The androgen-signaling pathway with the androgen receptor (AR) as its key molecule is widely understood to influence prostate tumor growth significantly even after androgen ablation. Under androgen-deprived conditions, the AR may be activated inappropriately through interaction with other molecules, including cyclic AMP-dependent protein kinase A (PKA). In a previous study, we have shown that knocking down the AR significantly inhibits prostate tumor growth. In this study, we show that combined inhibition of the AR and the regulatory subunit I alpha of PKA (RIα) with small interference RNAs significantly increased the growth-inhibitory and proapoptotic effects of AR knockdown. This treatment strategy was effective in androgen-sensitive and in androgen ablation-resistant prostate cancer cells. In addition, we report that downregulating PKA RIα was sufficient to inhibit PKA signaling and interestingly also impaired AR expression and activation. Vice versa, AR knockdown induced a decline in PKA RIα, associated with reduced PKA activity. This mutual influence on expression level was specific, because siRNAs against the AR did not affect expression of PKA RIα in AR negative DU-145 cells and a siRNA control did not affect protein expression. Another important finding of our study was that depletion of PKA RIα also potentiated the antiproliferative effect of the antiandrogen bicalutamide in androgen-sensitive LNCaP. We therefore concluded that combined inhibition of PKA RIα and AR may be a promising new therapeutic option for prostate cancer patients and might be superior to solely preventing AR expression.
Prostate cancer is a very common malignancy with significant complexity concerning the mechanisms that regulate its manifestation and progression. Development of normal prostate and early-stage prostate cancer depends on circulating androgens that elicit their effects through the androgen receptor (AR).1 The AR is a transcription factor, which—upon activation in the cytoplasm where it undergoes a cascade of activation processes—translocates to the nucleus where it finally binds to androgen-responsive elements of androgen-regulated genes such as the prostate-specific antigen (PSA). In prostate tumor cells, the activation of the androgen-signaling cascade is critical for survival and progression.2 Therefore, blocking AR activation by androgen ablation still represents 1 of the preferential treatment options for advanced prostate cancer.3 However, tumor regression as a response to androgen ablation is only palliative, because tumors eventually relapse due to transition of the malignancy from androgen-dependent to a hormone-refractory state.1 This transition results in poor prognostic outcome and limited treatment options, which are usually insufficient for tumor elimination.4 Consequently, after treatment failure, hormone-refractory prostate cancer becomes lethal.
The development of hormone-refractory prostate cancer is likely triggered by several different mechanisms. Apart from androgen-dependent prostate cancer, significance of AR as a key molecule is also evident in the hormone-refractory stage of the disease.5–7 Possible mechanisms by which the AR may contribute to prostate tumor growth include increased expression of AR6, 8, 9 or AR mutations that render the receptor promiscuous, allowing it to accept a broad spectrum of ligands that can bind to and activate the receptor.10–14 Under castrated conditions, it is also considered that low concentrations of androgens remaining in the tissue may contribute to the activation of AR.15 Besides, AR can be activated in a so-called ligand-independent manner through cross-talk with other cell-survival molecules such as cyclic AMP (cAMP)-dependent protein kinase A (PKA).2, 16–18 In particular, it was shown that PKA, as activated by forskolin, can phosphorylate the AR and thereby stimulate the expression of PSA.19
cAMP plays a critical role in the regulation of cell growth and differentiation thereby exerting a dual function via 2 different receptor subtypes, PKA-I and PKA-II, each consisting of 2 catalytic and 2 regulatory subunits, thereby forming a heterotetrameric enzyme. The 2 PKA subtypes have different regulatory subunits (RI and RII), whereas the catalytic subunits are identical for both.20 In addition, they differ in expression and function. Although PKA-II is found in normal nonproliferating tissues and in growth-arrested cells, PKA-I is overexpressed in growth-stimulated cells.20 In particular, the regulatory subunit RIα (PKA RIα) of PKA-I has been reported to be constitutively overexpressed in several tumor types, including colorectal, breast and lung cancers, in which it was associated with a poor prognosis.20, 21 Likewise, prostate tumor growth was arrested by blocking PKA with site-selective cAMP analogs20, 21 and downregulation of PKA-RIα with an antisense oligonucleotide.21, 22 These studies have shown that sequence-specifically knocking down the regulatory subunit RIα is sufficient to effectively reduce PKA-RIα mRNA and protein levels, resulting in growth inhibition and induction of apoptosis.
One of the valuable strategies to prevent inappropriate AR activation is by inhibiting AR itself. In fact, we have previously shown that the inhibition of AR expression by antisense molecules results in significant regression of prostate tumor growth in vitro and in vivo.23, 24 The aim of this study was to increase the growth inhibitory potential of AR knockdown by simultaneously inhibiting PKA-I signaling. We show that the combined use of small interference RNAs (siRNAs) against AR and PKA RIα-enhanced growth arrest of androgen-dependent and androgen ablation-resistant prostate cancer cells compared to single treatment. Targeting the regulatory subunit RIα of PKA-I is sufficient in effectively reducing PKA RIα protein levels as well as activity of PKA-I and interestingly also impaired AR signaling, indicating a possible cross-talk between the 2 pathways.
