Potential Therapeutic Implication for Synovial Sarcoma Parag P

Published OnlineFirst August 30, 2018; DOI: 10.1158/1535-7163.MCT-18-0319

Small Molecule Therapeutics Molecular Cancer Therapeutics Preclinical Evaluation of Nintedanib, a Triple Angiokinase Inhibitor, in Soft-tissue Sarcoma: Potential Therapeutic Implication for Synovial Sarcoma Parag P. Patwardhan1, Elgilda Musi1, and Gary K. Schwartz1,2

Abstract

Sarcomas are rare cancers that make up about 1% of all preclinical efficacy of nintedanib in soft-tissue sarcoma cell cancers in adults; however, they occur more commonly among lines. Nintedanib treatment resulted in significant antiproli- children and young adolescents. Sarcomas are genetically ferative effect in vitro in cell lines with high expression of RTK complex and are often difficult to treat given the lack of clinical drug targets. Furthermore, treatment with nintedanib showed efficacy of any of the currently available therapies. Receptor significant downregulation of downstream phosphorylated tyrosine kinases (RTK) such as c-Kit, c-Met, PDGFR, IGF-1R, as AKT and ERK1/2. Finally, treatment with nintedanib resulted well as FGFR have all been reported to be involved in driving in significant tumor growth suppression in mouse xenograft tumor development and progression in adult and pediatric model of synovial sarcoma. Notably, both the in vitro and soft-tissue sarcoma. These driver kinases often act as critical in vivo efficacy of nintedanib was superiortothatofimatinib, determinants of tumor cell proliferation and targeting these another multikinase inhibitor, previously tested with min- signal transduction pathways remains an attractive therapeutic imal success in clinical trials in sarcoma. Overall, the data approach. Nintedanib, a potent triple angiokinase inhibitor, from this study provide a strong rationale to warrant further targets PDGFR, VEGFR, and FGFR pathways critical for tumor clinical exploration of this drug in patients with synovial angiogenesis and vasculature. In this study, we evaluated the sarcoma. Mol Cancer Ther; 17(11); 2329–40. 2018 AACR.

Introduction and migration (5). Role of RTKs such as IGF-1R, PDGFR, c-Met, AXL, and FGFR as well as nonreceptor tyrosine kinases (non- Sarcomas are rare cancers with subtypes that are often genet- RTK) such as FAK and Src acting as tumor drivers has been ically and biologically distinct (1, 2). Sarcomas make up about 1% reported in various sarcoma subtypes over the past few years of all cancers in adults with approximately 14,000 people diag- (6–11). However, the translation of these preclinical studies nosed with sarcoma per year in the United States. However, they into effective molecularly targeted therapies has met with only are relatively more common among children making up to 12% to modest clinical success (12, 13). Gastrointestinal stromal 15% of cancers in children under the age of 20 (3). Some of the tumor (GIST) is one of the few sarcoma subtypes where more common and molecularly heterogeneous sarcoma subtypes oncogenic mutations in PDGFRa and c-Kit (14, 15) have been include liposarcoma, leiomyosarcoma, and synovial sarcoma targeted successfully using tyrosine kinase inhibitors such as among others (4). Localized disease is usually surgically resected imatinib and crenolanib (16, 17).In addition, in patients with followed by radiotherapy with or without adjuvant chemother- well- and dedifferentiated liposarcomas carrying CDK4 ampli- apy. However, lack of clinical efficacy coupled with toxicity often fication have shown promising results including progression- limits the treatment options especially in metastatic cases of the free survival in clinical trials (18, 19). disease. Although targeted RTK therapy presents an attractive approach Receptor tyrosine kinases (RTK) actaskeymediatorsofsignal to treat this disease, lack of clinical success using traditional transduction pathways that regulate cell proliferation, survival, therapies warrants exploration of novel agents that block multiple kinases such as PDGFR, FGFR, and VEGFR known to be critical for driving sarcoma progression (8, 20–23). 1Department of Medicine, Columbia University Medical Center, New York, In this study, we tested the in vitro and in vivo antiproliferative 2 New York. Herbert Irving Comprehensive Cancer Center, Columbia University efficacy of nintedanib (BIBF1120), a potent PDGFR, VEGFR, and College of Medicine, New York, New York. FGFR inhibitor, in sarcoma. Nintedanib was recently approved Note: Supplementary data for this article are available at Molecular Cancer for the treatment of non–small-cell lung cancer in the Europian Therapeutics Online (http://mct.aacrjournals.org/). Union (EU). Efficacy of other multikinase inhibitors in vitro and Corresponding Author: Parag P. Patwardhan, Columbia University Medical in clinical trials in sarcoma has been reported previously by our Center, Herbert Irving Comprehensive Cancer Center, 1130 St. Nicholas Avenue, laboratory (24) as well as others (12). However, not many Room 207, New York, NY 10032. Phone: 212-851-4904; Fax: 212-851-4901; effective molecularly targeted therapies are currently available E-mail: [email protected] for patients with sarcoma including those with synovial sarcoma doi: 10.1158/1535-7163.MCT-18-0319 which has a propensity to metastasize in many cases. This is the 2018 American Association for Cancer Research. first report describing efficacy of nintedanib in this disease

www.aacrjournals.org 2329

Patwardhan et al.

setting and our results strongly warrant further clinical explora- and visualized by Enhanced Chemiluminescence Reagent (GE tion of this drug in synovial sarcoma. Healthcare).

