FOXO3-induced lncRNA LOC554202 contributes to hepatocellular carcinoma progression via the miR-485-5p/BSG axis

Long non-coding RNAs (LncRNAs) have played very important roles in the malignancy behaviors of hepatocellular carcinoma (HCC). LncRNA LOC554202 (LOC554202) was a newly identified tumor-related lncRNA. However, its expression and function in HCC remained unknown. In this study, we firstly reported that LOC554202 expression was distinctly upregulated in HCC specimens and cell lines. Clinical assays indicated that increased LOC554202 expression had a diagnostic value for HCC patients and was positively associated with advanced stages and poor clinical prognosis. Additionally, forkhead box O3(FOXO3) could bind directly to the LOC554202 promoter region and activate its transcription. Functionally, we observed that knockdown of LOC554202 suppressed the proliferation, migration, invasion, and epithelial–mesenchymal transition (EMT) progress of HCC cells, and promoted apoptosis. Mechanistically, LOC554202 competitively bound to miR-485-5p and prevented the suppressive effects of miR-485-5p on its target gene basigin (BSG), which finally led to HCC metastasis, EMT, and docetaxel chemoresistance. Our data demonstrated that FOXO3-induced LOC554202 contributed to HCC progression by upregulating BSG via competitively binding to miR-485-5p, which suggested that the regulation of the FOXO3/LOC554202/miR-485-5p/BSG axis may have beneficial effects in the treatment of HCC.


Introduction
Hepatocellular carcinoma (HCC) acts as a common malignant disease and the major reason for death resulted from cancer worldwide with its incidence rapidly increasing in recent years [1]. It was reported that the number of new HCC cases and HCC death in the People's Republic of China (China) occupied 50% of the total in the world [2].
A lot of risk factors can lead to HCC, particularly the hepatitis B and C viruses as well as liver cirrhosis [3]. Despite progress in treatment protocols, including precious hepatectomy and liver transplantation, a 5-year overall survival (OS) for HCC is <5% because of its high recurrence and metastasis rate [4,5]. Under such background, researchers are suggested to figure out the underlying molecular mechanisms of HCC and seek for new biomarkers to screen and diagnose HCC in an early stage.
In recent studies, different genome-wide methods including the ENCODE project have been combined, finding the transcription of most mammalian genomes, with only about 1.5% representing protein-coding genes [6]. Long non-coding RNAs (lncRNAs) belong to a class of transcribed RNA with a length of over 200 nucleotides, which distinguishes them from small non-coding RNAs, such as miRNA, tRNA, and rRNA [7]. Due to the lack of protein-coding sequences, ncRNAs are the transcripts without coding potential. However, more and more evidence has revealed that lncRNAs participate in the regulatory processes of chromatin modification, transcriptional interference, transcriptional activation, selective splicing, and regulation of tumor-related gene activation, so as to regulate gene expression at epigenetic, transcriptional, or posttranscriptional levels [8,9]. Moreover, growing studies have demonstrated that lncRNAs, due to the special profiles in different human cancers, stand for the progress of diseases and can be used to predict the diagnosis and prognosis [10,11]. However, the functions of a large number of lncRNAs were still unclear in HCC.
Although more and more dysregulated lncRNAs have been identified in various tumors, the potential mechanisms remained unclear. In recent years, growing evidence suggested that transcription factors (TFs) could activate the transcription of lncRNAs [12,13]. Forkhead box O3 (FOXO3) belongs to the O subclass of the forkhead family of transcription factors and has been reported to exhibit multiple effects in tumors [14,15]. In HCC, FOXO3 has been reported to be distinctly upregulated and promote tumor progression [16,17]. However, whether FOXO3 could regulate the expression of lncRNAs was rarely reported.
LncRNA LOC554202 (LOC554202), located on 9p21.3, was a recently identified lncRNA related to tumors with a length of 2166 nt. In recent years, the distinct dysregulation of LOC554202 and its tumor-related effects have been reported in several tumors, such as breast cancer, gastric carcinoma, and colorectal carcinoma [18][19][20]. In gastric carcinoma, LOC554202 was shown to promote the proliferation and migration of gastric carcinoma cells via regulating p21 and E-cadherin [19]. In lung cancer and laryngeal squamous cell carcinoma, LOC554202 was also shown to serve as a tumor promoter [21,22]. Nevertheless, no studies have reported its expression and biological function in HCC.

