Prostate gland development and growth requires both androgen action and epithelial-stromal communications. In fact, androgen signaling through the androgen receptor (AR) may be important in both stromal and epithelial cells of the prostate. Because interaction of AR with the coactivator, Hic-5/ARA55, results in enhanced androgen-induced transcription, we analyzed Hic-5/ARA55 expression in prostate tissue sections from normal human donors and prostate cancer patients. In each sample, Hic-5/ARA55 expression was confined to the stromal compartment of the prostate. Furthermore, a prostate stromal cell line, WPMY-1 cells, expresses Hic-5/ARA55, which is localized both at focal adhesion complexes and within the soluble cytoplasmic compartment. The ability of Hic-5/ARA55 to shuttle between the nuclear and cytoplasmic compartments was revealed on inhibition of nuclear export with leptomycin B. Small interfering RNA ablation experiments established endogenous Hic-5/ARA55 as a coactivator for both viral and endogenous cellular AR-regulated genes. Finally, the mechanism of Hic-5/ARA55 coactivator activity in WPMY-1 cells was revealed by chromatin immunoprecipitation analysis that showed its androgen-dependent recruitment to the promoter of the stromal androgen-responsive keratinocyte growth factor gene. These data provide the first demonstration of a stromal-specific AR coactivator that has an effect on an androgen-regulated growth factor that is essential for stromal/epithelial cell communication in the prostate. (Cancer Res 2006; 66(14): 7326-33)
Communication between the epithelial and stromal compartments of the prostate that is mediated by growth factors and cytokines is crucial for the maintenance of prostate growth and function (1, 2). Examples of such growth factors in the prostate include transforming growth factor (TGF)-β, TGF-α, epidermal growth factor, keratinocyte growth factor (KGF), fibroblast growth factor-2 (FGF-2), and insulin-like growth factor (3, 4). Most of these growth factors enhance both growth and development of the prostate, although TGF-β may negatively affect prostate growth (5).
Circulating androgens derived primarily from the testes also contribute to normal growth and development of the prostate gland (6). Testicular Leydig cells secrete testosterone, which is converted in the prostate to dihydrotestosterone by 5α-reductase (7). Androgens action in both stromal and epithelial compartments of the prostate is mediated through the androgen receptor (AR), a member of the nuclear receptor superfamily of transcription factors (8). In the epithelial compartment, androgens are required for both proliferative and secretory functions of the prostate (9). For example, androgens induce the expression of prostate-specific secretory proteins, including human prostate acid phosphatase, prostate-specific antigen, and prostate-specific kallikrein (10, 11). In the stromal compartment, AR regulates the expression of various growth factors, such as KGF (FGF-7), a member of the FGF family (12–15). KGF is exclusively expressed in the stromal compartment of the prostate, whereas its receptor, a spliced variant of the FGF receptor 2IIIb, is expressed solely in the epithelial compartment (16). These characteristics suggest that KGF is a paracrine mediator of signaling from the stroma to the epithelial compartment that can influence epithelial cell growth (17, 18).
On ligand binding, AR translocates to the nucleus where it binds to specific DNA sequences termed hormone response elements of target genes. Furthermore, ligand-bound AR undergoes a conformational change that permits the recruitment and binding of coactivator complexes, thereby stimulating transcription through direct interactions with the basal transcription machinery or by inducing histone protein modifications or chromatin remodeling (19, 20). Many coactivators, including members of the steroid receptor coactivator (SRC) family, stimulate AR-mediated transcription through histone modification, including histone acetylation or methylation (21). However, many other AR coactivators, such as four-and-a-half LIM domain (FHL2) as well as the AR activator family, influence AR transactivation using distinct mechanisms that remain largely undefined (22–24). FHL2 is an AR-specific coactivator whose expression in the prostate overlaps that of AR (22). Furthermore, ARA70, another AR-specific coactivator, may play a role in uncovering the agonist activity of certain antiandrogens to activate AR activity (25). Finally, inactivation of ARA55, also known as hydrogen peroxide (H2O2)–inducible clone-5 (Hic-5), resulted in reduced AR activity in prostate cancer cell lines (26).
