Catechol-O-methyltransferase Gene Polymorphisms in Benign Prostatic Hyperplasia and Sporadic Prostate Cancer

Introduction

Prostate cancer incidence rates have ranked first, and death rates were second among all men with cancer in the United States over the past several years (1). Despite these high rates, the genetic basis of this disease is not well understood. Factors that may play a role in prostate carcinogenesis are estrogens and their metabolites (2, 3). Estrogens are implicated in the prostatic disease process by studies comparing racial differences in circulating levels of estrogens in men of high and low risk for prostate cancer. Serum estrone concentrations were 16% higher in healthy African American compared with lower risk Caucasian men (4). When comparing Dutch with Japanese, serum estradiol levels were significantly elevated in the higher risk Dutch men (5). Additionally, aromatase has been shown to be present in human prostate, which could provide a local source of estrogen via testosterone (6, 7), and epithelial levels of estrogens are higher compared with stromal levels in normal prostates (8). Interestingly within African Americans, prostate cancer patients displayed a significantly elevated level of estrone compared with matched controls (9). In addition to the malignant form, benign prostatic hyperplasia (BPH) has also been shown to have significantly elevated estradiol and estrone levels and is correlated with age (8); and this increase seems to be associated with prostate volume (10). On the contrary, other studies with BPH patients show estradiol levels to remain constant, whereas dihydrotestosterone declines with age, and this estrogen-dominant environment was suggested to lead to BPH (11).

The reports mentioned above thus suggest estrogen to play a role in the prostatic disease process. Although the classic role of estrogens in cells are to bind to its receptors and produce their biological effects, they may also undergo hydroxylation, and the resulting catechol estrogen may play a role in prostate carcinogenesis (12, 13). This conversion to catechol estrogens is due to cytochrome P450 (CYP) enzymes and include CYP1A1 (14) and CYP1B1 (15, 16) that have been shown to be expressed in human prostatic cells (17, 18). Indeed, catechol estrogen metabolites were observed to be present in the various lobes of the prostate in rats when 4-OH-estradiol was injected i.p. (13). As a consequence of their presence, these catechol estrogens and metabolites have been shown to induce adenocarcinoma in mice (19) as well as DNA single-strand breakage (20) and mutation (21). Thus, these findings show that catechol estrogens may play a causative role in the carcinogenesis process and may affect the prostate.

To counter the effects of these highly reactive and carcinogenic catechol estrogens, several enzymes are present in cells that can detoxify these compounds. One such enzyme is catechol-O-methyltransferase (COMT) that attaches a methyl group from the coenzyme S-adenosyl-l-methionine, to the hydroxyl group of the catechols, forming a methoxy compound (22). This methoxylation can either be at the 2-, 3-, or 4-positions of the catechol ring (23); thus, the formation of highly reactive catechol quinones and semiquinones that are damaging to DNA (24-27) or toxic to cells (28, 29) are prevented. These methoxy estrogens are also biologically inactive because they have little to no affinity for estrogen receptors and have no estrogenic effects on target tissues (30). On the contrary, 2-methoxy-estrogen has been shown to have antiangiogenic and antiproliferative effects on cancer cell lines in vitro (31-33), and this may be due to its ability to inhibit DNA synthesis and mitosis (31). Likewise, when given to mice, this methoxy estrogen has been shown to inhibit neovascularization and suppress the growth of solid tumors (32, 33). Thus, the COMT enzyme is shown to play a protective role in the body.

There are two types of COMT proteins. A soluble form (S-COMT) contains 221 amino acids with a mass of 24.4 kDa, whereas a membrane-bound form (MB-COMT) contains 271 amino acids with a mass of 30.0 kDa (34). The COMT gene is located on chromosome 22, band q11.2 (35, 36). There are six exons of which the first two are noncoding. On exon 3 are two distinct ATG start codons for promoters P1 and P2, which code for S-COMT and MB-COMT, respectively (37).

