Serum sex steroid hormone levels and polymorphisms of CYP17 and SRD5A2: implication for prostate cancer risk

Serum sex steroid hormone levels and polymorphisms of CYP17 and SRD5A2: implication for prostate cancer risk H Kakinuma 1, N Tsuchiya 1, T Habuchi 1 *, C Ohyama 1, S Matsuura 1, L Wang 1, A Nakamura 2 & T Kato 1 1 Department of Urology, Akita University School of Medicine, Akita, Japan; and 2 Department of Medical Information Science, Akita University School of Medicine, Akita, Japan (2004) 7, 333 337 & 2004 Nature Publishing Group All rights reserved 1365-7852/04 $30.00 www.nature.com/pcan Polymorphism of the steroid hormone-related genes might affect life-long androgen exposure, thus altering a risk of prostate cancer incidence. To evaluate the effect of the polymorphisms of CYP17 and SRD5A2 on serum steroid hormone levels, the 164 male Japanese cohort were tested for serum hormone levels and the genotype of the polymorphisms of CYP17 (T C base substitution in the promoter region) and SRD5A2 (V89L). The linear trends across the CYP17 genotypes in serum-free testosterone and androstenedione levels were found, suggesting the importance of the polymorphism of CYP17 in determining the circulating androgen levels. (2004) 7, 333 337. doi:10.1038/sj.pcan.4500753 Published online 12 October 2004 Keywords: CYP17 gene; SRD5A2 gene; polymorphism; steroid hormones Introduction The incidence and mortality rates of prostate cancer vary significantly among countries, with Japan having the lowest mortality rate in the world, reported as 4 per 100 000 between 1992 and 1995. 1,2 Genetic and environment backgrounds may be major factors accounting for its worldwide and ethnic variation. According to Ross et al, 3,4 both African and European American males had 41 51% higher preandrogen levels (androsterone glucuronide), compared with Japanese males. Furthermore, de Jong et al 5 also reported that 71% higher circulating levels of testosterone were seen in Caucasian males compared with those in Japanese males. The extensive metaanalysis of previously published studies on hormonal predictors revealed that males whose total testosterone levels were in the highest quartile were 2.34-fold more likely to develop prostate cancer. 6 Although the degree to which androgen and androgen metabolites contribute to prostate cancer risk remains a contentious issue, these findings support the hypothesis that the higher incidence of prostate cancer in a population might be substantially *Correspondence: T Habuchi, Department of Urology, Akita University School of Medicine, 1-1-1 Hondo, Akita 010-8543, Japan. E-mail: thabuchi@doc.med.akita-u.ac.jp Received 20 March 2004; revised 3 June 2004; accepted 6 July 2004; published online 12 October 2004 influenced by long-lasting impacts of higher circulating androgens. Several studies have addressed the hypothesis that the high/low activity alleles in the CYP17 and SRD5A2 polymorphisms could influence the incidence of prostate cancer and partially explain the observed ethnic/racial differences. 7 14 CYP17, located on chromosome 10q24, encodes cytochrome p450c17 that plays an important role at two key steps in testosterone biosynthesis. CYP17 harbors a single base pair mutation (T C) in the untranscribed 5 region of exon one. Conventionally, the T and C alleles are designated as the A1 allele and A2 allele, respectively. Although the results from casecontrol studies evaluating the association between the CYP17 genotype and prostate cancer risk have been contradictory, our previous study showed that the risk of prostate cancer was increased up to two-fold in males with A1 allele in a Japanese population. 7 11 On the other hand, SRD5A2 encodes for human type II 5a-reductase, which converts testosterone to the more potent, dihydrotestosterone (DHT), located on chromosome 2p23. Makridakis et al 15 reported that the presence of a valine (V) allele instead of a leucine (L) allele at codon 89 reduced significantly the conversion rate of testosterone to DHT, and the higher level of 5a-androstane-3a, 17bdiol glucuronide (A-diol-g) in V allele homozygous males compared with L allele homozygous and heterozygous males. In support of their observation, Nam et al 13

