Up-regulated LINC01234 promotes non-small-cell lung cancer cell metastasis by activating VAV3 and repressing BTG2 expression

A total of 45 paired NSCLC specimens and adjacent noncancerous tissues were obtained from 45 patients with a histopathological diagnosis of NSCLC who underwent surgery at the Second Affiliated Hospital of Nanjing Medical University between 2010 and 2013. These patients did not receive chemotherapy or radiotherapy prior to surgery. Tissue samples were immediately snap-frozen in liquid nitrogen and stored at − 80 °C until required. This study was approved by the Research Ethics Committee of Nanjing Medical University and informed consent was obtained from all patients.

Four NSCLC adenocarcinoma cell lines (A549, SPC-A1, H1299, and PC9), a NSCLC squamous carcinoma cell line (H226), and a normal human bronchial epithelial cell line (16HBE) were purchased from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences (Shanghai, China). A549, H1299, and H226 cells were maintained in RPMI 1640 basic medium, and SPC-A1, PC9, and 16HBE cells were maintained in DMEM medium with 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen, Carlsbad, CA, USA). All cells were maintained in a humidified 5% CO₂ atmosphere at 37 °C.

Total RNA was extracted from tissues or cultured cells using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Aliquots of 1 μg RNA were reverse transcribed to cDNA in a final volume of 20 μl under standard conditions using a PrimeScript RT Reagent Kit (TaKara, Dalian, China). Real-time PCR analyses were performed with SYBR Premix Ex Taq (Takara) on an Applied Biosystems 7500 Real-Time PCR System. Expression of LINC01234 and other genes was normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using the relative threshold cycle method, and then converted to fold changes. Specific primers are listed in Additional file 1: Table S1. Primers for miR-27b-3p, miR-340-5p, and U6 were purchased from GeneCopoeia (Rockville, MD, USA). Our quantitative reverse transcription-PCR (qRT-PCR) results were analyzed and expressed relative to threshold cycle (CT) values, and then converted to fold changes.

The LINC01234 sequence was synthesized according to the full-length cDNA of human LINC01234 and BTG2 sequences were synthesized according to its coding sequences. All sequences were cloned into the expression vector pCDNA3.1 (Invitrogen). Control and LINC01234-targeting short hairpin RNAs (shRNAs) were purchased from Invitrogen and inserted into the pLKO.1 vector. All final constructs were verified by sequencing. Plasmids were purified using DNA Midiprep Kits (Qiagen, Valencia, CA) and transfected into NSCLC cells using X-treme GENE HP DNA transfection reagent (Roche, Basel, Switzerland). Three LINC01234-targeting small interfering RNAs (siRNAs) obtained from Invitrogen and other gene-targeting siRNAs, miRNA mimics, and miRNA inhibitors (Genepharma, Shanghai, Chian) were transfected into NSCLC cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Nucleotide sequences for the siRNAs and shRNAs are listed in Additional file 1: Table S1. Cells were harvested 48 h after transfection and analyzed as indicated for the individual experiments.

Online bioinformatics databases DIANA Tools (http://​carolina.​imis.​athena-innovation.​gr/​diana_​tools/​web/​index.​php) and miRbase (http://​www.​mirbase.​org/​) were used to predict potential miRNA-binding sites in LINC01234. The putative binding sequences were synthesized, inserted into the pGL3-Basic luciferase reporter vector (Promega), and verified by sequencing. Vectors were transfected into human HEK293T cells for 48 h, and luciferase activity was then measured using a Dual Luciferase Kit (Promega), according to the manufacturer’s instructions. The data are presented as relative firefly luciferase activity normalized to the Renilla luciferase activity. All experiments were performed three times.

For cell migration and invasion assays, cells were collected at 48 h post-transfection, and 5 × 10⁴ (for migration assay) or 1 × 10⁵ (for invasion assay) cells in serum-free medium were placed into the upper chamber of an insert (8-μm pore size; Millipore, Billerica, MA, USA). Medium containing 10% fetal bovine serum was added to the lower chamber. After incubation for 24 h, cells in the upper chamber were removed with cotton swabs, and the cells on the lower membrane surface were fixed and stained with 0.5% crystal violet solution. Experiments were independently performed three times.