Patients and specimens
We collected 137 primary HCC specimens and corresponding non-tumor specimens in total from the Xianyang Central Hospital for RT-qPCR analysis between July 2014 and March 2017. Patients' ages ranged from 28 to 69 years, and the average age was (49.32 ± 12.51) years. All patients recruited in this study did not receive chemotherapy and/or preoperative radiotherapy. All the patients were followed up every three months by telephone until death or the end of follow-up. Study approval was obtained from independent Ethics Committees at the Xianyang Central Hospital. All participants provided written consent.

Quantitative real-time PCR (qRT-PCR)
The implementation steps of qRT-PCR analysis were based on the previous description [23]. Primers for RT-RCR are presented in Table 1.

Cell proliferation assays
HCC cells (3 × 10 3 cells) received 1 day of incubation by using culture medium (120 μL) in 96-multiwell plates at 37°C in 5% CO 2 . The plasmid was used to transfect cells for 24, 48, 72, and 96 h. Specifically, each well was added with Cell Counting Kit-8 (CCK-8) assay solution (10 mL), followed by 2 h of incubation. The microplate reader read the optical density exhibited by each well at 450 nm to estimate cell proliferation. During the colony-formation assay, we placed 500 Huh7 and LM3 cells in total in a fresh six-well plate as well as maintained them in RPMI-1640 medium (Pierce, Guangzhou, Guangdong, China) that contained 10% FBS, and then replaced the medium every 3 days. Twelve days later, we used methanol to fix cells and used 0.1% crystal violet (Sigma-Aldrich) to stain them. The resulting solution was then photographed.

5-Ethynyl-2′-deoxyuridine (EdU) assays
5-Ethynyl-2′-deoxyuridine assays assisted in determining the cellular proliferation by the use of a KeyFluor532 Click-iTEdU kit (KeyGEN, Hangzhou, Zhejiang, China) based on the instruction of the manufacturer. In brief, cells were first transfected by using corresponding vector cells, followed by 2 h of incubation by using EdU at 37°C. Then, 4% polyformaldehyde that contained phosphate buffer saline (PBS) was used for fixing these cells. Finally, fluorescence microscopy assisted in visualizing cells. We repeated experiments no less than three times.

Flow cytometry assay
Huh7 and LM3 cells were transiently transfected with the related factors. Flow cytometry (FACScan; BD Biosciences) was conducted to analyze cells by virtue of Cell-Quest software (BD Biosciences, Beijing, Haidian, China). Flowjo V10 software (Tree Star, Shenzhen, Guandong, China) assisted in analyzing all the apoptosis data regarding various cell lines.

Subcellular fractionation location
The PARIS Kit (Life Technologies, Hangzhou, Zhejiang, China) helped to separate nuclear from cytosolic fractions based on the instruction of the manufacturer.

Transwell invasion assays
The manufacturer guideline (BD Bioscience, Nanjing, Jiangsu, China) was prepared with the 48-well BD BioCoat Matrigel Invasion Chambers (Nanjing, Jiangsu, China). We added 3 × 10 5 cells to the upper wells which were separated by a PET membrane with a pore size of 8 μm that had a thin layer of matrigel basement membrane matrix. The lower chamber was added with a certain amount of FBS medium (600 μL, 10%). After cells received 24 h of incubation, 20% Giemsa was used to fix and stain those that migrated or invaded to membrane's lower side. Then 85% ethanol and crystal violet staining were used to fix and stain filters, respectively. An inverted microscope (Olympus, Tokyo, Japan) helped to count the five random fields in each chamber.

Wound-healing assay
Cells were cultured in a 6-well plate for 12 h and then wounded by a sterilized pipet tip to make a straight scratch. After cells were washed with PBS gently, 0.5% FBS together with 1% penicillin/streptomycin was used to culture cells in RPMI-1640 medium. Pictures were taken by an Olympus digital camera at 0 h and 24 h after wounding.

Biotinylated RNA-pulldown assay
The conduction of RNA-pulldown followed the previous description [24].

