The molecular mechanisms underlying progression of prostate cancer to the hormone-independent state are poorly understood. Signal transducer and activator of transcription 5a and 5b (Stat5a/b) is critical for the viability of human prostate cancer cells. We have previously shown that Stat5a/b is constitutively active in high-grade human prostate cancer, but not in normal prostate epithelium. Furthermore, activation of Stat5a/b in primary human prostate cancer predicted early disease recurrence. We show here that transcription factor Stat5a/b is active in 95% of clinical hormone-refractory human prostate cancers. We show for the first time that Stat5a/b synergizes with androgen receptor (AR) in prostate cancer cells. Specifically, active Stat5a/b increases transcriptional activity of AR, and AR, in turn, increases transcriptional activity of Stat5a/b. Liganded AR and active Stat5a/b physically interact in prostate cancer cells and, importantly, enhance nuclear localization of each other. The work presented here provides the first evidence of synergy between AR and the prolactin signaling protein Stat5a/b in human prostate cancer cells. [Cancer Res 2008;68(1):236–48]
- androgen receptor
- prostate cancer
Identification of molecular changes that lead to androgen-independent growth of prostate cancer cells is critical for the development of better therapeutic interventions for primary and advanced prostate cancer. Hormone-refractory prostate cancer is characterized by continued expression of androgen receptor (AR) and androgen-regulated genes, suggesting that the AR signaling pathway remains active despite low levels of circulatory androgens ( 1). AR is a cytoplasmic protein which, upon ligand binding, undergoes conformational changes, which leads to translocation of the AR to the nucleus. In the nucleus, AR undergoes intramolecular interactions that regulate its subsequent interactions with coactivators and influence its activity ( 2). Ligand-bound activated AR binds to specific DNA sequences at promoter regions of androgen-responsive genes to initiate transcription ( 3).
Adaptation of prostate cancer cells to androgen deprivation is known to involve several different molecular mechanisms. First, the AR gene is amplified in ∼25% to 30% of patients who experience disease recurrence ( 4). Second, somatic mutations of the AR in hormone-refractory prostate cancer result in receptors with elevated affinity for steroid hormones other than androgens and for pharmaceutical antiandrogens ( 5). In addition, mutations of the AR affect coactivator and corepressor binding ( 6). Third, protein kinase signaling pathways may promote proliferation and survival of prostate cancer cells directly or through stimulation of the AR in hormone-refractory prostate cancer.
One such protein kinase signaling pathway that may promote androgen-independent growth of prostate cancer via interaction with AR is the Janus-activated kinase (Jak2)–signal transducer and activator of transcription 5 (Stat5) signaling cascade. Stat5 belongs to the seven-member Stat gene family of transcription factors ( 7). There are two highly homologous isoforms of Stat5, 94-kDa Stat5a and 92-kDa Stat5b, which are encoded by separate genes ( 7). Stat5a and Stat5b (hereafter called Stat5a/b) are latent cytoplasmic proteins that act as both cytoplasmic signaling proteins and nuclear transcription factors. Stat5a/b becomes active by phosphorylation of a specific tyrosine residue in the carboxy-terminal domain ( 7) by a tyrosine kinase typically of the Jak protein family ( 8, 9). After phosphorylation, Stat5a and Stat5b homodimerize or heterodimerize and translocate to the nucleus, where they bind to specific Stat5a/b response elements of target gene promoters ( 7).
Stat5a and Stat5b are the principal signaling proteins that mediate the effects of prolactin (Prl) in normal and malignant prostate cells ( 10– 12). Prl is a mitogen and survival factor for prostate cells ( 13– 15), and transgenic mice overexpressing the Prl gene develop massive hyperplasia of prostate ( 16– 18). Correspondingly, Prl-null mice have smaller prostates than their wild-type (WT) counterparts ( 19). Prl protein and its receptor are expressed in malignant prostate epithelium ( 12, 13, 20, 21), and Prl protein expression in human prostate cancer is associated with high histologic grade ( 12). In line with these findings, transcription factors Stat5a and Stat5b are highly critical for the viability of human prostate cancer cells ( 22). This was later confirmed in the TRAMP mouse prostate cancer model ( 23). Prostate acinar epithelium is defective in Stat5a knockout mice ( 10), and Stat5a/b is constitutively active in human prostate cancer ( 12, 22, 24). Similar to autocrine Prl, expression of active Stat5a/b associates with high histologic grade of human prostate cancer (n = 114, P < 0.0001; ref. 12), a finding that we later confirmed in a larger, independent set of human prostate cancer specimens (n = 357, P = 0.03; ref. 24). Importantly, we also showed that active Stat5a/b in primary prostate cancer predicts early disease recurrence ( 24). These findings led us to propose that Stat5a/b may promote androgen-independent growth of prostate cancer. In this study, we hypothesized that Stat5a/b interacts with AR in prostate cancer cells.
Here, we show that Stat5a/b is in the active state in the majority of hormone-refractory clinical human prostate cancers. We further show that the active Stat5a/b signaling pathway increases transcriptional activity of AR in prostate cancer cells. Ligand-bound AR, in turn, increases transcriptional activity of Stat5a/b. The functional synergism of Stat5a/b and AR in prostate cancer cells involves direct physical interaction between the two. Liganded AR enhances nuclear localization of Stat5a/b, and active Stat5 promotes nuclear translocation of AR. In summary, this work establishes the novel concept that Stat5a/b interacts with AR signaling in human prostate cancer cells.
Materials and Methods
Primary human prostate cancers. Prostate cancer tissue specimens were from a total of 357 patients who were treated for clinically localized prostate cancer by radical prostatectomy or transurethral resection ( 25). The use of the deidentified archival tissue specimens in research was approved by the Institutional Review Boards of Kaiser Permanente and Thomas Jefferson University. Medical records for the entire cohort had been abstracted once (1999–2001) to ensure uniform criteria for diagnosis, progression, and staging. Hormone therapy was given to 64 of the patients, and the average length of androgen deprivation was 1.56 months ( Table 1 ).
