The lncRNA NEAT1 activates Wnt/β-catenin signaling and promotes colorectal cancer progression via interacting with DDX5

Fresh samples were obtained from 71 newly diagnosed CRC patients who underwent no preoperative therapy prior to surgical resection at Fudan University Shanghai Cancer Center (FDUSCC) between 2008 and 2009. This study was approved by the institutional review board of Shanghai Cancer Center. The median follow-up time was 80 months, and the longest follow-up was 87 months.

NEAT1 siRNAs (si1: sense 5’-GACCGUGGUUUGUUACUAUdTdT-3′, antisense 5′-AUAGUAACAAACCACGGUCdTdT-3′; si2: sense 5′-GUUGGUCAUUCCUAAAUCUTT-3′, antisense 5′-AGAUUUAGGAAUGACCAACTT-3′), si-DDX5 (sense: 5′-GCAAGUAGCUGCUGAAUAUUU-3′, antisense: 5′-pAUAUUCAGCAGCUACUUGCUU-3′), and a scramble siRNA were purchased from GenePharma (Shanghai, China). ShNEAT1 lentiviruses (5′-GACCGUGGUUUGUUACUAU-3′) and shNC (representing the sh-negative control, 5′-UUCUCCGAACGUGUCACGU-3′) were generated by Sbo-Bio (Shanghai, China). Transient transfection of pc3.1-NEAT1, pc3.1 (control vector), or siRNA was performed using Lipofectamine 3000 (Invitrogen, CA, USA) according to the manufacturer’s instructions. To establish stable cell lines, shNEAT1 lentiviruses were transduced into HCT116 and SW1116 cells with polybrene (5 μg/mL; Sigma-Aldrich, MO, USA). Then, the cells were selected with 0.5 μg/mL of puromycin for 14 days. The transfection efficiencies were verified by RT-qPCR and western blotting.

This study complied with the Animal Care guidelines of FDUSCC. Male BALB/c nude mice (6 weeks old) were housed under specific pathogen-free conditions. HCT116-shNC and HCT116-shNEAT1 cells were injected either subcutaneously (n = 4, 5 × 10⁶/mouse) or into the tail vein (n = 4, 2 × 10⁶/mouse). The mice were sacrificed after 4–6 weeks. The tumors and lungs of the mice were removed, fixed in 10% formalin, and stored at − 80 °C for the subsequent analyses. Each animal experiment was performed in triplicate.

The standard western blotting assay was performed as previously described [20]. The specific primary antibodies are listed in Additional file 2: Table S1.

RNA pull-down was performed using the Magnetic RNA-Protein Pull-Down kit (Pierce, MA, USA) in accordance with the manufacturer’s instructions. The protein bands on the gel were silver-stained. Bands of interest were identified by mass spectrometry (MS) and confirmed by western blotting. RNA immunoprecipitation (RIP) was performed using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, MA, USA) according to the manufacturer’s protocol.

The DDX5 promoter was cloned into the pGL3 basic luciferase reporter vectors (Promega, USA). In total, 5000 cells were seeded into each well of a 96-well plate and transfected with 100 ng of the TOP/FOP-flash reporter plasmids (Millipore, MA, USA), 100 ng of an expression vector (pGL3-DDX5 or pGL3-Basic) or 0.25 μl of siRNA. DDX5 promoter activity and TOP/FOP-flash were normalized by cotransfection with 10 ng of a Renilla luciferase reporter. After 24 h of incubation, the luciferase activity was detected using the Dual-Luciferase Reporter Assay System (Promega, USA).

Each experiment was independently repeated at least three times, and the data were presented as the mean ± SD. For the western blotting, EdU, wound healing, RIP, and Transwell assays, the representative results of three independent experiments are shown.

Comparisons between groups were analyzed using Student’s t tests for the mRNA levels, and clinicopathological parameters were compared using the χ² test. Correlations between the mRNA levels were calculated with Spearman’s rank correlation coefficients. Survival curves were plotted using the Kaplan-Meier method and compared using the log-rank test. p < 0.05 was considered significant. All statistical analyses were conducted using the SPSS 19.0 statistical software (SPSS Inc., IL, USA).

To investigate the clinical significance of NEAT1, we analyzed its expression levels in the publicly available TCGA dataset and in data from 71 CRC samples from FDUSCC. Both datasets showed that NEAT1 expression was significantly upregulated in the tumor tissues compared to the levels in the normal tissues (Fig. 1a, b). Correlations between the clinicopathological features of the CRC patients and the NEAT1 levels are summarized in Additional file 2: Table S3. NEAT1 was not related to age, gender, AJCC stage, T stage, N stage, and other features. The univariate Cox proportional hazards models for overall survival (OS) and disease-free survival (DFS) are summarized in Additional file 2: Table S4. Our results revealed that patients with high NEAT1 expression (divided by the mean value) showed obviously poorer OS than those with low NEAT1 expression [23/35 (65.7%) vs. 32/36 (88.9%), HR 4.457, 95% CI 1.267–15.685, p = 0.017, Fig. 1c]. In addition, 8 of the 28 (28.6%) patients with high NEAT1 expression experienced recurrence, whereas only 2 of the 31 (6.452%) patients with low NEAT1 relapsed. The Kaplan-Meier curves showed that patients with high NEAT1 expression has poorer DFS than those with relatively low NEAT1 expression (p = 0.028, Fig. 1d).

