RNA-seq data of HCC tissues and clinicopathological characteristics of these HCCs were retrieved from TCGA project using R2 (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi) and OncoLnc (http://www.oncolnc.org/). Two in silico tools Coding Potential Assessment Tool (CPAT) (http://lilab.research.bcm.edu/) and TestCode (https://www.bioinformatics.org/sms/testcode.html) were used to calculate coding potential of ADORA2A-AS1 (36). The proteins bound to ADORA2A-AS1 were predicted using three in silico tools, RBPmap (mapping binding sites of RNA binding proteins) (http://rbpmap.technion.ac.il/), ATtRACT (a database of RNA binding proteins and associated motifs) (https://attract.cnic.es/#), and RNA-Protein Interaction Prediction (RPISeq) (http://pridb.gdcb.iastate.edu/RPISeq/index.html) (37, 38).

Surgically resected HCC tissues (n = 76) were obtained from the Affiliated Hospital of Youjiang Medical University for Nationalities (Baise, China). Informed consents were acquired from all HCC patients. Clinicopathological characteristics of these 76 HCC cases were presented in Table 1 . The Ethics Committee of Affiliated Hospital of Youjiang Medical University for Nationalities approved this study.

Human HCC cell lines SNU-398, SK-HEP-1, and Hep3B were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). SNU-398 cells were cultured in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) added with 10% fetal bovine serum (FBS, Invitrogen). SK-HEP-1 and Hep3B cells were cultured in Eagle’s minimum essential medium (Invitrogen) added with 10% FBS. All cells were authenticated by short-tandem repeat analyses. All cells were routinely detected as mycoplasma-free. Where indicated, cells were treated with 50 µM α-amanitin (Sigma-Aldrich, St. Louis, MO, USA) or 10 µM LY294002 (Selleck, Houston, TX, USA) for indicated time.

RNA fluorescence in situ hybridization (FISH) was performed to detect subcellular localization of ADORA2A-AS1 in SNU-398 cells as previously described (31). Briefly, the ADORA2A-AS1 probes were designed and synthesized by Advanced Cell Diagnostics (Newark, NJ, USA). RNA FISH was undertaken using RNAscope Fluorescent Multiplex Detection Kit (Advanced Cell Diagnostics) following the provided instruction. FISH results were detected using confocal laser scanning microscopy (Leica, Wetzlar, Germany). Furthermore, subcellular localization of ADORA2A-AS1 was evaluated using cytoplasmic and nuclear RNA isolation, followed by quantitative real-time polymerase chain reaction (qRT-PCR). Cytoplasmic and nuclear RNA isolation was undertaken using the PARIS Kit (Invitrogen). qRT-PCR was undertaken as described below.

Total RNA from HCC tissues and cells was extracted using TRIzol Reagent (Invitrogen). First-strand complementary DNA (cDNA) was generated using the M-MLV Reverse Transcriptase (Invitrogen) and random primers. qRT-PCR was undertaken using the TB Green Premix Ex Taq II (TaKaRa, Tokyo, Japan) on StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The sequences of gene-specific primers were as follows: 5′-GGGCTTTCCTACATCCATC-3′ (sense) and 5′-CACAGCAAGTCCATTCCTCT-3′ (anti-sense) for ADORA2A-AS1, 5′-ATCAACCGCCCCATCATC-3′ (sense) and 5′-TGCCCACCGTCCAGTATTT-3′ (anti-sense) for FSCN1, 5′-CCGCAGTGACGGAAAGATG-3′ (sense) and 5′-CCTGGGACAGACGGAAGTT-3′ (anti-sense) for MMP2, 5′-TGGGACCCGTGGGAAGAA-3′ (sense) and 5′-GGCACTTTCAGACTGGACCTC-3′ (anti-sense) for BIRC7, 5′-TGGAGAAATAGTAGATGGC-3′ (sense) and 5′-GGTGAGGAAGTAAAAACAG-3′ (anti-sense) for Metastasis Associated Lung Adenocarcinoma Transcript 1 (MALAT1), 5′-ACACGGACAGGATTGACAGA-3′ (sense) and 5′-GGACATCTAAGGGCATCACA-3′ (anti-sense) for 18S rRNA, 5′-GTCGGAGTCAACGGATTTG-3′ (sense) and 5′-TGGGTGGAATCATATTGGAA-3′ (anti-sense) for glyceraldehyde 3-phosphate dehydrogenase (GAPDH). GAPDH was employed as an internal control. Relative expression levels of transcripts were calculated using the comparative Ct method.

