Does occupational exposure to solvents and pesticides in association with glutathione S-transferase A1, M1, P1, and T1 polymorphisms increase the risk of bladder cancer? The Belgrade case-control study.

OBJECTIVE: We investigated the role of the glutathione S-transferase A1, M1, P1 and T1 gene polymorphisms and potential effect modification by occupational exposure to different chemicals in Serbian bladder cancer male patients. PATIENTS AND METHODS: A hospital-based case-control study of bladder cancer in men comprised 143 histologically confirmed cases and 114 age-matched male controls. Deletion polymorphism of glutathione S-transferase M1 and T1 was identified by polymerase chain reaction method. Single nucleotide polymorphism of glutathione S-transferase A1 and P1 was identified by restriction fragment length polymorphism method. As a measure of effect size, odds ratio (OR) with corresponding 95% confidence interval (95%CI) was calculated.
RESULTS: The glutathione S-transferase A1, T1 and P1 genotypes did not contribute independently toward the risk of bladder cancer, while the glutathione S-transferase M1-null genotype was overrepresented among cases (OR = 2.1, 95% CI = 1.1-4.2, p = 0.032). The most pronounced effect regarding occupational exposure to solvents and glutathione S-transferase genotype on bladder cancer risk was observed for the low activity glutathione S-transferase A1 genotype (OR = 9.2, 95% CI = 2.4-34.7, p = 0.001). The glutathione S-transferase M1-null genotype also enhanced the risk of bladder cancer among subjects exposed to solvents (OR = 6,5, 95% CI = 2.1-19.7, p = 0.001). The risk of bladder cancer development was 5.3-fold elevated among glutathione S-transferase T1-active patients exposed to solvents in comparison with glutathione S-transferase T1-active unexposed patients (95% CI = 1.9-15.1, p = 0.002). Moreover, men with glutathione S-transferase T1-active genotype exposed to pesticides exhibited 4.5 times higher risk in comparison with unexposed glutathione S-transferase T1-active subjects (95% CI = 0.9-22.5, p = 0.067).
CONCLUSION: Null or low-activity genotypes of the glutathione S-transferase A1, T1, and P1 did not contribute independently towards the risk of bladder cancer in males. However, in association with occupational exposure, low activity glutathione S-transferase A1 and glutathione S-transferase M1-null as well as glutathione S-transferase T1-active genotypes increase individual susceptibility to bladder cancer.

Introduction

Bladder cancer is the second most common malignancy of the urinary tract and has the second highest mortality rate among urological neoplasms [1]. It affected 73,510 patients and lead to 14,880 deaths in 2012 worldwide [2]. Demographic characteristics associated with the greatest risk for bladder cancer include male gender, white race and the increasing age [3]. It is generally estimated that the male:female incidence ratio is 3.8∶1.0 [3]. The most frequent pathohistological type of bladder cancer is urothelial carcinoma, also called transitional cell carcinoma (TCC), accounting for approximately 90% of all bladder cancers [3]. It has been known that uroepithelial cells are most vulnerable to metabolic end products of different compounds, including carcinogens. This malignancy is characterized by multifactorial etiology, involving both genetic and environmental factors.

The well established risk factors for bladder cancer include cigarette smoking (50% cases in men, 30% cases in women), but also exposure to occupational agents [3]. Occupational exposures account for 5 to 25% of all bladder cancer cases. [4]. Over 40 occupations have been associated with an elevated risk of bladder cancer in epidemiologic studies, but the evidence is compelling for only a few. Those established at risk industries include the manufacturing of products such as synthetic dyes and paints, cables, textiles, leather works, and aluminum and the petrochemical, coal tar, and rubber industries [5], [6]. A number of specific occupations have also been identified to be associated with increased risk of bladder cancer. These include, but are not limited to, cooks and kitchen workers, electricians, hairdressers, leather workers, machinists, petroleum workers, rubber workers, coalminers, truckers, and vehicle mechanics, as summarized by Schulte et al. [7] in 1987, as well as coke oven workers, roofers, dry cleaners, chimney sweeps, and painters, as addressed by others in more recent literature [5], [8]–[10].

