ART-27, an androgen receptor coactivator regulated in prostate development and cancer.

Androgen receptor trapped clone-27 (ART-27) is a newly described transcriptional coactivator that binds to the N terminus of the androgen receptor (AR). Given the vital importance of AR signaling in prostate growth and differentiation, we investigated the role of ART-27 in these processes. Immunohistochemical studies indicate that ART-27 protein is expressed in differentiated epithelial cells of adult human prostate and breast tissue. In prostate, ART-27 is abundant in AR-positive prostate luminal epithelial cells, in contrast to the stroma, where cells express AR but not ART-27. The use of a rat model of androgen depletion/reconstitution indicates that ART-27 expression is associated with the elaboration of differentiated prostate epithelial cells. Interestingly, regulated expression of ART-27 in the androgen-sensitive LNCaP prostate cancer cell line inhibits androgen-mediated cellular proliferation and enhances androgen-mediated transcription of the prostate-specific antigen (PSA) gene. Consistent with a growth suppressive function, we show that ART-27 expression levels are negligible in human prostate cancer. Importantly, examination of ART-27 protein expression in early fetal prostate development demonstrates that ART-27 is detected only when the developing prostate gland has proceeded from a solid mass of undifferentiated cells to a stage in which differentiated luminal epithelial cells are evident. Thus, ART-27 is an AR cofactor shown to be subject to both cell type and developmental regulation in humans. Overall, the results suggest that decreased levels of ART-27 protein in prostate cancer tissue may occur as a result of de-differentiation, and indicate that ART-27 is likely to regulate a subset of AR-responsive genes important to prostate growth suppression and differentiation.

Androgen steroid hormones direct the genetic program dictating prostate development and maturation in male develop-ment. They exert biological effects by binding to the androgen receptor (AR), 1 a member of the steroid receptor family of transcription factors. Functional mapping of the androgen receptor shows that several regions are required for transcriptional activation (1,2). These include a C-terminal domain, AF-2, as well as two regions in the N terminus, AF-1a and AF-1b. Recent evidence suggests that the AR cell-and promoter-specific transcriptional response is generated through interactions with regulatory proteins termed coactivators and corepressors with AF-1 and AF-2 (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12).
Androgen stimulation results in cell proliferation in both the developing prostate and the malignant prostate. Current therapy for metastatic prostate cancer includes injections of luteinizing hormone-releasing hormone (LHRH) analogues to pharmacologically lower testosterone levels (androgen ablation), treatment with anti-androgens such as flutamide or bicalutamide, to block testosterone binding to the AR, or a maximal androgen blockade in which both treatments are combined. While such therapies are effective in the short term, cancers treated in this fashion will inevitably progress in an androgenindependent fashion.
An important determinant of AR transactivation function resides in the N terminus (1,13,14). The AR N-terminal residues 142-485 have been shown to activate a minimal promoter construct in a cell-free transcription system and to selectively interact with the transcription factors TFIIF and the TATA-binding protein (TBP), suggesting direct contact with general transcription factors (15,16). A growing list of proteins has also been reported to interact with the AR N terminus (15,(17)(18)(19)(20)(21)(22). The clinical importance of investigating the biological role of coactivators that bind the AR N terminus is underscored by a recent report describing a patient diagnosed with complete androgen insensitivity, but exhibiting no mutation in AR. The study suggested that a coactivator interacting with the Nterminal region of AR, apparently essential for genital development, was lacking in this patient (23).