ADORA2A-AS1 full-length sequences were PCR-amplified with the primers 5′-GGGGTACCTTCGCCTCTCGGTTAGC-3′ (sense) and 5′-CGGGATCCAAAGTTTTAAAACATTCCTGTTAAC-3′ (anti-sense). Next, the PCR products were cloned into the Kpn I and BamH I sites of pcDNA3.1(+) vector (Invitrogen) to construct ADORA2A-AS1 overexpression vector pcDNA3.1-ADORA2A-AS1. The PCR products were also cloned into the Kpn I and BamH I sites of pSPT19 vector (Roche, Basel, Switzerland) to construct ADORA2A-AS1 in vitro transcription vector pSPT19-ADORA2A-AS1. Two pairs of cDNA oligonucleotides targeting ADORA2A-AS1 were synthesized and cloned into the shRNA lentivirus expressing plasmid pLV3/H1/GFP&Puro (GenePharma, Shanghai, China). ADORA2A-AS1-specific shRNA lentivirus was produced by GenePharma. Scrambled non-targeting shRNA lentivirus was used as negative control (NC). The sequences of shRNA oligonucleotide were as follows: 5′-GATCCGGATAGAGATCTCAAAGAAGTTTCAAGAGAACTTCTTTGAGATCTCTATCCTTTTTTG-3′ (sense) and 5′-AATTCAAAAAAGGATAGAGATCTCAAAGAAGTTCTCTTGAAACTTCTTTGAGATCTCTATCCG-3′ (anti-sense) for shRNA-ADORA2A-AS1-1, 5′-GATCCGCCCATTGCTTCCAAGCAACATTCAAGAGATGTTGCTTGGAAGCAATGGGCTTTTTTG-3′ (sense) and 5′-AATTCAAAAAAGCCCATTGCTTCCAAGCAACATCTCTTGAATGTTGCTTGGAAGCAATGGGCG-3′ (anti-sense) for shRNA-ADORA2A-AS1-2, 5′-GATCCGTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAACTTTTTTG-3′ (sense) and 5′-AATTCAAAAAAGTTCTCCGAACGTGTCACGTTCTCTTGAAACGTGACACGTTCGGAGAACG-3′ (anti-sense) for shRNA-NC.

To generate HCC cells with ADORA2A-AS1 stable overexpression, ADORA2A-AS1 overexpression vector pcDNA3.1-ADORA2A-AS1was transfected into SNU-398 and SK-HEP-1 cells using Lipofectamine 3000 (Invitrogen). Empty vector pcDNA3.1 was concurrently transfected and used as negative control. Forty-eight hours after transfection, 800 µg/ml neomycin was added. Individual colonies were picked and expanded. The expression of ADORA2A-AS1 in these colonies was measured using qRT-PCR. We chose one colony with successful ADORA2A-AS1 overexpression for each cell line. To generate HCC cells with ADORA2A-AS1 stable silencing, SUN-398 and Hep3B cells were infected with ADORA2A-AS1-specific shRNA lentivirus. Ninety-six hours after infection, 2 µg/ml puromycin was added. Individual colonies were picked and expanded. The expression of ADORA2A-AS1 in these colonies was measured using qRT-PCR. We chose one colony with successful ADORA2A-AS1 downregulation for each shRNA lentivirus and each cell line. To generate HCC cells with ADORA2A-AS1 and FSCN1 overexpression, ADORA2A-AS1-overexpressing SUN-398 cells were infected with FSCN1 overexpression lentivirus (GenePharma). Ninety-six hours after infection, 2 µg/ml puromycin was added. Individual colonies were picked and expanded. The expression of FSCN1 in these colonies was measured using Western blot.

