EWS-FLI1-mediated tenascin-C expression promotes tumour progression by targeting MALAT1 through integrin α5β1-mediated YAP activation in Ewing sarcoma

Background The extracellular matrix has been critically associated with the tumorigenesis and progression of Ewing sarcoma (ES). However, the regulatory and prognostic roles of tenascin-C (TNC) in ES remain unclear. Methods TNC expression was examined in specimens by immunohistochemistry, and the association of TNC expression with ES patient survival was also analysed. TNC-knockout cell lines were constructed using CRISPR/Cas9 methods. In vitro experiments and in vivo bioluminescent imaging using BALB/c nude mice were conducted to evaluate the effect of TNC on ES tumour progression. RNA sequencing was performed, and the underlying mechanism of TNC was further explored. Results TNC was overexpressed in ES tissue and cell lines, and TNC overexpression was associated with poor survival in ES patients. TNC enhanced cell proliferation, migration and angiogenesis in vitro and promoted ES metastasis in vivo. The oncoprotein EWS-FLI1 profoundly increased TNC expression by directly binding to the TNC promoter region. Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) upregulation induced by Yes-associated protein (YAP) activation was responsible for TNC-regulated ES tumour progression. Activated integrin α5β1 signalling might be correlated with YAP dephosphorylation and nuclear translocation. Conclusions TNC may promote ES tumour progression by targeting MALAT1 through integrin α5β1-mediated YAP activation.

MALAT1 through integrin α5β1-mediated YAP activation. These findings reveal that TNC may serve as a prognostic and therapeutic candidate in ES.

ES specimens and clinical follow-up
Fifty-two ES specimens and thirty-seven adjacent normal tissue specimens were collected after approval by the hospital institutional review boards. This study was carried out in accordance with the Declaration of Helsinki, and written informed consent was obtained from all participants. The general characteristics and outcomes of ES patients are illustrated in Table 1. Metastasis-free survival (MFS) was defined as the period from the first day after operation to the date of metastasis confirmation or October 31 st , 2018. The overall survival (OS) was defined as the period from the first day after surgery until the date of disease-related death or October 31st, 2018.
Immunohistochemistry (IHC) and western blotting Paraffin-fixed tissue sections were stained via IHC to analyse the expression of TNC and CD99. Incubation with primary antibodies and horseradish peroxidase (HRP)-coupled secondary antibody was performed according to recommendations. The results of IHC were evaluated by experienced pathologists in a blinded manner.
Cells were lysed in NP40 cell lysis buffer (Invitrogen) containing a 1:1000 proteinase inhibitor cocktail (Biotool) according to the manufacturer's recommendations. After standard quantification, separation, and membrane transfer, the immunoperoxidase method was utilised with SuperSignal TM West Pico PLUS Chemiluminescent substrate (Thermo Scientific) to detect proteins. All the primary and secondary antibodies used are listed in Supplementary Table 1.

Cell culture ES cell lines (A673 [CVCL_0080] and SKNMC [CVCL_0530]) and
human bone marrow stromal cells (BMSCs, isolated from normal donors as previously described 19 ) were cultured in RPMI 1640 medium (Gibco) containing 10% foetal bovine serum (FBS, Gibco) and 100 U/ml penicillin/streptomycin (P/S, Gibco). Human umbilical vein endothelial cells (HUVECs [CVCL_2959]) and HEK293 cells (CVCL_0045) were grown in EGM-2 medium (Lonza) and DMEM (HyClone) with 10% FBS, respectively. Cultures were maintained in a 37°C incubator in 5% CO 2 . The purity and authenticity of the cell lines were assessed via short tandem repeat profiling. Cells were tested regularly to ensure they were free of mycoplasma infections. Cells were cultured to a proper density and sub-cultured every 3-4 days according to the methods of a previous study. 7 Exogenous recombinant TNC was purchased from Abcam. Verteporfin and saracatinib were purchased from Selleckchem.
