growth inhibition of human prostate cells in vitro by...

CANCER RESEARCH56. 4956-4964, November I, 19961

ABSTRACT

The long-standing strategy for the treatment ofmetastatic prostate cancerhas been to reduce androgenic stimulation oftumor growth by removal of thetestes, the primary site of testosterone synthesis. However, a low level ofandrogenic stimulation may continue, even after castration, by the conversionofadrenal androgefLs to Sa-dihydrotestosterone (DEfT) in the prostate tumorcells. Two important enzymes of the androgen biosynthetic pathway are17a-hydrose/C17@-lyase, which regulates an early step In the synthesis oftestosterone and other androgens in both the testes and adrenal glands, and5a-reductase, which converts testosterone to the more potent androgen,DHT, In the prostate. We have identified new inhibitors ofthese enzymes thatmay be of use in achieving a more complete ablation of androgeits in the

treatment ofmetastatic prostate cancer. Three derivatives ofandrostene wereshown to inhibit 17a-hydroxylase/C17,@-lyase with potencies 2.-20-foldgreater than that of ketoconazole, a previously established inhibiter of thisenzyme. Detivatives of pregnane and pregnene displayed activities against5a-reductase that were comparable to that of N-(1,1.dimethyl-ethyl).3-oxo

4-aza-Sa-andmst-1-ene-17fi-carboxamide. All of the 5a-reductase inhibitorswere able to at least partially inhibit the mitogeniC effect of testosterone in

either histocultures of human benign prostatic hypertrophic tissue or incultures ofthe LNCOP human prostatic tumor cell line@For these compounds,it appears that this inhibition can be attributed to a reduction of DHTsynthesis in these cultures, because no inhibitory effect was observed inDHT-treated cultures, and none of the compounds had a cytotoxic effecLSurprisingly, one of the inhibitors of 17a-hydroxylase/C17,@-lyase, 17@3-(4-hnidazolyl)[email protected], was also able to inhibit the mitogenic effect of

testosterone in both the histoculture and cell culture assays and had an effectagainst DHT as well. In transcriptional activation assays, it was found thatthis compound is an antagonist of both the wild-type androgen receptor andthe mutant androgen receptor,which Is present in LNCaP cells. In conduaba, the abilities of these compounds to inhibit androgen synthesis and, insome cases, to exert antiandrogen activity, did in fact translate to an inhibitory effect on the growth of human prostatic tissue in vitro, suggesting theirpotential utility in the treatment of prostatic cancer.

INTRODUCTION

The androgens testosterone and DHT5 play a key role in themaintenance of cell proliferation in the prostate gland (1). Accord

Received 6/3/96; accepted 9/3/96.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

I This work was supported by NIH Grant CA27440 (to A. M. H. B.) and by the

National Institute of Child Health and Development Center for Population ResearchGrants HD-169l0 and P30-HD18968 (to E. M. W.).

2 Present address: Research Institute of Life Science, Snow Brand Milk ProductsCompany, Ltd., Shimotsuga-gun, Tochigi 329-05, Japan.

3 Present address: Central Research, Pfizer, Inc., Groton, CT 06340.

4 To whom requests for reprints should be addressed, at Department of Pharmacology

and Experimental Therapeutics, University of Maryland School of Medicine, Room 4-002Bressler Building, 655 West Baltimore Street, Baltimore, MD 21201.

5 The abbreviations used are: DHT, Sa-dihydrotestosterone; AR, androgen receptor;

HPLC. high performance liquid chromatography; FBS, fetal bovine serum; BPH, benignprostatic hyperplasia; 1-33. 20a-hydroxy-4-pregnen-3-one; 1-34, 20@-hydroxy-4-pregnen3-one;1-41,17j3-(4-imidazolyl)-4-androsten-3-one;1-47,17@3-(4-imidazolyl)-5-pregnen3@-ol; 1-49, l7f3-(imidazolyl)-5,l6-pregnadien-3@-ol: L-l0, 20@-hydroxy-4,16-pregnadien-3-one; finasteride, N-(l,l-dimethyl-ethyl)-3-oxo-4-aza-5a-androst-1-ene-l7@-carboxamide; 4MA, N.N-diethyl-4-methyl-3-oxo-4-aza-5a-androstane-17@-carboxamide.

ingly, human prostatic tumors are often androgen dependent, andtreatments designed to reduce androgenic stimulation of these tumorsfrequently produce therapeutic responses. Testosterone, the majorcirculating androgen in the human male, is synthesized primarily inthe testes but also in the adrenal glands in much smaller amounts. Akey regulatory enzyme of the testosterone biosynthetic pathway issteroidal 17a-hydroxylase/C1720-lyase (EC 1.14.99.9), which catalyzes both the 17a-hydroxylation and the cleavage of the C 17.20sidechain during the conversion of the 21-carbon steroids pregnenoloneand progesterone to the 19-carbon androgens dehydroepiandrosteroneand androstenedione, respectively (Fig. I). It has been reported that asingle active site on this enzyme catalyzes both reactions (2), and thatthe l7a-hydroxylase/C1720-lyases present in human testes and adrenal glands have identical amino acid sequences (3). In the prostate,testosterone is converted by 5a-reductase to DHT (4), a more potentandrogen (5, 6) which is apparently the primary trophic hormone ofthe prostate (7). The growth effects of DHT and testosterone on theprostate are mediated by the AR, a Mr 120,000 protein (8) that isexpressed in both the epithelial and stromal cells of both normal andcancerous prostatic tissue (9).

The beneficial effect of androgen ablation therapy in patients withmetastatic prostate cancer was first established by Huggins et al. (10)in 1941 when they demonstrated that castration could often producean improvement in clinical status. Castration, performed either surgi

cally or medically by the administration of an luteinizing hormonereleasing hormone analogue, remains a standard treatment for metastatic prostate cancer. More recently, Labrie et a!. (1 1) proposed thatandrogen ablation therapy could be substantially improved by administration of an AR antagonist to castrate patients to block the effectsof residual androgens produced by the adrenal glands. The merit ofthis approach is supported by a large clinical trial conducted by theNational Cancer Institute in which combination treatment with theluteinizing hormone-releasing hormone analogue leuprolide and theantiandrogen flutamide was compared to treatment with leuprolideonly (12). Patients treated with the combination therapy had a longermedian progression-free period (16.5 months versus 13.9 months) andlonger median length of survival (35.6 months versus 28.3 months). Ina small subset of patients in this study who had minimal disease at theinitiation of therapy, the advantage of combination therapy was particularly impressive; the progression-free period was 58 months withcombination therapy versus 19 months with monotherapy (13). Thebenefit of combination therapy was also observed in a similar studyconducted by the European Organization for Research and Treatmentof Cancer (14). Thus, efforts to block the effects of adrenal androgensapparently result in worthwhile therapeutic gains.

An inhibitor of steroidal 17a-hydroxylase/C1720-lyase could p0-tentially produce the same effect as the combination of leuprolide andflutamide because it could eliminate the contribution of both testicularand adrenal androgens to the growth of the neoplastic prostate. Ketoconazole, an imidazole antifungal agent, is one such inhibitor of thisenzyme, and it has been used to treat metastatic prostate cancer. Someclinical success has been achieved with ketoconazole, both as a

4956

Growth Inhibition of Human Prostate Cells in Vitro by Novel Inhibitors ofAndrogen Synthesis

Gregory T. Kius, Junji Nakamura,2 Ji-song Li,3 Yang-zhi Ling, Chong Son, Jon A. Kemppainen,Elizabeth M. Wilson, and Angela M. H. Brodie4Department of Pharmacology and Experimental Therapeutics, University of Maryland School of Medicine, Baltimore, Maryland 21201 fG. T. K., J. N., f-s. L, Y-z. L, C. S..A. M. H. B.J, and the Laboratories for Reproductive Biology and the Departments of Pediatrics and Biochemistry, University of North Carolina, Chapel Hill. North Carolina27599 1). A. K., E. M. W.J

on July 7, 2019. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

GROWTH INHIB@0N OF HUMAN PROSTATE CELLS

IH3_______@ HO@ OH

HOX@@@@ C17@-Lyase l7@3-OHSD

C=0

HO@

Pregnenolone 17a-hydroxypregnenolone Dehydroepiandrosterone Androstenediol-- -- 3I@OHSD/Isomerase@@.1

1r@@

0@d::9Â±:@l7a-Hydroxy@@@@jOHC17@@-Lyase@ l7@_OHS@j@Â±@I0:0 C=O

Progesterone 17a-Hydroxyprogesterone Androstendione Testosterone

o@e@c:t:i9:f:@/i@1Reductase

Dihydrotestosterone

Fig. 1. Androgen biosynthesis. The microsomalenzyme 17a-hydroxylase/C1720-lyaseregulatesakeystepinthesynthesisof testosteroneinthetestesand adrenalglands.Testosteroneis takenup fromtheperipheralcirculationby theprostate,whereit isconverted to the more potent androgen DHT by

microsomal 5a-reductase. 313-OHSD, 3@3-hydroxysteroiddehydrogenase;17@3-OHSD,17@-hydroxysteroid dehydrogenase.

substitute (15) and adjuvant (16) for castration. However, the majorityof the clinical experience with this compound indicates that it haslimited usefulness in the treatment of metastatic prostate cancer because of its relative lack of efficacy and specificity, undesirable sideeffects, and short plasma half-life, which necessitates an inconvenientthree-times-daily dosing schedule (17). New, more potent inhibitorsof 17a-hydroxylase/C 1720-lyase could potentially overcome theseshortcomings of ketoconazole.

