• Open Access

Antitumor activity of NVP-AUY922, a novel heat shock protein 90 inhibitor, in human gastric cancer cells is mediated through proteasomal degradation of client proteins

Authors

  • Kyung-Hun Lee,

    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    Search for more papers by this author
    • These authors contributed equally as first author.

  • Ju-Hee Lee,

    1. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    Search for more papers by this author
    • Present address: Department of Cell Biology, Sloan-Kettering Institute, Memorial Sloan-Kettering Cancer Center, New York, New York, USA.

    • These authors contributed equally as first author.

  • Sae-Won Han,

    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    2. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    Search for more papers by this author
  • Seock-Ah Im,

    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    2. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    Search for more papers by this author
  • Tae-You Kim,

    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    2. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    3. Department of Molecular Medicine and Biopharmaceutical Sciences, WCU, Seoul National University, Seoul, Korea
    Search for more papers by this author
  • Do-Youn Oh,

    Corresponding author
    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    2. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    Search for more papers by this author
  • Yung-Jue Bang

    Corresponding author
    1. Department of Internal Medicine, Seoul National University Hospital, Seoul
    2. Cancer Research Institute, Seoul National University College of Medicine, Seoul
    3. Department of Molecular Medicine and Biopharmaceutical Sciences, WCU, Seoul National University, Seoul, Korea
    Search for more papers by this author

To whom correspondence should be addressed.
E-mail: bangyj@snu.ac.kr; ohdoyoun@snu.ac.kr

Abstract

Heat shock protein 90 (HSP90) is a molecular chaperone required for the stability of key regulators of cell survival and is an emerging target of cancer therapy. NVP-AUY922, a novel and potent inhibitor of HSP90, was evaluated against gastric cancer cell lines. NVP-AUY922 significantly inhibited the proliferation of all tested gastric cancer cell lines with 50% inhibitory concentration in the range of 2–40 nM and potently induced the degradation of growth factor receptors and other client proteins including HER-2, Akt and thymidylate synthase. HSP70 was induced by NVP-AUY922 and its binding with client proteins led to their proteasomal degradation. Moreover, the combination of NVP-AUY922 with cytotoxic chemotherapeutic agents such as 5-fluorouracil and oxaliplatin created a synergistic effect. Taken together, these preclinical data demonstrate the potent activity of NVP-AUY922 against gastric cancer cells and offer a rationale for clinical development of the agent alone or in combination with other chemotherapeutic drugs to effectively treat gastric cancer. (Cancer Sci 2011; 102: 1388–1395)

Heat shock protein 90 (HSP90) is an important molecular chaperone for protein folding, intracellular disposition and proteolytic turnover of key regulators of cell growth and survival.(1) Heat shock protein 90 is required for the stabilization of “client” proteins, including HER-2, Akt, Bcr-Abl and Raf,(2,3) and is considered to be an excellent target for anticancer therapy. Several HSP90 inhibitors have been introduced that simultaneously inhibit multiple signaling pathways: self-sufficiency in growth signals, insensitivity to anti-growth signals, evasion of apoptosis, limitless replicative potential, angiogenesis, and tissue invasion and metastasis.(4,5)

Gastric cancer is the second leading cause of cancer death worldwide and the prognosis of advanced gastric cancer is poor, with limited treatment options. Cytotoxic chemotherapeutic drugs including 5-fluorouracil (5-FU), platinum and taxane have been used in the treatment of gastric cancer. Often considered a chemosensitive cancer, gastric cancer responses are only partial in most sensitive cases and median survival is <1 year for patients with metastatic disease.(6) Novel agents and approaches are required for the treatment of gastric cancer and for overcoming resistance of chemotherapeutic agents. Several studies have reported new agents targeting HER-2, EGFR, VEGF or mTOR pathways for the treatment of gastric cancer.(7) Trastuzumab, a monoclonal antibody against HER-2, in combination with cytotoxic agents (e.g. fluoropyrimidine and cisplatin) showed a significant survival benefit, providing a median survival beyond 1 year in patients with HER-2-positive gastric cancers.(8)

