The XRCC1 399 glutamine allele is a risk factor for adenocarcinoma of the lung

Mutation Research 461 (2001) 273–278

The XRCC1 399 glutamine allele is a risk factor for adenocarcinoma of the lung Kevin K. Divine a , Frank D. Gilliland b , Richard E. Crowell c , Christine A. Stidley d , Therese J. Bocklage d , Dennis L. Cook a , Steven A. Belinsky a,∗ a

d

Lung Cancer Program, Lovelace Respiratory Research Institute, Albuquerque, NM 87185, USA b Department of Preventive Medicine, Division of Occupational and Environmental Health, University of Southern California, Los Angeles, CA 90033, USA c Department of Medicine, University of New Mexico, Albuquerque, NM 87131 and Pulmonary Section, New Mexico VA Health Care System, Albuquerque, NM 87108, USA Departments of Family and Community Medicine and Pathology, University of New Mexico, Albuquerque, NM, USA Received 3 July 2000; received in revised form 14 August 2000; accepted 15 August 2000

Abstract Defects in the repair and maintenance of DNA increase risk for cancer. X-ray cross-complementing group 1 protein (XRCC1) is involved with the repair of DNA single-strand breaks. A nucleotide substitution of guanine to adenine leading to a non-conservative amino acid change was identified in the XRCC1 gene at codon 399 (Arg/Gln). This change is associated with higher levels of aflatoxin B1 -adducts and glycophorin A somatic mutations. A case-control study was conducted to test the hypothesis that the 399Gln allele is positively associated with risk for adenocarcinoma of the lung. XRCC1 genotypes were assessed at codon 399 in 172 cases of lung adenocarcinoma and 143 cancer-free controls. Two ethnic populations were represented, non-Hispanic White and Hispanic. The distribution of XRCC1 genotypes differed between cases and controls. Among cases, 47.7% were Arg/Arg, 35.5% were Arg/Gln, and 16.9% were Gln/Gln. Among controls, XRCC1 allele frequencies were 45.5% for Arg/Arg, 44.8% for Arg/Gln, and 9.8% for Gln/Gln. Logistic regression analysis was used to assess the association between lung adenocarcinoma and the G/G genotype relative to the A/A or A/G genotypes. In non-Hispanic White participants, the lung cancer risk associated with the G/G genotype increased significantly after adjustment for age (OR = 2.81; 95% CI, 1.2–7.9; P = 0.03) and increased further after adjustment for smoking (OR = 3.25; 95% CI, 1.2–10.7; P = 0.03). Among all groups, a significant association was found between the G/G homozygote and lung cancer (OR = 2.45; 95% CI, 1.1–5.8; P = 0.03) after adjustment for age, ethnicity, and smoking. This study links a functional polymorphism in the critical repair gene XRCC1 to risk for adenocarcinoma of the lung. © 2001 Elsevier Science B.V. All rights reserved. Keywords: XRCC1; Lung adenocarcinoma; Lung cancer; Tobacco; Smoking; DNA repair enzymes; Polymorphism

1. Introduction Lung cancer is the leading cause of cancer-related deaths in the US with adenocarcinoma compris∗ Corresponding author. Tel.: +1-505-845-1165; fax: +1-505-845-1198. E-mail address: [email protected] (S.A. Belinsky).

ing the most frequent histologic diagnosis in both smokers (40%) and never-smokers (80%). Lifetime risk for lung cancer among smokers ranges from 13 to 25% depending on smoking duration [1]. Because lung cancer occurs in <25% of individuals exposed to cigarette smoke, variation in host defense mechanisms must play a signifi-

0921-8777/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 8 7 7 7 ( 0 0 ) 0 0 0 5 9 - 8

