CircRNA UBAP2 Serves as a Sponge of miR-1294 to Increase Tumorigenesis in Hepatocellular Carcinoma through Regulating c-Myc Expression

Background: Circular RNAs (circRNAs) are a class of regulatory RNAs with complex roles in healthy and diseased tissues. However, the oncogenic role of circRNAs in hepatocellular carcinoma (HCC) remains poorly understood, including the mechanisms by which the circRNA UBAP2 contributes to tumorigenesis. Methods: We analyzed the expression of circUBAP2 in 20 paired samples of HCC and healthy tissue as well as in seven HCC cell lines via quantitative real-time polymerase chain reaction (qRT-PCR). Functional experiments, such as CCK8 viability assays, colony formation assays, wound healing, transwell assays, and ow cytometry, were conducted to assess the effects of circUBAP2 in vitro. To further elucidate the mechanisms by which circUBAP2 acts, we conducted dual-luciferase assays, western blots, RNA pull-down assays, and rescue experiments. Results: circUBAP2 was highly upregulated in most HCC tissues and was associated with poor prognosis. HCC patients with high circUBAP2 expression had greater vascular invasion and worse differentiation. Functionally, circUBAP2 overexpression enhanced HCC cell proliferation, migration, and invasion and inhibited apoptosis. Furthermore, we found that circUBAP2 upregulated c-Myc expression by sponging miR-1294, thus contributing to hepatocarcinogenesis. Inhibiting circUBAP2 expression in HCC attenuated the oncogenic effects of c-Myc. Conclusions: These ndings suggest that circUBAP2 promotes HCC growth and metastasis. circUBAP2 may have value as an independent prognostic biomarker or as a new target for the treatment of HCC. EMT: Epithelial-mesenchymal transition; TMA: Tissue microarray analysis.


Background
Hepatocellular carcinoma (HCC) is responsible for about 75-85% of primary liver cancers, which are the fourth most common cause of cancer-related death around the world [1,2]. Despite great efforts of basic science and clinical research, HCC remains a highly deadly disease. Although HCC patients can be treated with surgery, tumor ablation, transarterial therapies, or systemic therapies (e.g. with tyrosine-kinase inhibitors or immune checkpoint inhibitors) [3], patients with advanced HCC lack effective options for therapeutic intervention, and prognoses are poor. A better understanding of the critical molecular events that occur in the pathogenesis of HCC may help uncover new biomarkers that can identify HCC patients who are at high risk of relapse. Further, a better understanding of the molecular process of disease may inform the rational design of effective therapeutics [4]. Circular RNAs (circRNAs), a type of noncoding, regulatory RNAs, are endogenous RNAs that have covalently linked ends [5]. These RNAs are mainly derived from the back-splicing of exons and are exceptionally stable due to their covalently closed ring structure [6][7][8]. Generally, circRNAs are expressed at low levels relative to their corresponding mRNAs and are primarily found in the cytoplasm of eukaryotic cells [9]. An increasing amount of evidence has suggested that circRNAs function as sponges that can protect speci c mRNAs from miRNA-mediated degradation and as scaffolds that facilitate the colocalization of enzymes and their substrates, indirectly regulating protein activity [9,10]. However, the function of many circRNAs remains unknown [11]. Moreover, some circRNAs have been implicated in the progression of diseases such as diabetes mellitus, chronic in ammatory diseases, cardiovascular diseases, and neurological disorders. Emerging studies have focused on the importance of circRNAs in solid tumors, wherein circRNAs are generally categorized as oncogenes or tumor suppressors. Studies of the molecular mechanisms of circRNAs in HCC are still in their infancy. Understanding the role of circRNAs in the tumorigenesis and progression of HCC may facilitate the identi cation of therapeutic targets or potential biomarkers.
In this study, we found that circUBAP2 was markedly upregulated in HCC compared to paired healthy liver tissues. Further clinical analysis revealed that HCC patients with higher circUBAP2 expression tended to have worse prognoses. Additionally, functional studies suggested that circUBAP2 promoted tumorigenesis and progression in liver cancer by altering the circUBAP2/miR-1294/c-MYC signaling pathway. In summary, we identi ed circUBAP2 as an important regulator of the growth and metastasis of HCC, making it a promising therapeutic target for the management of HCC.

