17β-Hydroxysteroid dehydrogenase in the human prostate. Properties and distribution between epithelium and stroma in benign hyperplastic tissue

Vol. 28, No. I, pp. 3542, Printed in Great Britain. All rights reserved

J. steroid Biochem.

1987

0022-4731/87

$3.00 + 0.00

Copyright 0 1987 Pergamon Journals Ltd

17/bHYDROXYSTEROID DEHYDROGENASE IN THE HUMAN PROSTATE: PROPERTIES AND DISTRIBUTION BETWEEN EPITHELIUM AND STROMA IN BENIGN HYPERPLASTIC TISSUE WILFRIED BARTSCH*, JOBST GREEVE~ and KLAUS-DIETER VOIGT Dept Clinical Chemistry, Medical Clinic, University of Hamburg, D-2000 Hamburg 20, F.R.G. (Received 13 November 1986)

order to delineate differences in the mechanism of androgen action in epithelium (E) and stroma (S) of the human prostate, we studied the 17b-hydroxysteroid dehydrogenase (17/?-HSDH) in these tissues of benign prostatic hyperplasia (BPH). Tissue was obtained by suprapubic prostatectomy. E and S were separated; samples were homogenized in buffer and incubated with [)H] steroids (4-androstenedione (Ae), estrone (E,), or dehydroepiandrosterone (DHEA)) and NADH (4.2 mmol/l) as cosubstrate for 60 min at 37°C. Separation and quantification of the metabolites were performed by TLC and LSC, respectively. The main results were: (1) Following incubation with DHEA and E, , only the metabolites S-androstene-38,17/?-diol and estradiol, respectively, were found. Following incubation with Ae, testosterone, Sa-dihydrotestosterone and 5a-androstane-3a-(&,17B-diol were detected as metabolites (the sum of these metabolites were used for calculations). (2) The Michaelis constants were identical in E and S (mean + SEM (n), pmol/l, Ae 6.92 rt 1.01 [9], E, 7.84 f 0.69 [9], DHEA 3.73 f 0.38 [14]). (3) The maximum velocity rate for the three substrates in E was 5-lo-fold that in S (P at least
INTRODUCTION

estrogen receptor and to induce estrogen dependent processes [8]. 2. It is well established that testosterone, which in males is derived almost exclusively from testicular secretion, is the main precursor of dihydrotestosterone (DHT) in the prostate [9]. DHT binds more strongly than testosterone to the androgen receptor and thus is the most potent androgen metabolite at the target organ level [lO,l 11. Recently [12] the question was raised concerning the extent to which adrenal androgens maintain prostatic proliferation in the castrated prostate cancer patient. DHEA and androstenedione (Ae) are derived to a considerable extent from adrenal secretion and can be converted in peripheral organs. Peripheral conversion to active androgen metabolites can be easily monitored by determination of blood levels of testosterone and DHT; however, the prostatic androgen activity of DHEA and Ae is additionally dependent on the activity of intraprostatic enzymes in utilizing these substrates for formation of active metabolites. For formation of testosterone from Ae 17/I-HSDH is necessary, whereas for use of DHEA 3/3-hydroxysteroid-AS-dehydrogenase (a A4-AS-isomerase) must be present.

For many years the presence of 17P-hydroxysteroid dehydrogenase (EC 1.1.1.63, 17,!?HSDH) in human prostatic tissue has been known [l-3]. Nevertheless, the kinetic properties of this enzyme have not been extensively studied in this organ and little is known with respect to the role it plays in determining the intracellular endocrine milieu. We considered it worthwhile to study 17P-HSDH for two biological reasons: 1. The role of estrogens in induction or enhancement of prostatic growth, particularly in benign prostatic hyperplasia (BPH), is still open [4-61. It is 178-HSDH which catalyzes the formation of the active estrogen estradiol (E2) from the less active compound estrone (E,). Additionally the enzyme converts dehydroepiandrosterone (DHEA) to 5-androstene-3,!?,17B-diol (A-Diol). This metabolite is present in relatively high concentrations in human prostatic tissue [7] and, although it is an androgen metabolite, has been reported to interact with the

*Correspondence to W. Bartsch at his present address: Fraunhofer Institute for Toxicology and Aerosol Research, Nikolai-Fuchs-Str. 1, D-3000 Hannover 61, F.R.G. tPart of doctoral thesis

