The effect of relevant genotypes on PAH exposure-related biomarkers

Transcription

1 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12, # 2002 Nature Publishing Group All rights reserved /02/$ The effect of relevant genotypes on PAH exposure-related biomarkers TERHI KULJUKKA-RABB, a LARS NYLUND, b RAIJA VAARANRINTA, a KIRSTI SAVELA, a PERTTI MUTANEN, d TOOMAS VEIDEBAUM, e MARJA SORSA, f AGNETA RANNUG c AND KIMMO PELTONEN a a Institute of Occupational Health, FIN Helsinki, Finland b National Product Control Agency for Welfare and Health, FIN Helsinki, Finland c Karolinska Institute and National Institute for Working Life, Stockholm, Sweden d Department of Epidemiology and Biostatistics, FIOH, FIN Helsinki, Finland e Institute of Experimental and Clinical Medicine, Tallinn, Estonia f Ministry of Education, FIN Helsinki, Finland Polycyclic aromatic hydrocarbons ( PAHs) in coke oven emissions cause a cancer risk to humans. In a comprehensive biomonitoring study among Estonian coke oven workers, we looked at the effect of genetic polymorphisms in metabolic enzymes on urinary mutagenicity, 1 - hydroxypyrene ( 1- OHP) concentration in urine, and aromatic DNA adducts in white blood cells ( WBCs). Coke oven workers were sampled twice ( samplings I and II), and controls only once at the time of sampling I. Urinary mutagenicity was measured using the Ames test. CYP1A1, microsomal epoxide hydrolase ( meh), and glutathione S - transferase ( GST) genotypes were analyzed by polymerase chain reaction ( PCR). Urinary mutagenicity did not differ between exposed and controls, but those coke oven workers who were smokers had significantly higher ( P=0.0002) mutagenic activity in urine than nonsmokers. Urinary mutagenicity was moderately correlated to levels of 1 - OHP and aromatic DNA adducts, the P values ranging from to Carriers of a variant allele in exon 4 of meh (Arg 139 ) had elevated urinary mutagenicity (sampling I ). In addition, urine mutagenicity of persons with predicted high meh activity was significantly higher. Smoking habit did not explain the differences observed in urinary mutagenicity between meh phenotype or genotype subgroups. Variation in exon 3 of meh (His 113 ) was related to a significantly ( P= 0.01) higher 1 - OHP concentration in exposed workers (sampling II). Workers from sampling I who had an Arg 139 variation in meh had lower levels of adducts in lymphocytes ( P= 0.01) than others, while airborne benzo[a]pyrene (B[ a]p ) and His 113 variation affected interactively on adduct levels. Our study shows that a comprehensive assessment of exposure is essential for elucidation of PAH exposure at a workplace. Even at high exposures metabolic polymorphisms seem to have some effect on biomarker levels, and should be assessed in biomonitoring studies. Journal of Exposure Analysis and Environmental Epidemiology ( 2002) 12, DOI: / sj / jea/ Keywords: aromatic DNA adducts, coke oven workers, genetic polymorphism, 1 - hydroxypyrene, urine mutagenicity. Introduction Epidemiological data show that exposure to polycyclic aromatic hydrocarbons ( PAHs) is possibly carcinogenic to humans and believed to be mainly responsible for an elevated lung cancer incidence among coke oven workers ( International Agency for Research on Cancer ( IARC), 1983;( International Agency for Research on Cancer ( IARC), 1984). In addition, other carcinogens, such as aromatic amines and heterocyclic compounds, are known to 1. Abbreviations: B[ a ]P, benzo[ a ]pyrene; 1- OHP, 1- hydroxypyrene; PAH, polycyclic aromatic hydrocarbon; CYP, cytochrome P450; GST, glutathione S - transferase; meh, microsomal epoxide hydrolase; BPDE, benzo[ a]pyrene- dihydrodiol- epoxide; WBC, white blood cell; PCR/ RFLP, polymerase chain reaction/ restriction fragment length polymorphism; DMSO, dimethyl sulfoxide; Ile, isoleucine; Val, valine; Ala, alanine; Tyr, tyrosine; His, histidine; Arg, arginine 2. Address all correspondence to: Dr. Kimmo Peltonen, Finnish Institute of Occupational Health, Topeliuksenkatu 41 a A, FIN Helsinki, Finland. Tel.: E- mail: kimmo.peltonen@ttl.fi Received 28 November be present in coke oven emissions ( IARC, 1984). Recent studies reveal both high and low mean exposure levels in operating coke oven plants. Most carcinogenic chemicals, such as PAHs, need to be activated to electrophilic reactants by metabolism in vivo (Miller and Miller, 1981). It is known at present that many of the genes coding enzymes involved in activation and detoxification of PAHs are polymorphic, which renders humans at variable risk of developing cancer if exposed (Bartsch and Hietanen, 1996; Daly et al., 1993; d Errico et al., 1996). The metabolism of PAHs is mediated in the initial oxidative step by cytochrome P450 gene products, mainly the isoenzyme CYP1A1 (Daly et al., 1993; Rannug et al., 1995). In phase II conjugation reactions, many carcinogens are detoxified by cytosolic glutathione S- transferases ( GSTs) ( Eaton and Bammler, 1999; Ketterer. 1988). On the other hand, epoxide hydrolases (EHs) convert biologically reactive epoxides to more watersoluble trans- dihydrodiols, which can have either beneficial or harmful effects, depending on subsequent metabolic pathways ( Daly et al., 1993; Hassett et al., 1994).