AR: androgen receptor; CREB: cAMP response element-binding protein; PKA RIα: protein kinase A regulatory subunit type I alpha; PKA RIIβ: protein kinase A regulatory subunit type II beta; PKA Cα: protein kinase A catalytic subunit alpha; PSA: prostate-specific antigen; siRNA: small interference RNA; pVASP: phosphorylated vasodilator stimulated phosphoprotein; TBP: TATA-binding protein
Material and Methods
Cell culture and chemicals
LNCaP and DU-145 prostate cancer cells (obtained from the American Type Culture Collection, USA) were cultured in RPMI 1640 (PAA, Austria) containing 10% fetal calf serum (FCS), 1% Glutamax (Invitrogen) and antibiotics, and maintained at 37°C in a humidified atmosphere of 5% CO2. VCaP cells (obtained from the American Type Culture Collection, USA) were cultured in DMEM medium (PAA, Austria) containing 10% FCS, 2% Glutamax (Invitrogen). VCaP prostate cancer cell line is derived from a vertebral metastatic lesion.25 These cells are androgen-dependent and express high levels of AR. The subline LNCaPabl was established from LNCaP by long-term culture in steroid-depleted medium.26 These cells were then routinely cultured in RPMI 1640 with 10% charcoal-stripped FCS (CS-FCS), 1% Glutamax and antibiotics. Characteristically, LNCaPabl cells become hypersensitive to androgen stimulation, show higher expression of AR than the parental cells and are able to grow in castrated mice, thereby representing a model to study hormone-refractory prostate cancer.
Five different siRNAs were used for the experiments. Three different siRNAs were applied to knockdown the AR, 1 against the polyglutamine region (siAR-1) [sense 5′-GCAGCAGCAGCAGCAGCAGdTdT-3′, antisense 5′-CUGCUGCUGCUGCUGCUGCd TdT-3′], 1 against a sequence 70 base pairs downstream the AUG start codon (siAR-2) [sense 5′-GACCUACCGAGGAGCUUU CdTdT-3′, antisense 5′-GAAAGCUCCUCGGUAGGUCdTdT-3′] and 1 against the ligand-binding domain of AR (siAR-3) [sense 5′-GCACUGCUACUCUUCAGCAdTdT-3′, antisense 5′-UGCU GAAGAGUAGCAGUGCdTdT-3′].27 The siRNA for PKA (siPKA) was designed against the PKA RIα subunit [sense 5′-CCAUGGA GUCUGGCAGUACdTdT-3′, antisense 5′-GUACUGCCAGACUCCAUGGdTdT-3′]. As a negative unspecific control, a siRNA designed against luciferase (siLUC) [sense 5′-CGUACGCGGAA UACUUCGAdTdT-3′, antisense 5′-UCGAAGUAUUCCGCGUACGdTdT-3′] was used. All siRNAs were purchased from GenXpress, Austria. Bicalutamide (Astra Zeneca, UK), an inhibitor of AR transcriptional activity, was applied in a concentration of 5 μM. Forskolin (Calbiochem, USA) was used as an inducer of PKA activity in a concentration of 1 nM, whereas the PKA inhibitor H89 (Calbiochem, USA) was used at 20-μM concentration. The synthetic androgen methyltrienolone (R1881) was purchased from Perkin Elmer, USA.
Semiconfluent cells were transfected with siRNAs using lipofectamine 2000 (Invitrogen), following the manufacturer's protocol. Transfection was performed in serum and antibiotic-free medium for the first 4 hr, followed by further incubation in antibiotic-free medium containing 5% FCS. Mock controls were treated with lipofectamine alone.
After 72 hr of treatment, cells were harvested from 6-well plates. Whole cell extracts were resuspended in lysis buffer [20 mM NaH2PO4, 1 mM EDTA, 10% glycerol, 0.1 nM PMSF, 0.5 nM NaF, 0.5% Protease Inhibitor Cocktail Set III (Calbiochem, Germany), 0.5% Phosphatase Inhibitor Cocktail 2 (Sigma, USA)] and shaken for 1 hr at 4°C. Nuclear fractions were extracted by using Ne-Per Nuclear and Cytoplasmic Extraction kit (Thermo scientific, USA) according to the manufacturer's protocol. After lysis of the cells, the supernatant was collected by centrifugation at 10,000 rpm for 10 min. Protein content was determined by Bradford Assay. Equal amounts of protein (20 μg) were loaded and resolved in 4–12% Bis–Tris gels (Invitrogen) and subsequently transferred onto nitrocellulose membranes (Invitrogen). Membranes were blocked for 1 hr by using Starting Block buffer (THP Medical Products, Austria), followed by overnight incubation with primary antibodies. The primary antibodies used for immunoblotting were as follows: AR (1:250, Biogenex, USA), PKA RIα (1:500, BD Transduction Laboratories, USA), PKA regulatory subunit type II beta (PKA RIIβ) (1:500, Sigma Life Science, USA), PKA catalytic subunit α (Cα) (1:500, Santa Cruz Biotechnology, USA), vasodilator-stimulated phosphoprotein (pVASP) (1:500, Cell Signaling Technology, USA), cAMP response element-binding protein (CREB) and phosphorylated CREB (pCREB) (1:1,000 Cell signaling, USA), GAPDH (1:50,000, Chemicon, USA) and TATA-binding protein (1:2,000, Abcam, UK). After 4 × 5 min of washing with TBS containing 0.05% Tween 20 (TBST), membranes were incubated for 1 hr with fluorescence-labeled secondary antibodies (Molecular Probes, USA) and then washed again with TBST as previously described. The membranes were finally scanned and quantified using the Odyssey infrared imaging system (LiCor Biosciences, USA).
