Genetic polymorphisms of glutathione S ... - Semantic Scholar

7 downloads 130 Views 82KB Size Report
Some of the glutathione S-transferases (GSTs) are poly- morphic and may play a role in lung cancer susceptibility. Our previous study in a French Caucasian ...
Carcinogenesis vol.23 no.9 pp.1475–1481, 2002

Genetic polymorphisms of glutathione S-transferases as modulators of lung cancer susceptibility

Isabelle Stu¨cker1,4, Ari Hirvonen2, Isabelle de Waziers3, Arnauld Cabelguenne3, Katja Mitrunen2, Sylvie Ce´ne´e1, Elisabeth Koum-Besson3, Denis He´mon1, Philippe Beaune3 and Marie-Anne Loriot3 1INSERM

U 170, IFR69, Epidemiologic and Statistical Research on Environment and Health, 16 Avenue Paul-Vaillant Couturier, 94807 Villejuif cedex, France, 2Department of Industrial Hygiene and Toxicology, Finnish Institute of Occupational Health, FIN-00250, Helsinki, Finland and 3INSERM U 490, Molecular Toxicology, University of Rene ´ Descartes, 75006 Paris, France 4To

whom correspondence should be addressed Email: [email protected]

Some of the glutathione S-transferases (GSTs) are polymorphic and may play a role in lung cancer susceptibility. Our previous study in a French Caucasian study population suggested GSTM1 null genotype as a moderate risk factor for lung cancer. Here we extended the study to investigate the potential role of GSTT1 and GSTP1 polymorphisms in susceptibility to lung cancer, either separately or in combination. The study population consisted of 268 controls and 251 cases. Nineteen percent of the controls and 15% of the cases had GSTT1 null genotype. The distribution of GSTP1*A/*A, *A/*B and *B/*B genotypes were 46.9, 45.5 and 7.6% in controls, and 47.8, 40.2 and 12.0% in cases, respectively. No statistically significant effects in the lung cancer risk were observed for the GSTT1 genotypes, but the GSTP1*B/*B genotype posed a 2-fold risk [odds ratio (OR) ⫽ 2.0, 95% confidence interval (CI) 1.0–4.1] of this malignancy compared with the GSTP1*A allele containing genotypes; this association was mainly attributable to small cell lung cancer (OR ⫽ 3.6, 95% CI 1.3–9.8). The most remarkable risk was seen for the small cell carcinoma among subjects with the GSTP1*B/*B genotype and concurrent lack of the GSTM1 gene (OR ⫽ 6.9, 95% CI 1.6– 30.2). The deficient genotypes for GSTM1 and GSTP1 seem thus to be important risk modifiers for lung cancer, especially in combination.

Introduction Tobacco smoking is a major cause of lung cancer. Among the constituents of tobacco smoke, the polycyclic aromatic hydrocarbons (PAHs), such as benzo[a]pyrene (B[a]P), play a major role in the chemical lung carcinogenesis (1). These carcinogens require metabolic activation to exert their carcinogenic effect. Many xenobiotic metabolising enzymes (XMEs) like cytochromes P450, glutathione S-transferases (GSTs) and epoxide hydrolase are involved in the metabolism of B[a]P, which can lead to or prevent the formation of its highly mutagenic metabolite 7,8-diol-9,10-epoxide (BPDE) (2). The Abbreviations: CI, confidence interval; GST, glutathione S-transferases; OR, odds ratio. © Oxford University Press

