SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Hepatocellular carcinoma (HCC) remains associated with a poor prognosis, but novel targeted therapies in combination with anti-angiogenic substances may offer new perspectives. We hypothesized that simultaneous targeting of tumor cells, endothelial cells, and pericytes would reduce growth and angiogenesis of HCC, which represents a highly vascularized tumor entity. Recently, because of their anti-angiogenic properties, inhibitors of mammalian target of rapamycin (mTOR) have entered clinical trials for therapy of HCC. However, treatment with mTOR inhibitors may lead to paradoxical activation of Akt signaling in tumor cells via insulin-like growth factor-I receptor (IGF-IR)–dependent and IGF-IR–independent mechanisms. Because we have recently identified heat shock protein 90 (Hsp90) antagonists to impair both oncogenic and angiogenic signaling cascades in tumor cells, including Akt and IGF-IR, we sought to investigate whether Hsp90 blockade could improve growth-inhibitory and anti-angiogenic effects of the mTOR inhibitor rapamycin. Human HCC cells, a murine hepatoma cell line, endothelial cells (ECs), and vascular smooth muscle cells (VSMC) were employed in experiments. Results show that dual inhibition of mTOR and Hsp90 leads to effective disruption of oncogenic signaling cascades and substantially improves growth-inhibitory effects in vivo. Importantly, blocking Hsp90 abrogated the rapamycin-induced activation of Akt and of the downstream effector nuclear factor kappa-B (NF-κB) in HCC tumors. Furthermore, Hsp90 inhibition reduced the expression of platelet-derived growth factor-receptor-β (PDGF-Rβ) on VSMCs, and diminished vascular endothelial growth factor-receptor 2 (VEGFR-2) expression on ECs, which further improves the anti-angiogenic capacity of this regimen. Conclusion: Blocking Hsp90 disrupts rapamycin-induced activation of alternative signaling pathways in HCCs and substantially improves the growth-inhibitory effects of mTOR inhibition in vivo. Hence, the concept of targeting tumor cells, ECs, and VSMCs by blocking Hsp90/mTOR could prove valuable for treatment of HCC. (HEPATOLOGY 2008.)

Hepatocellular cancer (HCC) is one of the most common solid tumor entities, rated fifth in incidence and third in mortality worldwide.1 To date, HCC is refractory to conventional chemotherapeutics, and thus surgical resection, or liver transplantation, remain the only potential curative treatment options. Therefore, new and effective strategies for treatment of HCCs are needed.

Within the last decade, novel therapeutic approaches have been developed for a variety of cancer entities. Among these, inhibition of mammalian target of rapamycin (mTOR) has proved anti-neoplastic efficacy in various preclinical tumor models and is being tested in trials (NCT00390195, NCT00494091; www.clinicaltrials.gov).2-5 The anti-neoplastic effect of rapamycin is mediated in part by anti-angiogenic effects via endothelial growth factor A (VEGF-A) signaling inhibition in endothelial cells (ECs).3 This is significant because hepatocellular cancer is characterized by hypervascularization.6 Additionally, mTOR is activated in a subset of HCC and is associated with poor prognosis, thereby representing a potential target for therapy.7 However, recent studies suggest that blocking mTOR can paradoxically activate other signaling pathways in tumor cells, revealing a potential counteractive effect to rapamycin in cancer therapy.2, 8 In particular, mTOR blockade by rapamycin involves mTOR complex 1 (mTORC1), composed of Raptor, mTOR, and mLST8, whereas mTOR complex 2 (mTORC2; composed of Rictor, mTOR and mLST8) is not affected.9 However, disrupting mTORC1 reduces the inhibitory effects of S6K on insulin receptor substrate-1 (IRS-1), which is an effector of phosphatidylinositol-3 kinase (PI-3K), and thus causing AktThr308 phosphorylation.9, 10 Moreover, mTORC1 blockade shifts activation toward the mTORC2 complex (also known as PDK2), which is important for Akt phosphorylation at Ser473, hence further enhancing Akt activity.8 Therefore, the obvious concern is that mTOR-targeting therapy increases Akt activity, which in turn is associated with poor prognosis in different cancer entities.11

Another promising target for therapy is heat-shock protein 90 (Hsp90), which is an essential chaperon for function and integrity of a wide range of oncogenic client proteins.12 Hsp90 is twofold to 10-fold overexpressed in tumors, compared with normal tissues.13 With respect to HCC, Hsp90 overexpression has been described and is associated with a poor prognosis.14 Importantly, oncogenic transcription factors and signaling intermediates have been identified as Hsp90 client proteins, including hypoxia-inducible-factor-1α (HIF-1α), signal transducer and activator of transcription-3 (STAT3), intracellular kinases (Akt, Erk), and growth-factor receptors [epidermal growth factor receptor, insulin-like growth factor receptor (IGF-R, EGF-R)].15-17 Interestingly, these Hsp90 clients have been implicated in hepatocarcinogenesis and are generally overexpressed in tumors.18 However, the value of targeting Hsp90 for HCC therapy is not known.