Recurrent human prostate cancers. All specimens were transurethral resections from local recurrences (patients, n = 198) and were obtained from Tampere University Hospital in Finland ( 26). The definition for local recurrence was clinical progression of prostate cancer (development of urethral obstruction after initial favorable response to the therapy). Androgen ablation therapy was given to 127 of the patients, whereas the rest (n = 71) had received no hormonal treatment ( Table 2 ).
Immunostaining of prostate cancer cells and paraffin-embedded tissue sections. Immunostaining was performed as described previously ( 11, 22, 27, 28). The primary antibodies used were the following: phosphorylated (Y694/Y699) Stat5a/b [monoclonal antibody (mAb); Advantex BioReagents; 0.6 μg/mL], total Stat5a/b (mAb; Santa Cruz Biotechnology; 2 μg/mL), and anti-AR (pAb; Santa Cruz Biotechnology; 1:400). Antigen-antibody complexes were detected using appropriate biotinylated goat secondary antibodies (Biogenex Laboratories) followed by streptavidin–horseradish peroxidase complex, using 3,3′-diaminobenzidine as chromogen and Mayer hematoxylin as counterstain.
Scoring of the levels of active Stat5a/b in primary and recurrent clinical human prostate cancers. Individual prostate tumor samples were scored (M.T.N., J.A., and H.L.) for active and nuclear Stat5a/b levels on a scale from 0 to 1, wherein 0 was undetectable and 1 represented positive immunostaining.
Luciferase reporter gene assays. PC-3 cells (0.5 × 105) in Figs. 1A–D and 2A,B and LNCaP prostate cancer cells (0.5 × 106) and T47D breast cancer cells (0.5 × 106) in Figs. 1E and 2C were transiently cotransfected using FuGENE6 (Roche) with 0.25 μg of each pStat5a or pStat5b, pPrlR, and/or pAR plasmid and either 0.5 μg of pGASx4T109-luciferase, pβ-casein–luciferase, pCIS-luciferase, pCyclin-D1–luciferase, pMMTV-luciferase, or pProbasin-luciferase, and 0.025 μg of pRL-TK (Renilla luciferase) as an internal control. The total amount of plasmid DNA per well was normalized by pMod-DNR to 1.275 μg per well. The cells were starved in serum-free medium for 20 h and stimulated with 10 nmol/L human Prl (hPrl) or vehicle in the presence or absence of 0.1, 1, or 10 nmol/L dihydrotestosterone (DHT) in the starvation medium for 16 h. The lysates were assayed for firefly and Renilla luciferase activities using the Dual-Luciferase reporter assay system (Promega). Three independent experiments were carried out in triplicate using at least two sets of plasmids that were prepared separately. The firefly luciferase activity was normalized to the Renilla luciferase activity of the same sample, and the mean was calculated from the parallels. From the mean values of each independent run, the overall mean and its SE were calculated.
In the experiments of Fig. 2D, LNCaP cells (0.5 × 106) were transfected using FuGENE6 (Roche) with a total of 0.525 μg of plasmid DNA (0.5 μg pMMTV-luciferase and 0.025 μg pRL-TK) as described above. After 4 h, the cells were transfected with a total of 200 pmol of control mismatch-antisense or STAT5a/b-antisense oligonucleotides at a final concentration of 100 nmol/L using jetPEI reagent (QBiogene, Inc.). Specifically, Stat5 antisense oligonucleotides (5′-GGG CCT GGT CCA TGT ACG TG-3′, a shared sequence within both human Stat5a and Stat5b transcripts; bp 2153–2173 in open reading frame) were synthesized using a phosphorothioate backbone with 2′-O-methoxyethyl modification of five terminal nucleotides (underlined) to increase their stability (ISIS 130826) as described before (ISIS Pharmaceuticals; ref. 29). Mismatch oligonucleotides for the same chemistry were synthesized as a mixture of all four bases. After 24 h, the medium was replaced with serum-free medium for 12 h. The cells were stimulated with 10 nmol/L hPrl and/or 1 nmol/L DHT for 16 h before they were harvested for the luciferase assay described above.
Solubilization of proteins, immunoprecipitation, and immunoblotting. For coimmunoprecipitations of Stat5a/b and AR ( Fig. 3A ), 1.25 × 107 PC-3 cells per immunoprecipitation were cotransfected with 2 μg of pPrlR, pAR, and pStat5a or pStat5b plasmids, serum-starved for 20 h, and then treated with 10 nmol/L hPrl and 1 nmol/L DHT for 16 h. Cells were lysed, and the protein concentrations of clarified tissue lysates were determined as described previously ( 12, 22). The cell lysates were immunoprecipitated for 3 h at 4°C with anti-Stat5a or anti-Stat5b pAb (both 1.2 μg/mL; Advantex Bioreagents), anti-AR pAb (1.2 μg/mL, C-19; Santa Cruz Biotechnologies), or normal rabbit serum (1.2 μg/mL; Advantex Bioreagents), followed by incubation for 60 min with protein A-Sepharose beads (Pharmacia Biotech). The filters were blotted with anti-Stat5a+b mAb (1:250; Transduction Laboratories, Inc.) or with mAb anti-AR (1:1,000; Biogenex). The immunoreaction was detected by horseradish peroxidase–conjugated secondary antibodies in conjunction with enhanced chemiluminescence substrate mixture (Amersham) and exposed to film.
For experiments shown in Fig. 3C, LNCaP cells (3 × 107 per immunoprecipitation) were treated with 1 nmol/L DHT in starvation medium for 16 h and then stimulated with hPrl (10 nmol/L) for 15 min. The cell lysates were immunoprecipitated and blotted as described above.
The concentrations of the primary antibodies used in the immunoblottings were the following: anti-AR mAb (1:1,000; Biogenex), anti-Stat5a pAb (Advantex BioReagents; 1:3,000), anti-Stat5b pAb (Advantex BioReagents; 1:3,000), anti-Stat5a+b mAb (1:250; Transduction Laboratories, Inc.), anti-PSA pAb (Dako; 1:1,000), anti-actin pAb (Sigma; 1:4,000), and anti–cyclin D1 (NeoMarkers; 1:200).
