Document LMLJJQZ1JL88MpEzqkXL4DVQ

Original article 789 Genetic polymorphisms and benzene metabolism in humans exposed to a wide Range of air concentrations Sungkyoon Kima, Qing Lanb, Suramya Waidyanathaa, Stephen Chanockb,c, Brent A. Johnsona, Roel Vermeulenb, Martyn T. Smithd, Luoping Zhangd, Guilan Lie, Min Shenb, Songnian Yine, Nathaniel Rothmanb and Stephen M. Rappaporta Using generalized linear models with natural-spline smoothing functions, we detected effects of specific xenobiotic metabolizing genes and geneenvironment interactions on levels of benzene metabolites in 250 benzene-exposed and 136 control workers in Tianjin, China (for all individuals, the median exposure was 0.512 p.p.m. and the 10th and 90th percentiles were 0.002 and 6.40 p.p.m., respectively). We investigated five urinary metabolites (E,E-muconic acid, S-phenylmercapturic acid, phenol, catechol, and hydroquinone) and nine polymorphisms in seven genes coding for key enzymes in benzene metabolism in humans {cytochrome P450 2E1 [CYP2E1, rs2031920], NAD(P)H: quinone oxidoreductase [NQO1, rs1800566 and rs4986998], microsomal epoxide hydrolase [EPHX1, rs1051740 and rs2234922], glutathione-S-transferases [GSTT1, GSTM1 and GSTP1(rs947894)] and myeloperoxidase [MPO, rs2333227]}. After adjusting for covariates, including sex, age, and smoking status, NQO1*2 (rs1800566) affected all five metabolites, CYP2E1 (rs2031920) affected most metabolites but not catechol, EPHX1 (rs1051740 or rs2234922) affected catechol and S-phenylmercapturic acid, and GSTT1 and GSTM1 affected S-phenylmercapturic acid. Significant interactions were also detected between benzene exposure and all four genes and between smoking status and NQO1*2 and EPHX1 (rs1051740). No significant effects were detected for GSTP1 or MPO. Results generally support prior associations between benzene hematotoxicity and specific gene mutations, confirm earlier evidence that GSTT1 affects production of S-phenylmercapturic acid, and provide additional evidence that genetic polymorphisms in NQO1*2, CYP2E1, and EPHX1 (rs1051740 or rs2234922) affect metabolism of benzene in the human liver. Pharmacogenetics and Genomics 17:789801 c 2007 Wolters Kluwer Health | Lippincott Williams & Wilkins. Pharmacogenetics and Genomics 2007, 17:789801 Keywords: benzene, CYP2E1, EPHX1, genetic polymorphisms, GSTT1, human, metabolism, MPO, NQO1, splines aSchool of Public Health, University of North Carolina, Chapel Hill, North Carolina, bNational Cancer Institute (NCI), National Institutes of Health (NIH), Department of Health and Human Services (DHHS), cCenter for Cancer Research, Bethesda, Maryland, dSchool of Public Health, University of California, Berkeley, California, USA and eInstitute of Occupational Health and Poison Control, Chinese Center for Disease Control and Prevention, Beijing, China Correspondence to Dr Stephen M. Rappaport, School of Public Health, University of North Carolina, Chapel Hill, NC 27599, USA Tel: + 1 919 966 5017; fax: + 1 919 966 0521; e-mail: stephen_rappaport@unc.edu Sponsorship: This research was supported by the National Institute for Environmental Health Sciences through Grants P42ES05948 and P30ES10126 (S.M.R.) and RO1ES06721, P42ES04705, and P30ES01896 (M.T.S.) and by funds from the intramural research program of the NIH, National Cancer Institute. M.T.S. has received consulting and expert testimony fees from law firms representing both plaintiffs and defendants in cases involving exposure to benzene. S.M.R. has received consulting and expert testimony fees from law firms representing plaintiffs' cases involving exposure to benzene. G.L. has received funds from the American Petroleum Institute for consulting on benzene-related health research. Received 23 August 2006 Accepted 26 October 2006 Introduction Benzene is an important industrial chemical that is ubiquitous in the environment owing to vaporization from petroleum products and combustion of hydrocarbons [13]. Occupational exposures to benzene can cause blood disorders, including aplastic anemia, myelodysplastic syndrome, and acute myelogenous leukemia [47]. Significant decreases in the numbers of white blood cells and platelets have been reported in workers exposed to less than one p.p.m. of benzene in air [8]. Although these toxic effects are related to metabolism of benzene in the liver, the particular metabolite(s) that damage bone marrow cells and the mode of toxic action are subjects of debate [911]. 1744-6872 c 2007 Wolters Kluwer Health | Lippincott Williams & Wilkins Since the pioneering work of Parke and Williams' [12,13], the metabolism of benzene has been extensively investigated (reviewed in [9,10]). The major metabolic pathways, shown in Fig. 1, begin with cytochrome P450 CYP2E1-mediated oxidation of benzene to benzene oxide (BO), which is in equilibrium with its tautomer, oxepin. BOoxepin is the source of all other major metabolites, namely, phenol (PH), E,E-muconic acid (MA), hydroquinone (HQ), and catechol (CA), and the minor product, S-phenylmercapturic acid (SPMA). All of these metabolites are excreted in urine, either free or in conjugated form. Additional metabolism of the primary metabolites produces additional electrophilic species, including the muconaldehydes (from CYP oxidation of Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 790 Pharmacogenetics and Genomics 2007, Vol 17 No 10 Fig. 1 Benzene CYP2E1 OCH HOOC S O CH2 C N CH CH3 COOH S-Phenylmercapturic acid via GST O Benzene oxide O Oxepin CYP CHO COOH E,E-muconaldehyde E,E-muconic acid Nonenzymatic rearrangement Epoxide hydrolase OH Phenol CYP2E1 OH OH CYP OH OH Benzene dihydrodiol OH O Benzene diol expoxide Dihyrodiol dehydrogenase OH HO Hydroquinone NQO1 MPO O CYP2E1 OH OH Catechol MPO NQO1 O O 1,4-Benzoquinone HO OH 1,2,4-Trihydroxybenzene O 1,2-Benzoquinone Simplified metabolic scheme for benzene showing major pathways and metabolizing genes. oxepin followed by ring opening), and 1,2-benzoquinone and 1,4-benzoquinone (BQ) (from spontaneous or peroxidase-mediated oxidation of CA and HQ, respectively). thought to catalyze transformations between CA and HQ and the corresponding quinones (1,2-BQ and 1,4-BQ, respectively) [10,14,19,20]. As shown in Fig. 1, numerous enzyme systems are involved in the metabolism of benzene and its metabolites. In addition to the CYP oxidations of benzene (to BO), oxepin (to the muconaldehydes and ultimately MA) and PH (to HQ) [1416], microsomal epoxide hydrolase (EPHX) catalyzes the hydrolysis of BO to initiate the CA pathway [14,17], various glutathione-Stransferases (GSTs) catalyze production of SPMA [18], and NAD(P)H: quinone oxidoreductase (NQO1) and peroxidases [notably myeloperoxidase (MPO)] are It has been speculated that polymorphic genes of the above enzymes predispose some individuals to benzene toxicity through metabolism [2123]. In particular, individuals with wild-type MPO (more active) and a variant of NQO1 (less active) were found to be at greater risk of reduced numbers of white blood cells at low levels of benzene exposure [8]. Yet, although GSTT1 polymorphisms have been shown to affect production of the minor metabolite SPMA [2426], there is only sketchy evidence that the major metabolites (PH, MA, HQ, Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 791 and CA) are affected by polymorphic forms of CYP2E1, EPHX, NQO1, or MPO. A major difficulty in elucidating the connections between genotypes of metabolizing genes and the corresponding in-vivo phenotypes has been the inability to control for the