Clusterin Inhibition Using OGX-011 Synergistically Enhances Hsp90 Inhibitor Activity by Suppressing the Heat Shock Response in Castrate-Resistant Prostate Cancer

Prostate cancer is the most common cancer and the third most common cause of cancer-related mortality in men in the United States (1). Androgen ablation remains the standard effective therapy for patients with advanced prostate cancer, inhibiting proliferation and inducing apoptosis in tumor cells (2). Unfortunately, after short-term remissions, surviving tumor cells recur with castrate-resistant prostate cancer (CRPC) and death usually within 3 years in most men (3). CRPC progression results from mechanisms attributed to reactivation of androgen receptor axis (4), alternative mitogenic growth factor pathways (5, 6), and stress-induced prosurvival gene (7, 8) and cytoprotective chaperone networks (9, 10). To significantly improve survival in men with prostate cancer, new therapeutic strategies to inhibit the appearance of this phenotype must be developed.

Hsp90 is an ATPase-dependent molecular chaperone required for protein folding, maturation, and conformational stabilization of many “client” proteins (11, 12). It interacts with several proteins involved in CRPC, including growth factor receptors, cell-cycle regulators, and signaling kinases such as Akt, androgen receptor (AR), or Raf-1 (13, 14). Tumor cells express higher Hsp90 levels than benign cells (12, 15), and Hsp90 inhibition has emerged as an exciting target in CRPC and other cancers. Many Hsp90 inhibitors were developed targeting its ATP-binding pocket, including natural compounds such as geldanamycin and its analogue 17-allylamino-17-demethoxy-geldanamycin (17-AAG) or synthetic compounds including PF-04928473. These agents inhibit Hsp90 function and induce apoptosis in preclinical studies of colon, breast, prostate, and other cancers (12, 16). Although promising, treatment resistance emerges early due to compensatory mechanisms involving activation of the heat shock factor HSF-1. Once released from Hsp90, HSF-1 translocates to the nucleus, binds to heat shock elements (HSE) of Hsp genes, and increases Hsp transcription activity (13). Therefore, Hsp90 inhibition induces a heat shock response with increased expression of several Hsps including Hsp70, Hsp27, and clusterin (CLU). The upregulation of these molecular chaperones has been reported to play a role in cellular recovery from stress by restoring protein homeostasis, promoting thermotolerance, cell survival, and treatment resistance (14, 17).

CLU is a stress-induced cytoprotective chaperone that inhibits protein aggregation in a manner analogous to small HSPs, and its promoter contains a 14-bp element recognized by the transcription factor HSF-1 (18). In human prostate cancer, CLU levels are low in Gleason grade 3 untreated hormone-naive tissues but increase with higher Gleason score (19) and within weeks after androgen deprivation (20). CLU expression correlates with loss of the tumor suppressor gene Nkx3.1 during the initial stages of prostate tumorigenesis in Nkx3.1 knockout mice (21). Experimental and clinical studies associate CLU with development of treatment resistance, where CLU suppresses treatment-induced cell death in response to androgen withdrawal, chemotherapy, or radiation (10, 20, 22, 23). Overexpression of CLU in human prostate LNCaP cells accelerates progression after hormone therapy or chemotherapy (10, 22), identifying CLU as an antiapoptotic gene upregulated by treatment stress that confers therapeutic resistance. OGX-011 is a second-generation phosphorothioate antisense oligonucleotide currently in late-stage clinical development that potently inhibits CLU expression and enhances the efficacy of anticancer therapies in various human cancers including prostate cancer (17, 24). Although targeting CLU synergistically enhances the cytotoxic effects of chemotherapy, a role for CLU has not been characterized in the context of Hsp90 inhibitor treatment and resistance.

We hypothesized that Hsp90 inhibition induces a heat shock response with increased HSF-1 activity and CLU expression, which functions to inhibit treatment-induced apoptosis and enhance emergence of treatment resistance. It follows that knockdown of CLU with OGX-011 will potentiate the effect of Hsp90 inhibitors in CRPC.