All protocols were approved by the Committee on the Ethics of Animal Experiments of the Nanjing Medical University and were carried out in strict accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. For metastasis assays, SPC-A1 and A549 cells stably transfected with control shRNA or sh-LINC01234 (3 × 10⁶) were injected intravenously via the tail vein. Eight weeks post-injection, the mice were sacrificed and the lungs were removed and photographed. Tumors visible on the lung surface were counted, and the lungs were then stored in formalin.

Cytoplasmic and nuclear RNA were isolated and purified from NSCLC cells using a PARIS Kit (Life Technologies), according to the manufacturer’s instructions.

RNA immunoprecipitation (RIP) assays were performed using an EZ Magna RIP kit (Millipore) using the manufacturer’s protocol. A549 and SPC-A1 cells were lysed in complete lysis buffer, and the extracts were incubated with magnetic beads conjugated with the appropriate specific antibodies or control IgGs (Millipore) for 3–6 h at 4 °C. The beads were washed, incubated with proteinase K to remove proteins, and the purified RNA was eluted and analyzed for the presence of LINC01234 by qRT-PCR. Details of the antibodies and primers are given in Additional file 1: Table S1.

LINC01234 or control RNAs were transcribed in vitro from pcDNA3.1-LINC01234 using T7 RNA polymerase (Ambion Life) and purified using an RNeasy Mini Kit (Qiagen). One aliquot of transcribed LINC01234 RNA was biotinylated with a Biotin RNA Labeling Mix (Ambion Life). Positive control, negative control, nonbiotinylated, and biotinylated RNAs were incubated with A549 cell lysates. Streptavidin-conjugated magnetic beads were then added and the samples were incubated at room temperature. The beads were then washed, and the eluted proteins were examined by western blot analysis.

Chromatin immunoprecipitation (ChIP) assays were performed using a MagnaChIP Kit (Millipore) according to the manufacturer’s instructions, as described previously [13].

A549 and SPC-A1 cells were lysed with RIPA extraction reagent (Beyotime) supplemented with a protease inhibitor cocktail (Roche). Proteins in cell lysates were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to 0.22 μm polyvinylidene fluoride membranes (Millipore). Membranes were probed with specific antibodies using standard methods. Specific protein bands were detected by incubation with ECL chromogenic substrate and quantified by densitometry (Quantity One software; Bio-Rad, Hercules, CA, USA). Antibodies against E-cadherin, N-cadherin, Vimentin, and GAPDH (1:1000) were purchased from Cell Signaling Technology; antibodies against VAV3, EZH2, LSD1, Ago2, and HuR were purchased from Millipore; antibody against BTG2 was purchased from Absin. GAPDH was probed as an internal control. Antibodies are listed in Additional file 1: Table S1.

Statistical analyses were performed using SPSS 20.0 (IBM, Armonk, NY, USA) and Prism software (GraphPad, La Jolla, CA, USA). LncRNA expression levels in primary solid tumors and normal solid tissue samples were compared using the Mann–Whitney U test. For the remaining assays, differences between groups were assessed by paired, two-tailed Student’s t test, Wilcoxon’s test, or χ² test, as appropriate. Spearman’s correlation analysis was used to calculate the correlations between clinical factors and LINC01234, miR-27b-3p, miR-340-5p, BTG2, and VAV3 expression. All tests were two-sided, and a P value < 0.05 was considered statistically significant.

We first analyzed lung adenocarcinoma and lung squamous cell carcinoma RNA sequencing datasets from TCGA and found LINC01234 was upregulated in NSCLC tissues compared with adjacent tissues (Fig. 1a). In addition, we found a significant correlation between LINC01234 expression and lung adenocarcinoma stage from TCGA dataset (Fig. 1b). Furthermore, we examined the expression level of LINC01234 in NSCLC tissues and cell lines. qRT-PCR analysis of 45 paired NSCLC and adjacent normal tissues indicated significant upregulation of LINC01234 (fold-change > 1; P < 0.01) in 78% (35/45) of cancerous tissues compared with normal tissues (Fig. 1c). Results also indicated that LINC01234 showed increased expression in NSCLC cell lines in comparison to normal bronchial epithelial cells (Fig. 1d).

