RETRACTED ARTICLE: YAP1-induced MALAT1 promotes epithelial–mesenchymal transition and angiogenesis by sponging miR-126-5p in colorectal cancer

Yes-associated protein 1 (YAP1) exerts significant effects in various malignancies. However, the oncogenic role of YAP1 remains controversial, and the mechanism by which YAP1 regulates non-coding RNAs is still largely unknown. The present study aimed to assess the effect of YAP1 on the malignant behaviors of colorectal carcinoma (CRC) and explore the underlying regulatory mechanism of the YAP1–MALAT1–miR-126-5p axis. YAP1 was highly expressed in CRC tissues as assessed by GSE20916 and its expression was negatively correlated with overall survival in 83 CRC cases. Meanwhile, YAP1 promoted proliferation, invasion, and migration in colon cancer cells, in vitro and in vivo. MALAT1 was obviously expressed, with differential expression of 11 lncRNAs in HCT116 cells after transfection with siYAP1 or si-Ctl. Based on bioinformatics prediction, immunoprecipitation (IP), and chromatin immunoprecipitation (ChIP), the interaction of YAP1 with TCF4/β-catenin was regulated by MALAT1. Bioinformatics prediction, dual luciferase assay, RNA-IP, and RNA pull-down assay demonstrated that YAP1-induced MALAT1 promoted the expression of metastasis-associated molecules such as VEGFA, SLUG, and TWIST, by sponging miR-126-5p in CRC. These findings indicated that the YAP1–MALAT1–miR-126-5p axis could control angiogenesis and epithelial–mesenchymal transition in CRC, providing potential biomarkers and therapeutic targets for CRC.


Introduction
Colorectal carcinoma (CRC) remains one of the most common malignancies worldwide, and is the sixth primary cause of death induced by cancer in China [1,2]. Moreover, the occurrence of CRC metastasis predicted a bad prognosis [3]. Thus, identifying metastasizing biomarkers in the early CRC stage would facilitate CRC diagnosis as well as the development of therapeutic targets.
Long non-coding RNAs (lncRNAs) are transcribed RNA molecules >200 nucleotides in length [4]. With the development of high sequencing technologies, an increasing number of lncRNAs have been shown to be involved in various biological functions, including tumor proliferation, apoptosis, invasion, metastasis, stemness, metabolism, and differentiation [5][6][7]. In CRC, recent studies reported that lncRNAs were involved in regulating tumor progression [8][9][10]. Moreover, increasing lncRNAs showed a potential to be diagnostic or prognostic markers [11][12][13][14]. However, the underlying mechanism of lncRNAs involved in CRC metastasis and clinical significance of lncRNAs in CRC remain largely unknown.
As we knew, YAP1 showed onco-roles in many cancer [15,16]. Our previous findings reported that miR-590-5passociated YAP1 downregulation results in declined CRC tumorigenesis, which illustrated that YAP1 as an oncogene was involved in CRC progression [17]. YAP1 was reported to regulate the maintain stemness, EMT process of cancer [18][19][20]. Moreover, there are a few studies on the regulation relationship between YAP1 and lncRNAs. In pancreatic cancer, Zhou et al. have found that lncRNA MALAT1 is upregulated and the results of si-MALAT1 showed a decline of YAP1 expression [21]. Meanwhile, the fact of lncRNA regulating YAP1 was proved by sponging miR-18a in breast cancer cells. Similarly, lncRNA LUST could directly interact with β-catenin for facilitating it to occupy more promoters related with WNT-signaling target genes including YAP1 in colon cancer stem-like cells [22]. Interestingly, contrary results displayed that si-YAP1 experiment results in a decreasing lncRNA H19 and si-H19 had no effect on YAP1 expression, but no further study [23]. These studies implied that there is a regulation relationship between YAP1 and lncRNAs in cancer. In this study, the role and mechanism of the YAP1 regulating lncRNA were explored in CRC. Based on a series of novel findings, this putative therapeutic target has a potential role in CRC diagnosis and treatment.

