¹Laboratory of Chemical Carcinogenesis and Pharmacogenetics, Faculty of Medicine, Biomedical Science Institute (ICBM), University of Chile, Independencia 1027, Santiago, Chile ²Department of Urology, National Cancer Corporation, Chile, Capell Abarz 027, Providencia, Santiago, Chile ³Epidemiology Division, School of Public Health, Faculty of Medicine, University of Chile, Independencia 939, Santiago, Chile

Correspondence to:Dr Dante D. Cáeres, Epidemiology Division, School of Public Health, University of Chile, Independencia 939, Santiago, Chile.
Tel: +56-2-678-6396, Fax: +56-2-737-7121
E-mail: dcaceres@med.uchile.cl
Received 2005-10-06 Accepted 2005-12-23

1 Introduction

Prostate cancer (PCa) is one of the most frequent malignant neoplasms in men; however, the rate of this disease shows remarkable worldwide variation. Many studies indicate that environmental and genetic factors play a significant role in the etiology of this disease [1,2]. Given the multicausal etiology of PCa, synergistic interactions among genetic and other risk factors might have significant effects on PCa risk, especially gene-gene (G×G) and gene-environment (G×E) interactions. It is well known that in the carcinogenic process there are multiple points at which genetically-determined host characteristics and/or environmental factors might influence individual susceptibility, through effects on metabolic activation, DNA-repair capacity, and other cellular processes. Polymorphic genes implicated in cancer etio-logy can have significant effects on increasing or redu-cing differential susceptibility to environmental cancer [3].

The p53 gene is one of the most mutated tumor-suppressor genes in human neoplasms, and it has been referred as the "emergency brake gene" because of its tumor-preventing apoptotic and cell-cycle-checkpoint functions in physiologically stressful situations [4]. The wild type p53 gene polymorphism at codon 72 (p53cd72) produces a protein with an arginine (Arg: CGC) or proline (Pro: CCC) genotype. This polymorphism related to changes in the function of the p53 protein is strongly associated with the tumor formation process. The wild-type p53 gene suppresses cellular transformation with activated oncogenes, therefore inhibiting the growth of malignant cells [5].

In contrast, human cytochrome P450 (CYP) phase I enzymes function in a wide variety of metabolic pathways involving endogenous and exogenous compounds, such as steroids and environmental xenobiotics. The CYP1A1 gene encodes by benzo(a)pyrene hydroxylase and it is primarily expressed in the liver but has also been detected in prostate tissue [6]. CYP1A1 activates benzo-(a)pyrene into epoxides and phenolic products that are mutagenic and carcinogenic; therefore, higher catalytic activity might predispose patients to cancer risk by increasing carcinogenic compounds in target sites such as the prostate and lung tissues [6, 7]. Three restriction fragment length polymorphic (RFLP) variants have received the most attention: MspI RFLP (CYP1A1*2A), MspI RFLP (CYP1A1*2C) and (CYP1A1*3). The rare Val and M2 alleles of the CYP1A1 gene might increase individual cancer risk by heightening aryl hydrocarbon hydroxylase (AHH) enzyme inducibility [8]. Alternatively, among the phase II enzymes involved in the metabolism of xenobiotic compounds the GST family catalyze the conjugation of glutathione to numerous potentially genotoxic compounds, including aliphatic aromatic he- terocyclic radicals, epoxides, or arene oxides. Individual differences in the detoxification of reactive chemicals through the GST pathways are frequently the result of deletion of the GST genes, particularly GSTM1 and GSTT1 [9]. Individuals who have inherited susceptible variants (homozygous deletions of GSTM1 or GSTT1) of the metabolizing genes might have an increased body burden of reactive metabolites from cigarette smoke, causing increased risk of the development of PCa [10, 11]. However, contradictory findings have been reported in recent studies [12-15].

It is possible that individual variations in biotransformation activities on both phase I and phase II enzymes in coordination with p53 activity regulate the effect of DNA toxic metabolites and might be partially responsible for host susceptibility to chemical exposure, which is related to PCa.

Our overall aim in the present study is to assess the role of several genetic factors in combination with an environment factor as modulators of PCa risk by focusing on allele variants of low-penetrance genes associated with cell control and detoxification processes, and smoking. We determine the relationships between GSTM1 deletion, Msp1 CYP1A1 polymorphism, p53cd72 polymorphism and smoking on PCa risk.

