Hippo pathway effectors YAP1/TAZ induce a EWS-FLI1-opposing gene signature and associate with disease progression in Ewing Sarcoma

YAP1 and TAZ (WWTR1) oncoproteins are the final transducers of Hippo tumor suppressor pathway. Deregulation of the pathway leads to YAP1/TAZ activation fostering tumorigenesis in multiple malignant tumor types, including sarcoma. However, oncogenic mutations within the core components of the Hippo pathway are uncommon. Ewing Sarcoma (EwS), a pediatric cancer with low mutation rate, is characterized by a canonical fusion involving EWSR1 gene, and FLI1 as the most common partner. The fusion protein is a potent driver of oncogenesis but secondary alterations are scarce, and little is known about other biological factors that determine the risk of relapse or progression. We have observed YAP1/TAZ expression and transcriptional activity in EwS cell lines. Analyses of 55 primary human EwS samples revealed that high YAP1/TAZ expression was associated with progression of the disease and predicted poorer outcome. We did not observe recurrent SNV or copy number gains/losses in Hippo pathway-related loci. However, differential CpG methylation of RASSF1 locus -a regulator of Hippo pathway- was observed in EwS cell lines compared with mesenchymal stem cells, the putative cell of origin of EwS. Hypermethylation of RASSF1 correlated with the transcriptional silencing of the tumor suppressor isoform RASFF1A, and transcriptional activation of the protumorigenic isoform RASSF1C promoting YAP1/TAZ activation. Knockdown of YAP1/TAZ decreased proliferation and invasion abilities of EwS cells, and revealed that YAP1/TAZ transcription activity is inversely correlated with the EWS-FLI1 transcriptional signature. This transcriptional antagonism could be partly explained by EWS-FLI1-mediated transcriptional repression of TAZ. Thus, YAP1/TAZ may override the transcriptional program induced by the fusion protein, contributing to the phenotypic plasticity determined by dynamic fluctuation of the fusion protein, a recently proposed model for disease dissemination in EwS.


INTRODUCTION
Ewing Sarcoma (EwS) represents the second most common primary malignant bone tumor in children and young adults [1]. Owing to multimodal treatment concepts, 2/3 of the patients with localized disease achieve sustained remission but approximately 30 % relapse. Patients at relapse or with advanced disease have limited chance to survive with a three-year event free survival of less than 25% [2,3]. While clinical prognostic markers such the presence of metastases or tumor volume are established, little is known about the biological factors determining the risk of progression, thus precluding risk-adapted therapeutic approaches. EwS was the first solid malignancy defined by the presence of tumor-specific EWSR1-ETS gene fusions [4], mainly EWSR1-FLI1 translocations, which are considered the main driver of the disease, but fusion type itself does not have any impact on disease progression [5]. As in most developmental cancers, additional recurrent mutations are scarce. The most common somatic mutations have been detected in STAG2, CDKN2A and TP53, associated with poor prognosis [6,7]. Copy number variation studies by the PROVABES consortium using samples supporting previous retrospective studies [7,8]. However, these secondary alterations occur with a frequency which does not account for the large proportion of patient who relapses.
The Hippo tumor suppressor pathway plays a critical role in tissue and organ size regulation by restraining cell proliferation and apoptosis under homeostatic conditions [9]. Central to Hippo pathway is a conserved cascade of adaptor proteins and inhibitory kinases that regulate the activity of the oncoproteins YAP1 and TAZ, the final effectors of this pathway in mammals. YAP1/TAZ do not directly bind DNA, but act as transcriptional coactivators of target genes involved in cell proliferation and survival through their interaction with transcriptional regulators such as TEAD factors [10]. The role of YAP1 and TAZ as important drivers in tumorigenesis has been extensively reported in carcinomas, and they also contribute to malignancies of mesenchymal origin [11][12][13]. In fact, given its key function in developmental processes, an important role has been inferred for Hippo signaling in pediatric cancer [14]. Despite this, somatic or germline mutations in Hippo pathway genes are uncommon, in comparison to other well-defined signaling pathways that are commonly disrupted in cancer [13,15]. Since secondary genetic alterations are scarce in EwS, and given the established role of YAP1 and TAZ in cancer without engaging mutation, we aimed to explore the contribution of these factors to oncogenesis in Ewing sarcoma. Herein we evaluated a series of 55 EwS patients by immunohistochemistry (IHC) for expression/activation of YAP1 and TAZ. We observed a significant association of YAP1/TAZ nuclear expression and disease progression, as well as a potential mechanism of dysregulation involving epigenetic regulation of RASSF1 locus. Moreover, we demonstrated an interesting interplay between TAZ/YAP1 function with the fusion protein, which fits into a recent model concept for metastatic spreading in EwS based on fluctuations of the expression of the fusion protein [16].

