LncRNA GAS6-AS1 facilitates tumorigenesis and metastasis of colorectal cancer by regulating TRIM14 through miR-370-3p/miR-1296-5p and FUS

GAS6-AS1 expression data and related clinical information were extracted from the TCGA-COAD dataset using TCGA database (https://​cancergenome.​nih.​gov/​). These data were analyzed using R software (3.5.1). The GAS6-AS1-targeted miRNAs and the binding site were predicted using StarBase (http://​starbase.​sysu.​edu.​cn/​index.​php). The binding site between miR-370-3p/miR-1296-5p and TRIM14 was predicted using TargetScan (http://​www.​targetscan.​org/​vert_​72/​).

Paired CRC and para-carcinoma samples were obtained from 40 patients who underwent tumor resection at the First Affiliated Hospital of Soochow University (November 2019 to May 2020). Each patient was CRC-positive at diagnosis and had not received preoperative chemoradiotherapy. The age of patients was range from 40 to 75-year-old. Informed consent was obtained from all patients. This study was approved by the ethics committee of the First Affiliated Hospital of Soochow University (No.2019138).

The human CRC cell lines HT29, LoVo, RKO, SW620, the human normal colon epithelial cells (NCM460), and the human embryo kidney cell line HEK-293T were purchased from the Cell Bank of the Chinese Academy of Sciences and the American Type Culture Collection (ATCC). The cells were cultured in RPMI-1640 (Hyclone, USA) or DMEM medium (Hyclone, USA) with 10% fetal bovine serum in a cell incubator at 37 °C and 5% CO₂.

Total RNA was extracted using a TRIzol reagent kit (Invitrogen, USA). Primers were designed by Sangon Biotech (Shanghai, China) and were listed in Additional file 1: Table S1. Quantitative real-time PCR was conducted with either 2X SYBR Green qPCR Master Mix (Abm, Canada) or miDETECT A Track miRNA qRT-PCR Starter Kit (RiboBio, China). β-actin (for mRNAs and lncRNAs) and U6 (for miRNAs) served as controls.

The PARIS Kit (Invitrogen, USA) was utilized to extract the cell nuclear and cytoplasmic RNA, for subsequent qRT-PCR. We used β-actin (for cytoplasm) and U6 (for nuclear) for normalizations.

The pCDH-GAS6-AS1, pCDH (blank plasmid), pLenti hU6/shRNA-GAS6-AS1, and pLenti hU6/shRNA-Control was established by Lingke Biotechnology (Shanghai, China). Then, plasmids were transfected into with HEK-293T cells psPAX2 and pMD2G. Virus particles were sterile-filtered, concentrated, and subsequently used for infecting HT29 and LoVo cells to produce corresponding stable cells.

The miR-370-3p mimics, miR-1296-5p mimics, miR-370-3p inhibitors, miR-1296-5p inhibitors and matched negative controls were synthetized by RiboBio (Guangzhou, China). Small interfering TRIM14 (si-TRIM14, CCA CAT GTG GGT ACT GCA T) were purchased from RiboBio. Lipofectamine 2000 (Invitrogen, USA) was applied for transfection following manufacturer's guide.

Cells were grown in 96-well plate with 5 × 10³ per well. Cell proliferation was detected by Cell Counting Kit (Beyotime, China) after transfected or not. Then the absorbance at 450 nm was measured.

BeyoClick EdU-555 Kits (Beyotime, China) were used for EdU assays. Cells were cultivated in medium containing 10 μM EdU before fixing with 4% paraformaldehyde and subsequent stained with EdU reaction buffer. To visualize the DNA, the cells were stained with Hoechst and observed with fluorescence microscope. The EdU-positive cells were counted.

Cells were seeded into 24-well plates and cultured in the incubator. 10 μl pipette tips were used to scratch on monolayer cells at multiple sites. The area of scratch was observed at 0 h, 24 h and 48 h after scratching. Resulting images were processed with ImageJ software.

