Inhibition of 5a-Reductase, Receptor Binding, and Nuclear Uptake of Androgens in the Prostate by a 4-Methyl-4-aza-steroid*

Transcription

1 THE JOURNAL OF BOLOGCAL CHEMSTRY Vol. 256, No. 15, ssue of August 1, pp , 191 Prrnted in 1J.S A nhibition of 5a-Reductase, Receptor Binding, and Nuclear Uptake of Androgens in the Prostate by a -Methyl--aza-steroid* Tehming Liang and C. Elizabeth Heiss (Received for publication, December 12, 19, and in revised form, April 13, 191) From the Department of Biochemical Endocrinology, Merck Sharp and Dohme Research Laboratories, Rahway, New Jersey /3-N,N-Diethylcarbamoyl--methyl--aza-5a-~n- the prostate, the seminal vesicle, the epididymis, and skin (1- drostan-3-one (DMAA) is a potent reversible inhibitor of 5a-reductase. The inhibition by DMAA of the conversion of testosterone to Sa-dihydrotestosterone by rat prostate 5a-reductase is competitive with testoster- one, the apparent Ki being 5 n ~ and, uncompetitive with NADPH. DMAA inhibited both membrane-bound and solubilized 5a-reductase. DMAA has moderate affinity for the prostate cytosol androgen receptor: 3 X lo- M gives 5% inhibition of the binding of lo- M 5a-[3H]dihydrotestosterone to this receptor. This affinity to the androgen receptor is 1,-, 5-, 12-, and O-fold lower than that of 5adihydrotestosterone, testosterone, spironolactone, and cyproterone acetate, respectively, and 7-fold higher than that of cimetidine. After incubation of [3H]testosterone with minced prostate, more than 9% of the radioactivity extracted from the nuclei co-chromatographed with Ba-dihydrotestosterone and the rest with testosterone. DMAA at low concentrations decreased the ratio of Ba-dihydrotestosterone to testosterone in the nuclei without significantly reducing the total uptake. DMAA at high concentrations also reduced the total radioactivity in the nuclei. This differential effect may reflect a higher affinity of DMAA for 5a-reductase than for the androgen receptor. When 6ru-[3H]dihydrotestosterone was used in the tissue incubations, all radioactivity extracted from nuclei co-chromatographed with 5a-dihydrotestosterone, regardless of whether or not DMAA was present. This nuclear uptake of Sa-dihydrotestosterone is inhibited only by high concentrations ofdmaa. n a cell-free system, the nuclear uptake of 5a-[3H]dihydrotestosterone prebound to the cytosol receptor was not inhibited by DMAA. These results suggest that DMAA may inhibit nuclear uptake of Sa-dihydrotestosterone by inhibiting the receptor binding. Sucrose gradient centrifugation of the radioactive KC nuclear extracts prepared from the tissue incubations showed that the nuclear rwtestosterone-recep- ). Testosterone, the major circulating androgen in adult males, is converted to 5a-dihydrotestosterone in the extranuclear compartment of these target cells. Sa-Dihydrotestosterone then binds to the androgen receptor protein in the cytoplasm, the 5a-dihydrotestosterone-receptor complex enters the nuclei, and RNA synthesis is eventually stimulated (5,6). n tissues which lack 5a-reductase, such as skeletal muscle (7), testosterone may bind to the androgen receptor and act directly. 5a-Reductase deficiency is associated with male pseudohermaphroditism accompanied by low levels of Sa-dihydrotestos- terone (,9). On the other hand, patients with benign prostatic hyperplasia have been reported to have a higher than normal concentration of Sa-dihydrotestosterone in the prostate (1, 11). 5a-Dihydrotestosterone has also been implicated in acne (12), female hirsutism (13), and male-pattern baldness (1). Thus, inhibition of 5a-reductase to reduce the formation of 5a-dihydrotestosterone may be a useful way of minimizing certain androgen-responsive conditions. A novel 5a-reductase inhibitor has recently been developed (15). 17~-N,N-Diethylcarbamoyl--methyl--aza-5a-androstan-3-one (Fig. 1) strongly inhibits 5a-reductase-mediated conversion of testosterone to 5a-dihydrotestosterone both in vitro and in uiuo. njection of DMAA into intact male rats results in a decrease in 5a-dihydrotestosterone and an increase in testosterone levels in the prostate. Pretreatment of castrated male rats (2 g body weight) with 1 mgofdmaa followed by testosterone propionate or dihydrotestosterone propionate causes a marked inhibition of 5a-dihydrotestosterone accumulation in the ventral prostates in rats given testosterone propionate but not in those given dihydrotestosterone propionate. DMAA at 1 mg/rat attenuates prostatic accumulation of both testosterone and 5a-dihydrotestosterone. n castrated rats, DMAA at low doses selectively inhibits the growth of the prostate and the seminal vesicles induced by testosterone or testosterone propionate, but not those induced by 5a-dihydrotestosterone or dihydrotestosterone propionate. However, DMAA at high doses inhibits tissue growth induced tor complex has a greater rate of dissociation than does by testosterone, 5a-dihydrotestosterone, and their propionate the nuclear 5a-[3H]dihydrotestosterone-receptor com- derivatives (15). plex. [3H]Testosterone prebound to the prostate cytosol receptor also dissociates faster than 5t~-[~H]dihydrotestosterone prebound to the cytosol receptor. The abbreviations and trivial names used are: testosterone (17phydroxy--androsten-3-one); 5a-dihydrotestosterone (5a-androstan- 17P-ol-3-one); DMAA, 7~-N,N-diethylcarbaoyl--methyl--aza- 5a-androstan-3-one; cortisol (-pregnen-ll~,l7a,21-triol-3,2-dione); 5a-Reductase (NADPH: A*-3-oxosteroid-5a-oxidoreduc- estradiol (1,3,5( lo)-estratriene-3,17fl-diol); progesterone (-pregnentase) is present in many androgen-sensitive tissues, such as 3,2-dione); cimetidine (N -cyano-n-methyl-n [2-(5-methyl-imidazol--yl)methylthioethyl]guanidine; spironolactone (7aacetylthio)- * The costs of publication of this article were defrayed in part by 17~-hydroxy-3-oxo-pregn--ene-2l-carboxylic acid); 5P-dihydrotesthe payment of page charges. This article must therefore be hereby tosterone (5/3-androstan-17P-ol-3-one); cyproterone acetate (17a-acemarked aduertisement in accordance with 1 U.S.C. Section 173 toxy-6-chloro-l,2a-methylene-,5-pregnadiene-3,2-dione~; HPLC, solely to indicate this fact. high pressure liquid chromatography. 799