Tumor samples
In this study we analyzed 88 formalin-fixed paraffin-embedded (FFPE) samples from 68 Ewing sarcoma patients (55 samples corresponding to primary tumor). We also analyzed a subset of 21 frozen samples from the same series. Clinical diagnosis of all the samples was performed according to the World Health Organization (WHO) classification [17], performing fluorescence in situ hybridization (FISH) to assess the presence of EwS translocation in tissue sections, which validates the immunohistochemical diagnosis. The only selection criteria were the availability of pathological data and tissue for tissue microarray (TMA) construction. Medical records were retrospectively reviewed and clinicopathologic information for 55 patients with primary tumor material were retrieved for further analyses (summarized in Table 1

TMA construction and Immunohistochemistry
Representative tumor areas of EwS samples were selected on H&E-stained sections and two 1-mm diameter tissue cores were obtained from each specimen to set up 4 different TMAs. IHC was carried out on TMA sections using the Envision method (Dako, CA, USA) with a step of heat-induced antigen retrieval and using a primary antibody against YAP1 and TAZ (Suppl. Table 1). IHC staining was separately evaluated by two pathologists. YAP1/TAZ expression was evaluated for nuclear staining, thus focusing in their transcriptional activity. Tissue was given a score which resulted of multiplying the nuclear staining intensity from 0 (no staining) to 3 (strong staining), by the extension based on the percentage of positive cells (from 0 to 3).

Migration assay
Migration assay was performed as described previously [23]. Cells were grown on 6 wells plates to 85-90% confluence and a wound was made by scratching the monolayer of the cells. Pictures of the same selected area were taken after the times indicated and the percentage of wound healing was calculated using the imageJ software.
These chambers were coated with BME (Cultrex Basement Membrane Extract), a natural extracellular matrix hydrogel. It was diluted at 0.2X in coating buffer and added to the inserts 24 h before cells were seeded. Cells were cultured in medium without serum for at least 24 h before the assay. Afterwards, a volume of 100 µl of serum-free medium containing 1x10 5 cells was deposited on the top of the chamber. Serum was used as chemoattractant in the lower part of the chamber. After 48h at the incubator at 37C, both serum-free medium from the top and pure serum from the lower part of the chamber were removed. The top of the membrane was carefully cleaned and the bottom retaining invading cells was washed with PBS. Cells were fixed to the membrane with methanol 100% for 5 minutes at -20ºC. After PBS washing, cells were stained with DAPI and membranes were mounted on a slide to be observed under a fluorescence microscope (Olympus BX-61). Cell nucleus of the respective conditions were counted and compared with controls.