balance between metabolic activation and detoxification pathways differ between individuals and is thought to affect cancer susceptibility. This balance may be modified by genetic variations in XMEs (3). GSTs are a multigene family of enzymes (cytosolic and membrane-bound), which catalyse the conjugation of glutathione to electrophilic xenobiotics in order to inactivate them and facilitate their excretion from the body (4). GSTs play an important role in the metabolism of potentially genotoxic compounds like BPDE, which is good substrate for GSTs. The cytosolic isoenzymes are divided into at least five major classes (α, µ, π, θ, ζ) among which polymorphisms have been detected in the genes encoding for GSTM1 and GSTM3 (µ class), GSTP1 (π class), GSTT1 (θ class) and GSTZ1 (ζ class) (5–7). Among them, the GSTM1, GSTP1 and GSTT1 genotypes have been extensively studied during recent years for their potential modulating role in individual susceptibility to environmentally induced diseases, including cancer (3). GSTM1 and GSTP1 detoxify more specifically active metabolites of polycyclic aromatic hydrocarbons, whereas GSTT1 is known to be involved in the metabolism of small compounds found in tobacco smoke like monohalomethanes and ethylene oxide (4). The GSTM1 locus is entirely deleted (null genotype) in ~50% of Caucasians. This deletion is associated with a slightly increased risk of lung cancer (8–10). A similar deletion at GSTT1 locus occurs in 10–20% of Caucasians. The results from the few studies conducted so far on the association between lung cancer risk and GSTT1 polymorphism do not consistently suggest a relationship between the GSTT1 genotypes and lung cancer (8,11,12). At least two different alleles, GSTP1*A and GSTP1*B, have been identified encoding for GSTP1, exhibiting A313G base change in the latter allele which results in an Ile104Val amino acid replacement within the active site of the enzyme. This residue lies in close proximity to the hydrophobic binding site for electrophilic substrates and the GSTP1*B allele has been demonstrated to exhibit altered specific activity and affinity for electrophilic substrates (13). We found previously that the GSTM1 null genotype was associated with moderately increased lung cancer risk (14). As recent reports indicated a significant gene–gene interaction between the combination of the deficient genotypes for GSTM1 and GSTP1 in individual risk of this malignancy (15–17), we extended our previous study to include also the GSTP1 genotype data. In addition, the GSTT1 genotypes were determined in the same material. The effect of GSTP1 and GSTT1 genotypes in individual lung cancer proneness was examined both separately and in combination with the previously analyzed GSTM1 data. We also addressed their interaction with cigarette smoking in lung cancer risk. Materials and methods The original study population included 310 ⬍75-year-old male patients with a histologically confirmed lung cancer diagnosis, 1475

I.Stu¨ cker et al.

Table I. Selected characteristics of the study subjects included in the analysis n

Age

Pack-years

Histological types

Diagnosis n

All Cases

310

59.3 ⫾ 9.6 (309)

38.8 ⫾ 19.7 (294)

Controls

301

59.6 ⫾ 9.9 (301)

23.0 ⫾ 22.9 (291)

Subjects included in GSTM1 analysis Cases 247 59.7 ⫾ 9.5 (247)

38.0 ⫾ 19.5 (240)

Controls

254

59.3 ⫾ 10.0 (254)

Subjects included in GSTT1 analysis Cases 251 59.5 ⫾ 9.6 (251)

Controls

268

59.4 ⫾ 10.0 (268)

Subjects included in GSTP1 analysis Cases 251 59.5 ⫾ 9.6 (251)

Controls

264

59.3 ⫾ 10.0 (264)

Squamous cell Small cell Adenocarcinoma Others

Squamous cell Small cell Adenocarcinoma Others

22.7 ⫾ 23.1 (246)

38.2 ⫾ 19.5 (243)

Squamous cell Small cell Adenocarcinoma Others

22.7 ⫾ 22.7 (259)

38.2 ⫾ 19.5 (243)

Squamous cell Small cell Adenocarcinoma Others

22.9 ⫾ 22.8 (255)

and no history of other cancer diagnosis or of radio- or chemotherapy (Table I). Patients were recruited in chest departments of three French hospitals, two in the Paris area and one in Besanc¸ on (eastern France). They were all Caucasians, born in France to native French parents. All new lung cancer patients who satisfied the selection criteria agreed to participate. The 301 controls were all male subjects hospitalized for different disorders except cancer. They were matched with the cases for age (⫾2.5 years) and hospital. In addition, similar to the cases they were all required to be born in France to native French parents. The controls were recruited in parallel 1476

135 64 72 39 310

113 49 57 28 247

115 48 60 28 251

115 48 60 28 251

%

n

%

44 21 23 13 Nervous System Circulatory disease Musculo-skeletal disease Skin disease Poisoning Other Missing

70 66 30 35 48 30 22 301

23 22 10 12 16 10 7

Nervous system Circulatory disease Musculo-skeletal disease Skin disease Poisoning Other

64 53 26 30 39 22 234

27 23 11 13 17 9

Nervous system Circulatory disease Musculo-skeletal disease Skin disease Poisoning Other

64 58 26 32 42 25 247

26 23 11 13 17 10

Nervous system Circulatory disease Muskulo-squeletal disease Skin disease Poisoning Other

62 57 26 31 42 25 243

26 23 11 13 17 10

46 20 23 11

46 19 24 11

46 19 24 11

to the cases; each time a new lung cancer patient was included in the study we sought a control meeting the matching criteria. It was also required that this control patient had no history of cancer and was in sufficiently good condition to provide a blood sample. The number of possible controls meeting all these criteria was of course limited. We systematically recruited the first person that met these criteria if he agreed to participate in the study: nearly all of them accepted. If the first person asked did not accept to participate, we recruited the next control that met the above criteria and accepted to participate. All the study subjects were interviewed with a standard