Alternatively, compounds that possess dual activity against 17a-hydroxylase/C1720-lyase and 5a-reductase or against one of theseenzymes and the human AR might also produce a therapeutic effectthat is comparable to or greater than that produced by the leuprolide/flutamide combination. The success of combination therapy withflutamide and the 5a-reductase inhibitor finasteride in preliminaryanimal and human studies (18, 19) lends support to the concept thatother combination therapies may also be of clinical use.

Here we report on the identification of novel inhibitors of 17a-hydroxylase/C1720-lyase and 5a-reductase, some of which exhibitpotencies against these enzymes that are comparable to or greater thanthose of ketoconazole and finasteride, and some which additionallydisplay antiandrogen activity. Also presented are the antiproliferativeeffects of selected compounds on histocultures of human BPH tissueand on cultures of the human prostatic tumor cell line, LNCaP.

MATERIALS AND METHODS

Materials

The LNCaP (human prostatic carcinoma) and CV1 (monkey kidney) celllines were obtained from the American Type Culture Collection (Rockville,MD). Eagle's MEM without phenol red was prepared from MEM amino acidsolution, MEM vitamins, and Earle's balanced salt solution and supplementedwith penicillin/streptomycinlamphotericinB, all ofwhich were purchased fromLife Technologies, Inc. (Grand Island, NY). RPM! 1640, DMEM (with highglucose, without phenol red), trypsinlEDTA (0.25%/0.02%), and PBS werepurchased from JRH Biosciences (Lenexa, KS). Calf serum was purchasedfrom HyClone Laboratories, Inc. (Logan, UT). Proteinase K was purchasedfrom Promega Corp. (Madison, WI). [7-3HjPregnenolone (25 Ci/mmol),[7-3H]testosterone(20 Ci/mmol), [3Hjthymidine(80 Cilmmol), and [â€˜4CJDHT(50 mCi/mmol) were obtained from DuPont NEN (Boston, MA). Si25OF-PAsilica gel thin layer chromatography plates, monobasic and dibasic sodiumphosphate, and HPLC grade chloroform, acetonitrile, water, methanol, andether were produced by J. T. Baker, Inc. (Phillipsburg, NJ) and purchased from

VWR Scientific (Bridgeport, NJ). All other reagents were obtained fromSigma Chemical Co. (St. Louis, MO) except where otherwise stated.

Candidate Inhibitors of Androgen Synthesis. 1-33, 1-34, flutamide, andketoconazole were purchased from Sigma Chemical Co. 1-41, 1-47, and 1-49

were synthesized in our laboratory as described by Li et a!.6 L-l0 wassynthesized as described by Ling et a!.7 Finasteride and 4MA were kindlyprovided by Dr. Linda Rhodes of Merck Research Laboratories (Rahway, NJ).Hydroxyflutamide was kindly provided by Dr. Rudolph Neri of ScheringPlough Research Institute (Kenilworth, NJ).

Preparation of Charcoal.stripped FBS

Steroids were removed from FBS by stirring with activated charcoal (4mg/mI) for 16 h (all steps were performed at 4Â°C), followed by centrifugation

for 1 h at 1600 x g. The supernatant was then centrifuged for 1 h at100,000 x g, and the resulting supernatant was either immediately sterilizedby passage through a cellulose acetate filter with a 0.22-j.@mpore size (CorningGlass Works, Coming, NY) or, when preparing triple-charcoal-stripped serum,subjected to the above procedure two more times and then filtered.

Human Tissue Specimens

Fresh specimens of testis (from untreated prostatic cancer patients undergoing orchiectomy) and BPH tissue were kindly provided by Dr. StephenJacobs (Division of Urology, Department of Surgery, University of MarylandMedical Center). The preparation of microsomes from these tissues wasperformed as described by Li et a!. (20).

Enzyme Inhibition Studies

17a-HydroxylasWC17,@-LyaseAssay. Incubations were carried out in atotal volume of 1.01 ml. Sample tubes were first supplied with propyleneglycol (â€”10 pi) and ethanol solutions containing [7-3Hlpregnenolone(280,000 dpm), pregnenolone (0.063 j.@g;final concentration in incubation of200 nM), and the indicated candidate inhibitors. The ethanol solutions wereevaporated under air, and then 850 @dof 0.1 M sodium phosphate buffer (with

DTT at 78 ,.&M, pH 7.4) and a NADPH generating system (50 @d of phosphate

6 J-s. Li, Y-z. Ling, G. T. KIus, C. Son, and A. M. H. Brodie. 17-Imidazolyl, pyrazolyl,

isoxazolylandrostenederivatives:novel steroidalinhibitors of humancytochrome17a-hydroxylase/C1720-lyase (P450@ submitted for publication.

7 Y-z. Ling, i-s. Li, G. T. Klus, C. Son, M. Kirk, and A. M. H. Brodie. Synthesis and

in vitro activity of some epimeric 20-hydroxy, 20-oxime and aziridine pregnene derivatives as inhibitors of human l7a-hydmxylase/C17@0-lyase and Sa-reductase, submittedfor publication.

4957



GROWTH INHIBITION OF HUMAN PROSTATE CELLS

buffer containing 6.5 mM NADP@,71 mt@iglucose-6-phosphate, 1.25 IU ofglucose-6-phosphate dehydrogenase) were added to each tube, and the tubeswere preincubated for 10 mm at 34Â°C.The reaction in each tube was initiatedby adding human testicular microsomes (100 @.tgof protein in 100 @.dofphosphate buffer), and the incubation was carried out for 5 mm under oxygenwith shaking in a water bath at 34Â°C. At the end of the incubation, authentic

markers of l7a-hydroxypregnenolone and dehydroepiandrosterone wereadded to each tube, the steroids were extracted with ether, and the etherevaporated. To remove unwanted lipids and particulate matter, which couldinterfere with the subsequent HPLC separation, the samples were resuspended

in acetonitrile and drawn by suction sequentially through a C18 Sep-Pak Vac

cartridge (100-mg size; Waters Corp., Milford, MA) and a Teflon filter

(0.22-sm pore size; Micron Separations, Inc., Westboro, MA). After evapo

ration of the acetonitrile, the samples were resuspended in 80 p1 of acetonitrile:

methanol (50:50), mixed with 80 @lof the HPLC initial mobile phase ofwater:acetonitrile:methanol (60:30: 10), and injected into the HPLC system

(Waters). HPLC was performed with a Nova-Pak C18 reverse phase column

(Waters) and a total flow rate of 1 mI/mm. Two pumps were used to producethe mobile phase: pump A, acetonitrile; and pump B, water:acetonitrile:methanol (60:30:10). For the elution ofeach sample, the initial condition of 1%

pump A and 99% pump B was maintained for the first 15mm. Conditions werethen changed as follows: from initial conditions at 15 mm to 80% pump A at20 mm; from 80% pump A at 20 mm to 90% pump A at 25 mm; from 90%pump A at 28 mm to 50% pump A at 30 mm; and finally, from 50% pump Aat 30 mm to initial conditions at 32 mm. A constant rate of change in pumpingconditions was maintained over the course of any single transition. The UVabsorbance at 216 nm and the radioactivity of each 0.1 ml of eluate weremeasured by an in-line spectrophotometer (Waters) and an in-line radioactiveflow detector (Packard Instrument Co., Meriden, CT), respectively. The UVabsorption peaks of the authentic standards were used to identify the corre

sponding peaks of radioactivity. The l7a-hydroxylase activity was determinedfrom the percentage of [7-3H]pregnenolone converted to [3H]17a-hydroxypregnenolone and [3H]dehydroepiandrosterone. C1720-lyase activity was

determined from the percentage conversion of substrate to [3H]dehydroepi

androsterone.5a.R@tjuc@ Assay. Incubations were carried out in a total volume of

1.01 ml. Ethanol solutions containing [7-3H]testosterone(600,000 dpm), testosterone (6.1 ng; final total concentration of testosterone in incubation of 30nM), and the indicated candidate inhibitors were added to 0.01 ml propylene

glycol in sample tubes. The ethanol was evaporated, and then 400 p1 of 0.1 Msodium phosphate buffer (with Dli' at 78 @.tM,pH 7.4) and a NADPHgenerating system (100 pi of phosphate buffer containing 6.5 mM NADP@, 71