Many polymorphisms have been identified in the COMT gene, and three of these in the coding region are as follows: codon 62 (C→T), codon 72 (G→T), and codon 158 (G→A; Fig. 1A; refs. 38, 39). The polymorphism at codon 62 is silent (His), whereas the variant genotype alters Ala to Ser at codon 72 and Val to Met at codon 158. Prior studies have shown that polymorphisms and mutations in many genes cause changes in the function of proteins (40, 41). In the case of COMT, the amino acid change at codon 158 (Val → Met) results in an enzyme activity that is up to four times less than the wild type (38, 42, 43). This codon 158 polymorphism has been extensively studied and found to have associations with neurologic/psychological diseases (44-49). Associations for the codon 158 polymorphism have also been investigated in cancers and include breast (50-53), bladder (54), liver (55), and endometrial (56, 57). In prostate cancer, Suzuki et al. (58) found no significant difference between cancer versus control for the codon 158 polymorphism. However, their study used familial prostate cancer patients that have a higher predisposition to cancer (59, 60). Therefore, in the present study, associations of COMT polymorphisms with sporadic prostate cancer patients were determined.

Because polymorphisms are inherited, interindividual differences in disease risk may occur and include prostate cancer. The present study was designed to investigate the polymorphisms of the COMT gene at three different loci in a Japanese population and determine their association with BPH and sporadic prostate cancer risk.

Materials and Methods

Samples

Prostate cancer samples (n = 178) were obtained from the Department of Urology at the hospital of Shimane Medical University, Izumo, Japan. The samples were collected from patients by radical prostatectomy and obtained during the period from 1997 to 2003. Samples were pathologically characterized in terms of their stage (Tumor-Node-Metastasis system) and grade (General Rules for Clinical and Pathological Studies on Prostate Cancer by Japanese Urological Association and Japanese Pathological Society). Normal region of tissues were carefully dissected under a microscope and used for analyses. Prostate-specific antigen was measured in patients, and mean levels were 11.5 ± 0.8. Age of cancerous patients was 68.6 ± 0.4 years. BPH specimens (n = 134) were collected via trans-urethral resection. Average age of BPH patients was 73.1 ± 0.7 years, and mean prostate-specific antigen levels were 6.4 ± 0.6. Control blood samples (n = 131) from volunteers were also obtained from male Japanese for genotyping. To ascertain that volunteers were healthy and free of cancer, they all underwent various tests that included physical exams, questionnaires about their health and history, chest X-rays, blood and urine tests for various tumor markers, abdominal ultrasound, gastric endoscopy, and colon enema. Based on doctor examinations and past history, all were confirmed to be free of cancer. Volunteers were closely matched with prostate cancer patients and had a mean age of 67.3 ± 1.05. Prostate cancer patients, BPH patients, and volunteers for the control group were unrelated and selected at random during the same period as the collection of prostate cancer samples.

DNA Extraction

DNA was extracted from all prostate cancer, BPH, and control samples by using a DNA extraction kit (Qiagen, Valencia, CA). Quantity and quality of DNA were measured at 260 and 280 nm by the use of a spectrophotometer.

Analyses of COMT Polymorphisms

A two-step PCR procedure was designed for the analysis of COMT polymorphisms. The primers of the three polymorphic sites studied and PCR conditions are summarized in Table 1. In the first PCR, DNA (10 ng) was amplified in a 20 μL reaction containing 1.5 mmol/L MgCl₂, 0.8 mmol/L deoxynucleotide triphosphate mix, PCR buffer, and 0.5 unit of Red-Taq polymerase (Sigma-Aldrich, St. Louis, MO), along with primer sets designed to contain the polymorphic sites (Table 1). In the sequence-specific PCR (SSP), each polymorphic fragment was further amplified under similar conditions as the first-step PCR except for the use of SSP primer sets (Table 1).

Gel Electrophoresis

Each of the SSP products were electrophoretically separated on 3% agarose gels using 200 V at ambient temperature. The products were then visualized by ethidium bromide staining under UV light.

