SS18-SSX–Dependent YAP/TAZ Signaling in Synovial Sarcoma

Published OnlineFirst February 27, 2019; DOI: 10.1158/1078-0432.CCR-17-3553

Translational Cancer Mechanisms and Therapy Clinical Cancer Research SS18-SSX–Dependent YAP/TAZ Signaling in Synovial Sarcoma Ilka Isfort1,2, Magdalene Cyra1,2, Sandra Elges2, Sareetha Kailayangiri3, Bianca Altvater3, Claudia Rossig3,4, Konrad Steinestel2,5, Inga Grunewald€ 1,2, Sebastian Huss2, Eva Eßeling6, Jan-Henrik Mikesch6, Susanne Hafner7, Thomas Simmet7, Agnieszka Wozniak8,9, Patrick Schoffski€ 8,9, Olle Larsson10, Eva Wardelmann2, Marcel Trautmann1,2, and Wolfgang Hartmann1,2

Abstract

Purpose: Synovial sarcoma is a soft tissue malignancy Results: Asignificant subset of synovial sarcoma characterized by a reciprocal t(X;18) translocation. The chi- showed nuclear positivity for YAP/TAZ and their tran- meric SS18-SSX fusion protein acts as a transcriptional dysre- scriptional targets FOXM1 and PLK1. In synovial sarco- gulator representing the major driver of the disease; however, ma cells, RNAi-mediated knockdown of SS18-SSX led to the signaling pathways activated by SS18-SSX remain to be significant reduction of YAP/TAZ-TEAD transcriptional elucidated to define innovative therapeutic strategies. activity. Conversely, SS18-SSX overexpression in SCP-1 Experimental Design: Immunohistochemical evaluation cells induced aberrant YAP/TAZ-dependent signals, mech- of the Hippo signaling pathway effectors YAP/TAZ was per- anistically mediated by an IGF-II/IGF-IR signaling loop formed in a large cohort of synovial sarcoma tissue specimens. leading to dysregulation of the Hippo effectors LATS1 SS18-SSX dependency and biological function of the YAP/TAZ and MOB1. Modulation of YAP/TAZ-TEAD–mediated Hippo signaling cascade were analyzed in five synovial sarco- transcriptional activity by RNAi or verteporfintreatment ma cell lines and a mesenchymal stem cell model in vitro. YAP/ resulted in significant growth inhibitory effects in vitro TAZ-TEAD–mediated transcriptional activity was modulated and in vivo. by RNAi-mediated knockdown and the small-molecule inhib- Conclusions: Our preclinical study identifies an elementary itor verteporfin. The effects of verteporfin were finally tested in role of SS18-SSX–driven YAP/TAZ signals, highlights the com- vivo in synovial sarcoma cell line-based avian chorioallantoic plex network of oncogenic signaling pathways in synovial membrane and murine xenograft models as well as a patient- sarcoma pathogenesis, and provides evidence for innovative derived xenograft. therapeutic approaches.

1Division of Translational Pathology, Gerhard-Domagk-Institute of Pathology, Munster€ University Hospital, Munster,€ Germany. 2Gerhard-Domagk-Institute of Pathology, Munster€ University Hospital, Munster,€ Germany. 3Department of Introduction € Pediatric Hematology and Oncology, University Children's Hospital Munster, Synovial sarcoma (SySa) is an aggressive malignancy com- Munster,€ Germany. 4Cells-in-Motion Cluster of Excellence (EXC 1003 – CiM), prising approximately 2% of all sarcomas with a predomi- University of Munster,€ Munster,€ Germany. 5Institute of Pathology and Molecular Pathology, Bundeswehrkrankenhaus Ulm, Ulm, Germany. 6Department of Med- nance in adolescents and young adults (1, 2). Wide surgical icine A, Hematology, Oncology and Respiratory Medicine, University Hospital resection, radiotherapy, and chemotherapy with ifosfamide Munster,€ Munster,€ Germany. 7Institute of Pharmacology of Natural Products & and doxorubicin represent established therapeutic options; Clinical Pharmacology, Ulm University, Ulm, Germany. 8Laboratory of Experi- however, prognosis in the metastatic situation is poor (3–5). mental Oncology, Department of Oncology, KU Leuven, Leuven, Belgium. ﬁ 9 Speci c molecularly targeted therapeutic approaches are cur- Department of General Medical Oncology, University Hospitals Leuven, Leuven rently limited. Cancer Institute, Leuven, Belgium. 10Departments of Oncology and Pathology, Karolinska Institute, Stockholm, Sweden. The molecular hallmark of SySa is a pathognomonic reciprocal t(X;18)(p11;q11) translocation, leading to the fusion Note: Supplementary data for this article are available at Clinical Cancer of SS18 (SYT) and one of the homologues SSX genes Research Online (http://clincancerres.aacrjournals.org/). (most frequently SSX1 or SSX2,inrarecasesSSX4), generating M. Trautmann and W. Hartmann contributed equally to this article. chimeric SS18-SSX fusion proteins (6–8). Although the SS18- Corresponding Authors: Wolfgang Hartmann and Marcel Trautmann, Division SSX protein is known to play a crucial role in SySa tumorigen- € of Translational Pathology, Gerhard-Domagk-Institute of Pathology, Munster esis, its speciﬁc biological function and its mechanism of þ University Hospital, Domagkstr. 17, 48149, Germany. Phone: 49 (0) 251-83- action remain to be elucidated. Neither SS18 and the SSX 58479 and -57623, Fax: þ49 (0) 251-83-57559. E-mail: [email protected] and [email protected] proteins, nor the chimeric SS18-SSX oncoproteins feature known DNA-binding motifs; however, they have been reported Clin Cancer Res 2019;25:3718–31 to contribute to the dysregulation of gene expression through doi: 10.1158/1078-0432.CCR-17-3553 association with SWI/SNF and Polycomb chromatin remodel- 2019 American Association for Cancer Research. ing complexes (8–12).