2 Kuljukka- Rabb et al. Effect of genotypes on PAH exposure-related biomarkers In 1990 a Japanese group found a close correlation between a polymorphic site in the 3 0 noncoding region of CYP1A1 and smoking-related lung cancer (Kawajiri et al., 1990). The striking relationship between mutations in the CYP1A1 gene and cancer have not been found in Caucasians, among whom the frequency of mutations is approximately 10- fold lower compared to Japanese people ( Hirvonen, 1995). There are three classes of dimeric cytosolic GSTs (,, ) in which polymorphisms have been identified (Eaton and Bammler, 1999). About 40% of Caucasians lack GSTM1 activity (Hirvonen, 1995; Ketterer et al., 1992). The most prevalent GST protein in human lung is GSTP1. Benzo[a]- pyrene ( B[ a]p)- diol epoxides are inactivated by both GSTM1 and GSTP1 enzymes (Ketterer et al., 1992). Four allelic variants of GSTP1 with differing activities towards components in PAH mixtures are known to exist in humans (Hu et al., 1998). The most compelling evidence for the association between alterations in xenobiotic- metabolizing enzymes and cancer is the up to 41-fold increased risk for lung cancer in individuals having simultaneously variant CYP1A1 and GSTM1 genes (Nakachi et al., 1993). The relationship between microsomal epoxide hydrolase ( meh) activity and susceptibility to smoking- induced cancers has recently been investigated ( Heckbert et al., 1992). However, in view of the role of meh in both toxication and detoxication reactions of PAH epoxides, the relationship between levels of enzyme activity and susceptibility to cancers is likely to be complex ( Daly et al., 1993). Urinary mutagenicity has long been used as a marker of exposure to environmental genotoxins of variable nature. Elevated urinary mutagenicity levels ( Clonfero et al., 1995) and DNA adducts in white blood cells (WBCs) of coke oven workers have been reported in several studies (Hemminki et al., 1990; Rojas et al., 1995). Overall, the findings concerning occupational PAH exposure and urinary mutagenicity are inconsistent ( see references in Clonfero et al., 1995). Besides the strong confounding effect of smoking on urine mutagenicity, its sensitivity as a biomarker for most occupational coal tar exposures has been criticized ( Clonfero et al., 1990). Fereira et al. ( 1994) concluded in a cross- sectional study on coke and graphiteelectrode- producing plant workers that, on a group level, urinary mutagenicity might contribute to the detection of workers exposed to genotoxic PAHs. Recently it has been demonstrated that PAH exposure- related urinary mutagenicity might vary depending on an individual s genetic characteristics ( Bartsch et al., 1995; Gabbani et al., 1996; Hirvonen et al., 1994). Many studies have revealed no or only a tentative effect of the GSTM1 null genotype on DNA adducts or urinary metabolites of PAHs ( Grinberg- Funes et al., 1994; Hemminki et al., 1997; Mooney et al., 1997; Motykiewicz et al., 1998; Nielsen et al., 1996). Instead, Binkova et al., 1998 reported a significant effect of the GSTM1 null genotype on urinary PAH metabolites and mutagenicity, but no effect on DNA adducts in WBCs of an environmentally exposed nonsmoking population. In studies reported by Hemminki et al. (1997) and Ovrebo et al. ( 1998), no significant correlation of GSTM1 or CYP1A1 genotypes to urinary 1-hydroxypyrene (1-OHP) was observed in high PAH exposure occupations. Rojas et al. ( 1998) reported a highly significant difference in B[ a]p- dihydrodiol- epoxide ( BPDE)- derived adducts in lung tissue of smoking lung cancer patients and in WBCs of coke oven workers depending on GSTM1 genotype. On the contrary, Rothman et al. (1995) saw a trend towards lower PAH DNA adduct level in WBCs in fire fighters with either inactive GSTM1 or inducible CYP1A1 genotype. Some studies have revealed an effect of the exon 7 mutation in CYP1A1 on DNA adducts in WBCs (Butkiewicz et al., 1998; Mooney et al., 1997). The objective of the present study is to investigate if genetic polymorphism in enzymes involved in PAH metabolism have any influence on biomarkers of exposure, i.e., urine mutagenicity, 1- OHP, and aromatic DNA adducts among coke oven workers. The study population is heavily exposed to PAHs (Kuljukka et al., 1997). Here we present studies on urinary mutagenicity and the influence of the CYP1A1, meh ( and predicted meh activity), GSTM1, GSTP1, and GSTT1 genotypes on markers of internal exposure. Materials and methods Subjects Forty-nine coke oven workers at an oil shale plant and 10 countryside controls participated in this study during the year Personal PAH exposure was monitored at the plant in wintertime and in fall, while the controls were monitored once in the winter. The group of exposed workers was ethnically heterogeneous, originating from different parts of the former Soviet Union, whereas the control population was mainly native Estonians. According to anamnesis in questionnaires, eight individuals took part in the study during both samplings at the coke ovens. For this reason, the analyses of genotype- associated differences were analyzed separately for the two sampling occasions. There were 12 smoking and 10 nonsmoking coke oven workers in the first sampling, and 16 smokers and 11 nonsmokers in the second sampling, respectively. Among controls, three were smokers and seven nonsmokers. Each group included both men and women. The sampling strategy and study group has been described in detail earlier ( Kuljukka et al., 1997; Kuljukka et al., 1998). 82 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1)

