The Influence of Metabolic Gene Polymorphisms on Urinary 1

中国科技论文在线 http://www.paper.edu.cn

Science of the Total Environment 381 (2007) 38–46

The influence of metabolic gene polymorphisms on urinary 1-hydroxypyrene concentrations in Chinese coke oven workers ⁎ ⁎ Bo Chen a,1, Yunping Hu a,c,1, Taiyi Jin a, , Daru Lu b, , Minhua Shao b, Lixing Zheng a, Qiangyi Wang a, Yue Shen a, Hongliang Liu b, Yanhong Liu b, Yuanfen Zhou a

a Department of Occupational Health and Toxicology, School of Public Health, Fudan University, Shanghai, 200032, PR China b State Key Laboratory of Genetic Engineering, Institute of Genetics, School of Life Sciences, Fudan University, Shanghai, 200433, PR China c Department of Pathology, Wake Forest University School of Medicine, Medical Center Boulevard, Winston-Salem, NC 27157, USA Received 21 November 2006; received in revised form 2 February 2007; accepted 19 February 2007 Available online 11 May 2007

Abstract

Urinary 1-hydroxypyrene (1-OHP), a biomarker of polycyclic aromatic hydrocarbons (PAHs) exposure, may be influenced by metabolic gene polymorphisms. Such knowledge could benefit us in understanding the inter-individual difference in the mechanism of PAHs-induced carcinogenesis. We investigated the influence of gene polymorphisms on urinary 1-OHP concentrations in 447 coke oven workers from two coking plants in south China. After adjustment for age, plant, level of occupational exposure, body mass index, level of education, alcohol consumption, cigarette smoking and respirator usage, AhR R554K (rs2066853), UGT1A1 −3263TNG (rs4124874) and GSTP1 I105V (rs1695) were associated with urinary 1-OHP excretion with the p-value of 0.053, 0.006 and 0.021, respectively. The concentrations of urinary 1-OHP (Geometric mean, μmol/mol creatinine) in the homozygous major variant carriers and homozygous minor variant carriers for AhR R554K, UGT1A1 −3263TNG and GSTP1 I105V were listed as follows: 4.20 and 5.12, 5.11 and 3.92, 4.93 and 2.91, respectively. GSTT1 present carriers had a significantly higher urinary 1-OHP level than that in null carriers in the case with AhR R554K GA/AA carriers (5.17 vs. 3.64 μmol/mol creatinine, p=0.038), as well as in the case with UGT1A1 −3263TNG TG/GG carriers (5.67 vs. 3.38 μmol/mol creatinine, p=0.001). These results showed that AhR, UGT1A1, GSTP1 and GSTT1 polymorphisms were associated with urinary 1-OHP concentrations in Chinese coke oven workers. No influence was found in the association between urinary 1-OHP and other genetic polymorphisms such as CYP1A1, CYP1A2, CYP1B1, CYP2E1, EPHX1, EPHX2 in this population. © 2007 Elsevier B.V. All rights reserved.

Keywords: 1-Hydroxypyrene; Polycyclic aromatic hydrocarbons; Coke oven workers; Gene polymorphism

1. Introduction

⁎ Corresponding authors. Jin is to be contacted at School of Public Coke oven workers are regularly exposed to coke oven Health, Fudan University, 138 Yixueyuan Road, Shanghai, 200032, emissions, which are mainly comprised of a wide variety PR China. Tel./fax: +86 21 6417 8160. Lu, State Key Laboratory of of volatile organic compounds and particulates, especially Genetic Engineering, School of Life Sciences, Fudan University, polycyclic aromatic hydrocarbons (PAHs) (IARC, 1984). Shanghai 200433, PR China. Tel./fax: +86 21 65642799. E-mail addresses: [email protected] (T. Jin), [email protected] Some PAHs are proved to be probably or possibly car- (D. Lu). cinogenic (IARC, 1983, 1987). Epidemiological studies 1 Bo Chen and Yunping Hu contributed equally to this work. have implicated an elevated incidence of lung cancer

