The RNA Helicase p68 Is a Novel Androgen Receptor Coactivator Involved in Splicing and Is Overexpressed in Prostate Cancer

Introduction

Prostate cancer (PCa) is the second leading cause of male cancer deaths with >670,000 men diagnosed worldwide each year. PCa incident rates for the United States, European Union, and the United Kingdom are currently 125, 106.2, and 107.3 cases per 100,000 men, respectively ( 1, 2). The onset and progression of PCa is driven by the transcriptional function of the androgen receptor (AR), a member of the nuclear hormone receptor family of transcriptional factors. Ablation of androgens, the major AR ligand, is a therapeutic modality in the early stages of the disease. However, PCa can progress to a hormone-refractory phenotype (HRPCa) due to mechanisms that are largely undefined. The majority of HRPCa still express AR and AR-regulated genes, indicating that an active AR-signaling cascade is maintained despite castrate levels of androgens ( 3). A suggested mechanism for maintenance of a functionally active AR is aberrant expression of AR cofactors leading to increased AR sensitivity or promiscuous activation of AR by other steroid hormones ( 4, 5). There are currently no curative therapeutic agents for HRPCa ( 6).

The p68 DEAD box RNA helicase (Ddx5) is a growth and developmentally regulated prototypic member of the DEAD box family of enzymes, and an established RNA helicase. p68 has been found to play a role in a wide range of cellular functions, including pre-mRNA, rRNA, and microRNA processing ( 7, 8), ribosome biogenesis, and cell proliferation ( 9), transcription (reviewed in refs. 10, 11) and transcriptional deactivation by promotion of mRNA transcript release in Drosophila ( 12). A role for p68 in transcription is underscored through its interaction with components of the transcription machinery and other transcription-related factors [RNA polymerase II (RNAPII), CBP, p300, HDAC1, and Smad3; ref. 10]. There is now considerable evidence indicating that p68 is a potent transcriptional coactivator of estrogen receptor α ( 13– 15); tumor suppressor p53 ( 16); MyoD ( 17), and β-catenin ( 18).

p68 has been shown to be overexpressed and polyubiquitylated in colorectal tumors ( 19), suggesting that posttranslational changes or modification of p68 expression may play a role in tumor development. In addition, modification of p68 by the small ubiquitin-like modifier SUMO-2 was found to modulate the activity of p68 as a transcriptional regulator, favoring an action as a repressor ( 20, 21). Furthermore, tyrosine phosphorylation on Y⁵⁹³ of p68 by c-Abl was shown to be associated with cellular transformation and epithelial-mesenchymal transition in colon cancer ( 22– 24)—although controversy exists as to whether this is through p68 nuclear translocation of β-catenin ( 25)—and tyrosine phosphorylation of p68 at this residue has also been shown to be important in platelet-derived growth factor (PDGF) induced cell proliferation through the up-regulation of cyclin D1 and c-Myc expression ( 24). Together, these findings indicate that the DEAD box RNA helicase p68 plays an important role in transcriptional regulation, and posttranslational modifications of the helicase may be important in tumor development. In this study, we identify p68 as a novel AR coactivator and explore the relationship between p68 with AR signaling in PCa and suggest a role for p68 in cancer progression.

Materials and Methods

Yeast two-hybrid assay. p68 was identified from a prostate library (Clontech) in a yeast two-hybrid screen using the AR bait clones pGBKT7-ARDS (containing the AR-DNA and steroid binding domains, but lacking the NH₂ terminal domain) and pGBT9-AGA (with a substituted AR-DNA binding domain for a GAL4 binding domain), as previously described ( 26).

Cell culture. All cells were grown at 37°C in 5% CO₂. LNCaP cells were cultured in RPMI 1640, COS-7, and HEK 293 cells in DMEM plus 1% l-glutamine (Invitrogen) supplemented with 10% FCS (Invitrogen) or 10% Dextran Charcoal Stripped FCS (Hyclone) to produce steroid-depleted media, as detailed in figure legends. Where indicated, cells were treated with 10 nmol/L R1881 (methyltrienolone).