3 Effect of genotypes on PAH exposure-related biomarkers Kuljukka- Rabb et al. Exposure Assessment Among coke oven workers and controls personal air monitoring was conducted during one shift per person by using a portable sampling device, equipped with a Tefloncoated filter unit to catch the particulate phase of PAHs and a XAD-2 adsorbent tube to collect the compounds in gas phase. A semiquantitative assessment of exposure through skin was performed with skin wiping from the wrist before and after the shift. Pyrene and B[a]P were measured from these samples by high- performance liquid chromatography with fluorescence detection ( Kuljukka et al., 1997). Analysis of Mutagenic Activity in Urine Urinary mutagenicity in Salmonella typhimurium TA98 and YG1024 strains was tested from samples collected during the day of occupational monitoring. An aliquot was separated for the measurement of creatinine concentration before deep freezing until analysis. Urine samples ( ml) were adjusted to ph 5 6 and treated with - glucuronidase/ sulphatase ( Sigma G0751, type H1, Sigma Chemical, St. Louis, MO) by adding 55 g of the enzyme preparation per milliliter urine ( 17.2 U - glucuronidase/ ml). Incubation was done by shaking at + 378C overnight. After that the samples were filtered and loaded onto equilibrated ( two bed volumes each of methanol and water) C 18 Megabondelut columns ( Varian, Harbor City, CA). After adsorption, the columns were washed with 10 ml of 5% methanol water. Desorption was performed with 20 ml methanol and the final sample volumes were adjusted to 20 ml. From each methanol extract 500 l was moved into sample vials and stored in a freezer. The rest of the methanol volume was evaporated to dryness and redissolved in 500 l of dimethyl sulfoxide (DMSO). For the Ames test the sample volumes were first adjusted to a volume of 1 ml using DMSO and were further diluted into five doses ( twofold dilution steps). The tested doses, corresponding to the original amounts of urine, ranged between 0.63 and 20 ml per plate. Samples were tested only with S9 mix, using two parallel plates per dose. The S. typhimurium strains TA98 and YG1024 were used. Mutagenic activities were calculated using the linear regression term B (when y=bx+a) indicating the slope of the mutagenic response. The best possible linear fit was always used. Analysis of Urinary 1-OHP 1- OHP concentration was analyzed from spot urine samples voided after the shift according to the fluorometric method of Jongeneelen and coworkers with in- house modifications ( Jongeneelen et al., 1987; Kuljukka et al., 1997). DNA Adduct Measurement Aromatic adducts were analyzed from lymphocyte DNA by the nuclease-p1-enhanced 32 P-postlabeling procedure, slightly modified from the original method of Randerath and coworkers ( Gupta et al., 1982; Reddy and Randerrath, 1986). Shortly, 4 g of DNA was digested with 0.2 U of micrococcal nuclease and 2 g of spleen phosphodiesterase. An aliquot of DNA digest was analyzed with HPLC/UV detection to confirm the DNA concentration and completion of digestion. For the enrichment reaction 3 g of nuclease P1 was added over 40 min. The samples were postlabeled with 40 Ci of [ - 32 P]ATP (ICN, Irvine, CA, 7000 Ci/ mmol) by 4.8 U T 4 polynucleotide kinase and chromatographed with commercial TLC plates. Details have been published earlier ( Kuljukka et al., 1998). Identification of Genotypes Cytochrome P4501A1 Two closely linked polymorphisms in the CYP1A1 gene were analyzed by PCR, essentially as described by Hayashi et al. (1991). The MspI polymorphism in the 3 0 end of CYP1A1 is due to a base substitution (T- C), 264 base pairs downstream the poly(a) signal. The absence or presence of an MspI restriction site identifies the m1 and the m2 alleles. The isoleucine/ valine ( Ile/ Val) polymorphism in the seventh exon, which arises from an A- G base change and results in replacement of Ile by Val at residue 462 in the heme-binding region of the enzyme, was analyzed by allele- specific PCR. Detailed information of the PCR assays is given elsewhere ( Alexandrie et al., 1994). Glutathione S- Transferases GSTM1 genotyping was performed essentially according to methods described by Brockmoller et al. ( 1992). GSTT1 genotypes were determined by PCR as described previously (Pemble et al., 1994; Warholm et al., 1995). -Actin (primers from Stratagene, La Jolla, CA) was used as positive control to verify the presence of amplified DNA. The PCR method used for the detection of individuals lacking the GSTM1 and GSTT1 genes ( the null genotypes) do not differentiate between the heterozygous and homozygous carriers of the functional gene. Analyses of the GSTP1 polymorphisms resulting in an Ile to Val substitution at residue 104 in exon 5 and an alanine (Ala) to Val substitution at residue 113 in exon 6 were performed as previously described ( Carstensen et al., 1999a). Microsomal Epoxide Hydrolase ( meh) The tyrosine/ histidine (Tyr/His) polymorphism in exon 3 (amino acid 113) of meh was analyzed by allele- specific PCR as previously described (Carstensen et al., 1999a). The His/ arginine (Arg) polymorphism in exon 4 (amino acid 139) was determined by PCR/ restriction fragment length polymorphism ( RFLP) as described by Hassett and coworkers Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1) 83

4 Kuljukka- Rabb et al. Effect of genotypes on PAH exposure-related biomarkers Table 1. Urine 1 - OHP concentration ( mol / mol creatinine) with respect to metabolic polymorphisms. Gene Genotype Controls PAH- exposed sampling I PAH- exposed sampling II n Median Range P value n Median Range P value n Median Range P value All All CYP1A1, MspI T/ T T6235C T/ C, C/ C CYP1A1, exon 7 Ile/ Ile Ile462Val Ile/ Val, Val / Val meh, exon 3 Tyr/ Tyr Tyr113His Tyr/ His, His/ His meh, exon 4 His/ His His139Arg His/ Arg, Arg/ Arg EH phenotype high or intermediate low GSTM1 + / +, + / / GSTP1, exon 5 Ile/ Ile Ile104Val Ile/ Val, Val / Val GSTP1, exon 6 Ala/ Ala Ala113Val Ala/ Val, Val / Val GSTT1 + / +, + / / (Hassett et al., 1994). In all PCR assays negative (containing water instead of DNA) and positive (containing DNA from individuals with known genotype) controls were included. Based on current knowledge of the in vitro functional expression of the variant alleles at residue 113 (exon 3, His 113 slow allele) and at residue 139 (exon 4, Arg 139 rapid allele) three predicted meh enzymatic activity levels were assigned ( Hassett et al., 1994). Low activity individuals were homozygous for His 113 and homozygous for His 139, homozygous for His 113 and heterozygous for His 139, or heterozygous for His 113 and homozygous for His 139. High activity individuals were homozygous for Tyr 113 and