MEK inhibition by cobimetinib suppresses hepatocellular carcinoma and angiogenesis in vitro and in vivo

a b s t r a c t
Hepatocellular carcinoma which is featured with the extensive vascularization is the third most frequent cause of cancer-related deaths with limited therapeutic options, particularly for advanced disease. Cobimetinib, a MEK inhibitor, has been approved for the treatment of melanomas with a BRAF mutation. In this work, we investigated the efficacy of cobimetinib in sensitive and resistant HCC cells. Using a panel of HCC cell lines and normal hepatocellular cells as control, we showed that cobimetinib is active against HCC cells and spare normal hepatocellular cells. Cobimetinib at nanomolar concentration inhibited proliferation and induced apoptosis in sorafenib-resistant HCC cells (Hep3B-r), suggesting its ability to overcome HCC resistance to standard of care. This was further demonstrated by our results that cobimetinib significantly augmented the inhibitory effects of sorafenib and doxorubicin in HCC cells. Notably, cobimetinib dose-dependently inhibited tumor angiogenesis by inhibiting HCC endothelial cell (HCCEC) growth, survival and capillary network work formation. Cobimetinib suppressed ERK/RSK without affecting JNK or p38 signaling pathways in Hep3B-r and HCCEC cells. In addition, cobimetinib negatively influenced the apoptosis pathways by increasing pro-apoptotic protein Bim and decreasing anti-apoptotic proteins Mcl-1 and Bcl-2. In addition, we validated the in vitro findings in HCC xenograft mouse model and demonstrated that cobimetinib inhibited ERK signaling, promoted apoptosis, and was active against resistant HCC growth and angiogenesis in vivo, without causing significant toxicity in mice. Our findings support the clinical trials of cobimetinib for HCC treatment and highlight the therapeutic value of inhibiting MEK/ERK/RSK to overcome HCC resistance.

1.Introduction
Hepatocellular carcinoma (HCC) is the third most common leading cause of cancer-related death in China and the incident of HCC is increasing [1]. The current treatments for HCC include sur- gical removal, transplantation, chemotherapy, radiation and tar- geted therapy [2]. However, most HCC patients were diagnosed at late stage when surgical resection and liver transplantation are not suitable [3,4]. Chemo- or radio-therapy is ineffective and the first line targeted agent sorafenib demonstrates minimal efficacy in advanced HCC [5]. Since HCC is characterized with the extensivevascularization, inter- and intra-tumor heterogeneity and aberrant activation of oncogenic pathways such as mitogen-activated pro- tein kinase (MAPK) pathway (RAF/MEK/ERK) [6,7], we hypothe- sized that agents targeting both endothelial cells and tumor cells might be effective in inhibiting HCC.BRAFV600E mutation was found in 14% of HCC that leads to activation of the RAF/MEK/ERK signaling pathway, inducing un- controlled cell proliferation [8]. Cobimetinib is a potent and highly selective MEK inhibitor which is FDA-approved for BRAF-mutated melanoma in combination with the BRAF inhibitor vemurafenib [9]. Apart from melanoma, several pre-clinical studies have shown the anti-cancer activity of cobimetinib as single drug or in combi- nation in other types of cancers, including colorectal cancer [10], high-grade serous ovarian cancer [11], neuroblastoma [12] and acute myeloid leukemia [13]. MEK1 inhibition has been identified as the mechanisms of the action of cobimetinib in cancer. Inhibition of MEK has been shown to suppress HCC and overcome HCC-resistance to sorafenib [14,15]. However, whether cobimetinib is effective in HCC is unknown.In this work, we determined the effect of cobimetinib on HCC cells and angiogenesis using in vitro cell culture and in vivo xeno- graft mouse model. We also analyzed the underlying mechanism of cobimetinib focusing on MEK/ERK pathway. We are the first to show that cobimetinib is effective in sensitizing HCC to sorafenib through targeting both tumor cells and angiogenesis regardless of BRAF mutation via inhibiting MEK/ERK/RSK.

