LINC02870 facilitates SNAIL translation to promote hepatocellular carcinoma progression

Exploring the roles of long noncoding RNAs (lncRNAs) in tumorigenesis and metastasis could contribute to the recognition of novel diagnostic and therapeutic targets. LINC02870 is a novel lncRNA, whose role in tumors has not been reported. Herein, we focused on the function and mechanism of LINC02870 in human hepatocellular carcinoma (HCC). We first carried out a pan-cancer study of LINC02870 expression and its relationship to prognosis, and LINC02870 was determined to be a possible oncogene in HCC. Upregulated expressions of LINC02870 were also found in our HCC samples compared to the para-tumor samples. Moreover, overexpression of LINC02870 promoted the growth, migration, and invasion of HCC cells. Subsequently, binding proteins of LINC02870 were identified by a number of in silico analyses, including correlation analysis, signaling network analysis, and survival analysis. Intriguingly, the most promising binding protein of LINC02870 was predicted and confirmed to be eukaryotic translation initiation factor 4 gamma 1 (EIF4G1), an important component of the eukaryotic translation initiation factor 4F complex that initiates cap-dependent translation. Further investigation showed that LINC02870 increased the translation of SNAIL to induce malignant phenotypes in HCC cells. Additionally, HCC patients with higher expression levels of LINC02870 and EIF4G1 had shorter survival times than those with lower expression levels. Thus, our findings suggested that LINC02870 induced SNAIL translation and correlated with poor prognosis and tumor progression in HCC.


Introduction
Primary liver cancer is one of the most common malignancies worldwide; hepatocellular carcinoma (HCC) accounts for approximately 90% of primary liver cancer [1]. The most common underlying cause of HCC is mostly as a consequence of hepatitis B or C virus (HBV/HCV) infection, cirrhosis, and fatty liver disease [2]. Although great improvements have been achieved in effective treatments for HCC in recent years, the five-year survival rate of HCC is still less than 5% [3], mainly because of the high recurrence rate and metastasis after surgical resection [4]. Therefore, clarification of the molecular mechanisms underlying HCC metastasis is critical to identify novel therapeutic targets and development alternative therapeutic strategies for HCC patients.
Previously, many studies have focused on the roles of dysregulated protein-coding genes in tumorigenesis and tumor progression [5]. However, with the help of high-throughput techniques, protein-coding genes have been revealed to Mengya Guo and Hao Zhuang contributed equally to this work. 1 3 occupy less than 2% of the mammalian genome, and the majority of genes are detectably transcribed with no or low protein-coding potential [6]. Long noncoding RNAs (lncR-NAs) have more than 200 nucleotides and are one type of those non-protein-coding transcripts [7]. Emerging evidence has indicated that abnormally expressed lncRNAs exert significant and comprehensive impacts on HCC progression via different molecular mechanisms, including the regulation of gene expression [8][9][10]. For instance, lncRNA-ICR interacts with ICAM-1 mRNA to form an RNA duplex, resulting in increased stability of ICAM-1 mRNA [11]; lncRNA-MCM3AP antisense RNA 1 competitively binds to miR-194-5p to rescue the miR-194-5p-mediated reduction in forkhead box A1 expression at the mRNA level, which contributes to increased proliferation of HCC cells [12]. LINC00665 directly interacts with double-stranded RNA (dsRNA)-activated protein kinase (PKR) to prevent proteasome-mediated degradation of PKR protein and subsequently activates the NF-κB pathway in HCC cells [13]. However, lncRNA biology and its potential clinical relevance remain underdeveloped.
Herein, we focus on a little-known lncRNA, LINC02870, noted in our previously reported RNA-sequencing data from HCC cells [14]. We show that LINC02870 has high expression in TCGA HCC samples and our clinical samples, which is related to adverse prognosis. Moreover, overexpression of LINC02870 enhances the proliferative and metastatic capacities of HCC cells. We further determine that eukaryotic translation initiation factor 4 gamma 1 (EIF4G1), an important component of the eukaryotic translation initiation factor 4F (EIF4F) complex that initiates cap-dependent translation, is the interacting protein of LINC02870. Furthermore, LINC02870 increases the translation of SNAIL to induce malignant phenotypes in HCC cells. Our findings suggest that LINC02870 induces SNAIL translation and correlates with poor prognosis and tumor progression in HCC.