SUMglucose-6-phosphate, and 2.5 IU of glucose-6-phosphate dehydrogenase)

were added to each tube; the tubes were then preincubated for 15 rain at 37Â°C.The reaction in each tube was initiated by adding human BPH microsomes(600 @igprotein in 500 @dof phosphate buffer), and the incubation was carriedout for 10 mm under oxygen with shaking in a water bath at 37Â°C.At the endof the incubation, approximately 3000 dpm of [â€˜4CJDHT and 50 @gof

nonradioactive DHT were added to each tube to serve, respectively, as arecovery standard and visualization marker for thin layer chromatography.Extracted steroids were then separated by thin layer chromatography (chloroform:ether, 80:20) and visualized in iodine vapor. The DHT spot of each

sample was scraped from the thin layer chromatography plate, the DHT wasextracted with ether, the ether evaporated, and the extract was analyzed forradioactivity by liquid scintillation spectrometry. The 5a-reductase activitywas determined from the percentage conversion of [7-3H}testosterone to[3H]DHT. IC50values were determined from plots of enzyme activity at fourdifferent concentrations of inhibitor (each sample in duplicate).

Histoculture

Human BPH tissue was collected in surgery and then washed in HESS(Sigma) and divided into pieces 1â€”2mm in diameter and 0.5â€”1.0-mmthick.Four to five pieces were placed on each of several absorbent gelatin sponges(Upjohn, Kalamazoo, MI). Each sponge was placed in a separate well of a24-well plate and incubated (37Â°C,5% C02) for 7 days in 1 ml of Eagle'sMEM (without phenol red, with penicillin, streptomycin, and amphotericin B)containing 5% charcoal-stripped FCS and the desired test compounds. (Tes

tosterone, DHT and candidate inhibitors were delivered in ethanol vehicle:

final ethanol concentration was 0.2%). The tissue pieces were then transferredon their sponges to a fresh 24-well plate and incubated for 3 days in a freshpreparation of the same medium/treatment containing 2 @Ci1mVwell[3H]thymidine. The tissue pieces on each sponge were then transferred to a l.5-ml

microfuge tube and incubated overnight at 37Â°C in 0.5 ml of collagenase

solution [10 mM Tris (pH 7.8), 0.1 mg/mI collagenase, and 5 mM EDTAI. Next,tissue pieces were transferred to a fresh microfuge tube and incubated for 2â€”3h at 37Â°Cin 0.5 ml of proteinase K solution [10 mM Tris (pH 7.8), 0.1 mg/mIproteinase K, 5 mMEDTA, and 0.5% SDS]. To extract DNA, 0.5 ml of phenolsolution [phenol mixed with an equal volume of 10 mr@iTris (pH 10) and 1 nmiEDTA] was added to each tube, and the contents were mixed gently until anemulsion was formed. The tube was then centrifuged at 1600 X g at roomtemperature for 15 mm or until the organic and aqueous phases had separatedwell. Approximately 0.45 ml of the aqueous phase was transferred to a freshmicrofuge tube, and 0.45 ml of chloroform:isoamyl alcohol (24:1, v/v) was

added. The tube was centrifuged at 12,000 X g for 15mm at room temperature,and then 0.4 ml of the supematant was transferred to a fresh microfuge tubeand mixed vigorously with 0.04 ml of 3 Msodium acetate. This was followedby an addition of 1.25ml of ice-cold ethanol, with gentle mixing. After storageof the tube at â€”70Â°Cfor at least 1 h, it was centrifuged at 12,000 X g at roomtemperature for 10 mm. The pellet was washed in 70% ethanol, and the tubewas centrifuged again. Finally, the pellet was resuspended in 0.4 ml of 10 mMTris (pH 7.8) and 5 mM EDTA. One-half of the 0.4 ml was used for aspectrophotometric determination of DNA concentration (A2@),and the otherone-half was used for measurement of the incorporation of [3H]thymidine into

DNA by liquid scintillation spectrometry.

Growth Effects on LNCaP Cell Cultures

Cultures of the human prostatic cancer cell line LNCaP were maintained in

RPMI 1640/5%FBS. Growth assays were performed in six-well plates (Nunc,Inc., Naperville, IL). LNCaP cells adhere only weakly to untreated tissueculture plastic. Thus, to enhance adherence, each well was exposed prior toseeding to 0.5 ml of a solution of poly-L-lysine (0.067 mg/mi of steriledouble-distilled water) for 30 mm; then the poly-L-lysine solution was removed, and the well was washed with 1 ml sterile water and allowed to drycompletely. Growth assays were then initiated by seeding six-well plates with60,000 LNCaP cells/well (passage number 20â€”28) in incubation medium(RPMI 1640/2% triple-charcoal-stripped FBS, with penicillin, streptomycin,and amphotericin B). After 48 h, the medium of each well was replaced withfresh incubation medium containing the inhibitors and androgens indicated inFig. 4 (added as concentrated ethanol solutions). Control wells were treatedwith ethanol only (final concentration, 0.25%), which was shown in severalprevious experiments to have no effect on growth. Medium/treatments werechanged every 3 days. After 9 days of treatment, cells were removed from thewells with trypsinlEDTA and counted in a Coulter counter (Coulter Electronics, Hialeah, FL).

Transcriptional Activation Assays

Agonist and antagonist properties of 1-33, 1-47, hydroxyflutamide, andfinasteride with human wild-type AR and with the mutant AR with codingsequence identical to that found in LNCaP cells were tested in transcriptionalactivation assays as described previously by Thou et al. (21) with some

modifications. Briefly, CV1 cells (monkey kidney cells) were plated in DMEM

containing 10% calf serum at 0.35 X 106 cells/60-mm dish and transientlycotransfected with 0.1 @gof either wild-type AR or LNCaP mutant ARexpression vector DNA and 5 ,.@gof luciferase reporter vector DNA. Theluciferase coding sequence in the reporter vector is under the control of theandrogen response element of the mouse mammary tumor virus promoter.DNA was precipitated in 4-mI aliquots and incubated for 20 mm at roomtemperature prior to the addition of 250 id/plate. Four ml of DMEM/l0% calf

serum were added to each plate and incubated for 4.5 h at 37Â°C.The mediumof each plate was then replaced with DMEM/0.2% calf serum, and the new

medium was incubated for 16h at 37Â°C.Finally, plates were incubated for 30 hat 37Â°Cwith serum-free/phenol red-free DMEM containing the indicatedconcentrations of inhibitors, with refreshment of the medium/treatment after24 h. Cells were harvested in lysis buffer (Ligand Pharmaceuticals, San Diego,CA), and luciferase activity was determined by adding AlP and luciferin and

4958



Compound 17a-Hydroxylase/C1720-lyase5a-reductase1-33

1547/78814â€•1-34

204/14390â€•1-41

47/42

1-47 11/7120NIâ€•1-49

4/4NTL-lO

NI21Ketoconazole

84/61NIFinasteride

NI144MA

NI10aOnly one experiment was performed.

b NI, no inhibition; NT, not tested.


Compound i.)UU@ULU@41L Base StructureR1 R2,R31-41,

and L-l0 were effective inhibitors of 5a-reductase. The relativepotencies of 1-33 and L-10 were similar to those of theMerckcompounds,

finasteride and 4MA. It is noteworthy thatcompoundsÂ¶@H3

1-33 HO-C-H H, H@ RI --. 3

c@H3

1-34 H-C-OH H,H@I1-34

and 1-41 inhibit both l7a-hydroxylase/C1720-lyase and 5a-reductase, indicating that they could potentially block androgen synthesis in all three tissues targeted in androgen ablation therapy: the testes,adrenals, and prostate.