Transwell chambers (Corning, USA) with (migration) and without (invasion) Matrigel (BD Biosciences, USA) were applied to perform transwell assays. Briefly, 1 × 10⁵ cells were cultured in the upper wells with 100 µL serum-free medium. The lower chamber was infused with 600 µL complete medium. After 48 h, the cells on the lower side of the membrane were fixed in 4% paraformaldehyde, stained with 0.1% crystal violet, and quantified under a microscope.

Animal experiments were conducted following the principles of the Animal Management and Use Committee of Soochow University and approved by the Medical Ethics Committee of The First Affiliated Hospital of Soochow University (Approval No.2017213). 24 BALB/c nude mice (4–5 weeks) were obtained from SLAC Laboratory Animal Center (Shanghai, China) and reared in a specific-pathogen-free environment at 23–25 °C. Xenograft tumors were established by subcutaneously injecting 5 × 10⁶ cells (HT29-GAS6-AS1, HT29-Vector, LoVo-shGAS6-AS1, and LoVo-shControl cells). The tumor size was determined weekly. The mice were euthanized after four weeks, and the tumors were excised and weighed.

For lung metastasis models, 2 × 10⁶ cells (HT29-GAS6-AS1 vs. HT29-Vector, LoVo-shGAS6-AS1 vs. LoVo-shControl) were injected through the tail vein. The mice were euthanized four weeks later. The lung tissue was dissected and fixed with formalin. Lung foci were counted using H&E staining.

Fluorescent in situ hybridization (FISH) kit (RiboBio, China) was applied for the in-situ detection of GAS6-AS1 in HT29 and LoVo cells following the guidelines. Cells were observed by the fluorescence microscope.

For dual-luciferase reporter assays, the wild-type (WT) 3′-UTR of GAS6-AS1/TRIM14 gene containing miR-370-3p and miR-1296-5p binding sites, and the mutant (MUT) 3′-UTR of GAS6-AS1/TRIM14 gene were obtained from RiboBio (Guangzhou, China). The WT/MUT 3′-UTR were co-transfected with mimics or negative control with Lipofectamine 2000 (Invitrogen, USA). Notably, 48 h post-transfection, cells were lysed. Then, we applied the Dual-Luciferase Reporter Assay Kit (Beyotime, China) to evaluate the luciferase activities.

RIP was performed using EZ-Magna RIP kits (Millipore, USA). Cells were collected and incubated with anti-Ago2 antibody (Abcam, UK) 24 h after transfection. IgG was used as the negative control. Quantitative RT-PCR was used to assess the co-precipitated RNAs.

Here, cells were solubilized in RIPA lysis buffer (Merck, China.). With the BCA method, we standardized the protein concentration. Lysates were resolved by 10–15% SDS polyacrylamide gels and transferred onto PVDF membranes. The membranes were incubated overnight at 4 °C with primary antibodies, including anti-TRIM14 (1:500, Proteintech, USA), anti-GAPDH (1:5000, CST, USA), and anti-E-cadherin (1:1000), anti-N-cadherin (1:1000), anti-Vimentin (1:1000) (Immunoway, USA), followed by another incubation with appropriate secondary antibodies (the anti-rabbit or anti-mouse antibodies were from Immunoway) at 25 °C 2 h. The blots were developed using ECL.