GEPIA database analysis
GEPIA (http:// gepia. cancer-pku. cn/) is a web server for cancer and normal profiling and interactive analyses, which is based on TCGA and Genotype-Tissue Expression (GETx) [15]. GEPIA was employed to determine LINC02870 and EIF4G1 expressions in various types of human cancer. P value < 0.05 was considered statistically significant.

Kaplan-Meier plotter analysis
Kaplan-Meier plotter (http:// kmplot. com/ analy sis/), an online database capable of accessing the effects of genes on survival in more than 20 cancer types including HCC [16], was employed to conduct survival analysis for LINC02870 and EIF4G1 in various types of human cancer. Logrank P value < 0.05 was statistically significant.

Plasmids, siRNA, and transfection
The LINC02870-expressing plasmid was synthesized by GENEWIZ and siRNAs against LINC02870 were synthesized by RIBOBIO. The plasmids and siRNAs were transfected into HCC cells with LipoGeneTM 2000 PLUS Transfection Reagent (US EVERBRIGHT ® INC, #L7003). Overexpression and knockdown efficiencies of LINC02870 were determined by qPCR. Sequences of the genes coding siRNAs for LINC02870 knockdown are listed in Supplemental Table S1.

Protein extraction and Western blotting
Cell lysates were prepared by incubation for 5 min on ice with lysis buffer (RIPA buffer with 1% SDS and 1 × protease inhibitor cocktail) and sonication for 3 × 8 s pulses in a Soniprep 150 MSE, 30% power. Cell lysates were centrifuged at 4 °C, 12,000 g for 15 min to remove debris. The protein concentration of the supernatants was determined by Pierce BCA Protein Assay Kit (Thermo Scientific, #23227). Equal amounts of proteins were separated in SDS-PAGE by electrophoresis and transferred to a PVDF membrane (Millipore, #ISEQ00010). The membrane was blocked with 5% skim milk (Solarbio, #D8340) at room temperature and incubated with indicated antibodies overnight at 4 °C. And then, the membrane was washed by TBST (1 × TBS containing 0.05% Tween-20). The membrane was incubated with a corresponding secondary antibody according to the primary antibody at room temperature for 1 h and washed as mentioned above. The blots were visualized using Super ECL Prime (US EVERBRIGHT ® INC, #S6008) by G-box (Syngene, Chemi XT). Intensity of Western blotting band was analyzed by ImageJ 1.46 software, and β-actin was used as a loading control to normalize the amount of protein.

Protein stability assay
Cells were treated with 20 μg/mL of cycloheximide to block protein synthesis and were treated with 10 μM of MG132 to inhibit proteasome function. Cells were incubated with cycloheximide alone for different times (0, 30, 60, 90, and 150 min) or with both cycloheximide and MG132 for 6 h, and DMSO was used as a control of the latter.

RNA isolation and qPCR
Total RNAs of cells and clinical samples were extracted with Trizol, and cDNAs were synthesized from total RNA using One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen, #AT311). Quantitative mRNA analysis was conducted using AugeGreenTM qPCR Master Mix (US EVERBRIGHT ® INC, #S2008) on a 7500 Real-Time PCR System (ABI, Foster City, CA, USA) according to the manufacture's instruction. 18 s-rRNA was severed as the reference gene to normalize the mean Ct of each sample. The qPCR primer sequences are listed in Supplemental Table S2.

Cytosolic and nuclear fractionation
Cytoplasmic and nuclear RNA were isolated and purified with Cytoplasmic & Nuclear RNA Purification Kit (Norgen, #21000) according to the manufacturer's instructions. GAPDH was used as the cytoplasmic endogenous control, and pre-rRNA45S as the nuclear endogenous control.

In vitro transcription
In vitro transcription templates were obtained by PCR and the primers of LINC02870 containing the T7 promoter sequence. The target products were verified with 2% agarose gel electrophoresis and cut off for purification with an agarose Gel DNA Extraction Kit Ver.4.0 (TaKaRa, #9762). Sense and antisense chains of LINC02870 were transcribed in vitro with a MAGEscript T7 Transcription kit (Invitrogen, #AM1333) and labeled with biotin-16-UTP (APExBIO, #B8154). According to the manufacture's instruction, the synthesized RNAs were purified by a MAGEclear kit (Invitrogen, #AM1908).