Materials and Methods Xenograft studies Chemicals and drugs Approximately 8-week-old athymic female mice were injected – Nintedanib (BIBF1120) was kindly provided by Boehringer- with 10 15 million cells in Matrigel of the cell line of interest and 3 Ingelheim Pharmaceuticals. Imatinib was purchased from LC allowed to grow until tumors reached 100 mm in volume prior to Laboratories. Selumetinib (MEKi) and MK-2206 (AKTi) were treatment with vehicle, imatinib (30 mg/kg i.p. once daily), or purchased from Selleck Chemicals. Drugs were dissolved in nintedanib (50 mg/kg orally once daily). Tumors were measured DMSO and aliquots were stored at 20C for long-term use. every 2 to 3 days using calipers and volumes calculated using the formula p/6 x (large diameter) (small diameter) where p stands for "pi". Animal weights were measured every 2 to 3 days as a Cell culture and reagents surrogate marker for overall toxicity. Animals were sacrificed 4 Cells were cultured in RPMI or DMEM with 10% FBS, 100 m/mL hours after the final dose of treatment (on day 21 for vehicle penicillin, and 100 mg/mL streptomycin, maintained at 37 Cin control or day 24 for imatinib and nintedanib treatment groups). 5% CO , and passaged for no more than 3 months. Synovial 2 Tumors were extracted from surrounding tissue and either flash- sarcoma cell lines (SYO-1 and HSSY-II) were obtained from Dr. frozen in liquid nitrogen for Western immunoblotting or placed Marc Ladanyi (Memorial Sloan Kettering Cancer Center, New in formalin or paraformaldehyde for IHC studies. Flash-frozen York, NY). Malignant peripheral nerve sheath tumor (MPNST) tissue was ground in RIPA lysis buffer and resin as per the cell line was supplied by Dr. Jonathan Fletcher (Dana Farber manufacturer's instructions using Sample Grinding Kit (GE Cancer Institute, Boston, MA). ST8814 cell line was a generous gift Healthcare) and analyzed via Western immunoblotting as from Dr. Karen Cichowski (BWS Biomedical Research Institute, described above. Formalin-fixed tissues were sectioned and Boston, MA), Ewing sarcoma cell lines (CHP100, A673) were stained with the indicated antibodies following standard proto- obtained from Dr. Melinda S. Merchant (Center for Cancer cols using HistoWiz automated histology services. Research, NCI/NIH, Bethesda, MD). Dedifferentiated liposarcoma (LS141, DDLS) were obtained from Dr. Samuel Singer (Memorial Sloan Kettering Cancer Center, New York, NY). Oste- Statistical analysis in vitro osarcoma cell line (SaOS2) was obtained from ATCC. Cell lines All experiments were carried out at least three times P CHP100 and A673 were authenticated using RT-PCR, and found unless otherwise indicated. values were calculated using Student t to have their expected characteristic chromosomal translocations. test, with values of 0.05 determined to be statistically signif- SYO-1 and HS-SY-II cell lines were authenticated by confirming icant. SE was calculated as the SD divided by the square root of the the expression of the pathognomonic SYT-SSX fusion gene by RT- number of samples tested. PCR. All cell lines were determined to be Mycoplasma free by using biochemical assay MycoAlert. Results Target RTKs for nintedanib are expressed in multiple sarcoma Cell viability assays cell lines Cell viability assays were carried out using the Dojindo Nintedanib is a small-molecule inhibitor that targets multiple Molecular Technologies (CCK-8) Kit as per the manufacturer's RTKs involved in tumor angiogenesis such as VEGFR, PDGFR, and instructions. Briefly, 2,000 cells were plated in 96-well plates FGFR (25, 26). It is an indolinone derivative that has been shown and then treated with the indicated drugs for 6 days unless to block RTKs, such as VEGFR and PDGFR families at low indicated otherwise. Media was replaced with 100 mLofmedia nanomolar concentrations, and also block downstream MAPK with 10% serum and 10% CCK-8 solution (Dojindo Molecular and AKT signaling in pericytes and endothelial cells (25). To test Technologies Kit). After 1 hour, the optical density was read at the basal levels of expression for nintedanib drug targets, we 450 nm using a Spectra Max 340 PC (Molecular Devices carried out a Western blot analysis using 9 different sarcoma cell Corporation) to determine viability. Background values from lines belonging to various subtypes such as Ewing sarcoma (A673, negative control wells without cells were subtracted for final CHP100), liposarcoma (DDLS, LS141), malignant peripheral sample quantification. Cell viability is expressed as a percentage nerve sheath tumor (MPNST, ST8814), synovial sarcoma (SYO- of untreated/DMSO-treated cells. 1, HS-SY-II), and osteosarcoma (SaOS2). Figure 1A shows that the cell lines tested express variable levels of PDGFR, FGFR, and Western immunoblotting VEGFR family of RTKs with malignant peripheral nerve sheath Cells and tissues were lysed with radioimmunoprecipitation tumor (MPNST, ST8814) and synovial sarcoma cell lines (SYO-1, assay (RIPA) buffer supplemented with protease inhibitor HS-SY-II) showing expression of maximum number of nintedanib drug targets. cocktail tablets (Roche Diagnostics) and 1 mmol/L Na3VO4. Equal amounts (20–30 mg) of protein were electrophoresed onto 4%–12% gradient gels (Life Technologies) and transferred Nintedanib exhibits potent antiproliferative activity and onto polyvinylidene difluoride or 0.45-mm nitrocellulose mem- inhibition of downstream signaling in sarcoma cell lines branes. Membranes were blocked with 5% nonfat dried expressing both PDGFR and FGFR milk and probed with primary antibodies. A complete list of To test the antiproliferative efficacy of nintedanib in multiple antibodies that were used in this study is included in the sarcoma cell lines, we carried out a cell viability assay following 6 Supplementary Table. Bound antibodies were detected with days of exposure to increasing concentrations of the drug. Ninte- horseradish peroxidase secondary antibodies (GE Healthcare) danib treatment resulted in inhibition of cell proliferation

2330 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Triple Angiokinase Inhibition by Nintedanib in Sarcoma

