1 Introduction
Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer-related death globally [
1]. Much progress has been obtained in the systemic treatments of advanced-stage HCC [
2]. Anti-angiogenic tyrosine kinase inhibitors (TKIs) are the landmark breakthrough in systemic treatments of HCC. Sorafenib and lenvatinib have been demonstrated to prolong the overall survival of patients with unresectable HCC and proved to be the first-line treatment. Nevertheless, the response rate and survival benefits provided by these TKIs are limited [
3]. Thus, more effective therapeutic strategies need to be developed.
Human growth factor (HGF)/cMet is a pivotal and clinically actionable pathway. Constitutive cMet activation drives a complex tumor progression program and is involved in sorafenib resistance in HCC [
4]. Cabozantinib is an oral small-molecule TKI that mainly blocks the phosphorylation of cMet and vascular endothelial growth factor receptor 2 (VEGFR2), as well as a number of other receptor tyrosine kinases (RTKs) including RET, KIT, AXL, and FLT3 [
5]. It exerts its anti-tumor effect via inhibiting angiogenesis, proliferation, and metastasis of HCC cells under hypoxia condition and has been proved to be the second-line treatment for advanced-stage HCC patients [
5–
7]. However, the lower response rate and the probabilities of drug resistance limit the enduring clinical benefits of cabozantinib [
5–
7].
Mammalian target of rapamycin (mTOR) is a common and key signal transducing molecule in downstream signaling pathways of RTKs. Sustained mTOR signal confers drug resistance to various targeted therapies. Rapamycin is an allosteric mTORC1 inhibitor and has been approved by the Food and Drug Administration in the treatment of renal cell carcinoma and neuroendocrine tumors [
8,
9]. However, the anti-tumor effect of rapamycin on patients with unresectable HCC is compromised [
10,
11].
Our previous study revealed that cMet-low HCC cells showed less sensitivity to cMet inhibitors than cMet-high HCC cells [
12]. In this study, we further demonstrated that the combination of rapamycin with cabozantinib had a synergistic suppressive effect on the tumor growth and hypoxia-induced metastasis of cMet inhibitor-resistant HCC cells. Mechanically, the combination treatment induced a more effective inhibition of downstream signal pathways of RTKs in cMet inhibitor-resistant HCC cells. This study provides a potential strategy to overcome the primary resistance of HCC to cabozantinib.
2 Materials and methods
2.1 Human clinical samples
Paraffin-fixed human HCC samples were collected from HCC patients who underwent liver resection at the Department of General Surgery, Huashan Hospital, Fudan University. All clinical samples were collected from patients after obtaining informed consent in accordance with a protocol approved by the Ethics Committee of Huashan Hospital, Fudan University (Shanghai, China).
2.2 Cell culture
Human HCC cell lines, MHCC-97H and MHCC-97L, were provided by the Liver Cancer Institute, Zhongshan Hospital, Fudan University. Human HCC cell lines including PLC/PRF/5 (PLC), Hep3B, and HepG2 cells were purchased from the American Type Culture Collection. Human umbilical vascular endothelial cells (HUVECs), Huh7, and mouse hepatocarcinoma cell line Hepa1-6 were obtained from the Cell Bank of Type Culture Collection of Chinese Academy of Sciences. All HCC cell lines were cultured in DMEM (Gibco) with 10% (v/v) FBS (Gibco) at 37 °C in a humidified incubator with 5% CO2. HUVECs were cultured in ECM (ScienCell) with 10% (v/v) FBS (Gibco).
2.3 Cell proliferation assay
Cell proliferation was measured by using the Cell Counting Kit-8 (CCK8) according to the manufacturer’s instructions (HY-K0301, MCE). Briefly, HCC cells were plated at a density of 3000 cells per well in 96-well plates and incubated for 12 h. Then, cells were treated with the indicated drug concentration for an additional 72 h. About 10 μL of CCK8 reagent was added and incubated for 1 h at 37 °C. The absorbance at 450 nm was measured with a microplate reader (Infinite M200 Pro NanoQuant, TECAN). The inhibitors used for cell proliferation assay included CC930 (S8490, Selleck), SB202190 (S1077, Selleck), sotrastaurin (S2791, Selleck), MK2206 (S1078, Selleck), SCH772984 (S7101, Selleck), rapamycin (HY-10219, MCE), SC75741 (S7273, Selleck), regorafenib (S1178, Selleck), gefitinib (S1025, Selleck), lapatinib (S2111, Selleck), BGJ398 (S2183, Selleck), dasatinib (S1021, Selleck), olaparib (S1060, Selleck), niraparib (S2741, Selleck), MK-1775 (S1525, Selleck), FK866 (S2799, Selleck), palbociclib (S1116, Selleck), obatoclax (S1057, Selleck), GSK525762A (S7189, Selleck), JQ1 (S7110, Selleck), cabozantinib (S1119, Selleck), and NZ001 (Nanjing Zhongrunyuan Pharmaceutical Company). Among these inhibitors, palbociclib was resolved using water, and all other inhibitors were resolved using DMSO for in vitro experiments.
