CircASPH expression is significantly upregulated in HCC tissue and related to patient prognosis
In order to identify the differentially expressed circRNAs between normal tissues and HCC tumor tissues, we used to two GEO databases (GSE28274 and GSE125469) (Figure 1A and 1B). The circRNAs with significant differential expression (fold change ≥2.0 and P < 0.05) between the groups were identified. From the results of two GEO databases, 1 circRNAs were preliminarily identified to be upregulated, while 2 circRNAs were downregulated in the HCC tissue samples compared to the paratumour normal samples (Figure 1C). Thereby, we chose the circASPH for further study.
To determine whether circASPH was associated with HCC, firstly, RT-qPCR indicated that the expression of circASPH was higher in HCC tissues (Figure 1D). Then, we used the FISH assay to identify the expression of circASPH in HCC tumor tissues. It showed that the level of circASPH expression was higher in tumor tissues (Figure 1E). We also investigated the level of circASPH in several HCC cell lines. The level of circASPH was higher in HCC cell lines (HepG2, SMMC-7721, Huh7 and HCCLM3) (Figure 1F). Then, we analyzed the relationship between the circASPH level and HCC patient prognosis. Multivariate Cox analysis revealed that the circASPH level in HCC tissue was an independent prognostic factor in HCC patients (Figure 1G). These data suggest that the downregulated circASPH level plays important roles in HCC tumorigenesis and progression.
CircASPH promotes cell proliferation, migration and invasion in HCC cell lines
To investigate the biological function of circASPH in HCC progression, circASPH was stably overexpressed and knocked down in HCC cell lines respectively (Figure 2A). CCK8 assay and EDU staining assay were carried out to verify the proliferation of HCC cells. The results indicated that the cell proliferation was enhanced by circASPH overexpression, while reduced by circASPH knockdown (Figure 2B-2C). Meanwhile, wound healing assays were performed to evaluate the cell migration ability and transwell assays were used to indicate the cell invasion. The results showed that the migration and invasion capacities were significantly increased after circASPH was overexpressed. However, the ability of cell migration and invasion capacities were reduced when the circASPH level was altered in vitro (Figure 2D-2E). To further determine whether circASPH induce the tumor growth in vivo, we injected control or shRNA-circASPH transfected cells into nude mice. Our results revealed that 6 weeks after transplantation, the tumors composed of shRNA-circASPH transfected cells were significantly smaller than those in control cells (Fig. 2F-2H). Therefore, we verified that circASPH functioned as a tumor promoter to enhance the proliferation, invasion and metastasis of HCC in vitro and in vivo.
CircASPH regulates the level of the DNMT3b/5mC Axis by sponging miR-370 in HCC cells
CircRNAs have been reported to function as sponges for miRNAs. Then, we investigated the ability of circASPH to function as a sponge for miRNAs. Candidate targets were determined using the target prediction. According to the predicted results, the miR-370 gained the highest score (Figure 3A). Then, we used the molecular biological methods to investigate the direct interaction between circASPH and miR-370, the dial luciferase assay was performed and it indicated that circASPH could interact with miR-370 (Figure 3B-3C). We also designed a linear biotinylated circASPH probe. Our qRT-PCR results showed that miR-370 was abundantly pulled down by the circASPH probe in HCC cells. In addition, we designed biotinylated miR-370 to pull down circASPH, it showed the same result (Figure 3D). We also used FISH assay to detect the interaction between the circASPH and miR-370. The result showed that circASPH and miR-370 could be co-localized (Figure 3E). At the same time, overexpressed the circASPH could decrease the miR-370 expression level, while reduction of miR-370 was induced by circASPH knockdown (Figure 3F).