Interestingly, HSP90 protein is abundantly expressed in gastric carcinoma, as well as other types of cancers. HSP90α protein is more highly expressed in gastric cancer tissue than its adjacent normal mucosa or in cases of gastritis.(9) Expression of heat shock proteins including HSP90 was associated with clinicopathological parameters of gastric cancer such as tumor size, nodal status, metastases, as well as patient survival.(10) Moreover, gastric cancer cells harbor multiple molecular targets that can be effectively disrupted by HSP90 inhibitors. Pharmacological inhibition of HSP90 led to a reduction in Akt, Erk, STAT3, HIF-1α, VEGF, EGFR and HER-2 expression, and significantly reduced angiogenesis and gastric cancer cell proliferation.(11) However, the effects of HSP90 inhibitors on gastric cancer have not been widely studied, despite their therapeutic potential.

We have previously shown that thymidylate synthase (TS), a key determinant of 5-FU resistance, interacts with HSP90. Furthermore, the inhibition of HSP90 with a histone deacetylase inhibitor enhances cytotoxicity of 5-FU-resistant cancer cells.(12) Moreover, treatment with histone deacetylase inhibitors induced a reduction in the binding of HSP90 to VEGFR1, VEGFR2 and Aurora-A kinase, followed by an increase in the binding of HSP70 to these client proteins, leading to their proteasomal degradation.(13,14)

NVP-AUY922 is a novel synthetic inhibitor of HSP90 that belongs to the family of resorcinylic isoxazoles. It is the most potent synthetic small-molecule inhibitor of HSP90 described to date and has shown preclinical activity against some tumor models.(15–17) In the present study, we analyzed the effect of NVP-AUY922 alone or in combination with other chemotherapeutic agents in human gastric cancer cell lines. We also investigated potential molecular mechanisms underlying the efficacy of NVP-AUY922.

Materials and Methods

Cell culture.  Human gastric cancer cells (SNU-1, -5, -16, -216, -484, -601, -620, -638, -668, -719 and NCI-N87) were purchased from the Korean Cell Line Bank (Seoul, Korea).(18) All cell lines were maintained in RPMI1640 media supplemented with 10% fetal bovine serum (FBS) (WELGENE Inc., Daegu, Korea) and gentamicin (10 μg/mL). All cell lines were incubated under standard culture conditions (20% O2 and 5% CO2 at 37°C).

BEAS-2B cells, non-cancer human bronchial epithelial cells, were maintained in LHC-9 medium (Invitrogen, Carlsbad, CA, USA) at 37°C in a humidified incubator containing 5% CO2.

Reagents.  NVP-AUY922 was kindly provided by Novartis (Basel, Switzerland). It was initially dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mM and aliquots were stored at −20°C. Working dilutions were freshly prepared.

The following monoclonal antibodies were used in our studies: anti-HSP90 and anti-HSP70 (Stress-Gen Biotechnologies Corp., Victoria, British Columbia, Canada), anti-p23 (Abcam Inc., Cambridge, MA, USA) and anti-TS (Neo Markers, Foremont, CA, USA). Polyclonal antibodies against Aurora-A, Aurora-B, PDGFR-α, VEGFR1, VEGFR2 and VEGFR3 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Polyclonal antibodies against EGFR, p-HER-2 (Tyr877), HER-2, p-AKT (Ser473) and Akt were purchased from Cell Signaling Technology (Beverley, MA, USA).

Oxaliplatin and docetaxel were purchased from Sanofi-Aventis Korea (Seoul, Korea) and 5-FU was purchased from Choongwae Pharma Corporation (Seoul, Korea). Proteasomal inhibitors (MG132, Lactacystin) and trypan blue solution (0.4%) were purchased from Sigma (St Louis, MO, USA). 17-AAG was purchased from Merck KGaA (Darmstadt, Germany).