We hypothesized that 17P-HSDH is an important enzyme in modulating the cellular androgen+strogen balance and have determined the kinetic properties and cellular localization of the enzyme in BPH tissue. 35

WILLIS

36

BARTSCH rf al

EXPERIMENTAL

Chemiculs

Radiolabelled steroids were obtained from Amersham International, Braunschweig, (11,2,6,7‘H]DHEA, 85 Cijmmol; ~2,4,6,7-3HJE~, 89 Cilmmoi) and from NEN, Dreieich (1,2,6,7-3H]Ae, 85 Ci/mmoi; [I ,2-“HIA-Dial, 55 Ci/mmol). Purity was checked by TLC. If less than 95% of radioactivity was found in the region of the reference steroids, the substance was purified using Sephadex-LH-20 chromatography. Stock solutions of 2 mmoljl in absolute ethanol were prepared and used directly for various purposes. Non-labelled steroids and other reagent grade chemicals were received from E. Merck, Darmstadt. Coenzymes NADH, NAD, NADPH, and NADP were delivered from Serva, Heidelberg. TkWCY

BPH tissue (weight: 42 f 23 g) was obtained from IO patients 70 f 9 years of age by suprapubic prostatectomy. After extirpation the tissue was immediately placed in ice-cold saline. A part was taken for histological examination by which the diagnosis was confirmed in all cases. The rest was transported to the laboratory. Preparation of tissue suefractions

Separation into epithelium and stroma was performed as reported earlier 113, 141. In brief, tissue was cut into small pieces with a scalpel, homogenized gently in 4 vol of Tris-buffer (10 mmol/l Tris-HCI, 2 mmol/l EDTA, 5 mmol/l NaN,, 10 mmol/l MgCl,, pH 7.4) in a Biihler homogenizer and filtered through nylon sieve. The filtrate was centrifuge and the pellet taken up in 4 vol (w/v) of T&,-buffer. This suspension was defined as the epithelium. The fraction remaining on the sieve was taken up in Tris-buffer, homogenized with an Ultraturrax and filtered again. The material remaining on the sieve was suspended in 4 vol (w/v) of Tris-buffer. This suspension was defined as the stroma. Whole tissue homogenates were prepared by homogenizing the original tissue with 4voi (w/v> of Tris-buffer using an Ultraturrax. One ml aliquots of the various fractions were quick-frozen in liquid nitrogen and stored until further use at -22°C. Subcellular fractionation of the epithelium was done by differentia1 ccnt~fugation. Epithel~um was centrifuged at 800 g and rehomogenized in Tris-buffer containing 0.88 mol/l sucrose. The 8OOg, the 800-l 00,000 g pellets, and the 100,000 g supernatant represented the nuclear fraction, the microsomal fraction. and the cytosol, respectively. Determination of enzyme activities incubation conditions. Aliquots of whole tissue, epithelium, and stroma were cooled to - 190°C and pulverized in a porcelain mortar. The samples were