The work described in this article investigates the biological role of the newly described androgen receptor coactivator, androgen receptor trapped clone-27 (ART-27). ART-27 binds to the AR N terminus and facilitates receptor dependent transcriptional activation. The ART-27 clone was isolated from an androgen-stimulated LNCaP cell cDNA library in a two-hybrid screen designed to identify proteins that interacted with the AR-N terminus (24) and was previously cloned in an exontrapping experiment (25). The ART-27 gene encodes a protein of 157 amino acids with a calculated molecular weight of 18 kDa. Functional characterization of ART-27 indicates that it is a nuclear protein that interacts with N-terminal rat AR amino acids 153-336, containing AF-1a and a part of AF-1b (24). Biochemical analysis of ART-27 indicates that it interacts with AR in nuclear extracts from LNCaP cells in a ligand-independent manner. In addition, ART-27 increases the transcriptional potency of AR when overexpressed in a variety of cultured mammalian cells. Mechanistically, ART-27 is likely to function as part of a transcription complex since analysis of ART-27 behavior in density gradient sedimentation of HeLa nuclear cell extracts indicates that it co-sediments as part of a larger complex (24). Although the mechanism by which ART-27 functions as a transcriptional activator is unclear, it is likely that ART-27 acts as an adapter to assemble protein complexes at androgen-responsive promoters.
Steroid receptor coactivators and corepressors are likely important determinants of the response of a given hormone receptor in a given cellular context. As androgen stimulation results in differing cellular responses at various stages of prostate development and malignant transformation, we investigated the role of ART-27 in prostate epithelial cells.

EXPERIMENTAL PROCEDURES
Construction of Plasmids-The ART-27 GST fusion construct was made by isolating full-length ART-27 as a EcoR1/XhoI fragment from pEG202 and subcloning into EcoR1/XhoI sites in pGEX-4T-1 (Amersham Biosciences). FLAG-tagged ART-27 and hemagglutinin (HA)-tagged ART-27 were made by subcloning the ART-27 EcoR1/XhoI fragment into a pcDNA3 vector with an N-terminal HA epitope (pcDNA3-HA) or into a pcDNA3 vector with a C-terminal FLAG epitope (pcDNA3-FLAG).
Generation of Anti-ART-27 Polyclonal Antibodies-Generation of Cterminal anti-ART-27 antibodies has been previously described (24). C-terminal antibodies were affinity-purified by isolation of the IgG fraction on a protein A Sepharose column, followed by affinity purification using Affi-gel 15 resin (Bio-Rad) coupled to the ART-27 peptide immunogen. Antibody was eluted with 100 mM glycine, pH 2.7, and immediately neutralized. Antibodies against the ART-27-GST fusion protein were made by immunizing rabbits with the purified protein (Covance Research Products, Denver, PA).
Immunohistochemistry-All human samples were used with approval of the New York University School of Medicine Institutional Review Board. Tissues utilized were obtained from paraffin-embedded archived radical prostatectomy or prostate needle biopsy samples. Immunohistochemistry was performed using either affinity purified Cterminal ART-27 antibody or AR antibody (Santa Cruz Biotechnology AR polyclonal N-20). Tissue sections were dewaxed in xylene, rehydrated, and washed in phosphate-buffered saline (PBS), pH 7.4. For antigen retrieval, paraffin sections were heated in a microwave oven (900 watts) in citrate buffer, followed by treatment with 3% H 2 O 2 and blocking with 20% normal goat serum. Sections were then incubated with affinity-purified ART-27 (1:100) or AR antibody (1:200) followed by incubation with a biotinylated rabbit secondary antibody (Vector Labs). Prostate specimens from 9 human fetuses 9.5-24 weeks of gestational age were obtained following surgical abortion performed for reasons unrelated to this investigation. Approval for the collection of specimens was obtained by the New York University School of Medicine Institutional Review Board. Informed consent was obtained for all specimens. Age of specimens were estimated by heel-toe length and based on the last menstrual period. Prostates were formalin-fixed and oriented in paraffin blocks.
Rat Model of Androgen Depletion/Reconstitution-Adult Sprague-Dawley rats (228 -250 grams) were obtained from Taconic (Germantown, NY). Two control rats were sacrificed on day 0 and ventral prostates harvested. The remaining rats were anesthetized and castrated on day 0. Subsequently, two ventral prostates were harvested at 96 h. In addition, at the 96-h postcastration time point, the remaining rats were implanted with 25 mg, 21-day release testosterone pellets (Innovative Research of America, Sarasota, FL). Two ventral prostates were then harvested at 24 and 48 h following testosterone supplemen-tation. The weights of the prostates indicated that tissue volume substantially diminished upon androgen depletion.