274

K.K. Divine et al. / Mutation Research 461 (2001) 273–278

cant role in determining susceptibility to this disease. Cigarette smoke contains >60 agents classified as carcinogens, tumor promoters, or initiators [2]. The inability to repair DNA damage induced by these chemical carcinogens appears to be one host factor that may influence cancer risk. Defects in the repair and maintenance of DNA increase risk for cancer [3–6]. DNA repair deficiency has been linked to mutations in some genes that result in a complete loss of function for the DNA repair protein, whereas more subtle differences in repair capacity occur through the inheritance of polymorphisms within the gene [7]. One candidate gene shown to be polymorphic is the X-ray cross-complementing group 1 protein (XRCC1). XRCC1 is involved in the repair of DNA single-strand breaks (SSBs), the most common alteration produced by DNA damage. XRCC1 is a multidomain protein that interacts with the nicked DNA and participates with at least three different enzymes, poly-ADP-ribose polymerase [PARP], DNA ligase III (Lig III), and DNA polymerase ␤ (Pol␤), to repair SSBs [8,9]. Shen et al. [8] recently identified three coding polymorphisms in the XRCC1 gene at codons 194 (Arg to Trp), 280 (Arg to His), and 399 (Arg to Gln). These polymorphisms code for non-conservative amino acid changes that could alter XRCC1 function. In particular, the 399Gln polymorphism resulting from a guanine to adenine nucleotide substitution occurs in the PARP binding domain and may affect complex assembly or repair efficiency. In support of this hypothesis, Lunn et al. [10] reported a significant increase in levels of placental aflatoxin B1 -DNA adducts and glycophorin A mutations in erythrocytes from individuals with the XRCC1 399Gln allele. No association was observed for the other two reported polymorphisms. The purpose of the current investigation was to determine whether the XRCC1 399Gln allele is a risk factor for adenocarcinoma of the lung.

2. Materials and methods 2.1. Study population XRCC1 genotypes were assessed in 172 subjects diagnosed with primary adenocarcinoma of the

lung and 143 cancer-free subjects. Classification of ethnicity was done by self-identification. The adenocarcinoma cases were comprised of incident adenocarcinomas diagnosed between 1973 and 1996 in residents of New Mexico and Southern Colorado ascertained through the statewide New Mexico Tumor Registry, a member of the SEER program (Albuquerque, NM) and the St. Mary’s Hospital Tumor Registry (Grand Junction, CO). All cases underwent tumor resection. Cancer-free controls were recruited primarily from the Albuquerque Veterans Administration Medical Center (Albuquerque, NM), a population largely comprised of men. 2.2. DNA source DNA samples for cases were mainly from paraffin-embedded, formalin-fixed adjacent normal tissue. If normal tissue was not available, tumor tissue was used. The cancer-free control DNA samples were from peripheral lymphocytes. DNA was isolated by digestion with proteinase K in sodium dodecyl sulfate (1%), followed by standard phenol–choloroform extraction, and ethanol precipitation. RNA was removed by digestion with RNase A and T1 . 2.3. XRCC1 genotyping XRCC1 genotypes were detected using a PCRrestriction fragment length polymorphism technique. PCR primers were designed around the coding sequence that includes codon 399. The primers were as follows: A) XRCC1-Forward 50 -CTA CTG GCA TCT TCA CTT CTG-30 and B) XRCC1-Reverse 50 -GAA TAG GAC ACG ACC CGT TAC-30 . A 50 ␮l reaction mixture containing ≈200 ng DNA, 1.5 mM MgCl2 , 200 ␮M of each dNTPS, 1 ␮M of each primer, and 2.5 U of Ampli Taq Gold (Perkin-Elmer, Norwalk, CN) in 1X PCR buffer (Perkin-Elmer) was used. The PCR program was initiated by a 10 min denaturation step at 94◦ C followed by 40 cycles of 94◦ C for 30 s, 62◦ C for 30 s, 72◦ C for 30 s, and a final elongation step of 72◦ C for 5 min. The Arg allele at codon 399 contains a MspI site that is lost upon conversion to the glutamine amino acid. The 198 bp PCR product was digested overnight with 20 U of MspI (New England BioLabs Inc., Beverly, MA) at 60◦ C and resolved on 3% agarose gels (FMC BioProducts, Rockland, MN).

K.K. Divine et al. / Mutation Research 461 (2001) 273–278

The homozygous Gln allele produced a single 198 bp product, the homozygous Arg allele produced 145 and 53 bp products (the 53 bp product produced was too small to accurately resolve), and the heterozygous Arg/Gln allele produced three products of 198, 145, and 53 bp. Allelotyping was successful for all samples evaluated.