B. Chen et al. / Science of the Total Environment 381 (2007) 38–46 39

among coke oven workers (WHO, 1987). 1-Hydroxypyr- 70s and had 4 and 6 coke ovens, respectively. 197 male ene has been recommended as a suitable biomarker of workers from Plant I and 250 from Plant II participated internal dose to assess recent exposure to PAHs (Jacob in this study. The study was undertaken with the and Seidel, 2002). Due to the high exposure to PAHs in permission of the local authority and the Ethics coke oven workers, an occupational exposure limit (OEL) Committee of School of Public Health, Fudan Univer- of 2.3 μmol/mol creatinine in urinary 1-OHP level has sity, China. Informed consent was obtained from each been recommended by Jongeneelen (1992), which was subject. Subjects completed a structured questionnaire corresponding to a unit risk of 1.3 on lung cancer by the to elicit information about demographic characteristics, epidemiological study (Jongeneelen, 1992). including age (empirical half, ≤36 or N36), BMI At least two major metabolic pathways are involved (normal, ≤25; overweight, N25), level of education in the biotransformation of PAHs (Grover, 1986). Phase I (high school or below, college or above), alcohol enzymes such as CYP1A1, CYP1A2 and CYP1B1 are consumption (nondrinker or drinker), cigarette smoking known to catalyze the formation of reactive intermedi- habit (nonsmoker; moderate-smoker, ≤10; heavy-smok- ates of some carcinogens, whereas phase II enzymes er, N10) and respirator usage (yes or no) and job such as glutathione S-transferases (GSTs) are mainly classification. Coke oven workers were classified into found to participate in the detoxification of these inter- four subgroups according to their job classifications as mediates through the conjugation with glutathione. follows: extremely high exposure group including Some of these genes are also known to be transcription- lidman, tar chaser and whistler; high exposure group ally up-regulated by the activation of Aryl hydrocarbon including larry car operator, benchman at coke side and Receptor (AhR) when exposure to PAHs. In addition, benchman at pusher side; moderate exposure group little but still available information has also been pro- including machine operator at coke side, pusher machine vided on other genes coding enzymes involved in the operator, quenching car operator, oven repairman, metabolism of PAHs. CYP2E1 is expressed at higher temperature controller and heater; low exposure group levels in Asians than in Caucasians, therefore was including screening station operator, supervisor and thought to be responsible for the metabolism of pyrene in electrician. All these job classifications had been Asian people (Nan et al., 2001). Epoxide hydrolases described in our previous study (Chen et al., revised by (EHs) convert epoxides to more water-soluble trans- American Journal of Industrial Medicine, AJIM-06- dihydrodiols, which has been found to be an important 0232.R1). Categorized by exposure situation, our step of the activation of benzo[a]pyrene (Seidegard and previous study had found a positive trend arising from Ekstrom, 1997). UDP-glucuronosyltransferase (UGT) is the referents to extremely high exposure group (linear another phase II enzyme detoxifying PAHs in vivo trend test, pb0.001). (Bock, 1991). It is known that most genes coding enzymes involved 2.2. Determination of 1-hydroxypyrene in the metabolism of PAHs are polymorphic. Therefore, the study on the polymorphisms of metabolizing enzymes Participants were asked to provide a sample of spot will benefit us in understanding the inter-individual dif- urine (20 mL) at the end of working shift. All samples ference in the rate of activation or detoxification of car- were kept at − 20 °C until analysis. Determination of cinogens related to PAHs and underlining the possible 1-OHP was performed using HPLC according to Li et al. molecular mechanism of their carcinogenesis. The present (2003), a method that was also described by Leng et al. study examined the associations between selected poly- (2004) and Liu et al. (2006). The detection limit of morphisms in a number of genes coding PAHs metabo- urinary 1-OHP was 0.15 ng/mL. All samples were above lizing enzymes and urinary 1-OHP concentrations in coke this limit. Urine specimen from each subject was oven workers from two large coking plants in south analyzed in duplicate. Quality control samples were China. analyzed after each twenty of measurements. The recovery rates from the entire procedure were 89.1%, 2. Material and methods 88.8% and 91.6% at 5, 20 and 100 μg/L, respectively. The intraday relative standard deviation (RSD) values 2.1. Subjects were 5.0%, 5.3% and 7.2% at 5, 20 and 100 μg/L, respectively, and the interday RSD values were 6.7%, This study was conducted in May, 2003 at Coking 8.3% and 10.2%, respectively. Urinary 1-OHP concen- plant I in Xinyu, Jiangxi province and in Dec, 2003 at trations were adjusted by urinary creatinine excretion Coking plant II in Shanghai, both of which started in the (μmol/mol creatinine). 中国科技论文在线 http://www.paper.edu.cn

40 B. Chen et al. / Science of the Total Environment 381 (2007) 38–46

2.3. Determination of genotype Table 1 Subject characteristics and their influence on urinary 1-hydroxypyrene levels Genomic DNA was isolated from the whole blood using a DNA isolating kit (Qiagen, USA). Genetic poly- N (%) GM (GSD) p-value morphisms were determined using Taqman® based assays Plant I 197 (44) 5.18 (3.87) (Applied Biosystems, USA) for all single nucleotide II 250 (56) 4.21 (2.46) 0.053 polymorphisms (SNPs). GSTM1 and GSTT1 deletion Occupational Low 31 (7) 1.56 (3.31) exposure Moderate 166 (37) 2.89 (2.53) 0.001 polymorphisms were analyzed using a multiple PCR High 144 (32) 5.88 (2.89) b0.001 procedure (Arand et al., 1996). Except for the CYP2E1 Extremely high 106 (24) 9.44 (2.42) b0.001 DraI- polymorphism (78%), GSTM1 and GSTT1 deletion Age (years) ≤36 217 (49) 4.96 (3.53) polymorphisms (88%), 93–99% of samples at each N36 230 (51) 4.31 (2.66) 0.186 2 ≤ polymorphism site were successfully genotyped. BMI (kg/m ) 25 342 (77) 4.93 (3.06) N25 105 (23) 3.71 (3.08) 0.023 Level of education High school 423 (95) 4.67 (3.00) 2.4. Statistic analysis and below College and 24 (5) 3.73 (4.65) 0.340 Urinary 1-OHP concentrations were ln-transformed to above normalize the variance. Arlequin software (version 3.01) Alcohol Nondrinker 159 (36) 4.58 (2.90) – consumption Drinker 288 (64) 4.63 (3.19) 0.911 was used to ensure the fitting with Hardy Weinberg Cigarette smoking Nonsmoker 163 (36) 4.16 (3.39) equilibrium in each SNP and explore the linkage (cig/day) ≤10 122 (27) 4.50 (2.77) 0.553 disequilibrium between polymorphisms (Excoffier et al., N10 162 (36) 5.22 (2.99) 0.069 2005). One-way ANOVA was used to explore the Respirator usage No 202 (45) 5.34 (3.72) associations between urinary 1-OHP concentrations and Yes 245 (55) 4.09 (2.53) 0.012 potential impact factors including age, BMI, level of education, alcohol consumption, cigarette smoking habit, respirator usage and occupational exposure classification. heavy cigarette smoking (N10 cig/day) and respirator Linear regression models were used to examine the usage were also found to affect urinary excretion of 1- association between 1-OHP excretion and genetic poly- OHP. morphisms. First, we examined 1-OHP excretion in a Twenty-two polymorphisms in 9 genes potentially crude model with no covariates except gene polymor- involved in PAHs metabolism were analyzed in the phism. Then an adjusted model was used with the addition investigated subjects. Table 2 shows urinary 1-OHP of the covariates mentioned above. All statistic tests were concentrations by each genetic polymorphisms examined two sided (α=0.05) and done by using SPSS 10.0 and in this study. Using the linear regression analysis, STATA/SE 8.0 for windows. significant crude association was found in CYP1A2 IVS4+43ANG, GSTP1 I105V and GSTT1 deletion. 3. Results Adjustment for sex, age, BMI, level of education, alcohol consumption, cigarette smoking habit and level of occu- Urinary 1-OHP in 447 coke oven workers from pational exposure strengthened the association with two coking plants in south China was measured during GSTP1 I105V, but dismissed the association with May–Dec, 2003. Table 1 presents the distributions of CYP1A2 IVS4+43ANG as well as GSTT1 deletion. some demographic variables and their influence on The adjusted model also showed a significant association urinary 1-OHP levels. The concentrations of urinary with AhR R554K as well as UGT1A1 −3263TNG. 1-OHP were 1.9, 3.8 and 6.1-fold higher in moderate, Adjusted by the covariates, this association with high and extremely high exposure group than that in low UGT1A1 −3263TNG as well as GSTP1 I105V was exposure group (2.89, 5.88 and 9.44 versus 1.56 μmol/ also found by a recessive model analysis (subjects mol creatinine, respectively). The percentage of subjects carrying ≥ 1 major allele were compared with those who with 1-OHP concentrations exceeding the occupational were homozygous for the minor allele, data not shown). exposure level of 2.3 μmol/mol creatinine was 45, 64, The recessive model analysis did not show the asso- 83 and 94% in low, moderate, high and extremely ciation with other polymorphisms. high exposure group, respectively (data not shown). The Using linear trend test across homozygous major allele concentrations of urinary 1-OHP in workers at Plant I carriers to homozygous minor allele carriers, the effects of were slightly higher than that in Plant II (5.18 versus 22 polymorphisms were tested in a single gene model 4.21 μmol/mol creatinine, p=0.053).BMI(N25 kg/m2), (Table 3). After adjustment for the above covariates, 中国科技论文在线 http://www.paper.edu.cn

