Crucial Role of the YAP/TEAD Axis in Osteosarcoma Tumor Growth: Effects of YAP Inhibitors Verteporn and CA3 on Tumor Growth in vivo

Osteosarcoma is the rst primary bone tumor in children and adolescents. Despite progress in the understanding of the biology of these tumors, survival rates have progressed very little in recent decades. In this context, although some studies suggested that disruption of the Hippo signaling pathway is associated with osteosarcoma progression, the molecular mechanisms by which YAP regulates primary tumor growth is not fully claried. In addition, the validation of YAP as a therapeutic target through the use of inhibitors in a preclinical model must be demonstrated. was 3.9 ± 0.7% in the absence of drug, and reached 17.9 ± 7.1% and 18.7 ± 6.9% after 72h treatment of cells with verteporn and CA3, respectively. The percentage of HOS death cells was 0.2 ± 0.1% in the absence of drug, and reached 23.7 ± 4.4% and 2.3 ± 0.3% after 72h treatment of cells with verteporn and CA3, respectively. these that TEAD


Immuno uorescence
Cells were seeded onto Ibidi µ-Slide 8 Well overnight and treated with or without vertepor n or CA3 for 48h, xed with 4% paraformaldehyde and permeabilized with 0.5% Triton. Samples were incubated with Anti-YAP antibody (Cell signaling Technology, Leiden, The Netherlands). F-actin and nucleus were stained using respectively Alexa-uor 488 phalloidin and DAPI. Images were acquired using a confocal microscope (NIKON A1 N-SIM) and processed using ImageJ.
RNA extraction and Real-time polymerase chain reaction RNA was extracted from cells and tumors using NucleoSpin®RNAplus (Macherey Nagel, Duren, Germany) and reverse transcribed using the Maxima H minus rst stand cDNA synthesis kit (Thermo Fisher, Courtaboeuf, France). Real-time monitoring of PCR ampli cation of complementary DNA was performed using DNA primers (primer sequences are available in Table 1 24 hours after transfection, media were changed with fresh DMEM containing 1% FCS for 24 hours. Transfected cells were then rinsed with ice-cold PBS and lysed in IP-lysis buffer (Invitrogen, Courtaboeuf, France). Equal amounts of proteins were precleared overnight at 4 °C using Protein-A/G-agarose (Santa Cruz Biotechnology, CA). Supernatants were incubated with primary antibody against Flag (Sigma Aldrich) and HA-tag (Cell signaling), for 2 h at 4 °C. 50 µL of Protein-A/G-agarose was then added and incubated overnight at 4°C . Beads were washed three times with IP-lysis buffer; thereafter, 30 µL of Laemmli buffer was added and boiled for 5 min. After centrifugation, supernatants were harvested and processed for SDS-PAGE and Western blot as described above.
In situ proximity ligation assay (PLA), immuno uorescence and confocal microscopy Duolink PLA ®: 5x10 3 OS cells were seeded in Ibidi µ-Slide VI 0.4. 24 later, media was changed to DMEM with 1% FBS. Cells were then xed with 4% PFA for 15 min at room temperature and incubated overnight at 4 °C with primary antibody against YAP (Cell signaling or Santa Cruz Biotechnology), TEF-1 (Santa Cruz). In situ PLA was performed using DuoLink in Situ Reagents (Sigma-Aldrich) according to the manufacturer's instructions.
Immuno uorescence assays: cells were seeded onto Ibidi µ-Slide 8 Well overnight, xed with 4% paraformaldehyde for 15 min and permeabilized with 0.5% Triton. Samples were incubated with Anti-Vinculin−FITC antibody (Sigma-Aldrich). F-actin and nucleus were stained using respectively Alexa-uor 488 phalloidin and DAPI. Images were acquired using a confocal microscope (NIKON A1 N-SIM) and processed using ImageJ.
RNA-seq analysis package. The p-values obtained were corrected for false positives by using Independent Hypothesis Weighting (package IHW) multiple testing adjustment method. Genes were considered signi cantly differentially expressed if log2 fold-change was over 1 or less than -1 and FDR was less than 0.05. For the differentially expressed genes, over-representation and gene set enrichment analysis (GSEA) were done using clusterPro ler package, and results were plotted using enrichPlot . GSEA was performed using GSEA software (http://software.broadinstitute.org/gsea/). Gene sets used are described in table 2.

