Lung cancer susceptibility in relation to combined polymorphisms of microsomal epoxide hydrolase and glutathione S-transferase P1

Cancer Letters 173 (2001) 155–162 www.elsevier.com/locate/canlet

Lung cancer susceptibility in relation to combined polymorphisms of microsomal epoxide hydrolase and glutathione S-transferase P1 Jordi To-Figueras*, Manuel Gene´, Jesu´s Go´mez-Catala´n, Esther Pique´, Natividad Borrego, Jacint Corbella Toxicology Unit, Hospital Clı´nic, IDIBAPS, Departament de Salut Pu´blica, Universitat de Barcelona, Villarroel 170, 08036 Barcelona, Spain Received 29 March 2001; received in revised form 31 May 2001; accepted 4 June 2001

Abstract Human microsomal epoxide hydrolase (mEH) catalyzes a key step in the biotransformation of benzo[a]pyrene that yields the highly mutagenic (1)-anti-7,8-diol-9,10 epoxide (BPDE). Two polymorphisms have been described in the coding region of the mEH gene (EPHX1) that produce two protein variants: 113Tyr ! 113His (exon 3) and 139His ! 139Arg (exon 4). We performed a case-control study among Northwestern Mediterranean Caucasians to investigate a possible association between these EPHX1 variants and lung cancer risk. Both EPHX1 polymorphisms were analyzed in a group of lung cancer patients (n ¼ 176) and in a control group of healthy smokers (n ¼ 187). The results showed a significantly decreased risk for the rare homozygous 113His/113His (adjusted odds ratio (OR): 0.44, 95% confidence interval (CI): 0.27–0.71) and 139Arg/139Arg (adjusted OR: 0.55, 95% CI: 0.33–0.91) compared with the major wild-types 113Tyr/113Tyr and 139His/139His, respectively, as the references. Thereafter, we analyzed the EPHX1 variants in combination with three glutathione S-transferase polymorphic genes (GSTM1, GSTT1, and GSTP1) and we found a significant overepresentation of cancer patients with a combination of exon 3 113Tyr/113Tyr EPHX1 and exon 5 105Ile/105Ile GSTP1 (adjusted OR: 2.34, 95% CI: 1.21–4.52). The polymorphic site within the exon 5 of GSTP1 results in a Ile ! Val substitution, and the isoleucine GSTpi isoform has been found in vitro to be less active than the valine isoform towards the conjugation of BPDE. The 113 Tyr/Tyr EPHX1 encodes for a high-activity mEH. Our results agree with these observations in vitro and suggest that a genetically determined combination of a high-activity mEH and a lowactivity GSTpi may increase lung cancer risk among smokers. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Lung cancer; Susceptibility; Microsomal epoxide hydrolase; Glutathione S-transferase

1. Introduction Several polymorphic genes encoding for enzymes involved in phase I and II reactions are currently investigated as possible modulators of lung cancer risk among smokers [1]. Microsomal epoxide hydrolase (mEH, EC 3.3.2.3) is an enzyme involved in the * Corresponding author. Tel.: 134-93-227-5419; fax: 134-93451-7252. E-mail address: [email protected] (J. To-Figueras).

metabolism of reactive intermediates including some aromatic hydrocarbons considered main procarcinogens of tobacco smoke [2]. The enzyme is a key element in mammals detoxication capacity since it catalyzes the conversion of highly reactive epoxides to inert dihydrodiols that are eliminated by feces and urine as water conjugates. However, in the case of some hydrocarbons such as benzo[a]pyrene, mEH intervenes on a critical sequence of reactions that yields the highly reactive (1)-anti-7,8-diol-9,10 epoxide (BPDE). This metabolite tends to form