effects of benzene exposure, and important physiological and lifestyle factors, in observational studies. Indeed, the relationship between metabolite levels and benzene exposure is highly nonlinear and is significantly affected by sex, age, smoking status, and body mass index (BMI) [27,28]. Using generalized linear models (GLM) with natural spline (NS) smoothing functions, we were able to elucidate the effects of sex, age, smoking status and BMI, after adjustment for benzene exposure (between 0.003 and 88.9 p.p.m.) in a sample of 326 individuals, who were exposed to benzene occupationally and environmentally in Tianjin, China [28]. In the current paper, we extend our application of GLM + NS models to investigate the effects of polymorphic forms of CYP2E1, EPHX, NQO1, MPO, and GSTs on urinary levels of PH, MA, HQ, CA, and SPMA in the same population. Materials and methods Study population, and air and biological sample collection Exposed and control study participants were recruited with informed consent from two shoe-making factories and three clothes-manufacturing factories, respectively, in Tianjin, China as described by Lan et al. [8]. Characteristics of the workplaces and levels of benzene exposure have been described previously in detail [8,27,29]. After excluding four controls, who were missing measurements of metabolites and/or exposures, the samples included 250 exposed individuals and 136 controls. Exposed and control participants were frequency-matched with respect to sex. Table 1 shows summary statistics for the sex, age and smoking status of participants. Methods of sampling air and urine were also described previously [8,27,29]. Briefly, personal full-shift air measurements were matched with urine samples after shift from exposed and control workers. Of the 386 participants in this analysis, 139 had repeated measurements of air and urine, making a total of 617 matched air/ urine samples. Among participants with repeated measurements, the median number of paired air and urine samples was 3 (range 24). Information about height, weight, smoking status and other relevant factors were obtained by questionnaire [8]. This study was approved by the Institutional Review Boards of the University of North Carolina, the University of California, Berkeley, the US National Cancer Institute and the Chinese Academy of Preventive Medicine. Measurements of air and urinary analytes Measurements of analytes in air and urine were described previously [2729]. Briefly, benzene and toluene were measured in air using passive personal monitors (Organic Vapor Monitors, 3M, St Paul, Minnesota, USA) followed by solvent desorption and gas chromatography [29]. Air measurements of benzene and toluene were below limits of detection [(LOD), nominally 0.2 p.p.m. for benzene and 0.3 p.p.m. for toluene] or were missing for all controls and for some exposed participants (missing values, n = 23; measurements below the LOD, n = 70 for benzene and n = 67 for toluene). Air levels were predicted in these censored and missing air samples from the corresponding urinary levels of benzene and toluene, as described previously for benzene [28]. As summarized in Table 1, the median air level of benzene was 0.512 p.p.m., the 10th percentile level was 0.002 p.p.m. and the 90th percentile was 6.40 p.p.m. The median air level of toluene was 1.77 p.p.m. Urinary benzene was determined by gas chromatography mass spectrometry (GCMS) using headspace solidphase microextraction according to Waidyanatha et al. Table 1 Demographics, benzene exposure and other characteristics of the study population (n = 386) Occupational exposurea Controls (%) Exposed (%) Benzene exposureb Air benzene (p.p.m.) Age (years)b BMI (kg/m2)b Current smoking statusa Nonsmoker (%) Smoker (%) Toluene exposurea Low [ r 1.77 p.p.m. (%)] High [ > 1.77 p.p.m. (%)] Women 84 (61.8) 164 (65.6) 0.517 (0.0026.63) 31 (2144) 21.9 (18.727.0) 240 (81.4) 8 (8.8) 130 (67.4) 118 (61.1) Men 52 (38.2) 86 (34.4) 0.487 (0.0023.78) 24 (2039) 22.0 (18.426.5) 55 (18.6) 83 (91.2) 63 (32.6) 75 (38.9) BMI, body mass index. aNumber (percent). bMedian (1090th percentiles). All 136 (100.0) 250 (100.0) 0.512 (0.0026.40) 28 (2143) 21.9 (18.726.9) 295 (100.0) 91 (100.0) 193 (100.0) 193 (100.0) Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 792 Pharmacogenetics and Genomics 2007, Vol 17 No 10 Table 2 Distributions of genetic polymorphisms among participants in the study Gene name, SNP region, SNP ID (notes) Genotype No. participants CYP2E1, 1054C-T, Rs2031920, (ascribed to RsaI) NQO1, Ex6 + 40C-T, Rs1800566, (NQO1*2) NQO1, Ex4 3C-T, Rs4986998, (NQO1*3) MPO, 642G-A, Rs2333227 GSTM1, del{GSTM1}, n/a GSTT1, del{GSTT1}, n/a GSTP1, Ex5 24A-G, Rs947894 EPHX1, Ex3 28T-C, Rs1051740, (Y113H) EPHX1, Ex4 + 52A-G, Rs2234922, (H139R) C/C C/T T/T C/C C/T T/T C/C C/T T/T G/G G/A A/A +/+ +/ / +/+ +/ / A/A A/G G/G T/T T/C C/C A/A A/G G/G 239 127 19 105 173 107 359 24 0 297 76 11 28 141 210 33 187 163 224 142 16 143 176 57 302 71 5 SNP, single nucleotide polymorphism. (%) (62.1) (33) (4.9) (27.3) (44.9) (27.8) (93.7) (6.3) (0.0) (77.3) (19.8) (2.9) (7.4) (37.2) (55.4) (8.6) (48.8) (42.6) (58.6) (37.2) (4.2) (38) (46.8) (15.2) (79.9) (18.8) (1.3) Presumed phenotype Active Less active Least active Active Less active Inactive Active Less active Least active Active Less active Least active Conjugator Reference [33] [34] [35] [36] [37] [38] Null Conjugator [38] Null Conjugator [39] Less active Normal Slow metabolizer Normal Rapid metabolizer [40] [41] [40] [41] [30]. Urinary PH, CA, HQ, MA, and SPMA were measured as trimethylsilylether derivatives by GCMS, after digestion of urine to release conjugates, according to Waidyanatha et al. [31]. Quantification of all urinary analytes was based on peak areas relative to the corresponding isotopically labeled internal standards. The minor metabolite, SPMA, wpasffiffinot detected in 30 urine specimens; a value of LOD= 2 0:591nmol=l was imputed to these samples [32]. Genotyping We selected nine polymorphisms in seven genes coding for key enzymes in benzene metabolism, on the basis of the evidence of functionality in experimental or human studies (described in the discussion section). As summarized in Table 2 [3341], the following genetic polymorphisms were chosen: CYP2E1 (rs2031920: CT), two alleles of NQO1 [NQO1*2 (rs1800566: C-T) and NQO1*3 (rs4986998: C-T)], MPO (rs2333227: G-A), GSTM1, GSTT1, and GSTP1 (rs947894: A-G), and two alleles of EPHX1 [(rs1051740: T-C) and (rs2234922: A-G), respectively] [42]. Genotyping was performed with an ABI 7900HT detection system using TaqMan end points as described on the website, http:// snp500cancer.nci.nih.gov [43]. The numbers of participants with each polymorphism of the various metabolizing genes are summarized in Table 2. Quality control procedures have been described previously [44]. In brief, blind replicate samples were randomly interspersed throughout the study sample plates and showed intrasubject agreement > 99% for all genotype assays. Statistical analyses Relationships between levels of the urinary metabolites and the corresponding air concentrations of benzene were examined using GLM + NS models, as described previously [28]. The smoothing functions were based upon 5-knot models for all metabolites, after comparing NS models with 57 knots by visual inspection and corrected Akaike's Information Criteria (AICc) [45]. (The candidate knots in the 5-knot model were 0.001, 0.009, 0.512, 1.54, and 11.3 p.p.m. of benzene in air, corresponding to the 5th, 27.5th, 50th, 72.5th, and 95th percentiles [46], respectively). To avoid overparameterization, nonsignificant knots for each exposuremetabolite relationship were removed by stepwise elimination using a value of P < 0.10 for retention (PROC REG of SAS; SAS Institute, Cary, North Carolina, USA) [28]. For participants with repeated measurements of air and urine, the estimated geometric mean air and metabolite concentrations were used in all statistical analyses. After establishing NS smoothing functions for each metabolite, we used GLM to investigate effects of genetic polymorphisms and their interactions with benzene exposure and smoking status, after adjusting for the following covariates: sex (0, women; 1, man), age (centered around the estimated mean value of 29.8 years, n = 386), smoking status (0, nonsmoker; 1, smoker), BMI (centered around the estimated mean value of 22.5 kg/m2, n = 384). We also investigated effects of toluene exposure (0, low exposure; 1, high exposure; median as a cutoff point, 1.77 p.p.m.) on levels of each metabolite. Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 793 Potential effects of genetic polymorphisms were screened in two stages to explore exposure-related interactions and smoking-related interactions, respectively. In both stages, the number of effects was restricted to less than 10% of observations to avoid overfitting. Then, important main effects and interactions were pooled from the two stages to build final models. Every candidate model for a given benzene metabolite was sorted by AICc, and the final model was selected from the best and second best candidates, using the following criteria: DAICc, evidence ratio, and the significance and biological plausibility of explanatory variables [45]. Modeling was performed using PROC GLMSELECT of SAS/STAT with the selection/ stop option of AICc [45,47]. In coding each genetic polymorphism, the homozygous wild-type was defined as the reference group. For EPHX1 (rs2234922: A-G), variant homozygotes and heterozygotes were combined in the analysis because the former contained only five participants (1.3%) [44]. Tests for HardyWeinberg equilibrium (HWE) among participants were conducted on the basis of observed genotype frequencies using PROC ALLELE of SAS/GENETICS (using a Pearson's w2 test with one degree of freedom). All genotypes were in HWE except MPO (rs2333227: G-A) (P = 0.04). The quality control data were rechecked and the precision of genotyping for this genetic polymorphism in blind replicates was confirmed; so this slight departure from HWE is likely due to chance. TukeyCramer adjustment was carried out for multiple comparisons of least-squares means among genetic polymorphisms in the final models. All statistical analyses were performed using SAS software for Windows ver. 9.13 (SAS Institute). Results GLM + NS models and covariate effects The following NS smoothing functions were used for the five benzene metabolites: MA :ElnYMA; jj lnXj 0:754 0:127lnXj 0:005lnXj x13 0:019lnXj x33 SPMA :ElnYSPMA; jj lnXj 6:65 0:119lnXj 0:030lnXj x13 0:040lnXj x23 PH:ElnYPH; jj lnXj 4:38 0:053lnXj 0:005lnXj x23 CA :ElnYCA; jj lnXj 3:02 0:098lnXj 0:005lnXj x13 0:011lnXj x23 and HQ:ElnYHQ; jj lnXj 1:72 0:016lnXj 0:004lnXj x13 0:123lnXj x53 where E[ln(Ym, j)|ln(Xj)] is the conditional mean value of ln(Ym, j) representing the natural log transform of the level of the mth metabolite level in the jth individual exposed at ln(Xj), the corresponding (logged) air concentration of benzene (p.p.m.), and xi is the location of the ith knot (in log-scale of benzene exposure): x1 = ln(0.001 p.p.m.), x2 = ln(0.009 p.p.m.), x3 = ln(0.512 p.p.m.), x4 = ln(1.54 p.p.m.), x5 = ln(11.3 p.p.m.). The Table 3 Parameter estimates for the final model of MA. [The dependent variable was the natural logarithm of the MA concentration (lmol/l); n = 382, R2 = 85.0%] Independent variable Description Parameter estimates Standard error P-value Cumulative DR 2 (%) Intercept Age (years) Sex Smoking BMI ln(Xj ) NQO1*2 NQO1*2 (rs1800566: C-T) ln(Xj ) CYP2E1 CYP2E1 (rs2031920: C-T) Centered at mean (29.8 years) Male Smoker Centered at mean (22.5 kg/m2) ln(Xj ) *1/*2, less active ln(Xj ) *2/*2, least active *1/*2, less active *2/*2, least active ln(Xj ) C/T, less active ln(Xj ) T/T, least active C/T, less active T/T, least active 0.937 0.017 0.276 0.177 0.019 0.076 0.070 0.174 0.184 0.001 0.150 0.037 0.406 0.197 0.005 0.099 0.107 0.011 0.027 0.029 0.099 0.110 0.024 0.042 0.087 0.191 < 0.0001 < 0.001 0.005 0.099 0.076 0.005 0.016 0.081 0.094 0.955 < 0.001 0.667 0.034 0.37 0.38 1.00 1.01 BMI, body mass index; MA, E,E-muconic acid. ln(Xj ) represents the natural logarithm of the air benzene concentration (p.p.m.) in the j th participant. Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 794 Pharmacogenetics and Genomics 2007, Vol 17 No 10 Table 4 Parameter estimates for the final model of SPMA. [The dependent variable was the natural logarithm of the SPMA concentration, (lmol/l); n = 365, R2 = 82.0%] Independent variable Description Parameter estimate Standard error P-value Cumulative DR 2 (%) Intercept Age (years) Sex Smoking BMI GSTM1 ln(Xj ) EPHX1 EPHX1 (rs2234922: A-G) ln(Xj ) NQO1*2 (rs1800566: C-T) ln(Xj ) CYP2E1 CYP2E1 (rs2031920: C-T) Smoking NQO1*2 NQO1*2 (rs1800566: C-T) ln(Xj ) GSTT1 GSTT1 Centered at mean (29.8 years) Male Smoker Centered at mean (22.5 kg/m2) + / , less active / , null ln(Xj ) (A/G or G/G, faster) A/G or G/G, faster ln(Xj ) *1/*2, less active ln(Xj ) *2/*2, least active ln(Xj ) C/T, less active ln(Xj ) T/T, least active C/T, less active T/T, least active Smoker *1/*2, less active Smoker *2/*2, least active *1/*2, less active *2/*2, least active ln(Xj ) + / , less active ln(Xj ) / , null + / , less active / , null 4.811 0.012 0.280 0.690 0.025 0.458 0.590 0.101 0.234 0.149 0.069 0.114 0.038 0.146 0.480 1.108 1.041 0.654 0.646 0.035 0.143 0.592 1.444 0.685 0.009 0.181 0.337 0.019 0.254 0.248 0.051 0.187 0.049 0.053 0.045 0.076 0.159 0.345 0.385 0.413 0.203 0.227 0.069 0.070 0.251 0.258 < 0.0001 0.161 0.123 0.042 0.189 0.072 0.018 0.049 0.213 0.003 0.199 0.011 0.617 0.357 0.164 0.004 0.012 0.001 0.005 0.614 0.041 0.019 < 0.0001 0.31 0.55 0.56 0.98 1.44 1.50 1.94 2.01 2.63 4.86 SPMA, S-phenylmercapturic acid; BMI, body mass index. ln(Xj ) represents the natural logarithm of the air benzene concentration (p.p.m.) in the j th participant. Table 5 Parameter estimates for the final model of PH. [The dependent variable was the natural logarithm of the PH concentration, (lmol/l); n = 382, R2 = 64.8%] Independent variable Description Parameter estimate Standard error P-value Cumulative DR 2 (%) Intercept Age (years) Sex Smoking BMI Smoking NQO1*2 NQO1*2 (rs1800566: C-T) ln(Xj ) CYP2E1 CYP2E1 (rs2031920: C-T) Centered at mean (29.8 years) Male Smoker Centered at mean (22.5 kg/m2) Smoker *1/*2, less active Smoker *2/*2, least active *1/*2, less active *2/*2, least active ln(Xj) C/T, less active ln(Xj) T/T, least active C/T, less active T/T, least active 4.645 0.010 0.313 0.204 0.007 0.311 0.538 0.196 0.337 0.026 0.141 0.037 0.613 0.100 0.004 0.090 0.167 0.010 0.192 0.204 0.085 0.097 0.022 0.038 0.079 0.174 < 0.0001 0.014 0.001 0.222 0.477 0.106 0.009 0.022 0.001 0.230 < 0.001 0.640 0.001 0.67 1.31 2.75 3.08 BMI, body mass index; PH, phenol. ln (Xj ) represents the natural logarithm of the air benzene concentration (p.p.m.) in the j th participant. function lnXj xi3 equals lnXj xi3 for positive values and equals zero otherwise. Final GLM + NS models are summarized in