2.4 Combination index (CI) assay
After treatment with the indicated inhibitors, the cell viability was measured by cell proliferation assay. The drug combination studies and their synergistic interactions were analyzed by CompuSyn software on the basis of the Chou–Talalay method [
13]. CI>1 indicated antagonism, CI= 1 indicated additivity, and CI<1 indicated synergism between two drugs.
2.5 Western blot
To prepare whole cell lysate, HCC cells were lysed with RIPA Lysis Buffer (P0013C, Beyotime) containing protease inhibitors (HY-K0010, MCE) and phosphatase inhibitors (HY-K0021, MCE), and the protein concentration was determined using bicinchoninic acid assay (23225, Thermo Fisher Scientific). Equal amounts of protein were separated by SDS-PAGE and analyzed with primary and secondary antibodies following the manufacturer’s protocols. The signals were obtained by using an enhanced chemiluminescence reagent (WBKLS0500, Millipore Corp). The primary antibodies used for Western blot included cMet (8198, Cell Signaling Technology), pMet (Tyr1234/1235) (3077, Cell Signaling Technology), cyclin D1 (ab134175, Abcam), PARP (9532, Cell Signaling Technology), cleaved-PARP (5625, Cell Signaling Technology), GAPDH (AC002, ABclonal), pAKT (Ser473) (4060, Cell Signaling Technology), AKT (4685, Cell Signaling Technology), pmTOR (Ser2448) (5536, Cell Signaling Technology), pmTOR (Ser2481) (2974, Cell Signaling Technology), mTOR (2983, Cell Signaling Technology), pERK1/2 (Thr202/Tyr204) (4370, Cell Signaling Technology), ERK1/2 (4695, Cell Signaling Technology), p-p65 (Ser536) (3033, Cell Signaling Technology), p65 (8242, Cell Signaling Technology), pP70S6K (Thr421/Ser424) (9204, Cell Signaling Technology), P70S6K (9202, Cell Signaling Technology), pEGFR (Tyr1068) (3777, Cell Signaling Technology), EGFR (4267, Cell Signaling Technology), pFGFR2 (Tyr653/654) (3476, Cell Signaling Technology), FGFR2 (23328, Cell Signaling Technology), pERBB2 (Tyr1221/1222) (2243, Cell Signaling Technology), ERBB2 (2165, Cell Signaling Technology), pPLCr (Ser1248) (8713, Cell Signaling Technology), PLCr (2822, Cell Signaling Technology), pMEK (Ser221) (2338, Cell Signaling Technology), MEK (4694, Cell Signaling Technology), pJAK2 (Tyr1007/1008) (3771, Cell Signaling Technology), and JAK2 (3230, Cell Signaling Technology). The secondary antibodies used for Western blot included HRP Goat Anti-mouse IgG (AS003, ABclonal) and HRP Goat Anti-rabbit IgG (AS014, ABclonal).
2.6 Cell cycle and apoptosis
After treatment with the indicated inhibitors for 48 h, the cells were washed with PBS buffer twice and collected (1×105). Cell cycle and apoptosis were analyzed using flow cytometry with propidium iodide/RNase staining buffer (550825, BD Biosciences, USA) and annexin V-FITC/PI apoptosis detection kits (556547, BD Biosciences, USA), respectively, according to the manufacturer’s instructions. The data were processed with FlowJoTM 10.6.1 software.
2.7 Cell invasion assay
Cell invasion assay was performed using a 24-well transwell chamber with Matrigel (356234, BD Pharmingen). After being starved in serum-free medium for 24 h, Huh7 cells were resuspended (1 × 105 cells in 200 μL of DMEM per well) in serum-free DMEM containing the indicated inhibitors and added to the upper chambers. The lower chambers were filled with 600 μL of DMEM containing 2% FBS with 1 or 10 ng/mL of HGF (100-39H, PeproTech) as chemoattractant and incubated at 37 °C for 48 h. Cells that migrated to the lower surface of chambers were fixed with methanol, stained with 0.5% crystal violet, and counted with a microscope (100×) (CTR6000, Leica). For hypoxia, cells were incubated at 5% CO2 and 95% N2 in a humidified incubator at 37 °C.