Generation of tet-ART-27 Cells-Tetracycline-inducible cell lines expressing a FLAG epitope-tagged ART-27 gene were made in LNCaP cells. LNCaP cells were first transfected with the pTet-On vector using DOTAP (Roche Applied Science), and resistant colonies were selected in 500 g/ml Geneticin (Invitrogen). Clones were transferred to 24-well dishes coated with fibronectin (10 g/ml; Invitrogen), expanded and screened for Tet-dependent activation by measuring pRevTRE-luciferase reporter gene activity. A LNCaP clone displaying tight Tet-dependent activation was transfected with pRevTRE:ART-27 (C-FLAG) prior to selection of resistant colonies in 150 g/ml hygromycin B. Approximately 70 clones were expanded and grown in the presence and absence of doxycycline. Cell lysates were made and subjected to Western blot analysis to determine which cell lines showed doxycycline-regulated ART-27 FLAG expression. Anti-FLAG antibody (Sigma) was used for Western blot analysis.
Thymidine Incorporation-LNCaP cell clones were plated in 24-well plates in RPMI 1640 ϩ 10% fetal bovine serum. The following day (T ϭ 0), cells were washed with phosphate-buffered saline and medium changed to RPMI 1640 (without phenol red) ϩ 10% charcoal-stripped fetal bovine serum (cFBS) Ϯ 1 g/ml doxycycline and Ϯ R1881. Media was changed every 24 h during the duration of the experiment.
[H 3 ]thymidine (2 Ci) was added to fresh medium and incubated with cells for 1 h at 37°C. Cells were then washed, methanol-fixed, and solubilized prior to scintillation counting.
Reverse Transcription (RT) and PCR (RT-PCR)-Total RNA was isolated from ART-27 FLAG clone 1 cells using the RNeasy Midi Kit (Qiagen Inc., Valencia, CA) according to the manufacturer's instructions. RT-PCR was performed using the C. therm. Polymerase One-Step RT-PCR System (Roche Applied Science). Primer pair sequences are as follows: PSA Forward, 5Ј-TGCGCAAGTTCACCCTCA; PSA Reverse, 5Ј-CCCTCTCCTTACTTCATC (26); RPL19 Forward, 5Ј-ATGTATCAC-AGCCTGTACCTG; RPL19 Reverse, 5Ј-TTCTTGGTCTCTTCCTCCTTG (27). The expected sizes for PSA and RPL19 are 752 bp and 233 bp, respectively. PCR parameters were optimized for each primer pair to ensure PCR products were in the linear range. Reverse transcription was performed at 60°C for 30 min for PSA and RPL19. 75 ng of RNA were used for the amplification of PSA. Reverse transcription was followed by denaturation at 94°C, annealing at 51°C and extension at 72°C for 25 cycles. For RPL19, 50 ng of RNA were used for reverse transcription followed by denaturation at 94°C, annealing at 60°C, and extension at 72°C for 26 cycles.

ART-27 Is Expressed in Epithelial Cells of Human Prostate and Breast
Tissues-To evaluate the biological role of ART-27, polyclonal antibodies were developed so that the cell and tissue-specific expression pattern of ART-27 could be determined in vivo. Characterization of anti-ART-27 polyclonal antibodies is shown in Fig. 1. Cells were transfected with tagged ART-27, and lysates were either immunoprecipitated with antibody against ART-27 or the tag or used directly for Western blot analysis. Fig. 1A (top panel) shows that a polyclonal antibody made against the C-terminal peptide of ART-27 recognizes overexpressed FLAG-tagged ART-27 (lanes 1 and 5), overexpressed HA-tagged ART-27 (lanes 2 and 6) as well as endogenous ART-27 protein (lanes 1, 2, 3, 4, and 7). A second polyclonal antibody made against full-length ART-27-glutathione S-transferase (GST) protein similarly recognizes these proteins (Fig. 1A, bottom panel) and confirms the identity of ART-27 protein. To verify detection of endogenous ART-27 protein, an untagged variant of ART-27 was expressed and compared with expression of endogenous ART-27 (Fig. 1B). The results establish that endogenous ART-27 is the same size as the untagged ART-27 and that the antibody is capable of specifically recognizing overexpressed and endogenous ART-27. As expected, the immune, but not the preimmune antisera recognizes ART-27 protein (Fig. 1C).