B. Chen et al. / Science of the Total Environment 381 (2007) 38–46 41

Table 2 Urinary 1-hydroxypyrene concentrations, regression coefficients and p-values for crude and adjusted models by genotype Polymorphic Genotype N GM Crude model Adjusted model c site a (%) (GSD) b β d p-value β d p-value AhR GG 166 (38) 4.20 (2.86) R554K GA 215 (49) 4.78 (3.15) 0.13 0.261 0.13 0.199 rs2068853 AA 61 (14) 5.12 (3.32) 0.20 0.236 0.26 0.067 CYP1A1 TT 140 (33) 4.56 (3.19) 3801TNC TC 223 (53) 4.85 (2.97) 0.06 0.597 0.01 0.914 rs4646903 CC 56 (13) 4.10 (3.05) −0.11 0.550 −0.09 0.546 CYP1A1 AA 260 (58) 4.67 (3.11) I462V AG 168 (38) 4.47 (3.06) −0.04 0.695 −0.01 0.925 rs1048943 GG 17 (4) 4.87 (3.01) 0.04 0.877 −0.06 0.809 CYP1A1 CC 301 (70) 4.49 (3.08) G45D CT 117 (27) 5.18 (3.12) 0.14 0.243 0.14 0.193 rs4646422 TT 14 (3) 5.65 (2.60) 0.23 0.454 0.15 0.578 CYP1A1 GG 152 (34) 4.51 (3.20) IVS1−728GNA GA 236 (53) 4.76 (3.06) 0.05 0.646 0.03 0.762 rs4646421 AA 55 (12) 4.13 (2.91) −0.09 0.621 −0.11 0.473 CYP1A2 GG 164 (39) 4.46 (3.38) EX1−5928GNA GA 194 (46) 4.81 (2.80) 0.08 0.519 0.06 0.528 rs2445618 AA 67 (16) 4.92 (3.01) 0.10 0.541 0.18 0.201 CYP1A2 TT 168 (40) 4.49 (3.38) − 163CNA TG 190 (45) 4.82 (2.79) 0.07 0.558 0.07 0.520 rs762551 GG 65 (15) 4.90 (3.06) 0.09 0.596 0.18 0.204 CYP1A2 GG 312 (74) 4.73 (3.01) IVS4+43ANG GA 106 (25) 4.26 (3.16) −0.11 0.399 −0.05 0.671 rs2472304 AA 6 (1) 15.39 (1.72) 1.18 0.010 0.63 0.114 CYP1A2 CC 317 (76) 4.67 (2.97) 5347CNT CT 94 (23) 4.89 (3.17) 0.04 0.732 0.08 0.477 rs2470890 TT 5 (1) 4.95 (5.32) 0.06 0.909 0.34 0.433 CYP1B1 CC 284 (64) 4.70 (3.09) R48G CG 138 (31) 4.44 (3.10) −0.06 0.629 0.02 0.862 rs10012 GG 22 (5) 4.53 (3.15) −0.04 0.885 0.11 0.595 CYP1B1 CC 321 (76) 4.52 (3.07) V432L CG 92 (22) 4.81 (3.03) 0.06 0.633 0.07 0.568 rs1056836 GG 7 (2) 7.16 (3.84) 0.46 0.285 0.32 0.386 CYP2E1 GG 292 (67) 4.52 (3.13) RsaI- GA 131 (30) 4.82 (2.99) 0.06 0.585 0.11 0.270 rs2031920 AA 12 (3) 3.93 (2.50) −0.14 0.673 0.05 0.872 CYP2E1 TT 215 (61) 4.54 (3.05) DraI- TA 122 (35) 5.30 (2.95) 0.15 0.215 0.08 0.449 rs6413432 AA 13 (4) 4.35 (2.03) −0.04 0.893 0.18 0.512 EPHX1 GG 206 (47) 4.57 (2.82) EX1−362GNA GA 184 (42) 4.80 (3.30) 0.05 0.677 0.13 0.176 rs2854451 AA 47 (11) 4.04 (3.17) −0.12 0.490 0.05 0.763 EPHX1 GG 231 (53) 4.23 (3.19) IVS2+20GNA GA 173 (40) 5.02 (2.87) 0.17 0.122 0.02 0.863 rs3738047 AA 30 (7) 5.04 (2.65) 0.18 0.413 0.47 0.210 EPHX1 AA 339 (77) 4.51 (3.06) H139R AG 95 (22) 4.67 (3.21) 0.03 0.800 0.16 0.103 rs2234922 GG 7 (2) 6.20 (2.49) 0.32 0.460 0.02 0.922 EPHX2 GG 282 (64) 4.54 (3.06) R287Q GA 147 (33) 5.03 (3.03) 0.10 0.371 0.08 0.438 rs751141 AA 15 (3) 3.14 (3.72) −0.37 0.216 −0.32 0.215 UGT1A1 TT 206 (48) 5.11 (2.96) −3263TNG TG 173 (40) 4.34 (3.29) −0.16 0.155 −0.18 0.069 rs4124874 GG 51 (12) 3.92 (2.82) −0.26 0.133 −0.37 0.013 UGT1A1 GG 298 (70) 4.58 (3.09) (continued on next page) 中国科技论文在线 http://www.paper.edu.cn