Inhibition of DNA Synthesis in Androgen-treatedHistoculturesNâ€”11

H,H1-41 1@,Nof

Human BPH Tissue. Because the primary trophic hormone of theprostate is DHT rather than testosterone (7), we hypothesized that thecompounds with activity against 5a-reductase might be able toinhibitCH3

double bondL-1O IH-C-OHthe

stimulatory effect of testosterone on DNA synthesis in prostatic

tissue as a result of their inhibition of the conversion oftestosteroneItoDHT. In addition, compounds possessing antagonist activity fortheNâ€”1

/c6@R1R21-47 R1,R2=H--.

1-49 R1,R2= doi.blebond @AR

would inhibit the growth-stimulatory effect of both testosteroneand DHT. Experiments were, therefore, conducted to determine theeffect of the various inhibitors on the rate of DNA synthesis inandrogen-treated histocultures of human BPH tissue (Fig. 3). Incultures treated with 1 p.Mtestosterone only or 10 nM DHT only,theFig.

2. Chemical structures of compounds with activity against 5a-reductase or17a-hydroxylase/C1720-lyase.rates

of DNA synthesis over a 3-day treatment period were, respectively, 2- and 3-fold that of cultures that were exposed only to theethanol vehicle (data not shown). These control values servedasTable

I Inhibition of human I 7a-hydroxylase/C,720-lyase and 5a-reductase bycandidate inhibitorsreference

standards for cultures treated with a combination of testosterone and inhibitor or DI-IT and inhibitor. Fig. 3 shows that at 1p@M,The

inhibitory effects of the listed compounds on the 17a-hydroxylase/C1720-lyase133, I34, 147, 4MA, and flutamide produced almost complete inhiactivity of human testicular microsomes and the 5a-reductase activity of human BPHmicrosomes were determined as described in â€œMaterialsand Methods.â€•The results froma singleexperimentforeachcompoundarepresented.Similarresultswereobtainedin atleastonerepeatexperiment.exceptwheredesignatedotherwise.bition

of the stimulatory effect of testosterone, and 1-49 was partiallyeffective. 1-33 and 4MA were also highly effective at 0.3 ,LM,and 1-34was partially effective. Only 1-47 and flutamide at 1 @LMdemonstrateda strong inhibition of DHT-stimulated DNA synthesis. None of thecompounds had an independent effect on DNA synthesis in theabsence of added androgens (data not shown). Thus, in no case wasthe inhibitory effect on androgen-stimulated DNA synthesis the resultof a direct toxic effect of thecompound.IC50

(nM)

3

measuring luminescence in a Monolight 2010 luminometer (Berthold, Inc.,Gaithersburg, MD). The average increases of luciferase expression induced by0.1 nMDHT in the control cultures of all experiments with wild-type AR andLNCaP AR coding sequences were 44- and 14-fold, respectively.

RESULTS

Enzyme Inhibition. Compounds were first evaluated for theirability to inhibit l7a-hydroxylase/C1720-lyase and 5a-reductase.Over 70 derivatives of pregnene or androstene were tested. Thestructures ofcompounds that exhibited strong inhibition of at least oneof these enzymes are shown in Fig. 2, and their relative potencies arelisted in Table 1. The inhibitory activities of the previously described5a-reductase inhibitors finasteride and 4MA and the 17a-hydroxylase/C17@-lyase inhibitor ketoconazole were also included in Table 1to serve as standards for comparison. The compounds 1-49, 1-47, and1-41 all displayed greater inhibition of l7a-hydroxylase/C1720.lyasethan did ketoconazole, with 1-49 producing an inhibition that was20-fold more potent than that of ketoconazole. Compounds 1-33, 1-34,

â€” withI @tMTestosteronewithlOnMDHT 2

110__ 10029Â°C8 80

@ 70

U) 60U)5) 50

.C

>â€˜U) 30

z 200 io

0

Conc. (jsM): 0.3 1.0 0.3 1.0 0.3 1.0 0.3 1.0 1.0 0.3 1.0 1.0

Inhibitor: 1-33 1-34 1-41 1-47 1-49 4MA FM.Fig. 3. Effect of compounds on DNA synthesis in histocultures of human BPH tissue

cotreated with androgens. The ability of selected compounds to inhibit the stimulatoryeffect of 1 @.LMtestosterone or 10 nM DHT on DNA synthesis in histocultures of humanBPH tissue was tested as described in â€œMaterialsand Methods.â€•Briefly, fresh humanBPH tissue was cut into small pieces, placed on gelatin sponges in 24-well plates, andcultured for 7 days in phenol red-free MEM/5% charcoal-stripped FBS with the indicatedcompounds and either testosterone or DHT. The medium/treatment of each well was thenreplacedwithfreshmedium/treatmentcontaining2 @.tCi/ml13H]thymidine.After3 days,the medium/treatment of each well was discarded, the tissue was digested, and DNA wasextracted. For each sample, the incorporation of [3H]thymidine into DNA was normalizedto DNA content and expressed as a percentage of the [3H]thymidine uptake in thecorrespondingcontrol samples.which were treatedwith testosteroneonly or DHT only.The numbers above the columns indicate the number of experiments in which thattreatment was tested, and the data are represented as the means of the [3Hjthymidineincorporation in all experiments; bars. SE. Each experiment was performed with BPHtissue from a different patient. Where no number is present, the column indicates the meanincorporation in duplicate wells of only one experiment. Flat., flutamide.

4959



GROWI@H1NHIB@0N OF HUMAN PROSTATE CELLS

The inhibition of testosterone-stimulated DNA synthesis by 1-33,1-34, and 4MA may be attributable to their inhibition of 5cr-reductase.These compounds failed to inhibit the stimulation of DNA synthesisin DHT-treated cultures (Fig. 3), ruling out the possibility that they areAR antagonists. 1-47, however, had no activity against 5a-reductase(Table 1) and yet still inhibited testosterone-stimulated DNA synthesis. This suggests that 1-47 has antagonist activity in addition to itsactivity against llcs-hydroxylase/C1720-lyase, a possibility which issupported by its almost complete inhibition at 1 @u@iof DHT-stimulated DNA synthesis in histoculture. Of the tested compounds, onlyflutamide, a well.known antagonist of the AR, also had effect againstDHT.

Inhibition of Growth in Androgen-freated Culiures of the LNCaP Human Prostatic Cancer Cell Line. The compounds effectiveat reversing the stimulatory effect of testosterone in histocultures ofhuman BPH tissue were further tested in a similar assay using theLNCaP human prostatic tumor cell line (22). Cells were cultured for9 days in a steroid-free medium (RPM! 1640/2% triple-charcoalstripped serum) so that the maximal stimulatory effect of androgentreatment on cell growth could be observed. In preliminary experiments (data not shown), it was determined that the concentrations atwhich testosterone and DHT produce their optimal effects on thestimulation of LNCaP growth are 0.1 and 0.03 flM, respectively, andtherefore, these were the concentrations used in all subsequent experiments. The lower concentrations of androgens required to stimulateLNCaP cell cultures as opposed to histocultures of human BPH tissuemay result from greater access of the compounds to the cell culturemonolayers.

Testosterone at 0.1 nM stimulated cell proliferation by approximately 2-fold compared to control cultures treated with drug vehicleonly (Fig. 4A). At a concentration of 5 @M,1-47, 1-49, L-lO, finasteride, and 4MA all completely reversed the stimulatory effect oftestosterone on cell proliferation. 1-47, finasteride, and 4MA were themost effective compounds, all producing at 1 p.M at least a 50%reduction in the growth effect of testosterone. The compounds 1-33,1-34, and hydroxyflutamide,8 which is the potent metabolite of flutamide in vivo (23), were not only ineffective at reversing the action oftestosterone but actually contributed their own stimulatory effects ongrowth, all of which were much stronger than that of testosterone.This may be attributable to a previously described mutation of the ARof LNCaP cells (24) that increases their responsiveness to progestins(see â€œDiscussionâ€•).

DHT at 0.03 nM stimulated LNCaP growth by 2.8.fold (Fig. 4B),and again, 1-47, 1-49, L-I0, finasteride, and 4MA were all at leastpartially effective at reversing this stimulatory effect. At a concentration of 1 @.LM,all of these compounds reduced the growth effect ofDHT by more than 50%.