2.Materials and methods
Two human normal hepatic cell lines (LO2 and THLE2) and six human HCC cell lines were maintained using the same culturing media as previously described in our study [16]. Cobimetinib and sorafenib were purchased from Selleckchem (Munich, Germany), doxorubicin (Sigma, USA). In all experiments performed on cell cultures, the DMSO concentration as control was 0.1% (v/v). Cell viability were determined using the MTT (3-(4,5-dimethylthiazol- 2-yl)-2,5-diphenyltetrazolium bromide) Assay Kit (Abcam, USA). Proliferation was determined using BrdU Cell Proliferation Assay Kit (Abcam, USA). Apoptotic cells were labelled with Annexin V and quantified using flow cytometer as the same described in our study [16]. Hep3B-r cell line was established by culturing Hep3B cells in the presence of gradually increased concentrations of sorafenib. The dose started with 0.5 mM, the next dose was increased by 2-fold and was given until the cells were stable in proliferation without significant death.Primary human umbilical vein endothelial cells (HUVEC) were purchased from Lonza Inc. USA. With written informed consent from HCC patients, human hepatocellular carcinoma endothelial cells (HCCEC) were isolated immediately from surgical HCC speci- mens after removal. HCC tissues were minced with scissors and then incubated with high-glucose Dulbecco’s Modified Eagle Me-dium containing 0.1% collagenase IV (Sigma, USA) at 37 ◦C.

After1.5 h collagenase digestion, the cell suspension was filtered through a graded series of meshes to isolate the cell components. HCCEC were purified from cell components using anti-CD105 antibody and then magnetic beads-based cell-sorting using the MoFlo XDP cell sorter (Beckman coulter, USA). Further around of purification using anti-CD105 antibody was performed and the purity of HUVEC cells were confirmed with flow cytometry. Primary HUVEC and HCCEC were maintained in CSC complete medium with serum and growth factors (Cell Systems, USA).Matrigel Matrix (Corning Inc, USA) were thawed overnight at 4 ◦C on ice and then added to 96-well plate and incubated at 37 ◦C for 1 h. After Matrigel Matrix solidification, endothelial cell sus-pension with drugs at different concentrations were added to the well. The plates were incubated for 8 h at 37 ◦C, 5% CO2 atmosphere. Capillary network were formed and observed under an inverted microscope (Zeiss Axiovert 200). Image J software was used to quantify the capillary length.Total proteins were isolated from cells and frozen tumor tissuesusing RIPA buffer (radioimmunoprecipitation assay buffer) con- taining protease and phosphatase inhibitors (Thermo Scientific, USA) according to standard protocol of WB samples. Protein con- centration was determined using the bicinchoninic acid protein assay kit (Thermo Scientific, USA). Proteins were resolved using denaturing SDSePAGE and analyzed by WB. Antibodies used in WB analyses were purchased from Cell Signaling Technology and are against p-ERK (Thr202/Tyr204, #4376), ERK (#9102), p-JNK (Thr183/Tyr185, #4511), JNK (#9258), p-90RSK (Thr359/Ser363,#9344), 90RSK (#9355), Bim (#2933), Mcl-1 (#4572), Bcl-2(#3498) and b-actin (#4967).

Image J software was used to quan- tify the band density.All procedures were conducted according to the guidelines approved by the Institutional Animal Care and Use Committee. Hep3B-r cells at five million in a volume of 100 ml PBS were sub- cutaneously injected into the flank of 6-week-old mice. Cobimeti- nib at 1 mg/kg/day was given by oral gavage to mice when tumor volume reached approximately 200 mm3 (n ¼ 5). Mice body weight and tumor size was monitored every two days. Tumor length andwidth were measured with a caliper and the volumes were calcu- lated using the formula: length x width2 x 0.5236. After 21 days drug treatment, mice were euthanized and tumors were isolated.Frozen tumor sections were immunostained with primary antibody against CD34 (Cell Signaling, USA) at 4 ◦C for overnight and furtherstained with fluorescence secondary antibody. The cell nuclei were counterstained with DAPI (Vector Laboratories, USA). The staining photos were imaged under Zeiss LSM510 confocal microscope and quantified using 5 randomly fields per sample.Statistical analyses between groups of cell assays were per- formed by unpaired Student’s t-test. Statistical analyses of the dif- ferences between two groups of samples were performed using the one-way analysis of variance (ANOVA) and subsequently by un- paired Student’s t-test. P-value < 0.05 was defined as statistically significant.