Transcriptome analysis
SK-N-MC cells were transfected with control or a combination of YAP1/TAZ siRNAs for 72h. Whole transcript expression analysis was conducted in four biological replicates of each sample. RNA was amplified and labeled using the GeneChip® WT PLUS Reagent Kit (Thermo Fisher Scientific, Inc.). Amplification was performed with 100 ng of total RNA input following procedures described in the WT PLUS Reagent Kit user manual. The amplified cDNA was quantified, fragmented, and labeled in preparation for hybridization to GeneChip® Human Transcriptome 2.0 Array (Thermo Fisher Scientific, Inc.) using 5.5 μg of single-stranded cDNA product and following protocols outlined in the user manual. Washing, staining (GeneChip® Fluidics Station 450, Thermo Fisher Scientific, Inc.), and scanning (GeneChip® Scanner 3000, Thermo Fisher Scientific, Inc.) were performed following protocols outlined in the user manual for cartridge arrays. The fluorescence signals scanned as DAT files were transformed to CEL files via the AGCC software (Thermo

YAP1/TAZ are expressed in EwS cell lines and tumor specimens, and are associated with the presence of metastasis and poor prognosis.
First, we examined YAP1/TAZ expression by WB in 13 EwS cell lines with different pathognomonic gene fusions (Fig. 1A). We observed heterogeneous expression of both proteins across the cell line panel. Some of the EwS cell lines showed YAP1/TAZ expression comparable to cell lines in which a relevant role has been described for these factors (i.e. MDA-MB-231, a triple negative breast cancer cell line with NF2-mutations leading to activation of TAZ/YAP1) [28]. TAZ/YAP1 expression was also detected in human mesenchymal stem cells (hMSC) derived from bone marrow, a proposed cell of origin of EwS. Importantly, nuclear expression was observed by subcellular fractionation and immunofluorescence ( Fig. 1B-C, Suppl. Fig. 1), suggesting functional transcriptional activity which was confirmed with luciferase reporter assays (Fig. 1D).
To test whether YAP1/TAZ abundance was associated with clinical variables in EwS, we analyzed their expression by IHC in a retrospective series of 55 primary tumors (Table 1). YAP1/TAZ strong expressing tumor cells exhibited intense nuclear staining with a variable signal in the cytoplasm ( Fig. 2A). YAP1/TAZ expression was also observed in endothelial cells in negative samples providing an internal positive control for the IHC determination (Fig. 2B). YAP1/TAZ strong expression was associated to disease progression (chisquare test, p<0.0054), whereas no significant association was observed with age at surgery or location (Table 1). We also observed increased YAP1/TAZ positivity in metastatic or relapsed tumors in 11 patients with paired samples (Fig. 2C-H, 2I, paired t-test, p = 0.0204). Additional non-paired metastatic or relapsed tumor samples showed preferential strong expression as well (Fig. 2J 2L). Accordingly, Cox regression univariate analyses determined that YAP1/TAZ strong expression was significantly correlated with the time to relapse but not with EwS specific survival, with the unadjusted hazard ratio (HR) being 3.354 (p=0.016) and 1.928 (p=0.167) respectively ( Table 2). A significant correlation with survival and time to relapse was also observed for metastasis ( Table 2). These variables were all included simultaneously, to assess the independent prognostic significance based on multivariate analysis. The adjusted HR of YAP1/TAZ strong expression for relapse did not reach significant confidence regarding time to relapse, after controlling the Coxˈs regression model for the effects of age, tumor location and metastasis.
However, a roughly significant HR for YAP1/TAZ was obtained in the multivariant analysis ( Table 2).