Tables 37 for the five benzene metabolites. Referring to the nongenetic effects, results are similar to those reported previously without adjustment for genetic polymorphisms [28]. Women participants had higher levels of MA, PH, CA, and HQ than men (P < 0.05) and younger individuals (below 30 years) had higher levels of MA, PH, and HQ than older individuals. Smokers generally had higher levels of benzene metabolites than nonsmokers, but the relationships were complicated by genesmoking interactions for SPMA, PH, and CA. No significant effects were observed on any of the benzene metabolites for either BMI or coexposure to toluene. Effects of genetic polymorphisms After adjusting for exposure and covariates, the following genetic polymorphisms were found to significantly affect levels of the various metabolites, either as main effects or as interactions with benzene exposure and/or smoking: NQO1*2 (rs1800566: C-T) for all metabolites, CYP2E1 (rs2031920: C-T) for all metabolites except CA, GSTT1 and GSTM1 for SPMA, EPHX1 (rs2234922: A-G) for SPMA and CA, and EPHX1 (rs1051740: T-C) for CA. (Note that negative values of estimated parameters indicate lower metabolite levels and vice versa). The interaction between CYP2E1 and benzene exposure accentuated the effects of polymorphic forms of this gene on levels of MA, PH, and HQ among individuals exposed to higher benzene concentrations (Tables 3, 5 and 7). For each of these metabolites, individuals having Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 795 Table 6 Parameter estimates for the final model of CA. [The dependent variable was the natural logarithm of the CA concentration, (lmol/l); n = 370, R2 = 57.2%] Independent variable Description Parameter estimate Standard error P-value Cumulative DR 2 (%) Intercept Age (years) Sex Smoking BMI ln(Xj ) NQO1*2 NQO1*2 (rs1800566: C-T) EPHX1 (rs2234922: A-G) ln(Xj ) EPHX1 (rs1051740: T-C) Smoking EPHX1 EPHX1 (rs1051740: T-C) Centered at mean (29.8 years) Male Smoker Centered at mean (22.5 kg/m2) ln(Xj) *1/*2, less active ln(Xj ) *2/*2, least active *1/*2, less active *2/*2, least active A/G or G/G, faster ln(Xj ) (T/C, slow) ln(Xj ) (C/C, slower) Smoker (T/C, slow) Smoker (C/C, slower) T/C, slow C/C, slower 3.072 < 0.001 0.272 0.567 0.007 0.037 0.058 0.186 0.237 0.143 0.039 0.069 0.163 0.633 0.019 0.096 0.285 0.004 0.087 0.127 0.009 0.023 0.025 0.089 0.097 0.076 0.021 0.028 0.156 0.206 0.086 0.118 < 0.0001 0.997 0.002 < 0.0001 0.448 0.113 0.021 0.036 0.015 0.061 0.063 0.014 0.299 0.002 0.827 0.420 0.67 0.96 1.43 2.22 3.32 3.36 BMI, body mass index; CA, catechol. ln(Xj ) represents the natural logarithm of the air benzene concentration (p.p.m.) in the j th participant. Table 7 Parameter estimates for the final model of HQ. [The dependent variable was the natural logarithm of the HQ concentration, (lmol/l); n = 382, R2 = 72.6%] Independent variable Description Parm. est. Standard error P-value Cumulative DR 2 (%) Intercept Age Sex Smoking BMI NQO1*2 (rs1800566: C-T) ln(Xj ) CYP2E1 CYP2E1 (rs2031920: C-T) Centered at mean (29.8 years) Male Smoker Centered at mean (22.5 kg/m2) 1/*2, less active *2/*2, least active ln(Xj ) C/T, less active ln(Xj ) T/T, least active C-T, C/T, less active C-T, T/T, least active 1.758 0.011 0.302 0.473 0.010 0.087 0.164 0.008 0.125 0.036 0.657 0.118 0.004 0.082 0.089 0.009 0.070 0.078 0.020 0.035 0.072 0.160 < 0.0001 0.006 < 0.001 < 0.0001 0.248 0.215 0.036 0.676 < 0.001 0.617 < 0.0001 0.34 1.25 1.77 BMI, body mass index; HQ, hydraquinone. ln(Xj ) represents the natural logarithm of the air benzene concentration (p.p.m.) in the j th participant. both variant alleles of CYP2E1 had lower metabolite levels than those with at least one wild-type allele. For example, the relationships for HQ, shown in Fig. 2a, indicate that homozygous variants produced appreciably less metabolite than heterozygotes or homozygous wildtypes at air concentrations greater than 0.1 p.p.m. (Tukey's test, P < 0.05). Similar behaviors were observed for PH, where significant departure was observed above 0.2 p.p.m., and for MA above 2 p.p.m. For SPMA, the effect of CYP2E1 was unclear. Individuals with both variant alleles of CYP2E1 had the lowest metabolite levels at benzene concentrations between 0.02 and 88.9 p.p.m.; however, the difference was not statistically significant (P > 0.05). Participants with at least one variant allele of NQO1*2 (rs1800566: C-T) had lower levels of all metabolites than homozygous wild-types or heterozygotes (Tables 37). This gene was also found to interact strongly with benzene exposure (for MA, SPMA, and CA) and/or smoking status (for PH and SPMA). These interactions with benzene exposure resulted in participants with at least one variant allele of NQO1*2 (rs1800566: C-T) having lower levels of CA above 0.8 p.p.m. (P < 0.05, Fig. 2b), lower levels of MA above 6 p.p.m. (P < 0.05), and lower levels of HQ over the entire range of exposure (significance, P < 0.1). Among nonsmokers, participants with NQO1*1/*1 produced more SPMA above 0.5 p.p.m. (Fig. 2c, P < 0.05) and more PH over the entire range of exposure (Fig. 2d, P < 0.01) than participants with NQO1*2/*2. Among smokers, however, these effects were diminished (Figs. 2d and e). No effects were observed for NQO1*3 (rs4986998: C-T) polymorphisms. Strong effects of GSTT1 and GSTM1 were observed on levels of SPMA, with homozygous variants producing the highest levels, followed by heterozygotes and homozygous null individuals (Table 4). A significant interaction was also observed between benzene exposure and GSTT1 null individuals, such that these workers produced increasingly less SPMA at higher air concentrations (see Fig. 2e). No effect of polymorphic forms of GSTP1 (rs947894: A-G) was observed on SPMA levels. Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 796 Pharmacogenetics and Genomics 2007, Vol 17 No 10 Fig. 2 (a) 1000 CYP2E1 (rs2031920: C T) 100 (b) 1000 NQO12 (rs1800566: C T) 100 Concentration of CA (mol/l) Concentration of HQ (mol/l) 10 10 C/C C/T T/T 1/1 1/2 2/2 11 0.01 0.1 1 10 100 0.01 0.1 1 10 100 Air benzene (p.p.m.) Air benzene (p.p.m.) (c) 100 NQO12 (rs1800566: C T) 10 by smoking status (d) 10 000 NQO12 (rs1800566: C T) by smoking status Concentration of PH (mol/l) Concentration of SPMA (mol/l) 1 1000 0.1 0.01 100 0.001 0.0001 0.01 Smoker, 1/1 Nonsmoker, 1/1 Smoker, 2/2 Nonsmoker, 2/2 0.1 1 10 100 Air benzene (p.p.m.) 10 0.01 Smoker, 1/1 Nonsmoker, 1/1 Smoker, 2/2 Nonsmoker, 2/2 0.1 1 10 100 Air benzene (p.p.m.) (e) 100 GSTT1 10 (f) 1000 EPHX1 (rs1051740: T C) by smoking status Concentration of CA (mol/l) Concentration of SPMA (mol/l) 1 100 0.1 0.01 0.001 0.0001 0.01 +/+ +/ 0.1 1 10 Air benzene (p.p.m.) / 100 10 1 0.01 Smoker, T/T Nonsmoker, T/T Smoker, C/C Nonsmoker, C/C 0.1 1 10 100 Air benzene (p.p.m.) Effects of genetic polymorphisms on urinary metabolites of benzene in humans (representative plots). Each panel depicts (log scale) effects of a particular genetic polymorphism on levels of a given metabolite versus the air concentration of benzene. Participants with variant alleles of EPHX1 (rs1051740: T-C) had lower levels of CA, particularly among smokers, than homozygous wild-types (see Table 6). Owing to the interaction of EPHX1 (rs1051740: T-C) with benzene exposure, this effect was accentuated among smokers at air concentrations above 0.8 p.p.m. (P < 0.05) (Fig. 2f). Participants with at least one variant allele of EPHX1 (rs2234922: A-G) had higher CA levels than homozygous wild-types (P = 0.06). Interestingly, participants with variant EPHX1 (rs2234922: A-G) also had higher levels of SPMA at benzene concentrations above 6 p.p.m. (P < 0.1). Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 797 No effects were observed for polymorphic forms of MPO (rs2333227: G-A). Discussion Although the hematotoxicity of benzene was reported more than a century ago, the mechanism is not yet fully understood [9]. It has been speculated that genetic and lifestyle factors can influence the toxic effects of benzene, but current evidence is far from conclusive [911]. Rothman et al. [23] reported that heavily benzene-exposed workers who were rapid chlorzoxazone metabolizers (a measure of CYP2E1 phenotype) and also possessed variant NQO1*2 (rs1800566: C-T), were at elevated risk of benzene poisoning. More recently, Wan et al. [48] reported increased benzene poisoning in workers with variant NQO1*2 (rs1800566: C-T) and in those with null-type GSTT1 or CYP2E1 DraI, and Lan et al. [8] found lower white blood cell counts in workers with variants of NQO1*3 (rs4986998: C-T) and MPO (rs2333227: G-A). Other studies of polymorphisms among benzene-exposed workers reported that individuals with variant NQO1*2 (rs1800566: C-T) had increased DNA single-strand breaks [49] but decreased aneuploidy [50] in peripheral lymphocytes. The latter study also reported increased aneuploidy in benzeneexposed workers with null deletions of GSTT1 and GSTM1 and with either of two CYP2E1 mutations (DraI or RsaI) [50]. In the present study, we focused upon the effects of metabolizing genes on production of five prominent metabolites (MA, SPMA, PH, CA and HQ). These metabolites are not `biological effect markers' [51] per se, but rather reflect primary metabolism in the liver, which appears to be a necessary prelude to benzene-induced toxicity in target organs [9,10]. Previous attempts to link levels of benzene metabolites with polymorphic forms of metabolizing genes have been hampered by methodological and practical problems, including low benzene exposures (below 0.1 p.p.m.) [18,25,26,52,53], no measurements of air exposure [38,48,54,55], small numbers of participants [25,26,50,53], and difficulties in adjusting for covariates and nonlinear effects of exposure [24,49]. As benzene exposures were typically very low in previous studies, only SPMA and MA (metabolites with high specificity for benzene) tended to be measured, and the only consistent effect of any genetic polymorphism was that of lower SPMA levels in individuals with GSTT1 (null deletion) [2426]. We detected several effects of genetic polymorphisms on benzene metabolite patterns and interactions with benzene exposure that have not been reported previously (Tables 37). Our study had substantially more participants (386) and therefore, greater power to detect such effects. Further strengths of our study were the ability to examine effects over a wide range of benzene exposures, determined in both benzene-exposed workers and controls, broad exploration of genetic variants in key genes, evaluation of all major benzene urinary metabolites, and use of GLM + NS models to adjust for exposure and covariates. At the same time, it is possible that some findings could be false positives, and these findings need to be replicated in other large studies. Individuals with NQO1*2 (rs1800566: C-T) had lower levels of all five metabolites in our study. As NQO1 catalyzes two-election or four-election reductions of quinones [34,5659], and the NQO1*2 polymorphism is associated with a lack of NQO1*2 protein [20], it is reasonable that the levels of CA and HQ would be lower in individuals with NQO1*2 (rs1800566: C-T) (Tables 6 and 7). When combined with evidence that less active forms of NQO1 are associated with benzene poisoning and DNA damage [8,23,48,49], this finding is also consistent with speculation that 1,4-BQ and/or 1,2-BQ (the oxidized forms of HQ and CA, respectively) play roles in benzeneinduced toxicity [10,14,6062]. The fact that levels of the other three metabolites (MA, SPMA, and PH) were also lower among participants with NQO1*2 (rs1800566: C-T), suggests a more general antioxidant role for NQO1 [20]. Furthermore, the interaction effects between NQO1*2 (rs1800566: C-T) and both benzene exposure and smoking status point to induction of NQO1 by reactive benzene metabolites or other reactive species [20,63,64]. Such induction could come about via either the antioxidant or xenobiotic response element in the NQO1 promoter region [19,65,66]. We found that NQO1*2 (rs1800566: C-T) but not NQO1*3 (rs4986998: C-T) was associated with lower levels of benzene metabolites and that single nucleotide polymorphisms (SNPs) of MPO appeared not to affect metabolism in liver. Considering our previous report of the presence of greater hematotoxicity in workers having the combination of variant NQO1*3 (C/T, less active) and wild-type MPO (A/A, more active) [8], the present results are interesting. They probably point to differences in the balance between NQO1 and MPO activities in liver (where metabolites are produced) and bone marrow (where metabolites are activated and deactivated in target hematopoietic cells) [60,67,68]. We observed significant effects of CYP2E1 (rs2031920: C-T) variants on levels of MA, PH, and HQ (Tables 3, 5, and 7). Controversy has surrounded the relationship between the genotype and phenotype of CYP2E1, an important gene that metabolizes many small molecules of toxicological interest, including benzene [21,22,6979]. There is some evidence that a variant type CYP2E1 (rs2031920: C-T, also referred to as RsaI ), is associated with decreased CYP2E1 activity in vivo [33,52,71,73,80,81]. In the present study, we found that participants with variant CYP2E1 (rs2031920: C-T) Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 798 Pharmacogenetics and Genomics 2007, Vol 17 No 10 produced lower levels of benzene metabolites at a given benzene exposure than homozygous wild-types, and that the effect was accentuated at higher benzene levels due to gene-environment interactions. The difference in metabolite levels between homozygous wild-types and homozygous variants was detected at benzene exposures in the range of 0.12 p.p.m. for HQ, PH and MA (Tables 3, 5 and 7 and Fig. 2a). This interaction effect could point to induction CYP2E1 (rs2031920: C-T) by benzene exposure, or to more rapid saturation of metabolism among homozygous variant individuals [22]. Therefore, our results substantially support evidence from previous studies [33,52,71,73,80,81] that CYP2E1 (rs2031920: C-T) mutations functionally reduce the metabolism of CYP2E1 substrates. Individuals with variant alleles of EPHX1 (rs1051740: T-C) produced lower levels of CA in our study (Table 6 and Fig. 2f). EPHX1 enzymes hydrolyze epoxides through the formation of hydroxyl alkyl-enzyme intermediates [82]. Although EPHX1 should logically be involved in benzene metabolism, notably in catalyzing BO to the dihydrodiol, the functional role of this enzyme has been ambiguous in benzene-exposed participants [56,83]. Two polymorphisms have been identified; one, in exon 3, decreased enzymatic activity 50% in vitro, whereas the other, in exon 4, increased activity 25% [84]. Our findings that workers with variant allele(s) of EPHX1 (rs1051740: T-C) had lower levels of CA than homozygous wildtypes, whereas those with at least one variant of EPHX1 (rs2234922: A-G) had marginally higher CA levels, support the in-vitro results. We also detected significant interactions between EPHX1 (rs1051740: T-C) and both benzene exposure and smoking status. The genesmoking interaction tended to accentuate differences in CA levels between smoking individuals who had different alleles of EPHX1 (rs1051740: T-C) and to obscure effects among nonsmokers (Fig. 2f). The gene-exposure interaction produced differences in CA levels that could be distinguished between smokers having homozygous wild-types and variant-types of EPHX1 (rs1051740: T-C) at benzene concentrations above 1 p.p.m. Our results are intriguing in light of recent epidemiologic studies indicating that, among smokers, fast EPHX1 metabolizers had greater risks of colorectal adenomas [85,86] than slow metabolizers. Interestingly, fast metabolizers of EPHX1 (rs2234922: A-G) also had marginally higher levels of SPMA at benzene concentrations above 6 p.p.m. (P < 0.10). Although the mechanism for production of SPMA is not yet established [87,88], the apparent effect of EPHX1 on SPMA levels may offer clues regarding formation of this minor benzene metabolite [82]. Among the GST isozymes, both GSTM1 and GSTT1 affected the production of SPMA (Table 4, Fig. 2e), with individuals having variant forms of these enzymes producing lower levels. Of the two isozymes, GSTT1 produced more substantial effects, based upon evidence ratios [45] (data not shown), and produced different profiles for each combination of alleles (Fig. 2e). This finding is consistent with previous studies [2426]. No effect of GSTP1 (rs947894: A-G) was detected. Finally, it is worth commenting upon the amounts of variability in metabolite levels that were explained by the observed genetic effects and the magnitudes of interindividual differences in metabolism that can be attributed to particular genes. The GLM + NS models summarized in Tables 37 had R2 (%) values of 85.0 for MA, 82.0 for SPMA, 64.8 for PH, 57.2 for CA, and 72.6 for HQ, among which benzene exposure and nongenetic covariates explained between 53 and 84% of the variability in metabolite levels. The corresponding percentages of variability explained collectively by all significant genes and gene-environment interactions were 1.0 for MA, 4.9 for SPMA, 3.1 for PH, 3.4 for CA, and 1.8 for HQ. Thus, although many significant genetic effects were detected, they collectively contributed rather little to the explained variation in benzene metabolism. Regarding interindividual differences in metabolism that would be expected for a given genetic polymorphism, Table 8 lists the ratios of predicted metabolite levels for homozygote variants to homozygous wild-types, based upon least-squares means of the models summarized in Tables 37. These ratios represent the mean fold ranges for variant/referent that would be expected for each genetic polymorphism after adjusting for benzene exposure as well as covariates and other genetic effects. The values shown in Table 8 suggest that interindividual differences in metabolite production were generally rather modest, with most ratios lying between about 0.3 and 2.0. Indeed, differences as great as two-fold to 3.5fold would only be anticipated for most metabolitegene combinations when persons were exposed to very high benzene concentrations (100 p.p.m.). The exception to this rule is the large effect of GSTT1 on SPMA production, where homozygous referents would typically have SPMA levels three-fold to eight-fold greater than those of homozygous variants. This large fold range undoubtedly contributed to the earlier reports of significant effects of GSTT1 on SPMA levels [2426]. In conclusion, we used GLM + NS regression to detect numerous effects of particular metabolizing genes and geneenvironment interactions on levels of benzene metabolites in 386 Chinese workers. Of the nine genetic polymorphisms investigated, NQO1*2 (rs1800566: C-T) affected all five metabolites, CYP2E1 (rs2031920: C-T) affected all metabolites but CA, EPHX1 (1051740: T-C or 2234922: A-G) affected CA and SPMA, and GSTT1 and GSTM1 affected SPMA. Significant interactions were detected between benzene exposure and all four genes [including CYP2E1 (rs2031920: C-T), NQO1*2 Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 799 Table 8 Effects of genetic polymorphisms on benzene metabolites at various levels of benzene exposure. (Least-squares-mean ratios of variant/variant to wild/wild) Air concentration (p.p.m.) Metabolite SNP 0.1 1 10 100 E,E-muconic acid (MA) S-Phenylmercapturic acid (SPMA) Phenol (PH) Catechol (CA) Hydroquinone (HQ) CYP2E1 (rs2031920: C-T) NQO1*2 (rs1800566: C-T) CYP2E1 (rs2031920: C-T) EPHX1 (rs2234922: A-G) GSTT1 GSTM1 NQO1*2 (rs1800566: C-T, nonsmokers) NQO1*2 (rs1800566: C-T, smokers) CYP2E1 (rs2031920: C-T) NQO1*2 (rs1800566: C-T, nonsmokers) NQO1*2 (rs1800566: C-T, smokers) EPHX1 (rs2234922: A-G) NQO1*2 (rs1800566: C-T) EPHX1 (rs1051740 : T-C, nonsmokers) EPHX1 (rs1051740 : T-C, smokers) CYP2E1 (rs2031920: C-T) NQO1*2 (rs1800566: C-T) 0.941 0.976 0.675 1.00 0.328 0.554 0.614 1.74 0.750 0.714 1.22 1.15 0.902 1.29 0.685 0.692 0.849 0.666 0.832 0.619 1.26 0.236 0.554 0.524 1.48 0.542 0.714 1.22 1.15 0.789 1.10 0.584 0.518 0.849 0.472 0.708 0.567 1.59 0.17 0.554 0.448 1.27 0.392 0.714 1.22 1.15 0.690 0.938 0.498 0.389 0.849 0.333 0.603 0.519 2.01 0.122 0.554 0.382 1.08 0.283 0.714 1.22 1.15 0.603 0.800 0.425 0.291 0.849 MA, E,E-muconic acid; PH, phenol; HQ, hydroquinone; CA, catechol; SPMA, S-phenylmercapturic acid; SNP, single nucleotide polymorphism. (rs1800566: C-T), EPHX1 (2234922: A-G), and GSTT1] and between smoking status and NQO1*2 (rs1800566: C-T) and EPHX1 (rs1051740: T-C). Results are generally consistent with previous reports of associations between benzene hematotoxicity and specific gene mutations and provide additional evidence regarding functionality of SNPs of NQO1, CYP2E1, and EPHX1 in humans exposed to benzene. References 1 IARC. IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans. Some Industrial Chemicals and Dyestuffs. Lyon, France: World Health Organization Inernation Agency for Research on Cancer; 1982. 2 ATSDR. Toxicological profile for benzene. In: DHHS US, editor. PB/98/ 101157/AS Agency for Toxic Substances and Disease Registry; 1997. 3 Wallace L. Environmental exposure to benzene: an update. Environ Health Perspect 1996; 104(Suppl 6):11291136. 4 Aksoy M. Malignancies due to occupational exposure to benzene. Am J Ind Med 1985; 7:395402. 5 Yin SN, Li GL, Tain FD, Fu ZI, Jin C, Chen YJ, et al. A retrospective cohort study of leukemia and other cancers in benzene workers. Environ Health Perspect 1989; 82:207213. 6 Rothman N, Li GL, Dosemeci M, Bechtold WE, Marti GE, Wang YZ, et al. Hematotoxocity among Chinese workers heavily exposed to benzene. Am J Ind Med 1996; 29:236246. 7 Savitz DA, Andrews KW. Review of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers. Am J Ind Med 1997; 31:287295. 8 Lan Q, Zhang L, Li G, Vermeulen R, Weinberg RS, Dosemeci M, et al. Hematotoxicity in workers exposed to low levels of benzene. Science 2004; 306:17741776. 9 Ross D. The role of metabolism and specific metabolites in benzene-induced toxicity: evidence and issues. J Toxicol Environ Health A 2000; 61:357372. 10 Snyder R. Benzene and leukemia. Crit Rev Toxicol 2002; 32:155210. 11 Snyder R. Overview of the toxicology of benzene. J Toxicol Environ Health A 2000; 61:339346. 12 Parke DV, Williams RT. Studies in detoxication. XLIX. The metabolism of benzene containing (14C1) benzene. Biochem J 1953; 54:231238. 13 Parke DV, Williams RT. Studies in detoxication. 