YAP1 is an upregulated oncogene in CRC
To assess the role of YAP1 in CRC, we first quantitated YAP1 gene expression levels in CRC biopsies and normal colorectal tissues using a microarray dataset, GSE20916, containing 24 normal colorectal tissues and 81 CRC tissue biopsies, with the Affymetrix HG U133 Plus 2.0 GeneChip platform. Significant analysis of microarray (SAM) [24] was carried out between normal colorectal and CRC tissue samples, and YAP1 was found to be upregulated in CRC samples compared with normal colorectal specimens (p < 0.001) (Fig. 1a). YAP1 overexpression in CRC tissues was confirmed by qRT-PCR and immunohistochemistry (Fig.  1b, c). In addition, the patients were divided into low and high expression groups based on YAP1 mean values. Univariate analysis showed that YAP1 expression was associated with TNM stage, lymphatic metastasis, and distant metastasis (all p < 0.05; Table S1). However, YAP1 expression was not associated with age, tumor location, tumor size, tumor general type, and tumor grade (all p > 0.05). Interestingly, patients with low YAP1 levels survived longer than those with high YAP1 expression (log-rank test, p < 0.05; Fig. 1d). These findings suggested that low YAP1 expression had a positive effect in the prognostic prediction of overall survival (OS) in CRC patients (Table S1). YAP1 was upregulated in colon cancer cells compared with normal colon mucosal cells (NCM460; figure S1a). The results for YAP1 depletion efficiency in vivo and in vitro are shown in figure S4. Subsequently, clonogenic and cell proliferation assays using CCK-8 and EdU staining showed that YAP1 knockdown inhibited HCT116 cell proliferation compared with the negative control group (all p < 0.05; Fig.  1e-g). In addition, Transwell and wound-healing assays showed that YAP1 knockdown suppressed the ability of invasion and migration of HCT116 cells, compared with the negative control group (all p < 0.05; Fig. 1h, i) and vice versa. The above results suggested that YAP1 acted as an oncogene in colon cancer cells in vitro. To assess the biological function of YAP1 in vivo, a nude mice xenograft model was established by subcutaneously injecting HCT116 cells. After 7 days, the subcutaneous tumors were discontinuously injected with RNAi-YAP1 (siYAP1 agomir) or agomir control for 28 days. Tumor volumes and weights in the siYAP1 agomir group were reduced compared with those of the agomir control group (Fig. 1j). YAP1 expression levels in subcutaneous tumors were lower in the siYAP1 agomir group compared with the agomir control group, as evaluated by immunohistochemistry and qRT-PCR (Fig. 1k, l).

YAP1 knockdown downregulates the MALAT1 oncogene in colon cancer cells
Although YAP1 was confirmed as an oncogene in CRC, the underlying mechanism remains undefined. First, gene microarray analysis (GSE92335) [25] revealed differential expression of 11 lncRNAs in HCT116 cells treated with siYAP1 or negative control (Fig. 2a). Similarly, MALAT1 was up-regulated in CRC tissues compared with normal colorectal tissues by analyzing the GSE9348 microarray dataset (Fig. 2b). In addition, MALAT1 was highly expressed in 83 CRC samples, as assessed by qRT-PCR (p < 0.001; Fig. 2c). Moreover, correlation analysis showed that YAP1 expression was closely related to MALAT1 in the GSE14095 dataset, GEPIA database, and 83 CRC samples assessed here (all p < 0.01; Fig. 2d, e). Furthermore, survival analysis indicated that higher expression of MALAT1 was associated with poor OS in all 83 CRC cases (p < 0.05; Fig. 2f); however, the cell localization and biological function of MALAT1 in CRC are largely unknown.

2628
Z. Sun et al.

A R T I C L E
MALAT1 was primarily localized in the cytoplasm of colon cancer cells, as estimated by cell fractionation assays (figure S1b). The results for MALAT1 depletion efficiency in vivo and in vitro are shown in figure S4c-d. Moreover, MALAT1 knockdown suppressed invasion and migration in HCT116 cells (figure S1c, d). These results indicated that MALAT1 played an oncogenic role in CRC and might be positively regulated by YAP1.

YAP1 regulates colon cancer cell proliferation and metastasis through MALAT1
Since MALAT1 expression is regulated by YAP1 interaction with TCF4/β-catenin, whether YAP1 completely or partially acts via MALAT1 is yet to be determined. As shown in Fig. 3, YAP1 overexpression promoted cell migration in HCT116 cells, whereas MALAT1 knockdown reversed the effects of YAP1 overexpression partially, in wound-healing (Fig. 3a) and Transwell migration (Fig. 3b) assays. Next, we established a subcutaneous xenograft model by injection with HCT116 cells stably transfected with control, YAP1, RNAi control, or YAP1 + RNAi-MALAT1. The results showed that YAP1 overexpression promoted the growth of in vivo subcutaneous colon tumors compared with the control group. In addition, RNAi-MALAT1 partially reversed the effect of YAP1 overexpression in promoting in vivo tumor growth (Fig. 3c, d). Moreover, YAP1 overexpression increased in vivo MALAT1 expression in subcutaneous colon tumors, while RNAi-MALAT1 reduced this increasing trend (Fig. 3e). However, RNAi-MALAT1 did not affect YAP1 expression in vivo (Fig. 3f). Overall, the YAP1-MALAT1 axis regulated biological activities in colon cancer cells.

