Preclinical assessment of the VEGFR inhibitor axitinib as a therapeutic agent for epithelial ovarian cancer

Axitinib, small molecule tyrosine kinase inhibitor, demonstrates anti-cancer activity for various solid tumors. We investigated anti-cancer effect of axitinib in epithelial ovarian cancer (EOC). We treated EOC cells (A2780, HeyA8, RMG1, and HeyA8-MDR) with axitinib to evaluate its effects on cell viabilty, apoptosis and migration. Western blots were performed to assess VEGFR2, ERK, and AKT levels, and ELISA and FACS to evaluate apoptosis according to axitinib treatment. In addition, in vivo experiments in xenografts using A2780, RMG1, and HeyA8-MDR cell lines were performed. We repeated the experiment with patient-derived xenograft models (PDX) of EOC. Axitinib significantly inhibited cell survival and migration, and increased apoptosis in EOC cells. The expression of VEGFR2 and phosphorylation of AKT and ERK in A2780, RMG1, and HeyA8 were decreased with axitinib treatment in dose-dependent manner, but not in HeyA8-MDR. In in vivo experiments, axitinib significantly decreased tumor weight in xenograft models of drug-sensitive (A2780), and clear cell carcinoma (RMG1) and PDX models for platinum sensitive EOC compared to control, but was not effective in drug-resistant cell line (HeyA8-MDR) or heavily pretreated refractory PDX model. Axitinib showed significant anti-cancer effects in drug-sensitive or clear cell EOC cells via inhibition of VEGFR signals associated with cell proliferation, apoptosis and migration, but not in drug-resistant cells.

with chemotherapeutic agents, and has become a standard therapy for EOC in selected patients 3 . However, the overall survival benefit of bevacizumab seems insignificant considering its high medical expense. In addition, relapse after bevacizumab treatment suggests that there remains a need for alternative, potent, and multiple-target agents to counter tumor escape mechanisms.
Axitinib is a highly selective inhibitor of vascular endothelial growth factor receptor (VEGFR) tyrosine kinase 1, 2, and 3, and is reported to have the potential to control tumors and metastases by inhibiting angiogenesis and lymphangiogenesis, as well as via effects on tumor cells by apoptosis 4 . Clinical studies demonstrated promising anti-cancer activity in phase 2 trials for the treatment of various solid tumors. Axitinib showed single agent activity in patients with thyroid cancer 5 , nasopharyngeal cancer 6 , and resulted in improved response rate in recurrent glioblastoma patients 7 . Axitinib also significantly elongated PFS compared with sorafenib in patients with persistent renal cell carcinoma (RCC) 8 . Combination immunotherapy plus axitinib for the treatment of RCC resulted in encouraging antitumor activity 9 . However, the effects of axitinib in EOC have not been investigated.
The purpose of this study was to evaluate the anti-cancer effects of axinitib in EOC by using cell line xenografts and patient-derived xenograft (PDX) models, and to investigate the possible underlying mechanisms.

Axitinib inhibits VEGFR2, AKT and ERK pathways in EOC cells.
To evaluate the anti-cancer mechanism of axitinib in EOC cells, we assessed the VEGFR2, AKT and ERK pathways with Western blot. In experiments examining cell proliferation and apoptosis, axitinib was effective in A2780, RMG1, and HeyA8, and relatively ineffective in HeyA8-MDR. These cell lines were compared to determine the mechanism of action. Treatment with various doses of axitinib markedly decreased the expression of phospho-VEGFR2 in A2780, RMG1 and HeyA8 in a dose-dependent manner (Fig. 3A,B,C), but not in HeyA8-MDR cells (Fig. 3D). Phosphorylation of AKT, and ERK, a direct binding partner of VEGFR2, was also examined. Phosphorylation of AKT, and ERK was inhibited 4 h after axitinib treatment in A2780, RMG1 and HeyA8 cells, but this change was not observed in HeyA8-MDR or drug-resistant EOC cells.  Apoptosis assay-active caspase 3 ELISA and FACS. Active caspase-3 ELISA (A-D) and flow cytometric determination (E) showed increased cell apoptosis in axitinib-treated cell lines. The significance of differences was determined by unpaired t-tests, and values of P < 0.05 (*) or P < 0.01 (**) were considered to be statistically significant. www.nature.com/scientificreports www.nature.com/scientificreports/ Axitinib inhibits cell migration in EOC cells. Based on the results of a study of fetal lung adenocarcinoma showing that axitinib affects cell migration 10 , we performed cell migration assays in EOC cells. These assays revealed that axitinib-treated (24 h in 2 and 4 uM) EOC cells were less proficient at migrating than controls, with less absorbance observed at 560 nm (Fig. 4). In axitinib treatment groups, the number of migrated cells significantly decreased per x200 fields in A2780 (p = 0.0022) and HeyA8 (p = 0.0022), but not in HeyA8-MDR cells.