CRISPR/Cas9-mediated TNC knockout (ko) For TNC-ko experiments, a CRISPR/Cas9-sgRNA genome-wide library (CRISPR-Pool™ KOUT, GeneChem) was used according to previous studies. 20,21 The targeted sequence was GTTGCCCCGA CCGCTACAGA. Generally, the viral vector was co-transfected with pVSVg and psPAX2 into HEK 293T cells with Lipofectamine 2000 (Invitrogen). After 48-72 h of co-transfection in DMEM, the viral supernatant was collected, filtered through a 0.45 µm filter, and then ultra-centrifuged at 25,000 rpm for 2 h at 4°C. Quality tests were performed to confirm the viral titre without contamination. Then, viral pellets were resuspended in RPMI 1640 medium and added to A673 and SKNMC cell cultures for 24 h, followed by puromycin (Gibco) selection for seven days. Stable cell lines were verified by Sanger sequencing and western blotting analysis at the gene and protein levels, respectively.
qRT-PCR Total RNA was extracted using TRIzol (Invitrogen) and reverse transcribed into cDNA with PrimeScript™ RT Master Mix (TaKaRa). Different gene transcripts were quantified on a 7900HT Fast Real-Time PCR System (Life Technologies) with SYBR Premix Ex Taq™ II (TaKaRa). The primers used for qRT-PCR in this study are listed in Supplementary Table 2. Cell viability and migration assays Cells were seeded in 96-well plates (5 × 10 3 /well) with or without exogenous TNC and cultured for 48 h. Cell Counting Kit-8 reagent (Dojindo Laboratories) (10 µl/well) was added and incubated for 1.5 h at 37°C. The absorbance at 450 nm was measured using an ELx800 microplate reader (BioTek Instruments, Inc.). The average absorbance of each sample was calculated from five duplicates.
For the cell migration assay, cells (1 × 10 5 /well) were added to the upper chamber of a Transwell migration chamber, with 0.1% FBS medium and 10% FBS medium loaded into the lower chamber (8 μm pore size Transwell migration chamber, Corning). After 24 h of incubation, cells were washed and fixed with 4% paraformaldehyde for 20 min. Non-invasive cells were removed from the upper compartment with a cotton swab. After crystal violet staining for 15 min, cells that had migrated through the membrane from four random fields in each membrane were counted with Image-Pro Plus 6.0 software.
Tube formation assay Matrigel (BD Bioscience) had previously been thawed at 4°C overnight. Then, 50 µl of chilled Matrigel was seeded in a 96-well plate and incubated at 37°C for 45 min. HUVECs (1 × 10 4 /well, 70-80% confluence) suspended in 100 µl of EGM-2 medium and conditioned medium from TNC-wild-type (wt) or TNC-ko ES cells were added to the solidified Matrigel. The rescue capacity was simultaneously evaluated in the presence of exogenous TNC. After  14,23 Briefly, the mice were anaesthetised (using isoflurane inhalation) and intraperitoneally injected with 150 mg/kg D-luciferin potassium salt (AOK Chem). After 10 min, the mice were imaged and assessed using Living Image 3.2 software (Caliper). The animal study was performed and assessed by two different investigators to minimise subjective bias.
Luciferase reporter assay Recombinant plasmid was established from the synthesised promoter region of TNC (nt −2000~0) (TsingKe) and the linearised pGL3-basic plasmid using a Quick-Fusion Cloning Kit (Biotool). The EWS-FLI1 and FLI1 overexpression plasmids were also constructed using the kit above. Dual-luciferase assays were conducted in a 24-well plate. The recombinant plasmid and phRL-TK100 Renilla reporter plasmid, together with the EWS-FLI1 or FLI1 overexpression plasmids, were transfected into 60-70% confluent HEK293 cells. After 48 h of co-transfection, a dual-luciferase assay kit (Promega) was used to quantify firefly and Renilla luciferase activities based on the manufacturer's instructions.
Chromatin immunoprecipitation (ChIP) assay ChIP assays were conducted with a ChIP assay kit (Millipore) according to the manufacturer's instructions. Formaldehyde (1%) was used to crosslink the chromatin in A673 cells (10 min) at room temperature, and glycine was generated to terminate the crosslinking. The lysis buffer (with SDS added) was then sonicated to generate 200 to 500 bp DNA fragments. Anti-FLI1 antibody was incubated with the DNA fragments, which were subjected to ChIP assays. Normal IgG was used as a nonspecific control. Then, the eluted DNA was enriched and investigated by PCR. The primers used in the ChIP assay are listed in Supplementary ICC was performed to detect YAP expression in ES cells using immunofluorescence (IF) staining. 9 The FISH and IF signals were monitored by two independent investigators with a BX51 fluorescence microscope (Olympus).