YAP1 interactions with TCF4 and β-catenin regulate MALAT1 expression in colon cancer cells
To further explore the mechanism by which YAP1 regulates MALAT1, bioinformatics analyses predicted that the MALAT1 promoter had three binding sequences for the transcription factor TCF4 (JASPAR, ifti.org, and UCSC). Interestingly, TCF4 and β-catenin were positively regulated by YAP1 in HCT116 cells as evaluated by qRT-PCR (Fig.  4a) and Western blotting (Fig. 4b). In addition, both TCF4 and β-catenin interacted with the YAP1 protein, as observed by IP (Fig. 4c). Luciferase activity assays indicated that gene knockdown of YAP1, TCF4, and β-catenin inhibited MALAT1 promoter activity in HCT116 and SW480 cells

A R T I C L E
( Fig. 4d), and downregulated MALAT1 in HCT116 cells (Fig. 4e). Further bioinformatics analysis (JASPAR) showed that TCF4 might have a DNA binding motif (Fig.   4f). Next, we analyzed the ChIP-seq data of TCF4 in HCT116 cells from ENCODE. The results showed high signals of TCF4 (TCF7L2) in the MALAT1 promoter YAP1-induced MALAT1 promotes epithelial-mesenchymal transition and angiogenesis by. . . 2631

A R T I C L E
region, which suggested a binding potential between TCF4 and the MALAT1 promoter ( Figure S3). The interactions among these three genes prompted us to identify the binding site of the MALAT promoter affected. Thus, we constructed luciferase reporters for different predicted binding-sites and observed that the suppression of YAP1, TCF4, or β-catenin reduced the activities of Luc-MAL-pro and Luc-MAL-pro-#1, both containing the first WT binding site away from the start site of MALAT1 transcription. Moreover, Luc-MALpro-#2 and Luc-MAL-pro-#3 did not contain the second and third WT binding sites, respectively ( Fig. 4g-i). Furthermore, YAP1, TCF4, or β-catenin knockdown affected the activity of MALAT1 promoter containing only the first WT-and non-mutant (MT) binding sites (Fig. 4j). These findings suggested that YAP1 and TCF4 were transcription co-activator and transcription factor, respectively. Then, qPCR-ChIP assays were carried out in HCT116 and SW480 cells. The results showed that anti-TCF4, anti-YAP1, and anti-β-catenin pulled-down the predicted binding-site sequence of the MALAT1 promoter in HCT116 and SW480 cells (Fig. 4k). These findings suggested that YAP1 interacted with TCF4 and β-catenin, and formed a compound regulating MALAT1 in CRC.
To validate the direct binding between MALAT1 and miR-126-5p at the physiological levels, an RNA pull-down assay was performed, with a biotin-labeled-specific MALAT1 probe (Fig. 5h). In this experiment, miR-125a-5p was used as a negative control because it has no binding sequence for MALAT1. An inverse pull-down assay evaluated whether miR-126-5p binds MALAT1. HCT116 or SW480 cells were transfected with biotinylated miR-126-5p, and harvested for pull-down assay. MALAT1 was pulled down; however, several site-specific mutations that disrupt the base pairing between MALAT1 and miR-126-5p prevented miR-126-5p from binding to MALAT1 (Fig. 5i). It is known that miRNAs bind their target genes and cause post-transcriptional suppression in an AGO2-dependent manner. To assess whether MALAT1 was regulated by miR-126-5p in this manner, we performed anti-AGO2 RIP in HCT116 cells transiently transfected with miR-126-5p. Endogenous MALAT1 was specifically enriched in miR-126-5p-overexpressing cells (Fig. 5j), indicating that it was recruited to AGO2-related RNA-induced silencing complexes, and functionally interacted with miR-126-5p. Next, we established a luciferase construct containing MALAT1 or a mutated form [MALAT1 MT (miR-126-5p)]. Luciferase assay showed that miR-126-5p overexpression significantly inhibited the luciferase activity of MALAT1, with reduced effects on the mutated form of MALAT1 (Fig. 5k).