Axitinib significantly inhibits tumor growth in cell line orthotopic xenografts of EOC.
To investigate the clinical relevance of our in vitro results, we conducted in vivo experiments using EOC orthotopic mouse models. A2780, RMG1, and HeyA8-MDR EOC cells were implanted into the peritoneal cavities of female nude mice, and therapy was started with axitinib (30 mg/kg twice daily p.o.) 7 days after cell injection. In A2780 and RMG1 models, the tumor weight of the axitinib-treated group had significantly decreased by 50% compared with controls ( Fig. 5A,B, p = 0.0078, and p = 0.0379, respectively), but the difference was not significant in HeyA8-MDR models (Fig. 5C). Daily monitoring of animals throughout the therapy showed acceptable tolerability with no untoward side effects such as changes in body weight, mobility, posture, or feeding habits.
To validate the results of in vitro studies, we evaluated the effects of axitinib therapy on cell proliferation and apoptosis by immunohistochemistry for Ki-67 staining and TUNEL assays, respectively. Also, effects of axitinib on angiogenesis were evaluated by immunohistochemistry for CD31. The numbers of Ki-67 positive cancer cells were significantly lower in tumors from mice treated with axitinib than in tumors from controls in A2780 and RMG1 (Fig. 5D,E, p < 0.0001, and p < 0.0001, respectively), but not in a HeyA8-MDR mouse model (Fig. 5F). TUNEL assays showed that the number of apoptotic cancer cells was significantly higher in A2780 and RMG1 mouse models following therapy with axitinib. However, in the HeyA8-MDR cell line, differences between Ki 67 positive cells and TUNEL positive cells were insignificant. In the axitinib treated group, p-VEGFR2 positive cells were decreased in A2780 and RMG1 cell lines, but not in HeyA8-MDR. Number of vessels by CD31were significantly decreased in axitinib treated group of A2780 and RMG1 cell lines, but not in HeyA8-MDR.

Axitinib inhibits tumor growth in EOC PDX models.
We also examined the effects of axitinib in PDX models of EOCs using subrenal implantation of human EOC tissue. Our group previously developed PDX models of EOC 11 We selected case numbers OV-89-M6, platinum-sensitive high grade serous carcinoma, OV-64-M9, clear cell carcinoma, and OV-40-M7, platinum-resistant recurrent high grade serous carcinoma. OV-89-M6 was a 53-year-old patient with FIGO stage IIIA2. She was treated with primary cytoreductive surgery followed by paclitaxel-carboplatin combination chemotherapy. There was no residual tumor after primary surgery, and her PFS was 28 months. OV-64-M9 was a 42-year-old patient with stage IIIC clear cell carcinoma with <1 cm residual disease after primary surgery. Progression of disease was detected during first-line chemotherapy consisting www.nature.com/scientificreports www.nature.com/scientificreports/ of paclitaxel-carboplatin, and the patient's overall survival was only 2.4 months. OV-40-M7 was a 61-year-old patient with stage IV high grade serous carcinoma. The residual disease status after primary surgery was less than 1 cm, and the patient underwent 6 cycles of adjuvant paclitaxel-carboplatin combination chemotherapy. This case was classified as platinum resistant, as disease recurred after 6 months from end of first-line of chemotherapy.
