circGLS2 inhibits hepatocellular carcinoma recurrence via regulating hsa-miR-222-3p–PTEN–AKT signaling

Abstract

Growing evidences indicate that circRNAs affect hepatocellular carcinoma (HCC) progression. 1 , 2 , 3 , 4 Understanding the roles of circRNAs in HCC recurrence helps develop innovated therapies. To identify novel HCC recurrence-related circRNA, thirteen HCC patients were subjected to a multi-omics analysis of mRNA, circRNA, and miRNA expression profiles by next-generation sequencing (Supplementary Fig. S1a ). Of these patients, four recurred within two years after the first curative resection (primary tumors, PT) and received a second radical resection (recurrent tumors, RT), while nine had not recurred in five years after the resection (non-recurrent tumors, NRT) (Fig. 1a , Supplementary Table S1 ). We enrolled 110 extra HCC patients as a validation cohort (Supplementary Table S2 ). GSEA among the three tumor types obtained more enriched gene sets than those enriched among the paired adjacent tissues (Supplementary Fig. S1b , Table S3 ). In the top enriched gene sets, HCC recurrence- and tumorigenesis-related gene sets were discovered between NRT and PT or RT (Supplementary Fig. S1c ). Thus, PT and RT were more malignant than NRT. Interestingly, gene sets enrichment between RT and PT suggested that recurrent HCCs were less malignant than primary counterparts (Supplementary Fig. S1c ). This was also evident in the circRNA data and miRNA data (Supplementary Fig. S1d, e ). Fig. 1 circGLS2 inhibits HCC recurrence. a The design of multi-omics analysis for HCC patients with and without postoperative recurrence. b Validation of circGLS2’s circular structure. PCR experiment and Sanger sequencing were performed sequentially. The back-splicing junction site was clearly observed. RNA templates diluted for 10 folds were applied for the examination of GAPDH (Left). After treating with RNase R, circGLS2 showed non-significant changes, while GLS2 with linear structure decreased significantly (Right). c The Cox regression analysis for the factors significantly affected the recurrence of postoperative HCC. d Luciferase reporter assay (Left) and ChIP-qPCR (Right) for KLF4 in parental and KLF4-overexpressing HepG2 cells. e circGLS2 was up-regulated upon ectopic express of KLF4 in HepG2 cells. f The weight and volume changes of tumors collected from mice injected with circGLS2 siRNA- and overexpression plasmid-transfected Huh7 cells. g Subcellular localization of circGLS2. circGLS2 was localized in parental and circGLS2-overexpressing Huh7 cells. 18 S and U6 served as controls localized in the cytoplasm and nucleus, respectively. h Competing endogenous RNA network of circGLS2. The edges were color-coded and proportional to the number of experimental data listed in the miRTarBase database. i Western blot assay of PTEN and AKT. j The IC 50 assay for sorafenib upon circGLS2 knockdown or overexpression in HepG2 cells. k A proposed model for circGLS2’s regulatory mechanism. The dotted arrow represents a possible and/or indirect regulation relationship. AT adjacent tissues; CI confidence interval; FC fold change; HCC hepatocellular carcinoma; HR hazard ratio; IC 50 , half-maximal inhibitory concentration; T tumors; The “si-circGLS2” was the mixture of siRNAs; * P value < 0.05; ** P value < 0.01; *** P value < 1 × 10 −3 ; **** P value < 1 × 10 −4 Full size image

Topics & Concepts