Cell proliferation was measured using Cell Counting Kit-8 (CCK-8) and 5-ethynyl-2’-deoxyuridine (EdU) incorporation assays as previously described (39). Briefly, for CCK-8 assay, HCC cells (2,000 per well) were plated into 96-well plate. After culture for 4 days, cell proliferation was measured using CCK-8 reagents (Dojindo, Kumamoto, Japan) following the provided protocol. EdU incorporation assays were performed using the Cell-Light EdU Apollo567 In Vitro Kit (RiboBio, Guangzhou, China) following the provided protocol. Cell apoptosis was measured using the Caspase-3 Activity Assay Kit (Cell Signaling Technology, Danvers, MA, USA) following the provided protocol. Cell migration was measured using transwell migration assay as we previously described (40). Cell invasion was measured using transwell invasion assay as we previously described (40).

Five-week-old male BALB/c nude mice were purchased from Shanghai SLAC Laboratory Animal Co. and maintained in specific pathogen-free conditions. SNU-398 cells with ADORA2A-AS1 overexpression or control were subcutaneously injected into the back flank of nude mice (5 × 10⁶ cells per site). Subcutaneous xenograft volumes were measured every week following the formula volume = 0.5 × length × width². At the 28th day after injection, subcutaneous xenografts were resected, weighed, and subjected to immunohistochemistry (IHC) staining using primary antibodies against Proliferating Cell Nuclear Antigen (PCNA) (#13110, 1:4,000; Cell Signaling Technology), Ki67 (#9027, 1:400; Cell Signaling Technology), or cleaved caspase-3 (#9664, 1:1,000; Cell Signaling Technology). The subcutaneous xenografts were also subjected to terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick end labeling (TUNEL) assay using the TUNEL Cell Apoptosis Detection Kit (Beyotime, Shanghai, China). SNU-398 cells with ADORA2A-AS1 overexpression or control were also intrasplenically injected into nude mice (2 × 10⁶ cells per mouse). At the 35th day after injection, the livers were resected and subjected to hematoxylin and eosin (H&E) staining to count the diameter and number of metastatic nodules. All animal experiments were approved by the Animal Ethics Committee of Affiliated Hospital of Youjiang Medical University for Nationalities.

ADORA2A-AS1 was in vitro transcribed from pSPT19-ADORA2A-AS1 using the MEGAscript^® Kit (Invitrogen) and Sp6 RNA polymerase (Invitrogen). ADORA2A-AS1 transcript was labeled using the Pierce™ RNA 3′ End Desthiobiotinylation Kit (Thermo Scientific, Waltham, MA, USA). RNA pull-down assays were performed in SNU-398 cells using ADORA2A-AS1 end-labeled with desthiobiotin and the Pierce™ Magnetic RNA-Protein Pull-Down Kit (Thermo Scientific) following the provided protocol. The protein present in the pull-down material was detected by Western blot.

Proteins extracted from indicated HCC cells or RNA pull-down material were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membrane. After blocking, the membranes were incubated with primary antibodies against HuR (Hu Antigen R) (#03-102, 1:500; Millipore, Burlington, MA, USA), FSCN1 (#54545, 1:1,000; Cell Signaling Technology), phosphor-AKT (66444-1-Ig, 1:2,000; Proteintech, Chicago, IL, USA), AKT (10176-2-AP, 1:1,000; Proteintech), MMP2 (#40994, 1:1,000; Cell Signaling Technology), BIRC7 (#5471, 1:1,000; Cell Signaling Technology), or GAPDH (ab8245, 1:5,000; Abcam) overnight at 4°C. The secondary antibodies used were IRDye 680RD Goat anti-Mouse IgG and IRDye 800CW Goat anti-Rabbit IgG (Li-Cor, Lincoln, NE, USA).