Chromatin immunoprecipitation (ChIP) assays
The EZ-Magna ChIP kit (Millipore, Pudong, Shanghai, China) was adopted in the ChIP assays for exploring the FOXO3 binding to the endogenous LOC554202 promoter. In brief, GC cells were cross-linked in 1% formaldehyde solution for 10 min at room temperature, followed by the addition of 130 mM of glycine for 8 min. Sonication helped to generate DNA fragments in the range of 200-300 bp. Antibodies of anti-FOXO3 (#9342, Cell Signaling Technology, Haidian, Beijing, China) and IgG were employed for each immunoprecipitation. The qRT-PCR analysis was conducted on input DNAs after immunoprecipitation.

Western blot
Frozen tissues were lysed in the lysis buffer, and the BCA protein assay kit (Beyotime, Pudong, Shanghai, China) assisted in measuring the concentration of protein. 12% SDS-PAGE gel was used to separate proteins which were then transferred to the PVDF membranes (polyvinylidene difluoride; Bio-Rad). After 2 h of blocking with 5% skim milk, the membrane received 2 h of incubation with primary antibodies and 1 h of incubation with the secondary antibody at room temperature. ImageJ software (National Institutes of Health, Bethesda, MD, USA) helped to analyze the western blot results. Primary antibodies of E-cadherin (ab241677, Epitomics Inc., Burlingame, CA, USA), Vimentin (ab45939, Epitomics Inc., Burlingame), N-cadherin (ab76011, Epitomics Inc., Burlingame), GAPDH (ab8245, Abcam Inc.), and BSG (ab188190, Epitomics Inc., Burlingame) were used in the immunoblotting assays.

Statistical analysis
SPSS statistical software package (standard version 18.0, SPSS Inc., Chicago, IL, USA) assisted in conducting the statistical analysis. Investigators were blinded to the sample group allocation during the experiment and analysis. Comparisons between the two groups were carried out by conducting t-test or chi-square test. Data among multiple groups were analyzed by a one-way analysis of variance (ANOVA), followed by a Tukey's post hoc test. Receiveroperating characteristic (ROC) curves were used to evaluate the performance of FOXO3, LOC554202, miR-485-5p, and BSG to discriminate HCC specimens from non-tumor tissues. The Kaplan-Meier method and the log-rank test were carried out for plotting the survival plots. Univariate and multivariate analyses were conducted regarding the prognostic significance of LOC554202 by virtue of the Cox proportional hazards regression model. All data are presented as the means ± SEM or SD, and p-value < 0.05 was considered with statistical significance.

The distinct upregulation of LOC554202 and its diagnostic value in HCC patients
We firstly performed RT-PCR to determine whether LOC554202 was dysregulated in HCC. As shown in Fig. 1A, we found that HCC specimens exhibited a distinctly higher LOC554202 expression relative to the matched non-tumor tissues (p < 0.01). Furthermore, patients in stage (III-IV) presented a significantly higher LOC554202 expression relative to those in stages (I-II) (p < 0.01, Fig. 1B). Then, we performed ROC assays which showed that high LOC554202 expression had an AUC value of 0.7958 (95% CI: 0.7417-0.8499) for distinguishing HCC specimens from normal tissues (Fig. 1C) and had an AUC value of 0.7033(95% CI: 0.6180-0.7887) or distinguishing specimens with stages (III-IV) from specimens with stages (I-II) (Fig. 1D). Moreover, we observed that relative to the LO2 cells, the expressions of LOC554202 were distinctly higher in HCC cell lines like HepG2, Hep3B, Huh7, SMMC-7721, and LM3 (Fig. 1E, p < 0.01). Moreover, Adriamycin (ADM)-resistant LM3 and Huh7 cells displayed a distinctly higher expression of LOC554202 than corresponding parental cells (Fig. 1F, G). Overall, the abnormal expression of LOC554202 may affect the HCC progression and ADM resistance of HCC.

Prognostic values of LOC554202 expression in HCC patients
For a better understanding of the clinical relevance of LOC554202 expressions in HCC, the 137 HCC patients were divided into two groups, group with a high expression (n = 70) and a group with a low expression (n = 67), according to the median expressions of LOC554202 in HCC samples. As presented in Table 2, high LOC554202 expression was associated with histologic grade (p = 0.025) and TNM stage (p = 0.008). Nevertheless, LOC554202 expression was not significantly related to other clinical features (All p > 0.05). Then, we collected 5-year survival data of 137 HCC patients and performed Kaplan-Meier analysis, finding that patients whose LOC554202 expression was high had shorter overall survival (OS, p < 0.0078; Fig. 1H) and disease-free survival (DFS, p < 0.0045; Fig. 1I) as compared with the LOC554202-low group. Univariate analyses revealed that TNM stage, histologic grade, as well as high LOC554202 expression could affect HCC patients' survival (Tables 3 and 4). More importantly, Multivariate analysis confirmed that high LOC554202 expression could be used to independently predict HCC patients regarding the OS (HR = 2.873, 95% CI: 1.237-4.327, p = 0.011) and DFS (HR = 2.893, 95% CI: 1.231-4.443, p = 0.009). Overall, LOC554202 could be used to predict the prognosis of HCC patients.

