E2F1-Induced lncRNA BAIAP2-AS1 Overexpression Contributes to the Malignant Progression of Hepatocellular Carcinoma via miR-361-3p/SOX4 Axis

Currently, plenty of researches have revealed that long noncoding RNAs (lncRNAs) can act as crucial roles during the progression of various tumors, including hepatocellular carcinoma (HCC). Here, we measured the expression of lncRNA BAIAP2 antisense RNA 1(BAIAP2-AS1) as well as its contribution to the developments of HCC. In this study, the expressions of BAIAP2-AS1 and SOX4 were distinctly upregulated in HCC cells and tissues, and high BAIAP2-AS1 may be a novel biomarker for HCC. E2F1 activated BAIAP2-AS1 expression. The silence of BAIAP2-AS1 inhibited the proliferation and metastasis of HepG2 and PLC5 cells. Assays for relationship verification showed that BAIAP2-AS1 regulated the expression of SOX4 and miR-361-3p. Rescue experiments further confirmed the positive interaction between miR-361-3p and BAIAP2-AS1 as well as between miR-361-3p and SOX4. Overall, BAIAP2-AS1 modulated the miR-361-3p/SOX4 axis to promote the development of HCC. Thus, our study offers a potential therapeutic target for treating HCC.


Introduction
Hepatocellular carcinoma (HCC) is a primary malignancy of the liver, now ranking the sixth most common malignant tumor and causing an estimated over half a million deaths annually [1,2]. It is more predominant in parts of Asia and Central Africa, and approximately 50 percent of all new cases are diagnosed in China [3]. Despite several advances in diagnostic strategies, chemotherapy, radiotherapy, and surgery, the survival time of many HCC patients remains poor due to drug resistance, tumor relapse, and late diagnosis [4,5]. Thus, the molecular mechanisms regarding HCC initiations and developments deserve further exploration.
The vast majority of human transcriptome could be classified as noncoding RNAs based on the advancements of high-throughput RNA sequencing technology [6]. As a class of noncoding RNAs, long noncoding RNAs (lncRNAs) have a length of >200 nucleotides (nt) [7]. Multiple model organisms have been used for the annotation of lncRNAs, and the results suggest that they are often expressed in a spatial-and temporal-specific pattern [8,9]. In recent years, emerging studies have suggested lncRNA involvements in diverse physiological and pathological processes, such as programmed cell death, cellular growth, transport through nuclear pore, stem cell pluripotency, and developments, via serving as epigenetic regulators at transcriptional, posttranscriptional, or epigenetic levels [10,11]. Growing evidence also confirms that lncRNAs exhibit important regulatory effects in the progression of various tumors via a series of complex mechanisms [12,13]. However, the potential functions of many lncRNAs in HCC development and metastasis have not been elucidated.
In this research, lncRNA BAIAP2 antisense RNA 1 (BAIAP2-AS1) was identified as a novel HCC-related lncRNA. It was observed to be highly expressed in HCC. Then, for the first time, the clinical significance and tumor-related functions of BAIAP2-AS1 in HCC were investigated. This research gained new insights into the metastatic mechanism of HCC and provided HCC treatment with a promising therapeutic target.

Materials and Methods
2.1. Clinic Samples. Totally, 10 patients receiving surgery in the Weifang People's Hospital between July 2019 and August 2020 were adopted to obtain 10 paired HCC and corresponding adjacent normal specimens. Before sample collection, local or systemic treatments were not applied to all patients. All specimens were snap-frozen and stored at -80 for the extraction of total RNAs. Two experienced pathologists identify the specimen types. Inform consent of all patients was obtained before the surgery, together with the approval of the Ethics Committee of the Weifang People's Hospital for the study.

RNA Extraction and Quantitative
Real-Time PCR (RT-qPCR). For the collection of total RNA from specimens and cells, TRIzol reagent which was purchased from Life Technologies (Haidian, Beijing, China) was used. Reverse EasyScript One Step gDNA Removal and cDNA Synthesis SuperMix (TAKALA, Hangzhou, Zhejiang, China) were applied to synthesize first-strand cDNA. ABI 7300 Sequence Detection System (Nanjing, Jiangsu, China) was applied for the performance of RT-qPCR. GAPDH or U6 snRNA was used for the endogenous control. Moreover, 2 −ΔΔCt methods were used to calculate the fold change in gene expressions. Primers were designed by Primer Express 3.0. Table 1 lists the sequences. . The CCK-8 solution (Dojindo, Gaithersburg, MD) was used for the cellular proliferation assays. First, cells were seeded in triplicate into 96-well plates at a concentration of 3 × 10 3 cells/well. Then, 15 μl/well of CCK-8 solution was applied for the treatments of the collected cells during the ultimate 4 h of culture. A microplate reader was applied to measure optical density.