2 Materials and methods

2.1 Sample subjects

Cases for this study were recruited in a voluntary screening performed in Santiago’s Metropolitan Area, Chile, by the National Cancer Corporation. Prostate specific antigen (PSA) and digital rectal examination (DRE) were carried out by urologists [16]. All people with suspected PCa (PSA=4mg/dL or altered DRE, or both) were biopsied and histologically confirmed. A total of 60 PCa cases and 117 controls were included in the present study. The controls were men attending the respiratory service of the Clinical Hospital of University of Chile, with similar demographic characteristics to the PCa cases. All study subjects provided informed consent for participation in this research under a protocol approved by the Ethics Committee for Studies on Human Beings at the University of Chile.

2.2 Blood samples

Blood samples were collected from all of the participants at time of inclusion. The samples were processed at the Laboratory of Chemical Carcinogenesis and Pharmacogenetics at the Faculty of Medicine of the University of Chile to obtain genomic DNA from peripheral leucocytes using the method of Miller [17]. The genomic DNA was checked for purity at 260/280mm absorption and re-purified with phenol/chloroform protocol if required. DNA was stored at -30°C until use.

2.3 Genotyping methods

After DNA extraction, DNA samples were analyzed for GSTM1, CYP1A1 and p53 genetic polymorphisms. Polymerase chain reaction (PCR) based RFLP was used to examine the polymorphisms of interest. All samples were submitted to separate amplifications followed by digestion with restriction enzymes.

2.4 Polymerase chain reaction (PCR) amplification detection

For the CYP1A1 MspI site, PCR amplification was carried out using previously described primers C44 and C47 yielding a fragment of 340 bp [7]. GSTM1 null variant was determined using primers described by Ambrosone et al. [18] simultaneously with Msp1 primers as internal control for amplification. p53cd72 genetic polymorphism was determined using the primers described by de la Calle-Martin et al. [19]. The different genotypes were observed using 2% agarose or 6% polyacrylamide gel electrophoresis.

2.5 Nomenclature used to genetic polymorphisms

The following nomenclature was used to describe the different polymorphic variants [20]. For the p53cd72 polymorphisms the possible genotypes are Pro/Pro, Arg/Pro and Arg/Arg. For the CYP1A1*2A, the reference allele is called wild type (Wt), and rare allele is called M1. The GSTM1 1*/1* and GSTM1 1*/*2 are referred to as present variant (homozygous and heterozygous), with GSTM1 2*/*2 genotype used to indicate the homozygous null variant. For GST, null and present denominations will be used. From this point forward, M1* and Pro* will be used for Wt/M1, M1/M1, and Pro/Pro and Arg/Pro genotypes, with the objective to increase precision.

2.6 Analytic methodology

Genotype frequencies and 95% confidence interval (CI) for GSTM1 were calculated as the proportion of individuals with a given genotype divided by the total number of participants. For p53cd72 and CYP1A1, allele frequencies and 95% CI were calculated as the number of alleles divided by the number of chromosomes, and the test for Hardy-Weinberg equilibrium was conducted. To explore the possible associations between GSTM1, CYP1A1 and p53cd72 genetic polymorphisms and PCa risk, and to evaluate the putative modification by these genotypes of the effect of smoking, we cross-classified the data using a 2 by 4 table, as described by different authors, for a case-control design [21]. The relationship between these polymorphic genes and smoking and PCa risk was examined using odds ratio (OR), with 95% CI using Woolf’s method in an unconditional logistic model. All associations were evaluated using a priori low-risk bivariate genotype combinations (Arg/Arg, Wt/Wt and GSTM1 present) in non-smokers as a common reference group. Finally, the ORs are presented unadjusted and adjusted by age. All statistical analyses were performed with stata version 7.0 software (STATA Corporation; College Station, TX, USA).

3 Results

Characteristics of participants, genotype and allelic frequencies and OR for these polymorphisms and PCa in the present study are described in Table2. Both the PCa and the control groups had a similar age distribution. Smoking frequency was higher in the PCa cases compared with that in the controls and PCa cases had a significant risk of PCa compared with the controls (OR: 2.59, 95% CI: 1.35-4.95). Allele frequencies for the Pro and M1 allele were higher in the PCa cases compared with that in the controls. We did not find significant differences in genotype frequencies for GSTM1 and CYP1A1 between the PCa cases and the controls subjects. A higher prevalence of Pro/Pro genotype in the PCa cases compared with that in the control subjects was observed. ORs for PCa associated with GSTM1 and CYP1A1 genotypes were close to the null value. For the different genotypes of p53cd72, only Pro/Pro genotype was positively associated with PCa (OR: 2.89, 95% CI: 1.17-7.10). We did not observe significant departures from the Hardy-Weinberg equilibrium from p53cd72 and CYP1A1 genotypes among the PCa cases (P=0.356; P=0.096) or the controls (P=0.621; P=0.706), respectively.