42 B. Chen et al. / Science of the Total Environment 381 (2007) 38–46

Table 2 (continued ) Polymorphic Genotype N GM Crude model Adjusted model c site a (%) (GSD) b β d p-value β d p-value G71R GA 110 (26) 5.38 (2.80) 0.16 0.191 0.16 0.128 rs4148212 AA 15 (4) 3.53 (3.18) −0.26 0.374 0.05 0.851 GSTP1 AA 297 (67) 4.93 (3.13) I105V AG 131 (30) 4.10 (2.73) −0.18 0.119 −0.11 0.271 rs1695 GG 14 (3) 2.91 (5.50) −0.53 0.087 −0.71 0.008 GSTM1 Present 158 (40) 4.63 (3.10) deletion Null 236 (60) 4.60 (2.93) −0.01 0.939 −0.14 0.171 GSTT1 Present 173 (44) 5.21 (2.88) deletion Null 223 (56) 4.22 (3.09) −0.21 0.060 −0.10 0.304 a Gene, trivial name of polymorphism site and dbSNP ID number. b Geometric mean (Geometric standard deviation), μmol/mol creatinine. c Adjusted model included variables for plant, level of occupational exposure, age, BMI, level of education, alcohol consumption, cigarette smoking and respirator usage. d Coefficient of the model.

significant association was found in AhR R554K, UGT1A1 −3263TNG and GSTP1 I105V, but not in the other 19 polymorphisms. Among these 22 polymorph- Table 3 Adjusted single or multi-gene regression models for the associations isms, strong linkage disequilibrium was found between between genotypes and urinary 1-hydroxypyrene concentrations CYP1A1 IVS1−728GNA and CYP1A1 3801TNC(D′= 1.00, r2 =0.98), CYP1A2 EX1−5928GNAandCYP1A2− Variables Single gene Multi-gene N ′ 2 N model model a 163C A(D =1.00, r =0.98), CYP1A2 IVS4+43A G and CYP1A2 5347CNT(D′=0.84, r2 =0.63), as well as β b p-value β b p-value CYP2E1 DraI- and CYP2E1 RsaI- (D′=0.93, r2 =0.67). AhR R554K c 0.13 0.053 0.12 0.098 − N − N − Excluded CYP1A1 IVS1 728G A, CYP1A2 EX1 CYP1A1 3801T C 0.03 0.657 0.05 0.627 N N CYP1A1 I462V −0.02 0.839 0.01 0.933 5928G A, CYP1A2 IVS4+43A GandCYP2E1DraI- CYP1A1 G45D 0.11 0.191 0.08 0.470 for the reason of linkage disequilibrium, the effects of CYP1A1 IVS1−728GNA −0.03 0.659 / d other 18 polymorphisms were tested in a combined model CYP1A2 EX1−5928GNA 0.08 0.206 / (Table 3). The combined model only slightly reduced the − N CYP1A2 163C A 0.08 0.207 0.11 0.194 associations with AhR R554K, UGT1A1 −3263TNGand CYP1A2 IVS4+43ANG 0.03 0.749 / CYP1A2 5347CNT 0.10 0.339 0.18 0.129 GSTP1 I105V. 1-OHP excretion was not associated with CYP1B1 R48G 0.04 0.649 0.02 0.797 other polymorphisms in the combined model. A model CYP1B1 V432L 0.09 0.369 0.13 0.204 without genetic polymorphisms had a total r2 of 0.286. CYP2E1 RsaI- 0.08 0.336 0.09 0.328 Adding these polymorphisms improved the total r2 to CYP2E1 DraI- 0.08 0.350 / 0.333, causing a slight increase (15%) in the explanatory EPHX1 EX1−362GNA 0.06 0.375 0.09 0.265 EPHX1 IVS2+20GNA 0.08 0.282 0.09 0.305 power of the model. EPHX1 H139R 0.07 0.461 −0.01 0.899 The gene–gene interactions were also examined in EPHX2 R287Q −0.01 0.938 −0.02 0.843 this study. It was found that GSTT1 deletion polymor- UGT1A1 –3263TNG −0.18 0.006 −0.17 0.033 phism interacted with AhR R554K or UGT1A1 − UGT1A1 G71R 0.11 0.218 0.01 0.958 3263TNG(Table 4). Adjusting for all the covariates had GSTP1 I105V −0.20 0.021 −0.19 0.040 GSTM1 deletion −0.14 0.171 −0.11 0.299 only a minor impact on the associations (data not GSTT1 deletion −0.10 0.304 −0.07 0.516 shown). The GSTT1 present carriers had 1.2-fold higher a r2 =0.333. Adjusted model included variables for plant, level of 1-OHP levels than GSTT1 null carriers, the difference occupational exposure, age, BMI, level of education, alcohol increased to 1.4-fold in the subjects with AhR GG consumption, cigarette smoking and respirator usage. carriers (5.17 vs. 3.64 μmol/mol creatinine, p=0.038), b Coefficient. and 1.7-fold in the subjects with UGT1A1 −3263TNG c Genotype across homozygous major allele, heterozygous to TG/GG carriers (5.67 vs. 3.38 μmol/mol creatinine, homozygous minor allele. d These polymorphisms were omitted from the multi-gene model for p=0.001). Also, 1-OHP levels were 1.2-fold higher in their strong linkage disequilibriums with other polymorphisms UGT1A1 −3263TNGTTcarriersthaninTG/GG (see details in the text). carriers, the difference increased to 1.6-fold in the 中国科技论文在线 http://www.paper.edu.cn