To further elucidate the molecular mechanism by which circASPH promotes HCC progression, we attempted to identify the target genes of miR-370. We used the bioinformatic analysis to indicate the function of target gene. It showed that the molecular function was enriched in chromatin binding (Figure 4A). And the biological process was enriched in the regulation of transcription, especially in apoptotic process (Figure 4B). KEEG pathway showed that target genes main functioned in pathways in cancer and transcription misregulation in cancer (Figure 4C). Previous studies revealed that DNMT3b played a key role in establishing de novo DNA methylation, and was upregulated in cancers. So, we used GEO databases (GSE124535) to analysis the expression of the family of DNMT. It showed that the family of DNMT was enhanced in HCC tissues (Figure 4D). Based on online target prediction algorithms, we found that DNMT3b was a potential candidate target gene of miR-370 (Figure 4E). The dial luciferase assay was performed and it indicated that miR-370 could interact with DNMT3b (Figure 4F). Overexpressed the miR-370 could decrease the DNMT3b expression level, while reduction of DNMT3b was induced by miR-370 knockdown (Figure 4G). Then, we detected changes in 5hmC and 5-mC levels to assess whether miR-370 can remodel the epigenetic landscape by targeting DNMT3b genes and regulating 5-mC levels in the genome. As expected, miR-370 knockdown exhibited lower levels of 5hmC and enhanced the level of 5-mC (Figure 4G). Next, we verified the expression of DNMT3b and level of 5-mC under the circASPH changed condition. The result showed that circASPH knockdown could decrease the DNMT3b expression level and increase level of 5-mC (Figure 4H). These data indicated that circASPH could sponge miR-370 and regulate the DNMT3b/5mc axis in HCC cells.
CircASPH promotes HCC process via the MiR-370/DNMT3b/5mc Axis
First, we used RT-qPCR to indicate that the level of DNMT3b was higher in HCC tumor tissues (Figure 5A). The Pearson analysis also showed the expression of DNMT3b level was positively related to the level of circASPH (Figure 5B). IHC assay also showed the 5mc level was enhanced in HCC tumor tissues (Figure 5C). These results prompted us to hypothesize that HCC tumor tissues may promote HCC progress via the MiR-370/DNMT3b/5mc axis. DNMT3b was knocked down in circASPH overexpression or miR-370 knockdown cells, and the proliferation ability of these cells was evaluated by CCK8 assays and EDU staining assays. We found that DNMT3b knockdown significantly reduced the cell proliferation, which was enhanced by circASPH overexpression or miR-370 knockdown (Figure 5D and 5E). In addition, the wound healing assays and transwell assays were also indicated DNMT3b knockdown significantly reduced the enhanced cell migration and invasion mediated by circASPH overexpression or miR-370 knockdown (Figure 5F and 5G). These results suggest that the DNMT3b/5mc axis plays an important role in the functions of miR-370 and circASPH in HCC cells.
CircASPH promotes HCC progression by regulating the DNA methylation and expression of HAO2
DNMT3b enzymes catalyzed the 5hmC to 5mC, leading to DNA methylation and gene expression regulation. Then, we used the methylation GEO database(GSE55752) in HCC tumor. It indicated the DNA methylation was enhanced in in HCC tumor and was enriched in promoter area (Figure 6A-6C). GSEA also showed the methylated genes also take part in negative regulation of growth. And HAO2 was one of high methylated genes in HCC tumor (Figure 6D). CHIP seq also showed that the methylation of HAO2 was occurred in promoter region (Figure 6F). Overexpressed the circASPH could increase the HAO2 expression level, while reduction of HAO2 was induced by circASPH knockdown (Figure 6G). Meanwhile, overexpressed the miR-370 could decrease the HAO2 expression level (Figure 6H). We also used CHIP assays to verified the 5mc enrichment on HAO2 promoter region (Figure 6I). It indicated that circASPH overexpression could enhance the enrichment and knockdown of circASPH reduce the binding.
Then, we detected the expression of HAO2 in HCC tumor. RT-qPCR indicated that the level of HAO2 was decreased in HCC tumor tissues (Figure 7A). The Pearson analysis also showed the expression of HAO2 level was negatively related to the level of circASPH (Figure 7B). HAO2 was overexpressed in circASPH overexpression or miR-370 knockdown cells, and the proliferation ability of these cells was evaluated by EDU staining assays. We found that HAO2 overexpression significantly reduced the cell proliferation, which was enhanced by circASPH overexpression or miR-370 knockdown (Figure 7C). In addition, the wound healing assays and transwell assays were also indicated HAO2 overexpression significantly reduced the enhaced cell migration and invasion mediated by circASPH overexpression or miR-370 knockdown (Figure 5D and 5E). These results demonstrated that circASPH promoted HCC progression by regulating the DNA promoter methylation and expression of HAO2.