2 nhibitor Sa-Reductase 7999 FG. 1. The chemical structure of 17/3-NJV-diethylcarbamoyl--methyl--aza-6a-androstan-3-one. n this communication, we report our studies on the kinetics of inhibition of 5a-reductase by DMAA, the effect of DMAA on the receptor binding of androgens, and the inhibition of nuclear uptake of androgens by DMAA. Our results suggest that the primary action of DMAA is the inhibition of 5areductase, although it also inhibits the receptor binding of androgens when used at high concentrations. EXPERMENTAL PROCEDURES Materials 5a-[3H]Dihydrotestosterone (5 to 2 Ci/mmol), [3H]testosterone ( to 1 Ci/mmol), and Aquasol-2 were from New England Nuclear. Cortisol and estradiol were purchased from Calbiochem. Progesterone, -androstene-3,17-dione, 5a-dihydrotestosterone, and testosterone were obtained from Steraloids. Cimetidine was given to us by Dr. W. A. Bolhofer of this institute. 17P-N,N-Diethylcarbamoyl-- methyl--aza-5a-androstan-3-one, 5P-dihydrotestosterone, and spironolactone were provided by Dr. G. H. Rasmusson of this institute. Cyproterone acetate was obtained from Schering. Cell culture Medium 199 and CMRL-166 were obtained from Grand sland Biological Co. Poly/Bed 12 was purchased from Polyscience. Preparation of Prostatic Sa-Reductase Twenty mature Sprague-Dawley male rats (body weight, 35 to g) were killed by cervical dislocation. The ventral prostates were dissected free of their capsules and their combined volume was measured by displacement in several milliliters of ice-cold medium A (.32 M sucrose,.1 mm dithiothreitol, and 2 mm sodium phosphate, ph 6.5). Unless specified, all the following procedures were carried out at - "C. The prostates were drained, minced, and then homogenized in 3 tissue volumes of medium A with either a Brinkmann Polytron homogenizer or a Potter-Elvehjem glass-glass homogenizer. The homogenate was fractionated by differential centrifugations at 1, X g for 1 min, and 1, X g for 6 min. The resulting pellets were washed with 2 pellet volumes of medium A, resuspended with a syringe in medium A, and combined. The combined suspension (6 mg of protein/ml) was used as the source of 5a-reductase and was stored at -7 "C. The 1, X g supernatant contained little 5areductase and was discarded. Solubilization of &Reductase Solubilization was achieved with digitonin and salt, similar to the procedure described by Moore and Wilson (16). Rat ventral prostates were obtained, homogenized, and the homogenate centrifuged as outlined above. The pellet fractions were suspended in 2 pellet volumes of a solution consisting of digitonin (5 mg/ml), 2 M NaCl, % glycerol, 1 mm dithriothreitol, 1 mm EDTA, and 1 mm sodium phosphate, ph 6.5, at "C for 6 min. The suspension was centrifuged at 15, X g for 6 min. The supernatant which contained solubilized 5a-reductase was stored at -7 "C after addition of NADPH to 5 mm. Sa-Reductase Assay A complete reaction mixture included 1 mm dithiothreitol, mm sodium phosphate, ph 6.5, [3H]testosterone (. pci,. x O-' M), nonradioactive testosterone, NADPH, and the enzyme preparation (1 mg of protein) in a final volume of.5 ml. The concentration of testosterone contributed by [3H]testosterone was negligible. The concentrations of nonradioactive testosterone and NADPH are specified in the individual experiments. The reaction was started by the addition of either the enzyme or NADPH. ncubation was carried out at 37 "C for 1 min, unless otherwise specified. To stop the reaction, 2 ml of ethyl acetate was added and the tube blended on a Vortex mixer. A carrier steroid mixture of testosterone, 5a-dihydrotestosterone, and -androstene-3,17-dione (O pg each) in 1 pl of ethanol was added and the tube was blended on a Vortex mixer again. The ethyl acetate phase was collected and flushed with Nz to dryness. The steroids were redissolved in 3 pl of ethyl acetate and spotted on a Whatman Silica GF plate. The plate was developed in ethyl acetate: cyclohexane (1:l) at room temperature. The plate was viewed under a UV lamp to locate testosterone and -androstene-3j7-dione. Since 5a-dihydrotestosterone cannot be visualized under UV, it was located by spraying the plate with a 1% CeS/1% HzSO, solution followed by heating on a hot plate to develop the color. Testosterone and - androstene-3j7-dione are also stained by this procedure. The mobility of these three steroids is in the following order: 5a-dihydrotestosterone > -androstene-3,17-dione > testosterone. After numerous experiments, we found that this staining procedure causes some vaporization of steroids. As a result, the recovery of radioactivity varied (2 to 7%) greatly. Since 5a-dihydrotestosterone migrates slightly faster than -androstene-3,17-dione, estimate of its location on the TLC plate can be made from the position of -androstene-3,17- dione. The staining procedure was continued on plates spotted with steroid standard to establish the relative mobility of these steroids, but was not used for experimental plates. At the initial phase of our work, each plate was scraped as.5-cm fractions and the radioactivity was determined. The recovery of radioactivity on the plate was 7 to 5% of the radioactivity in the incubation mixture. The amount of 5adihydrotestosterone formed was normalized to 1% recovery. Since essentially all radioactivity moved with testosterone and 5a-dihydrotestosterone, in most of our studies, only the testosterone and 5adihydrotestosterone areas and the zone between them were scraped to determine radioactivity. Androstane-3a,l7P-diol, androstane- 3p,17P-diol, and -androstene-3,17-dione chromatographed between testosterone and 5a-dihydrotestosterone and no significant amount of these steroids was formed. Both an HPLC analysis (see below) and a thin layer chromatography analysis, using silica plates predipped in AgN3 (% in water/acetone (1:9)), separated the diols from testos- terone and 5a-dihydrotestosterone with base-line resolution and showed that the amount of the diols formed was % of the 5adihydrotestosterone formed whether or not DMAA was present during the reaction. The data are presented as averages of duplicate determinations. Androgen