Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Toxicol Lett. Author manuscript; available in PMC 2010 May 22.
Published in final edited form as:
PMCID: PMC2680739

DNA repair genotype interacts with arsenic exposure to increase bladder cancer risk


Drinking water arsenic exposure has been associated with increased bladder cancer susceptibility. Epidemiologic and experimental data suggest a co-carcinogenic effect of arsenic with exposure to DNA damaging agents, such as cigarette smoke. Recent evidence further supports the hypothesis that genetic variation in DNA repair genes can modify the arsenic – cancer relationship, possibly because arsenic impairs DNA repair capacity. We tested this hypothesis in a population-based study of bladder cancer with XRCC3, ERCC2 genotype/haplotype and arsenic exposure data on 549 controls and 342 cases. Individual exposure to arsenic was determined in toenail samples by neutron activation. Gene-environment interaction with arsenic exposure was observed in relation to bladder cancer risk for a variant allele of the double-strand break repair gene XRCC3 T241M (adjusted OR 2.8 (1.1–7.3) comparing to homozygous wild type among those in the top arsenic exposure decile (interaction p-value 0.01). Haplotype analysis confirmed the association of the XRCC3 241. Thus, double-strand break repair genotype may enhance arsenic associated bladder cancer susceptibility in the U.S. population.

Keywords: DNA repair, bladder cancer, arsenic, polymorphism, interaction


Arsenic is an established carcinogen and is currently regulated by Federal and State standards for public water systems (Karagas et al. 1998b; Peters et al. 1999). Ingestion of high concentrations of inorganic arsenic is a cause of cancer of the skin, lung, and urinary bladder, and is a suspected cause of kidney cancer and other malignancies (Cantor et al. 2006). Drinking water arsenic exposure also has been associated with increased bladder cancer risk, possibly through a co-carcinogenic mechanism (Chen et al. 1986; Johansson and Cohen 1997; Karagas et al. 2004b). We hypothesized that individuals with impaired DNA repair capacity may represent a more sensitive subpopulation to arsenic carcinogenesis.

Arsenic’s biologic effects and the extent of disease risk in the human population remains unknown at the exposure levels commonly encountered in U.S. drinking water (Abernathy et al. 1999). Studies, both clinical and experimental, have demonstrated that the biological effects of arsenic exist at levels below the new 10 μg/L standard (Kapaj et al. 2006). These effects include endocrine disruption, suppression of hormone regulation and hormone mediated gene transcription, alteration of cell cycle kinetics, and alterations in cellular proliferative response associated with carcinogenesis (Kapaj et al. 2006; Rossman 2003).

Approximately 2% of the drinking water serving US households contains 2 μg/L arsenic or more (IARC 2004). Chronic exposure to arsenic through drinking water is a major problem in several regions of the U.S., particularly the Northeast and Southwest (Ayotte et al. 2003). While this problem is primarily the result of geologic contamination, drinking water arsenic levels in these areas can be well above the current recommended maximum contaminant level of 10 μg/L (Ayotte et al. 2005). Adding to this, approximately 40% of households in New Hampshire are served by un-regulated privately supplied drinking water wells, 10% of which contain arsenic at levels greater than 10 μg/L (Karagas et al. 1998b). Moreover, until recent studies revealed the extent of geologic arsenic contamination of drinking water, arsenic was not part of the standard laboratory water-testing panel (Andrew et al. 2006a).

Studies of bladder and skin cancer conducted in the New Hampshire population have detected evidence of elevated cancer risks (Karagas et al. 2004a; Karagas et al. 2001). For bladder cancer, an excess risk was observed primarily among smokers exposed to arsenic in the drinking water, supporting the hypothesis that these levels of arsenic are co-carcinogenic (Karagas et al. 2004a). The population of New Hampshire is quite stable. Subjects in this study reported using the same water system for a mean of 15 years prior to diagnosis (Karagas et al. 2004a).

There is also evidence that the process of arsenic carcinogenesis may be modulated by genetic differences in DNA repair. Specifically polymorphisms in the XRCC3 and ERCC2/XPD repair genes have been associated with arsenic-related skin cancer risk (Applebaum et al. 2007; Thirumaran et al. 2006). In Bangladesh, individuals with the nucleotide excision repair (NER) pathway variant ERCC2 genotype at codon 751 had an increased risk of arsenic-induced premalignant hyperkeratotic skin lesions (Ahsan et al. 2003).

