lncRNA THAP9-AS1 Promotes Pancreatic Ductal Adenocarcinoma Growth and Leads to a Poor Clinical Outcome via Sponging miR-484 and Interacting with YAP

Pancreatic ductal adenocarcinoma (PDAC) is still a highly aggressive cancer with a 5-year overall survival rate of only 3%–5%, which is, at least in part, due to its poor response to existing conventional or targeted treatments (1). Thus, novel and effective therapies are urgently required. Because of the endogenous genetic alterations including the mutations and/or amplification of KRAS oncogene and the mutations or losses of CDKN2A, TP53, and SMAD4 cancer suppressor genes, PDAC evolves through a series of histopathologic changes, referred to as increasingly dysplastic precursor lesions, or pancreatic intraepithelial neoplasias (PanIN), toward invasive and finally metastatic pancreatic cancer (2–6). The essential roles of these genetic aberrations in PDAC initiation and progression have been well characterized by several genetically engineered mouse models (7). In addition, the critical role of dysregulation of epigenetic modifiers, for example, noncoding RNAs in the development and progression of many human cancer types including PDAC is also increasingly emphasized.

Emerging evidence has regarded long noncoding RNAs (lncRNA) as major regulators of both development and disease. LncRNAs are an important set of endogenous transcripts that are longer than 200 nucleotides but do not encode proteins (8). LncRNAs function in a wide range of biological activities, including cell-cycle regulation, stem cell pluripotency, lineage differentiation, and cancer progression, by regulating gene expression by various mechanisms (9). Studies point out that lncRNAs play vital roles in mRNA transcription and translation, protein abundance and location, as well as chromatin and protein conformation (10–12). Recent evidence has shown abnormal lncRNA expression in a variety of cancers, and lncRNAs have been reported to play roles of oncogene, tumor suppressor or both, according to their multifaceted functions in a wide range of biological processes, such as proliferation, apoptosis, cell migration, and metastasis (13–15). Recently, although several lncRNAs have been shown to be involved in growth, invasion, metastasis, and stemness of PDAC (13, 16, 17), their specific roles and detailed mechanisms in pancreatic carcinogenesis still remain incompletely revealed.

The annotated potential lncRNA THAP9-AS1 (THAP9 antisense RNA 1) was previously reported to be lowly expressed in multiple human normal tissues, especially in pancreas (18). Via a bioinformatic analysis of publicly available data, we found that THAP9-AS1 is overexpressed in PDAC. In this study, we validated and characterized the upregulation of THAP9-AS1 in PDAC and corroborated the hypothesis that THAP9-AS1 is a key lncRNA involved in progression of PDAC.

This study was reviewed and approved by the Ethnics Committees of Guangzhou Medical University and Affiliated Cancer Hospital (Guangzhou, Guangdong, China) and Wayne State University (Detroit, MI). The “informed written consent” was obtained from each subject or each subject's guardian. All related procedures were performed with the approval of the internal review and ethics boards of the hospital. The study was conducted in accordance with the Declaration of Helsinki.

Pancreatic ductal adenocarcinoma cell lines PANC-1, SW1990, CFPAC-1, and BxPc-3, and HEK293T cells were obtained from ATCC. Cells were cultured in RPMI1640 or DMEM (Gibco) supplemented with 10% FBS. To establish stable transfectants with knockdown or overexpression, cell lines were transfected with psi-LVRU6GP vectors with THAP9-AS1 shRNAs (target sequence for sh-1#: 5′-AATGGGGAAGCTCTTGGCATG-3′, sh-2#: 5′- GTCTACAACTTCACTCTTTGC-3′, sh-3#: 5′- GCACTTTGGGTGGCTAAAGCA-3′), or with YAP shRNAs (target sequence for sh-1#: 5′- GGAAGCTGCCCGACTCCTTCT-3′, sh-2#: 5′- GCAGGTTGGGAGATGGCAAAG-3′), or with gene overexpression vectors EXP-LV203-THAP9-AS1 or pEZ-Lv203-YAP using Lipofectamine 3000 following the manufacturer's instructions (Invitrogen). To knockdown TEAD1, the shRNAs (target sequence for sh-1#: 5′-GGTTCTTGCCAGAAGGAAA-3′, sh-2#: 5′-GGATCAGACTGCAAAGGAT-3′, sh-3#: 5′-GCTTGAATCAGTGGACATT-3′, sh-4#: 5′-GCCGATTTGTATACCGAAT-3′) was used for transfection. To increase miR-484 level and inhibit miR-484 function, related cells were transfected with miR-484 mimics and miR-484 inhibitor (miRCURY LNA miRNA inhibitor for miR-484), respectively (Exiqon). The frozen-fresh and paraffin-embedded PDAC and noncancerous tissues were collected from patients at the Affiliated Cancer Hospital of Guangzhou Medical University (Guangzhou, Guangdong, China). The “informed written consent” was obtained from each subject or each subject's guardian. All related procedures were performed with the approval of the internal review and ethics boards of the hospital.