Despite the fact that occupations associated with bladder cancer have been well established, the question still arises why individuals with seemingly equal exposure to occupational carcinogens develop bladder cancer in an unpredictable manner. This is probably attributed to genetic polymorphisms of the genes coding for the xenobiotic metabolizing enzymes, particularly glutathione S-transferase (GST). GSTs catalyze the conjugation of glutathione on electrophilic substrates and are an important line of defense in the protection of cellular components against reactive species. The most well characterized GST classes have been named alpha (GSTA), mu (GSTM), pi (GSTP) and theta (GSTT). Appreciable GST activities are seen in bladder epithelium [11]. GST enzymes that belong to various classes have different, but sometimes overlapping, substrate specificities. Several types of allelic variations have been identified within GST classes, with that in the GSTM1, GSTT1 and GSTP1 genes receiving the most attention in genetic epidemiological studies [12]. Individuals homozygous for the GSTM1*0 and GSTT1*0 alleles (frequently referred to as GSTM1-null and GSTT1-null genotypes), which comprise for 50% and 11–18% of white population, respectively [13], [14], exhibit loss of GSTM1 and GSTT1 enzymatic activity. Single-nucleotide polymorphism (SNP) leading to amino acid substitution from isoleucine (Ile) to valine (Val) changes catalytic activity of the GSTP1 enzyme [15]. In healthy Caucasians, the frequencies of the genotype variants of GSTP Ile/Ile, -Ile/Val and -Val/Val are 51.5, 39.4, and 9.1%, respectively [15]. The role of GSTA1 polymorphism has emerged relatively recently in genetic epidemiological studies. It is represented by three, apparently linked, single nucleotide polymorphisms (SNPs): -567TOG, -69COT, -52GOA [16]. These substitutions result in differential expression with lower transcriptional activation of variant GSTA1*B (-567G, -69T, -52A) than common GSTA1*A allele (-567T, -69C,-52G) [16]. The relative frequencies of GSTA1-AA, AB and BB genotype in Caucasians are 38%, 48% and 14%, respectively [16].

Many of the well known occupational agents, such as polycyclic aromatic hydrocarbons, aromatic amines, halogenated hydrocarbons, associated with bladder cancer risk are substrates for GST. Although this reaction generally results in detoxification, in selected cases GST-mediated conjugation may lead to a more toxic or mutagenic metabolite. Still the data on association between GST gene variants and risk of occupational bladder cancer are scarce. We hypothesized that GST gene variants coding for enzymes involved in biotransformation of specific occupational agents may influence the risk of occupational bladder cancer. Therefore, in this case-control study we investigated the role of the polymorphisms GSTA1, GSTM1, GSTP1 and GSTT1 gene and potential of effect modification by occupational exposure to different chemicals in Serbian male TCC patients.

Methods and Materials

Ethics Statement

This study was approved by the Ethical Committee of Faculty of Medicine, University of Belgrade and conducted according to the principles expressed in the Declaration of Helsinki. All the participants provided written informed consent.

Study subjects

A hospital-based case-control study of urinary bladder cancer in men was carried out between September 2007 and January 2010. A total of 143 histologically confirmed incident urinary bladder carcinoma male cases were recruited from the Clinics of Urology and Nephrology, Clinical centre of Serbia, Belgrade. This is the national reference center for urology and nephrology and the majority of bladder cancer patients from Serbia are diagnosed and treated at this clinic. The control group consisted of 114 male subjects which were recruited from individuals with nephrolithiasis admitted to the same hospital during the same period of time and had no history of any malignant disease. Urinary bladder carcinoma patients and corresponding controls did not differ with respect to mean age (Table 1).

Table 1

Selected characteristics of male patients with bladder cancer and controls.

Characteristic	Cases	Controls	OR (95%CI)	P
	n (%)	n (%)
Group
Male	143	114
Age (years)	63.6±10.7	61.1±9.9		N.S.
Smoking habits
Never smokers	25 (18)	37 (34)	1.0 (reference group)
Current smokers	112 (82)	72 (66)	2.3 (1.3–4.1)	0.005
No of pack-years of smoking	46.4±28.1	41.9±30.3	1.3 (0.7–2.5)	0.357
Occupational exposure
No	77 (54)	80 (70)	1.0 (reference group)
Yes	66 (46)	34 (30)	3.2 (1.6–6.6)a	0.001
Organic solvents	48 (34)	22 (19)	3.4 (1.5–7.3)a	0.002
Pesticides	15 (10)	9(8)	3.5 (0.9–12.9)a	0.058
Other chemicals	3 (2)	3 (3)	2.6 (0.4–17.7)a	0.323

N.S. not significant, OR- odds ratio, CI-confidence interval,

OR adjusted by age and pack-years.