To assess the functional role of NEAT1 in colorectal cancer cells, first we examined the baseline NEAT1 RNA levels in eight CRC cell lines (RKO, CACO2, SW1116, LOVO, SW480, SW620, HT29, and HCT116) by RT-qPCR (Fig. 2a). NEAT1 was expressed at much higher levels in the HCT116 and SW1116 cells and relatively lower levels in the HT29 cells and was mainly located in the nucleus (Additional file 3: Figure S1A). Next, HCT116 and SW1116 cells were transfected with the NEAT1 siRNA and HT29 cells were transfected with the NEAT1 plasmid. The knockdown and overexpression efficiencies were verified by RT-PCR (Fig. 2b). The CCK8 assay showed that downregulation of NEAT1 significantly attenuated cell proliferation of the HCT116 and SW1116 cells, whereas forced NEAT1 expression had the opposite effect in the HT29 cells (Fig. 2c). This result was also confirmed by the EdU and colony formation assays (Fig. 2d, e, Additional file 3: Figure S1B-C).

Furthermore, our results showed that repression of NEAT1 induced late apoptosis (Additional file 4: Figure S2A-B) and G0/G1 phase arrest (Additional file 4: Figure S2D) in HCT116 and SW1116 cells. The western blotting analysis verified increased PARP1 and cleaved caspase 3 expression (Additional file 4: Figure S2C) and decreased cyclin D1 and p27 expression. In contrast, the G2/M transition-related markers cyclin B1 and CDC25B were not obviously altered (Additional file 4: Figure S2E). These results suggested that NEAT1 promoted CRC cell proliferation by reducing cell apoptosis and inducing the G1 to S phase cell cycle transition.

To further validate these effects in vivo, we injected HCT116-shNEAT1 and HCT116-shNC cells into the subcutis of nude mice. Consistent with our in vitro results, the volumes and weights of the tumors formed by the HCT116-shNEAT1 cells were significantly smaller than those formed by the control cells (Fig. 2f–h). The Ki-67 index was lower in the NEAT1 knockdown groups (Fig. 2i). These results suggested that NEAT1 promoted the tumorigenesis of CRC cells both in vitro and in vivo.

To determine the effect of altered NEAT1 expression on the migration and invasion of CRC cells, NEAT1-siRNA-transfected and NEAT1 plasmid-transfected CRC cells were wounded by scratching and maintained for 24 h. Knockdown of NEAT1 significantly inhibited the flattening and spreading abilities of the HCT116 and SW1116 cells (Fig. 3a, b, Additional file 5: Figure S3A), whereas overexpression of NEAT1 strongly promoted the flattening and spreading abilities of the HT29 cells (Fig. 3c, Additional file 5: Figure S3A). This result was also confirmed by the Transwell assay. We found that the invasive abilities of the NEAT1-siRNA-transfected HCT116 and SW1116 cells were significantly inhibited (Fig. 3d, Additional file 5: Figure S3B), whereas the invasive ability was higher in the NEAT1-transfected HT29 cells than in the control cells (Fig. 3e, Additional file 5: Figure S3B). Accordingly, the RT-qPCR and western blotting results showed that downregulation of NEAT1 resulted in higher E-cadherin expression and lower N-cadherin expression at both the mRNA and protein levels (Fig. 3f, Additional file 5: Figure S3C). The cell migration and invasion markers MMP2 and MMP9 were also decreased when NEAT1 was downregulated (Fig. 3f). However, overexpression of NEAT1 led to increased N-cadherin, MMP2, and MMP9 and decreased E-cadherin expression (Fig. 3f).

Next, we evaluated the functional role of NEAT1 in CRC cell metastasis in vivo. We injected HCT116-shNEAT1 and HCT116-shNC cells into the tail veins of mice in groups of four. The number of lung metastatic tumor nodules was decreased in the HCT116-shNEAT1 group compared with that of the HCT116-shNC group (Fig. 3g). Consistent with the effects of NEAT1 expression on migration and invasion in vitro, NEAT1 knockdown significantly abrogated metastasis both in vitro and in vivo.

As described above, NEAT1 plays an important role in CRC progression, although the detailed mechanism remains unknown. Because lncRNAs can interact with proteins, a biotin RNA-protein pull-down assay was performed to identify potential proteins binding to NEAT1 (Fig. 4a). Interestingly, DDX5 was identified as an interacting target of NEAT1 by MS analysis (Additional file 6: Table S6), and an immunoprecipitation assay further confirmed that DDX5 directly bound to NEAT1 (Fig. 4b). Intriguingly, the RIP assay confirmed the interaction between DDX5 and NEAT1 in extracts from HCT116 cells. SNRNP70 was used as the positive control (Fig. 4c). Collectively, these results demonstrated a direct interaction between DDX5 and NEAT1.

