Introduction

Prostate cancer is one of the most commonly diagnosed malignancies in men [1]. It is disproportionately common among individuals of African descent, irrespective of the place where they live in the world, and less common in Caucasian and Asian populations [2]. The reasons for these ethnic differences in incidence are largely unknown but probably involve a complex interplay between hormonal, environmental and genetic factors [2], [3].

Genetic polymorphism of metabolic enzymes has been studied to investigate the possible etiological role of carcinogenic agents in prostate cancer [4]. Promising candidates include the glutathione S-transferase (GSTs) family, a group of phase II detoxifying enzymes that catalyze reactions between cytosolic glutathione and electrophilic substrates, producing stable and more soluble compounds that can then be excreted or compartmentalized [5].

It was assumed that GST-mediated conjugation resulted in less toxic or inactive metabolites. For two GST genes, GSTM1 (GenBank: BC024005.2) and GSTT1 (GenBank: BC007065.1), the complete deletion of the gene eliminates the gene function, leading to the inability to eliminate electrophilic compounds as efficiently; this may potentiate the deleterious effects of various environmental and endogenous carcinogens [6]. The main role of GSTs is the conjugation of reactive metabolites, but they may also be involved in producing reactive derivatives from the metabolism of halogenated compounds such as haloalkanes [7],[8]. Thus, the final direction of the effect - protection or susceptibility - on carcinogenesis, if any, is difficult to predict, and exposure to particular chemicals may alter the effect of GSTs on cancer risk.

Extensive research has been carried out, mostly in Caucasian and Asian populations, studying the relationship between GSTM1 and GSTT1 polymorphisms and prostate cancer risk. However, the results obtained are conflicting, and the putative associations identified remain a matter of debate [9]–[13]. A recent multi-institutional case-control study including 10 studies (1715 cases and 2363 controls) among subjects of African descent (African-American, African-Caribbean and African) has provided new data for this ethnic group [14]: homozygous deletion of GSTM1 or of GSTT1 was found to be inversely associated with the risk of prostate cancer. This suggests a paradoxical inverse effect of the loss of GST function on the risk of prostate cancer in these populations.

A major limitation of these studies is that they could not discriminate between heterozygous and homozygous carriers of the non-deleted alleles. Classifying such genotypes as carrier or non-carrier implies a recessive model (one or two copies versus the absence of the risk allele), which may not reflect the true underlying genetic mechanisms involved and thus may not provide a valid or accurate estimate of the genetic risk [15]. Moreover, the GSTM1 and GSTT1 genes exhibit copy number variation (CNV), and a dose effect between gene copy number and enzymatic activity has been reported for both genes [16]–[19]. As a consequence, analyses based on gene dose are likely to provide a better description of any association with disease outcome [20].

We recently described a population-based case-control study in Guadeloupe among a Caribbean population of African descent showing that homozygous deletions of GSTM1 and those of GSTT1 are each, independently, significantly associated with a reduced risk of prostate cancer [14]. Here, we report the continuation of this case-control study by determining the exact gene copy numbers of GSTM1 and GSTT1 genes and investigating the associations between the CNV of each gene and the risk of prostate cancer.

Materials and Methods

Statistical Analysis

The odds ratio (OR) and 95% confidence intervals (CIs) for the association between CNV of GSTT1 and GSTM1 genes and prostate cancer were estimated using unconditional logistic regression. Genotypes were coded as categorical variables (0, 1, 2, 3 or more gene copies for individual GSTT1 or GSTM1 analysis; 0, 1, 2, 3, 4, 5 or more copies for the sum of GSTM1 and GSTT1 genes). We investigated whether covariates were confounding factors by looking at the association between covariates and subject status (Table 1), as well as that between covariates and exposure according to the carrier status for GSTM1, GSTT1 and the combined GSTM1/GSTT1 copy numbers (Tables S2, S3 and S4 respectively). Confounding covariates included in the logistic model if they were associated (P<0.20) with both subject status and exposure. Age was always included in the adjusted model. Because log linearity of age was not achieved, age was categorized as quartiles according to the age distribution of the controls. Tests for trends in risk were performed by entering the categorical variable into the model as an ordinal variable. Missing data for covariates varied from none to 3 (0.2%) for PSA screening history, 6 (0.5%) for smoking, 12 (1.0%) for alcohol, and 34 (2.7%) for education. Missing data were handled by substituting them by a missing value indicator variable. We also considered possible interactions between CNV of GSTT1 and GSTM1 genes and selected covariates (smoking, alcohol consumption, BMI, family history of prostate cancer) in relation to the risk of prostate cancer. The P value for interaction was calculated by the likelihood ratio test comparing the log-likelihood for the model with the interaction terms to the log-likelihood for the model without the interaction term. SAS software version 9.3 (SAS Institute, Cary, NC) was used for all statistical analyses. All tests were two-tailed, and P values less than 0.05 were considered significant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of the study population.