To determine the cell type specificity of ART-27 expression compared with AR expression, ART-27 and AR antibodies were separately incubated with freshly cut paraffin sections of human prostate and breast tissue. Consecutive sections from each tissue were analyzed. Both ART-27 and AR are predominantly expressed in nuclei of differentiated epithelial cells of human prostate tissue. AR is almost exclusively nuclear, while ART-27 immunoreactivity appears to be present in both the cytoplasm and the nucleus (see higher magnification insets, Fig. 2, A and B). In breast tissue, once again, AR and ART-27 are both expressed in nuclei of epithelial cells. However, AR appears to be more abundant in the distal regions of the gland (observed as the smaller round gland indicated by the arrow in Fig. 2E) than in the more proximal part of the gland (the large central gland), while ART-27 protein appears to be abundant in both regions. Neither protein is ubiquitously expressed in epithelial cells of all tissue types, however, because there is little positive immunoreactivity in epithelial cells of the colon (not shown). As a control for antibody specificity, tissues were incubated in the presence of the primary antibody plus the immunizing peptide; these tissues show no immunoreactivity indicating that all of the observable staining is due to antibody recognition of the ART-27 C-terminal peptide. Additionally, incubation of the sections with secondary antibody only (Fig. 2, C and F) indicates that the secondary antibody does not contribute to the observed staining pattern. Therefore, the pattern of ART-27 protein expression is epithelial cell-specific.
Closer examination of immunoreactivity with the ART-27 antibody in human prostate tissue sections shows strong epithelial cell staining and little to none in the stroma of adult human prostate tissue (Fig. 3A). To further verify the cell type specificity of ART-27 protein expression, Western blot analysis was performed using lysates from explant cultures of primary human epithelial and stromal cells (28). Consistent with the immunohistochemistry results, ART-27 is highly expressed in epithelial cells and expressed at low levels, if at all, in stromal cells (Fig. 3B).
ART-27 Is Expressed in Rat Prostate Epithelial Cells Undergoing Androgen-dependent Proliferation-Since studies in human prostate tissue (above) indicated that ART-27 and AR are both expressed in prostate epithelial cells, a rat model of androgen depletion/reconstitution was used to examine the effects of androgen-mediated prostate growth and differentiation on ART-27 protein expression (29). Similar to the expression of ART-27 in human prostate tissue, immunohistochemistry performed on rat prostate tissue samples indicated that ART-27 protein is expressed in prostate luminal epithelial cells, but not in stromal cells (Fig. 4A). To androgen-deprive the animals, rats were castrated to cause withdrawal of testicular androgens and regression of the prostate gland. Androgens were then re-administered for 24 and 48 h, resulting in cellular proliferation and growth of the prostate. Prostate glands were dissected from the animals at the indicated time points, and lysates were prepared from each gland and used for Western blot analysis. The filters were incubated with antibodies against ART-27 as well as proliferating cell nuclear antigen (PCNA, a marker for cellular proliferation), clusterin (a protein that increases upon prostate involution and apoptosis, Refs. 30 -32), and MAPK (used as an internal loading control). Following castration, PCNA expression is abolished and clusterin expression is greatly increased, supporting the induction of apoptosis in the prostate epithelium (Fig. 4B, cas). Upon administration of exogenous androgens, PCNA levels are increased with rapid reduction of clusterin expression, thereby confirming the mitotic role of androgens in this model (A24 and A48). ART-27 protein is dramatically reduced following androgen withdrawal (cas), but is abundant when androgens are available (con, A24 and A48). Although it is not evident in this figure, longer exposure of the film from the blot incubated with ART-27 antibody does detect ART-27 in the samples from the castrated animals (cas). Therefore, in this model, the expression of ART-27 is associated with an abundance of differentiated prostate epithelial cells and is reduced along with diminishing numbers of epithelial cells under conditions of androgen deprivation. Subsequent experiments were designed to determine if ART-27 may play a role in androgen-dependent prostate growth and/or differentiation.  1 and 2 versus lanes 3 and 4) indicating that androgens do not regulate the transcription of ART-27.

