Regulation of Aromatase and 5Α-Reductase by 25-Hydroxyvitamin

Journal of Steroid Biochemistry & Molecular Biology 94 (2005) 151–157

Regulation of aromatase and 5␣-reductase by 25-hydroxyvitamin D3, 1␣,25-dihydroxyvitamin D3, dexamethasone and progesterone in prostate cancer cells

Yan-Ru Lou a, ∗, 1, Teemu Murtola a, 1, Pentti Tuohimaa a, b

a Department of Anatomy, Medical School, University of Tampere, FIN-33014Tampere, Finland. b Department of Clinical Chemistry, Tampere University Hospital, Tampere, Finland.

Abstract

Estrogens and androgens are proposed to play a role in the pathogenesis of prostate cancer. The effective metabolites, estradiol and 5␣- dihydrotestosterone are produced from testosterone by aromatase and 5␣-reductase, respectively. Metabolites of vitamin D have shown to inhibit the growth of prostate cancer cells. The aim of the present study was to verify whether 25-hydroxyvitamin D3 (25OHD3), 1␣,25- dihydroxyvitamin D3 [1␣,25-(OH)2D3], dexamethasone, and progesterone regulate the expression of aromatase and 5␣-reductase in human prostate cancer cells. LNCaP and PC3 cells were treated with 25OHD3,1␣,25-(OH)2D3, dexamethasone, or progesterone. Aromatase and 5␣-reductase mRNA was quantiﬁed by real-time RT-PCR and aromatase enzyme activity was measured by the [3H] water assay. Aromatase enzyme activity in LNCaP and PC3 cells was increased by both 10 nM dexamethasone, 1–100 nM 1␣,25-(OH)2D3 and 100 nM–10 ␮M progesterone. The induction was enhanced when hormones were used synergistically. Real-time RT-PCR analysis showed no regulation of the expression of aromatase mRNA by any steroids tested in either LNCaP or PC3 cells. The expression of 5␣-reductase type I mRNA was not regulated by 1␣,25-(OH)2D3 and no expression of 5␣-reductase type II was detected in LNCaP. © 2005 Elsevier Ltd. All rights reserved.

Keywords: Aromatase; 5␣-Reductase; Vitamin D3; Dexamethasone; Progesterone

1. Introduction esized to influence prostate growth and possible incidence of prostate cancer. The hypothesis about the role of the lo- Prostate cancer is the leading cancer affecting men in the cal estrogen production in prostate carcinogenesis has been western world. The androgen-dependence of prostate cancer partly encouraged by results achieved with aromatase knock- is well documented [1,2]. However, the mechanism of an- out mice, which had no estrogen production of their own, drogen influence is still unclear [3]. The most potent andro- but very high androgen concentration in their blood. Knock- gen, dihydrotestosterone (DHT), is produced from the ma- out mice developed benign prostatic hyperplasia very fast, jor circulating androgen testosterone by 5␣-reductase [4]. but they did not have increased prostate cancer incidence There are two isoforms of the enzyme: 5␣-reductase type [10]. I (SRD5A1) and type II (SRD5A2) [5,6]. Inhibition of 5␣- The expression of aromatase in the prostate is a controver- reductase reduces androgen action and therefore has clinical sial issue. The aromatase immunoreactivity, enzyme activity significance in the treatment of benign prostatic hyperpla- and mRNA have been shown in tissues of benign prostate sia and prostate cancer [7,8]. On the other hand, testosterone hyperplasia and prostate cancer [11,12]. The prostate can- can be converted into estrogens by aromatase [9]. The lo- cer cell lines DU145 and PC3 were also reported to express cal production of estrogens in the prostate has been hypoth- aromatase mRNA [13]. An earlier study reported aromatase activity in LNCaP cells [14] and that estrogen synthesis was ∗ strongly dependent on cell culture conditions [15].Onthe Corresponding author. Tel.: +358 3 2158942; fax: +358 3 2156170. E-mail address: loyalo@uta.fi (Y.-R. Lou). other hand, benign prostate hyperplasia and LNCaP cells are 1 These authors contributed equally to this work. reported not to express aromatase mRNA [16].