Cell Counting
2.6. Colony Formation Assays. Cells were plated in six-well plates at a density of 600 cells per well and were cultured with RPMI 1640 (Sigma-Aldrich, Shenzhen, Guandong, China) supplemented with ten percent FBS (Cyagen Technology, Taicang, Jiangsu, China) for ten days. When the incubation period finished, PBS was used to wash the cells, which then were fixed in methanol and dyed by crystal violet (Tsbiochem, Jingan, Shanghai, China). Three independent experiments were performed.

EdU
Staining. The cell culture medium was added by EdU assay kit for three hours. Then, a Cell-Light™ EdU Apollo® 489 in vitro Imaging Kit was applied for the stain of the cells. DAPI staining was conducted on the cell nucleus without light. An inverted fluorescence microscope was used for the observation.      2.11. Chromatin Immunoprecipitation (ChIP) Assays. ChIP assays were performed with a Magna ChIP Kit (Millipore, Hangzhou, Zhejiang, China). To generate DNA-protein cross-links, formaldehyde was applied to HepG2 and PLC5 cells. Chromatin fragments of 200-300 bp were generated by sonication of cell lysates. It was followed by immunoprecipitation of lysates with specific antibodies. In addition, IgG was used as the control. For the examination of the precipitated chromatin DNA, RT-qPCR was carried out.

Luciferase Reporter Assays.
To predict the E2F1binding site in the promoter of BAIAP2-AS1, the JASPAR database was searched [19]. Then, we synthesized the diverse fragment sequences of the promoter of BAIAP2-AS1. The vectors generated by inserting the above sequences into the pGL3-basic vector were cotransfected with E2F1 expression plasmid into HepG2 and PLC5 cells.

Disease Markers
TTBS was used to wash the blots for six times. Then, the Beyo ECL Plus reagent was applied to develop the blots, and the Gel Imaging System was used to record the images. Abcam (Pudong, Shanghai, China) provided all antibodies. Normalization of data was conducted for expressions of GAPDH (Rabbit anti-GAPDH).
2.14. Statistical Analysis. The SPSS statistical software package (standard version 18.0, SPSS Inc., Chicago, IL, USA) was applied to statistical analysis, and mean ± SD was used as a form of values. Unless otherwise stated, the statistical significance for comparisons of 2 or more groups was identified by one-way ANOVA and Student's t-test. P < 0:05 was considered statistically significant.

BAIAP2-AS1 Was Upregulated in HCC.
We searched GEPIA which further showed BAIAP2-AS1 levels displayed an increasing trend in HCC specimens (n = 369) compared with normal liver specimens (n = 160) (Figure 1(a)). Moreover, Pan-cancer assays revealed that overexpression of BAIAP2-AS1 was a frequent event in many types of tumors (Figure 1(b)). To confirm the above results, we performed RT-qPCR in our cohort, finding that compared to matched nontumor specimens, the expressions of BAIAP2-AS1 were obviously decreased in HCC specimens (Figure 1(c)). We also observed high expression of BAIAP2-AS1 in four HCC cell lines compared to LO2 cells (Figure 1(d)). Besides, Anti-lgG      9 Disease Markers in cancers; nonetheless, the factors regulating these molecules' misregulation remain unclear [20,21]. More and more researchers found proof for the contribution of some crucial transcription factors to lncRNA dysregulation in human cancer cells [22,23]. To examine the mechanism of BAIAP2-AS1 upregulation in HCC, this research explored the promoter region of BAIAP2-AS1 by online bioinformatics software programs JASPAR, revealing six potential sites of E2F1 binding (Figure 2(a)). Then, the expressing pattern of E2F1 in HCC was determined by UALCAN, finding increased E2F1 expression in HCC specimens, especially in those with advanced stages (Figures 2(b) and 2(c)). Based on TCGA datasets, we also observed that E2F1 expression is positively associated with E2F1 in HCC specimens (Figure 2(d)). We also showed the increased levels of E2F1 in five HCC cell lines compared with LO2 cells (Figure 2(e)). To study whether the dysregulation of E2F1 may influence the expression of BAIAP2-AS1, we overexpressed or downregulated the expression of E2F1 and performed RT-qPCR which showed that BAIAP2-AS1 levels were distinctly decreased after the knockdown of E2F1, while E2F1 overexpression displayed an opposite result (Figures 2(f) and 2(g)). Clear E2F1-binding activity was implied by subsequent ChIP assays on the endogenous