2.8 Colony formation assay
Cells at the logarithmic growth phase were seeded at a density of 1000 cells per well in 6-well plates and cultured with or without inhibitors for 10 days. Fresh medium was added every 3 days. The colonies were fixed with methanol and stained with 0.5% crystal violet. The blue colonies were counted on five randomly chosen fields under a microscope (100×) (CTR6000, Leica).
2.9 Wound healing assay
HUVECs were plated and allowed to confluence on marked plastic dishes. After washing with PBS buffer twice and being starved for 8 h in serum-free ECM medium, a wound gap in HUVEC monolayer was created by scratching with a 200-μL pipette tip. Cells were washed with PBS buffer and then cultured with or without drugs for 24 h. The wounds were monitored and photographed at 0 and 24 h. The healing was quantified by measuring the distance between edges.
2.10 Endothelial tubule formation assay
Matrigel (356234, BD Pharmingen) was plated in 24-well plates and allowed to solidify in a 37 °C cell culture incubator. HUVECs were resuspended in serum-free ECM medium with or without drugs, plated on Matrigel layer (1×105/mL), and incubated at 37 °C for 4 h. Tube networks were photographed using a phase-contrast microscope (DM IL LED, Leica). The angiogenic index was calculated as the number of branch points in a randomly selected visual field.
2.11 Lentivirus-mediated knockdown of gene expression
Short hairpin RNAs (shRNAs) targeting MET were cloned into pLKO.1-puro vector (Addgene). shRNA sequences are available in Table S1. For lentivirus generation, 1 × 107 293T cells were seeded in a 10-cm dish in DMEM supplemented with 10% FBS the day before transfection. Cells were transfected by changing to 10 mL of DMEM containing 20 μL of LipoFiterTM Liposomal Transfection Reagent (HB-TRLF, Hanbio) and 10 μg of pLKO.1-puro vector or pLKO shMET combined with 7.5 μg of psPAX2 and 2.5 μg of pMD2.G. After 6 h, the medium was changed with DMEM 10% FBS. After 48 h, the supernatant was collected, filtered with 0.45 mm filters, and used to infect HCC cell lines. The stable cell lines were obtained by puromycin selection for 1 week.
2.12 In vivo xenograft tumor model and immunohistochemistry (IHC)
Subcutaneous implantation models were established using Hepa1-6 cells. A total of 5 × 105 cells were resuspended in 80 μL of PBS and injected subcutaneously into the right side of the posterior flank of male C57BL/6 mice (4 weeks old, Shanghai Institute of Materia Medica, Chinese Academy of Sciences). After injection for 1 week and tumor volume reached about 100 mm3, the mice were randomized into four groups: vehicle group (1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), cabozantinib group (30 mg/kg of cabozantinib in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), rapamycin group (2 mg/kg of rapamycin in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), and combination group (30 mg/kg of cabozantinib and 2 mg/kg of rapamycin in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days). The volume of tumors was calculated by using the following formula: a × b2/2 (a and b refer to the largest and smallest diameter, respectively) every 3 days. After 2 weeks of treatment, the mice were sacrificed and tumors were collected.
Tumors were either homogenized in tumor lysis buffer for Western blot analysis or fixed with 4% formaldehyde solution and embedded in paraffin. Sections were deparaffinized and rehydrated in alcohol gradient, and subjected to microwave antigen retrieval. Sections were treated with 0.3% hydrogen peroxide to block the endogenous peroxidase. After blocking with 2% BSA for 1 h at RT, sections were incubated with primary antibodies and secondary antibodies following the manufacturer’s instructions. Color development was performed with HRP (DAB Peroxidase Substrate Kit (Vector Laboratories, SK-4100)).
The primary antibodies used for IHC included cMet (8198, Cell Signaling Technology), CD31 (77699, Cell Signaling Technology), cyclin D1 (ab134175, Abcam), PCNA (13110, Cell Signaling Technology), pmTOR (Ser2448) (ab109268, Abcam), and pAKT (Ser473) (ab81283, Abcam). The secondary antibodies used for IHC include Peroxidase AffiniPure Goat Anti-Rabbit IgG (H+ L) (111-035-003, JACKSON).