After the informed consent was obtained, each subject was interviewed by well-trained interviewers using a standard questionnaire to collect information including demographic characteristics, history of cigarette smoking and occupational exposure. Response rate was 92% and the most frequent reason for no participation was personal.

In our study, smokers were defined as persons who reported every day smoking during a minimum of 60-day period prior to completing the questionnaire. Participants were asked about the number of cigarettes smoked per day and duration of smoking. The amount of pack-years was calculated using the following formula: pack-years  =  (cigarettes/day÷20)×(smoked years).

The life-time occupational history listed all jobs (including official jobs and jobs done outside normal working hours) lasting more than six months and consisted of the job title, the industry or type of business, employment dates and duration, company name and location, tasks as well as the exposure to at least one of the categories of agents under study, solvents and pesticides. In order to analyze occupational exposure occupational reports of patients were evaluated by experienced specialist in occupational medicine (author, P.B.). The exposure categories were defined as no exposure and exposure. Based on the evaluation patients were exposed to the following organic solvents: tetrachloroethylene, toluene, xylene, ethyl acetate, acetone, petrol ether and ethanol, as well as pesticides: organophosphate, carbamates, aminophosphonic analogues, chloroacetanilides, derivative of benzoic acid. All exposure data referred to a time period prior to the diagnosis of bladder cancer for the cases, and a corresponding period for the controls.

DNA extraction and genotyping

Genomic DNA was isolated from whole blood using the QIAGEN QIAmp (Qiagen, Inc., Chatsworth, CA, USA) 96-spin blood protocol according to the manufacturer's instructions. Blood was collected when patients were admitted to the clinic.

GSTM1 genotyping was performed by multiplex PCR method [17]. Primers used were GSTM1 forward: 5′-GAACTCCCTGAAAAGCTAAAGC-3′ and GSTM1 reverse: 5′-GTTGGGCTCAAATATACGGTGG-3′. Exon 7 of the CYP1A1 gene was co-amplified and used as an internal control using the following primers: CYP1A1 forward: 5′-GAACTGCCACTT CAGCTGTCT-3; and CYP1A1 reverse: 5′-CAGCTGCATTTG GAAGTGCTC-3′. The presence of the GSTM1-active genotype was detected by the band at 215 bp, since the assay does not distinguish heterozygous or homozygous wild-type genotypes. Internal positive control (CYP1A1) PCR product corresponded to 312 bp.

GSTT1 genotyping was performed by multiplex PCR method [17]. Primers used were GSTT1-forward: 5′-TTCCTTACTGGTCCTCACATCTC-3′ and GSTT1-reverse: 5′-TCACGGGATCATGGCCAGCA-3′. Exon 7 of CYP1A1 genes were co-amplified and used as an internal control. The assay does not distinguish between heterozygous or homozygous wild-type genotypes; therefore, the presence of 480 bp bands was indicative for the GSTT1-active genotype. Internal positive control (CYP1A1) PCR product corresponded to 312 bp.

GSTP1 Ile105Val polymorphism was analyzed using the polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) [18]. Primers used were: GSTP1 Ile105Val forward: 5′-ACCCCAGGGCTCTATGGGAA-3′ and GSTP1 Ile105Val reverse: 5′-TGAGGGCACAAGAAGCCCCT-3′. The amplification 176 bp products (20 µl) were digested by 10 U of restriction endonuclease Alw261at 37°C over night. The presence of restriction site resulting in two fragments (91 and 85 bp) indicated mutant allele (Val/Val), while if Ile/Val polymorphism incurred, it resulted in one more fragment of 176 bp.

GSTA1 C-69T polymorphism was determined by polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) according to Coles et al [16]. The primers used were GSTA1 C-69T forward:5′-TGTTGATTGTTTGCCTGAAATT-3′ and GSTA1 C-69T reverse: 5′-GTTAAACGCTGTCACCCGTCCT-3′. The amplification 481 bp products (20 µl) were digested by 10 U of restriction endonuclease Ear1 at 37°C over night. The presence of restriction site resulting in two fragments (385 and 96 bp) indicated mutant allele (B/B) and if A/B polymorphism incurred, it resulted in one more fragment of 481 bp.