Cell growth inhibition assay.  Cells (5–7 × 103 in 50 μL/well) were seeded in 96-well plates and incubated for 24 h at 37°C followed by drug treatment (i.e. NVP-AUY922, 5-FU, oxaliplatin or docetaxel) for 1–3 days at 37°C. After treatment, cells were incubated with 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl-tetrazolium bromide solution and analyzed by microplate reader (Molecular Devices, Sunnyvale, CA, USA).(12) In some experiments of sequential treatment, proliferating cancer cells were incubated in the presence of indicated doses of the first drug for 24 h, and fresh medium containing indicated doses of the second drug was subsequently added. The cells were then incubated for an additional 2 days. Graphs were generated by nonlinear regression of the data points to a four parameters logistic curve using SigmaPlot software (Statistical Package for the Social Sciences, Inc., Chicago, IL, USA).(12)

Combination index (CI) calculation.  The CI index was calculated according to the Chou-Talalay method.(19) Data were analyzed using the Calcusyn software (Biosoft, Ferguson, MO, USA). The CI index has been used for data analysis of two-drug combinations. CI < 1, CI = 1 and CI > 1 indicate synergism, addictive effect and antagonism, respectively.

Immunoblot analysis.  Cells were collected, washed with ice-cold phosphate-buffered saline (PBS), and suspended in an extraction buffer (20 mM Tris-Cl [pH 7.4], 100 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 5 mM MgCl2, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM pepstatin A, 0.1 mM antipain, 0.1 mM chymostatin, 0.2 mM leupeptin, 10 mg/mL aprotinin, 0.5 mg/mL soybean trypsin inhibitor and 1 mM benzamidine) on ice for 15 min. Lysates were cleared by centrifugation at 16000g for 20 min. Equal amounts of cell extracts were resolved on 10% SDS-polyacrylamide denaturing gels, transferred onto nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany), and probed with an appropriate primary antibody and horseradish peroxidase-conjugated secondary antibody. Detection was performed using an enhanced chemiluminescence system (Amersham Biosciences, Piscataway, NJ, USA).(20)

Immunoprecipitation.  Cells were collected and washed with PBS as described above. They were then suspended in an immunoprecipitation (IP) extraction buffer (50 mM Tris-Cl [pH 7.5], 250 mM NaCl, 0.1% NP40, 5 mM EDTA, 50 mM NaF, 0.1 mM NaVO4, 100 mM phenylmethylsulfonyl fluoride, 0.2 mM leupeptin, 10 μg/mL aprotinin, 0.1 mM pepstatin A and 0.1 mM antipain). Immunoprecipitation assays were performed and analyzed as previously described.(12)

Cell cycle analysis.  Cells (5 × 105 cells/6-cm plates) were seeded and treated with NVP-AUY922 alone or in combination with other chemotherapeutic agents for 24, 48 and 72 h. Cells were washed once with PBS and then resuspended in 70% ethanol for fixation. After treating with 100 μg/mL RNase at 37°C for 10 min, followed by treatment with 50 μg/mL propidium iodide, the cell cycle distribution was analyzed with a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA, USA).

Annexin-V assay.  Cells were treated with NVP-AUY922 for 72 h and then collected and stained with Annexin V-phycoerythrin and propidium iodide (Becton Dickinson). Apoptotic cell death was determined by the number of cells that stained positive for Annexin V-phycoerythrin and negative for propidium iodide using fluorescence-activated cell sorting analysis.

Statistical analyses.  Data are expressed as mean ± standard error (SE). Comparisons used Student’s t test or anova as appropriate. P values of <0.05 (two-sided) were considered significant.

Results

NVP-AUY922 has a growth inhibitory effect on gastric cancer cell lines.  To determine the sensitivity of gastric cancer cells to NVP-AUY922, we performed MTT assays with concentrations raging from 0 to 1 μM for 72 h in 11 human gastric cancer cell lines. The IC50 values of NVP-AUY922 fell in the nanomolar range of 2–40 nM in these gastric cancer cells (Fig. 1). The IC50 values for the cell lines NCI-N87 and SNU-216, cells with HER-2 amplification, were 3.23 nM and 11.99 nM, respectively. The nanomolar cytotoxicity of NVP-AUY922 has previously been reported in other types of cancer cells (e.g. multiple myeloma, non-small-cell lung cancer, breast cancer and glioblastoma).(15,21–23)

Figure 1.