0304-3835/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(01)00626-7

156

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

adducts in DNA hotspots and is commonly viewed as benzo[a]pyrene ultimate carcinogen [3]. The mEH enzyme is encoded by a single gene (EPHX1) in 1p11-qter [4,5] and two polymorphic sites have been reported within the coding region. An exon 3 polymorphism results in a Tyr ! His substitution in residue 113 and a exon 4 polymorphic site results in a 139 His ! Arg substitution [6,7]. According to in vitro studies these exon 3 and 4 substitutions may, respectively, decrease and increase mEH activity in tissues [8]. Some studies have addressed a possible association between EPHX1 and the risk of lung cancer among smokers [9–13]. The results have been conflicting, although in some cases the ‘slow’ 113 His/His mEH variant has been suggested to be protective and decrease cancer risk [10,11]. In this study we have investigated the possible role of mEH as a lung cancer risk modifier. We have determined the frequencies of both the 113 Tyr/His and 139 His/Arg variants among a healthy smokers control group and among lung cancer patients. Given the complex metabolic pathways of the tobacco pro-carcinogens and the possible interaction between different loci encoding for activation/deactivation enzymes, we have analyzed EPHX1 alone and in combination with three glutathione S-transferase polymorphic genes involved in detoxification reactions (GSTM1, GSTT1, and GSTP1) and previously genotyped in the same population [14–17]. The objective was to find specific EPHX1/GST combinations that may enhance or decrease lung cancer risk among smokers. Additionally, in order to asses the frequencies of EPHX1 in the general population, an

independently accruited group of blood donors was also genotyped.

2. Materials and methods The main characteristics of the study population are summarised in Table 1. The case group included 175 patients consecutively diagnosed of bronchogenic carcinoma in the ‘Hospital Clı´nic’ of Barcelona (Spain) during 1995–1999. They were classified according to the tumour histology (WHO guidelines) into four main subgroups: squamous cell carcinoma (n ¼ 59), large cell carcinoma (n ¼ 14), small cell carcinoma (n ¼ 56), and adenocarcinoma (n ¼ 46). Patients with a mixed or dubious tumour histology were not included in the present study. They were all Caucasians as judged by their names and place of birth (Spain) and living in Catalonia. There were 160 men and 15 women. A detailed clinical history was obtained in each case, including past and present smoking consumption, possible occupational exposure to chemical carcinogens, dietary and drinking habits and cancer in family members. All but four were active smokers with a mean consumption of 53 ^ 28 pack-years (range 0–170). None reported an occupational exposure to asbestos or other carcinogenic chemicals. The mean age at the time of the diagnose and blood extraction was 60 years (range 32–87). The main control group were healthy volunteers (n ¼ 187), all current smokers, who were recruited with the collaboration of a tobacco detoxification program (Preventive Medicine Department, Univer-

Table 1 Main characteristics of the study population

Gender Ethnicity Age (years) Tobacco

Recruitment

Male/female White, caucasians (%) Mean ^ SD Range Active smokers (%) Pack-years Mean (SD) Range

Cancer patients (n ¼ 175)

Healthy smokers (n ¼ 187)

General population individuals (n ¼ 128)

