Ligand Activation of the Androgen Receptor Downregulates E-Cadherin-Mediated Cell Adhesion and Promotes Apoptosis of Prostatic Cancer Cells

Abstract

Androgen independence is the major cause of endocrine therapy failure in advanced prostate cancer (PC). To examine the effects of human androgen receptor (AR) expression on growth of human PC cells, transfection of full-length AR cDNA in an androgen-insensitive human prostatic adenocarcinoma cell line (DU145) was performed. Transcriptional activity of AR was confirmed by the MMTV luciferase assay and AR expression was assessed by reverse transcriptase polymerase chain reaction, Western blotting, and immunocytochemistry. Two stable transfectant cell lines expressing functional AR were established and passaged over 60 times. Under standard culture conditions, AR expression in transfected cells was predominantly cytoplasmic. Exposure to dihydrotestosterone (DHT; 60 pM-10 nM) resulted in a rapid (maximal at 30 minutes) translocation of AR to the nucleus. Treatment with DHT (5 nM) caused a significant reduction in cell-cell adhesion and aggregation accompanied by a decrease in E-cadherin expression. This was associated with up to 40% inhibition of proliferation and approximately two-fold increase in apoptosis. These results suggest that gene transfer-mediated AR expression in DU145 cells confers sensitivity to DHT, modulates cell-cell adhesion through E-cadherin, and suppresses cell growth by inhibiting proliferation and promoting apoptosis. This provides a model for studies of AR-regulated cell signalling and identification of novel androgen-regulated genes in PC.

Introduction

Prostate cancer (PC) is the second commonest cause of cancer-related deaths in males in the western world and a steep rise in its incidence has occurred in the last two decades [1]. Prostatic epithelial cells are androgen-dependent and eliminating or reducing endogenous androgens by surgical or chemical castration results in subjective and objective improvements in approximately 80% of patients with PC [2,3]. This effect, however, lasts only for a median period of ∼2 years, after which death follows within 6 to 12 months [4,5]. Androgen ablation results in the inhibition of cell proliferation and apoptotic cell death in androgen-dependent PC cells [6–8].

Androgens mediate their effect on target cells through binding to the high-affinity intracellular androgen receptor (AR), a member of the steroid receptor superfamily that belongs to a larger family of ligand-dependent transcription factors [9]. AR is essential for normal male sexual differentiation and prostatic epithelial cell proliferation [10]. Upon ligand binding, cytoplasmic AR undergoes dimerization, leading to exposure of its DNA-binding domain. This receptor-ligand complex is then translocated to the nucleus, where it regulates the transcription of target genes through androgen-responsive elements in the promoter regions [11,12]. Structurally or functionally abnormal ARs may be responsible for altered sensitivity to androgens or ligand-independent activation, resulting in the development of androgen independence (AI) [13,14]. The underlying molecular mechanisms governing the development of AI in human PC are still not clearly understood [15]. Possible contributing factors include amplification, deletions, or mutations in the AR gene [16–23].

Sex steroids (both androgens and estrogens) are also known to modulate cell-cell adhesion in a tissue-specific and cell type-specific manner [24–28]. There are few data investigating the role of androgens in the modulation of expression of cell-cell adhesion molecules, including E-cadherin [24]. E-cadherin, a calcium-dependent transmembrane protein, is the dominant epithelial cell adhesion molecule and a member of a large family of cell-cell adhesion molecules. It is found in adherens junctions where it undergoes homophilic binding [29,30]. E-cadherin interacts with a group of cytoplasmic proteins including α-, β-, and γ-catenin. β- and γ-catenin bind directly to the E-cadherin cytoplasmic tail and α-catenin binds to either β- or γ-catenin in a mutually exclusive complex. α-Catenin links the E-cadherin complex to the cytoskeleton through its interaction with actin filaments [29–33]. In human PC, a decrease in E-cadherin expression is associated with tumor progression [34,35].

Experimental studies into the mechanisms underlying AR-mediated control of gene transcription are limited because of the lack of human PC cell lines that contain stable and functional (nonmutated) AR. Most data are derived from studies on the LNCaP human PC line, which was derived from a lymph node metastasis [36]. This contains a mutated AR [13,37], which is activated not only by androgens but also by estrogen, progesterone, and antiandrogens [14].

The aim of this study was, therefore, to develop an in vitro model that would enable investigation of the mechanism of androgen-mediated fate of PC cells. Specifically, our aims were to: 1) establish an androgen-responsive phenotype in a hitherto AI cell line (DU 145); 2) assess the effect of dihydrotestosterone (DHT) on AR localization; and 3) determine the influence of ligand-mediated activation of AR on cell behavior including cell-cell adhesion, cell growth, and apoptosis.

Materials and Methods

All chemicals were of molecular biology grade and were purchased from Sigma (Poole, UK), unless otherwise stated.

Cell Culture and Reagents

The DU145 cell line and the LNCaP cell line were purchased from the American Type Culture Collection (ATCC; Rockville, MD). DU145 cells and its derivatives were maintained in minimal essential medium (MEM) containing Earle's salts (Gibco BRL, Paisley, UK). LNCaP cells were cultured in RPMI medium (Cancer Research UK, London, UK). Both types of media were supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, and 1% antibiotics/antimycotics (Gibco BRL). Cell cultures were maintained at 37°C in a humidified 10% CO₂ atmosphere.

Steroid-stripped FBS (ssFBS) was prepared by adding 1 g of dextran-coated charcoal into 100 ml of FBS and agitated on a rolling platform overnight at room temperature. After centrifugation at 6000g for 15 minutes, the serum was transferred to a fresh tube and 1 g of dextran-coated charcoal was added. Following agitation for 6 to 8 hours, the serum was spun again and filtered through a 0.22-µm filter.

DHT was purchased form Sigma-Aldrich (Poole, UK). 4-Hydroxyflutamide (4-OH-F), a nonsteroidal AR antagonist, was a kind gift from AstraZeneca (Lund, Sweden).

A mouse monoclonal antibody (MAB) against E-cadherin (HECD1; supernatant or affinity-purified) was obtained from Cancer Research UK; MABs against AR (clone F39.4.1) and α-, β-, and γ-catenins were purchased from Biogenex (Wokingham, UK) and Transduction Laboratories (Calne, UK), respectively.