Statistical analysis SPSS v. 19.0 software and GraphPad Prism v.6.01 software were used for statistical analysis. Quantitative data are described as the mean/median (standard deviation) and were compared by Student's t-test and ANOVA with Tukey's multiple comparisons test when appropriate. Qualitative data are reported as counts/ percentages and were compared with the chi-square test or Fisher's exact test when appropriate. Correlation analysis was performed by using Spearman rank correlation analysis. Survival probability was estimated by the Kaplan-Meier method, and logrank tests were used to compare metastasis-free survival (MFS) and overall survival (OS) between groups. All in vitro experiments were repeated three times independently with at least three replicates per experiment, and representative experiments are shown. Differences with p values < 0.05 were considered statistically significant (p-values: *p < 0.05; **p < 0.01; ***p < 0.001).

RESULTS
TNC overexpression in specimens is associated with ES patient survival TNC overexpression was observed in clinical ES tissues (×10 and ×40) and compare to that in adjacent normal tissue via IHC staining (Fig. 1a). The IHC results were assessed by two investigators independently, and thirty-seven (71.2%) specimens tested positive for TNC compared to sixteen (43.2%) para-tumour tissue specimens that tested positive for TNC. The difference in TNC expression (positive/negative ratio) between tumour and adjacent normal tissue was statistically significant (p < 0.01) (Fig. 1b). As shown in Supplementary Table 1, the mean followup time of the enrolled patients was 21.9 ± 13.3 months (range 3.0-68.0 months). Thirty-eight (73.1%) patients developed metastasis, and 35 (67.3%) patients died because of the disease. The estimated MFS and OS rates were 25.2 and 24.6%, respectively, according to the Kaplan-Meier method. The mean MFS and OS were 23.3 ± 3.8 and 30.0 ± 3.5 months, respectively. IHC revealed that the percentage of specimens strongly positive for TNC was 30.8% (n = 16). When analysing the follow-up data, the log-rank test was further used to evaluate the impact of TNC positivity on patient disease progression and overall survival. The Kaplan-Meier curve showed that strong positivity for TNC was correlated with a worse MFS (p = 0.012) (Fig. 1c), and ES patients with strong positivity for TNC also had a worse OS (p = 0.02) than those who were not strongly positive for TNC (Fig. 1d). Moreover, since CD99 is a specific marker for ES found in the cytoplasmic membrane of 100% ES tumour cells, 25,26 the IHC double-staining method was applied to detect the co-expression of TNC and CD99 in ES tissue. TNC (brown) and CD99 (purple) were co-expressed at high levels in ES tissue under ×10 and ×40 magnification (Fig. 1e). TNC overexpression in ES tissue and cells was then confirmed using western blotting (Fig. 1f, g).
TNC depletion inhibits ES cell proliferation, migration and angiogenesis TNC-ko cell lines were initially established using the CRISPR/Cas9 genome editing technique as described above. 20,21 After puromycin selection for seven days, clones were generated by multiple proportion dilution and cultured in 96-well plates. DNA was extracted when cells in 24-well plates were 90-100% confluent and sequenced using the Sanger method. A frameshift mutation around "AGG" in both A673 and SKNMC cells was confirmed ( Supplementary Fig. 1), and the knockdown of TNC expression was further validated via western blotting (Fig. 2a).
A CCK8 assay to assess A673 and SKNMC cell viability showed that the proliferation rates of the TNC-ko groups were significantly lower than those of the TNC-wt groups (p < 0.01) and that the addition of exogenous TNC failed to rescue tumour cell growth in both the TNC-wt and TNC-ko groups (Fig. 2b). As shown by migration experiments, the knockout of TNC inhibited ES  migration in both A673 (p < 0.01) and SKNMC cells (p < 0.001) (Fig. 2c). However, the addition of exogenous TNC had no significant impact on the migration capacity of A673 and SKNMC cells. This is probably due to the use of recombinant TNC (amino acids 181-290) instead of full-length purified TNC. 10 Due to the pivotal role of the blood supply in tumorigenesis, we further performed tube formation experiments to evaluate the role of TNC in ES-induced angiogenesis. The knockout of TNC impaired angiogenesis compared with the TNC-ko groups, but the difference in capillary-like structures after the addition of exogenous TNC was no different between the TNCko and TNC-wt groups (in both A673 and SKNMC cells). Representative images and relative statistics are illustrated in Fig. 2d, e.

