Liver cancer poses great threat to human health. Annually, approximately 800,000 new cases are diagnosed with and over 700,000 deaths are linked to liver cancer worldwide[1]. Understanding its underlying pathogenesis mechanism may provide therapeutic targets. In this study, we found that circSOD2, derived from SOD2 was highly expressed in HCC patient tumor tissues and liver cancer cells than normal liver tissues and normal liver cells, respectively. Depletion of circSOD2 significantly ceased liver cancer cell growth and impaired in vivo tumorigenesis. Mechanically, we showed that, by acting as a sponge for miR-502-5p, circSOD2 suppressed miR-502-5p expression, which in turn upregulated DNMT3a expression. Elevated DNMT3a suppressed SOCS3 expression and further activates JAK2/STAT3 signaling pathway. In addition, we also found that increased STAT3 in liver cancer cells also regulates circSOD2 expression in a feedback way.
Complexed roles of miR-502 have been revealed in different cancers. For example, hsa-miR-502-5p (miR-502) was reported to inhibit autophagy and tumor growth in colon cancer by suppression of Rab1B[35]. Exogenously introduce miR-502-5p mimic into breast cancer cells enhanced early cell apoptosis and inhibited cell proliferation[36]. However, upregulation of miR-502 has been reported to accelerate esophageal cancer cell TE1 proliferation by promoting AKT phosphorylation[37]. In this study, we found that, similar to breast cancer and colon cancer, miR-502-5p was also downregulated in HCC. Further studies on circSOD2 and miR-502-5p showed that miR-502-5p binds to circSOD2 exon 3, depletion of circSOD2 significantly upregulated miR-502-5p expression. These results suggest that in HCC, miR-502-5p was suppressed by circSOD2 through direct interaction.
Bioinformatic analysis showed that miR-502-5p binds to DNMT3a 3’UTR. Introducing miR-502-5p into liver cancer cells greatly suppressed DNMT3a expression, suggesting DNMT3a is a target of miR-502-5p in HCC. The importance of DNMT3a in HCC progression has been extensively studied. Zhao et, found that depletion of DNMT3a suppressed HCC cell proliferation[38]. Chen et, reported that miR-30a-3p inhibits the proliferation of liver cancer cells by targeting DNMT3a[39]. Similarly, in this study we also demonstrated that downregulation of DNMT3a by silencing circSOD2 impaired liver cancer cell growth and migration.
DNMT3a mediated DNA hypermethylation suppressed gene expression and has been associated with many cancer developments. DNA hyper-methylation at NF2 and KIBRA promoter mediated by MOC2 and DNMT3a complexes suppresses their gene expression and Hippo signaling pathway activation, which further promotes HCC cancer stemness and tumorigenesis[40]. In Triple Negative Breast Cancer, coupled with MYC, DNMT3a promotes the epithelial to mesenchymal transition and mammosphere formation of TNBC cells by upregulating miR-200b promoter methylation[41]. SUV39H1/DNMT3A-dependent methylation of the RB1 promoter stimulates PIN1 expression and melanoma development[42]. Here, by knocking down DNMT3a, we discovered that SOCS3 is regulated by DNMT3a. Further analysis and experimental validation demonstrated that upregulated DNMT3a in liver cancer cells promoted SOCS3 promoter hyper-methylation and suppressed SOCS3 expression.
Abnormal activation of IL-6/STAT3 signaling pathway has been implicated in the progression of HCC. In HBx induced hepatocellular carcinoma mice model, IL-6/STAT3, and Wnt/β-catenin signaling pathways are constitutively activated[43]. Tumor cell-intrinsic Tim-3 promotes liver cancer via NF-κB/IL-6/STAT3 Axis[34]. Long non-coding RNA DILC regulates liver cancer stem cells via IL-6/STAT3 axis[32]. In this study, we uncovered a new mechanism through which STAT3 is activated, we showed that, SOCS3 suppression mediated by DNMT3a methylation, activated JAK2/STAT3 signaling pathway. Due to the complexity of IL-6/STAT3 signaling pathway, more studies should be performed to clarify the crosstalk among different pathways.
We also evaluated the mechanism through which circSOD2 was activated in HCC. We found that by tethering to circSOD2 promoter, EP300 and WDR5 facilitated circSOD2 promoter H3K27ac and H3K4me3 modification respectively, and increased circSOD2 expression. Moreover, STAT3 binding sites on circSOD2 promoter recruited STAT3 binding to circSOD2 promoter and upregulated circSOD2 expression in a feedback way. Although, for now, we do not have any idea whether STAT3 recruits EP300 or WDR5 into circSOD2 promoter or vice versa, more studies should be performed in the future to clarify this.