Construction of Human AR Encoding Mammalian Expression Plasmid

A full-length wild-type human AR cDNA containing a plasmid, psVARo, was a kind gift from Dr. A. O. Brinkmann (Department of Endocrinology, Erasmus University, The Netherlands). Full-length AR cDNA was excised at SalI sites and ligated into the pSP73 multiple cloning vector (cat. no. P2221; Promega, Southampton, UK) at SalI sites. A clone containing AR cDNA in the correct orientation was identified by restriction enzyme analysis and named pSP73-AR. pSP73-AR was digested at BamHI and XhoI sites to obtain a ∼3096-bp AR cDNA-containing fragment, which was then ligated between BamHI and XhoI sites of pcDNA3.1/zeo (Invitrogen, Paisley, UK) to achieve an AR-encoding expression plasmid, pcAR3.1. This plasmid contains a CMV enhancer promoter and zeocin-resistant gene, which allows selection of stable transfectants in the presence of cytotoxic concentrations of zeocin. The integrity of pcAR3.1 and the presence and orientation of insert were confirmed by extensive restriction analysis and plasmid sequencing. For transfection experiments, plasmid DNA was prepared and purified using a Maxi prep DNA kit (Qiagen, Crawley, UK).

Stable Transfection of DU145 Cell Line

A modified calcium phosphate method (Stratagene, Amsterdam, The Netherlands) was employed for stable transfection of the DU145 cell line. Briefly, cells were plated overnight in 5-cm tissue culture dishes (Nunc, Nottingham, UK) at 30% to 40% confluency. The following day, cells in each dish were fed with 5 ml of 10% FBS medium, and either 5µg of pcAR3.1 or pcDNA3.1/zeo (control vector) was added (in mixture with transfection reagents). Transfections were carried out for 7 hours in a humidified atmosphere at 37°C and 5% CO₂. Thereafter, cells were washed twice with warm (37°C) phosphate-buffered saline (PBS) and refed with fresh 10% FBS medium. Zeocin selection (300 µg/ml) was applied after another 24 hours. After 3 weeks of continuous zeocin exposure, a number of zeocin-resistant colonies were ring-cloned and expanded. Initial screening for AR expression was performed using immunocytochemistry (ICC). One control transfectant, Dzeo1, and two AR transfectants, DAR17 and DAR19 (which expressed AR immunoreactivity in close to 100% of cells) were selected for further studies.

RNA Extraction From Cell Lines

Total cellular RNA was extracted from DAR17, DAR19, and Dzeo1 cell lines and a positive control, LNCaP, using RNAzol B (Biogenesis, Poole, UK) according to a protocol described by Stubbs et al. [38]. Briefly, cells grown to approximately 80% confluency in 90-mm tissue culture dishes were washed with three changes of PBS and drained, and 3 ml of RNAzol B was added for 5 minutes. The resultant lysates were transferred into Eppendorf tubes, 200 µl of chloroform was added, and the mixture vortexed and iced for 5 minutes. After centrifugation at 10,000g at 4°C for 15 minutes, the upper aqueous (RNA containing) phase was transferred into a fresh tube and an equal volume of isopropanol was added, mixed gently, placed on ice for 15 minutes, and centrifuged again under the same conditions. The supernatant was aspirated and the RNA pellet washed with 75% ethanol, dried at 65°C for 5 minutes, and resuspended in 20 µl of diethyl pyrocarbonate (DEPC)-treated water. The yield was determined spectrophotometrically at 360 nm.

Reverse Transcriptase Polymerase Chain Reaction (RT-PCR)

Total RNA (2 µg) was denatured at 70°C for 10 minutes and chilled on ice for 2 minutes. cDNA was synthesized from 2 µg of denatured RNA in 50 mM Tris-Cl, pH 8.3, 75 mM KCl, 10 mM dithiothreitol, 3 mM MgCl₂ (Gibco BRL), 40 U of RNAsin (ribonuclease inhibitor; Promega), 500 mM dNTPs (Promega), 1 ml of oligo(dT)_12–18 (Gibco BRL), and 50 U of StrataScriptase (Stratagene, Amsterdam, Netherlands) in a final volume of 20 µl. This mixture was incubated at 37°C for 1 hour and the resultant RNA/cDNA was stored at -20°C until required. PCR for both AR and glyceraldehyde phosphate dehydrogenase (GAPDH) consisted of 100 ng of cDNA, 10 mM Tris-Cl (pH 9), 50 mM KCl, 0.01% gelatin wt/vol, Triton X-100 (Sigma), and 5 mM of each primer, either AR-specific (forward primer dAGCTACTCCGGACCTTACG and reverse primer dAGGTGCCATGGGAGGGTTAG) or GAPDH-specific (forward primer dGCCACATCGCTCAGACACCA and reverse primer dGATGACCCTTTTGGCTCCCC), 200 mM dNTPs (Promega), 1.5 mM MgCl₂, and 1.5 U of Taq polymerase (Promega) in a final volume of 25 µl. Thermal cycling parameters for AR were 94°C for 1 minute; 60°C for 1 minute; 72°C for 3 minutes with a final extension of 72°C for 10 minutes for 20, 25, 30, or 35 cycles; and for GAPDH were 94°C for 1 minute; 65°C for 1 minute; and 72°C for 3 minutes for 15, 20, 25, or 30 cycles in a Techne-PHC3 thermal cycler. PCR products were separated on either a 1% or 1.5% TAE agarose gel, visualized with ethidium bromide (0.5 mg/ml), and photographed under UV light.

ICC

For ICC, cells were grown on four spot glass slides (Hendley, Loughton, UK) in standard culture medium or medium containing ssFBS. Twenty-four hours later, the medium was aspirated gently and replaced with medium containing varying concentrations of the AR ligand, DHT (20 pM-10 nM), or the anti-androgen, 4-OH-F (1 µM). Cells were incubated for variable lengths of time (15 minutes–144 hours), washed with PBS, and fixed in a chilled (-20°C) methanol:acetone (1:1) mixture.

Light microscopy An indirect avidin-biotin-peroxidase complex (ABC) method was used according to the published protocol by Hsu et al. [39]. Briefly, after blocking of nonspecific binding using normal rabbit serum (1:20 dilution in PBS; Dako, Ely, UK) for 20 minutes at room temperature, primary antibody [anti-AR (1:100), anti-E-cadherin (supernatant; neat), anti-α-catenin (1:50), anti-β-catenin (1:200), or anti-γ-catenin (1:100)] diluted in PBS containing 0.1% (wt/vol) bovine serum albumin (BSA) (Sigma) (PSA/BSA) was applied for 2 hours at room temperature in a humidified chamber. After three washes in PBS, biotinylated rabbit antimouse antibody (1:200; Dako) was applied for 1 hour and, subsequently, slides were incubated with avidin-biotin-HRP complex (1:200; Dako) for 45 minutes at room temperature. Peroxidase-H₂O₂ reaction was developed in 0.3% H₂O₂ containing 3,3-diamino-benzidine (DAB; 0.5 mg/ml) solution in PBS. Slides were counterstained with Cole's hematoxylin.