Figure 1. Inhibition of cell proliferation and downstream signaling pathways by nintedanib in PDGFR and FGFR expressing sarcoma cell lines. A, Basal levels of target RTK expression in 9 sarcoma cell lines. Cells were plated in 100 mm tissue culture plates (Costar) and grown for 24 hours in media containing 10%FBS. Cell lysates were obtained by scraping in RIPA lysis buffer and 30 mg of lysates were run on SDS/PAGE gels. Western blot analysis was carried out using indicated antibodies. B, Indicated cell lines were plated in 96-well plates and treated in triplicate per condition with increasing doses of nintedanib for 6 days. Cell viability was measured using Dojindo Cell Counting Kit 8 according to the manufacturer's instructions. Dose response graphs were generated as a percentage of DMSO-treated control. Error bars represent SEM. Combined data from two replicates from a single experiment representative of three independent experiments are shown. C, Cell lines were grown to 60% conﬂuency in serum-free media overnight. Next day, cells were treated with indicated concentrations of nintedanib in serum- free media for 3 hours. Following drug treatment, cells were stimulated in drug-free media containing 10%FBS. 30 mg of RIPA lysates were loaded on SDS/PAGE and immunoblotted using indicated antibodies. Representative western blots from two independent experiments are shown.

especially in cell lines that express both PDGFR and FGFR nintedanib in 4 cell lines (selected on the basis oftheir anti- family of RTKs (Fig. 1B). As shown in Fig. 1B, inhibition of proliferative efficacy). Figure 1C shows that treatment with cell proliferation in the nanomolar range was observed for increasing doses of nintedanib results in significant inhibition MPNST, ST8814, as well as both synovial sarcoma cell lines of p-AKT in both SYO-1 and ST8814 cell line. Inhibition of (SYO-1, HS-SY-II) compared with the two Ewing sarcoma cell p-ERK1/2 was also observed for the cell line SYO-1 but no such lines (A673 and CHP100). Modest activity was observed at inhibition of p-ERK1/2 was seen in ST8814 cell line. This could micromolar concentrations for liposarcoma cell lines (DDLS, partly be due to the fact that ST8814 cell line has higher basal LS141) and for osteosarcoma cell line, SaOS2. SYO-1 and levels of p-ERK1/2,whicharedifficult to suppress owing to the ST8814 cell lines showed IC50 values < 1 mmol/L, MPNST, NF1 (Neurofibromatosis Type 1) loss and increased Ras activity HS-SY-II, and LS141 cell lines had IC50 values between 1–2 in this cell line (27). On the other hand, cell lines A673 and mmol/L, whereas, SaOS2, DDLS, A673, and CHP100 cell lines CHP100 showed no change in p-AKT and p-ERK1/2 levels after showed IC50 values > 5–10 mmol/L (Table 1). Approximate IC50 nintedanib treatment. This correlated with the observation that values for all the cell lines tested are summarized in Table 1. To both these cell lines did not show any significant inhibition of test for the inhibition of downstream signaling pathways, we proliferation even at micromolar concentrations of the drug carried out Western blot analysis with increasing dose of (Fig. 1B).

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2331

Patwardhan et al.

Table 1. Table showing approximate IC50 values for all the cell lines tested we tested the efficacy of MEK inhibitor (MEKi) selumetinib m Cell Line [IC50] mol/L (28) and AKT inhibitor (AKTi) MK-2206 (29) alone or in A673 >10 combination and compared the inhibitory effects against nin- > CHP100 10 tedanib and imatinib, another well-known c-Kit and PDGFR DDLS >10 LS141 1.266 inhibitor (30). Cell proliferation assay in Fig. 3A and B shows SaOS2 7.120 that inhibitory effects of nintedanib in synovial sarcoma cell SYO-1 0.525 line SYO-1 were significantly (P < 0.01 and P < 0.05, respec- HS-SY-II 1.397 tively) better than imatinib alone or a combination of MEKi MPNST 1.431 and AKTi. On the other hand, no significant inhibition of ST8814 0.734 proliferation was observed for the CHP100 (Ewing sarcoma) cell line (Fig. 3B). In ST8814 cell line, the inhibition of cell To test whether the drug treatment results in inhibition of target proliferation by a combination of MEKi and AKTi was signif- RTKs in various cell lines, we carried out phospho-RTK proteome icantly better than nintedanib treatment alone (Fig. 3A). For the profiler array analysis in the 4 cell lines tested above. Nintedanib A673 cell line, nintedanib treatment did not result in any treatment (at 500 nmol/L) resulted in significant downregulation significant inhibition of cell proliferation compared with no of RTK targets including PDGFRa and PDGFRb in all the cell lines drug control; however, a combination of MEKi and tested including A673 and CHP100 (Fig. 2). This suggests that AKTi treatment resulted in significant inhibition of prolifera- even though the phosphorylation of PDGFRb is blocked by tion (Fig. 3B). Western blot analysis after 24 hours of drug nintedanib, the lack of antiproliferative efficacy in the two Ewing treatment (Fig. 3C) to test pathway inhibition shows that MEKi sarcoma cell lines A673 and CHP100 is possibly due to the fact alone or in combination with AKTi did not result in suppres- that pathways other than PDGFR such as Ins-R and IGF-1R are sion of p-ERK1/2. Nintedanib treatment, on the other hand, driving cell proliferation. Although we were able to observe resulted in significant inhibition of p-AKT and p-ERK1/2 levels phosphorylated levels of RYK in SYO-1 and ST8814 cell lines, it in SYO-1 cell line. When nintedanib treatment was combined is highly unlikely that it plays any significant role in driving with MEKi, the combination treatment resulted in further sarcoma cell growth especially because it lacks any catalytic inhibition of cell proliferation (Fig. 3A) compared with ninte- activity of its own (24). danib alone (P < 0.05) as well as further inhibition of p-ERK1/2 In the SYO-1 cell line, which expressed high levels of VEGFR2 levels (Fig. 3C). No such downstream signaling pathway inhi- (Fig. 1A), we were unable to detect any phosphorylated VEGFR2 bition was observed in Ewing sarcoma cell lines, A673 and either by RTK array (Fig. 2) or by Western blot analysis (Supple- CHP100 (Fig. 3C). It must be noted though that when the cell mentary Fig. S1A). AXL, a well-characterized RTK, has been shown viability was assayed after 3 days of drug treatment, the inhi- to be induced as a result of resistance to imatinib in GISTs. To test bition of cell proliferation for nintedanib versus MEKi and AKTi whether nintedanib treatment results in induction of AXL, we combination treatment seemed comparable in the SYO-1 cell carried out Western blot analysis. Our results indicate that SYO-1 line (Supplementary Fig. S2). However, when the assay was cells lacked total as well as phosphorylated levels of AXL (Sup- performed after 6 days of drug exposure, nintedanib treatment plementary Fig. S1A). CHP100 cell line showed presence of total resulted in further inhibition of cell proliferation (Supplemen- AXL receptor on the Western blot analysis; however, phosphor- tary Fig. S2, viability at day 3 vs. day 6) compared with MEKi ylated AXL was unable to be detected irrespective of whether the and AKTi combination. cells were treated with nintedanib (Supplementary Fig. S1A). Next, we wanted to test whether nintedanib treatment resulted Antiproliferative activity of nintedanib is a result of combined in downregulation of non-RTKs such as SRC family kinases, a inhibition of both PDGFR and FGFR pathways known downstream target of nintedanib (25). Treatment with To test whether the inhibitory effects of nintedanib are medi- 500 nmol/L nintedanib did not result in activation of SRC family ated through inhibition of PDGFR and/or FGFR pathways, we kinases as determined by phospho-SRC Tyr416 Western blot carried out cell proliferation assay and Western blot analysis after analysis (Supplementary Fig. S1A). No changes in total levels of nintedanib or pan-FGFR inhibitor, PD173074 (31) treat- SRC were observed after nintedanib treatment. We also carried out ment. Figure 4A shows that inhibition by pan-FGFR inhibitor, analysis of non-RTK phosphorylation using non-RTK protein PD173074, resulted in significant (P < 0.001) inhibition of cell profiler array kit (R&D Systems). The majority of the non-RTKs proliferation compared with DMSO control. However, inhibition that were detected on the array in both SYO-1 and CHP100 cell of cell proliferation by nintedanib was significantly better than lines included JNK, GSK3b, c-Jun, PRAS40, and HSP60 kinases PD173074 (P < 0.01) especially at 500 and 1,000 nmol/L con- (Supplementary Fig. S1B). Phosphorylated FAK was detected only centrations (Fig. 4A). Western blot analysis (Fig. 4B) showed in the SYO-1 cell line; however, no significant changes in phos- that PD173074 treatment resulted in significant inhibition of phorylation of FAK were observed post-nintedanib treatment. In p-ERK1/2 signaling in SYO-1 cell line; however, no decrease in addition, in both the cell lines, no major change in phosphory- p-AKT signal was observed (Fig. 4B, lanes 4 and 5). Nintedanib lation of any of the non-RTKs was observed other than phosphor- treatment, on the other hand, showed significant reduction in ylated p53 in SYO-1 cell line (Supplementary Fig. S1B). We were p-AKT as well as p-ERK1/2 signaling at 500 nmol/L concentration unable to detect any phosphorylated SRC on the non-RTK array. (Fig. 4B, lanes 2 and 3). Results of nintedanib versus pan-FGFR inhibition in SYO-1 cell line strongly suggest a role for combined Antiproliferative activity of nintedanib is a result of sustained inhibition of PDGFR and FGFR in downregulating p-AKT signal- p-AKT and p-ERK1/2 inhibition ing compared with inhibition of FGFR activity alone in down- To determine whether the inhibitory effect of nintedanib is a regulating p-ERK1/2 signaling. No significant inhibition of cell result of combined inhibition of both p-AKT and p-ERK pathways, proliferation or any decrease in downstream p-AKT signaling was