Metastasis enhancement is driven by TNC
The metastatic colonisation abilities of the TNC-wt and TNC-ko groups (A673) determined by bioluminescence imaging were compared in in vivo experiments. The mice in our study had similar weights before experiments, and no mouse died due to severe adverse effects during the experiments. The mean total flux of the mice (n = 8 per group) injected with TNC-wt cells was higher than that of mice injected with TNC-ko cells 19 days after injection (p < 0.05), and this difference was highest at day 26 after inoculation (p < 0.001). Representative bioluminescence images at day 26 and growth curves are shown in Fig. 2f-h.

EWS-FLI1 significantly activates TNC expression in ES cells
Flag-EWS-FLI1 and Flag-FLI1 plasmids were transfected into ES cells, and the overexpression of both EWS-FLI1 and FLI1 was found to promote TNC expression according to immunoblotting analysis (A673 and SKNMC cells, Fig. 3a). EWS-FLI1 expression was knocked down by transfecting cells with siRNA encoding EF-2RNAi and Ctrl-RNAi sequences (referring to a previous study 27 ). We then used anti-FLI1 and anti-EWS antibodies to detect EWS-FLI1 expression and found that anti-FLI1 could detect the EWS-FLI1 fusion oncoprotein without FLI1, while anti-EWS could detect both the EWS-FLI1 and EWS proteins. Consistent TNC downregulation was confirmed in both A673 and SKNMC cells (Fig. 3b). qRT-PCR was then performed to detect TNC mRNA expression, and the differences in TNC mRNA expression between the EWS-FLI1overexpressed vs. ctrl groups (p < 0.001 in A673 cells, p < 0.001 in SKNMC cells), FLI1-overexpressed vs. ctrl groups (p < 0.01), and EWS-FLI1-deficient vs. ctrl groups (p < 0.01) were statistically significant (Fig. 3c). Meanwhile, a dual-luciferase reporter assay revealed higher relative luciferase activities in the presence of EWS-FLI1 (p < 0.001) or FLI1 (p < 0.01) compared with those in the control groups (A673 and SKNMC cells). Moreover, EWS-FLI1 seemed to have a higher affinity for the TNC promoter than FLI1 did (Fig. 3d). To determine whether TNC is transactivated by EWS-FLI/FLI, we searched for the EWS-FLI/FLI consensus sequence in the promoter region of the TNC gene from 2 kb upstream to exon 1. We identified five putative binding sites and examined the relative luciferase activities at different sites with a series of luciferase reporter plasmids (Fig. 3e). We built the different recombinant plasmids by inserting the synthesised promoter region into the pGL3-basic plasmid, and TK100 plasmids were transfected together as an internal standard for transfection efficiency normalisation. TNC promoter activity was greatly enhanced by the region encompassing nt −360~−368 (site 4) (p < 0.01) and nt −121~−129 (site 5) (p < 0.001) (Fig. 3f), which indicated that EWS-FLI1/FLI1 might promote TNC expression via binding to sites 4 and 5 of the TNC promoter. A ChIP assay was also performed to further confirm the interaction between EWS-FLI1 and TNC, and the signal for site 4 was higher than that for site 5; thus, site 4 might be the EWS-FLI1-binding site in the TNC promoter (Fig. 3g). MALAT1 is overexpressed in ES and associated with TNC RNA-seq was performed in TNC-wt and TNC-ko A673 cells, and the results were compared to explore the possible downstream genes regulated to TNC. Differential gene expression was delineated with a heat map (Fig. 4a). According to the volcano plot, 3532 genes were upregulated, and 1744 genes were downregulated. Among the differentially expressed genes, MALAT1 expression differed by >16-fold between groups and was one of the most statistically significant differentially expressed genes. KEGG pathway enrichment analysis revealed that the mechanism by which TNC regulates ES tumour progression might be correlated with Hippo signalling pathway (Fig. 4c). We detected and compared MALAT1 expression in ES tissue and adjacent normal tissue by using a fluorescent-labelled oligonucleotide probe. The results showed prevalent MALAT1 overexpression in ES tissue with respect to normal tissue (Fig. 4d). By combining the TNC expression results, we found that MALAT1 was positively expressed in thirty-three specimens, of which twenty-eight were TNC-positive (including sixteen specimens with strong positivity and twelve specimens with moderate positivity). Ten of the remaining nineteen specimens were negative for both MALAT1 and TNC expression. The Spearman chi-square test was further used to analyse the correlation between MALAT1 and TNC expression, and the heat map revealed the robust correlation between MALAT1 and TNC expression in ES specimens (p < 0.01) (Fig. 4e). Further experiments implied that the overexpression of TNC can promote MALAT1 expression in A673 (p < 0.01) and SKNMC (p < 0.001) cells, and the difference in MALAT1 expression between the TNC-wt and TNC-ko cell lines was statistically significant (p < 0.001). These results indicated that MALAT1 is associated with TNC-regulated ES tumour progression.