2.4. Statistical analysis Summary statistics were calculated for genotype, smoking history, age, ethnicity, and gender for both cases and controls. Cases and controls were compared using two-sample t-tests for continuous variables and Fisher’s exact test for categorical variables. To assess the effect of genotype with and without adjustment for smoking, logistic regression models were developed with case-control status as the binary outcome variable. To adjust for the effect of smoking, various models were examined, including ones that used ever versus never smoker as a covariate, restricted the sample to smokers, or included a term differentiating heavy (≥30 pack years) from moderate (<30 pack years) smoking. Detailed smoking information on pack years of smoking was available for 100 and 74% of controls and cases, respectively. Potential interactions between genotype and smoking were assessed through the addition of interaction terms in the logistic regression models. However, most cases and controls were smokers, so separate analyses were also conducted restricted to smokers only. Adjusted models included age and ethnicity. Separate models were developed for each ethnic group. All statistical computing was done in SAS (SAS Institute, Cary, NC) and LogXact (Cytel Software Corp., Cambridge, MA).

275

3. Results An equal proportion of non-Hispanic Whites and Hispanics were present in the control group, whereas approximately 75% of cases were non-Hispanic White. This difference reflects the higher rates of lung cancer in this group that have been ascribed to cigarette smoking patterns of these two groups [11]. The median age was 65 years (range, 33–83) for cases and 63 years (range, 34–83) for controls. Women comprised 48 and 3% of cases and controls, respectively, reflecting the fact that the control population was recruited from the Veterans Medical Center. Gender was not considered as a confounder for the current study, because two previous investigations [10,12] reported no significant difference in the distribution of the 399Gln allele between males and females. The distribution frequency for adenocarcinoma of the lung by gender agrees with historical data reported in the 1973–1997 SEER report [13]. Among cases and controls, 94 and 83% were smokers, respectively. Heavy smokers (≥30 pack years) were more prevalent among cases (72 versus 43%). All three XRCC1 genotypes were represented among cases and controls (Fig. 1), and their distribution is shown in Table 1. The distribution of the A/A genotype (Table 1) did not appear to differ between cases and controls. The prevalence of A/G heterozygotes was 35.5 and 44.8%, whereas the prevalence for G/G homozygotes was 16.9 and 9.8% among cases and controls, respectively (Table 1). The proportion of male and female cases exhibiting the G/G genotype was the same (not shown). The association between the XRCC1 genotype and lung cancer risk is shown in Table 2. All ORs are calculated as G/G relative to A/G or A/A since the distribution of A/G heterozygotes and A/A homozygotes did not differ

Fig. 1. Genotyping the XRCC1 gene in lung adenocarcinoma and control subjects. That portion of the XRCC1 gene encompassing codon 399 was amplified from DNA isolated from 172 lung adenocarcinoma cases and 143 control cases. DNA was digested with MspI overnight and the DNA fragments resolved by agarose gel electrophoresis. Subjects homozygous for 399Gln (198 bp product) are seen in lanes 10, 11, and 17; homozygous for 399Arg (146 bp product) are seen in lanes 1 and 18; and heterozygous for 399Arg/Gln are seen in the remaining lanes.

276

K.K. Divine et al. / Mutation Research 461 (2001) 273–278

Table 1 Distribution of XRCC1 alleles in adenocarcinoma cases and controls Ethnicity

n

Distribution of genotype (%) A/A

All Non-Hispanic White Hispanic

315 200 115

A/G

G/G

Case

Control

Case

Control

Case

Control

47.7 42.5 62.2

45.5 49.3 41.4

35.5 37.8 28.9

44.8 42.5 47.1

16.9 19.7 8.9

9.8 8.2 11.4

Table 2 Observed odds ratios for XRCC1 genotypes in cases and controls Model 1a

n Odds ratioc 95% CI P-value

Model 2b

All

Non-Hispanic White

Hispanic

All

Non-Hispanic White

Hispanic

315 1.79 0.9–3.7 0.10

200 2.81 1.2–7.9 0.03

115 0.79 0.2–2.8 0.72

234 2.45 1.1–5.8 0.03

155 3.25 1.2–10.7 0.03

79 1.40 0.3–5.9 0.65

a

Adjusted for age and ethnicity. Adjusted for age, ethnicity, and smoking. Adjustment for smoking was restricted to smokers with detailed smoking information and included a term differentiating heavy from moderate smoking. c GG relative to AA and AG. b