The current study takes advantage of a population-based study of bladder cancer conducted in New Hampshire, U.S. to investigate the hypothesis that the XRCC3 and ERCC2 DNA repair gene polymorphisms interact with arsenic exposure to increase bladder cancer susceptibility.


Study group

We identified all cases of bladder cancer diagnosed among New Hampshire residents, ages 25 to 74 years, from July 1, 1994 to June 30, 1998 from the State Cancer Registry. Within 15 days of diagnosis, the state mandated rapid reporting system requires submission of an initial report of cancer, and a definitive report within 120 days. To be eligible for the study, subjects were required to have a listed telephone number and speak English. We sought physician consent before contacting eligible bladder cancer patients. We interviewed a total n = 459 bladder cancer cases, which was 85% of the cases confirmed to be eligible for the study. Non-participants included (n = 10) whose physician denied patient contact, (n = 63) were reported as deceased by a household member or physician, (n = 3) no answer after 40 attempts distributed over day, evenings and weekends, (n = 75) declined participation and (n = 8) were too ill to take part. A standardized histopathology review was conducted by the study pathologist, and from this review we excluded eleven subjects who were initially reported to the cancer registry as bladder cancer (Karagas et al. 2004a).

For efficiency, we shared a control group with a study of non-melanoma skin cancer conducted covering a diagnostic period of July 1, 1993 to June 30, 1995 (Karagas et al. 1998a). We selected additional controls for bladder cancer cases for the intervening diagnostic period, frequency matched to these cases on age (25–34, 35–44, 45–54, 55–64, 65–69, 70–74 years) and gender. All controls less than 65 years of age were selected using population lists obtained from the New Hampshire Department of Transportation. The file contains the names and addresses of those holding a valid driver’s license for the state of New Hampshire. Controls 65 years of age and older were chosen from data files provided by the Centers for Medicare & Medicaid Services (CMS) of New Hampshire. The method of control selection used in our study has been successfully employed in other case-control studies conducted in the region (e.g. Karagas et al. (Karagas et al. 1998a)). We interviewed a total n = 665 controls, which was 70% of the controls confirmed to be eligible for the study. Of the potential participants, (n = 18) were reported as deceased by a member of the household, (n = 17) no answer after 40 attempts distributed over day, evenings and weekends, (n = 261) declined, (n = 29) were mentally incompetent or too ill to take part.

Personal interview

Informed consent was obtained from each participant and all procedures and study materials were approved by the Committee for the Protection of Human Subjects at Dartmouth College. Consenting participants underwent a detailed in-person interview, usually at their home. Questions covered sociodemographic information (including level of education), lifestyle factors such as use of tobacco (including frequency, duration and intensity of smoking), family history of cancer, and medical history prior to the diagnosis date of the bladder cancer (cases) or reference date assigned to controls. Recruitment procedures for both the shared controls from the non-melanoma skin cancer study and additional controls were identical and ongoing concomitantly with the case interviews. Case-control status and the main objectives of the study were not disclosed to the interviewers. To ensure consistent quality of the study interviewer, interviews were tape recorded with the consent of the participants and routinely monitored by the interviewer supervisor. To assess comparability of cases and controls, we asked subjects if they currently held a driver’s license or a Medicare enrollment card. Subjects were asked to provide a blood sample (buccal sample was requested in the case of a refusal) and a toenail sample. Samples were maintained at 4°C and sent via courier to the study laboratory at Dartmouth within 24 hours for processing and analysis. More than 90% of the toenail samples were collected at the time of interview (>95% within 12 months of interview) and the majority of cases were interviewed within 14 months of diagnosis.