Materials And Methods
Patients and follow-up HCC samples were collected from 125 patients diagnosed with HCC based on World Health Organization criteria at the Liver Cancer Institute, Zhongshan Hospital, Fudan University (Shanghai, China). All patients underwent curative resection at Zhongshan Hospital between February 2014 and January 2015. Followups were performed after surgery and were completed by December 2019. Additionally, 20 pairs of HCC tissue and adjacent liver tissues were randomly selected to compare relative circUBAP2 levels. All collection of HCC tissues was performed according to a protocol approved by the institutional review board of Zhongshan Hospital, and each patient provided informed consent.

Cell lines and culture conditions
The human HCC cell lines used in these studies included MHCC97H, MHCC97L,HCCLM3, Huh-7, PLC/PRF/5, Hep-G2, and Hep-3B. The MHCC97H, MHCC97L, and HCCLM3 cell lines were developed at our institute. Huh-7 and PLC/PRF/5 cell lines were purchased from the Chinese Academy of Sciences.
Hep-G2, and Hep-3B cell lines were purchased from the American Type Culture Collection and preserved at our institute. All cell lines were maintained in Dulbecco's modi ed Eagle's medium (DMEM) supplemented with 10% FBS in a humidi ed incubator at 37°C with 5% CO 2 , as previously described.
CircRNA pull-down assay CircRNA pull-down assays were conducted in accordance with the manufacturer's instructions. Biotinlabeled circUBAP2 and control probes were designed and synthesized by Sangon Biotech. Brie y, cell extracts were xed with 1% formaldehyde, lysed, and sonicated. Then, each cell extract sample was centrifuged, and 50 μL of supernatant was retained. Next, the streptavidin-coupled Dynabeads were added to the supernatant and incubated at 25°C for 1 hour. The complexes were then washed and incubated with lysis buffer and proteinase K. Finally, the RNA isolated from the pull-down bead sample was studied via qRT-PCR.

RNA immunoprecipitation (RIP)
RNA immunoprecipitation experiments were performed using an EZ-Magna RIP Kit in accordance with the manufacturer's instructions (Millipore, Billerica, MA, USA). Brie y, about 2 × 10 7 cells were collected and lysed in RIPA Lysis Buffer that contained proteinase and RNase inhibitors. Then, samples were incubated with the anti-AGO2 antibody (ab3238, Abcam) or IgG primary antibody (Cell Signaling Technology, USA). Twenty-four hours later, the complexes were washed ve times with cold PBS. Finally, the immunoprecipitated RNAs were extracted and qRT-PCR was performed to detect RNA enrichment.
Luciferase assays MHCC97H and HCCLM3 were cultured at 37°C with 5% CO 2 and were co-transfected with miR-1294 mimic, a luciferase reporter plasmid, and corresponding controls using Lipofectamine 2000. Forty-eight hours after transfection, luciferase activity was quanti ed using a Dual Luciferase Reporter Assay Kit (Yeasen) according to the manufacturer's protocol.
RNA extraction, qRT-PCR and western blot analysis RNA isolation and qRT-PCR were performed as previously described [12]. Brie y, total RNA was extracted with Trizol reagent (Thermo Fisher, US). cDNA was synthesized using a PrimeScript RT reagent kit (Takara, Japan). Target genes were quanti ed using SYB Premix Ex Taq II (Takara, Japan). DNA ampli cation was performed using an ABI 7900 system. Western blots (WB) and immunohistochemistry staining were also performed as previously described [13].
Cell proliferation,colony formation, and invasion assays To analyze cell proliferation and viability, we used Cell Counting Kit-8 (CCK-8) assays (Dojindo, Kumamoto, Japan) according to the manufacturer's instructions. Brie y, HCC cells were seeded into wells of 96-well plates (1 × 10 3 cells per well) and cultured at 37°C and 5% CO 2 . Plates were monitored at the indicated times at an absorbance of 450 nm [14].
Colony formation assays were conducted as previously described [15]. Brie y, HCC cells (1 × 10 3 cells per well) were aliquoted into 6-well plates and cultured for 14 days. Colonies were then xed with 100% methanol and stained with Giemsa staining solution (Sigma, USA). Visible colonies were counted using Image-Pro Plus 5.0 (Media Cybernetics, USA).
Invasion assays were conducted using transwell plates (Millipore, USA) coated with Matrigel (BD Biosciences, USA). DMEM containing 1% FBS was added to the upper chamber and DMEM containing 10% FBS was added to the lower chamber. HCC cells (2 × 10 5 per well) from each treatment group were added to the upper chambers. After 48 hours of incubation, cells that invaded the lower chamber surface were xed, stained with crystal violet (Sigma), and manually counted. All functional experiments were conducted in triplicate.