Next, we explored the relationship between LINC01234 expression and clinicopathological features of patients with NSCLC. For this, the 45 NSCLC tumor tissues were assigned to two groups (high and low LINC01234 expression) based on the median value (Fig. 1e). Higher LINC01234 expression level was also significantly correlated with tumor size (Chi-square test, P = 0.042), TNM stage (Chi-square test, P = 0.015), and lymph node metastasis (Chi-square test, P = 0.011) in NSCLC patients, but not with other factors, including sex and age (Fig. 1f, g, Additional file 1: Table S2). Univariate Cox regression analyses identified histologic grade, lymphatic metastasis, TNM stage, and LINC01234 expression level as prognostic factors. Other clinicopathologic features such as sex and age were not statistically significant prognosis factors. Multivariate Cox regression analyses showed that LINC01234 expression level were independent prognostic factors for NSCLC patients (Additional file 1: Table S3). Kaplan–Meier survival analysis showed that the OS rates were lower for the high vs. low LINC01234 expression groups, which was supported by Kaplan–Meier Plotter analysis (www.​kmplot.​com) [14] (Fig. 1h).

To assess the mechanisms by which LINC01234 promotes the progression of NSCLC, we firstly performed transwell migration and invasion assays. We found that LINC01234 knockdown significantly decreased the migration and invasion ability of A549 and SPC-A1 cells compared with cells expressing si-NC, while overexpression of LINC01234 promoted A549 and SPC-A1 cells migration and invasion (Fig. 2a–c). To validate these results in vivo, we examined the metastatic potential of A549 and SPC-A1 cells stably transfected with empty vector or sh-LINC01234 after injection into nude mice. Consistent with the in vitro analyses, LINC01234 knockdown reduced the number of metastatic lung nodules compared with the control group. H&E staining of excised lung sections confirmed the lower frequency of metastases in the LINC01234-deleted tumors (Fig. 2d, e). However, we did not find metastases in the liver, kidney, intestine, spleen, and other organs (Additional file 1: Figure S1A). Interestingly, expression of the epithelial–mesenchymal transition (EMT) marker E-cadherin was increased, whereas N-cadherin and Vimentin were decreased in LINC01234 downregulated cells (Fig. 2f). In addition, E-cadherin protein levels were also found to be upregulated in A549 and SPC-A1 cells by immunofluorescence assay (Fig. 2g). These in vivo data therefore complement the results of the in vitro functional studies and establish LINC01234 as a regulator of NSCLC metastasis.

To better understand the underlying mechanism of LINC01234 in NSCLC, we examined its distribution in NSCLC cells via subcellular fractionation. Interestingly, qRT-PCR analysis of RNA purified from nuclear and cytoplasmic cellular fractions revealed that LINC01234 RNA was localized both in nucleus and cytoplasm (Fig. 3a). We also performed ISH to detect the distribution of LINC01234 in NSCLC tissues. The results showed that LINC01234 was expressed both in nucleus and cytoplasm in NSCLC tissues (Additional file 1: Figure S1B). Multiple studies have indicated that lncRNAs could regulate gene expression at transcription level by interacting with RNA binding proteins, such as EZH2, SUZ12, and LSD1. lncRNAs could also affect gene expression at post-transcription level by binding with Ago2. RIP assays confirmed the presence of LINC01234 in EZH2, LSD1, and Ago2 immunoprecipitates from A549 and SPC-A1 cell lysates (Fig. 3b). We next performed RNA pull-down assays to identify LINC01234-associated proteins. We found that biotinylated LINC01234 RNA, but not labeled negative control or antisense RNAs, bound to EZH2, LSD1, and Ago2 (Fig. 3c). Collectively, these data demonstrate that LINC01234 binds to EZH2, LSD1, and Ago2 proteins in NSCLC cells.