In the current study, we hypothesized that blockade of Hsp90 would improve anti-tumoral and anti-angiogenic effects of rapamycin at least in part by preventing paradoxical mTOR-inhibitor–induced activation of Akt, in addition to blocking oncogenic signaling molecules, in tumor cells. We provide evidence that concomitant targeting of tumor cells, ECs, and vascular smooth muscle cells (VSMCs) by combined Hsp90/mTOR blockade is a novel strategy for HCC therapy.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Cell Culture.

Human HCC lines HepG2 and Huh-7, and ECs and VSMCs, were purchased from American Type Culture Collection (Manassas, VA). Hep3B cells were provided by Dr. Claus Hellerbrand (University of Regensburg, Germany), and murine Hepa129 cells from Dr. Volker Schmitz (University of Bonn, Germany). HepG2 and Huh-7 were cultured in Dulbecco's modified Eagle's medium (DMEM) (Gibco, Karlsruhe, Germany) supplemented with 10% fetal bovine serum, Hep3B, and Hepa129 were maintained in Roswell Park Memorial Institute (Gibco). The Hsp90 inhibitor 17-(dimethylaminoethylamino)-17-demethoxygeledanamycin was purchased (17-DMAG; Invivogen Cayla-Invivogen, Toulouse, France) and dissolved in water. Rapamycin (Wyeth, Madison, NJ) was obtained from our hospital pharmacy and dissolved in water for in vivo use. Recombinant hepatocyte growth factor (HGF), IGF-I, VEGF-A, and platelet-derived growth factor (PDGF)-B were purchased from R&D Systems (Wiesbaden, Germany). In vitro results shown are from HepG2 cell experiments (otherwise indicated), but results were confirmed in Huh-7 and Hep3B cells.

Methylthiazole Tetrazolium Assay.

To evaluate effects of 17-DMAG or rapamycin, cells (0.5-1 × 103/well) were seeded into 96-well plates and exposed to various drug concentrations for 24 and 48 hours. We used the methylthiazole tetrazolium assay to assess cell viability.5

Cell Death Detection.

The Cell Death Detection ELISAPlus kit (Roche, Germany) was used, as described,19 to measure DNA fragmentation. Tumor cells (4 × 103 cells/well) were seeded into 96-well plates. After 24 hours, cells were treated and grown for an additional 48 hours. Cytoplasmic fractions of control and treated cells were used in the assay. Specific enrichment of mononucleosomes and oligonucleosomes released into the cytoplasm of the treated cells was calculated by: absorbance (dying/dead cells)/absorbance of corresponding negative control.

Fluorescence-activated cell sorting analyses for cell death were performed as previously described.20 In brief, cells were grown in the presence of vehicle, rapamycin (10 ng/mL), 17-DMAG (100 nM), or a combination of both. After 24 and 48 hours, cells were trypsinized, labeled with propidium iodide (PI)/annexin V, and analyzed by flow cytometry to determine apoptotic rates.

Cell Motility.

Migration assays were performed using modified Boyden chambers, as described.15 Briefly, 5 × 104 cells were resuspended in 1% fetal bovine serum–Dulbecco's modified eagle's medium and seeded into 8-μm filter pore inserts (Becton Dickinson Bioscience, Heidelberg, Germany). Either 10% fetal bovine serum or HGF (50 ng/mL) was used for chemoattraction. Conditioned media from HepG2 or Huh-7 was used for experiments with VSMCs and ECs. Cells were allowed to migrate for either 4 (VSMCs, ECs) or 24 hours (tumor cells). Migrated cells were Wright-Giemsa stained and counted in four random fields.

Western Blotting.

Whole cell lysates, nuclear extracts, and lysates from subcutaneous tumors were prepared as described.15 Protein samples (50 μg) were subjected to western blotting on a denaturating 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis. Membranes were probed with antibodies against phospho-AktSer473, phospho-AktThr308, Akt, phospho- glycogen synthase kinase (GSK) 3Ser21/9, phospho-ErkThr202/Tyr204, Erk, phospho-STAT3Tyr705, STAT3, STAT5, phospho-FAKTyr925, FAK (Cell Signaling Technologies, Beverly, MA), and VEGF-A, VEGFR-2, PDGF-Rβ and β-actin (Santa Cruz Biotechnologies, Santa Cruz, CA); antibodies to HIF-1α (Novus Biologicals, Merck, Darmstadt, Germany) and to Hsp90 were also used (Stressgen, Ann Arbor, MI). Antibodies were detected by enhanced chemiluminescence (Amersham Bioscience, Piscataway, NJ).