Coimmunoprecipitations of FLAG-tagged Stat5a and MYC-tagged AR. Full length Stat5a was amplified by PCR and subcloned to pCMV-3FLAG vector (Stratagene) with an EcoR1 and a SalI sites. The full-length AR was amplified by PCR and subcloned to pCMV-3MYC vector (Stratagene) with a BamHI and a XhoI sites. Both constructs were subjected to sequencing analysis.
Plasmid pCMV3Flag-Stat5, pCMV3MyC-AR, and pPrlR were cotransfected using FuGENE6 (Roche) into PC-3 cells seeded 24 h earlier (2 μg of each plasmid per 1 × 107 cells; Fig. 3B). The cells were starved for 20 h and then stimulated with DHT (1 nmol/L) and hPrl (10 nmol/L) in RPMI 1640 without serum for 16 h. The cell lysates were immunoprecipitated with 25 μL anti-FLAG M2 polyclonal affinity gel (2 μg/mL; Sigma), anti-MYC pAb (1 μg/mL; Upstate), or normal rabbit serum. The primary antibodies were used in the immunoblottings at the following concentrations: anti-FLAG pAb (1:1,000; Stratagene), anti-MYC mAb (1:1,000; Sigma), anti-Stat5a+b mAb (1:250; Transduction Laboratories), and anti-AR mAb (1:1,000; Biogenex) detected by horseradish peroxidase–conjugated secondary antibodies.
Electromobility shift assay. LNCaP cells were transfected with plasmids (0.25 μg of each) expressing AR (pAR), Stat5a (pStat5a), and Prl receptor (PrlR; pPrlR) using FuGENE6 as indicated in Fig. 4C . The total amount of plasmid DNA per well was normalized with pMod-DNR. After 18 h, the cells were serum starved for 20 h and then stimulated with 1 nmol/L DHT for 0, 10, or 30 min, or with 1 nmol/L DHT and 10 nmol/L hPrl for 0, 10, or 30 min as indicated. The cell pellets were lysed in 1.5 mL of buffer A ( 11) and centrifuged (4°C) at 2,300×g for 2 min, and the nuclear pellets were resuspended in 100 μL of high-salt buffer ( 11) by homogenizing the pellets in a glass homogenizer. After centrifugation at 16,000×g, the supernatants were dialyzed in a Slide-a-lyszer mini dialysis unit (Pierce) against 200 mL of dialysis buffer [20 mmol/L HEPES (pH 7.5), 20% glycerol, 100 mmol/L KCl, 0.2 mmol/L EDTA, 5 mmol/L MgCl2, 2 mmol/L DTT, 0.5 mmol/L 2-mercaptoethanol, and 0.5 mmol/L phenylmethylsulfonyl fluoride] for 30 min. Double-stranded AR response element 1 (ARE-1) oligonucleotides (AGCTTGTCTGGTACAGGGTGTTCTTTTTGTCGA) were end-labeled with 50 μCi of P32 P-ATP (5,000 Ci/mmol) and incubated (1 ng per reaction) with 10 μL of nuclear extracts in the final volume of 20 μL of the binding mixture [50 mmol/L Tris-HCl (pH 7.4), 25 mmol/L MgCl2, 500 mmol/L KCl, 5 mmol/L DTT, 50% glycerol] and 1 μL of 1 mg/mL poly d[(I:C)]. The samples were preincubated with anti-AR antibody (BD BioSciences) or normal rabbit serum as indicated. Polyacrylamide gels (5%) containing 5% glycerol and 0.25× Tris-borate/EDTA were prerun in 0.25× Tris-borate/EDTA buffer at 4 to 10°C for 1.5 h at 300 V. The gels were run at room temperature for 3 h at 250 V, dried, and exposed to X-ray films (X-Omat, Eastman Kodak Co.).
Generation of adenoviruses for gene delivery of wild-type (WT), dominant-negative (DN) Stat5a/b and WT AR. pcDNA-CMV–WT Stat5a/b, pcDNA-CMV–(DN) Stat5a/b, and pAR were cloned to adenoviral vector using BD Adeno-X Expression System 2 (BD Biosciences Clontech) according to the manufacturer's protocol. This specific cloning system was chosen because it uses Cre-loxP–mediated recombination, which reduces the likelihood of development of replication-competent adenovirus over time. The recombinant adenoviruses were purified, linearized by PacI digestion, and transfected to QBI-293A cells to produce infectious recombinant adenoviruses. Viral stocks were expanded in large-scale cultures, purified by double cesium chloride gradient centrifugation, and titered side-by-side by a standard plaque assay method in QBI-293A cells as per the manufacturer's instructions.
Double immunofluorescence cytochemistry of AR and Stat5. PC-3 cells were infected with AdWTStat5a, AdWTPrlR, and AdWTAR each at MOI 4. After 24 h, the cells were plated on collagen-coated culture dishes (MatTek) and serum starved the next day for 12 h before stimulation with 10 nmol/L hPrl for 30 min and/or 1 nmol/L DHT for 60 min. The cells were fixed with 4% PFA and permeabilized with 0.5% Triton X-100, and unspecific staining was blocked with 2% bovine serum albumin in PBS for 1 h. The fixed cells were incubated with anti-Stat5 pAb (Santa Cruz Biotechnology) and anti-AR mAb (Santa Cruz Biotechnology; both at 1:200) followed by goat anti-rabbit fluorescein IgG (Vector Laboratories) and horse anti-mouse Texas Red IgG (Vector Laboratories), respectively (1:150). The immunofluorescence staining was observed by a Zeiss LSM 510 laser scanning microscope with an Apochromat X63/1.4 oil immersion objective.
Statistics. The probabilities of having active Stat5a/b in primary prostate cancer for patients with or without adjuvant hormone were calculated, and independence between Stat5a/b and hormone therapy was tested using Pearson χ2 test.