Treatment with axitinib significantly decreased tumor weight in two PDX models compared with the control group (P = 0.0005 for OV-89-M6 and P < 0.0001 for OV-64-M9, respectively) ( Fig. 6A,B). The inhibitory effect of axitinib on tumor growth was not seen in the OV-40-M7 model, which is heavily-pretreated and platinum-resistant (Fig. 6C). Immunohistochemistry staining of Ki-67, p-VEGFR2, and TUNEL assay yielded similar results to those obtained for the xenograft model. Significantly higher numbers of TUNEL positive cells, and lower numbers of Ki-67 positive cells, were observed with axitinib treatment in high grade serous and clear cell carcinoma PDX (Fig. 6D,E, p < 0.0001). In platinum resistant ovarian cancer PDX, differences between controls and the axitinib-treated group for Ki-67 positive cells and apoptotic cells were not significant (Fig. 6F). In addition, in the axitinib-treated group, the number of p-VEGFR2 positive cells was lower in platinum sensitive high grade serous and clear cell carcinoma PDX, but this difference was not observed in platinum resistant high grade serous cases. Also, in the axitinib-treated group, the number of vessels by CD31 was decreased in platinum sensitive high grade serous and clear cell carcinoma PDX, but this difference was not observed in platinum resistant high grade serous cases.

Discussion
Axitinib, a highly selective VEGFR tyrosine kinase inhibitor, is known as one of the most effective substance for the metastatic renal cell carcinoma treatment 12 . In this study, we exhibited anti-angiogenesis and anti-tumor activity of axitinib that has potential for use in the treatment of EOC. In an in vitro study, axitinib significantly inhibited proliferation and migration, and increased apoptosis, of EOC cells in a dose-dependent manner. Initially, cell viability experiments presented that axitinib showed cytotoxic activity in all EOC cells. In addition, axitinib-induced apoptosis was confirmed in EOC cell lines. However, in Western blot confirming expression of VEGFR and its downstream signaling in EOC cell lines, axitinib-induced inhibitory effects in VEGFR2, phosphorylation of AKT, and ERK were not observed in HeyA8-MDR. Unlike A2780 and HeyA8, the migration assay showed no effect of axitinib on HeyA8-MDR. Based on these results, we hypothesize that axitinib inhibits EOC cells by targeting multiple pathways including angiogenesis, AKT, and ERK signaling pathways. Additionally, invasion-related MMP2/ MMP9 ELISA was performed for further explanation of the differences between cell viability experiments and cell signal assay, but no differences were found in HeyA8-MDR and other cell lines. This may be due to differences in tumor microenvironment, but the exact mechanism through the experiment could not be presented. In orthotopic mouse models, tumor reduction by axitinib in platinum sensitive cell line and Figure 5. In vivo EOC cell line mouse models. Axitinib inhibits the tumor growth of ovarian cancer xenografts. Mice treated with axitinib had significantly lower tumor weight than control mice (by 50%; P < 0.005 in A2780 and RMG1), but the difference was not significant in drug-resistant EOC models (HeyA8-MDR). The expression of apoptosis,cell proliferation, and angiogenesis in xenografts was also analyzed by IHC with p-VEGFR2, TUNEL assay, Ki-67 staining, and CD31 staining. (2020) 10:4904 | https://doi.org/10.1038/s41598-020-61871-w www.nature.com/scientificreports www.nature.com/scientificreports/ ovarian clear cell line was observed, but axitinib was not effective in tumor of drug resistant cell line. The same result was observed in the PDX model. Thus, in our experiments, we concluded that antitumor effects of axitinib as a single agent were significant only in drug-sensitive EOC models, but were not remarkable in drug-resistant EOC cell line xenograft or PDX models.