B. Chen et al. / Science of the Total Environment 381 (2007) 38–46 43

Table 4 Effect of gene–gene interaction on urinary 1-hydroxypyrene levels GSTT1 deletion Present Null Total p-value b N GM (GSD) a N GM (GSD) N GM (GSD) Total 173 5.21 (2.88) 223 4.22 (3.09) 0.060 AhR R554K GG 57 5.17 (2.61) 91 3.64 (2.77) 166 4.20 (2.86) 0.038 GA/AA 116 5.23 (3.03) 131 4.67 (3.31) 276 4.86 (3.18) 0.445 p-value c 0.950 0.106 0.186 UGT1A1 −3263TNG TT 77 4.70 (3.17) 110 5.46 (2.73) 206 5.11 (2.96) 0.349 TG/GG 87 5.67 (2.74) 109 3.38 (3.23) 224 4.24 (3.18) 0.001 p-value c 0.272 0.001 0.084 a Geometric mean (Geometric standard deviation), μmol/mol creatinine. b The corresponding GSTT1 present carriers compared to null carriers. c The corresponding minor allele carriers compared to homozygous major allele carriers in AhR R554K or in UGT1A1 −3263TNG polymorphism.

subjects of GSTT1 null carriers (5.46 vs. 3.38 μmol/mol reasonable sample size (447 subjects) in this study also creatinine, p=0.001). allows us to explore the possible gene–gene interactions on urinary 1-OHP levels. 4. Discussion We did not found the modulations of CYP polymorphisms on urinary 1-OHP levels, except that six CYP1A2 We measured urinary 1-OHP in 447 healthy subjects IVS4+43ANG AA carriers showed a significantly higher from two large coking plants in south China. Each plant level compared to GG carriers. This significance was also was similar in coke manufacturing and had equal repre- disappeared when adjusting other factors including level sentative occupational exposure categories. The signif- of occupational exposure and cigarette smoking. Basical- icant difference in 1-OHP excretions from two plants ly, 1-OHP is the main metabolite of pyrene by one step could be partly explained by higher PAHs exposure in biotransformation of CYP450. CYP1A1, CYP1A2 and Plant I than that in Plant II. Area sampling data showed CYP1B1 are three major CYP450 enzymes that are that the mean airborne benzene soluble fraction was capable of the hydroxylation of pyrene in vitro (Kim et al., higher in Plant I than that in Plant II (data not shown). 2004). However, the modulations of CYP450 polymorph- An unusual result from our data was the negative effect isms on urinary 1-OHP levels are still controversial. In the on 1-OHP excretion by BMI. This finding was con- case of CYP1A1, most investigations found negative sistent with Roggi et al. (1997) and Mcclean et al. results except one study showed increased 1-OHP (2004). One possible explanation for this phenomenon excretions related to 3801C variant (Wu et al., 1998), is that increased BMI may result in increased storage of three others 462Val variant (Alexandrie et al., 2000; pyrene in fatty tissue so that less of pyrene is metab- Nerurkar et al., 2000; Lee et al., 2001). CYP1A2 olized into 1-OHP. 3860ANG (rs2069514) and CYP1B1 7644GNC We found a positive influence of AhR, UGT1A1, (rs10012) polymorphisms were reported to be associated GSTP1 and GSTT1 polymorphisms on the levels of with urinary 1-OHP glucuronide (1-OHPG) levels in 1-OHP in urine after controlling relevant confoundings Abnet et al.'s (2007) study. However, there were no posi- including occupational exposure level. However, no other tive associations between urinary 1-OHP (or 1-OHPG) polymorphisms were found to be associated with urinary levels and CYP1A2 polymorphisms or CYP1B1 poly- 1-OHP levels in coke oven workers. Some researchers morphisms based on other studies (Yang et al., 2003; Rihs have investigated the urinary 1-OHP concentrations in et al., 2005). Nan et al. (2001) found a significant coke oven workers and the possible modulations by gene modulation on 1-OHP excretion by CYP2E1 RsaI- polymorphisms (Ovrebo et al., 1998; Wu et al., 1998; polymorphism. This association was not found in our Zhang et al., 2001; Kuljukka-Rabb et al., 2002; Siwinska results. et al., 2004; Wu et al., 2004). Inconsistent results were A variant causing an amino acidic exchange (His- observed in those studies, which could be related to the 139-Arg, H139R) in EPHX1 was associated with 25% relatively small study size. Up to our knowledge, the decrease of enzyme activity (Ssett et al., 1994) and present study is one of the largest cross-sectional related to the significant decrease of urinary hydro- investigations in PAHs-exposed workers in the world. A xyphenanthrenes levels in German PAHs-exposed 中国科技论文在线 http://www.paper.edu.cn