TNC may regulate MALAT1 through integrin α5β1-mediated YAP activation Further analysis showed that the expression of key downstream target genes of the Hippo signalling pathway (CTGF and CYR61) was downregulated in TNC-ko cells compared to TNC-wt cells (Fig. 4b). Since YAP, a key transcriptional cofactor in the Hippo signalling pathway, has been reported to regulate MALAT1 expression by binding to the MALAT1 promoter, 17,28 we investigated whether TNC upregulates MALAT1 through YAP. Although the total YAP expression levels were unchanged by TNC expression, phosphorylated YAP levels in A673 cells were decreased, and CTGF and CYR61 expression levels were positively correlated with the TNC expression level. In addition, the YAP inhibitor serine/arginine-rich splicing factor 1 (SRSF1) was downregulated regardless of TNC overexpression or depletion (Fig. 5a). Furthermore, IF staining revealed that YAP nuclear localisation was decreased in TNC-ko A673 cells compared to TNC-wt A673 cells, while TNC-overexpressing cells exhibited enhanced YAP nuclear translocation compared to TNC-wt cells (Fig. 5b). Meanwhile, MALAT1 expression in A673 cells was decreased significantly after 48 h of pharmacological administration of verteporfin, a potent YAP inhibitor (p < 0.01) (Fig. 5c).
Subsequently, western blotting for TNC-associated receptors showed the increased expression of integrins α5, αV, and β1 with respect to normal tissue (Fig. 5d), but no obvious differences in integrins α2, α7, α8, α9, β3, and β6; Notch 1, 2, 3, and 4; or EGFR expression were found. The qRT-PCR results also indicated a high level of integrin α5 and integrin β1 mRNA expression (p < 0.001), followed by integrin αV mRNA expression (p < 0.01) (Fig. 5e). Further experiments revealed that the blockade of integrin α5 using monoclonal antibody increased YAP phosphorylation in both A673 and SKNMC cells, while the inhibition of integrin αV or EGFR did not influence the phosphorylated status of YAP compared with the control group (Fig. 5f). Then, qRT-PCR showed that MALAT1 expression was downregulated significantly when integrin α5 was blocked in both ES cell lines (p < 0.01), and antibody against integrin αV caused a decrease in MALAT1 expression in only A673 cells (p < 0.05) (Fig. 5g). Moreover, unlike the IgG isotope control, marked YAP cytoplasmic retention was observed after the blockade of integrin α5 but not integrin αV or EGFR (Fig. 5h).
We further detected the expression of kinases downstream of integrin by western blotting. The levels of phosphorylated Src and v-myc myelocytomatosis viral oncogene homologue (MYC) were significantly decreased in TNC-ko A673 cells compared to those in TNC-wt A673 cells, while the level of phosphorylated focal adhesion kinase (FAK) and spleen tyrosine kinase (SYK) seemed not to have changed (Fig. 6a). Meanwhile, pharmaceutical administration of a potent Src inhibitor (saracatinib) resulted in the prominent inhibition of YAP nuclear translocation, as shown by IF (Fig. 6b), and decreased MYC expression, as shown by immunoblotting (Fig. 6c), in contrast with those control A673 cells. These results implied that Src activation is partly responsible for integrin α5β1-mediated YAP activation and that MYC is partly responsible for YAP-mediated MALAT1 upregulation.