Polymorphisms of GST’s and lung cancer susceptibility

questionnaire for their demographic characteristics, previous family history of cancer and lifetime occupational history. Detailed information was collected about cigar, pipe, cigarette consumption and smoking habits. For each brand and type of tobacco patients had ever smoked, they were asked to specify their age at starting to smoke and, if applicable, their age at quitting smoking. They were also asked for their usual daily consumption, use of filter tips (for cigarette smokers), and if they inhaled usually, sometimes or not at all. Blood samples (20 ml) were taken as soon as lung cancer diagnosis was histologically confirmed, before the initiation of any chemo- or radiotherapy. DNA was isolated from the lymphocyte preparations using standard techniques. The DNA bank was established for only 552 of the 611 subjects (89.5%), as blood was not taken from the first 19 subjects who allowed us to refine the recruitment protocol or from 23 other patients for medical reasons. In addition, samples from 17 patients were lost or damaged, and due to technical problems the genotype data could not be achieved for all study subjects. Ultimately the GSTM1 data were therefore available for 247 cases and 254 controls, GSTT1 data for 251 cases and 268 controls and GSTP1 data for 251 cases and 264 controls. Genotyping analyses Multiplex polymerase chain reaction (PCR)-based methods were used for the detection of GSTM1 and GSTT1 genotypes as described previously (14,18). Briefly, in the GSTM1 PCR analysis lack of GSTM1-specific fragment, amplified by two combinations of the primers, revealed the null genotypes. Similarly, in the GSTT1-genotyping analysis specific primers for the GSTT1 gene were used together with another primer pair for β-globin. The absence of GSTT1-specific PCR product indicated the corresponding null genotype whereas the β-globin-specific fragment confirmed proper functioning of the reaction. The GSTP1 genotypes were determined by a PCR–RFLP (restriction fragment length polymorphism) analysis as described previously (15). The A313G base change creates a restriction site recognized by the BsmAI enzyme (Biolabs, France) enabling differentiation between the wild-type GSTP1*A allele and the variant GSTP1*B allele. Statistical analysis All statistical analyses were performed on a RS6000 IBM (UNIX system), using the SAS Version 6.12 (Cary, NC, USA). The estimates we calculated were odds ratios (OR) with their 95% confidence intervals (95% CI). Potential confounding by the matching variables of age and hospital was accounted for by including them in an unconditional logistic regression, as the strata were large enough not to require matched analysis (19). Tobacco smoking was accounted for as two independent continuous variables: mean number of cigarettes smoked and total duration of tobacco smoking. All results were systematically adjusted for these two variables as well as the matching variables (i.e. age and hospital). Statistical interactions (OR and 95% CI) between two variables (either gene–gene or gene–environment) were assessed using unconditional logistic regression by adding in the usual model a multiplicative term of the two variables. When the interaction proved not to be significant, a second model without the interaction term furnished the parameters for estimating the OR of lung cancer associated with the variable. Stata software (version 6.0) was used to estimate a metaOR, based on a random effect model.

Results The distribution of the three GST genotypes and their respective ORs of lung cancer are presented in Table II. No statistically significant associations were observed between the GSTM1 and GSTT1 genotypes and overall lung cancer risk, whereas a 2-fold risk of lung cancer (OR ⫽ 2.0, 95% CI 1.0–4.1) was seen for the GSTP1*B/*B genotype carriers compared with the subjects with the GSTP1*A allele containing genotypes; the heterozygous and homozygous carriers of the GSTP1*A allele were grouped together as the corresponding OR was close to 1 (OR ⫽ 0.9, 95% CI 0.6–1.4) for the GSTP1*A/*B genotype, whereas it was almost 2 (OR ⫽ 1.9, 95% CI 0.9– 4.0) for the GSTP1*B/*B genotype (data not presented in Table II). When stratified by histological type of lung cancer, the strongest relation was observed between the GSTP1*B/*B genotype and risk of small cell lung cancer (OR ⫽ 3.6, 95% CI 1.3–9.6). A borderline significant association was also observed between the GSTP1*B/*B genotype and squamous cell carcinoma (OR ⫽ 2.1, 95% CI 0.9–4.8) whereas it appeared to have no effect in the risk of adenocarcinoma (OR ⫽ 0.9, 95% CI 0.3–2.9). On the other hand, the GSTM1 null genotype posed a significantly elevated risk of adenocarcinoma (OR ⫽ 2.2, 95% CI 1.1–4.2). A borderline significant association was also seen between GSTM1 null genotype and risk of small cell carcinoma (OR ⫽ 1.8, 95% CI 0.9– 3.6), whereas no association was seen for the squamous cell carcinoma (OR ⫽1.0, 95% CI 0.6–1.6). The results of the different combinations of the GST at-risk genotypes are presented in Table III. The presence of one of the GSTM1 or GSTP1 at-risk genotypes resulted in a statistically significant OR of 1.5 (95% CI 1.0–2.3). Concurrent presence of at-risk genotypes of both of these genes almost doubled the risk with borderline of significance (OR ⫽ 2.8, 95% CI 0.9– 8.5). Test for interaction, however, revealed no statistically significant interaction between the GSTM1 and GSTP1 genotypes. No statistically significant association was observed between the overall lung cancer risk and combinations of GSTM1 and GSTT1 genotypes, or GSTP1 and GSTT1 genotypes. Neither were there any statistically significant association between overall lung cancer risk and any of the combinations of the at-risk genotypes of all three GST genes (Table III). The combinations of GSTM1 and GSTP1 at-risk genotypes showed stronger relations with the increased number of deficient genotypes among the small cell carcinoma group; the OR of lung cancer was particularly high for the subjects carrying deficient genotypes for both GSTM1 and GSTP1 (OR ⫽ 7.0, 95% CI 1.6–30.2). Among the adenocarcinoma patients, on the other hand, a significant OR of 2.1 (95% CI 1.1–4.2) was only seen in the category of subjects with one deficient genotype, whereas among the squamous cell carcinoma group none of the genotype combinations were associated with significant ORs. When the relation between lung cancer and the combined deficient GST genotypes was examined according to the level of tobacco smoking (Table IV), we did not observe any significant overall relationship between lung cancer risk and different combinations of GST genotypes among the light smokers (⬍30 pack-years). Results were generally slightly stronger among the heavy smokers (艌30 pack-years), but no clear effects in the lung cancer risk were observed for the 1477