To determine whether any of the above compounds reverse thegrowth-stimulating effects of testosterone or DHT on LNCaP cells byexerting a direct toxic effect, a cytotoxicity assay was performed. Thenumber of cells in LNCaP cultures treated for 9 days with candidateinhibitors at concentrations of up to 20 ELM,but not with addedandrogens, were compared with the cell number in control culturestreated only with drug vehicle (Fig. 5). The data show that only 1-49is toxic to LNCaP cells in this assay system and only at concentrationsabove 2.5 /.LM.1-47, finasteride, and 4MA have no independent effecton LNCaP growth; thus, their ability to reverse the growth stimulationproduced by testosterone and DHT does not result from their intrinsictoxicity. At a concentration of 1 p.M. LNCaP growth was increased9-fold by hydroxyflutamide and, interestingly, 2.5-fold by L-l0. Thus,

S Note that cell cultures were treated with hydroxyflutamide (Figs. 4â€”6), whereas

histocultures were treated with flutamide (Fig. 3).

A)Testosteronestimulation [T]= 0.1nM1200 -

1100 - . .Inhibitor

I000 - concen@@@bon

@0.3i@M

@1@tM

@ 2.5 @sM

600- :.i:@

@@@@ ::@@@ :::::::;

:::@ [email protected]==JIII@@I@@ .@..

although L- 10 is by itself mitogenic, it has an overall negative effecton growth when administered in combination with testosterone orDHT in the LNCaP system (compare with Fig. 4).

The effectiveness of the 5a-reductase inhibitors L-lO, finasteride,and 4MA in reversing the effect of testosterone on LNCaP growth wasexpected, due to their inhibitory effect on DHT synthesis. However, asin the histoculture experiments, it was unexpected that 1-47, which

@- 900 -

2.@ 800 -

8 700-

a,.0EC

a,C)

I I I I I IControl T 1.33 1-34 1-47 1-49

only + + + +T T T T

Treatment

I I IL.10 FinastOH-FL 4MA

+ + + +

T T T I

B)DHTstimulation [DHT]= 0.03nM350 -

300 -

.@ 250 -

@-200 -

I 150-:@ ioo -0

50 -

0-

Control DHT 1-47 1-49 L-10 Finast. 4MAonly + + + + +

DHT DHT DHT DHT DHT

TreatmentFig. 4. Effect of compounds on growth of LNCaP cultures cotreated with androgens.

Theabilityof selectedcompoundsto inhibitthestimulatoryeffectof 0.1nMtestosterone(A) or 0.03 nM DHT (B) on the growth of LNCaP cells was tested as described inâ€œMaterialsand Methods.â€•Briefly,LNCaPcells wereseededinto six-wellplatesat adensity of 60,000 cells/well (approximately 6,120 cells/cm2) and cultured for 9 days inRPMI 1640/2% triple-charcoal-stripped FBS with the indicated compounds and eithertestosterone or DHT, or with testosterone or DHT only. Medium/treatment was changedevery 3 days. Control wells were treated with ethanol vehicle only (final concentration,0.25%). The columns represent the means of the cell number in duplicate wells of eachtreatment at the end of the 9-day incubation, expressed as a percentage of the mean cellnumber in the control wells; bars, SE. For each compound, a similar result was obtainedin at least one repeat experiment. T, testosterone; Finast.. finasteride; OH-FL hydroxyflutamide.

4960




less dependent on concentration. Both compounds reduced transcriptional activity by more than 50% at a concentration of 0.1 p.M.However, unlike 1-47, neither 1-33 nor hydroxyflutamide behaved asantagonists of the LNCaP AR. Interestingly, although finasteride didnot show an antagonist effect on the wild-type AR, it did antagonizethe LNCaP AR in a dose-dependent manner similar to that of 1-47;both compounds reduced transcriptional activity by approximately40% at 0.3 p.M and by 50% at 1 p.M.

In an assay for agonist activity using the wild-type AR, none of thecompounds induced transcriptional activity to more than 25% of thelevel of activity induced by 0.1 nM DHT (Fig. 6B). In culturestransfected with the LNCaP AR sequence, however, 1-33 and hydroxyflutamide displayed agonist activity that was similar, and, atsome concentrations, greater than that of DHT. This agonist effectexplains their lack of antagonism against LNCaP AR in the previousassay (Fig. 6A). 1-47 and finasteride generally displayed minimalagonist activity with the LNCaP AR, with the exception of finasterideat 0.3 p.M. which produced an agonist effect that was 46% of the

â€”A OH-FL\ â€”Vâ€”â€”â€¢0 -47@SS\\\s\sss%%.@Ii@ FinasteÃ±de

, I â€˜ I â€˜ I â€˜ I

0 5 10 15 20

Inhibitorconcentration(pM)Fig. 5. Cytotoxicity assay of compounds in LNCaP cell cultures. LNCaP cells were

seeded into six-well plates at a density of 180,000 cells/well. Beginning 48 h later, eachof the indicated compounds, at concentrations ranging from 0.3â€”20@tu,was applied toduplicate wells for 9 days in RPMI 164012% triple-charcoal-stripped FBS, with medium!treatment changes every 3 days. In contrast to the experiments presented in Fig. 4, therewas no co-addition of testosterone or DHT in this assay. Control wells were treated withethanolvehicleonly(finalconcentration,0.25%).Attheendof the9-daytreatment,cellswere removed from the wells with trypsinlEDlA and counted in a Coulter counter. Themean of cell number of duplicate wells ofeach treatment, expressed as a percentage of themeancell numberin thecontrolwells, is plottedversusinhibitor concentration;bars,SE.Similarresultswereobtainedin a repeatexperiment.OH-FL@hydroxyflutamide.

â€”0.1@â€”0.3pM

@1pM

@5pM

- +

WIARTreatment

does not have activity against 5a-reductase, would have an effectagainst growth stimulation by testosterone, or that any of the compounds would have effect against DHT. To understand better theinhibitory effects of some of the compounds on androgen stimulationof growth, their antagonist and agonist properties were investigated infurther detail.

Antagonist and Agonist ACtIVItYof Compounds with Wild-TypeAR and LNCaP AR. The antagonist and agonist effects ofl-47, 1-33,hydroxyflutamide, and finasteride on the wild-type human AR and onthe mutant human AR present in LNCaP cells were evaluated intranscriptional activation assays. An expression plasmid containingthe coding sequence for either the wild-type AR or LNCaP AR wascotransfected into CY1 cells along with a luciferase reporter vector inwhich the luciferase coding sequence is under the control of theandrogen response element of the mouse mammary tumor viruspromoter. Thus, the effects of the various treatments on AR-mediatedtranscriptional activity could be assessed by measuring luciferaseactivity. Low basal activity was observed in cultures transfected withthe parent pCMV expression vector lacking AR sequence (p5) and incultures transfected with wild-type AR sequence in the absence ofDHT (Fig. 6A). DHT at 0.1 ni@iinduced a 44-fold increase in transcriptional activity in cells transfected with wild-type AR codingsequence. In cultures transfected with the LNCaP AR sequence, thebasal activity in the absence of DHT was somewhat higher, and, as aresult, the induction by DHT was only 14-fold.

1-47 exhibited a dose-dependent antagonism of the wild-type AR,producing a 46% reduction in transcriptional activity at a concentration of 1 /.LM.1-33 and hydroxyflutamide displayed even strongerantagonist activity against the wild-type AR, and their effects were

LNCaPAR

Fig. 6. Antagonist and agonist effects of compounds on the wild-type and LNCaP AR.CV1 cells were transiently transfected with either wild-type human AR expression vectorDNA, PCMVhAR, or an expression vector with the mutant human AR coding sequencepresent in LNCaP cells, pCMVLNCaPAR. Cells were also transfected with a reportervectorcomprisedof the luciferasecodingsequenceunderthe controlof the androgenresponse element of the mouse mammary tumor virus promoter. Stimulation of ARmediatedtranscriptionalactivityby the varioustreatmentswas determinedusing aluminescenceassayas describedin â€œMaterialsandMethods.â€•A,antagonisteffectsof thecompounds were assessed by comparing transcriptional activities observed in culturestreated with 0.1 nM DHT with or without the indicated concentrations of compounds. Thedata represent the means of 3â€”5experiments; bars. SE. Included in each experiment asnegative controls were cultures not treated with DHT and transfected with either wild-typepCMVhAR, pCMVLNCaPAR, or the parent expression vector lacking AR sequence,pCMV5 (p5). B, agonist effects were determined by comparing the transcriptional activityin cultures treated with the indicated compounds (without cotreatment with DHT) with theactivity in cultures treated with 0.1 ass DHT. The data represent the means of threeexperiments; bars, SE. For the agonist experiments, the relative transcriptional activitiesobserved in the negative control cultures were similar to those presented in A and are notshown. wi: wild-type; OH-FL hydroxytlutamide; Finast., finasteride.

