Genetic Polymorphisms in XRCC1, APE1, ADPRT, XRCC2, and XRCC3 and Risk of Chronic Benzene Poisoning in a Chinese Occupational Population

As an important component of many organic solvents, benzene is also commonly used to synthesize organic chemicals. The major industries using benzene are those involved in the production of rubber, paint, shoes, lubricants, dyes, detergents, drugs, and pesticides (1). Humans are exposed to benzene via either occupational exposure or environmental exposure to cigarette smoke, gasoline emissions, or products of incomplete combustion (2). Exposure to benzene may result in chronic benzene poisoning (CBP) characterized by hematotoxicities, leading to pancytopenia, aplastic anemia, myelodysplastic syndrome, and acute myeloid leukemia (3-7).

In biological systems, inhaled benzene is converted to benzene oxide, which is ultimately metabolized to phenol, catechol, hydroquinone, and benzoquinone (8, 9), and it is in this bioactivation process that substantive reactive oxygen species, such as hydroxyl radicals, are formed (10, 11). Benzene metabolites bind covalently to macromolecules, such as proteins and DNA in the tissue, and therefore cause genotoxicity (10, 12) as a result of oxidative damage to DNA or DNA double-strand breaks caused by reactive oxygen species.

Previous mechanistic studies of CBP showed that p-benzoquinone, a stable metabolite of benzene, formed two-ring benzetheno exocyclic base adducts with the bases C, A, and G of DNA, and the p-benzoquinone-C adducts were the substrate for apurinic/apyrimidinic endonuclease (APE1; refs. 13-15), the rate-limiting enzyme in base-excision repair (16). After excision of these adducts, other components involved in base-excision repair [i.e., X-ray repair cross-complementing group 1 (XRCC1), ADP ribosyltransferase (ADPRT), DNA polymerase β, and DNA ligase III] will complete this repair process (17). Strongly induced by the presence of single DNA strand breaks, ADPRT interacts with XRCC1 protein to participate in DNA repair and in the recovery of cells from DNA damage (18-21).

In cultured HL-60 cells treated with trans, trans-muconaldehyde, hydroquinone, and their mixtures, the formation of DNA-protein cross-links and DNA strand breaks was observed (12), and increased homologous recombination repair was detected in cultured Chinese hamster ovary cells administrated with phenol, catechol, and benzoquinone (22). These results indicate that DNA repair genes involved in base-excision repair, such as XRCC1, ADPRT, and APE1, and in homologous recombination repair, such as X-ray repair cross-complementing group 2 (XRCC2) and X-ray repair cross-complementing group 3 (XRCC3) genes, play a role in repair of benzene-induced DNA damage. Therefore, genetic variation in these genes may be associated with risk of developing CBP in the exposed population.

A number of single-nucleotide polymorphisms have been identified in XRCC1, APE1, ADPRT, XRCC2, and XRCC3, some of which are relatively common, including single-nucleotide polymorphisms at codons 194 (Arg to Trp), 280 (Arg to His), and 399 (Arg to Gln) of XRCC1 (23), codon 148 (Asp to Glu) of APE1 (24), codon 762 (Ala to Val) of ADPRT (25, 26), codon 188 (Arg to His) of XRCC2 (27), and codon 241 (Thr to Met) of XRCC3 (27, 28). These single-nucleotide polymorphisms have been studied for their roles in the development of cancers of the breast, bladder stomach, and lung (29-33). However, whether or not these common genetic polymorphisms may affect risk of CBP is still unknown. Therefore, using the case-control design and PCR-RFLP technique, we investigated the associations between these genetic polymorphisms and risk of developing CBP in a Chinese occupational population who had been exposed to benzene in the workplace.