Activation of YAP1 /TAZ in Ewing Sarcoma
We tried to determine the mechanisms that contribute to YAP1/TAZ activation in EwS. To do so, we interrogated public datasets for somatic mutations in the Hippo pathway-related genes, but we did not find any recurrent SNV (Supp. Fig. 2). Next, we analyzed copy number alterations in a series of 24 EwS by SNP arrays (Fig. 3). Gross chromosomal alterations were similar to previous reports, i.e. gains of whole chromosomes 8 and 12 [7]. Copy number gain in WWTR1 locus, with complete gain of chromosome 3 was detected in a single case. Gain at YAP1 locus was detected in another case with an almost tetraploid genotype. None of the two cases showed incremented mRNA expression associated to the copy number event. Regarding the core regulatory kinases of the Hippo pathway and other negative regulators of YAP1/TAZ function, no significant copy-loss events were observed (Fig. 3). Focal copy number aberration events in Hippo-related loci were also precluded after inspecting the data with the STAC algorithm (Suppl.  (Fig. 4B).
Moreover, expression of YAP1/TAZ target genes positively correlated with RASSF1C expression in the cell line panel, as well as in EwS tumor specimens (Fig. 4C, D). Interestingly, TAZ but not YAP1 seems to be transcriptionally regulated since CTGF expression correlate with TAZ mRNA expression (Fig. 4D). Correlation of TAZ mRNA levels with Hippo target genes was also observed in larger EwS series in public repository expression data (Suppl. Fig. 3).
There is extensive evidence that Src can promote YAP/TAZ activity through a variety of mechanisms, i.e. Src, and other SFKs can directly phosphorylate YAP1 and TAZ promoting their activity and stability [31]. Therefore, since RASSF1C activates SFKs in RASSF1-methylated cells, we blocked SFK activity by exposing EwS cells to dasatinib. Inhibition of SFKs resulted in reduced cell viability (SK-N-MC IC50=6,55 μM; TTC-466 IC50=2,11 μM) and upregulation of YAP1/TAZ target genes (Fig. 5A). Upon dasatinib treatment mRNA levels of YAP1 and TAZ remained unaffected, but TAZ protein expression was decreased and YAP1 inactivating phosphorylation increased in both cell lines (Fig. 5A). As an alternative approach of pharmacologic blockade of YAP1/TAZ activity we tested pitavastatin. Statins prevent nuclear localization of YAP1/TAZ via inhibition of the enzyme HMG-CoA reductase, ultimately affecting the metabolic control of YAP1/TAZ by the mevalonate pathway [32]. We also observed antiproliferative effect upon pitavastatin treatment (SK-N-MC IC50= 1,83 μM; TTC-466 IC50=1,86 μM), with mild reduction of YAP1/TAZ target genes and TAZ protein downregulation ( Fig   5A). Neither dasatinib nor pitavastatin treatments affected EWS-FLI1 expression in SK-N-MC cell line, thus precluding the antiproliferative effect of these drugs to be mediated by the fusion protein.

YAP1/TAZ loss-of-function affects cell proliferation and invasion capacity in EwS cells
To assess the oncogenic properties of YAP1 and TAZ in EwS cells, we induced transient knockdown of YAP1, TAZ or simultaneous depletion of both factors, and evaluated cell proliferation, invasion and migration capacity of the silenced cells. We observed inhibition of proliferation in knockdown cells for every individual or combined siRNA transfection. Individual depletion of YAP1 inhibited cell growth more efficiently than TAZ silencing (Fig 5B). YAP1/TAZ silenced cells showed a significantly reduced invasive capacity as well (Fig 5C).
Migration capacity of EwS cells upon YAP1/TAZ silencing was not significantly altered as compared to the control, but a slight trend towards diminished migration was observed in the doble-silenced cells (Suppl. Fig.   6).

YAP1/TAZ transcription activity is anti-correlated with EWS-FLI1 transcriptional signature
To evaluate the transcriptome modulation by YAP1/TAZ we conducted gene expression profiling by Affymetrix microarrays in SK-N-MC cells upon simultaneous silencing of both factors. We observed differential expression of 938 coding genes (Suppl. Table 3) including well-stablished YAP1/TAZ target genes, such as CYR61, CTGF or AMOT, which were confirmed by qPCR analyses in two EwS cell lines with different gene fusions (Fig. 6A). Similar results were obtained with individual silencing of each factor (Suppl. Fig. 5). Of note, expression levels of EWS-FLI1 were not affected in SK-N-MC (Fig. 6A) and other EwS cell lines tested (Suppl. Fig. 5).
Next, we collated this transcriptional profile with previously published curated gene sets. Interestingly, we found significant enrichment for several EwS-related gene signatures both in YAP1/TAZ-correlated and anticorrelated genes (Fig. 6C). YAP1/TAZ-anticorrelated genes were significantly over-represented among EwS induced gene sets, and inversely YAP1/TAZ-correlated genes overlapped with EwS repressed genes.
Thus, suggesting opposite transcriptional activity of EWS-FLI1 fusion gene and YAP1/TAZ factors.
Accordingly, depletion of the EWS-FLI1 protein in the A673 EwS cell line resulted in the induction of YAP1/TAZ-regulated genes, as well as TAZ but not YAP1 factor (Fig. 6C, D).  Fig. 8).