Hic-5/ARA55 is a nuclear receptor coactivator that belongs to the group III LIM domain protein family and contains four separate LIM domains in its COOH terminus (27). LIM domains are cysteine-rich motifs that may provide critical surfaces for protein-protein interactions (28). LIM domain–containing proteins are localized within a variety of subcellular compartments. For example, Hic-5/ARA55 is associated with focal adhesion complexes as well as localized within the nucleus (29). Furthermore, nuclear accumulation of Hic-5/ARA55 occurs in response to oxidants, such as H2O2 (30). Once in the nucleus, Hic-5/ARA55 is associated with a variety of hormone-responsive promoters, including c-fos and p21, and may participate in the recruitment of histone acetyltransferase–containing coactivators, such as TIF-2, RAC3, CBP, and p300 (31).
In this study, we examined Hic-5/ARA55 expression and activity in the prostate. In both normal and tumor-derived prostates, Hic-5/ARA55 expression is confined to the stromal compartment. We therefore focused our analysis of Hic-5/ARA55 function in a prostate stromal, myofibroblast cell line, WPMY-1 cells. We show here that Hic-5/ARA55 functions as an AR coactivator in WPMY-1 cells and is necessary for effective androgen induction of the stromal paracrine factor, KGF.
Materials and Methods
Antibodies. Antibodies used in this study included anti-Hic-5/ARA55 (BD Transduction Laboratories, Los Angeles, CA), anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH; CSA-335, Stressgen, Victoria, British Columbia, Canada), anti-AR (N-20, Santa Cruz Biotechnology, Santa Cruz, CA), anti-actin (H-196, Santa Cruz Biotechnology), and anti–histone 3 (Upstate Biotechnology, Lake Placid, NY).
Cell culture and transient transfection. CV-1 and WPMY-1 cells were routinely maintained in DMEM and RPMI (phenol red–free; Life Technologies, Grand Island, NY), respectively, supplemented with 100 μg/mL penicillin/streptomycin and 5% fetal bovine serum (FBS) at 37°C under 5% CO2. The WPMY-1 cells were isolated from a normal human prostate specimen and immortalized using SV40 large T antigen (32). Cells were seeded on 12-well cell culture dishes at a density of 7.5 × 104 per well for 24 hours before transfection. Transfections using myc-tagged human Hic-5/ARA55 (huHic-5/pcDNA 3.1), hemagglutinin (HA)–tagged mouse Hic-5/ARA55 (moHic-5/pcDNA 3.1), an androgen-responsive mouse mammary tumor virus (MMTV)-luciferase reporter, pLC-Luc, and the Renilla-luciferase control were done using Opti-MEM (Life Technologies) and the LipofectAMINE transfection reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. After 3 hours, fresh medium was added to the cells and hormone treatments were initiated where relevant.
Cytoplasmic and nuclear extract preparation. WPMY-1 cells were cultured on 10 cm plates and treated with 100 nmol/L leptomysin B (LC Laboratory, Woburn, MA). After 3 hours, the medium was removed and cells were harvested in PBS by centrifugation at 3,000 rpm at 4°C in a microfuge. Resulting cell pellets were resuspended in hypotonic buffer [10 mmol/L HEPES (pH 7.9), 1.5 mmol/L MgCl2, 10 mmol/L KCl, 0.2 mmol/L phenylmethylsulfonyl fluoride (PMSF), protease inhibitors] and incubated on ice for 30 minutes. Cells were disrupted using a Dounce homogenizer and the cytoplasmic extract was collected after 15 minutes of centrifugation at 3,000 rpm at 4°C in a microfuge. The remaining nuclei were resuspended in low-salt buffer [20 mmol/L HEPES (pH 7.9), 25% glycerol, 1.5 mmol/L MgCl2, 20 mmol/L KCl, 0.2 mmol/L EDTA, 0.2 mmol/L PMSF, protease inhibitors] and an equal volume of high-salt buffer [20 mmol/L HEPES (pH 7.9), 25% glycerol, 1.5 mmol/L MgCl2, 1.2 mol/L KCl, 0.2 mmol/L EDTA, 0.2 mmol/L PMSF, protease inhibitors] for 30 minutes on ice. Nuclear extracts were collected after centrifugation at 13,000 rpm for 30 minutes at 4°C in a microfuge.