All genotyping was performed by laboratory personnel blinded to case-control status, and blinded quality control samples were inserted to validate genotyping identification procedures; concordance for blinded samples was 100%.

Statistical analysis

The distribution of the GSTA1 and GSTP1 polymorphisms for the case and control populations was tested for the Hardy–Weinberg equilibrium by χ2 test. As a measure of effect size, odds ratio (OR) with corresponding 95% confidence interval (95%CI) was used to describe the strength of association between the genotypes and bladder cancer modified by occupational exposure. Unconditional logistic regression analysis is applied. Bearing in mind that age and smoking are well established risk factors for bladder cancer, we adjusted OR by these variables as potential confounders. Interactions between GST polymorphisms and occupational exposure were included in the logistic regression models and also adjusted by potential confounding variables. The probability level of ≤0.05 was considered statistically significant. For statistical analysis the SPSS 17.0 statistical software package (SPSS Inc, Chicago, IL, USA.) was used.

Results

Table 1 shows selected characteristics of male patients with bladder cancer and their controls. The smoking prevalence among cases was higher (82%) than the prevalence found in controls (66%) with the smokers being at 2.3-fold higher risk for TCC than non-smokers (95% CI = 1.3–4.1,p = 0.005). Furthermore, occupationally exposed men had 3.2 times higher risk for TCC than those unexposed (95% CI = 1.6–6.6, p = 0.001). We observed the significantly higher risk in those men occupationally exposed to organic solvents (OR = 3.4, 95% CI = 1.5–7.3, p = 0.002).

Genotyping was conducted for all recruited patients (Table 2). The GSTA1 and GSTP1 genotype frequencies were in Hardy-Weinberg equilibrium both for cases and controls (p>0.05). The observed genotype frequencies in controls were not significantly different from frequencies previously described among Caucasians. However, the frequency of GSTT1-null genotype in control group (28%) was higher than values reported among Caucasians (18.1%). As shown in Table 2, the frequencies of GST null/low-activity genotypes were higher in cases compared to controls with the exception of the GSTT1-null genotype. Although GST A1, T1 and P1 genotypes did not contribute independently toward the risk of TCC, the GSTM1-null genotype was overrepresented among cases (56%) compared to M1-active genotype with an adjusted OR of 2.1 reaching a statistical significance (95% CI = 1.1–4.2, p = 0.032).

Table 2

GSTA1, GSTM1, GSTT1 and GSTP1 genotypes in relation to bladder cancer risk in male patients.

GST genotype	Cases	Controls	OR (95%CI)	p
	n (%)	n (%)
GSTA1
AA	45 (31)	41 (36)	1.0 (reference group)
AB	81 (57)	54 (47)	1.9 (0.9–4.2)	0.094
BB	17 (12)	19 (17)	1.1 (0.4–2.9)	0.875
AB+BB	98 (69)	73 (64)	1.7 (0.8–3.5)	0.171
GSTM1
active a	63 (44)	58 (51)	1.0 (reference group)
null b	80 (56)	56 (49)	2.1 (1.1–4.2)	0.032
GSTT1
active a	101 (74)	82 (72)	1.0 (reference group)
null b	36 (26)	32 (28)	1.0 (0.5–2.2)	0.999
GSTP1
Ile/Ile	62 (43)	49 (43)	1.0 (reference group)
Ile/Val	65 (46)	48 (42)	0.92 (0.5–1.9)	0.918
Val/Val	16 (11)	17 (15)	0.6 (0.2–1.9)	0.401
Ile/Val+Val/Val	81 (57)	65 (47)	0.9 (0.4–1.7)	0.876

Active (present) if at least one active allele present.

Inactive (null) if no active alleles present. OR- odds ratio adjusted for age and pack-years. CI- confidence interval.