4961

1000-

@.140 A) Antagonism of 0.1 nM DHT

@ 120

a.!.ba

@80

.@ 60@ H

@[email protected] r,@

@ 20 @Lj It@I! p5 h@I-.@@ @L; @L: .L Li

Inhibtor - - - -47 1-33 OH-FL AnaStDHT(0.1nM): - - + + + + +

Li'@1-47 -33 OH-FLFinast

+ + + +

160 B)Agonisteffect@ 140

@ _o.1,$M@â€˜a@120 O.3@tM

@100@ _5_ -

cc

@ 60@ H130@ I.

@@40 â€¢:

@ @:@ ii@ i:t@ I@ ii@@ .@@

-47 1-33 OH-FL Finast. 1-47 -33 OH-FL Fmast

wTARTreatment



GROWTH 1NHIB@ON OF HUMAN PROSTATE CELLS

magnitude of the effect of DHT. In summary, 1-33 and hydroxyflutamide are antagonists of the wild-type AR but are agonists of theLNCaP AR, whereas 1-47 is an antagonist of both receptors. Atconcentrations up to 5 p.M. finasteride is an antagonist of the LNCaPAR but not of the wild-type AR. Although it inhibits the stimulatoryeffect of DHT on the LNCaP AR, finasteride itself exerts a smallagonist effect on the LNCaP AR (Fig. 6B).

DISCUSSION

It has been recognized that adrenal androgens may play an important role in the progression of prostatic tumors since the early 1940s,when Huggins and Scott first performed bilateral adrenalectomies onpatients who had relapsed after an initial response to castration therapy (25). Thirty years later, Harper et a!. (26) demonstrated thatdehydroepiandrosterone and androstenedione, the major adrenal androgens, are in fact metabolized in the prostate to other androgens,including DHT. Although this study showed that the percentage ofadrenal androgens converted to DHT is low, the authors proposed thatthe quantities of DHT produced in the prostate in this way arephysiologically significant because of the high plasma concentrationof dehydroepiandrosterone. The influence of adrenal androgens hasbeen confirmed in the past few years by two large clinical trials thatdemonstrated the therapeutic advantage of maximal androgen blockade therapy (12â€”14)in which castration is supplemented with anantiandrogen to block the effects of adrenal androgens. The importance of maximal androgen ablation has been further verified in arecent study by Visakorpi et a!. (27) in which AR gene amplificationwas found in 7 of 23 human prostatic tumors that had recurred duringandrogen deprivation therapy. This suggests that some prostatic tumors recur not by becoming truly androgen independent but rather byundergoing an adaptation that permits them to respond to very lowlevels of androgen.

These findings have prompted our efforts to develop new inhibitorsof 17a-hydroxylase/C1720-lyase and Scr-reductase. Compounds withactivity against l7cx-hydroxylase/C1720-lyase may be of particularvalue in that they can potentially suppress the synthesis of bothtesticular and adrenal androgens. Inhibitors of 5a-reductase may beuseful when used in combination with treatments designed to reduceextraprostatic androgen synthesis, since these treatments usually failto produce 100% inhibition. By preventing the conversion to DHT ofthe small amount of androgens that remain in spite of these treatments,5a-reductase inhibitors may be able to produce a more completereduction of the androgenic stimulation of prostatic tumor cells.

We have identified several compounds with activity against theseenzymes. 1-33, 1-34, 1-41, and L-l0, as well as the reference compounds finasteride and 4MA, all displayed activity against 5a-reductase (Table 1). Because DHT has a stronger mitogenic effect on theprostate than does testosterone, it was expected that all of thesecompounds would be able to diminish the stimulatory effect of testosterone on histocultures of human BPH tissue by virtue of theirabilities to block the conversion of testosterone to DHT. Of thecompounds that were tested in the histoculture assay (L-lO andfinasteride were not tested), all did in fact at least partially reduce themitogenic effect observed in testosterone-treated cultures (Fig. 3). Thecompounds L-lO, finasteride, and 4MA produced a similar effect intestosterone-treated cultures of LNCaP cells. At a concentration of 5p.M. these compounds caused almost complete reduction of the growth

effect of testosterone (Fig. 4A). Some compounds, however, althoughactive in histoculture, were not effective in LNCaP cell cultures.Rather than reducing the stimulatory effect of testosterone in LNCaPcultures, the progesterone analogues 1-33 and 1-34 both produced astimulation of growth that greatly exceeded the growth effect of

testosterone (Fig. 4A). This result is not entirely unexpected, given thefact that the AR of LNCaP cells contains a mutation at amino acid868@in theligand-bindingdomain(24)thatincreasestheaffinityofthe receptor for estrogens, progestins, and even some antiandrogens(24, 28). Moreover, this increased binding affinity is associated witha greatly enhanced agonist effect on the receptor, even for the antiandrogens. It is well documented that the growth rate of LNCaP cellscan be stimulated by a variety of steroidal and nonsteroidal compounds, such as estradiol, progesterone, and the antiandrogens cyproterone acetate, nilutamide, and hydroxyflutamide (28â€”31).Activationof the LNCaP AR can apparently account for the growth effect onLNCaP cells of some of these compounds, but not all. Althoughseveral earlier studies had failed to detect steroid receptors other thanthe AR in LNCaP cells (32â€”34),it has recently been reported that bothestrogen and progesterone receptors are present in LNCaP, and furthermore, that the mitogenic effect of estradiol on these cells ismediated by their estrogen receptors (35). Unlike the mechanism ofestradiol stimulation, however, it remains unclear whether the stimulation of LNCaP growth by progestins is mediated by progesteronereceptors or by the mutant AR of LNCaP or by a combination of both.Using a transcriptional activation assay, we have verified that 1-33 isan agonist of the LNCaP AR (Fig. 6B). Although we have not tested1-34, it is likely that it is also an agonist of the LNCaP AR, as it differsfrom 1-33 only in the orientation of the hydroxyl group at the C20position. Thus, the stimulation of LNCaP growth by 1-33 and 1-34may be attributable, at least in part, to an agonist effect on the LNCaPAR.

An interesting aspect of the LNCaP experiments is that L-l0 andthe reference compounds finasteride and 4MA were also able toreduce the mitogenic effect not only of testosterone but also of DHTin LNCaP cultures (Fig. 4B), suggesting that these compounds maypossess antagonist activity for the LNCaP AR. Transcriptional activation assays verified that finasteride is an antagonist of the LNCaPAR but not of human wild-type AR (Fig. 6). The lack of antagonistactivity for the human wild-type AR in our assay is consistent with theprevious finding by Liang et a!. (36) that finasteride does not bind theAR of rat prostatic cytosol, because the rat and human AR haveidentical hormone-binding domains (8). However, to our knowledge,it has not been reported previously that finasteride is an antagonist ofthe LNCaP AR. Because finasteride is not toxic to LNCaP cells atconcentrations up to 20 p.M (Fig. 5), this antagonist effect probablyaccounts for the effectiveness of finasteride in diminishing the mitogenic effect of DHT in LNCaP cultures. This antagonist activityprobably also contributes, together with the inhibition of 5a-reductase, to the effectiveness of finasteride against testosterone (Fig. 4A).Whether this is also true for 4MA and L-l0 remains to be determined.4MA has been shown previously to bind to rat prostate AR (35), butit has not yet been determined whether 4MA and L-lO are antagonistsof the LNCaP AR.

As a 5a-reductase inhibitor, 1-47 has minimal activity, and 1-49 hasnot been tested. However, both of these compounds are potent inhibitors of 17a-hydroxylase/C1720-lyase. Although these compoundswere not expected to reduce the mitogenic influence of testosterone onthe prostate, they were tested in the histoculture assay and the LNCaPcell culture assay to determine if they might have antagonist activity.1-47 was effective against both testosterone and DHT in the histoculture assay, whereas 1-49 was ineffective against DHT and was onlypartially effective against testosterone. This suggested that 1-47 couldbe an antagonist. The data for 1-49 was harder to interpret because ofthe lack of effect against DHT. In the LNCaP assays, 1-47 continued

9 This mutation has been described either as a codon 868 mutation in a 910-codon AR

cDNA or as a codon 877 mutation in a 919-codon AR cDNA.