Regulated Expression of ART-27 in the Androgen-sensitive LNCaP Cell Line Causes Decreased Cell Proliferation-To
Since subsequent experiments were designed to look at the effect of exogenous ART-27 on androgen mediated cell growth in LNCaP cells it was first determined if endogenous ART-27 protein levels are regulated by androgens, as might be suggested by the depletion of ART-27 protein levels in the castrated rat prostate cell lysates (Fig. 4B). For this purpose, ART-27 protein levels were determined in parental LNCaP cells in the presence and absence of R1881 (Fig. 5B). Western blot analysis of lysates made from androgen starved or androgen stimulated LNCaP prostate cancer cells is shown in Fig.  5B. The abundance of ART-27 protein in androgen-starved cells (Fig. 5B, lane 1) is identical to the levels of ART-27 in androgen-treated LNCaP cells (Fig. 5B, lane 2) suggesting that androgens do not affect the steady state levels of ART-27 protein.
Analysis of ART-27-FLAG protein expression in response to doxycycline is shown in Fig. 6A. As expected, treatment of an ART-27-FLAG clone 1 with increasing concentrations of doxycycline results in higher levels of ART-27 protein expression. As a preliminary step to the thymidine uptake experiments described below, it was also determined that endogenous levels of AR expression are not affected by treatment of the cell clones with doxycycline (Fig. 6B).
To determine the effect of ART-27 on androgen-mediated LNCaP cell proliferation, DNA synthesis was determined by incorporation of [ 3 H]thymidine into cellular DNA. Fig. 6 (C and D) shows analysis of ART-27-FLAG clone 1 (C) and clone 2 (D) in the presence and absence of doxycycline and R1881. Both clones demonstrate regulation of ART-27-FLAG in response to doxycycline. Fig. 6C shows a representative experiment in which [ 3 H]thymidine was added to the cells 1 h prior to sample analysis. As expected, there is very little incorporation of thymidine into DNA in the absence of androgens whether the cells are treated with doxycycline or not (Fig. 6C, 0 nM conditions). In the presence of R1881 (Fig. 6C, 0.05 nM)  midine incorporation compared with those that do not (0.05 nM ϪDox) at the 48-h and 72-h treatment with R1881. Analysis of an additional clone (ART-27-FLAG clone 2) shows a similar trend (Fig. 6D). Again the cells show thymidine incorporation in response to androgens (compare 0 nM ϪDox to 0.05 nM ϪDox). In this clone, expression of ART-27-FLAG in the presence of 0.05 nM R1881 (0.05 nM ϩDox) also reduces thymidine uptake compared with the cells grown in the absence of ART-27-FLAG (0.05 nM ϪDox). In fact, in clone 2, the presence of ART-27-FLAG thymidine uptake is reduced to levels seen in the 0 nM R1881 condition (compare 0.05 nM ϩDox to 0 nM ϩDox). As a control it was shown that doxycycline alone at the concentration utilized does not inhibit parental LNCaP cell thymidine uptake (not shown).
The decreased ability of ART-27-FLAG LNCaP cell clones to synthesize DNA in the presence of doxycycline induced ART-27 gene expression is corroborated by the fact that we were not able to make stable LNCaP cell lines that overexpressed ART-27 constitutively. In these experiments ART-27 was subcloned into the pIRESneo2 vector (Clontech) that utilizes an internal ribosome entry site (IRES) to express the protein of interest (ART-27) on the same transcript as a selectable marker (neomycin). While hundreds of LNCaP colonies grew upon selection when transfected with pIRESneo2 alone, no colonies were evident on plates transfected with ART-27-pIRESneo2, again suggesting that increased levels of ART-27 protein inhibit or prevent LNCaP cell growth.