DISCUSSION
ECM proteins are pivotal in regulating tumorigenesis and progression in most tumour types through matrix-cell interactions. 15 TNC is one of the most prominent ECM proteins, and the  Fig. 3 Promotion of TNC expression driven by EWS-FLI1. a TNC overexpression was induced by EWS-FLI1 and FLI1 in A673 and SKNMC cells, respectively. b EWS-FLI1 expression was knocked down by transfection with siRNA that encoded EF-2-RNAi sequences (as detected by both anti-FLI1 and anti-EWSR/EWS antibodies), and TNC deficiency was detected EWS-FLI1 silencing. c Statistically significant differences in TNC mRNA expression were found between the EWS-FLI1-overexpressed vs. ctrl groups (p < 0.001 for A673 cells, p < 0.001 for SKNMC cells), FLI1overexpressed vs. ctrl groups (p < 0.01), and EWS-FLI1-deficient vs. ctrl groups (p < 0.01). d Luciferase activity was measured between the groups above (n = 3 replicates per group). e, f Putative EWS-FLI1/FLI1-binding sites in the TNC promoter region (n = 3 replicates per group). f ChIP assay indicated that site 4 (−360~−368) is the probable binding site for the interaction of EWS-FLI1 with TNC. Representative data from one of three independent experiments are shown overexpression of TNC is associated with many aspects of tumour cell behaviours in various tumours. 8 However, the regulatory role of TNC and the underlying mechanism of this role in ES remain obscure. In the current study, we demonstrate that the EWS-FLI1 oncoprotein, which is characteristic of ES, enhances TNC expression by directly binding to its promoter region and that integrin α5β1-mediated YAP activation may be responsible for MALAT1 upregulation in TNC-induced ES tumour progression (Fig. 6d).
TNC is overexpressed in ES tissue and cell lines compared to its expression in normal tissue and cell lines. In vivo experiments revealed that TNC may contribute to ES metastasis, which is consistent with the results of a previous study. 1 Indeed, autocrine TNC expression enables tumour cells to survive as a metastatic niche component at the early stages of colonisation. 14 Meanwhile, in a clinical study, the presence of metastasis was determined to be the strongest predictor of prognosis, with the overall survival rate decreased from 65% to less than 30% with the presence of metastasis. 5 Our survival analysis found a robust correlation between TNC expression and ES patient survival, and this correlation was also previously reported in breast cancer and osteosarcoma. 9,14 These findings imply that TNC expression predicts ES survival and can serve as a potential prognostic marker for ES in clinical practice.
Furthermore, the knockout of TNC inhibited ES cell proliferation, migration, and angiogenesis in vitro, which indicates that TNC promotes ES tumour progression in an autocrine manner. However, the addition of exogenous TNC failed to antagonise this repression in our study. The use of recombinant TNC (amino  Integrin α5 Integrin α7 Integrin α8 Integrin α9 Integrin αV

Integrin β1
Integrin β3 Integrin β6  acids 181-290) but not full-length purified TNC might account for the failure of TNC to rescue these effects. 10 Notably, we first reported the role of TNC in facilitating ES angiogenesis, which may be related to its regulation of vascular endothelial growth factor (VEGF) expression. 29 Nonetheless, the dual angio-modulatory role of TNC in glioblastoma was previously reported by Rupp and colleagues because TNC regulates both pro-and anti-angiogenic signalling. 13 In addition, in vivo imaging revealed that TNC overexpression promoted distant metastases in immunodeficient mouse models, which indicated that ES patients with TNC overexpression may be more likely to develop metastases. Due to the prevalence of the EWS-FLI1 oncoprotein in ES and its crucial role as an ES biomarker, 5,30 we first detected the putative role of EWS-FLI1 in promoting TNC transcription through bioinformatics analysis and further explored this role by the dual-luciferase reporter and ChIP assays. EWS-FLI1 profoundly promoted TNC expression by putatively binding to site 4 (−360− 368) of the TNC promoter, and Watanabe G and colleagues 31 reported that EWS-ETSs upregulated the TNC gene through four ETS-binding sites between nt −30~−130 in the TNC promoter. This binding indicates that TNC may be closely involved in ES tumorigenesis and at least partly accounts for the EWS-FLI1mediated oncogenic effect. This finding also provides a possible explanation for TNC overexpression observed in most of the clinical ES tumour tissues. Indeed, TNC overexpression driven by EWS-FLI1 in the tumour-associated microenvironment may play an enhanced role in persistently promoting ES to metastasise to distant organs. 14 RNA sequencing and FISH revealed MALAT1 overexpression in ES cells and tumour tissues, which implies that MALAT1 is associated with TNC regulation in ES. Although MALAT1 has been proven to be of the most significant lncRNAs that regulates various tumour types, 17,18,[32][33][34] little information regarding the relationship between TNC and MALAT1 is available. We confirmed the strong correlation between MALAT1 overexpression and TNC in ES tumour tissue. In addition, overexpression of TNC dramatically increased MALAT1 expression, while TNC depletion decreased MALAT1 expression (Fig. 4f). These results collectively indicate that TNC may regulate ES tumour progression by targeting MALAT1.