Transcription factor Stat5a/b is active in hormone-refractory recurrent human prostate cancers and during androgen deprivation of primary prostate cancers. Based on our previous results showing that active Stat5a/b in primary human prostate cancer predicted early prostate cancer recurrence ( 24), we hypothesized that Stat5a/b is constitutively active during androgen deprivation of prostate cancer. To test this hypothesis, we determined Stat5a/b activation during androgen deprivation of both primary and recurrent prostate cancer. First, we assessed the frequency of Stat5a/b activation in primary human prostate cancers from 357 patients. Of these patients, 64 had received androgen deprivation adjuvant therapy for 1.5 months before radical prostatectomy ( Table 1). Activation of Stat5a/b was analyzed by immunohistochemical detection of phosphorylated Stat5a/b in paraffin-embedded tissue sections (Supplementary data) as shown previously ( 11, 12, 22, 24, 27, 28).
In primary prostate cancer, Stat5a/b was more likely to be active if the patient had been treated with androgen ablation before radical prostatectomy (P = 0.02) compared with patients who did not receive adjuvant androgen deprivation therapy ( Table 1). Specifically, Stat5a/b was almost thrice more likely to be active in prostate cancer of a patient who had been treated with androgen deprivation (Stat5 active, 73% versus Stat5 not active, 27%; P = 0.02; Table 1) compared with prostate cancers from patients who did not receive androgen deprivation therapy before radical prostatectomy (Stat5 active, 58% versus Stat5 not active, 42%; P > 0.05). There was no significant difference in the distribution of Gleason grades between the two groups. These results show that Stat5a/b is more frequently active in primary prostate cancer during androgen deprivation compared with primary prostate cancers in the presence of circulating androgens.
The observation that Stat5a/b is frequently active in androgen-deprived primary prostate cancers prompted us to ask whether Stat5a/b is also active in hormone-refractory recurrent prostate cancers. To address this question, we first analyzed Stat5a/b activation in tissue specimens obtained by transurethral resection from 198 patients with clinical prostate cancer recurrence (development of urethral obstruction). Significant activation of Stat5a/b was detected in 74% (146 of 198) of all recurrent human prostate cancer specimens ( Table 2). Of these patients, 127 had been treated with androgen deprivation before recurrence (see Materials and Methods). Stat5a/b was active in 95% of the 127 recurrent prostate cancer specimens treated with androgen ablation therapy ( Table 2). Collectively, these results showed that Stat5a/b is active in the majority of hormone-refractory recurrent prostate cancers.
AR increases transcriptional activity of Stat5 in prostate cancer cells. AR expression persists in prostate cancer cells regardless of prostate cancer progression to hormone refractory stage ( 30). Our finding showing that Stat5a/b is active in 95% of hormone refractory human prostate cancers led us to test the hypothesis that AR and Stat5a/b signaling pathways interact in prostate cancer cells. First, we tested whether AR affects transcriptional activity of Stat5a/b in prostate cancer cells. Stat5a/b activity was determined using a transiently transfected luciferase reporter gene driven by an artificial promoter containing four Stat5 response elements upstream of HSV thymidine kinase promoter (4xGASpT109; sodium taurocholate cotransporting protein GAS-like elements; ref. 31). Reconstruction of AR and Stat5a/b signaling in PC-3 human prostate cancer cells was chosen as the experimental model for these studies because PC-3 cells do not endogenously express AR or Stat5a/b and because PC-3 cells are of prostate origin, expressing prostate cell–specific coactivators/repressors. Therefore, PC-3 cells provide an unambiguous model for determining the interactions between WT AR and Stat5a/b in prostate cancer cells. In the first set of experiments, PC-3 cells were cotransfected with 4xGASpT109-luciferase, PrlR, Stat5a, or Stat5b, and/or AR, serum-starved for 20 h, and stimulated with hPrl (10 nmol/L) and/or DHT at concentrations of 0.1, 1, and 10 nmol/L for 16 h. Upon coexpression of active Stat5a/b with DHT-stimulated AR, Prl-induced transcriptional activity of Stat5a was increased by 7-fold to 9-fold and of Stat5b by 5-fold to 6-fold in prostate cancer cells ( Fig. 1A, lanes 7–9) compared with cells negative for AR ( Fig. 1A, lane 6).
Because activity of transcription factors and nuclear receptors is dependent on a specific promoter context, we extended the determination of Stat5a/b activity from the artificial promoter to genomic promoters regulated by Stat5a/b. We chose for the next set of experiments β-casein–luciferase, cyclin D1–luciferase, and CIS (cytokine inducible SH2-containing protein)–luciferase reporter genes. β-casein and cyclin D1 promoters both contain a Stat5-binding site flanked by a nonconsensus Stat5a/b site, and therefore, they constitute strong tetrameric Stat5-binding promoters ( 32, 33). The promoter of CIS contains four consensus Stat5a/b recognition sequences. In PC-3 cells coexpressing liganded AR with active Stat5a or Stat5b and PrlR, ligand-induced transcriptional activity of Stat5a/b in β-casein–luciferase assay was increased by 3-fold to 6-fold and in CIS-luciferase assay by 7-fold to 10-fold ( Fig. 1B and C, lanes 7–9) compared with cells negative for AR ( Fig. 1B and C, lane 6). When assayed with cyclin D1–luciferase, Prl-induced transcriptional activity of Stat5a/b was increased by 2-fold to 3-fold in cells expressing liganded AR in addition to Stat5a or Stat5b compared with cells negative for active AR ( Fig. 1D, compare lanes 7–9 to lane 6). To extend the findings to the protein level, we used cyclin D1 protein expression as a quantifiable indicator of Stat5 transcriptional activity in LNCaP cells transfected with PrlR, Stat5a/b, and/or AR ( Fig. 1D, bottom). In line with the results obtained from the Stat5-regulated gene assays, levels of cyclin D1 protein were increased if prostate cancer cells expressed both active Stat5a or Stat5b and liganded AR ( Fig. 1D, bottom, lane 5) compared with DHT-stimulated AR only (lane 3) or cells expressing Prl-activated Stat5a/b only (lane 4). In addition to PC-3 cells, synergy between active Stat5a/b and AR in induction of Stat5a/b driven transcription in β-casein promoter reporter gene assay was also evident in LNCaP cells cotransfected with AR, PrlR, and Stat5a or Stat5b ( Fig. 1E, compare lane 6 to lane 5 and lane 9 to lane 8). In contrast, no coaction between Stat5a/b and AR was observed in T47D human breast cancer cells transfected with the same plasmids. This result suggested that prostate cell–specific factors may be required for the transcriptional synergy between AR and Stat5a/b. Taken together, the data presented here indicate for the first time that liganded AR, even at low levels of androgens, increases transcriptional activity of Stat5a/b in prostate cancer cells.