Data on lncRNA GAS6-AS1 expression and corresponding clinical information were obtained from the TCGA-COAD dataset in the TCGA database, including 41 normal colon tissues and 446 colon cancer tissues. Detailed information on colon cancer samples is shown in Additional file 2: Table S2, including sex, age, T stage, N stage, M stage, clinical stage, and microsatellite instability (MSI) status (MSH: high microsatellite instability; MSS: microsatellite instability low). Expression differences and clinical prognostic analyses were performed using R software. The GAS6-AS1 level in colon cancer tissues was significantly higher than that in normal tissues (Fig. 1A). In addition, the relative GAS6-AS1 level of patients with T4 stage disease was higher than that of patients with T1-3 stage disease (Fig. 1B). GAS6-AS1 levels were higher in lymphatic metastasis-positive patients than those in lymphatic metastasis-negative patients (Fig. 1C). Furthermore, GAS6-AS1 levels in patients with distant metastasis were significantly higher than those in patients without distant metastasis (Fig. 1D). Although no statistical significance was detected between GAS6-AS1 in stage III and stage IV patients, stage IV patients exhibited higher GAS6-AS1 levels than stage I and II patients. GAS6-AS1 expression in patients with stage III-IV was higher than that in stage I and II patients (Fig. 1E). Patients with MSS and MSL exhibited higher GAS6-AS1 expression levels than patients with MSH (Fig. 1F). Survival analysis revealed that patients with relatively high GAS6-AS1 expression had shorter survival times than those with relatively low GAS6-AS1 expression (P = 0.028) (Fig. 1G). Receiver operating characteristic (ROC) analysis of GAS6-AS1 was performed and the area under the curve (AUC) was 0.929 (Fig. 1H), which indicates the diagnostic value of GAS6-AS1. After examining GAS6-AS1 levels in CRC, 40 pairs of operable CRC and para-cancer matched samples were collected and analyzed via qRT-PCR. The results indicated remarkably higher GAS6-AS1 levels in the CRC samples than in the para-cancerous samples (Fig. 1I). These findings suggested that GAS6-AS1 expression was elevated in CRC and that GAS6-AS1 was positively associated with tumor progression and poor prognosis.

GAS6-AS1 levels in human CRC cell lines (HT29, SW620, LoVo, and RKO cells) and NCM460 cells were evaluated using qRT-PCR. Elevated GAS6-AS1 expression was observed in CRC cells (Fig. 2A). LoVo cells, which had the highest GAS6-AS1 levels, and HT29 cells, which had the lowest GAS6-AS1 levels, were selected for in vitro experiments. To investigate whether GAS6-AS1 influences CRC cell viability and mobility, we overexpressed GAS6-AS1 in HT29 cells (Fig. 2B) and inhibited GAS6-AS1 expression in LoVo cells (Fig. 2C). CCK8 assays demonstrated that the viability of GAS6-AS1-overexpressing HT29 cells was significantly increased (Fig. 2D), whereas GAS6-AS1 knockdown significantly suppressed LoVo cell proliferation (Fig. 2E). EdU experiments demonstrated that HT29 cell viability was greatly increased after GAS6-AS1 overexpression (Fig. 2F), whereas LoVo cell viability greatly decreased after GAS6-AS1 knockdown (Fig. 2G).

To further uncover the function of GAS6-AS1 in CRC cell mobility, scratch and transwell assays were performed. In scratch assays, we found that GAS6-AS1 upregulation significantly increased the migration of HT29 cell (Fig. 2H), whereas GAS6-AS1 downregulation reduced the migration of LoVo cell (Fig. 2I). In line with these results, transwell migration experiments revealed that the number of migrating HT29 cells increased after GAS6-AS1 overexpression (Fig. 2J), while the number of migrating LoVo cells was reduced after GAS6-AS1 knockdown (Fig. 2K). Furthermore, transwell invasion assays demonstrated that GAS6-AS1 overexpression promoted the invasion of HT29 cells (Fig. 2L), whereas downregulating GAS6-AS1 impeded LoVo cell invasion (Fig. 2M). These results indicated that GAS6-AS1 promoted CRC cell viability and mobility in vitro.