2332 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Triple Angiokinase Inhibition by Nintedanib in Sarcoma

observed in CHP100 cell line (Fig. 4A and B). Interestingly, basal Pazopanib, another multitargeted kinase, was recently levels of p-ERK1/2 in CHP100 cell line were lower compared with approved for the treatment of soft-tissue sarcoma (32). To test SYO-1 cell line and showed only a modest decrease after ninte- whether nintedanib treatment fares better than pazopanib treat- danib or PD173074 treatment. ment in SYO-1, we carried out a cell viability assay and Western

Figure 2. Inhibition of phosphorylated RTK targets by nintedanib. Approximately, 106 cells were plated in 60 mm plates in serum-free media for 24 hours. Next day, DMSO or 500 nmol/L nintedanib was added in serum-free media and treated for 3 hours. Cells were then stimulated in drug-free media containing 10% FBS for 1 hour. Following stimulation, lysates were prepared according to the manufacturer's instructions (R&D Systems, ARY001B) and 200 mg of lysates were applied to phospho-RTK membranes overnight. Arrows indicate phosphorylated RTK spots in duplicate.

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2333

Patwardhan et al.

Figure 3. Combined inhibition of p-AKT and p-ERK can mimic inhibitory effects of nintedanib. A and B, Cell lines were plated in 96-well plates and treated in triplicate per condition with indicated drugs at 1,000 nmol/L for 6 days. Cell viability was measured using Dojindo Cell Counting Kit 8 according to the manufacturer's instructions. MEKi ¼ selumetinib, AKTi ¼ MK-2206. Dose response graphs were generated as a percentage of DMSO-treated control. Error bars represent SEM. Combined data from two replicates from a single experiment representative of three independent experiments are shown. P values indicated as , ¼ P < 0.05, , ¼ P < 0.01. C, Approximately, 5 105 cells were plated in 60 mm plated in serum-free media for 24 hours. Next day, cells were treated with 500 nmol/L of indicated drugs in media containing 10%FBS for another 24 hours. Cell lysates were prepared using RIPA lysis buffer and 30 mg lysates were loaded on SDS/PAGE and immunoblotted using indicated antibodies. Representative western blots from two independent experiments are shown.

2334 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Triple Angiokinase Inhibition by Nintedanib in Sarcoma

Figure 4. Inhibition of cell proliferation and downstream signaling by nintedanib is a result of combined inhibition of PDGFR and FGFR. A and C, Cell lines were plated in 96-well plates and treated in triplicate per condition with increasing doses of indicated drugs for 6 days. Cell viability was measured using Dojindo Cell Counting Kit 8 according to the manufacturer's instructions. Dose response graphs were generated as a percentage of DMSO-treated control. Error bars represent SEM. Combined data from two replicates from a single experiment representative of three independent experiments are shown. P values indicated as , ¼ P < 0.05, , ¼ P < 0.01. B and D, Approximately, 5 105 cells were plated in 60 mm plated in serum-free media for 24 hours. Next day, cells were treated with 500 nmol/L of indicated drugs in media containing 10%FBS for another 24 hours. Cell lysates were prepared using RIPA lysis buffer and 30 mg lysates were loaded on SDS/PAGE and immunoblotted using indicated antibodies. Representative western blots from two independent experiments are shown. www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2335

Patwardhan et al.