RNA immunoprecipitation (RIP) assay was performed to validate the interaction of THAP9-AS1 with miR-484, and of YAP mRNA 3′UTR with miR-484; HEK293T cells were cotransfected with MS2bs vectors cloned with related DNA sequences (MS2bs, MS2bs-THAP9-AS1-WT, MS2bs-THAP9-AS1-Mut, MS2bs-YAP-3′UTR-WT, or MS2bs-YAP-3′UTR-Mut) and MS2bp-GFP overexpression vector (Addgene). To validate the interaction between THAP9-AS1 and YAP protein, HEK293T cells were cotransfected with pReceiver-M11 vector with N-terminal Flag-tag–expressing YAP truncated fragments and EXP-LV203 vector-expressing THAP9-AS1, or EXP-LV203 vector-expressing THAP9-AS1 truncated fragments and pEZ-Lv203-YAP. At 48 hours; related cells were used to perform RIP assay according to the manufacturer's instructions of the EZ-Magna RIP Kit (Millipore), using anti-GFP, anti-Flag or anti-YAP antibodies, or normal rabbit IgG. After extraction of RNAs, miR-484 and THAP9-AS1 truncated fragments were examined by qRT-PCR.

The animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of Guangzhou Medical University (Guangzhou, Guangdong, China). Standard animal care and laboratory guidelines were followed according to the IACUC protocol. BxPc-3/sh-Con, BxPc-3/sh-THAP9-AS1, BxPc-3/sh-THAP9-AS1+Vector, and BxPc-3/sh-THAP9-AS1+YAP cell lines were injected at 1 × 10⁶ cells in 200 μL, respectively, subcutaneously into the armpit of female Balb/C athymic nude mice to generate xenograft tumors (6 mice/group). The tumor growth was measured every 5 days. The wet weight of tumor was recorded after excision at the experimental endpoint.

qRT-PCR, plasmid constructs, proliferation assay, tumor sphere assay, Western blot, immunofluorescence, transactivation of transcription factors, luciferase reporter assay, ChIP assay, co-immunoprecipitation, in situ hybridization, IHC, and primers. These methods used are described in the Supplementary Experimental Procedures.

All data were presented as the mean ± SD. The Student t test and the χ² test were used to compare the differences among different groups. Survival curves were plotted using the Kaplan–Meier method and compared using the log-rank test. Statistical analyses were performed using GraphPad Prism 6. P < 0.05 was considered statistically significant.

To explore lncRNAs involved in pancreatic ductal adenocarcinoma (PDAC), a publicly accessible microarray dataset from patients with PDAC was analyzed between PDAC tissues and normal pancreatic tissues (GEO: GSE16515). Among the annotated potential lncRNAs, THAP9-AS1 levels were found to be increased in PDAC tissues compared with those in normal pancreatic tissues (Fig. 1A). Next, online analysis based on TCGA and GTeX data (http://gepia.cancer-pku.cn/) indicates that there are also significantly higher THAP9-AS1 levels in PDAC tissues compared with normal pancreatic tissues (Fig. 1B). Furthermore, the THAP9-AS1 levels were determined in 11 pairs of PDAC tissues and matched noncancer tissues collected from patients at the Affiliated Cancer Hospital of Guangzhou Medical University (Guangzhou, Guangdong, China). The THAP9-AS1 levels were also significantly increased in primary PDAC tissues as compared with their matched adjacent normal tissue (Fig. 1C). Moreover, THAP9-AS1 levels in PDAC cell lines were also increased compared with that in the adjacent normal pancreatic tissues (Fig. 1D).