Xenograft mouse model
All in vivo experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals, as published by the US National Institutes of Health. All animal protocols were approved by the Animal Care Committee of Zhongshan Hospital. Brie y, 6-week-old nude mice, purchased from Charles River Laboratories (Beijing, China), were subcutaneously injected with 3 × 10 6 tumor cells. Tumor volumes and tumor weights were measured every 3 days. After 6 weeks, all mice were sacri ced for subsequent analyses.

Statistical analysis
All data were shown as the mean ± standard deviation of three independent experiments and were analyzed using SPSS 20.0 (SPSS, Chicago, IL, USA). We used t-tests, ANOVA, Mann-Whitney, chi-squared, and Fisher's exact tests to analyze differences between groups when appropriate. Time to reccurence (TTR) and overall survival (OS) were analyzed using Kaplan-Meier survival curves and log-rank tests, respectively. We also conducted univariate and multivariate analyses using Cox proportional hazard regression models. P < 0.05 was considered statistically signi cant.

CircUBAP2 was upregulated in HCC tissues and indicated poor prognosis
To evaluate the relationship between circUBAP2 and outcomes among patients with HCC, we leveraged the StarBase database to perform bioinformatic analyses of the expression of circUBAP2 in HCC tissues. We found that circUBAP2 expression was signi cantly higher in HCC tissues relative to healthy liver tissues (Fig. 1B). In order to con rm the ndings from the public database, we compared the relative expression of circUBAP2 in 20 paired formalin-xed, para n-embedded tissue samples of liver cancer and adjacent healthy liver tissues via qRT-PCR (Fig. 1C). To investigate the diagnostic and prognostic value of circUBAP2 in HCC, we next measured the expression of circUBAP2 in 125 HCC tissues. Unexpectedly, we found that circUBAP2 was associated with markedly worse OS (p < 0.001) and TTR (p < 0.001) after surgical resection ( Fig. 1D-E).
We investigated the association between the expression of circUBAP2 in HCC tissues and clinicopathological parameters through univariate and multivariate Cox proportional regression analyses. Multivariate subgroup analyses showed that high circUBAP2 expression was an independent factor for predicting both TTR [ Indeed, patients with higher circUBAP2 expression were more likely to have microvascular invasion (MVI) and worse differentiation (Tables 2-3).
We measured the relative expression of circUBAP2 in seven different characteristic HCC cell lines (Fig.   1F). In good agreement with our ndings in patient samples, we found that circUBAP2 expression was higher in the high-grade malignancy cell lines HCCLM3 and MHCC97H relative to MHCC97L, a less malignant cell line, suggesting the cicrUBAP2 may facilitate the progression of HCC. Thus, we considered HCCLM3 and MHCC97H cells to be suitable for further experiments. These results collectively suggested that circUBAP2 is of signi cant importance in the pathogenesis of HCC.