Receptor-binding Assay Mature Sprague-Dawley or Holtzmann male rats were castrated for 2 to h before being killed. Ventral prostates were removed, minced, and homogenized as described for the preparation of 5areductase, except that the buffer was replaced with ET buffer (1 mm EDTA and 2 mm Tris.HC1, ph 7.5). The ph of Tris buffer was measured at 25 "C unless specified. Experiments were then carried out with one of the following procedures. Sucrose Gradient Centrifugation Method-The prostate homogenate was centrifuged at 1, X g for 1 min. Aliquots of the resulting supernatant (3 mg of protein) were incubated with 1 nm 5a-[3H]dihydrotestosterone with or without competitors for receptor binding at "C for 6 min in ET buffer in a final volume of 2 ml. Competitors were dissolved and diluted to desired concentrations with ethanol and 1 pl were added. n control tubes, 1 pl of ethanol alone was added. The incubation mixture was then centrifuged at 1,ooO X g for 6 min to yield cytosol. Saturated ammonium sulfate in ET buffer, final ph 7. at "C,was added to cytosol to 35% saturation (1. ml) to precipitate 5a-[3H]dihydrotestosterone-receptor complex. After incubating at "C for 3 min with occasional shaking, the precipitate was collected by a centrifugation at 1, X g for 1 min. Ammonium sulfate precipitation separates 5a-["H] dihydrotestosterone bound to the androgen receptor from most free steroid and from that bound to a low affinity non-receptor protein (17). The pellet was redissolved in.1 ml of ET buffer and layered on a 5 to 1% linear sucrose gradient containing. M KC1 and ET buffer. After centrifuging at 3, X g for 1 h at -2 "C, the sucrose gradient was fractionated into.2-ml fractions with an SCO gradient fractionator and the radioactivity of each fraction was determined. The sedimentation coefficient of 5a-[3H]dihydrotestosterone-receptor complex was estimated from its relative mobility with bovine serum albumin by the method of Martin and Ames (1) using bovine serum

3 nhibitor 5a-Reductase albumin =.) as the standard (19). Dextran-Charcoal Method-The prostate homogenate was centrifuged at 1, X g for 2 min. Aliquots of the resulting supernatant (2 mg of protein) were incubated in duplicate with 1 nm 5a-r3H] dihydrotestosterone (123 Ci/mmol) with or without competitors. The incubation was at "C for 9 min in a final volume of 2 ml containing ET buffer. Saturated ammonium sulfate in ET buffer was added to 35% saturation and the mixture was further incubated for 9 min for precipitation of the 5a-[3H]dihydrotestosterone-receptor complex. As mentioned above, this step eliminates the nonspecific binding. The precipitate was collected by a centrifugation at 1, X g for 1 min. The pellet was redissolved with. ml of ET buffer and was mixed with.2 ml of a charcoal suspension ( mg) containing dextran (.2 mg), gelatin (.2 mg), and ET buffer. After incubation at "C for 1 min, the mixture was centrifuged at 1,5 X g for 5 min to pellet the charcoal. One-half ml of the supernatant was taken to determine the receptor-bound radioactivity. This combination of using a subsaturating level of 5a-[3H]dihydrotestosterone, ammonium sulfate fractionation, and charcoal treatment essentially eliminates all the nonspecific binding (the binding that is not inhibited by micromolar nonradioactive 5cr-dihydrotestosterone). ncubation of Minced Prostate Mature male Sprague-Dawley rats (35 to g) were castrated 1 to 2 h before being killed. Ventral prostates were removed and minced. DMAA was dissolved and diluted with ethanol to the desired concentrations. Five p1 of DMAA or ethanol (control) was added to a 25 Erlenmeyer flask containing 3.7 ml of cell culture Medium 199. Aliquots (1 ml) of the minced prostate were allocated to flasks with a wide mouth pipette. The tissue suspension was incubated in a shaking water bath at 37 "C for 5 min. The flasks were divided into two sets. One set received 1 pci of 5a-[3H]dihydrotestosterone while 1 pci of [3H]testosterone was added to the other. The radioactive steroids (1 pci in 1 p.l of ethanol) were fmt diluted with.3 ml of Medium 199 before they were added to the flasks. The final incubation volume was 5 ml. n some experiments, cell culture Medium 199 was replaced with cell culture medium CMRL-166 or medium B (.32 M sucrose, 3 mmmgc12 and 2 mm Tris.HC1, ph 7.5). The flask was flushed with 95% 2/5%COZ, covered with parafilm, and further incubated in a shaking water bath at 37 "C for min. The tissue suspension was chilled in an ice bath and then transferred to a conical test tube. The tubes were centrifuged in a clinical centrifuge at 1, x g for 1 min. The supernatant was decanted. The tissue pellet was washed twice with 1 ml of medium B and homogenized in ml of medium B using a glass-glass homogenizer. The purification of nuclei was similar to that described by Liao et al. (2). The homogenate was mixed with 3 ml of 2.2 M sucrose containing 1 mm MgC12 and 2 m~ Tris.HC1, ph 7.5, and the mixture was layered on top of 5 ml of the same solution placed in a centrifuge tube. The tubes were centrifuged at 1, X g for 6 min in a Beckman SW 27 rotor. The supernatant was decanted and the tube wall was carefully wiped with tissues. The nuclear pellet was suspended in 5 ml of.25% Triton X-1 in medium B and the suspension was centrifuged at 1, X g for 1 min in a Sorvall centrifuge. The pellet was then suspended with 3 ml of ET buffer and the suspension was centrifuged at 11,7 X g for 1 min. The resulting pellet was placed in. ml of.6 M KC1 containing ET buffer and dispersed by a syringe bearing an 1-gauge needle. The suspension was incubated at "C for min and then centrifuged at 11,7 X g for 1 min. The radioactivity in the supernatant (KC1 nuclear extract) and in the pellet was determined. The KC1 nuclear extract contained 6 to 5% of the radioactivity in nuclei. ncubation with DMAA did not appear to affect the extent of extraction. To examine the purity of the nuclear preparation at the steps before and after Triton X-1 washing, the nuclear pellets were fixed in cold 3% phosphate-buffered glutaraldyhyde, rinsed in.1 M phosphate buffer (ph 7.2), and post-fixed in cold 1% osmium tetroxide. The nuclei