 Growth-inhibitory effects of AUY922 in gastric cancer cells. The effects of NVP-AUY922 on the proliferation of various human gastric cancer cell lines were determined using the MTT assay. Cells were seeded into 96-well culture plates and treated with NVP-AUY922 for 3 days, followed by treatment with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide for 4 h. Cell viability was determined by measuring the absorbance. Each value represents the mean of three experiments; bars ± SD. *P < 0.5. **P < 0.01. IC50 values are as follows: 2.80 ± 0.08 nM (SNU-5); 3.03 ± 0.01 (SNU-719); 3.23 ± 0.03 (NCI-N87); 4.90 ± 0.12 (SNU-668); 5.56 ± 0.01 (SNU-638); 7.19 ± 1.47 (SNU-1); 10.28 ± 0.11 (SNU-601); 10.79 ± 0.18 (SNU-484); 11.99 ± 0.25 (SNU-216); 19.28 ± 0.41 (SNU-16); and 38.73 ± 0.20 (SNU-620).

Growth inhibitory activity of NVP-AUY922 was also compared with that of 17-AAG, a well established HSP90 inhibitor.(11) IC50 values for NCI-N87 cells were 3.48 nM for NVP-AUY922 and 11.36 nM for 17-AAG. Additionally, IC50 values for SNU-216 cells were 11.99 nM for NVP-AUY922 and 294.07 nM for 17-AAG. Moreover, to determine the different activity levels of NVP-AUY922 between cancer cells and normal cells, we performed identical experiments on the BEAS-2B cell line, which is a non-cancerous bronchial epithelial cell line. IC50 values for the BEAS-2B cells were 28.49 nM for NVP-AUY922 and 53.42 nM for 17-AAG.

NVP-AUY922 induced HSP70 protein and its interaction with HSP90.  To further investigate the effects of NVP-AUY922 on the HSP chaperone complex, we performed immunoprecipitation assays using an antibody for HSP90. Treatment with NVP-AUY922 did not affect the expression of HSP90, but HSP70 expression was increased by NVP-AUY922 treatment. As shown in Figure 2(a), the binding of HSP70 to HSP90 was increased by NVP-AUY922 treatment, whereas immunoprecipitation was conducted in the same input amount of HSP90. At the same time, treatment with NVP-AUY922 induced p23 dissociation from the HSP90 complex, which is consistent with a previous report.(16) p23 is required as a co-chaperone for the stabilization of the ATP-HSP90-client protein complex; thus, the dissociation of the p23-HSP90 complex reveals the action of NVP-AUY922 as a HSP90 inhibitor in gastric cancer cells. Taken together, these data suggest that treatment with NVP-AUY922 can recruit HSP70 to the HSP90 complex.

Figure 2.

 Effects of NVP-AUY922 on the heat shock protein (HSP) chaperone complex and expression of client proteins. (a) NVP-AUY922 increases the binding of HSP70 to HSP90. SNU-484 cells were treated with 100 nM NVP-AUY922 for the indicated times. HSP90 immunoprecipitates from cell lysates were analyzed by immunoblot with HSP90, HSP70 and p23 antibodies. IP, immunoprecipitation. Twenty micrograms of protein was used as an input control. (b) NVP-AUY922 decreases client protein expression in gastric cancer cells. For immunoblot analysis, cells were treated with the indicated concentrations of NVP-AUY922 for 24 h. PDGFR, platelet-derived growth factor; TS, thymidylate synthase; VEGFR, vascular endothelial growth factor receptor.

Decrease of client oncoproteins by inhibition of HSP90.  To determine the effects of NVP-AUY922 on client proteins of HSP90, we measured the levels of client proteins after treatment with NVP-AUY922 in gastric cancer cells. NVP-AUY922 treatment decreased expression of receptor tyrosine kinases, such as VEGFR1, 2, 3 and PDGFR-α. Moreover, Akt and phospho-Akt decreased in a dose-dependent manner (Fig. 2b). NCI-N87 and SNU-216 are gastric cancer cell lines with amplifications of HER-2, but SNU-484 is a HER-2-negative cell line.(24,25) NVP-AUY922 treatment resulted in decreased expression of HER-2 in NCI-N87 cells (Fig. 3a). These results suggest that NVP-AUY922 reduces expression of client proteins, but the sensitivity of NVP-AUY922 is not limited to HER-2-amplified cells.

Figure 3.