Analysis shows that THAP9-AS1 is composed of five exons, with a transcript of poly (A)-negative (Ensembl: ENSG00000251022). THAP9-AS1 is located in both cytoplasm and nucleus (Fig. 1E). The protein coding potentiality of THAP9-AS1 was analyzed by online software, that is, PhyloCSF and CPAT. The LNCipedia database based on both softwares failed to predict its coding potentiality (Supplementary Fig. S1A). Although a Bazzini small ORF is predicted, there is no translation initiation site in the THAP9-AS1 transcript (Supplementary Fig. S1A). In addition, RegRNA 2.0 online software also predicted two short ORFs, but there is no ribosome-binding site in the THAP9-AS1 transcript (Supplementary Fig. S1B), supporting that no protein-coding potential of THAP9-AS1 is discovered. To further experimentally verify the protein-coding ability of THAP9-AS1 or the activation of ORFs, a series of constructs was generated in which a GFPmut ORF (in which the start codon ATGGTG is mutated to ATTGTT) was fused to the full-length THAP9-AS1 transcript or predicted ORFs (Fig. 1F). No substantial expression of GFP was observed in HEK293T cells transfected with full-length GFPmut or ORFs-GFPmut construct (Fig. 1G). Western blot analysis using anti-GFP antibody further confirmed that THAP9-AS1 is of no protein-coding ability (Fig. 1H). These data show that expression of THAP9-AS1, as a lncRNA, is upregulated in PDAC cells.

Given that THAP9-AS1 is overexpressed in PDAC, to determine the roles of THAP9-AS1 in PDAC cells, loss- and gain-of-function approaches were employed. First, THAP9-AS1 was knocked down using shRNAs. We found that sh-1# and sh-3# could effectively knockdown THAP9-AS1 levels in CFPAC-1 and BxPc-3 cells after transient transfection (Supplementary Fig. S2A). Then, the stable THAP9-AS1-knockdown CFPAC-1 and BxPc-3 cell lines expressing sh-1# or sh-3# shRNAs were established (Fig. 2A). Their proliferation ability was determined by MTS assay. As shown, THAP9-AS1 knockdown decreased the proliferation of PDAC cells (Fig. 2B). The colony-forming assay also showed that THAP9-AS1 knockdown significantly inhibited colony-forming capacity of PDAC cells (Fig. 2C; Supplementary Fig. S2B). In addition, to confirm the role of THAP9-AS1 in PDAC cells, THAP9-AS1 was overexpressed in PANC-1 cells that exhibit relatively low level of endogenous THAP9-AS1 expression (Fig. 2D). THAP9-AS1 overexpression markedly promoted cell proliferation and colony-forming capacity (Fig. 2D; Supplementary Fig. S2C). Moreover, THAP9-AS1 knockdown inhibited tumor sphere formation of PDAC cells represented in sphere number and sphere diameter (Fig. 2E; Supplementary Fig. S2D), whereas, ectopic overexpression of THAP9-AS1 enhanced tumor sphere–forming ability (Fig. 2F; Supplementary Fig. S2E). These data show that lncRNA THAP9-AS1 drives PDAC cells proliferation and growth.

Given the upregulation and oncogenic role of THAP9-AS1 in PDAC cells, to identify the downstream signaling pathway of THAP9-AS1, we examined the transactivation of a series of transcription factors after THAP9-AS1 overexpression in PANC-1 cells, but there was no significant change observed (Supplementary Fig. S3A). Unexpectedly, we found that the expression of YAP downstream target genes CTGF and CYR61 was significantly increased after THAP9-AS1 overexpression (Supplementary Fig. S3B), whereas THAP9-AS1 knockdown decreased the expression of CTGF and CYR61 in PDAC cells (Supplementary Fig. S3C), suggesting that THAP9-AS1 level was potentially correlated with YAP activation.