Statistical analysis
Histogram and data are shown as mean +/-S.D. of a minimum of three independent experiments.
Statistical analyses were performed using GraphPad Prism version 6 for Windows (GraphPad Software, La Jolla, CA), www.graphpad.com. The Wilcoxon matched test was used to compare the expression levels between OS and matched normal tissue. The Mann-Whitney test was used to compare the difference between two groups. A p-value under or equal to 0.05 was considered statistically signi cant. Database: RNA sequencing data of OS patient and matched normal tissue were downloaded from the Gene Expression Omnibus database (GSE99671, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE99671).
Kaplan Meier analysis of osteosarcoma patient tumor samples was performed using the R2 Genomics Analysis and Visualization Platform. Genome-wide gene expression analyses of high-grade osteosarcoma are from GSE42352 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE42352).

Results
The elevation of Hippo target genes expression in OS patients is associated with the overall survival of patients.
To resolve outstanding questions regarding the role of the Hippo signaling pathway in OS development, we attempted to identify for a hypothetical Hippo signature in OS patients, using publicly available databases. Gene set enrichment analysis (GSEA) of expression data obtained by using RNAseq assays from a cohort of OS patients reveals a Hippo conserved signature in OS samples as compared to normal bone samples from the same patient ( Figure 1A, Human Osteosarcoma vs. Matched normal tissue, and Figure 1B), suggesting a hyperactivity of the Hippo signaling pathway in OS patients. Indeed, multiple Hippo-regulated genes are signi cantly overexpressed in OS compared with normal bone tissue. These genes include, for example, CYR61, THBS1, PAI-1, and BIRC5, previously described as YAP target genes in tumor tissues ( Figure 1A and Figure S1). Consistently, YAP is overexpressed in OS samples compared with normal tissue from the same patient ( Figure 1C). Interestingly, high YAP transcripts signi cantly correlate with poor survival outcome in OS patients as illustrated in the Kaplan-Meier plot in Figure 1D Taken together, these results highlight that the Hippo/YAP signature correlates with a poor survival outcome in OS patients.

YAP/TEAD interactions are crucial to promote YAP-driven TEAD transcriptional activity in OS cells
Since some previous studies indicated that activation of the Hippo/YAP signaling pathway induces tumor progression through the recruitment of YAP to DNA by the TEAD transcription factor family, we began our analysis by examining the relationship between YAP and TEAD in a panel of three human OS cell lines: HOS, MG63 and G292 cells. In situ PLA assays clearly demonstrate that YAP and TEAD interact (proteins localized within 40 nm of each other) in the nucleus of OS cell lines ( Figure 2A and Figure S2A). To elucidate the role of TEAD in YAP-driven OS development, we then probed the consequences of YAP activation able or unable to interact with TEAD, using overexpression of either YAPS127A (constitutively active, TEAD-binding YAP protein) or YAPS94A (TEAD-binding de cient YAP protein). We rst veri ed the expression and functionality of these mutated YAP expression vectors in transient transfection assays. As anticipated, unlike YAPS127A, YAPS94A proteins are unable to bind TEAD1 ( Figure 2B) and do not induce a transcriptional response in OS cells ( Figure 2C and Figure S2B). Using retroviral infection, we then established K-HOS clones stably overexpressing YAPS127A, YAPS94A or empty vector. As shown in Figure 2D, YAPS94A-and YAPS127A-transfected cells express high levels of YAP mRNA (left panel) and YAP protein (right panel). In situ PLA assays demonstrate increased YAP-TEAD interactions in the nucleus of K-HOS cells in YAPS127A cells compared to YAPS94A-or mock-transfected cells ( Figure 2E). In addition, unlike YAPS94A-and mock-transfected cells, YAPS127A cells exhibit an increased-TEAD transcriptional response as measured by luciferase reporter gene assay with the TEAD-speci c reporter construct (TEAD) 8 -lux ( Figure 2F).
Taken together these results highlight that YAP/TEAD interactions are crucial in the ability of YAP to drive transcriptional activity in OS cells.