I.Stu¨ cker et al.

Table II. Odds ratio of lung cancer according to GST genotypes Genotypes Controls

All lung cancer

Small cell

n

(%)

n

(%)

GSTM1 Positive

254 136 118

(100) (53.5) (46.5)

247 118 129

GSTT1ı Positive Null

268 216 52

(100) (80.6) (19.4)

251 213 38

GSTP1c *A/*A *A/*G *G/*G

264 124 120 20

(100) (46.9) (45.5) (7.6)

251 120 101 30

(100) (47.8) 1.0 (Ref) (52.2) 1.4 (0.9–2.1)a 1.3 (0.9–1.8)b (100) (84.9) 1.0 (Ref) (15.1) 0.8 (0.4–1.3) 0.7 (0.5–1.2) (100) (47.8) 1.0 (Ref) (40.2) (12.0) 2.0 (1.0–4.1) 1.7 (0.9–3.0)

aOdds

OR (95% CI)

n

(%)

49 20 29

(100) (40.8) (59.2)

48 42 6

(100) (87.5) (12.5)

48 23 17 8

(100) (47.9) (35.4) (16.7)

Squamous cell OR (95% CI) 1.0 (Ref) 1.8 (0.9–3.6) 1.7 (0.9–3.2) 1.0 (Ref) 0.6 (0.2–1.6) 0.6 (0.2–1.5) 1.0 (Ref) 3.6 (1.3–9.6) 2.5 (1.0–6.0)

n

(%)

113 63 50

(100) (55.7) (44.3)

115 97 18

(100) (84.3) (15.7)

115 54 46 15

(100) (47.0) (40.0) (13.0)

Adenocarcinoma OR (95% CI) 1.0 (Ref) 1.0 (0.6–1.6) 0.9 (0.6–1.4) 1.0 (Ref) 0.9 (0.5–1.8) 0.8 (0.4–1.4) 1.0 (Ref) 2.1 (0.9–4.8) 1.8 (0.9–3.7)

n

(%)

57 21 36

(100) (36.8) (63.2)

60 52 8

(100) (86.7) (13.3)

60 32 23 5

(100) (53.4) (38.3) (8.3)

OR (95% CI) 1.0 (Ref) 2.2 (1.1–4.2) 2.0 (1.1–3.6) 1.0 (Ref) 0.7 (0.3–1.6) 0.6 (0.3–1.4) 1.0 (Ref) 0.9 (0.3–2.9) 1.1 (0.4–3.1)

ratio adjusted for age, hospital and tobacco smoking.

bOdds ratio adjusted for age, hospital. cThe ORs are for GSTP1*G/*G versus

GSTP1*(A/*A and *A/*G).