The detailed information of the study population has been reported previously (34). Briefly, 112 of 152 CBP patients in this study were from four factories where clusters of the cases were reported in Shanghai, Hangzhou, Maanshan, and Guangzhou, China. Another 40 patients, who returned to the hospital periodically for health examination, were from other 12 small factories that had been closed down. The benzene poisoning diagnoses were made between 1980 and 1998 by the Authorized Local Occupational Disease Diagnostic Team. The diagnostic criteria for occupational CBP, according to the Ministry of Health, China, include (a) total WBC counts <4,000/μL or WBC counts between 4,000 and 4,500/μL and platelet counts <80,000/μL, with repeated confirmation of these counts in a few months in a peripheral blood examination; (b) documented benzene exposure as a result of an employment for at least 6 months in the factory; and (c) exclusion of other known causes of abnormal blood counts such as chloromycetin use and ionizing radiation. The medical records of these patients were independently reviewed by at least two hemopathologists, especially for those with WBC counts >3,500 to confirm the CBP diagnosis. One hundred fifty-two healthy workers that had been occupationally exposed to benzene in the four factories were selected as controls. The cases and controls were frequency matched on age within 5 years, sex, exposure duration within 3 years, and exposure level at the workplace. The calendar range for start of employment of the cases and the controls was from 1958 to 1998 and from 1963 to 1998, respectively. All the eligible individuals agreed to participate in this study after an informed consent was orally explained and obtained.

The subjects were interviewed by a trained personnel and a questionnaire was used to obtain general information including ethnic background, nutrition, cigarette smoking, alcohol consumption, protective measures, self-reported symptoms, medical history, and occupational history, such as work unit (department), type of work, and exposure duration. Exposure estimation was based on monitoring data obtained by industrial hygienists and long-term employees' evaluation considering historical changes in the workplace (35). The intensity of benzene exposure (milligrams per cubic meter) for the patients was taken as the benzene level of workplaces when diagnoses were made; the intensity of benzene exposure for the controls was taken as the current level monitored by organic vapor passive dosimetry badges during collection of the blood samples from the controls. Those who smoked at least one cigarette per day for more than 1 year were considered regular smokers. Alcohol consumption was defined as drinking at least 7 standard units of alcohol on average per week [1 standard unit = 10 g of alcohol equivalent; e.g., a glass/can/bottle (330 mL) of regular beer (5%), a measure (40 mL) of liquor, a glass (120 mL) of wine] for more than 6 months. Under the criteria, the smokers and alcohol users in cases satisfied it when they were diagnosed CBP whereas those in controls meant current status when they enrolled in the study. The subjects were administrated a rigorous physical examination at a local occupational disease hospital. Alanine aminotransferase level in serum was also examined to assess liver functions.

Blood samples of the subjects enrolled in this study were collected only after informed consent was obtained. Blood was immediately frozen at −80°C after collection and was sent to the laboratory later on dry ice.

Genomic DNA was extracted from blood samples by a routine phenol-chloroform method (36). The genotypes were determined by the PCR-RFLP method and primers, annealing temperature, length of amplified fragments, restriction pattern, and restriction enzymes used are listed in Table 1. The PCR was done using 50 ng of genomic DNA, 0.2 μmol/L of each primer, 1 × PCR buffer, 0.2 μmol/L of each deoxynucleotide triphosphate, 2.0 mmol/L MgCl₂, and 0.625 units of Taq in a 25-μL reaction volume. PCR program was a 5-minute denaturation step at 95°C followed by 35 cycles of 94°C for 50 seconds, 55°C to 69.5°C for 30 to 50 seconds, 72°C for 30 to 50 seconds (according to the length of fragment amplified), respectively, and a final extension step at 72°C for 10 minutes.

The variant alleles at codons 194, 280, and 399 of XRCC1 create MspI, RsaI, and MspI sites, respectively (37, 38); the variant alleles at codon 148 of APE1, codon 762 of ADPRT, codon 188 of XRCC2, and codon 241 of XRCC3 create BfaI, Bsh1236I (FnuDII, BstI), PmacI (Eco721), and NlaIII sites (29), respectively. After digestion at 37°C for 16 hours (MspI, RsaI, BfaI, Bsh1236I, and PmacI are from Fermentas, Inc., Vilnius, Lithuania; BfaI and NlaIII are from New England Biolabs, Inc., Beverly, MA) and resolved on 2% to 3% Metaphor agarose gels (BBI, Toronto, Canada). The digested PCR products were observed under UV image system (Gel Doc 2000, Segrate, Milan, Italy). All genotypes were evaluated and agreed on by at least two persons independently. Ten percent of DNA samples were selected randomly for repeat and the concordance was 100%.