VEGF-A Measurement.

Total RNA was isolated using Trizol reagent, purified by ethanol precipitation, and 1-μg aliquots of RNA were reverse transcribed. VEGF165-A primers were: (5′-GCACCCATGGCAGAAGGAGGAG; 3′-AGCCCCCGCATCGCATCAG). Reverse transcription polymerase chain reaction was performed using the LightCycler system and Roche Fast-Start Light Cycler-Master Hybridization Probes master mix (Roche Diagnostics, Mannheim, Germany).

We used an enzyme-linked immunosorbent assay (ELISA) kit (BioSource Europe, Nivelles, Belgium) to measure human VEGF-A secretion by tumor cells. HCC cells were plated at 40% to 50% density and incubated with or without rapamycin, 17-DMAG, or the combination of both substances under either nonhypoxic (20% O2), or hypoxic (1% O2) conditions. VEGF-A was measured in culture supernatants.

AktSer473 Phosphorylation.

An enzyme immunometric assay (BD Bioscience) was used to quantify phosphorylated AktSer473 as described.8 HCC cells or ECs were incubated for 24 hours with or without rapamycin, 17-DMAG, or the combination of both substances. Analyses of cell lysates were performed according to the manufacturer's protocol.

Subcutaneous Tumor Models.

Huh-7 cells (1 × 106) were subcutaneously injected into nude mice (Balb-c nu/nu, Charles River, Sulzfeld, Germany). Mice were randomized and assigned to treatment groups and intraperitoneally injected daily (100 μL) with diluent, 17-DMAG (10 mg/kg), rapamycin (0.1 mg/kg), or 17-DMAG + rapamycin, starting on day 4 after tumor cell implantation. Tumor diameters were measured every other day and volumes calculated (width2 × length × 0.5). The experiment was terminated on day 16 and tumors harvested. Experiments were approved and performed according to regional authorities.

Nuclear Factor Kappa B DNA Binding Activity.

Nuclear factor kappaB (NF-κB) activity was evaluated in subcutaneous tumors using a p65 DNA binding activity assay, as described.21 Briefly, nuclear extracts were prepared from frozen tumor sections. DNA binding activity was tested using the NF-κB TransFactor Kit (Becton Dickinson Bioscience).

Orthotopic Tumor Models.

An orthotopic syngenic tumor model was used as described.22 Briefly, 5 × 104 Hepa129 cells were injected into the liver lobe of C3H mice. Treatment was initiated on day 4 after tumor cell implantation, with mice receiving: (1) 17-DMAG 25 mg/kg, 3×/wk, (2) rapamycin 0.2 mg/kg/day, (3) combination therapy, or (4) diluent. Treatment was continued for 10 days. Mice were observed daily and sacrificed when tumor-related symptoms occurred. Tumors were measured and the incidence determined at the endpoint (day 14).

Immunohistochemistry.

CD-31–positive vessel area was measured and assessed as described previously.15 Rat anti-mouse CD31/PECAM-1 antibody (Pharmingen, San Diego, CA) and peroxidase-conjugated goat anti-rat immunoglobulin G (Jackson Research Laboratories, West Grove, PA) were applied. To determine amounts of proliferating tumor cells, mice received intraperitoneal injections of bromodeoxyuridine (BrdU) (1.0 mg/mouse; Sigma) 2 hours before termination of studies. A commercially available BrdU detection kit (Becton Dickinson) was used to visualize BrdU uptake of proliferating cells in sections. BrdU-positive tumor cells were counted in four fields at 20× magnification and averages calculated.15

A terminal deoxynucleotidyl transferase-mediated nick-end labeling detection kit (TUNEL; Promega Corp., Mannheim, Germany) was used to detect cell apoptosis.19 Four fields at 40× magnification were selected at the proliferation front in each tumor, and TUNEL-positive cells were counted.

Statistics.

Statistics were performed using SigmaStat (Version 3.0). Results of in vivo experiments were analyzed for significant outliers using Grubb's test (www.graphpad.com). Tumor-associated variables in in vivo experiments were tested for significance using the Mann-Whitney U test. The two-sided Student t test was applied for analysis of in vitro data. All results are expressed as the mean ± standard error of the mean (SEM).

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Impact of Combining Hsp90 and mTOR Inhibitors on HCC Cells.