Table III. Odds ratio of lung cancer related to combination of GST genotypes according to histological type No. of at-risk genotypes

M1/P1 0 1

n

n

(%)

10 (4.1)

M1/T1 0 1 2

(%)

13

250 (100) 245 111 (44.4) 98 114 (45.6) 128 25 (10.0)

P1/T1 0 1

19

263 (100) 251 197 (74.9) 188 62 (23.6) 58

2

bOdds

All lung cancer

246 (100) 245 124 (50.4) 99 112 (45.5) 133

2

aOdds

Controls

4 (1.5)

5

Small cell OR (95% CI)

(100) (40.4) 1.0 (Ref) (54.3) 1.5 (1.0–2.3)a 1.5 (1.0–2.2)b (5.3) 2.8 (0.9–8.5) 1.6 (0.7–3.8) (100) (40.0) 1.0 (Ref) (52.2) 1.5 (0.9–2.2) 1.3 (0.9–1.9) (7.8) 0.9 (0.4–1.9) 0.9 (0.5–1.7) (100) (74.9) 1.0 (Ref) (23.1) 1.1 (0.7–1.7) 1.0 (0.6–1.5) (2.0) 1.3 (0.3–6.1) 1.3 (0.3–4.9)

n

(%)

48 15 29

(100) (31.3) (60.4)

4

(8.3)

48 16 29

(100) (33.3) (60.4)

3

(6.3)

48 34 14

(100) (70.8) (29.2)

0

(0)

OR (95% CI) 1.0 2.3 2.3 7.0 3.1

(Ref) (1.1–4.8) (1.1–4.4) (1.6–30.2) (0.9–11.4)

1.0 2.0 1.8 0.8 0.9

(Ref) (1.0–4.1) (0.9–3.5) (0.2–3.5) (0.2–3.4)

1.0 (Ref) 1.5 (0.7–3.2) 1.3 (0.7–2.7)

n

(%)

112 52 55

(100) (46.4) (49.1)

5 112 51 55 6 115 86 25 4

(4.5) (100) (45.5) (49.1) (5.4) (100) (74.8) (21.7) (3.5)

Adenocarcinoma OR (95% CI) 1.0 1.2 1.2 1.4 1.1

(Ref) (0.7–2.1) (0.7–1.9) (0.4–5.6) (0.4–3.5)

1.0 1.2 1.1 0.6 0.5

(Ref) (0.7–2.1) (0.7–1.7) (0.2–1.6) (0.2–1.4)

1.0 1.1 0.9 2.6 2.3

(Ref) (0.6–2.0) (0.6–1.6) (0.5–13.9) (0.6–9.4)

n

(%)

57 18 37

(100) (31.6) (64.9)

2

(3.5)

57 18 34

(100) (31.6) (59.6)

5

(8.8)

60 48 11

(100) (80.0) (18.3)

1

(1.7)

OR (95% CI) 1.0 2.1 2.2 1.8 1.5

(Ref) (1.1–4.2) (1.2–4.1) (0.3–10.2) (0.3–7.2)

1.0 2.1 1.8 1.4 1.2

(Ref) (1.0–4.1) (1.0–3.4) (0.4–4.7) (0.4–3.6)

1.0 0.7 0.7 0.8 1.1

(Ref) (0.3–1.6) (0.3–1.4) (0.07–9.4) (0.1–10.6)

ratio adjusted for age, hospital and tobacco smoking. ratio adjusted for age, hospital.

different combinations of the GST genotypes. Furthermore, only very few subjects had at least two deficient GST genotypes, leading to imprecise OR estimates. The results therefore do not suggest any statistically significant interaction between level of tobacco exposure and GST genotypes in the risk of lung cancer. Discussion We observed previously an association between lung cancer risk and GSTM1 genotype (14). The analysis suggested that the joint effect of tobacco exposure and GSTM1 polymorphism did not significantly differ from a multiplicative model. Here we extended the study to investigate whether the GSTT1 and GSTP1 genotypes, either separately or in combination with the GSTM1 genotype, were also related to individual lung cancer susceptibility. 1478

Squamous cell

The distributions of GST genotypes in this study agreed with those reported earlier in other Caucasians study populations (8,20). The homozygous GSTP1 deficient allele containing genotypes posed a 2-fold risk of lung cancer. The effect of the GSTP1*B/*B genotype was biggest in the risk of developing small cell lung cancer; it posed almost a 4-fold risk for this histological type of lung cancer. Moreover, if the GSTM1 gene was concurrently lacking, the risk for small cell carcinoma increased into almost 7-fold. The crude ORs for the association between the GSTM1 and GSTT1 genotypes and lung cancer were very close to the adjusted ones. For the GSTP1 genotype, the crude OR for the variant allele containing genotypes was somewhat smaller (OR 1.7, 95% CI 0.9–3.0) than the adjusted OR (OR 2.0, 95% CI 1.0–4.1). This difference was most pronounced in the small

Polymorphisms of GST’s and lung cancer susceptibility

Fig. 1. Odds ratio of lung cancer associated to GSTP1 polymorphism in studies previously published.