were then rinsed in phosphate buffer, and dehydrated in increasing concentrations of ethanol (5% to absolute ethanol), cleared in propylene oxide, and infdtrated with increasing concentrations of Poly/Bed 12 and propylene oxide (5:5,91, 1% Poly/ Bed 12). Labeled nuclei in capsules were prepared and the epoxyembedding medium polymerized at 37 "C for 17 h followed by 6 "C for 2 h. The sample was sectioned, stained with a uranyl acetate and lead citrate solution, and examined under a Zeiss-95 transmission electron microscope. Before being washed with Triton, the outer layer of the nuclear membrane and some cytoplasmic membranes were clearly seen. These membranes were essentially removed by washing with the detergent, confiing the observation of Anderson et al. (21). Nuclear Uptake of 5a-fH]Dihydrotestosterone-Receptor Complex in Cell-free System The 5a-[3H]Dihydrotestosterone-receptor complex was prepared as described for the "sucrose gradient centrifugation method" except lo-* M 5a-[3H]dihydrotestoster~ne was used in the binding incubation, and after ammonium sulfate fractionation, the complex was desalted by chromatography on a Sephadex G-25 column equilibrated with ET buffer. The nuclei were isolated from a prostate homogenate from castrated rats as described above, but the nuclei were not washed with Triton. The 5a-[3H]dihydrotestosterone-receptor complex (5, cpm, 2.3 mgof protein) and DMAA in 1 pl of ethanol (final concentration M and M) were mixed and incubated with the purified nuclei (7.5 mg of protein) in a solution containing.2 M sucrose, 1.9 mm MgCl,,. r n EDTA, ~ and 2 mm Tris-HCl, ph 7.5, in a final volume of 1.1 ml at 2 "C for 5 to 15 min. To the control tubes, 1 p1 of ethanol was added without DMAA. The nuclei were reisolated by a centrifugation at 1,5 X g for 1 min and washed with 2 ml of medium B three times. The nuclear pellet was then transferred to a vial to determine radioactivity. High Pressure Liquid Chromatography [3H]Steroids in 1 p1 of the eluting solvent (ethyl acetate:cyclohexane 1:l) were injected into a Spectra-Physics, using a Waters Associate pporasil column (3.9 mm inside diameter X 3 cm). The samples were then eluted at a flow rate of 2 ml/min at a column temperature of 31 "C and.2-ml fractions were collected to determine radioactivity. Radioactivity Determination The radioactivity was determined in 1 ml of Aquasol-2 using a Packard scintillation counter. The counting efficiency for tritium was 5%. Determination of Protein Concentrations Protein concentration was determined by the method of Bradford (22)usingCoomassie Brilliant Blue G (Bio-Rad Laboratories) for protein binding and bovine y-globulin as the standard. RESULTS DMAA decreases the rate of the 5a-reductase-catalyzed conversion of testosterone to 5a-dihydrotestosterone. At M testosterone and 5 X M NADPH, the formation of 5adihydrotestosterone was linear for at least 3 min at 37 "C both in the presence and absence of DMAA. DMAA inhibited the enzyme activity dose dependently and the inhibition reached 96% at low6 M of DMAA. For the kinetic studies to be described below, the reactions were all carried out at 37 "C for 1 min to ensure that the reaction rate was within the linear range. The inhibition of 5a-reductase activity by DMAA was ex- amined at varying concentrations of testosterone or NADPH. A Lineweaver-Burk plot of these data (Fig. Za) indicates that DMAA is a competitive inhibitor with testosterone for 5areductase. The apparent Ki for DMAA, determined from this and three similar experiments, was 5.3 f.6 X lo-' M (mean & S.D.). The apparent K, for testosterone was X M. Fig. 2b shows that DMAA is uncompetitive with NADPH in its inhibition of Sa-reductase. The apparent K, for NADPH was 2.1 f.3 X M (four experiments). These results suggest that DMAA may inhibit 5a-reductase by interacting with the testosterone binding site, but not the NADPH binding site. This may be expected as both testosterone and DMAA are 3-oxo-steroids. We also investigated DMAA inhibition of solubilized 5a- reductase. Using W 7 M [3H]testosterone, DMAA at concentrations of lo-' M, 3 X lo-' M, lop7 M, and 3 X M inhibited [3H]dihydrotestosterone formation by the solubilized enzyme

4 5a-Reductase nhibitor 1 2o - Control l l l l l l " " +5a-DHT +DMAA BSA Testosterone (PM-? -(pm") NADPH FG. 2. Lineweaver-Burk plots of the inhibition by DMAA of rat prostate 5a-reductase-catalyzed conversion of testosterone to 5a-dihydrotestosterone (5a-DHT). (a) nhibition by DMAA at varying concentrations of testosterone. The reactions were carried out in the absence (, control) and in the presence of 5 nm () and 15 nm (A) DMAA. A saturating level of NADPH (5 p ~ was ) used in the reaction. (b) nhibition by DMAA at varyingconcentrations of NADPH.Thereactionswerecarriedoutwith M testosterone in the absence (, control) and in the presence of 15 DM DMAA (). by 67%, 6%, 9%, and loo%, respectively. Thus, DMAA effectively inhibits the membrane-bound as well as the solubilized 5a-reductase. DMAA was evaluated for its effect on 5a-[3H]dihydrotestosterone binding by the androgen receptor protein of prostate cytosol. 5a-[3H]Dihydrotestosterone was incubated with prostate cytosol and the 5a-[3H]dihydrotestosterone-receptor complex was partially purified by ammonium sulfate precipitation. As previously reported (17,2), the 5a-[3H]dihydrotestosterone-receptor complex sedimented as a 3 to S radioactive peak in a sucrose gradient containing. M KC1. The receptor binding is steroid-specific. The receptor binding of 5~~-[~H]dihydrotestosterone (lo-' M) was inhibited by % by a 1-fold excess (lo-' M) of nonradioactive Sa-dihydrotestosterone, and it was not affected by the same concentration of 5/?