The active forms of vitamin D, 25-hydroxyvitamin 2.3. Treatment with hormones and assay of aromatase D3 (25OHD3) and 1␣,25-dihydroxyvitamin D3 [1␣,25- enzyme activity (OH)2D3] are suggested to be important factors in the de- velopment of prostate cancer. Both inhibit the growth of both The method used to measure aromatase enzyme activity prostatic epithelial cells [17] and stromal cells [18]. Only few was modified from that described previously [24]. Aromatase studies on the effect of vitamin D metabolites on 5␣-reductase enzyme activity was calculated based on the loss of [1␤-3H] and aromatase enzyme have been published. 1␣,25-(OH)2D3 androstenedione into the aqueous phase of the reaction mix- alone has no effect on aromatase enzyme activity but it stabi- ture during aromatization. The conversion rate was deter- 3 lizes aromatase mRNA synergistically with dexamethasone mined by quantification of [ H] H2O. depending on the vitamin D receptor (VDR) levels in hu- LNCaP and PC3-cells were seeded in 6-well plates at a man primary osteoblasts [19]. Dexamethasone is a synthetic density of 1.5 × 105 cells per well in phenol red DMEM/F12 steroid that belongs to the group of glucocorticoids. The anti- medium containing 5–10% FBS. At confluence, cells were inflammatory potential of dexamethasone is 10 times stronger gently washed with warm phenol red free DMEM/F12 than that of cortisol [20]. It has been reported to increase medium, and then treated with hormones at appropriate con- the VDR levels in prostate cancer cells together with 1␣,25- centrations together with 5 nM [1␤-3H] androstenedione in (OH)2D3 [21]. Glucocorticoids clearly increase aromatase phenol red free and serum free DMEM/F12 in a volume enzyme activity in different cell lines depending on the serum of 1 ml per well for 24 h. The plates were then placed on 3 concentrations used in growth medium [9]. Progesterone has ice for 15 min to condensate any evaporated [ H] H2O. been reported to inhibit aromatase enzyme activity in the Five hundred microliters of medium was taken from each ◦ 3 stroma of endometrium and in breast cancer cells [22].It well and stored at −70 C for later isolation of [ H] H2O. has also been reported to reduce the action of cortisol in Cells were immediately fixed with 11% glutaraldehyde inducing aromatase enzyme activity in adipose fibroblasts and stored at −5 ◦C for the subsequent measurement of [23]. cell density with crystal violet staining as described be- The goal of our study was to determine whether 25OHD3, low. 3 1␣,25-(OH)2D3, dexamethasone or progesterone may affect After storage samples were thawed and [ H] H2O was iso- aromatase and/or 5␣-reductase expression in prostate cancer lated by filtering the samples through Sep-Pak C18 Cartridges cells. (Waters Corporation, Dublin, Ireland), which removed 97.8% of radioactivity caused by unconverted aromatase substrate (data not shown). Radioactivity of filtered samples was then 2. Materials and methods measured with a 1450 Microbeta Plus Liquid Scintillation Counter (Wallac, Turku, Finland) after 1 ml of scintillation 2.1. Materials liquid was added into each sample to enhance the scintillation measurement. 