Figure 5. siRNA-mediated knockdown of PDGFR and FGFR can mimic inhibition of cell proliferation and downstream signaling by nintedanib. A, Approximately, 2.5 105 cells were plated and transfected with 100 nmol/L of indicated siRNAs (pooled siRNAs, GE Dharmacon) after 24 hours. Twenty-four hours after transfection, media were changed and cells were transfected again with 100 nmol/L siRNA. After another 48 hours, cells were harvested and approximately 2,000 cells per well were plated in 96-well plates in triplicates. Cell viability was measured after 6 days using Dojindo Cell Counting Kit 8. Error bars represent SEM. Combined data from two replicates from a single experiment representative of three independent experiments are shown. P values indicated as , ¼ P < 0.05. B, Cell lysates were prepared using RIPA lysis buffer 72 hours post siRNA transfection and 30 mg lysates were loaded on SDS/PAGE and immunoblotted using indicated antibodies. Representative western blots from two independent experiments are shown.

blot analysis to compare their efficacy. Figure 4C shows that superior in the synovial sarcoma setting. Furthermore, it has been nintedanib treatment resulted in significantly better inhibition shown recently that nintedanib has a more potent inhibition (P < 0.05) compared with pazopanib in the SYO-1 cell line. As profile across the triple angiokinase panel when compared head- seen earlier, nintedanib treatment (at 1,000 nmol/L) resulted in to-head with other in-class competitor molecules such as suniti- inhibition of p-AKT as well as p-ERK1/2 (Fig. 4D), whereas nib, pazopanib, and vandetanib (26). pazopanib treatment (at 1,000 nmol/L) showed downregulation To further confirm the role of both PDGFR and FGFR in driving of p-AKT signaling but failed to show any downregulation of p- cell proliferation, we carried out siRNA-mediated analysis of cell ERK1/2 signaling strongly suggesting that nintedanib treatment is proliferation in SYO-1 cell line. siRNA-mediated knockdown of

2336 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Triple Angiokinase Inhibition by Nintedanib in Sarcoma

Figure 6. Effect of nintedanib treatment in SYO-1 xenografts. A, Tumor growth of SYO-1 xenografts (n ¼ 7/group) treated with the indicated drugs is shown. Drug treatment was stopped on day 21 for Vehicle control group (when the tumors grew more than 2,000 mm3 in size) and on day 24 for the remaining two treatment groups. P values indicated as , ¼ P < 0.05, , ¼ P < 0.01, , ¼ P < 0.005. B, Thirty micrograms of RIPA lysates obtained using sample grinding kit (GE Healthcare) from xenograft tissues at the end of drug treatment were loaded on SDS/PAGE and immunoblotted using indicated antibodies. Representative blots from two animals sacrificed at the end of drug treatment are shown. C, Xenograft tissues obtained from mice at the end of drug treatment with the indicated drugs were stained immuno-histochemically using CD31 (HistoWiz). Magnification of 40 is shown for each image. Brown CD31 staining is indicated with red arrows in the zoomed portion of the image. Representative images from two animals sacrificedattheendofdrug treatment are shown.

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2337

Patwardhan et al.