allowed to thaw in a cold room at 4°C. To allow correction for nonenzymatic conversion and impurities of the tracer, duplicate samples of heat denatured samples were run with each series. For each incubation 150 ~1 of Tris buffer, 2~1 of both, radiolabelled and unlabelled substrate (dissolved in absolute ethanol), and 200~1 of tissue suspension were transferred into conical glass test tubes with stopper. The reaction was started by adding 50 ~1 of cofactor and transferring the vial into a 37°C shaking waterbath. After the incubation time the reaction was stopped by the addition of 6ml of ether. All incubations were performed in duplicate, the mean of which was used for further calculations. Serial determinations of Km and V,,, values in the tissues were performed under standard conditions using 8 substrate concentrations in the range of 0.35-35 pmol/l and NADH of 4.1 mmol/l (final con~ntrations~, 200 ~1 of the original tissue suspension and an incubation time of 1 hr. When no K,,, was determined two concentrations of substrate (20 and 35pmol/l) were used, each in duplicate. Extraction and separation of metabolites: The incubation mixture was extracted successively 1 x with 6 ml, 2 x with 2 ml ether, and finally 1 x with 2 ml chloroform. The organic layers were pooled and evaporated to dryness. After adding 50~1 of an ethanolic reference steroid mixture the walls of the vials were repeatedly washed with ether and evaporated to dryness in order to concentrate the steroids in the tip of the vial. They were transferred with 3 times 75 ~1 ether to the start lane of a thin layer plate (20 x 20 cm, 0.33 mm Silica GF [Merck]). In experiments with DHEA or Ae as substrate the reference steroid mixture was composed of androstanedione (R, = 0.84), Ae (RF= 0.72), DHT (R, = OS@, DHEA (R, = 0.54), testosterone (R, = O&t), 5a-androstane3a(/?),17fi-diols (R, = 0.32), and A-Diol (R, = 0.34). The plates were developed in chloroform-acetone, 9:l (v/v). In experiments with E, as substrate the reference steroid mixture consisted of E, (R, = 0.76), E, (R,= 0.38), and E, (R,= 0.05). The plates were developed in dichloromethane~ther, 9 : 1 (v/v). The reference steroids were marked with U.V.light or with iodine vapour. Eight reference steroid and intermediate fractions were scraped off and the silica gel was transferred to scintillation vials. After adding lOm1 of scintillator (Biofluor, NEN, Dreieich) and mechanical shaking for 2 h, radioactivity was measured in a Packard 460CD scintillation counter with quench correction by use of an external standard. Determination of K,,, and I’,,,,,: the regression line of the double reciprocal plot according to Lineweaver-Burk[ 151was calculated by the methods of least squares. Km and V,,, were obtained from the intercepts with the 1,/V and l/S axis, respectively. In cases where no Km was determined, V,,,,, was obtained by the Michaelis-Menten equation V,,, = V(k, -tS)/S using for K,,, the means of all determinations for the respective substrate (Table 3).

l78-hydroxysteroid dehydrogenase IdentiJication

of metabolites

Using the substrate DHEA, TLC separates neither the substrate from the possible metabolites DHT and androsterone nor the metabolite A-Diol from other diols with 5a-H configuration. Discrimination between 5-unsaturated and 4-unsaturated as well as saturated steroids can be achieved by epoxidation with m-chloroperbenzoic acid [ 161. For epoxidation the respective fractions were scraped off, transferred with suction into a Pasteur pipette, and eluted with 2 ml of ether. An aliquot of the eluate was evaporated to dryness. 200 pg of reference steroids DHT and DHEA or A-Diol and 3b-Diol and 1 ml of a 1% (w/v) solution of m-chloroperbenzoic acid in dichloromethane were added. The mixture was left at room temperature for 2 h. In order to eliminate excess peracid, 0.5 ml of 10% (w/v) aqueous KJ solution, 0.5 ml of 10% (w/v) aqueous NaHCO, solution and 4 ml dichloromethane were added successively and the mixture was shaken vigorously after each addition. The aqueous layer was discarded and the organic solution washed with 1 ml of water. It was evaporated to dryness and transferred to a thin layer plate as described above. On separate lanes of the same plate the second aliquot of the identically handled (without epoxidation) sample and identically handled samples (epoxidized and non-epoxidized) of the reference steroids were applied and developed in cyclohexane-ethylacetate, 4: 6 (v/v). Isolation of fractions and counting of radioactivity were performed as described above. Recrystallization. Testosterone and estradiol product fractions were identified using recrystallization to constant specific activity. The respective fractions were scraped off the thin layer plate, and eluted with ether. An aliquot was evaporated to dryness. 100 mg of the crystalline authentic material was added. The mixture was dissolved by heating in ethanol-water and crystallized by cooling. The crystals were collected and dried at 90-100°C. The radioactivity was determined in an aliquot, and the remainder was subjected to the same procedure 3 times. Radioactivity per 100 mg of crystalline material was computed and compared with the radioactivity measured in the aliquot of the original sample. Epoxidation.