OS cell proliferation and in vivo OS tumor growth critically depend on YAP-TEAD interactions
Using these OS cellular tools, we then examined the functional role of TEAD in YAP-driven OS cell proliferation and in vivo tumor growth ( Figure 3A). Real-time proliferation assays demonstrate an increase of OS cell proliferation when YAPS127A is overexpressed compared with the ability of OS cells to proliferate when YAPS94A or an empty vector is overexpressed ( Figure 3B). A preclinical experimental model of OS induced by orthotopic injection of either YAPS94A-, YAPS127A-or mock-transfected OS cells demonstrates the crucial role of TEAD in YAP-driven in vivo OS tumor growth ( Figure 3C). Indeed, 29 days after cell injection, the tumor volume is signi cantly increased when TEAD-interacting YAP is overexpressed (YAPS127A cells versus mock-transfected cells and YAPS127A cells versus YAPS94A cells) ( Figure 3C). In contrast, the tumor volume is signi cantly reduced when TEAD-binding de cient YAP is overexpressed (YAPS94A-versus mock-transfected cells and YAPS94A-versus YAPS127A-transfected cells). Speci cally, the mean tumor size at day 29 was 1043 ± 333 mm3 in the control group (mock-transfected cells), 1517 ± 330 mm3 when TEAD-interacting YAP is overexpressed (YAPS127A cells), and only 599 ± 140 mm3 when TEAD-binding de cient YAP is overexpressed ( Figure 3C).
Taken together, these results demonstrate the crucial role of TEAD in YAP-driven cell proliferation and in vivo tumor growth in OS preclinical models.

Role of TEAD in YAP-driven cell cycle genes expression
To gain more insights into the crucial role of TEAD in YAP-driven cell proliferation and tumor growth, we then compared the RNA sequencing transcriptional pro les of YAPS127A-, YAPS94A-and mock-transfected cells. As shown in Figure S3, mock-, YAPS94A-and YAPS127A-transfected cells display distinct transcriptional pro les, with multiple genes signi cantly differentially expressed.
Transcriptional analysis thus identi es 1,617 genes whose expression is regulated by the overexpression of the YAP mutated proteins able to bind TEAD (YAPS127A) or not (YAPS94A). Of these, 559 genes require the interaction between YAP and TEAD ( Figure 4A). RNAseq analysis identi es 128 genes related to positive regulation of cell proliferation ( Figure 4B) that are signi cantly overexpressed in YAPS127A cells (compared to mock-or YAPS94A-transfected cells). These include genes directly involved in the control of cell cycle, such as CDC25B, and cell proliferation, such as Gli1 or AKT ( Figure 4B, right panel). In contrast, in YAPS94A cells, the expression of some genes related to inhibition of cell proliferation is increased, such as CDKN1A, CDKN1C, CDKN2D and LATS1 ( Figure 4B, right panel). Interestingly, quantitative PCR analysis indicates that the expression of gli1 and AKT genes by tumor cells from mice biopsies and from cultured cells are upregulated when YAP able to interact with TEAD is over-expressed ( Figures 4C and 4D). In addition, GSEA analysis indicates that overexpression of YAP-S127A increase the expression of genes involved on positive regulation of cell cycle G1-S phase transition ( Figure 4E). This strongly demonstrates the crucial role of TEAD in YAP-driven gene expression, which is implicated in the regulation of both OS cell proliferation and in vivo tumor growth. Finally, to investigate the clinical importance of the role played by TEAD in OS tumor development, we analyzed TEAD gene expression using data extracted from the GSE99671 database [26].
Analysis of OS RNAseq data demonstrates that TEAD is overexpressed in OS biopsies compared with control samples from the same patient ( Figure 4F).
Taken together, these results demonstrate: a) the crucial role of TEAD in YAP-driven cell cycle genes expression, and b) that TEAD is overexpressed in OS samples compared with normal tissue from the same patient.
Vertepor n and CA3 inhibit OS primary bone tumor.
To validate YAP/TEAD signaling as a potential therapeutic target for OS treatment, we evaluated the effect of vertepor n and CA3, two Hippo/YAP inhibitors, on primary tumor growth in a preclinical model of OS.
We rst validated that vertepor n and CA3 block the YAP/TEAD cascade, as they inhibit transactivation of the TEAD-speci c reporter construct (TEAD) 8 Figure S4A) two target genes of the YAP/TEAD cascade. In situ PLA assays clearly demonstrate that YAP and TEAD interactions is signi cantly reduced when the cells are treated with vertepor n or CA3 ( Figure 5C and Figure S4B). To elucidate the mechanism underlying the effect of vertepor n and CA3 on YAP/TEAD transcriptional activity, we then evaluated the expression of YAP by immuno uorescence. A shown in Figure 5D and S4C, vertepor n and CA3 reduce the expression of YAP. These results, con rmed by Western-Blot analysis ( Figure 5E and S4D) suggest that vertepor n and CA3 reduce TEAD transcriptional activity mainly by their ability to reduce YAP expression and thus YAP/TEAD interaction.
Importantly, experiments using an orthotopic preclinical model of OS demonstrate that injection of vertepor n or CA3 inhibit the OS tumor growth in vivo ( Figures 6A). Indeed, respectively 30 and 33 days after tumor cell injection, the bone tumor volume is signi cantly decreased in mice treated with vertepor n or CA3. Regarding vertepor n experiments, the mean tumor size at day 30 was 2122 ± 618 mm3 when the mice were treated with vehicle (control group) and only 1258 ± 334 mm3 when the mice were treated with 20 mg/kg of vertepor n ( Figure 6A, right and upper panel). Regarding CA3 experiments, the mean tumor size at day 33 was 2123 ± 535 mm3 when the mice were treated with vehicle (control group) and only 1179 ± 319 mm3 when the mice were treated with 10 mg/kg of CA3 ( Figure   6A, right and lower panel). In vitro assays demonstrate that vertepor n and CA3 affect the cell viability of the three OS cell lines used; HOS, G292, and MG63, in a dose-dependent manners ( Figures 6B). Flow cytometric Annexin V/PI assay showed that vertepor n and CA3 induce early and late apoptotic events, and cell death ( Figure 6C and S5). For example, the percentage of HOS cells in early apoptosis (Annexin V+/PI-) was 3.2 ± 0.6% in the absence of drug, and was 2.1 ± 0.1% and 31.6 ± 2.8% after 72h treatment of cells with vertepor n and CA3, respectively. The percentage of HOS cells in late apoptosis (Annexin V+/PI-) was 3.9 ± 0.7% in the absence of drug, and reached 17.9 ± 7.1% and 18.7 ± 6.9% after 72h treatment of cells with vertepor n and CA3, respectively. The percentage of HOS death cells was 0.2 ± 0.1% in the absence of drug, and reached 23.7 ± 4.4% and 2.3 ± 0.3% after 72h treatment of cells with vertepor n and CA3, respectively.
Together, these results demonstrate that vertepor n and CA3 i) inhibit TEAD transcriptional activity mainly via their ability to reduce YAP expression and thus YAP/TEAD interactions, ii) inhibit in vitro OS cell lines viability, iii) reduce in vivo primary tumor growth and iv) suggest that these later effects are mainly due to their ability to induce cell apoptosis.