individual RTKs (PDGFRa, FGFR1, and FGFR2) resulted in sig- tive, disease-specific treatment is often needed given the com- nificant inhibition of cell proliferation compared with control plexities of various subtypes. siRNA (P < 0.01; Fig. 5A). However, combined knockdown of Recent evidence has suggested involvement of RTKs such as PDGFRa, FGFR1, and FGFR2 (thus, mimicking nintedanib treat- PDGFR (8, 34), c-Met (11) as well as IGF1-R (35), and ErbB4 ment) resulted in further inhibition of cell proliferation compared (9) among others in various sarcoma subtypes. Phosphopro- with any of the single or double siRNA-mediated RTK silencing teomic analysis using a mass spectrometry approach has further (P < 0.05; Fig. 5A). Western blot in Fig. 5B shows that siRNA revealed that RTKs such as PDGFR, IGF-1R, c-Met, and ALK play treatment resulted in significant downregulation of RTKs. As seen a role in driving sarcoma development and progression (6, 7). with pan-FGFR inhibitor PD173074 (Fig. 4B), siRNA-mediated Previous studies from our laboratory (8) and others (20) have knockdown of FGFR1 and FGFR2 resulted in downregulation of also elucidated the role of PDGFR and FGFR signaling pathways pERK1/2, whereas, only combined knockdown of PDGFRa and in driving synovial sarcoma proliferation. Despite many of FGFR1, and FGFR2 resulted in downregulation of p-AKT signal these preclinical and clinical advances in the treatment of (Fig. 5B). sarcomas, the survival rate for this disease remains poor and novel therapeutic approaches are needed to treat this disease. To determine the efficacy of nintedanib in various sarcoma cell Nintedanib treatment results in significant tumor growth lines, we first analyzed the basal levels of expression of many of suppression compared with imatinib in vivo in SYO-1 xenograft the RTK drug targets such as PDGFRa, b as well as FGFR and model VEGFR family of kinases. Many sarcoma subtypes including To test the efficacy of nintedanib treatment in vivo, we carried MPNST and synovial sarcoma cell lines expressed multiple drug out mouse xenograft experiments using SYO-1 xenograft model. targets specific for nintedanib. When we determined the antipro- We compared the efficacy of nintedanib against imatinib in vivo as liferative efficacy of nintedanib in a cell proliferation assay, we tested earlier in vitro (Fig. 3A and B). Both imatinib and nintedanib observed increased sensitivity with increasing doses of the drug treatment resulted in significant suppression (P < 0.05 and P < mostly in the cell lines with high levels of target RTK expression 0.005) of tumor growth compared with vehicle control (Fig. 6A). (such as SYO-1, HS-SY-II). Western blot analysis confirmed inhi- However, tumor suppression after nintedanib treatment was bition of downstream signaling pathways, p-AKT and p-ERK1/2,in significantly better than imatinib treatment (P < 0.01). Western synovial and MPNST cell line (ST8814), but not in Ewing sarcoma blot analysis of tissue samples obtained at the end of 3 weeks of cell lines (A673, CHP100), which also showed the lowest expres- nintedanib treatment showed induction of cleaved PARP, a sur- sion of drug targets. Although nintedanib treatment resulted in rogate marker for apoptosis, as well as greater inhibition of significant inhibition of phosphorylated RTKs when tested in a phosphorylation of FGFR, p-AKT, and p-ERK1/2 (Fig. 6B) com- phospho-RTK array experiment, little-to-no inhibition of cell pared with imatinib treatment. Because nintedanib blocks VEGFR proliferation was observed in the two Ewing sarcoma cell lines in addition to PDGFR and FGFR, we tested the effect of nintedanib (A673, and CHP100). Interestingly, phosphorylated ErbB4 as treatment on tumor vasculature as determined by CD31 IHC. In well as phosphorylated IGF-1R and insulin receptor were detected accordance with previous reports (25), nintedanib treatment on the phospho-RTK array. This strongly suggests that RTK path- resulted in inhibition of tumor vessel density seen as complete ways other than PDGFRa, b such as IGF-1R (10), or ErbB4 (9) are absence of brown IHC staining compared with vehicle control involved in driving Ewing sarcoma progression. (Fig. 6C, red arrows in the zoomed image pointing to the brown Multikinase inhibitors such as imatinib, which block RTKs signal for CD31 staining). Imatinib treatment, on the other hand, including c-Kit and PDGFR, have been used in clinical trials with did not show any reduction in CD31 staining (Fig. 6C). In vivo modest success (12). When we tested the efficacy of nintedanib tumor growth data as well as Western blot analysis and IHC results against imatinib using in vitro cell proliferation assay and western obtained in SYO-1 xenografts confirmed our in vitro findings that immunoblotting, nintedanib treatment showed significantly bet- nintedanib can be explored further as a potential therapy for ter inhibition of cell proliferation as well as blockade of signaling synovial sarcoma. pathways compared with imatinib. Nintedanib has been shown previously to inhibit MAP kinase as well as AKT signaling (25). To Discussion test whether the observed inhibition of cell proliferation in the In this study, we report preclinical evaluation of nintedanib synovial sarcoma cell line SYO-1 is indeed a result of p-ERK (25), a triple angiokinase inhibitor, in multiple sarcoma cell lines. and p-AKT inhibition, we combined a MEK and AKT inhibitor Nintedanib, a potent inhibitor of PDGFR, VEGFR, and FGFR (selumetinib and MK-2206, respectively) to mimic nintedanib- family of kinases that are involved in tumor angiogenesis, has mediated inhibition of downstream signaling. As hypothesized, undergone considerable preclinical and clinical evaluation in combined inhibition of p-ERK and p-AKT resulted in inhibition of cancers such as non–small-cell lung cancer, ovarian cancer, cell proliferation comparable with nintedanib treatment. Inter- as well as in diseases like idiopathic pulmonary fibrosis (33); estingly, combined treatment with nintedanib and selumetinib however, it has largely been untested as a potential therapy for resulted in further inhibition of cell proliferation than nintedanib orphan diseases like sarcoma. Pazopanib, another multikinase alone. This is probably the result of further suppression of MAPK inhibitor that blocks VEGFR and PDGFR has been approved for signaling (seen as p-ERK downregulation in Fig. 3C) after the the treatment of soft-tissue sarcoma albeit with limited clinical addition of selumetinib as confirmed by western immunoblot- success (32). Recent studies (24) from our laboratory have also ting. Notably, in the SYO-1 cell line, we were unable to observe evaluated another multikinase inhibitor, sitravatinib, for its anti- any significant downregulation of p-ERK1/2 when treated with proliferative efficacy in sarcoma and is currently in clinical trial selumetinib alone or in combination with MK-2206. Treatment for liposarcoma (ClinicalTrials.gov Identifier: NCT02978859). with selumetinib alone was also accompanied by an increase in Although broad-spectrum RTK inhibition approach seems attrac- p-AKT which could be a result of feedback inhibition after MAPK

2338 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Triple Angiokinase Inhibition by Nintedanib in Sarcoma