To determine if ART-27-mediated LNCaP cell growth suppression is linked to the up-regulation of androgen-responsive genes involved in differentiated cell function, we examined the effect of PSA gene transcription in the presence and absence of doxycycline and R1881 by RT-PCR. PSA is an AR target gene and product of differentiated epithelial cells. Fig. 6E shows the DNA analysis of PCR products generated in RT-PCR reactions using primers for either the PSA gene or the ribosomal protein L19 (RPL19) gene included as a reaction and loading control. The doxycycline-inducible ART-27 FLAG clone 1 was grown in 0.05 nM R1881 and treated with either doxycycline or R1881 for 72 or 120 h as indicated. Cells were treated with 0.05 nM R1881 to be consistent with conditions under which ART-27 suppression of cell growth was observed (Fig. 6, C and D). Since R1881 is typically used at concentrations between 1-100 nM to elicit robust PSA transcription, it was anticipated that under conditions of 0.05 nM R1881, PSA transcript levels would be low. Fig.  6E shows that PSA is marginally detectable in all the samples and that the PSA gene product is increased in response to ART-27 overexpression (compare ϪDox ϩR1881 with ϩDox ϩR1881 at both the 72-and 120-h time points). We note that this experiment does not indicate the mechanism by which ART-27 enhances AR-mediated PSA gene transcription, and experiments are underway to determine if ART-27 plays a direct role. However, consistent with the hypothesis that ART-27 plays a role in AR-mediated cellular differentiation and growth suppression, transcription of PSA is enhanced un- ART-27 Expression Is Decreased in Prostate Cancer-As the above results suggested that ART-27 inhibits androgen-mediated prostate cell growth, we evaluated the expression of ART-27 protein in benign (non-cancerous) and malignant (cancerous) human prostate tissue through immunohistochemical staining of paraffin-embedded radical prostatectomy (n ϭ 22) specimens. Consecutive tissue sections stained with affinity purified ART-27 antibody or with an AR antibody show nuclear staining of normal prostate glandular epithelium (Fig. 7, A and  B) as previously noted. Additionally, insets in A and B show that neither ART-27 nor AR exhibit immunoreactivity in prostate basal cells (see vertical arrows). Immunohistochemical staining of sections containing regions of adenocarcinoma (glandular epithelial cancer) indicate that ART-27 staining in prostate cancer is decreased (indicated by the black box in Fig.  7, C and E) relative to that observed in adjacent benign prostatic glands (indicated by arrows in Fig. 7, C and E). AR staining of a consecutive section shows uniform staining in benign and malignant glands, thereby providing a positive control for the integrity of the tissue specimen (Fig. 7, D and F). A higher magnification comparison of AR and ART-27 expres-sion in an additional prostate cancer tissue specimen again confirms robust levels of AR and negligible levels of ART-27 (Fig. 8, A and B).
Decreased ART-27 staining was observed in all 22 cancer specimens examined. Decreased ART-27 signal could not be attributed to tissue damage (caused by overfixation, underfixation, cauterization, etc.) since only slides that showed positive immunoreactivity with anti-AR antibody were included in the study. The majority of samples tested were moderately differentiated (Gleason score 5-7), with a minority representing poorly differentiated disease (Gleason 8 -10). In comparing the samples tested, however, staining patterns of the malignant glands appear similar regardless of the degree of differentiation.
In evaluating sections containing high grade prostatic intraepithelial neoplasia (HG-PIN), a condition thought to be a precursor to prostate cancer, strong ART-27 staining was noted within the dysplastic (abnormal looking, but non-cancerous) epithelium (Fig. 8, C and D). Thus, reduction in ART-27 expression appears to be a relatively late event in prostate carcinogenesis. Based collectively upon immunohistochemical staining patterns in prostate cancer and effect on prostate cancer cell growth, we speculate that ART-27 may play a role in promoting androgen-dependent epithelial differentiation thereby reducing tumor progression or proliferation rate. Its potential role in carcinogenesis remains to be defined.