LOC554202 is activated by FOXO3 in HCC cells
Although the distinct dysregulation of lncRNAs has been reported in several tumors, the regulators involved in the dysregulation of these molecules are not properly understood. In recent years, more and more studies indicated that several key transcription factors were involved in the overexpression of lncRNAs in tumor cells [8,26]. Using JASPAR, which is an online tool for the exploration of transcription factors, we found several FOXO3-binding sites in the LOC554202 promoter regions with high scores ( Fig. 2A). Then, we performed RT-PCR and found that FOXO3 presented an obviously upregulated level in HCC specimens from our cohort (Fig. 2B). ROC assays revealed that FOXO3 could be used to distinguish HCC tissues from non-tumor tissues with an AUC of 0.7878 (p < 0.001, Fig. 2C). FOXO3 was also highly expressed in five HCC cell lines (Fig. 2D). To explore whether the dysregulation of FOXO3 could influence LOC554202 expression, we performed RT-PCR using LM3 and Huh7 cells, which received the transfection of si-FOXO3-#1, si-FOXO3-#1, or pcDNA3.1-FOXO3, finding that knockdown of FOXO3 suppressed LOC554202 expression in the above cell lines, while overexpression of FOXO3 displayed an opposite trend (Fig. 2E, F). Subsequent ChIP assays documented an apparent FOXO3-binding activity on the endogenous LOC554202 promoter region around E2 (Fig. 2G). More importantly, based on luciferase reporter assays, FOXO3 only bound to the E2 (−1148 bp) binding site, and did not bind to the other two sites (Fig. 2H-J). Overall, our findings indicated that some TFs can contribute to the tumorigenesis and progression of various tumors not only via modulating the expression of protein-coding genes but also through affecting lncRNAs transcription.

LOC554202 loss of function markedly impairs HCC cell malignant phenotypes
To determine the function possessed by LOC554202 in HCC cells, LM3 and Huh7 cells were used to construct stable cells with LOC554202 knockdown due to the higher expression of LOC554202 in the above two cells (Fig. 3A). Next, as shown in CCK-8 assays, LOC554202 knockdown suppressed the proliferation of LM3 and Huh7 cells (Fig. 3B). We also explored the effect of LOC554202 on chemotherapy resistance of HCC cells, finding that inhibition rate increased with the increase of concentration of ADM, and knockdown of LOC554202 decreased the inhibition effect of ADM on HCC cells (Fig. 3C). Also, based on the clone formation assays, LM3 and Huh7 cells presented a decreasing clone forming ability due to LOC554202 knockdown (Fig. 3D). Besides, as confirmed by the 5-ethynyl-2′-deoxyuridine (EdU) proliferation assay, LOC554202 depletion caused the cell to grow slowly (Fig. 3E). Importantly, we also showed that knockdown of LOC554202 promoted the apoptosis of LM3 and Huh7 cells (Fig. 3F). On the other hand, we also demonstrated that the activity of Caspase 3/9 was suppressed in HCC cells transfected with si-LOC554202-#1 and si-LOC554202-#2 (Fig. 3G). On the other hand, we explored the possible function of knockdown of LOC554202 in the metastatic potential of HCC cells. As presented in Fig. 4A, using the wound-healing assays, we observed that LOC554202  depletion caused significant inhibition of HCC cell migration. Besides, according to the transwell assays results, LOC554202 knockdown suppressed the invasive ability of LM3 and Huh7 cells compared with control cells (Fig. 4B).
Since epithelial-mesenchymal transition (EMT) progress could affect the migration and invasion ability, the possible functions of LOC554202 on the EMT process were studied. As expected, LOC554202 knockdown inhibited the EMT  process in HCC cells characterized by decreased expressions of mesenchymal marker (Vimentin and N-cadherin) and increased expressions of epithelial marker (E-cadherin) (Fig. 4C). Overall, our findings suggested LOC554202 as an oncogenic lncRNA in HCC progression.