-dihydrotestosterone or cortisol and was only slightly inhibited by estradtol (2%) and by progesterone (15%). DMAA was titrated for its ability to inhibit 5~1-[~H]dihydrotestosterone binding to the prostate cytosol androgen re- ceptor. Using M 5a-['H]dihydrotestosterone, DMAA up to M did not significantly inhibit the receptor binding of 5a-[3H]dihydrotestosterone. However, at higher concentrations, DMAA inhibited 5a-[3H]dihydrotestosterone binding to the receptor (Figs. 3 and ). The concentration of DMAA required to give 5% inhibition (C,) was estimated from the titration curves to be 2. k 1. X M (mean k S.D., four experiments). The C, for 5a-dihydrotestosterone was 2.9 k.9 X lo-' M. Thus, the androgen receptor affinity for DMAA is about 1,-fold lower than that of 5a-dihydrotestosterone. We also compared DMAA withsome antiandrogens for their ability to inhibit 5a-[3H]dihydrotestosterone binding to the androgen receptor. The results are summarized in Fig.. The C5 was estimated to be 2. X lo-' M for spironolactone, 6.7 X lo-' M for cyproterone acetate, and 2 X M for cimetidine. Thus, the androgen receptor affinity of DMAA is about 12 and times lower than that of spironolactone and cyproterone acetate, respectively, and is 7-fold higher than that of cimetidine. The antiandrogenic activity and androgen receptor-binding activity of spironolactone (23, 2), an aldosterone antagonist, cimetidine (25), a histamine type-2 receptor antagonist, and cyproterone acetate (17) have been documented. The inhibition of the nuclear uptake of 5~u-[~H]dihydrotestosterone and [3H]testosterone by DMAA was investigated in tissue incubations. Table shows that DMAA reduces the nuclear uptake of the [3H]steroid dose dependently and the E, N Fraction Number FG. 3. nhibition of prostate cytosol androgen receptor binding of 5~-[~~dihydrotestosterone by DMAA and Sa-dihydrotestosterone. The assay was carried out using the sucrose gradient centrifugation method described under "Experimental Procedures.'' The incubation for receptor binding of 5a-[3H]dihydrotestosterone (1 nm, 123 Ci/mmol) was carried out in the absence (control) and in the presence of nonradioactive 5a-dihydrotestosterone (5a- DHT, lo-' M and M) or DMAA (O-" M and M). Sedimentation was from left to right. Bovine serum albumin (BSA, 1 mg in.2 ml) was layered on one of the sucrose gradients and centrifuged in the same rotor. The peak position (measured by absorbance at 2 nm wave length) is indicated in the figure. (J.- U.- 9 b c a a al Competitor (M FG.. nhibition of prostate cytosol androgen receptor binding of 5a-[3H]dihydrotestosterone by selected compounds. The receptor binding was determined by the dextran-charcoal method as described under "Experimental Procedures." The receptor binding of 5a-[3H]dihydrotestosterone in the absence of a competitor was taken as 1%. Each point was the average of two to four experiments. 5a-DHT, 5a-dihyclrotestosterone. extent of inhibition of DMAA appears to be the same whether 5a-[3H]dihydrotestosterone or [3H]testosterone is used in the tissue incubation. HPLC analysis of the radioactive steroids in the nuclei showed that when 5a-[3H]dihydrotestosterone was used in the tissue incubation, all radioactivity co-chromatographed with the 5a-dihydrotestosterone standard. Fig. 5 shows some representative chromatograms. Sucrose gradient centrifugation of the nuclear KC1 extracts showed that most of the radioactivity sedimented as a 3.3 S radioactive peak. Some representative results are shown in Fig. 6. These

5 ~~ 2 Sa-Reductase nhibitor TABLE The effect of DMAA on the nuclear uptake of 5a-L3 Hdihydrotestosterone and [3H]testosterone in minced prostate incubations Ventral prostates from castrated rats were minced and then incubated with 1 pci of 5a-[3H]dihydrotestosterone (5a-r3H]DHT, 2 Ci/mmol) or [3H]testosterone (r3h]t, 1 Ci/mmol) in the presence or absence of DMAA. Nuclei were purified and then extracted with.6 M KC1 containing 1 rnm EDTAand 2 mm Tris.HC1, ph 7.5. Aliquots of the KC1 - extracts " were taken & for determining the a radioac-. tivityandproteinconcentrationsand for HPLCanalysis of r3h] steroids. Some representative HPLC chromatograms are shown in Fig. 5. An aliquot of the KC1 extract was also analyzed by sucrose gradient centrifugation (see text). Nuclear extract Tissue incubation ~cx-[~h]dht M) 5a-C3H]DHT DMAA 5 X Oxs M 5a-r3H]DHT DMAA 1 X M 5a-C3H]DHT DMAA 3 X M 5a-C3H]DHT + DMAA 1 X M r3h]t (1.1 x O-' M) [3H]T DMAA 5 x O-' M t3h]t DMAA 1 x O-' M [3H]T DMAA 3 X M T3H1T + DMAA 1 x Oa6 M ' Duplicate incubation. Per cent radiocpm/lw pg protein activity 5o-DHT T 1,552 f 121" 1 9, 1,6 1 5,5 1,7 1 11,37 f 351" 96, , , E, u Tg " 5a- DHT T li - ['H]5a-DHT+DMAA 5~16'M ['H 5a-DHT+DMAA 3xlCT'M results suggest that most of the 5a-[3H]dihydrotestosterone in the nuclei is associated with the receptor, and is not free. The presence of DMAA in the tissue incubation decreased the amount of the 5a-[3H]dihydrotestosterone-receptor complex (Table ). This decrease in the nuclear uptake of the 5a-13H) dihydrotestosterone-receptor complex by DMAA in the tissue incubation may be a result of inhibition of the receptor binding of 5~~-[~H]dihydrotestosterone, but not of the translocation or binding of the 5a-[3H]dihydrotestosterone-receptor complex to the nuclear acceptor site. Consistent with this suggestion, when 5a-[3H]dihydrotestosterone was prebound to the cytosol e ['H T + DMAA 5xlO'M receptor and incubated with isolated nuclei, DMAA ( W 7 M and M) had no effect on the nuclear uptake of the ~LY-[~H) dihydrotestosterone-receptor complex (data not shown). Fig. 5 and Table show that when [3H]testosterone was used in the tissue incubation, 96% of the nuclear radioactivity co-chromatographed with 5a-[3HJdihydrotestosterone on HPLC and only % eluted with [3H]testosterone. No other radioactive metabolites were found in the nuclei. The conversion of ['H]testosterone to 5a-[3H]dihydrotestosterone was FG. 5. High pressure liquid chromatography of [3H)steroids less extensive in the experiments in which cell culture Medium from nuclei of prostates incubated with ['mtestosterone or 199 was replaced with the sucrose/mg2'/tris.hc1 buffer. DMAA effectively inhibits [3H]testosterone conversion to 5a-[3H]dihydrotestosterone in the tissue incubation. This was indicated by the observation that in incubates containing 1.1 X lo-* M [3H]testosterone, DMAA (5 X LO-' to M) caused a decrease in the concentration of 5a-['H]dihydrotestosterone in the nucleiwhile the nuclear [3HJtestosterone increased (Fig. 5, Table ). Table also shows that at a given concentration ofdmaa there is a greater decrease in the 5a$H] dihydrotesto~terone/[~h]testosterone ratio than in the total nuclear radioactivity. These results indicate that 5a-reductase is more sensitive to DMAA than is the receptor binding of androgens. The DMAA inhibition of the conversion of [3H]testosterone to 5a-[3H]dihydrotestosterone was associated with a dramatic change in the sedimentation profile of the radioactive nuclear KC1 extracts. The radioactivity in the KC1 extract of the nuclei prepared from tissue incubated with ['Hltestosterone Sa-['mdihydrotestosterone with or without DMAA. The KC1 nuclear extract (.25 ml) from the experiment in Table was extracted with 1 ml of ethyl acetate and the ethyl acetate phase was flushed with N2 to dryness. r3h]steroids were then redissolved with 2 p1 of ethyl acetate:cyclohexane (1:l) and 1 p1 wm analyzed with HPLC. The [3H]steroidsandtheconcentration of DMAA in the tissue incubation are indicated in the figure. 