Table2 shows the results of the distribution of the joint effect between gene-gene polymorphisms and smoking risk factor on PCa. The joint age-adjusted OR for smokers carrying Pro* and M1* variants was 13.13 (95% CI: 2.41-71.36) and for non-smokers was 2.25 (95% CI: 0.44-13.48), compared with the reference group (Arg/Arg=Wt/Wt=non-smokers ). In contrast, those smokers and non-smokers carrying only Pro* variants had a lower, non-significant risk (OR: 4.16, 95% CI: 0.75-22.96 and OR: 1.51, 95% CI: 0.21-10.59). However, smokers carrying M1* and Arg/Arg genotypic variants had a higher significant risk compared with the reference group (OR: 8.74, 95% CI: 1.58-48.39), but was not significant in non-smokers (OR: 2.71, 95% CI: 0.48-15.35). The joint age-adjusted OR for smokers carrying Pro* and GSTM1 null polymorphism was 3.97 (95% CI: 1.13-13.95) and for non-smokers was 0.80 (95% CI: 0.19-3.28) compared with the reference group (Arg/Arg=GSTM1 null=non-smokers). Conversely, those smokers carrying Pro* and GSTM1 present variants had a significant risk, 3.07 (95% CI: 1.01-9.37), but non-smokers had a non-significant risk, 0.95 (95% CI: 0.28-3.19). Smokers carrying Arg/Arg and GSTM1 null genotypes showed an increased but non-significant risk compared with the reference group (OR: 4.73, 95% CI: 0.89-5.18). Similarly, the risk for non-smokers was not significant (OR: 0.57, 95% CI: 0.10-3.29). However, the joint effect between the GSTM1 null and CYP1A1 M1* in smokers was significantly associated with PCa risk: 6.87 (95% CI: 1.68-27.97). In contrast, in non-smokers the risk was not significant: 1.37 (95% CI: 0.35-5.46). The age-adjusted OR for smokers carrying the GSTM1 null and Wt/wt genotypes were high but not significant: 2.69 (95% CI: 0.52-14.08). Those subjects carrying GSTM1 present and M1* genotypes had a significant Pca risk for smokers but not for non-smokers (OR: 5.00, 95% CI: 1.47-17.05 and OR: 1.29, 95% CI: 0.36-4.58, respectively).

We performed the test of homogeneity for p53cd72-CYP1A1, p53cd72-GSTM1, GSTM1-CYP1A1 and smoking status, which clearly indicated that the Mantel-Henszel OR for Pca differs depending on whether an individual smoked and on polymorphisms combinations (OR: 3.44, 95% CI: 1.79-6.62; OR: 1.68, 95% CI: 0.98-2.56; OR: 1.86, 95% CI: 1.11-3.11, respectively).

4 Discussion

The p53 Pro allele has recently been reported to be associated with genetically determined susceptibility to smoking-related lung cancer in the Chilean population [22]. Few studies have reported an association of p53cd72 polymorphism with PCa risk [23,24], however the findings have been inconsistent. Phase I enzymes on cytochrome P450 (CYP) is the major enzyme system in xenobiotic metabolism and plays a critical role in metabolic activation of many environmental chemicals. Together with phase II metabolizing enzymes on glutathione S-transferase (GST) are responsible for detoxification process, even though they might also be involved in bioactivation of some carcinogenic compounds. This is true in the case of polycyclic aromatic hydrocarbons (PAH), which are important carcinogenic components of tobacco smoke. Individual variation in the genes encoding these enzymes could be modifying the effect of specific environmental risk factors and, therefore, could influence susceptibilities to cancer [25].

In this study, we used a case-control design to assess the joint effects of p53cd72, CYP1A1, GSTM1 polymorphism and smoking habit on PCa risk. We observed that those subjects who are smokers carrying high-risk genotypic variants have an increased PCa risk compared with non-carrying susceptible variant subjects. In general terms, smoking has a synergistic effect on overall risk, which can be explained by the carcinogenic compounds of cigarette smoke that can be differentially biotransformed by CYP1A1 and/or GSTM1 (e.g. benzo-[a]pyrene).