Growth and cytotoxicity assays
For recording cell proliferation, 1 × 104 cells were treated in 96-well plates for 72 hr. Thereafter, [3H]thymidine (1 μCi/well) was added to the cells. Plates were incubated overnight at 37°C and then frozen at −20°C. After thawing, DNA was harvested on 96-well filter plates (UniFilter; Perkin Elmer, USA), and 50 μl of scintillation fluid was added. Radioactivity was quantified using Chameleon 5025 liquid scintillation counter (HVD Life Sciences, Austria).
For measuring cytotoxicity, cells were seeded and treated in 96-well plates for 72 hr. Subsequently, EZ4U kit (Biomedica, Austria) was used according to the manufacturer's instructions for measuring the cell-viability status of the cells under different treatments. Absorbance was measured at 450 nm.
For quantifying PSA secretion, LNCaP were seeded in 6-well plates in RPMI 1640 supplemented with 3% CS-FCS and 1 nM R1881. After overnight incubation, cells were transiently transfected with the siRNAs. Another 24 hr later, medium was changed (using again the medium described earlier), and cells were incubated for 48 hr until PSA measurement. Secreted PSA was measured in supernatant by using Advia CentaurXP Immunoassay System (Siemens, Munich, Germany). PSA values (ng/ml) were normalized with total cell mass (μg total protein).
Caspase-3 was quantified by the EnzChek® Caspase-3 Assay Kit #2 (Invitrogen), according to the manufacturer's protocol. After transfection in 6-well plates for 72 hr, cells were lysed, and the supernatant was collected by centrifugation at 10,000 rpm for 10 min. A mixture containing a reaction buffer and a caspase substrate containing the amino acid sequence D-E-V-D was prepared according to the protocol and added to each sample for monitoring activity of caspase-3 and caspase-7. After 30 min of incubation at room temperature, activity rates of caspase-3 and caspase-7 were recorded by measuring the fluorescence at 520 nm. For normalization of the values obtained, results were adjusted to protein content.
PKA activity assay
PKA activity was measured in cells after treatment with a nonradioactive PKA kinase activity kit (Assay Designs, USA). This assay is based on a solid-phase enzyme-linked immuno-sorbent assay (ELISA) that uses a specific synthetic peptide as a substrate for PKA and a polyclonal antibody that recognizes the phosphorylated form of the substrate. Cells were seeded and transfected for 48 hr as described earlier. After lysis of the cells, the samples were placed in provided wells precoated with the substrate peptide for PKA. The reaction was initiated by addition of 10-μl diluted adenosine triphosphate to each well except the blank. The wells were incubated for 90 min at 30°C. Then, the supernatant from each well was discarded, and the phosphospecific substrate antibody was added for 1 hr at room temperature, followed by 4 washing steps with the buffer provided. A horseraddish-peroxidase-conjugated secondary antibody was then added to each well except the blank for 30 min at room temperature. After washing, a tetramethylbenzidine substrate was added for 1 hr, followed by the addition of an acid stop solution. Absorbance was measured at 450 nm. PKA activity values were finally normalized to protein content.
Immunohistochemistry was performed with 3-μm paraffin tissue sections using the Ventana Discovery—XT-staining automate (Roche). Tris–EDTA pretreatment for pVASP and α-methylacyl-CoA racemase (AMACR), and citrate buffer pretreatment for AR was followed by incubation with primary antibody solution for 1 hr. No pretreatment was applied for PSA staining. The following antibodies were used: AR (1:80, Biogenex, USA), pVASP (1:35, Cell Signaling Technology, USA) and PSA (1:5,000, DAKO, Denmark), AMACR (1:300, DAKO, Denmark). Specificity of staining was controlled by including an unspecific control antibody (DAKO, Denmark).
For manual immunohistochemistry of PKA RIα, antigen retrieval was achieved by autoclaving the tissue samples in citrate buffer at 115°C for 3 min. After blocking with TBST-M (TBST containing 5% nonfat dry milk), slides were incubated with primary antibody overnight at 4°C (PKA RIα 1:400). The signal was developed with the Zymed Polymer Detection System (Invitrogen) according to the manufacturer's instructions. For double staining of AR and PKA RIα, we first stained for PKA RIα manually and then performed AR staining using the Ventana Discovery—XT staining automate. AMACR was used to stain tumor epithelial cells.
All values were expressed as mean ± standard deviation. Each value is the mean of at least 3 independent experiments in each group. Significant differences among groups were analyzed by 1-way ANOVA and Dunnett t-tests. p values below 0.05 were defined as statistically significant (*p < 0.05; **p < 0.001 for comparison between single-targeting treatment and control and +p < 0.05; ++p < 0.001 for comparison between binary targeting treatment and combination of each siRNA with siLUC).
Progression of prostate cancer to an androgen-independent state does not involve loss of AR expression, hence allowing the AR to be activated inappropriately. In a previous study, we therefore proposed that inhibiting AR expression with short antisense oligonucleotides or siRNA might be a very efficient method to inhibit androgen-sensitive as well as castration-resistant prostate cancer and provide a treatment advantage over standard androgen ablation. In this study, we aimed at potentiating the effect of this AR knockdown by simultaneously targeting an additional survival pathway in prostate cancer. For this purpose, we chose PKA-I, which is not only considered to be an important survival molecule in prostate cancer, but also as 1 of the proteins that are able to activate the AR in a ligand-independent manner. We therefore hypothesized that this multitargeting approach would not only have a stronger overall antitumor effect but also exclude the possibility that cells, which weakly express AR, could escape therapy.