Table IV. Odds ratio of lung cancer related to combination of GST genotypes according to level of smoking No. of at-risk genotypes

M1/P1 0 1 2 M1/T1 0 1 2 P1/T1 0 1 2 aOdds

Light smokers (PA ⬍30)

Heavy smokers (PA ⬎30)

Cases

Cases

Controls

n

(%)

n

(%)

91 40 46 5 91 39 49 3 91 72 19 0

(100) (44.0) (50.5) (5.5) (100) (42.9) (53.8) (3.3) (100) (79.1) (20.9) (0)

154 74 74 6 158 66 75 17 166 122 42 2

(100) (48.05) (48.05) (3.9) (100) (41.8) (47.5) (10.7) (100) (73.5) (25.3) (1.2)

OR (95% CI)a 1.0 (Ref) 1.1 (0.7–1.9) 1.5 (0.4–5.5) 1.0 (Ref) 1.1 (0.6–1.9) 0.3 (0.08–1.1) 1.0 (Ref) 0.8 (0.4–1.4) –

Controls

n

(%)

n

(%)

147 55 84 8 147 54 77 16 152 109 38 5

(100) (37.4) (57.2) (5.4) (100) (36.7) (52.4) (10.9) (100) (71.7) (25.0) (3.3)

84 47 34 3 84 42 34 8 88 67 19 2

(100) (56.0) (40.5) (3.5) (100) (50.0) (40.5) (9.5) (100) (76.1) (21.6) (2.3)

OR (95% CI) 1.0 (Ref) 2.2 (1.2–3.8) 2.2 (0.5–8.9) 1.0 (Ref) 1.8 (1.0–3.2) 1.3 (0.5–3.8) 1.0 (Ref) 1.2 (0.6–2.2) 1.2 (0.2–6.8)

ratio adjusted for age and hospital.

cell lung cancer group where the crude and adjusted ORs were 2.5 (95% CI 1.0–6.0) and 3.6 (95% CI 1.3–9.6), respectively. This is due probably to the slight decrease in the pack-years of cigarettes smoked by controls with the GSTP1 genotype. Accordingly, even though no association has been reported between GSTP1-related metabolism and tobacco consumption, we considered smoking to be a potential confounding factor. Consequently we adjusted for cumulative cigarette smoking dose and total duration of smoking in all statistical analyses.

The present results on GSTM1 genotypes agree well with the recent meta-analysis suggesting a moderately increased risk of lung cancer for GSTM1 null individuals (9). The results on GSTP1 genotypes also agree with some previous studies on this topic. Similarly to us, Ryberg et al. (17) found a 2fold higher risk of lung cancer among the GSTP1*B/*B genotype carrying subjects as compared with the other genotypes. A couple of other studies have also found a moderately increased, although non-significant, risk associated with the 1479

I.Stu¨ cker et al.

GSTP1*B/*B genotype (15,16). However, four studies exist, which failed to find any association between the GSTP1 genotypes and lung cancer risk (18,21–23). One reason for the discrepant results could be that all of the studies were conducted in different populations. However, none of the main characteristics of the subjects explain satisfactorily these apparent discrepancies (i.e. race, histological type of lung cancer, level of smoking). Furthermore, the inter-study comparison confirms the homogeneity of the results (P of heterogeneity 0.65), allowing us to estimate a pooled OR of 1.3 (95% CI 1.1–1.6) for the GSTP1*B/*B genotype (Figure 1). In contrast to the GSTM1 and GSTP1 genotypes, no relationship between lung cancer and GSTT1 genotypes was observed in the present study, neither separately nor in association with the GSTM1 or GSTP1 genotypes. These results are in agreement with previously published data (11,18,24–26). Among the gene–gene combinations tested, only the combination of GSTM1 and GSTP1 at-risk genotypes was associated with an increased risk of lung cancer according to the number of deficient genotypes (P of tendency 0.02). This result was in total agreement with a multiplicative effect of the two deleterious genotypes in the risk of lung cancer. However, the very small number of subjects having jointly deficient genotypes for all the three enzymes hampered us to examine this relationship with lung cancer with sufficient statistical power. The combined effect of the deficiency in two GST genes has been addressed in several studies. In most cases the combined effect of GSTM1 and GSTT1 genotypes have been evaluated, but almost all of these studies failed to show any significant association between the joint deficiency of these genes and overall lung cancer risk (18,24–27); only one study conducted in migrant populations (Hispanic and AfricanAmerican) found a significant excess of lung cancer among carriers of null genotypes for both GSTM1 and GSTT1 genes (28). Some studies on the other genotype combinations have shown that the concurrent lack of GSTM1 and GSTP1 genes posed a statistically significant two-fold risk of lung cancer (15–17), but this was not seen in other studies (21,23). None of these five latter studies reported any statistically significant multiplicative interaction between both deficient genotypes. Such gene–gene interaction was observed between GSTP1 and GSTT1 (21), but we were not able to reproduce this result. We also investigated the joint effect of tobacco smoking and GST genotypes in the risk of lung cancer. Overall, we did not observe any significant interaction between the genetic and the environmental risk factors, in accordance with the majority of the literature (11,18,26). However, one study suggested the existence of an interaction between the combination of GSTM1, GSTM3 and GSTP1 genotypes and smoking. This result may, however, also result from associations between smoking and genotypes among the controls (21). The cases included in this study had incident lung cancer. It is therefore unlikely that survival bias has occurred in our analysis. Hospital controls were subjects suffering from different disorders. Although we cannot rule out a relationship between these pathologies and GST genes, no such association is known and it is thus unlikely that our results are due to such a relationship. Furthermore, in order to avoid any potential differences in cultural and genetic variations by countries or by French regions, we decided to include only individuals born in France to native French parents, and matched our cases and controls by the hospital. All cases and almost all 1480