4962



GROWTh INHIBITION OF HUMAN PROSTATECELLS

to exhibit the characteristics of an antagonist, and 1-49 also displayedantagonistic properties, because both compounds were effectiveagainst both testosterone and DHT. However, it was subsequentlyshown that the behavior of 1-49 could be explained by its cytotoxiceffect on LNCaP cells at concentrations above 2.5 p.M (Fig. 5). Incontrast, 1-47 is not toxic, and transcriptional activation assays confirmed that it is an antagonist of both the wild-type AR and theLNCaP AR.

A wide variety of compounds that are agonists of the LNCaP AR,but not of the wild-type AR, will stimulate the growth of LNCaP cells(29â€”31).If tested only in this system, compounds that would have anegative effect on the growth of prostatic tumor cells containingwild-type AR would be erroneously classified as having either noeffect or possibly even a stimulatory effect on growth if they happento activate the LNCaP AR (e.g., flutamide). However, the LNCaPsystem is not without use. LNCaP is the only androgen-responsivehuman prostatic tumor cell line that will grow continuously in culture.The mutant AR responsible for its aberrant hormonal responsivenesshas been found in a number of other human prostatic tumors (37â€”39).One study reports finding this mutation in the tumors of 6 of 24prostate cancer patients (38). With the LNCaP system, the opportunityexists to test whether a candidate drug would stimulate the growth ofthis subset of tumors. The importance of this concept is evident whenone considers the phenomenon of the flutamide withdrawal syndrome.Patients who have relapsed after an initial period of positive responseto the combination therapy of castration and flutamide often experience significant clinical improvement upon the discontinuation offlutamide (40â€”42). It has been proposed that in these patients, theprolonged exposure to flutamide has resulted in the selective proliferation of tumor cells containing a mutant AR, such as the LNCaPAR, that recognizes flutamide and its metabolites as agonists (40).The withdrawal of flutamide from these patients would, therefore,equate to removal of a tumor growth factor. Thus, in the evaluation ofnew compounds as potential treatments for prostatic cancer, an assessment of their effects on the LNCaP AR or on the growth ofLNCaP cells clearly now has new relevance. In this regard, 1-47 mayoffer an advantage over hydroxyflutamide. Although hydroxyflutamide is somewhat more potent than 1-47 as an antagonist of thewild-type AR, 1-47 is also an antagonist of the LNCaP AR, whereashydroxyflutamide is actually an agonist of this mutant receptor. Forthe treatment of prostatic cancer patients with tumors that possess theLNCaP AR, which may be a substantial percentage (38), a compoundthat is an antagonist of both the LNCaP AR and the wild-type AR,such as 1-47, may be preferable to hydroxyflutamide. Relapsed patients who experience a positive response to the withdrawal of flutamide from their treatment regimen might further benefit from theconcurrent addition of a compound such as 1-47.

Another clinical situation in which a compound such as 1-47 mayfind special use is in the treatment of prostatic tumors that display anincreased responsiveness to adrenal androgens and their metabolites.Recent studies have identified specific mutations in the AR of humanprostate tumors that confer this increased responsiveness (43, 44). Ifthese mutations do indeed contribute to the relapse of some patientsduring androgen deprivation therapy, an inhibitor of adrenal androgensynthesis may be useful in reestablishing disease remission.

In conclusion, we have identified several compounds that couldpotentially achieve a more complete ablation of androgens in patientswith metastatic prostate cancer. The compound 1-47 is particularlypromising in that it has the potential to inhibit testosterone synthesisin both the testes and adrenal glands and also to exert an antiandrogeneffect in tumors containing the LNCaP AR as well as the wild-typeAR. 1-47 may be useful in the treatment of tumors containing theLNCaP AR or a similar mutant AR that recognizes hydroxyflutamide

as an agonist. Alternatively, by helping to produce a more thoroughandrogen ablation, 1-47, and possibly also 1-41 and L-I0, may proveuseful in the treatment of tumors that may have recurred as a result ofAR overexpression, which allows them to respond to very low levelsof androgens, or as a result of an AR mutation, which enhances theirmitogenic response to adrenal androgens. The data and screeningstrategy presented here demonstrate the importance of evaluatingpotential new treatments by several methods, including enzyme assay,histoculture of patient tissue, cell culture, and transcriptional activation assay with both wild-type and mutant AR.

REFERENCES

1. Cheng, E., Lee, C., and Grayhack, J. Endocrinology of the prostate. In: H. Leper andR. K. Lawson (eds. ), Prostate Diseases, pp. 57-73. Philadelphia: W. B. Saunders,1993.

2. Nakajin, S., Hall, P. F., and Onoda, M. Testicular microsomal P-450 for C21 steroidside chain cleavage: spectral and binding studies. J. Biol. Chem.. 256: 6134â€”6139,1981.

3. Chung, B., Picado-Leonard, J., Haniu, M., Bienkowski, M., Hall, P. F., Shively, J. E.,and Miller, W. L. Cytochrome P450cl7 (steroid l7a-hydroxylasef1720 lyase): doning of human adrenal and testis cDNAs indicates the same gene is expressed in bothtissues. Proc. Nail. Acad. Sci. USA, 84: 407â€”411, 1987.

4. Bmchovsky, N., and Wilson, J. D. The conversion of testosterone to 5a-androstan-17@-ol-3-one by rat prostate in vivo and in vitro. J. Biol. Chem., 243: 2012â€”2021,1968.

5. Saunders, F. J. Some aspects of relation of structure of steroids to their prostatestimulating effects. in: E. P. Vollmer (ed), Prostate and Related Tissues (NationalCancer Institute Monograph No. 12), pp. 139â€”159.Washington, DC: U. S. Government Printing Office, 1963.

6. Dorfman,R. I., and Kind, F. A. Relativepotencyof varioussteroidsin anabolicandrogenic assay using the castrated rat. Endocrinology, 72: 259â€”266, 1963.

7. Petrow, V. The dihydrotestosterone (DH'l') hypothesis of prostate cancer and itstherapeutic implications. Prostate. 9: 343â€”361,1986.

8. Thou, Z-X., Wong, C-I., Sar, M., and Wilson, E. M. The androgen receptor: anoverview. Recent Prog. Horm. Res., 49: 249â€”274, 1994.

9. Chodak, 0. W., Kranc, D. M., Puy, L. A., Takeda, H., Johnson, K., and Chang, C.Nuclear localization of androgen receptor in heterogeneous samples of normal,hyperplastic and neoplastic human prostate. J. Urol., 147: 798â€”803, 1992.

10. Huggins, C., Stevens, R. E., and Hodges, C. V. Studies on prostatic cancer. II. Theeffects of castration on advanced carcinoma of the prostate gland. Arch. Surg., 43:209â€”223,1941.

I 1. Labrie, F., Dupont, A., Belanger, A., Lacoursiere, Y., Raynaud, J. P., Husson, J. M.,Gareau, J., Fazekas, A. T. A., Sandow, J., Monfette, 0., Girard, J. G., Emond, J., andHoule, J. G. New approach in the treatment of prostate cancer: complete instead ofpartialwithdrawalof androgens.Prostate,4: 579â€”594,1983.

12. Crawford, E. D., Eisenberger, M. A., McLeod, D. G., Spaulding, J. I., Benson, R.,DolT, F. A., Blumstein, B. A., Davis, M. A., and Goodman, P. J. A controlled trial ofleuprolide with and without flutamide in prostatic carcinoma. N. Engl. J. Med., 321:419â€”424,1989.

13. Crawford, E. D., Eisenberger, M., and McLeod, D. Treatment of newly diagnosedstage D2 prostate cancer with leuprolide and flutamide or leuprolide alone, phase III:prognostic significance of minimal disease. J. Urol., 147: 417A, 1992.

14. Denis, L. Role of maximal androgen blockade in advanced prostate cancer. Prostate,5 (Suppl.): 17â€”22,1994.

15. Trachtenberg, J., and Pont, A. Ketoconazole therapy for advanced prostatic cancer.Lancet, 2: 433â€”435, 1984.

16. Williams, 0., Kerle, D. J., Doble, A., Dunlop, H., Smith, C., Allen, J., Yeo, T., andBloom, S. R. Objective responsesto ketoconazole therapy in patients with relapsedprogressive prostatic cancer. Br. J. Urol., 58: 45â€”51,1986.