54. The metabolism of benzene. (a) The formation of phenylglucuronide and phenylsulphuric acid from [14C]benzene. (b) The metabolism of [14C]phenol. Biochem J 1953; 55:337340. 14 Snyder R, Hedli CC. An overview of benzene metabolism. Environ Health Perspect 1996; 104(Suppl 6):11651171. 15 Witz G, Zhang Z, Goldstein BD. Reactive ring-opened aldehyde metabolites in benzene hematotoxicity. Environ Health Perspect 1996; 104(Suppl 6):11951199. 16 Gilmour SK, Kalf GF, Snyder R. Comparison of the metabolism of benzene and its metabolite phenol in rat liver microsomes. Adv Exp Med Biol 1986; 197:223235. 17 Penning TM, Sharp RB. Characterization of dihydrodiol dehydrogenase in human liver and lung. Carcinogenesis 1990; 11:12031208. 18 Verdina A, Galati R, Falasca G, Ghittori S, Imbriani M, Tomei F, et al. Metabolic polymorphisms and urinary biomarkers in subjects with low benzene exposure. J Toxicol Environ Health A 2001; 64:607618. 19 Smith MT. Benzene, NQO1, and genetic susceptibility to cancer. Proc Natl Acad Sci U S A 1999; 96:76247626. 20 Ross D. Functions and distribution of NQO1 in human bone marrow: potential clues to benzene toxicity. Chem Biol Interact 2005; 153154:137146. 21 Seaton MJ, Schlosser PM, Bond JA, Medinsky MA. Benzene metabolism by human liver microsomes in relation to cytochrome P450 2E1 activity. Carcinogenesis 1994; 15:17991806. 22 Nedelcheva V, Gut I, Soucek P, Tichavska B, Tynkova L, Mraz J, et al. Metabolism of benzene in human liver microsomes: individual variations in relation to CYP2E1 expression. Arch Toxicol 1999; 73:3340. 23 Rothman N, Smith MT, Hayes RB, Traver RD, Hoener B, Campleman S, et al. Benzene poisoning, a risk factor for hematological malignancy, is associated with the NQO1 609C-T mutation and rapid fractional excretion of chlorzoxazone. Cancer Res 1997; 57:28392842. 24 Qu Q, Shore R, Li G, Su L, Jin X, Melikian AA, et al. Biomarkers of benzene: urinary metabolites in relation to individual genotype and personal exposure. Chem Biol Interact 2005; 153154:8595. 25 Sorensen M, Poole J, Autrup H, Muzyka V, Jensen A, Loft S, et al. Benzene exposure assessed by metabolite excretion in Estonian oil shale mineworkers: influence of glutathione s-transferase polymorphisms. Cancer Epidemiol Biomarkers Prev 2004; 13:17291735. 26 Sorensen M, Skov H, Autrup H, Hertel O, Loft S. Urban benzene exposure and oxidative DNA damage: influence of genetic polymorphisms in metabolism genes. Sci Total Environ 2003; 309:6980. 27 Kim S, Vermeulen R, Waidyanatha S, Johnson B, Lan Q, Rothman N, et al. Using urinary biomarkers to elucidate dose-related patterns of human benzene metabolism. Carcinogenesis 2006; 27:772781. 28 Kim S, Vermeulen R, Waidyanatha S, Johnson B, Lan Q, Rothman N, et al. Modeling human metabolism of benzene following occupational and environmental exposures. Cancer Epidemiol Biomarkers Prev 2006; 15:22462256. 29 Vermeulen R, Li G, Lan Q, Dosemeci M, Rappaport SM, Bohong X, et al. Detailed exposure assessment for a molecular epidemiology study of benzene in two shoe factories in China. Ann Occup Hyg 2004; 48:105116. Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. 800 Pharmacogenetics and Genomics 2007, Vol 17 No 10 30 Waidyanatha S, Rothman N, Fustinoni S, Smith MT, Hayes RB, Bechtold W, et al. Urinary benzene as a biomarker of exposure among occupationally exposed and unexposed subjects. Carcinogenesis 2001; 22:279286. 31 Waidyanatha S, Rothman N, Li G, Smith MT, Yin S, Rappaport SM. Rapid determination of six urinary benzene metabolites in occupationally exposed and unexposed subjects. Anal Biochem 2004; 327:184199. 32 Hornung RW, Reed LD. Estimation of average concentration in the presence of nondetectable values. Appl Occup Environ Hyg 1990; 5:4651. 33 Marchand LL, Wilkinson GR, Wilkens LR. Genetic and dietary predictors of CYP2E1 activity: a phenotyping study in Hawaii Japanese using chlorzoxazone. Cancer Epidemiol Biomarkers Prev 1999; 8:495500. 34 Ross D, Kepa JK, Winski SL, Beall HD, Anwar A, Siegel D. NAD(P)H:quinone oxidoreductase 1 (NQO1): chemoprotection, bioactivation, gene regulation and genetic polymorphisms. Chem Biol Interact 2000; 129:7797. 35 Pan SS, Forrest GL, Akman SA, Hu LT. NAD(P)H:quinone oxidoreductase expression and mitomycin C resistance developed by human colon cancer HCT 116 cells. Cancer Res 1995; 55:330335. 36 Eguchi-Ishimae M, Eguchi M, Ishii E, Knight D, Sadakane Y, Isoyama K, et al. The association of a distinctive allele of NAD(P)H: quinone oxidoreductase with pediatric acute lymphoblastic leukemias with MLL fusion genes in Japan. Haematologica 2005; 90:15111515. 37 Williams JA. Single nucleotide polymorphisms, metabolic activation and environmental carcinogenesis: why molecular epidemiologists should think about enzyme expression. Carcinogenesis 2001; 22:209214. 38 Hsieh LL, Liou SH, Chiu LL, Chen YH. Glutathione S-transferase (GST) M1 and GST T1 genotypes and hematopoietic effects of benzene exposure. Arch Toxicol 1999; 73:8082. 39 Morgan GJ, Smith MT. Metabolic enzyme polymorphisms and susceptibility to acute leukemia in adults. Am J Pharmacogenomics 2002; 2:7992. 40 Jourenkova-Mironova N, Mitrunen K, Bouchardy C, Dayer P, Benhamou S, Hirvonen A. High-activity microsomal epoxide hydrolase genotypes and the risk of oral, pharynx, and larynx cancers. Cancer Res 2000; 60:534536. 41 Lodovici M, Luceri C, Guglielmi F, Bacci C, Akpan V, Fonnesu ML, et al. Benzo(a)pyrene diolepoxide (BPDE)-DNA adduct levels in leukocytes of smokers in relation to polymorphism of CYP1A1, GSTM1, GSTP1, GSTT1, and mEH. Cancer Epidemiol Biomarkers Prev 2004; 13:13421348. 42 den Dunnen JT, Antonarakis SE. Mutation nomenclature extensions and suggestions to describe complex mutations: a discussion. Hum Mutat 2000; 15:712. 43 Packer BR, Yeager M, Staats B, Welch R, Crenshaw A, Kiley M, et al. SNP500Cancer: a public resource for sequence validation and assay development for genetic variation in candidate genes. Nucleic Acids Res 2004; 32:D528D532. 44 Lan Q, Zhang L, Shen M, Smith MT, Li G, Vermeulen R, et al. Polymorphisms in cytokine and cellular adhesion molecule genes and susceptibility to hematotoxicity among workers exposed to benzene. Cancer Res 2005; 65:95749581. 45 Burnham KP, Anderson DR, Burnham KP. Model selection and multimodel inference: a practical information-theoretic approach. 2nd ed. New York: Springer; 2002. 46 Harrell FE. Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis. Springer series in statistics. New York: Springer; 2001. 47 SAS Institute. The GLMSELECT Procedure. [updated 2005 Aug.]. Available from: http://www.sas.com/apps/demosdownloads/glmselect_Exp_sysdep. jsp?packageID = 000344/. 48 Wan J, Shi J, Hui L, Wu D, Jin X, Zhao N, et al. Association of genetic polymorphisms in CYP2E1, MPO, NQO1, GSTM1, and GSTT1 genes with benzene poisoning. Environ Health Perspect 2002; 110:12131218. 49 Garte S, Popov T, Georgieva T, Bolognesi C, Taioli E, Bertazzi P, et al. Biomarkers of exposure and effect in Bulgarian petrochemical workers exposed to benzene. Chem Biol Interact 2005; 153154:247251. 50 Kim SY, Choi JK, Cho YH, Chung EJ, Paek D, Chung HW. Chromosomal aberrations in workers exposed to low levels of benzene: association with genetic polymorphisms. Pharmacogenetics 2004; 14:453463. 51 Wogan GN. Molecular epidemiology in cancer risk assessment and prevention: recent progress and avenues for future research. Environ Health Perspect 1992; 98:167178. 52 Fustinoni S, Consonni D, Campo L, Buratti M, Colombi A, Pesatori AC, et al. Monitoring low benzene exposure: comparative evaluation of urinary biomarkers, influence of cigarette smoking, and genetic polymorphisms. Cancer Epidemiol Biomarkers Prev 2005; 14:22372244. 