Expression and localization of AR and PKA RIα in prostate cancer tissue
Because the efficacy of targeted therapies is dependent on sufficient expression of the target molecules, we first assessed the expression pattern of AR and PKA RIα in human prostate cancer. For this purpose, we performed immunohistochemical staining of PKA RIα and AR on paraffin-embedded tissue sections. As shown in Figure 1, both target molecules were found to be significantly expressed throughout tumor cells (Fig. 1b), which were verified by staining with the tumor marker AMACR (Fig. 1a). Besides, staining of both targets was found in benign areas as well as in stroma. The staining pattern of all 4 molecules, in general, was heterogeneous, a feature that is characteristic for prostate tissue. The AR was localized in the nucleus as well as in the cytoplasm (Fig. 1c). Because the AR translocates from the cytoplasm to the nucleus upon ligand binding, this expression pattern is expected in tumors that had not undergone prior androgen ablation treatment. Likewise, significant amounts of PKA RIα (Fig. 1e) were detected in the cytoplasm of tumor cells. In addition, we revealed expression of the downstream effector molecules of AR and PKA-I, PSA and pVASP (Figs. 1d and 1f), indicating that AR and PKA-I pathways are activated in prostate cancer cells in vivo. Hence, simultaneous targeting of these 2 pathways seems to be a reliable approach.
Inhibition of AR and PKA RIα by siRNA-mediated knockdown
To block target expression sequence-specifically as well as efficiently, we tested the use of short siRNAs and assessed their efficacy in androgen-sensitive LNCaP and VCaP and the androgen ablation-resistant subline LNCaPabl. This subline has been previously established in our laboratory by long-term culture in an androgen-deprived medium, thereby gaining characteristic features of androgen ablation-resistant prostate cancer. Downregulation of the AR was achieved with 3 different siRNAs (siAR-1, siAR-2 and siAR-3). For siRNA-mediated knockdown of PKA RIα, we used a target sequence, which corresponds to a previously described antisense oligonucleotide that has already reached clinical evaluation.22, 28, 29 Concentrations of 20 nM of each siRNA were sufficient to significantly reduce protein expression of AR and PKA RIα, respectively, in all 3 cell lines, LNCaP (Fig. 2a), LNCaPabl (Fig. 2b) and VCaP (Fig. 2c). Hence, target knockdown worked in androgen-sensitive as well as in androgen ablation-resistant prostate cancer cells irrelevantly of AR expression levels. The siLUC negative control did not affect AR expression, suggesting the specificity of the siRNAs.
Downregulation of PKA RIα is sufficient to inhibit PKA-I signaling
Because PKA-I forms a tetramer but the siRNA in this study was only directed against the regulatory subunit of the protein, we further assessed whether depletion of PKA RIα is sufficient to inhibit PKA signaling. To do so, we determined phosphorylation of VASP (pVASP), a specific downstream substrate molecule of PKA and phosphorylation of cAMP response element-binding protein (pCREB) another main target of activated PKA in LNCaP cells. As summarized in Figure 3, pVASP was significantly diminished by siRNA-mediated knockdown of PKA RIα by 75% versus control (Fig. 3a, p < 0.001). Similarly, levels of pCREB were decreased in response to PKA RIα downregulation (Fig. 3b). Moreover, PKA activity was significantly reduced to 53% in comparison with mock control (Fig. 3d, p = 0.001). In contrast to siPKA, the siLUC control had only a minor effect on pVASP expression or pCREB and did not affect PKA activity. Within the same setting, forskolin (5 μM for 48 hr) was used as a positive control showing induction of PKA activity up to 157% (Fig. 3d, p < 0.001).
We next examined expression levels of another PKA regulatory subunit, PKA RIIβ, that has been previously associated with differentiation and growth arrest phenotype.20–22 As shown in Figure 3b, PKA RIIβ levels did not decrease but even slightly increased in response to siRNA treatment. In addition, nuclear expression of the catalytic α subunit of PKA (PKA Cα) was found to be even higher in response to downregulating PKA RIα when compared with controls (Fig. 3c).
Downregulation of PKA RIα results in reduction of AR expression and vice versa
To our surprise, we observed that PKA RIα and AR expression levels were correspondingly inhibited after treatment of the cells with siRNAs against the AR (siAR-1, siAR-2 and siAR-3). In addition, AR was shown to be downregulated in the presence of siPKA. This phenomenon was observed in LNCaP, LNCaPabl (Figs. 2a and 2b) and VCaP, which express higher levels of AR than LNCaP and LNCaPabl (Fig. 2c). siLUC did not affect the expression levels of AR or PKA RIα in any cell line. In contrast, PKA RIα expression was not affected by the siRNAs directed against the AR in AR-negative DU145 cells (Fig. 2d). Because of the absence of AR in DU145, we concluded that the siRNAs against AR were able to decrease PKA RIα levels only in the presence of AR, as shown in LNCaP, LNCaPabl and VCaP, thus signifying a potential impact between AR and PKA pathways at the protein expression level.