Western blot analysis. Cell lysates were collected in radioimmunoprecipitation assay buffer [10 mmol/L Tris (pH 8), 1 mmol/L EDTA, 0.5 mmol/L EGTA, 140 mmol/L NaCl, 1% Triton X-100, 0.1% deoxycholic acid, 0.1% SDS, 1 mmol/L PMSF, protease inhibitors] and boiled in sample buffer [62.5 mmol/L Tris (pH 6.8), 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.001% bromophenol blue] for 5 minutes. Proteins were then separated on 7.5% SDS-PAGE and transferred to polyvinylidene difluoride transfer membrane (Bio-Rad, Hercules, CA) in transfer buffer (20% methanol, 48 mmol/L Tris, 39 mmol/L glycine, 1.3 mmol/L SDS) at 15 V for 30 minutes. Membranes were then incubated in blocking buffer [5% dry milk in TBS (pH 7.4)] for 2 hours to overnight. Next, the membranes were incubated with Hic-5/ARA55-specific antibodies diluted 1:1,000 in blocking solution for 2 hours at room temperature. After extensive washing, the membranes were then probed with horseradish peroxidase–conjugated goat anti-mouse IgG antibodies (Santa Cruz Biotechnology) diluted in blocking solution for 1 hour. Finally, the membranes were washed and developed using the Renaissance Western Bolt Chemiluminescence Reagent (Perkin-Elmer Life Sciences, Boston, MA) according to the manufacturer's instructions. The membranes were stripped with Re-Blot Plus Strong Solution following the manufacturer's instructions (Chemicon International, Temecula, CA) and reprobed for GAPDH as a loading control. Where indicated, quantification of scanned images was done using the NIH Image software.
RNA interference. RNA interference (RNAi) plasmids were constructed following the manufacturer's guidelines (pSilencer hygro, Ambion, Austin, TX). The DNA template sequences for human Hic-5/ARA55 were RNAi-A 5′-GGAGGACCAGTCTGAAGAT-3′, RNAi-B 5′-GAAAAGACCCAGCCTCCCT-3′, RNAi-C 5′-GCATCACAGATGAAATCAT-3′, and RNAi-D 5′-GTGGATTGCACACAGACAA-3′. The oligonucleotides were annealed to form dsDNA and inserted into the p2.1-U6 plasmid after enzymatic digestion by HindIII and BamHI. The green fluorescent protein (GFP) template sequence was used as a positive control during the plasmid construction and a negative control for transfection analyses.
Immunofluorescence. WPMY-1 cells were cultured on glass coverslips treated with ethanol-vehicle or 100 nmol/L leptomysin B for 3 hours, fixed with 4% paraformaldehyde in PBS for 30 minutes at 4 C, washed with PBS, and permeabilized in PBS containing 1% bovine serum albumin and 0.5% Triton X-100 for 5 minutes at room temperature. The cells were incubated overnight at 4°C with anti-Hic-5/ARA55-specific antibodies diluted 1:1,000 in PBS. After extensive washing with PBS, the cells were incubated 2 hours at room temperature with anti-mouse IgG-Cy2-conjugated antibodies (Molecular Probes, Portland, OR) diluted 1:300 in PBS. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; 10 μg/mL; Molecular Probes, Carlsbad, CA) according to the manufacturer's instructions.