Tissue microarray and immunohistochemistry. Immunohistochemistry was performed using a tissue microarray (TMA) of benign and malignant prostate biopsies derived from transrectal biopsy, transurethral resection, and radical prostatectomy as previously described ( 27). All materials were used in accordance with approval granted by the Northumberland, Tyne and Wear NHS Strategic Health Authority Research Ethics Committee (reference 2003/11; The Freeman Hospital). The average core loss was 28%, and the final study included 147 cancer biopsies and 44 benign biopsies. Antigen retrieval was achieved by immersion in 10 mmol/L citric acid buffer (pH 6.0), followed by microwaving for 15 min (at 1,000 W) in a pressure cooker. Sections were immunostained with PAb204 (mouse monoclonal antibody against p68, Upstate) on a DAKO autostainer using Vectastain ABC kits (Vector Labs), according to the manufacturer's protocol. Sections known to stain positively were included in each batch, and negative controls were prepared by replacing the primary antibody with TBS buffer. p68 expression was scored blindly by intensity of staining in each biopsy core as being absent (0), weak (+), moderate (++), or strong (+++). Sections were viewed on a Nikon Eclipse 600 microscope, and images were captured using a Nikon OXM1200 camera with Eclipse net software.

Plasmids, antibodies, and drugs. The plasmids p(ARE)₃ Luc, pCMV-β-gal, pcDNA₃-AR_WT ( 26), GFP-AR_WT ( 28), pcDNA₃-p68_WT, pcDNA₃-p72_WT ( 16, 29), pCMV-c-Abl ( 30), pGFP3-Sam68 ( 31), pcDNA₃-p68_GLT, and pSG5-p68_Y593F, described previously, were generated by site-directed mutagenesis of p68_WT and ligated into appropriate vectors. The YFP-p68 construct, CDM and CD44v5 minigenes, were kind gifts from Angus Lamond (University of Dundee), Bert O'Malley (Baylor College of Medicine), and Stefan Stamm (University of Kentucky), respectively.

The antibodies p68 goat polyclonal, AR rabbit polyclonal C-19, and prostate-specific antigen (PSA) goat polyclonal (Santa Cruz); p68 mouse monoclonal (Upstate); AR mouse monoclonal (BD Pharmingen); TATA binding protein mouse monoclonal (TBP-Abcam); α-tubulin mouse monoclonal (Sigma); c-Abl mouse monoclonal (Chemicon); and P-Tyr-100 and myc mouse monoclonal (Cell Signaling) were used as stated in figure legends.

The tyrosine kinase inhibitor imatinib was obtained from the Pharmacy Department of the Royal Victoria Infirmary Hospital.

Immunocytochemistry. Immunocytochemistry was performed on HEK 293 cells transfected with 1 μg of GFP-AR and/or YFP-p68 construct using Superfect (Qiagen), as described previously ( 28). Transfected cells were placed in steroid depleted media for 24 h and then treated with 10 nmol/L R1881 for a further 16 h before fixing them in methanol. Slides were visualized using laser scanning confocal microscopy (Leica).

Luciferase reporter and minigene splicing assays. Luciferase reporter assays were performed on COS-7 cells cotransfected in triplicate with 0.1 μg of p(ARE)₃Luc, 0.1 g of pcDNA₃-AR_WT, 0.1 μg of pCMV-β-gal, and p68^WT/Y593F, p72, and c-Abl pcDNA fusion constructs using Superfect (Qiagen), as described previously ( 30). The range of plasmid levels indicated in figures (+, ++, +++, and ++++), corresponds to 25, 50, 75 and 100 ng, respectively. Luciferase activity was corrected for the corresponding β-gal activity to give a relative activity.

Minigene splicing assays were performed in HEK 293 cells in six-well plates, as previously described ( 31, 32). Transfected cells were placed in steroid depleted media for 24 h and treated with 10 nmol/L R1881, as detailed in figure legends. RNA was extracted using Trizol (Invitrogen) and reverse transcription–PCR (RT-PCR) performed using the One-Step RT-PCR kit (Qiagen). Densitometric band quantification was performed as previously described ( 33). The identities and sizes of bands were verified by sequencing.

All reporter and minigene assays shown are the mean of at least three independent experiments ± SE.

