RETRACTED ARTICLE: miR-3928v is induced by HBx via NF-κB/EGR1 and contributes to hepatocellular carcinoma malignancy by down-regulating VDAC3

Twenty pairs of human hepatic carcinoma and normal adjacent tissues were obtained from the Cancer Center of Sun Yat-sen University. Sixty serum samples from patients with HCC, 30 serum samples from healthy people and 10 pairs of slides of human hepatoma tissues and adjacent non-tumor tissues for immunohistochemistry analysis were obtained from TangShan People’s Hospital. Written informed consent was obtained from each patient included in the study, and the samples were used in accordance with the standards of the center’s ethics committee. The HCC cell lines are described in Additional file 1.

The miR-3928v expression vector (pri-miR-3928v), VDAC3 expression vector (pcDNA3-Flag-VDAC3), shR-VDAC3 vector, dual-luciferase reporter plasmids and pGL3-luciferase-miR-3928v promoter were constructed. The details of these constructs are provided in Additional file 1. The sequences of the primers used to generate the above plasmids are listed in Additional file 2. A 2′-O-methyl-modified antisense oligonucleotide of miR-3928v (ASO-miR-3928v) was commercially synthesized by Genpharm (Shanghai, China).

RNA was extracted from serum and tissues using a miRcute miRNA Isolation Kit (Tiangen Biotech, China) according to the manufacturer’s instructions. The relative mRNA level was examined by RT-qPCR as detailed in Additional file 1. The primer sequences are provided in Additional file 2.

Huh7, HepG2 and HepG2.2.15 cells were seeded in 48-well plates and cultured for approximately 20 h after co-transfection with different plasmids. The procedures are described in detail in Additional file 1.

Following transfection, cells were counted and seeded in the appropriate wells, and the density, cell viability, colony formation rate, migration and invasion were determined as described previously [25].

To examine the effect of miR-3928v and VDAC3 on VM in HCC cells, the cells were transfected with pri-miR-3928v, ASO-miR-3928v, pVDAC3 or shR-VDAC3 and the corresponding controls. A 24-well cell culture plate was coated with 50 μl of Matrigel (DB Biosciences), which was allowed to solidify at 37 °C for 1 h. After the Matrigel solidified, 4 × 10⁵ or 5 × 10⁵ cells were seeded on it. After 24 h of culture at 37 °C, tube-like structures were observed and photographed using a microscope.

Transfected cancer cells were seeded into 6-well plates and incubated for 24 h in complete medium until the cells became 60% confluent. The culture medium was then replaced with serum-free culture medium for starvation. After 24 h of starvation, the cells were cultured in complete medium for an additional 24 h. The cells were then collected for cell cycle and apoptosis analyses. Details can be found in Ref. [26].

Chromatin immunoprecipitation (ChIP) assays were performed using an EZ-ChIP™ Chromatin Immunoprecipitation Kit (Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. Huh7 cells were seeded in 10-cm cell culture plates. The cells were lysed, sonicated to shear the DNA and immunoprecipitated with anti-EGR1 (Shengyang Wanleibio, China) or control antibodies (IgG and GAPDH). qPCR primers were designed using PrimerBLAST (Additional file 2), and DNA purified from EGR1/DNA crosslinks was analyzed by qPCR.

Electrophoretic mobility shift assays (EMSAs) were performed using a LightShift Chemiluminescent EMSA Kit (Thermo) according to the manufacturer’s protocol. Details can be found in Ref [12], with some modifications as described in Additional file 1.

For the in vivo tumor growth study, 1 × 10⁷ Huh7 cells transfected with pri-miR-3928v or control vector were suspended in 110 μL of serum-free RPMI or DMEM for each mouse. Each female, BALB/c-nu/nu SCID mouse (5–6 weeks old, 6 per group) was injected subcutaneously in the flank. All mice were killed 8 days after implantation. The tumors were isolated, and the weights were recorded. All studies were performed according to the American Association for the Accreditation of Laboratory Animal Care Guidelines for the Humane Treatment of Animals and adhered to national and international standards.

Two tailed Student’s t-tests and analysis of variance (ANOVA) were used to analyze the significance of differences between sample means obtained from three independent experiments. P ≤ 0.05 indicated statistical significance (*: P < 0.05, **: P < 0.01, ***: P < 0.001).