Combined effects of GSTA1, GSTM1, GSTP1 and GSTT1 polymorphisms and occupational exposure on bladder cancer risk in male patients are shown in Table 3. When both cases and controls were dichotomized according to both genotype and occupational exposure, exposed subgroup was at TCC risk regardless of GST genotype. We found that occupationally exposed individuals with GSTT1-active genotype exhibited 4.3-fold increased risk compared to the unexposed T1-active subjects (95% CI = 1.7–10.6, p = 0.002). However, only for the GSTP1 gene is there evidence of a gene–occupational exposure interaction (p = 0.017).

Table 3

Combined effect of occupational exposure and GST genotype on bladder cancer risk in male male patients.

GST/exposure	Cases	Controls	OR (95%CI)	p
	n (%)	n (%)
GSTA1
AA/unexposed	21 (15%)	32 (28%)	1.0 (reference group)
AB+BB/unexposed	56 (39%)	48 (42%)	2.4 (0.8–7.3)	0.121
AA/exposed	24 (17%)	9 (8%)	6.2 (1.4–27.1)	0.015
AB+BB/exposed	42 (29%)	25 (22%)	6.4 (2.0–20.2)	0.002
P interaction between genotype and occupational exposure  = 0.104
GSTM1
active a/unexposed	35 (24%)	44 (39%)	1.0 (reference group)
null b/unexposed	42 (29%)	36 (32%)	3.3 (1.2–9.4)	0.023
active/exposed	28 (20%)	14 (12%)	5.4 (1.9–15.8)	0.002
null/exposed	38 (27%)	20 (17%)	6.0 (2.2–16.5)	0.001
P interaction between genotype and occupational exposure  = 0.601
GSTT1
active a/unexposed	54 (40%)	57 (50%)	1.0 (reference group)
null b/unexposed	22 (16%)	23 (20%)	1.3 (0.5–3.9)	0.577
active/exposed	47 (34%)	25 (22%)	4.3 (1.7–10.6)	0.002
null/exposed	14 (10%)	9 (8%)	2.6 (0.8–8.9)	0.124
P interaction between genotype and occupational exposure  = 0.770
GSTP1
Ile/Ile/unexposed	31 (22%)	32 (28%)	1.0 (reference group)
Ile/Val+Val/Val/unexposed	46 (32%)	48 (42%)	0.8 (0.3–2.1)	0.605
Ile/Ile/exposed	31 (22%)	17 (15%)	2.8 (1.0–7.9)	0.049
Ile/Val+Val/Val/exposed	35 (24%)	17 (15%)	2.8 (1.0–8.0)	0.049
P interaction between genotype and occupational exposure  = 0.017

Active (present) if at least one active allele present.

Inactive (null) if no active alleles present. OR- odds ratio adjusted for age and pack-years. CI- confidence interval.

In order to test whether GST-occupational exposure interaction is modified by the specific type of exposure, cases and controls were further stratified into exposed to solvents and exposed to pesticides. Combined effect of occupational exposure to solvents and GST genotype on bladder cancer risk in male patients is shown on Table 4. The results of gene-occupational exposure to solvents interaction analyses indicated a significant effect between occupational exposure to solvents and all common GST polymorphisms tested. The most pronounced effect regarding occupational exposure to solvents and GST genotype on bladder cancer risk was observed for the GSTA1 genotype, since men exposed to solvents with GSTA1-low activity genotype had 9 times higher risk of bladder cancer than GSTA1-active unexposed men (95% CI = 2.4–34.7, p = 0.001). Similarly to that observed for GSTA1-low activity, the GSTM1-null genotype enhanced the risk of TCC among subjects exposed to solvents compared to the unexposed GSTM1-active individuals (OR = 6.5, 95% CI = 2.1–19.7, p = 0.001). These results point to the importance of antioxidant GSTA1 and GSTM1 activity protection against free radicals produced during solvent metabolism. The risk of TCC development was 5.3–fold elevated among GSTT1-active patients exposed to solvents in comparison with GSTT1-active unexposed patients (95% CI = 1.9–15.1, p = 0.002). Significant association was also found for GSTP1 Ile/Ile individuals who had 3.3 higher TCC risk compared to the unexposed Ile/Ile individuals (95% CI = 1.0–10.8, p = 0.047). However, only for GSTP1 statistically significant interaction between genotype and occupational exposure to solvents was found (p = 0.044)

Table 4

Combined effect of occupational exposure to solvents and GST genotype on bladder cancer risk in male patients.