Concurrently, it was revealed that AR knockdown in LNCaP also resulted in reduced pVASP by about 60% in comparison with mock control (Fig. 3a). Likewise, pCREB and PKA activity were significantly reduced upon depletion of AR, whereas PKA RIIß and nuclear PKA Cα expression was not affected (Figs. 3b and 3d). In particular, PKA activity was reduced by 30% with siAR-1 (p = 0.043) and by 42% with siAR-2 (p = 0.002). These data further supported the notion that there is a mutual influence of AR and PKA pathways on expression of the proteins.
We next investigated the effect on the AR target molecule PSA. For this, we transiently transfected LNCaP cells with the respective siRNAs in the presence of 1 nM R1881 and measured PSA levels in the cell-culture supernatant 72 hr afterward. As shown in Figure 3e, knockdown of the AR with siAR-1, siAR-2 or siAR-3 induced a 60%, 65% and 56% reduction of PSA compared to mock control (p < 0.001), respectively (Fig. 3e). Furthermore, inhibition of PKA RIα resulted in a 42% decrease of PSA (p < 0.05). Again, PSA levels were not affected by the siLUC control, indicating that these effects were specific.
To further strengthen our outcome that there might be a mutual influence of the 2 molecular pathways, we activated AR and PKA in LNCaP cells by different concentrations of androgen or forskolin and dbcAMP for 72 hr, respectively. Cells were treated with R1881 in 3% (CS-FCS) RPMI 1640, whereas forskolin treatment was done in 10% FCS RPMI 1640. PKA RIα expression levels should be upregulated by stimulation of the cells with androgens, thus resulting in increased levels and activity of PKA and AR. In fact, after treatment of LNCaP with increasing concentrations of R1881, not only AR but also PKA RIα protein levels raised (Fig. 4a). Moreover, when forskolin or dbcAMP was added to the cells, PKA RIα and AR protein levels were similarly increased (Fig. 4b). We then investigated the activity status of the 2 pathways in LNCaP after addition of R1881, forskolin or dbcAMP. As shown in Figures 4c–4f, PSA levels were increased and, similarly, PKA activity was enhanced in all cases.
Knockdown of PKA RIα enhances the growth-inhibitory effect of AR downregulation in androgen-sensitive and androgen ablation-resistant prostate cancer cells
We next assessed whether the depletion of PKA RIα could improve the growth-inhibitory effects of AR knockdown. Androgen-sensitive LNCaP and VCaP and androgen ablation-resistant LNCaPabl were treated with different siRNA combinations for 3 days, followed by measuring [3H] thymidine incorporation and cell cytotoxicity. Final concentration in double targeting was 20 nM, whereas concentration of each siRNA for single treatment was 10 nM.
In androgen-sensitive LNCaP cells, simultaneous knockdown of AR and PKA RIα resulted in significantly reduced cell proliferation by more than 70% when compared with mock control (siAR-1 + siPKA 75%; p <0.05, siAR-2 + siPKA 72%; p < 0.05, siAR-3 + siPKA 72%; p < 0.05) (Fig. 5a). By comparison, downregulation of the AR alone with siAR-1, siAR-2 or siAR-3 inhibited cell proliferation only by 43%, 40% and 48% versus mock control (p < 0.05), respectively. Silencing of PKA RIα with siPKA also resulted in decreased proliferation (25% decrease over mock control), although the effect was weaker than with AR knockdown. Similar results were obtained with VCaP, although the growth-inhibitory effects were slightly weaker than in LNCaP cells (Fig. 5b). These results showed that combined inhibition of AR and PKA RIα increases significantly the growth inhibitory effect of AR targeting alone. To further confirm these results, we determined cell viability by MTT assay (Supporting information Fig. 1). Again, combined knockdown of AR and PKA RIα resulted in enhanced growth inhibition (siAR-1 + siPKA 63%; p < 0.05, siAR-2 + siPKA 51%; p < 0.05 and siAR-3 + siPKA 80%; p < 0.001) versus AR depletion alone (siAR-1 40%, siAR-2 30% and siAR-3 32%). Moreover, when cells were treated over a longer period of 6 days, the growth-inhibitory effect was even more pronounced resulting in inhibition of LNCaP cells by 97% (siAR-1 + siPKA) and 92% (siAR-2 + siPKA), respectively (data not shown).
Because dual targeting was performed with a total siRNA concentration of 20 nM, one could argue that the enhanced effect on cell proliferation was just by doubling the amount of siRNA. However, the combination of each siRNA together with the unspecific negative control siLUC (siAR-1 + siLUC, siAR-2 + siLUC, siAR-3 + siLUC and siPKA + siLUC) did not further decrease cell proliferation when compared with single treatment. In addition, treatment of VCaP cells with 20 nM siLUC did not significantly influence cell proliferation. We therefore concluded from these data that combined inhibition of AR and PKA RIα significantly enhances the growth inhibitory effect of AR knockdown in androgen-sensitive prostate cancer cells.
We next investigated whether this combined treatment strategy would also be effective in androgen ablation-resistant LNCaPabl cells (Fig. 5c). Similarly to androgen-sensitive LNCaP cells, simultaneous downregulation of AR and PKA RIα inhibited LNCaPabl cells to 62% (siAR-1 + siPKA), 65% (siAR-2 + siPKA) and 82% (siAR-3 + siPKA) when compared with AR knockdown alone or in combination with siLUC (siAR-1 + siLUC, siAR-2 + siLUC and siAR-3 + siLUC). Again, these results were further confirmed through measuring cell viability by MTT assay (Supporting information Fig. 1), indicating that combined targeting of AR and PKA RIα is also effective in androgen ablation-resistant prostate cancer cells.
