Twenty-seven HCC specimens were collected from the Department of Hepatobiliary and Gastrointestinal Surgery, Sichuan Provincial Cancer Hospital (Chengdu, China). Tissue samples were immediately frozen in liquid nitrogen after isolation. Informed consent was obtained from each patient. The Ethics Commitment of Sichuan Provincial Cancer Hospital and the Ethic Commitment of Sichuan University approved this study. The senior pathologists of Sichuan Provincial Cancer Hospital evaluated the histological features of the specimens.

HCC cell lines HepG2 and SMMC-7721 and, L-02 cells were purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA). Human HCC HepG2, SMMC-7721 and L-02 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS; Invitrogen Corporation, Grand Island, NY, USA).

Total RNAs of HCC tissues and cells were extracted by TRIzol reagent (Invitrogen Corporation) according the manufacturer's instruction. The total RNAs was reverse-transcribed to cDNA by using All-in-One™ miRNA First-Strand cDNA Synthesis kit (Invitrogen Corporation). The primers were designed and synthesized by Shengong Company (Shengong, Shanghai, China). Real-time PCR assay was performed as described previously (24,27).

miR-4782-3p antisense oligonucleotides (miR-4782-3p ASO), miR-4782-3p mimics and negative control miRNA were purchased from RiboBio (Guangdong, China). miR-4782-3p ASO or miR-4782-3p mimics were transfected into cells by using Lipofectamine (Invitrogen, Shanghai, China), according to the manufacturer's instruction (27).

HepG-2 and SMMC-7721 cells (5×10³/well) were seeded into 96-well plates. Then MTT experiments were performed as previously described (28,29). Absorbance in each well was measured by using a microplate reader set at 570 nm.

The apoptosis of HepG-2 and SMMC-7721 cells transfected with miR-4782-3p ASO or miR-4782-3p mimics were analyzed by using PI/Annexin V staining and fluorescence-activated cell-sorting (FACS) flow cytometer (Biosciences, Beijing, China). Details in brief, HepG2 and SMMC-7721 cells were seeded into 12-well plates. Twelve hours after transfection, cells were cultured with serum-depleted medium for 12 h. Then cells were suspended in binding buffer followed by staining with PI/Annexin V-FITC for 15 min. The apoptotic rates were calculated on a flow cytometer (30).

Bioinformatics methods were applied for the prediction of the targeted genes of miR-4782-3p. Moreover, the bioinformatic algorithms from TargetScanHuman were used (31–34).

To assess and confirm whether miR-4782-3p binds USP14 directly, a dual-luciferase assay was performed (25,30). The 3′UTR of USP14 was amplified by using PCR from genomic DNA. The production was inserted downstream of USP14 3′UTR reporter plasmids (pRL-USP14; Biotech, Chengdu, China), and mutants of USP14 3′UTR were generated by Site-Directed Mutagenesis kit (Shanghai, China). Then the whole plasmid was confirmed by sequencing. Mutation in the miR-4782-3p binding site of the USP14 3′UTR was constructed by Shengong Company (Shengong, Chengdu, China). The luciferase reporter containing mutant was also constructed. For luciferase assays, HepG2 cells were transfected with luciferase reporter plasmid along with miR-4782-3p mimics or negative control by using Lipofectamine 2000 (Invitrogen). Twenty-four hours after transfection, these cells were analyzed by using a Luciferase assay kit (Promega, Madison, WI, USA) (35).

Two-tailed Student's t-test was used to analyze the difference between two groups. ANOVA was used to analyze the difference of three groups. Kaplan-Meier analysis was employed to evaluate the overall survival of HCC patients. The correlation between miR-4782-3p and USP14 was performed by Spearman's correlation analysis. All statistical analysis was performed by SPSS version 19. Values are expressed as the mean ± SD from three tests. All values of P<0.05 are marked with an asterisk in the figures, and was considered statistically significantly different.

We first examined the miR-4782-3p level in HCC cell lines; the miR-4782-3p level in normal liver tissue was treated as a negative control. We found that the miR-4782-3p levels in HepG2 and SMMC-7721 cells were lower than in normal liver tissue and L-02 cells (Fig. 1A). Then we assayed the miR-4782-3p levels in 27 HCC tissues and adjacent normal liver tissues. We found that the mean value of miR-4782-3p levels in tumor tissues is lower than the mean value of adjacent normal tissues (Fig. 1B). The relative expression of miR-4782-3p in each HCC tissue and the matched adjacent normal liver tissue are shown in Fig. 1C. There were only 4 pairs, in which HCC tissues showed higher levels of miR-4782-3p. Next, the 27 HCC patients were separated into two groups by the median of the miR-4782-3p expression, and their progress was followed for about 100 weeks. We found that HCC patients with higher miR-4782-3p level had a longer survival period (Fig. 1D).

To investigate the effect of miR-4782-3p in cell growth and apoptosis, HepG2 and SMMC-7721 cells were transiently transfected with miR-4782-3p mimics, and then the cell growth and apoptosis were evaluated. Forty-eight hours after transfection of the miR-4782-3p mimic, the miR-4782-3p level in HepG2 and SMMC-7721 cells was tested by qRT-PCR. It showed that the miR-4782-3p levels in both cell types were increased (Fig. 2A). We then detected the effect of miR-4782-3p on cell growth by MTT. With transfection of miR-4782-3p mimic, HepG2 and SMMC-7721 cells showed significant decrease in cellular growth (Fig. 2B). To verify whether miR-4782-3p could influence HCC cells apoptosis, PI/Annexin V double staining were performed to evaluate apoptosis. The results showed that miR-4782-3p mimic transfection increased apoptosis about 3–5% (Fig. 2C).

Additionally, HepG2 and SMMC-7721 cells were transfected with miR-4782-3p ASO to downregulate the miR-4782-3p levels. Then the miR-4782-3p ASO transfection was tested by qRT-PCR. The miR-4782-3p levels in both cell lines decreased following the transfection (Fig. 3A). With miR-4782-3p ASO transfection, HepG2 and SMMC-7721 cells showed significant increase in cellular growth (Fig. 3B). However, the miR-4782-3p ASO showed no significant effect on cell apoptosis (Fig. 3C).

The gene targets of miR-4782-3p were predicted by TargetScanHuman, this prediction showed that USP14 could be targeted by miR-4782-3p. Then we mutated the binding site of miR-4782-3p in USP14 (Fig. 4A). The effect of miR-4782-3p on USP14 translation was tested by a luciferase reporter assay. The luciferase reporter plasmid with wild-type 3′UTR of USP14 or the mutated version as indicated in Fig. 4A showed that upregulation of miR-4782-3p reduced the luciferase activity of the reporter gene with wild-type, but not with the mutant (Fig. 4B). Next the USP14 mRNA expressions in 10 HCC tissues were examined by qRT-PCR. We found that the miR-4782-3p level and the USP14 mRNA level were negatively correlated (Fig. 4C).