Circular RNA hsa_circ_0003141 promotes tumorigenesis of hepatocellular carcinoma via a miR-1827/UBAP2 axis

Circular RNAs (circRNAs) play an important role in the tumorigenesis of hepatocellular carcinoma (HCC), but their specific functions in HCC remain largely unknown. Using bioinformatics analysis, we have found that the expression of circRNA hsa_circ_0003141 is significantly increased in HCC tissues. Ubiquitin-associated protein 2 (UBAP2) is the parent gene for hsa_circ_0003141, and its high expression correlates with poor overall survival rates in HCC patients. In addition, our results show that miR-1827 is a binding target of hsa_circ_0003141, and indicate that hsa_circ_0003141 regulates UBAP2 expression by sponging miR-1827 in HCC cells. Downregulation of hsa_circ_0003141 suppresses UBAP2 expression, induces apoptosis, and inhibits proliferation and invasion by HCC Huh-7 cells. Importantly, downregulation of hsa_circ_0003141 inhibits tumorigenesis in a xenograft mouse model of HCC. Together, our results indicate that hsa_circ_0003141 functions as an oncogene in HCC cells, and suggest that the hsa_circ_0003141/miR-1827/UBAP2 axis might represent a novel therapeutic option for the treatment of HCC.

In the present study, we analyzed two GEO datasets to identify differentially expressed circRNAs (DEcircRNAs) between HCC tissues and matched normal tissues. We found that the circRNA hsa_circ_ 0003141 is significantly increased in HCC tissues, and promotes HCC tumorigenesis.

GO and KEGG analysis of DEcircRNAs
Next, the DEcircRNAs were analyzed using the gene ontology (GO) enrichment and KEGG pathway analyses. The GO results showed that DEcircRNAs were mainly enriched in "proteasome regulatory pathway", "proteasome-activating ATPase activity", and "one-carbon metabolic processes" (Figure 2A). The KEGG analysis showed enrichment in "one carbon pool by folate", "glycosylphosphatidylinositol (GPI)-anchor biosynthesis", "proteasome", and "cysteine and methionine metabolism" ( Figure 2B). In addition, the prognostic value of the overlapping circRNAs' parent genes was analyzed using the Kaplan-Meier method from TCGA data. Ubiquitin associated protein 2 (UBAP2) is the parent gene of hsa_circ_0003141. As shown in Supplementary Figure 1A and 1B, the level of UBAP2 was much higher in Hep3B2.1-7 and Huh-7 cells than that in HepG2 and HCCLM3 cells, thus, Huh-7 and Hep3B2.1-7 cells were chose to conduct the following experiments. In addition, the analysis of the TCGA dataset revealed that UBAP2 levels in 369 liver hepatocellular carcinoma (LIHC) tissues were higher than that of in 160 normal tissues ( Figure 2C). Moreover, high UBAP2 expression correlated with poor overall survival rates in patients with HCC ( Figure 2D), indicating that increased hsa_circ_0003141 levels could predict worse survival outcomes in HCC patients.

Downregulation of hsa_circ_0003141 inhibits proliferation of HCC cells
To investigate the role of hsa_circ_0003141 in HCC cells, we established HCC cell lines (Huh-7 and Hep3B2.1-7) with hsa_circ_0003141 stable knockdown. As shown in Figure 3A, the level of hsa_circ_0003141 was significantly downregulated in Huh-7 and Hep3B2.1-7 cells following transfection with hsa_circ_0003141-shRNA1. In addition, CCK-8 proliferation assay showed that downregulation of hsa_circ_0003141 suppressed proliferation of Huh-7 and Hep3B2.1-7 cells, compared with the shRNA-NC group ( Figure 3B and 3C). Moreover, silencing of hsa_circ_0003141 greatly induced apoptosis of Huh-7 and Hep3B2.1-7 cells, compared with the shRNA-NC group ( Figure 3D-3G). These data indicate that downregulation of hsa_circ_0003141 induces apoptosis and inhibits proliferation of HCC cells.

Downregulation of hsa_circ_0003141 inhibits invasion of HCC cells
To investigate the effect of hsa_circ_0003141 on invasion of HCC cells, transwell invasion assay was performed. As shown in Figure 4A and 4B, the invasion ability of Huh-7 cells was significantly reduced following transfection with hsa_circ_0003141-shRNA1. In addition, downregulation of hsa_circ_0003141 markedly increased the level of mesenchymal-toepithelial transition (MET) marker E-cadherin, and decreased the levels of epithelial-mesenchymal AGING transition (EMT) markers N-cadherin and α-SMA in Huh-7 cells ( Figure 4C-4F). These results suggest that downregulation of hsa_circ_0003141 inhibits the invasion ability of HCC cells via inhibiting the EMT process.