homozygous for Arg 139, homozygous for Tyr 113 and heterozygous for Arg 139, or heterozygous for Tyr 113 and homozygous for Arg 139. The remaining genotypes were classified as intermediate meh activity (Benhamou et al., 1998). Statistical Analysis Nonparametric and parametric statistical methods ( Spearman correlation, Kruskal Wallis test, Student s t- test, multiple regression analysis) were performed using SAS Table 2. Aromatic DNA adducts in lymphocytes ( adducts per 10 8 normal nucleotides) with respect to metabolic polymorphisms. Gene Genotype Controls PAH- exposed sampling I PAH- exposed sampling II n Median Range P value n Median Range P value n Median Range P value All All CYP1A1, MspI T/ T T6235C T/ C, C/ C CYP1A1, exon 7 Ile/ Ile Ile462Val Ile/ Val, Val / Val meh, exon 3 Tyr/ Tyr Tyr113His Tyr/ His, His/ His , meh, exon 4 His/ His His139Arg His/ Arg, Arg/ Arg EH phenotype high or intermediate low GSTM1 + / +,+ / / GSTP1, exon 5 Ile/ Ile Ile104Val Ile/ Val, Val / Val GSTP1, exon 6 Ala / Ala Ala113Val Ala / Val, Val/ Val GSTT1 + / +,+ / / Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1)

5 Effect of genotypes on PAH exposure-related biomarkers Kuljukka- Rabb et al. Table 3. Mutagenic activity ( revertants /100 ml urine) in urine of coke oven workers (sampling I and II ) and controls. S. typhimurium Controls PAH- exposed sampling I PAH- exposed sampling II n Median Range P value n Median Range P value n a Median Range P value TA98 All Smokers Nonsmokers YG1024 All Smokers Nonsmokers a Unknown smoking habit of one person. software ( SAS, Cary, NC). The validity of statistical methods was assessed by examining the residuals, normality, and outlying values. A nonparametric method was used when necessary. Group differences in urine mutagenicity between smoking and nonsmoking individuals were tested with the nonparametric Kruskal Wallis test. Correlation between external exposure parameters and urine mutagenicity, 1-OHP or DNA adducts was calculated by Spearman correlation. Group difference in 1- OHP levels between genotype subgroups of meh ( exon 3) was tested with Student s t- test. Genotype dependence of 1- hydroxypyrene and aromatic adduct levels in lymphocytes was examined by Kruskal Wallis test. Multiple regression analysis was carried out with urine mutagenicity, 1- OHP and DNA adducts as the dependent variable, each at a time. The independent variables were particulate pyrene and B[ a]p concentrations. In addition, the independent variables for genotypes were defined as variant=1 and wild-type =0. CYP1A1 (MspI); B and C=1 or A=0, CYP1A1 (exon 7); Ile/ Val and Val/ Val= 1 or Ile/ Ile= 0, EH (exon 3); Tyr/His and His/His=1 or Tyr/Tyr=0, EH (exon 4); His/Arg and Arg/Arg=1 or His/His=0, GSTM1; / =1 or +/ and +/+=0, GSTP1(exon 5); Ile/Val and Val/Val=1 or Ile/ Ile=0, GSTP1(exon 6); Ala/Val and Val/Val=1 or Ala/ Ala=0, GSTT!; / =1 and / + or + / + =0. The role of tobacco smoking was analyzed by adding a smoker/ nonsmoker parameter to the multiple regression model, and unless it was significant the results of the model without tobacco parameter were used. The effect of smoking was checked further, by excluding any other statistically nonsignificant independent variable from the model due to the limited number of observations in the study. Results Background Data on Exposure Assessment The coke oven workers studied were on average heavily exposed to PAHs (Kuljukka et al., 1998). The airborne pyrene concentration ranged from 0.01 to 69.6 g/m 3 (median 0.2 and mean 8.1), and the B[a]P concentration from 0.02 to 39.6 g/m 3 (median 0.4 and mean 5.7) in personal air samples. Exposure was considerably higher during wintertime ( sampling I). Urinary 1- OHP ranged from 0.2 to 69.5 mol/mol creatinine (median 3.4 and mean 8.7) in the first sampling set, and from 0.28 to 21.6 mol/mol creatinine (median 1.5 and mean 3.7) during the second sampling. However, the level of aromatic DNA adducts in lymphocytes, median 1.5 adducts/10 8 nucleotides and mean 1.6 adducts/10 8 nucleotides, did not statistically differ between the controls and exposed ( Kuljukka et al., 1998). Levels of urinary 1- OHP and aromatic DNA adducts in lymphocytes with respect to metabolic polymorhisms are presented in Tables 1 and 2. Those associations of urine mutagenicity, 1- OHP, and DNA adducts with genotypes having at least moderate degree of Table 4. Spearman correlation of urine mutagenicity in TA98 and YG1024 strains to airborne particulate pyrene and B[ a]p concentrations, urine 1 -OHP and aromatic adducts measured in lymphocytes by the 32 P-postlabeling method.these values are calculated from the results of exposed workers and controls. Rev /100 ml urine Pyrene B[a]P 1 - OHP DNA adducts Mutagenicity in TA98 r = 0.38 r = 0.35 r =0.47 r =0.64 P= P=0.025 P=0.002 P= n =41 n =41 n =42 n =33 Mutagenicity in YG1024 r = 0.37 r = 0.36 r =0.51 r =0.56 P= P=0.02 P= P= n =34 n =42 n =43 n =34 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1) 85

6 Kuljukka- Rabb et al. Effect of genotypes on PAH exposure-related biomarkers Figure 1. Dependence of urine mutagenicity in Salmonella typhimurium TA98 and YG1024 strains ( with metabolic activation) on microsomal epoxide hydrolase ( meh) genotype (His/Arg 139 ) in exon 4. Smoking and nonsmoking coke oven workers were monitored during the first sampling period. in exon 4, the P value being The genotype of meh in exon 3 did not significantly effect on urinary mutagenicity tested with the TA98 strain. There was a corresponding trend in YG1024 mutagenicity, although statistically nonsignificant, towards lower urinary mutagenicity in persons having the His/His genotype in exon 4 of meh gene ( P=0.098). The variation of mutagenic activity in urine depending on meh exon 4 is illustrated in Figure 1. Figure 2A and B shows the urinary mutagenicity dependence on B[a]P concentration in air and the predicted phenotypic activity of meh. Those coke oven workers (n=16) who were classified as phenotypically highly or intermediately active had significantly higher level of mutagenicity in urine than those having a genotype combination leading to presumably low enzyme activity ( P= for