Cell lysate extraction and coimmunoprecipitation. LNCaP cells were cultured in steroid depleted media for 48 h before 10 nmol/L R1881 treatment for 8 h. Cells were nuclear extracted using CelLytic NuCLEAR extraction kit with an extraction buffer salt concentration of 330 mmol/L in the presence of PhosStop phosphatase inhibitors (Roche) and RNase A (100 μg/mL; Sigma). AR and p68 coimmunoprecipitation experiments were performed as described previously ( 34).

RNA interference and quantitative real-time PCR. LNCaP cells seeded in six-well plates were transfected with p68 or nonsilencing (NS) small interfering RNA (siRNA) using RNAifect (Qiagen), according to manufacturer's instructions. The p68 and NS siRNA oligonucleotides have been described previously ( 16, 35). Transfected cells and nontreated controls (C) were grown in steroid-depleted media for 24 h and then treated with 10 nmol/L R1881 for a further 16 h. Cells were harvested in either SDS sample buffer (protein) or TRIzol reagent (Invitrogen, followed by RT-PCR) to study mRNA expression by quantitative real-time PCR (QPCR) relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression as previously described ( 35). Oligonucleotides were used to determine p68 (CACATCAATCATCAGCCATTCCT and CAAACAAATAGGCCCATCGC) and AR (GGACTTGTGCATGCGGTACTCA and CCTGGCTTCGGCAACTTACAC) mRNA expression. The PSA and GAPDH oligonucleotides have been described previously ( 34).

Chromatin-immunoprecipitation assays. Chromatin immunoprecipitation (ChIP) assays were performed in LNCaP cells, as described previously ( 35), in the presence of RNase A (100 μg/mL). p68 and AR bound chromatin was immunoprecipitated using 2 μg of p68 goat or AR rabbit polyclonal C19 antibody, respectively. QPCR was performed on inputs and recovered material to assess the co-occupancy of AR and p68 to the proximal ARE I within the PSA promoter and distal ARE III within the PSA enhancer, as described previously ( 34). The results of all ChIP assays shown are the mean of at least three independent experiments ± SE.

Statistical analysis. The independent sample t test was used to compare differences in the ratio of exon skipping/inclusion products between groups in the minigene assays. For immunohistochemistry, differences in the frequency and expression of p68 between groups were examined using Fisher's exact test and Mann-Whitney U test, respectively. Correlations with Gleason grade (GG) were undertaken using the Kruskal-Wallis test. Patient survival was analyzed using the Kaplan-Meier method with log-rank testing, and multivariate analysis was performed using the Cox regression model. All tests were undertaken using SPSS version 11.0 computer software (SPSS, Inc.). All tests were two-sided and a P value of <0.05 was taken to indicate statistical significance.

Results

p68 is a novel nuclear AR-interacting protein in PCa cells. The conserved helicase core (amino acids 203–353) of the p68 DEAD box RNA helicase was identified as a positive clone during a yeast two-hybrid screen for novel AR-interacting partners. Subsequent immunoblot analysis showed nuclear localization of p68 in three different cell lines ( Fig. 1A ), that was unaffected by androgen treatment in the PCa LNCaP cell line (Supplementary Fig. S1). Accordingly, nuclear colocalization of cotransfected p68_WT-YFP and AR_WT-GFP tagged constructs in HEK 293 cells grown in the presence of androgen was observed by confocal microscopy ( Fig. 1B). Coimmunoprecipitation experiments with nuclear extracts of LNCaP cells further showed that the endogenous p68 and AR interaction was enhanced in the presence of androgens ( Fig. 1C). However, it is notable that the AR gene is itself androgen responsive and so the enhanced interaction is most likely due to an increase in AR levels and not due to a higher interaction affinity in the presence of androgens. No difference in the intensity of the AR protein band was seen between ±RNase-treated samples (in the presence of androgens), suggesting that the AR-p68 interaction was not RNA-dependent. Together, these data provide evidence that p68 interacts with the AR in the nucleus of prostate cells and could thus function as a potential AR coregulator.