To access whether THAP9-AS1 regulates YAP activity, we examined the protein levels of the genes regulating YAP activity. As shown, THAP9-AS1 knockdown decreased YAP protein level, but did not influence the protein levels of YAP upstream kinases LATS1 and MST1 (Fig. 3A). YAP mRNA level was also decreased after THAP9-AS1 knockdown (Fig. 3B). Phosphorylation at YAP serine 127 by LATS1 is one of the most common posttranslational modifications resulting in cytoplasmic localization of YAP to repress its activity (19). THAP9-AS1 knockdown increased YAP serine 127 phosphorylation (Fig. 3A). Immunofluorescence staining revealed that THAP9-AS1 knockdown markedly alleviated protein level and nuclear translocation of YAP (Fig. 3D; Supplementary Fig. S3D). On the other hand, ectopic overexpression of THAP9-AS1 positively regulated YAP expression at both protein (Fig. 3A) and mRNA levels (Fig. 3C), but negatively regulated YAP serine 127 phosphorylation (Fig. 3A). Immunofluorescence staining further confirmed that THAP9-AS1 overexpression enhanced YAP protein level and nuclear translocation (Fig. 3E).

To further verify this potential regulation ex vivo, we further examined YAP expression in the aforementioned eleven pairs of fresh-frozen primary PDAC tissues and their matched adjacent nontumoral (NT) pancreatic tissues in which THAP9-AS1 level has been examined. We found that YAP expression was also increased in the primary PDAC tissues as compared with their matched normal tissue (Fig. 3F). THAP9-AS1 level was positively correlated with YAP mRNA level (Supplementary Fig. S3E). Moreover, online analysis based on TCGA and GTeX data indicates that the expression of YAP and YAP target gene CTGF and CYR61 was upregulated in PDAC tissues as compared with the normal pancreatic tissue (Supplementary Fig. S3F). THAP9-AS1 level was positively correlated with YAP, CTGF, and CYR61 levels (Supplementary Fig. S3G).

To access whether YAP acted as the downstream effector to mediate the role of THAP9-AS1 in PDAC, we overexpressed YAP in BxPc-3 cells in which THAP9-AS1 has been stably knocked down (Supplementary Fig. S3H). YAP overexpression significantly overcame the effects of THAP9-AS1 knockdown on proliferation (Fig. 3G; Supplementary Fig. S3I) and tumor sphere formation (Fig. 3H; Supplementary Fig. S3J). Inversely, we knocked down YAP expression in PANC-1 cells with THAP9-AS1 overexpression (Fig. S3H). YAP knockdown abolished the effects of THAP9-AS1 overexpression on proliferation (Fig. 3I; Supplementary Fig. S3K) and tumor sphere formation (Fig. 3J; Supplementary Fig. S3L). These data indicated that THAP9-AS1 exerted its effects in PDAC cells via regulating YAP activity.

Recent evidences demonstrate that lncRNAs are involved in cancer progression by acting as competing endogenous RNAs (ceRNA; ref. 20). That means lncRNAs act as molecular sponges to regulate the levels of mRNAs by competitively binding their same miRNAs targeting mRNAs. Given that knockdown of THAP9-AS1 decreased, whereas induction of THAP9-AS1 increased both YAP mRNA and protein levels, we wonder whether THAP9-AS1 could act as a miRNA sponge to regulate YAP expression. To test the possibility, we first performed the bioinformatic analysis using the online tool DIANA (21) and miR-484 was predicted to target THAP9-AS1. The RNA targeting by miRNAs in animals primarily relies on a “seed region” mapping to positions 2–7 at the miRNA's 5′end. DIANA tools predicted three putative complementary sequences at THAP9-AS1 for miR-484, including one 10-mer site (Fig. 4A), one 7-mer-A1 site, and one 6-mer site (Supplementary Fig. S4A). The RegRNA 2.0 online tool also predicted a 10-mer site at the same position as what DIANA tools predicted, which has suitable RNA secondary structure for miRNA binding (Supplementary Fig. S4B).

The interaction between THAP9-AS1 and miR-484 was further predicted and the minimum free energy at the binding site was calculated by using RNAhybrid (Fig. 4B; ref. 22). Concurrently, the 3′ untranslated region (3′-UTR) of the YAP mRNA contains one 7-mer-m8 site that matches to the miR-484 seed region predicted by online server TargetScan (Fig. 4A; Supplementary Fig. S4C).