160/15 100 60 ^ 10 32–87 98

155/32 100 50 ^ 10 28–82 100

100/28 100 43 ^ 11 26–77 Unknown

53 ^ 28 0–170 Hospital patients

47 ^ 27 7–196 Detoxification program

Unknown Unknown Blood donors

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

sity of Barcelona). They were also Northwestern Mediterranean Caucasians living in Catalonia. These volunteers were interviewed regarding their smoking history, dietary habits, alcohol consumption, health status, and possible exposure to carcinogenic chemicals. Those presenting a diagnosed pathology of any kind were excluded. The group finally selected contained 155 men and 32 women; the mean age at the time of inclusion in the study and blood extraction was 50 ^ 10 years (range 28–82); the mean tobacco consumption was 47 ^ 27 pack-years (range 7–196). The EPHX1 variants were also genotyped in a group of healthy blood donors from the general population. They were 128 volunteers of the same origin (Catalonia), 100 men and 28 women with and a mean age at the time of DNA extraction of 43 ^ 10 years (range 26– 77). The smoking status of this volunteers was in most cases unknown. In all cases (patients and controls), a personal consent for inclusion in the study, blood extraction, and DNA genotyping was obtained, and the whole study design was approved by the ethical committees (IDIBAPS, Hospital Clı´nic). 2.1. Analysis of EPHX1 polymorphisms 2.1.1. Tyr/His polymorphism The exon 3 polymorphic site (T ! C) of EPHX1 was analyzed by polymerase chain reaction (PCR) as reported by Persson et al. [13], (described as ‘HYL1*2’ polymorphism) with slight changes. A PCR amplification of specific alleles (PASA) method was used with two primers (ex3Rwt 5 0 -agt ctt gaa gtg agg gtg-3 0 and ex3Rmut 5 0 -agt ctt gaa gtg agg gta-3 0 ) with a single different terminal base corresponding to the polymorphic site and a common primer of the other strand (ex3F 5 0 -ttt gct ctt gtg ctc tgt-3 0 ). The amplification reaction was carried out in a 20ml volume containing 100 ng of genomic DNA as a template, 8 pmol of each oligonucleotide, 200 mM of each deoxynucleotide triphosphate, 10 mM Tris–HCl (pH 9.0 at 258C), 50 mM KCl and 0.1% Triton X-100 and 1.2 (ex3Rwt) or 1.5 (ex3Rmut) mM MgCl2, and about 0.5 units of Taq DNA polymerase. After 1 min at 948C, 30 temperature cycles were used: 1 min at 948C, 1 min at 568C (ex3Rwt) or 528C (ex3Rmut), and 1 min 15 s at 728C. The last elongation step was extended to 10 min. All reactions were performed in a 9700 PE Thermocycler. Negative

157

(without DNA) and positive control samples were included in each amplification series. Each 232 bp PCR product, which was subjected to electrophoresis in 2.5% agarose gel, was obtained only when primer ex3Rwt and/or ex3Rmut was complementary to the template DNAs with respect to the terminal bases (two electrophoresis lanes were used for typing each individual). 2.1.2. His/Arg 139 polymorhism The exon 4 (A ! G) polymorphic site was analyzed by restriction fragment length polymorphism of PCR products (RFLP-PCR). Hot-start PCR reactions were carried out in a 20-ml volume containing 40 ng of genomic DNA template, 12 pmol of each oligonucleotide, 200 mM of each deoxynucleotide triphosphate, 50 mM KCl, 10 mM Tris–HCl (pH 8.3), 0.001% (w/v gelatin) at 1.5 mM MgCl2, 5% DMSO and about 0.5 units of AmpliTaq Gold polymerase. Exon 4 EPHX1 locus amplification was achieved using oligonucleotides described by Smith and Harrison [9], Epo 3 5 0 aca tcc act tca tcc acg t-3 0 and Epo 4 5 0 -atg cct ctg aga agc cat-3 0 . Temperature cycling was identical to that described above for exon 3 EPHX1 polymorphism. 2.2. Statistical analysis A chi-square test was used to compare the frequency distribution of EPHX1 and GST alleles between patients and controls. The influence EPHX1 and combined EPHX1/GST genotypes on lung cancer susceptibility was determined as adjusted odds ratios by logistic regression analysis. Sex, age, and smoking status (quantified as pack-years) were included as variables in the regression model. Age and pack years were used as numerical variables. The possible interaction between EPHX1 and GSTs (GSTM1, GSTP1, and GSTT1) was studied carrying out different logistic regression models: (a) introducing EPHX1 or each GST alone, as variables; (b) introducing combinations of two genes; (c) including two genes and an interaction term between them (second order regression model). The odds ratio for the interaction variable can be interpreted as the additional risk associated to the simultaneous presence of risk genotypes in both loci. The calculations were made using the Statgraphics Plus software.