significantly between cases and controls (Table 1). Among all groups, the association between the G/G homozygote and lung cancer after adjusting for age and ethnicity was 1.79 (95% CI, 0.9–3.7; P = 0.10; Table 2). After additional adjustment for smoking, there was a significant association (OR, 2.45; 95% CI, 1.1–5.8; P = 0.03) between the G/G genotype and lung cancer risk. The distribution of alleles between cases and controls differed by ethnicity suggesting that lung cancer risk associated with the G allele was different between non-Hispanic White and Hispanic participants. The frequency of the G/G genotype in non-Hispanic White cases was 2.4-times that in controls (Table 1). In contrast, no difference was observed between Hispanic cases and controls. When analyses were stratified by ethnicity, lung cancer risk associated with the G/G genotype was increased in non-Hispanic White participants (OR, 2.81; 95% CI, 1.2–7.9; P = 0.03). After adjustment for smoking, the lung cancer risk associated with the G/G genotype increased to 3.25 (95% CI, 1.2–10.7; P = 0.03). The risk for lung cancer after adjustment for smoking in Hispanic participants was lower (OR, 1.40; 95% CI, 0.3–5.9;

P = 0.65), but was not statistically different from risk in non-Hispanic Whites. Smoking intensity (packs per day) did not modify the association between the 399Gln allele and lung adenocarcinoma (not shown); however, there was limited power to detect an interaction between smoking and the XRCC1 genotype.

4. Discussion This study is the first to link a functional polymorphism in the critical DNA repair gene XRCC1 and risk for adenocarcinoma of the lung. A role for this polymorphism as a risk factor for tobacco-associated cancer is consistent with a previous study that demonstrated increased risk for squamous cell carcinoma of the head and neck in non-Hispanic Whites [12]. These findings suggest that elevated risk associated with the 399Gln allele may be linked to a deficit in repair of DNA damage initiated by the carcinogens within tobacco. The two major carcinogens present in tobacco, benzo(a)pyrene and 4-methylnitrosamino-1-3-pyridyl1-butanone (NNK), both produce SSBs, which if

K.K. Divine et al. / Mutation Research 461 (2001) 273–278

not repaired can lead to gene mutation [14–17]. In addition, persistent SSBs can be converted to double-strand breaks [9] leading to homozygous deletion, an alteration detected in adenocarcinoma particularly within the short arm of chromosome 3 [18]. The XRCC1 pathway has been implicated as an important factor in single nucleotide excision repair [9,19]. This conclusion is based in part on the fact that deficiencies in XRCC1 result in genomic instability as evident by increased frequency of sister chromatid exchange and increased sensitivity to alkylating agents and/or ionizating radiation [20–22]. In addition, Chinese hamster ovary cell lines containing mutations in critical domains of the XRCC1 gene exhibit reduced SSB repair capacity. XRCC1 is thought to function as a molecular chaperone or scaffold protein that stabilizes and/or modifies the activity of other polypeptides. XRCC1 interacts with three other proteins, PARP, Pol␤, and Lig III, all of which possess enzymatic activity and are involved in repair of SSBs. The XRCC1 399 polymorphism occurs at the COOH-terminal side of the PARP interaction domain at residues conserved in hamsters, mice, and humans. The XRCC1/PARP interaction is likely an important step for XRCC1 function. XRCC1 interacts with PARP through a BRCA1 C-terminal (BRCT) domain specifically at amino acids 384–476. BRCT domains are autonomous folding units of approximately 100 amino acids that are widespread in a superfamily of DNA damage-response and cell-cycle check-point proteins [23,24]. Several studies [25–27] have demonstrated that inhibition or deficiencies in PARP produce results similar to although not as severe as those observed with XRCC1 deficiencies. Cells deficient in both PARP and XRCC1 have long-lived or excessive numbers of SSBs. Lunn et al. [10] demonstrated that people homozygous for the 399Gln allele have significantly higher levels of glycophorin A mutations than heterozygotes or persons homozgyous for the 399Arg allele. In addition, persons homozygous for the 399Gln allele also have higher frequencies of another marker for DNA damage, sister chromatid exchange, than those carrying the Arg/Arg genotype [28]. Thus, these studies support the hypothesis that alteration in the BRCT region as a result of the XRCC1 399Gln polymorphism may disrupt the binding efficiency of PARP to XRCC1, thereby reducing the fidelity of this DNA repair pathway.