3718 Clin Cancer Res; 25(12) June 15, 2019

YAP/TAZ Signals in Synovial Sarcoma

Translational Relevance Materials and Methods Tumor specimens and tissue microarray Acting as a powerful transcriptional dysregulator, the chi- SySa tissue microarrays (TMA, containing two representative meric SS18-SSX fusion protein constitutes the major onco- 1-mm cores) were prepared from formalin- fixed, paraffin- genic driver of synovial sarcoma. Given the notorious diffi- embedded tumor specimens selected from the archive of the culty to target the fusion protein itself, functional insights into Gerhard-Domagk-Institute of Pathology, Muenster University SS18-SSX–shaped tumor biology are essential to decipher Hospital, essentially as described previously (19). Two areas druggable tumor vulnerabilities. We here describe a molecu- within each tumor were selected by experienced pathologists larly based mechanism of YAP/TAZ activation in synovial (E. Wardelmann, W. Hartmann) to represent potential heteroge- sarcoma effected by the SS18-SSX fusion protein that involves neity. Necrobiotic areas and their neighborhood were excluded an IGF-II/IGF-IR signaling loop, leading to dysregulation of from TMA sampling to avoid the detection of secondary (e.g., the Hippo effectors LATS1 and MOB1 and provide evidence of hypoxia-induced) alterations. All diagnoses were reviewed by high efficacy of a YAP/TAZ–directed therapeutic approach in three experienced pathologists (S. Huss, S. Elges and W. Hart- synovial sarcoma in vitro and in vivo. Our study highlights the mann) according to the current WHO classification of tumours of complex network of oncogenic signaling pathways in synovial Soft Tissue and Bone (2), based on morphologic and immuno- sarcoma pathogenesis and refines the concept of biologically histochemical criteria, FISH or RT-PCR analysis. In total, 65 SySa founded molecular strategies to inhibit SS18-SSX–driven tissue specimens were included in the study (30/46.2% female, tumorigenesis. 35/53.8% male; median age at diagnosis was 45 years, range 8–78 years). Forty-eight tumors belong to the monophasic subtype, 14 to the biphasic subtype, and 3 tumors were classified as poorly differentiated SySa. Median tumor size was 5 cm (range 0.3–20 The YAP/TAZ Hippo signaling pathway is an evolutionarily cm). In all cases, FISH or RT-PCR analysis confirmed the diagnosis conserved pathway essential in the control of tissue homeostasis of SySa, revealing the pathognomonic t(X;18) translocation as and organ size through the regulation of cell proliferation, apo- described previously (20). Clinicopathologic characteristics of the ptosis, and stem cell self-renewal (13–15). The central component cohort are summarized in Table 1. The study was approved by the is a kinase module comprising the serine-threonine kinases Ethics Committee of the University of Munster€ (2015-548-f-S) MST1/2 and LATS1/2, complemented by the adaptor proteins SAV1 and MOB1 to control the transcriptional co-activators YAP and TAZ. While in their non-phosphorylated state YAP and TAZ translocate to the nucleus and interact with TEAD1-4 transcription Table 1. Clinicopathologic characteristics of the cohort of SySa patients (n ¼ 65) factors to induce expression of target genes such as CTGF, CYR61, Age (years) PLK1 and FOXM1, LATS1/2-mediated phosphorylation of YAP/ Mean (SD) 41 (17) – TAZ results in their cytoplasmic retention and proteasomal deg- Median (range) 45 (8 78) <41 28 (43.1%) radation (14, 16). 41 37 (59.9%) Although convincing data on the oncogenic function of YAP/ Type TAZ signals in several epithelial tumors are available (13), only Primary tumor 43 (66.2%) little is known about their role in malignant soft tissue tumors. Metastasis 9 (13.8%) First evidence for a function of YAP/TAZ in mesenchymal tumor- Recurrence 7 (10.8%) igenesis was gathered by St John and colleagues who demonstrat- ND 6 (9.2%) fi Histology ed that approximately 15% of LATS1/2-de cient mice develop Monophasic 48 (73.9%) metastasizing spindle-cell sarcomas (17). Analyzing genetic data Biphasic 14 (21.5%) from The Cancer Genome Atlas (TCGA) dataset, Eisinger-Matha- Poorly differentiated 3 (4.6%) son and colleagues detected frequent copy number losses in the Size (cm) Hippo pathway components and activators LATS2, NF2, and Mean (SD) 7 (5) – SAV1 in several high-grade sarcomas with complex karyotypes, Median (range) 5 (0.3 20) <7 32 (49.2%) leading to an activation of YAP signals (16). Interestingly, they 7 18 (27.7%) fi did not report signi cant genomic alterations within the group ND 15 (23.1%) of soft tissue tumors driven by genomic translocations. In con- Gender trast, Fullenkamp and colleagues found high nuclear expression Female 30 (46.2%) levels of YAP (50%) and TAZ (66%) among 159 different Male 35 (53.8%) soft tissue malignancies, including a (minor) set of 12 SySa FISH SS18 (break-apart) positive 52 (80.0%) specimens (18). This raises the question whether activation of ND 13 (20.0%) YAP/TAZ in SySa might be based on a functional fusion protein- t(X;18) translocation type dependent mechanism in place of genomic alterations affecting SS18-SSX1 28 (43.1%) regulatory elements of the Hippo pathway. SS18-SSX2 20 (30.8%) The current study was performed to explore the prevalence and ND 17 (26.1%) functional relevance of YAP/TAZ signals in a large set of SySa, Grading (FNCLCC) G2 16 (24.6%) including its molecular dependence on the pathognomonic SS18- G3 24 (36.9%) SSX fusion protein, and to test a molecularly targeted approach ND 25 (38.5%) employing a small-molecule YAP/TAZ inhibitor in a preclinical Abbreviations: FNCLCC, Fed eration Nationale des Centres de Lutte Contre le setting. Cancer; ND, not determined.

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3719

Isfort et al.

and conducted in accordance with current ethical standards Extraction Reagents Kit (Thermo Fisher Scientiﬁc) according to (Declaration of Helsinki, 1975). the manufacturer's instructions.

Immunohistochemistry (IHC) RT-PCR Following primary antibodies were applied: YAP (monoclonal The AllPrep DNA/RNA Mini Kit (Qiagen) was applied for RNA rabbit, D8H1X, 1:100, #14074 Cell Signaling Technology), TAZ isolation followed by cDNA synthesis with the ProtoScript II First (polyclonal rabbit, 1:200, #HPA007415 Sigma-Aldrich), FOXM1 Strand cDNA Synthesis Kit (NEB) according to the manufacturer's (monoclonal mouse, G-5; 1:1000, #376471 Santa Cruz Biotech- instructions. PCR primer sequences are given in Supplementary nology), PLK1 (monoclonal rabbit, 208G4, 1:25, #4513 Cell Table S2. Signaling Technology). IHC staining was performed with a Bench- Mark ULTRA Autostainer (VENTANA/Roche) on 3-mm TMA sec- Immunoblotting tions. In brief, the staining procedure included heat-induced The following primary antibodies were used according to the epitope retrieval (HIER) pretreatment using Tris-Borate-EDTA manufacturer's instructions: CTGF (#6992) obtained from buffer (pH 8.4; 95-100C, 32–72 minutes) followed by incuba- Abcam, b-Actin (#A5441) from Sigma-Aldrich, SS18/SYT tion with respective primary antibodies for 16 to 120 minutes and (#8819) and FOXM1 (#500) both obtained from Santa Cruz employment of the OptiView DAB IHC Detection Kit (VENTANA/ Biotechnology, AKT (#9272), phospho-AKT (Ser473) (#4060), Roche), essentially as described previously (19). Positive and GAPDH (#5174), Histone H3 (#4499), phospho-Histone H3 negative control stainings using an appropriate IgG subtype (Ser10) (#9701), IGF-I Receptor b (#9750), phospho-IGF-I (DCS) were included. Immunoreactivity was assessed using a Receptor b (Tyr1135/1136)/phospho-Insulin Receptor b semiquantitative score (0, negative; 1, weak; 2, moderate; and (Tyr1150/1151) (#3024), LATS1 (#3477), phospho-LATS1 3, strong) defining the staining intensity in the positive control (Thr1079) (#8654), MOB1 (#13730), phospho-MOB1 (Thr35) (hepatocellular carcinoma) as strong. Only TMA tissue cores with (#8699), PLK1 (#4513), SS18 (#21792), phospho-TAZ (Ser89) at least moderate staining (semiquantitative score 2) and 30% (#59971), YAP (#14074), phospho-YAP (Ser127) (#13008) and (YAP, TAZ, and FOXM1) or 5% (PLK1) positive cells were YAP/TAZ (#8418) all obtained from Cell Signaling Technology. considered positive for the purposes of the study (Supplementary The SS18-SSX fusion protein was detected with an antibody Table S1). The IHC readers were blinded to outcome data, and the targeting the N-terminus of SS18 (which is retained in the score cut-off point (positive ¼ semiquantitative score 2) was SS18-SSX fusion oncoprotein). Secondary antibody labeling prespecified without prior analyses of the clinical course. (Bio-Rad Laboratories) as well as immunoblot development was performed using an enhanced chemiluminescence detection kit Cell culture and cell lines (SignalFire ECL Reagent; Cell Signaling Technology) and the The human SySa cell lines HS-SY-II (expressing SS18-SSX1), Molecular Imager ChemiDoc system (Image Lab Software; Bio- FUJI, 1273/99, CME-1, and SYO-1 (all expressing SS18-SSX2) Rad Laboratories), as described previously (19). were cultured as described previously (21). For the purpose of cell line authentication, presence of the pathognomonic SS18-SSX Immunofluorescence (IF) translocation was confirmed by RT-PCR and subsequent Sanger Immunofluorescence staining procedure was as follows: sequencing using specific primers for the t(X;18) translocation Cells were grown on poly-L-lysine (0.001%) coated chamber subtypes. The human mesenchymal stem cell line SCP-1 was slides (both Sigma-Aldrich), fixed in 4% paraformaldehyde kindly provided by Dr. Attila Aszodi, Munich, Germany (22). (15 minutes), pre-treated in blocking buffer (5% BSA, 0.3% The human rhabdomyosarcoma cell line RD was obtained from Triton X-100 in 1 PBS) and incubated with primary YAP the American Type Culture Collection (ATCC) and authenticated (monoclonal rabbit, D8H1X, 1:100, #14074 Cell Signaling by STR-PCR. All monolayer cell cultures were grown under Technology) or TAZ (polyclonal rabbit, 1:200, #HPA007415 standard incubation conditions (37C, humidified atmosphere, Sigma-Aldrich) antibody solution (4C, overnight). After sec- 5% CO2) and maintained in DMEM (HS-SY-II, SYO-1), ondary antibody incubation (2 hours), cells were mounted in RPMI1640 (FUJI, CME-1 and RD), F-12 (1273/99), or Minimum Vectashield mounting medium with DAPI (Vector Laborato- Essential Media (SCP-1) supplemented with 10% to 15% FBS ries). To detect YAP/TAZ localization upon BMS-754807 (or (Life Technologies). Mycoplasma testing was performed quarterly DMSO) treatment (4–12 hours), cells were grown in medium by standardized PCR, and cells were passaged for a maximum of supplemented with 2% FBS. IF analyses were performed with a 20 to 30 culturing cycles between thawing and use in the described Leica DM5500 B microscope. experiments. To study the effects of the YAP/TAZ-TEAD inhibitor verteporfin (23, 24) and the specific IGF-IR inhibitor BMS- Stable lentiviral transduction 754807 (25), SySa cells were grown in medium supplemented Lentiviral transductions were performed in SCP-1 cells essen- with 2% FBS. Short-term and long-term treatments with BMS- tially as described previously (27) using pLenti6.2/V5- 754807 were performed for 10 minutes and 36 hours, respec- DEST_SS18-SSX1 or pLenti6.2-GW/EmGFP (Gateway cloning tively. Treatments with verteporfin (0.25–1 mmol/L) to analyze system, Life Technologies) and blasticidin (10 mg/mL; Life effects on the protein level were performed for 14 hours. Prior to Technologies) for cell selection. EmGFP expression was verified stimulation with human recombinant Insulin-like growth factor by fluorescence microscopy and SS18-SSX1 expression by (IGF)-II (#50342 Biomol, 200 ng/mL IGF-II for 30 minutes), cells immunoblotting. were incubated with medium devoid of FBS supplement for 4 hours. Cell lysis, protein extraction, and immunoblotting were RNA interference (RNAi) performed as described previously (26). Subcellular fractionation To exclude unspecific off-target effects, a set of prevalidated was performed using the NE-PER Nuclear and Cytoplasmic siRNAs for YAP (Set of 3): #1 ¼ HSS115942, #2 ¼ HSS115944,