Immunohistochemistry. Paraffin-embedded prostate tissue sections were obtained from the Shadyside Hospital Tissue Bank (University of Pittsburgh institutional review board approval 0507135, expiration date 08/09/08). Tissue sections were deparaffinized and rehydrated and antigens were retrieved by boiling in 10 mmol/L citrate buffer (pH 5.5) for 10 minutes. The sections were incubated overnight at 4°C in a humid chamber with anti-Hic-5/ARA55 antibodies diluted 1:2,000 in PBS. Immunostaining was done using a Vecta stain kit (Vector Laboratories, Inc., Burlingame, CA) following the manufacturer's instructions. Normal mouse IgG was used as controls.
Luciferase assays. Luciferase activity in cell-free lysates was measured using a Victor (Perkin-Elmer, Wellesley, MA). Cells were washed with PBS and lysed in Reporter Lysis Buffer (Promega, Madison, WI) followed by a freeze and thaw incubation to ensure proper cell lysis. The lysate was incubated with luciferase assay reagent followed by a 10-second relative luciferase unit measurement. Firefly luciferase activity was normalized to Renilla-luciferase activity. All experiments were done three or more times.
Chromatin immunoprecipitation. WPMY-1 cells were grown to 80% to 90% confluence in medium supplemented with charcoal dextran-treated FBS for 24 hours. The cells were treated with ethanol-vehicle or 10 nmol/L mibilerone for 1 hour, cross-linked with 1% formaldehyde at room temperature for 30 minutes, lysed, and sonicated as described previously (33). Chromatin fragments were immunoprecipitated with specific antibodies overnight at 4°C. After immunoprecipitation, 30 μL protein A/G-Sepharose (Upstate Biotechnology) was added and the incubation was continued for 1 hour. After extensive washing, precipitates were eluted with reverse cross-linking buffer (1% SDS, 0.1 mol/L NaHCO3) at 65°C for 4 hours followed by a proteinase K treatment for 1 hour at 45°C. Eluted DNA was isolated using the QIAquick PCR purification kit (Qiagen, Valencia, CA). PCRs were done using the Platinum PCR Supermix (Invitrogen), 2 μL DNA, and 32 cycles of amplification. Primers were forward 5′-CCCTTTCCCCTTCTAACTGC-3′ and reverse 5′-ACCTTTGCTGACCTCATTGG-3′ for the KGF promoter and forward 5′-TGCCACTCAGTTCAACATCACA-3′ and reverse 5′-CACTTGCCCAGCAATAGGTT-3′ for the KGF gene. PCR products were resolved on a 12% polyacrylamide gel and visualized with ethidium bromide. Semiquantitation was done using densitometric analysis of the resolved gels using the Kodak Imaging System (Perkin-Elmer Life Sciences, Boston, MA). Data points were subtracted for background and normalized to the Input data.
Reverse transcription-PCR. Total RNA was isolated from WPMY-1 cells using the RNAqueous RNA isolation kit (Ambion) following the manufacturer's instructions. For reverse transcription-PCR (RT-PCR), 1 μg RNA was incubated with 100 μL reaction mix containing 25 mmol/L MgCl2, 25 mmol/L deoxynucleotide triphosphates (Perkin-Elmer, Wellesley, MA), 10× PCR II buffer (Life Technologies), 40 units/μL RNAsin RNase inhibitor (Promega), 45 μmol/L random hexamers (IDT, Coralville, IA), 200 units/μL Superscript reverse transcriptase (Life Technologies), and nuclease-free water (Ambion). Parallel reactions were done without reverse transcriptase to control for the presence of contaminant DNA. The samples were incubated at 25°C for 10 minutes, 48°C for 30 minutes, and 95°C for 5 minutes followed by 4°C for 5 minutes to inactivate the reverse transcriptase.