We next determined whether THAP9-AS1 potentially interacted with miR-484 and whether miR-484 potentially targeted YAP via the predicted binding motif (Fig. 4A). To validate the direct binding between THAP9-AS1 and miR-484, the THAP9-AS1 with or without deletion mutations in miR-484 targeting site were cloned into MS2b plasmid to transcribe RNA combined with MS2-binding sequences and then cotransfected with the MS2bp-GFP expression plasmid and miR-484 mimics into HEK-293T cells. Subsequently, we performed RIP assays to pull down miRNAs associated with THAP9-AS1 via GFP antibody and demonstrated via qRT-PCR analysis that miR-484 associated with THAP9-AS1 via targeting site, while the nontargeting miR-191–5p used as negative control, was not associated with THAP9-AS1 (Fig. 4C). RIP assays also indicated that miR-484 also associated with YAP 3′-UTR specifically (Fig. 4D). To further confirm the association of THAP9-AS1 with miR-484, based on pMir-reporter plasmid we constructed luciferase reporters containing the 5′-500nt of THAP9-AS1, which contains wild-type or deletion-mutated miR-484 targeting site. These reporters were cotransfected with miR-484 mimics into HEK-293T cells. We found that miR-484 mimics reduced the reporter activity of the construct with wild-type THAP9-AS1 (Fig. 4E).

To determine whether YAP is a bona fide target of miR-484, based on pMir-reporter plasmid we also constructed luciferase reporters containing the 500nt of the 3′-UTR of YAP mRNA, which contains wild-type or mutated miR-484 targeting site. These reporters were cotransfected with miR-484 mimics into HEK-293T cells. We found that miR-484 mimics reduced the reporter activity of the construct with wild-type 3′-UTR of YAP mRNA (Fig. 4F), suggesting that 3′-UTR of YAP mRNA was inhibited by miR-484. We further performed function gain- and loss-experiments and found that treatment by miR-484 mimics resulted into decrease of endogenous YAP mRNA and protein levels in BxPc-3 and CFPAC-1 cells, whereas miR-484 inhibitor showed the contrary effect in PANC-1 cells (Fig. 4G; Supplementary Fig. S4D).

Given that miR-484 efficiently targets both THAP9-AS1 and YAP, we wonder whether THAP9-AS1 decoys the miR-484 from YAP to enhance YAP upregulation. We detected YAP expression change upon simultaneous interference of THAP9-AS1 expression and miR-484 function in PDAC cells. We found that function loss of miR-484 by its inhibitor greatly overcame THAP9-AS1 knockdown–induced decrease of YAP mRNA and protein levels in BxPc-3 cells (Fig. 4H; Supplementary Fig. S4E). On the contrary, induction of miR-484 by mimics markedly reversed THAP9-AS1 overexpression–mediated increase of YAP mRNA and protein levels in PANC-1 cells (Fig. 4H; Supplementary Fig. S4E). Taken together, these data indicated that THAP9-AS1 acts as a miR-484 sponge and attenuates the inhibitory effect of miR-484 on YAP, thereby resulting into increase of YAP expression in PDAC cells.

Given that THAP9-AS1 not only increased YAP expression but also its activity, and that lncRNAs can exert their effects through RNA–protein interactions, we asked whether THAP9-AS1 also modulated YAP activity beyond expression regulation. Here, we used inhibitor of miR-484 to eliminate the expression interference on YAP protein level and to validate the role of THAP9-AS1 on YAP activity. There were comparable protein levels of LAST1 in selected PDAC cell lines (Fig. 3A). Function inhibition of miR-484 with inhibitor endogenously increased levels of total YAP and phosphorylated YAP, which is possibly due to the stable LAST1 level in PANC-1 cells (Fig. 5A), while THAP9-AS1 overexpression reduced YAP phosphorylation, but without influence on total YAP and LAST1 levels (Fig. 5A). In THAP9-AS1 highly expressed BxPc-3 and CFPAC-1 cell lines, to exclude the effect of miR-484 upon THAP9-AS1 knockdown, inhibitor of miR-484 was applied. Transfection with inhibitor of miR-484 showed no significant effect on YAP and YAP phosphorylation, potentially due to the endogenous functional exhaustion of miR-484 by high THAP9-AS1 level (Fig. 5B). However, THAP9-AS1 knockdown increased YAP phosphorylation, but without influence on total YAP and LAST1 levels (Fig. 5B).