The human prostate cancer cell line PC-3 was purchased from the American Type Culture Collection [(ATCC), 2008, ATCC authentication by isoenzyme analysis] and maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen-Life Technologies, Inc.) supplemented with 5% FBS and 2 mmol/L l-glutamine. LNCaP cells were kindly provided by Dr. Leland W.K. Chung (1992, MD Anderson Cancer Center, Houston, TX) and tested and authenticated by whole-genome and whole-transcriptome sequencing on an Illumina Genome Analyzer IIx platform in July 2009. LNCaP cells were maintained RPMI 1640 (Invitrogen-Life Technologies, Inc.) supplemented with 5% FBS and 2 mmol/L l-glutamine. All cell lines were cultured in a humidified 5% CO₂/air atmosphere at 37°C. All cell lines were passaged for less than 3 months after resurrection. Western blotting and/or real-time PCR was carried out for AR and prostate-specific antigen (PSA) each time when LNCaP cells were resurrected.

Hsp90 inhibitor, PF-04928473 [4-(6,6-dimethyl-4-oxo-3-trifluoromethyl-4,5,6,7-tetrahydro-indazol-1-yl)-2-(4-hydroxy-cyclohexylamino)-benzamide] and its prodrug PF-04929113 were kindly provided by Pfizer and used for in vitro and in vivo studies, respectively. These compounds are novel, synthetic, small-molecular-weight inhibitors that bind the N-terminal adenosine triphosphate binding site of Hsp90, and PF-04929113 is orally bioavailable. For the in vitro studies, PF-04928473 was dissolved in dimethyl sulfoxide (DMSO) at 10 mmol/L stock solutions and stored at −20°C. For the in vivo studies, PF-04929113 was dissolved in PBS containing 0.5% carboxymethylcellulose and 0.5% Tween 80 (Invitrogen-Life Technologies, Inc.) at 15 mg/mL and stored at 4°C.

17-AAG was kindly provided by NIH and used for in vitro and in vivo studies. For the studies, 17-AAG was dissolved in DMSO at 10 mmol/L stock solutions and stored at −20°C.

siRNAs were purchased from Dharmacon Research, Inc., using the siRNA sequence corresponding to the human CLU initiation site in exon 2 and a scramble control as previously described (25). Second-generation antisense (OGX-011) and scrambled (ScrB) oligonucleotides with a 2′-O-(2-methoxy)ethyl modification were supplied by OncoGenex Pharmaceuticals. OGX-011 sequence (5′-CAGCAGCAGAGTCTTCATCAT-3′) corresponds to the initiation site in exon II of human CLU. The ScrB control sequence was 5′-CAGCGCTGACAACAGTTTCAT-3′. Prostate cells were treated with siRNA or oligonucleotides, using protocols described previously (25).

Prostate cancer cells lines were plated in appropriate media (DMEM or RPMI) with 5% FBS and treated with PF-04928473 or 17-AAG at indicated concentration and time and cell growth was measured using the crystal violet assay as described previously (26). Detection and quantitation of apoptotic cells were done by flow cytometry (described later) and Western blot analysis. Each assay was repeated in triplicate.

The combination index (CI) was evaluated using CalcuSyn dose–effect analysis software (Biosoft). This method, based on the multiple drug effect equation of Chou–Talalay (27), is suitable for calculating combined drug activity over a wide range of growth inhibition: CI = 1, additivity; CI > 1, antagonism; CI < 1, synergism. CI was calculated at ED₅₀ and ED₇₅.

Caspase-3 activity was assessed 3 days after treatment with the CaspACE Assay System, Fluorometric (Promega). Incubation of 50 μg of total cell lysate was done with the caspase-3 substrate AC-DEVD-AMC at room temperature for 4 hours, and caspase-3 activity was quantified in a fluorometer with excitation and emission at 360 and at 460 nm, respectively.

Prostate cancer cell lines were incubated in the absence or presence of 1 μmol/L PF-04928473 or 17-AAG for 72 hours, trypsinized, washed twice, and incubated in PBS containing 0.12% Triton X-100, 0.12 mmol/L EDTA, and 100 μg/mL ribonuclease A; 50 μg/mL propidium iodide was then added to each sample for 20 minutes at 4°C. Cell-cycle distribution was analyzed by flow cytometry (Beckman Coulter Epics Elite; Beckman, Inc.), based on 2N and 4N DNA content. Each assay was done in triplicate.