Immunofluorescence For IF, fixed slides were air-dried, washed twice in PBS, blocked for 2 hours in antibody buffer [0.5 M NaCl, 0.05% Tween 20, 1% BSA, and 3% donkey serum (Dako) in PBS], and incubated with MABs against AR (1:10; Biogenex) or E-cadherin (affinity-purified; 0.5 µg/ml) overnight at 4°C. The second layer of Cy-3-conjugated donkey antimouse IgG (1:300; Jackson ImmunoResearch Laboratories, Soham, UK) was applied for 1 hour at room temperature. Slides were then washed and mounted in Vectashield containing nuclear dye DAPI (Vector Laboratories, Peterborough, UK). Images were captured using a digital camera system (Axiocam, Kingston, UK).

For both techniques, a negative control consisted of the application of nonimmune serum as a first layer.

Western Blotting

For preparation of protein lysates, 2x10⁶ cells (in duplicate) in 90-mm tissue culture dishes were plated in 10 ml of culture medium containing ±5 nM DHT (final concentration) for 24 or 48 hours. Cells were washed twice with PBS and lysed in situ by addition of 0.5 ml of lysis buffer (715 mM 2-β-mercaptoethanol, 10% glycerol, 2% SDS, 40 mM Tris, pH 6.8, 1 mM EDTA). The resulting lysate was boiled for 10 minutes with periodic vortexing, iced for 5 minutes, and centrifuged at 12,000 rpm at 4°C for 5 minutes. Clear supernatant was aspirated carefully and stored at -70°C until used. Approximately 20 µg of total cellular proteins (quantified spectrophotometrically at 595 nm) from each sample was subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (100 V for 2 hours), proteins were transferred to nitrocellulose membranes, and the membranes were blocked with 5% nonfat milk in Tris-buffered saline solution containing 0.5% Tween 20 (TBST) for 1 hour at room temperature. The blots were then probed overnight with anti-AR MAB (1:500 in TBST), washed in TBST for 1 hour, and probed with HRP-conjugated rabbit antimouse immunoglobulins (1:1500 dilution; Dako). The blots were placed in Enhanced Chemoluminescent solution (ECL; Amersham, Little Chalfont, UK) and exposed to X-ray film (Hyperfilm; Amersham).

Densitometric Analyses

Protein bands were analyzed using Personal Densitometer SI (Molecular Dynamics, Hercules, CA) linked to a Macintosh (Power Macintosh 8100). Gel images were analyzed with the software program IPLAB gel. Briefly, separate readings were taken for each band; each reading is the sum of pixel intensity for a fixed pixel area (rectangular). For each band, pixel intensities were averaged and background readings were subtracted (measurement taken as above) to give average band intensity.

Transient Transfections and Analysis of Reporter Gene Expression

Twenty four hours after subculturing in 50-mm dishes AR and control transfectants (1x10⁵ cells of each cell line) were incubated with 2.5 µg of pSV β-galactosidase (β-gal; Promega) and 5 µg of MMTV luc (a kind gift from Dr. A. O. Brinkmann, Department of Endocrinology, Erasmus University, The Netherlands) plasmids (using above mentioned protocol for transfection) for 7 hours in a humidified incubator at 37°C and 5% CO₂, washed twice in PBS, and refed with standard medium. The following day, cells were fed with media containing dextran charcoal stripped FBS for 24 hours, and then either 5 nM DHT or 1 µM 4-OH-F was added for a further 24 hours. The cells were harvested by incubating in 300 µl of reporter lysis buffer (Promega) for 15 minutes and scraping all the cells and cellular debris into an Eppendorf. The sample was vortexed for 15 seconds and spun at 16,060g (13,000 rpm) for 2 minutes at 4°C in minifuge. Samples (supernatants) were either used immediately or stored at -70°C.

For β-gal assay (Promega), cell extract (150 µl) was mixed with 150 µl (2x) of assay buffer (120 mM Na₂HPO₄, 80 mM NaH₂PO₄, 2 mM MgCl₂, 100 mM β-mercaptoethanol, and 1.33 mg/ml ONPG (O-nitrophenyl-β-d-galactopyranoside) by vortexing and incubated at 37°C overnight. The reaction was stopped by the addition of 500 µl of 1 M sodium carbonate and the absorption was read at 420 nm in a spectrophotometer. A sample that had not been transfected with pSV β-gal was used as a negative control. The values were used to assess transfection efficiency and to normalize samples (data not shown).

For luciferase (luc) assay (Promega), 5 to 20 µl of cell extracts was mixed with 100 µl of luc assay buffer (Promega) (20 mM tricine; 1.07 mM (MgCO₃)·4Mg(OH)·2.5H₂O; 2.67 mM MgSO₄, 0.1 mM EDTA, pH 7.8, 33.3 mM DTT, 270 µM coenzyme A, 470 µM luciferin, 530 µM ATP) and immediately counted on a scintillation counter (Beckman LS 1801). The light produced was measured for 1 minute and the data were analyzed as counts per minute. The value obtained from a sample that had not been transfected with MMTV luc was used to calculate the background reading, which was then deducted from each result.

MTT Cell Proliferation/Viability Assay

MTT cell proliferation/viability assays were conducted in 96-well tissue culture plates (Nunc). Briefly, 5x10³ cells/well in 200 µl of medium with or without 5 nM DHT (final concentration) were plated in quadruplicate. Plates were kept at 37°C in 10% CO₂ in a humidified incubator. Cells were assayed daily for reduction of MTT for five consecutive days. Fifty microliters of MTT (2.5 mg/ml) was added into each well and plates were kept for 4 hours at 37°C in a humidified incubator. Medium was removed by gently inverting the plates. A mixture of 100 µl of dimethylsulphoxide (DMSO; BDH, Lutterworth, UK) and 25 µl of glycine buffer (pH 10.5) was added into each well and plates were kept on an orbital shaker for 10 minutes to develop and homogenize color. Plates were read in an ELISA plate reader (BioRad 550, Hemel Hempstead, UK) at 540 nm.

FACS Analysis

For cell cycling, the cell lines Dzeo1, DAR17, and DAR19 were seeded into 5-cm culture dishes and grown in the presence of 5 nM DHT for 0, 1, 3, and 6 days. The cells were washed twice in PBS, harvested in versene only, and washed in PBS. The cells were fixed in 70% ice-cold ethanol on ice for 1 hour, washed in PBS, and counted, and 2x10⁶ cells were used. The cells were incubated with 1 ml of 5 µg/ml propidium iodide (PI) and 0.25 mg/ml RNAse 1 for 30 minutes at 37°C before being transferred to Falcon tubes and analyzed on a cell sorter (Beckman Coulter XL flow cytometer, High Wycombe, UK).