3720 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

#3 ¼ HSS173621, TAZ (Set of 3): #1 ¼ HSS119545, #2 ¼ prior to the MTT cell viability assay. At least three independent HSS119546, #3 ¼ HSS119547, IGF-IR:#1¼ s7211, #2 ¼ 74 and experiments were performed (each in quintuplicates). a non-targeting negative control siRNA (BLOCK-iT Alexa Fluor Red Fluorescent Control; all purchased by Life Technologies) were Flow cytometry applied. To target the SS18-SSX fusion transcript, a set of pub- Effects of verteporfin on the apoptotic rate were assessed by lished and prevalidated duplex oligos (28, 29) was employed (#1 flow cytometric analyses. In brief, SySa cells were grown in 175 sense: 50-AAC CAA CUA CCU CUG AGA AGA-30; antisense: 50- cm2 cell culture flasks (medium supplemented with 2% FBS) and UCU UCU CAG AGG UAG UUG GUU-30, #2 sense: 50-CAA GAA treated with verteporfin (0.25–0.5 mmol/L; 72 hours). Adherent GCC AGC AGA GGA ATT-30; antisense: 50-UUC CUC UGC UGG cells were detached using 0.025% Trypsin (Life Technologies), CUU CUU GTT-30 and #3 sense: 50-AUA UGA CCA GAU CAU fixed in 2% paraformaldehyde (10 minutes on ice), washed in 1 GCC CAA GAU U-30; antisense: 50-UCU UGG GCA UGA UCU PBS, collected by centrifugation and incubated in 1 PBS (sup- GGU CAU AUU U-30). CME-1 and FUJI cells were transfected with plemented with 0.25% Triton X-100) for 5 minutes on ice. After the indicated siRNA using Lipofectamine RNAiMAX (Life Tech- an additional washing step, cells were incubated with the cleaved nologies). Medium containing transfection reagent was replaced Poly-(ADP-ribose)-polymerase (PARP, Asp214) antibody (BD after 3 to 6 hours with medium supplemented with 2% FBS. After Biosciences; phycoerythrin-labelled) antibody for 60 minutes at incubation for 24 to 72 hours, siRNA-transfected cells were lysed room temperature. Fluorescence intensity was measured using a and knockdown efficiency was documented by immunoblotting. FACSCanto II analytical flow cytometer and cytometric data were Experiments were repeated at least three times. analyzed using the FACSDiva software (both BD Biosciences). Each experiment was performed at least in duplicates. Luciferase reporter assay To assess YAP/TAZ-mediated transcriptional activity, SySa cells Chicken chorioallantoic membrane studies were transfected with a YAP/TAZ-responsive TEAD luciferase Chorioallantoic membrane (CAM) assays were performed as reporter plasmid (8xGTIIC; Addgene #34615; ref. 30) using either described previously (34). In brief, SySa cells (1.5 106 cells/egg; Lipofectamine 2000 (Life Technologies) or Viromer RED (Lipo- dissolved in medium/Matrigel 1:1, v/v) were xenografted onto the calyx) transfection reagents. Luciferase reporter assays were egg CAM (7 days after fertilization) and incubated at 37 C with performed using the Dual-Luciferase reporter assay system (Pro- 60% relative humidity. On day 8 of incubation, verteporfin(1–2 mega) according to the manufacturer's instructions. For experi- mmol/L) or vehicle control (0.2% DMSO in NaCl 0.9%) were mental activation, cells were cotransfected with mutant YAP topically applied (6/group). The identical treatment protocol DS127A (Addgene #27370) or TAZ DS89A (Addgene #32840) was recapitulated on 2 consecutive days. Three days after treat- plasmid DNA (31, 32). The amount of plasmid DNA in each ment initiation, tumor-containing CAM xenografts were transfection was kept constant by addition of the non-coding explanted and fixed in 5% paraformaldehyde. Tumor volume backbone plasmid. After incubation for 24 hours, cells were lysed (mm3) was calculated according to the following formula: (35). and luciferase activity was measured in quintuplicates (GloMax- fl Multi Detection System, Promega). Fire y luciferase activity was h p i normalized to Renilla luciferase activity (cotransfected pRL-TK Tumor volume ¼ ðÞsmall diameter 2 large diameter 6 control plasmid; Promega) to account for potential differences in transfection efficiency. Cells were transfected with the 8xGTIIC TEAD luciferase reporter plasmid to evaluate the effects of YAP/ SySa cell line and patient-derived xenograft (PDX) studies TAZ nuclear activation and to determine the suppressive effects of To establish SySa cell line xenografts, SYO-1 cells (5 106 cells/ verteporfin (17–24 hours) and BMS-754807 (17–24 hours) on 100 mL1 PBS) were subcutaneously injected into the lower YAP/TAZ-TEAD–mediated transcriptional activity. SySa cells flanks of NOD scid gamma (NSG) mice. Once tumors reached a cotransfected with DS127A YAP or DS89A TAZ expression plas- volume of approximately 100 mm3 (11 days after cell injection), mids were incubated with verteporfin (0.075 mmol/L) for 72 tumor-bearing mice were homogeneously distributed across four hours. For experiments including cytokine stimulation, prior to treatment groups (3 mice/group), receiving: (i) verteporfin (75 IGF-II incubation (200 ng/mL IGF-II for 5 minutes), cells mg/kg, every other day), (ii) doxorubicin (3 mg/kg, once a week), were grown in medium devoid of FBS supplement for 19 hours. (iii) combination of verteporfin (75 mg/kg, every other day) and At least three independent experiments were performed (each in doxorubicin (3 mg/kg, once a week), or (iv) DMSO vehicle quintuplicates). control (every other day) by intraperitoneal administration. Because of its established role in conventional therapeutic regi- Cell viability assay (MTT) mens applied for SySa treatment, doxorubicin was included as To document cell viability after drug treatment or RNAi-medi- positive control; drug combinations were tested to screen for ated knockdown, the MTT Cell Proliferation Kit (Roche) was potential additive or complementary effects. Tumor volumes were applied as described previously (19). In brief, SYO-1 (8 calculated according to the same formula as described above and 103), FUJI (8 103), 1273/99 (7 103), CME-1 (6 103), normalized to the individual volume at treatment initiation. The HS-SY-II (15 103), and RD rhabdomyosarcoma (6 103) SySa PDX model (UZLX-STS7) was described previously (36). In cells (33) were seeded in 96-well cell culture plates (medium brief, human SySa tissue was subcutaneously transplanted into supplemented with 2% FBS) and exposed to increasing concen- the lower flanks of NMRI nu/nu mice. On day 27 after transplan- trations of verteporfin (0.1875-3 mmol/L) for 72 hours. An appro- tation, the identical treatment protocol was initiated (4 mice/ priate DMSO vehicle control was included. In CME-1 and FUJI group) as described for the SYO-1 xenografts. For all in vivo mouse cells, YAP and TAZ siRNA transfection was performed 72 hours studies, permission was obtained from the local authorities, and

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3721

Isfort et al.