17. Jubelirer, S. J., and Hogan, T. High dose ketoconazole for the treatment of hormonerefractory metastatic prostate cancer: 16 cases and review of the literature. J. Urol.,142: 89â€”91,1989.

18. Fleshner, N. E., and Trachtenberg, J. Sequential androgen blockade: a biologicalstudy in the inhibition of prostatic growth. J. Urol., 148: 1928â€”1931,1992.

19. Fleshner,N.E.,andTrachtenberg,J. Treatmentof advancedprostatecancerwiththecombination of finasteride plus flutamide: early results. Eur. Urol., 24 (Suppl. 2):106â€”112,1993.

20. Li, i-S., Li, Y., Son, C., and Brodie, A. M. H. Inhibition of androgen synthesis by22-hydroximino-23,24-bisnor-4-cholen-3-one. Prostate, 26: 140â€”150,1995.

21. Thou, Z-X., Sar, M., Simental, J. A., Lane, M. V., and Wilson, E. M. A liganddependentbipartitenucleartargetingsignalin thehumanandrogenreceptor.J. Biol.Chem.,269:13115â€”13123,1994.

22. Horoszewicz, i. S., Leong, S. S., Kawinski, E., Karr, i. P., Rosenthal, H., Chu, I. M.,Mirand,E. A., and Murphy,G. P. LNCaPmodelof humanprostaticcarcinoma.CancerRes., 43: 1809â€”1818, 1983.

23. Neri, R., Peets, E., and Watnick, A. Anti-androgenicity of flutamide and its metaboliteSch 16423.Biochem.Soc.Trans.,7: 565â€”569,1979.

24. Veldscholte, J., Ris-Stalpers, C., Kuiper, G. G. J. M., Jenster, 0., Berrevoets, C.,Claassen, E., VanRooij, H. C. J., Trapman, J., Brinkmann, A. 0., and Mulder, E. Amutation in the ligand binding domain of the androgen receptor of human LNCaPcellsaffectssteroidbindingcharacteristicsandresponseto anti-androgens.Biochem.Biophys. Res. Commun., 173: 534â€”540, 1990.

4963



GROWTh INHIBITION OF HUMAN PROSTATE CELLS

25. Huggins, C., and Scott, W. W. Bilateral adrenalectomy in prostatic cancer. Ann.Surg., 122: 1031â€”1041,1945.

26. Harper,M. E., Pike,A.,Peeling,W. B.,andGriffiths,K. Steroidsof adrenaloriginmetabolized by human prostatic tissue both in vivo and in vitro. I. Endocrinol., 60:117â€”125,1974.

27. Visakorpi, T., Hyytinen, E., Koivisto, P., Tanner, M., Keinanen, R., Palmberg, C.,Palotie, A., Tammela, T., Isola, J., and Kallioniemi, 0-P. In viva amplification of theandrogen receptor gene and progression of human prostate cancer. Nat. Genet., 9:401â€”406,1995.

28. Schuurmans,A. L.0., Bolt,J., Veldscholte,i., andMulder,E.Stimulatoryeffectsof antiandrogenson LNCaPhumanprostate tumor cell growth,EGF-receptorlevel and acid phosphatase secretion. J. Steroid Biochem. Mol. Biol., 37:849â€”853,1990.

29. Wilding,0., Chen,M., andGelmann,E. P. Aberrantresponsein vitroof hormoneresponsive prostate cancer cells to antiandrogens. Prostate, 14: 103â€”115,1989.

30. Veldscholte,1.,Berrevoets,C.A.,Brinkmann,A.0., Grootegoed,i. A.,andMulder,E.Anti-androgensandmutatedandrogenreceptorofLNCSPcells:differentialeffectson binding affinity, heat-shock protein interaction, and transcription activation. Bio

chemistry, 31: 2393â€”2399, 1992.31. Simard, J., Veilleux, R., de Launoit, Y., Haagensen, D. E., and Labrie, F. Stimulation

of apolipoprotein D secretion by steroids coincides with inhibition ofcell proliferation

in human LNCaP prostate cancer cells. Cancer Res., 51: 4336â€”4341, 1991.32. Schuurmans, A. L. 0., Bolt, J., Voorhorst, M. M., Blankenstein, R. A., and Mulder,

E. Regulationof growthand epidermalgrowthfactorreceptorlevelsof LNCaPprostate tumor cells by different steroids. hit. i. Cancer, 42: 917â€”922,1988.

33. Bems, E. M. J. J., de Boer, W., and Mulder, E. Androgen-dependent growthregulationof andreleaseof specificprotein(s)by the androgenreceptorcontaininghuman prostate tumor cell line LNCaP. Prostate, 9: 247â€”259,1986.

34. Brolin,I., Skoog,L., and Ekman,P. Immunohistochemistryand biochemistryindetection of androgen, progesterone and estrogen receptors in benign and malignanthuman prostatic tumors. Prostate, 20: 281â€”295,1992.

35. Castagnetta,L A.,Micei, M.D.,Sorci,C.M.0., Pfeffer,U.,Farruggio,R.,Oliveri,G.,Calabro,M., andCarruba,G. Growthof LNCaPhumanprostatecancercellsis

stimulated by estradiol via its own receptor. Endocrinology, 136: 2309â€”2319, 1995.36. Liang, I., Cascieri, M. A., Cheung, A. H., Reynolds, G. F., and Rasmusson, G. H.

Speciesdifferencesin prostaticsteroid5a-reductasesof rat,dog,andhuman.Endocrinology,117:571â€”579,1985.

37. Suzuki, H., Sato, N., Watabe, Y., Masai, M., Seino, S., and Shimazaki, J. Androgenreceptorgenemutationsinhumanprostatecancer.J.SteroidBiochem.Mol.Biol.,46:759â€”765, 1993.

38. Gaddipati, J. P., McLeod, D. 0., Heidenberg, H. B., Sesterhenn, I. A., Finger, M. J.,Moul, 1. W., and Srivastava, S. Frequent detection of codon 877 mutation in theandrogen receptor gene in advanced prostate cancers. Cancer Res., 54: 2861â€”2864,1994.

39. Taplin,M-E.,Bubley,G. i., Shuster,I. D., Frantz,M. E., Spooner,A. E., Ogata,G. K., Keer,H. N., and Balk, S. P. Mutationof the androgen-receptorgene inmetastatic androgen-independent prostate cancer. N. Engl. J. Med., 332: 1393â€”1398,1995.

40. Sartor, 0., Cooper, M., Weinberger, M.. Headlee, D., Thibault, A., Tompkins, A.,Steinberg, S., Figg, W. D., Linehan, W. M., and Myers, C. E. Surprising activity offlutamide withdrawal, when combined with aminoglutethimide, in treatment of â€œhormone-refractoryâ€• prostate cancer. i. Natl. Cancer Inst., 86: 222â€”227,1994.

41. Scher, H. I., and Kelly, W. K. Flutamide withdrawal syndrome: its impact on clinicaltrials in hormone-refractory prostate cancer. 1. Clin. Oncol., 11: 1566â€”1572, 1993.

42. Dupont, A., Gomez, i-L., Cusan, L., Koutsilieris, M., and Labrie, F. Response tofiutamide withdrawal in advanced prostate cancer in progression under combinationtherapy. J. Urol., 150: 908â€”913, 1993.

43. Culig, Z., Hobisch, A., Cronauer, M. V., Cato, A. C. B., Hittmair, A., Radmayr, C.,Eberle,I., Bartsch,G., and Klocker,H. Mutantandrogenreceptordetectedin anadvanced-stage prostatic carcinoma is activated by adrenal androgens and progesterone.Mol.Endocrinol.,7: 1541â€”1550,1993.

44. Peterziel, H., Culig. Z., Stober, 1., Hobisch, A., Radmayr, C., Bartsch, 0., Klocker,H., and Cato,A. C. B. Mutantandrogenreceptorsin prostatictumorsdistinguishbetween amino-acid sequence requirements for transactivation and ligand binding.Int. J. Cancer, 63: 544â€”550, 1995.

4964



1996;56:4956-4964. Cancer Res Gregory T. Klus, Junji Nakamura, Ji-song Li, et al. Inhibitors of Androgen Synthesis

by Novelin VitroGrowth Inhibition of Human Prostate Cells

Updated version

http://cancerres.aacrjournals.org/content/56/21/4956

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/56/21/4956To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



growth inhibition of human prostate cells in vitro by...

Documents