Discussion
The lack of response to drugs is a major obstacle to the effectiveness of treatment against OS. Although chemotherapy signi cantly improved the prognosis of patients with osteosarcoma after the introduction of neoadjuvant therapy in the early 1980s [27], the results have not improved since then and are approximately 70% for 5-year survival. The remaining 30% of patients have resistance to several types of chemotherapy. In this context it seems essential to develop new approachs to improve survival.

YAP/TEAD signaling as a target therapy against primary tumor growth
High YAP expression and/or YAP activation have been described in several solid tumor types and correlated with poor prognosis [19]. It has been proposed that YAP acts as an oncogene through activation of target genes that especially promote stimulation of tumor cell proliferation [28,29]. Despite the emerging importance of YAP in many cancers, the exact mechanisms driving key functions in cancer progression still remain to be resolved. While YAP expression was reported in OS, the molecular mechanisms underlying primary tumor growth have not been established in this pathology.
In this work, we rst demonstrate that YAP is highly expressed in biopsies from OS patients and con rm a previous study reporting that YAP expression predicts a poor prognosis in this pathology [23]. We demonstrate the crucial role of YAP in the control of OS cell proliferation and tumor growth. Indeed, the overexpression of a constitutively active YAP (YAPS127A) promotes both the in vitro proliferation of OS cells and the in vivo growth of primary bone tumors. In several cancers, it has been demonstrated that YAP stimulates cell proliferation largely by controlling the expression of a broad number of cell cycle regulators or the expression of oncogenes, for example MYC and AP-1 family members [19]. In this work, we identify genes directly involved in the control of cell cycle, such as CDC25B, or involved in the regulation of oncogene expression, such as Gli1 previously described as a pro-proliferation factor and thus as a potential therapeutic target in OS [30]. Furthermore, using overexpression of a mutant YAP protein unable to interact with TEAD1-4 (YAPS94A), we clearly demonstrate that the TEADs transcriptional factors are crucial in YAP-driven OS growth both in vitro and in vivo as previously described in other cancers [21]. Reinforcing these results, TEAD1 has been found to be highly expressed in OS patients. Together with the previous observation that TEAD1 plays a crucial role in the regulation of OS cell proliferation [25], these results strongly support the hypothesis that the YAP/TEAD axis could represent a promising target to inhibit primary OS tumor growth. Regarding the main role of TEAD transcriptional factor in the expression of YAP-driven genes, we further performed transcriptomic analysis in OS cells that identi ed TEAD-independent and TEAD-dependent modulation of YAP target genes in OS.