all experiments were performed in accordance with the standards was reduced upon RNAi-mediated SS18-SSX knockdown as dem- of the National and European Union guidelines. onstrated by immunofluorescence (Fig. 2A, top) and immunoblotting after subcellular protein fractionation (Fig. 2A, bottom). SS18-SSX fi Compounds Depletion of additionally resulted in signi cant suppression of YAP/TAZ-responsive TEAD luciferase reporter activity Verteporfin (VP; C41H42N4O8; CAS#: 129497-78-5; Targetmol; (Fig. 2B) and reduced protein levels of YAP, TAZ, FOXM1, CTGF, refs. 23, 24), and BMS-754807 (C23H24FN9O; CAS#: 1001350- 96-4, Biomol; ref. 25) were dissolved in DMSO (Sigma-Aldrich). and PLK1 (Fig. 2C). Overexpression of SS18-SSX in mesenchymal Doxorubicin hydrochloride (Doxo; C H NO HCl; CAS#: SCP-1 stem cells resulted in nuclear accumulation of YAP/TAZ 27 29 11 fi 25316-40-9; Actavis) was dissolved in Aqua ad iniectabilia (Braun). protein levels (Fig. 2D), signi cantly induced TEAD luciferase reporter activity (Fig. 2E) and consistently increased FOXM1, CTGF, and PLK1 expression levels (Fig. 2F). Statistical analysis Two-group comparisons were analyzed by unpaired, two-tailed IGF-II/IGF-IR signaling links the SS18-SSX fusion protein to Student t test (GraphPad Software). Statistical differences were YAP/TAZ activation considered significant at P < 0.05 ( ), P < 0.01 ( ), and P < 0.001 Given several kinase signaling pathways being activated by ( ). Outlier testing was performed for every measurement of SS18-SSX fusion oncoprotein in SySa (20, 26, 37–39) and com- quintuplicates and for the xenograft data set according to the plex signaling networks modulating YAP/TAZ activation (13), we formula (n ¼ count of values): set out to analyze whether a functional connection between SS18- SSX driven IGF-IR/PI3K/AKT signals might be involved in dysre- SD SD Mean 2 pffiffiffi > Outlier > Mean þ 2 pffiffiffi gulation of the Hippo signaling pathway. In agreement with our n n previous findings in T-REx-293 (26), stable expression of SS18- SSX in mesenchymal SCP-1 stem cells resulted in promoter- specific induction of IGF2 expression (Fig. 3A; Supplementary The concentration of verteporfin required for 50% growth Fig. S2A). In two SySa cell lines, IGF-II stimulation significantly inhibition (IC value), was calculated by non-linear regression 50 enhanced TEAD luciferase reporter activation, indicating elevated analysis. YAP/TAZ-mediated transcriptional activity (Fig. 3B). On the protein level, IGF-II stimulation of SySa cells (expectedly) induced Results phosphorylation of IGF-IR (Tyr1135/6) and AKT (Ser473) while phosphorylation of LATS1 (Thr1079), MOB1 (Thr35), YAP Expression of YAP/TAZ in human SySa tumor tissue and cell (Ser127), and TAZ (Ser89) was reduced. The cytoplasmic local- lines ization of phosphorylated YAP (Ser127) and TAZ (Ser89) was – To determine the involvement of YAP/TAZ mediated signal confirmed by subcellular fractionation (Supplementary Fig. S2B), transduction in SySa tumorigenesis, expression levels of nuclear demonstrating that these phosphorylation events are associated YAP and TAZ (corresponding to the transcriptionally active pool) with YAP/TAZ transcriptional inactivation and that a decrease in along with the downstream targets FOXM1 and PLK1 were phosphorylation levels is associated with YAP/TAZ activation. examined in a set of 65 SySa tissue specimens using IHC These data imply an IGF-II-dependent modulation of YAP/TAZ (Table 1; Fig. 1A; Supplementary Table S1). Moderate to strong phosphorylation levels through dysregulation of the Hippo effec- nuclear staining of YAP and TAZ was detected in 80% (52/65) and tors LATS1 and MOB1, resulting in YAP/TAZ activation (Fig. 3C). 35% (23/65) of SySa tissue specimens, respectively. Overall, 11 of Accordingly, RNAi-mediated knockdown of IGF-IR reduced YAP/ 65 tumors (17%) showed neither YAP nor TAZ nuclear staining TAZ, FOXM1 and CTGF protein levels in CME-1 cells (Fig. 3D). In positivity. In total, 100% (65/65) of SySa showed a moderate to order to substantiate these results with a pharmacological strong staining intensity for FOXM1 and 72% (47/65) for PLK1 approach allowing shorter incubation periods, we performed (Fig. 1B; Supplementary Fig. S1A). Concordance of nuclear YAP additional experiments applying the specific IGF-IR inhibitor and/or TAZ immunoreactivity and expression of FOXM1 and/or BMS-754807 (Supplementary Fig. S2C; ref. 25) in three SySa cell PLK1 was demonstrated in 83% (54/65) of SySa (Fig. 1C). Strong lines. Inhibition of IGF-IR led to inactivation of YAP/TAZ as YAP/TAZ and downstream target (FOXM1, CTGF, and PLK1) mirrored by significantly reduced TEAD luciferase reporter activity expression levels were detectable in total protein extracts of all (Fig. 3E), accompanied by elevated phosphorylation levels of fi ve SySa cell lines (Fig. 1D; top). Expression of the pathogno- LATS1 (Thr1079), MOB1 (Thr35), YAP (Ser127), and TAZ SS18-SSX fi monic fusion transcript was con rmed by RT-PCR (Ser89; Fig. 3F) in short-term incubations as well as consistently (Fig. 1D; bottom). The transcriptionally active pool of nuclear reduced YAP/TAZ, FOXM1, CTGF, and PLK1 protein levels in fl YAP/TAZ was demonstrated by immuno uorescence staining long-term treatments (Fig. 3G) going along with diminished (Fig. 1E; Supplementary Fig. S1B) and subcellular fractionation nuclear localization of transcriptionally active YAP/TAZ (Fig. 1F). (Fig. 3H; Supplementary Fig. S2D).

Nuclear localization of YAP/TAZ and TEAD-associated Knockdown of YAP and/or TAZ affects target gene expression transcriptional activity is driven by the chimeric SS18-SSX and SySa cell viability fusion protein To document the functional role of YAP and TAZ in SySa in a To evaluate whether YAP/TAZ-mediated transcriptional activity nonpharmacologic approach, two different SySa cell lines (CME- is dependent on the presence of the chimeric SS18-SSX fusion 1 and FUJI) were transfected with prevalidated siRNAs-directed protein, RNAi-mediated SS18-SSX loss-of-function analyses were against human YAP and TAZ. RNAi-mediated depletion resulted performed in two SySa cell lines. Nuclear localization of YAP/TAZ in decreased FOXM1, CTGF, and PLK1 protein levels and reduced

3722 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

Figure 1. Nuclear expression and activation of YAP/TAZ signaling in a comprehensive cohort of SySa tissue specimens (n ¼ 65) and five SySa cell lines. A, IHC stainings show strong nuclear localization of YAP and TAZ in a representative SySa tissue specimen. Strong expression levels of FOXM1 and PLK1 indicate YAP/TAZ- mediated transcriptional activity (original magnification: x10, inset x20). B, IHC staining intensity of nuclear YAP/TAZ, FOXM1, and PLK1 summarized as bar chart (intensity score). C, Venn diagram indicating the concordance of nuclear YAP and/or TAZ immunoreactivity in line with FOXM1 and/or PLK1 expression. D, Immunoblotting demonstrates strong expression levels of YAP, TAZ, FOXM1, CTGF, and PLK1 in total protein extracts of five SySa cells lines (b-Actin used as loading reference). Detection of t(X;18) SS18-SSX fusion gene transcripts in all SySa cell lines by RT-PCR (28S rRNA used as loading reference). E, Immunocytochemical detection of YAP and TAZ (left), DAPI-stained nuclei (middle) and merged fluorescence images (right; original magnification 40), demonstrating predominant nuclear YAP/TAZ localization, implying transcriptional YAP/TAZ activity. F, Detection of "N" nuclear and "C" cytoplasmic fractions of YAP and TAZ in all five SySa cell lines (Histone H3 and GAPDH used as loading references for the nuclear and cytoplasmic fractions, respectively).

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3723

Isfort et al.

Figure 2. The SS18-SSX fusion protein stimulates YAP/TAZ-mediated transcriptional activity. A, Immunofluorescence staining demonstrates diminished nuclear localization of YAP and TAZ upon RNAi-mediated knockdown of SS18-SSX (top, original magnification 40). Subcellular protein fractionation shows reduced nuclear localization of YAP/TAZ and diminished expression of the downstream targets FOXM1, CTGF, and PLK1 upon RNAi-mediated SS18-SSX depletion (lower panel, Histone H3 and GAPDH used as loading references for the nuclear and cytoplasmic fractions, respectively). B, Significantly decreased YAP/TAZ- responsive TEAD luciferase reporter activity upon RNAi-mediated SS18-SSX knockdown in CME-1 and FUJI SySa cells (mean of quintuplicates þ SD; , P < 0.001). C, Immunoblotting of total protein lysates shows reduced YAP/TAZ levels and diminished downstream target expression of FOXM1, CTGF and PLK1 upon RNAi-mediated SS18-SSX knockdown (b-Actin used as loading reference). D, Stable expression of SS18-SSX in mesenchymal SCP-1 stem cells increased nuclear protein levels of YAP and TAZ compared with EmGFP-transduced SCP-1 cells as demonstrated by subcellular protein fractionation (Histone H3 and GAPDH used as loading references for the nuclear and cytoplasmic fractions, respectively). E, Significantly increased TEAD luciferase reporter activity upon stable SS18-SSX expression in SCP-1 cells (mean of quintuplicates þ SD; , P < 0.001). F, Enhanced YAP/TAZ, FOXM1, CTGF, and PLK1 total protein levels in SS18-SSX-transduced SCP-1 cells (b-Actin used as loading reference).