To further evaluate the role of GAS6-AS1 in CRC tumorigenicity in vivo, xenograft models were established. Notably, GAS6-AS1 overexpression promoted subcutaneous xenograft tumor growth (Fig. 3A), whereas GAS6-AS1 knockdown significantly reduced tumor volume (Fig. 3B). Tumors of the GAS6-AS1 overexpression group weighed more than control tumors (Fig. 3C). Lower tumor weight was observed in the GAS6-AS1 knockdown group compared to the control group (Fig. 3D). To further reveal the effect of GAS6-AS1 on CRC metastasis in vivo, we developed lung metastasis models of CRC and lung metastatic foci were counted. Compared to the control group, GAS6-AS1 overexpression resulted in more metastatic pulmonary nodules (Fig. 3E). GAS6-AS1 knockdown resulted in fewer metastatic pulmonary nodules (Fig. 3F). Collectively, these observations demonstrated that GAS6-AS1 overexpression facilitated CRC growth and metastasis in vivo.

To uncover cellular GAS6-AS1 functions, quantitative RT-PCR and fluorescent in situ hybridization (FISH) were performed to predict its subcellular localization. qRT-PCR demonstrated that GAS6-AS1 was localized both in the cytoplasm and nuclei of CRC cells. This observation was verified using FISH (Fig. 4A, B). Based on these results, we hypothesized that GAS6-AS1 functioned as a ceRNA. Thus, RNA immunoprecipitation (RIP) assays were conducted using an anti-Ago2 antibody. The results demonstrated that endogenous GAS6-AS1 was preferentially enriched in Ago2-RIPs compared to control IgG-RIPs (Fig. 4C). These findings suggested that GAS6-AS1 may serve as a ceRNA that promoted CRC development and progression.

GAS6-AS1 targeted miRNAs were predicted using the starbase database and we obtained 15 miRNAs that may bind with GAS6-AS1. We made use of TCGA database to analyze the expression difference and prognosis of the 15 miRNAs (Additional file 3: Fig. S1A). The results showed that five of them were highly expressed in normal tissues compared with colon cancer tissues, including miR-370-3p, miR-3173-5p, miR-1296-5p, miR-324-3p and miR-491-5p. Further survival analysis suggested that miR-370-3p and miR-1296-5p may play an anti-cancer role in colon cancer (Additional file 3: Fig. S1B). Therefore, we chose miR-370-3p and miR-1296-5p for further experiments. To verify whether GAS6-AS1 could bind with miR-370-3p and miR-1296-5p, Ago2-RIP assays and dual-luciferase reporter assays were performed. The Ago2-RIP assay demonstrated GAS6-AS1 higher enrichment in the miR-370-3p or miR-1296-5p mimic groups than in the NC mimic group (Fig. 4D, E). Dual-luciferase reporter assays demonstrated that upregulating miR-370-3p/miR-1296-5p suppressed the luciferase activity of the wild-type GAS6-AS1 reporter plasmid, but not the GAS6-AS1-370Mut or GAS6-AS1-1296Mut vectors (Fig. 4F, G). qRT-PCR also showed that the miR-370-3p and miR-1296-5p levels in CRC cells were negatively regulated by GAS6-AS1 (Fig. 4H, I). These findings confirmed that GAS6-AS1 directly bind to miR-370-3p and miR-1296-5p.

To verify whether GAS6-AS1 regulated CRC cell viability and mobility via miR-370-3p and miR-1296-5p, we performed rescue experiments. In vitro functional experiments demonstrated that CRC cell growth, migration, and invasion decreased after miR-370-3p/miR-1296-5p upregulation. In addition, miR-370-3p/miR-1296-5p partially reversed GAS6-AS1-mediated cell proliferation, migration, and invasion (Fig. 5A–E, Additional file 4: Fig. S2A, C). Moreover, inhibiting miR-370-3p or miR-1296-5p promoted CRC cell proliferation, migration, and invasion, whereas downregulating miR-370-3p or miR-1296-5p reversed the effect of GAS6-AS1 knockdown (Fig. 5F–J, Additional file 4: Fig. S2D, F). These results suggested that GAS6-AS1 promoted the CRC oncogenesis via sponging miR-370-3p/miR-1296-5p.

