Combined inhibition of Wee1 and Hsp90 activates intrinsic apoptosis in cancer cells

Abstract

Heat shock protein 90 (Hsp90) is an essential, evolutionarily conserved molecular chaperone. Cancer cells rely on Hsp90 to chaperone mutated and/or activated oncoproteins, and its involvement in numerous signaling pathways makes it an attractive target for drug development. Surprisingly, however, the impact of Hsp90 inhibitors on cancer cells is frequently cytostatic in nature, and efforts to enhance the antitumor activity of Hsp90 inhibitors in the clinic remain a significant challenge. In agreement with previous data obtained using Wee1 siRNA, we show that dual pharmacologic inhibition of Wee1 tyrosine kinase and Hsp90 causes cancer cells to undergo apoptosis in vitro and in vivo. Gene expression profiling revealed that induction of the intrinsic apoptotic pathway by this drug combination coincided with transcriptional downregulation of Survivin and Wee1, an outcome not seen in cells treated separately with either agent. At the translational level, expression of these two proteins, as well as activated Akt, was completely abrogated. These data support the hypothesis that Wee1 inhibition sensitizes cancer cells to Hsp90 inhibitors; they establish combined Wee1/Hsp90 inhibition as a novel therapeutic strategy; and they provide a mechanistic rationale for enhancing the pro-apoptotic activity of Hsp90 inhibitors.

Introduction

Heat shock protein 90 (Hsp90) is an essential molecular chaperone that is utilized by cancer cells to protect a number of overexpressed or mutated oncoproteins from misfolding and degradation.^1-3 Several Hsp90 inhibitors have been evaluated in cancer clinical trials, and single-agent activity is seen in certain indications in which the tumor is driven by a highly Hsp90-dependent client protein (e.g., HER2-positive breast cancer or EML4-ALK-positive non-small cell lung cancer).⁴ However, in most cases, single-agent Hsp90 inhibitors have proven to be less efficacious than expected, given the central involvement of the chaperone in numerous signaling pathways whose activity is essential for cancer proliferation and survival.¹ Among possible causes contributing to this outcome is the fact that the cellular consequences of Hsp90 inhibition are frequently cytostasis and not cytotoxicity.⁵ Therefore, strategies to enhance tumor cell death in response to Hsp90 inhibitors are being actively sought.⁶

The tyrosine kinase Wee1 regulates the G₂/M cell cycle checkpoint and is an Hsp90 “client.”^7,8 Wee1 also phosphorylates a conserved tyrosine residue in the Hsp90 N-domain and alters chaperone activity to favor stabilization of a number of Hsp90-dependent kinases, including Wee1 itself.^9,10 Pharmacologic inhibition or molecular silencing of Wee1 has been reported to synergize with a number of DNA damaging agents.^11,12 We reported recently that similar treatment sensitizes prostate (PC3) or cervical (HeLa) cancer cells in vitro to the Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin (17-AAG).^9,10 Here, we expand these in vitro studies to include additional clinically evaluated Hsp90 inhibitors, and we show that Wee1 inhibition also sensitizes tumor xenografts to Hsp90 inhibition. Using microarray analysis, we identify several pathways that are uniquely sensitive to Wee1/Hsp90 inhibitor combination, and by examining additional cancer models, we show that activation of the intrinsic apoptotic pathway is primarily responsible for the enhanced apoptosis caused by this drug combination. These data provide a novel strategy to augment the apoptosis-inducing activity of Hsp90 inhibitors.

Results and Discussion

Wee1 tyrosine kinase phosphorylates and regulates Hsp90.^9,10 Both proteins are evolutionarily conserved in eukaryotes. Using yeast as a model organism, we found that deletion of Wee1 (or expression of non-phosphorylatable Hsp90) hypersensitized the cells to a structurally diverse panel of Hsp90 inhibitors, including the clinically evaluated drugs 17-AAG, SNX-2112 and STA-9090 (Ganetespib) (Fig. 1A). We observed similar results when we pharmacologically inhibited both Wee1 and Hsp90 in PC3 prostate carcinoma cells (Fig. 1B–F). Both the percentage of apoptotic cells and the abundance of the apoptotic markers cleaved caspase-3 and cleaved poly ADP-ribose polymerase (PARP) were significantly increased in dually treated cells, and this effect was abrogated by addition of the caspase inhibitor Z-VAD-fmk.

