Fig 2: Targeted binding of circCCNB1 VS miR-10106b-5p and miR-106b-5p VS GPM6A. (A, C) Schematic diagram of report vector construction of dual-luciferase reporter assay. (A) The constructed vector for the dual-luciferase reporter of GPM6A 3' UTR VS miR-106b-5p. (C) The constructed vector for the dual-luciferase reporter of circCCNB1 vs. miR-106b-5p. (B) Quantification from the right-upper of A. (D) Quantification from the right-lower of C. (E) circCCNB1 co-localized with miR-106b-5p in HepG2 and HCCLM3 cells detected by FISH. miR-106b-5p was pulled down and enriched with a circCCNB1 specific probe in HepG2 (F) and HCCLM3 (G) cells then detected by qPCR. Data were derived from the results of three independent repeated experiments (Mean ± SD). The bar of the pictures (E) represents 50 µm. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig 3: Silent expression of circCCNB1 promotes the cell cycle of HCC cells by down-expression of GPM6A and up-expression of DYNC1I1 and activation of AKT and ERK signaling pathway. (A) The PPI networks. The PPI network was constructed by GPM6A (green rhombus), ten genes interacting with GPM6A predicted by STRING (pink rectangles), and 145 cell cycle-related genes (yellow Ellipse) derived from PathCards using Cytoscape software. (B) CircCCNB1 silencing inhibited GPM6A expression and up-regulated DYNC1I1 expression by activating the AKT/ERK signaling pathway in HepG2 cells. (C) CircCCNB1 overexpressing promoted GPM6A expression and reduced DYNC1I1 expression by inhibiting the activation of AKT/ERK signaling pathway in HCCLM3 cells. (D) Quantification from B. (E) Quantification from C. Data of the three independent experiments was represented as Mean ± SD. The two-independent t-test was applied for comparing the difference of the gray value of the WB band between two groups (D, E). ns represents not significant, **P<0.01, ****P<0.0001.

Fig 4: The expression role of circCCNB1, miR-106b-5p, and GPM6A in the cells and tissues of HCC. The expression level of circCCNB1 (A), miR-106b-5p (B), and GPM6A (C) in two HCC cell lines (HepG2 and HCCLM3) and one normal liver cell line (HL-7702) detected by qRT-PCR. The expression level of circCCNB1 (D), miR-106b-5p (E), and GPM6A (F) in twenty pairs of carcinoma and para-carcinoma tissues detected by qRT-PCR. (G). The GPM6A expression levels of three pairs of HCC carcinoma and para-carcinoma tissues detected by WB. (H) Quantification from G. (I) The GPM6A expression levels in the carcinoma and para-carcinoma tissues of three pairs of HCC patients detected by IHC. (J) Quantification from I. One-way ANOVA analysis was used to compare the three RNAs expression between four cell lines. The two-independent t-test was used to compare the difference between any two cell lines. paired t-test was used to compare the difference between carcinoma and para-carcinoma tissues. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig 5: Aberrant expression of circCCNB1 significantly regulated cell cycle and clonality of HepG2 and HCCLM3 cells. (A) The G1-S transition of the cell cycle was, respectively, promoted and blocked in HepG2 (circCCNB1 silent expression) and HCCLM3 (circCCNB1 overexpression) cells determined by PI analysis. (B, C) Quantification from A. (D) The underexpression and overexpression of circCCNB1 significantly enhanced and inhibited the clonal formation ability of HepG2 and HCCLM3 cells. (E, F) Quantification from D. (G-J) The efficiency of circCCNB1 silencing and miR-106-5p intervention in hepG2 and HCCLM3 cells was determined by GPM6A WB analysis. After silencing of circCCNB1, miR-106b-5p inhibitor or mimics was added to hepG2 (G) and HCCLM3 (I) cell culture. The GPM6A protein expression level was analyzed by Western blot. (H) Quantification from G. (J) Quantification from I. Data was derived from the results of three independent repeated experiments (Mean ± SD). The two independent t-tests were utilized to compare cell cycle and clonal formation ability between any two groups. ns represents not significant, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.