BAIAP2-AS1 Knockdown Suppressed the Proliferation and Metastasis of HCC Cells.
For examining the biological influence of BAIAP2-AS1 on HCC growth, we transfected HepG2 and PLC5 cells with sh-BAIAP2-AS1-1, BAIAP2-AS1-2 or sh-NC. RT-qPCR confirmed a distinct reduction in the expression level of BAIAP2-AS1 in HepG2 and PLC5 cells treated with sh-BAIAP2-AS1-1 and BAIAP2-AS1-2 (Figure 3(a)). Consequently, the results of CCK-8 assays revealed that cell viability was significantly decreased in sh-BAIAP2-AS1-1 and BAIAP2-AS1-2 transfected cells than in sh-NC transfected cells (Figure 3(b)). Also, the clonogenic survival of HepG2 and PLC5 cells was shortened by BAIAP2-AS1 knockdown as found from colony formation assays (Figure 3(c)). Edu assays also confirmed the BAIAP2-AS1 knockdown suppressed HCC cells' proliferation (Figure 3(d)). Furthermore, cell invasion abilities were investigated by transwell invasion. Compared to the sh-NC group, the invasion capacities of HepG2 and PLC5 cells transfected with sh-BAIAP2-AS1-1 or sh-BAIAP2-AS1-2 were significantly decreased (Figure 4(a)). To explore the mechanisms involved in the effects of BAIAP2-AS1 on metastasis capacities of HCC cells, we observed that silence of BAIAP2-AS1 promoted the expression of E-cadherin while suppressing the expression of N-cadherin and vimentin in HepG2 and PLC5 cells (Figure 4(b)). Based on the above results, BAIAP2-AS1 was concluded as a tumor promoter in HCC.

BAIAP2-AS1
Regulates miR-361-3p to Modulate SOX4 in HCC Cells. The influence of BAIAP2-AS1 on miR-361-3p expression was revealed above. Afterward, the function was taken into account. SOX4's involvements in HCC progression and the ability of miRNAs to target downstream genes have been reported previously [25,26]. In the 3 ′ UTR of SOX4, potential miR-361-3p binding sites were confirmed by bioinformatics tools (Figure 6(a)). The results of TCGA database revealed that SOX4 was highly expressed in HCC (Figure 6(b)). Moreover, a positive association between BAIAP2-AS1 expression and SOX4 expression was found from correlation analysis (Figure 6(c)). We also observed that both mRNA and protein levels of SOX4 expression were distinctly increased in HCC cells compared to LO2 cells (Figure 6(d)). The higher the SOX4 expression, the shorter the OS and DFS, as revealed by survival assays based on TCGA datasets (Figure 6(e)). Previously, SOX4 has been demonstrated to be an oncogene in several tumors, including HCC. Luciferase activity reporter assays were conducted for verifying whether SOX4 was a direct target of miR-361-3p. Substantial reduction in the luciferase activity was found when the cells were cotransfected with miR-361-3p mimic and the 3 ′ UTR-SOX4-WT reporter vector. However, the cells cotransfected with the 3′UTR-SOX4-MUT reporter vector and miR-361-3p mimic in HepG2 and PLC5 cells displayed no changes in this aspect ( Figure 6(f)). Also, when miR-361-3p was overexpressed, the expression of SOX4 was reduced at both protein and 14 Disease Markers mRNA levels ( Figure 6(g)). To study whether BAIAP2-AS1 promoted the progression of HCC via regulating miR-361-3p/SOX4 axis, we performed rescue experiments. As shown in Figure 7(a), miR-361-3p inhibitors were found to partly relieve the role of BAIAP2-AS1 knockdown in reducing SOX4's expression in HepG2 and PLC5 cells. Subsequently, we performed functional experiments, finding that the proliferative and invasive abilities of HepG2 and PLC5 which were suppressed by BAIAP2-AS1 knockdown were partly rescued by miR-361-3p downregulation (Figures 7(b)-7(e)). Moreover, we also found that SOX4 knockdown could reverse the enhanced proliferation and invasion ability of