TA98 and for YG1024, respectively). Smoking habit did not explain the differences in urinary mutagenicity observed between meh phenotype or genotype subgroups. statistical significance, based on regression analysis, are chosen for further analyses. Urinary Mutagenicity Variation of urinary mutagenicity in the groups of coke oven workers and controls is shown in Table 3. In Kruskal Wallis test the difference in urinary mutagenicity between the exposed and controls was not significant, the P value being for TA98, and P= for YG1024, respectively. Within the group of exposed workers urine was significantly more mutagenic ( P= for TA98, and P= for YG1024) among smoking individuals than among nonsmokers. Multiple regression analysis revealed that among exposed workers ( n =38) only smoking had a statistically significant effect on urinary mutagenicity. Correlation of urinary mutagenicity to parameters of external exposure and biomarkers, among the exposed workers and controls together ( n =59), is presented in Table 4. Due to various missing values, the number of observations is variable. As shown in Table 4, the mutagenic activity in urine correlated weakly to pyrene and B[ a]p concentrations at breathing zone, and there was a more pronounced correlation observed between urinary mutagenicity and urinary 1- OHP concentration, and DNA adduct levels in lymphocytes. Influence of Genotypes on Urinary Mutagenicity Multiple regression analysis on data from the 17 exposed workers from sampling I showed that urinary mutagenicity levels, tested with the strain TA98, were somewhat lower among coke oven workers with the meh His/ His genotype Figure 2. Urinary mutagenicity depending on predicted microsomal epoxide meh activity and airborne B[ a]p among coke oven workers monitored during the first sampling period. Mutagenicity with TA98 in ( A) and in YG1024 in ( B), respectively. 86 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1)

7 Effect of genotypes on PAH exposure-related biomarkers Kuljukka- Rabb et al. Influence of Genotypes on Urinary 1- OHP In the group of coke oven workers monitored during the second sampling period a significant effect of meh genotype was observed on 1- OHP level, expressed in micromoles per mole creatinine. According to Student s t- test, persons ( n = 23) having the Tyr/ His or His/ His genotype in exon 3 had significantly ( P=0.01) higher urinary 1- OHP levels than those having the Tyr/ Tyr genotype. This difference is illustrated in Figure 3. According to multiple regression analysis, airborne pyrene ( P= 0.007), tobacco smoking ( P= 0.054) and EH3 genotype ( P= 0.076) all affected 1- OHP concentration in urine. Influence of Genotypes on DNA Adducts in Lymphocytes We saw some significant differences in adduct levels depending on metabolic genotype in the group of coke oven workers from the first sampling ( n= 16). The polymorphism in meh (exon 3) and the B[a]P concentration in breathing zone air affected interactively ( P=0.01) the level of aromatic DNA adducts in lymphocytes. At lower exposures, those who had the Tyr to His mutation in meh gene in exon 3 had lower adduct levels than others. Another observation was that among persons with the Tyr/ Tyr genotype the measured adduct levels were independent of airborne B[ a]p concentration, whereas among those with the Tyr/ His or His/ His meh genotype, lower B[ a]p exposure was accompanied by lower adduct levels, respectively. This rather complicated relationship is visualized in Figure 4. Also the meh phenotype and airborne B[ a]p were interactive ( P= 0.03) in determining the aromatic adduct level in lymphocytes. With lower ( <20 g/m 3 ) B[a]P levels the persons with low level of meh activity had Figure 3. Urinary 1- OHP concentration depending on meh genotype (Tyr /His 113 ) in exon 3 among coke oven workers monitored during the second sampling. Figure 4. Level of aromatic DNA adducts in lymphocytes among coke oven workers monitored during the first sampling, dependence on meh genotype (Tyr/His 113 ) in exon 3 and airborne B[a]P. lower adduct levels than those with the intermediate or high activity. On the contrary, the meh His to Arg (exon 4) polymorphism and airborne B[ a]p were not interactive. Instead, there was a significant group difference ( P=0.0088) in adduct levels: Persons with the His/ His genotype in exon 4 had somewhat lower level of adducts. Discussion The coke oven workers of this study were highly exposed to PAHs. The mean concentration of B[ a]p, also measured from personal air samples, in a modern Finnish cokery is approximately five times lower than the average exposure observed at this plant. The use of protective equipment and clothing and benefits of technical measures in processes are the most obvious reasons for low and even decreasing exposure levels in the course of time (Pyy et al., 1997). Measurement of biomarkers of exposure is essential when assessing human exposure to mixtures of chemicals ( e.g., PAHs), not least because it reflects the integrated dose from multiple absorption routes. The contribution of dermal absorption for pyrene in coke oven work has been evaluated to be as much as 75% (VanRooij et al., 1993). We have also reported earlier of skin contamination during coke oven work ( Kuljukka et al., 1997). Among this study population urinary 1- OHP correlated very well with the pyrene concentration in breathing zone air ( Kuljukka et al., 1997). The two sampling occasions turned out to be different, which was evident both in exposure levels and in the observed effects of metabolic genotypes on biomarker levels. A longer period of monitoring would provide a more reliable picture of the average occupational exposure. Secondly, a larger dataset would be needed to evaluate why the effect of genotypes was more evident in the first sampling. Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1) 87