GST/exposure	Cases	Controls	OR (95%CI)	p
	n (%)	n (%)
GSTA1
AA/unexposed	21 (1%)	32 (32%)	1.0 (reference group)
AB+BB/unexposed	56 (46%)	48 (49%)	2.4 (0.8–7.3)	0.121
AA/solvents	14 (11%)	6 (6%)	5.9 (1.0–33.1)	0.046
AB+BB/solvents	31 (25%)	13 (13%)	9.2 (2.4–34.7)	0.001
P interaction between genotype and occupational exposure to solvents  = 0.228
GSTM1
active a/unexposed	35 (28%)	44 (43%)	1.0 (reference group)
null b/unexposed	42 (34%)	36 (35%)	3.3 (1.2–9.4)	0.023
active/solvents	21 (17%)	10 (10%)	4.7 (1.6–13.8)	0.006
null/solvents	27 (22%)	12 (12%)	6.5 (2.1–19.7)	0.001
P interaction between genotype and occupational exposure to solvents  = 0.896
GSTT1
active a/unexposed	54 (46%)	57 (56%)	1.0 (reference group)
null b/unexposed	22 (18%)	23 (22%)	1.3 (0.5–3.9)	0.577
active/solvents	34 (29%)	15 (15%)	5.3 (1.9–15.1)	0.002
null/solvents	8 (7%)	7 (7%)	1.7 (0.4–7.3)	0.470
P interaction between genotype and occupational exposure to solvents  = 0.224
GSTP1
Ile/Ile/unexposed	31 (25%)	32 (31%)	1.0 (reference group)
Ile/Val+Val/Val/unexposed	46 (37%)	48 (47%)	0.8 (0.3–2.1)	0.605
Ile/Ile/solvents	22 (18%)	9 (9%)	3.3 (1.0–10.8)	0.047
Ile/Val+Val/Val/solvents	26 (21%)	13 (13%)	2.6 (0.9–7.9)	0.089
P interaction between genotype and total occupational exposure to solvents  = 0.044

Active (present) if at least one active allele present.

Inactive (null) if no active alleles present. OR- odds ratio adjusted for age and pack years; CI- confidence interval.

Combined effect of occupational exposure to pesticides and GST genotype on bladder cancer risk in male patients is shown on Table 5. Men with GSTT1-active genotype exposed to pesticides exhibited 4.5 times higher risk in comparison with unexposed GSTT1-active subjects (95% CI = 0.9-22.5, p = 0.067).

Table 5

Combined effect of occupational exposure to pesticides and GST genotype on bladder cancer risk in male patients.

GST/exposure	Cases	Controls	OR (95%CI)	p
	n (%)	n (%)
GSTA1
AA/unexposed	21 (22%)	32 (36%)	1.0 (reference group)
AB+BB/unexposed	56 (60%)	48 (54%)	2.4 (0.8–7.3)	0.121
AA/pesticides	8 (9%)	3 (3%)	4.2 (0.5–36.0)	0.190
AB+BB/pesticides	8 (9%)	6 (7%)	2.0 (0.5–7.9)	0.239
P interaction between genotype and occupational exposure to pesticides  = 0.957
GSTM1
active a/unexposed	35 (37%)	44 (49%)	1.0 (reference group)
null b/unexposed	42 (45%)	36 (41%)	3.3 (1.2–9.4)	0.023
active/pesticides	7 (8%)	3 (3%)	2.9 (0.7–12.2)	0.138
null/pesticides	9 (10%)	6 (7%)	1.9 (0.5–6.7)	0.264
P interaction between genotype and occupational exposure to pesticides  = 0.125
GSTT1
active a/unexposed	54 (59%)	57 (64%)	1.0 (reference group)
null b/unexposed	22 (24%)	23 (25%)	1.3 (0.5–3.9)	0.577
active/pesticides	11 (12%)	7 (8%)	4.5 (0.9–22.5)	0.067
null/pesticides	5 (5%)	2 (3%)	2.6 (0.4–20.6)	0.264
P interaction between genotype and occupational exposure to pesticides  = 0.508
GSTP1
Ile/Ile/unexposed	31 (33%)	32 (36%)	1.0 (reference group)
Ile/Val+Val/Val/unexposed	46 (49%)	48 (53%)	0.8 (0.3–2.1)	0.605
Ile/Ile/pesticides	9 (10%)	6 (7%)	2.9 (0.6–13.6)	0.181
Ile/Val+Val/Val/pesticides	7 (8%)	3 (4%)	2.4 (0.5–10.1)	0.231
P interaction between genotype and occupational exposure to pesticides  = 0.320

Active (present) if at least one active allele present.