Suppression of primary tumor growth by YAP/TEAD inhibitors
To validate YAP/TEAD axis as a potential therapeutic target in OS, we evaluated the effect of two YAP inhibitors, vertepor n and CA3, on OS tumor growth [31]. Vertepor n is a light-activated drug used in photodynamic therapy for the treatment of choroidal neovascular membranes [32]. CA3 is a novel YAP inhibitor recently selected and identi ed through chemical library screening [33]. We speci cally demonstrate that vertepor n and CA3 inhibit primary OS tumor growth. In this context, we show that vertepor n and CA3 block in vitro cell proliferation and induce in vitro cell apoptosis. In accordance with our results, vertepor n was subsequently reported to inhibit the growth of malignant cells without light activation, such as in human glioma [34]. CA3 was seen to strongly inhibits esophageal adenocarcinoma cell growth in vitro and exerts antitumor activity in xenograft model [33]. Both inhibitors suppress mesothelioma cancer stem cell phenotype and tumor formation [35]. Initially described as a YAP/TEAD interaction inhibitor [32], vertepor n was recently described as able to induce the degradation of YAP protein, demonstrating its capacity to target the YAP cascade via different modes of action [35,36]. Here we demonstrate that vertepor n reduces YAP expression and thus YAP-driven TEAD transcriptional activity in OS cell lines. We cannot exclude that vertepor n also inhibit YAP/TEAD direct interaction and therefore YAP/TEAD transcriptional response [37]. Regarding CA3, as previously described in only one recent work using mesothelioma cells, we demonstrated that CA3 induce the decrease of YAP production in OS cell lines [35]. Whatever the mechanisms of action, our work clearly shows that treatment with vertepor n or CA3 reduce YAP/TEAD signaling in OS cells as demonstrated using speci c TEAD promoter/gene reporter assays. In addition, we have shown that vertepor n and CA3 induce cell death by apoptosis in the three cell lines tested, and therefore induce OS cell death and thus inhibits the tumor growth in vivo. These nding are in accordance with previous observations using cultured tumor cells in that vertepor n or CA3 induce apoptosis of tumor cells [35,38].

Conclusion
Our data clearly demonstrated that i) the Hippo/YAP signature correlates with a poor survival outcome in OS patients, ii) the crucial role of TEAD in YAP-driven cell proliferation and in vivo tumor growth in OS, and iii) Vertepor n and CA3, two YAP/TEAD transcriptional inhibitors, signi cantly reduce in vivo primary tumor growth mainly due to their ability to induce cell apoptosis.
In this context, this work forms the basis for the development of better approaches to improve the survival of osteosarcoma patients by identifying the YAP/TEAD axis as a promising therapeutic target. In addition, we demonstrated that YAP/TEAD transcriptional inhibitors, such as vertepor n and CA3, represent promising therapeutic drugs in OS.    [26]. D) Kaplan-Meier analysis of the survival outcome of patients dichotomized into high and low YAP levels, following analysis of the RNAseq dataset GSE42352 [39] from an OS patient cohort comprising 88 samples. Analysis was performed using R2 (http://r2.amc.nl). P value is from log-rank tests.

Figure 2
Role of TEAD in YAP-driven transcriptional activity A) Localization of endogenous YAP/TEAD1 complexes by in situ PLA in HOS cells.
The red signal was obtained using Alexa555-labeled hybridization oligo nucleotides targeting ampli ed in situ PLA products. DAPI