Discussion
In the present study, we have shown that YAP1/TAZ expression associates with disease progression and poor  [45,46]. Notwithstanding, aberrant activation of YAP1/TAZ in cancer is often promoted by mechanisms not involving somatic alterations. We have observed that epigenetic regulation of the RASSF1 locus could affect the expression of YAP1/TAZ target genes in EwS cell lines (Fig. 4). This result may explain previous observations describing a correlation of hypermethylation of RASSF1 and RASSF2 with worse clinical outcome in EwS [47,48]. Moreover, Src kinase activation of invadopodia in response to stress in EwS [49] could be related to SFK-mediated activation of YAP1/TAZ by RASSF1C (Fig. 5A). However, YAP1/TAZ activation does not seem to rely on RASSF1 hypermethylation in hMSC (Fig. 4), the putative cell of origin of EwS, which exhibits high expression levels of YAP1 and TAZ (Fig. 1). Unaffected expression levels of YAP1 and derepression of TAZ upon EWS-FLI1 silencing (Fig. 6C, D) also supports the notion that both factors are maybe expressed in the cell of origin, as proposed for ZEB2, an EMT (epithelial-mesenchymal transition) inducer like YAP1 and TAZ [50].
The association of YAP1/TAZ with metastatic spread could be arguably related to the relative levels of the fusion protein, recently reported to promote phenotypic plasticity of EwS cells [16]. In this scenario, YAP1/TAZ may promote a mesenchymal phenotype in EWS-FLI1 depleted EwS cells together with Wnt/beta-catenin [51], since it is well-established that the crosstalk between Hippo and wnt signaling is essential for tumor progression in several types of cancer [52]. As it has been described for Wnt/beta-catenin [51], the opposing transcriptional signature between YAP1/TAZ and EWS-FLI1 could partly contribute to the metastatic process.
I.e. We found strong downregulation or upregulation of LOX (a mediator of metastasis [16]) in YAP1/TAZ- . Another mechanism contributing to the opposing gene signatures might involve Ewing sarcoma-associated transcript 1 (EWSAT1), which was found to be significantly induced in YAP1/TAZ-silenced SK-N-MC cells (Suppl. Table 3). EWSAT1 is a long noncoding RNA that mediates EWS-FLI1 gene repression via interaction with a heterogeneous nuclear ribonucleoprotein [54]. In addition, we have observed inhibition of TAZ expression associated to the presence of EWS-FLI1, which also binds DNA at WWTR1 locus (Suppl.

Fig. 8).
Indeed, regulation of TAZ seems to occur at the transcriptional level, whereas YAP1 activity is not correlated with mRNA levels (Figs. 4C, 6E).
In summary, our study reveals that the interplay between Hippo pathway effectors YAP1/TAZ and the function of the gene fusion is relevant to shape the transcriptional program in EwS. The transcriptional output elicited by these factors deserves further characterization since our observations provide clinical evidence that YAP1/TAZ expression associates with disease progression in EwS patients. Studies with larger prospective series are needed in order to corroborate our observations and to stablish whether YAP1/TAZ could serve as reliable biomarkers to stratify and identify patients who could benefit from targeted therapies.

Acknowledgments
This research has been conducted using samples from the Hospital Universitario Virgen del Rocío-Instituto de