To identify the genes that shared complementary binding sites of miR-370-3p and miR-1296-5p with GAS6-AS1, TargetScan was used to predict the common target genes of miR-370-3p and miR-1296-5p. Combined with a literature search, the TargetScan results predicted the common target gene, TRIM14. Previous reports revealed that higher TRIM14 levels in CRC contribute to disease progression [21, 22]. Therefore, TRIM14 was chosen for further analysis. Luciferase reporter assays indicated that miR-370-3p or miR-1296-5p upregulation reduced the luciferase activity of the wild-type TRIM14 reporter plasmid, but not of the TRIM14-370Mut or TRIM14-1296Mut vectors (Fig. 6A, B).

Moreover, TRIM14 mRNA levels in CRC cells were elevated after miR-370-3p or miR-1296-5p knockdown (Fig. 6C). Western blot analysis confirmed higher TRIM14 expression after miR-370-3p or miR-1296-5p knockdown (Fig. 6D), suggesting that miR-370-3p/miR-1296-5p negatively regulate TRIM14 levels in CRC cells. We also assessed the expression of E-cadherin, N-cadherin, and vimentin, which are EMT-associated markers. Notably, downregulating miR-370-3p or miR-1296-5p caused significant de-repression of E-cadherin and increased N-cadherin and vimentin levels. In addition, TRIM14 knockdown rescued the effects of miR-370-3p/miR-1296-5p inhibition in CRC cells (Fig. 6D). Thus, miR-370-3p/miR-1296-5p directly targeted TRIM14 to inhibit EMT in CRC cells.

We also observed high TRIM14 levels in CRC cells upon GAS6-AS1 overexpression, whereas TRIM14 knockdown or miR-370-3p/miR-1296-5p upregulation partially reversed the effect of GAS6-AS1 in CRC cells. In addition, GAS6-AS1 overexpression significantly decreased E-cadherin levels and increased N-cadherin and vimentin expression. Likewise, knocking down TRIM14 or upregulating miR-370-3p/miR-1296-5p reversed the changes caused by GAS6-AS1 (Fig. 6E, F). Thus, the results indicated that GAS6-AS1 could adsorb miR-370-3p/miR-1296-5p to upregulate TRIM14 and promote EMT in CRC cells.

Emerging evidence shows that RNA-binding proteins (RBPs), including FUS, maintain mRNA stability by interacting with lncRNAs [23]. Using the Starbase database, we predicted that GAS6-AS1 and TRIM14 both bind to FUS. RIP assays were performed to confirm this prediction. Both GAS6-AS1 and TRIM14 were harvested using FUS co-immunoprecipitation (Fig. 6G). Moreover, the level of TRIM14 mRNA bound to FUS significantly increased when GAS6-AS1 was upregulated but decreased when GAS6-AS1 knockdown (Fig. 6H). Meanwhile, qRT-PCR showed that overexpressing GAS6-AS1 increased the mRNA level of TRIM14, which was partially offset by downregulating FUS (Fig. 6I). In addition, GAS6-AS1 upregulation delayed TRIM14 mRNA degradation, whereas FUS knockdown partially rescued this effect (Fig. 6J). Based on these findings, GAS6-AS1 maintained TRIM14 mRNA stability by recruiting FUS.

To evaluate the correlation between GAS6-AS1 and TRIM14 expression in CRC tissues, TRIM14 levels in 40 CRC cases were analyzed by qRT-PCR for correlation analysis. The results showed a positive correlation between GAS6-AS1 and TRIM14 (Fig. 7A). Further rescue experiments were performed to verify whether GAS6-AS1 promoted tumor progression in a TRIM14-dependent manner. CCK8 and EdU assays revealed that TRIM14 downregulation partially reversed the effects of GAS6-AS1 overexpression on CRC cell proliferation (Fig. 7B, C). Scratch, transwell migration, and transwell invasion experiments demonstrated that TRIM14 knockdown rescued the effects of GAS6-AS1 overexpression on CRC cell migration and invasion (Fig. 7D–F). These findings demonstrated that GAS6-AS1 promoted CRC progression via TRIM14.