RNA fluorescent in situ hybridization (FISH) assay
Cells were maintained in 12-well plates containing sterile coverslips for 48 h, fixed with 37% formaldehyde in PBS for 10 min at room temperature, and permeabilized with 70% ethanol for 1 h at 4 °C. Next, 1 pmol of antisense probes were dissolved in the hybridization buffer for slides hybridized. After 4 h of hybridization at 37 °C, slides were washed in 2 × SSC for 30 min at 37 °C. Then the slides were hybridized with streptavidin Alexa 546 (Invitrogen, #A20183) and washed as mentioned above. The slides were counterstained with DAPI, followed by mounting with Fluorescence Mounting Medium (DAKO, #S3023) and photographed by confocal laser scanning microscopy (Olympus, Tokyo, Japan). Primer sequences for the sense and antisense LINC02870 are listed in Supplemental Table S3.

RNA pull-down
Cell lysate was extracted by lysis buffer (40 mM Tris, 1% Triton X-100, 120 mM NaCl, 1 mM NaF, and 1 mM Na 3 VO 4 ) containing 1 × protease inhibitor cocktail (APE × BIO, #K1007) and 1 U/mL RNase inhibitor (TAKARA, #2313A). Then, 40 μL of MyOneTM Streptavidin C1 Dynabeads (Invitrogen, #65801D) were washed and incubated with 4 μg of biotinylated sense or antisense LINC02870 RNA in a binding buffer (0.01% Tween-20, 300 mM NaCl, and 50 mM Na 3 PO 4 , pH = 8.0) at 4 °C for 3 h. Next, the beads were incubated with cell lysates which protein concentrations were measured with a BCA assay at 4 °C for 4 h. After being washed ten times with lysis buffer, the beads were mixed with a loading buffer for the SDS-PAGE sample, and analyzed by Western blotting.

RNA immunoprecipitation
Cell lysate was extracted according to the mentioned in RNA pull-down assay. Same amounts of interested proteins were quantified with BCA assay. Enriched RNAs were isolated using protein A/G magnetic beads with the corresponding primary antibody, and IgG antibody was simultaneously assayed as the negative control. Co-precipitation RNAs were extracted by Trizol. The amount of LINC02870 in the eluate was determined by qPCR.
Gradients were divided into fifteen fractions. Each fraction was 1 ml, and absorbance at 260 nm was measured for each fraction. RNA isolation and quantitation were conducted as mentioned above. The mRNA distribution across the polysome fractions was graphically presented as the percentage.

Cell proliferation assays
The growth curve assay was conducted according to Kratzat's method [18]. Cells were evenly plated at a density of 2 × 10 5 cells/well in 6-well plates. At each time point, cells were counted and replaced at the same amount as the start point.
Cell counting Kit-8 (US EVERBRIGHT® INC, #C6005) was used according to the product instructions. In simple terms, cells were seeded in 96-well plates. 90 μL of complete culture medium and 10 μL of CCK-8 reagent were added to each well. The absorbance at 450 nm was measured after 3 h of incubation at 37 °C.
For the colony formation assay, 500, 800, and 1000 cells were evenly planted in 6-well plates for 14 days, and the culture medium was replaced every three or four days. Then, cell colonies were stained with 0.01% crystal violet and counted manually.

Wound healing, migration, and invasion assays
In wound healing assay, cells were planted evenly in 6-well plates and cultured until 80%-90% confluence. The cells were scratched with a pipette tip in the middle of the well, washed gently with PBS to remove the exfoliated cells, and incubated in the culture medium with 1% FBS. The wound widths were measured microscopically at different time points and photographed at 0 and 48 h, respectively.
Invasion and migration assays were performed using transwell chambers containing polycarbonate membranes with 8 μm pores (Corning Costar, #3422) paved with or without Matrigel matrix (BD Biosciences, #346234). A total of 500 μL cell suspension in the DMEM (1 × 10 6 cells/mL) was added to the upper chamber and 750 μL DMEM with 20% FBS was added to the lower chamber.