3.Results
To determine whether cobimetinib is active against HCC cells, we examined the effects of cobimetinib using not only sensitive but also resistant HCC cells to standard of care drug. Six parental cell lines which are representatives of in vitro HCC models presenting different cellular origin, genetic background and HBV infection status [17] were used as sensitive HCC cells. Two normal hepato- cellular cell lines LO2 and THLE-2 were used as normal control. We found that cobimetinib at concentration ranging from 50 nM to 400 nM significantly decreased viability of all six sensitive HCC cell lines in a dose dependent manner (Fig. 1A). In addition, cobimetinib at the same concentrations were less effective in decreasing viability of normal hepatocellular cells (Fig. 1A), suggesting that cobimetinib has selective anti-HCC activity.We next generated sorafenib-resistant Hep3B-r cells by pro-longed exposure of Hep3B cells to sorafenib. We showed that Hep3B-r exhibited significantly higher resistance to sorafenib, with the IC50 in Hep3B-r at least 8-fold higher than that in Hep3B (Fig. 1C). Notably, cobimetinib was also effective in decreasing proliferation and inducing apoptosis in Hep3B-r cells (Fig. 1D andE), suggesting the ability of cobimetinib to overcome HCC resis- tance to sorafenib. To confirm this finding, we further investigated the combinatory effects of cobimetinib with sorafenib or doxoru- bicin.

We found that the combination resulted in greater efficacy than sorafenib or doxorubicin alone in inhibiting proliferation and inducing apoptosis in Hep3B, HuH7 and SK-HEP-1 cells (Fig. 1F and G). Taken together, these results demonstrate the selective anti- HCC activity of cobimetinib and its ability overcome HCC resis- tance to standard of care drugs.It is well known that angiogenesis plays a vital role in HCC progression and metastasis and most currently approved treat- ments for advanced HCC in the first- and second-line settings target angiogenic pathways [18]. We also investigated the effects of cobimetinib in angiogenesis using two types of endothelial cells: HUVEC which is isolated from umbilical cord vein and HCCEC which is isolated from HCC specimen. We performed in vitro angiogenesis assay by plating endothelial cells onto the Matrigel Matrix which are enriched rich in extracellular matrix proteins (eg, collagens and laminin), endothelial cell growth factors and cyto- kines that support endothelial cells to differentiate and form capillary network [19]. As shown in Fig. 2A control, both HUVEC and HCCEC were able to form tubular structure on the MatrigelMatrix within 8 h of incubation. However, hardly any tube forma- tion was observed in the presence of cobimetinib. Quantification of capillary length indicated that cobimetinib at 100e400 nM signif- icantly inhibited HUVEC and HCCEC tube formation in a dose- dependent manner (Fig. 2B). We further found that cobimetinib also decreased growth and increased apoptosis in endothelial cells (Fig. 2C and D).Cobimetinib is a MEK inhibitor and its anti-cancer activities have been identified to attribute to the inhibition of MEK pathway [12,20].

To investigate whether MEK inhibition is the mechanism of the action of cobimetinib in overcoming HCC resistance and inhibiting angiogenesis, we examined the phosphorylation state of ERK, a MEK downstream effector, in Hep3B-r and HCCEC cells. We found that cobimetinib dose-dependently decreased the phos- phorylation of ERK and its downstream effector RSK, demon- strating the inhibition of ERK signaling by cobimetinib in Hep3B-r and HCCEC cells (Fig. 3). In contrast, cobimetinib did not change the level of p-JNK and p-p38 in Hep3B-r and HCCEC cells (Fig. 3), suggesting that other signaling pathways, including the JNK and p38, were not affected by cobimetinib. Furthermore, pro-apoptotic molecule Bim was increased and anti-apoptotic molecules Mcl-1and Bcl-2 were decreased in Hep3B-r and HCCEC cells exposed to cobimetinib (Fig. 3).We next validated our in vitro data using resistant HCC xenograftmouse model. We subcutaneously implanted Hep3B-r cells into the flank of SCID mice to form tumor. We monitored the mice body weight and tumor size throughout the duration of cobimetinib treatment. We examined the blood vessel density and level of p- ERK, p-RSK, Bim and Mcl-1 on Hep3B-r tumors after cobimetinib treatment. We did not observe body weight change in mice receiving cobimetinib by oral gavage at 1 mg/kg/day (Fig. 4A),suggesting that mice tolerate the treatment well. In contrast, cobimetinib significantly delayed Hep3B-r tumor growth beginning at 6 days of the initial treatment (Fig. 4B). Immunofluorescence staining of tumor tissue sections with anti-CD31 antibody (which stains tumor endothelial cell) revealed a decreased tumor vascu- larization compared to control tumors (Fig. 4C). The average tumor vascular density of each tumor group, which included all lumenand non-lumen CD34 + structures, was decreased by more than70% in groups treated with cobimetinib (Fig. 4D). Western blot analysis indicated that cobimetinib decreased the level of p-ERK, p- RSK and Mcl-1, and increased Bim level (Fig. 4E). Taken together, these results clearly indicate that cobimetinib inhibits ERK signaling, promotes apoptosis, and is active against resistant HCC growth and angiogenesis in vivo.