Miscellaneous

Protein and DNA were determined as reported earlier [14, 173. Statistical comparisons were performed using the two-tailed t-test.

conversion rates than NADPH. Therefore, in all further experiments NADH was used. Acceptable conversion rates were obtained in the substrate concentration range 0.440 pmol/l. Further experiments demonstrated linear relations between reaction velocity and the incubation time (Fig. 1, upper panels) or tissue dilution (Fig. 1, middle panels). Maximum velocity rates (Fig. I, lower panels) were obtained with a NADH concentration of about 4 mmol/l. Similar qualitative patterns were seen when incubations were performed with samples from whole tissue (WT), epithelium (E) and stroma (S) (Fig. 1, left, middle, and right panel). Results obtained with the substrates Ae and E, were similar to those of DHEA (data not shown). From these data the standard conditions for the quantitative determinations were depicted (see Experimental). Ident$cation

of optimal incubation

of metabolites

Identification of substrate and product fractions was done by cochromatography with authentic material. In case of the substrate DHEA all radioactivity migrated with the DHEA and A-Diol fractions when unspecific radioactivity was substracted. Since the chromatographic system did not separate DHEA and A-Dial from other potential metabolites, i.e. Scl-saturated androstanolones and androstanediols, respectively, both fractions were further examined by epoxidation. As indicated in Table 1, no significant radioactivity remained in the DHEA or A-Diol fractions following epoxidation, indicating that no See-reduced metabolites were formed from DHEA under the prevailing conditions. In case of the substrates Ae and E, the authenticity of the testosterone and E, fractions, respectively, was proven by recrystallization to constant specific activity. The data given in Table 2 indicate that the testosterone and Ez fractions were at least 8 1% pure. Determination

of K,,, and V,,,

Using the standard conditions, K, and V,,, data were determined using the substrates DHEA, Ae and E, . Typical examples are shown in Figs 24. Data are summarized in Table 3. All data for the substrates AC and E, were very similar, whereas K,,, and V,,,,, were lower for the substrate DHEA than for the other Table I. Percentual distribution of radioactivity in the DHEA and A-Diol fraction before (c) and after (e) epoxidation with m-chloroperbenzoic acid

Control

RESULTS

Evaluation

37

conditions

In initial experiments 200 ~1 whole tissue homogenate was incubated with the substrate DHEA (final concentration range: 1-lOmmol/l) and NADH or NADPH as cosubstrate (final concentration: 1.6 mmol/l) for 17 min at 37°C. Under these conditions, the cofactor NADH led to about 40% higher

Epithelium c e

c

e

DHEA fraction: DHEA Epoxide

91.8 2.5

2.5 93.6

91.8 1.6

A-Diol fraction: A-Diol Epoxide

97.5 1.2

3.6 85.5

84.3 13.3

Data

Stroma c

e

2.1 94.4

91.7 6.5

4.5 80.5

4.8 84.9

90.0 5.6

6.0 87.0

represent per cent radioactivity related to total radioactivity found in the chromatographic lane. Control = authentic radioactive material.

WILFIUEDBARTSCHet al.

38

JII Homogenate

E

L

L

12

S

8

L

B

NADH (mmolll)

Fig. 1. Dependence of 17b-hydroxysteroid dehydrogenase from time (upper panels), tissue dilution (middle panels), and cofactor concentration (lower panels) in whole tissue (WT), epithelium (E), and stroma (S) of benign prostatic hyperplasia. Homogenates were incubated with t3H]DHEA under standard conditions (arrows) unless other conditions are indicated. Products formed were measured by TLC and LSC.

substrates. Regardless of whether V,,, was expressed in terms of pmol/mg protein x h or nmol/mg DNA x h, in each case the activity in the epithelium was distinctly higher than in the stroma (P at least Table 2. Crystallization of the testosterone (substrate androstenedione) and estradiol (substrate estrone) fractions to constant specific activity Epithelium Stroma (dpm/lOO mg solid material)

Testosterone: Initial value

28,480

9870

1st Crystallization 2nd Crystallization 3rd Crystallization

24,740 24,740 25,770

9060 9440 9150

Estradiol: Initial value 1st Crystallization 2nd Crystallization 3rd Crystallization

56,240 50,050 46,960 47,700

21,640 19,680 19,520 20,780

~0.01). This difference was more pronounced when the data were related to protein than to DNA. In each case the activity in the whole tissue was found to lie between that of epithelium and stroma. Further experiments were performed to localize the enzyme in question intracellularly. Since the bulk of enzyme activity was found in the epithelium, these experiments were restricted to this fraction and were performed with the substrate DHEA. 30% of the activity was recovered in the 800g and 10% in the 800-100,OOOg pellets, whereas no activity was found in the 100,OOOgsupernatant. The question whether active androgens can be formed in the prostate via the oxidative pathway from the substrate DHEA was evaluated by incubating whole tissue, epithelium, or stroma with DHEA (10 nmol/l-1.5 pmol/l) and NAD or NADP (4 mmol/l, final concentrations). The formation of