Samples containing equal amounts of protein (depending on the antibody, 5–50 μg) from lysates of cultured tumor prostate cell lines underwent SDS-PAGE and were transferred to nitrocellulose filters. The filters were blocked in Odyssey Blocking buffer (LI-COR Biosciences) at room temperature for 1 hour, and blots were probed overnight at 4°C with primary antibodies (Supplementary Materials) to detect proteins of interest. After incubation, the filters were washed 3 times with washing buffer (PBS containing 0.1% Tween) for 5 minutes. Filters were then incubated for 1 hour with 1:5,000 diluted Alexa Fluor secondary antibodies (Invitrogen) at room temperature. Specific proteins were detected using Odyssey IR imaging system (LI-COR Biosciences) after washing.

Total RNA was extracted from cultured cells after 48 hours of treatment using TRIzol reagent (Invitrogen-Life Technologies, Inc.). Two micrograms of total RNA was reversed transcribed using the Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Science). Real-time monitoring of PCR amplification of cDNA was done using DNA primers (Supplementary Table S1) on the ABI PRISM 7900 HT Sequence Detection System (applied Biosystems) with SYBR PCR Master Mix (Applied Biosystems). Target gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels in respective samples as an internal standard, and the comparative cycle threshold (C_t) method was used to calculated relative quantification of target mRNAs. Each assay was carried out in triplicate.

LNCaP and C4-2 cells (2.5 × 10⁵) were plated on 6-well plates and transfected using lipofectin (6 μL per well; Invitrogen-Life Technologies, Inc.). The total amount of HSE plasmid DNA used was normalized to 1 μg per well by the addition of a control plasmid. PF-04928473 or 17-AAG (1 μmol/L) was added 4 hours after the transfection for a total of 48 hours. HSE-luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega) with the aid of a microplate luminometer (EG&G Berthold). All experiments were carried out in triplicate wells and repeated 3 times using different preparations of plasmids.

Tumor cells were grown on coverslips and treated with different concentration of PF-04928473 or 17-AAG for 48 hours. After treatment, cells were fixed in ice-cold methanol completed with 3% acetone for 10 minutes at −20°C. Cells were then washed thrice with PBS and incubated with 0.2% Triton/PBS for 10 minutes, followed by washing and 30 minutes of blocking in 3% nonfat milk before the addition of antibody overnight to detect HSF-1 (1:250). Antigens were visualized using anti-mouse antibody coupled with fluorescein isothiocyanate (1:500; 30 minutes). Photomicrographs were taken at 20× magnification with a Zeiss Axioplan II fluorescence microscope, followed by analysis with imaging software (Northern Eclipse; Empix Imaging, Inc.).

Male athymic nude mice (Harlan Sprague-Dawley, Inc.) were injected subcutaneously with 2 × 10⁶ LNCaP cells (suspended in 0.1 mL Matrigel; BD Biosciences). The mice were castrated once tumors reach between 300 and 500 mm³ or the PSA level increased above 50 ng/mL. Once tumors progressed to castrate resistance, mice were randomly assigned to vehicle, PF-04929113 alone, PF-04929113 + ScrB ASO, or PF-04929113 + OGX-011. PF-04929113 (prodrug, 25 mg/kg; formulation in 0.5% carboxymethylcellulose + 0.5% Tween 80) is orally administered 3 times per week, and OGX-011 or ScrB ASO (15 mg/kg) was injected intraperitoneally once daily for the first week and then 3 times per week. Each experimental group consisted of 10 mice. Tumor volume was measured twice weekly (length × width × depth × 0.5432). Serum PSA level was determined weekly by enzymatic immunoassay (Abbott IMX). PSA doubling time (PSA dt) and velocity were calculated by the log-slope method (PSA_t = PSA_initial × e^mt). Data points were expressed as average tumor volume ± SEM or average PSA concentration ± SEM.

To establish PC-3 tumors, 2 × 10⁶ PC-3 cells were inoculated subcutaneously in the flank region of 6- to 8-week-old male athymic mice (Harlan Sprague-Dawley, Inc.). When tumors reached 100 mm³, usually 3 to 4 weeks after injection, mice were randomly selected for treatment with 17-AAG (25 mg/kg) + control ScrB ASO (15 mg/kg) or 17AAG + OGX-011 (15 mg/kg). 17-AAG was injected intraperitoneally 3 times per week, and OGX-011 or ScrB was injected intraperitoneally once per day for the first week and then 3 times per week. Each experimental group consisted of 7 mice. Tumor volume was measured twice weekly. Data points were expressed as average tumor volume ± SEM.

When tumor volume reached 10% or more of body weight, mice were sacrificed and tumors were harvested for evaluation of protein expression by Western blot analyses and immunohistochemistry. All animal procedures were carried out according to the guidelines of the Canadian Council on Animal Care and appropriate institutional certification.