Hsa_circ_0003141 functions as a ceRNA of miR-1827 in HCC
Circular RNA interactome (https://circinteractome.nia. nih.gov) was used to identify the miRNAs interacting with hsa_circ_0003141. As shown in Figure 5A and 5B, the potential binding sites for miR-1827 in hsa_circ_0003141 were predicted, indicating that miR-1827 is a potential binding target of hsa_circ_0003141. RT-qPCR assay showed that miR-1827 mimics significantly increased the level of miR-1827 in Huh-7 cells, and miR-1827 inhibitor significantly decreased the level of miR-1827, compared with NC group ( Figure 5C). In addition, dual luciferase reporter assay showed that the luciferase activity was significantly lower in Huh-7 cells co-transfected with WT-hsa_circ_0003141 segment and miR-1827 mimics compared to vector-control group, while no significant difference was observed in Huh-7 cells co-transfected with MT-hsa_circ_0003141 segment and miR-1827 mimics ( Figure 5D). Moreover, FISH assay revealed that hsa_circ_0003141 and miR-1827 are co-localized in the cytoplasm ( Figure 5E). These data suggest that hsa_circ_0003141 may function as a competing endogenous RNA (ceRNA) of miR-1827 in HCC.

UBAP2 is a direct binding target of miR-1827
To identify the potential target of miR-1827, we used the TargetScan dataset (http://www.targetscan.org/ vert_71/). The data analysis indicated that ubiquitinassociated protein 2 (UBAP2) might be a potential target of miR-1827 ( Figure 6A). In addition, dual luciferase reporter assay showed that the luciferase activity was markedly lower in Huh-7 cells cotransfected with the WT-UBAP2 segment and miR-1827 mimics, compared to vector-control group ( Figure  6B), indicating that UBAP2 is a direct binding target of miR-1827.

Hsa_circ_0003141 functions as oncogene in HCC by sponging miR-1827, thus increasing UBAP2 expression
Next, we investigated the role of hsa_circ_0003141 and miR-1827 in the regulation of UBAP2 protein AGING expression. RT-qPCR assay showed that miR-1827 mimics significantly increased the level of miR-1827 in Hep3B2.1-7 cells, and miR-1827 inhibitor significantly decreased the level of miR-1827, compared with NC group (Supplementary Figure 2A). As shown in Figure  7A-7C, and Supplementary Figure  2B-2D, downregulation of hsa_circ_0003141 significantly decreased the level of UBAP2, and increased the level of cleaved caspase 3 in Huh-7 and Hep3B2.1-7 cells, compared with shRNA-NC group. However, when Huh-7 or Hep3B2.1-7 cells were treated with hsa_circ_0003141-shRNA1 together with miR-1827 inhibitor, the effect of hsa_circ_0003141 on UBAP2 protein expression and cleaved caspase 3 was reversed by miR-1827 knockdown. These data indicate that hsa_circ_0003141 promotes HCC tumorigenesis by sponging miR-1827, thus increasing UBAP2 expression.