Role of p68 as an AR coactivator. The LXXLL motif is observed in cofactors that interact with ligand-activated nuclear hormone receptors ( 36). We found that p68 contained an LXXLL motif at amino acid 146, which was a strong indicator that p68 would act as a transcriptional coregulator of the AR. To further explore this, we undertook ChIP assays to confirm p68 and AR cooccupancy at androgen-responsive elements (ARE) within the promoter and enhancer regions of the androgen-responsive PSA gene. Both p68 and AR were dynamically recruited to the ARE I and ARE III regions of the PSA promoter and enhancer, respectively ( Fig. 2A and B ), with maximal observed recruitment to both regions at 80 min after androgen stimulation. These results clearly show the presence of p68 at the active promoter of an AR-regulated gene, suggesting a putative role for p68 in AR-dependent transcriptional initiation.

To further examine whether p68 coregulates AR activity, luciferase reporter assays driven by androgen-responsive fragments of the PSA promoter were used to monitor AR transcriptional activity. AR-mediated reporter activity was robustly stimulated by the addition of androgens, and in the presence of p68; we observed a dose-dependent coactivation of AR-mediated activity above AR-alone levels ( Fig. 2C). p68 did not directly affect luciferase reporter activity in the absence of AR (Supplementary Fig. S2), illustrating that p68 was not acting via AR-independent pathways.

ATP-dependent helicase enzymes, such as p68, use the binding of ATP to increase the affinity and specificity of binding to RNA. Using ATPase mutants, we sought to investigate whether the ATP binding/hydrolysis activity of p68 had a functional role in AR coactivation. An ATPase A (p68_GLT) mutant maintained AR coactivation above levels of AR alone ( Fig. 2D), similar to findings with a mutant directed to the ATPase B (p68_NEAD) domain of p68 (data not shown), confirming that neither a conformational change upon p68 ATP binding nor helicase activity was required for p68 coactivation of the AR. This supports previous findings with other p68 interacting partners that showed ATPase/RNA helicase activity was not required for transcriptional function ( 13, 16, 17).

To determine whether post-translational modifications were necessary for AR coactivation, we used a p68_Y593F mutant that was not tyrosine-phosphorylated at Y⁵⁹³ in luciferase reporter assays. We found that the p68_Y593F mutant was unable to increase the activation of AR over AR-alone levels compared with p68_WT, indicating that the phosphorylation of p68 at Y⁵⁹³ was important for coactivation of the AR ( Fig. 3A ). In addition, coimmunoprecipitation experiments showed that p68 was tyrosine phosphorylated and associated with c-Abl in the PCa LNCaP cell line ( Fig. 3B), although LNCaP cells transiently transfected with the p68_Y593F mutant also showed an interaction with c-Abl, indicating that sites other than Y⁵⁹³ may be phosphorylated by c-Abl. Further luciferase reporter assays showed that increasing amounts of c-Abl in the presence of a low concentration of p68_WT enhanced coactivation of AR over p68_WT alone ( Fig. 3C). c-Abl had no enhanced effect on luciferase reporter activity in the presence of the p68_Y593F mutant or in the absence of AR and/or p68_WT, indicating that c-Abl was neither enhancing the effect of AR directly nor acting via AR-independent pathways. Furthermore, luciferase reporter assays showed a decrease in coactivation of the AR by p68_WT in the presence of the tyrosine kinase inhibitor imatinib (10 μmol/L) compared with untreated cells ( Fig. 3D). Collectively, these data show that p68 is a coactivator of the AR, independent of the helicase function of p68, and c-Abl activity enhances coactivation through the tyrosine phosphorylation of p68 at Y⁵⁹³.

The highly homologous p72 DEAD box RNA helicase (Ddx17) has been shown to interact with p68 and coactivate both estrogen receptor α and MyoD ( 14, 17). Therefore, we examined the effect of p72 on AR-mediated luciferase reporter activity. p72 did not enhance AR-regulated transcriptional activity above levels seen with AR alone (Supplementary Fig. S3). Moreover, increasing amounts of p72 had no effect on p68-mediated coactivation of AR transcriptional activity (at a low p68 concentration), and increasing p68 and p72 amounts (at the same concentration) did not enhance AR-mediated transcriptional activity over levels observed with p68 alone. These data show that p72 does not function as an AR transcriptional coregulator either alone or in conjunction with p68.