5a-DHT and Ton the top of the figure indicate the fractions where 5a-[JH]dihydrotestosterone the and the [3H]testosterone standards eluted. in the absence of DMAA sedimented as a peak indistinguishable from that incubated with 5a-[3H]dihydrotestosterone (Fig. 6). However, when DMAA (LO-' to M) was present in the tissue incubation, the [3H]androgen-receptor complex peak decreased in proportion to the decrease in the amount of 5a-[3H]dihydrotestosterone in the nuclei (data not shown). By 5 X lo-* M DMAA, the [3H]androgen-receptor complex had largely disappeared. The disappearance of the [3H]androgen-receptor complex peak may be due to the dissociation of

6 5a-Reductase nhibitor E, u ro ' - 2 E Q m [3H]testosterone from the receptor during sucrose gradient centrifugation. We have prebound [3H]testosterone and 5a- [3H]dihydrotestosterone to the prostate cytosol androgen receptor, partially purified the [3H]steroid-receptor complexes by ammonium sulfate precipitation, and analyzed the complexes by sucrose gradient centrifugation. As shown in Fig. 7, [3HJtestosterone gradually dissociated from the receptor during the Centrifugation, whereas 5a-[3H]dihydrotestosterone remained associated with the receptor. The possibility that DMAA may act preferentially by inhibiting the androgen receptor binding of [3H]testosterone but not that of 5&H] dihydrotestosterone can be ruled out. Fig. shows that DMAA inhibited the cytosol androgen receptor binding of [3H]testosterone and 5a-[3H]dihydrotestosterone to the same extent. The concentrations of DMAA required to inhibit the receptor binding of 5a-[3H]dihydrotestosterone and [3H]tes- tosterone by 5% was 2 X lo-" M and 2.5 X O-@ M, respectively, not a significant difference. Furthermore, at the concentrations of DMAA which nearly completely inhibited testoster- 1 t TOP Froction Number FG. 7. Sucrose gradient centrifugation of [3H]testosterone and 5a-[3H]dihydrotestosterone prebound to the prostate cytosol receptor. Ventral prostates were obtained from male Holtzman rats (3 to 375 g bodyweight) castrated 2 h. Thetissue was homogenized in ET and the homogenate was centrifuged at 1, X g for 2 min. Cytosol (O mg of protein) was incubated with either 2 nm [3H]testosterone (C3H]T)or 2 nm 5a-[3H]dihydrotestosterone Fraction Number FG. 6. Sucrose gradient centrifugation of KC1 nuclear extracts of prostates incubated with either 5a-[SH]dihydrotestosterone or [3H]testosterone in the presence or absence of DMAA. Tissue incubation, purification of nuclei, and KC1 extraction ([3H]DHT) at "Cfor 9 minfor receptor binding. The specific of nuclei were as described under "Experimental Procedures" with activity of both C3H]steroid was 123 Ci/mmol. The [3H]steroid-recepthe followingchanges.medium 199 in the tissueincubationwas tor complexes were precipitated by ammonium sulfate (35% saturareplaced with medium CMRL-166. Eight pci of 5a-C3H]dihydrotes- tion) and redissolved in.1 ml of ET. The solution was layered on a tosterone (2 Ci/mmol) and [3H]testosterone (152 Ci/mmol) were 5 to 1% linear sucrose gradient containing. M KC1 and ET buffer used in the tissue incubation and the final concentrations were X and the centrifugation was at 3, X g for 1 h. The sedimentation lo-' M and 1 X lo-' M, respectively. When indicated, the concentration was from left to right. of DMAAin the tissueincubationwas W 7 M. The KC1 nuclear extract (1.1 mg of protein in.1 ml) was layered on a 5 to 1% sucrose one conversion to 5a-dihydrotestosterone, nuclear retention gradient containing. M KC1 and ET. The centrifugation was at 3, X g for 1 h at -2 "C. The sedimentation was from left to of [3H]testosterone and 5a-[3H]dihydrotestosterone were inright. Bovine serum albumin (BSA,.2 mg) was layered on a gradient hibited similarly (Table ). Fig. also shows that 5a-dihydrocontaining no nuclear extract and was centrifuged in the same rotor. testosterone is slightly (twice) more potent than testosterone The peak position of bovine serum albumin is indicated in the figure. in inhibiting the receptor binding of 5a-[3H]dihydrotestosterone and [3H]testosterone. DSCUSSON DMAA is a unique new anti-androgen with dual actions. t has high affinity for Sa-reductase (C&, 5 X lo-' M) and moderate affinity for the androgen receptor (&, 3 X 1O"j M). The differential affinity of DMAA for 5a-reductase and the androgen receptor suggests that DMAA may have a greater effect on 5a-reductase than on the receptor binding of androgens. Our studies on the effect of DMAA on the nuclear uptake of 5a-[3H]dihydrotestosterone and [3H]testosterone in prostate incubation and the results of Brooks et al. (15) on the effect of DMAA in vivo on the concentration of 5a-dihydrotestos- terone and testosterone in the rat prostate are consistent with this suggestion. However, in vivo the difference in the DMAA dosage (about lox) which selectively inhibits 5a-reductase and that which inhibits both 5a-reductase and the receptor binding is not as great as the affinity differences (about 6X) determined in the cell-free systems. t is possible that the actual concentration of DMAA in the prostate may not be