Prediction and analysis of RNA-binding proteins
RNA-binding proteins (RBPs) were predicted in RNAct (https:// rnact. crg. eu/), a website supporting Protein-RNA interaction predictions for model organisms with supporting experimental data [17]. For network construction, the protein-protein interaction (PPI) information was obtained from String 11.5 (https:// string-db. org/), a database for protein-protein interaction networks and functional enrichment analysis [18], and it was imported into Cytoscape 3.8.2 for network construction. Interactions between proteins were shown as nodes and edges. The node size represented the interaction degree, and the color represented different biological functions obtained from a protein information hub, UniProt (https:// www. unipr ot. org/). 35 paired fresh HCC tissues and para-tumor tissues were collected from Henan Cancer Hospital affiliated to Zhengzhou University (Zhengzhou, China) with the informed consent of the patients and ethics approval from the Ethics Committee (No. 2016CT054) of Henan Cancer Hospital. The para-tumor tissues were 2 cm away from the HCC tissue margin. All tissue samples were pathologically confirmed. Total RNAs of these tissues were analyzed for RT-qPCR.

Statistical analysis
Wilcoxon rank-sum test was used for pan-cancer analysis, Overall survival was calculated by Kaplan-Meier survival analysis and logrank test. Two-sided Student's t-test was used for two group comparison, and one-way analysis of variance (ANOVA) was performed among multiple group comparison. Growth curve, CCK-8, and wound healing data were analyzed by ANOVA for repeated measures. GraphPad 9.0 was used for statistical analyses and general plots. P values less than 0.05 were considered statistically significant.

Pan-cancer analysis of LINC02870 expression
To explore the potential function of LINC02870 in tumorigenesis, we first observed its expression pattern in twenty of the most frequently diagnosed forms of human cancer from the TCGA database. LINC02870 was found to be significantly upregulated in tissues of eleven different types of cancer compared with the corresponding normal tissues, namely, LIHC (P < 0.001), BRCA (P < 0.001), COAD (P < 0.001), ESCA (P < 0.001), HNSC (P < 0.001), LUAD (P < 0.001), LUSC (P < 0.001), READ (P < 0.001), STAD (P < 0.001), THCA (P < 0.001), and UCEC (P < 0.001). However, LINC02870 was expressed at lower levels in GMB (P < 0.05) and PRAD (P < 0.001). In the seven other forms of cancer, the expression levels of LINC02870 did not vary considerably (Fig. 1A). Next, we analyzed the expression of LINC02870 in the eleven abovementioned cancer types using the GEPIA database. LINC02870 expression was confirmed to be elevated in LIHC, BRCA, COAD, ESCA, HNSC, LUAD, LUSC, READ, STAD, THCA, and UCEC ( Fig. 1B-L). Moreover, using survival rate analysis, the upregulated expression levels of LINC02870 showed unfavorable prognosis in LIHC, KIRC, KIRP, LUAD, PAAD, PCPG, and UCEC ( Fig. 2A-G). By combining pan-cancer analysis and overall survival (OS), LINC02870 was identified as a cancer-enriched prognostic biomarker. Then, LINC02870 expression levels were determined in 35 paired HCC clinical samples by quantitative PCR (qPCR). As expected, LINC02870 expression was higher in HCC tissues than in the corresponding para-tumor tissues (Fig. 2H). Thus, we chose to focus our subsequent investigation on the role of LINC02870 in HCC.

LINC02870 enhances the proliferative and metastatic abilities of HCC cells
To explore the potential biological functions of LINC02870 in HCC cells, we overexpressed LINC02870 in PLC and HepG2 cells. The overexpression efficiency was determined using qPCR (Fig. 3A). The growth curve assay showed that LINC02870 overexpression promoted cell proliferation (Fig. 3B). In accordance with this result, the CCK-8 and clonogenic assays also showed the effects of LINC02870 overexpression on growth acceleration ( Fig. 3C and D). Moreover, upregulation of LINC02870 resulted in a markedly increased migratory and invasive ability in wound-healing and Transwell assays with and without Matrigel ( Fig. 3E and F). We then knocked down LINC02870 expression with the transfection of LINC02870-specific siRNAs in HCCLM3 and Huh7 cells (Fig. 4A). The cells showed a dramatic decrease in cell proliferation with the downregulation of LINC02870 by growth curve, CCK-8, and colony formation assays (Fig. 4B-D). Additionally, the migration and invasion of LINC02870-knockdown cells were significantly suppressed compared with those of vector cells (Fig. 4E and F). Collectively, these results showed that LINC02870 promotes the proliferation, migration, and invasion in HCC cells.