Stat5a/b signaling increases transcriptional activity of liganded AR in prostate cancer cells. Our observations of transcriptional activity of Stat5a/b being regulated by AR suggested an interaction between AR and Stat5a/b in prostate cancer cells. This made us focus our investigation next on determining whether active Stat5a/b signaling affects transcriptional activity of AR in prostate cancer. We tested this in PC-3 cells using a transiently transfected luciferase reporter gene driven by androgen-regulated promoters mouse mammary tumor virus (MMTV; ref. 34) or a genomic probasin promoter ( 35). PC-3 cells cotransfected with MMTV-luciferase, PrlR, Stat5a or Stat5b, and/or AR were serum starved for 20 h and stimulated with hPrl (10 nmol/L) and/or DHT (1 nmol/L) in serum-free medium for 16 h. Compared with Stat5a/b-negative cells, coexpression of active Stat5a or Stat5b with PrlR and AR in prostate cancer cells led to an 8-fold increase in DHT-induced transcriptional activity of AR ( Fig. 2A, compare lanes 7 and 11 to lane 3) using MMTV-luciferase as a quantifiable indicator of AR transcriptional activity. Similarly, probasin-luciferase assay showed a 4-fold increase in ligand-induced transcriptional activity of AR in the cells coexpressing active Stat5a or Stat5b in addition to PrlR and AR versus AR only–expressing cells ( Fig. 2B, compare lanes 7 and 11 to lane 3). Synergy between transfected Stat5a/b and AR was observed in LNCaP prostate cancer cells as well, but not in T47D human breast cancer cells cotransfected with probasin-luciferase, PrlR, AR, and Stat5a or Stat5b ( Fig. 2C) suggesting prostate cell specificity for the cooperative interaction between Stat5a/b and AR on AR-regulated gene transcription.
Next, we tested synergy between endogenously expressed Stat5 and AR in human prostate cancer cells ( Fig. 2D). LNCaP cells, transfected with MMTV-luciferase, served as a model system in these experiments because LNCaP cells express both Stat5a/b and AR. After 24 h, the cells were serum starved for 12 h and stimulated with hPrl (10 nmol/L) and/or DHT (1 nmol/L) for 16 h before the MMTV-luciferase assay. Treatment of the cells with both DHT and Prl increased transcriptional activity of AR by 80% ( Fig. 2D, left, compare lane 3 to lane 2), suggesting a positive coaction between endogenously expressed Stat5a/b and AR in prostate cancer cells. To verify that Stat5a/b is critical for the enhancement of AR transcriptional activity by Prl in LNCaP cells, we inhibited Stat5a/b expression by Stat5a/b antisense oligonucleotides, which was shown by Western blotting of the whole cell lysates ( Fig. 2D, top right). Mismatch oligonucleotides were used as control, and the filters were reblotted with antiactin pAb to show equal loading of the gels. Importantly, inhibition of Stat5a/b expression by Stat5a/b antisense oligonucleotides blocked the increase of Prl-induced transcriptional activity of AR ( Fig. 2D, left, lanes 5 and 6).
To validate the significance of Stat5a/b signaling on transcriptional activity of AR at the protein expression level, we tested whether Stat5a/b contributes to PSA protein expression in prostate cancer cells. LNCaP cells were chosen as the experimental model because they are positive for Stat5a/b and AR and express PSA under the regulation of androgens. LNCaP cells cultured in the presence of DHT (1 nmol/L) were transiently transfected with Stat5a/b antisense oligonucleotides for 24 and 48 h. The level of PSA protein expression in LNCaP cells was decreased by ∼90% if Stat5a/b expression was inhibited by Stat5a/b antisense oligonucleotides at 48 h versus control (mismatch oligonucleotides; Fig. 2E). These results suggest that functional synergy between Stat5a/b and AR is reflected on AR-driven protein expression in LNCaP prostate cancer cells.
AR and Stat5a/b physically interact in prostate cancer cells. To identify the molecular mechanisms underlying the mutually stimulatory interaction between Stat5a/b and AR in prostate cancer cells, we first examined whether Stat5a/b and AR physically interact in human prostate cancer cells. PC-3 cells were cotransfected with PrlR, AR, and Stat5a or Stat5b, serum starved for 20 h, and then stimulated with hPrl (10 nmol/L) and DHT (1 nmol/L) for 16 h. AR was immunoprecipitated from the cell lysates and blotted with a mAb recognizing both Stat5a and Stat5b. Stat5a (94 kDa) and Stat5b (92 kDa) formed a complex with AR in prostate cancer cells ( Fig. 3A, i, lanes 1 and 2). Control immunoprecipitations with normal rabbit serum gave a band at 94 kDa, which was significantly weaker than the 94-kDa band yielded with Stat5a-specific antibody. In the converse experiments, immunoprecipitated Stat5a and Stat5b were blotted with a mAb recognizing AR yielding a band at 110 kDa. This suggested that liganded AR formed a complex with active Stat5a and Stat5b in prostate cancer cells ( Fig. 3A, ii, lanes 5 and 6). The weaker band under the 110-kDa AR in lane 5 ( Fig. 3A, ii) likely resulted from in vitro proteolytic cleavage of AR, which has been described in prostate cancer cells previously ( 36).