Knockdown of hsa_circ_0003141 inhibits tumorigenesis in HCC xenografts in vivo
To investigate the role of hsa_circ_0003141 in the regulation of HCC tumor growth in vivo, we established a mouse Huh-7 subcutaneous xenograft model. As shown in Figure 8A-8C, downregulation of hsa_circ_0003141 significantly inhibited the tumor volume and tumor weight, compared with shRNA-NC group. In addition, the expression of UBAP2 was significantly decreased, and the expression of cleaved caspase 3 was increased in tumor tissues following transfection with hsa_circ_ 0003141-shRNA1; however, this was reversed after transfection with miR-1827 inhibitor ( Figure 8D-8F). These results indicate that hsa_circ_0003141 suppression inhibits the HCC tumorigenesis in vivo.  AGING DISCUSSION CircRNAs may serve as disease-specific biomarkers for disease diagnosis [20], and play a critical role in HCC [21]. For example, circ-FOXP1 inhibits proliferation and invasion of HCC cells via sponging miR-875-3p and miR-421 [22]. The study by Lin et al has indicated that hsa_circ_0091570 acts as a sponge for miR-1307, and suppresses HCC progression [23]. However, the specific function of most circRNAs in HCC tumorigenesis is unknown.
Wang et al only identified one overlapping DEcircRNA, hsa_circ_0000517, in GSE97332 and GSE94508 datasets using Venn diagram [20]. We reanalyzed these two datasets, whereas the thresholds for the screening of DEcircRNAs we defined were different with the previous study. Using bioinformatics analysis, we have identified a novel circRNA named hsa_circ_0003141, which is significantly upregulated in human HCC tissues. Our results demonstrate that downregulation of hsa_circ_0003141 inhibits proliferation and invasion of HCC cells, and induces their apoptosis. In addition, knockdown of hsa_circ_0003141 reduces EMT in HCC cells.
Recent studies have shown that circRNAs may function as competing endogenous RNAs (ceRNAs) to recruit miRNAs, thus leading to posttranscriptional downregulation of miRNA targets [24]. The circRNA-miRNA-mRNA axis may act as a network of gene expression regulators, and the dysregulated circRNAs may contribute to HCC progression [25]. For example, a previous study by Zhan et al has indicated that hsa_circRNA_103809 promotes the progression of HCC via regulating the hsa_circRNA_103809/ miR-377-3p/FGFR1 axis [26]. In this study, we have found that miR-1827 is a potential binding target of hsa_circ_0003141 in HCC. Moreover, our data indicate that hsa_circ_0003141 functions as a ceRNA of miR-1827 in HCC, and that UBAP2 is a potential binding target of miR-1827. Yao et al indicated that overexpressed FZD4 could contribute to HCC progression, while overexpression of miR-1827 could suppress the expression of FZD4 in HCC cells, indicating that miR-1827 may display an antitumor AGING effect in HCC [27]. We have found that miR-1827 has a complementary sequence to both hsa_circ_0003141 and UBAP2. Since downregulation of hsa_circ_0003141 reduces UBAP2 expression, but this is reversed by miR-1827 inhibitor, these data indicate that hsa_circ_ 0003141 promotes HCC tumorigenesis by sponging miR-1827, resulting in the increased UBAP2 expression. Feng et al have shown that circRNA ITCH inhibits HCC progression by sponging miR-214, thus upregulating linear ITCH expression; this is consistent with our results [28]. Our current data show that downregulation of hsa_circ_0003141 inhibits the progression of HCC via regulating the hsa_circ_ 0003141/ miR-1827/UBAP2 axis.
Ubiquitin associated protein 2 (UBAP2) participates in the regulation of cellular process including apoptosis, metastasis, and tumor growth [29]. A previous study has suggested that UBAP2 acts as a tumor suppressor in HCC, since downregulation of UBAP2 promoted proliferation and invasion of HCC cells [29]. However, our results show that high UBAP2 expression correlates with poor overall survival in patients with HCC, indicating that UBAP2 serves as an oncogene in HCC. The oncogenic role of UBAP2 is supported by the study of Latonen et al that has shown that the expression of UBAP2 is significantly upregulated in prostate tumor samples [30].
In summary, our results demonstrate that the expression of hsa_circ_0003141 is significantly increased in human HCC tissues. Downregulation of hsa_circ_0003141 inhibits proliferation, invasion, and EMT in HCC cells in vitro, and reduces tumor growth in a mouse HCC xenograft model. Our data indicate that hsa_circ_0003141 promotes the HCC progression by sponging miR-1827, thus increasing the UBAP2 expression. Therefore, targeting the hsa_circ_0003141/miR-1827/UBAP2 axis might represent a novel therapeutic option for the treatment of HCC.

Data acquisition and identification of DEcircRNAs
The GSE97332 and GSE94508 datasets were downloaded from the Gene Expression Omnibus database (GEO, https://www.ncbi.nlm.nih.gov/geo/). GSE97332 dataset contains circRNAs expression data of 7 HCC tissues and 7 matched normal tissues. The GSE94508 dataset contains circRNAs expression data of 5 HCC tissues and 5 matched normal tissues. The DEcircRNAs in HCC and normal tissues were identified using R language. CircRNAs with adjusted P<0.001 and |log2 (FC)|>2 were considered as significant DEcircRNAs. The overlapping DEcircRNAs from the two datasets (GSE97332 and GSE94508) were identified using the VennDiagram package in R language.

Survival analysis
TCGA dataset (http://tcga-data.nci.nih.gov/tcga) was used to assess the prognostic value of the overlapping DEcircRNAs' parental genes between HCC tissues and matched normal tissues; P<0.05 was considered statistically significant.

Cell culture
Human hepatoma cell line Huh7 was purchased from Procell (Wuhan, China). Human hepatoma cell line Hep3B2.1-7 was purchased from Beyotime Biotechnology (Beijing, China). Human hepatocellular carcinoma cell line HepG2 was purchased from American Type Culture Collection (ATCC, Rockville, MD, USA). Human hepatocarcinoma cell line AGING HCCLM3 was obtained from the China Center for Type Culture Collection (CCTCC, Wuhan, China). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM, Thermo Fisher Scientific) with 10% fetal bovine serum (FBS, Hyclone, Logan, UT, USA) at 37°C in a 5% CO2 humidified incubator.
MiR-1827 mimics, miR-1827 inhibitor and NC, obtained from GenePharma (Shanghai, China), were transfected into Huh-7 and Hep3B2.1-7 cells using Lipofectamine 2000 (Thermo Fisher Scientific) respectively. The culture medium containing 10% FBS was changed 6 h after transfection, and the cells were incubated for another 48 h at 37˚C. Cells of the blank control group were not transfected.