Prediction and analysis of interacting proteins of LINC02870
Previous molecular mechanistic studies of lncRNAs have found that lncRNAs function depending on their interacting proteins and cellular localization [19]. We first predicted proteins that could potentially bind to LINC02870 in RNAct (https:// rnact. crg. eu/), a website supporting protein-RNA interaction predictions, and found 13 known RNA-binding proteins (RBPs) and 8 predicted RBPs (Fig. 5A). To construct the protein interactome network of LINC0287, the LINC02870 RBPs were analyzed against the STRING protein-protein interaction (PPI) database, and the nodes were grouped according to their known functions based on GO analysis. As shown in Fig. 5B, a LINC02870-protein regulatory network was mapped using Cytoscape software. These RBPs were classified into five subclusters, including translation, mRNA processing, transcription, DNA related, and other (Fig. 5B). These results showed the potential roles of LINC02870 in diverse biological processes.
We next determined the cellular localization of LINC02870. According to the results of the RNA fluorescent in situ hybridization (FISH) assay and expression measurement from cytoplasmic and nuclear fractions, LINC02870 was predominantly localized in the cytoplasm (Fig. 5C  and D). Given that the cytoplasmic RBPs colocalized with LINC02870, they were chosen to further determine their interaction probabilities with LINC02870 via the RNA-protein interaction prediction website (http:// pridb. gdcb. iasta te. edu/ RPISeq/) [20]. Interestingly, the scores of the RF classifier and SVM classifier for EIF4G1 were both greater than 0.5, suggesting that EIF4G1 has a promising probability of interacting with LINC02870 (Fig. 5E). We then constructed in vitro-transcribed biotin-labeled LINC02870 sense and antisense probes. RNA pulldown combined with Western blotting showed that the LINC02870 sense probes achieved great enrichment of EIF4G1 than the LINC02870 antisense probes (Fig. 5F). Reciprocally, LINC02870 enrichment with the EIF4G1 antibody was more than tenfold greater than that with the IgG control (Fig. 5G). These results suggested that LINC02870 might bind to EIF4G1 in HCC cells.
Additionally, we carried out a pan-cancer study for EIF4G1 expression. The expression levels of EIF4G1 were remarkably increased in 10 different types of cancer, including LIHC, BRCA, COAD, GBM, HNSC, LUAD, LUSC, READ, STAD, and UCEC (Fig. 6A-J). Moreover, Kaplan-Meier survival analyses revealed that high expressions of EIF4G1 were associated with poor prognosis for cancer patients in LIHC, HNSC, KIRP, LUAD, PAAD, SARC, and UCEC (Supplemental Fig. S1A-G). High expression of either LINC02870 or EIF4G1 was associated with poor prognosis in LIHC, KIRP, LUAD, PAAD, and UCSC patients ( Fig. 2A-G), we sought to determine whether cancer patients with upregulated expressions of LINC02870 and EIF4G1 had shorter survival times. Indeed, short survival times were found in patients with LIHC (Fig. 6K), KIRP, LUAD, PAAD, and UCEC (Supplemental Fig. S2A-D), which showed that LINC02870 and EIF4G1 might work together to promote cancer progression. Taken together, these findings suggested that EIF4G1 is the binding protein of LINC02870, and we chose to investigate the association of LINC02870 and EIF4G1 in HCC cells.