MTOR inhibitors have shown anti-neoplastic potential in preclinical models, including HCC.3, 5, 7, 23 The rationale for combining an mTOR inhibitor with an Hsp90 antagonist comes from studies demonstrating that targeting mTOR elicits potent anti-angiogenic effects through inhibition of VEGF-signaling in ECs, whereas blocking Hsp90 predominantly affects tumor cells.2, 13, 24 We first determined that HCC tumor cells express higher amounts of Hsp90 protein than do ECs and VSMCs (data not shown). Next, methylthiazole tetrazolium analysis shows that combination of rapamycin with 17-DMAG leads to improved growth inhibitory effects in HCC cells (Supporting Fig. 1). Moreover, a cell death detection ELISA shows that the apoptotic tumor cell rate (HepG2) is markedly increased after 48 hours by combination therapy (Fig. 1A). Similar results were obtained by fluorescence-activated cell sorting analysis for PI/annexin V (data not shown). In addition, combination therapy also increased the apoptotic rate of ECs, as compared with either substance alone (Fig. 1B). However, these additional effects were observed to a lesser extent in VSMCs. These experiments suggest that, in addition to proposed anti-angiogenic effects of rapamycin, combination with 17-DMAG results in substantially improved direct anti-tumoral/anti-angiogenic effects of these drugs.

thumbnail image

Figure 1. Impact of simultaneous Hsp90 blockade and mTOR inhibition on apoptosis and signaling. (A) Apoptosis ELISA. Combined inhibition of Hsp90 (17-DMAG, 100nM) and mTOR (Rapa, 10 nM) induces apoptosis in the HCC cell line HepG2 after 48 hours. (B) Similar effects were observed in endothelial cells after 48 hours. (C) Hsp90 blockade led to inhibition of constitutive activation of Akt, Erk, and STAT3. mTOR inhibition with rapamycin induced phosphorylation of Akt and the downstream intermediate GSK-3β. When cells were treated with rapamycin and 17-DMAG for 24 hours, this induction of Akt and GSK-3 was markedly reduced. (D) Treatment of HepG2 HCC cells with rapamycin led to increased Akt activity as determined by AktSer473 ELISA. This effect was blunted by simultaneous inhibition of Hsp90. (E) Pretreatment with 17-DMAG (24 hours) diminished IGF-I–induced activation of Akt. Moreover, treatment with 17-DMAG (24 hours) reduced Akt activation induced by rapamycin and IGF-I stimulation. Results are shown for HepG2; similar results were obtained with Hep3B and Huh-7 cells.

Download figure to PowerPoint

Because 17-DMAG and rapamycin are both known to suppress multiple signaling pathways, we next examined signaling intermediates affected by Hsp90 or mTOR inhibition. Blocking Hsp90 diminished constitutive phosphorylation of Akt, Erk, and STAT3 in HCC cells (Fig. 1C). Interestingly, expression of the transcription factor STAT5 was also reduced in 17-DMAG–treated tumor cells. However, similar to reported studies, treatment with rapamycin paradoxically induced AktSer473 and AktThr308 phosphorylation, as well as its downstream effector GSK-3β (Fig. 1C). Activation of other signaling intermediates was not detectable on rapamycin exposure. Importantly, inhibition of Hsp90 reduced rapamycin-induced phosphorylation of Akt (both sites). This effect was confirmed by measurement of AktSer473 activity (Fig. 1D). In addition, total-Akt expression was reduced by Hsp90 blockade in HepG2 and Hep3B cells, but not in Huh-7 cells, a variable effect reported previously.25 Otherwise, the effects of this dual-inhibition in terms of pathway inhibition were similar to 17-DMAG treatment alone (Fig. 1C).

Furthermore, the IGF-I/IGF-IR system is involved in rapamycin-mediated activation of Akt and has been connected to HCC progression.26 Therefore, we investigated the effect of Hsp90 inhibition on IGF-I–induced Akt phosphorylation in HCC cells. Importantly, IGF-I–mediated Akt phosphorylation could be suppressed by pretreating HCC cells with 17-DMAG (Fig. 1E). Moreover, treatment of tumor cells with rapamycin followed by stimulation with IGF-I further enhanced Akt phosphorylation. This induction was partially reduced by combination with 17-DMAG. These data indicate Hsp90 blockade suppresses rapamycin-induced and growth factor–induced activation of Akt in HCC cells, in addition to disrupting oncogenic signaling intermediates.

Effects of Hsp90/mTOR Inhibition on HIF-1α and VEGF-A Expression.

The transcription factor HIF-1α plays a pivotal role in tumor growth and angiogenesis,27 HIF-1α expression is associated with tumor progression in many tumor entities, including hepatocellular carcinoma.28 Therefore, we investigated the effects of 17-DMAG and rapamycin on HIF-1α in HCC cells. Consisting with previous results, rapamycin and 17-DMAG reduced hypoxia-mediated (1% O2) HIF-1α expression.5, 15 However, combination of both compounds did not further reduce HIF-1α (Fig. 2A). This diminished HIF-1α activation by 17-DMAG translated into significantly reduced VEGF-A messenger RNA (mRNA) levels on combinational treatment, whereas rapamycin or 17-DMAG treatment alone did not affect VEGF-A mRNA expression. However, when cells were stimulated with hypoxia, both 17-DMAG and combination therapy significantly inhibited VEGF-A mRNA expression (Fig. 2B). Similarly, VEGF-A secretion on hypoxia was reduced predominantly by Hsp90 inhibition (Fig. 2C). We conclude that Hsp90 inhibition can reduce tumor angiogenesis by impairing hypoxia-induced activation of HIF-1α and VEGF-A secretion.