44 B. Chen et al. / Science of the Total Environment 381 (2007) 38–46

workers (Rihs et al., 2005). Although pyrene 4,5-oxide carriers which were highly exposed to PAHs. The GSTP1 could be hydrolyzed by microsomal EHs (Jacob and 105V variant of the enzyme was found to have higher Seidel, 2002), we did not find the association with catalytic activity towards PAHs intermediates compared to EHPX1 and EPHX2 on urine 1-OHP excretions. Our 105I variant (Harries et al., 1997). The increased GSTP1 data showed that EHs might not be an important enzyme activity may reduce pyrene derivatives that are otherwise in the metabolism of pyrene. going through glucuronidation pathway. GSTs have been considered to be important enzymes A polymorphism in the promoter region of UGT1A1 in the detoxification of PAHs in vivo. An association (−3263TNG) was found to affect urinary 1-OHP levels in between urinary 1-OHP level and GSTM1 deletion this study. 1-OHP is excreted in urine as a glucuronide polymorphism had been reported in aluminum potroom conjugate, 1-OHPG. A recent study has shown that − workers (Alexandrie et al., 2000) and aircraft mainte- 3263TNG was related to the decreased transcriptional nance workers (Lee et al., 2001). This association was activity of UGT1A1 (Sugatani et al., 2002). Deficiency of not observed either in non-occupational individuals glucuronidation pathway may stimulate other pyrene (Nerurkar et al., 2000; Yang et al., 2003) or in other metabolism pathways, resulting of the up-regulation highly exposed workers such as coke oven workers of other pyrene metabolites and the down-regulation of (Zhang et al., 2001; Wu et al., 2004). The deficiency in 1-OHPG. However, more work need to be done in the GSTM1 activity was suspected to stimulate the glucur- influence of UGT on PAHs metabolism considering onidation pathway and the induction of CYP1A1 by that UGT1A7–UGT1A10 showed higher conjugation of accumulating PAHs derivatives that were otherwise 1-OHP than UGT1A1 in vitro (Luukkanen et al., 2005) conjugated to glutathione (Vaury et al., 1995). In and UGT1A1 −3263TNG may be linked to other impor- Alexandrie et al.'s (2000) and Lee et al.'s (2001) studies, tant polymorphisms, which may cause the functional CYP1A1 polymorphisms were also found to be asso- disturbance of other UGT enzymes involved in the ciated with urinary 1-OHP levels. In fact, this association metabolism of PAHs. with CYP1A1 polymorphisms was negative in most Our data also showed that AhR R554K polymor- researches of GSTM1 deletion polymorphism (Nan phism could slightly affect the excretion of 1-OHP in et al., 2001; Zhang et al., 2001; Apostoli et al., 2003; urine. This trend was not found in Abnet et al.'s (2007) Rihs et al., 2005). It seems that the modulation of study using 1-OHPG as the bioindicator. However, GSTM1 deletion polymorphism on urinary 1-OHP another SNP named 2290GNA (rs4986826) was asso- levels was related to the association with CYP1A1 ciated with 1-OHPG excretion in Abnet et al.'s study. polymorphisms. Some PAHs can bind and activate the AhR which We found a higher urinary 1-OHP level in GSTT1 subsequently up-regulates the transcription of other present carriers compared to null carriers. This trend was genes including CYP450 (Ramadoss et al., 2005) and also found in a non-occupationally exposed population UGT (Zhou et al., 2005). In the present study, we found with a huge study size (n=661)(Yang et al., 2003), but not that urinary 1-OHP levels were associated with in most other reports based on small populations (nb200) UGT1A1, but not with CYP450. It is possible that (Alexandrie et al., 2000; Lee et al., 2001; Nan et al., 2001; AhR R554K polymorphism can change the inducibility Zhang et al., 2001). However, previous studies did not of UGT as well as CYP450, resulting the increased explore the interactions between GSTT1 and other excretion of pyrene metabolites. polymorphisms. In this study, we found that GSTT1 When gene–gene interactions were examined, we significantly enhanced urinary 1-OHP excretions when found the potential interaction between GSTT1 deletion taking into account AhR R554K or UGT1A1 −3263TNG polymorphism and UGT1A1 −3263TNG on 1-OHP ex- polymorphism, which indicates that GSTT1 may be an cretions, as well as the interaction between GSTT1 dele- important enzyme in the metabolism of PAHs in highly tion polymorphism and AhR R554K (Table 4). Giving the exposed workers. fact in our data that increased 1-OHP excretions were There is a lack of information on the role of GSTP1 related to AhR 554K variant, decreased 1-OHP excretions in the metabolism of PAHs. No significant influence of were related to GSTT1 null type and UGT1A1 −3263G GSTP1 was found in Ovrebo et al.'s (1998), Alexandrie variant, one would expect the synergistic effect on 1-OHP et al.'s (2000), Nerurkar et al.'s (2000), Zhang et al.'s excretions by GSTT1 null type and UGT1A1 −3263G (2001), and Rihs et al.'s (2005) studies. However, Schoket variant (3.38 μmol/mol creatinine), or by GSTT1 present et al. (2001) found a slightly lower 1-OHP level in GSTP1 type and AhR 554K variant (5.23 μmol/mol creatinine). I105V VV carriers. Here we reported a significantly dec- However, the biological mechanisms of these interactions reased urinary 1-OHP level in GSTP1 I105V V variant are still under cover. One possible explanation would be 中国科技论文在线 http://www.paper.edu.cn