334 and Li et al 16 suggested that V allele homozygote males had a significantly decreased risk of prostate cancer. The biological importance and mechanisms of those polymorphisms have not been elucidated completely, and it has been of much interest to analyze the association between serum steroid hormone levels and these polymorphisms in view of the possible relation between the serum steroid hormone levels and a prostate cancer risk. Furthermore, those polymorphisms could be utilized as a marker of an individual s life-long hormone exposure and cancer risk susceptibility, if they really reflected serum hormone levels. In this study, five major circulating hormones, namely, dehydroepiandrosterone sulfate (DHEA S), andreostenedione, free testosterone (fte), testosterone (Te), and estradiol (E2) in 164 healthy Japanese males were measured to assess the biological relevance of those polymorphisms. Materials and methods Subjects This study group was drawn from 170 consecutive healthy males (60 80 y old) who visited a community health center for an annual health check between January and May in 2000. The 164 samples from males with normal serum PSA levels (o4.0 ng/ml, Tandem-R assay) were used for this study. In total, 12 ml of whole blood for a hormone assay and DNA analysis was donated after receiving written informal consent. The vein puncture was performed between 8:00 a.m. and 10:00 a.m. to minimize the effect of daily fluctuation. Those blood samples were immediately stored at 701C until the hormone assay. DNA extraction and hormone assays Genomic DNA was extracted using a QIAamp Blood Kit (QIAGEN, Valencia, CA, USA). DHEA S, androstenedione, fte, Te, and E2 were measured by a radioimmunoassay kit provided by the Diagnostic Products Corporation (LA, USA) in collaboration with SRL Inc. Co. (Tokyo, Japan). Samples were randomly assorted into batches. Genotype analyses SRD5A2 genotyping was performed by the PCR-RELP method as described previously. 16 The PCR primer set was designed to amplify the part of SRD5A2 that includes the polymorphic site (forward: 5 0 -CCACCTGG GACGGTACTTCT, reverse: 5 0 -CTCCACGCTGCGCTCC TGGA). PCR was carried out in a 25 ml aliquot containing 20 ng of genomic DNA template, 50 pmol of each primer, 125 mm deoxynucleotidetriphosphates, 1 U of Ampli-Taq Gold DNA polymerase (PE Applied Biosystems, Foster City, CA, USA), and 2.5 ml of 10 reaction buffer supplied by the manufacturer. PCR amplification conditions were 10 min initial denaturation at 951C followed by 35 cycles of 30 s at 941C, 30 s at 601C, and 90 s at 721C, 7 min final extension at 721C. Products were digested with 10 U of RsaI (NEB, Beverly, MA, USA) and separated by electrophoresis on Gene Gel Excel 12.5/24 (Amersham Biosciences, Inc., Piscataway, NJ, USA). DNA bands were detected by silver staining. The valine allele, which was sensitive to RsaI, resulted in two bands (35 and 89 bp), while the leucine allele was seen as a single band (124 bp). The genotypes were designated as V for the valine allele and L for the leucine allele. CYP17 genotyping was performed by the PCR-RELP method as described previously. 11 In brief, a single base pair mutation (T C) in the A2 allele creates the MspAI recognition site. The C allele was designated as A2 when the PCR product was digested by MspAI, while it was designated as A1 when the product was undigested. Statistical methods All data were entered into an access database and analyzed using Excel 2000 and SPSS (version 10.0J, SPSS) software (SPSS Institute, Inc., Chicago, IL, USA). The w 2 test was used to compare the allele and genotype frequency. Hardy Weinberg equilibrium analyses were performed to compare observed and expected genotype frequencies using a w 2 test (d.f. ¼ 1). Linear regression models were used to evaluate associations between three genotypes and circulating steroid hormone levels. 