 NVP-AUY922 promotes proteasomal degradation of client proteins. (a) NVP-AUY922 promotes proteasomal degradation of human epidermal growth factor receptor 2 (HER-2). Samples for immunoblot analysis were prepared from NCI-N87 cells treated with 100 nM NVP-AUY922 and/or 1 μM lactacystin for 12 h. (b) NVP-AUY922 enhances the binding of heat shock protein 70 (HSP70) to HER-2. NCI-N87 cells were treated with 100 nM NVP-AUY922 and/or 1 μM lactacystin for 12 h. HER-2 immunoprecipitates from cell lysates were analyzed by immunoblot with antibodies to HSP90, HSP70, p23 or HER-2. IP, immunoprecipitation; VEGFR, vascular endothelial growth factor receptor.

Blocking HSP90 leads to proteasomal degradation of client proteins.  To investigate the mechanism involved in the downregulation of client proteins, we conducted immunoblot analyses after exposure to NVP-AUY922 with or without lactacystin, a proteasomal inhibitor. Decreased levels of HER-2 and VEGFR1 by NVP-AUY922 were recovered by treatment with lactacystin (Fig. 3a). This data reveals that NVP-AUY922 downregulates client proteins by a proteasome-dependent pathway. To examine the binding of the HSP chaperone complex with HER-2, we performed immunoprecipitation assays with HER-2 antibody after treatment with NVP-AUY922. Since NVP-AUY922 induced degradation of HER-2 protein, lactacystin was used to block proteasomal degradation, followed by immunoprecipitation. The binding of HSP70 with HER-2 was induced by NVP-AUY922 treatment, whereas binding of HSP90 was not induced (Fig. 3b). Taken together, NVP-AUY922 treatment seems to induce the binding of HSP70 to client proteins and targets the client proteins for degradation by the proteasome.

Effects of NVP-AUY922 on cell cycle progression and apoptosis are cell-line dependent.  NVP-AUY922 induced accumulation in sub-G1 and G1 phases in NCI-N87 cells in a dose-dependent manner (Fig. 4a). G2 phase cell cycle arrest was observed in SNU-216 cells with decreasing G1 and S cell cycle fraction. No significant change in cell cycle distribution was observed in SNU-5 cells, although SNU-5 cells were highly sensitive to NVP-AUY922 treatment with IC50 of 2.8 nM. These data collectively show that the anti-proliferative activity of NVP-AUY922 is not dependent on cell cycle distribution (Figs 1,4a). As shown in Figure 4(b), there were significant reductions of cell cycle regulating proteins (e.g. CDC2, cyclin B, cyclin D, CDK 4 and CDK6) in both NCI-N87 and SNU-216 cells. No change in the expression of CDK2 and CDC25C was observed. Marked reductions of cyclin E and cyclin A were shown in NCI-N87 cells. A potent induction of apoptosis by NVP-AUY922 in NCI-N87 cells was verified by Annexin-V assay (Fig. 4c). Cleavage of caspase-3, -7 and PARP in NCI-N87 cells confirmed the induction of the intrinsic apoptotic pathway by NVP-AUY922 (Fig. 4d). However, there was no cleavage of caspase or PARP in SNU-216 cells. Taken together, these data suggest that NVP-AUY922 can affect cell growth by altering multiple signaling pathways that are cell-line dependent.

Figure 4.

 Apoptosis and cell cycle arrest after exposure to NVP-AUY922. (a) Changes in cell cycle distribution after treatment with NVP-AUY922. Cells were treated with the indicated doses of NVP-AUY922 for 3 days and cell cycle distribution was analyzed by propidium iodide staining and flow cytometry. Graphs represent the results of cell cycle distribution in NCI-N87, SNU-216 and SNU-5. (b) Changes in cell cycle regulatory molecules after treatment with NVP-AUY922. Cells were treated with indicated doses of NVP-AUY922 for 3 days and cell lysates were analyzed by immunoblot with indicated antibodies. (c) Apoptosis is induced in NCI-N87 cells with NVP-AUY922. Cells were treated with the indicated doses of NVP-AUY922 for 3 days and the percentage of apoptotic cells was determined by Annexin-V staining and flow cytometry. (d) Cleavage of caspases and poly(ADP-ribose)-polymerase-1 (PARP) shows apoptosis in NCI-N87. After exposure to the indicated doses of NVP-AUY922 for 3 days, cell lysates were analyzed by immunoblot with the indicated antibodies.