Figure 1. (A) Impact of Wee1 (Swe1) deletion in yeast and of non-phosphorylatable mutation of the Hsp90 Wee1 phosphorylation site (Y24 in yeast Hsp90 and Y38 in human Hsp90) on yeast sensitivity to Hsp90 inhibitors (radicicol, RD; geldanamycin, GA; 17-AAG; SNX-2112; STA-9090) is shown. Each row represents serially diluted yeast cultures treated with either 40 μM or 60 μM Hsp90 inhibitor. PC3 cells were treated as in Figure 1, with Wee1 inhibitor followed by the Hsp90 inhibitors SNX-2112 (B) or STA-9090 (D) as shown. At the end of the experiment, percent apoptosis was determined by counting the number of trypan blue-positive cells. PC3 cells were treated with Wee1 inhibitor (2.5 μM) followed by the Hsp90 inhibitors SNX2112 (C), STA-9090 (E) or 17-AAG (F), as in Figure 1, and total proteins were extracted and immunoblotted for cleaved PARP and cleaved caspase-3. As indicated, cells in (F) were treated with the caspase inhibitor Z-VAD-fmk (10 μM) for 1 h prior to other treatments. Actin or tubulin are shown as loading controls.

Although pro-apoptotic in combination, at the concentrations used here, neither Wee1 inhibitor nor Hsp90 inhibitor caused significant apoptosis when administered as single agents (Fig. 1B and D). In order to explore the mechanism underlying the synergistic activity of this drug combination, we performed microarray analysis using PC3 cells treated with Wee1 inhibitor alone, 17-AAG alone or the two drugs in combination (see Materials and Methods). Principal component analysis (PCA) of our data revealed a unique gene expression signature for individual and combination treatments (Fig. S1). Although single-agent treatment affected a number of overlapping genes, combination treatment caused a greater than 2-fold change in more than 1,300 unique genes (Fig. S2). Upregulated genes include those associated with cell morphology, amino acid metabolism and the stress response, while downregulated genes include those associated with the cell cycle and the DNA damage response (Fig. 2A).

Figure 2. Identification of genes in prostate cancer cells that respond uniquely to combined Wee1 and Hsp90 inhibition. (A) Expression profile of PC3 cells treated with Wee1 inhibitor (2.5 µM) for 24 h followed by 17-AAG (40 nM) for an additional 48 h. Single-agent treatments are shown for comparison; mRNA from each sample was subjected to microarray analysis. (B) Heat maps showing Survivin and Wee1 gene expression following single agent and combined treatments are shown. (C) Changes in Survivin and Wee1 expression shown by microarray analysis were confirmed by RT-qPCR. Error bars were generated based on a confidence interval of 95%. (D) Western blot analysis of Survivin, Wee1, Akt and phospho-Ser473 Akt proteins in PC3 cells treated as above. Hsp90 is shown as loading control.

We identified Wee1 and Survivin among apoptosis-inhibiting genes that were uniquely downregulated by the combination treatment (Fig. 2B; Fig. S3). These results were confirmed by RT-qPCR (Fig. 2C). Finally, we demonstrated that Wee1 and Survivin protein expression was completely abolished in cells treated with the combination of Wee1 and Hsp90 inhibitors (Fig. 2D). Although both Survivin and Wee1 are Hsp90 clients, our data show that the Hsp90 inhibitor concentration we chose had no impact on protein expression. In contrast, expression of both Wee1 and Survivin was significantly affected at the transcriptional and translational level following dual inhibition of Wee1 and Hsp90. While Akt was completely inactivated in cells exposed to the drug combination, this effect was mediated posttranslationally since Akt transcription and total protein expression were unaffected or minimally affected (Fig. 2C and D).

Wee1 phosphorylates Cdc2 and inhibits its pro-apoptotic activity. Elevated Survivin expression prevents caspase-3-mediated Wee1 degradation, establishing a link between Wee1, Survivin and resistance to apoptosis.¹³ Further, activated Akt has been reported to indirectly enhance the transcription and translation of Survivin.^14,15 Thus, the enhanced caspase-3 activation caused by dual inhibition of Wee1/Hsp90 is consistent with the downregulation of both Survivin/Wee1 and activated Akt/Survivin signaling pathways.

Because of the dramatic impact of this drug combination on caspase-3 activation, we examined its effect on cytochrome c release from mitochondria, since cytochrome c release into the cytoplasm plays a major role in caspase-3 cleavage.¹⁶ Total PC3 cell lysate was fractionated into mitochondrial and cytosolic fractions (see Materials and Methods), and cytochrome c was visualized by western blotting (Fig. 3A). Consistent with our previous data, cytoplasmic cytochrome c levels were markedly increased in cells treated with both Wee1 and Hsp90 inhibitors.