https://doi.org/10.1371/journal.pone.0107275.t001

Results

The general characteristics of the study participants and the frequencies of the numbers of gene copies of GSTM1 and GSTT1 among the 629 cases and 622 controls are summarized in Table 1.

The frequency of homozygous deletion (0 copies) of GSTM1 gene in the control group was 0.32 (95% CI, 0.28–0.35) and that of GSTT1 was 0.31 (95% CI, 0.27–0.34). The frequency of homozygous deletions at both loci in the control group was 0.07 (95% CI, 0.05–0.09).

Crude and adjusted ORs for prostate cancer according to the copy number of GSTM1 and GSTT1 genes are given in Table 2. According to the adjusted model GSTM1 carriers were at higher risk of prostate cancer (OR: 1.31, 95% CI: 1.01–1.71) than men with the homozygous GSTM1 deletion. Men with at least 3 GSTM1 genes were at a higher, but not significantly higher, risk of prostate cancer (OR: 2.55, 95% CI: 0.78–8.39); the trend for GSTM1 gene dose association was inconclusive (P_trend 0.17). According to adjusted models, GSTT1 carriers were at a higher risk of prostate cancer (OR: 1.40, 95% CI: 1.07–1.83) than men with the homozygous GSTT1 deletion. Men with at least 2 and those with 3 or more GSTT1 genes were at significantly increased risk of prostate cancer (OR: 1.55, 95% CI: 1.11–2.16 and OR: 4.89, 95% CI: 1.71–13.99, respectively), and there was a significant gene dose relationship of the GSTT1 gene (P_trend<0.001). Using subjects with homozygous deletion of both GSTM1 and GSTT1 genes as the reference group, adjusted models indicated that subjects carrying both genes were at significantly higher risk of prostate cancer (OR: 1.88, 95% CI: 1.11–3.21). Finally, men with 3, 4, and 5 or more copies of GSTM1 and GSTT1 genes were at significantly increased risk of prostate cancer (OR: 2.18, 95% CI: 1.21–3.91, OR: 3.24, 95% CI, 1.63–6.46, and OR: 5.77, 96% CI: 1.40–23.8, respectively) and there was a significant gene dose relationship (P_trend<0.001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association between copy number of the GSTM1 and GSTT1 genes and risk of prostate cancer.

https://doi.org/10.1371/journal.pone.0107275.t002

An analysis of the interaction between smoking, alcohol consumption, body mass index, family history of prostate cancer and GSTM1, GSTT1, and the sum of GSTM1/GSTT1 copy numbers is shown in Tables 3 to 5. The interaction between GSTT1 copy number and smoking status resulted in a tendency towards a higher risk of prostate cancer risk only in individuals who had never smoked (Table 4). By contrast, the interaction between GSTM1 or the sum of GSTM1/GSTT1 copy numbers and alcohol consumption resulted in a tendency towards an increase in the risk of prostate cancer restricted to former or current drinkers (Tables 3 and 5). However, these interactions were not significant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Association between copy number of the GSTM1 gene and risk of prostate cancer according to smoking, alcohol consumption, body mass index and family history of prostate cancer.

https://doi.org/10.1371/journal.pone.0107275.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Association between copy number of the GSTT1 gene and risk of prostate cancer according to smoking, alcohol consumption, body mass index and family history of prostate cancer.

https://doi.org/10.1371/journal.pone.0107275.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Association between copy number of the sum of GSTM1/GSTT1 genes and risk of prostate cancer according to smoking, alcohol consumption, body mass index and family history of prostate cancer.

https://doi.org/10.1371/journal.pone.0107275.t005

Discussion

To the best of our knowledge, this is the first study reporting the contribution of exact copy number variation of GSTM1 and GSTT1 genes to prostate cancer susceptibility in a population of African descent.