For E-cadherin, cells were harvested using versene only after reaching 80% confluency, spun at 400g, and resuspended in PBS containing 1% FBS. After obtaining a single cell suspension, cell viability and number were counted using trypan blue and 1x10⁶ viable cells were used per reaction and always kept on ice. The cells were then washed three times using 1% FBS/PBSA and spun at 400g for 1 minute to pellet the cells. Affinity-purified anti-E-cadherin (HECD1) was added at 10 µg/ml in 50 µl of PBS containing 1% FBS and incubated on ice for 45 minutes prior to washing three times in 1% FBS in PBSA and pelleting the cells at 400g. An FITC-labelled rabbit antimouse secondary antibody (Dako) was added to the samples at 1:40 dilution in 500 µl of PBS containing 1% FBS and incubated for 45 minutes on ice. After three washes in 1% FBS/PBSA, the cells were resuspended in 1 ml of 1% FBS/PBSA, transferred to Falcon tubes, and analyzed on a cell sorter (Beckman Coulter XL flow cytometer). The integrity of E-cadherin was maintained by addition of 1 mM calcium chloride to the FBS/PBSA throughout the experiment [40]. To assess the effects of 5 nM DHT and 1 µM 4-OH-F on cell adhesion molecule expression, these agents were added to the standard medium 24 hours prior to harvesting the cells. The LNCaP cell line known to express E-cadherin was used as a positive control.

Trypan Blue Exclusion Assay

For the determination of viable and nonviable cell count, cells grown in the presence and absence of 5 nM DHT were harvested at 1, 3, and 6 days. Both adherent and floaters were pelleted by centrifugation and resuspended in 3-ml culture medium. A 10-µl aliquot of this suspension was mixed with an equal volume of trypan blue solution (0.4%; Sigma). The viable and nonviable cells were counted using a hemocytometer.

Cell Death Detection ELISA

A cell death ELISA kit (Boehringer Mannheim, Bracknell, UK) was used to detect the cytoplasmic fraction of histone-bound oligonucleosomal fragments as described by Lalani et al. [41]. Briefly, 10⁵ cells from DHT-treated (5 nM) and nontreated cultures were taken at 1, 3, and 6 days and subjected to lysate preparation according to manufacturer's instructions. Experiments were conducted in 96-well microtiter plates. Triplicate wells for each lysate were coated with 100 µl of antihistone antibody for 1 hour at room temperature, washed three times with wash buffer, and blocked with incubation buffer for 30 minutes. After removing the incubation buffer, the plates were incubated with 100 µl of cell lysates per well (representing 10⁴ cells) for 90 minutes at room temperature. Wells were washed three times, 3 minutes each, with wash buffer and then incubated with an anti-DNA antibody conjugated with peroxidase (anti-DNA POD) for 1 hour. After three washes with the wash buffer, color was developed using 2,2′-azino-di-[3-ethylbenzothiazoline sulphonate] (ABTS) (a chromogenic substrate for peroxidase) for 20 minutes and the plate was read on the ELISA plate reader at 405 nm.

Cell Aggregation Assay

Single cell suspensions of the transfected cell lines were obtained using versene only, as described previously. The cells were washed in MEM/1% FBS and counted. About 1 x10⁵ cells were resuspended in 3 ml of MEM/1% FBS (contains Ca²⁺) or PBSA/1% FBS (no Ca²⁺) and incubated in a shaking incubator at 37°C for 1 hour. The number of single cells was counted using a hemocytometer and the results were presented as the cell aggregation index (number of cells remaining single divided by total) over time. To assess the effects of DHT and 4-OH-F on cell aggregation, 5 nM DHT and/or 1 µm 4-OH-F was added to the media for 18 hours before initiation of the experiment and was included in the MEM/1% FBS during the incubation. The absence of DHT was tested by growing the cells in media containing 10% ssFBS and repeating the experiment in this media. To test whether the aggregation effect was mediated by E-cadherin, the cells were incubated with 10 µg/ml affinity-purified HECD1 for 30 minutes prior to starting the aggregation.

Immunoprecipitation

Dzeo1, DAR17, and DAR19 were seeded into 150-mm culture dishes and grown to 80% confluency in standard medium, standard medium supplemented with 5 nM DHT for 18 hours, or standard medium supplemented with 1 µm 4-OH-F for 18 hours prior to being harvested. The cells were washed three times in PBSA, which was completely removed, and 5 ml of nondenaturing immunoprecipitation buffer (0.5% Triton-X, 50 mM NaCl, 10 mM PIPES, pH 6.9, 3 mM MgCl₂, 300 mM sucrose) was added. The dishes were placed on ice for 30 minutes with occasional rocking. The dishes were scraped and the cells were placed in 15-ml Falcon tubes, which were placed on ice for a further 30 minutes with gentle vortexing. The samples were transferred to Eppendorfs and spun at 13,000 rpm for 10 minutes at 4°C. The supernatant was transferred to a fresh Eppendorf and stored at -70°C, or processed immediately. The samples were spun again for 30 minutes at 13,000 rpm at 4°C and the supernatant transferred to a fresh Eppendorf. To each 1-ml sample, 1 µg of antibody (either affinity-purified HECD1, anti-α-catenin, anti-β-catenin, anti-γ-catenin, or rabbit antimouse IgG; Dako) was added and the tubes were rotated at 4°C overnight. The following morning, Sepharose G beads were prepared by washing twice in buffer 2 (0.1% Triton-X, 15 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, pH 8.0, 2.5 mM EGTA, pH 8.0). Each sample (supernatant/antibody mix) was added to 50 µl of prepared beads and placed on a rotator at 4°C for 2 hours after which the beads/antibody complexes formed were harvested by spinning at 13,000 rpm at 4°C. The beads were washed and retained twice in buffer 2, once in buffer 3 (15 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 5 mM EDTA, 2.5 mM EGTA), and once in buffer 4 (50 mM Tris-HCl, pH 6.8, 5 mM EDTA). After removal of buffer 4 from the beads, 25 µl of 1x lysis buffer (715 mM 2-β-mercaptoethanol, 10% glycerol, 2% SDS, 40 mM Tris, pH 6.8, 1 mM EDTA) was added to the beads and the samples boiled for 5 minutes. The samples were spun and the supernatant containing the dissociated immunocomplexes was subjected to Western blotting and probed with anti-E-cadherin (Transduction Laboratories).

The data were assessed for statistical significance by t-tests and a P value of <.05 was considered to be significant.

Results

Morphological Assessment of Transfectants

The only observable difference between the AR transfectants (DAR17 and DAR19) compared to the control, Dzeo1, was that the cells did not grow as tightly packed together (cell-cell), which was more noticeable when DHT was added to the growth medium (Figure 1, A and B). All the transfectants have been passaged over 60 times and maintained the above pattern of growth.

Figure 1.