LOC554202 directly interacts with miR-485-5p in HCC cells
As confirmed by studies, lncRNAs function as the endogenous RNAs competing for microRNAs, thereby regulating the underlying targets [27]. Subcellular fractionation assays assisted in confirming the potential mechanisms regarding the LOC554202-related modulation of biological features in HCC cells. As shown in Fig. 5A, we observed that LOC554202 was distributed in both the cytoplasm and nucleus. Using starbase, we predicted that LOC554202 could be targeted by miR-485-5p (Fig. 5B). Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) assays revealed that the potential targeting genes of miR-485-5p (n = 77) were positively associated with tumorrelated pathways, suggesting it as a functional regulator in the progression of HCC (Fig. 5C-E). Then, we determined miR-485-5p expression in HCC in TCGA data sets, finding miR-485-5p expression exhibited a decreased trend in HCC specimens relative to non-tumor liver tissues (Fig. 5F). In our cohort, we further confirmed that miR-485-5p expression in HCC specimens was distinctly lower than that in matched non-tumor specimens (Fig. 5G). The diagnostic assays using the expression data from our cohort revealed that high miR-485-5p expression had an AUC value of 0.8365 (95% CI: 0.7873-0.8857) for distinguishing HCC specimens from normal tissues (Fig. 1H). In addition, the distinct downregulation of miR-485-5p also occurred in  Five HCC cell lines (Fig. 5I). As found by the functional assays, miR-485-5p overexpression significantly hindered LM3 and Huh7 cells in terms of proliferation and invasion according to the results of CCK-8 assays and transwell assays (Fig. 5J, K). For figuring out how LOC554202 is associated with miR-485-5p, RNA-pulldown assays were conducted, finding that LOC554202 had the function of precipitating miR-485-5p in HCC cells (Fig. 5L). The luciferase assay showed that miR-485-5p mimic could exert an obvious inhibiting effect on the reporter activities of wt-LOC554202 but not mut-LOC554202 in LM3 and Huh7 cells (Fig. 5M). Moreover, we further performed RT-PCR to confirm the above results and found that knockdown of LOC554202 promoted the expression of miR-485-5p ( Fig. 5N), while miR-485-5p overexpression led to a decreased expression of LOC554202 in LM3 and Huh7 cells (Fig. 5O). Overall, our findings strongly suggest that LOC554202 targeted and negatively regulated miR-485-5p in HCC cells.