thumbnail image

Figure 2. Effect of Hsp90/mTOR blockade on HIF-1α/VEGF-A expression. (A) Western blot analysis of full protein extract. HIF-1α expression was reduced in HepG2 cells on hypoxic stimulation for 24 hours and simultaneous inhibition of Hsp90 and mTOR. Similar results were obtained in Hep3B and Huh-7. (B) Treatment with 17-DMAG (100 nM) or the combination of 17-DMAG (100 nM) with rapamycin (10 nM) significantly down-regulated VEGF-A mRNA levels under nonhypoxic (£P < 0.05) and even under hypoxic conditions (1% O2) after 24 hours (*P < 0.05). Rapamycin alone had no significant effect on constitutive VEGF-A mRNA. Ratios of VEGF-A mRNA to β-actin are shown from three independent experiments. Results are given for HepG2 cells. (C) Similar results were observed on a protein level, as determined by ELISA. Constitutive VEGF-A levels were diminished by rapamycin, 17-DMAG, or the combination of both agents after 24 hours (£P < 0.05). Protein levels on 17-DMAG treatment were also lower under hypoxic stimulation (1% O2) (*P < 0.05). Bars = SEM.

Download figure to PowerPoint

Effects of Hsp90 and mTOR Inhibition on Cancer Cell Motility.

Hsp90 client proteins also may regulate tumor cell motility.17 In particular, focal adhesion kinase (FAK) is involved in cancer cell migration and invasiveness, and HGF represents an important growth factor for mediating FAK activation. When we determined the effects of Hsp90 and mTOR inhibition on constitutive and HGF-mediated FAK phosphorylation, we found that 17-DMAG markedly reduces FAK activation, whereas mTOR inhibition has no effect on FAK phosphorylation (Fig. 3A). Moreover, blocking Hsp90 reduced HGF-mediated tumor cell motility (Fig. 3B). In contrast, rapamycin had only a minor impact on tumor cell motility. These results suggest that Hsp90 inhibitors reduce HCC cell migration, which could promote anti-metastatic effects of rapamycin therapy in vivo.

thumbnail image

Figure 3. Effects of Hsp90 blockade on HGF-induced tumor cell motility. (A) Pre-incubation with 17-DMAG for 24 hours diminishes HGF-induced phosphorylation of FAK (30 minutes). Effects are shown for HepG2 cells; similar results were found for Huh-7 and Hep3B cells. (B) HGF (50 ng/mL) was used as a chemoattractant to HCC cells, which significantly induced tumor cell migration (#P < 0.05). Hsp90 inhibition with 17-DMAG therapy led to a significant down-regulation of constitutive and HGF-mediated tumor cell motility after 24 hours (*P < 0.05).

Download figure to PowerPoint

Effects of Hsp90 and mTOR Inhibition on Endothelial Cells and Pericytes.

In view of the fact that activation of angiogenic signaling intermediates in ECs and pericytes is essential for tumor angiogenesis,29 we next investigated the effects of Hsp90 and mTOR inhibition on their proliferation and motility. Human VSMCs were used as representative cell type for pericytes. In methylthiazole tetrazolium assays, both rapamycin and 17-DMAG effectively inhibited EC and VSMC proliferation (Supporting Fig. 2A, B and Supporting Fig. 3A, B), which was enhanced by combining both agents. Importantly, in migration assays using conditioned media from Huh-7 and HepG2 cells, rapamycin, but not 17-DMAG, significantly inhibited EC and VSMC migration after 4 hours (Fig. 4A, B). When both agents were combined, EC motility was further inhibited, but VSMCs showed no migratory change compared with rapamycin alone. Nevertheless, blocking Hsp90 also elicited anti-angiogenic properties in ECs by inhibiting VEGF-A–induced Erk and Akt activation, as well as down-regulating total-Akt expression (Fig. 4C). In contrast to tumor cells, rapamycin did not induce AktSer473 activity in ECs, whereas addition of 17-DMAG substantially decreased AktSer473 activity (data not shown). Moreover, VEGFR-2 protein expression on ECs decreased on treatment with 17-DMAG (Fig. 4C). Regarding PDGF-B signaling in VSMCs, Hsp90 blockade impaired Akt and Erk phosphorylation, and down-regulated Akt protein. Furthermore, blocking Hsp90 also diminished receptor expression on pericytes, as we detected PDGF-Rβ down-regulation on VSMCs (Fig. 4D). Therefore, targeting both mTOR and Hsp90 potentially reduces tumor vascularization and angiogenesis by direct effects on ECs and VSMCs.