Arsenic analysis

Toenail arsenic levels are an internal biomarker of exposure over a 6–12 month period that correlates with drinking water arsenic concentration (Karagas et al. 2000; Slotnick et al. 2007). Toenail clipping samples collected at the time of interview were analyzed for arsenic and other trace elements by Instrumental Neutron Activation Analysis (INAA) at the University of Missouri Research Reactor, using a standard comparison approach as described previously (Karagas et al. 1998b; Wang et al. 2007). The detection limit for arsenic measured by INAA is approximately 0.001 μg/g. A total of 1068 participants provided toenail samples for analysis. Data from our previous analyses of cancer risk in this U.S. population suggested that the risk did not increase linearly across the observed arsenic concentrations. Rather the graphs suggested that risk elevation began in the more highly exposed individuals (Karagas et al. 2004a; Karagas et al. 2001). Thus, groups were created for high versus low arsenic exposure analyses using the 90th percentile as a cutoff, which is a toenail arsenic level of 0.191 μg/g. The mean toenail arsenic level in the lower 90th percentile was 0.085 (range 0.009–0.191), while the upper 10th percentile mean was 0.363 (range 0.192–2.484).


DNA was isolated from peripheral circulating blood lymphocyte specimens harvested at the time of interview using Qiagen genomic DNA extraction kits (QIAGEN Inc., Valencia, CA). Genotyping for non-synonymous SNPs XRCC3 C/T at position 241, ERCC2 A/C at position 751 and G/A at position 312 was performed by Qiagen Genomics using their SNP mass-tagging system(Kelsey et al. 2004). Genotyping for other SNPs was performed using the GoldenGate Assay system through Illumina’s Custom Genetic Analysis service (Illumina, Inc., San Diego, CA). Of the 1113 participating cases with confirmed bladder cancer and controls, we genotyped DNA on 915 (82%) for the current study.

Statistical analysis

Analysis was performed on subjects with both genotype and toenail arsenic data (549 controls and 342 cases). Characteristics of this population are similar to that of the overall study population. To assess the relation between DNA repair gene SNPs and bladder cancer risk, we conducted logistic regression analyses for individuals with one or two variant alleles in comparison to those homozygous wild type for each individual SNP. We assessed linkage disequilibrium in homozygotes using a chi-square test and inferred haplotypes using PHASE 2.1(Stephens et al. 2001). Analyses were stratified by age, gender, and smoking status (never, former, current). Interactions were evaluated by including interaction terms in the regression model. Statistical significances of the interactions were assessed using likelihood ratio tests comparing the models with and without interaction terms.


Demographic characteristics of the study population are shown in Table 1. A large percentage of the subjects were male. In addition, participants tended to be mainly Caucasian with a mean age of 62. Our population-based study enrolled cases diagnosed at a variety of facilities in the state of New Hampshire with 71% of the participants being diagnosed with non-invasive tumors. Current smoking was more common among cases than controls.

Table 1
Population characteristics

Table 2 shows bladder cancer risk stratified by toenail arsenic exposure level for XRCC3 and ERCC2. Elevated bladder cancer risk was associated with toenail arsenic levels >90th percentile in XRCC3 heterozygotes/variants compared with wildtype (adjusted OR 2.8 (1.1–7.3), but not among those with low arsenic exposure (OR 0.8 (0.6–1.1)) (gene-environment interaction p-value 0.01). The risk associated with variant XRCC3 genotype was even more elevated in the top 5th percentile, but with limited statistical precision (toenail arsenic level above 0.278 μg/g adjusted OR 3.5 95%CI (0.9–14)). Haplotype analysis (Table 3) further supported the hypothesis that XRCC3 genotype modifies arsenic-related bladder cancer risk. The risk for the 100 haplotype varies by arsenic exposure level above and below the 90th percentile (interaction p-value = 0.04). Results were not altered when our analysis was restricted to Caucasians which comprised 97% of our study population.

Table 2
Bladder cancer risk associated with XRCC3 and ERCC2 polymorphisms by toenail arsenic level dichotomized at the 90th percentile.
Table 3
Haplotypes and interaction p-values by gene.

We further explored arsenic-genotype relationships stratified by smoking status. Risk elevation for XRCC3_241 was appeared largely restricted to current and former smokers (≥90th percentile toenail arsenic level, adjusted OR 2.9 (1.0– 8.2)) as compared to never smokers ((≥90th percentile toenail arsenic level, adjusted OR 1.3 (0.1– 12))(Table 4). With limited statistically precision, we additionally observed a 2.3 fold higher risk associated with the ERCC2 D312N polymorphism in arsenic exposed individuals (OR 2.3 (0.8–7.0) (Table 2), that also appeared higher among smokers (OR 2.7 (0.8–9.0)) than non-smokers (OR 1.4 (0.1–19)). Odds ratios for the other XRCC3 or ERCC2 SNPs did not appear to differ by arsenic exposure status above or below the top decile.