signaling inhibition by selumetinib as has been reported in proliferation and significant tumor growth suppression that is othercancersasaresultofreliefofnegativefeedbackloop(36). superior to imatinib which has previously been tested in clinical Combined treatment with nintedanib and selumetinib, how- trials in synovial sarcoma with no clinical benefit (ClinicalTrials. ever, prevented this feedback activation of p-AKT as well as gov Identifier: NCT00031915; ref. 12). Data obtained in this study resulted in almost complete downregulation of p-ERK1/2 signal strongly warrant further preclinical and clinical evaluation of on the Western blot analysis. It must also be noted that nintedanib in synovial sarcoma. nintedanib treatment did not result in any significant inhibition of cell proliferation in A673 cell line which carries a BRAF Disclosure of Potential Conflicts of Interest mutation (V600E; ref. 37); however, combined inhibition of No potential conflicts of interest were disclosed. MEKi and AKTi resulted in inhibition of cell proliferation which could be a result of greater suppression of p-ERK by Authors' Contributions MEKi treatment as seen on Western blot analysis. Conception and design: P.P. Patwardhan, G.K. Schwartz Previous studies from our laboratory have reported that syno- Development of methodology: P.P. Patwardhan, G.K. Schwartz vial sarcomas often show uniquely high levels of PDGFRa expres- Acquisition of data (provided animals, acquired and managed patients, a provided facilities, etc.): P.P. Patwardhan, E. Musi, G.K. Schwartz sion and PDGFR is primarily involved in rapamycin induced Analysis and interpretation of data (e.g., statistical analysis, biostatistics, feedback activation of AKT (8). Expression of multiple FGF genes computational analysis): P.P. Patwardhan, E. Musi, G.K. Schwartz 1/2 and growth stimulatory effect of FGFR-mediated p-ERK in Writing, review, and/or revision of the manuscript: P.P. Patwardhan, E. Musi, synovial sarcomas has also been documented previously (20). G.K. Schwartz Combined siRNA-mediated knockdown of PDGFRa and FGFR1, Administrative, technical, or material support (i.e., reporting or organizing and 2 in the SYO-1 cell line resulted in inhibition of cell prolif- data, constructing databases): P.P. Patwardhan, G.K. Schwartz Study supervision: P.P. Patwardhan eration significantly higher than siRNA knockdown of any of the RTKs individually. This strongly supported our hypothesis that Acknowledgments nintedanib-mediated inhibition of PDGFR, FGFR pathways, and The authors would like to thank Dr. Frank Hilberg and Dr. Ulrike Weyer- 1/2 downstream p-AKT and p-ERK pathways is indeed responsible Czernilofsky from Boehringer Ingelheim for valuable feedback and helpful for the observed antiproliferative effect in SYO-1 cell line. Finally, scientific discussions. Funding for this work was provided by the Herbert Irving the SYO-1 xenograft model confirmed that nintedanib-mediated Comprehensive Cancer Center, Columbia University (to G.K. Schwartz).Finan- inhibition of RTK signaling pathways results in significant sup- cial support for this work was provided by Herbert Irving Comprehensive pression of tumor growth, decreased tumor vasculature which is Cancer Center, Columbia University College of Medicine (New York, NY). far better than the effects observed for imatinib treatment. The costs of publication of this article were defrayed in part by the payment of Taken together, data from this study strongly suggest that page charges. This article must therefore be hereby marked advertisement in nintedanib-mediated inhibition of PDGFR and FGFR in synovial accordance with 18 U.S.C. Section 1734 solely to indicate this fact. sarcoma could have potential therapeutic implication given the role of these RTKs in this disease setting (8, 20). Downregulation Received March 27, 2018; revised July 25, 2018; accepted August 21, 2018; of p-AKT and p-ERK1/2 by nintedanib results in decreased cell published first August 30, 2018.

References 1. Borden EC, Baker LH, Bell RS, Bramwell V, Demetri GD, Eisenberg BL, et al. sarcoma: lack of consistent upregulation or recurrent mutation and a Soft tissue sarcomas of adults: state of the translational science. Clin Cancer review of the clinical trial literature. Sarcoma 2013;2013:450478. Res 2003;9:1941–56. 11. TorresKE,ZhuQS,BillK,LopezG,GhadimiMP,XieX,etal.Activated 2. Clark MA, Fisher C, Judson I, Thomas JM. Soft-tissue sarcomas in adults. MET is a molecular prognosticator and potential therapeutic target for N Engl J Med 2005;353:701–11. malignant peripheral nerve sheath tumors. Clin Cancer Res 2011;17: 3. Anderson JL, Denny CT, Tap WD, Federman N. Pediatric sarcomas: trans- 3943–55. lating molecular pathogenesis of disease to novel therapeutic possibilities. 12. Chugh R, Wathen JK, Maki RG, Benjamin RS, Patel SR, Meyers PA, et al. Pediatr Res 2012;72:112–21. Phase II multicenter trial of imatinib in 10 histologic subtypes of sarcoma 4. Daugaard S. Current soft-tissue sarcoma classifications. Eur J Cancer using a bayesian hierarchical statistical model. J Clin Oncol 2009;27:3148– 2004;40:543–8. 53. 5. Lemmon MA, Schlessinger J. Cell signaling by receptor tyrosine kinases. 13. Ho AL, Schwartz GK. Targeting of insulin-like growth factor type 1 receptor Cell 2010;141:1117–34. in Ewing sarcoma: unfulfilled promise or a promising beginning? J Clin 6. Bai Y, Li J, Fang B, Edwards A, Zhang G, Bui M, et al. Phosphoproteomics Oncol 2011;29:4581–3. identifies driver tyrosine kinases in sarcoma cell lines and tumors. Cancer 14. Heinrich MC, Corless CL, Duensing A, McGreevey L, Chen CJ, Joseph N, Res 2012;72:2501–11. et al. PDGFRA activating mutations in gastrointestinal stromal tumors. 7. Fleuren EDG, Vlenterie M, van der Graaf WTA, Hillebrandt-Roeffen MHS, Science 2003;299:708–10. Blackburn J, Ma X, et al. Phosphoproteomic profiling reveals ALK and MET 15. Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, Ishiguro S, et al. as novel actionable targets across synovial sarcoma subtypes. Cancer Res Gain-of-function mutations of c-kit in human gastrointestinal stromal 2017;77:4279–4292. tumors. Science 1998;279:577–80. 8. Ho AL, Vasudeva SD, Lae M, Saito T, Barbashina V, Antonescu CR, et al. 16. Demetri GD, von Mehren M, Blanke CD, Van den Abbeele AD, Eisenberg B, PDGF receptor alpha is an alternative mediator of rapamycin-induced Akt Roberts PJ, et al. Efficacy and safety of imatinib mesylate in advanced activation: implications for combination targeted therapy of synovial gastrointestinal stromal tumors. N Engl J Med 2002;347:472–80. sarcoma. Cancer Res 2012;72:4515–25. 17. Heinrich MC, Griffith D, McKinley A, Patterson J, Presnell A, Ramachan- 9. Mendoza-Naranjo A, El-Naggar A, Wai DH, Mistry P, Lazic N, Ayala FR, dran A, et al. Crenolanib inhibits the drug-resistant PDGFRA D842V et al. ERBB4 confers metastatic capacity in Ewing sarcoma. EMBO Mol Med mutation associated with imatinib-resistant gastrointestinal stromal 2013;5:1087–102. tumors. Clin Cancer Res 2012;18:4375–84. 10. O'Neill A, Shah N, Zitomersky N, Ladanyi M, Shukla N, Uren A, et al. 18. Dickson MA, Schwartz GK, Keohan ML, D'Angelo SP, Gounder MM, Chi P, Insulin-like growth factor 1 receptor as a therapeutic target in ewing et al. Progression-Free survival among patients with well-differentiated or

www.aacrjournals.org Mol Cancer Ther; 17(11) November 2018 2339

Patwardhan et al.