277

If the binding efficiency of PARP to XRCC1 is reduced, then the penetrance of the G/G genotype may be greater in persons exposed to higher amounts of carcinogens compared to persons with low exposure. Too few subjects were never smokers to assess the modifying effect of smoking; however, the effect of exposure intensity among smokers could be examined. The effect of the G/G genotype was larger when restricted to smokers. Because non-Hispanic White and Hispanic participants had significant differences in pack years of smoking (57 ± 3 versus 36 ± 5, P < 0.01), the lower OR in Hispanics is consistent with the hypothesis that the penetrance depends upon smoking intensity. Larger case-control studies are needed to better define whether smoking intensity interacts with the 399Gln allele in adenocarcinoma and to determine if this allele is a risk factor in other ethnic populations. In conclusion, this study implicates the XRCC1 399Gln/Gln genotype as a marker of genetic susceptibility for adenocarcinoma of the lung. The high mortality associated with lung cancer could be reduced through early detection and by the identification of individuals at elevated risk. Genotyping people for the XRCC1 399Gln allele could be used as one marker for defining risk, thereby identifying individuals who should participate in chemoprevention trials.

Acknowledgements Supported by Grants NIEHS ES08801, ES05836, and P20 ES09871, and NIH CA70190 under US DOE Cooperative Agreement No. DE-FC04-96AL-76406; P01 ES09581; and by the New Mexico VA Health Care System. References [1] M.R. Law, J.K. Morris, H.C. Watt, N.J. Wald, The dose-response relationship between cigarette consumption, biochemical markers and risk of lung cancer, Br. J. Cancer 75 (1997) 1690–1693. [2] DDHS, Reducing the health consequences of smoking: 25 years of progress. A report of the Surgeon General — 1989, DDHS Publ. no. (CDC) 89-8411, US Department of Health and Human Services, Washington, 1989. [3] L. Grossman, Q. Wei, DNA repair capacity (DRC) as a biomarker of human variational responses to the environment, in: J.M. Vos (Ed.), DNA Repair Mechanisms: Impact on

278

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11]

[12]

[13]

[14]

[15]

[16]

K.K. Divine et al. / Mutation Research 461 (2001) 273–278 Human Diseases and Cancer, Landes, Georgetown, TX, 1994, pp. 329–347. K.J. Helzlsouer, E.L. Harris, R. Parshad, H.R. Perry, F.M. Price, K.K. Sanfors, DNA repair proficiency: potential susceptibility factor for breast cancer, J. Natl. Cancer Inst. 88 (1996) 754–755. X. Wu, J. Gu, W.K. Hong, J.J. Lee, C.I. Amos, J.H. Winn, K.K. Fu, J. Cooper, M.R. Spitz, Benzyo(a)pyrene diol expoxide and bleomycin sensitivity and susceptibility to cancer of upper aerodigestive tract, J. Natl. Cancer Inst. 90 (1998) 1393–1399. Q. Wei, L. Cheng, W.K. Hong, M.R. Spitz, Reduced DNA repair capacity in lung cancer patients, Cancer Res. 56 (1996) 4103–4107. M.R. Shen, M.Z. Zdzienicka, H. Morenweiser, L.H. Thompson, M.P. The, Mutations in hamster single-strand break repair gene XRCC1 causing defective DNA repair, Nucleic Acids Res. 26 (1998) 1032–1037. M.R. Shen, I.M. Jones, H. Morenweiser, Nonconservative amino acid substitution variants exist at polymorphic frequency in DNA repair genes in healthy humans, Cancer Res. 58 (1998) 604–608. L.H. Thompson, M.G. West, XRCC1 keeps DNA from getting stranded, Mutat. Res. 459 (2000) 1–18. R.M. Lunn, R.G. Langlois, L.L. Hsieh, C.T. Thompson, D.A. Bell, XRCC1 polymorphisms: effects on alfatoxin B1 -DNA adducts and glycophorin A variant frequency, Cancer Res. 59 (1999) 2557–2561. C.G. Humble, J.M. Samet, D.R. Pathak, B.J. Skipper, Cigarette smoking and lung cancer in ‘Hispanic’ Whites and other Whites in New Mexico, Am. J. Public Health 75 (1985) 145–148. E.M. Sturgis, E.J. Castillo, L. Li, R. Zheng, S.A. Eicher, G.L. Clayman, S.S. Strom, M.R. Spitz, Q. Wei, Polymorphisms of DNA repair gene XRCC1 in squamous cell carcinoma of the head and neck, Carcinogenesis 20 (1999) 2125–2129. L.A.G. Ries, M.P. Eisner, C.L. Kosary, B.F. Hankey, B.A. Miller, L. Clegg, B.K. Edwards (Eds.), SEER Cancer Statistics Review, 1973–1977, National Cancer Institute, Bethesda, MD, 2000. K.M. Yang, C.G. Shin, S.Y. Choe, D.H. Kim, Detection of DNA single-strand breaks induced by procarcinogens in Chinese hamster ovary cells cocultured with rat hepatocytes, J. Toxicol. Environ. Health 13 (1984) 133–143. H. Vaghef, A.C. Wisen, B. Hellman, Demonstration of benzo(a)pyrene-induced DNA damage in mice by alkaline single cell gel electrophoresis: evidence for strand breaks in liver but not in lymphocytes and bone marrow, Pharmacol. Toxicol. 78 (1996) 37–43. A.B. Weitberg, D. Corvese, The effect of epigallocatechin galleate and sarcophytol A on DNA strand breakage induced by tobacco-specific nitrosamines and stimulated human phagocytes, J. Exp. Clin. Cancer Res. 18 (1999) 433–437.