158

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

brium. The healthy smokers were used as control group for all the analyses to follow. The frequency distribution of both EPHX1 genotypes among the patients and healthy smokers are shown in Table 2. The results showed that the 113Tyr allele tended to be more frequent in patients (frequency 0.75) than in controls (frequency 0.69). The calculation of adjusted odds ratio (OR) showed a statistically significantly decreased risk for the rare homozygous 113His/113His (adjusted OR: 0.44, 95% confidence interval (CI): 0.27–0.71, Table 2) when compared with the major wild-type 113Tyr/113Tyr. This result suggests a protective effect against lung cancer of the 113His variant of mEH. An alternative analysis can be performed grouping in the reference category the heterozygous 113Tyr/His113 with the homozygous 113His/113His and computing the risk associated with the homozygous wild-type. The odds ratio for the wild-type 113Tyr/113Tyr was 1.34 but not statistically significant (Table 3, analysis 1). A similar analysis with the exon 4 His/Arg 139 polymorphism showed a significant decreased risk for the rare homozygous 139Arg/139Arg (adjusted

3. Results In the general population group the frequencies obtained for the Tyr/His 113 genotypes were as follows: 113Tyr/113Tyr 47.6%; 113Tyr/113His 42.2%; 113His/113His 9.4% and the Hardy–Weinberg equilibrium (x2 ¼ 0:23; P ¼ 0:67). Among the control group of healthy smokers the frequencies were: 46.5% (113Tyr/113Tyr); 45.5% (113Tyr/113His) and 8.0% (113His/113His) and also fitted the Hardy–Weinberg equilibrium (x2 ¼ 0:84; P ¼ 0:36). The frequencies obtained for the 113Tyr and 113His alleles among the general population group were 0.684 and 0.316, respectively. Among the healthy smokers the frequencies were 0.693 and 0.307. The allele frequencies showed no significant differences between both groups (x2 ¼ 1:01; P ¼ 0:6; df ¼ 2). The allele frequencies for the exon 4 EPHX1 polymorphism were 0.82 and 0.18 for 139His and 139Arg, respectively, in both the healthy smokers group and the general population group, which is in accordance with several reports among Caucasians [10], and the genotype frequencies fit the Hardy–Weinberg equili-

Table 2 Genotype frequencies of microsomal epoxide hydrolase exon 3 and exon 4 polymorphisms EPHX1 genotypes

Controls a

Cases n

(%)

n

Adjusted OR

(95% CI) b

(%)

Exon 3 Tyr113/Tyr 113 Tyr113/His113 His113/His113

97 70 8

55.4 40.0 4.6

87 85 15

46.5 45.5 8.0

1 0.80 0.44

(0.50–1.29) (0.27–0.71)

Exon 4 His139/His139 His139/Arg139 Arg139/Arg139

119 53 3

68.0 30.3 1.7

126 54 7

67.4 28.9 3.7

1 1.14 0.55

(0.68–1.89) (0.33–0.91)

55 87 33

31.4 49.7 18.9

70 88 29

37.4 47.1 15.5

1 1.28 1.46

(0.77–2.14) (0.87–2.43)

Epoxide hydrolase activity c Low Intermediate High Total a

175

187

Healthy smokers control group. Odds ratios were adjusted including terms for sex, age and smoking (natural logarithm of pack-years). c Classification based on Benhamou et al. [10]: Low: His113/His113 and His139/His139, His113/His113 and His139/Arg139, Tyr113/His113 and His139/His139. Intermediate: Tyr113/Tyr113 and His139/His139, Tyr113/His113 and His139/Arg139, His113/His113 and Arg139/Arg139. High: Tyr113/Tyr113 and His139/Arg139, Tyr113/Tyr113 and Arg139/Arg139, Tyr113/His113 and Arg139/Arg139. b

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

OR: 0.55, 95% CI: 0.33–0.91, Table 2). The alternative analysis grouping the 139Arg containing genotypes did not show an increased risk for the wildtype homozygous (OR: 1.05, 95% CI: 0.47–2.35, Table 3, analysis 2). Based on in vitro activities of the different variants, some authors have combined the exon 3 and exon 4 mutations to estimate a predicted mEH activity in vivo. Following the classification of Benhamou et

159

al. [10], we have assigned the combined genotypes to three levels of activity. The criteria for grouping genotypes and the distribution among cases and controls are shown in Table 2. This distribution did not show significant differences between patients and controls, although there was a tendency for the individuals with predicted high activity (subjects with 113Tyr/113Tyr and 139His139/139Arg or 113Tyr/ 113Tyr and 139His/139His or 113Tyr/113His and