The increased risks observed for smokers carrying susceptible genotypes of CYP1A1 and p53 (OR: 13.13), GSTM1 and p53 (OR: 3.97), and CYP1A1 and GSTM1 (OR: 6.87) might be explained by the metabolic function of these biotransformation enzymes, which might act in a coordinated but contrary pathway. Whereas CYP1A1 produces the reactive benzo(a)pyrene diol epoxide, which can initiate a tumoral process, GSTM1 detoxifies it by GSH conjugation [6, 26]. However, people who have M1 allele and GSTM1 deletion cannot properly detoxify the carcinogenic metabolites. This situation could be worse if p53 function is decreased or deleted, which apparently occurs with the p53cd72 Pro allele [27]. Our results support this asseveration (OR: 8.87, 95% CI: 1.25-62.71). In contrast, there is evidence that benzo(a)-pyrene diol epoxide is able to inactivate p53 antioncogen, providing indirect evidence of the potential relationship between CYP1A1 and GSTM1 biotransformation enzymes and p53 antioncogen [27]. An interesting additional research hypothesis is related to the role of CYP1A1 and GSTM1 in steroid metabolism, taking account the structural similarities between these hormones and PAH, and the participation of testosterone in prostate cancer. This issue can explain, in part, the observed positive associations with the CYP1A1 gene, even though this topic should be further investigated.

A limitation of the present study is the small numbers of cases of PCa. Hence, it is likely that the relationships between these polymorphisms and smoking can be explained by chance. Therefore, the presence of positive associations for CYP1A1, GSTM1 and p53 polymorphisms in smokers and PCa risk must be determined in a bigger study.

In conclusion, our results suggest that a combination of p53cd72, CYP1A1, GSTM1 genetic polymorphisms and smoking play a significant role in modified prostate cancer risk on the study population, which means that smokers carrying susceptible genotypes might have a significantly higher risk of PCa than those carrying non-susceptible genotypes.

Acknowledgment

The National Cancer Corporation of Chile supported this work. We thank Dr Jorge Soto for reviewing this manuscript.

References

1 Adami SL. Prostate Cancer. In: Hans-Olov Adami DH, Trichopoulos D, editors. Textbook of Cancer Epidemiology. New York: Oxford University Press, 2002; p400-28.

2 Coughlin, SS, Hall IJ. A review of genetic polymorphisms and prostate cancer risk. Ann Epidemiol 2002; 12: 182-96.

3 Daly AK, Fairbrother KS, Smart J. Recent advances in understanding the molecular basis of polymorphisms in genes encoding cytochrome P450 enzymes. Toxicol Lett 1998; 102-103: 143-7.

4 Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997; 88: 323-31.

5 Greenblatt, MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suprressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994; 54: 4855-78.

6 Agundez JA, Martinez C, Olivera M, Gallardo L, Ladero JM, Rosado C, et al. Expression in human prostate of drug- and carcinogen-metabolizing enzymes: association with prostate cancer risk. Br J Cancer 1998; 78: 1361-7.

7 Quiñnes L, Lucas D, Godoy J, Cáeres D, Berthou F, Varela N, et al. CYP1A1, CYP2E1 and GSTM1 genetic polymorphisms. The effect of single and combined genotypes on lung cancer susceptibility in Chilean people. Cancer Lett 2001; 174: 35-44.

8 Hirvonen, A. Combinations of susceptible genotypes and individual responses to toxicants. Environ Health Perspect 1997; 105 (Suppl 4): 755-8.

9 Seidegard J, Vorachek WR, Pero RW, Pearson WR. Hereditary differences in the expression of the human glutathione transferase active on trans-stilbene oxide are due to a gene deletion. Proc Natl Acad Sci USA 1988; 85: 7293-7.

10 Nakazato H, Suzuki K, Matsui H, Koike H, Okugi H, Ohtake N, et al. Association of genetic polymorphisms of glutathione-S-transferase genes (GSTM1, GSTT1 and GSTP1) with familial prostate cancer risk in a Japanese population. Anticancer Res 2003; 23: 2897-902.