For amplification, a PCR containing a cDNA aliquot along with AmpliTaq Gold DNA polymerase in a volume of 25 μL was used according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). Primers used for gene expression analysis were KGF forward 5′-CAATCTAGAATTCACAGATAGGAGGAGGC-3′ and reverse 5′-ACAATTCCAACTGCCACTGTCCTGATTTC-3′ and GAPDH forward 5′-CATCACCATCTTCCAGGAGCGAGA-3′ and reverse 5′-GTCTTCTGGGTGGCAGTGATGG-3′. Thermocycling conditions involved an initial denaturation step at 95°C for 12 minutes followed by 28 to 30 cycles at 95°C for 30 seconds, 56°C for 30 seconds, and 72°C for 30 seconds. PCR products were resolved on a 12% polyacrylamide gel and visualized with ethidium bromide. Semiquantitation was done using densitometric analysis of the resolved gels using the Kodak Imaging System. All KGF data points were divided by GAPDH data points and then normalized to the control.
The predominant form of human Hic-5/ARA55 functions as an AR coactivator. Hic-5/ARA55 was initially cloned based on its induction by H2O2 in mouse osteoblast cells and subsequently identified in screens for AR and glucocorticoid receptor–interacting proteins (23, 27, 29). Minor sequence variations have been noted between human and mouse Hic-5/ARA55 cDNAs (Fig. 1). Furthermore, one particular sequence variant of human Hic-5/ARA55 (23) was shown to differ at six positions from the predominant human Hic-5/ARA55 cDNA sequence deposited in the human genome database (27). In fact, a dominant-negative form of human Hic-5/ARA55 was identified using the variant Hic-5/ARA55 as the source of screening, which contains a single point mutation that converts an alanine at position 413 to a threonine (26). However, as shown in Fig. 1, both the predominant human and mouse Hic-5/ARA55 cDNA sequences contain a threonine at position 413. Therefore, to confirm that the predominant human Hic-5/ARA55 (413 T) is an AR coactivator, we examined its effects on AR transactivation in transiently transfected CV-1 cells. This predominant human form of Hic-5/ARA55 has never been tested for its effects on AR. As shown in Fig. 2, both the mouse and the predominant human Hic-5/ARA55 function as AR coactivators. Thus, it is unlikely that a single substitution at position 413 of Hic-5/ARA55 confers a dominant-negative phenotype (23). Other amino acid differences in the variant human Hic-5/ARA55 protein (Fig. 1) must be required along with a threonine at position 413 to generate dominant-negative activity.
Regional expression of Hic-5/ARA55 in human prostate tissue. Given Hic-5/ARA55 effects on AR and its role in hormone-dependent growth of prostate cancer cell lines, several studies have examined its expression in prostate tissue. For example, Hic-5/ARA55 mRNA expression, as quantified using PCR, is reduced in hormone-refractory prostate cancer tissue compared with normal or benign prostatic hypertrophy tissue (34, 35). Because these tissue samples contained a heterogeneous population of cells, potential cell type–specific differences in Hic-5/ARA55 expression would not be revealed. Therefore, we analyzed prostate tissue from both normal human donors and prostate cancer patients for Hic-5/ARA55 expression using immunohistochemistry. As shown in Fig. 3B and D, Hic-5/ARA55 expression is confined to the stromal compartment of the prostate. The epithelial compartment was consistently negative for Hic-5/ARA55 expression. This profile of Hic-5/ARA55 expression in the prostate is consistent with that previously published using normal human prostates (36). In tumor-derived prostate tissue, Hic-5/ARA55 was also restricted to stromal cells and not detected in the prostate tumor cells (Fig. 3, bottom). In both cases, the Hic-5/ARA55 staining appeared diffuse and its precise subcellular compartmentalization was not apparent.