8 Kuljukka- Rabb et al. Effect of genotypes on PAH exposure-related biomarkers Urinary mutagenicity has been frequently used as an indicator of exposure to environmental genotoxins, but smoking is known to be a strong confounding factor ( Fereira et al., 1994). This disadvantage was obvious also in our study, as tobacco but not occupational B[a]P exposure was a significant determinant of urinary mutagenicity. The bacterial strains used in our study differ in their sensitivity towards genotoxic compounds having a different structure. As an O- acetyltransferase overproducing S. typhimurium strain, YG1024 detects the mutagenicity of aromatic nitro, amino and hydroxylamino compounds more efficiently than its sister strain TA98 ( Watanabe et al., 1990). Furthermore, YG1024 has been found to be more sensitive in detecting mutagenicity in human urine caused by cigarette smoke than the conventionally used strain TA98 in the presence of S9 mix. The results suggest that part of the mutagenicity detected in urine from smokers is caused by the presence of amino aromatic compounds ( Einistö et al., 1990). The difference in urinary mutagenicity between all exposed and controls was not significant (P=0.0825), possibly due to low number of controls, and smoking coke oven workers had significantly ( P=0.0002) higher mutagenic activities in urine than nonsmokers. Corresponding results have been reported also by others recently ( Clonfero et al., 1995; Mielzynska et al., 1997). Gabbani et al. (1996) also found that despite high level of PAH exposure at work, only one sample out of 14 mutagenic urine samples belonged to a nonsmoker. Despite high external exposure to PAHs, which was also reflected as high 1- OHP concentrations in coke oven workers, aromatic DNA adducts did not differ significantly between the exposed and controls, when smokers and nonsmokers were considered together ( Kuljukka et al., 1998). Identification of DNA adducts by 32 P-postlabeling is one of the most relevant markers of biologically effective dose, but measurement of 1-OHP holds its position as a sound and sensitive biomarker of pyrene exposure, and as such, an indicator of PAH exposure from multiple absorption routes. Our results on adduct levels are in accordance with other studies, which also show nonlinearity in adduct formation in high exposure situations (Binkova et al., 1998; Carstensen et al., 1999b; Godschalk et al., 1998; Lewtas et al., 1997; Van Schooten et al., 1997). Many reasons may lie behind the absence of dose response in DNA adduct formation in high level PAH exposure. For example, saturation of activation, and on the other hand, induction of detoxification mechanisms by varying components of the exposure mixture. During the past decade genetically determined host factors have become likely candidates as an important reason for the observed variance in human response to drugs and other chemicals. A number of polymorphic alleles have been identified in many genes mediating the metabolism of parent chemicals both in phase I and phase II enzymes. Molecular epidemiology studies mainly focus on the association of genetic polymorphisms and cancer risk. A meta- analysis of relevant epidemiological studies ( until 1995) revealed that GSTM1 polymorphism causes an overall odds ratio for lung cancer of 1.32 ( ) in Caucasians. In the case of CYP1A1 ( combined MspI variant genotype) and lung cancer, the overall odds ratio is 1.1 ( ), with estimates of 1.0 ( ) in Caucasians and 1.4 ( ) in Japanese studies ( d Errico et al., 1996). The genes of relevance in PAH exposure-related DNA damage include CYP1A1, meh, and GSTs, although the reactions catalyzed by their enzyme products constitute only a few examples of several possible pathways. After oxidation by CYP1A1, the 7,8- oxide of B[ a]p is converted by mehs to its trans-7,8-diol, which is further metabolized to the ultimate carcinogenic diol epoxide. It has been postulated that the limited sensitivity of genotyping to find susceptible individuals might be due to incomplete knowledge of relevant mutations in the complex alternatives in metabolism ( d Errico et al., 1996). In addition, the frequency of mutations varies considerable depending on the ethnicity of the population. We did not see any effect of GSTM1 mutation on biomarker levels in this study. The high exposure to PAHs may have masked the effect of genetic polymorphism, or could not be seen due to low number of observations. Others have also reported negative findings concerning GSTM1 polymorphism and effect on PAH exposure markers lately (Binkova et al., 1998; Hemminki et al., 1997). On the other hand, among potroom workers urinary 1- OHP was associated with CYP1A1 and GSTM1 polymorphisms ( Alexandrie et al., 2000). GSTM1- positive coke oven workers have even been reported to have higher 1-OHP levels than those lacking the gene (Ovrebo et al., 1998). Results concerning the role of meh Tyr to His (residue 113, exon 3) and His to Arg ( residue 139, exon 4) variations in the level of biomarkers are not easy to interpret. These mutations are believed to alter enzymatic function by modifying protein stability ( Hassett et al., 1994). Based on in vitro studies the phenotypic effect of His variation in exon 3 is estimated to cause a decrease in activity of the enzyme, and an Arg variation in exon 4 causes an increase in activity, respectively (Hassett et al., 1994). The net effect of amino acid exchanges in both exons should result in intermediate, somewhat lowered activity. We used this classification when dividing coke oven workers to low or high and intermediate metabolizer groups. Our results revealed that neither of the meh mutations had a statistically significant effect on urinary mutagenicity. However, coke oven workers with the predicted high and intermediate meh activity had significantly higher level of urinary mutagenicity, but no effect was seen in urinary 1- OHP concentrations. Instead, among coke oven workers monitored during the 88 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1)