MALAT1-miR-126-5p regulates VEGFA, SLUG, and TWIST in colon cancer cells
Since miR-126-5p was confirmed to be sponged by MALAT1, we next assessed its downstream cascade. MicroRNAs regulate biological functions primarily by inhibiting the expression of target genes. As shown above, miR-126-5p was downregulated and exerted a tumor suppressor function in CRC. Therefore, bioinformatics tools The effect of the YAP1-MALAT1 axis on migration was assessed in HCT116 cells after transfection with pEGFP-YAP1, pEGFP-YAP1 + MALAT1, or vector, by Transwell assay. c, d In vivo subcutaneous tumor xenografts stably transfected HCT116 cells in nude mice after injection of control, YAP1, RNAi control, or YAP1 + RNAi-MALAT1. After 28 days, the nude mice were sacrificed and the subcutaneous tumors were assessed macroscopically. Tumor weight reflected the proliferation ability of HCT116 cells in vivo. Four representative mice from each group are shown. e MALAT1 expression in various groups with in vivo subcutaneous tumors, as assessed by qRT-PCR. f Immunohistochemical detection of YAP1 protein levels in subcutaneous tumors from nude mice. Three independent experiments were performed. *p < 0.05, **p < 0.01 compared to the control group, by Student's t-test Z. Sun et al.

R E T R A C T E D A R T I C L E
(MIRDB, microRNA.org, and TargetScan.org) were used to predict that miR-126-5p potentially interacted with the 3′ UTR regions of SLUG, TWIST, and VEGFA mRNAs through complementary sequences (Fig. 6a). The mRNA , and βcatenin expression levels in HCT116 cells transfected with pEGFP-c1-YAP1, siYAP1, vector or negative control by qRT-PCR. b TCF4 and β-catenin expression levels in HCT116 cells treated with pEGFP-c1-YAP1, siYAP1, vector or negative control by Western blotting. c IP confirmed YAP1 interactions with TCF4 and β-catenin in HCT116 cells (GAPDH, IgG as a negative control). d Promoter activities in HCT116 and SW480 cells co-transfected with luciferase reporters for MALAT1 and silencing plasmids (siYAP1, siTCF4, siβ-catenin or negative control, as indicated) for 48 h. e The expression levels of MALAT1 in HCT116 cells treated with silencing plasmids (siYAP1, siTCF4, siβ-catenin or negative control, as indicated) for 48 h. f The TCF4 motif predicted from JASPAR matrix models. g-j Promoter activities from HCT116 cells co-transfected with luciferase reporters of differently predicted binding-site plasmids (Luc-MAL-pro, Luc-MAL-pro-#1, Luc-MAL-pro-#2, or Luc-MAL-pro-#3) and treated with the silencing plasmids siYAP1, siTCF4, siβ-catenin, or negative control, were measured by dual luciferase reporter assay. Luc-MALpro, containing three WT binding sites; Luc-MAL-pro-#1, Luc-MALpro-#2, and Luc-MAL-pro-#3, containing the first, second, and third WT binding sites, respectively, away from the transcription start site of MALAT1; E4F1 was assessed as the negative control. k ChIP assays using anti-TCF4, anti-YAP1, and anti-β-catenin antibodies were performed in HCT116 and SW480 cells. A control IgG was used as the negative control for IP. Semi-qPCR was used to assess ChIP signals. BRP binding region primer, URP unrelated region primer. Three independent experiments were performed. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group, by Student's t-test YAP1-induced MALAT1 promotes epithelial-mesenchymal transition and angiogenesis by. . . 2633

R E T R A C T E D A R T I C L E
expression levels of SLUG, TWIST, and VEGFA, detected by qRT-PCR, were increased ( Figure S2a) and negatively correlated with miR-126-5p expression in 83 cases of CRC ( Fig. 6b-d). VEGFA, SLUG, and TWIST expression levels were positively associated with MALAT1 expression in the 83 above cases of CRC, GEO data (GSE10495), and GEPIA database, respectively ( Fig. 6e-g, S2b, d). In addition, YAP1 expression showed positive associations with VEGFA, SLUG, and TWIST expression levels in the GEO database (GSE10495) and GEPIA database ( Figure S2c, d).
MiR-126-5p overexpression or MALAT1 knockdown could simultaneously reduce the expression levels of SLUG, 2634 Z. Sun et al.