17/l-hydroxysteroid

dehydrogenase

39

0 16 K"al "Wl0X'

0 E

I2

3 1 C

Fig. 2. Substrate saturation (left panel) and double reciprocal plot (right panel) of 17/7-hydroxysteroid dehydrogenase incubated under standard conditions (see Fig. 1) with [‘HIDHEA. Abbreviations-see Fig. 1.

250

/’ 0

0

150

50

Fig. 3. Substrate saturation (left panel) and double reciprocal plot (right panel) of 17/I-hydroxysteroid dehydrogenase incubated under standard conditions (see Fig. 1) with [3H]androstenedione. Abbreviations-see Fig. 1.

WILFRIED BARTSCH et

40

al.

Fig. 4. Substrate saturation (left panel) and double reciprocal plot (right panel) of l7~-hydroxystcroid dehydrogenase incubated under standard conditions (see Fig. I) with [)H]estrone. Abbreviations-see Fig. I.

the testosterone and G-reduced metabolites couId not be demonstrated, whereas in each experiment more than 95% of the substrate was recovered. DISCUSSION

Before discussing the results a few remarks concerning the methods are necessary. They are primarily concerned with two points: (1) difficulties in the interpretation of enzyme activities obtained in vim from organ homogenates or subcellular fractions, and (2) problems arising from the mechanical separation of stromal from epithelia1 cells.

(1) Physical-chemical laws used for the characterization of enzymes apply in principle to pure, isolated enzymes. Applied to optimized measurements in homogenates, however, only K,,, values, i.e. Michaelis-Menten constants, can be used for that purpose as it is an independent parameter of the enzyme. V,,, values under the prevailing conditions, however, cannot be used as constants but are rather a measure of enzyme concentrations, provided that the dete~ination has been performed correctly. The artificial milieu presented during in aitro measurement, i.e. substrate or coenzyme saturation, does not prevail in oivo. Furthe~ore, the direction of

Table 3. I7@-Hydroxysteroid dehydrogenase in human benign hyperplastic prostate V

pmol/(mg prot h) S~ndrosteron~ Epithelium Stroma Whole tissue

132 i 67 (IO)* 261 ll(l0) 361: 12(9)

Substrate androstenedione: Epithelium 383 i: 126 (5)* Stroma 40 & 7 (5) Whole tissue 75 & 29 (5) Substrate ___.__ estrone: Epithelium stroma Whole tissue

362 i I58 (5)’ 33 i 9 (5) 63 i: IX (5)

KS

m~m~lj(mgDNA h)

if moljl

3.14& 2.54(10)* 1.~~0.57(10) 2.26 + 0.70 (9)

3.14+1.12(s) 3.72 ri: 1.83 (5) 4.15-k 1.55(4)

9.05 k 4.30(5)’

1.86rt 0.28(5) 5.45 _t 1.70 (5)

9.60 i: 3.06 (3) 3.93 i. 0.84 (3) 7.32 + 1.65 (3)

8.95 i S.03 (5)* 1.32I 0.38 (S) 3.69 & 1.08 (5)

6.99 4 0.64 (3) 6.38 rt 0.38 (3) 10.1 & 2.06(3)