9 Effect of genotypes on PAH exposure-related biomarkers Kuljukka- Rabb et al. second sampling, those having a Tyr/ His or His/ His genotype in exon 3, had significantly (P=0.01) higher 1- OHP concentration in urine, possibly implying more efficient detoxification reactions when the activity of meh is lowered. Concerning the association of meh genotypes with DNA adduct formation the picture is even more complicated. His 113 in exon 3 of meh and the predicted meh activity interactively affected DNA adducts with exposure to B[a]P. Expectedly, those who had a mutation in exon 3, or low phenotypic activity, had somewhat lower adduct levels compared to others, as the decreased enzyme activity might provide less substrates for further metabolism to diol epoxide formation. Likewise, the effect of Arg allele in exon 4 was in accordance with the assumption based on in vitro enzyme activity studies. On the other hand, low net activity of meh resulted in lower adduct levels, if B[a]P exposure was less than 20 g/m 3. It is not wise to make strong conclusions, but at least these results confirm that the role of mehs seems important in PAH exposure. The ultimate faith of meh reaction products to proximate carcinogens, or to more hydrophilic excretion products, depends on complicated stereochemistry and yet unrevealed interplay of meh alleles. The net effect of meh activity is also affected by the composition of PAH mixture. There are indications of varying faiths of the components of PAH mixture due to different enzyme affinities. Obviously, the exposure level itself is important. Benhamou et al. ( 1998) reported recently a significant effect of meh activity on lung cancer risk, but our observation on DNA adduct formation and meh polymorphism is still far from cancer risk assessment. In a biomonitoring study it is possible to measure or control only a limited number of factors affecting total PAH exposure and internal dose. As one of the most important confounding factors, the role of tobacco smoking was studied in multiple regression analysis. It appeared that the differences seen in urinary mutagenicity or DNA adduct levels between meh genotype or phenotype subgroups could not be explained by tobacco smoking. In addition, B[ a]p concentration in air was statistically nonsignificant in determining urinary mutagenicity, leaving the main determinants of urinary mutagenicity unrevealed in this study. On the contrary, the measured B[ a]p concentration and meh together affected DNA adduct levels in some subpopulations in this study. In conclusion, this comprehensive biomonitoring study of coke oven workers has confirmed the usefulness of combining external exposure assessment to measurement of urinary markers of exposure and DNA adducts. Individual genetic susceptibility, assessed by identification of relevant genotypes involved in metabolism, is a valuable tool in interpreting interindividual differences in response to exposure. We identified only a few of the genetic polymorphisms affecting PAH metabolism, which certainly impairs the sensitivity of genotyping as a susceptibility marker. Acknowledgments The authors thank the volunteers of the RAS Kiviter plant who participated in this study and other employees whose contribution was essential in this study. The personnel of Institute of Experimental and Clinical Medicine, in Tallinn are thanked for practical arrangements on site. Diari Ghafouri is thanked for skillful performance of the genotype analyses. All the employees of Finnish Institute of Occupational Health are thanked for their contribution during this study. The research was supported by the Academy of Finland, and the CEC project ERBCIPA CT References Alexandrie A.K., et al. Genetic susceptibility to lung cancer with special emphasis on CYP1A1 and GSTM1: a study on host factors in relation to age at onset, gender and histological cancer types. Carcinogenesis 1994: 15: Alexandrie A.K., et al. CYP1A1 and GSTM1 polymorphisms affect urinary 1 - hydroxypyrene levels after PAH exposure. Carcinogenesis 2000: 21: Bartsch H. et al. Metabolic polymorphism affecting DNA binding and excretion of carcinogens in humans. Pharmacogenetics 1995: 5: S Bartsch H., and Hietanen E. The role of individual susceptibility in cancer burden related to environmental exposure. Environ Health Perspect 1996: 104: ( Suppl 3 ): Benhamou S., et al. Association between lung cancer and microsomal epoxide hydrolase genotypes. Cancer Res 1998: 58: Bieniek G. Aromatic and polycyclic hydrocarbons in air and their urinary metabolites in coke plant workers. Am J Ind Med 1998: 34: Binkova B., et al. Biomarker studies in northern Bohemia. Environ Health Perspect 1996: 104 ( Suppl 3 ): Binkova B., et al. Coke oven workers study: the effect of exposure and GSTM1 and NAT2 genotypes on DNA adduct levels in white blood cells and lymphocytes as determined by 32 P-postlabelling. Mutat Res 1998: 416: Brockmoller J., et al. Correlation between trans - stilbene oxide - glutathione conjugation activity and the deletion mutation in the glutathione S - transferase class mu gene detected by polymerase chain reaction. Biochem Pharmacol 1992: 43: Butkiewicz D., et al. Modulation of DNA adduct levels in human mononuclear white blood cells and granulocytes by CYP1A1 CYP2D6 and GSTM1 genetic polymorphisms. Mutat Res 1998: 415: Carstensen U., et al. Influence of genetic polymorphisms of biotransformation enzymes on gene mutations, strand breaks of deoxyribonucleic acid, and micronuclei in mononuclear blood cells and urinary 8 - hydroxydeoxyguanosine in potroom workers exposed to polyaromatic hydrocarbons. Scand J Work Environ Health 1999: 25: Carstensen U., et al. Genotoxic exposures of potroom workers. Scand J Work Environ Health 1999: 25: Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1) 89