A R T I C L E
TWIST, and VEGFA in colon cancer cells, as assessed by qRT-PCR and Western blotting (Fig. 6h, i). Hence, MALAT1 could regulate SLUG, TWIST, and VEGFA by sponging miR-126-5p in order to exert its biological function in CRC. To further evaluate the ability of miR-126-5p in targeting SLUG, TWIST, and VEGFA, WT and MT plasmids of psiCHECK™-2-SLUG 3′UTR-Luc, psi-CHECK™-2-VEGFA XII 3′UTR-Luc, and psiCHECK™-2-TWIST 3′UTR-Luc were constructed. As shown in Fig.  6j, miR-126-5p directly targeted SLUG, TWIST, and VEGFA in a dual luciferase reporter assay system, thereby exhibiting a tumor suppressor role in CRC.

Discussion
At present, CRC is one of the most common malignant cancers [1,2]. Developing an effective treatment for CRC patients is an indispensable prerequisite. In this study, we found that YAP1 inhibition downregulated MALAT1, which exhibited tumor-promoting effects in colon cancer cells via interaction with TCF4 and β-catenin. Further assessment showed that MALAT1 promoted the biological characteristics of colon cancer cells by sponging miR-126-5p for upregulating its targets VEGFA, SLUG, and TWIST. Meanwhile, YAP1-induced MALAT1 induced EMT and angiogenesis in colon cancer cells. a Predicted sequences of miR-126-5p-binding sites within MALAT1, and MALAT1 MT and miR-126-5p MT sequences at the top and bottom, respectively, were used. b, c miR-126-5p was lowly expressed in CRC samples as assessed by the GSE83924 and GSE18392 datasets. d Analysis of miR-126-5p expression levels in colorectal tissues as detected by qRT-PCR. e Analysis of the expression patterns of miR-126-5p in different TNM stages of CRC by qRT-PCR. f Correlation analysis of miR-126-5p and MALAT1 expression levels in CRC tissues by Spearman's rank correlation coefficient. g Overall survival analysis was carried out to assess the prognostic effect of miR-126-5p in patients with CRC by the Kaplan-Meier method. h HCT116/SW480 cell lysates were incubated with biotinlabeled MALAT1, and miR-126-5p expression levels were determined by qPCR after pull-down. i HCT116/SW480 cell lysates were incubated with biotin-labeled miR-126-5p, and MALAT1 expression was assayed by qPCR after pull-down. j AGO2-RIP was followed by MALAT1 qPCR to assess MALAT1 levels after miR-126-5p overexpression. k Luciferase activity in HCT116 cells co-transfected with miR-126-5p and luciferase reporters containing MALAT1 or mutant transcript. Data were presented as the relative ratio of firefly luciferase activity to Renilla luciferase activity. Three independent experiments were performed. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the control group, by Student's t-test YAP1-induced MALAT1 promotes epithelial-mesenchymal transition and angiogenesis by. . . 2635

R E T R A C T E D A R T I C L E
YAP1 is a potent oncogene upregulated in most cancers. Our previous meta-analysis of YAP1 revealed that both overall and nuclear overexpression levels of YAP1 were intimately associated with adverse OS and disease-free survival (DFS) in several cancers, including CRC [26][27][28]. Currently, YAP1 was proved to promote cell proliferation, invasion, and migration in colorectal cancer [29][30][31]. To explore the underlying regulation of YAP1, we previously 2636 Z. Sun et al.