RNA immunoprecipitation (RIP) assays were undertaken in indicated HCC cells using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) and a primary antibody against HuR (#03-102, 5 µg per reaction; Millipore) following the provided instruction. The precipitated RNAs were measured by qRT-PCR.

All statistical analyses were performed using the GraphPad Prism 6.0 Software. Log-rank test, Pearson chi-square test, Student’s t-test, one-way ANOVA followed by Dunnett’s multiple comparisons test, and Pearson correlation analysis were used as indicated in the figure and table legends. p < 0.05 was considered statistically significant.

To explore clinical significances of ADORA2A-AS1 in HCC, we analyzed the RNA-seq data of HCC tissues from TCGA project. The results revealed that low ADORA2A-AS1 expression was associated with worse overall survival of HCC patients ( Figure 1A ). Furthermore, correlation analyses between ADORA2A-AS1 expression and clinicopathological characteristics revealed that low ADORA2A-AS1 expression was associated with advanced American Joint Committee on Cancer (AJCC) pathologic t stage and pathologic tumor stage ( Table 2 ). To further confirm the clinical significances of ADORA2A-AS1 in HCC, we randomly collected 76 HCC samples, measured ADORA2A-AS1 expression in these 76 HCC tissues, and analyzed the correlation between ADORA2A-AS1 expression and clinicopathological characteristics. Kaplan–Meier survival analyses revealed that low ADORA2A-AS1 expression was also associated with worse overall survival in our HCC cohort ( Figure 1B ). In this HCC cohort, low ADORA2A-AS1 expression was found to be correlated with bad encapsulation, microvascular invasion, and advanced Barcelona Clinic Liver Cancer (BCLC) stage ( Table 1 ). These findings suggested ADORA2A-AS1 as a potential prognostic biomarker for HCC. Furthermore, the expressions of ADORA2A-AS1 in immortalized human liver cell THLE-2 and human HCC cells SNU-398, SK-HEP-1, and Hep3B were measured, and the results revealed that ADORA2A-AS1 was lowly expressed in HCC cells compared with that in immortalized liver cell ( Figure 1C ).

Two in silico tools CPAT (http://lilab.research.bcm.edu/) and TestCode (https://www.bioinformatics.org/sms/testcode.html) both indicated that ADORA2A-AS1 was a noncoding RNA ( Figures S1A, B ). RNA FISH results showed that ADORA2A-AS1 was mainly distributed in the cytoplasm ( Figure S1C ). Subcellular fractionation followed by qRT-PCR results also showed that ADORA2A-AS1 was mainly distributed in the cytoplasm ( Figure S1D ).

Due to the correlation between low ADORA2A-AS1 expression and bad encapsulation, microvascular invasion, advanced pathologic t stage, advanced pathologic tumor stage, advanced BCLC stage, and poor overall survival, we next detected the potential roles of ADORA2A-AS1 in HCC. We stably overexpressed ADORA2A-AS1 in SNU-398 and SK-HEP-1 cells via stable transfecting of ADORA2A-AS1 overexpression vector ( Figure 2A ). We also stably silenced ADORA2A-AS1 expression in SNU-398 and Hep3B cells via stable infecting of ADORA2A-AS1 shRNA lentiviruses ( Figure 2B ). CCK-8 assay results showed that ADORA2A-AS1 overexpression repressed cell proliferation of SNU-398 and SK-HEP-1 cells ( Figure 2C ). Conversely, CCK-8 assay results also showed that ADORA2A-AS1 silencing promoted cell proliferation of SNU-398 and Hep3B cells ( Figure 2D ). EdU incorporation assay results showed that SNU-398 and SK-HEP-1 cells with ADORA2A-AS1 overexpression had less EdU-positive cells compared with their control cells ( Figure 2E ), indicating that ADORA2A-AS1 overexpression repressed cell proliferation of SNU-398 and SK-HEP-1 cells. Conversely, SNU-398 and Hep3B cells with ADORA2A-AS1 silencing had more EdU-positive cells compared with their control cells ( Figure 2F ), indicating that ADORA2A-AS1 silencing promoted cell proliferation of SNU-398 and Hep3B cells. Caspase-3 activity assay results showed that ADORA2A-AS1 overexpression increased caspase-3 activity in SNU-398 and SK-HEP-1 cells ( Figure 2G ), indicating that ADORA2A-AS1 overexpression promoted cell apoptosis of SNU-398 and SK-HEP-1 cells. Conversely, caspase-3 activity assay results also showed that ADORA2A-AS1 silencing decreased caspase-3 activity in SNU-398 and Hep3B cells ( Figure 2H ), indicating that ADORA2A-AS1 silencing repressed cell apoptosis of SNU-398 and Hep3B cells. These findings demonstrated that ADORA2A-AS1 repressed HCC cellular proliferation and promoted HCC cellular apoptosis.