To further exclude off target effects of the siRNAs on cell proliferation, we performed another control experiment in AR negative DU-145 cells. When PKA was arrested by the siPKA, cell growth was inhibited by 30% compared to the mock control (p = 0.039) (Fig. 5d). In addition, when DU145 cells were treated with a combination of siPKA and either siRNA against AR, cell viability could not be further decreased compared to siPKA single treatment. siAR-2 and siLUC did not affect cell viability, whereas siAR-1 had only a minor effect on DU145 (12% decrease in cell viability, Supporting information Fig. 1).
Increased apoptosis by combined knockdown of AR and PKA RIα
We have previously shown that siRNA-mediated silencing of the AR induces apoptosis in prostate cancer cells.30 This result could be reproduced in the present study, where individual treatment of LNCaP (Fig. 6a) with either siAR-1, siAR-2 for 72 hr resulted in a 9-fold (siAR-1) and 11-fold (siAR-2) induction of caspase-3 activity rates compared to the mock control, as assessed by ELISA (p < 0.001). Downregulating PKA RIα also induced caspase activity to 7.5-fold over the mock control (p < 0.001), although the effect was weaker than that obtained by AR knockdown.
Intriguingly, these effects could be strongly enhanced to 17-fold (AR-1 + siPKA) and 18-fold (AR-2 + siPKA) when LNCaP were simultaneously treated with siRNAs against both targets (Fig. 6a). This increase in caspase activity was significantly higher than the value obtained by treatment with each siRNA alone and also over the combination of each siRNA together with the unspecific control siLUC.
In VCaP cells that express dramatically higher AR and PKA RIα levels than LNCaP, the effect of the combinatorial treatment in inducing apoptosis was relatively low. Single treatment of VCaP (Fig. 6b) with either siAR-1, siAR-2 or siAR-3 for 72 hr resulted in 43%, 64% and 65% induction of caspase-3 activity rates respectively in comparison with the mock control, while siPKA increased caspase-3 activity by 47%. Nevertheless, the combination of siPKA with each of the 3 siRNAs against AR resulted in 2–2.5-fold induction of apoptosis, an effect that was also significantly higher than combination of each siRNA with negative control siLUC.
A similar result, as in LNCaP, was obtained in androgen ablation-resistant LNCaPabl cells (Fig. 6c). Combined targeting of AR-1 and PKA RIα resulted in 11-fold (siAR-1 + siPKA) and 8-fold (siAR-2 + siPKA) induction of caspase activity, respectively, while treatment with each siRNA alone did not increase the levels of caspase activity for more than 4.5-fold. Thus, combined inhibition of AR and PKA RIα significantly enhanced the proapoptotic effect of single AR knockdown (p < 0.05).
Depletion of PKA RIα potentiates the growth inhibitory effects of the antiandrogen bicalutamide in androgen-sensitive LNCaP cells
We finally tested if enhanced growth inhibitory effects were also achieved by inhibiting the 2 pathways with siRNA knockdown in combination with the chemical inhibitors bicalutamide and H89, respectively. In prostate cancer patients, standard treatment to inhibit AR signaling is by androgen ablation or/and the use of antiandrogens such as bicalutamide, also called Casodex®. When androgen-sensitive LNCaP cells were treated with bicalutamide for 3 days, cell proliferation was decreased by 25% compared to mock control (p = 0.01; Fig. 7a). This growth-inhibitory effect of bicalutamide was strongly enhanced when it was combined with the siPKA resulting in a 60% inhibition versus mock control. Hence, we concluded that the siRNA-mediated inhibition of PKA-I signaling significantly potentiates the antiproliferative effect of the antiandrogen bicalutamide in androgen-sensitive LNCaP cells.
In androgen ablation-resistant LNCaPabl cells, the combination of bicalutamide and siPKA by contrast only resulted in a minor enhancement of growth arrest (28% growth inhibition versus mock control, p = 0.004) compared with PKA-I knockdown alone (23% growth inhibition versus mock control, p = 0.008) (Fig. 7b). This was probably due to the fact that bicalutamide failed to inhibit cell proliferation in this cell line. By contrary, we even observed a moderate increase (23% versus mock control) in proliferation following bicalutamide treatment. This result corresponds well with the previous findings where bicalutamide was shown to loose its growth inhibitory activity and to acquire agonistic properties in androgen ablation-resistant LNCaPabl.26
We next investigated the effect of H89, an assigned inhibitor for PKA, in combination with siRNAs for AR or bicalutamide (Supporting information Fig. 2). LNCaP and LNCaPabl cells were treated with 20 μM H89 for 72 hr in the presence of 5 μM bicalutamide. It was revealed that H89 could inhibit proliferation of LNCaP and LNCaPabl by 20 and 33%, respectively, whereas combination of H89 with bicalutamide further decreased LNCaP proliferation by 58%. In LNCaPabl, the agonistic effect of bicalutamide was abolished in the presence of H89 as in the case of siPKA. When cells were treated with H89 together with siAR-1 or siAR-2, further reduction of proliferation was recorded in both LNCaP and LNCaPabl, whereas H89 treatment combined with siLUC was not more effective in decreasing proliferation than H89 single treatment in either cell line.