7 Sa-Reductase nhibitor roc '-\ - ", o.[~h]dht FG.. nhibition of prostate cytosol androgen receptor binding of 5a-[' rjdihydrotestosterone and [3H]testosterone by DMAA, testosterone, and Sa-dihydrotestosterone. Ventral prostates were obtained from rats (5 g body weight) castrated for 19 h. Cytosol (3 mg of protein) was incubated in duplicate with 1 nm 5a- [3H]dihydrotestosterone (C3H]DHT, 2 Ci/mrnol) and 1 nm [3H]testosterone (C3H]T, 152 Ci/mmol) with or without competitors. The amount of [3H]steroids bound to the androgen receptor was determined by the dextran-charcoal method as described under "Ex- perimental Procedures." The receptor binding in the absence of a competitor was taken as 1%. The data presented were the average of two experiments and the variation was indicated with a bar. Some concentrations of DMAA were tested in only one of the experiments and the average of duplicates was shown. The receptor-bound radioactivity in the absence of a competitor for 5a-[3H]dihydrotestosterone and for [3H]testosterone was 62 cpm and 21 cpm, respectively, for one of the experiments and was 365 cpm and 22 cpm, respectively, for the other experiment. proportional to the amount ofdmaa administered. The effectiveness of DMAA to inhibit 5a-reductase and the androgen receptor in target cells may also be further complicated by the relative concentrations of the enzyme, the receptor, 5adihydrotestosterone, and testosterone in the target cells. DMAA does not have a high affinity to proteins in rat or human serum. ncubation of [1,2-3H]DMAA (>2 Ci/mmol) with serum (16X and ~ dilution) from a mature male rat and human followed by treatment with charcoal, showed no binding of [3H]DMAA to serum proteins from either species. Under the same assay conditions, we observed good binding of [3H]progesterone,[3H]testosterone, and 5a-[3H]dihydrotestosterone to human serum proteins and little binding of these steroids to rat serum proteins. This is consistent with previous reports that human serum contains a corticosteroid-binding globulin, which also binds progesterone (26) and a sex steroidbinding globulin which binds both testosterone and 5a-dihydrotestosterone (27). DMAA is the most potent inhibitor of 5a-reductase reported to date. The most unique feature of this compound as a 5areductase inhibitor is the -methyl--aza substitution. Among the most potent inhibitors of 5a-reductase, 2a-hydroxy-- pregnen-3-one (2) has a Ki 1 order of magnitude higher than that of DMAA, and epitestosterone and progesterone (2) both have a Ki 1 to 2 orders of magnitude higher than that of DMAA. The Ki of -androstene-3-one-l7~-carboxylic acid to human skin Sa-reductase has been reported to be 1O"j M (). -Androsten-3-one-17~-carboxylic acid ( pg) has been reported to inhibit the growth of female hamster flank organ induced by testosterone propionate ( pg), but not 5a-dihydrotestosterone upon topical application (29). mplantation of this compound with testosterone in castrated rats reduced testosterone-stimulated growth of the prostate and the seminal vesicle (3). n intact male rats (2 g body weight), DMAA injected at 1 mg/rat resulted in an increase of testosterone and a decrease of 5a-dihydrotestosterone concentrations in the prostates, whereas progesterone and -androstene- 3-one-17P-carboxylic acid, even at 1 mg/rat, did not significantly affect the concentrations of testosterone and 5a-dihydrotestosterone in the prostates (15).(5a,2R)--Diazo-21-hy- droxy-2-methylpregnan-3-one has been reported to be an irreversible inhibitor of rat prostate 5a-reductase, with apparent Kc being 3.5 X lo-' M (31). The inhibition of 5a-reductase by DMAA permits a direct evaluation of relative potency between 5a-dihydrotestosterone and testosterone in the prostate. Brooks et al. (15) showed that DMAA at low dosage (1 mg/rat/day) selectively inhibited testosterone (5 to pg/rat/day)-induced growth of the ventral prostate and seminal vesicles of the castrated rats, but that it had no effect on sex accessory gland growth of 5adihydrotestosterone (5 to pg/rat/day)-treated animals. These observations support the idea that 5a-dihydrotestosterone is more potent than testosterone. This may in part be due to the fact that the receptor has higher affinity for 5a-dihydrotestosterone than testosterone, which varied from 2- to 1- fold among reports (3, 2, 32, 33). We confi the previous reports (2, 32) that the rate of dissociation of the prostate cytoplasmic testosterone-receptor complex is greater than the cytoplasmic 5a-dihydrotestosterone-receptor complex. By the use of DMAA to block the conversion of testosterone to 5adihydrotestosterone, we showed that the prostate nuclear testosterone-receptor complex also dissociates faster than the nuclear 5a-dihydrotestosterone-receptor complex (Fig. 6). By competition experiments, we observed a small but sigmficantly higher affinity of the receptor for Sa-dihydrotestosterone than for testosterone (Fig. ). This slightly higher affinity of the receptor for 5a-dihydrotestosterone than for testosterone also has been reported for other tissues including testis (3, 32), epididymis (3, 32), seminal vesicle (3, 3) and muscle (33). n some other tissues (3, 35, 36), it has been reported that testosterone has an equal or slightly higher affinity for the androgen receptor than does 5a-dihydrotestosterone. Al- though [3H]testosterone dissociates from the prostate receptor during the sucrose gradient centrifugation, the [3H]testosterone-receptor complexes from some other tissues such as pituitary (37,3) and muscle (39) have been shown to sediment as a unique peak. The inhibition of 5a-reductase by DMAA probably would not affect the biological actions of cortisol, progesterone, and aldosterone. These three A*-3-oxo-steroids appear to exert biological effects without 5a-reduction. They have a high affinity for their respective receptors and are the major steroids found in the nuclei of their target cells (5), although in the nuclei of some species dihydroprogesterone was found to be in equal proportion as progesterone (). DMAA should be useful for investigating the relative importance of 5a-dihydrotestosterone and testosterone in androgen-sensitive tissues. t may also have potential for the treatment of benign prostatic hyperplasia, acne, female hirsutism, and male-pattern baldness.