Next, we analyzed the regulatory effects of NEAT1 on DDX5. We found that downregulation of NEAT1 effectively reduced the DDX5 protein level but not the mRNA level in the HCT116 and SW1116 cells (Fig. 4d, e). In addition, immunohistochemistry analysis of xenografts showed that nude mice in the NEAT1-knockdown group exhibited lower DDX5 expression (Fig. 2i). To further clarify the mechanism underlying the regulation of DDX5 expression by NEAT1, first we examined whether NEAT1 could directly regulate the transcriptional activity of DDX5. Our results showed that DDX5 promoter activity was not increased in NEAT1-transfected HCT116 cells (Fig. 4f), suggesting that NEAT1 might participate in regulation of DDX5 at the posttranscriptional level. Therefore, we used the protein synthesis inhibitor cycloheximide (CHX) to observe the effect of NEAT1 on DDX5 degradation. The western blotting results showed that overexpression of NEAT1 in HT29 cells enhanced DDX5 protein stability (Fig. 4g). Moreover, the 26S proteasome inhibitor MG132 rescued the reduction of DDX5 caused by repression of NEAT1 in HCT116 cells (Fig. 4h), suggesting that NEAT1 elevated DDX5 by reducing its degradation. Taken together, our data indicated that NEAT1 directly bound the DDX5 protein and enhanced its stability in CRC cells.

Recently, decreased NEAT1 expression was reported to inhibit Wnt/β-catenin signaling pathway activity in glioblastoma [21]. In our study, we performed a TOP/FOP-flash luciferase assay; the results revealed that Wnt/β-catenin signaling was inhibited by NEAT1 depletion (Fig. 5b). However, the Dual-Luciferase Reporter Assay showed that NEAT1 did not influence the activity of the β-catenin 3′-UTR (Fig. 5a), which indicated that NEAT1 did not regulate β-catenin directly. DDX5 reportedly forms a complex with β-catenin and promotes its transcriptional ability to activate gene transcription [22]. Co-IP assays detected an interaction between endogenous DDX5 and β-catenin (Fig. 5d) in HCT116 cells, and knockdown of DDX5 reduced the β-catenin level (Fig. 5e). We speculated that NEAT1 might activate Wnt/β-catenin signaling by targeting DDX5. To explore the effects of NEAT1 on Wnt/β-catenin signaling, we examined the levels of the Wnt/β-catenin signaling targets Axin2, c-myc, and cyclin D1. The RT-qPCR and western blotting analyses showed that downregulation of NEAT1 reduced the mRNA (Additional file 5: Figure S3D) and protein (Fig. 5c) levels of Wnt/β-catenin signaling target genes. Furthermore, the increased TOP/FOP-flash luciferase activity resulting from NEAT1 overexpression was significantly suppressed by si-DDX5 (Fig. 5b), suggesting that NEAT1-induced activation of Wnt/β-catenin signaling was dependent on DDX5 expression. The RT-qPCR and western blotting analyses showed that downregulation of DDX5 did not reduce the NEAT1 RNA level but did abrogate the increase in the Axin2, c-myc, and cyclin D1 levels (Fig. 5f, Additional file 5: Figure S3E). Collectively, our results suggest that NEAT1 indirectly promotes β-catenin transcriptional activation by binding to DDX5.

To elucidate whether NEAT1 functioned in CRC cells in a DDX5-mediated manner, we performed CCK-8 and EdU assays. The results showed that overexpression of NEAT1 recovered the proliferation potential of HCT116 and SW1116 NEAT1 stable knockdown cells, which nevertheless was impaired by the simultaneous downregulation of DDX5 (Fig. 6d, e). Similarly, the effect of NEAT1 on the invasion and migration of CRC cells was also partially attenuated by repression of DDX5 (Fig. 6c, d). Therefore, we hypothesized that NEAT1 facilitated tumor proliferation and metastasis in a DDX5-mediated manner.

Finally, we tested the clinical association between NEAT1 and DDX5 in the FDUSCC dataset. Immunohistochemistry analysis of CRC samples showed that DDX5 was located in the cell nucleus and was overexpressed in cancerous tissues compared to normal tissues (Fig. 7a). In addition, DDX5 expression was correlated with NEAT1 expression in 71 CRC samples (p < 0.01, Fig. 7b). The survival analysis demonstrated that CRC patients with positive DDX5 expression (H score ≥ 50) had poorer OS and DFS than those with negative DDX5 expression (Fig. 7c, d). Next, the patients were divided into three groups based on their NEAT1 and DDX5 expression levels. Patients with positive NEAT1 and DDX5 expression had the poorest OS and DFS. In contrast, those with negative NEAT1 and DDX5 expression had the best OS (p = 0.008) and DFS (p = 0.016) (Fig. 7e, f). In addition, the multivariate COX analysis showed that the combination of NEAT1 and DDX5 was an independent prognostic indicator of OS (p = 0.024, HR = 6.916, 95% CI 1.291–37.051, Additional file 2: Table S5).