Figure 3. Wee1 inhibitor/17-AAG combination stimulates intrinsic apoptosis. (A) PC3 cells were treated with Wee1 inhibitor and 17-AAG (as in Fig. 1), and mitochondrial were isolated. Cytosolic and mitochondria fractions were subjected to SDS-PAGE and immunoblotted for cytochrome c. Tom70 and β-actin were used to verify the purity of mitochondrial and cytosolic fractions, respectively. (B) Jurkat cells stably overexpressing Bcl-2 or Bcl-xL, and (C) Jurkat cells lacking expression of caspase-8 or FADD were treated with Wee1 inhibitor and 17-AAG (as above). Cleaved caspase-3 and cleaved PARP were visualized by immunoblotting. Jurkat cells stably expressing empty vector were used as control. Tubulin is shown as loading control.

Cytochrome c release from mitochondria is a consequence of activating both the intrinsic and extrinsic apoptotic pathways.¹⁷ Therefore, we further examined the contribution of each pathway to the apoptosis induced by the Wee1 inhibitor/Hsp90 inhibitor drug combination. Jurkat cells stably overexpressing the anti-apoptotic mitochondria-associated proteins Bcl-2 or Bcl-xL were fully resistant to caspase-3 and PARP cleavage caused by combined Wee1/Hsp90 inhibition (Fig. 3B). In contrast, Jurkat cells stably lacking expression of either caspase-8 or FADD, two proximal effectors of the extrinsic apoptotic pathway, retained full sensitivity to the drug combination (Fig. 3C). We conclude from these data that combined inhibition of Wee1 and Hsp90 primarily engages the intrinsic apoptotic pathway to increase cell death.

Based on these in vitro observations, we examined whether Wee1 inhibition enhanced the in vivo antitumor activity of Hsp90 inhibitor. Athymic mice bearing established (100–200 mm³ at treatment initiation) subcutaneous PC3 xenografts received intraperitoneal injections of either Wee1 inhibitor (5 mg/kg) or the Hsp90 inhibitor 17-AAG (40 mg/kg) alone or a combination of both drugs three times per week. Mice were sacrificed when tumors reached 2 cm in size or became necrotic. While 17-AAG has demonstrated significant single-agent activity toward PC3 xenografts at higher dose levels,¹⁸ for this study we purposefully chose a dose level and schedule without single-agent activity (Fig. 4A). After performing a pilot study with Wee1 inhibitor (data not shown), we selected a dose that inhibited Wee1 in xenograft tumor tissue (data not shown), was well-tolerated by the mice, but also had no single-agent antitumor activity (Fig. 4A). However, when the two drugs were administered in combination, we observed significant inhibition of tumor growth (80% inhibition at day 21, p < 0.001) and marked prolongation of survival (Fig. 4A and B).

Figure 4. Effect of Wee1/Hsp90 inhibitor combination on PC3 xenograft growth and host survival. PC3 tumor-bearing athymic mice were randomly assigned to indicated treatment groups (ten mice per group). Mice were sacrificed when tumor size reached 2 cm (control and single treatment arms), or if tumors became necrotic (combined treatment arm). (A) Tumor size was measured with calipers following therapy initiation. Statistics were obtained using 2-way ANOVA analysis; we observed 80% inhibition of tumor growth on day 21 in mice receiving combined treatment compared with other groups (***p < 0.001). (B) Kaplan-Meier survival curves for mice in each treatment arm. Mice in the combined treatment group were sacrificed when tumors became necrotic (day 33). (C) Tumors were removed at time of sacrifice and were analyzed by immunocytochemistry for TUNEL, cleaved caspase-3, Survivin, Wee1, Akt and phospho-Akt (pSer473). Representative images and mean positive cell number/high power field are shown. Statistics were obtained using Student’s t-test (**p < 0.01; ***p < 0.001). Error bars represent SEM.

After sacrificing the mice on either day 21 (due to excessive tumor size, control and single-agent treatment arms) or on day 33 (due to tumor necrosis, combined treatment arm), tumors were excised and processed for immunohistochemistry (IHC). Representative IHC images and staining quantification are shown in Figure 4C. Intratumoral apoptosis was monitored by assessing DNA fragmentation (TUNEL) and cleaved caspase-3. Expression of Survivin, Wee1, total Akt and phospho-Akt was also assessed. In each case, dual inhibition of Wee1 and Hsp90 was significantly more effective than single-agent treatment. TUNEL and cleaved caspase-3 staining were increased more than 3-fold compared with other treatment groups, while xenograft expression of Survivin, Wee1 and phospho-Akt was reduced by more than 90% in the combination treatment group compared with the other groups. Similarly, H&E staining of tumor tissue from each treatment group revealed significant areas of necrosis only in dually treated mice (Fig. S4).