controls agreed to participate. Although the genotype data was lacking from ~100 subjects as shown in Table I, no significant deviation existed in mean age, mean pack-years of cigarette smoking, distribution of histological type of cancer or type of disease, between subjects included in the genotype analysis and the initial study population. To conclude, in agreement with experimental and epidemiological data, our results suggest an important modulating role for the genetic polymorphism of the PAH-metabolizing enzymes GSTM1 and GSTP1 in individual risk for lung cancer, especially in combination and for certain histological subtypes of lung cancer. Future studies including examination of other genotypes also involved either in the metabolic activation or detoxication of carcinogens are anticipated to further improve our ability to find genetic factors contributing to individual cancer susceptibility. Acknowledgements This work is supported by the Association contre le Cancer (ARC), La Ligue contre le cancer, La Re´ gion Ile-de-France (programme SESAME), the Caisse Nationale d’Assurance Maladie des Travailleurs Salarie´ s (Collaboration INSERM-CNAMTS).

References 1. Hecht,S.S. (1999) Tobacco smoke carcinogens and lung cancer. J. Natl Cancer Inst., 91, 1194–1210. 2. Guengerich,F.P. and Shimada,T. (1998) Activation of procarcinogens by human cytochrome P450 enzymes. Mutat. Res., 400, 201–213. 3. Hirvonen,A. (1999) Polymorphisms of xenobiotic-metabolizing enzymes and susceptibility to cancer. Environ. Health Perspect., 107 (suppl. 1), 37–47. 4. Hayes,J.D. and Pulford,D.J. (1995) The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit. Rev. Biochem. Mol. Biol., 30, 445–600. 5. Ali-Osman,F., Akande,O., Antoun,G., Mao,J.X. and Buolamwini,J. (1997) Molecular cloning, characterization and expression in Escherichia coli of full-length cDNAs of three human glutathione S-transferase Pi gene variants. Evidence for differential catalytic activity of the encoded proteins. J. Biol. Chem., 272, 10004–10012. 6. Blackburn,A.C., Woollatt,E., Sutherland,G.R. and Board,P.G. (1998) Characterization and chromosome location of the gene GSTZ1 encoding the human Zeta class glutathione transferase and maleylacetoacetate isomerase. Cytogenet. Cell Genet., 83, 109–114. 7. Ketterer,B., Harris,J.M., Talaska,G., Meyer,D.J., Pemble,S.E., Taylor,J.B., Lang,N.P. and Kadlubar,F.F. (1992) The human glutathione S-transferase supergene family, its polymorphism and its effects on susceptibility to lung cancer. Environ. Health Perspect., 98, 87–94. 8. D’Errico,A., Malats,N., Vineis,P. and Boffetta,P. (eds) (1999) Review of studies of selected metabolic polymorphism and cancer. Metabolic Polymorphism and Susceptibility to Cancer, Vol. 148. IARC Scientific Publications, IARC, Lyon, pp. 323–393. 9. Houlston,R.S. (1999) Glutathione S-transferase M1 status and lung cancer risk: a meta-analysis. Cancer Epidemiol. Biomarkers Prev., 8, 675–682. 10. McWilliams,J.E., Sanderson,B.J., Harris,E.L., Richert-Boe,K.E. and Henner,W.D. (1995) Glutathione S-transferase M1 (GSTM1) deficiency and lung cancer risk. Cancer Epidemiol. Biomarkers Prev., 4, 589–594. 11. Jourenkova,N., Reinikanen,M., Bouchardy,C., Husgafvel-Pursiainen,K., Dayer,P., Benhamou,S. and Hirvonen,A. (1997) Effects of glutathione S-transferases GSTM1 and GSTT1 genotypes on lung cancer risk in smokers. Pharmacogenetics, 7, 515–518. 12. Malats,N., Camus-Radon,A.M., Nyberg,F. et al. (2000) Lung cancer risk in nonsmokers and GSTM1 and GSTT1 genetic polymorphism [In Process Citation]. Cancer Epidemiol. Biomarkers Prev., 9, 827–833. 13. Watson,M.A., Stewart,R.K., Smith,G.B., Massey,T.E. and Bell,D.A. (1998) Human glutathione S-transferase P1 polymorphisms: relationship to lung tissue enzyme activity and population frequency distribution. Carcinogenesis, 19, 275–280. 14. Stucker,I., de Waziers,I., Cenee,S., Bignon,J., Depierre,A., Milleron,B., Beaune,P. and Hemon,D. (1999) GSTM1, smoking and lung cancer: a case-control study. Int. J. Epidemiol., 28, 829–835.