10 The least-square geometric mean hormone levels were estimated, and the differences in hormone levels between genotypes were evaluated with the A1/A1 group or the V/V group as a reference category. We also evaluated associations between the combination of the CYP17 and SRD5A2 genotypes and plasma hormone levels with the A1/A1 group and V/V group as a reference category. The interaction of the CYP17 SRD5A2 genotype was assessed including the gene gene interaction term in the linear regression models. All statistical results in this study were not age adjusted since the range and standard deviation (s.d.) of the study subjects were so narrow for age adjustment. No data concerning diet, body mass index, or life style were available. The results were considered to be significant when Po0.05. Results The age and serum PSA levels of 164 samples were 66. 474.7 y (mean7s.d.) and 1.170.05 ng/ml, respectively. The frequencies of variant alleles of CYP17 and SRD5A2 are shown in Tables 1 and 2, respectively, and are in Hardy Weinberg equilibrium, with the expected frequencies for the CYP17 genotypes being A1/A1 ¼ 50, A1/A2 ¼ 81, and A2/A2 ¼ 32 (P ¼ 0.268); and for SRD5A2 genotypes, V/V ¼ 52, V/L ¼ 80, and L/L ¼ 31 (P ¼ 0.981) (Tables 1 and 2). The prevalence of the CYP17 A2 allele and SRD5A2 L allele were 44.5 and 43.3%, respectively. DHEA S, androstenedione, fte, Te, and E2 were measured by radioimmunoassay. Intra-assay coefficients of variation were 4.9% at 760.0 ng/ml for DHEA S, 8.5% at 1.22 ng/ml for andreostenedione, 7.7% at 10.84 pg/ml for fte, 4.1% at 5.49 pg/ml for Te, and 2.5% at 87.92 pg/ml for E2. The values of each hormone

Table 1 Mean hormone levels (95% CI) by the CYP17 genotype 335 Genotype A1/A1 A1/A2 A2/A2 Hormone Number (%) 54 (33%) 74 (45%) 36 (22%) P for trend DHEA S (ng/ml) 1518.8 1356.7 [ 10.6%] b 1,519.3 [+0.06%] NS (1316.4 1721.2) a (1183.8 1529.5) (1288.3 1750.3) Androstenedione (ng/ml) 1.58 1.50 [ 12.9%] 1.31 [ 16.7%] 0.020 (1.38 1.76) (1.40 1.60) (1.20 1.41) Free testosterone (pg/ml) 11.6 11.4 [ 2.7%] 10.5 [ 9.9%] 0.045 (10.8 12.4) (10.7 12.0) (9.8 11.1) Testosterone (ng/ml) 5.00 5.41 [+8.1%] 4.76 [ 4.8%] NS (4.66 5.34) (5.03 5.79) (4.37 5.15) Estradiol (pg/ml) 19.8 20.4 [+3.1%] 18.2 [ 8.0%] NS (17.8 21.5) (18.4 22.3) (15.9 20.3) a 95% CI. b Differences (%) compared with A1/A1 genotype. Table 2 Mean hormone level (95% CI) by the SRD5A2 genotype Genotype V/V V/L L/L Hormone Number (%) 52 (32%) 82 (50%) 30 (18%) P for trend DHEA S (ng/ml) 1531.6 1390.9 [ 9.1%] b 1446.7 [ 5.5%] NS (1338.9 1724.3) a (1243.7 1538.1) (1094.9 1798.5) Androstenedione (ng/ml) 1.53 1.43 [ 7.1%] 1.56 [+1.6%] NS (1.36 1.70) (1.31 1.53) (1.40 1.71) Free testosterone (pg/ml) 11.3 11.0 [ 2.5%] 11.8 [+5.2%] NS (10.5 12.0) (10.4 11.5) (10.7 13.0) Testosterone (ng/ml) 5.07 5.11 [+0.8%] 5.32 [+5.1%] NS (4.65 5.48) (4.81 5.40) (4.70 5.93) Estradiol (pg/ml) 21.4 18.3 [ 14.4%] 20.3 [ 4.7%] NS (19.2 23.5) (16.8 19.8) (17.1 23.6) a 95% CI (confidence interval). b Differences (%) compared with V/V genotype. level were 1445.7757.1 ng/ml (DHEA S, mean7s.e.), 1.4870.04 ng/ml (androstenedione), 11.270.2 pg/ml (fte), 5.1370.11 pg/ml (Te), and 19.770.5 pg/ml (E2), respectively. The genotype-serum steroid hormone analysis revealed a linear decreasing trend across the CYP17 genotypes for fte and androstenedione levels, resulting in 9.9% (fte) and 16.7% (androstenedione) lower levels of each hormone in the A2 homozygote (Table 1). The P- values for the linear decreasing trend across the CYP17 genotype were 0.04 for fte, and 0.02 for androstenedione, respectively. As for the SRD5A2 genotype, no associations were found between genotypes and any hormone fractions (Table 2). In the analysis of the CYP17 SRD5A2 gene interaction, compared with the males homozygous for both the A1 and V allele, those with other genotype combinations did not have significantly increased or decreased mean levels of fte and androstenedione (Tables 3, 4). However, we observed 15% increase of fte across the SRD5A2 genotype among the A1/A1 genotype and 17% increase of androstenedione among the A1/A2 