Discussion
The obvious symptoms do not display in many HCC patients until an advanced stage, and an increasing occurrence of HCC was observed in Africa and China [27]. Clinical data have demonstrated that early diagnosis and individual treatments based on the advanced prediction of clinical outcomes can enhance long-term survivals of many HCC patients [28,29]. However, sensitive biomarkers are limited. In recent years, more and more researches suggested the potential of lncRNAs to be new biomarkers for prognosis and diagnosis of HCC patients [30,31]. Here, significantly higher-level expression of BAIAP2-AS1 was found in HCC patients, suggesting it as a potential regulator in HCC progression. However, TCGA datasets indicated that the dysregulation of BAIAP2-AS1 expression did not influence the OS and DFS of HCC patients. Thus, the results should be confirmed by larger clinical trials. Overall, our data revealed BAIAP2-AS1 to be a possible biomarker for HCC patients.
Although the abnormal expressions of many lncRNAs were observed in HCC by the use of microarray assays, the potential mechanisms remained largely unclear. Recently, just like some protein-coding genes, the transcription of lncRNAs has been reported to be modulated by transcription factors [32]. For instance, the metastasis of HCC cells was promoted by STAT3-mediated overexpression of lncRNA HOXD-AS1 [33]. Then, we searched Jaspar database and focused on E2F1 which presented a relatively higher score than other regulators. Our group also observed that E2F1 was overexpressed in HCC, which was consistent with previous studies. The role of E2F1 in regulating lncRNA transcription was also reported previously. Subsequently, we performed luciferase reporter assays and ChIP-qPCR analysis, confirming the ability of E2F1 to induce the positive transcription of BAIAP2-AS1 by binding to its promoter region. According to the combined results of the literature and this research, the abnormal activation of transcription factors may be involved in the overexpression of lncRNAs in tumor cells.
Growing studies have demonstrated that lncRNA could influence tumor cells' metastasis and proliferation. Then, we confirmed that under knockdown of BAIAP2-AS1, the proliferation and metastasis of HCC cells were suppressed. It has been reported that EMT can be an important driver of tumor growth and metastasis [34]. Then, Western blot demonstrated that knockdown of BAIAP2-AS1 can suppress the activity of EMT pathway. These findings suggested BAIAP2-AS1 played a role in promoting tumor development in HCC. To further explore the underlying molecular mechanisms, its localization in HCC cells was identified considering the dependence of lncRNA functions on its subcellular localization. Our results confirmed BAIAP2-AS1 as a cytosolic lncRNA, suggesting it can act as a microRNA sponge. The bioinformatics assays and luciferase reporter technologies suggested that BAIAP2-AS1 might bind with miR-361-3p. In many types of tumors, miR-361-3p was observed to exhibit a decreased level and act as an antioncogene, including HCC [35,36]. In our cohort, we also observed that miR-361-3p levels were decreased in HCC, and the negative association of its levels with BAIAP2-AS1was found from the correlation analysis. Thus, we suggested BAIAP2-AS1 may exhibit its oncogenic roles via sponging miR-361-3p.
In general, as a ceRNA, the functions of lncRNAs depend on the mRNA targets of miRNAs [37]. Thus, the targeting genes of miRNAs are a crucial part of the regulatory network. Using StarBase 2.0, we found SOX4 may be a potential target of miR-361-3p. As a critical developmental transcription factor, SOX4 has been demonstrated to exhibit regulatory effects on progenitor developments, cytodifferentiation, and stemness [38,39]. In addition, the activity of several developmental pathways is also regulated by this gene, such as TGFβ, Wnt, and PI3K signaling [40]. In over twenty types of malignancies, SOX4 is distinctly overexpressed, and more and more functional experiments reveal SOX4 as an oncogenic gene [26]. In TCGA datasets, SOX4 was found overexpressed in HCC and a poor outcome of HCC patients was predicted. Further, luciferase reporter assays and Western blot assays identified SOX4 to be a potential target of miRNA-361-3p. To further explore whether lncRNA modulated HCC progression through the modulation of miRNA-361-3p/SOX4, we performed rescue experiments and found that when miR-361-3p decreased, the suppression of SOX4 expression mediated by the silence of BAIAP2-AS1 was reduced to a certain extent. Moreover, a series of functional experiments revealed that the suppression of HCC cell proliferation and metastasis mediated by the inhibition of BAIAP2-AS1 was relieved by miR-361-3p knockdown. In addition, SOX4 knockdown could reverse the enhanced proliferation and invasion ability of HepG2 and PLC5 cells caused by miR-361-3p inhibition. Hence, the integrated pathway of BAIAP2-AS1-miR-361-3p-SOX4 offered new insights into HCC tumorigenesis, suggesting a potential therapeutic approach via targeting the BAIAP2-AS1-miR-361-3p-SOX4 axis.

Conclusions
The results of the present research demonstrated that BAIAP2-AS1 was upregulated and activated by an E2F1 regulator in HCC cells. BAIAP2-AS1 could promote the proliferation and metastasis of HCC cells via sponging miR-361-3p and releasing SOX4. This research contributed to the development of lncRNA-directed diagnostics and therapeutics against HCC.

Data Availability
The analyzed datasets generated during the study are available from the corresponding author on reasonable request.