To access how THAP9-AS1 regulates YAP activity, we first predicted that THAP9-AS1 would physically interact with YAP protein by online tool RPISeq (RNA–protein interaction prediction) analysis via inputting protein and RNA sequences, respectively (Fig. 5C). Experimentally, we performed a RIP assay by using YAP antibody to pull down mRNAs interacting with YAP protein. RIP result presented that THAP9-AS1 but not GAPDH mRNA could interact with YAP in PDAC cells (Fig. 5D). First, to identify the domain of YAP interacting with THAP9-AS1, we constructed truncated YAP mutants fused with flag tag. In HEK293T cells cotransfected with YAP truncations and THAP9-AS1–overexpressing vectors, RIP assay using flag antibody unraveled that the WW1/2 domain of YAP (residues 155–263) was responsible for its interaction with THAP9-AS1 (Fig. 5E). Second, to predict the region of THAP9-AS1 responsible for interacting with YAP, a series of THAP9-AS1 truncations were analyzed and the probabilities to interact with YAP were predicted (Supplementary Fig. S5A). To validate this hypothesis, a series of THAP9-AS1 truncations were constructed to map its fragment interacting with YAP. The experimental results from RIP assay using YAP antibody solidly confirmed the interaction between YAP and THAP9-AS1 at two sites within the 3′-region in HEK293T cells cotransfected with truncations of THAP-AS1 and YAP-overexpressing vectors (Fig. 5F). Third, the WW1/2 domain of YAP is required to interact with LATS1 to induce YAP phosphorylation and subsequent cytoplasmic retention. Thus, we speculated whether THAP9-AS1 interaction with YAP blocks LATS1–YAP interaction. We performed coimmunoprecipitation (co-IP) assay and showed that THAP9-AS1 knockdown facilitated the interaction between YAP and LATS1 upon the presence of miR-484 inhibitor to maintain the high endogenous YAP protein level in BxPc-3 PDAC cells (Fig. 5G). These results indicated that THAP9-AS1 could interact with YAP to block its interaction with LATS1 and then promote its activity.

Given the high expression and role of THAP9-AS1 in PDAC, and the involvement of regulatory feed-forward loops in tumor (23), we wonder whether THAP9-AS1–regulated YAP would have a feedback regulation for THAP9-AS1 expression. Expectedly, THAP9-AS1 levels were decreased upon YAP knockdown (Fig. 6A), while THAP9-AS1 levels were increased upon ectopic YAP overexpression in PDAC cells (Fig. 6A), indicating YAP might positively regulate THAP9-AS1 transcription. YAP as a transcriptional coactivator without DNA-binding domains usually binds with transcription factors such as TEAD1-4 to modulate target genes expression. We performed bioinformatic analysis (The JASPAR database) to screen potential transcription factors that are located within a 3-kb region upstream of the THAP9-AS1 transcription start site and identified a TEAD1 binding site at the potential promoter (Fig. 6B). Chromatin immunoprecipitation (ChIP) assay indicated the occupancy of YAP on THAP9-AS1 promoter in PDAC cells (Fig. 6C). TEAD1 was consistently expressed in PDAC cells (Fig. 3A). Here, we found that TEAD1 knockdown reduced the enrichment of YAP on THAP9-AS1 promoter (Fig. 6C). Also, treatment with verteporfin, a drug selectively disrupts YAP–TEAD interaction, significantly repressed THAP9-AS1 transcription in BxPc-3 and CFPAC-1 cells (Fig. 6D). Moreover, TEAD1 knockdown disrupted the increase of THAP9-AS1 expression induced by YAP overexpression (Fig. 6E). Verteporfin treatment also normalized the increased expression of THAP9-AS1 induced by YAP overexpression in PANC-1 cells (Fig. 6E). These results indicated that YAP bound to THAP9-AS1 promoter via TEAD1 and promoted THAP9-AS1 transcription via a positive feedback regulatory loop in PDAC cells.