3724 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

Figure 3. IGF-II/IGF-IR signaling pathway as a link between SS18-SSX fusion protein expression and YAP/TAZ activation. A, In RT-PCR analyses, SS18-SSX expressing SCP- 1 cells demonstrate a transcriptional induction of IGF2 (promoter P3 transcript, 28S rRNA used as loading reference). B, Short-term IGF-II stimulation of SySa cells significantly increases TEAD luciferase reporter activity (mean of quintuplicates þ SD; , P < 0.001). C, IGF-II stimulation of SySa cells induces phosphorylation of IGF-IR (Tyr1135/6) and AKT (Ser473) while phosphorylation of LATS1 (Thr1079), MOB1 (Thr35), YAP (Ser127), and TAZ (Ser89) is reduced (b-Actin used as loading reference). D, RNAi-mediated knockdown of IGF-IR in SySa cells reduces phosphorylation of AKT (Ser473) as well as total YAP/TAZ, FOXM1, and CTGF protein levels (b-Actin used as loading reference). E, Significant reduction of TEAD luciferase activity upon incubation with increasing concentrations of BMS- 754807 in SySa cells (, P < 0.001; , P < 0.01). F, Short-term incubation of BMS-754807 in SySa cells leads to decreased phosphorylation levels of IGF-IR (Tyr1135/6) and AKT (Ser473), while phosphorylation of LATS1 (Thr1079), MOB1 (Thr35), YAP (Ser127), and TAZ (Ser89) is increased (b-actin used as loading reference). G, Long-term incubation of BMS-754807 in SySa cells leads to decreased phosphorylation levels of IGF-IR (Tyr1135/6) and AKT (Ser473) as well as total YAP/TAZ, FOXM1, CTGF, and PLK1 levels (b-Actin used as loading reference). H, Immunofluorescence of CME-1 cells demonstrates diminished nuclear YAP and TAZ levels upon BMS-754807 treatment (original magnification, 63).

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3725

Isfort et al.

Figure 4. Suppressive effects of RNAi-mediated knockdown of YAP and TAZ or verteporfin treatment on YAP/TAZ-TEAD transcriptional activity and SySa cell viability in vitro. A, RNAi-mediated knockdown of YAP (siRNA#3) and TAZ (siRNA#3) reduces target protein levels of FOXM1, CTGF and PLK1 as well as phospho Histone H3 (Ser10), serving as a marker of cellular proliferation (b-Actin used as loading reference). B, Luciferase reporter assays demonstrate a reduction of YAP/TAZ- TEAD activity upon RNAi-mediated YAP/TAZ depletion (mean of quintuplicates þ SD; , P < 0.001). C, Significant reduction of cell viability (MTT assay) upon YAP or TAZ knockdown in SySa cells (mean of quintuplicates þ SD; , P < 0.001). D, Viability of five SySa cell lines is significantly decreased upon treatment with increasing concentrations of verteporfin (MTT assay). RD cells are included as verteporfin-sensitive control (mean of three independent experiments SD). E, SySa cells treated with increasing concentrations of verteporfin show a reduction of YAP, TAZ, FOXM1, CTGF and PLK1 protein levels (b-Actin used as loading reference). F, Verteporfin treatment results in a significant reduction of TEAD luciferase reporter activity in SySa cells (mean of quintuplicates þ SD; , P < 0.001). G, In flow cytometric analyses, significantly increased rates of apoptosis (cleaved PARP) are detected upon treatment with increasing concentrations of verteporfin in SySa cells (mean of two independent experiments SD; , P < 0.01; , P < 0.05). H, In CME-1 cells, TEAD luciferase reporter activity is significantly increased upon overexpression of constitutively active DS127A YAP or DS89A TAZ, which is reverted by incubation with verteporfin (mean of quintuplicates þ SD; , P < 0.001).

3726 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

phosphorylation of Histone H3 (Ser10; Fig. 4A). To exclude treated with verteporfin and doxorubicin showed a significant unspecific off-target effects, two additional siRNAs were tested reduction of SySa tumor growth (Fig. 5D and E). As expected, and (Supplementary Fig. S3A). In both SySa cell lines, YAP/TAZ consistent with the in vitro data, treatments with mono verteporfin knockdown significantly reduced TEAD luciferase reporter activity and the combination of verteporfin and doxorubicin resulted in a (Fig. 4B) and significantly suppressed SySa cell viability in MTT reduction of YAP, TAZ, FOXM1, CTGF, and PLK1 protein levels assays (Fig. 4C), pointing to an essential role of YAP/TAZ-medi- compared with the control and mono doxorubicin treatment ated signals in SySa. (Supplementary Fig. S4).

Verteporfin reduces cell viability, inhibits YAP/TAZ-mediated transcriptional activity and induces apoptosis in SySa cells in Discussion vitro Synovial sarcoma (SySa) is a malignant soft tissue tumor of To investigate the growth-suppressive effects of YAP/TAZ-TEAD adolescence and young adulthood with a particular propensity to fi inhibition, ve SySa cells were exposed to increasing concentra- develop local relapse and distant metastases (2, 8). Prognosis in fi fi tions of vertepor n. In MTT assays, vertepor n was effective in the metastasized situation is poor (4). SySa is molecularly char- suppressing SySa cell viability with IC50 values ranging from 0.13 acterized by a SS18-SSX gene fusion encoding an aberrant tran- m to 0.51 mol/L, showing a dose-dependent mode of action scriptional regulator that has been shown to be sufficient to drive (Fig. 4D; Table 2). YAP/TAZ-driven rhabdomyosarcoma RD cells tumorigenesis in a conditional MYF5-CRE transgenic mouse included as positive control cells (33) were less sensitive to model with 100% penetrance (42). Although it has been con- fi vertepor n treatment compared to SySa cells. As demonstrated vincingly shown that SS18-SSX associates with SWI/SNF and fi in epithelial cells before (40, 41), in all ve analyzed SySa cell Polycomb chromatin complexes to dysregulate gene expression fi fi lines, vertepor n led to a signi cant and dose-dependent reduc- in a genome-wide fashion through the disruption of epigenetic tion of YAP/TAZ and their downstream targets (Fig. 4E; Supple- regulation (9, 11), knowledge about the actual oncogenic signals mentary Fig. S3B). Consistently, TEAD luciferase reporter assays effected by SS18-SSX is still limited. We and others have previ- indicated a dose-dependent suppression of YAP/TAZ-TEAD tran- ously shown that Insulin-like growth factor-dependent signals, fl scriptional activity (Fig. 4F; Supplementary Fig. S3C). In ow the subsequent PI3K/AKT cascade, the SRC kinase as well as WNT/ fi cytometric analyses, treatment with vertepor n induced a signif- b-catenin signals contribute to SySa tumorigenesis (20, 21, 26, icant and dose-dependent increase of PARP (Asp214) cleavage in 29, 38, 39). Since targeting of the chimeric SS18-SSX fusion three SySa cell lines (Fig. 4G; Supplementary Fig. S3D). TEAD protein itself represents a particular challenge, it appears reason- luciferase reporter activity induced through overexpression of able to therapeutically address the signaling pathways which are D D constitutively active S127A YAP or S89A TAZ in SySa cells known to be functionally dependent on the primary oncogenic fi fi decreased upon vertepor n treatment, indicating speci c suppres- driver, that is, the SS18-SSX fusion protein. In-depth insights into sion of YAP/TAZ-mediated transcriptional activity (Fig. 4H). the dysregulated signaling pathways are therefore elementary to design tailored therapeutic concepts. Given first data docu- In vivo efficacy of verteporfin in SySa cell line–based CAM, menting a role of YAP/TAZ in mesenchymal tumorigenesis mouse xenografts, and a SySa PDX model (16–18), we set out to analyze in detail the prevalence and As a final step of our preclinical evaluation of the efficacy of functional relevance of YAP/TAZ signals in SySa tumorigenesis. YAP/TAZ-TEAD inhibition on SySa tumor growth, we applied Aiming at a truly biologically motivated therapeutic approach, we several in vivo models. First, we performed chick CAM assays put a particular focus on the analysis of the functional dependency xenografting SYO-1 cells to initiate SySa tumor formation. Topical of YAP/TAZ signals on the chimeric SS18-SSX fusion protein verteporfin administration resulted in a significant and dose- before testing a small molecule YAP/TAZ-TEAD inhibitor in a dependent reduction of tumor volumes compared to the DMSO preclinical setting in different model systems. vehicle control group (Fig. 5A). In the next step, SYO-1 cells were In a large cohort of tumor specimens and a representative set of subcutaneously xenografted into the lower flank region of NSG SySa cell lines, we detected a high prevalence of nuclear YAP/TAZ mice, receiving: (i) mono verteporfin, (ii) mono doxorubicin, (iii) expression, associated with concordant expression of the down- combination of verteporfin and doxorubicin, or (iv) DMSO stream targets FOXM1 and/or PLK1 in 83% of primary SySa and all vehicle control. Intraperitoneal administration of verteporfin and SySa cell lines (Fig. 1). FOXM1 expression has previously been doxorubicin resulted in a significant reduction of tumor volumes investigated in SySa and was found to be prognostically rele- compared with the DMSO vehicle control group. Highest growth- vant (43). Addressing the impact of SS18-SSX on YAP/TAZ signals suppressive effects were observed under combined administra- in two different model systems, we unveiled nuclear translocation tion of verteporfin and doxorubicin (Fig. 5B and C). In the final of YAP/TAZ and associated YAP/TAZ-TEAD signaling to be func- step, human SySa tissue was subcutaneously transplanted into the tionally dependent on the chimeric SS18-SSX fusion protein, both lower flanks of NMRI nu/nu mice, receiving the identical treat- in RNAi-based experiments in SySa cells and through stable over- ment protocol. Consistent with the SYO-1 xenografts, PDX mice expression of SS18-SSX in SCP-1 mesenchymal stem cells (Fig. 2).