8 5a-Reductas,e nhibitor 5 Acknowledgments-We are indebted to Dr. E. H. Cordes for encouragements and thank Mrs. M. Spencer for typing the manuscript. We thank Mr. s. Scott for performing electron microscopy. REFERENCES 1. Anderson, K. M., and Liao, S. (196) Nature 219, Bruchovsky, N., and Wilson, J. D. (196) J. Biol. Chem. 23, Attramadal, A,, Weddington, S. C., Naess, O., Dj~seland, O., and Hansson, V. (1976) in Prostate Diseases (Marberger, H., ed) pp , Alan R. Liss, nc., New York. Voigt, W., and Hsia, S. L. (1973) J. Biol. Chem. 2, Liao, S. (1975) nt. Rev. Cytol Liang, T., Tymoczko, J. L., Chan, K. M. B., Hung, S. C., and Liao, S. (1977) in Androgens and Antiandrogens (Martini, L., and Motta, M., eds) pp. 77-9, Raven Press, New York 7. Krieg, M., and Voigt, K. D. (1976) J. Steroid Biochem. 7, mperato-mcginley, J., Peterson, R. E., Gautier, T., and Sturla, E. (1979) J. Steroid Biochem. 11, Griffin, J., and Wilson, J. D. (19) New Eng2. J. Med. 32, Siiteri, P. K., and Wilson, J. D. (197) J. Clin. nvest. 9, Geller, J., Albert, J., Lopez, D., Geller, S., and Niwayama, G. (1976) J. Clin. Endocrinol. Metab. 3, Sansone, G., and Reisner, R. M. (1971) J. nvest. Derrnatol. 56, Kuttenn, F., Mowszowicz,., Schaison, G., and Mauvais-Jarvis, P. (1977) J. Endocrinol. 75, Bingham, K. D., and Shaw, D. A. (1973) J. Endocrinol. 57, Brooks, J. R., Baptista, E. M., Berman, C., Ham, E. A., Hichens, M., Johnston, D. B. R., Primka, R. L., Rasmusson, G. H., Reynolds, G. F., Schmitt, S. M., and Arth, G. E., (191) Endocrinology, in press 16. Moore, R. J., and Wilson, J. D. (197) Biochemistry 13, Fang, S., and Liao, S. (1971) J. Biol. Chem. 26, Martin, R. G., and Ames, B. N. (1961) J. Biol. Chem. 236, Yphantis, D. A,, and Waugh, D. F. (1956) J. Phys. Chem. 6, Liao, S., Liang, T., Fang, S., Castaiieda, E., and Shao, T.-C. (1973) J. Biol. Chem. 2, Anderson, K. M., Lee, F. H., and Miyai, K. (197) Exp. Cell Res. 61, Bradford, M. M. (1976) Anal. Biochem. 72, Bonne, C., and Raynaud, J. P. (197) Mol. Cell. Endocrinol. 2, Pita, Tr. C., Lippman, M. E., Thompson, E. B., and Loriaux, D. L. (1975) Endocrinology 97, Funder, J. W., and Mercer, J. E. (1979) J. Clin. Endocrinol. Metab., Milgrom, E. (197) in Receptors and Hormone Action (O'Malley, W. O., and Birnbaumer, L., eds) Vol. 11, pp. 73-9, Academic Press, New York 27. Rosner, W., and Smith, R. N. (1975) Biochemistry 1, Frederiksen, D. W., and Wilson, J. D. (1971) J. Biol. Chem. 26, Voigt, W., and Hsia, S. L. (1973) Endocrinology 92, Kao, L. W. L., and Weisz, J. (1979) J. Endocrinol. 1, Blohm, T. R., Metcalf, B. W., Laughlin, M. E., Sjoerdsma, S., and Schatzman, G. L. (19) Biochem. Biophys. Res. Commun. 95, :. Wilson, E. M., and French, F. S. (1976) J. Biol. Chem. 251, Krieg, M., and Voigt, K. D. (1977) Acta Endocrinol. 6, Suppl. 21, Zakar, T., and Toth, M. (19) J. Steroid Biochem. 13, Gustafsson, J.-A,, Pousette, A., and Svensson, E. (1976) J. Biol. Chem. 251, Lieberburg,., and McEwen B. W. (1979) in Biochemical Action of Hormones (Litwack, G., ed) pp , Academic Press, New York 37. Tieulant, M.-L., and Pelletier, J. (1979) J. Steroid Biochem. 1, Armstrong, E. G., and Villee, C. A. (1977) J. Steroid Biochem., Dionne, F. T., Dube, J. Y., Lesage, L., and Tremblay, R. R. (1979) Acta Endocrinol. 91, OMalley, B. W., and Means, A. R. (197) MTP (Med. Tech. Publ. Co.) nt. Rev. Sci. Biochem. Ser.,17-21