To evaluate the biological and clinical consequences of the function abnormality of THAP9-AS1 in cancer, we explored the role of THAP9-AS1 in tumor growth. We implemented subcutaneous injection of BxPc-3 cells with THAP9-AS1 knockdown and sh-Con cells into nude mice. THAP9-AS1 knockdown repressed tumor volume growth and tumor weight growth, but restoration of YAP expression rescued the effect of THAP9-AS1 knockdown on tumor growth in vivo (Fig. 7A).

To further define the role of THAP9-AS1 clinically and to further verify its correlation with YAP in clinical samples, we measured THAP9-AS1 and miR-484 expression via in situ hybridization (ISH) in a cohort of PDAC specimens (N = 57). A THAP9-AS1 expression signal was detected in about 82% of the specimens. THAP9-AS1–positive samples also exhibited a signal in both cytoplasm and nucleus (Fig. 7B). Whereas, a miR-484 expression signal was detected in all the specimens (Fig. 7B). Moreover, we performed IHC staining of YAP in PDAC samples, which have been subjected to ISH analysis of THAP9-AS1. As expected, the expression of THAP9-AS1 was positively correlated with YAP protein level (Fig. 7C). Importantly, either high THAP9-AS1 or YAP protein level in PDAC significantly predicts a poor outcome of patients with PDAC (Fig. 7D). In addition, online analysis based on TCGA data (http://gepia.cancer-pku.cn/) also showed that THAP9-AS1 and YAP levels were negatively correlated with overall survival of patients with PDAC (Fig. 7E). These data indicate that THAP9-AS1/YAP axis is strongly correlated with PDAC tumor growth and poor outcome in patients with PDAC.

In this study, we for the first time delineated the critical role of the lncRNA THAP9-AS1 in PDAC. Our findings provided several advanced insights into the underlying mechanisms for PDAC growth: (i) THAP9-AS1 is overexpressed in PDAC to (ii) promote PDAC growth via enhancing YAP activity; (iii) THAP9-AS1 acts as a competing endogenous RNA (ceRNA) to upregulate YAP expression; (iv) THAP9-AS1 interacts with YAP to block LATS1-mediated YAP inactivation; and (v) YAP/TEAD1 complex transcriptionally regulates THAP9-AS1 expression to form a positive feedback loop in PDAC cells (Fig. 7F).

Emerging evidence has shown the important roles of lncRNAs in cancers, especially their effects on cancer growth and malignant transformation. Dysregulation of some lncRNAs have been reported to regulate important cancer biological processes, such as proliferation (24), metabolism (25), metastasis (26) and cancer cells stemness (15). Recent evidence suggests that lncRNAs can act as ceRNAs to regulate miRNAs, subsequently to regulate expression of target genes. For instance, LncRNA-ATB regulated by TGFβ activates the invasion-metastasis cascade through competitively binding miR-200s with ZEB1 and ZEB2 to inducing EMT process in hepatocellular carcinomas (14). HOTAIR promotes gastric cancer progression by sponging miR-331-3p to upregulate HER2 expression (27). Exosome-transmitted lncARSR functions as a sponge of miR-34/miR-449 to induce c-MET and AXL expression to mediate sunitinib resistance in renal cell carcinoma (28). In this study, we found that THAP9-AS1 shared miR-484 response elements with YAP and facilitated YAP expression by sponging miR-484. YAP was experimentally validated to be a bona fide target of miR-484. Function inhibition of miR-484 effectively rescues the decreased expression of YAP mRNA and protein that are induced by THAP9-AS1 knockdown in PDAC cells, indicating that THAP9-AS1 acts as a ceRNA. Recent studies reported the important involvement of miR-484 in cancer progression. Reduced expression of miR-484 in cervical cancer promoted proliferation, invasion, and EMT process, due to upregulation of its targets ZEB1 and Smad2 (29). In colorectal cancer with MSI (microsatellite instability), miR-484 was decreased because of the CpG island methylation. miR-484 suppresses MSI colorectal cancer cell growth via inhibiting CD137L/IL8 axis (30). Our results also extended the regulatory mechanism for miR-484 function.