Table 3 Odds ratios for some selected genotypes or genotype combinations. ORs were adjusted for sex, age and smoking habit (natural logarithm of pack-years) Analysis 1 2 3

4

5

6

7

8

9

10 11

12

a

Category

Cases

Controls

OR

(95% CI)

EPHX1 exon3 Tyr/Tyr EPHX1 exon3 Tyr/His or His/His EPHX1 exon 4 His/His EPHX1 exon 4 Arg/Arg or His/Arg GSTM1 null EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTM1 positive EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTT1 null EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTT1 positive EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTP1 Ile/Ile EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTP1 Val/Val or Ile/Val EPHX1 Tyr/Tyr EPHX1 Tyr/His or His/His GSTP1 Ile/Ile and EPHX1 Tyr/Tyr GSTP1 Ile/Ile and EPHX1 Tyr/His or His/His GSTP1 Val/Val or Ile/Val and EPHX1 Tyr/Tyr GSTP1 Val/Val or Ile/Val and EPHX1 Tyr/His or His/His GSTP1 Ile/Ile £ EPHX1 Tyr/Tyr a GSTP1 Val/Val or Ile/Val and EPHX1 Tyr/His or His/His GSTP1 Ile/Ile mEH activity High mEH activity Medium mEH activity Low GSTP1 Val/Val or Ile/Val mEH activity High mEH activity Medium mEH activity Low

97 78 119 56

87 100 126 61

1.34 1 1.05 1

(0.85–2.13)

59 43

40 44

1.21 1

(0.65–2.26)

37 35

39 46

1.46 1

(0.75–2.83)

20 20

19 13

1.26 1

(0.60–2.65)

67 56

44 56

1.59 1

(0.95–2.68)

58 31

44 50

2.19 1

(1.12–4.28)

39 47 58 31 39 47 58 47

43 50 44 50 43 50 44 50

0.89 1 1.41 0.69 0.88 1 2.34 1

(0.45–1.77)

17 51 21

12 46 36

2.48 2.49 1

(1.15–5.32) (1.16–5.35)

16 36 34

17 42 34

0.96 0.70 1

(0.45–2.08) (0.33–1.52)

Interaction term (second order regression) estimated including both individual factors in the model.

(0.47–2.35)

(0.73–2.73) (0.36–1.33) (0.45–1.70) (1.21–4.52)

160

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

Arg139/Arg 139) to be over-represented among the cases. Cases and controls were also stratified according to cigarette consumption into heavy smokers (pack-years . 50) and medium/light smokers (packyears # 50). Adjusted OR for predicted mEH high activity were 1.43 (95% CI: 0.66–3.11) and 1.42 (95% CI: 0.70–2.87), respectively, thus suggesting no interaction between EPHX1 and tobacco smoking. Thereafter, we investigated possible interactions combining EPHX1 with GST genes previously genotyped in these patients and controls [14–17]. EPHX1 variants were combined with GSTM1, GSTP1, and GSTT1 genotypes. The analyses were carried out grouping rare EPHX1 homozygotes and heterozygotes in a single category. The results are shown in Table 3. Only in the case of GSTP1 we found a significant interaction with EPHX1. The distribution of the combined exon 3 EPHX1 genotypes and exon 5 Ile105ValGSTP1 genotypes in patients and controls showed a significant overrepresentation of patients homozygous for both exon 3 wild-type EPHX1 (113Tyr/113Tyr) and wild-type exon 5 GSTP1 (105Ile/105/Ile) (Table 4). If only the subjects with the 105Ile/105Ile GSTP1 were considered, a significant risk (OR: 2.19, 95% CI: 1.12–4.28) is associated with the homozygous EPHX1 wild-type. However, considering only the subjects with one or two Val GSTP1 alleles, no significant risk is associated with the homozygous EPHX1 wild-type (OR: 0.89, 95% CI: 0.45–1.77). An alternative analysis considering all the subjects, the four genotype combinations, and taking GSTP1 (Val/Val or Ile/Val) and EPHX1 (Tyr/His or His/His)