Given that low ADORA2A-AS1 expression was associated with bad encapsulation and microvascular invasion, we next detected the potential roles of ADORA2A-AS1 in HCC cellular migration and invasion. Transwell migration assay results showed that SNU-398 and SK-HEP-1 cells with ADORA2A-AS1 overexpression had less migrated cells compared with their control cells ( Figure 3A ). Conversely, SNU-398 and Hep3B cells with ADORA2A-AS1 silencing had more migrated cells compared with their control cells ( Figure 3B ). Transwell invasion assay results showed that SNU-398 and SK-HEP-1 cells with ADORA2A-AS1 overexpression had less invasive cells compared with their control cells ( Figure 3C ). Conversely, SNU-398 and Hep3B cells with ADORA2A-AS1 silencing had more invasive cells compared with their control cells ( Figure 3D ). These findings demonstrated that ADORA2A-AS1 repressed HCC cellular migration and invasion.

To investigate the roles of ADORA2A-AS1 in mouse xenograft model in vivo, SNU-398 cells with ADORA2A-AS1 stable overexpression or control were subcutaneously injected into nude mice. Subcutaneous xenograft growth curve showed that ADORA2A-AS1 overexpression slowed xenograft growth rate ( Figure 4A ). At the 28th day after injection, subcutaneous xenografts were resected and weighed. Obviously, the xenografts formed by ADORA2A-AS1-overexpressed SNU-398 cells were lighter and smaller than the xenografts formed by control SNU-398 cells ( Figures 4B, C ). The xenografts were further subjected to PCNA and Ki67 IHC staining to evaluate cell proliferation in vivo. Proliferation markers PCNA and Ki67 IHC staining both showed that ADORA2A-AS1 overexpression repressed cell proliferation in vivo ( Figures 4D, E ). Furthermore, cleaved caspase-3 IHC staining of the xenografts showed that ADORA2A-AS1 overexpression promoted cell apoptosis in vivo ( Figure 4F ). TUNEL assay results further confirmed that ADORA2A-AS1 overexpression promoted cell apoptosis in vivo ( Figure 4G ). To investigate the roles of ADORA2A-AS1 in HCC metastasis, SNU-398 cells with ADORA2A-AS1 stable overexpression or control were intrasplenically injected into nude mice. At the 35th day after injection, liver metastases were evaluated by H&E staining. As shown in Figures 4H–J , ADORA2A-AS1-overexpressed SNU-398 cells formed significantly smaller and less liver metastatic nodules. These findings demonstrated that ADORA2A-AS1 repressed HCC xenograft growth and metastasis in vivo.