Immunohistochemical staining was done on formalin-fixed, paraffin-embedded 4-μm sections of tumor samples, using adequate primary antibody (Supplementary Materials) and the Ventana autostainer Discover XT (Ventana Medical System) with enzyme-labeled biotin streptavidin system and solvent-resistant 3,3′-diaminobenyidine map kit. All comparisons of staining intensities were made at 200× magnifications.

All in vitro data were assessed using the Student t test and the Mann–Whitney test. Tumor volumes of mice were compared using the Kruskal–Wallis test. Overall survival was analyzed using Kaplan–Meier curves, and statistical significance between the groups was assessed with the log-rank test (GraphPad Prism). Levels of statistical significance were set at P < 0.05.

Dose- and time-dependent effects of 17-AAG or PF-04928473 on the expression of CLU, Hsp90, Hsp70, and Akt protein and mRNA levels were evaluated in LNCaP and PC-3 cells. Both 17-AAG and PF-04928473 increased Hsp70 and CLU protein levels 3-fold in a dose- and time-dependent manner (Fig. 1A–C). Hsp90 inhibition induced a dose- and time-dependent decline of Akt expression as previously reported (28). mRNA levels of CLU, Hsp70, and Hsp90 also increased after Hsp90 inhibitor treatment (Fig. 1D).

Next, we assessed the effects of PF-04928473 treatment on CLU expression in vivo in CRPC LNCaP xenografts by immunohistochemistry and Western blot (Fig. 2). CLU expression increased 4-fold after treatment with PF-04929113 (***, P < 0.001) compared with vehicle-treated tumor (Fig. 2A and B). Similarly, Hsp70, considered a pharmacodynamic measure of Hsp90 inhibition (16, 29), increased 2.3-fold after treatment with PF-04929113 (***, P < 0.001; Fig. 2A).

Because HSF-1 is the predominant regulator of the heat shock response (30, 31), we evaluated the effect of Hsp90 inhibition on HSF-1 activity and expression of HSPs. As expected, 17-AAG or PF-04928473 significantly induced CLU (Fig. 1) and HSF-1 activity in a dose-dependent manner (***, P ≤ 0.001; Fig. 3A). CLU overexpression protected PC-3 tumor cells from PF-04928473–induced apoptosis (**, P ≤ 0.01; Supplementary Fig. S1A). Moreover, HSF-1 knockdown using siRNA decreases CLU expression, sensitizing tumor cells to apoptosis induced by PF-04928473 (Supplementary Fig. S1B), confirming that the protective effect of CLU is mediated by HSF-1. Surprisingly, overexpression of CLU also increased HSF-1 activity (***, P ≤ 0.001, Fig. 3B), whereas CLU knockdown using siRNA or OGX-011 significantly decreased HSF-1 activity (*, P ≤ 0.05; ***, P ≤ 0.001; Fig. 3C), identifying novel feed-forward regulation of HSF-1 by CLU. Indeed, silencing of CLU inhibited HSF-1 transcriptional activity induced by 17-AAG or PF-04928473 (Fig. 3C), and HSF-1 regulated genes such as Hsp27 and Hsp70 (Fig. 3D). CLU knockdown also sequesters HSF-1 in the cytoplasm (Supplementary Fig. S1C), suggesting that CLU plays a role in HSF-1 nuclear translocation and transactivation.

Because Hsp90 inhibitors induce upregulation of CLU and CLU functions as a mediator in treatment resistance (17, 24, 32), we next evaluated whether CLU knockdown potentiated the effect of Hsp90 inhibition. LNCaP cells were treated with OGX-011 and subsequently treated with indicated concentrations of 17-AAG or PF-04928473. OGX-011 significantly enhanced 17-AAG or PF-04928473 activity, reducing cell viability by an additional 20% at 100 and 1,000 nmol/L (*, P < 0.05) compared with cells treated with control ScrB and Hsp90 inhibitor (Fig. 4A). To determine whether this effect was additive or synergistic, the dose-dependent effects with constant ratio design and the CI values were calculated according to the Chou and Talalay median effect principal (27). Figure 4B shows the dose–response curve (combination treatment, OGX-011, or PF-04928473 monotherapy) and the CI plots, indicating that OGX-011 synergistically enhances the effect of Hsp90 inhibitor on tumor cell growth.