Table 2. IC50 values for verteporﬁn in synovial sarcoma cell lines

IC50 (mmol/L) Compound SYO-1 (SS18-SSX2) FUJI (SS18-SSX2) 1273/99 (SS18-SSX2) CME-1 (SS18-SSX2) HS-SY-II (SS18-SSX1) Verteporﬁn 0.51 0.14 0.18 0.09 0.27 0.16 0.13 0.03 0.41 0.08 Cytotoxic effects on synovial sarcoma cell viability were assessed in MTT assays (72 hours). Results are represented as mean SD of at least three independent experiments performed in quintuplicates.

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3727

Isfort et al.

Figure 5. In vivo efficacy of verteporfin in SySa cell line–based CAM, mouse xenografts and a SySa PDX model. A, SYO-1 cells were xenografted on the CAM of chick eggs. Tumor-bearing eggs were randomized and topically treated with verteporfin or vehicle control (DMSO). Significantly reduced tumor volumes þ SD and representative CAM explants are shown (scale bar: 1 mm, , P < 0.001; , P < 0.05). B, Intraperitoneal mono- or combined administration of verteporfinand doxorubicin (indicated by the ^ arrow) results in a significant reduction of SYO-1 xenograft tumor volumes compared with the DMSO vehicle control group (normalized tumor volumes are indicated as mean SEM with 3 mice/group). C, Normalized SYO-1 mice xenograft tumor volumes on day 19 (top, mean þ SD; , P < 0.01; , P < 0.05; ns, not significant) and representative tumor explants (bottom, diameter: 20 mm; scale bar, 10 mm). D, Intraperitoneal mono- or combined administration of verteporfin and doxorubicin (indicated by the ^ arrow) results in a significant reduction of SySa PDX tumor volumes compared with the DMSO vehicle control group in NMRI nu/nu mice (normalized tumor volumes are indicated as mean SEM with 4 mice/group). E, Normalized SySa PDX tumor volumes on day 39 (top, mean þ SD; , P < 0.001; , P < 0.01) and representative PDX tumor explants (bottom; scale bar: 10 mm).

3728 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

In view of the known example of an "indirect" fusion-dependent the case of WNT signaling but to be rather (indirectly) mediated mechanism of YAP activation in alveolar rhabdomyosarcoma through the SySa-specific SS18-SSX fusion protein (21, 29). (exerted via PAX3-FOXO1-dependent upregulation of RASSF4 Because there is convincing evidence of a simultaneous activation leading to inhibition of MST1 and MOB1) (33), we wondered if of several different signaling pathways contributing to tumori- a comparably complex molecular network might be effective in genesis in SySa, future therapeutic concepts need to be based on SySa cells. Integrating the established role of SS18-SSX as a molec- an integrated signaling network concept. ular driver of IGF2 expression (26, 37) resulting in an activation of On the basis of our data, SySa joins alveolar rhabdomyosar- IGF-IR/PI3K/AKT signal transduction (20, 38), we here disclosed coma, in which the PAX3-FOXO1 fusion protein was also shown that an IGF-II/IGF-IR signaling loop contributes to aberrant YAP/ to promote tumorigenesis by dysregulation of YAP (33), to make TAZ activation through dysregulation of the Hippo effectors up a subgroup of soft tissue tumors in which a specific chimeric LATS1 and MOB1. Consistently, molecular inhibition of IGF-IR fusion protein effects its oncogenic signal through dysregulation reconstituted the function of the "negative" Hippo effectors, of the Hippo pathway. In another fusion-driven soft tissue tumor, resulting in phosphorylation of YAP and TAZ, thereby contribut- epithelioid hemangioendothelioma, Hippo pathway disruption ing to downregulation of YAP/TAZ signals (Fig. 3). This molecular is realized through oncogenic translocations which directly circuit represents a novel mechanism to regulate "active" nuclear involve the YAP and WWTR1 (encoding TAZ) genes, leading to YAP/TAZ levels; however, it is in good agreement with the concept an aberrant signaling pathway activation (49, 50). These two of regulatory inputs of diverse signaling pathways on the Hippo oncogenic mechanisms add to what has been reported by axis (13, 44, 45). We are aware of the limitations of the signaling Eisinger-Mathason and colleagues in non fusion-driven, karyo- concept proposed here, as it naturally presents only an excerpt of a typically complex sarcomas, in which genetic events affecting the much more complex system - however, we render a conclusive tumor-suppressive Hippo regulators LATS2, NF2, and SAV1 were model integrating diverse signaling pathways driven by the SS18- found in 39% of all 261 sarcomas deposited in the TCGA SSX fusion. Yet, it appears probable to us that further signaling database (16). The diversity of activation mechanisms of YAP/ inputs beyond IGF-IR-mediated signals exist and even direct TAZ signals in soft tissue sarcomas represents a particular chal- effects of SS18-SSX on YAP/TAZ are still conceivable. lenge for molecular tumor diagnostics because the identification Targeting the transcriptionally active YAP/TAZ-TEAD com- of an appropriate predictive biomarker is mandatory before plex, either by specific RNAi-mediated functional depletion of entering molecularly based therapeutic studies. With regard to YAP/TAZ or through pharmacologic inhibition with vertepor- the current state of knowledge, IHC screening for nuclear YAP/ fin,ledtoasignificant and dose-dependent impact on SySa cell TAZ expression could provide such a biomarker which should be growth, coupled with the expected regulation of YAP/TAZ- prospectively assessed. TEAD transcriptional activity. Although verteporfinisknown In conclusion, the results of our study demonstrate that acti- to have off-target effects so that our data need to be interpreted vation of YAP/TAZ signals is a common pattern in SySa and is with caution, it may be regarded as a proof of specificity in the functionally dependent on the chimeric SS18-SSX fusion protein. context of SySa that we could show that verteporfinisableto Disruption of YAP/TAZ-mediated signal transduction via a small- revert experimental overactivation of YAP/TAZ-dependent gene molecule inhibitor may provide an effective and novel therapeutic transcription, as measured in TEAD luciferase reporter assays. approach for the treatment of SySa. Aiming at a further insight in the concrete mode of action of verteporfin, we demonstrated a significant dose-dependent Disclosure of Potential Conflicts of Interest induction of apoptosis in SySa cells due to YAP/TAZ-TEAD K. Steinestel reports receiving speakers bureau honoraria from Boehringer inhibition (Fig. 4). Consistent with our in vitro results, topical Ingelheim, and is a consultant/advisory board member for AstraZeneca. administration of verteporfin to SYO-1 CAM xenografts led E. Wardelmann is a consultant/advisory board member for Novartis, Milestone, fl to a significant suppression of SySa tumor growth in vivo.This PharmaMar, Roche, Nanobiotix, Bayer, and Lilly. No potential con icts of finding could eventually be translated to SYO-1 NSG mouse interest were disclosed by the other authors. xenografts and a human SySa PDX model, in which we observed significant therapeutic effects of mono-treatment with Authors' Contributions verteporfin and in combination with doxorubicin (Fig. 5). Conception and design: I. Isfort, M. Trautmann, W. Hartmann The enhanced therapeutic efficacy of the combination of verte- Development of methodology: I. Isfort, M. Cyra, A. Wozniak, M. Trautmann, fi W. Hartmann por n and doxorubicin observed in the SySa PDX model might Acquisition of data (provided animals, acquired and managed patients, imply a role of YAP/TAZ inhibition in counteracting mechan- provided facilities, etc.): I. Isfort, S. Elges, S. Kailayangiri, B. Altvater, isms of chemoresistance as reported in epithelial tumors (46, C. Rossig, K. Steinestel, I. Grunewald,€ S. Huss, J.-H. Mikesch, S. Hafner, 47); however, further systematic experiments are required to A. Wozniak, P. Schoffski,€ E. Wardelmann, W. Hartmann evaluate this therapeutic option. In summary, our findings Analysis and interpretation of data (e.g., statistical analysis, biostatistics, argue in favor of a relevant functional role of YAZ/TAZ signals computational analysis): I. Isfort, M. Cyra, S. Kailayangiri, B. Altvater, C. Rossig, S. Huss, S. Hafner, M. Trautmann, W. Hartmann in the oncogenic translation of SS18-SSX fusion protein-driven Writing, review, and/or revision of the manuscript: I. Isfort, M. Cyra, S. Elges, effects and disclose a novel, molecularly based therapeutic S. Kailayangiri, B. Altvater, C. Rossig, K. Steinestel, I. Grunewald,€ S. Huss, target in SySa. J.-H. Mikesch, A. Wozniak, P. Schoffski,€ E. Wardelmann, M. Trautmann, Given the described lack of genomic alterations including W. Hartmann effectors of the Hippo pathway in SySa (16, 48), the setting of Administrative, technical, or material support (i.e., reporting or organizing € Hippo YAP/TAZ signaling in SySa to some extent parallels what data, constructing databases): I. Isfort, K. Steinestel, P. Schoffski, O.L. Larsson, b M. Trautmann, W. Hartmann we and others have shown for the activation of the WNT/ -catenin Study supervision: M. Trautmann, W. Hartmann pathway in SySa. Both effects seem mainly not to be due to genetic Other (discussion of content and initial ideas): E. Eßeling alterations in pathway effectors such as APC and/or b-catenin in Other (designed and supervised the CAM experiments): T. Simmet