Student's t test was used to compare differences between continuous variables, such as age and exposure duration, and χ² test was used to examine the differences in the distribution of genetic polymorphisms between cases and controls. The odds ratio (OR) and 95% confidence interval (95% CI) for estimating the associations between genotypes of XRCC1, APE1, ADPRT, XRCC2, and XRCC3 and risk of CBP were obtained from both univariate and multivariate unconditional logistic regression models with adjustment for potential confounding factors, such as age, sex, exposure duration, and intensity of benzene exposure (adjusted OR and 95% CI). Haplotype analysis was conducted with PHASE 2.0.2 software (University of Washington, Seattle, WA). All tests were done by using SPSS 10.0 software (SPSS, Inc., Chicago, IL).

The distributions of age, sex, exposure duration, intensity of benzene exposure, type of work, cigarette smoking, and alcohol consumption in cases and controls are presented in Table 2. Overall, there was slightly difference in the distributions of these variables between cases and controls but none of them were statistically significant. For example, the median age and exposure duration were 36.78 (range, 21.00-61.00) and 12.97 (range, 1.00-38.00) in the 152 CBP cases, respectively, and 38.25 (range: 19.00-57.00) and 11.10 (range: 1.00-36.00) in the 152 controls (data not shown), respectively. There was no significant difference in the distribution of age groups (≤25, 26-35, 36-45, >45 years), exposure duration (≤5, 6-10, 11-15, 16-20, >25 years), and intensity of benzene exposure (≤40, 41-100, >100 mg/m³; P > 0.05 for all). However, the percentage of female subjects (59.9%) was higher than that of males (40.1%) in both cases and controls (62.5% versus 37.5%). This higher female ratio may explain the relatively low frequency of cigarette smoking (17.4%) and alcohol consumption (11.5%) in the subjects than expected from the general population because the frequency of cigarette smoking and alcohol consumption among men is much higher than among women (39-41). The range for total WBC counts in the controls was from 4,500/μL to 5,900/μL; the mean was 4,800 ± 260/μL.

The frequencies of XRCC1, APE1, ADPRT, XRCC2, and XRCC3 variant alleles and genotypes in cases and controls are presented in Table 3. No XRCC2 codon 188 variant genotype was detected, the frequency of XRCC3 241Met allele was low (0.06), and no Met/Met genotype was detected. However, the variant allele and genotype frequency of XRCC1 Arg280His polymorphism were significantly different between cases and controls (P < 0.01) but the variant allele and genotype frequencies of XRCC1 Arg194Trp, XRCC1 Arg399Gln, APE1 Asp148Glu, ADPRT Val762Ala, and XRCC3 Thr241Met were not (P > 0.05 for all). The distributions of the genotypes of these genetic polymorphisms in the controls were all in Hardy-Weinberg equilibrium (P > 0.05, χ² test).