Mean f SD (number of samples); *Significant difference compared to stroma
P

at least

17/I-hydroxysteroid dehydrogenase reversible transformations should critically depend on the amounts of reduced and oxidized forms of substrate and coenzyme available. Since we did not study the reverse reactions, i.e. the formation of 1‘I-keto derivatives from the respective 17fl-hydroxysteroids, and we do not know actual cellular substrate and coenzyme concentrations, conclusions regarding the physiological meaning of high or low enzyme activities must be made with great caution and are only of relative value. (2) In our laboratory the separation of epithelium and stroma is performed according to Cowan ef a1.[13]. This technique leads to inevitable damage of epithelial cells, the cytoplasm of which often appears broken with holes and leakages. The stromal cells seem to be more intact. However, stromal cells are firmly embedded in the collagenous matrix and their damage is therefore difficult to evaluate. Nevertheless, we assume that these procedural insufficiencies do not affect our main conclusions with respect to the enzyme distribution between epithelium and stroma: the enzyme in question was not found in the cytosol, but mostly bound to cellular particle fractions. Cell damage therefore does not necessarily lead to the release of the enzyme. Furthermore, did the epithelial supernatant which is discarded during the preparation of epithelium and stroma and which should contain most of released soluble and particle bound enzyme, contained only 15% enzyme activity of the whole tissue homogenate. This loss of enzyme activity compares well to the loss of tissue mass in the same fraction (data not shown). We therefore assume that artificial redistribution does not occur to a considerable extent and that cross contamination between the epithelial and stromal enzyme should compare to the cross contamination of the two fractions. Data concerning the cross contamination of epithelium and stroma favour the assumption that the difference in enzyme activity observed between the two fractions is even more pronounced than indicated by Table 3. From experiments [14] in which we checked the purity of the cell preparations with two biochemical markers, we can assume an overall contamination of epithelial cells by stromal elements of about 15% and of stromal by epithelial structures of about 6%. In view of the large differences in 17/I-HSDH between the two compartments the latter contamination is of particular importance. It indicates that a large percentage of the stromal enzyme activity represents epithelial contamination. From these data it is questionable if the stroma contains 17/I-HSDH at all. It is obviously localized principally in the epithelium of BPH. Comparing Km and V,,,,, values obtained with the three substrates DHEA, Ae, and E, it becomes obvious that the data are very similar and we assume that the reactions are catalyzed by the same enzyme. Although absolute proof cannot be given by our

41

experiments, inhibition studies (unpublished) further support this assumption. The properties of 17/?-HSDH with respect to the conversions of Ae and E, are more similar to each other than to DHEA. The differences between the substrates Ae and E, compared to DHEA might demonstrate differences in kinetic properties. With respect to conversion rates in the biologically important linear part of the velocity-substrate curve, however, which is characterized by the ratio V,,, /K,,, , all three substrates appear very similar. To our knowledge, in human prostate epithelium and stoma 178-HSDH has not been previously characterized. Data from whole tissue have been reported very recently [18] using the “reverse” reaction, i.e. testosterone as substrate and monitoring the formation of androstenedione in the presence of NAD as the cofactor. Km and V,,, values agree very well with our data, supporting the reversibility of the reaction. The distinct differences found in 17/?-HSDH activity between epithelium and stroma of BPH indicate the necessity to discriminate between the two compartments when comparing prostatic tissue specimens of different origin: when comparing 17B-HSDH in whole tissue homogenates from BPH, normal prostate, and prostate carcinoma it otherwise cannot be concluded if differences are either due to the altered composition of epithelium and stroma, or to altered specific activity of particular cells. From a biological standpoint the following conclusions can be drawn: (1) Our data do not favour a role of 17J-HSDH in sustaining growth of BPH by formation of compounds with estrogenic properties (E,, A-Diol). 178-HSDH is found predominantly in the epithelium. This compartment tends to be decreased in BPH, and accumulation of E, has been found exclusively in stromal nuclei [6]. Furthermore, the ratio between A-Diol and DHEA did not change when comparing blood and prostatic tissue [7] ruling out extensive conversion of both metabolites in preference to one of them. (2) Otherwise our data underline that Ae can be converted to testosterone and subsequently to DHT in the human prostate. Indeed, this finding is most important for the orchidectomized prostate cancer patient where adrenal Ae secretion remains considerable. Furthermore, it has to be kept in mind that prostate carcinoma is an epithelial tumour. Measurements of 17b-HSDH in this type of tissue, however, are mandatory before drawing further conclusions with respect to this disease. (3) The finding of minor (not detectable by our method) or absent conversion of DHEA to testosterone, i.e. the absence of 3/?-hydroxy-A5steroid dehydrogenase indicates that Ae and not DHEA is the most important adrenal precursor for active androgens in the prostate, although one should keep in mind that this statement is derived from BPH

WILFRIED BARTSCH et al.