In a parallel set of experiments, we cloned Stat5a and AR into FLAG-tagged and MYC-tagged expression vectors, respectively. PC-3 cells were transfected with pAR-MYC and pStat5a-FLAG in addition to pPrlR as described in the Materials and Methods. After serum starvation for 20 h, the cells were treated with DHT (1 nmol/L) and hPrl (10 nmol/L) for 16 h. Complex formation between Stat5 and AR was shown by immunoprecipitation of 3FLAG-STAT5a with anti-FLAG M2 antibody coupled gel and by blotting with mAb recognizing c-MYC tag ( Fig. 3B, i, lane 1). The immunoprecipitation with normal rabbit serum was loaded in lane 2. The immunoprecipitation of 3FLAG-STAT5a and coimmunoprecipitation of 3MYC-AR were validated by stripping and reblotting the membrane with anti-STAT5a mAb ( Fig. 3B, ii, lane 1) and anti-AR mAb ( Fig. 3B, iii, lane 1), respectively. Conversely, immunoprecipitation of the cell lysates with anti-MYC pAb yielded a 118-kDa protein corresponding to the FLAG-tagged STAT5a, as shown by immunoblotting with anti-FLAG mAb ( Fig. 3B, iv, lane 3). The immunoprecipitation with normal rabbit serum was loaded in lane 4. The immunoprecipitation of 3Myc-AR and coimmunoprecipitation of 3FLAG-STAT5a was validated by stripping and reblotting the membrane with anti-AR mAb ( Fig. 3B, v, lane 3) and anti-STAT5a ( Fig. 3B, vi, lane 3) mAb, respectively.
In the third set of experiments ( Fig. 3C), we investigated whether endogenously expressed Stat5a/b and AR in human prostate cancer cells physically interact when activated. LNCaP cells were treated with DHT (1 nmol/L) in starvation medium for 16 h and with hPrl (10 nmol/L) for 15 min before harvesting the cells. Stat5a/b was immunoprecipitated using anti-Stat5a/b pAb with normal rabbit serum as control and blotted with a monoclonal antibody against AR. Immunoprecipitation of Stat5a/b pulled down a 110-kDa protein recognized by anti-AR antibody ( Fig. 3C), which was absent in the cell lysates immunoprecipitated with normal rabbit serum indicating physical interaction between endogenously expressed Stat5a/b and AR in LNCaP cells upon ligand stimulation. The results of these studies indicate that activated Stat5a/b and liganded AR physically interact in prostate cancer cells.
AR increases nuclear localization of Stat5a/b in human prostate cancer cells. Neither Stat5a/b or AR possess kinase activity and therefore are not likely to promote transcriptional activity of each other through increased protein phosphorylation. To search for alternative molecular mechanisms underlying synergy of AR and Stat5a/b in prostate cancer cells, we hypothesized that nuclear translocation of active Stat5a/b is altered by the presence of liganded AR in prostate cancer cells. To test this, PC-3 cells were cotransfected with PrlR and Stat5a with or without cotransfection of AR, serum-starved for 20 h, treated with DHT (1 nmol/L) for 16 h, and stimulated with hPrl (10 nmol/L) for 0, 10, or 60 min. In the absence of added hPrl, Stat5a was predominantly found in the cytoplasm of PC-3 cells transfected with PrlR and Stat5a ( Fig. 4A, i). As expected, addition of hPrl (10 nmol/L) led to nuclear translocation of Stat5a at 10 min with maximal immunostaining of Stat5a in the nucleus at 60 min ( Fig. 4A, ii and iii). However, if the cells coexpressed liganded AR in addition to Stat5a and PrlR, stimulation of the cells with Prl resulted in remarkably more intense nuclear immunostaining for Stat5a at 10 and 60 min ( Fig. 4A, v and vi) compared with the cells stimulated with Prl but expressing only Stat5a and PrlR ( Fig. 4A, ii and iii). Unexpectedly, there was also evidence of active Stat5 in the nuclei in the absence of hPrl stimulation if the cells coexpressed liganded AR ( Fig. 4A, iv). These results suggest that nuclear translocation of Stat5a/b is enhanced by liganded AR.
To further investigate the effect of AR on nuclear translocation of Stat5a/b, we setup double immunostaining of Stat5a/b and AR by indirect immunofluorescence in prostate cancer cells. Stat5a, AR, and PrlR were expressed in PC-3 cells using replication-deficient adenovirus as an expression vector, as previously shown ( 12, 22). After serum starvation for 12 h, the cells were stimulated with hPrl (10 nmol/L) for 30 min and/or DHT (1 nmol/L) for 60 min. As shown in Fig. 4B(i–iv), when PC-3 cells were not stimulated with hPrl or DHT, Stat5a (green) was located in the cytoplasm of the cells ( Fig. 4B, i). As expected, when PC-3 cells were stimulated with hPrl for 30 min, Stat5 was translocated to the nuclei of the cells ( Fig. 4B, ii). Importantly, if the cells were simultaneously stimulated with both hPrl and DHT, the nuclear immunostaining of Stat5a was the most intense ( Fig. 4B, iv). Intriguingly, stimulation of the cells with DHT induced nuclear translocation of Stat5a in the absence of hPrl stimulation of the cells ( Fig. 4B, iii). In summary, the results of these experiments shown in Fig. 4B(i–iv) support the novel concept that liganded AR enhances nuclear translocation of Stat5a/b in prostate cancer cells.
Active Stat5a/b increases nuclear localization of AR in human prostate cancer cells. In the next set of experiments, we asked whether accumulation of DHT-induced AR in the nuclear compartment is affected by active Stat5a/b in prostate cancer cells ( Fig. 4C). PC-3 cells were cotransfected with AR with or without transfection of PrlR and Stat5a, serum-starved for 20 h, treated with hPrl (10 nmol/L) for 16 h, and then stimulated with DHT (1 nmol/L) for 0, 10, or 60 min. In the absence of the ligand (DHT), AR was predominantly located in the cytoplasm of PC-3 cells transfected with AR only ( Fig. 4C, i). Weak nuclear AR immunostaining was detected at 10 min with maximal nuclear immunostaining detected at 60 min after stimulation with DHT as expected ( Fig. 4C, ii–iii). In cells expressing activated Stat5a and PrlR in addition to AR, strong punctate nuclear immunostaining of AR was already evident at 10 min after DHT stimulation of PC-3 cells ( Fig. 4C, v). At 60 min after DHT stimulation, immunostaining of AR in subnuclear foci ( 37) was significantly more intense in cells expressing active Stat5a and AR compared with cells expressing AR only ( Fig. 4C, vi). A weak nuclear immunostaining of unliganded AR was visible in cells expressing ligand-induced active Stat5 ( Fig. 4C, iv).
