cancer-free white men controls in the study by Stern et al.[14]. Similarly, the frequencies of the XRCC1 280Arg/Arg, Arg/His, and His/His in our controls were 82.1%,16.6%, and 1.3%, respectively, compared with 90.7%,9.3%, and 0%, respectively, of 312 white men in a studyby Moullan et al. [18]. For the XRCC1 Arg399Glnpolymorphisms, we found the frequencies of Arg/Arg,Arg/Gln, and Gln/Gln (65.1%, 30.6% and 4.3%,respectively) among controls were also different to thestudy by Olshan et al. [19] (38.5%, 50.9% and 10.6%,respectively). This variation needs to be investigatedfurther.

Tobacco smoking might be a risk factor for PCa.We found that the OR for XRCC1 Arg399Gln combinedvariant genotypes was 2.04 (95% CI: 1.03–4.05) amongheavy-smokers (pack-years of smoking ≥ 20). But wedid not find any evidence for interactions between thepolymorphisms of the XRCC1 gene and other risk factors,such as alcohol drinking and family history of cancer.There was a significantly increased risk associated withthe Arg399Gln variant genotypes among non-drinkers orsubjects without a family history of cancer. This ob-

served association might suggest that the cancer patientswithout the established risk factors may have some otherunknown exposures, or have other genetic factors thatare linked to the putative XRCC1 risk genotypes. An-other possibility is that the results are purely chance asthe numbers in the subgroups compared were relativelysmall. Again, larger studies are needed to verify thefindings.

In conclusion, the XRCC1 Arg194Trp and Arg399Glnpolymorphisms have a major effect on the risk of PCa.The variant genotypes may modulate the risk of PCa as-sociated with smoking. However, these results may bebiased by the relatively small number of subjects in thevarious subgroups analyzed and therefore need to bevalidated by larger studies. Future molecular epidemio-logical studies of the three and other sequence variantsof the XRCC1 gene and their association with DNA re-pair phenotypes will help us understanding the role ofthe XRCC1 gene in the etiology of PCa.

References

1 Greenlee RT, Hill-Harmon MB, Murray T, Thun M. Cancerstatistics, 2001. CA Cancer J Clin 2001; 51: 15–36.

2 Quinn M, Babb P. Patterns and trends in prostate cancerincidence, survival, prevalence and mortality. Part I: interna-tional comparisons. BJU Int 2002; 90: 162–73.

3 Quinones LA, Irarrazabal CE, Rojas CR, Orellana CE, AcevedoC, Huidobro C, et al. Joint effect among p53, CYP1A1,GSTM1 polymorphism combinations and smoking on pros-tate cancer risk: an exploratory genotype-environment inter-action study. Asian J Androl 2006; 8: 349–55.

4 Wei Q, Cheng L, Amos CI, Wang LE, Guo Z, Hong WK, et al.Repair of tobacco carcinogen-induced DNA adducts and lungcancer risk: a molecular epidemiologic study. J Natl CancerInst 2000; 92: 1764–72.

5 Chen S, Tang D, Xue K, Xu L, Ma G, Hsu Y, et al. DNArepair gene XRCC1 and XPD polymorphisms and risk of lungcancer in a Chinese population. Carcinogenesis 2002; 23:1321–5.

6 Kubota Y, Nash RA, Klungland A, Schar P, Barnes DE, LindahlT. Reconstitution of DNA base excision-repair with purifiedhuman proteins: interaction between DNA polymerase betaand the XRCC1 protein. EMBO J 1996; 15: 6662–70.

7 Shen MR, Jones IM, Mohrenweiser H. Nonconservative aminoacid substitution variants exist at polymorphic frequency in DNArepair genes in healthy humans. Cancer Res 1998; 58: 604–8.

8 Marintchev A, Mullen MA, Maciejewski MW, Pan B, GrykMR, Mullen GP. Solution structure of the single-strand breakrepair protein XRCC1 N-terminal domain. Nat Struct Biol1999; 6: 884–93.

9 Lunn RM, Langlois RG, Hsieh LL, Thompson CL, Bell DA.XRCC1 polymorphisms: effects on aflatoxin B1-DNA ad-

.338.



ducts and glycophorin A variant frequency. Cancer Res 1999;59: 2557–61.

10 Duell EJ, Wiencke JK, Cheng TJ, Varkonyi A, Zuo ZF, AshokTD, et al. Polymorphisms in the DNA repair genes XRCC1and ERCC2 and biomarkers of DNA damage in human bloodmononuclear cells. Carcinogenesis 2000; 21: 965–71.

11 Casse C, Hu YC, Ahrendt SA. The XRCC1 codon 399Glnallele is associated with adenine to guanine p53 mutations innon-small cell lung cancer. Mutat Res 2003 528: 19–27.

12 Hu JJ, Smith TR, Miller MS, Lohman K, Case LD. Geneticregulation of ionizing radiation sensitivity and breast cancerrisk. Environ Mol Mutagen 2002; 39: 208–15.

13 Fan J, Otterlei M, Wong HK, Tomkinson AE, Wilson DM3rd. XRCC1 co-localizes and physically interacts with PCNA.Nucleic Acids Res 2004; 32: 2193–201.

14 Stern MC, Umbach DM, van Gils CH, Lunn RM, Taylor JA.DNA repair gene XRCC1 polymorphisms, smoking, and bladdercancer risk. Cancer Epidemiol Biomarkers Prev 2002; 11: 939–43.

15 Ritchey JD, Huang WY, Chokkalingam AP, Gao YT, Deng J,Levine P, et al. Genetic variants of DNA repair genes andprostate cancer: a population-based study. Cancer EpidemiolBiomarkers Prev 2005; 14: 1703–9.

16 Chen L, Ambrosone CB, Lee J, Sellers TA, Pow-Sang J, ParkJY. Association between polymorphisms in the DNA repairgenes XRCC1 and APE1, and the risk of prostate cancer inwhite and black Americans. J Urol 2006; 175: 108–12.

17 Lee JM, Lee YC, Yang SY, Yang PW, Luh SP, Lee CJ, et al.Genetic polymorphisms of XRCC1 and risk of the esophagealcancer. Int J Cancer 2001; 95: 240–6.

18 Moullan N, Cox DG, Angele S, Romestaing P, Gerard JP, HallJ. Polymorphisms in the DNA repair gene XRCC1, breastcancer risk, and response to radiotherapy. Cancer EpidemiolBiomarkers Prev 2003; 12: 1168–74.

19 Olshan AF, Watson MA, Weissler MC, Bell DA. XRCC1polymorphisms and head and neck cancer. Cancer Lett 2002;178: 181–6.

Edited by Dr Sidney R. Grimes

relationship between xrcc1 polymorphisms and susceptibility to prostate cancer in men from han,...

Documents