The most common mechanism of action of lncRNAs is to bind proteins and regulate the function of the interacted proteins (20). To investigate whether ADORA2A-AS1 also binds protein and which proteins are bound to ADORA2A-AS1, two in silico tools RBPmap (mapping binding sites of RNA binding proteins) (http://rbpmap.technion.ac.il/) and ATtRACT (a database of RNA binding proteins and associated motifs) (https://attract.cnic.es/#) were used to predict the proteins bound to ADORA2A-AS1. Intriguingly, RNA-binding protein HuR was predicted to bind ADORA2A-AS1 by both tools ( Tables S1 , S2 ). The binding of ADORA2A-AS1 to HuR was further predicted by another in silico tool RPISeq (RNA-Protein Interaction Prediction) (http://pridb.gdcb.iastate.edu/RPISeq/index.html) with a score of 0.8. RNA pull-down assays using ADORA2A-AS1 end-labeled with desthiobiotin showed that HuR was specifically enriched in ADORA2A-AS1 group ( Figure 5A ). RIP assays using anti-HuR antibody showed that ADORA2A-AS1 was specifically enriched in anti-HuR antibody group ( Figure 5B ). These results demonstrated that ADORA2A-AS1 specifically bound to HuR. HuR is reported to bind and stabilize several mRNAs, including FSCN1, VEGFC, MAT2A, MDM2, CCND1, CASP3, USP7, FAS, PDGFA, CTSS, GATA3, and EGFR (41–46). The RNA-seq data of HCC tissues from TCGA project showed that among all these HuR targets, FSCN1 has the most significant correlation with ADORA2A-AS1 with r value of -0.4627 ( Figure 5C and Figure S2 ). The significantly negative correlation between ADORA2A-AS1 and FSCN1 expression levels in HCC tissues was also found in our HCC cohort ( Figure 5D ). Therefore, we further explore the potential effects of ADORA2A-AS1 on FSCN1. RIP assays using anti-HuR antibody in SNU-398 cells with ADORA2A-AS1 stable overexpression or control showed that HuR bound to FSCN1 transcript and ADORA2A-AS1 overexpression repressed the binding of HuR to FSCN1 transcript ( Figure 5E ). Conversely, RIP assays using anti-HuR antibody in SNU-398 cells with ADORA2A-AS1 stable silencing or control showed that ADORA2A-AS1 silencing promoted the binding of HuR to FSCN1 transcript ( Figure 5F ). These findings suggested that ADORA2A-AS1 repressed the binding of HuR to FSCN1 transcript via competitively binding HuR. Given that HuR could bind and stabilize FSCN1 transcript, we next investigated the effects of ADORA2A-AS1 on FSCN1 transcript stability. SNU-398 cells with ADORA2A-AS1 stable overexpression or silencing were treated with α-amanitin to block new RNA transcription. FSCN1 transcript stability over time was measured, and the results showed that ADORA2A-AS1 overexpression shortened the half-life of FSCN1 transcript ( Figure 5G ). Conversely, ADORA2A-AS1 silencing elongated the half-life of FSCN1 transcript ( Figure 5H ). As expected, FSCN1 transcript level was decreased in SNU-398 cells with ADORA2A-AS1 overexpression and increased in SNU-398 cells with ADORA2A-AS1 silencing ( Figures 5I, J ). FSCN1 protein level was also decreased in SNU-398 cells with ADORA2A-AS1 overexpression and increased in SNU-398 cells with ADORA2A-AS1 silencing ( Figures 5K, L ). These findings suggested that ADORA2A-AS1 bound to HuR, repressed the binding of HuR to FSCN1 transcript, reduced FSCN1 transcript stability, and downregulated FSCN1 expression.