25OHD3 and 1␣,25-(OH)2D3 were obtained from Leo Pharmaceuticals (Ballerup, Denmark), and proges- 2.4. Measurement of cell density terone from Merck (Darmstadt, Germany). DMEM/F12 medium, RPMI-1640 medium, dexamethasone, and 4- Crystal violet staining was used to measure cell den- hydroxyandrost-4-ene-3,17-dione (formestane, 4-OHA) sity in samples treated with hormones [25]. After the me- were purchased from Sigma (St. Louis, MO, USA), [1␤-3H] dia were taken, cells were fixed with 50 ␮l of 11% glu- androstenedione from NEN Life Science (PerkinElmer Life taraldehyde, washed with distilled water, air-dried, stained and Analytical Sciences, Boston, MA, USA), fetal bovine with 1 ml of 0.1% filtered crystal violet, washed, and air- serum (FBS) and penicillin-streptomycin from Gibco BRL dried. Then 1 ml of 10% acetic acid was added, and the (Groningen, The Netherlands), TRIzol Reagent from Gibco absorbance was measured by using a Victor 1420 Multil- BRL (Life Technologies, Grand Island, New York, USA). abel Counter (Wallac, Turku, Finland) at a wavelength of MPV-2213ad (Finrozole) was a kind gift from Hormos 590 nm. medicals (Turku, Finland). 2.5. Counting the aromatase enzyme activity 2.2. Cell culture All measured radioactivity and cell density values were Human prostate cancer cells LNCaP clone FGC and PC3 normalized by a no cell-control. Aromatase enzyme activ- were obtained from the American Type Culture Collec- ity appeared to be very labile, the absolute value changing tion and routinely maintained in 75 cm2 flasks with RPMI- strongly between repeats of experiments despite of identi- 1640 medium or DMEM/F12 medium, supplemented with cal culture conditions. This resulted in a skewed distribution 5–10% FBS, 100 units/ml penicillin and 100 ␮g/ml strepto- of enzyme activity values, with many outlying values. As ◦ mycin at 37 C in a humidified atmosphere of 5% CO2 in a consequence, median was chosen to describe the middle air. value of enzyme activities instead of mean, and interquar- Y.-R. Lou et al. / Journal of Steroid Biochemistry & Molecular Biology 94(2005) 151–157 153 tile range between upper (75%) and lower (25%) quartiles, components was prepared in a 20 ␮l volume: 300 ng total which consists 50% of all observed values, was used as the RNA, 0.5 ␮M HMBS primers or 0.3 ␮M SRD5A1 primers, descriptor of variability of the enzyme activity values. Since and 3.5 mM Mn2+ for HMBS or 3.25 mM Mn2+ for SRD5A1. both median and interquartile range are less dependent on the Nucleotides, Tth DNA polymerase (DNA polymerase and normal distribution of observations and outlying values, they reverse transcriptase activity), SYBR Green I, and reaction give a more truthful picture in the skew distribution of val- buffer were included in the LightCycler-RNA Master SYBR ues. Two-sided p-values were calculated using Wilcoxon’s Green I kit (Roche Diagnostics, Basel, Switzerland). HMBS test, suitable for related samples with nonparametric distri- was used as the endogenous control. The RT-PCR protocol butions of observations. The relative enzyme activities com- was as follows: 20 min reverse transcription at 61 ◦C and pared with control are expressed as medians and interquar- 30 s denaturation at 95 ◦C followed by 40 cycles with a tile ranges of at least three independent samples. All exper- 95 ◦C denaturation for 1 s, 62 ◦C for HMBS or 57 ◦C for iments were performed in triplicate. Differences of p > 0.05 SRD5A1 annealing for 7 s and 72 ◦C extension for 12 s. were considered not significant, and p < 0.05 significant (*). The accumulated fluorescent products were detected at The enzyme kinetics and response to aromatase enzyme in- the end of the extension step of each cycle. To verify hibitors was studied in one to two experiments, each compris- the specific products, melting curve analysis and gel ing three samples of each treatment. Median and range of all electrophoresis were performed. The data were quantified values were chosen to describe the distribution of measured by the Fit Points method with LightCycler Data Analysis values in these studies due to low numbers of observations. software. All statistical analyses were carried out using SPSS 10.1- The expression of aromatase mRNA was detected on program. an ABI Prism 7000 sequence detection system (Perkin- Elmer Applied Biosystems, Foster City, CA, USA). Primer 2.6. Treatments with steroid hormones and RNA design, the procedures for cDNA synthesis and quanti- isolation tative real-time PCR were as described previously [18]. The primers for human acidic ribosomal phosphoprotein LNCaP and PC3 cells were routinely cultured in phenol P0 (RPLP0) were that used earlier [18]. For aromatase red free RPMI 1640 medium containing 10% FBS. To de- (NM 000103) amplification, the forward primer was 5- plete endogenous steroids, the medium was changed to one CCAGGTCCTGGCTACTGCAT-3, corresponding to base containing 10% DCC-FBS 2–3 days before treatments. The 283–302 (exons 8 and 9) and the reverse primer was 5- cells were incubated with either vehicle (ethanol, final con- GATCCCCATCCACAGGAATCT-3, corresponding to base centration 0.05%) or hormones for 24 h. The ethanol concen- 351–331 (exon 9). All sequence-specific oligonucleotide tration was equal in controls and hormone-treated samples. primers were synthesized by TAG Copenhagen A/S (Copen- Total cellular RNA was isolated by using TRIzol Reagent fol- hagen, Denmark). lowing the manufacturer’s instructions. Total RNA amounts Results obtained from both LightCycler and ABI Prism were quantified by measuring absorbance at 260 nm. The 7000 are expressed as means (±S.D.) of three independent A260 nm/A280 nm absorption ratio was greater than 1.95. De- experiments performed in duplicate. Statistical analysis was naturing agarose gel electrophoresis was performed to verify performed by Student’s t-test. the integrity of RNA. The intensity of the 28S rRNA band was more than twice that of the 18S rRNA band stained by ethidium bromide. 3. Results 2.7. Real-time RT-PCR analysis 3.1. Kinetics of aromatase in LNCaP and PC3 cells The expression of 5␣-reductase mRNA was detected on a LightCycler instrument (Roche Diagnostics, To study the enzyme kinetics of aromatase in PC3 and Basel, Switzerland). The following primers synthe- LNCaP, cells were incubated with different concentrations sized by Amersham Pharmacia Biotech (Amersham, of aromatase substrate and/or two enzyme inhibitors, non- UK) were used for one-step RT-PCR: hydroxymethyl- competitive Finrozole and competitive 4-OHA. Saturation bilane synthase (HMBS, NM 000190) forward primer of aromatase activity was reached with substrate at 50 nM 5-AAGTGCGAGCCAAGGACCAG-3, corresponding in both cell lines (Fig. 1). In the presence of 5 nM substrate, to base 819–838 (exon 10) and the reverse primer 5- 1200 nM Finrozole caused a 25.7% decrease in the median TTACGAGCAGTGATGCCTACCAAC-3, corresponding aromatase enzyme activity (range 4.6–30.8%) and 1 ␮M4- to base 1116–1093 (exon 12); SRD5A1 (NM 001047) OHA decreased the median enzyme activity 32.1% in LNCaP forward primer 5-GGCGATTATGTTCTGTACCTG-3, cells (range 7.6–33.5%) (Fig. 2A). In PC3 cells, 800 nM of corresponding to base 484–504 (exon 2) and the reverse Finrozole decreased the median enzyme activity most, 85% primer 5-GCATAGCCACACCACTCC-3, corresponding (range 69.8–96.6%); 1 ␮M 4-OHA caused a 74% decrease to base 756–739 (exon 4). A master mix of the following (range 31.8–74.1%) (Fig. 2B). 154 Y.-R. Lou et al. / Journal of Steroid Biochemistry & Molecular Biology 94(2005) 151–157