In our study, axitinib showed weaker effects in the platinum-resistant group than the platinum-sensitive group. Drug resistance can be explained by changes of intracellular active drug concentrations, drug-target interactions, target-mediated cell damage, damage-induced apoptotic signaling, or apoptotic effectors 13 , which may influence response to axitinib. In a previous study analyzing the biological characteristics of platinum-resistant cells 14 , resistant cell lines exhibited decreased levels of DNA platination and faster repair of damaged DNA, suggesting that drug uptake, detoxification, and excretion, along with the DNA repair pathway play central roles in resistant phenotypes. In the current study, axitinib was effective for treatment against a drug-resistant cell line (HeyA8-MDR) in cell viability assay and apoptosis assay, but not in cell signal assay and in vivo experiments with a drug-resistant xenograft model. We additionally retried Western blot to confirm expression of VEGFR and its downstream signaling for HeyA8 and HeyA8-MDR. Expression of phospho-VEGFR2, phosphorylation of AKT, and ERK was inhibited 4 h after axitinib treatment in HeyA8 cells, but in contrast, this change was not observed in HeyA8-MDR. To explain the difference between result of cell viability test and cell signal analysis, we performed invasion-related MMP2/9 ELISA in HeyA8, and HeyA8-MDR. However, result for MMP2 was not measured for HeyA8-MDR, and expression inhibition of MMP2 by axitinib showed no difference between HeyA8 and HeyA8-MDR as shown in Supplementary Data. Unfortunately, these additional experiments could not reveal a clear mechanism. These differences may be explained by differences in reactions between cell lines and tissue. In cancer research, in vitro experiments are mainly performed to study gene regulation and signaling that lead to uncontrolled cell growth. In vivo experiments are performed to evaluate cancer cell interactions with the environment, and result in more informative outcomes because the microenvironment is a critical determinant of the migration strategy and the efficiency of cancer cell invasion 15 .
VEGF-mediated angiogenesis plays an important role in ovarian function, and there is a well-investigated association between VEGF overexpression, increased angiogenesis, and the development and progression of ovarian cancer 16 . Previously in clinical studies, high serum VEGF levels were correlated with higher risks of recurrence and death of EOC in a review of nine studies including 529 EOC patients 17 . Serum VEGF was considered an independent prognostic factor for survival after multivariate analysis in five studies. Associations between increased angiogenesis and progression of EOC led to the investigation of a number of anti-angiogenic agents as potential treatment options for EOC. Bevacizumab gained approval for first-line treatment for advanced EOC patients, and is included in the National Comprehensive Cancer Network (NCCN) guidelines for EOC treatment. Other anti-angiogenic agents, including trebananib, aflibercept, nintedanib, cediranib, imatinib, pazopanib, sorafenib and sunitinib, are currently in phase II/III development 18 . However, the effects of axitinib, part of a new Figure 6. In vivo EOC PDX models. Axitinib inhibits tumor growth of ovarian cancer xenografts. Mice treated with axitinib had significantly lower tumor weight than control mice (P = 0.007 for OV-89-M5 and P < 0.0001 for OV-64-M9), but the effect was not significant in platinum-resistant OV-40-M7. The expression of apoptosis and cell proliferation in these xenografts was analyzed by IHC with p-VEGFR2(x400), Ki-67 staining, CD31 staining, and TUNEL assay (x200).
Strategies for inhibiting angiogenesis are key to prevent the survival, proliferation, invasion, and metastasis of ovarian cancer cells 20 . In this study, we found that axitinib is effective as a single agent in a drug-sensitive EOC cell line mouse and PDX models, but not in drug-resistant EOC cell models. In clinical trials using anti-angiogenic agents for the treatment of EOC, the response rate was not significant in drug-resistant recurrent EOC. A study of other VEGFR inhibitor, sorafenib, showed no anti-tumor activity in patients with possibly drug resistant EOC or primary peritoneal carcinoma after multiple use of chemotherapy 21 . Previous study of bevacizumab as a single agent in patients with platinum-resistant relapsed EOC and peritoneal serous carcinoma also showed poor response (response rate 15.9% (7/44) and median response duration 4.2 months (range, 1.7 to 9.2 months)) 22 .