Table 4
Bladder cancer risk associated with XRCC3 241 and toenail arsenic by smoking status.


Consistent with previous studies in other cancers, we observed evidence of arsenic modification of bladder cancer risk by DNA repair gene polymorphisms. In particular we found evidence of gene-environment interaction between the XRCC3 variant genotype and high toenail arsenic levels on bladder cancer risk, especially among smokers. ERCC2 variants with toenail arsenic levels in the top decile also had a slightly higher risk of bladder cancer, but our estimates were imprecise. These data are consistent with our previous report of an elevation in overall bladder cancer risk among smokers with high levels of arsenic exposure (Karagas et al. 2004b) and extend our earlier finding of an increased risk of bladder cancer among those with a variant XRCC3 genotype (Andrew et al. 2007). Together these data support the hypothesis that other sources of DNA damage may interact with arsenic to induce tumorigenesis. This damage can be caused by high levels of smoking-induced lesions or an accumulation of lesions due to polymorphisms that decrease DNA repair function.

The importance of DNA damage and repair pathways for arsenic carcinogenicity is highlighted by its co-carcinogenic potential in the presence of UV irradiation and other mutagenic-carcinogens (Karagas et al. 2004b; Pott et al. 2001; Rossman et al. 2001). Arsenic has also been shown to inhibit DNA repair enzymes, possibly through a mechanism that involves displacement of zinc ions from zinc fingers (Andrew et al. 2003; Hartwig 1998; Hartwig et al. 1997; Li and Rossman 1989; Yager and Wiencke 1997) (Andrew et al. 2006b; Andrew et al. 2003). DNA repair pathways may have some redundancy and be able to compensate for one another to ensure that lesions are repaired despite impaired activity of a single enzyme. An interaction could occur between a polymorphism that hampered DNA repair gene function combined with arsenic inhibition of other DNA repair capabilities. In a study of human lymphocytes, XRCC3 241 variants had higher levels of damage measured as the frequency of micronuclei (Mateuca et al. 2008). Additional mechanistic data and replication in other populations are needed to clarify the nature of the interaction between arsenic exposure and genetic variations in DNA repair genes, particularly with respect to the ability to repair tobacco associated DNA damage. Benzo(a)pyrene and arsenic synergize to induce DNA adducts in the lungs and skin of exposed mice providing evidence of arsenic – polycyclic aromatic hydrocarbon (PAH) interactions (Evans et al. 2004).

Limitations of this study include the fact that there were a small number of highly exposed individuals. While our arsenic measurements does not indicate the specific source of the exposure, we believe that toenail arsenic serves as a good indicator for arsenic exposure because it integrates all sources of exposure and metabolic differences between individuals over a long period of time. Future studies are needed to replicate these findings in other arsenic exposed populations and elucidate the mechanism by which arsenic interacts with the DNA double-strand break repair gene XRCC3.