The AR-signaling pathway is the main mediator of prostate cancer development and progression. It has been indicated that the AR is expressed significantly in androgen ablation-resistant tumors and is also able to control the transcription of critical genes contributing to tumor progression.5–7 Activation of AR may occur through residual hormone levels within the tumors that can still be detected following androgen ablation.15 It has also been strongly supported that AR activation, in the absence of androgens, occurs due to genetic alterations within the AR gene, allowing the receptor to accept a broad spectrum of nonandrogenic ligands.10–14 Another possible mechanism of inappropriate AR activation is through ligand-independent interaction with other signaling pathways including cAMP-dependent PKA.2, 4, 16, 19 Two different isoforms of PKA are known, PKA-I and PKA-II, which not only differ in terms of their regulatory subunits but also have a strikingly different impact on tumor growth and progression.20 PKA-II is found preferentially in normal nonproliferating tissues and in growth-arrested cells, whereas PKA-I with its regulatory subunit PKA RIα is associated with cell proliferation and survival.20 Overexpression of PKA RIα was revealed in different tumor types and therefore strongly associated with manifestation and progression of cancer cells.20, 21 PKA RIα has been selected in numerous studies as a target for treatment intervention, and an antisense oligonucleotide targeting PKA RIα has already reached clinical evaluation.28, 29
In a previous study, we proposed that inhibiting AR expression with short antisense oligonucleotides23, 24 or siRNAs30 is an efficient strategy to inhibit androgen-sensitive as well as androgen ablation-resistant prostate cancer cells. The key finding of this study was that simultaneous blockade of PKA RIα significantly enhances the effect of AR knockdown in terms of cell-growth arrest and induction of apoptosis. Importantly, this enhanced effect of combined inhibition of the 2 pathways, AR and PKA-I, was efficient in androgen sensitive (LNCaP, VCaP) and androgen ablation-resistant (LNCaPabl) prostate cancer cells. We believe that such a combined targeting strategy includes several major advantages over single-treatment regimens. First, an enhanced growth-inhibitory effect is achieved without increasing drug doses thereby preventing higher costs of treatment and toxic side effects. A second advantage is that by targeting more than 1 molecule a broader spectrum of tumor cells can be reached, a fact that is especially important in tumor types such as prostate cancer which in general exhibits strong heterogeneity. Targeted therapies are strongly dependent on the strong and broad expression of the target molecules. Immunohistochemical staining has revealed that both target molecules, AR and PKA RIα, are strongly expressed in human prostate cancer. Moreover, we considered both pathways to be active, because we also revealed the downstream effector molecules of PKA and AR, pVASP and PSA, which showed a closely similar expression pattern as PKA RIα and AR. These data on in vivo expression pattern suggested that targeting these 2 pathways in prostate cancer represents a reliable approach. AR and PKA RIα are also found to be expressed in benign tissue, but we believe that targeted treatment against them will affect most extensively the prostate cancer sites where the 2 target proteins are overexpressed.
To inhibit AR expression, we used 3 different siRNAs, 1 targeting the N-terminal polyglutamine region (siAR-1) that has previously been shown to efficiently inhibit prostate tumor growth, a second 1 targeting a specific region nearby the start codon of the AR gene (siAR-2) and a third 1 against the ligand-binding domain of AR (siAR-3). All siRNAs efficiently inhibited AR expression and exhibited a strong antiproliferative and proapoptotic effect in androgen-sensitive LNCaP and VCaP as well as in androgen ablation-resistant LNCaPabl cells. In VCaP cells, which express higher levels of AR and PKA RIα than LNCaP cells, siRNA-mediated target knockdown was less efficient. Correspondingly, the effects on proliferation and apoptosis were slightly less prominent in VCaP than in LNCaP cells, pointing out that the therapeutic effect is strongly dependent on the level of target expression. Nevertheless, combined treatment also had an additive antiproliferative effect in this cell line.
An important finding of our study was that siRNA-mediated AR knockdown is much more efficient in terms of tumor growth arrest than the antiandrogen bicalutamide, which acts via inhibition of AR activation. This may be of advantage especially in tumors that escaped from androgen-ablation therapy. To investigate the efficacy of our combination therapy in advanced prostate cancer cells, we used a subline of LNCaP, LNCaPabl, which has acquired specific features of advanced androgen ablation-resistant prostate cancer, including increased AR levels, hypersensitivity to androgen stimulation and the ability to grow in castrated mice.26 Moreover, these cells are less susceptible to growth inhibition and induction of apoptosis than LNCaP,31 a phenomenon that could also be confirmed in this study. When LNCaPabl cells were treated with siRNAs to inhibit AR expression, they were inhibited by 35% versus mock control. By contrast, blocking AR activation with the antiandrogen bicalutamide failed to inhibit the cells, but even resulted in an increase of cell proliferation by 23%. This loss of antagonistic activity of bicalutamide, which reflects the clinical situation of therapy resistance, has previously been shown by our group.26 Hence, we concluded that especially in advanced therapy-resistant prostate cancer, the target-specific prevention of AR expression could have higher tumor inhibition efficacy than blocking AR activation through antiandrogens.