Moreover, enrichment of the Hippo signalling pathway and the downregulation of its downstream targeted genes (CTGF, CYR61) implies that Hippo signalling is also involved in TNCmediated regulation of ES. We thus explored the potential role of YAP, one of the key transcriptional coactivators of the Hippo signalling pathway, in regulating ES pathogenesis. The overexpression of TNC markedly decreased YAP phosphorylation and enhanced YAP nuclear localisation, as confirmed by ICC. Dephosphorylated YAP further activates itself and translocates to the nucleus by negatively regulating Hippo signalling. 35 Activated YAP also transcriptionally activates MALAT1 by binding to a region of the MALAT1 promoter, 17,28 and this enhanced MALAT1 expression may trigger the expression of downstream genes to positively regulate ES tumour progression. 36 Consistently, in our study, MALAT1 expression was significantly decreased by the pharmacological inhibition of YAP with verteporfin. Altogether, these results suggest that TNC-induced YAP dephosphorylation and nuclear translocation mediate MALAT1 overexpression and negatively regulate the canonical Hippo signalling pathway.
Further experiments revealed that TNC might regulate YAPmediated MALAT1 expression by activating integrin α5β1. Integrin family members, including integrins α9β1, αvβ1, α5β1, and αvβ3, 8 are crucial receptors of TNC that regulate matrix-cell interactions in the cancer microenvironment. Among these integrins, integrin α5β1 is one of the most significant members of the integrin family that influences tumour survival, proliferation, migration, and chemo-resistance. [37][38][39] TNC was reported to regulate various cell behaviours by interacting with integrin α5β1 in both a syndecan-4-dependent and syndecan-4independent manner. For instance, the TNC-integrin α5β1 interaction was shown to block cell adhesion, stimulate tumour cell proliferation, 40 regulate the chemotaxis of human monocytes and polymorphonuclear leucocytes, 41 inhibit fibroblast cell cycle progression, 42 protects cancer stem-like cells from immune surveillance, 43 and enhance cell survival and proliferation in a platelet-derived growth factor-dependent way. 44 We further discovered decreased phosphorylation of Src (but not FAK) induced by knockout of TNC and that YAP cytoplasmic retention was driven by an Src inhibitor. Indeed, integrin α5β1-Src signalling plays a crucial role in different cancer types, [45][46][47] and activated Src is capable of inducing YAP dephosphorylation and nuclear accumulation to regulate the downstream target gene expression. [48][49][50] These results indicate that activated Src, but not FAK, may be partially responsible for TNC/integrin α5β1mediated YAP activation in ES. Moreover, MYC was downregulated ( Fig. 6a) in parallel with high levels of YAP phosphorylation (Fig. 5a) after TNC depletion in A673 cells. MYC, which was reported to be a transcriptional target of YAP in both gastric cancer and osteosarcoma, 51,52 can transcriptionally activate MALAT1 by directly binding to the MALAT1 promoter region. 36 These results indicate that MYC may be partly responsible for YAP-mediated MALAT1 activation. In summary, our study demonstrates that MALAT1 upregulation induced by integrinα5β1-mediated YAP activation may be responsible for TNC-regulated tumour progression in ES and that TNC overexpression is associated with poor survival in ES patients. These findings may provide new insights into the role of TNC in the treatment and prognosis of ES.