ART-27 in Human Fetal Development-The disappearance of ART-27 protein in prostate cancer progression and consequent de-differentiation suggests that ART-27 may play a role in luminal epithelial cell differentiation. To examine ART-27 protein expression during developmental differentiation, we assessed the pattern of ART-27 protein in the region of the urogenital sinus from which the prostate develops in human fetal tissue. Sections through the urogenital sinus region of a 15-week-old fetus show a cell free region in the center which is the lumen of the urethra (indicated by the U in Fig. 9, A and B). Adjacent to the lumen are columns of epithelial cells that are surrounded by mesenchymal or stromal tissue (Str). Epithelial outgrowths from the urethra grow and branch into the surrounding mesenchyme to form the prostate buds (PB) or tu-  A and B, insets). The horizontal arrow (B, inset) indicates an ARpositive staining stromal cell. A tissue region typical of prostate cancer containing small irregularly shaped glands is demarcated by the black box in C-F. While AR is present in the cancerous regions (D and F), ART-27 is not expressed in the cancerous glands (C and E). Both ART-27 and AR are expressed, however, in adjacent non-cancerous glands that are either normal or benign hyperplastic glands (indicated by arrows in C, E, and F). Images are shown at ϫ200 magnification.

ART-27 in Epithelial Cell Biology
bules. Incubation of a tissue section with ART-27 antibody shows ART-27 protein in the layer(s) of the epithelial cells adjacent to the lumen of the urethra, but not in stromal cells or in the prostatic buds at 15 weeks of development (Fig. 9A). Incubation with anti-AR antibody however, shows staining of the epithelial cells surrounding the lumen, as well as staining in stromal cells, and epithelial cells in the central region of the prostatic buds (Fig. 9B). Tissue from the urogenital region of a 21-week-old fetus was also stained with ART-27 antibody. Since the urethra is much larger at this point in development only a portion of the urethra is visible in the photograph (Fig.  9C). Again ART-27 is present in the luminal cells adjacent to the lumen of the urethra. At higher magnification, (Fig. 9D, top) a single cell layer of columnar shaped cells adjacent to the lumen stain strongly. This is in contrast to the pattern of AR protein expression. Incubation of a consecutive section with anti-AR antibody shows that AR is strongly expressed in stromal cells as well as in multiple layers of epithelial cells adjacent to the lumen of the urethra (Fig. 9D, bottom). The arrowheads indicate identical cells that are positive for the presence of both ART-27 and AR. These results suggest that in the urogenital sinus, ART-27 is present in differentiated cells with an apical surface adjacent to the lumen rather than in proliferative cells closer to the basal cell layer. This idea is supported by the fact that in tissue from the urogenital region of a 15week-old fetus (Fig. 9A) there is little expression of ART-27 in the prostatic buds. At 21 weeks however, there is expression in prostatic buds that have begun to form a lumen (see Fig. 9C, arrowheads). By 24 weeks most of the prostatic buds have developed a lumen and nuclear expression of ART-27 is now observable in most of the prostatic buds (not shown). Thus, expression of ART-27 strongly correlates with differentiation of prostate luminal epithelial cells in human fetal development. DISCUSSION Attempts to understand the role of AR in prostate cell growth or differentiation suggest that its effects are complex and dependent upon cellular context (33)(34)(35)(36). Androgen receptor coactivators such as ART-27, are likely to modulate AR activity and determine differentiation and proliferation programs of prostate cell gene transcription.
Concise transcriptional regulation is a prerequisite for the development of highly specialized cell types that comprise multicellular organisms. One mechanism underlying such specificity is the restricted expression pattern of components of the transcriptional machinery to specialized cell types. For example, TAF II 105, a component of the transcription factor TFIID, is restricted to the granulosa cells of the ovary follicle and regulates a subset of genes necessary for normal ovarian follicular development (37). The striking cell type specific pattern of ART-27 protein localization during fetal urogenital development suggests that ART-27 regulates a subset of AR-respon- sive genes in specialized cell types. For example, while AR is expressed in multiple cell layers of the urogenital sinus, ART-27 is expressed in a single cell layer adjacent to the lumen. In prostate, ART-27 is not present in prostatic buds at week 15, but becomes detectable at later stages that correlate with the differentiation of luminal epithelial cells and the appearance of a lumen. As the majority of proteins expressed by the differentiated prostate epithelium are secretory in nature, the presence of a lumen indicates cellular differentiation. Thus, ART-27 protein expression strongly correlates with prostate cell differentiation during organ development.