2.13 In vivo lung metastasis model
Lung metastasis models were established by intravenously injecting Hepa1-6 cells. Briefly, 2 × 106 cells were resuspended in 100 μL of PBS and injected into the tail vein of male C57BL/6 mice (4 weeks old, Shanghai JieSijie Laboratory Animals Co., Ltd.). After injection for 2 weeks, the mice were randomized into four groups: vehicle group (1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), cabozantinib group (30 mg/kg of cabozantinib in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), rapamycin group (2 mg/kg of rapamycin in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days), and combination group (30 mg/kg of cabozantinib and 2 mg/kg of rapamycin in 1.25% polyethylene glycol, 2.5% Tween-80, and 5% DMSO, orally, once every 2 days). Mice were sacrificed after 2 weeks of treatment. The lungs were removed, fixed in paraformaldehyde (4%), and embedded in paraffin. Consecutive sections were made for every lung tissue block and stained with hematoxylin and eosin (HE). The number of lung metastases in the maximal section of the metastatic lesion was calculated and evaluated.
2.14 Statistical analysis
All data are presented as mean±SD as indicated in the figure legends. Statistical analysis was performed with SPSS10.0 software. Depending on the data, Student’s t-test and one- or two-way ANOVA were used to analyze quantitative variables. Differences were considered statistically significant at P<0.05.
3 Results
3.1 cMet-low HCC cells are resistant to cMet inhibitors
Consistent with our previous report that
MET amplification and cMet overexpression were positively associated with the sensitivity of HCC cells to cMet inhibitors [
12], low concentration (2 μmol/L) of cabozantinib showed no obvious inhibitory effect on the phosphorylation of cMet and expression of cyclin D1 and cleaved-PARP in Huh7 cells with low cMet expression (Fig. 1A). We then assessed the responses of different HCC cells to cMet inhibitors, cabozantinib, and NZ001. NZ001 is a novel dual cMet and VEGFR2 inhibitor. MHCC-97H and MHCC-97L cells, which had high cMet expression, were more sensitive to cabozantinib and NZ001 than HCC cells with low cMet expression (Fig. 1B). The cell growth curves were similar among HCC cells treated with cabozantinib and NZ001 (Fig. 1B), indicating their similar direct inhibitory effects on HCC cells. Therefore, NZ001 was used to verify the result of cabozantinib in this study. Moreover, we retrieved the Genomics of Drug Sensitivity in Cancer database for more drug response data about cabozantinib on HCC cells. We found that cells with IC50>5 μmol/L were identified as resistant to cabozantinib, and 13 out of 17 cell lines screened were cabozantinib-resistant (Fig. 1C). Then, we analyzed cMet expression in 10 human HCC tissues by IHC staining and found that cMet expression exhibited extensive heterogeneity among different regions within the same HCC tissues (Fig. 1D). This molecular trait reflects different sensitivity of distinct colonies in HCC to cMet inhibitors. Collectively, these results indicate that a spatial heterogeneity of cMet expression exists in HCC tissues and that cMet-low HCC cells show primary resistance to cMet inhibitors.
3.2 Rapamycin increases the sensitivity of cMet inhibitor-resistant HCC cells to cabozantinib
As RTKs are the main targets for cabozantinib, we next examined the activations of RTKs in three cMet inhibitor-resistant HCC cell lines and observed the diverse expression of activated RTK signaling pathways (Fig. S1). Moreover, cabozantinib treatment did not affect the phosphorylation of these RTKs (Fig. S1). Therefore, we hypothesized that the inhibition of common and key downstream signal molecules of RTKs could be an alternative to cover as many cMet inhibitor-resistant HCC cells as possible. We examined the inhibitory effect of cabozantinib in combination with inhibitors targeting the key points of MAPK–ERK and PI3K–AKT–mTOR cascades on cMet inhibitor-resistant Huh7 and PLC cells using cell proliferation assay. We found that, in combination with cabozantinib, mTOR inhibitor rapamycin exhibited the most effective inhibition of Huh7 and PLC cell proliferation (Fig. 2A). We further expanded the combination treatment screening to the common target drugs, and the inhibitory efficiency of these drugs varied between Huh7 and PLC cells. The EGFR inhibitor gefitinib and Wee1 inhibitor MK-1775 could notably enhance the anti-proliferation effect of cabozantinib on PLC cells but only showed modest inhibition on Huh7 cells, which was consistent with the diverse expression of activated RTKs among cMet inhibitor-resistant HCC cells (Fig. S1). Moreover, among the three cMet-low HCC cell lines we examined, mTOR signals were all activated, which could not be effectively inhibited by cabozantinib (Fig. S1). The effectiveness of small molecular inhibitors was confirmed by examining the phosphorylation of molecular targets (Fig. 2B–2E). These results indicate that cabozantinib in combination with rapamycin is a potential choice for treating cMet inhibitor-resistant HCC.