CircUBAP2 promotes HCC cell proliferation and induces cell apoptosis
To investigate the biological effects of circUBAP2 in HCC cells, we generated two circUBAP2-speci c shRNAs to target the back-splice sequence. As expected, these shRNAs led to downregulation of circUBAP2 expression in HCCLM3 and MHCC97H cell lines, as validated by qRT-PCR ( Fig. 2A). We next performed a series of functional experiments. Here, we found that reduced circUBAP2 expression markedly suppressed cell proliferation (CCK8 assay, Fig. 2B-C). Consistently, ow cytometry assays showed that downregulation of circUBAP2 led to increased apoptosis in circUBAP2 high (Fig. 2D-E).
To further evaluate the function of circUBAP2 in vivo, we generated a xenograft tumor model using NOD/SCID/γc(null) (NOG) mice. We found that tumor volumes and weights were decreased in mice with down-regulated circUBAP2 relative to the control group (Fig. 2F). Further, the rate of tumor growth was signi cantly decreased (Fig. 2F).

CircUBAP2 promotes HCC cell migration and invasion
Pathological epithelial-mesenchymal transition (EMT) is a signi cant stage in cancer progression, in which cancer cells gain invasive potential [16]. Many important drivers of EMT, such as SNAIL1, bind to and suppress the activity of E-cadherin promoters and are correlated with increased chance of relapse and decreased survival in patients with HCC [17]. Notably, we found that, after knockdown of circUBAP2, HCCLM3 and MHCC97H had decreased migration in both transwell and wound healing assays ( Fig. 3A-D). Moreover, in vitro transwell assays found decreased invasion of cells in the circUBAP2-KD group compared to the control group. Collectively, the above experiments con rmed that circUBAP2 promotes migration and invasion of HCC cells in vitro. To identify whether EMT was at the root of these malignant phenotypic changes, qRT-PCR and WB assays were performed to analyze changes in epithelial markers such as E-cadherin and mesenchymal markers such as N-cadherin and SNAIL1 (Fig. 3E-G). We found that circUBAP2-KD cell lines had decreased expression of N-cadherin and SNAIL1 compared with the mocktransfected cell lines. In accordance with the results from functional and molecular experiments, we found that circUBAP2 was associated with EMT and promoted migration and invasion of HCC cells.
circUBAP2 upregulated c-MYC expression by sponging miR-1294 Fluorescence in situ hybridization (FISH) and nuclear-plasma extraction assays suggested that circUBAP2 was primarily localized in the cytoplasm of the two HCC cell lines (Fig. 4A-B). A number of prior studies have reported that some circRNAs act as microRNA sponges. Accordingly, we conducted dual-luciferase reporter assays to explore this possibility for circUBAP2. We constructed a luciferase report vector of wild-type (WT) and mutant (Mut) circUBAP2 and transfected that vector into both HCCLM3 and MHCC97H cell lines (Fig. 4C). After transfection, we found that overexpression of the miR-1294 mimic signi cantly reduced luciferase activity relative to the negative control (Fig. 4F). Furthermore, we performed RNA immunoprecipitation (RIP) studies with anti-Ago2 antibodies in the same two HCC cells and found that endogenous circUBAP2 was enriched (Fig. 4D). The results of a microRNA pull-down assay using biotin-labeled miR-1294 mimics showed signi cant enrichment of circUBAP2 compared with the control group (Fig. 4E). Meanwhile, a circRNA pull-down assay using a probe speci c for circUBAP2 detected a clear increase in miR-1294 (Fig. 4D-E). Taken together, these results suggested that circUBAP2 functions as a sponge for miR-1294.
Next, we sought to determine the regulatory relationship and binding properties between miR-1294 and c-MYC through qRT-PCR (Fig. 4G), dual-luciferase reporter assays (Fig. 4H), and WB (Fig. 4I). We found that circUBAP2-KD cells had increased miR-1294 levels (Fig. 4G). Conversely, overexpression of miR-1294 reduced the expression of c-MYC protein (Fig. 4I). This result was con rmed by the reduced activity of a WT LUC-c-MYC reporter gene (Fig. 4H), however overexpression of miR-1294 had no effect on the activity of a LUC-c-MYC-mutant reporter gene (Fig. 4H). Collectively, these studies suggested that circUBAP2 may function as a sponge of miR-1294 to upregulate c-MYC expression.
miRNA inhibitor and overexpression of mRNA prevented the effects of circUBAP2 in liver cancer cells.
To further illustrate the regulatory interactions between circUBAP2 and miR-1294 that affect the c-MYC signaling pathway, we used a miR-1294 inhibitor to attenuate the induced expression in the initially circUBAP2 knockdown cells. qRT-PCR results con rmed that, as expected, addition of the miR-1294 inhibitor decreased the expression of miR-1294 in both MHCC97H and HCCLM3 cells (Fig. 5B). Of interest, we found that treatment with the miR-1294 inhibitor partially rescued the expression of c-MYC mRNA and protein relative to the control group ( Fig. 5C-D). However, qRT-PCR analysis found that neither the addition of miR-1294 inhibitor nor the overexpression of c-MYC affected the expression of circUBAP2 (Fig. 5A).
Functional experiments revealed that the addition of miR-1294 inhibitor reversed the circUBAP2knockdown-mediated increase in apoptosis and decrease in the proliferation, migration, and invasion in HCCLM3 and MHCC97H cells (Fig. 5E-H). Of note, the overexpression of c-MYC led to a reversal of the suppressive effects caused by circUBAP2 knockdown. circUBAP2-KD cells overexpressing c-MYC had signi cantly increased rates of proliferation and apoptosis, as determined by CCK8 and ow cytometry assays (Fig. 5E-F). Transwell migration and invasion assays showed attenuated motility and invasive potential in circUBAP2-KD cells, however these effects were partially reversed by overexpression of c-MYC (Fig. 5G-H). In summary, we determined that the addition of miR-1294 inhibitor and the overexpression of c-MYC reversed the oncogenic effects of circUBAP2 in HCC cells. mRNA overexpression abolished the effects of miRNA on liver cancer cells.
In order to understand the exact interactions between miR-1294 and c-MYC, we transfected a miR-1294 mimic into MHCC97H and HCCLM3 cell lines. The effects of the miR-1294 mimic were con rmed by qRT-PCR in those two cell lines compared with a scrambled control miRNA mimic. Next, qRT-PCR was combined with WB assays to show that the overexpression of miR-1294 in MHCC97H and HCCLM3 cells led to decreased c-MYC expression (Fig. 6B-C). However, c-MYC overexpression did not alter miR-1294 expression in the same two cell lines (Fig. 5B). Therefore, we next tested whether c-MYC overexpression would attenuate the effects of miR-1294 in HCC cells. Indeed, overexpression of miR-1294 inhibited proliferation, migration, and invasion in MHCC97H and HCCLM3 cell lines (Fig. 6E-G). Conversely, when we also introduced c-MYC into the HCC cells, the tumor suppressive role of miR-1294 was signi cantly weakened. These ndings matched with the nding of decreased cell viability, migration, and invasion ( Fig. 6E-G). In summary, overexpression of c-MYC reversed the tumor suppressive role of miR-1294.