genotype (Tables 3, 4). For androstenedione, the increasing trend was statistically significant (P ¼ 0.047). In other hormone levels (DHEA S, Te, and E2), no significant gene interaction between the genotypes was observed. Discussion Recent molecular studies have suggested that the genetic polymorphisms of CYP17 and SRD5A2, which encode the key enzymes at androgen synthesis and metabolism, might affect sex hormone production and serum levels. Carey et al 17 first reported a positive association between the A2 allele of CYP17 and the risk of hyperandrogenic disease (polycystic ovary syndrome and male pattern baldness). Following this report, two groups revealed that the A2 allele was associated with elevated concentrations of estradiol in premenopausal and postmenopausal women. 18,19 So far, a few studies suggested that the A2 allele might increase the risk of breast cancer, while other studies suggested no such association. 20,21 On the other hand, in prostate cancer, the relation between the CYP17 polymorphism and prostate cancer risk has been conflicting. Although a few groups including ours reported that the A1 allele of CYP17 increased the risk of prostate cancer, other studies reported a negative association or converse results. 7 11 In the present study, the linear decreasing trends across the CYP17 genotypes were observed with significance for fte and androstenedione, a precursor of testosterone. The mean levels of fte and androstenedione in A2 homozygous males were 9.9 and 16.7% lower when

336 Table 3 Mean free testosterone level by the CYP17 and SRD5A2 genotype SRD5A2 genotype CYP17 genotype V/V V/L L/L P trend for the SRD5A2 genotype A1/A1 11.3 (18) 11.2 (27) 13.3 (9) NS A1/A2 11.5 (24) 11.0 (36) 12.0 (14) NS A2/A2 10.5 (10) 10.7 (19) 9.7 (7) NS P trend for the CYP17 genotype NS NS 0.02 P interaction, NS () ¼ number. Table 4 Mean androstenedione level by the CYP17 and SRD5A2 genotype SRD5A2 genotype CYP17 genotype V/V V/L L/L P trend for the SRD5A2 genotype A1/A1 1.78 (18) 1.47 (27) 1.47 (9) NS A1/A2 1.43 (24) 1.46 (36) 1.72 (14) 0.047 A2/A2 1.31 (10) 1.30 (19) 1.34 (7) NS P trend for the CYP17 genotype 0.049 NS NS P interaction, NS () ¼ number. compared with A1 homozygous males, while total testosterone and DHEA S did not appear to be related. In contrast, Haiman et al 10 and Allen et al 22 recently failed to find a significant relation between the CYP17 genotype and circulating steroid hormone levels regardless of their larger sample sizes, entering 377 control males and 622 control males. There might be a possibility that the influence of the CYP17 genotypes on enzyme activity could be more prominent in Asians than in Caucasians. Allen et al 22 used a cohort derived from a Caucasian population, while Haiman et al 10 did not mention the ethnicity of their cohort. In addition, the mean age of the cohort in Allen s study was nearly 20 y younger than the current study (47 y vs 66 y). 22 Thus, further investigation of the cohort, which should be well stratified by both ethnicity and age, may be needed to support that hypothesis. Although the role of estrogen for prostate cancer development has not been well understood, CYP17 is also accountable for regulation of estrogen and some studies showed that estrogen may support androgen s effect on prostate growth. 23 The present result showed no significant difference between E2 level and the polymorphism of CYP17. So far, a few findings supported the hypothesis that the polymorphism of CYP17 may have an effect on estrogen biosynthesis in female subjects, while in male subjects the relation was tenuous and limited. 10,18,19 The difference in the metabolic pathway by gender may partly explain this discrepancy. The present finding that the CYP17 A1 allele was related to lower free testosterone level corroborated the results for our previous observation in a Japanese case-controlled study in which the A1 allele was associated with an increased risk of prostate cancer with a gene dosage effect. 