No statistical difference was found in the distribution of genotypes of XRCC1 Arg399Gln, APE1 Asp148Glu, ADPRT Val762Ala, and XRCC3 Thr241Met (P > 0.05). However, the proportion of individuals carrying XRCC1 194Arg/Trp+Trp/Trp was lower in cases (50.7%) than in controls (62.2%; P = 0.041) and the proportion for individuals carrying XRCC1 280Arg/His+His/His genotypes was higher in cases (52.1%) than in controls (36.0%; P = 0.009). Correspondingly, the risk of CBP for individuals carrying of XRCC1 194Arg/Trp+Trp/Trp genotypes was 1.67-fold lower (OR_adj, 0.60; 95% CI, 0.37-0.98; χ² = 4.18, P = 0.041) compared with those carrying the Arg/Arg genotype after adjustment for age, sex, exposure duration, and intensity of benzene exposure, respectively. In contrast, individuals with XRCC1 280Arg/His and XRCC1 280Arg/His+His/His genotypes had 1.67-fold (OR_adj, 1.67; 95% CI, 1.02-2.74; χ² = 4.18, P = 0.04) and 1.91-fold (OR_adj, 1.91; 95% CI, 1.17-3.10; χ² = 6.74, P = 0.009) increased risk of CBP, respectively, compared with those with the Arg/Arg genotype (Table 3). The risk of CBP stratified by smoking and alcohol use suggested that significantly increased risk associated with the XRCC1 280Arg/His+His/His genotypes was confined to nonsmokers (OR_adj, 1.96; 95% CI, 1.14-3.38) and nonalcohol users (OR_adj, 1.78; 95% CI, 1.05-3.03) whereas reduced risk of CBP associated with the APE1 148Asp/Glu+Glu/Glu genotypes was confined to alcohol users (OR_adj, 0.11; 95% CI, 0.02-0.69). However, homogeneity test of ORs for these effects stratified by smoking and alcohol consumption suggested that there was potential interaction merely between polymorphism of APE1Asp148Glu and alcohol consumption (χ_H² = 6.93, P = 0.01).

Because we genotyped for three single-nucleotide polymorphisms of XRCC1 and they are reportedly in linkage disequilibrium with each other, we calculated their haplotypes based on the observed genotypes. As shown in Table 4, there were seven haplotypes in this study population. There were a statistically significant difference for the distribution of haplotypes of HAP3 [i.e., (XRCC1 194Arg, 280His, and 399Arg), HAP4 (XRCC1 194Trp, 280Arg, and 399Arg), and HAP7 (XRCC1 194Trp, 280His, and 399Gln); P < 0.05 for all]. Compared with those carrying the HAP1 haplotype, there was a 2.96-fold increased risk of CBP for subjects with the HAP3 haplotype (OR, 2.96; 95% CI, 1.60-5.49; χ² = 12.39, P = 0.001; Table 4).

Previous studies have shown that benzene toxicity mainly resulted from its intermediate reactive metabolites that cause DNA damage by directly oxidative damage and covalently binding to DNA. Previous study with the comet assay showed that the damage levels of lymphocyte and granulocytes were significantly higher in peripheral blood of benzene-exposed group than in the unexposed control group (42). It is postulated that DNA damage induced by benzene exposure and resultant cellular dysfunction, apoptosis, and lethality, which are correlated with the intensity of DNA damage and an individual's susceptibility, may explain the manifestations of CBP such as pancytopenia (43). In the present study, our data further suggested that genetic polymorphisms in DNA repair genes, such as XRCC1, may have played a role in benzene-induced CBP in the workplace in this Chinese population.

The XRCC1 protein is an important component involved in base-excision repair. After excision of a damaged base, it stimulates endonuclease action and acts as a scaffold in the subsequent restoration of the site by complexing with DNA ligase III via a BRCT domain in its COOH terminus and with DNA polymerase β via the NH₂-terminal domain (44, 45) Both XRCC2 and XRCC3 proteins are essential for homologous recombination to recognize and repair the double-strand breaks on DNA strands; lack of XRCC2 or XRCC3 gene will result in decreased function of homologous recombination (46-48).

Genetic polymorphisms are common in the general population. Most polymorphisms are located outside gene boundaries and thus thought not to have any apparent effects on protein functions. If a polymorphism is within the coding region of a gene, amino acid substitution may occur and result in change of protein activity whereas polymorphisms in the promoter will alter the efficiency of transcription and polymorphisms located at an intron/exon boundary in a gene may produce incorrect mRNA splicing, which results in incomplete or inactive proteins (49). Thus, in this study, although no association of polymorphisms in XRCC1 Arg399Gln, APE1 Asp148Glu, ADPRT Val762Ala, XRCC2 Arg188His, and XRCC3 Thr241Met with risk of CBP was found, other polymorphisms in the promoter and coding region of these genes need to be identified, which may also contribute to risk of CBP.