R E T R A C T E D A R T I C L E
knock down the YAP1 expression for the following highthroughput sequencing. Interestingly, we not only found the dysregulated mRNA, but also the most lncRNA as we showed, especially MALAT1. It has been reported that overexpressed MALAT1 was involved in regulating tumor progression [32,33]. The lncRNA MALAT1 shows a positive effect on cancer tumorigenicity and metastasis by regulating the vasculogenesis-like process and angiogenesis [34]. The loss of MALAT1 could lead to the reduction of metastasis in mammary tumors [35]. There are some studies supporting the current findings, which implied that MALAT1 could be as an oncogene in CRC [36,37]. MALAT1 was also involved in promoting tumor growth and metastasis in CRC. And these researchers found that MALAT1 could be as a liberator to release oncogene PTBP2 from SFPQ/PTBP2 complex [38]. However, the promotion of YAP1 in CRC by regulating MALAT1 still remains elusive. In our study, we reported that high MALAT1 expression was identified in colorectal cancer tissues and showed an oncogene role of MALAT1 in colon cancer cell. Based on these results, we supposed that YAP1 may positively regulate the MALAT1 to promote the CRC progression.
It is known to us, YAP1 is a transcriptional coactivator. A study reported that β-catenin/YAP/TBX5 transcription complex could enhance the expression of RNA component of mitochondrial RNA processing endoribonuclease (RMRP), which has the function in mitochondrial and ribosomal RNA processing, especially noncoding RNA in colon cancer cells [39]. Previous evidence suggests that lncRNA H19 is aberrantly expressed in osteosarcoma and Hh signaling/Yap1 overexpression could induce H19 expression by gene knockdown [40]. However, the regulation mechanism of YAP1 by control lncRNA is still confused. In this study, we used dual luciferase and ChIP experiment to explore a direct binding between MALAT1 promoter and YAP1 interacting with TCF4 and β-catenin, which provides forceful evidence that YAP1 regulating MALAT1 plays an important role in CRC metastasis. However, the mechanism of regulators targeting MALAT1 in cancer is still to be explored.
Increasing studies showed a ceRNA mechanism regulated by MALAT1 in several malignancies. MALAT1 promotes the development of osteosarcoma by regulating HMGB1 via miR-142-3p and miR-129-5p [41]. Downregulation of miR-146b-5p by MALAT1 in HCC promotes cancer growth and metastasis [42]. The lncRNA MALAT1 interacts with miR-204 to modulate proliferation, migration, and invasion in human hilar cholangiocarcinoma by targeting CXCR4 [43]. A handful of studies have found the function of miR-126-5p in CRC [44,45]. It showed that low miR-126-5p is related with metastatic colorectal cancer treated with bevacizumab [46]. Furthermore, miR-126 downregulation promotes cell proliferation, migration, and invasion in CRC and induces cell cycle arrest by targeting IRS-1 [47]. Upstream, few studies showed that miR-126 could be regulated by lncRNA to exert its tumor suppressor role. In lung cancer, miR-126 was sponged by lncRNAPVT1-5, which promoted the cell proliferation [48]. Similarly, miR-126 was inhibited by the lncRNA, HOTAIR to enhance the cisplatin resistance in gastric cancer. In this study, we found that MALAT1 had two miRNA response elements (MREs) of miR-126-5p by bioinformatics analysis. Dual luciferase, RNA-IP, and biotin-RNA pull-down assays showed that MALAT1 potentially sponged miR-126-5p. β-catenin/TCF4 complex in colorectal cancer cells could decrease several of these miRNAs (miR-145, miR-126, miR-30e-3p, and miR-139-5p), which markedly suppressed CRC cell growth in vitro. In our study, we proved that YAP1 bound with β-catenin and TCF4 had a positive effect on MALAT1 promoter. High MALAT1 sponged miR-126-5p to enhance the CRC metastasis. These findings strongly suggested that the presence and onco-role of YAP1-induced MALAT1-miR-126-5p axis in CRC. As we all know, miRNAs suppress the expression of target messenger RNAs (mRNAs) in post-transcriptional gene regulatory pathways and initiate either translational suppression or cleavage [49]. Then, the current findings have proved that miR-126-5p could target three mRNAs, including VEGFA, SLUG, and TWIST in CRC. As shown previously, VEGFA promotes angiogenesis [50], while SLUG and TWIST significantly accelerate the EMT [51].
Take together, we found that the oncogene, YAP1 forming a complex with β-catenin/TCF4 bound to the MALAT1 promoter, which could sponge miR-126-5p to promote VEGFA, SLUG, and TWIST expression. Our findings suggested that YAP1-induced MALAT1-miR-126-5p axis may regulate angiogenesis and the EMT to promote metastasis in CRC (Fig. 8) Fig. 6 MALAT1-miR-126-5p regulates VEGFA, SLUG, and TWIST expression in colon cancer cells. a Predicted sequences of miR-126-5p binding sites within the 3′-UTRs of VEGFA, SLUG, and TWIST, and sequences of VEGFA, SLUG, and TWIST 3′-UTR mutants (Mut) used in this study (top). b-d Correlation analysis of miR-126-5p and predicted target gene (VEGFA, SLUG, and TWIST) expression levels in CRC tissues, assessed by Spearman's rank correlation. e-g Correlation analysis of MALAT1 and expression levels of predicted target gene (VEGFA, SLUG, and TWIST) in CRC tissues, assessed by Spearman's rank correlation. h, i Analysis of miR-126-5p and MALAT1 regulatory effects on VEGFA, SLUG, and TWIST in HCT116 cells by qRT-PCR and immunoblot. j Luciferase activity in HCT116 cells co-transfected with miR-126-5p mimics and luciferase reporters containing the indicated 3′-UTR-driven reporter constructs or mutant transcripts. Data were presented as the relative ratio of firefly luciferase activity to Renilla luciferase activity. Three independent experiments were performed. *p < 0.05, **p < 0.01 compared to the control group, by Student's t-test YAP1-induced MALAT1 promotes epithelial-mesenchymal transition and angiogenesis by. . . 2637