as the reference category was performed (Table 3, analysis 9). No significant relative risk was associated to any of the three other categories. However, applying a second order logistic regression model, using GSTP1, EPHX1, and the interaction term as independent variables, and considering the same reference category as in the previous analysis, a significant risk was found to be associated with the interaction term (OR: 2.34, 95% CI: 1.21–4.52, Table 3, analysis 10). The odds ratio for the interaction variable can be interpreted as the additional risk associated with the simultaneous presence of both wildtype homozygous genotypes. The rest of exon 3 EPHX1-GST combinations did not show significant interactions (Table 3). No interaction was found between the exon 4 139His/Arg EPHX1 polymorphism and the different GST alleles (data not shown). The interaction between GSTP1 and EPHX was also analysed using the predicted mEH activity for calculations. The results (Table 3) showed a significant increased risk for the combination 105Ile/105Ile GSTP1 and high and medium activity mEH.

4. Discussion Our results on Caucasians of Catalonia (Spain) support the hypothesis that the EPHX1 polymorphisms may be related to lung cancer risk. The frequency of exon 3 113His/113His EPHX1 was significantly increased among the controls suggesting a slightly decreased risk for the individuals with this homozygous rare genotype. Also, the frequency of the rare

Table 4 Distribution of EPHX1 exon 3 and GSTP1 exon 5 genotypes in cases and controls a GSTP1 exon 5

EPHX1 exon 3 genotypes Lung cancer patients

Ile/Ile Ile/Val Val/Val Total

Controls

Tyr/Tyr

Tyr/His

His/His

Total

Tyr/Tyr

Tyr/His

His/His

Total

58 31 8 97

28 32 10 70

3 4 1 8

89 67 19 175

44 31 12 87

42 36 7 85

8 6 1 15

94 73 20 187

a A chi-square analysis was performed grouping EPHX1 Tyr/His with His/His and GSTP1 Ile/Val with Val/Val, and showed a significant association between both polymorphisms among the patients (x2 ¼ 6:95; P ¼ 0:008) with over-representation of individuals EPHX1 Tyr/ Tyr 1 GSTP1 Ile/Ile.

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162

exon 4 139Arg/139Arg EPHX1 was also increased among the controls. The exon 3 113His/113His encodes for a low-activity mEH while exon 4 rare genotype encodes for a high-activity enzyme. This may seem contradictory, however, given the very small counts (three cases/seven controls, Table 2), the exon 4 difference should be confirmed by a larger study. The small counts of rare homozygotes in both EPHX1 variants also precluded an analysis stratified by smoking status and tumour histology. Some previous reports have studied the relation between lung cancer risk and the mEH predicted activity. Smith and Harrison [9] studied a group of lung cancer patients of Caucasian origin and found an association between predicted ‘low-activity’ mEH and susceptibility to emphysema but they did not confirm a similar association with lung cancer risk. Conversely, Benhamou et al. [10] reported an association between predicted ‘high-activity’ mEH and lung cancer risk among French Caucasians. The study of London et al. [11] in Los Angeles county found an association between susceptibility to lung cancer and predicted ‘high’ activity among African-Americans although the same study reported that a similar association was not found among white Caucasians. Lin et al. [12] also found a significant risk excess of squamous cell carcinoma (but not other histological types of lung cancer) with high mEH activity. In all these cases ‘low’ and ‘high’ activity was predicted from combinations of exon 3 and exon 4 EPHX1 genotypes and the supposed activities of the different protein variants as estimated in vitro by Hasset et al. [8]. However, the genetically determined activity of mEH in human tissues may not be completely predicted from these in vitro estimations since a genetic polymorphism in the 5 0 flanking region has also been described that may modify the expression of the EPHX1 gene [18,19]. The analysis of a possible interaction between EPHX1 and GSTs was performed using for calculations both the separate frequencies of the two EPHX polymorphisms and the predicted mEH activity. No interaction was found between EPHX1 and GSTM1 or GSTT1 but the study of combined alleles of EPHX1 and GSTP1 yield an overrepresentation of cases with both exon 3 113His/113His EPHX1 and 105Ile/105Ile GSTP1. The statistical interaction suggests that this particular genotype combination may produce an