Fig. 1. Effect of aromatase substrate concentration on aromatase enzyme activity in LNCaP (A) and PC3 (B) cells. Cells were incubated with aromatase substrate [1␤-3H] androstenedione for 24 h. Aromatase activity was assayed by incubation with 0.5–60 nM of substrate. Displayed is the median and the range of measured absolute of aromatase activity values from three Fig. 2. Effect of two aromatase enzyme inhibitors on aromatase enzyme ac- to six samples. tivity in LNCaP (A) and PC3 (B) cells. Cells were incubated simultaneously with substrate and inhibitors for 24 h. Aromatase activity was assayed by ␣ incubation with 5–15 nM substrate, 800–1500 nM Finrozole (F-zole), and 3.2. Effects of 1 ,25-(OH)2D3, dexamethasone, and 1 ␮M 4-OHA. Displayed is the median and the highest value of measured progesterone on aromatase activity in LNCaP and PC3 absolute aromatase activities from three to six samples. cells in Fig. 4. Induction of aromatase activity by combined Dexamethasone at 1 and 5 nM did not significantly af- treatment with 1␣,25-(OH)2D3 and 10 nM dexamethasone fect aromatase enzyme activity, whereas 10 nM increased was significant in LNCaP at all concentrations of 1␣,25- the enzyme activity 2.7-fold (interquartile range 1.1–4.5, (OH)2D3; 1 nM increased enzyme activity 3.1-fold (in- p = 0.018) in LNCaP cells and 4.5-fold (interquartile range terquartile range 2.4–5.9, p = 0.018), 10 nM 5.3-fold (in- ␣ 0.6–11.9) in PC3 cells, respectively (Fig. 3). 1 ,25-(OH)2D3 terquartile range 3.0–7.8, p = 0.028), 100 nM 2.1-fold (in- induced aromatase activity in both cell lines, more clearly terquartile range 1.6–4.3, p = 0.018). In PC3 the induc- in LNCaP. Significance was reached at 100 nM in LNCaP tion by 10 nM dexamethasone together with different con- cells, causing a 3.1-fold increase (interquartile range 1.7–4.7, centrations of 1␣,25-(OH) D was not so clear; statisti- ␣ 2 3 p = 0.018). All concentrations of 1 ,25-(OH)2D3 increased cal significance was not reached in any incubation. An the median enzyme activity non-significantly in PC3, 1 nM 11.5-fold increase of the median enzyme activity was ob- caused the strongest induction (8.6-fold, interquartile range served with 1 nM of 1␣,25-(OH)2D3 (interquartile range 0.7–16.5) in PC3 cells. Progesterone increased aromatase en- 2.0–22.2). Increase of the enzyme activity was greater ␮ zyme activity dose-dependently in both cell lines; at 10 M in both cell lines with combined treatment with dexam- the activity increased 4.8-fold in LNCaP cells (interquar- ethasone and 1␣,25-(OH)2D3, than with either treatment tile range 3.7–5.9) and 39.8-fold in PC3 (interquartile range alone. 25.8–53.8). Increasing concentrations of progesterone along with 10 nM dexamethasone increased the median enzyme ac- 3.3. Combined effect of dexamethasone with tivity strongly though non-significantly in both cell 1␣,25-(OH)2D3 and progesterone on aromatase activity lines; 1 ␮M progesterone increased the median activ- in LNCaP and PC3 cells ity 13.4-fold in LNCaP (interquartile range 2.9–23.8), 10 ␮M progesterone increased the median activity 213.6- The results of experiments using dexamethasone to- fold in PC3 (interquartile range 11.5–415.6). The en- gether with 1␣,25-(OH)2D3 and progesterone are shown zyme activity increased in a dose-dependent manner Y.-R. Lou et al. / Journal of Steroid Biochemistry & Molecular Biology 94(2005) 151–157 155

Fig. 3. Effect of dexamethasone (dexa), 1␣,25-(OH)2D3 (1,25-VD) and progesterone (prog.) on aromatase enzyme activity in LNCaP (A) and PC3 (B) cells. Activity was assayed after 24 h simultaneous incubation with hor- Fig. 4. Combined effect of dexamethasone (dexa) together with 1␣,25- mones and aromatase substrate, and was subsequently divided by the value (OH)2D3 (1,25-VD) and progesterone (prog.) on aromatase enzyme activity produced in cell density measurement. Medians of enzyme activity values in LNCaP (A) and PC3 (B). Activity was assayed after 24 h of simultaneous are reported on a relative scale compared to absolute activity value of control incubation with hormones and aromatase substrate, and was subsequently sample, i.e. activity of untreated control sample is 1. Each column represents divided by the value produced in