3.3 Combination of rapamycin and cabozantinib shows synergistic inhibition of cMet inhibitor-resistant HCC cell proliferation
To further evaluate the effect of cabozantinib and rapamycin combination on cMet inhibitor-resistant HCC cells, we performed cell proliferation assay and calculated their CI value in Huh7 and PLC cells. Rapamycin alone showed only moderate inhibition on cell growth. The cell survival percentage of Huh7 and PLC cells was 50%–60% and around 50%, respectively, when the rapamycin concentration reached 2 μmol/L or more (Fig. 3A and 3B). However, the inhibitory effect of cabozantinib or NZ001 on HCC cell growth increased in a dose-dependent manner (Fig. 3A and 3B). The combination of cabozantinib and rapamycin synergistically inhibited cell growth (Fig. 3A and 3B), with CI values less than one at all concentrations examined (Fig. 3C and 3D). Similar synergistic inhibitory effect was also observed in the combination of NZ001 with rapamycin (Fig. 3A–3D). We further found that the colony formation ability of Huh7 and PLC cells treated with cabozantinib and rapamycin was much lower than that of HCC cells treated with each inhibitor alone (Fig. 3E and 3F). Moreover, to evaluate the anti-tumor effects of cabozantinib and rapamycin on HCC cells with high cMet expression when cMet was knocked down, we generated two MET-specific shRNAs to silence cMet expression (shMET). The shMET#1 induced more significant knockdown effects than the other in MHCC-97H cells that highly expressed cMet, and was adopted for knocking down cMet expression (Fig. S2A). We performed cell proliferation assay of shNC and shMET MHCC-97H treated with cabozantinib and rapamycin and found that the inhibitory effect on shMET MHCC-97H was compromised compared with shNC MHCC-97H (Fig. S2B). The underlying mechanism was that because cMet overexpression was positively associated with the sensitivity of HCC cells to cMet inhibitors (Fig. 1B), knocking down cMet expression in MHCC-97H with high cMet expression allowed them to acquire resistance to cMet inhibitors. Collectively, these data indicated a synergistic interaction between cabozantinib and rapamycin in inhibiting cMet inhibitor-resistant HCC cells growth in vitro.
3.4 Rapamycin enhances the inhibitory effect of cabozantinib on angiogenesis
On the basis of the known anti-angiogenic function of rapamycin and VEGFR2 inhibitor, we determined whether cabozantinib and rapamycin have synergistic anti-angiogenesis effect. Cabozantinib or rapamycin alone suppressed the migration ability of HUVECs, and this effect was significantly enhanced by combination treatment of rapamycin and cabozantinib or NZ001 (Fig. 4A and 4B). In addition, we analyzed the angiogenic index measured by tubule formation assay. Similarly, combination treatment induced the most remarkable inhibition of tubule formation. The angiogenic index in the combination treatment group was significantly smaller than that of the cabozantinib, NZ001, or rapamycin alone group (Fig. 4C and 4D). To further validate the anti-angiogenesis effect of combination treatment in vivo, we detected CD31-positive blood vessels in the subcutaneous implantation models of Fig. 6A by IHC staining. Compared with vehicle, cabozantinib and rapamycin alone significantly reduced the blood vessel density of tumor tissues. Furthermore, the anti-angiogenesis effect of combination treatment was greater than that of each single agent alone (Fig. 4E and 4F). These data indicate the enhanced anti-angiogenesis effect of combination treatment of cabozantinib and rapamycin.