Moreover, OGX-011 potentiates the effect of Hsp90 inhibitor to induce apoptosis (Fig. 4C and D). Flow cytometric analysis shows that apoptotic rates (sub-G₁ fraction) increased significantly (P < 0.001) when OGX-011 is combined with 17-AAG (53%) or PF-04928473 (65.4%), compared with control ScrB (4.2%), OGX-011 (17.4%), control ScrB ASO + 17-AAG (18.3%), or control ScrB + PF-04928473 (24.8%; Fig. 4D). Moreover, the combination of OGX-011 with 17-AAG or PF-04928473 increased caspase-dependent apoptosis compared with Hsp90 inhibitor or OGX-011 monotherapy, as shown by cleaved PARP and caspase-3 expression (Fig. 4C). The significant increase in caspase-3 activity confirms that OGX-011 sensitizes cells to Hsp90 inhibition with increased apoptotic rates (Fig. 4D). Reduced cell viability from combined CLU plus Hsp90 inhibition results, in part, from decreases in p-Akt levels in both PC-3 and LNCaP cells, as well as AR (and PSA) expression in LNCaP cells (Fig. 4C).

We next evaluated effects of combining OGX-011 with 17-AAG in PC-3 tumors in vivo. Male nude mice bearing PC-3 xenografts were randomly assigned for treatment with OGX-011 + 17-AAG (n = 7) versus control ScrB + 17-AAG (n = 7). OGX-011 significantly enhanced the antitumor effect of 17-AAG in vivo, reducing mean tumor volume from 2,935.3 to 1,176.9 mm³ after 68 days (**, P ≤ 0.01) compared with control ScrB (Fig. 5A). Cancer-specific survival was significantly prolonged with combined OGX-011 + 17-AAG compared with controls (71.4% vs. 14.3% at day 72, respectively; *, P ≤ 0.05; Fig. 5B). Immunohistochemical analysis reveals decreased CLU, Ki67, and Akt expression after treatment with OGX-011 + 17-AAG compared with other groups (Fig. 5C). In addition, OGX-011 + 17-AAG–treated tumors had higher apoptotic rates, as shown by increased terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining, than other groups (Fig. 5C).

We next assessed the effects of combined treatment with OGX-011 and PF-04929113 in castrate-resistant LNCaP tumors. Mice bearing LNCaP tumors were castrated when PSA values exceeded 50 ng/mL. Once PSA levels relapsed above precastration levels, mice were randomly assigned to vehicle control, PF-04929113 alone, PF-04929113 + control ScrB, or PF-04929113 + OGX-011 (n = 10 in each group). Mice treated with OGX-011 + PF-04929113 had significant delays in tumor growth compared with all other groups (Fig. 6A). By 7 weeks posttreatment, all mice in the control had been euthanized; tumor volume in the OGX-011 + PF-04929113 group was 517.4 mm³ compared with 2,483.6 mm³ for PF-04929113 alone and 2,176.4 mm³ for PF-04929113 + control ScrB (***, P < 0.001; Fig. 6A)

Serum PSA levels were also significantly lower (∼4-fold) in the OGX-011 + PF-04929113 group than in other groups (***, P < 0.001; Fig. 6B). The combination OGX-011 + PF-04929113 group had a mean PSA level of 120 ng/mL after 42 days compared with 418.7 ng/mL in vehicle, 527 ng/mL in PF-04929113 alone, or 480.3 ng/mL in ScrB + PF-4929113 groups. The combination OGX-011 + PF-04929113 group had a significantly prolonged PSA doubling time (33.6 weeks; *, P <0.05) and decreased PSA velocity (13.78 ng/mL/wk; *, P < 0.05) compared with other groups (PSA doubling time: ∼2.4 weeks; velocity: ∼85 ng/mL/wk; Fig. 6C).

Overall survival was significantly prolonged in mice treated with combined OGX-011 + PF-04929113 (Fig. 6D). By day 57, all mice died or were euthanized because of high tumor burden in control, PF-04929113 alone, or control ScrB + PF-04929113 groups compared with the combined OGX-011 + PF-04929113 group, where all mice were still alive (P < 0.001) after 62 days. These data show that targeting CLU with OGX-011 potentiates the effects of PF-04929113 to significantly inhibit tumor growth and prolong survival in human CRPC xenograft models.