[17] S.S. Hecht, DNA adduct formation from tobacco-specific N-nitrosamines, Mutat. Res. 424 (1999) 127–142. [18] J. Hung, Y. Kishimoto, K. Sugio, A. Virmani, D.D. McIntire, J.D. Minna, A.F. Gazdar, Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinoma, JAMA 273 (1995) 558–563. [19] M. Masson, C. Niedergang, V. Schreiber, S. Muller, J. Menissier-De Murcia, G. De Murcia, XRCC1 is specifically associated with poly(ADP-ribose) polymerase and negatively regulates its activity following DNA damage, Mol. Cell Biol. 18 (1998) 3563–3571. [20] L.H. Thompson, K.W. Brookman, L.E. Dillehay, A.V. Carrano, J.A. Mazimas, C.L. Mooney, J.L. Minkler, A CHO-cell strain having hypersensitivity to mutagens, a defect in strand-break repair, and an extraordinary baseline frequency of sister chromatid exchange, Mutat. Res. 95 (1982) 427–440. [21] A.V. Carrano, J.L. Minker, L.E. Dillehay, L.H. Thompson, Incorporated bromodeoxyuridine enhances the sisterchromatid exchange and chromosomal aberration frequencies in EMS-sensitive Chinese hamster cell line, Mutat. Res. 162 (1986) 233–239. [22] M.Z. Zdzienicka, G.P. van der Schans, A.T. Natarajan, L.H. Thompson, I. Nueteboom, J.W.I.M. Simons, A Chinese hamster ovary cell mutant (EM-C11) with sensitivity to simple alkylating agents and a very high level of sister chromatid exchanges, Mutagenesis 7 (1992) 265–269. [23] P. Bork, K. Hofmann, P. Bucher, A.F. Neuwald, S.F. Altschul, E.V. Koonin, A superfamily of conserved domains in DNA damage-responsive cell cycle checkpoint proteins, FASEB J. 11 (1997) 68–76. [24] I. Callebaut, J.P. Mornon, From BRCA1 to RAP1: a widespread BRCT module closely associated with DNA repair, FASEB Lett. 400 (1997) 25–30. [25] J.M. de Murcia, C. Niedergang, M. Trucco, B. Ricoul, M. Dutrillaux, F.J. Mark, M. Oliver, M. Masson, A. Dierich, M. LeMur, C. Walztinger, P. Chambon, G. de Murcia, Requirement of poly(ADP-ribose) polymerase in recovery from DNA damage in mice and in cells, Proc. Natl. Acad. Sci. U. S. A. 94 (1997) 7303–7307. [26] Z.Q. Wang, L. Stingl, C. Morrison, M. Jantsch, M. Los, K. Schulze-Osthoff, E.F. Wagner, PARP is important for genomic stability but dispensible in apoptosis, Genes Dev. 11 (1997) 2347–2358. [27] M. Masutani, T. Nozaki, E. Nishiyama, T. Shimokawa, Y. Tachi, H. Suzuki, H. Nakagama, K. Wakabayashi, T. Sugimura, Function of poly(ADP-ribose) polymerase in response to DNA damage: gene-disruption study in mice, Mol. Cell Biochem. 193 (1999) 149–152. [28] E.J. Duell, J.K. Wiencke, T.J. Cheng, A. Varkonyi, Z.F. Zuo, T.D.S. Ashok, E.J. Mark, J.C. Wain, D.C. Christiani, K.T. Kelsey, Polymorphisms in the DNA repair genes XRCC1 and ERCC2 and biomarkers of DNA damage in human blood mononuclear cells, Carcinogenesis 21 (2000) 956–971.