thumbnail image

Figure 4. Effects of Hsp90/mTOR blockade on endothelial cells and pericytes. (A) After 4 hours, rapamycin significantly reduced migration of ECs on exposure to conditioned media derived from human HCC cell lines (*P < 0.05). Hsp90 blockade had no substantial effect on EC motility. (B) Similar significant results were found for VSMCs (*P < 0.05). (C) Treatment with 17-DMAG (100 nM, 24 hours) down-regulated Akt, and VEGFR-2 expression in EC and diminished Erk phosphorylation. (D) In VSMCs, Hsp90 blockade (24 hours) reduced expression of Akt and PDGF-Rβ. In addition, activation of Akt and Erk on stimulation with PDGF-B was blunted. Data in A and B are presented as the mean ± SEM (n = 3 experiments).

Download figure to PowerPoint

Combination of Hsp90 and mTOR Inhibitors for HCC Therapy.

To determine the efficacy of Hsp90 and mTOR targeting therapy in vivo, we first used a subcutaneous model (Huh-7). Treatment with either 17-DMAG (10 mg/kg/day) or rapamycin (0.1 mg/kg/day) reduced tumor growth, as compared with controls (Fig. 5A). However, the combination of both agents significantly improved growth-inhibitory effects, as was reflect by final tumor weights on day 16 (Supporting Fig. 4). Analyses of CD31-positive vessel area in tumors revealed that either 17-DMAG or rapamycin reduces HCC vascularization (Fig. 5B). This effect was improved with dual Hsp90/mTOR inhibition. In addition to reducing tumor vascularization, blocking Hsp90 or mTOR also reduced the number of BrdU-positive tumor cells (Fig. 5C), with combination therapy again demonstrating an improved effect. Furthermore, combination therapy showed a significant increase in apoptotic (TUNEL-positive) cells, as compared with either monotherapy arm, or controls (Fig. 5D).

thumbnail image

Figure 5. Impact of dual Hsp90 and mTOR inhibition on tumor growth in a subcutaneous xenograft model. Huh-7 human HCC cells (106) were injected into the right flank of nude mice. Animals (n = 9-10/group) received either 17-DMAG (10 mg/kg/day) ± rapamycin (0.1 mg/kg/day). (A) Single therapy with either 17-DMAG or rapamycin significantly inhibited tumor growth, as compared with controls (#P < 0.05 versus control); the effect was markedly enhanced by combining both agents (*P < 0.01 versus control). (B) Densitometric analyses of CD31 positive vessel area reveals reduced angiogenesis in 17-DMAG or rapamycin monotherapy groups, and this was further enhanced on dual inhibition therapy (*P < 0.01 versus control; #P < 0.05 versus monotherapy). C) Similar to these effects tumor cell proliferation was significantly reduced (*P < 0.01 versus control; #P < 0.05 versus monotherapy). (D) The number of apoptotic (TUNEL-positive) cells in tumors was elevated by combination therapy (*P < 0.01 versus control; #P < 0.05 versus monotherapy). (E) Western blot analysis of signaling pathways in tumor tissues in vivo. Rapamycin induced Akt phosphorylation at both phosphorylation sites; combination of 17-DMAG with rapamycin abrogated this effect. In addition, activation of STAT3 was markedly reduced with combinational therapy. Furthermore, expression of VEGF-A was diminished in 17-DMAG–treated groups. (F) Rapamycin treatment led to a significant induction of p65 DNA binding activity (*P < 0.05 versus. control). Addition of 17-DMAG significantly lowered NF-κB activity levels (#P < 0.05, compared to rapamycin). Data are presented as mean ± SEM.

Download figure to PowerPoint

Moreover, similar to in vitro experiments, activation of Akt occurred in rapamycin-treated HCC tumors (Fig. 5E). Akt activation was blunted by blocking Hsp90 with 17-DMAG therapy in vivo. Interestingly, Hsp90 blockade did not reduce total Akt expression in this environment. This does not conflict with our in vitro results, because we found that Hsp90 inhibition leads to a total Akt reduction in only some cell types; the in vivo situation certainly adds another confounding factor. Notably, STAT3 phosphorylation was markedly reduced in tumors receiving combination therapy. Furthermore, in keeping with in vitro results, 17-DMAG reduced expression of VEGF-A in tumors, whereas rapamycin had no effect on VEGF-A. More importantly, the activity of NF-κB, one downstream mediator of Akt, was significantly increased in rapamycin-treated tumors, suggesting an mTOR-independent biological downstream impact of rapamycin-induced Akt activation. Treatment with 17-DMAG markedly reduced NF-κB activity, and the combination therapy was able to lower rapamycin-induced DNA p65-binding levels (Fig. 5F). These results demonstrate that dual inhibition of Hsp90 and mTOR reduces growth and vascularization of HCC tumors and effectively blunts rapamycin-induced activation of both Akt and NF-κB in vivo.