11 Although the relation between the serum testosterone levels and prostate cancer may be controversial, the trend of increasing prostate cancer development was observed with increasing serum-free ( ¼ bio-available) testosterone level as assessed by the ratio of testosterone to serum hormone binding protein (SHBG). 24 In addition, a single measurement of testosterone was suggested to reliably reflect mean annual androgen concentrations. 25 Thus, it is suggested that males with the CYP17 A1 allele could be exposed to higher levels of androgen throughout their life and end up being amenable to prostate cancer. Epidemiologic findings have also been suggesting ethnicity-specific differences in genotype frequency for SRD5A2. 7,15 The assumption that the L allele of the V89L in the SRD5A2 might be protective to prostate cancer has been gaining favor. 12 14,16 In support of this view, we recently reported that the V allele may dominantly increase the risk of prostate cancer (vs L/L genotype; OR, 1.69; 95% CI; 1.07 2.65). 16 Studies of analyses between the polymorphism of SRD5A2 (V89L) and 5areductase enzyme activity were originally conducted by Makridakis et al. 15,26 In their study, they showed that males with the L homozygous allele had lower 5areductase activity compared with males with the V homozygous allele, as evidenced by the finding that L homozygous males showed a 38% lower serum level of A-diol-g, which is a direct metabolite of DHT and may be an indication of SRD5A2 activity. They also revealed a significant difference in 5a-reductase enzyme activity between the L allele and V allele in vitro. 26 Recently, Hsing et al 14 and Allen et al 22 reported that males with the V/L and L/L genotypes showed a lower level of A- diol-g, compared with males with the V/V genotypes, among Chinese (24%) and Caucasian populations (2.4%). These findings may suggest that the impact of the SRD5A2 polymorphism on 5a-reductase activity appears more prominent in low-risk populations, but the serum testosterone level in the L/L genotype was 17% higher than V/V genotype in the study of Hsing et al. 14 In the

present study, we found neither a single effect of the SRD5A2 polymorphism nor a definite combined influence of the CYP17 and SRD5A2 polymorphisms on circulating hormone levels. Although there might be a small relation between serum androgen levels and the SRD5A2 genotype, the impact of the SRD5A2 polymorphism on serum androgen level would not be as robust as that of the CYP17 polymorphism. Since the 5areductase activity may play a major role in prostate tissue levels, studies of the effect of the SRD5A2 polymorphism on androgen activity in prostate tissue levels are required. Serum hormone levels may be affected by environmental factors such as life style, dietary habit, and fat intake, and it will be challenging to find a difference by genetic polymorphism alone. 27 Asian ethnic populations may be more suitable to evaluate the differences in mean hormone levels across the CYP17 and SRD5A2 genotypes, because the Japanese are less affected by environmental factors for prostate cancer than Caucasians. 28 However, the present study might have a flaw in that no information of these life styles on subjects were obtained. Furthermore, because our study subjects were limited to a relatively small number of normal men aged 60 y or over with normal serum PSA, it remains unknown whether the findings of this study apply to prostate cancer patients. To further delineate the roles of the polymorphisms of the two genes in prostate carcinogenesis, a larger study including both normal and prostate cancer subjects is required and warranted. The present limited findings suggested that the significance of the CYP17 polymorphism in determining levels of some androgens therefore affected the risk of prostate cancer, while the SRD5A2 polymorphism may have little effect on the serum level of testosterone among Japanese populations. A larger study will be needed to examine gene gene interactions in both normal males and prostate cancer patients. References 1 Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1998. CA Cancer J Clin 1998; 48: 6 29. 