p68 knockdown by RNA interference reduces AR and PSA protein and mRNA expression. To determine a physiological role for p68 in AR transcriptional activity, we examined the effects of RNA interference (siRNA)–mediated silencing of endogenous p68 gene expression on the expression of AR and the AR-regulated androgen-responsive PSA gene in vivo in the LNCaP cell line. In the presence of androgens, p68 targeted siRNA (but not NS siRNA) reduced the expression of p68 mRNA (>60%; Fig. 4B ) and protein ( Fig. 4A) compared with control (non–siRNA transfected) LNCaP cells. Furthermore, a reduction in the expression of PSA and AR mRNA (∼40% and ∼60%, respectively; Fig. 4C and D) and protein ( Fig. 4A) was also seen with p68 targeted siRNA compared with NS controls. The AR is an androgen-responsive gene itself, and these findings clearly show that p68 targeted RNAi in PCa cells in the presence of androgens reduces protein expression of both the AR and the AR-regulated PSA gene. These data are strong evidence that p68 is an important coactivator of AR-mediated genes.

p68 enhances AR-dependent repression of splicing of a hormone-responsive CD44 variable exon minigene. Although p68 plays key roles in alternative splicing in some contexts, e.g., in splice site stability, spliceosome assembly, and H-ras splicing (reviewed in ref. 10), it is unable to directly affect CD44 variable exon inclusion (ref. 37; Supplementary Fig. S4). However, transcriptional coregulators of nuclear hormone receptors have been shown to affect splicing events ( 32, 38, 39). The AR has recently been shown to repress splicing of a CD44 variable exon minigene downstream of a hormone-responsive promoter ( 31, 39). Because p68 is recruited to the PSA promoter and coactivates AR-dependent transcription, we considered the possibility that p68 may also modulate AR-dependent splicing. We used a CD44 variable exon v4 and v5 minigene under the transcriptional control of the hormone-responsive mouse mammary tumor virus (MMTV) promoter (CDM) transfected into HEK 293 cells with expression constructs encoding AR and/or p68 ( Fig. 5A ). RT-PCR and band densitometry ( Fig. 5B and C), to study alternatively spliced transcripts, showed that p68 enhanced AR-mediated exon skipping in the presence of androgens (compare the ratio of exon skipping/inclusion product in lanes 1 and 2 with lanes 5 and 6; P = 0.008).

p68 protein expression in clinical prostate biopsies. Because other AR transcriptional coregulators have been shown to be overexpressed in PCa ( 40), we decided to examine p68 protein expression by immunostaining of a prostate TMA containing 147 cancer and 44 benign biopsies. The incidence of p68 expressing biopsies was significantly greater in the PCa cohort (66 of 147; 45%) compared with benign prostatic hyperplasia (BPH; 11 of 44, 25%, P = 0.003, two-sided Fisher's exact test; Fig. 6A ). In addition, an increasing frequency of p68 expression was observed with increasing GG tumors [GG3 = 14 of 38 (37%), GG4 = 23 of 51 (45%), GG5 = 26 of 55 (47%)], although the trend did not reach statistical significance (P = 0.3, Kruskal-Wallis test; Fig. 6B and C). Up-regulation in the level of p68 expression by nearly 3-fold was seen in PCa biopsies compared with BPH (P = 0.008, Mann-Whitney test; Fig. 6D). No statistical correlation with local stage, PSA, metastases, time treatment failure, or survival was seen with p68 expression in either multivariate or univariate analyses.

Discussion

A number of AR transcriptional coregulators have been identified by our group and others that play a role in PCa ( 40). In this study, we reveal a novel role for the DEAD box RNA helicase p68 as an AR coactivator in PCa. Our results show that p68 is an AR-interacting protein and is associated with the AR at the promoter region of an androgen-responsive gene (PSA). By overexpression of p68 in luciferase reporter assays, we show a functional role for p68 as an AR transcriptional coactivator, dependent on Y⁵⁹³ phosphorylation, but independent of ATP-binding. Furthermore, p68 coactivator function was verified in vivo in PCa cells by p68 targeted RNAi demonstrating reduced expression of androgen-responsive endogenous protein (AR and PSA).