161

increased cancer risk. The alternative calculation using the predicted mEH activity showed also an excess of cancer cases with high or medium mEH activity clearly restricted to those individuals with a 105Ile/Ile GSTP1 genotype. Given the dual role of mEH on the bioactivation/detoxification of carcinogenic hydrocarbons it is probable that the generation of mutagenic intermediates may be highly dependent not only of mEH activity but also of the activity of other biotransformation enzymes as GSTP1. The relation between the 105Ile/Val GSTP1 polymorphism and cancer risk has not been elucidated [17,20]. In a previous lung cancer study performed on the same population, no risk was found to be associated with this GSTP1 polymorphism alone (OR: 1.04, 95% CI: 0.66–1.63) [17]. However, the polymorphic site within the exon 5 of GSTP1 results in a Ile ! Val substitution at residue 105 of GSTpi protein and the valine 105 GSTpi isoform has been found in vitro to be more active than the isoleucine isoform towards the conjugation of BPDE [21,22]. In this case, the combination of a 113Tyr EPHX1 (encoding for a high-activity mEH) with a 105Ile GSTP1(encoding for a low-activity GSTpi) may produce some shift in the formation/elimination ratio of BPDE in tissues and facilitate the formation of adducts in DNA hotspots. The results of our study tend to agree with this mechanistic basis suggesting an increased lung cancer risk for this genotype combination. However, more studies are needed to assess the genotype/phenotype correlation of both mEH and GSTp1, the effect of modulators of enzyme activity present in the diet, and the combined effect in vivo of both enzymes in the bioactivation of carcinogenic hydrocarbons.

Acknowledgements Supported by Spanish ‘Fondo de Investigacio´ n Sanitaria’ (grant no. 00/0504).

References [1] A. d’Errico, E. Taioli, X. Che, P. Vineis, Genetic metabolic polymorphisms and the risk of cancer: a review of the literature, Biomarkers 1 (1996) 149–173. [2] F. Oesch, Mammalian epoxide hydrolases: inducible enzymes catalyzing the inactivation of carcinogenic and cytotoxic

162

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

J. To-Figueras et al. / Cancer Letters 173 (2001) 155–162 metabolites derived from aromatic and olefinic compounds, Xenobiotica 3 (1973) 305–340. P. Sims, P.L. Grover, A. Swaisland, K. Pal, A. Hewer, Metabolic activation of benzo[a]pyrene proceeds by a diol epoxide, Nature 252 (1974) 326–328. J.K. Hartsfield Jr., M.J. Sutcliffe, E.T. Everett, C. Hassett, C.J. Omiecinski, J.A. Saari, Assignment of microsomal epoxide hydrolase (EPHX1) to human chromosome 1q42.1 by in situ hybridization, Cytogenet. Cell Genet. 83 (1998) 44–45. C. Hassett, K.B. Robinson, N.B. Beck, C.J. Omiecinski, The human microsomal epoxide hydrolase gene (EPHX1): complete nucleotide sequence and structural characterization, Genomics 23 (1994) 433–442. C. Hassett, J. Lin, C.L. Carty, E.M. Laurenzana, C.J. Omiecinski, Human hepatic microsomal epoxide hydrolase: comparative analysis of polymorphic expression, Arch. Biochem. Biophys. 337 (1997) 275–283. C.J. Omiecinski, C. Hassett, V. Hosagrahara, Epoxide hydrolase polymorphism and role in toxicology, Toxicol. Lett. 112113 (2000) 365–370. C. Hassett, L. Aicher, J.S. Sidhu, Human microsomal epoxide hydrolase: genetic polymorphism and functional expression in vitro of amino acid variants, Hum. Mol. Genet 3 (1994) 421– 428. C.A. Smith, D.J. Harrison, Association between polymorphism in gene for microsomal epoxide hydrolase and susceptibility to emphysema, Lancet 350 (1997) 360–363. S. Benhamou, M. Reinikainen, C. Bouchardy, P. Dayer, A. Hirvonen, Association between lung cancer and microsomal epoxide hydrolase genotype, Cancer Res. 58 (1998) 5291– 5293. S. London, J. Smart, A.K. Daly, Lung cancer risk in relation to genetic polymorphisms of microsomal epoxide hydrolase among African-Americans and Caucasians in Los Angeles County, Lung Cancer 28 (2000) 147–155. P. Lin, S.L. Wang, H.J. Wang, K.W. Chen, H.S. Lee, K.J. Tsai, C.Y. Chen, H. Lee, Association between CYP1A1 and microsomal epoxide hydrolase polymorphisms with lung squamous cell carcinoma, Br. J. Cancer 82 (2000) 852–857. Y. Persson, Y. Johansson, Y.C. Lou, M. Ingelman-Sundberg, Genetic polymorphism of xenobiotic metabolizing enzymes among Chinese lung cancer patients, Int. J. Cancer 81 (1999) 325–329.