DNA Sequencing

To confirm genotyping, products of the first PCR were subjected to direct DNA sequencing. DNA was purified from gels using a QIAquick PCR purification kit (Qiagen). Sequence analysis of purified products was then determined by using the first PCR primers and ABI 377 Sequencer and Dye Terminator Cycle sequencing kit (Applied Biosystems, Inc., Foster City, CA). Confirmation of DNA sequence was done on at least three representative samples for each of the polymorphic types.

Statistical Analyses

Frequencies of the various genotypes and allele types of COMT polymorphisms in the different categories of samples were determined and tabulated. χ² analysis was used to test each of the polymorphisms for differences in genotypic and allelic frequencies among BPH, prostate cancer, and control, as well as between stages and grades of cancer. Relative risk associated with a particular genotype was estimated by calculating odds ratios along with 95% confidence intervals. Linkage disequlibrium between polymorphic sites and haplotype frequency differences between disease categories were calculated using the SNPAlyze version 2.2 software (DYNACOM, Tokyo, Japan).

Results

The COMT gene structure is shown in Fig. 1A. Two polymorphic sites are located on exon 3 at codon 62 (C→T) and codon 72 (G→T), and a third is located on exon 4 at codon 158 (G→A). Representative gels displaying the polymorphic genotypes at these codons are shown in Fig. 1B. Upper gels show wild type, and lower gels show variant type of each polymorphic locus. Lanes 1, 2, 3 and 4 show 100-bp DNA ladder marker, homozygous wild type, heterozygous, and homozygous variant genotype, respectively. The upper band within gels is the first PCR product, and the SSP products are the lower band. SSP products were confirmed by DNA sequencing (data not shown).

Genotypic and allelic frequencies of the three different polymorphisms of the COMT gene in Japanese controls, BPH, and prostate cancer patients are shown in Table 2A and B, respectively. All polymorphic frequencies for healthy controls follow the Hardy-Weinberg equilibrium. Interestingly, in this Japanese control population, the homozygous variant is detected in only one sample at codon 72 (T/T), and frequency is considerably low at codon 62 (T/T; 3.8%) and codon 158 (A/A; 5.3%). Among BPH patients, no differences in genotype or allele frequency were detected when compared with controls for all polymorphisms measured. However, risks for prostate cancer were significant for the variant genotype and allele at these sites. At codon 62, 11.2% of patients had the variant T/T versus 3.8% of healthy controls (χ², P = 0.060). The odds ratio compared with C/C (reference) was 3.24 with a 95% confidence interval of 1.38 to 7.61. The variant genotype at codon 158 was also observed to be a risk factor (χ², P = 0.047), as more than twice the cancer patients (13.5%) had the variant A/A compared with controls (5.3%). Odds ratio and 95% confidence interval were 3.00 and 1.38 to 6.54, respectively. Concordantly, the variant allele A proved to be a risk for cancer (P = 0.034). No association was found for the heterozygous variant G/T on codon 72.

Linkage disequlibrium values were calculated among each of the three polymorphic sites in the normal healthy control group and are shown in Table 3. Interestingly, codon 62 was in linkage disequlibrium with codon 158 and had a D of 0.158. Codon 72, however, was not in linkage disequlibrium with either codon.

Haplotype frequencies of the two polymorphic sites (codons 62 and 158) were calculated, and Table 4 shows the distributions for healthy controls, BPH, and prostate cancer. The codon 62 to 158 C-G haplotype is observed in 72.1% of healthy controls, 72.4% of BPH, and 64.3% of prostate cancer patients. Interestingly, when compared with C-G, the other haplotypes (C-A, T-G, and T-A) were observed to be significantly higher in frequency among prostate cancer patients compared with controls (P = 0.040). However, no significant differences were observed in haplotype frequencies between BPH and controls.

The correlation of the polymorphisms of COMT with clinical stage of cancer is shown in Table 5. Prostate cancer tissues were classified as either ≤T₂ (n = 118) or ≥T₃ (n = 58) with two samples of unknown status. No significant differences between stages were found due to genotype for all codons. Allele frequencies also did not differ between stages except for a tendency (P = 0.096) for the codon 158 variant A to be correlated with higher stages (≥T3) of prostate cancer.