Polymorphisms of GST’s and lung cancer susceptibility 15. Harris,M.J., Coggan,M., Langton,L., Wilson,S.R. and Board,P.G. (1998) Polymorphism of the Pi class glutathione S-transferase in normal populations and cancer patients. Pharmacogenetics, 8, 27–31. 16. Kihara,M. and Noda K. (1999) Lung cancer risk of the GSTM1 null genotype is enhanced in the presence of the GSTP1 mutated genotype in male Japanese smokers. Cancer Lett., 137, 53–60. 17. Ryberg,D., Skaug,V., Hewer,A. et al. (1997) Genotypes of glutathione transferase M1 and P1 and their significance for lung DNA adduct levels and cancer risk. Carcinogenesis, 18, 1285–1289. 18. Saarikoski,S.T., Voho,A., Reinikainen,M., Anttila,S., Karjalainen,A., Malaveille,C., Vainio,H., Husgafvel-Pursiainen,K. and Hirvonen,A. (1998) Combined effect of polymorphic GST genes on individual susceptibility to lung cancer. Int. J. Cancer, 77, 516–521. 19. Rothman,K.J. (1986) Modern Epidemiology. Little, Brown & Co., Boston, MA. 20. Rebbeck,T.R. (1997) Molecular epidemiology of the human glutathione Stransferase genotypes GSTM1 and GSTT1 in cancer susceptibility. Cancer Epidemiol. Biomarkers Prev., 6, 733–743. 21. Jourenkova,N., Wikman,H., Bouchardy,C., Voho,A., Dayer,P., Benhamou,S. and Hirvonen,A. (1998) Role of glutathione S-transferase GSTM1, GSTM3, GSTP1 and GSTT1 genotypes in modulating susceptibility to smoking-related lung cancer. Pharmacogenetics, 8, 495– 502. 22. Katoh,T., Kaneko,S., Takasawa,S., Nagata,N., Inatomi,H., Ikemura,K., Itoh,H., Matsumoto,T., Kawamoto,T. and Bell,D.A. (1999) Human glutathione S-transferase P1 polymorphism and susceptibility to smoking

related epithelial cancer; oral, lung, gastric, colorectal and urothelial cancer. Pharmacogenetics, 9, 165–169. 23. To-Figueras,J., Gene,M., Gomez-Catalan,J., Pique,E., Borrego,N., Carrasco,J.L., Ramon,J. and Corbella,J. (1999) Genetic polymorphism of glutathione S-transferase P1 gene and lung cancer risk. Cancer Causes Control, 10, 65–70. 24. el-Zein,R., Zwischenberger,J.B., Wood,T.G., Abdel-Rahman,S.Z., Brekelbaum,C. and Au,W.W. (1997) Combined genetic polymorphism and risk for development of lung cancer. Mutat. Res., 381, 189–200. 25. Salagovic,J., Kalina,I., Stubna,J., Habalova,V., Hrivnak,M., Valansky,L., Kohut,A. and Biros,E. (1998) Genetic polymorphism of glutathione S-transferases M1 and T1 as a risk factor in lung and bladder cancers. Neoplasma, 45, 312–317. 26. To-Figueras,J., Gene,M., Gomez-Catalan,J., Galan,M.C., Fuentes,M., Ramon,J.M., Rodamilans,M., Huguet,E. and Corbella,J. (1997) Glutathione S-transferase M1 (GSTM1) and T1 (GSTT1) polymorphisms and lung cancer risk among Northwestern Mediterraneans. Carcinogenesis, 18, 1529–1533. 27. Deakin,M., Elder,J., Hendrickse,C. et al. (1996) Glutathione S-transferase GSTT1 genotypes and susceptibility to cancer: studies of interactions with GSTM1 in lung, oral, gastric and colorectal cancers. Carcinogenesis, 17, 881–884. 28. Kelsey,K.T., Spitz,M.R., Zuo,Z.F. and Wiencke,J.K. (1997) Polymorphisms in the glutathione S-transferase class mu and theta genes interact and increase susceptibility to lung cancer in minority populations (Texas, United States). Cancer Causes Control, 8, 554–559. Received March 5, 2002; revised May 12, 2002; accepted May 20, 2002

1481