(A and B) Effect of 5 nM DHT on cellular morphology of AR and control transfectants. Phase contrast microscopy of Dzeo1 control transfectant (A) and DAR17 AR transfectant (B) grown on plastic in the presence of 5 nM DHT. Dzeo1 cells show typical epithelioid morphology with close cell-cell contacts, whereas in DAR17, the cells do not grow as tightly together. (C) Ethidium bromide-stained PCR products. RT-PCR was performed by utilizing AR- and GAPDH-specific primers on total cellular RNA extracted from DAR17, DAR19, Dzeo1, and LNCaP cell lines. A single AR-specific band is observed in the DAR17, DAR19, and in the positive control LNCaP. The GAPDH bands indicate equal loading and optimal RT-PCR. (D) A Western blot probed with anti-AR monoclonal antibody shows a single 110-kDa AR reactive band in AR transfectants, DAR17 and DAR19, and LNCaP (control) cell line. No AR-specific protein band was observed in the control transfectant (Dzeo1) and the DU145 cell line.

Analysis of AR Expression

RT-PCR performed on total cellular RNA extracted from cell lines confirmed the presence of the predicted size AR PCR products in AR transfectants but not Dzeo1 control (Figure 1C). AR-specific primers were designed to amplify a cDNA fragment corresponding to ligand-binding and DNA-binding domains of the AR spanning seven exons. RT-PCR for GAPDH expression served as a positive control. Western blotting demonstrated a single band of 110 kDa corresponding to the molecular weight of AR protein in both AR transfectants but not in Dzeo1 control or the parental line DU145 (Figure 1D). The AR transfectants expressed similar levels of AR protein. Under standard growth conditions, both AR transfectants (DAR17 and DAR19) expressed cytoplasmic and nuclear AR. The control, Dzeo1, did not express any immunoreactivity to AR (data not shown).

Nuclear Translocation of AR in Response to DHT

Addition of 5 nM DHT to the culture medium resulted in a rapid (from 15 minutes, maximal at 30 minutes) nuclear translocation of the AR with minimal residual cytosolic expression (Figure 2, A and B)—this effect could not be observed in either Dzeo1 or parental line DU145 (not shown). The minimum concentration of DHT required for complete translocation under standard conditions was 0.1 and 0.5 or 0.4 nM DHT for DAR17 or DAR19 transfectants, respectively, when cells were grown in ssFBS (data not shown). Treatment with physiological concentration (1 µM) of 4-OH-F for 1 hour also resulted in nuclear translocation of AR in DAR17 and DAR19 AR transfectants (data not shown).

Figure 2.

(A and B) Immunocytochemical localization of AR in DAR19 cells. In cells grown in the absence of DHT, AR expression is localized mainly to the cytoplasm (A). Thirty minutes of treatment with 5 nM DHT results in a complete translocation of AR to the nucleus (B). All cells express nuclear AR but the intensity of staining is heterogeneous. Nuclei stained with DAPI. (C) Immunoblotting using whole cell lysates containing equivalent quantity proteins (20 µg). A single 110-kDa AR reactive band was observed in all lanes. Addition of 5 nM DHT (DAR/DHT) to the culture medium did not alter total AR protein content in either DAR17 or DAR19. (D) DHT-dependent transactivation of the MMTV luc reporter plasmid in DAR17 and DAR19. In the absence of DHT, luc activity was only 2.5-fold compared to Dzeo1. A 5-nM DHT treatment increased this 23- to 25-fold. Treatment with 1 µM 4-OH-F resulted in levels of luc activity in DAR17 and DAR19 similar to controls (i.e., in the absence of DHT). Error bars represent mean±SD.

Western blotting of DHT-treated (5 nM for 24 or 48 hours) and nontreated AR transfectants did not show any demonstrable alteration in the total cellular AR protein content (Figure 2C) (i.e., both transfectants expressed similar levels of AR protein after addition of DHT as before treatment).

Transfected AR is Transcriptionally Functional

Treatment with 5 nM DHT increased luc production in the AR transfectants up to 25-fold compared to that of Dzeo1 (DAR17: P = 0.003; DAR19: P = 0.04) (Figure 2D). Addition of 1 µM 4-OH-F did not induce any comparable activation of MMTV luc in DAR17 and DAR19 (DAR17: P = 0.08; DAR19: P = 0.15) (Figure 2D).

Ligand-Activated AR Results in Decreased E-Cadherin Expression

ICC and FACS analysis demonstrated membranous with some cytoplasmic E-cadherin expression in the majority of Dzeo1 cells (∼70%). Addition of 5 nM DHT or 1 µM 4-OH-F had no effect on expression (DHT: P = 0.78; 4-OH-F: P = 0.8). Up to 55% of the cell population of DAR17 and DAR19 expressed membranous with some cytoplasmic E-cadherin when grown in standard culture conditions (Figure 3A). Addition of 5 nM DHT to the cultures resulted in a decrease of E-cadherin expressing cells to 30% (DAR17: P = 0.02; DAR19: P = 0.045) (Figure 3, B and C). Addition of 4-OH-F did not affect the percentage of cell surface expression of E-cadherin (Figure 3C). Results of FACS analysis were normalized to levels of E-cadherin expression in the LNCaP cell line.

Figure 3.

(A and B) E-cadherin expression in DAR19 cells in the absence (A) and presence (B) of 5 nM DHT in the growth medium. Note a decrease in the number of immunoreactive cells as well as a disruption of membranous expression of E-cadherin caused by 30-minute treatment (arrows). (C) The figure summarizes the results of FACS analysis for E-cadherin expression in AR and control transfectants. The expression of E-cadherin in LNCaP, Dzeo1, DAR17, and DAR19 was examined in cells grown in the absence or presence of DHT and 4-OH-F by FACS analysis. Percentage of positive cells from 5000 counted is summarized as a graph. The graph shows a reduction in surface E-cadherin expression in DAR17 and DAR19 cells treated with 5 nM DHT (error bars±SD from three independent experiments).

Dzeo1, DAR17, and DAR19 cells exhibited identical expression of α-catenin, which was found to be localized mainly to the cell membrane with some cytoplasmic expression, and addition of DHT did not affect either the staining or the number of cells expressing α-catenin (data not shown). A similar pattern was observed for β- and γ-catenin (data not shown).

AR Transfectants Exhibit DHT-Dependent Decreased Cellular Aggregation

After 1 hour, 60% maximal aggregation was observed in the Dzeo1 cell line, which was unaffected by treatment of cells with DHT or 4-OH-F, alone or in combination (Figure 4A). Similarly, no effect on cellular aggregation was observed with the use of ssFBS during experiments.

Figure 4.