LINC02870 induces SNAIL translation
Epithelial-to-mesenchymal transition (EMT) is essential for cancer cell metastasis [21], and we anticipated that LINC02870 might influence EMT. We detected the expressions of the EMT markers E-cadherin and SNAIL [22]. As shown in Fig. 7A, the Western blotting results showed that E-cadherin expression was significantly suppressed, while SNAIL expression was significantly increased in LINC02870-overexpressing PLC cells compared to control cells. Whereas, LINC02870 knockdown increased E-cadherin expression and decreased SNAIL expressions in HCCLM3 cells. Given that LINC02870 interacts with EIF4G1 ( Fig. 5F and G), and that EIF4G1 is an important component of the EIF4F complex that initiates cap-dependent translation [23], we hypothesized that LINC02870 regulates the cap-dependent translation of SNAIL. Polysome analysis was employed to detect the distribution of SNAIL mRNA within the different sucrose gradients to assess the translational status of SNAIL. Interestingly, quantities of SNAIL mRNAs were detected in the high-molecularweight polysomes with LINC02870 overexpression compared to the vector control, indicating higher translational efficiency of SNAIL mRNA in LINC02870-overexpressing PLC cells. To evaluate the specificity of increased SNAIL translation, GAPDH mRNA was examined as an irrelevant RNA control in the polysome assay (Fig. 7B). Next, when cells were treated with rapamycin, an inhibitor that interferes with translation initiation [24,25], the increase in the 1 3 protein level of SNAIL with LINC02870 overexpression was abolished (Fig. 7C). Moreover, when a protein elongation inhibitor, cycloheximide (CHX), was used to treat cells, no significant difference in the steady-state level of SNAIL was observed between LINC02870-overexpressing cells and control cells (Fig. 7D). Cells were treated with MG132, a proteasome inhibitor. Similarly, the increased SNAIL level with LINC02870 overexpression was not rescued by the addition A Proteins predicted as LINC02870 RBPs using RNAct. Red color for the cytoplasmic protein, B PPI network of LINC02870 and its predicted RBPs, C The cellular localization of LINC02870 in PLC cells by RISH assay, D Expression level of LINC02870 in the cytoplasmic and nuclear fractions of PLC cells by qPCR. Pre-rRNA45s was used as nuclear marker and GAPDH as cytoplasmic marker, E Interaction probability scores between LINC02870 and its predicted cytoplasmic RBPs from RPISeq database, F In vitro transcribed biotinylated LINC02870 was incubated with HepG2 lysate was examined by RNA pulldown assay. Levels of EIF4G1 in the pulldown were determined by western blotting. Biotinylated LINC02870 antisense was used as control, G The enrichment of LINC02870 in EIF4G1 immunoprecipitates   (Fig. 7E), suggesting that the overexpression of LINC02870 did not influence the protein stability of SNAIL. Finally, we determined the mRNA expression levels of SNAIL. The SNAIL mRNA levels showed little change with LINC02870 overexpression (Fig. 7F). Together, these results indicated that LINC02870 could increase the cap-dependent translation of SNAIL. In addition, the protein expression levels of Bcl-2, Bcl-xL, and cyclin D1 whose translations are known to be regulated by EIF4G1 were also enhanced with overexpression of LINC02870 (Supplemental Fig. S3A), further suggesting the regulatory role of LINC02870 in EIF4G1 function [26][27][28][29].

LINC02870 enhances the metastatic ability of HCC cells through an increase in SNAIL translation
To test whether LINC02870 exerts its effects on metastatic phenotypes by regulating SNAIL cap-dependent translation, loss-of-function experiments were carried out. First, we treated cells with rapamycin. As shown by the Transwell assays without/with Matrigel, rapamycin treatment abolished the function of cell migration and invasion in LINC02870-overexpressing cells (Fig. 8A). The woundhealing assay also demonstrated a similar trend (Fig. 8B).
Moreover, knockdown of SNAIL reduced the increase in cell migration and invasiveness induced by LINC02870 overexpression ( Fig. 8C and D, Supplemental Fig. S3B). Taken together, these results suggested that LINC02870 promotes migration and invasion by increasing the cap-dependent translation of SNAIL.