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3729

Isfort et al.

Acknowledgments Stiftung (2016.099.1; to W. Hartmann and M. Trautmann, and K. Steinestel), The authors thank Charlotte Sohlbach, Inka Buchroth, Christin Fehmer, and and the "Innovative Medical Research" funding program of the University of € Christian Bertling for excellent technical support and Jasmien Wellens for Munster Medical School (IMF; TR121716 and TR221611 to M. Trautmann; excellent technical help with PDX in vivo experiments. Dr. Attila Aszodi, HU121421, to M. Trautmann and S. Huss). The experimental PDX work was Experimental Surgery and Regenerative Medicine, Department of General, partially supported by research grant from Kom op tegen Kanker (Stand up to € Trauma, and Reconstructive Surgery, Ludwig-Maximilians-University of Cancer), the Flemish Cancer Society (grant to P. Schoffski). Munich (Munich, Germany), kindly provided SCP-1 cells. Dr. Shinya Tanaka, Laboratory of Molecular & Cellular Pathology, Hokkaido University Graduate The costs of publication of this article were defrayed in part by the payment of School of Medicine (Sapporo, Japan) kindly provided FUJI cells. Dr. Hiroshi page charges. This article must therefore be hereby marked advertisement in Sonobe, Department of Laboratory Medicine, Chungoku Central Hospital accordance with 18 U.S.C. Section 1734 solely to indicate this fact. (Fukuyama, Hiroshima, Japan) kindly provided HS-SY-II cells. Dr. Akira Kawai, Division of Orthopaedic Surgery, National Cancer Center Hospital (Tokyo, Japan) kindly provided SYO-1 cells. The study was supported by grants from the Received November 28, 2017; revised December 2, 2018; accepted February DFG (HA 4441/2-1; to W. Hartmann and M. Trautmann), the Wilhelm-Sander- 21, 2019; published ﬁrst February 27, 2019.

References 1.LurkinA,DucimetiereF,VinceDR,DecouvelaereAV,CellierD,Gilly 18. Fullenkamp CA, Hall SL, Jaber OI, Pakalniskis BL, Savage EC, Savage JM, FN, et al. Epidemiological evaluation of concordance between initial et al. TAZ and YAP are frequently activated oncoproteins in sarcomas. diagnosis and central pathology review in a comprehensive and pro- Oncotarget 2016;7:30094–108. spective series of sarcoma patients in the Rhone-Alpes region. 19. Trautmann M, Menzel J, Bertling C, Cyra M, Isfort I, Steinestel K, et al. FUS- BMC Cancer 2010;10:150. DDIT3 fusion protein-driven IGF-IR signaling is a therapeutic target in 2. Suurmeijer AJH, de Bruijn D, Geurts van Kessel A, Miettinen MM. myxoid liposarcoma. Clin Cancer Res 2017;23:6227–38. World Health Organization Classification of Tumours of soft tissue and 20. Friedrichs N, Kuchler J, Endl E, Koch A, Czerwitzki J, Wurst P, et al. Insulin- Bone. In:Fletcher CDM, Bridge JA, Hogendoorn PCW, Mertens F, editors. like growth factor-1 receptor acts as a growth regulator in synovial sarcoma. World Health Organization Classification of Tumours of soft tissue and J Pathol 2008;216:428–39. Bone. 4 edition. Geneva, Switzerland: WHO Press; 2013. p.213–5. 21. Trautmann M, Sievers E, Aretz S, Kindler D, Michels S, Friedrichs N, et al. 3. Ferrari A, De Salvo GL, Brennan B, van Noesel MM, De Paoli A, Casanova M, SS18-SSX fusion protein-induced Wnt/beta-catenin signaling is a thera- et al. Synovial sarcoma in children and adolescents: the European Pediatric peutic target in synovial sarcoma. Oncogene 2014;33:5006–16. Soft Tissue Sarcoma Study Group prospective trial (EpSSG NRSTS 2005). 22. Bocker W, Yin Z, Drosse I, Haasters F, Rossmann O, Wierer M, et al. Ann Oncol 2015;26:567–72. Introducing a single-cell-derived human mesenchymal stem cell line 4. Scheer M, Dantonello T, Hallmen E, Vokuhl C, Leuschner I, Sparber-Sauer expressing hTERT after lentiviral gene transfer. J Cell Mol Med 2008;12: M, et al. Primary metastatic synovial sarcoma: experience of the CWS Study 1347–59. Group. Pediatr Blood Cancer 2016;63:1198–206. 23. Gibault F, Corvaisier M, Bailly F, Huet G, Melnyk P, Cotelle P. Non- 5. Spurrell EL, Fisher C, Thomas JM, Judson IR. Prognostic factors in advanced photoinduced biological properties of verteporfin. Curr Med Chem 2016; synovial sarcoma: an analysis of 104 patients treated at the Royal Marsden 23:1171–84. Hospital. Ann Oncol 2005;16:437–44. 24. Liu-Chittenden Y, Huang B, Shim JS, Chen Q, Lee SJ, Anders RA, et al. 6. Haldar M, Randall RL, Capecchi MR. Synovial sarcoma: from genetics to Genetic and pharmacological disruption of the TEAD-YAP complex sup- genetic-based animal modeling. Clin Orthop Relat Res 2008;466:2156–67. presses the oncogenic activity of YAP. Genes Dev 2012;26:1300–5. 7. Ladanyi M. Fusions of the SYT and SSX genes in synovial sarcoma. 25. Carboni JM, Wittman M, Yang Z, Lee F, Greer A, Hurlburt W, et al. BMS- Oncogene 2001;20:5755–62. 754807, a small molecule inhibitor of insulin-like growth factor-1R/IR. 8. Nielsen TO, Poulin NM, Ladanyi M. Synovial sarcoma: recent discov- Mol Cancer Ther 2009;8:3341–9. eriesasaroadmaptonewavenuesfor therapy. Cancer Discov 2015;5: 26. Michels S, Trautmann M, Sievers E, Kindler D, Huss S, Renner M, et al. SRC 124–34. signaling is crucial in the growth of synovial sarcoma cells. Cancer Res 9. Banito A, Li X, Laporte AN, Roe JS, Sanchez-Vega F, Huang CH, et al. The 2013;73:2518–28. SS18-SSX oncoprotein hijacks KDM2B-PRC1.1 to Drive Synovial Sarcoma. 27. Arteaga MF, Mikesch JH, Qiu J, Christensen J, Helin K, Kogan SC, et al. The Cancer Cell 2018;33:527–41. histone demethylase PHF8 governs retinoic acid response in acute pro- 10. Garcia CB, Shaffer CM, Eid JE. Genome-wide recruitment to Polycomb- myelocytic leukemia. Cancer Cell 2013;23:376–89. modified chromatin and activity regulation of the synovial sarcoma 28. Lubieniecka JM, de Bruijn DR, Su L, van Dijk AH, Subramanian S, van de oncogene SYT-SSX2. BMC Genomics 2012;13:189. Rijn M, et al. Histone deacetylase inhibitors reverse SS18-SSX-mediated 11. Kadoch C, Crabtree GR.Reversible disruption of mSWI/SNF (BAF) com- polycomb silencing of the tumor suppressor early growth response 1 in plexes by the SS18-SSX oncogenic fusion in synovial sarcoma. Cell 2013; synovial sarcoma. Cancer Res 2008;68:4303–10. 153:71–85. 29. PrettoD,BarcoR,RiveraJ,NeelN,GustavsonMD,EidJE.Thesynovial 12. Soulez M, Saurin AJ, Freemont PS, Knight JC. SSX and the synovial- sarcoma translocation protein SYT-SSX2 recruits beta-catenin to the sarcoma-specific chimaeric protein SYT-SSX co-localize with the human nucleus and associates with it in an active complex. Oncogene 2006;25: Polycomb group complex. Oncogene 1999;18:2739–46. 3661–9. 13. Moroishi T, Hansen CG, Guan KL. The emerging roles of YAP and TAZ in 30. Mohamed A, Sun C, De Mello V, Selfe J, Missiaglia E, Shipley J, et al. The cancer. Nat Rev Cancer 2015;15:73–9. Hippo effector TAZ (WWTR1) transforms myoblasts and TAZ abundance is 14. Pan D. The hippo signaling pathway in development and cancer. Dev Cell associated with reduced survival in embryonal rhabdomyosarcoma. 2010;19:491–505. J Pathol 2016;240:3–14. 15. Zhao B, Li L, Lei Q, Guan KL. The Hippo-YAP pathway in organ size 31. Lei QY, Zhang H, Zhao B, Zha ZY, Bai F, Pei XH, et al. TAZ promotes cell control and tumorigenesis: an updated version. Genes Dev 2010;24: proliferation and epithelial-mesenchymal transition and is inhibited by 862–74. the hippo pathway. Mol Cell Biol 2008;28:2426–36. 16. Eisinger-Mathason TS, Mucaj V, Biju KM, Nakazawa MS, Gohil M, Cash TP, 32. Zhao B, Wei X, Li W, Udan RS, Yang Q, Kim J, et al. Inactivation of YAP et al. Deregulation of the Hippo pathway in soft-tissue sarcoma promotes oncoprotein by the Hippo pathway is involved in cell contact inhibition FOXM1 expression and tumorigenesis. Proc Natl Acad Sci U S A 2015;112: and tissue growth control. Genes Dev 2007;21:2747–61. E3402–11. 33. Crose LE, Galindo KA, Kephart JG, Chen C, Fitamant J, Bardeesy N, et al. 17. St John MA, Tao W, Fei X, Fukumoto R, Carcangiu ML, Brownstein DG, Alveolar rhabdomyosarcoma-associated PAX3-FOXO1 promotes et al. Mice deficient of Lats1 develop soft-tissue sarcomas, ovarian tumours tumorigenesis via Hippo pathway suppression. J Clin Invest 2014; and pituitary dysfunction. Nat Genet 1999;21:182–6. 124:285–96.