Nucleocytoplasmic shuttling of Hic-5/ARA55 in the WPMY-1 prostate stromal cell line. Hic-5/ARA55 coactivator function has been analyzed in a variety of epithelial-derived prostate cancer cell lines, including LNCaP, PC3, and DU145 cells (23). However, because its expression is confined to the stromal compartment of the prostate, prostate cancer cell lines may not provide a suitable model system with which to analyze Hic-5/ARA55 activity. Given the apparent stromal expression of Hic-5/ARA55, we considered it more appropriate to examine Hic-5/ARA55 effects on AR in stromal cells rather than prostate cancer–derived epithelial cells. We therefore employed WPMY-1 cells, a prostate myofibroblastic cell line that was derived from a human prostate primary stromal culture that was immortalized using the SV40 large T antigen (32). Hic-5/ARA55 expression in WPMY-1 cells was confirmed by Western blot analysis (Fig. 4A). Immunofluorescence analysis revealed that most Hic-5/ARA55 is localized within the cytoplasm and at focal adhesions (Fig. 4B,, top). Although the extent of Hic-5/ARA55 expression within the nucleus of WPMY-1 cells was not apparent from the immunofluorescence images, leptomycin B treatment of the cells resulted in robust nuclear accumulation of Hic-5/ARA55 (Fig. 4B,, bottom). Leptomycin B is an inhibitor of a major nuclear export receptor and has been used to reveal the nucleocytoplasmic shuttling properties of a variety of proteins, including Hic-5/ARA55 and the Hic-5/ARA55 family member, zyxin (30, 37). This result was corroborated by subcellular fractionations that showed both the low level of Hic-5/ARA55 present in nuclei of WPMY-1 cells and its enhanced retention within nuclei on leptomycin B treatment (Fig. 4C). The effects of leptomycin B on Hic-5/ARA55 localization in WPMY-1 cells establish that Hic-5/ARA55 is indeed capable of shuttling between the cytoplasmic and nuclear compartments and must therefore have some residence time within the nucleus. Furthermore, Hic-5/ARA55 staining was not depleted from focal adhesions on leptomycin B treatment (data not shown), indicating that shuttling of Hic-5/ARA55 most likely occurs between the cytoplasmic and nuclear compartments.
AR and Hic-5/ARA55 recruitment to the KGF promoter. Because Hic-5/ARA55 enhances AR-mediated gene expression in transient transfection assays, we set out to reveal whether Hic-5/ARA55 was localized to endogenous androgen-responsive promoters in WPMY-1 cells. KGF is a well-established marker of androgen induction in prostate stromal cells (13, 14). As shown in Fig. 5A and B, KGF mRNA expression is induced 2-fold in response to treatment with various androgens. In primary stromal cell cultures, KGF expression is similarly induced ∼2-fold by androgen treatment (14). Furthermore, androgen induction of KGF mRNA expression is blocked by the AR antagonist, hydroxyflutamide, indicating that androgen induction of KGF mRNA in WPMY-1 cells is indeed mediated by AR.
Because Hic-5/ARA55 has been found to directly interact with AR, we sought to determine if Hic-5/ARA55 is bound to the KGF promoter using chromatin immunoprecipitation assays and whether this binding is affected by androgen treatment. Detailed analysis of the human KGF promoter revealed that promoter proximal sequences located between −225 and +190 were necessary for basal KGF expression (38). Although a canonical hormone response element is not located within this region, there is a half-palindromic glucocorticoid response element located between positions −178 and −183 around which we designed our primers for chromatin immunoprecipitation analysis (38). After stimulation with androgens for 1 hour, both AR and Hic-5/ARA55 were localized to the KGF promoter but not to the coding region of the KGF gene (Fig. 5C). Given the high degree of background obtained in chromatin immunoprecipitation assays with nonimmune IgG, it is difficult to assess the extent of AR and/or Hic-5/ARA55 binding to the KGF promoter in the absence of androgen treatment. Nonetheless, the chromatin immunoprecipitation results from hormone-treated cells suggest that Hic-5/ARA55-mediated coactivation of AR activity in WPMY-1 cells may be mediated in part by their recruitment to endogenous androgen-responsive promoters.