Effects of Hsp90 and mTOR Inhibition on Hepatic Growth of Syngenic HCC Cells.

Tumor microenvironment has a substantial impact on growth rates and efficacy of targeted therapies. Therefore, we validated our results in a C3H syngenic orthotopic tumor model using Hepa129 cells.22 Because this tumor model elicits an aggressive growth behavior, we stepped up the treatment schedule by administering 17-DMAG three times per week (25 mg/kg) and rapamycin at 0.2 mg/kg/day. Single therapy with either 17-DMAG or rapamycin had no effect on hepatic tumor incidence (Fig. 6A). However, combination therapy reduced the incidence of HCC tumor–bearing mice to 40% (2/5); this effect was only a trend and did not reach statistical significance. Nevertheless, in addition to reducing the tumor incidence, combination therapy again led to improved growth-inhibitory effects in this syngenic model (Fig. 6B). Notably, blood sample analysis of rapamycin-treated mice showed rapamycin levels within a biologically relevant range (5-20 ng/mL; Table 1). Rapamycin use did not cause any wound-healing disturbances in these mice. We conclude that combining an mTOR inhibitor with an Hsp90 antagonist potently inhibits HCC tumor growth, providing a novel therapeutic approach for HCC treatment.

thumbnail image

Figure 6. Effect of Hsp90/mTOR blockade on orthotopic tumor growth in a syngenic model. Hepa129 tumor cells (105) were injected into the left liver lobe of C3H mice (n = 5-8/group) and animals received 17-DMAG (25mg/kg; 3×/week) ± rapamycin (0.2 mg/kg/day), starting at day 4 after tumor cell implantation. (A) Combination therapy with 17-DMAG and rapamycin appeared to reduce the incidence of hepatic tumors, albeit not to a level reaching statistical significance (P = 0.11). (B) Again, a significant growth inhibition was achieved by dual targeting of mTOR plus Hsp90, as reflected by final hepatic tumor volumes (*P < 0.05 versus control). Data are presented as mean ± ± SEM.

Download figure to PowerPoint

Table 1. Weight and Rapamycin Blood Measurements in an Orthotopic Mouse Model
 Mouse Weight (g)Rapamycin Blood Level (ng/mL)
  • *

    P < 0.05.

Control19.8 ± 1.9* 
Rapamycin23.9 ± 1.515 ± 3.6
17-DMAG24.7 ± 1.8 
17-DMAG+rapamycin24.0 ± 1.39 ± 4.3

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

In the current study we demonstrate that blocking Hsp90 is effective for reducing HCC growth and for improving anti-neoplastic efficacy of rapamycin in preclinical models. Importantly, inhibition of Hsp90 abrogates the rapamycin-induced paradoxical activation of Akt and NF-κB in vivo and exerts direct effects on ECs and VSMCs by reducing receptor expression and inhibiting signaling pathways. These results support a rationale for combining mTOR inhibitors with an Hsp90 inhibitor for targeted HCC therapy.

The rationale for adding an Hsp90 inhibitor to rapamycin is based on the following facts. First, anti-angiogenic properties of mTOR inhibitors are basically mediated via inhibition of cell growth, motility, and survival signaling in ECs.2, 3, 30 Second, despite their anti-angiogenic activity, mTOR inhibitors can also trigger a negative feedback loop that leads to Akt phosphorylation by release of S6K inhibition on IRS-1 and a shift of signaling toward mTORC2.2, 8, 9, 31 Because Akt phosphorylation is associated with a poor prognosis in different cancer entities, including HCC, the obvious concern is that mTOR blockade could increase Akt activity in patients and thus negate its anti-neoplastic efficacy.11 Importantly, Akt and the IRS-1/IGF-IR system are known client proteins of Hsp90, making Hsp90 inhibitors potentially useful for reducing these signaling systems.15, 17, 19 Indeed, in vitro rapamycin led to Akt activation in tumor cells, as might be predicted. This induction was completely abrogated by combining rapamycin with 17-DMAG, in part by reducing phosphorylation and down-regulation of Akt content in tumor cells. These in vitro results were confirmed in vivo, as 17-DMAG completely blunted rapamycin-induced activation of Akt at both phosphorylation sites. However, total Akt expression remained unaffected, as tumors consist of various cell types that do not all respond to Hsp90 inhibition in terms of down-regulating Akt. More importantly, mTOR inhibition by rapamycin induces DNA binding activity of NF-κB in vivo, which is one downstream effector of the PI3K/Akt signaling axis. This is of particular interest because NF-κB itself has been associated with enhanced tumor growth and angiogenesis.32 Induction of p65 activity was blunted on combination of rapamycin with 17-DMAG, reflecting reduced Akt activity in these tumors. Induction of NF-κB activity on mTOR inhibition and its abrogation by additionally targeting Hsp90 has not been previously described. These results demonstrate that combined inhibition of Hsp90 and mTOR is an attractive strategy to reduce paradoxical mTOR-inhibitor signaling in HCCs.