The XRCC1 codon 194 and 280 polymorphisms are located in the vicinity of two sequences mediating the protein-protein interactions with ADPRT and polymerase β (19-21). Thus, the variant alleles may encode a twisted XRCC1 protein, resulting in increased or decreased affinity to other proteins, which may affect normal DNA repair capability. Our results indicated that there was significant difference in distribution of genotypes of XRCC1 Arg194Trp and XRCC1 Arg280His and the risk was reduced in those with XRCC1 194Arg/Trp+Trp/Trp genotypes but increased in those with XRCC1 280Arg/His and XRCC1 280Arg/His+His/His genotypes. These data were consistent with previous reports on risk of cancers of the breast (50), bladder (51), and lung (52, 53). In the analysis of haplotypes of polymorphisms in XRCC1, we also found that workers carrying genotypes of XRCC1 194Arg/Arg, XRCC1 280Arg/His+His/His, and XRCC1 399Arg/Arg simultaneously are more susceptible to CBP. Thus, variants at codons 194 and 280 of XRCC1 may be associated with altered function of the coding products and thus contribute to the individual's genetic susceptibility to CBP, and more attention should be paid on workers occupying jobs with benzene exposure and carrying genotypes of XRCC1 194Arg, XRCC1 280His, and XRCC1 399Arg simultaneously, especially in the health surveillance work. This hypothesis needs to be tested in larger studies.

Most chronic and complex diseases are likely caused by interactions among environmental exposure, genetic polymorphism, and lifestyles (34) such as smoking and alcohol consumption, which have been identified as risk factors for CBP (54, 55). In this study, reduced risk of CBP for alcohol users with APE1 148Asp/Glu+Glu/Glu genotypes was found; however, the matching design may have masked the effect of smoking and alcohol consumption on risk of CBP. The relative low frequency of alcohol consumption and smoking may contribute to results in this study for the majority of subjects involved in this study were females who were less likely than men to be indulged in alcohol and cigarette use, and further study with larger sample is needed to elucidate interactions between polymorphisms in these DNA repair genes and smoking or alcohol use. Furthermore, accurate exposure estimation is quite important for evaluating exposure levels between the case and control groups; however, because not all cases and controls are in the same exposure environment and due to difference in types of work, we evaluated intensity of benzene exposure according to method described by Dosemeci et al. (34) and the exposure levels were measured differently between cases and controls that may not completely reflect personal exposure. Thus, more accurate methods for exposure estimation are needed to be introduced in further study such as individual sampling, which is more accurate than traditional area sampling. Those benzene metabolites reflecting the exposure levels are also needed to be applied in accurate exposure estimation.

In summary, we found for the first time that genetic variants in XRCC1 may contribute to the development of CBP in occupational exposure to benzene in a Chinese population. This finding, once verified by large studies, will have important implication in the prevention of CBP in the susceptible workers. The strengths of this study include a homogeneous ethnic background of the subject, well-documented exposure history to benzene in the workplace, and a frequency-matching design. However, joint action between genetic polymorphisms and environmental exposure on special diseases such as CBP is complicated and small studies like the present study do not have enough statistical power to detect gene-environment interactions. Thus, a more comprehensive, larger-scale study is needed to further explore the effects of gene-environment interaction on genetic susceptibility to CBP.

We thank Duozhi Cao (Institute of Occupational Health attached to Maanshan Steel & Iron Group, Maanshan, Anhui, China), Weiwei Liu (Guangzhou Hospital for Occupational Diseases, Guangzhou, Guangdong, China), and Rong Ye and Jiru Guan (Hangzhou Center for Diseases Control and Prevention, Hangzhou, Zhejiang, China) for their contribution to field research and physical examination in this study; and Dr. Qingyi Wei (Department of Epidemiology, The University of Texas M.D. Anderson Cancer Center, Houston, TX) for his critical review and scientific editing.