Consistent with in vitro findings, immunohistochemical analysis reveals decreased CLU, Ki67, Akt, and AR expression after treatment with combined OGX-011 + PF-04929113 compared with other groups (Fig. 7A). The immunostaining results were corroborated by Western blots (Fig. 7B). In addition, tumors treated with combination OGX-011 + PF-04929113 had higher apoptosis rates than other groups as shown by increased TUNEL staining (Fig. 7A). These data suggest that delays in tumor progression in OGX-011 + PF-04929113–treated mice result from both reduced proliferation rates and increased apoptosis rates.

Development of treatment resistance is a common feature of most malignancies and the underlying basis for most cancer deaths. Treatment resistance evolves, in part, from selective pressures of treatment that collectively increase the apoptotic rheostat of cancer cells. Survival proteins upregulated after treatment stress include antiapoptotic members of the Bcl-2 protein family, survivin, and molecular chaperones such as CLU and other HSPs (33).

Molecular chaperones help cells cope with stress-induced protein aggregation and play prominent roles in cell signaling and transcriptional regulatory networks. Chaperones act as genetic buffers, stabilizing the phenotype of various cells and organisms at times of environmental stress, and enhance Darwinian fitness of cells during cancer progression and treatment resistance (13). Heat shock chaperones are key components of the heat shock response, a highly conserved, stress-activated protective mechanism also associated with oncogenic transformation and thermotolerance (34). Chaperones are particularly important in regulating misfolded protein and endoplasmic reticular stress responses, an emerging area of interest in treatment stress and resistance. A growing enthusiasm for therapeutic modulation of this proteostasis network highlights Hsps and CLU as rational targets because of their multifunctional roles in signaling and transcriptional networks associated with cancer progression and treatment resistance. Cancer cells express higher levels of molecular chaperones and pirate the protective functions of HSF-1 to support their transformation (34). Indeed, inhibitors of Hsp90, Hsp70, Hsp27, or CLU have all been reported to induce cancer cell death and sensitize chemotherapy (28, 35).

Several Hsp90 inhibitors including PF-04928473 have potent antitumor activity in various preclinical models (28, 36, 37) and are in clinical trials (28, 38). Consistent with prior reports (28, 39), here we report that Hsp90 inhibitors induce a stress response with activation of the transcription factor HSF-1 and subsequent increased levels of Hsp90 itself, Hsp70, and CLU. This heat shock response likely enhance emergence of treatment resistance, as inhibition of transcription using actinomycin D attenuates 17-AAG–mediated Hsp70 and Hsp27 expression and potentiates the effect of 17-AAG in vitro (39). In addition, inhibition of the stress response by silencing HSF-1 increases the activity of Hsp90 inhibitors (40). In this study, we set out to evaluate the role of CLU in this heat shock response because CLU is dramatically induced by Hsp90 inhibitor treatment and CLU inhibitors are in late-stage clinical development.

CLU is associated with many varied pathophysiologic processes including reproduction, lipid transport, complement regulation, and apoptosis (17, 41). CLU expression is rapidly upregulated in various tissues undergoing apoptosis, including normal and malignant prostate and breast tissues, following hormone withdrawal (2, 42). Previous studies have also linked CLU expression with induction and progression of many cancers, including CRPC (17). Furthermore, CLU upregulation following androgen ablation in xenograft tumor models accelerates progression to castrate resistance and renders cells resistant to other apoptotic stimuli, including taxane chemotherapy (10, 43). Consistent with these accumulated findings (43), inhibition of CLU with OGX-011 synergistically enhances conventional and molecular targeted therapies in prostate cancer preclinical models (25). Indeed, OGX-011 is now in phase III trials, as phase II studies reported more than 90% inhibition of CLU in human prostate cancer tissues (44) and 7 months prolonged survival when OGX-011 is combined with docetaxel in CRPC (45, 46).