Grant support: This publication was funded in part by grant numbers CA102327, CA121382, CA099500, CA82354, CA57494, 5 P42 ES05947, RR018787, CA078609, and ES007373 from the National Cancer Institute, NIH and from the National Institute of Environmental Health Sciences, NIH.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Abernathy CO, Liu YP, Longfellow D, Aposhian HV, Beck B, Fowler B, Goyer R, Menzer R, Rossman T, Thompson C, Waalkes M. Arsenic: health effects, mechanisms of actions, and research issues. Environ Health Perspect. 1999;107:593–597. [PMC free article] [PubMed]
  • Ahsan H, Chen Y, Wang Q, Slavkovich V, Graziano JH, Santella RM. DNA repair gene XPD and susceptibility to arsenic-induced hyperkeratosis. Toxicol Lett. 2003;143:123–31. [PubMed]
  • Andrew A, Burgess J, Meza M, Demidenko E, Waugh M, Hamilton J, Karagas M. Arsenic Exposure Is Associated with Decreased DNA Repair in Vitro and in Individuals Exposed to Drinking Water Arsenic. Environmental Health Perspectives. 2006a;114:1193–1198. [PMC free article] [PubMed]
  • Andrew AS, Burgess JL, Meza MM, Demidenko E, Waugh MG, Hamilton JW, Karagas MR. Arsenic exposure is associated with decreased DNA repair in vitro and in individuals exposed to drinking water arsenic. Environ Health Perspect. 2006b;114:1193–8. [PMC free article] [PubMed]
  • Andrew AS, Karagas MR, Hamilton JW. Decreased DNA repair gene expression among individuals exposed to arsenic in United States drinking water. Int J Cancer. 2003;104:263–268. [PubMed]
  • Andrew AS, Karagas MR, Nelson HH, Guarrera S, Polidoro S, Gamberini S, Sacerdote C, Moore JH, Kelsey KT, Demidenko E, Vineis P, Matullo G. DNA Repair Polymorphisms Modify Bladder Cancer Risk: A Multi-factor Analytic Strategy. Hum Hered. 2007;65:105–118. [PMC free article] [PubMed]
  • Applebaum KM, Karagas MR, Hunter DJ, Catalano PJ, Byler SH, Morris S, Nelson HH. Polymorphisms in nucleotide excision repair genes, arsenic exposure, and non-melanoma skin cancer in New Hampshire. Environ Health Perspect. 2007;115:1231–6. [PMC free article] [PubMed]
  • Ayotte JD, Baris D, Cantor KP, Colt J, Jr, Lubin GRR, Karagas JH, Hoover M, Joseph RN, Fraumeni F, Jr, Silverman DT. Bladder Cancer Mortality and Private Well Use in New England: An Ecological Study. Journal of Epidemiology Community Health. 2005;60:168–172. [PMC free article] [PubMed]
  • Ayotte JD, Montogomery DL, Flanagan SM, Robinson KW. Arsenic in groundwater in eastern New England: occurrence, controls, and human health implications. Environ Sci Technol. 2003;37:2075–2083. [PubMed]
  • Cantor KP, Ward MH, Moore L, Lubin J. Water Contaminants. Oxford Univeristy Press; 2006.
  • Chen CJ, Chuang YC, You SL, Lin TM, Wu HY. A retrospective study on malignant neoplasms of bladder, lung and liver in blackfoot disease endemic area in Taiwan. Br J Cancer. 1986;53:399–405. [PMC free article] [PubMed]
  • Evans CD, LaDow K, Schumann BL, Savage RE, Jr, Caruso J, Vonderheide A, Succop P, Talaska G. Effect of arsenic on benzo[a]pyrene DNA adduct levels in mouse skin and lung. Carcinogenesis. 2004;25:493–497. [PubMed]
  • Hartwig A. Carcinogenicity of metal compounds: possible role of DNA repair inhibition. Toxicol Lett. 1998;102–103:235–239. [PubMed]
  • Hartwig A, Groblinghoff UD, Beyersmann D, Natarajan AT, Filon R, Mullenders LH. Interaction of arsenic(III) with nucleotide excision repair in UV-irradiated human fibroblasts. Carcinogenesis. 1997;18:399–405. [PubMed]
  • IARC. Some drinking-water disinfectants and contaminants, including arsenic. IARC monographs on the evaluation of carcinogenic risks to humans; Lyon, France. 2004. pp. 39–267.
  • Johansson SL, Cohen SM. Epidemiology and etiology of bladder cancer. 1997;13:291–298. [PubMed]