The requirement of ATP dependency in p68-mediated transcription is conflicting. A recent study suggested that ATP binding was required for the transcriptional activation of cyclin D1 by tyrosine phosphorylated p68 ( 24). However, helicase inactive ATPase mutants have been shown to coregulate transcription as efficiently as wild type for several other transcription factors ( 13, 14, 16, 17, 29), suggesting that the function of p68 as a transcriptional coregulator is independent of its RNA helicase activity. In agreement with these findings, we also show p68 ATPase mutants maintain coactivator function in ARE luciferase reporter assays demonstrating that neither ATPase activity nor a conformational change upon p68 ATP binding is required for AR coactivation. At present, it is unclear why the ATP-dependent activity of tyrosine phosphorylated p68 is specifically required for the transcription of the cyclin D1 promoter ( 24), but it may reflect differences in the way p68 functions as a direct transcriptional activator and as a coregulator.

The COOH terminal domain of p68 can bind single-stranded RNA and has been shown to be a target for protein kinase C ( 41). Phosphorylation of the COOH terminal domain of p68 consequently abolished RNA binding, and it was hypothesized that phosphorylated p68 may act as a molecular switch between transcriptional functions and splicing ( 42). In this study, we show that tyrosine phosphorylation of p68 on Y⁵⁹³ by c-Abl increased transcriptional coactivation of the AR. The Y⁵⁹³ phosphorylation of p68 has been implicated in promoting epithelial-mesenchymal transition via PDGF- and c-Abl–dependent pathways ( 23). Paracrine signaling and the role of stromal-epithelial interactions leading to loss of cell-cell adhesion in PCa has been well documented ( 43). Hence, a model where p68 may directly coactivate AR transcriptional activity and promote epithelial-mesenchymal transition-driven PCa progression would be of notable interest.

Recent evidence suggests that p68 may function as an “adaptor” or “coupling” protein that coordinates the tightly integrated processes of transcription and mRNA processing ( 10, 44, 45). p68 has key roles in U1snRNP-5′ splice site complex stability, spliceosome assembly, pre-mRNA splicing ( 10), and mRNA transcript clearance ( 12). In this study, we show that p68 potentiates AR-regulated repression of splicing of a CD44 variable exon minigene. In contrast, the homologous DEAD box RNA helicase p72 promotes CD44 exon inclusion downstream of a hormone-responsive promoter ( 32). These results suggest that the effect of p68 on AR-dependent splicing may be via transcriptional coregulatory mechanisms, possibly regulated by the post-translational status of p68. Alternatively, recruitment of p68 to endogenous AR-responsive genes may facilitate spliceosome assembly and increase the rate of RNAPII elongation, affecting splice site recognition and promotion of exon skipping in nascent transcripts ( 46). AR is thought to enhance RNAPII elongation by interacting with transcription factor IIH (TFIIH) and positive transcription elongation factor b ( 47), which phosphorylate the COOH terminal domain of RNAPII switching from a nonprocessive to a processive form ( 48– 50). In this study, we show p68 is involved in both AR transcriptional coactivation and AR-mediated repression of splicing of an androgen-responsive gene. Therefore, p68 may serve as a common link between transcription and splicing machinery. RNA processing activities are thought to be strongly connected to efficient transcriptional coregulation, and this study provides evidence for p68 as part of a “molecular link” that involves the coupling of transcriptional initiation to mRNA processing.

Finally, we show that p68 is overexpressed in PCa biopsies compared with benign prostatic hyperplasia controls. There is an increase in the incidence of p68 expression by 2-fold and the level of expression increases by 3-fold (P = 0.008), highlighting the clinical significance of p68 in PCa. Not all PCa samples express p68, and this may represent a specific mechanism of disease progression in a discrete cohort. Although, there was limited correlation with other clinicopathologic variables (grade, stage, PSA, and prognosis), a putative role of p68 in PCa progression is shown by the functional data and the increased levels of expression in clinical samples.

Identifying cofactors that influence the development of hormone escape of PCa will be important in the development of effective therapies and prevention of hormone refractory PCa. This study provides strong evidence of the up-regulation of a novel AR coactivator in PCa, which may play a significant role in progression to androgen independence, as well as a role in epithelial-mesenchymal transition; activation of c-Abl pathways may enhance p68 function as a coactivator of AR-dependent gene expression in PCa. Hence, tyrosine kinase inhibitors, such as imatinib, may be of therapeutic benefit if posttranslational modifications are shown to be important in modulating p68 function and affect disease phenotypes.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.