2 Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999; 49: 8 31. 3 Ross R et al. Serum testosterone levels in healthy young black and white men 5-alpha-reductase activity and risk of prostate cancer among Japanese and US white and black males. J Natl Cancer Inst 1986; 76: 45 48. 4 Ross RK et al. 5-Alpha-reductase activity and risk of prostate cancer among Japanese and US white and black males. Lancet 1992; 339: 887 889. 5 de Jong FH et al. Peripheral hormone levels in controls and patients with prostatic cancer or benign prostatic hyperplasia: results from the Dutch Japanese case control study. Cancer Res 1991; 51: 3445 3450. 6 Shaneyfelt T, Husein R, Bubley G, Mantzoros CS. Hormonal predictors of prostate cancer: a meta-analysis. J Clin Oncol 2000; 18: 847 853. 7 Lunn RM, Bell DA, Mohler JL, Taylor JA. Prostate cancer risk and polymorphism in 17 hydroxylase (CYP17) and steroid reductase (SRD5A2). Carcinogenesis 1999; 20: 1727 1731. 8 Gsur A et al. A polymorphism in the CYP17 gene is associated with prostate cancer risk. Int J Cancer 2000; 87: 434 437. 9 Wadelius M et al. Prostate cancer associated with CYP17 genotype. Pharmacogenetics 1999; 9: 635 639. 10 Haiman CA et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and prostate cancer. Cancer Epidemiol Biomarkers Prev 2001; 10: 743 748. 11 Habuchi T et al. Increased risk of prostate cancer and benign prostatic hyperplasia associated with a CYP17 gene polymorphism with a gene dosage effect. Cancer Res 2000; 60: 5710 5713. 12 Febbo PG et al. The V89L polymorphism in the 5 alpha-reductase type 2 gene and risk of prostate cancer. Cancer Res 1999; 59: 5878 5881. 13 Nam RK et al. V89L polymorphism of type-2, 5-alpha reductase enzyme gene predicts prostate cancer presence and progression. Urology 2001; 57: 199 204. 14 Hsing AW et al. Polymorphic markers in the SRD5A2 gene and prostate cancer risk: a population-based case-control study. Cancer Epidemiol Biomarkers Prev 2001; 10: 1077 1082. 15 Makridakis N et al. A prevalent missense substitution that modulates activity of prostatic steroid 5 alpha-reductase. Cancer Res 1997; 57: 1020 1022. 16 Li Z et al. Association of V89L SRD5A2 polymorphism with prostate cancer development in a Japanese population. JUrol 2003; 169: 2378 2381. 17 Carey AH et al. Polycystic ovaries and premature male pattern baldness are associated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet 1994; 3: 1873 1876. 18 Feigelson HS et al. Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 1998; 58: 585 587. 19 Haiman CA et al. The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer. Cancer Res 1999; 59: 1015 1020. 20 Ye Z, Parry JM. The CYP17 MspA1 polymorphism and breast cancer risk: a meta-analysis. Mutagenesis 2002; 17: 119 126. 21 Feigelson HS et al. A polymorphism in the CYP17 gene increases the risk of breast cancer. The CYP17 MspA1 polymorphism and breast cancer risk: a meta-analysis. Cancer Res 1997; 57: 1063 1065. 22 Allen NE, Forrest MS, Key TJ. The association between polymorphisms in the CYP17 and 5 alpha-reductase (SRD5A2) genes and serum androgen concentrations in men. Cancer Epidemiol Biomarkers Prev 2001; 10: 185 189. 23 Walsh PC, Wilson JD. The induction of prostatic hypertrophy in the dog with androstanediol. J Clin Invest 1976; 57: 1093 1097. 24 Gann PH et al. Prospective study of sex hormone levels and risk of prostate cancer. J Natl Cancer Inst 1996; 88: 1118 1126. 25 Vermeulen A, Verdonck G. Representativeness of a single point plasma testosterone level for the long term hormonal milieu in men. J Clin Endocrinol Metab 1992; 74: 939 942. 26 Makridakis NM, di Salle E, Reichardt JK. Biochemical and pharmacogenetic dissection of human steroid 5 alpha-reductase type II. Pharmacogenetics 2000; 10: 407 413. 27 Ross RK, Henderson BE. Do diet and androgens alter prostate cancer risk via a common etiologic pathway? J Natl Cancer Inst 1994; 86: 252 254. 28 Oishi K, Yoshida O, Schroeder FH. The geography of prostate cancer and its treatment in Japan. Cancer Surv 1995; 23: 267 280. 337