11 Ferreira PM, Medeiros R, Vasconcelos A, Costa S, Pinto D, Morais A, et al. Association between CYP2E1 polymorphisms and susceptibility to prostate cancer. Eur J Cancer Prev 2003; 12: 205-11.

12 Kidd LC, Woodson K, Taylor PR, Albanes D, Virtamo J, Tangrea JA. Polymorphisms in glutathione-S-transferase genes (GST-M1, GST-T1 and GST-P1) and susceptibility to prostate cancer among male smokers of the ATBC cancer prevention study. Eur J Cancer Prev 2003; 12: 317-20.

13 Gsur A, Haidinger G, Hinteregger S, Bernhofer G, Schatzl G, Madersbacher S, et al. Polymorphisms of glutathione-S-transferase genes (GSTP1, GSTM1 and GSTT1) and prostate-cancer risk. Int J Cancer 2001; 95: 152-5.

14 Kote-Jarai Z, Easton D, Edwards SM, Jefferies S, Durocher F, Jackson RA, et al. CRC/BPG UK Familial Prostate Cancer Study Collaborators.Relationship between glutathione S-transferase M1, P1 and T1 polymorphisms and early onset prostate cancer. Pharmacogenetics 2001; 11: 325-30.

15 Rebbeck TR, Walker AH, Jaffe JM, White DL, Wein AJ, Malkowicz SB. Glutathione S-transferase-mu (GSTM1) and -theta (GSTT1) genotypes in the etiology of prostate cancer. Cancer Epidemiol Biomarkers Prev 1999; 8 (4 Pt 1): 283-7.

16 Acevedo C, Opazo JL, Huidobro C, Cabezas J, Iturrieta J, Quinones Sepulveda L. Positive correlation between single or combined genotypes of CYP1A1 and GSTM1 in relation to prostate cancer in Chilean people. Prostate 2003; 57: 111-7.

17 Miller SA, Dykes DD, Polesky HF. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 1988; 16: 1215.

18 Ambrosone CB, Freudenheim JL, Graham S, Marshall JR, Vena JE, Brasure JR, et al. Cytochrome P4501A1 and glutathione S-transferase (M1) genetic polymorphisms and postmenopausal breast cancer risk. Cancer Res 1995; 55: 3483-5.

19 de la Calle-Martin O, Fabregat V, Romero M, Soler J, Vives J, Yague J. AccII polymorphism of the p53 gene. Nucleic Acids Res 1990;18: 4963.

20 Garte S, Boffetta P, Caporaso N, Vineis P. Metabolic gene allele nomenclature. Cancer Epidemiol Biomarkers Prev 2001; 10: 1305-6.

21 Botto, LD, Khoury MJ. Commentary: facing the challenge of gene-environment interaction:the two-by-four table and beyond. Am J Epidemiol 2001; 153: 1016-20.

22 Irarrazabal, CE, Rojas C, Aracena R, Marquez C, Gil L. Chi-lean pilot study on the risk of lung cancer associated with codon 72 polymorphism in the gene of protein p53. Toxicol Lett 2003; 144: 69-76.

23 Huang SP, Wu WJ, Chang WS, Wu MT, Chen YY, Chen JY, et al. p53 Codon 72 and p21 Codon 31 Polymorphisms in Prostate Cancer. Cancer Epidemiol Biomarkers Prev 2004; 13: 2217-24.

24 Henner WD, Evans AJ, Hough KM, Harris EL, Lowe BA, Beer TM. Association of codon 72 polymorphism of p53 with lower prostate cancer risk. Prostate 2001; 49: 263-6.

25 Seidegard J, Ekstrom G. The role of human glutathione transferases and epoxide hydrolases in the metabolism of xenobiotics. Environ Health Perspect 1997; 105 (Suppl 4): 791-9.

26 Scheckenbach K, Lieven O, Gotte K, Bockmuhl U, Zotz R, Bier H, et al. p53 codon 72 polymorphic variants, loss of allele-specific transcription, and human papilloma virus 16 and/or 18 E6 messenger RNA expression in squamous cell carcinomas of the head and neck. Cancer Epidemiol Biomarkers Prev 2004; 13 (11 Pt 1): 1805-9.

27 Dong, H, Bonala RR, Suzuki N, Johnson F, Grollman AP, Shibutani S. Mutagenic potential of benzo[a]pyrene-derived DNA adducts positioned in codon 273 of the human P53 gene. Biochemistry 2004; 43: 15922-8.