Graphs show the results of aggregation assays conducted on AR transfectants and control transfectant over a period of 1 hour after exposing them to different culture conditions (see Materials and Methods section). The cells were counted every 20 minutes and the ratio (number of single cells divided by the total number of cells) is plotted against time. Error bars represent ±SEM. (A) The graph shows no significant difference in Dzeo1 cellular aggregation under different culture conditions. (B and C) Results of aggregation assays for DAR17 (B) and DAR19 (C) AR transfectants show significant differences (see text) between aggregation of cells, which were incubated under different culture conditions. (D) Graphical representation of the cellular aggregation assay conducted after incubating the control transfectant and AR transfectants in the presence (30 minutes prior to initiation of the assay) or absence of anti-E-cadherin blocking antibody (HECD1). Cellular aggregation was reduced to less than 30% in Dzeo1 (P = 0.003), DAR17 (P = 0.03), and DAR19 (P = 0.03).

Aggregation of DAR17 and DAR19 was diminished in comparison with Dzeo1 (approximately 45% after 1 hour vs 65% in Dzeo1; Figure 4, A–C). Cellular aggregation of DAR17 and DAR19 was further reduced when the cells were incubated with 5 nM DHT (about 35% during the same time range; DAR17: P = 0.018; DAR19: P = 0.01) (Figure 4, B and C). AR transfectants treated with 4-OH-F alone exhibited up to 65% aggregation (DAR17: P = 0.0007; DAR19: P = 0.003) and treatment for 6 hours with 4-OH-F prior to incubation with 5 nM DHT abrogated the DHT-mediated reduction of aggregation with approximately 60% of the population aggregated compared to 35% (when treated alone) (DAR17: P = 0.002; DAR19: P = 0.0002). Use of ssFBS in the growth media increased the cellular aggregation of both DAR17 and DAR19 transfectants, with approximately 65% of cells aggregated at the end of 1 hour (P = 0.01 and P = 0.04, respectively) (Figure 4, B and C).

As E-cadherin-mediated cell-cell adhesion requires calcium ions, the experiments were repeated in Ca²⁺-free conditions where there was virtually no aggregation of the cells (data not shown). To further confirm that aggregation was mediated by E-cadherin, blocking studies with HECD1 anti-E-cadherin antibody were performed and aggregation was reduced to less than 30% in all cell lines (Dzeo1: P = 0.003; DAR17: P = 0.003; DAR19: P = 0.03) (Figure 4D).

DHT-Mediated Decrease of E-Cadherin in Immunocomplexes of β- and γ-Catenin

In Dzeo1, E-cadherin was found to be present in the immunoprecipitates formed with E-cadherin, β-catenin, and γ-catenin antibodies, but not in those formed by α-catenin antibody (Figure 5), suggesting that E-cadherin was complexed with β- and γ-catenin but not α-catenin. The relative quantities of each protein were not changed by the addition of DHT (Figure 5) or 1 µM 4-OH-F (data not shown).

Figure 5.

E-cadherin/catenin complexes were immunoprecipitated using E-cadherin, α-catenin, β-catenin, and γ-catenin antibodies after growth in normal media and media supplemented with 5 nM DHT. Subsequent Western blotting used the alternative anti-E-cadherin (Transduction Laboratories) antibody. In Dzeo1, E-cadherin levels were unaffected by addition of DHT, whereas in DAR17, a reduction in E-cadherin was observed in immunocomplexes of γ-catenin and, in DAR19, a reduction in E-cadherin was observed in immunocomplexes of β- and γ-catenin following incubation with DHT. The antibodies used in immunoprecipitation act as internal controls for protein concentration determination.

Similarly, E-cadherin was identified in the immunoprecipitates formed by E-cadherin, β-catenin, and γ-catenin antibodies in DAR17 and DAR19 AR transfectants (Figure 5). Addition of DHT to the growth medium resulted in a decrease in the quantitative expression of E-cadherin associated with γ-catenin in immunoprecipitates of DAR17 transfectants and β- and-γ catenin in immunoprecipitates of DAR19, suggesting a disruption of these complexes (Figure 5). Addition of 1 µM 4-OH-F gave results identical to standard growth conditions (data not shown).

DHT-Mediated Growth Inhibition and Apoptosis in AR Transfectants

Under standard growth conditions, there was no difference in cellular proliferation among DAR17, DAR19, and Dzeo1. In the presence of 5 nM DHT, up to 40% decrease in cellular proliferation was observed in DAR17 and DAR19 (DAR17 and DAR19: P = 0.003) (Figure 6), with no change in Dzeo1 (P = 0.92).

Figure 6.

Growth profiles of AR transfectants and controls were determined over a 5-day period in the presence and absence of 5 nM DHT using the MTT assay. Results of cell proliferation are presented as percentage of DHT-free controls. There were no appreciable growth differences between AR transfectants and controls in the absence of DHT. Addition of 5 nM DHT resulted in decreased cellular proliferation (up to 40%) in DAR17 and DAR19 compared to DHT-free controls. The DU145 and Dzeo1 control transfectants exhibit a similar growth profile in the presence and absence of 5 nM DHT. Error bars represent ±SD over three independent experiments.

FACS analysis revealed similar numbers of cells in each stage of the cell cycle for DAR17, DAR19, and Dzeo1. Supplementing the growth media with 5 nM DHT resulted in an increase in the percentage of the cell population of DAR17 and DAR19 in the G₁ phase of the cell cycle and a concomitant decrease in the numbers in the S and G₂/M phases (P = 0.048 and P = 0.032, respectively) but had no effect on the Dzeo1 cells (P = 0.57). These data suggest that the DAR17 and DAR19 cells are still proliferating but are held in G₁ for a longer time than when grown under standard culture conditions (Table 1).

Table 1.

Percentage of Cells in the Different Phases of the Cell Cycle, ±SD, as Determined by FACS Analysis.

	Dzeo 1	Dzeo1+5 nM DHT	DAR17	DAR17+5 nM DHT	DAR19	DAR19+5 nM DHT
G1	60.7±0.1	59.8±4.9	74.2±4.5	81.6±1.5	78.6±5.2	86.2±4
S	13.7±0.4	13.2±2.8	7.6±1.1	6.8±1	7.2±0.7	3.6±0.1
G2/M	24.9±0.8	26.6±2.1	18±3.4	11.4±2.7	13.9±4.2	9.9±3.9

The data show that there is little difference in the percentage of cells in each stage of the cell cycle under standard growth conditions. However, addition of 5 nM DHT to the media for 72 hours resulted in an increase in the percentage of DAR17 and DAR19 cells in the G₁ phase of the cycle with a concomitant drop in S and G₂/M (P = 0.048 and P = 0.032, respectively), whereas Dzeo1 cells were unaffected (P = 0.57).

Occurrence of cell death, evidenced by nuclear fragmentation, was observed in DAR17 and DAR19 when grown in the presence of 5 nM DHT for over 144 hours (Figure 7, A and B).

Figure 7.