Table 6 shows the frequencies of the various genotypes and alleles of different COMT polymorphic sites with respect to pathologic grade of cancer. Three types of grade classifications were made and include ≤6 (n = 96), 7 (n = 47), and ≥8 (n = 31) with four samples of unknown status. No significant associations for both genotype and allele type of COMT polymorphic sites were observed between these grades of cancer. Classifications based on grade ≤7 versus grade ≥8 or grade ≤6 versus grade ≥7 also show no correlation with either genotype or allele at these polymorphic sites.

The results of these experiments suggest that COMT polymorphisms may be a risk factor for prostate cancer.

Discussion

The bioactivation of estrogens and carcinogens via CYP1A1 or CYP1B1 (12, 16) have been associated with several tumors in various tissues, including prostate (12, 13, 27). One of the resultant chemical structures responsible for the tumorigenicity of CYPs is catechol estrogens. This form of estrogen leads to the formation of semi-quinones and quinones that are known to react with DNA (61, 62). Therefore, detoxification of catechol estrogens in the cells of the body is important to prevent mutations, and the COMT enzyme is responsible for this defense.

Proper enzymatic function of COMT in cells is thus essential to prevent genetic changes. Polymorphisms of genes, however, have been shown to alter enzymatic activity. In the case of CYP2A6, a polymorphism results in reduced activity of the enzyme (63). Thus, polymorphism is a potential mechanism by which COMT activity is reduced, thereby causing an accumulation of mutagenic catechol compounds. The COMT gene consists of six exons, and at least three polymorphisms in the coding region (codons 62, 72, and 158) have been identified that are located on exons 3 and 4 (Fig. 1A). Exons 1 and 2 are noncoding and contain no polymorphism. The polymorphism at codon 62 is silent and results in no amino acid change (His; ref. 39). The other two sites, however, result in amino acid substitutions as follows: codon 72 (Ala → Ser) and codon 158 (Val → Met; ref. 39).

In the present study, the three polymorphic sites mentioned above were analyzed to determine if the variant polymorphism influences the risk of prostatic disease. No differences were observed in genotypic or allelic frequency at any of these sites between BPH and normal healthy controls in this Japanese population. Interestingly, the homozygous variant T/T on codon 72 is only found in one sample, and this rarity is observed in another Asian population (64). Thus, these polymorphic sites are not responsible for the development of the benign form of prostate disease. In support of this finding, other estrogen-related gene polymorphisms were also found to have no risk for BPH, such as steroid 5-reductase type II (65) and CYP17 (66).

When determining the effects of COMT polymorphisms on prostate cancer, the heterozygous genotype on codon 72 was also observed to have no association. Thus, Ala/Ser⁷² of the COMT enzyme does not seem to play a role in prostatic carcinogenesis in this Japanese population. On the other hand, differences in genotypic distributions were found at codon 62 (P = 0.060) and codon 158 (P = 0.047) of COMT between prostate cancer patients and healthy controls in this population, and these two polymorphic sites were determined to be in linkage disequlibrium with each other. The odds ratio and 95% confidence interval of variant genotype T/T on codon 62 and A/A on codon 158 were significantly higher in cancer compared with normal controls. Respective allele frequencies also differed at codon 158 (P = 0.034). Concordantly, haplotype analyses among these two sites also show that the presence of a variant (codon 62-158, C-A,T-G, T-A) are significantly associated with prostate cancer. Thus, the silent variant His⁶² and variant Met¹⁵⁸ seem to be of importance in the pathogenesis of prostate cancer. These results for Met¹⁵⁸ are in agreement with studies done on breast cancer as Huang et al. (67) and Saczi et al. (52) reported an association for Met¹⁵⁸ with cancer. However, in another study done on prostate cancer in Japan, Suzuki et al. (58) reported no association for Met¹⁵⁸ with cancer. Differences between their studies and the present one may possibly be due to the type of cancerous patients as Suzuki et al. used familial cancer patients whereas the present consisted of sporadic cancer patients. Familial prostate cancer has been shown by Woolf (59) to have a greater expectancy to develop the cancer with a risk factor of three times that of controls, and Steinberg et al. (60) reported up to a 11-fold risk among patients with three first-degree relatives. Another difference between these studies is the N size used as the present study consisted of 44% more patients. Thus, Suzuki et al. had 114 control and 101 cancer patients compared with the present study with 131 control and 178 patients. Interestingly, in Suzuki et al.'s study, the risk of the homozygous variant A/A on codon 158 for cancer was essentially none with a P of 0.99. However, the heterozygous G/A in their study had a slight indication toward a risk for familial prostate cancer with a P of 0.18 and perhaps, a greater N size might have shown the heterozygous variant to achieve significance in their study. As a consequence of this pattern of results, their data display a highly unique genotype distribution among prostate cancer patients.