For knocking down PKA RIα, we used a siRNA targeting, a sequence that corresponds to the target sequence of a previously described efficient antisense oligonucleotide.28, 29 The regulatory subunit of PKA-I, PKA RIα, is essential for the formation and subsequent activation of the tetrameric core protein. It undergoes dramatic conformational changes upon complex formation with the catalytic subunits, thus resulting in stabilization of the whole complex.32 It has also been shown that overexpression of PKA RIα in PC3 prostate cancer cells results in elevated PKA catalytic alpha subunit expression but with no increase in mRNA, indicating the posttranscriptional stabilization of the catalytic subunit via formation of the tetramer with PKA RIα.33 In the same context, overexpression of PKA catalytic alpha subunit resulted in concurrent PKA RIα mRNA induction as PKA RIα is the only PKA subunit whose promoter is known to contain cAMP response element (CRE).33, 34 Thus, PKA RIα is thought to be essential as a mediator of PKA signaling and for the stabilization of PKA catalytic subunits via the holoenzyme formation, and in addition, it is assumed that PKA signaling might be autoregulated through the regulation of PKA RIα levels. Here, we have shown that downregulating PKA RIα was sufficient to significantly reduce PKA activity, resulting in cell-growth arrest and induction of apoptosis.
When using siRNA technology, the use of controls is extremely important to exclude unspecific off target effects. Therefore, we used an unspecific siLUC control to demonstrate specificity of our results. The question regarding off target effects arose from the finding that knocking down the AR considerably reduced PKA RIα protein levels. This was associated with the reduction of PKA activity and reduced phosphorylation of the PKA downstream substrates VASP35, 36 and CREB37 in LNCaP cells. On the other hand, siPKA was found to reduce AR expression. A possible interaction between the 2 pathways has been postulated previously, suggesting an activation of the AR via cAMP-dependent PKA activation.2, 4, 16 It has also been shown that androgens cause rapid cAMP-dependent activation of PKA, thereby indirectly stimulating activation of AR signaling in prostate cancer cells.17 Our study provides evidence that, in AR positive prostate cancer cells, protein expression levels of AR and PKA RIα are possibly indirectly dependent on each other. This was assumed from the results illustrating consistently that AR knockdown by 3 different siRNAs significantly reduced PKA RIα expression and PKA activity, whereas PKA RIα downregulation similarly decreased AR protein levels and signaling in LNCaP. In contrast, knockdown of AR or PKA RIα did not decrease but even slightly increase PKA RIIβ levels. PKA RIIβ is a regulatory subtype of PKA-II, which has been associated with cell differentiation and growth-arrest phenotype. This inverse correlation of RIα and RIIβ expression was shown in this study, therefore agreeing with previously published reports.20–22, 33 In the same context, PKA catalytic α subunit has been reported to translocate to the nucleus after inhibition of PKA RIα expression.38 This event also occurred in our study after treatment of LNCaP with either siRNA against AR or PKA RIα.
To further assess if this mutual interaction of the 2 molecules is due to siRNA-mediated off target effects, we further tested our siRNAs in AR negative DU145 prostate cancer cells. In this cell line, PKA RIα was significantly inhibited by siPKA, however, none of the 3 used siRNAs against the AR affected PKA RIα levels, again suggesting that expression of PKA RIα is dependent on AR signaling. Moreover, cell-viability assays in DU145 showed that silencing PKA RIα induced growth arrest, whereas AR knockdown with siAR-2 did not affect viability. Even siAR-1, the siRNA targeting the polyglutamine region of the AR, which is thought to target various CAG repeat containing genes, only had a minor effect on DU145.
To follow the opposite direction and investigate this possible mutual influence between the 2 proteins, we induced AR and PKA signaling by stimulating LNCaP cells with androgen, forskolin and dbcAMP. In earlier studies, it has been proposed that androgen regulation of AR expression occurs in a tissue- and cell-type-specific fashion.39 In particular, we showed here that androgen upregulates AR protein in LNCaP cells. These results are in agreement with the previous studies indicating that androgen stimulation results in increase of AR stability in LNCaP.40 Within the same setting, we showed that androgen increased expression of PKA RIα in LNCaP, a consistent and relevant result in comparison with our knockdown studies. The increase of AR and PKA RIα in the presence of androgen was further associated with increased PSA levels and PKA activity in LNCaP. We then assessed AR and PKA RIα levels in LNCaP after stimulation with forskolin and dbcAMP in parallel. We recorded upregulation of AR expression and activity (PSA induction) in the presence of either forskolin or dbcAMP along with expected induction of PKA activity and higher PKA RIα protein levels.33, 34 This outcome also confirmed an earlier study, in which the authors demonstrated AR induction in LNCaP by cAMP analogues and also identified a potential CRE in the promoter of AR possibly being responsible for this induction.41 Moreover, our result is consistent with the postulation that the activation of AR can occur through PKA signaling,2, 4, 16, 19 and confirms PSA as a common transcriptional target in response to stimulation of both pathways as previously reported.5, 19, 42, 43 We additionally identified candidate androgen-response elements that might be responsible for the regulation of PKA RIα by the AR pathway; however, further experiments are required to elucidate the mechanism(s) by which AR signaling might regulate PKA RIα expression in prostate cancer.
In summary, this is the first study showing that simultaneous inhibition of PKA RIα and AR significantly enhances the antiproliferative and proapoptotic effects of AR knockdown alone in androgen-sensitive and androgen ablation-resistant prostate cancer cells. We therefore conclude that combined targeting of AR and PKA RIα may be a promising new therapeutic option for prostate cancer patients with malignancies of early small tumors, especially when applied as a focal molecular therapy, as well as for advanced tumor stages and might be superior to solely preventing AR expression.
We thank Mrs. Carolin Achammer and Mrs. Gerda Hölzl for helpful assistance with PSA measurements and Mr. Christof Seifarth for technical support in immunohistochemistry.