Our results indicate: first, that higher copy number of GSTM1 tends to be associated, although not significantly, with an increased risk of prostate cancer; and second, that higher copy number of GSTT1 is significantly associated with an increased risk of the disease. Furthermore, a higher combined GSTM1 and GSTT1 copy number appears to be significantly associated with an increased risk of prostate cancer.

We detected more than two copies of GSTM1 or GSTT1 genes in 3.4% of our study population: 1.2% for GSTM1 and 2.2% for GSTT1. Evidence for duplication of GSTM1 or GSTT1 has been reported in Caucasian populations but at a substantially lower prevalence than in our study population. Among 10271 Danish subjects, only 24 individuals (0.2%) carried more than two copies of GSTM1 or GSTT1 [23]; in two studies of German subjects including 1320 [24] and 3602 [25] individuals, the frequency of GSTM1 duplication was between 0.08% and 0%, and the frequency of GSTT1 duplication 0% and 0.14%, respectively.

Previous association studies based on comparing carriers (irrespective of the copy number of the genes) and homozygous non-carriers of GSTM1 and GSTT1 genes with prostate cancer have given inconsistent results, leading to divergent conclusions [9]–[14]. There are several possible reasons for these discrepancies: differences in ethnic background, geographic origin, and/or in the environment of the populations studied; different definitions of control groups; and both small numbers of cases included and small effects of the genes leading to a lack of power. The most recent meta-analysis, mostly grouping studies conducted with Caucasian and Asian men, suggest that homozygous deletion of the GSTM1 and GSTT1 genes is associated with increased risk of prostate cancer [11]–[13]. Only one previous study assessed the relationship between the GSTM1 and GSTT1 copy numbers and prostate cancer [23]: it was a prospective study (The Copenhagen City Heart Study) of a Caucasian population of Danish descent and included 128 cases of prostate cancer. Lower GSTT1 copy number was significantly associated with increasing cumulative incidence of prostate cancer and decreasing cumulative 5-year survival; no association was found with GSTM1 copy number. Not only did our study not confirm these findings, but it even leads to the opposite conclusion for this population of African descent [14]: there was a positive association between GSTM1 or GSTT1 copy number and the risk of prostate cancer. These associations appear to be significant in among African-Caribbean and native African populations [14]. Our analysis of a Caribbean population of African descent suggests a relationship between the risk of prostate cancer and GSTM1 gene copy number and shows a significant relationship between the risk and GSTT1 gene copy number; consequently, it provides further evidence for the previously described inverse association between deletions of both GST genes and the risk of prostate cancer in men of African descent. These discrepancies between the findings of different studies remain unclear, but several explanations can be suggested.

The conjugation of glutathione prevent damage resulting from exposure to toxic chemicals and to normal oxidative products of cellular metabolism, the association between GSTM1 and GSTT1 copy number and increased risk of cancer is, at first glance, unexpected. However, GSTs and particularly GSTT1 may also be involved in producing reactive derivatives with higher reactivity [7], [8]. Indeed, the increase in risk associated with higher GST gene copy number, for which a dose effect on the concentration of both enzymes has been shown [16]–[19], is consistent with the idea that functional GST alleles may result in the production of genotoxic metabolites from particular endogenous or environmental agents. It is entirely plausible that the men included in our study have been exposed to chemicals that are “activated” by GSTM1 and/or GSTT1 rather than “detoxified”. In such a scenario, carriers of large numbers of copies of the GST gene would generate and expose cells to higher concentrations of genotoxic products, thereby increasing the likelihood of carcinogenesis promotion. Alternatively, it is also possible that exogenous protective factors are substrates of GST. Were this to be the case, these compounds would be more rapidly eliminated through conjugation in GSTM1- and/or GSTT1-carriers, reducing the bioavailability of the putative anti-cancer compound.