The correlation between the COMT genotypes and alleles with clinical stages of prostate cancer were analyzed. Although cases of samples at higher stages (≥T₃) were greater compared with lower stages (≤T₂) for the variant A/A at codon 158 (17.2% versus 11.9%, respectively), no significant difference was found and may be due to a small sample size (n = 118 for ≤T₂; n = 58 for ≥T₃). The variant allele at this codon, however, was found to approach a significant association with advanced stages of cancer (P = 0.096). Thus, polymorphisms of the COMT gene at codon 158 may play a role in the progression of sporadic prostate cancer. In agreement, codon 158 polymorphism has been shown to be correlated with advanced stages of breast cancer in a Japanese population (68). On the contrary in familial prostate cancer, Suzuki et al. (58) found no correlation between the COMT 158 polymorphism with stage of cancer. These authors, however, did not show their data regarding stage and had a much smaller number of cases of prostate cancer (n = 101) compared with the present study (n = 178).

The correlation between the COMT genotypes and alleles with pathologic grades of prostate cancer were also analyzed. No association for grade types was found at all polymorphic sites studied. Suzuki et al. (58) also observe no association for the codon 158 variant with grade of cancer in familial prostate cancer. Thus, polymorphisms of the COMT gene do not seem to play a role with pathologic grade of prostate cancer. This lack of association with grade of prostate cancer was also found in other estrogen-related gene polymorphisms, such as CYP17 (69, 70) and CYP19 (70).

The mechanism by which polymorphisms on codons 62 and 158 of the COMT gene may play a role in the carcinogenesis process is not fully understood. The polymorphism on codon 62 is silent because the amino acid remains unchanged as His. However, gene expression could be affected as a result of a structural change in the RNA due to the variant, leading to an alteration in processing or efficiency of translation (71). In contrast, the polymorphism on codon 158 results in an amino acid change from Val to Met. As a consequence of this change, properties of the resultant variant enzyme have been shown to reduce its activity with the heterozygous type displaying intermediate activity (38, 42, 43). COMT activity of the variant form has been shown to be up to 4-fold less than normal (38, 42, 43) and to be thermolabile as well (38, 43, 72, 73). In addition to its role in the regulation of various biological substrates (i.e., dopamine and norepinephrine; refs. 74, 75), the COMT enzyme also is important in the elimination of toxic or carcinogenic compounds that include catechol estrogens. These catechol compounds have been shown to induce adenocarcinoma in mice (19), DNA single-strand breakage (20), and mutation (21). Thus, the variant polymorphism on codon 158 leading to a reduced COMT activity may cause an increase in mutagenic compounds and lead to prostate cancer.

This is the first report to investigate the polymorphisms in the COMT gene at three different locations (codons 62, 72, and 158) in BPH and prostate cancer. Interestingly, these polymorphisms were observed to be a risk factor for prostate cancer but are not associated with BPH. Because polymorphisms are inherited, the Met¹⁵⁸ variant that can lead to reduced methoxylation activity and along with the codon 62 polymorphism may be responsible for interindividual differences in prostate cancer risk associated with catechol estrogens or carcinogens. In conclusion, these findings suggest that polymorphisms of COMT may be important in understanding the pathogenesis of prostate cancer.