Within the mature prostate, ART-27 expression appears to remain unique to differentiated epithelial cells. Little to no expression is observed within the proliferative (basal cell) compartment of the glandular epithelium or the stroma. Within the luminal epithelium, variable intensity of ART-27 nuclear staining is often observed within cells uniformly staining for AR. This suggests a functional role for ART-27, perhaps dependent upon cell cycle, stage of cellular maturation, or secretory function.
As ART-27 is present in both benign and premalignant (HG-PIN) epithelium, but relatively absent in prostate cancer, the loss of ART-27 expression appears to be a relatively late event in carcinognesis. Reintroduction of ART-27 into a prostate cancer cell line by stable transfection results in markedly decreased androgen-stimulated proliferation. Further, complete cessation of growth appears to occur upon attempts to constitutively express ART-27 within the cell line. These data are consistent with the hypothesis that ART-27 functions in initiation or maintenance of an AR mediated program of prostate cell differentiation and are supported by evidence that increased expression of ART-27 results in up-regulation of PSA gene transcription (Fig. 6), a gene expressed by differentiated prostate luminal epithelial cells.
Androgen receptor coactivators and corepressors are likely to play a role in androgen dependent and androgen-independent prostate cancer. The most effective means of treating prostate cancer is by androgen ablation therapy, which causes prostate tissue along with prostate tumors, to regress. While suppressing the supply of androgens is beneficial in the short term, some prostate cancer cells inevitably begin to grow in a hormone independent fashion, and at this point treatment options are poor. Multiple mechanisms might explain hormone refractory disease including androgen receptor mutation and/or upregulation and altered recruitment of AR coactivators or corepressors to the AR (38,39). The idea that steroid receptor coactivators might be instrumental in cancer progression has been realized with the discovery that Tamoxifen, the estrogen antagonist widely used to treat breast cancer, acts by recruiting the nuclear receptor corepressor, NcoR and preventing co-activator binding to ER (40 -42). A recent report indicates that recurrent prostate cancers express higher than normal levels of the steroid hormone coactivators, transcriptional intermediary factor 2 (TIF2), and SRC-1 coactivator (steroid receptor coactivator-1) genes (43). In addition, in situ analysis of multiple androgen receptor coactivator expression levels showed that mRNA expression of the coactivators varied between normal and malignant prostate tissue samples (44). Effects on normal prostate cell growth have also been demonstrated in animal studies by targeted mutagenesis of the SRC-1 gene (45). However, these effects are relatively minor and therefore suggest that compensatory mechanisms exist and/or that additional co-regulators are required for androgen mediated cell growth.
Although the AR N terminus has been shown to be critical for AR transactivation, the expression of AF-1-binding proteins has not been evaluated during prostate cancer progression.
Our data indicate that ART-27 expression is altered during prostate cancer progression, and therefore may play an important role in this disease. It is currently unknown how ART-27 is regulated in normal prostate cells or in prostate cancer. The cell type specificity of ART-27 expression suggests that it is not directly regulated by androgens since ART-27 is expressed in a far more restricted pattern than AR. In addition, androgen treatment does not induce ART-27 mRNA or protein in LNCaP cells and there are no decipherable androgen response elements in the promoter of ART-27. 2 However, ART-27 protein expression may be indirectly regulated by localized secretion of cytokines or growth factors in response to androgen signaling.
A growing body of evidence suggests that co-regulators of steroid hormone receptors impart functional flexibility, thus allowing the receptors to respond differentially, contingent upon their biological context (46). Our results are consistent with the idea that the androgen receptor coactivator ART-27, is a cell type-restricted cofactor, and promotes androgen-dependent prostate epithelial cell differentiation, thereby inhibiting cellular proliferation.