B. Chen et al. / Science of the Total Environment 381 (2007) 38–46 45

that glutathione conjugation pathway and glucuroni- Excoffier LG, Laval, Schneider S. Arlequin ver. 3.0: an integrated dation pathway can interact with each other, so that the software package for population genetics data analysis. Evol Bioinformatics Online 2005;vol. 1:47–50. deficiency of one pathway will stimulate the other one by Grover PL. Pathways involved in the metabolism and activation of accumulating PAHs derivatives that were otherwise going polycyclic hydrocarbons. Xenobiotica 1986;16(10–11):915–31. through the first one. Harries LW, Stubbins MJ, Forman D, Howard GC, Wolf CR. The present study did not give an individual cumu- Identification of genetic polymorphisms at the glutathione S- lative exposure assessment of PAHs at the coking plants, transferase Pi locus and association with susceptibility to bladder, testicular and prostate cancer. Carcinogenesis 1997;18(4): 641–4. since routine surveillance of PAHs in workplace was IARC, editor. Polynuclear aromatic compounds. Part 1. Chemicals, limited by technology and data on PAHs levels were environmental and experimental data. IARC Monogr Eval lack. We also did not investigate other sources of PAHs Carcinog Risks HumLyon, France: International Agency for exposure such as dermal and dietary intake. Therefore, a Research on Cancer; 1983. comprehensive external exposure assessment should be IARC. International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risk of chemicals to humans polynuclear done in the future study to support the job classification aromatic compounds. Lyon, France: IARC; 1984. p. 101–31. in this study. In addition, our findings in the associations IARC. Overall evaluations of carcinogenicity: an updating of IARC between genotype and urinary 1-OHP were based on the monographs, vols 1 to 42. IARC Monogr Eval Carcinog Risks cross-section study, the cohort study will help to further Hum. Supplement 7. Lyon, France: International Agency for unravel the contribution of genetic polymorphism to Research on Cancer; 1987. Jacob R, Seidel A. Biomonitoring of polycyclic aromatic hydro- urinary 1-OHP in coke oven workers. carbons in human urine. J Chromatogr B 2002;778(1–2):31–47. In conclusion, our data showed that urinary 1-OHP Jongeneelen FJ. Biological exposure limit for occupational exposure to concentrations in Chinese coke oven workers were coal–tar pitch volatiles at cokeovens. Int Arch Occup Environ associated with AhR, GSTT1, GSTP1 and UGT1A1 Health 1992;63(8):511–6. polymorphisms, but not with CYP450, mEH and GSTM1 Kim YD, Todoroki H, Oyama T, Isse T, Matsumoto A, Yamaguchi T, et al. Identification of cytochrome P450 isoforms involved in polymorphisms after adjusting for relevant covariant 1-hydroxylation of pyrene. Environ Res 2004;94(3):262–6. factors. Kuljukka-Rabb T, Nylund L, Vaaranrinta R, Savela K, Mutanen P, Veidebaum T, et al. The effect of relevant genotypes on PAH Acknowledgements exposure-related biomarkers. J Expo Anal Environ Epidemiol 2002;12(1):81–91. Lee CY, Lee JY, Kang JW, Kim H. Effects of genetic polymorphisms We thank Dr. Jusheng Pan, Anti-epidemic and of CYP1A1, CYP2E1, GSTM1, and GSTT1 on the urinary levels Prevention Station of Jiangxi Xinyu Iron-Steel Compa- of 1-hydroxypyrene and 2-naphthol in aircraft maintenance ny, and Dr. Xiong Sun, Shanghai Coke Oven Company, workers. Toxicol Lett 2001;123(2–3):115–24. for their kind assistants. This study was supported by Leng SG, Dai YF, Niu Y, Pan ZF, Li XH, Cheng J, et al. Effects of genetic “973” project from National Department of Science and polymorphisms of metabolic enzymes on cytokinesis-block micro- nucleus in peripheral blood lymphocyte among coke-oven workers. Technology, P. R. China (2002CB512902). Cancer Epidemiol Biomark Prev 2004;13(10):1631–9. Li X, Leng SG, Guo J, Zheng YX. An improved high performance References liquid chromatography method for determination of 1-hydroxypyrene in urine. J Hyg Res 2003;32(6):616–7 (in Chinese). Liu AL, Lu WQ, Wang ZZ, Chen WH, Lu WH, Yuan J, et al. Elevated Abnet CC, Fagundes RB, Strickland PT, Kamangar F, Roth MJ, Taylor levels of urinary 8-hydroxy-2′-deoxyguanosine, lymphocytic PR, et al. The influence of genetic polymorphisms in Ahr, CYP1A1, micronuclei, and serum glutathione S-transferase in workers CYP1A2, CYP1B1, GST M1, GST T1, and UGT1A1 on urine 1- exposed to coke oven emissions. Environ Health Perspect hyrdoxypyrene glucuronide concentrations in healthy subjects from 2006;114(5):673–7. Rio Grande do Sul, Brazil. Carcinogenesis 2007;28(1): 112–7. Luukkanen L, Taskinen J, Kurkela M, Kostiainen R, Hirvonen J, Finel Alexandrie AK, Warholm M, Carstensen U, Axmon A, Hagmar L, M. Kinetic characterization of the 1A subfamily of recombinant Levin JO, et al. CYP1A1 and GSTM1 polymorphisms affect human UDP-glucuronosyltransferases. Drug Metab Dispos urinary 1-hydroxypyrene levels after PAH exposure. Carcinogen- 2005;33(7):1017–26. esis 2000;21(4):669–76. Mcclean MD, Rinehart RD, Ngo L, Eisen EA, Kelsey KT, Wiencke Apostoli P, Neri G, Lucas D, Manno M, Berthou F. Influence of JK, et al. Urinary 1-hydroxypyrene and polycyclic aromatic genetic polymorphisms of CYP1A1 and GSTM1 on the urinary hydrocarbon exposure among asphalt paving workers. Ann Occup levels of 1-hydroxypyrene. Toxicol Lett 2003;144(1):27–34. Hyg 2004;48(6):565–78. Arand M, Muhlbauer R, Hengstler J, Jager E, Fuchs J, Winkler L, et al. Nan H, Kim H, Lim HS, Choi JK, Kawamoto T, Kang JW, et al. Effects A multiplex polymerase chain reaction protocol for the simulta- of occupation, lifestyle and genetic polymorphisms of CYP1A1, neous analysis of the glutathione S-transferase GSTM1 and CYP2E1, GSTM1 and GSTT1 on urinary 1-hydroxypyrene and GSTT1 polymorphisms. Anal Biochem 1996;236:184–6. 2-naphthol concentrations. Carcinogenesis 2001;22:787–93. Bock KW. Roles of UDP-glucuronosyltransferases in chemical Nerurkar PV, Okinaka L, Aoki C, Seifried A, Lum-Jones A, Wilkens carcinogenesis. Crit Rev Biochem Mol Biol 1991;26(2):129–50. LR, et al. CYP1A1, GSTM1, and GSTP1 genetic polymorphisms 中国科技论文在线 http://www.paper.edu.cn