[14] J. To-Figueras, M. Gene´ , J. Go´ mez-Catala´ n, M.C. Gala´ n, J. Firvida, M. Fuentes, M. Rodamilans, E. Huguet, J. Estope´ , J. Corbella, Glutathione-S-transferase M1 and codon 72 p53 polymorphisms in a Northwestern mediterranean population and their relation to lung cancer susceptibility, Cancer Epidemiol. Biol. Prev. 5 (1996) 337–342. [15] J. To-Figueras, M. Gene´ , M. Go´ mez-Catala´ n, M.C. Gala´ n, M. Fuentes, J.M. Ramo´ n, M. Rodamilans, E. Huguet, G. Corbella, Glutathione S-transferase M1 (GSTM1) and T1 (GSTT1) polymorphisms and lung cancer risk among Northwestern Mediterraneans, Carcinogenesis 18 (1997) 1529–1533. [16] J. To-Figueras, M. Gene´ , J. Go´ mez-Catala´ n, E. Pique´ , N. Borrego, G. Marfany, R. Gonzalez-Duarte, J. Corbella, Polymorphism of glutathione S-transferase M3: interaction with glutathione S-transferase M1 and lung cancer susceptibility, Biomarkers 5 (2000) 73–80. [17] J. To-Figueras, M. Gene´ , J. Go´ mez-Catala´ n, E. Pique´ , N. Borrego, J.L. Carrasco, J. Ramo´ n, J. Corbella, Genetic polymorphism of glutathione S-transferase P1 gene and lung cancer risk, Cancer Causes Control 10 (1999) 65–70. [18] S. Raaka, C. Hassett, C.J. Omiencinski, Human microsomal epoxide hydrolase: 5 0 -flanking region genetic polymorphisms, Carcinogenesis 19 (1998) 387–393. [19] E.M. Laurenzana, C. Hassett, C.J. Omiecinski, Post-transcriptional regulation of human microsomal epoxide hydrolase, Pharmacogenetics 2 (1998) 57–67. [20] M.J. Harris, M. Coggan, L. Langton, S.R. Wilson, P.G. Board, Polymorphism of the Pi class glutathione S-transferase in normal populations and cancer patients, Pharmacogenetics 8 (1998) 27–31. [21] K. Sundberg, A.S. Johansson, G. Stenberg, M. Widerstein, A. Seidel, B. Mannervile, B. Jernstro¨ m, Differences in the catalytic efficiencies of allelic variants of glutathione S-transferase P1-1 towards carcinogenic diol epoxides of polycyclic aromatic hydrocarbons, Carcinogenesis 19 (1998) 433– 436. [22] P. Zimniak, B. Nanduri, S. Pikula, J. Bondorowicz-Pikula, S.S. Singhal, S.K. Srivastava, S. Awasthi, Y.C. Awasthi, Naturally occurring human glutathione S-transferase GSTP1-1 isoforms with isoleucine and valine in position 104 differ in enzymic properties, Eur. J. Biochem. 224 (1994) 893–899.