Following stratification of the data for smoking status, the significant increase in prostate cancer risk associated with GSTT1 copy number was found to be restricted to individuals who had never smoked. This finding is consistent with a previous observation of an association between homozygous GSTT1 deletion and a lower risk of prostate cancer in Caribbean men who had never smoked [14]. Such observations contrast with those reported for African-American men, in whom homozygous GSTM1 deletion was associated with an increase in the risk of prostate cancer among smokers, whereas homozygous GSTT1 deletion was not [14]. The observed differences in such associations between Caribbean and African-American men suggest possible differences in various factors between these populations, including lifetime exposure to other carcinogens that saturate the GST system, different levels of tobacco consumption and possible differences in the composition of cigarettes between countries.

We found that current or former drinkers with more than two GSTM1 or combined GSTM1/GSTT1 gene copies had a greater risk of prostate cancer than teetotalers. No reports have yet been published concerning this genetic s × environment interaction and prostate cancer risk. Moreover, toxicological studies carried out in vivo and in vitro have reported conflicting results concerning the possible effects of alcohol on GST expression [26] and most epidemiologic studies have suggested that neither the amount nor the type of alcohol is clearly associated with a risk of developing prostate cancer [27], [28]. These observations should therefore be interpreted with caution until more data become available.

Guadeloupe is a Caribbean department of France characterized by the adoption of a Western lifestyle, including, in particular, eating habits that may be risk factors for prostate cancer [2], [3]. Moreover, since the middle of the 20th century, intensive banana farming in Guadeloupe led to the use of large amounts of chlordecone, an organochlorine insecticide, which has since been banned. This pesticide undergoes no significant biotic or abiotic degradation in the environment, so permanently polluted soils and waters remain a major source of contamination of foodstuffs, such that human beings continue to be exposed to this chemical [29]. Chlordecone is a potential carcinogen and has been associated with increased risk of prostate cancer in Guadeloupe [21]. Nevertheless, it is difficult to suggest that the exposure of our population to environmental agents is such that it can cause the association between prostate cancer and GSTs in the direction we observe.

We cannot exclude the possibility that our findings are due to the ethnic background of our population if a common inherited genetic trait is simultaneously associated with the disease and with a factor that may alters the balance between bioactivation and detoxification in the body. GST enzymes are involved in steroid metabolism, which is thought to be involved in the initiation or progression of prostate cancer [2], [3]. It has already been suggested that ethnic differences in the incidence of prostate cancer may be related to differences in lifelong androgen or estrogen exposure [2].

We are aware of the inherent limitations of patient-control studies. Several factors potentially generating bias must be considered, particularly those relating to differential errors in the measurement of disease or exposure. Patient identification was based on unequivocal histologic criteria and controls were selected on the basis of strict criteria, including normal findings on digital rectal examination and PSA in the normal range for age, taking the ethnic origin of the population into account. However, we cannot exclude the possibility that some control individuals had latent disease that was not detected by PSA analysis or digital rectal examination. However, undetected prostate cancer in control subjects would be expected to bias estimates toward the null hypothesis, so the positive association observed may be an underestimate. We recruited incident rather than prevalent patients, and controls were selected from a representative sample of the male Guadeloupean population during the study period. Differential misclassification of the GST genotypes with respect to case status is unlikely because the staff responsible for genotyping were blind to the case/control status of the subjects. Ethnic identification is always difficult and misclassification can never be excluded, particularly because mixed ancestry is very likely. At least 90% of the inhabitants of Guadeloupe are descended from slaves and immigrants from West and Central Africa. The remaining 10% of the population are descended from Indian immigrants during the XIX century, from more recent immigrants from the middle-east or Europeans. Our selection criteria, including only subjects whose parents were born in French West Indies or on any Caribbean island with a population of predominantly African descent (Haiti, Dominica), gave us some confidence in the homogeneity of our study population. Moreover, the frequency of homozygous deletions of GSTM1 and GSTT1 in our control population is consistent with that previously reported for Afro-American, Afro-Caribbean, Native African and Brazilian men of African descent [14], [30], [31]. Nevertheless, we cannot exclude the possibility that some unknown confounding factors remain that may account for the associations observed or that they are chance findings.

In summary, our study suggests that copy number of GSTT1 and combined GSTM1/GSTT1 copy number are associated to prostate cancer risk in men of African descent with gene dose relationship. Replication of these observations in other populations and mechanistic studies are needed before any causal link can be established.

Supporting Information

Acknowledgments

We thank H. Delacroix-Maillard and the staff of the Centre d’Examens de Santé Sainte Genevieve for recruiting subjects.