3.5 Rapamycin augments the inhibitory effect of cabozantinib on the metastasis of HCC cells
Our previous study indicated that hypoxia induced by VEGFR blockade could increase HCC cell invasion induced by HGF [
12]. We detected the invasion ability of HCC cells treated with 1 or 10 ng/mL of HGF under normoxia or hypoxia condition. We found that 1 ng/mL of HGF slightly enhanced the invasion of Huh7 cells under normoxia condition, and the number of invading cells increased when treated with 10 ng/mL of HGF (Fig. 5A and 5C). Consistent with our previous findings, hypoxia augmented the sensitivity of HCC cells to HGF, as indicated by the significantly enhanced invasion of HCC cells treated with the same level of HGF under hypoxia condition (Fig. 5A and 5C). We further determined whether rapamycin augmented the inhibitory effect of cabozantinib on HGF-induced Huh7 invasion under hypoxia condition using the chemotactic index, and the results suggested that the chemotactic index of Huh7 cells with combination treatment was significantly less than that of HCC cells treated with cabozantinib alone. Moreover, the inhibitory effect of combination treatment on HGF-induced invasion of Huh7 cells notably increased dose-dependently (Fig. 5B and 5D). Finally, to evaluate the combined effects on the metastasis of HCC
in vivo, we injected Hepa1-6 cells into the tail vein of male C57BL/6 mice to establish a lung metastasis model. After 2 weeks, vehicle, cabozantinib, rapamycin, and combined drugs were administered for another 2 weeks. Then, mice were sacrificed, and their lungs were removed, sliced, and stained with HE to evaluate lung metastasis. We found that compared with the vehicle group, a remarkable decrease in the number of lung metastatic lesions was observed in other groups, and no lung metastasis lesion was found in C57BL/6 mice under the combination treatment of cabozantinib and rapamycin (Fig. 5E and 5F). These data indicate that combination treatment of cabozantinib and rapamycin exhibits more effective suppressive effect on the metastasis of HCC cells than cabozantinib alone.
3.6 Anti-tumor effects of combination treatment on the growth of HCC xenograft in vivo
The synergistic effects of cabozantinib in combination with rapamycin were further confirmed in immunocompetent subcutaneous tumor model. In our previous work, Hepa1-6 was identified as a cMet-low murine HCC cell line that was insensitive to cMet inhibitor [
12]. We first detected the inhibitory effect of combination treatment on the proliferation and colony formation of Hepa1-6 cells
in vitro. Similar to the studies above, cMet inhibitors (cabozantinib, NZ001, and PF04217903) or rapamycin alone showed no inhibitory effect, whereas the synergistic effects of combination treatment on inhibiting proliferation and colony formation were observed in Hepa1-6 cells (Fig. S3A and S3B). Next, vehicle, cabozantinib, rapamycin, and combined drugs were administered to C57BL/6 mice 1 week after the establishment of the subcutaneous tumor model of Hepa1-6 cells (tumor volume reached about 100 mm
3). We found that compared with vehicle, all other treatments delayed tumor growth significantly, among which combination treatment exerted the most effective inhibitory effect (Fig. 6A–6C). In addition, no significant weight loss was observed among all groups, which indicated that the combination treatment of cabozantinib and rapamycin did not significantly increase toxicity and could be well tolerated (Fig. 6D). To determine whether the cell cycle is affected by combination treatment, we examined the cyclin D1 expression in subcutaneous transplantation tumors by IHC staining. A modest decrease of cyclin D1 in the cabozantinib treatment group and a significant decrease in rapamycin-treated tumors were observed, while the combination of cabozantinib and rapamycin resulted in a much more obvious decrease of cyclin D1 expression than the other treatments (Fig. 6E and 6F). In addition, we used PCNA, another indicator of cell proliferation, to validate the effect of combination treatment on cell proliferation. We found that the proliferation index decreased using each inhibitor alone, and the biggest decrease occurred in the combination treatment group (Fig. 6E and 6G). Finally, to assess whether combination therapy suppressed AKT/mTOR signaling, we examined pmTOR and pAKT expression in subcutaneous xenograft by IHC staining. Cabozantinib alone failed to suppress the phosphorylation of mTOR, which could be inhibited by rapamycin alone. Conversely, rapamycin led to feedback activation of AKT, which could be attenuated by cabozantinib. The suppression of pmTOR and pAKT was achieved only in the combination of cabozantinib and rapamycin (Fig. 6E, 6H, and 6I). Taken together, our
in vivo results reveal that rapamycin augments the anti-tumor effects of cabozantinib on the growth of cMet inhibitor-resistant HCC.