Patient tissue specimens and follow-up
Clinical data were obtained from 83 patients who underwent radical resection for colorectal cancer during January 2010-June 2011 at the First Affiliated Hospital of Zhengzhou University (Zhengzhou, China). All enrolled patients (16-80 years old) were diagnosed with colorectal adenocarcinoma with no other tumors, had the nonemergency operation treatment, and were undergoing no chemotherapy before operation. The 83 CRC cases were 59.49 ± 8.36 years old, and included 47 males and 36 females. There were 9 stage I, 21 stage II, 38 stage III, and 15 stage IV cases. They included 21 left-sided colon, 16 right-sided colon, and 46 rectal carcinomas. Tumor diameters were <4 cm (n = 21), between 4 and 6 cm (n = 44), and >6 cm (n = 18). There were 23 ulcerating, 50 timorous, and 10 infiltrating carcinomas, with 15 distant and 68 nondistant metastasis cases. The patients were divided into high (n = 59) and low YAP1 (n = 24) expression groups based on the median of YAP1 expression in CRC tissues. Then, the associations of malignant phenotype and prognosis with YAP1 expression in these two groups were analyzed. The patients underwent radical surgery for CRC and the malignancy was confirmed by histopathology. According to NCCN Guidelines, the patients were classified as stage II disease and above and administered postoperative chemotherapy. Adjacent normal tissue samples were obtained from normal colorectal tissues localized 5 cm distal to the tumors. None of the patients had a history of radiotherapy or chemotherapy before surgery. Follow-up was performed monthly for the first 6 months, then quarterly until 2 years, every 6 months until 5 years, and once a year thereafter. For follow-up, telephonic interviews and medical examination were performed for 5 years after discharge post-systematic treatment, until June 2016. Survival was calculated by the month. OS was defined as the time from tumor excision to death by any cause. The protocols were approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans; all patients provided written informed consent.

In situ hybridization and immunohistochemistry
In situ hybridization was performed to detect MALAT1 expression in tissue specimens with three 30-base nucleotide probes from different MALAT1 regions. The MALAT1 probes were: 5′-TTGAACTCGCTTTCCATGGCGATTT GCCTTGTGAG-3′, 5′-CTTTAATCTTCCTTCAAAAGG TGGTAAACTATACC-3′, and 5′-TTAAACCGAAGGTG ATTAAAAGACCTTGAAATCCA-3′, and synthesized and labeled with DIG-dUTP at the 3′ end by Boster (Wuhan, China; No. 01533TW424056050DL). In situ hybridization was performed as described previously [52]. The semi-quantitative scoring criterion for in situ hybridization was used, wherein both the staining intensity and the number of positive cells were recorded. Protein expression levels of YAP1 were determined by the Elivision TM twostep immunohistochemical method in normal colorectal and colorectal cancer tissue samples (supplementary material).

Gene microarray analysis
HCT116 cells were transfected with scramble siRNA (siCtrl) or siRNA targeting YAP1 (siYAP1) or MALAT1 (siMALAT1) for 36 h. Total RNA was extracted and assessed with Affymetrix Human Genome U133 Plus 2.0 Array (Thermo, MA, USA). The data were deposited in NCBI's Gene Expression Omnibus (#GSE92335). Gene set enrichment analysis (GSEA) was used to evaluate expression differences in gene sets between the two groups.

Western blotting
Cell lysis, electrophoresis, and target protein visualization were performed as described previously [60]. See supplementary material for details.