Discussion
Noncoding genes account for the vast majority of RNA transcripts [25]. Increasing evidence has suggested that lncRNAs have vital functions in multiple events in human cancers, such as proliferation, metastasis, angiogenesis, drug resistance, microenvironment regulation, and self-renewal of cancer cells [30,31]. Some lncRNAs have been reported to be upregulated in HCC tissues. For example, lncRNA LNC-HUR1 is upregulated to induce cell proliferation by blocking p53 activity [32]. Therefore, the functions of lncR-NAs in HCC development require further elucidation. In the current study, LINC02870 was indicated to be a cancerenriched prognostic biomarker and found to be involved in HCC. LINC02870 overexpression was strongly correlated with poor outcomes in HCC patients. Moreover, LINC02870 was suggested to interact with EIF4G1 and to induce the cap-dependent translation of SNAIL, resulting in an increase in the cell proliferation and metastasis of HCC cells. Thus, LINC02870 might be an oncogene in HCC.
LncRNAs, along with their biomacromolecules or micro-RNA partners, regulate gene expression at the transcriptional, posttranscriptional, translational, and posttranslational levels [33][34][35]. In addition, lncRNAs in the cytoplasm dominantly modulate the stability and translation of mRNAs [36,37]. LncRNAs can influence each step of translation by regulating translation factors. The known mechanisms include acting as microRNA sponges to prevent the degradation of translation factors or directly binding to either ribonucleoproteins (RNPs) or components of the EIF4F complex and then positively or negatively regulating protein translation [38]. The EIF4F complex binds to a 5'-terminal mRNA to activate it, resulting in protein translation, and consists of EIF4E, EIF4G, and EIF4A. EIF4G functions as a scaffold protein in the EIF4F complex [23,39]. The components of the EIF4F complex are associated with lncRNAs to regulate protein translation. For example, the binding of EIF4G by EIF4E could be interfered by the competitive binding of EIF4E by lncRNA RP1, thus contributing to decreased translation of p27kip1 in breast cancer cells [40]. The expression level of the lncRNA UCA1 is positively correlated with EIF4G1 expression in prostate cancer, since UCA1 induces EIF4G1 expression by sponging miR-331-3p [41]. Here, upregulation of EIF4G1 was identified to be associated with poor prognosis in cancer patients in our pan-cancer analysis, consistent with previous reports [42]. We also showed that EIF4G1 bound to LINC02870, resulting in induced SNAIL translation. The mTOR1 complex functionally regulates cap-dependent translation, and rapamycin is a specific inhibitor of mTOR pathway. Here, we observed that rapamycin treatment impaired LINC02870induced aggressiveness in HCC cells. In addition, EIF4G1 has a complex role in translation initiation. For instance, EIF4G1 can regulate the translation of a subset of proteins important for cell proliferation and apoptosis, such as cyclin D1, Bcl-2, and Bcl-xL [26][27][28][29]. We also showed that the protein expressions of Bcl-2, Bcl-xL, and cyclin D1 were enhanced with the overexpression of LINC02870, which was consistent with previous reports and further supported the regulatory role of LINC02870 in EIF4G1 function. To the best of our knowledge, there are no reports on the function of LINC02870 and its potential mechanism. Further investigation of LINC02870/EIF4G1-mediated regulation of translation is still needed for clarification.
EMT is essential for cell escape and colonization in cancer metastasis [43]. EMT induces increased expression of mesenchymal markers and decreased expression of epithelial markers in cells. Downregulation of E-cadherin is a key event during the EMT process. SNAIL, one of the EMT-inducing transcription factors, is important for E-cadherin repression and EMT activation [44][45][46]. The switch from E-cadherin to N-cadherin expression might be induced by the SNAIL-activated signaling pathway in the progression of breast cancer [44], colorectal cancer [45], and melanoma [46]. SNAIL can bind to the promoter region of E-cadherin and simultaneously increase the transcription of mesenchymal markers, such as N-cadherin [47]. In HCC cells, the expression of SNAIL can be regulated by lncRNAs [48,49]. For example, lncRNA-MUF (mesenchymal stem cell upregulated factor) acts as a ceRNA for miR-34a to prevent SNAIL downregulation, resulting in EMT in SMMC-7721 cells [48]. Our findings on the regulation of SNAIL expression are in accordance with previous studies in HCC. LINC02870 increased SNAIL expression at the translational level. Moreover, transient transfection of SNAIL siRNAs impaired the LINC02870-induced metastatic phenotypes of HCC cells. Therefore, LINC02870 is an oncogene in HCC cells, based on its regulatory role in SNAIL expression and subsequent EMT activation. Further studies should explore whether LINC02870 contributes to the progression of other cancers, and whether its expression can be induced under other conditions, such as DNA damage.
Collectively, our findings showed the role of LINC02870, a cancer-enriched prognostic biomarker and an oncogene, in promoting cell metastasis through increased cap-dependent translation of SNAIL in HCC cells. Considering the tumor-and tissue-specific expression characteristics of lncRNAs, more details on the role of LINC02870 in HCC and other types of cancer are needed for further clarification. Our findings provide insight for better understanding the mechanism underlying the functions of LINC02870 and provide a potential novel therapeutic strategy for HCC.