(A and B) Photomicrographs showing nuclear fragmentation in 5 nM DHT-treated AR transfectants (AR expression is predominantly nuclear). Cells were grown in the presence of DHT for 3 days (A) (x400) and 6 days (B) (x400) and ICC stained for AR. Black arrows indicate cells undergoing apoptosis. (C) AR transfectants and controls were cultured in the presence and absence of 5 nM DHT for 1, 3, and 6 days before trypan blue (0.4%) dye exclusion testing. Cell were counted on a hemocytometer. Up to three-fold higher nonviable, trypan blue-positive cells were present in DAR17 and DAR19 incubated with 5 nM DHT compared to DHT-free controls at days 3 and 6. The control transfectant and the parental DU145 cell line did not exhibit any appreciable differences. Results are presented as a percentage of DHT-free controls. Error bars represent ±SD over three independent experiments. (D) Cell death detection ELISA was conducted on the lysates prepared from AR transfectants and controls (parental DU145 and vector control transfectant) at 1, 3, and 6 days (incubated in the presence or absence of 5 nM DHT). Higher color absorbance values for DHT-treated AR transfectants at days 3 and 6 indicated the presence of high levels (up to two-fold at days 3 and 6) of histone-bound oligonucleosomes compared to DHT-free controls. Results are presented as percentage of control absorbance of DHT-free controls. Errors bars represent ±SD over three independent experiments.

Trypan blue results indicated up to three-fold increase in nonviable cells in DHT-treated AR transfectants at days 3 and 6 compared to DHT-free controls (DAR17 and DAR19: P = 0.001) (Figure 7C). Dzeo1 did not exhibit any difference in nonviable cell count in the presence and absence of 5 nM DHT.

Results of cell death detection ELISA indicate the presence of significantly higher levels (up to two-fold) of histone-bound oligonucleosomes in DHT-treated AR transfectants compared to nontreated controls (DAR17 and DAR19: P = 0.001) (Figure 7D). The control transfectant did not show comparable difference in the levels of oligonucleosomes in the presence or absence of 5 nM DHT (P = 0.84). These results substantiated those obtained by trypan blue exclusion assay.

Discussion

We describe the generation of two stable androgen-responsive AR transfectants from the AI DU145 human prostatic cell line. AR transfectants expressed a functionally active AR, which exhibited ligand (DHT)-mediated nuclear translocation on exposure to physiological concentrations of DHT. AR nuclear translocation was seen over a range (60 pM-10 nM) of DHT, with translocation observed at 15 minutes and completed by 30 minutes. Nuclear transport of AR has been previously demonstrated in studies utilizing prostatic and nonprostatic cells [42,43]. Jenster et al. [43] demonstrated cytosolic expression of unliganded AR in transiently transfected Cos-1 and nuclear localization in HeLa cells. With the addition of ligand, AR translocated to the nucleus in the Cos-1 cells. They also demonstrated that the signal responsible for nuclear import was encoded by the amino acid residues 608 to 625 and was functionally similar to the bipartite nucleoplasmin nuclear localization signal. AR transfected into Cos-7 cells also resulted in a cytoplasmically localized AR, and nuclear transport occurred after addition of 50 nM DHT [42]. In comparison, we have shown that 60 pM DHT was required for translocation and 100 pM DHT (∼500x less) was sufficient to induce complete nuclear translocation of the AR in our stable transfectants.

We have also demonstrated that 5 nM DHT resulted in AR-dependent activation of the MMTV luc reporter plasmid and that 4-OH-F could not induce any comparable activation of this reporter plasmid, despite the fact that the same dose of 4-OH-F induced complete AR nuclear translocation. 4-OH-F has the capacity to bind to the AR, resulting in nuclear translocation without transcriptionally activating the receptor. This has previously been demonstrated by Jenster et al. [43] in transiently transfected Cos-1 and HeLa cells in which antiandrogens and a range of steroid hormones induced nuclear localization of the wild-type AR in a dose-dependent way, without transcriptional activity. Collectively, these studies performed in epithelial cells of differing origins (kidney, cervix, and prostate) demonstrate that the mechanisms underlying ligand-mediated AR cytoplasmic to nuclear shuttling may be similar between androgens and antiandrogens. Furthermore, nuclear translocation per se is insufficient for transcriptional activity and is dependent on other factors (conformational change, cofactors, etc.).

Following ligand activation, the presence of nuclear AR protein could have been due to: 1) translocation of preformed AR located in the cytoplasm, or 2) translocation of preformed and de novo synthesized AR. Does ligand-mediated nuclear translocation result in de novo protein synthesis? In order to address this, cells were incubated in the presence of 50 µg/ml cycloheximide (a protein synthesis inhibitor, also used in the investigations of nucleocytoplasmic shuttling) [44,45]. Growth of DAR17 and DAR19 in the presence of ligand (DHT) and 50 µg/ml cycloheximide resulted in cytoplasm nuclear shuttling without de novo AR protein synthesis (data not shown). The accumulation of the AR in the nucleus was that of preformed cytoplasmic AR. Our observations are consistent with those reported by Lin and Shain [46] and Dai et al. [47]. Dai et al. showed increased AR protein levels in transiently transfected PC3 cells but not in DU145 cells despite increased mRNA in both cell lines. Lin and Shain showed that exogenous ligand (testosterone) either increased AR levels (two-fold to three-fold) or had no effect in clonally derived rat PC cell lines. It is conceivable that transcriptional and translational ligand-mediated regulation of AR is cell line-specific.

In our study, both AR transfectants exhibited decreased cohesive growth and addition of DHT increased this effect. This was shown to be due to a decrease in E-cadherin expression in these cells. Steroids (including androgens and estrogens) have been reported to modulate cell-cell adhesion in a tissue- and cell type-specific manner. For example, estradiol increases E-cadherin mRNA levels in the ovaries of immature mice and increases cadherin expression in rat granulosa cells [48,49]. In endometrium, when cells in the secretory phase are treated with estradiol, the levels of E-cadherin and catenins are reduced [50]. The cell-to-cell aggregation activity and the expressions of E-cadherin, and α- and β-catenin mRNA in Ishikawa cells (derived from well-differentiated endometrial cancer) were significantly suppressed by estrogen, resulting in detachment of these cells from the cultured surfaces [51].