Inactive (null) if no active alleles present. OR- odds ratio adjusted for age and pack years. CI- confidence interval.

Discussion

Our results showed that occupationally exposed men had 3 times higher risk for TCC. This result confirms the occupational exposure as a TCC risk factor [4]. Furthermore, the analysis of gene-occupational exposure interaction indicated a significant effect between occupationally exposed men and GSTP1 polymorphism. GSTP1 seems to play a role of particular importance in the detoxification of inhaled toxicants in occupationally exposed individuals since it is the most abundant GST isoform in the lung [19]. The mutated GSTP1 seems to be less effective in detoxification than the wild genotype [20]. Thus, Heuser et al. [18] showed that the mutated genotype (Ile/Val or Val/Val) was associated with greater DNA damage in Brazilian footwear workers than the wild (Ile/Ile) genotype [21]. These studies point to an interaction between the exposure and GSTP1 genotype. In our study, the most significant TCC risk was found for solvents. Epidemiologic evidence on the relationship between solvents and various cancers, such as gastrointestinal cancers, lung cancer and lymphohematopoietic malignancies, is well established [22]. Among compounds that have carcinogenic role halogenic aliphatic solvents have been mostly described. There are few reports about relationship between urinary bladder risk and solvents. Previous case-control studies reported significantly increased risks (between 3.1 and 8.8 times) among workers in the dyestuffs industry [23], [24]. Several other investigators have reported elevated risks for spray painters [25], [26], who have been reported to be exposed to many known or suspected carcinogens, including solvents. On the other hand, Lohi and others [27] found that among Finnish workers exposure to solvents was positively associated with the incidence of bladder cancer in women, but not in men.

It is important to note that risk imposed by occupational hazards was modified by GST polymorphism. We observed that individuals occupationally exposed to solvents with at least one low activity GSTA1 allele had the highest risk (about 9 times), while GSTM1-null carriers had 6.5 times higher bladder cancer risk when compared to unexposed GSTA1 AA and GSTM1-active persons, respectively. This result was expected since in several malignant diseases, such as colorectal, prostate and hepatocellular cancer, GSTA1*B allele with lower transcriptional activity was associated with increased risk. GSTA1 protein belongs to the most promiscuous GSTs that acts upon a broad range of substrates which bind to its active site [28]. Our findings that low-activity GSTA1 and GSTM1-null genotype increase susceptibility to bladder cancer in occupationally exposed men can be explained by the role of GST enzymes in detoxification and in antioxidant defense. Namely, GSTA1 and GSTM1 possess strong peroxidase activity and are key components in cellular defense against free radicals [29]. It may be speculated that free radicals are produced during solvent metabolism [30]. Regarding potential place of solvent detoxification, it is important to note that uroepithelial cells do not express GSTA1, while their GSTM1 protein level is also relatively low [31]. On the other hand, liver cells abundantly express GSTA1 and GSTM1 and thus participate in GSTA1 and GSTM1 mediated conjugation of different metabolites with glutathione, thereby enhancing their excretion in urine [32]. Taken together, these data suggest that liver, by its GSTs conjugating and peroxidase activity plays a key role in protection against bladder carcinogens present in halogenated solvents. On the other hand, GSTT1-active individuals occupationally exposed to solvents exhibited 5 times higher risk of TCC in comparison with GSTT1-active unexposed subjects. These results are biologically plausible since GST-mediated conjugation with halogenated substrates may lead to a more toxic or mutagenic metabolites. Namely, substrates with ≥2 halogenes are activated because the conjugated product is instable, leading to reactions with nucleophiles, particularly DNA and proteins [33]. The human polymorphic GSTT1 catalyze conjugation of halomethanes, dihalomethanes, ethylene oxide and a number of other industrial compounds. Our results confirm the assumption of Avima M Ruder et al. [34] that humans with fully functional GST genes produce enzymes that metabolize some solvents to cytotoxic metabolites; while those with less functional or nonfunctioning genes have little or no enzyme and apparently do not produce cytotoxic metabolites from solvent exposure. Until now, the association between GST polymorphism and occupationally related cancers has been studied mostly in renal cell carcinoma. Results of these studies showed that GSTT1-active genotype enhanced the risk of renal cell carcinoma among subjects exposed to solvents. Our results on higher bladder carcinoma risk in GSTT1-active individuals occupationaly exposed to solvents are in accordance with previously published results in renal cell carcinoma [35], [36]. Regarding the potential mechanism of solvent metabolism by GST, it is generally assumed that the main site is liver, followed by a mandatory transfer of conjugates to the kidney. However, the initial bioactivation step of halogenated solvents, can take place in the kidney itself [37]. Uroepithelium is also capable of metabolizing some procarcinogens to inactive or genotoxic metabolites, and is, therefore, not exposed only to preformed reactive metabolites in the urine [38]. As the renal parenchyma and uroepithelium are exposed to the same broad range of potentially genotoxic compounds, the potential genotoxicity of carcinogens also depends on the biotransformation capacity of these tissues. As a result of GST polymorphism, great interindividual differences in GST isoenzyme profiles exist, in both renal parenchyma and uroepithelial cells [37].