NVP-AUY922 shows synergistic interaction with cytotoxic agents.  Next, we examined the effect of combined treatment with NVP-AUY922 and chemotherapeutic cytotoxic agents, which are commonly used to treat gastric cancer. Interestingly, the combined treatment with NVP-AUY922 and 5-FU increased sensitivity in 5-FU-resistant cell lines. SNU-216 (IC50 of 5-FU was 4.45 μM), NCI-N87 (>20 μM) and SNU484 (13.01 μM) showed CI values <1 with the combination of NVP-AUY922 and 5-FU, indicating a synergistic interaction (Table 1). IC50 values of 5-FU in these cell lines dropped to <1 μM in the presence of 10 nM of NVP-AUY922 (Fig. 5a). Likewise, CI values for NVP-AUY922 and oxaliplatin were <1, indicating synergism (Fig. 5b, Table 1). On the contrary, we could not find a synergistic relationship between NVP-AUY922 and docetaxel, as CI values were >1 (71.0 for NCI-N87, 1.02 for SNU-216 and 6.73 for SNU-484 at 1 nM NVP-AUY922 and 5 nM docetaxel).

Table 1.   Combination index (CI) values of concurrent treatment with NVP-AUY922 in gastric cancer cells
Cell line5-FU (1 μM) + AUY (1 nM)5-FU (1 μM) + AUY (10 nM)OXP (1 μM) + AUY (1 nM)OXP (1 μM) + AUY (5 nM)
  1. IC50 values for 5-fluorouracil (5-FU) or oxaliplatin (OXP) were obtained and the CI index was calculated according to the Chou-Talalay method using Calcusyn software. CI < 1, CI = 1 and CI > 1 indicate synergism, addictive effect and antagonism, respectively. AUY, NVP-AUY922.

SNU-2160.82 ± 0.090.30 ± 0.090.33 ± 0.010.46 ± 0.03
NCI-N870.87 ± 0.10.04 ± 0.050.76 ± 0.020.08 ± 0.1
Figure 5.

 Effects of combined treatment with NVP-AUY922 and cytotoxic agents. (a,b) Cells were treated with NVP-AUY922 and 5-fluorouracil (5-FU) (a) or oxaliplatin (b) for 3 days, followed by MTT assay. Each value represents the mean of three experiments; bars ± SD. (c) NVP-AUY922 promotes the proteasomal degradation of thymidylate synthase (TS). Samples for immunoblot were prepared from SNU-484 cells treated with NVP-AUY922 (100 nM) and/or proteasomal inhibitor at the indicated times.

Our previous report showed that thymidylate synthase (TS) is often overexpressed in gastric cancer cells and increased TS expression and activity are viewed as mechanistic drivers of 5-FU resistance.(12) SNU-484 cells show high basal expression of TS and is resistant to 5-FU.(12) Treatment with NVP-AUY922 decreased the expression of TS in a time- and dose-dependent manner (Figs 2b,5c). Treatment with the proteasome inhibitor, MG132, restored the expression of TS, which was decreased by NVP-AUY922. This suggests that proteasomal degradation plays a role in the decrease of TS by NVP-AUY922 (Fig. 5c).

Discussion

The present study shows the efficacy of NVP-AUY922, a novel HSP90 inhibitor, in gastric cancer cell lines. The small molecule NVP-AUY922 effectively binds to the HSP90 N-terminal domain and inhibits HSP90 activity.(15) Inhibition of HSP90 has emerged as a powerful therapeutic strategy that interacts specifically with a single molecular target and eventually causes the degradation and inactivation of numerous client proteins and pathways that play a central role in cancer cells.(4) Recent evidence suggested that heat shock proteins such as HSP90 are highly expressed in gastric cancers and play an important role for cancer survival.(9,10) Client proteins of HSP90, such as EGFR and HER-2, are effective targets of HSP90 inhibitors in gastric cancer cells and multiple angiogenic signaling pathways are impaired by the inhibition of HSP90.(11) These features suggest the possibility for HSP90 inhibitors to act as effective anti-cancer agents in gastric cancer, but this has not been widely studied so far. The present study demonstrates that targeting HSP90 with a novel inhibitor, NVP-AUY922, is an efficacious approach for gastric cancer therapy.