Cell proliferation assay
Huh-7 and Hep3B2.1-7 cells (5000 cells/well) were seed onto 96-well plates and incubated overnight at 37 °C. Cells were then transfected with hsa_circ_0003141-shRNA1 and shRNA-NC for 0, 24, 48, and 72 h, and cell proliferation was measured using the Cell Counting Kit-8 (CCK8, Beyotime, Shanghai, China) according to the manufacturer's protocol. Absorbance was measured at a wavelength of 450 nm using a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Cell apoptosis assay
Huh-7 and Hep3B2.1-7 cells were washed with cold PBS, and then incubated in 75% ethanol at -20°C overnight. Cells were stained with 5 μL of propidium iodide (PI) and Annexin V-FITC (Thermo Fisher Scientific) for 20 min in dark at room temperature (RT), and apoptotic cells were analyzed by a flow cytometer (FACScan™, BD Biosciences, Franklin Lakes, NJ, USA).

Transwell invasion assay
Transwell invasion assay was performed using 24-well transwell chambers (0.8 μm; Corning New York, NY, USA) with Matrigel coating. Huh-7 cells (4 x 10 4 ) were suspended in 200 μL of serum-free DMEM, and seeded onto the upper chamber. After that, 600 μL of DMEM medium supplemented with 10% FBS was added into the lower chamber. Cells that invaded through the membrane were fixed with 4% paraformaldehyde, stained with 0.2% crystal violet, and photographed using a laser confocal microscope (Olympus CX23 Tokyo, Japan).

Dual-luciferase reporter assay
The binding sequences of miR-1827 in hsa_circ_0003141 were cloned into psiCHECK-2 vector (Applied Biosystems, USA). After that, cells were co-transfected with WT-hsa_circ_0003141 or MT-hsa_circ_0003141 plasmid, with miR-1827 mimics respectively using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Moreover, cells were co-transfected with WT-hsa_circ_0003141 or MT-hsa_circ_0003141 plasmid, with mimics NC using Lipofectamine 2000, which was considered as the vector-control group. Meanwhile, cells were transfected with WT-hsa_circ_0003141 or MT-hsa_circ_0003141 plasmid, which was considered as the control group. For UBAP2 reporter assay, UBAP2 segment was synthesized with either wild-type (WT) or mutant (MT) seed region and cloned into the psiCHECK-2 vector. Cells were then co-transfected with WT-UBAP2 or MT-UBAP2 plasmid, with miR-1827 mimics or mimics NC respectively using Lipofectamine 2000. After 48 h, the Dual Luciferase Reporter Assay System (Promega, Madison, USA) was used to detect luciferase activity with renilla luciferase activity as endogenous control.

Fluorescence in situ hybridization analysis
The oligonucleotide probe for hsa_circ_0003141 was synthesized by Biosense (Guangzhou, China). Huh-7 cells were incubated with 10 μL of the probe mixture (Biosense) at 37°C in the dark overnight. Cell nuclei were stained 20 min with 4,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific) for 20 min, and visualized using a fluorescence microscope (Olympus) as previously described [19].

Animal study
BALB/c nude mice (n = 18, 6 -8 weeks old) were obtained from the Charles River (Beijing, China). Animals were randomized into three groups: blank, shRNA-NC, and hsa_circ_0003141-shRNA1 group. 5 x 10 6 Huh-7 or Huh-7 cells stably expressing shRNA-NC, hsa_circ_0003141-shRNA1 cells in 100 μL PBS were injected into the left flank of nude mice. The tumor volume was calculated by the following formula: V = (length x width 2 )/2. After 5 weeks, animals were sacrificed under anesthesia by following the recommended procedures of the National Institutes of Health guide for the care and use of laboratory animals. All animal experiments were approved by the Institutional Animal Care and Use Committee at the Union Hospital, Tongji Medical College, Huazhong University of Science and Technology.

Statistical analysis
All experiments were repeated three times. Data are presented as mean ± standard deviation (S. D.). Graphs were generated using GraphPad Prism software (version 7.0, La Jolla, CA, USA). One-way analysis of variance (ANOVA) and Tukey's tests were carried out for multiple group comparisons. A Pvalue < 0.05 was considered as statistically significant.