To the best of our knowledge, this is the first assessment of the efficacy and mechanism of axitinib as an anti-cancer therapeutic in preclinical models of EOC. In this study, we demonstrated marked anti-tumor effects of axitinib that were associated with anti-angiogenesis in drug-sensitive EOC cells, xenograft, and PDX models. Our findings have important clinical implications for the administration of axitinib as a single agent in the treatment of drug-sensitive EOC patients who are highly likely to experience toxicity if treated with conventional taxaneand platinum-based chemotherapy. However, the effects of axitinib were not promising against drug-resistant EOC, so clinical trials evaluating combination therapies of axitinib with other target agents, including immunotherapy, are needed. In summary, axitinib may be one of the most promising VEGFR tyrosine kinase inhibitors available, exhibiting significant antitumor activity when used as a single agent for the treatment of EOC.

Materials and Methods
Chemicals and cell culture. Axitinib  Cell viability assay. Cells were plated in culture medium in 96-well plates at 3 × 10 3 cells/well. After 24 h, cells were treated with axitinib, and assays performed at 24, 48, and 72 h. For cell viability assays, cells were stained with 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Amresco, Solon, OH, USA); after 4 h of additional incubation, the medium was discarded, 100ul of acidic isopropanol (0.1 N HCL in absolute isopropanol) was added, and the plate was shaken gently. Absorbance was measured on an enzyme linked immunosorbent assay (ELISA) reader at a wavelength of 540 nm. Experiment was conducted as our previous study 23 . Active caspase-3 ELISA. For the apoptosis assay, we used an active caspase-3 ELISA assay (#KHO1091; Invitrogen). Cells were seeded in 6-well plates (1 × 10 4 cells in 3 ml of media per well), and incubated overnight to allow the cells to attach to the plate. After 24 h of treatment with 0, 1, 2, and 4 uM axitinib, the medium was removed by suction and cells were lysed with lysis buffer. Apoptotic activity was determined for each well according to the manufacturer's protocol, as described previously 24 . FACS analysis. Cell apoptosis was measured at 48 h after treatment using the FITC Annexin-V apoptosis Detection Kit-1 (BD Pharmingen, San Diego, CA, USA) according to the manufacturer's protocol. Each sample was assayed in triplicate. A minimum of 5,000 cells were then analyzed by FACScan with Cell Quest software (Beckton Dickinson) for acquisition and analysis, as described previously 23 . Western blot. Cells were lysed in PRO-PRE-Protein Extraction Solution (Intron Biotechnology, Seongnam, Korea). Protein concentrations were determined using a Bradford assay kit (BIO-RAD, Hercules, CA, USA). Cell lysates (50 μg of total protein) were separated in 8% acrylamide gels by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to Hybond-ECL nitrocellulose filter paper (Amersham Biosciences, Buckinghamshire, UK). Membranes were blocked with 5% BSA in Tris-buffered saline containing 0.1% Tween-20 for 1 h at room temperature. Protein bands were probed with VEGFR2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at a 1:500 dilution; phospho-VEGFR2 (Tyr951, Tyr1175), total-ERK (Thr202/Tyr204), phosphor-ERK (Thr202/Tyr204), total-AKT, phospho-AKT total-p38, phospho-p38 (Cell Signaling, USA) at 1:1000 dilutions; β-actin antibody at a 1:4000 dilution (Santa Cruz Biotechnology, Santa Cruz, USA) and then labeled with horseradish peroxidase-conjugated anti-rabbit antibody (GE Healthcare, Piscataway, USA). Bands were visualized by enhanced chemiluminescence using an ECL kit (Amersham Biosciences, Buckinghamshire, UK) according to the manufacturer's protocol, as described previously 23 .
Migration assay. The migration assay was performed with a Cytoselect 24-well cell migration kit according to the manufacturer's protocol (Cell Biolabs, San Diego, USA), as described previously 25 .

Animal care and development of in vivo models including established cell lines and PDX.
In vivo experiments were performed to confirm the anti-tumor effect of axitinib in orthotopic cell-lines or patient-derived xenograft (PDX) mouse models. Female BALB/c nude mice were purchased from ORIENT BIO (Sungnam, Korea). This study was performed in accordance with all relevant guidelines and regulations. This study was reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Samsung Biomedical Research Institute (SBRI). SBRI is an Association for Assessment and Accreditation of Laboratory