3.7 Combination treatment inhibits the phosphorylation of AKT, mTOR, and ERK and induces cell cycle arrest
We next investigated the mechanism underlying the synergetic inhibitory effects of the combination treatment. Cabozantinib and NZ001 alone failed to inhibit the phosphorylation of AKT, ERK, and mTOR, the core signal molecules of RTK downstream pathways, in cMet inhibitor-resistant Huh7 cells (Fig. 7A). By contrast, rapamycin suppressed the phosphorylation of mTOR, p70S6K, and ERK but led to feedback activation of AKT and slightly enhanced cMet phosphorylation, which could be attenuated by cabozantinib and NZ001 (Fig. 7A). Therefore, the combination treatment could block the PI3K–AKT–mTOR and MAPK–ERK pathways downstream of RTKs at the same time. Considering that the PI3K–AKT–mTOR and MAPK–ERK pathways play a role in cell cycle regulation, we detected the expression of cyclin D1, a critical regulator of G1–S phase transition, and investigated cell cycle by flow cytometry. The results indicated that cabozantinib, NZ001, or rapamycin alone showed no effect on cell cycle, while the combination treatment induced a significant downregulation of cyclin D1 and an obvious cell cycle arrest compared with the control group (G1%: 50.65 or 52.71 vs. 34.31, P<0.001) (Fig. 7B–7D). Moreover, we detected the cleaved-PARP level, an indicator of apoptosis, and conducted annexin V-FITC/PI apoptosis assay using flow cytometry after treatment with inhibitors alone or in combination. The results showed no obvious cleaved-PARP alteration or changes in the percentage of apoptotic Huh7 cells treated with single drug or combined drugs (Fig. S4A–S4C). These studies indicate that combination treatment synergistically inhibits the phosphorylation of AKT, mTOR, and ERK and induces cell cycle arrest in cMet inhibitor-resistant HCC cells.
4 Discussion
The cMet signaling pathway has been implicated in HCC progression, including proliferation, invasion, and hypoxia resistance [
14]. Elevated expressions of cMet and its ligand HGF have been proved to indicate poor prognosis in HCC patients [
15,
16]. Our previous results indicated that cMet-low HCC cell lines (Huh7, PLC, HepG2, Hepa1-6) were relatively resistant to cabozantinib treatment compared with cMet-high HCC cell lines (MHCC-97H and MHCC-97L), with oral administration of cMet inhibitor leading to more tumor growth inhibition of MHCC-97H xenografts than Huh7 xenografts [
12]. Nevertheless, several clinical trials have indicated that the IHC score of cMet expression or
MET gene amplification showed limited power to predict the response of advanced-stage HCC patients to cMet inhibitors [
17].
HCC harbors remarkable spatial heterogeneity—an individual tumor might comprise diverse cancer cells with distinct molecular driver alterations, thus providing the fuel for drug resistance and relapse [
18,
19]. Herein, we found a remarkable heterogeneity of cMet expression among distinct regions within the same HCC tumor tissue. On the basis of our previous data showing that the level of cMet expression or
MET amplification of HCC cell lines was correlated with the sensitivity to cMet inhibitors [
12], we hypothesized that under selection pressure, cMet-high HCC cells that are sensitive to cMet inhibitors are obliterated, while HCC cells with low cMet expression survive and evolve more progressively. In other words, the heterogeneity of cMet expression within HCC tumor tissue mediates the primary resistance of HCC to cMet inhibitors. Moreover, we found the primary resistance of a majority of HCC cell lines to cMet inhibitors
in vitro in this study and the low ratio of cMet expression in HCC samples previously reported [
12]. On this basis, we proposed a combination strategy to target cMet inhibitor-resistant HCC cells in this study.
The targeted drugs proved to improve the prognosis of HCC patients mainly targeted RTKs [
7]. One of the approaches to reduce the primary resistance is to combine with other RTK inhibitors on the basis of the IHC-staining-detected abnormal activation of RTK signaling [
20,
21]. However, it is unfeasible to cover all the abnormally activated RTKs detected at the same time. On one hand, the activated RTK signaling pathways in cMet inhibitor-resistant HCC cells are diverse [
22]. On the other hand, it is hard to obtain clinical tissue biopsy of unresectable HCC patients. Thus, cMet inhibitors combined with small molecular inhibitors targeting common and druggable molecules downstream of RTK signaling pathways may be a rational alternative [
23–
25]. In this regard, we screened inhibitors targeting these main downstream proteins listed in Fig. 2A. Herein, our results indicated that the combination of cabozantinib and rapamycin, a specific inhibitor of mTOR, synergistically led to significantly enhanced inhibition of cMet inhibitor-resistant HCC cells in terms of proliferation, colony formation, cell cycle regulation, invasion, and angiogenesis.