Emerging evidence suggests that some lncRNAs regulate target gene expression by functioning as ceRNAs, or sponges, for miRNAs, thereby preventing miRNA-mediated regulation of target mRNAs. Interestingly, one of the LINC01234-binding proteins identified above is Ago2, a component of the RNA-induced silencing complex involved in miRNA-mediated repression of mRNAs. Therefore, we investigated the possibility that LINC01234 may also function as a ceRNA. To test this hypothesis, we searched the online bioinformatics databases LncACTdb and miRcode for putative miRNA-binding sites in the LINC01234 sequence. Both databases identified nine putative sites (Fig. 3d), from which we analyzed five with known tumor suppressor functions (miR-27b-3p, miR-340-5p, miR-490-3p, miR-100-5p, and miR-153-5p). Luciferase reporter assays showed that co-transfection with miR-27b-3p and miR-340-5p significantly decreased the luciferase activity of HEK293T cells harboring a LINC01234-driven reporter construct compared with control miRNAs (Fig. 3d). Mutation of the putative miR-27b-3p and miR-340-5p binding sites in LINC01234 abolished their ability to suppress luciferase expression, confirming that the miRNAs specifically interact with LINC01234 (Fig. 3e, f). Consistent with this, Ago2 immunoprecipitates from A549 and SPC-A1 cell lysates were enriched in LINC01234, miR-27b-3p, and miR-340-5p compared with control IgG immunoprecipitates (Fig. 3g). Finally, miR-27b-3p and miR-340-5p levels in A549 and SPC-A1 cells were significantly increased and decreased by LINC01234 depletion and overexpression, respectively (Fig. 3h, i), whereas overexpression of miR-27b-3p or miR-340-5p had no effect on LINC01234 levels (Fig. 3j). These data indicate that LINC01234 acts as a sponge to directly interact with miR-27b-3p and miR-340-5p.

To identify LINC01234-regulated target genes in NSCLC in an unbiased manner, we performed RNA-seq to compare the gene expression profiles of A549 cells expressing LINC01234 siRNA or si-NC (Fig. 4a). To identify target genes regulated by both miRNAs and LINC01234, we performed an integrative analysis of the miRDB database and the gene expression profiles of LINC01234-depleted and control A549 cells. Among the 266 genes significantly downregulated by LINC01234 silencing (log₂fold-change > 1 and P < 0.05), 12 and 10 were target genes of miR-340-5p and miR-27b-3p, respectively. Interestingly, VAV3 was identified as a target gene for both miRNAs (Fig. 4b). We validated the LINC01234, miR-27b-3p, miR-340-5p, and VAV3 network using luciferase reporter assays. We cloned the wild-type 3′-untranslated region (3′UTR) of VAV3 (wt-VAV3) or one carrying mutations in the miR-27b-3p/miR-340-5p-binding sites (mut-VAV3) into the luciferase vector and co-transfected them with control miRNA or miR-27b-3p/miR-340-5p mimics into HEK293T cells. We found that miR-27b-3p and miR-340-5p significantly decreased luciferase activity driven by wt-VAV3 but not by mut-VAV3 (Fig. 4c), indicating that VAV3 mRNA levels are directly regulated by miR-27b-3p/miR-340-5p in NSCLC cells. Consistent with this, VAV3 mRNA levels in NSCLC cells were reduced or elevated by expression of miR-27b-3p/miR-340-5p mimics or inhibitors, respectively (Fig. 4d). Additionally, E-cadherin was increased, whereas Vimentin was decreased, in VAV3 knockdown cells (Fig. 4e). Knockdown of LINC01234 significantly reduced VAV3 mRNA and protein levels in A549 and SPC-A1 cells, which is consistent with the loss of miR-27b-3p and miR-340-5p sponging (Fig. 4f). In luciferase reporter assays, miR-27b-3p and miR-340-5p-mediated suppression of wt-VAV3-driven luciferase activity was partly reversed by co-expression of the wild-type LINC01234 sequence, but not by LINC01234 mutated in the miR-27b-3p/miR-340-5p binding sites (Fig. 4g). Finally, the suppression of VAV3 protein levels induced by LINC01234 silencing was effectively reversed by co-expression of miR-27b-3p/miR-340-5p inhibitors in NSCLC cells (Fig. 4h). Moreover, the expression levels of miR-27b-3p/miR-340-5p in NSCLC tissues were negatively related with VAV3 expression, while LINC01234 expression was positively related to VAV3 expression. Additionally, miR-27b-3p/miR-340-5p were negatively correlated with LINC01234 (Fig. 4i). Collectively, these data indicate that LINC01234 modulates the expression of VAV3 by acting as a ceRNA for miR-27b-3p and miR-340-5p in NSCLC cells.