cell density measurement. Medians of en- the median and upper quartile of measured aromatase activity values from zyme activity values are reported on a relative scale compared to absolute 9 to 18 samples. Two-sided p-values were calculated using Wilcoxon’s test activity value of control sample, i.e. activity of untreated control sample (*p < 0.05). is 1. Each column represents the median and upper quartile of measured aromatase activity values from 9 to 18 samples. Two-sided p-values were along with increasing concentrations of progesterone in calculated using Wilcoxon’s test (*p < 0.05). PC3. 4. Discussion 3.4. Effects of 25OHD3,1␣,25-(OH)2D3, dexamethasone, or progesterone on aromatase mRNA This study shows that both aromatase and 5␣-reductase expression in LNCaP and PC3 cells type I are expressed in both prostate cancer cell lines LNCaP and PC3. Varioussteroid hormones affect the enzyme activity To verify whether the hormone regulation of aromatase is of aromatase. at transcriptional level, the expression of aromatase mRNA The level of absolute aromatase enzyme activity value was analyzed by real-time RT-PCR. Since 25OHD3, has re- showed great lability between experiments despite of iden- cently been suggested to be an active hormone [18],wein- tical culture conditions and procedures, as reported before cluded it in this experiment. Both cell lines expressed de- [15]. This lability is well demonstrated in comparison be- tectable aromatase mRNA, which however was not altered tween Figs. 1 and 2; the same substrate concentration seemed ␣ by 500 nM 25OHD3,10nM1 ,25-(OH)2D3, 10 nM dexam- to induce a different absolute enzyme activity value in sep- ethasone or 10 nM progesterone (Fig. 5). arate experiments. However, relative changes in aromatase activity caused by hormones remained fairly constant. 3.5. Effect of 1␣,25-(OH)2D3 on SRD5A1 mRNA Glucocorticoid effect was similar to that in other cell types, expression in LNCaP cells increasing the aromatase enzyme activity [9]. We demonstrated here that 1␣,25-(OH)2D3 has an inducing effect on To study whether 1␣,25-(OH)2D3 regulates the expres- aromatase activity, and this effect was enhanced in the pres- sion of two isoenzymes of 5␣-reductase mRNA, LNCaP ence of dexamethasone. The latter situation resembles the sit- cells were treated with 10 nM 1␣,25-(OH)2D3 for 2, 6, 12, uation in vivo, as glucocorticoids are normally present in the 24, 30, and 48 h. Real-time RT-PCR showed no expression human body, and 1␣,25-(OH)2D3 cannot affect the prostate of SRD5A2 (data no shown) and unaltered expression of of a living man without the simultaneous effect of circulating SRD5A1 (Fig. 6). glucocorticoids. Therefore, it can be hypothesized that 1␣,25- 156 Y.-R. Lou et al. / Journal of Steroid Biochemistry & Molecular Biology 94(2005) 151–157

activity and the local estrogen production similarly in vivo. However, variation in the magnitude of the effect was also great, thus the subject needs further studies. RT-PCR studies showed that hormones tested do not affect the aromatase mRNA levels in prostate cancer cells. There- fore the regulation is not at the transcriptional level, unlike in fibroblasts from human adipose tissue, where induction of the aromatase gene by cortisol has been described [26], depending on the presence of serum or serum-derived growth factors [27]. It is likely that hormones tested affect the enzyme directly or by inactivating an inhibitory factor of aromatase enzyme activity. Our results show no regulation of 5␣-reductase mRNA expression by 1␣,25-(OH)2D3. This is in agreement with a recent report showing that a calcitriol analog failed to inhibit 5␣-reductase types I and II activities [28]. This suggests that a combined treatment of 