LOC554202 modulated malignant behaviors of HCC via miR-485-5p/BSG axis
We first predicted the potential targets regarding miR-485-5p, for further investigating the function exhibited by miR-485-5p in HCC. Using TCGA data sets with UALCAN, we identified the top 50 overexpressed lncRNAs in HCC, which was shown in the heat map (Fig. 6A). In addition, the dysregulated genes analyzing TCGA data sets were shown in the Volcano plot (Fig. 6B). Then, we searched mi-Randa, Starbase, and TargetScan, finding that miR-485-5p may directly target BSG which was among the top 50 overexpressed lncRNAs in HCC (Fig. 6C). Previous studies reported the functional role of BSG in the development of many tumors [28,29]. However, the potential functions of BSG in HCC remained largely unclear. Using UALCAN, we performed pan-cancer expression, finding that BSG exhibited a high level in many types of tumors based on the TCGA data sets (Fig. 6D). Then, we also observed the high expression of BSG and its positive relation to advanced stages and tumor metastasis (Fig. 6E). In our cohort, we also confirmed the mRNA expression of BSG was upregulated in HCC specimens and had a potential diagnostic in distinguishing HCC specimens from non-tumor specimens (Fig. 6F). Besides, we performed a similar analysis, finding that BSG exhibited a higher level in HCC specimens than in   non-tumor tissues and its diagnostic value was also confirmed (Fig. 6G). Moreover, we provided evidence that BSG mRNA exhibited an obviously upregulated expression in five HCC cell lines (Fig. 6H). On the other hand, we explored the prognostic value of BSG expression in HCC patients. 137 HCC patients were divided into two groups, a group with a high expression (n = 76) and a group with a low expression (n = 61), according to the median  FXYD2  GMNN  STMN1  LYZ  CD34  CD24  CKB  CLUL   UBE2S  PAFAH1B3  HIST2H4A  ACSL4  PKM2  COL1A1  TYMS  RNASEH2A  CRNA00152  COL4A1  CYP7A1  BCAT2  TPM2  TP53I3  CLDN15  HMGA1  LAPTM4B   Log2(TPM+1)   0  5  10  15   PDLIM7  Clorf35  GPAA1  SCNM1  Cl9orf48  GOLM1  UCK2  BOLA2  SCRIB  NDRG1  DTYMK  ZNF219  PYGO2  CAPG  TMEM98  HIST1H1C  SAC3D1  SCAMP3  Cllorf9  CKS1B  C8orf33  IRAK1   BSG   expressions of BSG in HCC samples. Based on the results of GEPIA, we observed that patients whose BSG expression was high exhibited a shorter OS (p = 0.0014) and DFS (p = 0.0097) relative to patients whose BSG expression was low (Fig. 6I), which was also confirmed in our cohort (Fig. 6J). As suggested by those finding, together with previous findings, BSG could be a tumor promoter in HCC. Importantly, based on luciferase reporter assays, miR-485-5p overexpression remarkably suppressed the BSG wildtype (wt) luciferase reporters (Fig. 6K). Furthermore, miR-485-5p overexpression in LM3 and Huh7 cells significantly lowered the mRNA and protein levels of BSG (Fig. 6L, M).
On that account, miR-485-5p may suppress the progression of HCC via targeting BSG. To further explore whether miR-485-5p/BSG axis could affect the HCC progression mediated by LOC554202, si-NC, si-LOC554202-#1, si-LOC554202-#1, and anti-miR-NC or anti-miR-485-5p were used to transfect the LM3 and Huh7 cells. As presented in Fig. 7A, BSG levels decreased due to LOC554202 knockdown and it was obviously reduced by transfection of anti-miR-485-5p. Besides, the proliferation, migration, and invasion suppressions caused by silencing LOC554202 were distinctly attenuated by miR-485-5p exhaustion in LM3 and Huh7 cells (Fig. 7B-F). Taken together, our study revealed that FOXO3-induced upregulation of LOC554202 exerted its function as a ceRNA through sponging miR-485-5p to regulate downstream BSG expression, and therefore contributed to the proliferation and metastasis of HCC cells (Fig. 8).