To determine whether miR-27b-3p and miR-340-5p have tumor suppressor activities, we transfected A549 and SPC-A1 cells with miRNA mimics or inhibitors and examined the effects on cell function (Fig. 5a). Expression of miR-27b-3p and miR-340-5p mimics significantly reduced cell invasion (Fig. 5b). Moreover, co-transfection with miR-27b-3p/miR-340-5p inhibitors partially reversed the inhibition of cell invasion induced by LINC01234 knockdown (Fig. 5c). Next, we silenced VAV3 expression in A549 and SPC-A1 cells by transfection with VAV3 siRNA, which was confirmed to be effective by qRT-PCR (Fig. 5d). Inhibition of VAV3 significantly inhibited NSCLC cell invasion (Fig. 5e). In addition, co-transfection with VAV3 siRNA partially reversed the promotion of A549 and SPC-A1 cell invasion induced by transfection with miR-27b-3p/miR-340-5p inhibitors (Fig. 5f). The increasing in VAV3 protein levels in miR-27b-3p/miR-340-5p inhibitor transfected A549 and SPC-A1 cells was also partly rescued by co-transfection with VAV3 siRNA (Fig. 5g). Collectively, these results suggest that miR-27b-3p, miR-340-5p, and VAV3 are involved in mediating the functional effects of LINC01234 in NSCLC cells.

According to the RNA-seq results, we selected eight genes with altered expression involved in cancer cell migration and invasion for further study (Fig. 4a). The identified gene expression changes were validated in A549 and SPC-A1 cells by qRT-PCR. BTG2 is of particular interest because of its remarkable expression fold change upon LINC01234 knockdown (Fig. 6a). Western blot analysis showed that knockdown of LINC01234 significantly increased BTG2 protein levels in NSCLC cells (Fig. 6b), which is consistent with the qRT-PCR results (Fig. 6a) and suggests that BTG2 may play a role in mediating the functional effects of LINC01234. Our findings indicate that, in addition to acting as a ceRNA in the cytoplasm, LINC01234 may also act via EZH2 and LSD1 to regulate transcription of target genes (such as BTG2) in the nucleus. To test this hypothesis, we examined the behavior of A549 and SPCA1 cells expressing EZH2 or LSD1 siRNA. Interestingly, siRNA-mediated knockdown of the LINC01234-binding proteins EZH2 and LSD1 induced BTG2 expression at both the mRNA and protein levels (Fig. 6c, d). Importantly, ChIP assays showed that EZH2 and LSD1 bound directly to the promoter regions of BTG2 and induced deposition of the trimethylated histone 4 lysine 37 (H3K27) and demethylated H3K4 marks (Fig. 6e), which are associated with transcriptional repression. These data suggest that LINC01234 acts, at least in part, by interacting with EZH2 and LSD1, which, in turn, repress transcription of BTG2.

Finally, we examined the potential role of BTG2 as a tumor suppressor in NSCLC. In support of this possibility, BTG2 protein was expressed at lower level in tumor tissues compared with normal tissues (Fig. 6f). Kaplan–Meier analysis of patient survival showed that BTG2 expression positively correlated with prognosis, which was supported by Kaplan–Meier Plotter analysis (www.​kmplot.​com) [14] (Fig. 6g). We analyzed the correlation between BTG2 and LINC01234 expression, and found that BTG2 expression in paired NSCLC tissues and normal lung tissues was negatively correlated with LINC01234 expression (Fig. 6h).

To better understand the biological function of BTG2, we performed gain-of-function assays. A549 and SPC-A1 cells were transfected with BTG2 overexpression vectors or siRNAs, and the mRNA and protein levels were confirmed to be significantly upregulated or downregulated by qRT-PCR and Western blotting (Fig. 7a, c). Importantly, overexpression of BTG2 led to impairment and invasion by A549 and SPC-A1 cells (Fig. 7b). Interestingly, expression of EMT marker E-cadherin was increased, whereas Vimentin was decreased, in BTG2 overexpressed cells (Fig. 7c). We next performed rescue experiments to determine whether LINC01234-mediated regulation of invasion is dependent on repression of BTG2. A549 cells were cotransfected with si-LINC01234 and si-BTG2, and SPC-A1 cells were cotransfected with LINC01234 and BTG2 vector. Indeed, co-transfection of cells with si-BTG2 partially rescued si-LINC01234-mediated impairment of invasion and EMT process. Conversely, BTG2 overexpression partially rescued the LINC01234-induced promotion of cell invasion and EMT progress (Fig. 7d, e). Taken together, these findings demonstrate that LINC01234 affects NSCLC cell development and progression through the epigenetic repression of BTG2, at least in part.