FSCN1, also named as Fascin, is an actin-bundling protein (47). High FSCN1 expression has been revealed to be associated with worse outcomes in HCC (48). Increasing evidence showed that FSCN1 is involved in cell proliferation, apoptosis, motility, chemotherapeutic resistance, tumor growth, and metastasis of several malignancies (49, 50). PI3K/AKT pathway was revealed by several reports to be a critical downstream target of FSCN1 (51, 52). Thus, we further investigate the effects of ADORA2A-AS1 on PI3K/AKT pathway. AKT phosphorylation level was decreased in SNU-398 cells with ADORA2A-AS1 overexpression compared with control SNU-398 cells ( Figure 6A ). Conversely, AKT phosphorylation level was increased in SNU-398 cells with ADORA2A-AS1 silencing ( Figure 6B ). MMP2 is a downstream target of AKT activation, which is frequently reported to be involved in cell motility (53). BIRC7, known as Livin, is another critical downstream target of AKT activation, which exerts antiapoptotic roles (51). MMP2 and BIRC7 expression levels were detected in SNU-398 cells with ADORA2A-AS1 overexpression or silencing. As shown in Figures 6C, D , mRNA and protein levels of MMP2 and BIRC7 were both decreased in SNU-398 cells with ADORA2A-AS1 overexpression compared with control SNU-398 cells. Conversely, mRNA and protein levels of MMP2 and BIRC7 were both increased in SNU-398 cells with ADORA2A-AS1 silencing ( Figures 6E, F ). The RNA-seq data of HCC tissues from TCGA project showed that MMP2 expression level was negatively correlated with ADORA2A-AS1 expression level ( Figure 6G ). The negative correlation between MMP2 and ADORA2A-AS1 expression levels was also confirmed in our HCC cohort ( Figure 6H ). BIRC7 expression level was also negatively correlated with ADORA2A-AS1 expression level in HCC tissues, analyzed using the RNA-seq data from TCGA project ( Figure 6I ). Our HCC cohort further supported the negative correlation between BIRC7 and ADORA2A-AS1 expression levels ( Figure 6J ). These findings suggested that ADORA2A-AS1 repressed PI3K/AKT pathway via downregulating FSCN1.

To explore whether FSCN1 is the critical mediator of the roles of ADORA2A-AS1 in HCC, we stably overexpressed FSCN1 in SNU-398 cells with ADORA2A-AS1 stable overexpression ( Figure 7A ). CCK-8 assay results showed that FSCN1 overexpression reversed the decreased cell proliferation caused by ADORA2A-AS1 overexpression ( Figure 7B ). EdU incorporation assay results showed that FSCN1 overexpression reversed the decreased EdU-positive cell percentage caused by ADORA2A-AS1 overexpression ( Figure 7C ), indicating that FSCN1 overexpression rescued cell proliferation that was repressed by ADORA2A-AS1. Caspase-3 activity assay results showed that FSCN1 overexpression reversed the increased caspase-3 activity caused by ADORA2A-AS1 overexpression ( Figure 7D ), indicating that FSCN1 overexpression reversed cell apoptosis that was induced by ADORA2A-AS1. Transwell migration assay results showed that FSCN1 overexpression reversed the decreased migrated cell number caused by ADORA2A-AS1 overexpression ( Figure 7E ). Transwell invasion assay results showed that FSCN1 overexpression reversed the decreased invasive cell number caused by ADORA2A-AS1 overexpression ( Figure 7F ). These findings suggested that FSCN1 reversed the roles of ADORA2A-AS1 in repressing cell proliferation, inducing cell apoptosis, and repressing cell migration and invasion.

To further explore whether PI3K/AKT pathway is also the critical mediator of the roles of ADORA2A-AS1 in HCC, SNU-398 cells with ADORA2A-AS1 stable silencing were treated with PI3K/AKT pathway inhibitor LY294002. CCK-8 assay results showed that LY294002 reversed the increased cell proliferation caused by ADORA2A-AS1 silencing ( Figure 8A ). EdU incorporation assay results showed that LY294002 reversed the increased EdU-positive cell percentage caused by ADORA2A-AS1 silencing ( Figure 8B ), indicating that LY294002 reversed the increased cell proliferation induced by ADORA2A-AS1 silencing. Caspase-3 activity assay results showed that LY294002 reversed the reduced caspase-3 activity caused by ADORA2A-AS1 silencing ( Figure 8C ), indicating that LY294002 reversed the reduced cell apoptosis caused by ADORA2A-AS1 silencing. Transwell migration assay results showed that LY294002 reversed the increased migrated cell number caused by ADORA2A-AS1 silencing ( Figure 8D ). Transwell invasion assay results showed that LY294002 reversed the increased invasive cell number caused by ADORA2A-AS1 silencing ( Figure 8E ). These findings suggested that the oncogenic roles of ADORA2A-AS1 depletion in HCC were abolished by PI3K/AKT pathway inhibition.