1␣,25-(OH)2D3 and 5␣-reductase in- hibitor could be considered in prostate cancer therapy. We did not detect any expression of 5␣-reductase type II in LNCaP cells. Our study brings a bit more enlightenment to the complex issue of ways of androgen metabolism in the prostate cancer. However, further studies are needed. Fig. 5. Effects of various hormones on aromatase mRNA expression. LNCaP (A) and PC3 (B) cells were treated with vehicle (0.05% ethanol), 500 nM 25OHD3 (25VD), 10 nM 1␣,25-(OH)2D3 (1,25-VD), 10 nM dexamethasone (dexa) or 10 nM progesterone (prog.) for 24 h. The levels of aromatase Acknowledgements (CYP19) mRNA were quantified by real-time RT-PCR on an ABI Prism 7000 sequence detection system. The relative expression level of CYP19 mRNA in control sample is 1, n =3. This study was supported by grants from the Medical Re- search Fund of Tampere University Hospital, the Academy (OH)2D3 together with adrenal glucocorticoids increases the of Finland and the Tampere Graduate School in Biomedicine local estrogen production by inducing aromatase activity in and Biotechnology. We wish to thank Ms. Hilkka Makinen¨ the prostate similarly in vivo. and Ms. Taina Eskola for excellent technical assistance. We The effect of progesterone on aromatase enzyme activity also thank Leo Pharmaceuticals (Ballerup, Denmark) for genin prostate cancer cells is the opposite compared to breast erous gifts of 1␣,25-(OH)2D3 and 25OHD3 and Hormos cancer cells and the stroma of endometrium [22]. Proges- Medicals (Turku, Finland) for donating Finrozole. terone increased the aromatase enzyme activity strongly in both prostate cancer cell lines, even in a dose-dependent manner. Combining progesterone and dexamethasone treatments References heightened this effect. Therefore it can be hypothesized that progesterone would affect the prostatic aromatase enzyme [1] A.W. Hsing, J.K. Reichardt, F.Z. Stanczyk, Hormones and prostate cancer: current perspectives and future directions, Prostate 52 (2002) 213–235. [2] T. Shaneyfelt, R. Husein, G. Bubley, C.S. Mantzoros, Hormonal predictors of prostate cancer: a meta-analysis, J. Clin. Oncol. 18 (2000) 847–853. [3] M. Algarte-Genin, O. Cussenot, P. Costa, Prevention of prostate cancer by androgens: experimental paradox or clinical reality, Eur. Urol. 46 (2004) 285–294, discussion 294–285. [4] J.D. Wilson, M.W. Leihy, G. Shaw, M.B. Renfree, Androgen physiology: unsolved problems at the millennium, Mol. Cell. Endocrinol. 198 (2002) 1–5. [5] S. Andersson, D.W. Russell, Structural and biochemical properties of cloned and expressed human and rat steroid 5 alpha-reductases, Proc. Natl. Acad. Sci. U.S.A. 87 (1990) 3640–3644. Fig. 6. Effects of 1␣,25-(OH)2D3 on 5␣-reductase type I mRNA expres- [6] S. Andersson, D.M. Berman, E.P. Jenkins, D.W. Russell, Deletion sion. LNCaP cells were treated with vehicle (0.1% ethanol) or 10 nM 1␣,25- of steroid 5 alpha-reductase 2 gene in male pseudohermaphroditism, (OH)2D3 (1,25-VD) for 2, 6, 12, 24, 30, and 48 h. The levels of 5␣-reductase Nature 354 (1991) 159–161. type I mRNA were quantified by real-time RT-PCR on a LightCycler instru- [7] C.L. Foley, R.S. Kirby, 5 alpha-reductase inhibitors: what’s new? ment, n =3. Curr. Opin. Urol. 13 (2003) 31–37. Y.-R. Lou et al. / Journal of Steroid Biochemistry & Molecular Biology 94(2005) 151–157 157