4.Discussion
Sorafenib is the standard of care for patients with advanced unresectable HCC. Unfortunately, after an initial satisfactory response to sorafenib, patients develop resistance and succumb to the disease [21]. Although sorafenib inhibits the activity of several tyrosine kinases including vascular endothelial growth factor re- ceptor (VEGFR-2/3), platelet-derived growth factor receptor (PDGF- R), Flt3 and c-Kit, and also targets Raf kinases involved in the MAPK/ ERK pathway, increasing evidence suggests that the activation of an escape pathway from RAF/MEK/ERK results in resistance [22]. In- hibition of RAF kinases can induce a dose-dependent “paradoxical” upregulation of the downstream MAPK/ERK pathway in cancer cells [23]. In this work, we demonstrate that MEK inhibition by cobimetinib is effective in overcoming resistance to not only sorafenib but also doxorubicin. Cobimetinib was active against all the HCC cell lines we tested with IC50 at nanomolar concentration range regardless of cellular origin and genetic profile (Fig. 1A). SK-HEP-1 is the most sensitive cell line to cobimetinib compared to other HCC cell lines. This is expected because SK-HEP-1 harbors BRAFV660E mutation which leads to constitutively active MEK/ERK pathway and therefore is more sensitive to MEK inhibitor than BRAFWT cell lines [15].

In addition, cobimetinib display less inhibitory activity in normal than malignant liver cells (Fig. 1A), suggesting that therapeutic window of cobimetinib in HCC. Using two different approaches, we demonstrate that cobimetinib at nanomolar concentration range is also active against sorafenib- and doxorubicin-resistant HCC cells (Fig. 1BeF), suggesting that cobimetinib has potential to overcome HCC resistance to standard of care drugs. The above results are consistent with our in vivo findings that cobimetinib at nontoxic dose effectively inhibits sorafenib-resistant HCC growth in mice (Fig. 4A and B). Cobimetinib is newly approved MEK inhibitor for the treatment of metastatic melanoma with BRAF mutation. The efficacy of cobimetinib has been investigated in other cancers using cell culture and mouse models [10e13]. Our pre-clinical findings support the anti-cancer activity of cobimetinib and add HCC to the growing list of cobimetinib-targeted cancers. Furthermore, we show that cobimetinib at nanomolar concentration range inhibits capillary network formation of HCCEC and HUVEC on Matrigel matrix, and suppresses growth and survival of endothelial cells (Fig. 2). This is important because angiogenesis plays a critical role in HCC progression and targeting angiogenesis by sorafenib does show clinical improvement in patients with advanced HCC [21]. We are the first to show the anti-angiogenic activity of cobimetinib in vivo (Fig. 4C and D), which makes cobi- metinib has advantage over many other anti-cancer agents that only target tumor cells alone.We confirmed that MEK/ERK inhibition is the mechanism of the action of cobimetinib in HCC and tumor angiogenesis. We observed the decreased phosphorylation of ERK and p90RSK in both sorafenib-resistant and HCCEC (Fig. 3). RAF/MEK/ERK pathway can regulate the activity and expression of members of the BCL-2 protein family to promote cell survival [24]. This is consistent with our findings that cobimetinib increase Bim and decreases Bcl- 2 level (Figs. 3 and 4E). ERK activation and signaling plays a key role in the resistance of chemotherapy and sorafenib [15,25]. We show that ERK inhibition by cobimetinib is an alternative therapeutic strategy to sensitize HCC to standard of care.

In conclusion, our work demonstrates the selective anti-HCC activity and anti-angiogenic activity of cobimetinib and highlights the potential of cobimetinib to overcome sorafenib resistance, through MEK/ERK inhibition. A large number of phase I, II or III clinical trials are currently evaluating the dose, toxicity or efficacy of cobimetinib as combination with other anti-cancer agents in ovarian cancer (ClinicalTrials.gov No: NCT03695380), gastrointes- tinal cancer (ClinicalTrials.gov No: NCT02876224), triple-negative breast cancer (ClinicalTrials.gov No: NCT02322814) and many other cancers. In line with these efforts, the findings in our work will accelerate the initialization of clinical trial to evaluate the po- tential of cobimetinib for HCC treatment.