  • Kapaj S, Peterson H, Liber K, Bhattacharya P. Human health effects from chronic arsenic poisoning--a review. J Environ Sci Health A Tox Hazard Subst Environ Eng. 2006;41:2399–2428. [PubMed]
  • Karagas M, Tosteson T, Morris S, Demidenko E, Mott L, Heaney J, Schned A. Incidence of Transitional Cell Carcinoma of the Bladder and Arsenic Exposure in New Hampshire. Cancer Causes and Control. 2004a;15:465–472. [PubMed]
  • Karagas MR, Stukel TA, Morris JS, Tosteson TD, Weiss JE, Spencer SK, Greenberg ER. Skin cancer risk in relation to toenail arsenic concentrations in a US population-based case-control study. Am J Epidemiol. 2001;153:559–565. [PubMed]
  • Karagas MR, Tosteson TD, Blum J, Klaue B, Weiss JE, Stannard V, Spate V, Morris JS. Measurement of low levels of arsenic exposure: a comparison of water and toenail concentrations. Am J Epidemiol. 2000;152:84–90. [PubMed]
  • Karagas MR, Tosteson TD, Blum J, Morris JS, Baron JA, Klaue B. Design of an epidemiologic study of drinking water arsenic exposure and skin and bladder cancer risk in a U.S. population. Environ Health Perspect. 1998a;106 Suppl 4:1047–1050. [PMC free article] [PubMed]
  • Karagas MR, Tosteson TD, Blum J, Morris JS, Baron JA, Klaue B. Design of an epidemiologic study of drinking water arsenic exposure and skin and bladder cancer risk in a U.S. population. Environ Health Perspect. 1998b;106(Suppl 4):1047–50. [PMC free article] [PubMed]
  • Karagas MR, Tosteson TD, Morris JS, Demidenko E, Mott LA, Heaney J, Schned A. Incidence of Transitional Cell Carcinoma of the Bladder and Arsenic Exposure in New Hampshire. Cancer Causes Control. 2004b;15:465–472. [PubMed]
  • Kelsey KT, Park S, Nelson HH, Karagas MR. A population-based case-control study of the XRCC1 Arg399Gln polymorphism and susceptibility to bladder cancer. Cancer Epidemiol Biomarkers Prev. 2004;13:1337–1341. [PubMed]
  • Li JH, Rossman TG. Inhibition of DNA ligase activity by arsenite: a possible mechanism of its comutagenesis. Mol Toxicol. 1989;2:1–9. [PubMed]
  • Mateuca RA, Roelants M, Iarmarcovai G, Aka PV, Godderis L, Tremp A, Bonassi S, Fenech M, Berge-Lefranc JL, Kirsch-Volders M. hOGG1(326), XRCC1(399) and XRCC3(241) polymorphisms influence micronucleus frequencies in human lymphocytes in vivo. Mutagenesis. 2008;23:35–41. [PubMed]
  • Peters SC, Blum J, Klaue B, Karagas MR. Arsenic occurrence in New Hampshire groundwater. Environmental Science and Technology. 1999;33:1328–1333.
  • Pott WA, Benjamin SA, Yang RS. Pharmacokinetics, metabolism, and carcinogenicity of arsenic. Rev Environ Contam Toxicol. 2001;169:165–214. [PubMed]
  • Rossman T. Mechanism of Arsenic Carcinogenesis: An Integrated Approach. Mutation Research. 2003;533:37–65. [PubMed]
  • Rossman TG, Uddin AN, Burns FJ, Bosland MC. Arsenite is a cocarcinogen with solar ultraviolet radiation for mouse skin: an animal model for arsenic carcinogenesis. Toxicol Appl Pharmacol. 2001;176:64–71. [PubMed]
  • Slotnick MJ, Meliker JR, Avruskin GA, Ghosh D, Nriagu JO. Toenails as a biomarker of inorganic arsenic intake from drinking water and foods. J Toxicol Environ Health A. 2007;70:148–58. [PubMed]
  • Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet. 2001;68:978–989. [PubMed]
  • Thirumaran RK, Bermejo JL, Rudnai P, Gurzau E, Koppova K, Goessler W, Vahter M, Leonardi GS, Clemens F, Fletcher T, Hemminki K, Kumar R. Single nucleotide polymorphisms in DNA repair genes and basal cell carcinoma of skin. Carcinogenesis. 2006;27:1676–81. [PubMed]
  • Wang X, Zhang S, MacLennan GT, Eble JN, Lopez-Beltran A, Yang XJ, Pan CX, Zhou H, Montironi R, Cheng L. Epidermal growth factor receptor protein expression and gene amplification in small cell carcinoma of the urinary bladder. Clin Cancer Res. 2007;13:953–7. [PubMed]
  • Yager JW, Wiencke JK. Inhibition of poly(ADP-ribose) polymerase by arsenite. Mutat Res. 1997:345–351. [PubMed]