mTOR is an essential integrator of growth factor-activated pathways to regulate various functions, including cell proliferation, survival, and metabolism [
7]. mTORC1 is mainly activated through the phosphoinositide 3-kinase (PI3K)/AKT signaling pathway [
26,
27]. mTOR inhibitors have been investigated in cancer treatment for decades and were tested in several ongoing clinical trials for patients with advanced HCC. A previous study revealed that the mTOR signaling pathway was abnormally activated in nearly half of HCC tissues, and the inhibition of mTOR could extend the survival of mice harboring HCC xenograft tumors [
28]. However, the efficiency of mTOR inhibitors could be impaired by the feedback activation of AKT, which provided survival signal to cancer cells. Our results indicated that this side effect could be attenuated by cabozantinib, which was in line with a previous report in epithelioid sarcoma showing that AKT activation induced by mTOR inhibition was dependent on the HGF–cMet signaling pathway [
29]. For cMet inhibitor-resistant HCC cells, rapamycin could in turn inhibit mTOR signaling and ERK phosphorylation, which are core intermediate molecules of RTK signaling transduction pathways. These results could account for the synergistic effects of combination treatment
in vitro.
Apart from its direct anti-tumor activity, cabozantinib also suppresses angiogenesis by inhibiting VEGFR2 [
5]. Likewise, mTOR signaling also acts as a pivotal mediator of angiogenesis, especially under hypoxia condition [
8]. We examined the potential synergistic effect of rapamycin and cabozantinib on HUVECs. As expected, we found that combination treatment exhibited enhanced inhibition of vascular cell migration and tubule formation
in vitro and angiogenesis
in vivo than single drug. Moreover, activated mTOR signaling was found in anti-angiogenesis therapy-resistant cells [
30], which provides more evidence for the synergistic effects of combination therapy.
Hypoxia induced by anti-angiogenic therapies initially retards cancer cell proliferation. Nevertheless, it also triggers the upregulation of a series of downstream genes of HIF1α, such as
VEGF and
MET [
31,
32]. Our previous data indicated that hypoxia induced the upregulation of cMet expression in cMet-low HCC cells, which increased the sensitivity to HGF stimulation and enhanced invasion ability. In this regard, cMet inhibitors could disrupt the aggressive phenotype induced by hypoxia [
12]. This study further revealed that, under hypoxia condition, rapamycin could further enhance the inhibitory effect of cabozantinib on HCC invasion even at a low level of cabozantinib.
Most of the previous preclinical trials examining therapeutic effects of cabozantinib on HCC were based on immunodeficient mouse models [
5,
33]. Considering that cMet and mTOR signaling are implicated in the establishment of a tumor microenvironment [
11,
34], we used immunocompetent mice and Hepa1-6 cells, proved to be resistant to cabozantinib
in vitro, to explore the synergistic inhibitory effects of combination treatment
in vivo. In Hepa1-6 xenografts, combination treatment showed significantly enhanced inhibition of tumor growth and angiogenesis than single inhibitor. In addition to the inhibitory effects on HCC cells, the mechanism underlying immune microenvironment regulation of the combination therapy needs further investigation.
Furthermore, besides cMet and VEGFR2, cabozantinib also targets other RTKs including RET, KIT, AXL, and FLT3 [
5]. For instance, the growth arrest-specific protein 6/AXL signaling pathway and RET mutations have been implicated in the promotion of tumor cell proliferation, survival, migration, invasion, angiogenesis, and immune evasion [
35,
36]. Therefore, the synergistic effect of combination therapy on HCC cells may also be attributed to the modulation of these RTKs, which needs to be elucidated by further investigation.
In summary, our study identifies the mTOR inhibitor rapamycin as a feasible option to increase the sensitivity of cMet inhibitor-resistant HCC cells to cabozantinib. Rapamycin and cabozantinib exhibit synergistic inhibition of the proliferation and hypoxia-induced invasion of cMet inhibitor-resistant HCC cells and angiogenesis. Rapamycin along with other mTOR inhibitors is being evaluated in advanced-stage HCC clinical trials, and some have been approved in the treatment of different types of solid tumors [
37]. Our data provide mechanistic evidence for the combination treatment of rapamycin and cabozantinib in treating HCC patients, especially for cMet inhibitor-resistant HCC patients.
PDF
(9161KB)