Discussion
In recent years, an increasing number of lncRNAs have emerged as critical regulators in tumor progression [30,31]. However, the function of many lncRNAs remains unclear in HCC. In this study, we identified a possible HCC-related lncRNA, LOC554202 which was confirmed to exhibit an obvious upregulation in HCC tissues and HCC/ADM cells, indicating that LOC554202 may contribute to chemotherapy resistance. The diagnostic value of LOC554202 in distinguishing HCC specimens from non-tumor specimens was also confirmed. Clinical assays revealed that high LOC554202 expression was associated with advanced clinical progress and poor prognosis of HCC patients. In addition, we performed functional assays, finding that LOC554202 knockdown hindered HCC cells regarding the proliferation, migration, and invasion while facilitated the apoptosis. Therefore, the findings in the study proved that LOC554202 may be used as a novel biomarker to predict the diagnosis and prognosis of HCC patients, and also a therapeutic target.
Although studies have confirmed the abnormal expressions of lncRNAs in different tumors, the potential mechanisms involving lncRNAs dysregulation in tumor cells remained unclear. Interestingly, many studies have proved that some TFs, as well as epigenetic regulators, could affect the transcription of some lncRNAs [32]. For instance, STAT1-mediated lncRNA LINC00174 overexpression promoted the proliferation and invasion regarding colorectal carcinoma cells via regulating the miRNA-1910-3p/TAZ axis [33]. HIF1-induced upregulation of lncRNA RAET1K sponged miRNA-100-5p, thus promoting HCC cells in terms of their proliferation and metastasis [34]. In this study, FOXO3 was predicted to be bound to the LOC554202 promoter region by the use of the JASPAR online database. Previously, FOXO3 overexpression and its oncogenic roles were reported in many tumors, including HCC [15,35]. Mechanism experiments demonstrated that FOXO3 could bind to the LOC554202 promoter region and induce its transcription. Thus, our findings revealed FOXO3 may promote HCC progression via upregulating LOC554202.
In recent years, the expression and function of LOC554202 have been reported in several tumors. For instance, high LOC554202 expression was reported to partially represent the poor prognosis of colorectal cancer patients as well as could predict a favorable outcome in patients treated with oxaliplatin [36]. LOC554202 was shown to be highly expressed in lung cancer and promote acquired gefitinib resistance in lung cancer via sponging miRNA-31 expressions [21]. In laryngeal squamous cell carcinoma, based on studies, high LOC554202 levels could regulate miRNA-31, thereby promoting tumor cells in terms of proliferation and metastasis [22]. However, whether LOC554202 may influence HCC progression had not been reported. Given the distinct overexpression of LOC554202 in HCC and its clinical value, we further performed loss-of-function assays to explore the potential effects of LOC554202 on HCC cells, finding that knockdown of LOC554202 suppressed the proliferation, metastasis, and EMT progress of HCC cells. Overall, our findings suggested LOC554202 as an oncogenic lncRNA in HCC progression.
It has been confirmed that one of the major functions of lncRNAs is to harbor miRNAs to act as a "sponge" and decrease the expressions and activities of miRNAs [37,38]. For figuring out the potential molecular mechanism for LOC554202 to regulate the downstream functional factors in HCC, the localization of LOC554202 in LM3 and Huh7 cells was first identified because the potential functions of lncRNAs were associated with their subcellular localization. The results of subcellular fractionation revealed that LOC554202 was expressed in the cytoplasm and nucleus. StarBase v2.0 indicated that LOC554202 may be a target of miR-485-5p. In previous studies, miR-485-5p was reported to promote the progression of several tumors, including HCC, which was also demonstrated in our functional assays when miR-485-5p was overexpressed [39,40]. In addition, in HCC, several lncRNAs such as lncRNA RHPN1-AS and LINC00460 were reported to promote the proliferation and invasion of HCC cells via sponging miR-485-5p [39,41]. These findings suggested miR-485-5p as a common target of different lncRNAs. In this research, we also provided evidence that LOC554202 acted as a ceRNA for miR-485-5p in the cytoplasm and sponged miR-485-5p. Thus, our findings suggested that LOC554202 may promote HCC progression via sponging miR-485-5p.
Growing basic evidences have indicated that basigin (BSG, also named TCSF, CD147, EMMPRIN) is a transmembrane-type glycoprotein that is strongly expressed on the cell membranes of several tumor cells [42]. It has been confirmed that BSG could stimulate para-cancerous fibroblast secreting matrix metalloproteinases (MMP) which can help cancer cells easily penetrate the basement membrane of blood vessels, therapy resulting in metastasis [43]. Previously, large-scale studies have found the obvious BSG upregulation in many tumors [44,45]. Importantly, BSG has been reported to promote the proliferation and invasion of HCC cells [46]. In this study, we also observed that BSG was highly expressed in HCC and predicted a poor prognosis of HCC patients. Interestingly, by the use of bioinformatics analysis, we found that BSG may act as a targeting protein of miR-485-5p, which was confirmed through the luciferase assays. Moreover, BSG was overexpressed by LOC554202, while downregulated by miR-485-5p. In addition, rescue experiments further confirmed knockdown of miR-485-5p reversed the oncogenic effects of LOC554202 overexpression on the proliferation, migration, and invasion of HCC cells. Overall, our findings revealed that FOXO3induced upregulation of LOC554202 exhibited its oncogenic effects at least in part by modulating the miR-485-5p/BSG axis. The FOXO3/LOC554202/miR-485-5p/BSG axis may be a promising therapeutic target for HCC.

Conclusions
The study is the first one that showed the distinct upregulation of LOC554202 expression in HCC specimens, which is related to the poor clinical outcome and can be used to predict the diagnosis and prognosis of HCC patients. FOXO3 can be directly bound to the LOC554202 promoter region, thereby activating its transcription. Silencing LOC554202 can target the miR-485-5p/BSG axis, thus hindering the proliferation and metastasis of HCC cells and inducing apoptosis. To sum up, the data in the study prove the possibility of targeting LOC554202/miR-485-5p/BSG axis being an effective alternative treatment, which can provide an essential implication for HCC treatment and diagnosis.

Data availability
The data used to support the findings of this study are available from the corresponding author upon request