10 Kuljukka- Rabb et al. Effect of genotypes on PAH exposure-related biomarkers Clonfero E., et al. Biological monitoring of human exposure to coal tar. Urinary mutagenicity assays and analytical determination of polycyclic aromatic hydrocarbon metabolites in urine. In: Vainio H., Sorsa M., McMichael A.J. (Eds.). Complex Mixtures and Cancer Risk IARC Scientific Publication, no. 104, Lyon 1990, pp Clonfero E., et al. Urinary excretion of mutagens in coke oven workers. Carcinogenesis 1995: 16: Daly A.K., et al. Metabolic polymorphisms. Pharmacol Ther 1993: 57: d Errico A., et al. Genetic metabolic polymorphisms and the risk of cancer: a review of the literature. Biomarkers 1996: 1: Eaton D.L., and Bammler T.K. Concise review of the glutathione S- transferases and their significance in toxicology. Toxicol Sci 1999: 49: Einistö P., et al. Sensitivity of Salmonella typhimurium YG1024 to urine mutagenicity caused by cigarette smoking. Mutat Res 1990: 245: Fereira Jr. M., et al. Determinants of urinary thioethers, D - glucaric acid and mutagenicity after exposure to polycyclic aromatic hydrocarbons assessed by air monitoring and measurement of 1 - hydroxypyrene in urine: a cross- sectional study in workers of coke and graphiteelectrode- producing plants. Int Arch Occup Environ Health 1994: 65: Gabbani G., et al. GSTM1 and NAT2 genotypes and urinary mutagens in coke oven workers. Carcinogenesis 1996: 17: Godschalk R.W., et al. Aromatic DNA adducts in human white blood cells and skin after dermal application of coal tar. Cancer Epidemiol Biomarkers Prev 1998: 7: Grinberg - Funes R.A., et al. Polycyclic aromatic hydrocarbon DNA adducts in smokers and their relationship to micronutrient levels and the glutathione- S- transferase M1 genotype. Carcinogenesis 1994: 15: Gupta R.C., et al. 32 P-postlabeling analysis of non-radioactive aromatic carcinogen DNA adducts. Carcinogenesis 1982: 3: Hassett C., et al. Human microsomal epoxide hydrolase: genetic polymorphism and functional expression in vitro of amino acid variants. Hum Mol Genet 1994: 3: Hayashi S.I., et al. Genetic linkage of lung cancer- associated MspI polymorphisms with amino acid replacement in the heme binding region of the human cytochrome P4501A1 gene. J Biochem 1991: 110: Heckbert S.R., et al. Glutathione S- transferase and epoxide hydrolase activity in human leukocytes in relation to risk of lung cancer and other smoking- related cancers. J Natl Cancer Inst 1992: 84: Hemminki K., et al. DNA adducts in humans environmentally exposed to aromatic compounds in an industrial area of Poland. Carcinogenesis 1990: 11: Hemminki K., et al. Aromatic DNA adducts in foundry workers in relation to exposure, life style and CYP1A1 and glutathione transferase M1 genotype. Carcinogenesis 1997: 18: Hirvonen A. Genetic factors in individual responses to environmental exposures. J Occup Environ Med 1995: 37 ( 1 ): Hirvonen A., et al. Modulation of urinary mutagenicity by genetically determined carcinogen metabolism in smokers. Carcinogenesis 1994: 15: Hu X., et al. Catalytic efficiencies of allelic variants of human glutathione S - transferase P1-1 toward carcinogenic anti - diol epoxides of benzo[ c]phenanthrene and benzo[ g ]chrysene. Cancer Res 1998: 58: International Agency for Research on Cancer ( IARC ). Polynuclear aromatic compounds: Part 1. Chemical, environmental and experimental data. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans Vol. 32, pp IARC, Lyon, International Agency for Research on Cancer ( IARC ). Polynuclear aromatic compounds: Part 3. Industrial exposures in aluminum production, coal gasification, coke production, and iron and steel founding. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans Vol. 34, pp IARC, Lyon. Jongeneelen F.J., et al. Determination of hydroxylated metabolites of polycyclic aromatic hydrocarbons in urine. J Chromatogr 1987: 413: Kalina I., et al. Cytogenetic monitoring in coke oven workers. Mutat Res 1998: 417: Kawajiri K., et al. Identification of genetically high risk individuals to cancer by DNA polymorphisms of the cytochrome P450Ia1 gene. FEBS Lett 1990: 263: Ketterer B. Protective role of glutathione and glutathione transferases in mutagenesis and carcinogenesis. Mutat Res 1988: 202: Ketterer B., et al. The human glutathione S- transferase supergene family, its polymorphism, and its effects on susceptibility to lung cancer. Environ Health Perspect 1992: 98: Kuljukka T., et al. Assessment of occupational exposure to PAHs in Estonian coke oven plant- correlation of total external exposure to internal dose measured as 1- hydroxypyrene concentration. Biomarkers 1997: 2: Kuljukka T., et al. Low response in white blood cell DNA adducts among workers in a highly polluted cokery environment. J Occup Environ Med 1998: 40: Lewtas J., et al. Air pollution exposure DNA adduct dosimetry in humans and rodents: evidence for non-linearity at high doses. Mutat Res 1997: 378: Miel_zyńska D., et al. Exposure of coke - oven workers to polycyclic aromatic hydrocarbons based on biological monitoring results. Am Ind Hyg Assoc J 1997: 58: Miller E.C., and Miller J.A. Mechanisms of chemical carcinogenesis. Cancer 1981: 47 ( 5 ): Mooney L.A., et al. Contribution of genetic and nutritional factors to DNA damage in heavy smokers. Carcinogenesis 1997: 18: Motykiewicz G., et al. A molecular epidemiology study in women from Upper Silesia, Poland. Toxicol Lett 1998: 96 97: Nakachi K., et al. Polymorphisms of the CYP1A1 and glutathione S- transferase genes associated with susceptibility to lung cancer in relation to cigarette dose in a Japanese population. Cancer Res 1993: 53: Nielsen P.S., et al. Environmental air pollution and DNA adducts in Copenhagen bus drivers effect of GSTM1 and NAT2 genotypes on adduct levels. Carcinogenesis 1996: 17: Ovrebo S., et al. Glutathione S - transferase M1 and P1 genotypes and urinary excretion of 1 - hydroxypyrene in coke oven workers. Sci Total Environ 1998: 220: Pemble S., et al. Human glutathione S - transferase theta ( GSTT1): cdna cloning and the characterization of a genetic polymorphism. Biochem J 1994: 300: Pyy L., et al. Ambient and biological monitoring of exposure to polycyclic aromatic hydrocarbons at a coking plant. Sci Total Environ 1997: 199: Rannug A., et al. Genetic polymorphism of cytochromes P450 1A1, 2D6 and 2E1: regulation and toxicological significance. J Occup Environ Med 1995: 37 ( 1): Reddy M.V., and Randerrath K. Nuclease P1- mediated enhancement of sensitivity of 32 P-postlabeling test for structurally diverse DNA adducts. Carcinogenesis 1986: 7: Rojas M., et al. Anti - benzo[ a ]pyrene diolepoxide DNA adduct levels in peripheral mononuclear cells from coke oven workers and the enhancing effect of smoking. Carcinogenesis 1995: 16: Rojas M., et al. High benzo[a ]pyrene diol-epoxide DNA adduct levels 90 Journal of Exposure Analysis and Environmental Epidemiology (2002) 12(1)