AR expression in prostatic carcinoma cells has been reported to suppress cell adhesion molecules [52]. Bonaccorsi et al. [52] have reported that AR expression in androgen-insensitive PC3 cells resulted in decreased expression of α₆β₄ integrin. Androgens have been reported to repress an epithelial cell adhesion molecule, C-CAM, in the rat ventral prostate and seminal vesicle but not in other androgen-dependent organs (the coagulating gland and the dorsolateral prostate), or other organs (the liver and kidney) [25]. These observations suggest that regulation of C-CAM expression by androgen is tissue-specific. The exact mechanism by which AR and DHT are involved in E-cadherin regulation needs further evaluation. It could be that AR mediates transcriptional regulation of the E-cadherin promoter in the AR transfectants. Hennig et al. have examined the in vivo properties of the genomic E-cadherin promoter in well and poorly differentiated carcinoma cell lines in order to gain insights into the mechanisms of E-cadherin downregulation in tumors. In vivo footprinting analysis revealed that positive regulatory elements of the E-cadherin promoter (a GC-rich region, the CCAAT box, and a palindromic element) are specifically bound by transcription factors in E-cadherin-expressing but not in E-cadherin-nonexpressing cells [53]. Bussemakers et al. have isolated and characterized the human E-cadherin gene promoter and have studied the transcriptional regulation of this gene in two human PC cell lines, one expressing E-cadherin (PC3) and the other one not expressing E-cadherin (TSU-pr1). They showed that the E-cadherin promoter is not active in the nonexpressing cells and that this may be due to the binding of a repressor protein to the promoter [54]. Transcriptional regulation of steroid-responsive genes by steroid hormone transcription factors, including the AR, is complex and also dependent on accessory factors (i.e., coactivators or corepressors). In the present study, AR-mediated downregulation of E-cadherin could be due to: 1) AR-dependent direct transcriptional repression of E-cadherin promoter; 2) AR-mediated transcriptional regulation through the binding of a repressor protein (or repressor protein could itself be a part of AR transcriptional complex); or 3) AR-dependent transrepression of another gene presumably involved in transcriptional regulation of E-cadherin. Further studies are required to elucidate DHT-mediated reduction in E-cadherin levels in DAR17 and DAR19 AR transfectants by E-cadherin promoter analysis.

From the immunoprecipitation data, it can be seen that addition of DHT to the standard medium resulted in a decrease in the quantitative expression of E-cadherin associated with γ-catenin in DAR17 and with β- and γ-catenin in DAR19. This suggests that there is a disruption of the E-cadherin/catenin complex normally located at the cell surface. Catenins are now known to involved in signalling pathways including the Wnt-1 pathway and, generally, these lead to either proteosomal degradation of β- and γ-catenin [55,56] or cytoplasmic stabilization and subsequent translocation to the nucleus, where they participate in transactivation of Wnt-1-responsive genes such as C-MYC [57–59]. However, in DAR17 and DAR19, there is no reduction of expression of the catenins at the cell surface, implying that they are neither cytoplasmically stabilized nor degraded. It is possible that β- and γ-catenin interact with other molecules at the cell surface (e.g., MUC1 [60]), or are diverted to other adhesion complexes (e.g. desmosomes [61]), which would explain their continued presence at the cell surface.

Clinical studies and both in vivo and in vitro experimental works have clearly demonstrated that androgen withdrawal results in inhibited proliferation and increased apoptosis of androgen-sensitive prostatic cells [5,62,63]. Our results are paradoxical. Ligand stimulation of the DAR17 and DAR19 resulted in: 1) decreased cellular proliferation, and 2) increased apoptotic cell death. A slower cellular proliferation in response to DHT has been reported in three subclones (LNCaP-r and LNCaP-LNO and LNCaP-FGC) of the LNCaP prostatic adenocarcinoma cell line [64,65]. PC3 cells transfected with AR on exposure to physiological concentration of DHT also showed decreased cellular proliferation [66,67]. FACS analysis data show that there is an increased number of DAR17 and DAR19 in G₁ when treated with 5 nM DHT in comparison to when grown under standard conditions, and this could be due to an increase in cell cycle molecules that negatively control G₁-S transition such as p21^CIP1 and p27^KIP1 [68–70]. Kokontis et al. observed that at concentrations that repressed growth, androgen transiently induced the expression of the cyclin-dependent kinase (cdk) inhibitor p21waf1/cip1 in 104-R1 cells, whereas expression of the cdk inhibitor p27KIP1 was persistently induced by androgen in both 104-R1 and 104-R2 cells derived from LNCaP-FGC prostatic cancer cell line by prolonged androgen deprivation [69]. Induced expression of murine p27^KIP1 in 104-R2 cells resulted in G₁ arrest. Lapointe and Labrie [70] have found a dose-dependent DHT-induced inhibition of CAMA-1 breast cancer cells resulting in increase in p27^KIP1. Recently, Geck et al. [71] have demonstrated that androgen-induced proliferative quiescence in PC cells is related to expression of AS3 and its mediator, which belongs to a novel gene family and a homologue of fungal proteins involved in proliferation arrest.

Our results of androgen-induced apoptosis in AR transfectants are consistent with the observations by Heisler et al. [67], who reported that DHT treatment resulted in apoptotic cell death of PC3 derived AR transfectants. However, in comparison to their finding of apoptosis in all DHT-treated cells, we observed cell death in only a fraction of DHT-treated AR transfectants. This may be an adaptive response in the continuous presence of ligand in the culture medium. Dai et al. [47] have reported the desensitization of AR transcriptional activity in LNCaP cells after prolonged (96 hours) exposure to DHT. Androgen-induced apoptosis in AR transfectants may also be secondary to decrease in E-cadherin protein expression following DHT treatment. Previous studies have suggested that the loss or downregulation of E-cadherin is associated with subsequent signalling for apoptosis in cultured cells in vitro. Day et al. [72] have reported that E-cadherin mediates aggregation-dependent survival of prostate and mammary epithelial cells. They have shown increased E-cadherin expression in surviving cells in response to protein kinase C-induced apoptosis, and blocking with E-cadherin-specific antibodies leads to synergistic increase in apoptosis. Kantak and Kramer [73] have shown that HSC3 human squamous carcinoma cells die by apoptotic cell death when are they were grown in single cell suspension cultures. However, if these cells are allowed to form E-cadherin-mediated multicellular aggregates, they not only survive but proliferate. These anchorage-independent survival and growth of HSC3 cells required high levels of extracellular calcium and were inhibited with anti-E-cadherin blocking antibody. These observations indicate that E-cadherin-mediated intercellular adhesions generate a compensatory mechanism that promotes growth and suppresses apoptosis.

Our study highlights a new mechanism of androgen-mediated regulation of cell adhesion through E-cadherin in human PC cells in vitro. This may well be relevant to androgen-mediated human PC progression (invasion and metastasis) in vivo. Further studies exploring downstream signalling of AR and its relation to other pathways involved in regulation of cell-cell adhesion, cell growth, and cell death are needed to be undertaken to clarify the role and mechanisms of AR-mediated signalling in PC.

Abbreviations

AI

androgen independence

PC

prostate cancer

AR

androgen receptor

DHT

dihydrotestosterone

ICC

immunocytochemistry

IF

immunofluorescence

RT-PCR

reverse transcriptase polymerase chain reaction

ssFBS

serum-stripped FBS

4-OH-F

4-hydroxyflutamide

GAPDH

glyceraldehyde phosphate dehydrogenase