We next analyzed the AR immunofluorescence staining for AR as described in Fig. 4B. In the absence of DHT stimulation, AR (red) was located in the cytoplasm of the cells ( Fig. 4B, v). When PC-3 cells were stimulated with DHT, AR was translocated to the nuclei of PC-3 cells as expected ( Fig. 4B, vii). Importantly, the nuclear immunostaining of Stat5 and AR was the most intense if the cells were simultaneously stimulated with both hPrl and DHT ( Fig. 4B, viii). Stimulation of the cells with hPrl induced nuclear translocation of not only Stat5 ( Fig. 4A, ii) but also AR ( Fig. 4B, vi). In conclusion, the results of these experiments ( Fig. 4B and C) provide evidence that activated Stat5a/b enhances nuclear translocation of AR in prostate cancer cells.
Finally, to examine whether DNA binding of AR is affected by active Stat5a/b, we performed EMSA analysis of AR in LNCaP cells. LNCaP cells were transfected with AR alone or with plasmids encoding AR, PrlR, and Stat5a, and the total amount of plasmid DNA was normalized by addition of the empty vector. After transfection (18 h), the cells were serum-starved for 20 h, stimulated with DHT (1 nmol/L) and hPrl (10 nmol/L) for 0, 10, and 30 min, and the nuclear extracts were isolated ( Fig. 4D). The ARE-1 was used as a probe, and the ability of anti-AR antibody to supershift DNA-bound AR in LNCaP cells was verified ( Fig. 4D, lanes 7 and 8). Stimulation of LNCaP cells with DHT induced binding of AR to DNA at 10 and 30 min as expected ( Fig. 4D, lanes 2 and 3). When LNCaP cells were transfected with not only AR but also with Stat5a and PrlR and stimulated with both DHT and hPrl for 30 min, the binding of AR to DNA was increased by 2-fold to 3-fold as determined by EMSA ( Fig. 4D, compare lanes 2 and 3 to lanes 5 and 6). This result further supports the concept that active Stat5 enhances transcriptional activity of AR and that active Stat5 and AR increase nuclear localization of each other in human prostate cancer cells.
The median duration of response to hormonal ablation therapy of primary prostate cancer in patients is <3 years ( 30). Although the use of androgen blockade in the treatment of prostate cancer has expanded in the last decade, the molecular mechanisms underlying the development of androgen-independent disease are still largely unclear and no effective therapy for hormone-refractory prostate cancer presently exists. Importantly, we establish here cooperative interaction between AR and Stat5a/b in prostate cancer cells. We propose that Stat5a/b and AR are two survival signaling pathways that transcriptionally synergize with one another in human prostate cancer.
We have shown previously that Stat5a/b is critical for the viability of human prostate cancer cells ( 22) and that Stat5a/b is activated in prostate cancer but not in normal human prostate epithelium ( 22). Furthermore, activation of Stat5a/b in primary prostate cancer predicted early prostate cancer recurrence ( 12, 24). The key result of the work presented here is the cooperative interaction between Stat5a/b and AR signaling in prostate cancer. Specifically, we showed that active Stat5a/b increased transcriptional activity of AR in prostate cancer cells. We further showed that liganded AR, in turn, increased transcriptional activity of Stat5a/b in human prostate cancer cells. Our findings of synergic interaction between Stat5a/b and AR were consistent in a battery of natural Stat5-responsive and AR-responsive promoters, suggesting that Stat5a/b and AR interaction is independent of the promoter context. The likelihood that this synergy was due to the presence of AREs or Stat5-binding sites on the Stat5-regulated cyclin D1 and β-casein promoters or MMTV and probasin promoters, respectively, was ruled out by verification searches using the P-Match pattern matrix and transcription-binding site search ( 38). Moreover, experiments in LNCaP cells showed coaction between endogenously expressed Stat5a/b and AR as well. When Stat5a/b and AR were cotransfected to both LNCaP cells and T47D human breast cancer cells, no synergic interaction of Stat5a/b and AR was observed in T47D human breast cancer cells versus LNCaP cells. This suggests the transcriptional coaction between Stat5a/b and AR is specific for prostate cancer cells or it may be specific for androgen-regulated cells in general. The enhancement of transcriptional activity of liganded AR by Stat5a/b in reporter gene assays was equal to or stronger than what has previously been reported for promotion of AR by Stat3, Akt, or mitogen-activated protein kinase signaling pathways ( 39– 43). Importantly, the functional synergy between AR and Stat5a/b was reflected in protein expression levels of the AR target gene PSA and a Stat5a/b target gene cyclin D1, further suggesting biological significance of Stat5a/b interaction with AR in prostate cancer cells. It is possible that prostate cancers with activated Stat5a/b have higher PSA expression as a surrogate of active AR, which is an interesting question to address in a future study. Limitations of the work presented here include exclusive use of plasmid reporter gene assays. Therefore, future studies using gene expression arrays need to determine the extent of transcriptional synergy between Stat5 and AR in prostate cancer cells in vitro and in vivo.