Although it has been postulated that exposure to pesticides and/or fertilizers might be responsible for higher urinary bladder risk, the evidence is still conflicting. Some studies have shown that TCC risk was significantly elevated among men in the landscape and horticultural services industry, as well as in gardeners, and lawn care service employees [39], [40], while others did not [41]. Some suggestions of a possible relation between GST status and early markers of genotoxic effects in humans exposed to pesticides are available. An increased frequency of micronuclei in cultured peripheral lymphocytes has been found among pesticide exposed greenhouse workers with the GSTM1-active genotype [42]. Significantly higher levels of sister hromatid exchanges were also found among GSTT1-active individuals exposed to pesticides when compared to GSTT1-null workers similarly exposed [43]. Until now only one study investigated association between GST polymorphism and occupational exposure to pesticide with respect to risk of carcinoma of urinary tract. Namely, Karami and others reported that renal cell carcinoma risk associated with pesticide exposure was highest among individuals with active GSTM1/T1 genotypes [44]. Although we did not observe significant effect between exposure to pesticides and GST polymorphisms we found borderline significance for GSTT1-active genotype. One of the reasons for non-significant association between GSTT1-active genotype may be the relatively small number of pesticide exposed participants in both case and control groups. Nevertheless, it is well known that pesticides produced from halogenated alkanes, alkenes undergo bioactivation in the liver and kidney after conjugation to glutathione by GSTT1 [41]. Therefore, an active GSTT1 enzyme will be required to conjugate substrates and form more reactive intermediates that directly damage tissues. Conversely, the deleted variant of GSTT1-genotype will form an inactive enzyme and therefore metabolism of halogenated compounds will occur through oxidation, without formation of reactive intermediates [44].

The principal limitations of this study are the relatively small sample size which limiting the precision of the odds ratios, hospital-based control group and qualitative evaluation of occupational exposure. Concerning the actual sample size (143 cases and 114 controls), the statistical power is 66%. Furthermore, it is well known that relatively small numbers of both study participants and GST polymorphisms studied might be sources of potential biases which may influence the study findings. However, we tested effects of four GST polymorphisms and occupational exposure on bladder cancer risk and therefore significantly decreased chance for publication bias. Additionally, we cannot entirely rule out the possibility that some of our results could be caused by confounding, although we included only men and adjusted all results by age and smoking status. Further studies with larger samples and more rigorous designs are needed to investigate the gene effects and the potential effect modification by environmental factors.

Conclusions

GSTM1-null genotype increased the risk of bladder cancer in males. Null or low-activity genotypes of the GSTA1, GSTT1, and GSTP1 did not contribute independently towards the risk of bladder cancer in males. However, in association with occupational exposure, both low activity GSTA1 and GSTM1-null genotype increase individual susceptibility to bladder cancer suggesting the protective role of these detoxification and antioxidant enzymes in metabolism of occupational hazards, specifically organic solvents. On the other hand, the presence of GSTT1-active genotype in occupationally exposed subjects, resulting in GSTT1 protein expression and GSTT1 mediated bioactivation, increases the risk of bladder cancer.