46 B. Chen et al. / Science of the Total Environment 381 (2007) 38–46

and urinary 1-hydroxypyrene excretion in non-occupationally Sugatani J, Yamakawa K, Yoshinari K, Machida T, Takagi H, Mori M, exposed individuals. Cancer Epidemiol Biomark Prev 2000;9(10): et al. Identification of a defect in the UGT1A1 gene promoter and 1119–22. its association with hyperbilirubinemia. Biochem Biophys Res Ovrebo S, Ryberg D, Haugen A, Leira HL. Glutathione S-transferase Commun 2002;292(2):492–7. M1 and P1 genotypes and urinary excretion of 1-hydroxypyrene in Vaury C, Laine R, Noguiez P, de Coppet P, Jaulin C, Praz F, et al. coke oven workers. Sci Total Environ 1998;220(1):25–31. Human glutathione S-transferase M1 null genotype is associated Ramadoss P, Marcus C, Perdew GH. Role of the aryl hydrocarbon with a high inducibility of cytochrome P450 1A1 gene receptor in drug metabolism. Expert Opin Drug Metab Toxicol transcription. Cancer Res 1995;55(23):5520–3. 2005;1(1):9–21. WHO. Air quality guidelines for Europe. WHO Regional Publication, Rihs HP, Pesch B, Kappler M, Rabstein S, Rossbach B, Angerer J, European SeriesCopenhagen: World Health Organization; 1987. et al. Occupational exposure to polycyclic aromatic hydrocarbons Wu MT, Huang SL, Ho CK, Yeh YF, Christiani DC. Cytochrome P450 in German industries: association between exogenous exposure 1A1 MspI polymorphism and urinary 1-hydroxypyrene concentra- and urinary metabolites and its modulation by enzyme poly- tions in coke-oven workers. Cancer Epidemiol Biomark Prev morphisms. Toxicol Lett 2005;157(3):241–55. 1998;7(9):823–9. Roggi C, Sciarra GF, Apostoli P, Maccarini L, Magnaghi S, Cenni A, Wu MT, Pan CH, Chen CY, Chen CJ, Huang LH, Tsai LY, et al. Lack et al. Urinary 1-hydroxypyrene as a marker of exposure to pyrene: of modulating influence of CSTM1 and CSTT1 polymorphisms on an epidemiological survey on a general population group. Sci Total urinary biomonitoring markers in coke-oven workers. Am J Ind Environ 1997;199(3):247–54. Med 2004;46:112–9. Schoket B, Papp G, Levay K, Mrackova G, Kadlubar FF, Vincze I. Yang MH, Jang JY, Kim S, Lee SM, Chang SS, Cheong HK, et al. Impact of metabolic genotypes on levels of biomarkers of Genetic effects on urinary 1-hydroxypyrene levels in a Korean genotoxic exposure. Mutat Res Fundam Mol Mech Mutagen population. Carcinogenesis 2003;24(6):1085–9. 2001;482(1–2):57–69. Zhang J, Ichiba M, Hara K, Zhang S, Hanaoka T, Pan G, et al. Urinary Seidegard J, Ekstrom G. The role of human glutathione transferases 1-hydroxypyrene in coke oven workers relative to exposure, and epoxide hydrolases in the metabolism of xenobiotics. Environ alcohol consumption, and metabolic enzymes. Occup Environ Health Perspect 1997;105(Suppl4):791–9. Med 2001;58(11):716–21. Siwinska E, Mielzynska D, Kapka L. Association between urinary Zhou J, Zhang J, Xie W. Xenobiotic nuclear receptor-mediated 1-hydroxypyrene and genotoxic effects in coke oven workers. regulation of UDP-glucuronosyl-transferases. Curr Drug Metab Occup Environ Med 2004;61(3):e10. 2005;6(4):289–98. Ssett 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(3):421–8.