These data demonstrate that combined inhibition of Wee1 and Hsp90 activates the intrinsic apoptotic pathway in cancer cells both in vitro and in vivo, likely as a consequence, at least in part, of suppressing Wee1 and Survivin expression, and inhibiting Akt activation. Survivin is considered to be a global target of intrinsic tumor suppression networks, contributes to chemoresistance and is almost uniformly highly expressed in cancer compared with non-embryologic normal tissues.^19,20 Survivin has been validated as a therapeutic target in cancer, and small molecules that inhibit Survivin transcription show activity in some preclinical cancer models, including prostate cancer.^21,22 However, initial clinical evaluation of the Survivin inhibitor YM155 reported only modest single-agent activity in patients with refractory, advanced non-small cell lung cancer, suggesting that combination with other chemotherapy or targeted agents may be clinically warranted.²³ In addition, more robust suppression of Survivin expression, perhaps as a consequence of targeting several pathways that impact Survivin, as we have shown here, may be beneficial.

The Wee1 inhibitor MK-1775, combined with paclitaxel and carboplatin, is currently under clinical evaluation in ovarian cancer (www.clinicaltrials.gov/ct2/show/NCT01357161). Preclinical studies have shown that MK-1775 enhanced the antitumor efficacy of DNA-damaging drugs including doxorubicin, camptothecin, pemetrexed and gemcitabine.^11,12 Similarly, Wee1 inhibition, or its molecular silencing, sensitized breast cancer cells to TRAIL-induced apoptosis.^24,25 Unlike these reports, however, our current data establish synergy between Wee1 inhibition and inhibition of a non-DNA damage/non-apoptosis inducing molecular target, Hsp90. Indeed, our microarray data predicts that this drug combination may sensitize cancer cells to DNA damaging agents, since DNA damage response genes comprise a significant proportion of targets uniquely downregulated by this strategy.

Unexpectedly, we found that combined inhibition of Wee1 and Hsp90 results in the transcriptional suppression of both Wee1 and Survivin, providing a novel strategy to attack this key cell cycle checkpoint/anti-apoptotic signaling pathway. Connectivity pathway mapping (Fig. S3) extends the potential impact of this strategy to a network of apoptosis-related genes and suggests that further molecular analysis of the consequences of Wee1/Hsp90 inhibition may uncover additional drug targets and drug combinations.

Materials and Methods

Yeast strains and growth media

The wild-type and Y24F mutant yeast strains used in this study were derived from pp30 (MAT a, trp1–289, leu2–3,112, his3–200, ura3–52, ade2–101, lys2–801, hsc82KANMX4, hsp82KANMX4) and previously reported.⁹

Yeast was grown on YPD [2% (wt/vol) Bacto peptone, 1% yeast extract, 2% glucose, 20 mg/liter adenine]. Yeast cultures were diluted to an optical density at 600 nm of 0.5, and 5~µl aliquots of a 10-fold dilution series were spotted onto YPD–2.0% agar plates supplemented with the indicated level of Hsp90 inhibitors. Growth was monitored over 3–5 d at 25°C.

Flow cytometric analysis (FACS analysis)

Apoptosis was monitored by FACS analysis. Detailed methodology can be found in Supplemental Material.

Cell culture

The human prostate cancer cell line, PC3, was obtained from American Type Culture Collection. The human T-cell lymphoma cell line Jurkat, and its stably transfected and knocked down variants were kind gifts from J. D. Robertson (University of Kansas Medical Center). Please see Supplemental Material for additional details.

Western blotting and co-immunoprecipitation

Total protein extracts were prepared and analyzed as previously described.⁹ Additional methods are detailed in Supplemental Material.

Wee1 inhibitor synthesis

Wee1 inhibitor was prepared as reported previously.²⁶ Additional information can be found in Supplemental Material.

Microarray gene expression profiling

Total RNA was isolated with the RNAeasy Kit (Qiagen). RNA quality was checked on an Agilent Bioanalyzer. Please see Supplemental Material for additional details.

Tumor xenograft studies

In vivo experiments were conducted as previously described.¹⁸ Detailed methodology is presented in Supplemental Material. All animal experiments were performed in accordance with an animal protocol that was approved by the NIH Animal Care Committee.

Immunohistochemistry

Resected tumors were fixed in fresh 4% paraformaldehyde overnight and transferred to 70% ethanol prior to paraffin embedding and hematoxylin and eosin staining (American HistoLabs, Inc.). Immunohistochemistry was performed as previously described.²⁷ Detailed staining techniques and method used to quantify positive signals are presented in Supplemental Material.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.