Quantitative RT-PCR (qRT-PCR) and luciferase activity assay
Total RNA was isolated with TRIzol reagent (Thermo, MA, USA) as described previously [61]. The primer sequences are shown in Table S3. Luciferase activity assays were performed with Dual-Luciferase Reporter Assay System (Promega, Beijing, China). Three promoter sequences of MALAT1 are shown in Table S5. See supplementary material for details.

Cell proliferation, invasion, and metastasis assays
Cell proliferation was assayed with CCK-8 (Dojin Laboratories, Tokyo, Japan) and Click-iT® EdU Microplate Assay Kit (Ribobio, Guangzhou, China), according to the manufacturer's instructions. Wounds were generated with 10 μL pipette tips. Matrigel invasion assay was performed as described previously [62]. See supplementary material for details.

Subcellular fractionation
To determine the cellular localization of MALAT1 in colon cells, cytoplasmic and nuclear fractions were collected according to the manufacturer's instructions using PARIS Kit (Life Technologies, MA, USA). Data analysis was performed as described previously [63].

Immunoprecipitation (IP) and chromatin immunoprecipitation (ChIP)
Cell lysates were incubated with 2 μg antibody for 16 h at 4°C . Then, protein A/G Sepharose beads were added to the mixture and incubated for 4 h at 4°C. Subsequently, the generated immuno-complexes were washed with lysis buffer four times and separated by SDS-PAGE.

RNA-binding protein immunoprecipitation (RIP)
RIP assays were performed with the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, MA, USA) following the manufacturer's protocol. HCT116 cells were transfected with miR-126-5p control or miR-125a-5p mimic for 48 h, followed by lysis with the RIP lysis buffer. Then, 100 μL of cell extract was incubated with the RIP buffer containing magnetic beads conjugated with human anti-Ago2 antibodies (Abcam, Cambridge, UK) or the negative control (Millipore, MA, USA). After antibody recovery by the protein A/G beads, qRT-PCR was performed to assess the expression levels of MALAT1 in the precipitates. lgG was used to normalize RNA-IPs. The RIP-qPCR Data Analysis (ΔΔCt method) was performed as described previously [65,66]. Normalize each RIP RNA fractions' Ct value to the input RNA fraction Ct value for the same qPCR Assay (ΔCt) to account for RNA sample preparation differences.

RNA pull-down assay
RNA pull-down was performed as described previously [67]. SW480 and HCT116 cells were transfected with biotinylated miRNA (40 nmol/L) (Ribobio, Guangzhou, China). After 48 h, the cells were harvested, and lysates were incubated with M-280 streptavidin magnetic beads (Sigma, Shanghai, China). The bound RNAs were purified with TRIzol reagent, and MALAT1 mRNA levels were evaluated by qPCR.
The biotinylated DNA probe complementary to MALAT1 was synthesized (GenePharma, Shanghai, China). SW480 and HCT116 cells were overexpressed with MALAT1, MALAT1-mut (miR-126-5p), or LINC00152 containing plasmids. Then, cell lysates were prepared after 48 h and incubated with the biotinylated probe and M-280 streptavidin magnetic beads (Sigma, Shanghai, China). The miR-126-5p levels were analyzed by qPCR; miR-125a-5p was used as a negative control since it has no binding sequence for MALAT1.

Tube formation assay
Tube formation assay was performed as described previously [68]. In these assays, 96-well plates were coated with 50 μL Matrigel. Then, 2 × 10 4 HUVEC were coincubated with 200 μL of supernatants from each group for 10 min, followed by incubation at 37°C for 6-8 h. Images were acquired under a microscope (Nikon, Tokyo, Japan). Antiangiogenic activity was quantified by measuring the length of the tube walls formed between discrete endothelial cells in each well relative to the control.

A R T I C L E
agomir, was used as long-acting siRNA because it embeds specific chemical modifications to protect the siRNAs against the immune system in vivo.

Statistical analysis
All statistical analyses were carried out with SPSS version 18.0 (MT, USA) and the GraphPad Prism 5.0 software (CA, USA). Data are mean ± SEM. Group pairs were compared by Student's t-test. The Pearson's coefficient was used to assess the correlations between variables. Survival data were obtained by the Kaplan-Meier method, with significance assessed by the log-rank test. The associations of YAP1 expression and clinicopathologic variables were assessed by the Chi-square test or Fisher's exact test. Median YAP1, MALAT1, and miR-126-5P expression levels were cutoff points for determining high and low expression. p < 0.05 was considered statistically significant.