Discussion
Although the approach to management of HCC has changed dramatically over the last 5 years, outcomes remain poor. This may be because the main molecular drivers behind the transformation and progression of HCC have not yet been recognized [18]. CircRNAs are relatively new members of the noncoding cancer genome that have gene-regulatory potential and have been reported to drive malignant transformation and progression [19]. In recent years, there has been major progress in our understanding of the biogenesis and roles of circRNAs in HCC. An increasing number of studies indicate that circRNAs can play both oncogenic and tumor suppressive roles in HCC [20,21]. Nevertheless, the vast majority of circRNAs in HCC have unknown functions, and the list of circRNAs involved in HCC continues to grow at a steady pace [19]. Previous studies have reported that circUBAP2 is important in the tumorigenesis and progression of esophageal squamous cell carcinoma [22], pancreatic adenocarcinoma [23], ovarian cancer [24], and osteosarcoma [25]. In this study, we found that circUBAP2 was aberrantly expressed in HCC tissues and was positively correlated with vascular invasion and worse differentiation. circUBAP2 functions as a sponge to inhibit the function of miR-1294, thus promoting HCC proliferation, invasion, and migration through the c-Myc signaling pathway. These ndings suggested that circUBAP2 may hold promise as a biomarker or target for the management of HCC. However, further prospective cohort studies are required to validate the clinical value of circUBAP2.
Previous studies have reported that miR-1294 plays a critical role in osteosarcoma [26], gastric cancer [27], esophageal cancer [28], ovarian cancer [29], pancreatic ductal adenocarcinoma [30], clear cell renal cell carcinoma [31], and many other types of cancers. However, no prior studies have examined the role of miR-1294 in HCC. In this work, we found that knockdown of circUBAP2 enhanced the expression of miR-1294, and that upregulation of miR-1294 inhibited the growth and metastasis of HCC in vitro and in vivo. Moreover, we validated that the c-Myc oncogene was the target of miR-1294 through dual-luciferase reporter assays. Overexpression of c-Myc resulted in a reversal of the effects of miR-1294 in HCC cells. Future work will be needed to validate the clinical impacts of miR-1294 in HCC.
It has been rmly established that c-Myc acts as a pleiotropic transcription factor to control cell proliferation, apoptosis, metabolism, adhesion, DNA replication, differentiation, and angiogenesis [32][33][34]. The c-Myc oncogene is broadly overexpressed in many late-stage cancers and is often associated with tumorigenesis by causing inappropriate gene expression [35][36][37]. An increasing number of reports have recently focused on the role of c-Myc in HCC [38]. Mechanistically, our results suggested that knockdown of circUBAP2 led to a signi cant decrease in the expression of c-Myc in HCC and inhibited cellular DNA synthesis. Further, rescued expression of c-Myc in circUBAP2-KD cells promoted cellular proliferation and inhibited apoptosis. Previous reports have shown that c-Myc is a crucial factor in driving the transition from G0/G1 to the S phase in hepatocytes [39,40]. We believe that our results suggest that circUBAP2 regulates HCC cell survival by affecting the c-Myc signaling pathway. Thus, inhibiting the expression of circUBAP2 may abolish the tumorigenesis of c-Myc, a previously undruggable target[41].