Our data show that NVP-AUY922 possesses a high cytotoxicity in gastric cancer cells (Fig. 1). The growth inhibition of the 11 gastric cancer cell lines occurred in the nanomolar range (2–40 nM range) of NVP-AUY922, and no cell line examined in the present study was resistant to a single treatment with this agent. This is comparable with the results of NVP-AUY922 in breast cancer cells, where six of the seven breast cancer cell lines were inhibited with IC50 values of 3.1–8.8 nM of NVP-AUY922.(16)

Although 17-AAG is the most widely studied HSP90 inhibitor in both preclinical and human clinical studies,(26) NVP-AUY922 has been suggested as a more potent HSP90 inhibitor, which can regulate the level of prostate-specific antigen, a novel biomarker of HSP90 inhibition.(27) The IC50 values of NVP-AUY922 were three- to 200-fold lower than those of 17-AAG in breast cancer cells.(16) As for gastric cancer, a previous study reported that treatment with 100 nM of 17-AAG for 48 h was effective, but only a modest gain in cytotoxic effects was observed with increasing doses of 17-AAG up to 10 μM.(11) In the present study, we compared the anti-proliferative effect of NVP-AUY922 with that of 17-AAG in gastric cancer cells. IC50 values for NCI-N87 cells were 3.48 nM for NVP-AUY922 and 11.36 nM for 17-AAG. Similarly, IC50 values for SNU-216 cells were 11.99 nM for NVP-AUY922 and 294.07 nM for 17-AAG. Moreover, in identical experiments on the non-cancerous BEAS-2B cells, IC50 values were 28.49 nM for NVP-AUY922 and 53.42 nM for 17-AAG. These data support a more potent and effective anti-proliferative effect of NVP-AUY922 over 17-AAG in both normal and cancer cell lines, and the anti-proliferative effect of NVP-AUY922 is more prominent in cancer cells compared with normal cells.

The antitumor effects of NVP-AUY922 in various xenograft models including breast cancer, melanoma, colon cancer, glioma, ovarian cancer and gastric cancer were previously reported.(15) In the case of gastric cancer, the xenograft model using GTL-16 was treated with 63 mg/kg NVP-AUY922 intravenously for 2 days out of a 7 day schedule. The tumor volume treated with NVP-AUY922 was reduced up to 63% when treated with NVP-AUY922 compared with that of vehicle. Moreover, when NVP-AUY922 was given at levels below the maximum tolerated dose, glioblastoma cell xenografts exhibited tumor regression and induced apoptosis, whereas 17-AAG used at maximum tolerated doses was less effective.(23)

In gastric cancer cells, NVP-AUY922 treatment clearly downregulated HSP90 client oncoproteins such as HER-2, VEGFR and Akt and induced the expression of HSP70 protein as well as the binding of HSP70 with client protein. The dissociation of p23 from HSP90 by NVP-AUY922 also provides evidence of the instability of client proteins (Fig. 2a).(16) Our results demonstrate that increased binding of HSP70 to client proteins leads these proteins to proteasomal degradation, as confirmed by co-treatment with proteasomal inhibitors (Figs 3,5c). This data suggests that NVP-AUY922 activates the proteasome degradation pathway and affects the stability of client proteins in agreement with previous studies of 17-AAG.(28,29)

Although NVP-AUY922 induced a reduction of client proteins in all cell lines tested, there were different effects of NVP-AUY922 on cell cycle distribution and apoptosis. Although IC50 values fell in a rather narrow spectrum in the nanomolar range, the cell cycle analysis revealed diverse features dependent on each cell line. Cells arrested were mainly in G2 (SNU-216) and G1 (NCI-N87). Apoptosis and cleavage of apoptosis-related molecules were observed only in NCI-N87 cells. Inhibition of HSP90 by 17-AAG arrested human breast cancer SKBr-3 and BT-474 cells in the G1 phase.(30) However, we found that NVP-AUY922 did not specifically induce G1 arrest.