Here, we show that Hsp90 inhibitors increase CLU levels both in vitro and in vivo whereas OGX-011 inhibits PF-04928473 or 17-AAG induces CLU. As expected (39, 40), PF-04928473 or 17-AAG induces HSF-1 transcriptional activity leading to upregulation of HSP expression. Surprisingly, we found that CLU silencing abrogates, whereas CLU overexpression enhances, Hsp90 inhibitor–induced HSF-1 transcription activity, identifying a role for CLU in the regulation of HSF-1 and the heat shock response itself. CLU knockdown blocks the translocation to HSF-1 to the nucleus following treatment with Hsp90 inhibitors. This effect of CLU on HSF-1 activity is biologically relevant because CLU overexpression protects, whereas CLU silencing enhances, cytotoxicity of Hsp90 inhibitors. Consistent with these in vitro results, synergistic effects were also observed in vivo in PC-3 and LNCaP models when OGX-011 was combined with Hsp90 inhibitors. Combination OGX-011 plus Hsp90 inhibitor significantly delay CRPC tumor growth and prolonged survival in PC-3 and LNCaP models. Increased apoptotic rates with combined Hsp90 and CLU inhibition suggest that delayed tumor progression resulted from enhanced treatment-induced apoptosis. Collectively, these results highlight, for the first time, a biologically relevant feed-forward regulation loop of CLU on HSF-1 and the heat shock response.

In addition to the effects of CLU inhibition on the heat shock response, observations in the castrate-sensitive, AR-positive LNCaP model highlight another possible benefit of combined CLU and Hsp90 suppression involving AR activity. Hsp90 inhibition is known to destabilize and degrade the AR with decreased PSA expression (16, 47). In vivo, serum PSA levels, as well as PSA doubling time and velocity, were significantly reduced with combination OGX-011 therapy compared with PF-04929113 monotherapy. Serum PSA level is an established and useful AR-regulated biomarker (48) and a valuable tool in assessing efficacy of chemotherapy. Interestingly, at the low doses of Hsp90 inhibitor used in this in vivo study, no effect on serum PSA level was apparent. Lower PSA levels with combination therapy correlated with lower AR levels. This correlation between CLU inhibition and lower AR levels may involve the regulation loop of CLU on HSF-1 and the role of HSF-1 in regulating expression of other AR chaperones (e.g., Hsp27, Hsp70, Hsp90, FKBP5.2), and we are actively exploring the molecular basis in ongoing experiments. Although CLU is known to be transcriptionally activated by HSF-1 (17), in this study, we also show that CLU exerts a feed-forward loop that in turn activates HSF-1. CLU knockdown decreases HSF-1 transcriptional activity and abrogates its nuclear translocation, which subsequently leads to decreased Hsp27, Hsp70, and Hsp90 expression, similar to that observed after HSF-1 knockdown (49). Consequently, AR stability is reduced because of lowered chaperone levels.

In addition to synergistically enhancing antitumor activity, combination therapy may also allow dose reduction strategies to reduce toxicity that has been associated with Hsp90 inhibitors in clinical trials. For example, 17-AAG induced hepatotoxicity as monotherapy at 60 mg/kg/d (50) whereas PF-04929113 caused body weight loss at 50 mg/kg/d. In a previous study, 50 mg/kg PF-04929113 as monotherapy inhibited LNCaP CRPC tumor progression (28). At subtherapeutic doses of 25 mg/kg/d used in the present study, PF-04929113 monotherapy showed marginal, nonsignificant decreases in tumor volume and no effect on serum PSA levels; however, significant delays in tumor progression were seen at this lower dose when PF-04929113 was combined with OGX-011, with no toxicity observed.

In summary, this article helps define how stress induced by Hsp90 inhibitors regulates CLU by induction of HSF-1 activity and, in turn, how CLU regulates HSF-1 activity, cell survival, and treatment resistance. We showed, for the first time, that CLU inhibition abrogates the heat shock response–induced Hsp90 inhibitors. These observations are clinically relevant because CLU inhibitors are in phase III clinical trials and provide a framework for building new drug combinations based on mechanism-based interventions to overcome drug resistance. The present study supports for the first time the development of targeted strategies employing OGX-011 in combination with Hsp90 inhibitors to improve patient outcome in CRPC.

The University of British Columbia has submitted patent applications, listing M.E. Gleave as inventor, on the antisense sequence described in this article. This international patent has been licensed to OncoGenex Technologies, a Vancouver-based biotechnology company that M.E. Gleave has founding shares in. M.-J. Yin is an employee of Pfizer. The other authors disclosed no potential conflicts of interest.

The authors thank Gerrit Los for his support and discussion and Virginia Yago and Estelle Li for technical assistance.

This study was supported by Pfizer Worldwide Research & Development (United States), The Pacific Northwest Prostate SPORE (National Cancer Institute CA097186), L'Association pour la Recherche sur le Cancer (France), the Canadian Institutes of Health Research (F. Lamoureux), and the Terry Fox Prostate Cancer Program from the National Cancer Institute of Canada.

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.