Ablation of Hic-5/ARA55 expression results in decreased AR transactivation. Given the localization of Hic-5/ARA55 to the KGF promoter in response to androgen treatment, we ascertained whether Hic-5/ARA55 was necessary for androgen induction of KGF expression using a small interfering RNA (siRNA) approach to ablate Hic-5/ARA55 expression in WPMY-1 cells. WPMY-1 cells were analyzed for Hic-5/ARA55 expression by Western blot analysis following transfection with either a control GFP siRNA or Hic-5/ARA55 siRNA. Densitometric analysis revealed that there was ∼60% less Hic-5/ARA55 in cells transfected with the Hic-5/ARA55 siRNA oligonucleotide D compared with control (Fig. 6A). Although this was the maximal amount of Hic-5/ARA55 ablation that could be attained with Hic-5/ARA55 siRNA transfections, it was sufficient to reduce androgen induction of the transfected androgen-inducible MMTV promoter (Fig. 6B) and of endogenous KGF mRNA expression (Fig. 6C). Basal levels of KGF expression were reproducibly higher on Hic-5/ARA55 ablation, suggesting that Hic-5/ARA55 may negatively regulate basal KGF promoter activity. Finally, androgen-inducible KGF expression was recovered after the transfection of human Hic-5/ARA55 expression plasmid to cells whose Hic-5/ARA55 expression was silenced (Fig. 6E). Because the Hic-5/ARA55 siRNA construct targets the 3′ untranslated region of the gene, transfection of the human Hic-5/ARA55 cDNA construct was able to effectively rescue the silencing of Hic-5/ARA55. This indicates that modest changes in Hic-5/ARA55 expression affect KGF expression directly. Irrespective of these effects of Hic-5/ARA55 on basal KGF promoter activity, the siRNA ablation experiments establish a role for Hic-5/ARA55 in androgen regulation of both a viral androgen-responsive promoter and an endogenous growth factor gene.
Epithelial-stromal interactions are mediated by growth factors and cytokines in the prostate (reviewed in ref. 39). During prostate development, the urogenital sinus mesenchyme (UGM) stimulates epithelial differentiation, ductal branching, and proliferation (40). Conversely, signals derived from the UGM promote mesenchymal differentiation (9). Along with epithelial-stromal communication, androgen signaling through AR is crucial for prostate development (41). Reconstitution experiments with isolated stromal and epithelial cells done by Cunha and Lung illustrated the necessity for functional AR in the stromal compartment for the development of a functional prostate (12). In these studies, prostate tissue was derived from either normal or testicular feminized mice, which have a loss-of-function mutation in the AR gene. Whereas combining testicular feminized epithelia and normal stroma produced normal prostate organ formation in the presence of androgen, analogous tissue reconstitutions with normal epithelia and testicular feminized stroma did not support prostate organogenesis (12). These data suggest that AR controls the expression of a factor in the stromal compartment that is necessary for proper epithelial differentiation and thus prostate development.
As coactivators play an important role in the physiologic action of nuclear receptors in various target tissues, there are likely to be prostate stromal-specific coactivators that influence AR-regulated gene expression and thereby contribute to stromal/epithelial cell communication (42–44). Our results identify Hic-5/ARA55 as one such compartment-specific AR coactivator that has an effect on androgen regulation of a paracrine growth factor, KGF, in the stroma. The importance of KGF in prostate gland development has been revealed by various approaches. For example, KGF-neutralizing antibodies inhibit prostate epithelial growth as well as ductal branching (18). Furthermore, exogenous KGF was sufficient to produce epithelial growth as well as ductal branching of the prostate in the absence of androgen (18).
Analysis of Hic-5/ARA55 expression in intact animals had revealed its limited tissue distribution. For example, immunohistochemical analysis was used to reveal selective expression of Hic-5/ARA55 in smooth muscle and myoepithelial cells (45). Hic-5/ARA55 was not detected in epithelial cells of the tissues examined, including the stomach, colon, liver, skin, and mammary gland (45). Additionally, Hic-5/ARA55 is expressed in the stromal but not epithelial compartment of the prostate (36). Here, we confirmed the stromal-specific expression of Hic-5/ARA55 in normal prostate. Furthermore, Hic-5/ARA55 expression was also restricted to the stromal compartment in prostate tumor tissue.