dedifferentiated liposarcoma treated with CDK4 inhibitor palbociclib: a 27. Barkan B, Starinsky S, Friedman E, Stein R, Kloog Y. The Ras inhibitor phase 2 clinical trial. JAMA Oncol 2016;2:937–40. farnesylthiosalicylic acid as a potential therapy for neurofibromatosis type 19. Dickson MA, Tap WD, Keohan ML, D'Angelo SP, Gounder MM, Antonescu 1. Clin Cancer Res 2006;12:5533–42. CR, et al. Phase II trial of the CDK4 inhibitor PD0332991 in patients with 28. Ciombor KK, Bekaii-Saab T. Selumetinib for the treatment of cancer. Expert advanced CDK4-amplified well-differentiated or dedifferentiated liposar- Opin Investig Drugs 2015;24:111–123. coma. J Clin Oncol 2013;31:2024–8. 29. Pal SK, Reckamp K, Yu H, Figlin RA. Akt inhibitors in clinical development 20. Ishibe T, Nakayama T, Okamoto T, Aoyama T, Nishijo K, Shibata KR, et al. for the treatment of cancer. Expert Opin Investig Drugs 2010;19:1355–66. Disruption of fibroblast growth factor signal pathway inhibits the growth 30. Jones RL, Judson IR. The development and application of imatinib. Expert of synovial sarcomas: potential application of signal inhibitors to molec- Opin Drug Saf 2005;4:183–91. ular target therapy. Clin Cancer Res 2005;11:2702–12. 31. Pardo OE, Latigo J, Jeffery RE, Nye E, Poulsom R, Spencer-Dene B, et al. The 21. Zhou WY, Zheng H, Du XL, Yang JL. Characterization of FGFR signaling fibroblast growth factor receptor inhibitor PD173074 blocks small cell pathway as therapeutic targets for sarcoma patients. Cancer Biol Med lung cancer growth in vitro and in vivo. Cancer Res 2009;69:8645–51. 2016;13:260–8. 32. Nguyen DT, Shayahi S. Pazopanib: approval for soft-tissue sarcoma. J Adv 22. Zhang L, Hannay JA, Liu J, Das P, Zhan M, Nguyen T, et al. Vascular Pract Oncol 2013;4:53–7. endothelial growth factor overexpression by soft tissue sarcoma cells: 33. Roth GJ, Binder R, Colbatzky F, Dallinger C, Schlenker-Herceg R, Hilberg F, implications for tumor growth, metastasis, and chemoresistance. Cancer et al. Nintedanib: from discovery to the clinic. J Med Chem 2015;58: Res 2006;66:8770–8. 1053–63. 23. Movva S, Wen W, Chen W, Millis SZ, Gatalica Z, Reddy S, et al. Multi- 34. Zietsch J, Ziegenhagen N, Heppner FL, Reuss D, von Deimling A, Holtkamp platform profiling of over 2000 sarcomas: identification of biomarkers and N. The 4q12 amplicon in malignant peripheral nerve sheath tumors: novel therapeutic targets. Oncotarget 2015;6:12234–47. consequences on gene expression and implications for sunitinib treatment. 24. Patwardhan PP, Ivy KS, Musi E, de Stanchina E, Schwartz GK. Significant PLoS One 2010;5:e11858. blockade of multiple receptor tyrosine kinases by MGCD516 (Sitravati- 35. Yang J, Ylip€a€a A, Sun Y, Zheng H, Chen K, Nykter M, et al. Genomic and nib), a novel small molecule inhibitor, shows potent anti-tumor activity in molecular characterization of malignant peripheral nerve sheath tumor preclinical models of sarcoma. Oncotarget 2016;7:4093–109. identifies the IGF1R pathway as a primary target for treatment. Clin Cancer 25. Hilberg F, Roth GJ, Krssak M, Kautschitsch S, Sommergruber W, Res 2011;17:7563–73. Tontsch-Grunt U, et al. BIBF 1120: triple angiokinase inhibitor with 36. Turke AB, Song Y, Costa C, Cook R, Arteaga CL, Asara JM, et al. MEK sustained receptor blockade and good antitumor efficacy. Cancer Res inhibition leads to PI3K/AKT activation by relieving a negative feedback on 2008;68:4774–82. ERBB receptors. Cancer Res 2012;72:3228–37. 26. Hilberg F, Tontsch-Grunt U, Baum A, Le AT, Doebele RC, Lieb S, et al. Triple 37. Slotkin EK, Patwardhan PP, Vasudeva SD, de Stanchina E, Tap WD, angiokinase inhibitor nintedanib directly inhibits tumor cell growth and Schwartz GK. MLN0128, an ATP-competitive mTOR kinase inhibitor with induces tumor shrinkage via blocking oncogenic receptor tyrosine kinases. potent in vitro and in vivo antitumor activity, as potential therapy for bone J Pharmacol Exp Ther 2018;364:494–503. and soft-tissue sarcoma. Mol Cancer Ther 2015;14:395–406.

2340 Mol Cancer Ther; 17(11) November 2018 Molecular Cancer Therapeutics

Preclinical Evaluation of Nintedanib, a Triple Angiokinase Inhibitor, in Soft-tissue Sarcoma: Potential Therapeutic Implication for Synovial Sarcoma

Parag P. Patwardhan, Elgilda Musi and Gary K. Schwartz

Mol Cancer Ther 2018;17:2329-2340. Published OnlineFirst August 30, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-18-0319

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2018/08/30/1535-7163.MCT-18-0319.DC1

Cited articles This article cites 37 articles, 19 of which you can access for free at: http://mct.aacrjournals.org/content/17/11/2329.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/17/11/2329. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.