42

tissue and might not be valid for carcinoma

tissue.

Finally it should be stressed that hydrogenation by 17p-HSDH is an important step in the pathway of fo~ation of active androgens and estrogens. Effective inhibitors of this enzyme would therefore inhibit androgenic and estrogenic actions and might be of benefit in androgen and estrogen dependent tumours. Compounds demonstrating such potencies are presently under investigation in our laboratory.

REFERENCES I. Acevedo H. F. and Goldzieher J. W.: Further studies on

the metabolism of 4(4-‘4C)androstene-3,17-dione by normal and pathological human tissue. Biochem. biophys. Acta 97 (1965) 564-570.

2. Acevedo H. F. and Goldzieher J. W.: The meta~lism of (“C) estrone by hypertrophic and carcinomatous human prostate tissue. Biochem. biophys. Acta 97 (1965) 571-578.

3. Harper M. E., Pike A., Peeling W. B. and Grifliths K.: Steroids of adrenal origin metabolized by human prostatic tissue both in uivo and in vitro. J. Endocr. 60 (1974) 117-125. 4. Belis J. A.: Methodological basis for the radioimmunoassay of endogenous steroids in human prostatic tissue. Invest. Ural. 17 (1980) 332.-,336. 5. Ghanadian R. and Puah C. M.: Relationship between estradiol-l7fi, testosterone, dihydrotestosterone, and 5sc-androstane-3a,l7/?-diol in human benign hypertrophy and carcinoma of the prostate. J. Endocr. gS (1981) 255-262. 6. Kozak I., Bartsch W.. Krieg M. and Voigt K. D.: Nuclei of stroma: site of highest estrogen concentration in human benign prostatic hyperplasia. Prostate 3 (1982) 433-438.

I. Bartsch W., Kozak I., Gorenflos P., Becker H. and Voigt K. D.: Concentration of 3/?-hydroxy androgens in

epithelium and stroma of benign hyperplastic and norma1 human prostate. Prostate 8 (1986) 3-10. 8. Rochefort H. and Garcia M.: The estrogenic and antiestrogenic activities of androgens in female target tissues. Pharmac. Ther. 23 (1984) 193-216. 9. Farnsworth W. E. and Brown J. R.: Metabolism of testosterone by the human prostate. J. Am. med. Ass. 183 (1963) 140-143. 10. Anderson K. M. and Liao S.: Selective retention of dihydrotestosterone by prostatic nuclei. Nafure 219 (1968) 5953-5960. 11. Bruchovsky N. and Wilson J. D.: The intranuclear binding of testosterone and 5a-androstan-~7~-01-3-one by rat prostate. J. biol. Chem. 243 (1968) 5953-5960. 12. Labrie F., DuPont A., Belanger A. and Members of the Lava1 University Prostate Cancer Study Group: Spectacular response to combined antihormonal treatment in advanced prostate cancer. In Endocrinology (Edited by F. Labrie and L. Proulx). Excerpta Med., Amsterdam (1984) pp. 45%453. 13. Cowan R. A., Cowan S. K.. Giies C. A. and Grant J. K.: Prostatic distribution of sex ho~one-binding globulin and cortisol-binding globulin in benign hyperplasia. J. Endocr. 74 (1976) 121-131. 14. Krieg M., Kloetzl G., Kaufmann J. and Voigt K. D.: Stroma of benign prostatic hyperplasia: preferential tissue for androgen metabolism and oestrogen binding. Acta endocr., Copenh. 96 (1981) 422-432.

of 15. Lineweaver H. and Burk D.: The dete~ination enzyme dissociation constants. 1. Am. them. Sac. 56 (1934) 658-666. 16. Lisboa B. P.: Thin layer chromatography of steroids, sterols and related compounds. Meth. Enzym. 15 (1969) 3-73. 17. Bartsch W., Krieg M., Becker H., Mohrmann J. and Voigt K. D.: Endogenous androgen levels in epithelium and stroma of human benign prostatic hyperplasia and normal prostate. /Icta endoer., Cope& 100 (1982) 634640. 18. Brendler C. B., Follausbee A. L. and Isaacs J. T.: Discrimination between normal, hyperplastic and malignant human prostatic tissue by enzymatic profiles. J. &of. 133 (1985) 495-501.