The precise details underlying the synergy between active Stat5a/b and AR signaling in prostate cancer cells remain to be elucidated. We currently favor a model where functional synergy between Stat5a/b and AR may involve shared nuclear translocation mechanisms between the two. This hypothesis is supported by the findings of increased nuclear accumulation of ligand-induced AR in cells with an active Stat5 signaling pathway, whereas liganded AR enhanced nuclear accumulation of active Stat5a/b in PC-3 cells expressing PrlR, Stat5a, and AR. Importantly, double immunostaining of AR and Stat5 with indirect fluorescence further supported the concept of active Stat5 and liganded AR enhancing nuclear localization of each other. Our results further show that active Stat5a/b and liganded AR physically complexed in prostate cancer cells, which was shown by coimmunoprecipitation of Stat5a/b and AR. This complex was shown in PC-3 cells cotransfected with Stat5a/b and AR, as well as in LNCaP cells that endogenously express Stat5a/b and AR. Stat proteins are transported to the nucleus by karyopherins ( 44). Provided that importin-α and importin-β are involved in nuclear import of both Stats ( 45) and AR ( 37) and that our results indicated physical interaction between active Stat5a/b and AR in coimmunoprecipitation experiments, it is possible that Stat5-AR protein complex involves karyopherin proteins. Finally, coexpression of active Stat5a with liganded AR led to enhanced binding of ligand-induced AR to DNA in AR EMSA analysis. We propose that enhanced DNA-binding of AR is due to higher abundance of AR in the nuclei when the Stat5a/b signaling pathway is active simultaneously with AR signaling in prostate cancer cells. Alternatively, the stability of DNA binding of AR may be increased by the presence of active Stat5 in the nuclei. Future studies using Chip assays need to determine whether Stat5a/b and AR stay in a complex on AR or Stat5-regulated promoters in prostate cancer cells. The fluorescence immunostaining of Stat5 and AR revealed that liganded AR induced nuclear translocation of Stat5 in the absence of Prl stimulation. At the same time, Prl-activated Stat5 induced nuclear translocation of unliganded AR in prostate cancer cells. The function and importance of the nonphosphorylated Stat5a/b or unliganded AR brought into the nuclei by liganded AR or phosphorylated Stat5a/b, respectively, remain to be determined. Interestingly, direct DNA binding and transcriptional regulation of certain promoters by unphosphorylated Stat monomers have been shown previously ( 46). We have shown that Stat5a/b is constitutively active in human prostate cancer cells ( 12, 22). It is possible that this pool of constitutively phosphorylated Stat5 dimers in prostate cancer cells translocate to the nucleus complexed with liganded AR in the absence of exogenous Prl stimulation.
Stat5a/b is in the active state in 95% of hormone-refractory recurrent human prostate cancers. In addition, the likelihood of Stat5a/b being active in primary human prostate cancer was higher (P = 0.02) if the patient had been treated with androgen deprivation as an adjuvant therapy before radical prostatectomy. It is currently unclear whether increased frequency of Stat5a/b activation in primary prostate cancer followed by adjuvant hormone therapy reflects the effect of androgen ablation on prostate cancer cells or rather a selection of tumor cells with progressive features. The finding of a mutually cooperative interaction between active Stat5a/b and AR in prostate cancer cells may have relevance for growth of prostate cancer during androgen deprivation in the presence of low levels of circulating androgens. This is because the components required for the interaction of both these two prostate growth promoting signaling pathways are present in hormone-independent prostate cancers. Similarly to active Stat5a/b, AR expression persists in prostate cancer despite the progression to hormone-refractory state ( 30, 47). The effects of the synergy between Stat5a/b and AR signaling on the growth of prostate tumors in vivo will be crucial for the understanding of the full biological significance of Stat5-AR interaction. In addition, it will be important to determine whether Stat5a/b and AR transcriptionally synergize on AR-driven and Stat5a/b-driven transcription when AR is liganded by AR ligands that are present during clinical androgen ablation therapy, such as antiandrogens or adrenal androgens.
The molecular mechanisms underlying constitutive activation of Stat5a/b in hormone-refractory human prostate cancer are currently unclear. Such mechanisms may involve autocrine Prl ( 12, 13, 21). Our previous results showed that autocrine Prl expression in prostate cells was associated with high-grade prostate cancer ( 12). Moreover, we have also shown that Jak2 is the predominant kinase that activates Stat5a/b in prostate cancer cells ( 12), and activating mutations of Jak2 have been recently described in hematopoietic malignancies resulting in constitutive activation of Stat5 ( 48). Such Jak2 mutations may also occur in advanced prostate cancer. A third potential mechanism for constitutive activation of Stat5a/b in prostate cancer is amplification of Stat5a/b genes. This is a subject worthwhile of a study because Stat5a/b genes are located on chromosome 17 ( 49), which is frequently altered in prostate cancer ( 50).
In summary, the findings presented in this work show for the first time transcriptional synergy between AR and Stat5 signaling in prostate cancer. This finding is important because Stat5a/b and AR are transcription factors that inhibit apoptosis and promote growth of prostate cancer cells. Ongoing work aims to determine the extent of transcriptional synergy between Stat5a/b and AR on Stat5a/b-driven and AR-driven gene transcription in prostate cancer, as well as the molecular mechanisms underlying the coaction between Stat5a/b and AR. Future studies will also establish the effect of Stat5a/b-AR synergy on prostate tumor growth in vivo. Finally, identification of small molecule inhibitors for Prl-Jak2-Stat5 signaling pathway may be explored as a new strategy combined with androgen deprivation for therapy of hormone-refractory prostate cancer.
Grant support: American Cancer Society grant RSG-04-196-01-MGO, Department of Defense Prostate Cancer grants W81XWH-05-01-0062 and W81XWH-07-01-0411, and NIH National Cancer Institute grant 1RO1CA113580-01A1. This project was funded, in part, under a grant with the Pennsylvania Department of Health (the Department specifically disclaims responsibility for any analyses, interpretations, or conclusions). Shared resources of Kimmel Cancer Center are partially supported by NIH Cancer Center Support grant CA56036-08.
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. Hallgeir Rui for Ad-WT-PrlR construct, Dr. Mary Vore for the sodium taurocholate cotransporting protein–luciferase construct, Dr. Munkhtsegtseg Dorj for help in immunostainings, Jacqueline Lutz for critical reading of the manuscript, and Drs. Chenguang Wang and Richard Pestell for providing the cyclin D1 promoter–luciferase construct.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received August 3, 2007.
- Revision received September 14, 2007.
- Accepted October 30, 2007.
- ©2008 American Association for Cancer Research.