Conclusions
These experiments suggested that circUBAP2 has oncogenic potential in HCC. Speci cally, circUBAP2 acts as a sponge of miR-1294 to upregulate c-Myc expression, which subsequently contributes to the activation of EMT signaling pathways and consequently promotes tumorigenesis and progression. Of note, circUBAP2 expression was a signi cant independent prognostic factor for TTR and OS. These ndings may provide additional insights into the molecular events responsible for hepatocarcinogenesis. Further prospective studies will help determine the potential value of circUBAP2 as a prognostic indicator or therapeutic target for patients with HCC.      qRT-PCR. (G) An RNase R treatment assay was conducted. Although circUBAP2 showed strong exonuclease resistance, UBAP2 did not. Data are shown as the mean ± standard deviation of three independent experiments, and t-tests were used to compare group averages. **p < 0.01; ***p < 0.001. or loss-of-function on in vivo HCC tumor size. Data are shown as the mean ± standard deviation of three independent experiments, and t-tests were used to compare group averages. **p < 0.01; ***p < 0.001.   CCK8 assays found that an miR-1294 inhibitor reversed the effect of circUBAP2 knockdown on HCC proliferation. (F) Flow cytometry studies found that miR-1294 inhibitor attenuated the effects of circUBAP2 knockdown on HCC viability. Bar graphs represent the rate of apoptosis in treated cells. (G-H) MHCC97H and HCCLM3 cells were used to evaluate the effects of the miR-1294 inhibitor on HCC cell migration and invasion after circUBAP2 knockdown. Data are shown as the mean ± standard deviation of three independent experiments, and t-tests were used to compare group averages. **p < 0.01; ***p < 0.001.