3730 Clin Cancer Res; 25(12) June 15, 2019 Clinical Cancer Research

YAP/TAZ Signals in Synovial Sarcoma

34. Syrovets T, Gschwend JE, Buchele B, Laumonnier Y, Zugmaier W, Genze F, 42. Haldar M, Hancock JD, Coffin CM, Lessnick SL, Capecchi MR. A condi- et al. Inhibition of IkappaB kinase activity by acetyl-boswellic acids tional mouse model of synovial sarcoma: insights into a myogenic origin. promotes apoptosis in androgen-independent PC-3 prostate cancer cells Cancer Cell 2007;11:375–88. in vitro and in vivo. J Biol Chem 2005;280:6170–80. 43. Maekawa A, Kohashi K, Kuda M, Iura K, Ishii T, Endo M, et al. Prognostic 35. Tomayko MM, Reynolds CP.Determination of subcutaneous tumor size significance of FOXM1 expression and antitumor effect of FOXM1 inhi- in athymic (nude) mice. Cancer Chemother Pharmacol 1989;24: bition in synovial sarcomas. BMC Cancer 2016;16:511. 148–54. 44. Fan R, Kim NG, Gumbiner BM. Regulation of Hippo pathway by mitogenic 36. Cornillie J, Wozniak A, Pokreisz P, Casazza A, Vreys L, Wellens J, et al. In growth factors via phosphoinositide 3-kinase and phosphoinositide- vivo antitumoral efficacy of PhAc-ALGP-doxorubicin, an enzyme-activated dependent kinase-1. Proc Natl Acad Sci U S A 2013;110:2569–74. doxorubicin prodrug, in patient-derived soft tissue sarcoma xenograft 45. Kim D, Shu S, Coppola MD, Kaneko S, Yuan ZQ, Cheng JQ. Regulation of models. Mol Cancer Ther 2017;16:1566–75. proapoptotic mammalian ste20-like kinase MST2 by the IGF1-Akt path- 37. de Bruijn DR, Allander SV, van Dijk AH, Willemse MP, Thijssen J, van way. PLoS One 2010;5:e9616. Groningen JJ, et al. The synovial-sarcoma-associated SS18-SSX2 fusion 46. Bartucci M, Dattilo R, Moriconi C, Pagliuca A, Mottolese M, Federici G, et al. protein induces epigenetic gene (de)regulation. Cancer Res 2006;66: TAZ is required for metastatic activity and chemoresistance of breast cancer 9474–82. stem cells. Oncogene 2015;34:681–90. 38. Friedrichs N, Trautmann M, Endl E, Sievers E, Kindler D, Wurst P, et al. 47. Pan W, Wang Q, Zhang Y, Zhang N, Qin J, Li W, et al. Verteporfin can reverse Phosphatidylinositol-30-kinase/AKT signaling is essential in synovial sar- the paclitaxel resistance induced by YAP over-expression in HCT-8/T cells coma. Int J Cancer 2011;129:1564–75. without photoactivation through inhibiting YAP expression. Cell Physiol 39. Sun Y, Gao D, Liu Y, Huang J, Lessnick S, Tanaka S. IGF2 is critical for Biochem 2016;39:481–90. tumorigenesis by synovial sarcoma oncoprotein SYT-SSX1. Oncogene 48. Barretina J, Taylor BS, Banerji S, Ramos AH, Lagos-Quintana M, Decarolis 2006;25:1042–52. PL, et al. Subtype-specific genomic alterations define new targets for soft- 40. Garcia-Rendueles ME, Ricarte-Filho JC, Untch BR, Landa I, Knauf JA, Voza tissue sarcoma therapy. Nat Genet 2010;42:715–21. F, et al. NF2 loss promotes oncogenic RAS-induced thyroid cancers via YAP- 49. Errani C, Zhang L, Sung YS, Hajdu M, Singer S, Maki RG, et al. A novel dependent transactivation of RAS proteins and sensitizes them to MEK WWTR1-CAMTA1 gene fusion is a consistent abnormality in epithelioid inhibition. Cancer Discov 2015;5:1178–93. hemangioendothelioma of different anatomic sites. Genes Chromosomes 41. Wang C, Gu C, Jeong KJ, Zhang D, Guo W, Lu Y, et al. YAP/TAZ- Cancer 2011;50:644–53. mediated upregulation of GAB2 leads to increased sensitivity to growth 50. Tanas MR, Ma S, Jadaan FO, Ng CK, Weigelt B, Reis-Filho JS, et al. factor-induced activation of the PI3K pathway. Cancer Res 2017;77: Mechanism of action of a WWTR1(TAZ)-CAMTA1 fusion oncoprotein. 1637–48. Oncogene 2016;35:929–38.

www.aacrjournals.org Clin Cancer Res; 25(12) June 15, 2019 3731

SS18-SSX−Dependent YAP/TAZ Signaling in Synovial Sarcoma

Ilka Isfort, Magdalene Cyra, Sandra Elges, et al.

Clin Cancer Res 2019;25:3718-3731. Published OnlineFirst February 27, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-17-3553

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2019/02/26/1078-0432.CCR-17-3553.DC1

Cited articles This article cites 49 articles, 16 of which you can access for free at: http://clincancerres.aacrjournals.org/content/25/12/3718.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://clincancerres.aacrjournals.org/content/25/12/3718. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.