The stromal-specific expression of Hic-5/ARA55 suggests that this AR coactivator is unlikely to contribute directly to the development or progression of prostate cancer. However, Hic-5/ARA55 may play a role in stromal/epithelial cell communication in the prostate because androgen induction of a stromal-selective paracrine factor (i.e., KGF) is lost on partial ablation of Hic-5/ARA55. The increase in basal KGF expression on Hic-5/ARA55 ablation suggests that this coactivator may interact with other transcription factors or corepressors to negatively regulate KGF expression in the absence of androgens (46). Thus, by serving opposing roles to positively or negatively regulate KGF gene transcription, Hic-5/ARA55 may be critical in sculpting tissue organization in the prostate, responding to various hormonal and paracrine cues that serve to maintain the appropriate balance between distinct stromal and epithelial cell layers. Furthermore, given the role of stromal-derived KGF on the growth, transformation, and invasive properties of prostate cancer cells, coactivators, such as Hic-5/ARA55, may influence tumor development and progression through its effect on either basal and/or androgen-induced expression of KGF in stroma during both androgen-dependent and androgen-independent phases (47, 48).
Detailed sequence analysis of the human KGF promoter revealed a half-palindromic steroid hormone response element between positions −178 and −183, implicating AR regulation of KGF transcription (38). Chromatin immunoprecipitation analysis of the KGF promoter revealed both AR and Hic-5/ARA55 recruitment to this region of the KGF promoter in response to androgens. Although most Hic-5/ARA55 is found at focal adhesions and the cytoplasm in WPMY-1 cells, Hic-5/ARA55 shuttles between the cytoplasm and the nucleus. Interestingly, Hic-5/ARA55 was localized to the KGF promoter in the absence of leptomycin B. This indicates that there may be a small pool of Hic-5/ARA55 resident in the nucleus that may not be apparent by immunofluorescence analysis but which nonetheless participates in transcriptional regulation of androgen-regulated promoters.
Amplified in breast 1/ACTR/SRC-3/RAC3 is a nuclear receptor coactivator whose expression is amplified in 10% of primary breast tumor biopsies (49). Many investigators have subsequently examined whether altered expression, activity, or mutation of other nuclear receptor coactivators is associated with tumors of different stages (49–51). Analysis of coactivator expression in prostate cancer has yielded various results. For example, Gregory et al. report increased SRC-1 and TIF-2 expression in prostate tumors, whereas Fujimoto et al. report similar SRC-1 expression in benign, intermediate, and high-grade cancers (52, 53). Interestingly, Hic-5/ARA55 expression was reduced in tumor compared with normal tissue (34), although it is difficult to ascertain whether this represents reduced expression of Hic-5/ARA55 within individual cells or alterations in the balance between stromal and epithelial cell content in tumors. A thorough analysis of the presence of Hic-5/ARA55 mutations in prostate cancer is also lacking. Although minor sequence variations have been noted between human and mouse Hic-5/ARA55 clones, we show here that the predominant form of human Hic-5/ARA55 is an AR coactivator. Furthermore, one variant human Hic-5/ARA55 that has been isolated (ref. 23; see Fig. 1B) most likely represents a rare form of this coactivator because it was not detected in at least 12 normal and cancerous prostate tissue samples.3
In endocrine target tissues, nuclear receptor activity seems critical for tumor development and/or progression (54). For example, elevated AR expression may be necessary for prostate cancer progression to an androgen-independent state (55). Furthermore, increased AR expression correlates with a higher probability of recurrence (56). In response to elevated expression of AR, mutations in AR, or altered expression of coactivators, the activity of nuclear receptors within individual cells may exhibit a heightened response to nonsteroidal signals. Our work suggests the existence of another mechanism for coactivator effects on prostate cancer that does not function in a cell autonomous context. AR action in cells that support the tumor, located in the stromal compartment, may also be affected by stromal-specific coactivators.
Grant support: NIH grant CA43037.
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 with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Mukta Webber (Michigan State University) for providing the WPMY-1 cells, Dr. Motoko Shibanuma (Showa University School of Pharmaceutical Sciences) for the human Hic-5/ARA55 expression plasmid, and Dr. Robert Getzenberg (Johns Hopkins University School of Medicine) for helpful discussions.