The expression of cell cycle regulating proteins, G1 cell cycle regulators, CDK4, CDK6 and CDK2, were decreased in gastric cancer cell lines. G2/M cell cycle regulators, cdc2 and cdc25c, were downregulated by 17-AAG in human glioblastoma cell lines.(31) Thus, both G1 and G2 arrest are biochemically plausible with HSP90 inhibitors.

It is still not clear which factors contribute the different effects of NVP-AUY922 in each cell line. Our results suggest two features of the anti-tumor effects of NVP-AUY922. First, its effects are not solely attributable to the cell cycle effect, and second, the effects are not limited to cells harboring specific genetic aberrations.

Fluoropyrimidine (such as 5-FU, capecitabine and S-1) still remains one of the main anticancer agents used to treat gastric cancer, despite tumors that are initially sensitive to 5-FU eventually becoming resistant to this drug. Thus, agents that can abrogate and overcome resistance to 5-FU would have great clinical significance. Overexpression of TS protein is considered a main mechanism of resistance to various TS inhibitors, including 5-FU. Moreover, cancer cells exposed to TS inhibitors acutely upregulate TS, which might eventually lead to drug resistance.(32) The combined treatment with NVP-AUY922 showed significant improvement of cytotoxicity to 5-FU, accomplished by a reduction of TS protein by NVP-AUY922 treatment. This unique result has significant implications for gastric cancer therapy.

Several mechanistic explanations for the inhibition of TS are possible. HSP90 can directly induce proteasomal degradation of TS (Fig. 5c). This has not been reported in previous studies, and offers an explanation of the synergism between 5-FU and NVP-AUY922. Alternatively, HSP90 inhibition can induce hypophosphorylation and activation of tumor-suppressor retinoblastoma gene product (RB),(33) and activation of RB attenuates the expression of TS.(34)

We previously demonstrated that histone deacetylase inhibitors can reverse 5-FU resistance by downregulating TS.(12) In terms of inhibition of HSP90, histone deacetylase inhibitors induce acetylation of HSP90, while NVP-AUY922 blocks the ATP-binding pocket of HSP90. Although there are different mechanisms between histone deacetylase inhibitors and NVP-AUY922, the present study is in line with the previous one(12), showing that downregulation of TS by the inhibition of HSP90 chaperone is a rational and effective approach to overcome 5-FU resistance.

In gastric cancer cells, we found a synergistic effect between NVP-AUY922 and oxaliplatin, which was also shown in colon cancer cells in vivo.(35) Previous studies in breast cancer cells reported a synergistic effect between 17-AAG and taxane.(30) However, there was no synergistic effect between NVP-AUY922 and docetaxel in our gastric cancer cell lines. Possible explanations for these results would be the structural difference of 17-AAG and NVP-AUY922, in addition to the tissue-specific effect.(30) As observed in the present study, the effect of cell cycle arrest in G1 or G2 might abrogate the efficacy of taxane, which is cell cycle specific. Further investigation regarding the combination of HSP90 inhibitors and taxanes is warranted.

Expression of HER-2 and its downstream Akt/phospho-Akt was diminished by NVP-AUY922 treatment. This is an intriguing point because a subset of gastric cancer is considered to be driven by HER-2 signaling, making HER-2 an effective target for the treatment of these tumors.(8,24,25)

In conclusion, NVP-AUY922 shows a potent activity as a single agent in the low nanomolar range against gastric cancer cells, as well as in combination with cytotoxic drugs. We determined that the IC50 value of NVP-AUY922 on gastric cancer cells is lower than that of normal cells, and NVP-AUY922 is more potent in gastric cancer cells than 17-AAG. We also found that the client proteins are downregulated by a proteasomal degradation mechanism. While the TS downregulation by NVP-AUY922 confers a synergism with 5-FU, there was no synergism of NVP-AUY922 with docetaxel in contrast to other HSP90 inhibitors such as 17-AAG. Our preclinical data support the clinical development of NVP-AUY922 for effective treatment of gastric cancer.

Acknowledgments

This research was supported in part by grant No. R31-2008-000-10103-0 from the WCU project of the MEST and the NRF.

Disclosure Statement

The authors have no conflict of interest.

Ancillary