Hypervascularization in HCCs prompted us to investigate combinational therapy effects on angiogenesis.6 A crucial point of tumor angiogenesis is recruitment of ECs and VSMCs to create functional vessels.29 Rapamycin is known to have anti-angiogenic potential by acting on ECs and VSMCs. As expected, mTOR inhibition reduced EC and VSMC motility and survival in vitro. Regarding Hsp90 inhibitors, anti-angiogenic properties have also been proposed in part via inhibition of angiogenic signaling in tumor cells, thereby reducing VEGF-A secretion. Moreover, direct anti-angiogenic effects by down-regulation of VEGFR-1, VEGFR-2, and VEGFR-3 on endothelial cells have been reported.33 In our experiments, a combination of rapamycin and 17-DMAG reduced the expression of VEGFR-2 and PDGF-Rβ in ECs and VSMCs, respectively. Interestingly, down-regulation of PDGF-Rβ in VSMC by Hsp90 inhibitors has not been reported. Matei and coworkers34 found PDGF-Rα destabilization in ovarian, glioblastoma, and lung cancer cells, but no effects on PDGF-Rα expression on VSMCs were observed.34 The expression of PDGF-Rβ was not investigated in their study. This is particularly interesting because the tumor cell lines used in our experiments express neither PDGF-Rα nor PDGF-Rβ (data not shown), suggesting a direct effect of Hsp90 blockade on VSMCs by PDGF-R interference. In addition, PDGF-B–induced signaling in VSMCs was substantially diminished on treatment with 17-DMAG, again suggesting that combining rapamycin with an Hsp90 inhibitor has potent anti-angiogenic properties through additive effects on ECs and VSMCs. Indeed, anti-angiogenic properties from combination therapy were observed in our in vivo experiments, as reflected by substantial reductions in CD31 vessel area and diminished VEGF-A expression in tumors. Together these data clearly demonstrate the effectiveness of combined Hsp90/mTOR inhibition on vascularization in HCC tumors.

A similar approach of combining an mTOR inhibitor with a geldanamycin derivative for blocking Hsp90 was very recently published for multiple myeloma cells, because both agents have entered clinical trials for malignant hematopoietic disorders.35 Combination of the Hsp90 inhibitor 17-allylamino-17-demethoxygeldanamycin with rapamycin inhibited proliferation and induced apoptosis and cell cycle arrest in multiple myeloma cell lines. These observations were linked to cleavage of poly(ADP-ribose) polymerase (PARP)/caspase-8/caspase-9, and to effects on the PI3K/Akt/mTOR and cyclin D1/Rb pathways. They also found inhibition of angiogenesis and osteoclast formation, but only in in vitro assays. These results are consistent with our in vitro findings, pointing toward a synergistic effect of a combined Hsp90/mTOR blockade. Our study goes one step further by demonstrating significant inhibition of tumor cell proliferation in vivo. Moreover, we suggest that this effect is at least in part mediated through blunting rapamycin-mediated Akt/NF-κB induction.

Lastly, it is well known that the efficiency of anti-neoplastic and anti-angiogenic therapies is dependent on the tumor microenvironment and the immune system. Therefore, the observed effects on tumor growth were subsequently evaluated in an orthotopic model. Because of tumor aggressiveness in this model, we modified our treatment schedule by slightly increasing rapamycin doses (0.2 mg/kg/daily) to achieve therapeutic rapamycin levels. This dosage is still considered to be “low” and is indefinitely sustainable in humans.3, 5 Sustaining our previous results, we found improved anti-neoplastic efficacy and a low tumor incidence when combined treatment was applied. To our knowledge, this is the first study demonstrating efficacy of an Hsp90 inhibitor or mTOR inhibitor in a syngenic orthotopic tumor model of HCC. The value of this novel concept is further supported by recent reports indicating a suppressive effect of Hsp90 inhibition on HCV-virus replication, which could be beneficial for patients with HCC arising in such a setting.36 In summary, our study provides preclinical evidence for the use of Hsp90 inhibitors to improve mTOR inhibitor efficacy against HCC.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
hep_22685_sm_SupFig1.tif4979KSuppl. Fig 1
hep_22685_sm_SupFig2.tif4980KSuppl. Fig 2
hep_22685_sm_SupFig3.tif4980KSuppl. Fig 3
hep_22685_sm_SupFig4.tif4978KSuppl. Fig 4

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.