[8] W.D. Steers, 5 alpha-reductase activity in the prostate, Urology 58 [18] Y.R. Lou, I. Laaksi, H. Syval¨ a,¨ M. Blauer,¨ T.L. Tammela, T. Ylikomi, (2001) 17–24, discussion 24. P. Tuohimaa, 25-Hydroxyvitamin D3 is an active hormone in human [9] E.R. Simpson, G.E. Ackerman, M.E. Smith, C.R. Mendelson, Estro- primary prostatic stromal cells, FASEB J. 18 (2004) 332–334. gen formation in stromal cells of adipose tissue of women: induc- [19] S. Tanaka, M. Haji, R. Takayanagi, Y. Sugioka, H. Nawata, 1,25- tion by glucocorticosteroids, Proc. Natl. Acad. Sci. U.S.A. 78 (1981) Dihydroxyvitamin D3 enhances the enzymatic activity and expres- 5690–5694. sion of the messenger ribonucleic acid for aromatase cytochrome [10] G.P. Risbridger, J.J. Bianco, S.J. Ellem, S.J. McPherson, Oestrogens P450 synergistically with dexamethasone depending on the vitamin and prostate cancer, Endocr. Relat. Cancer 10 (2003) 187–191. D receptor level in cultured human osteoblasts, Endocrinology 137 [11] M. Hiramatsu, I. Maehara, M. Ozaki, N. Harada, S. Orikasa, H. (1996) 1860–1869. Sasano, Aromatase in hyperplasia and carcinoma of the human [20] A.C. Guyton, J.E. Hall, Functions of the glucocorticoids, in: A.C. prostate, Prostate 31 (1997) 118–124. Guyton, J.E. Hall (Eds.), Textbook of Medical Physiology, W.B. [12] M. Tsugaya, N. Harada, K. Tozawa, Y. Yamada, Y. Hayashi, S. Saunders Company, Philadelphia, 1996, pp. 962–967. Tanaka, K. Maruyama, K. Kohri, Aromatase mRNA levels in be- [21] C.S. Johnson, P.A. Hershberger, D.L. Trump, Vitamin D-related ther- nign prostatic hyperplasia and prostate cancer, Int. J. Urol. 3 (1996) apies in prostate cancer, Cancer Metastasis Rev. 21 (2002) 147–158. 292–296. [22] C.L. Clarke, R.L. Sutherland, Progestin regulation of cellular prolif- [13] P. Negri-Cesi, A. Colciago, A. Poletti, M. Motta, 5alpha-reductase eration, Endocr. Rev. 11 (1990) 266–301. isozymes and aromatase are differentially expressed and active in [23] M. Schmidt, C. Renner, G. Loffler, Progesterone inhibits the androgen-independent human prostate cancer cell lines DU145 glucocorticoid-dependent aromatase induction in human adipose fi- and PC3, Prostate 41 (1999) 224–232. broblasts, J. Endocrinol. 158 (1998) 401–407. [14] J.L. Block, N.L. Block, B.L. Lokeshwar, Inhibition of aromatase ac- [24] E.D. Lephart, E.R. Simpson, Assay of aromatase activity, Methods tivity and growth suppression by 4-methoxy-4-androstene-3,17-dione Enzymol. 206 (1991) 477–483. in an androgen sensitive human prostatic carcinoma cell line, Cancer [25] W. Kueng, E. Silber, U. Eppenberger, Quantification of cells cultured Lett. 101 (1996) 143–148. on 96-well plates, Anal. Biochem. 182 (1989) 16–19. [15] L.A. Castagnetta, O.M. Granata, V. Bellavia, R. Amodio, E. Scac- [26] E.R. Simpson, M.S. Mahendroo, G.D. Means, M.W. Kilgore, M.M. cianoce, M. Notarbartolo, M.R. Follari, M.D. Miceli, G. Carruba, Hinshelwood, S. Graham-Lorence, B. Amarneh, Y. Ito, C.R. Fisher, Product of aromatase activity in intact LNCaP and MCF-7 human M.D. Michael, Aromatase cytochrome P450, the enzyme responsible cancer cells, J. Steroid Biochem. Mol. Biol. 61 (1997) 287–292. for estrogen biosynthesis, Endocr. Rev. 15 (1994) 342–355. [16] P. Negri-Cesi, A. Poletti, A. Colciago, P. Magni, P. Martini, M. [27] M. Schmidt, G. Loffler, Induction of aromatase in stromal vascu- Motta, Presence of 5 alpha-reductase isozymes and aromatase in hu- lar cells from human breast adipose tissue depends on cortisol and man prostate cancer cells and in benign prostate hyperplastic tissue, growth factors, FEBS Lett. 341 (1994) 177–181. Prostate 34 (1998) 283–291. [28] C. Crescioli, P. Ferruzzi, A. Caporali, R. Mancina, A. Comerci, M. [17] D.M. Peehl, R.J. Skowronski, G.K. Leung, S.T. Wong, T.A. Stamey, Muratori, M. Scaltriti, G.B. Vannelli, S. Smiroldo, R. Mariani, D. D. Feldman, Antiproliferative effects of 1,25-dihydroxyvitamin D3 Villari, S. Bettuzzi, M. Serio, L. Adorini, M. Maggi, Inhibition of on primary cultures of human prostatic cells, Cancer Res. 54 (1994) spontaneous and androgen-induced prostate growth by a nonhyper- 805–810. calcemic calcitriol analog, Endocrinology 144 (2003) 3046–3057.