Our findings that THAP9-AS1 knockdown increases the phosphorylation of YAP at serine 127 are new. In mammalian cells, the major components of the Hippo signaling pathway contain the transcriptional coactivator YAP, nuclear transcription factors TEAD1-4, and their upstream kinases (MST1/2 and LATS1/2; ref. 31). Beyond transcriptional regulation due to amplification (32), YAP is also tightly regulated posttranscriptionally by kinases-mediated degradation or cytoplasmic sequestration. In response to unfavorable extracellular- or intracellular signal, MST1/2 phosphorylates and activates LATS1/2. Activated LATS1/2 in turn phosphorylates YAP at serine 127, which as one of the most common posttranslational modifications providing the binding site for the 14–3-3 protein and resulting in cytoplasmic location of YAP to repress its activity (19). Otherwise, unphosphorylated YAP translocates into the nucleus and functions as a transcriptional coactivator of TEAD (33). Here, to confirm the function of THAP9-AS1 on YAP protein, we used inhibitor of miR-484 to exclude the change of total YAP protein level. Our results indicated that THAP9-AS1 negatively regulated the phosphorylation at YAP serine 127, irrespective of LATS1 level in PDAC cells. Recent evidence indicates that lncRNAs could directly bind to proteins to regulate proteins modification and activity. Bian and colleagues reported that FEZF1-AS1 could bind the PKM2 protein to increase cytoplasmic and nuclear PKM2 protein levels, resulting in increased aerobic glycolysis and STAT3 activation (25). Here, we identified the physical binding between THAP9-AS1 and YAP protein. This binding prevents YAP from interaction and phosphorylation by LATS1. Our data indicated that THAP9-AS1 regulated YAP signaling at two aspects: promoting expression and modifying function in PDAC cells. YAP as oncogene is known to be involved in cell proliferation, invasion, EMT, and metastasis. YAP activation has been shown to correlate with poor outcome in several cancer types (34). Evenly, YAP activation mediated the tumor relapse of PDAC with addiction to Kras oncogene (35). Furthermore, we showed YAP could transcriptionally regulate THAP9-AS1 expression via YAP/TEAD1 complex, forming a feed-forward loop to maintain YAP signaling.

In conclusion, our work presented here shows that as lncRNA THAP9-AS1 is upregulated in PDAC, which is associated with poor clinical outcome. THAP9-AS1 promotes PDAC growth via YAP signaling. THAP9-AS1 promotes YAP activity by sponging miR-484 to upregulate YAP expression and by binding YAP protein to enhance its activation. Our findings provide insight into the THAP9-AS1/YAP axis as promising therapeutic target against PDAC, implying important translational implications. However, further studies should be performed to develop precise strategies targeting THAP9-AS1/YAP signaling.

No potential conflicts of interest were disclosed.

Conception and design: N. Li, H. Liu, W. Liu, G. Zheng

Development of methodology: N. Li, G. Yang, L. Luo, L. Ling, J. Lan, X. Jia, Q. Zhang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. Li, G. Yang, L. Luo, X. Wang, L. Shi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Li, L. Luo, X. Jia, Z. Long, G. Zheng

Writing, review, and/or revision of the manuscript: N. Li, W. Hu, W. Liu, G. Zheng

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G. Yang, L. Luo

Study supervision: J. Liu, Z. He, G. Zheng

This study was supported by the National Natural Science Foundation of China (grant no. 81872197, to G. Zheng; grant no. 81672616, to G. Zheng; grant no. 81401989, to N. Li; and grant no. 81602016, to X. Jia); Guangdong Natural Science Funds for Distinguished Young Scholars (grant no. 2016A030306003, to G. Zheng); Guangdong Special Support Program (grant no. 2017TQ04R809, to G. Zheng); Guangzhou Key Medical Discipline Construction Project Fund; Science and Technology Program of Guangzhou, China (grant no. 201710010100, to G. Zheng; grant no. 201804010001, to N. Li; and grant no. 201804010077, to W. Hu); and Guangzhou Municipal University Scientific Research Project (grant no. 1201610027, to G. Zheng).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.