53 Rossi AM, Guarnieri C, Rovesti S, Gobba F, Ghittori S, Vivoli G, et al. Genetic polymorphisms influence variability in benzene metabolism in humans. Pharmacogenetics 1999; 9:445451. 54 Avogbe PH, Ayi-Fanou L, Autrup H, Loft S, Fayomi B, Sanni A, et al. Ultrafine particulate matter and high-level benzene urban air pollution in relation to oxidative DNA damage. Carcinogenesis 2005; 26:613620. 55 Testa A, Festa F, Ranaldi R, Giachelia M, Tirindelli D, De Marco A, et al. A multi-biomarker analysis of DNA damage in automobile painters. Environ Mol Mutagen 2005; 46:182188. 56 Nebert DW, Roe AL, Vandale SE, Bingham E, Oakley GG. NAD(P)H:quinone oxidoreductase (NQO1) polymorphism, exposure to benzene, and predisposition to disease: a HuGE review. Genet Med 2002; 4:6270. 57 Traver RD, Siegel D, Beall HD, Phillips RM, Gibson NW, Franklin WA, et al. Characterization of a polymorphism in NAD(P)H: quinone oxidoreductase (DT-diaphorase). Br J Cancer 1997; 75:6975. 58 Siegel D, McGuinness SM, Winski SL, Ross D. Genotype-phenotype relationships in studies of a polymorphism in NAD(P)H:quinone oxidoreductase 1. Pharmacogenetics 1999; 9:113121. 59 Brunmark A, Cadenas E, Lind C, Segura-Aguilar J, Ernster L. DT-diaphorasecatalyzed two-electron reduction of quinone epoxides. Free Radic Biol Med 1987; 3:181188. 60 Subrahmanyam VV, Doane-Setzer P, Steinmetz KL, Ross D, Smith MT. Phenol-induced stimulation of hydroquinone bioactivation in mouse bone marrow in vivo: possible implications in benzene myelotoxicity. Toxicology 1990; 62:107116. 61 Subrahmanyam VV, Kolachana P, Smith MT. Metabolism of hydroquinone by human myeloperoxidase: mechanisms of stimulation by other phenolic compounds. Arch Biochem Biophys 1991; 286:7684. 62 McDonald TA, Yeowell-O'Connell K, Rappaport SM. Comparison of protein adducts of benzene oxide and benzoquinone in the blood and bone marrow of rats and mice exposed to [14C/13C6]benzene. Cancer Res 1994; 54:49074914. 63 De Long MJ, Santamaria AB, Talalay P. Role of cytochrome P1450 in the induction of NAD(P)H:quinone reductase in a murine hepatoma cell line and its mutants. Carcinogenesis 1987; 8:15491553. 64 Moran JL, Siegel D, Ross D. A potential mechanism underlying the increased susceptibility of individuals with a polymorphism in NAD(P)H:quinone oxidoreductase 1 (NQO1) to benzene toxicity. Proc Natl Acad Sci U S A 1999; 96:81508155. 65 Jaiswal AK. Regulation of genes encoding NAD(P)H:quinone oxidoreductases. Free Radic Biol Med 2000; 29:254262. 66 Ross D. Quinone reductases. In: Guengerich FP, editor. Comprehensive toxicology. [CD-ROM] New York, NY, USA: Pergamon; 1997. 67 Ross D, Siegel D, Gibson NW, Pacheco D, Thomas DJ, Reasor M, et al. Activation and deactivation of quinones catalyzed by DT-diaphorase. Evidence for bioreductive activation of diaziquone (AZQ) in human tumor cells and detoxification of benzene metabolites in bone marrow stroma. Free Radic Res Commun 1990; 8:373381. 68 Ross D, Siegel D, Schattenberg DG, Sun XM, Moran JL. Cell-specific activation and detoxification of benzene metabolites in mouse and human bone marrow: identification of target cells and a potential role for modulation of apoptosis in benzene toxicity. Environ Health Perspect 1996; 104(Suppl 6):11771182. 69 Hayashi S, Watanabe J, Kawajiri K. Genetic polymorphisms in the 50-flanking region change transcriptional regulation of the human cytochrome P450IIE1 gene. J Biochem (Tokyo) 1991; 110:559565. 70 Lucas D, Menez C, Girre C, Berthou F, Bodenez P, Joannet I, et al. Cytochrome P450 2E1 genotype and chlorzoxazone metabolism in healthy and alcoholic Caucasian subjects. Pharmacogenetics 1995; 5:298304. 71 Carriere V, Berthou F, Baird S, Belloc C, Beaune P, de Waziers I. Human cytochrome P450 2E1 (CYP2E1): from genotype to phenotype. Pharmacogenetics 1996; 6:203211. 72 Kim RB, Yamazaki H, Chiba K, O'Shea D, Mimura M, Guengerich FP, et al. In vivo and in vitro characterization of CYP2E1 activity in Japanese and Caucasians. J Pharmacol Exp Ther 1996; 279:411. 73 Wang SL, Lee H, Chen KW, Tsai KJ, Chen CY, Lin P. Cytochrome P4502E1 genetic polymorphisms and lung cancer in a Taiwanese population. Lung Cancer 1999; 26:2734. 74 Wormhoudt LW, Commandeur JN, Vermeulen NP. Genetic polymorphisms of human N-acetyltransferase, cytochrome P450, glutathione-S-transferase, and epoxide hydrolase enzymes: relevance to xenobiotic metabolism and toxicity. Crit Rev Toxicol 1999; 29:59124. 75 Le Marchand L, Donlon T, Seifried A, Wilkens LR. Red meat intake, CYP2E1 genetic polymorphisms, and colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 2002; 11:10191024. 76 Carere A, Andreoli C, Galati R, Leopardi P, Marcon F, Rosati MV, et al. Biomonitoring of exposure to urban air pollutants: analysis of sister Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Genetic polymorphisms and benzene metabolism Kim et al. 801 chromatid exchanges and DNA lesions in peripheral lymphocytes of traffic policemen. Mutat Res 2002; 518:215224. 77 Ingelman-Sundberg M, Daly AK, Nebert DW. Human Cytochrome P450 (CYP) Allele Nomenclature Committee at http://www.imm.ki.se/CYPalleles/. 2006. 78 Seaton MJ, Schlosser P, Medinsky MA. In vitro conjugation of benzene metabolites by human liver: potential influence of interindividual variability on benzene toxicity. Carcinogenesis 1995; 16:15191527. 79 Bolt HM, Roos PH, Thier R. The cytochrome P-450 isoenzyme CYP2E1 in the biological processing of industrial chemicals: consequences for occupational and environmental medicine. Int Arch Occup Environ Health 2003; 76:174185. 80 Powell H, Kitteringham NR, Pirmohamed M, Smith DA, Park BK. Expression of cytochrome P4502E1 in human liver: assessment by mRNA, genotype and phenotype. Pharmacogenetics 1998; 8:411421. 81 Choi JY, Lee KM, Cho SH, Kim SW, Choi HY, Lee SY, et al. CYP2E1 and NQO1 genotypes, smoking and bladder cancer. Pharmacogenetics 2003; 13:349355. 82 Morisseau C, Hammock BD. Epoxide hydrolases: mechanisms, inhibitor designs, and biological roles. Annu Rev Pharmacol Toxicol 2005; 45:311333. 83 Lovern MR, Maris ME, Schlosser PM. Use of a mathematical model of rodent in vitro benzene metabolism to predict human in vitro metabolism data. Carcinogenesis 1999; 20:15111520. 84 Hassett C, Aicher L, Sidhu JS, Omiecinski CJ. Human microsomal epoxide hydrolase: genetic polymorphism and functional expression in vitro of amino acid variants. Hum Mol Genet 1994; 3:421428. 85 Cortessis V, Siegmund K, Chen Q, Zhou N, Diep A, Frankl H, et al. A casecontrol study of microsomal epoxide hydrolase, smoking, meat consumption, glutathione S-transferase M3, and risk of colorectal adenomas. Cancer Res 2001; 61:23812385. 86 Huang WY, Chatterjee N, Chanock S, Dean M, Yeager M, Schoen RE, et al. Microsomal epoxide hydrolase polymorphisms and risk for advanced colorectal adenoma. Cancer Epidemiol Biomarkers Prev 2005; 14:152157. 87 Snyder R, Chepiga T, Yang CS, Thomas H, Platt K, Oesch F. Benzene metabolism by reconstituted cytochromes P450 2B1 and 2E1 and its modulation by cytochrome b5, microsomal epoxide hydrolase, and glutathione transferases: evidence for an important role of microsomal epoxide hydrolase in the formation of hydroquinone. Toxicol Appl Pharmacol 1993; 122:172181. 88 Henderson AP, Barnes ML, Bleasdale C, Cameron R, Clegg W, Heath SL, et al. Reactions of benzene oxide with thiols including glutathione. Chem Res Toxicol 2005; 18:265270. Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.