To identify novel miRNAs that may play key roles in hepatocarcinogenesis, the miRNA expression profile in HBV (+) HCC tissues was analyzed by deep sequencing. We selected TLRC-m0008_3p for further analysis because of the higher read count and the lowest minimum free energy (Additional file 3). Using the PubMed search tool blastn, we found that TLRC-m0008_3p may be a variant of miR-3928. We therefore named this miRNA miR-3928v and further investigated its role in HCC. The predicted secondary structure of the miR-3928v precursor and its mature miRNA sequence are shown in Fig. 1a. RT-qPCR was used to examine miR-3928v expression in hepatoma cell lines. Compared with the immortalized normal liver cell line L02, the HBV (−) cell line HepG2 and the HBV (+) cell line HepG2.2.15 showed a significant up-regulation of miR-3928v (Fig. 1b). Furthermore, overexpression of pHBV1.3copy (a plasmid containing a 1.3-unit-length genome of HBV in pUC18 for the expression and replication of HBV [27]) in HBV (−) cell lines increased miR-3928v expression (Fig. 1c). To determine whether the non-structural HBV protein HBx is responsible for regulating miR-3928v expression, an expression plasmid harboring HBx was transfected into HepG2 and Huh7 cells and found to increase miR-3928v expression levels (Fig. 1d). These data indicate that HBV, and HBx in particular, induces miR-3928v expression.

In addition, we examined miR-3928v expression levels in 20 pairs of liver cancer and adjacent non-tumor tissues. RT-qPCR analysis showed that miR-3928v was highly expressed in HCC tissues (Fig. 1e). Moreover, to assess the correlation between the miR-3928v level and the clinical progression of HCC, we examined miR-3928v serum levels in 60 HBV (+) HCC patients and 30 healthy controls. miR-3928v serum levels were obviously higher in HCC patients than in healthy controls (Fig. 1f), suggesting that miR-3928v may be involved in HCC pathogenesis.

To determine the function of miR-3928v in HBV-associated HCC, pri-miR-3928v and ASO-miR-3928v were used to overexpress or block miR-3928v expression in HepG2, Huh7 and HepG2.2.15 cells. After the transfection efficiency was confirmed (Additional file 1: Figure S1A), MTT assays suggested that miR-3928v promoted cell viability (Fig. 2a). To address whether the promotion of cell viability is associated with apoptosis, FACS analysis was used to examine cell apoptosis. Annexin-V staining showed that miR-3928v decreased the apoptosis rate of HCC cells (Fig. 2b). Next, colony formation assays showed that miR-3928v promoted colony formation (Fig. 2c). FACS analysis was used to determine whether cell cycle progression is responsible for the promotion of cell growth by miR-3928v; the results showed that miR-3928v accelerated the G1/S transition and increased the proliferation index (PI) (Fig. 2d). Taken together, these data indicate that miR-3928v suppresses apoptosis and accelerates cell cycle progression to promote HCC cell viability and proliferation.

To further investigate the effect of miR-3928v on migration/invasion and VM in HCC cells, we performed migration/invasion and tube formation assays in HepG2, Huh7 and HepG2.2.15 cells. As shown in Fig. 2e-f, forced miR-3928v expression promoted migration/invasion and tube formation, but ASO-mediated inhibition of miR-3928v repressed these effects. To determine the mechanisms underlying the effects of miR-3928v on cell mobility and VM, epithelial-mesenchymal transition (EMT) and VM markers were detected by Western blot. As expected, miR-3928v promoted vimentin expression and suppressed E-cadherin protein expression (Fig. 2e, right panel), indicating that miR-3928v promotes EMT. Moreover, miR-3928v increased matrix metallopeptidase 2 (MMP2) and MMP9 protein expression (Fig. 2f, right panel), further indicating that miR-3928v promotes VM. These results suggest that miR-3928v induces EMT to promote migration/invasion and enhances MMP2 and MMP9 expression to promote VM.

To determine the effect of miR-3928v on HCC tumor growth, xenograft tumor experiments were performed in mice. The average volume and weight of tumors derived from Huh7 cells were higher in the miR-3928v group than in the control group (Fig. 2g). In short, miR-3928v functions as an oncogene to contribute to malignancy in vivo and in vitro.

To identify the mechanism by which miR-3928v promotes the malignancy of HCC cells, bioinformatics analysis (http://​www.​targetscan.​org/​vert_​71/​) was used to predict target genes of miR-3928v. Among over 100 candidate target genes, we noticed that the 3′UTR of VDAC3 contains the miR-3928v binding site, and previous reports have found that HBx can directly interact with VDAC3. Therefore, we chose VDAC3 as a candidate target gene in HCC and asked whether HBx modulates miR-3928v to indirectly regulate VDAC3 expression. As shown in Fig. 3a, the seed sequence of miR-3928v is complementary to a sequence within the VDAC3 mRNA 3′UTR. In addition, dual-luciferase reporter assays showed that miR-3928v overexpression in HepG2 cells decreased luciferase reporter activity, which was blocked by ASO in HepG2.2.15 cells, whereas these effects were abolished when the binding site in the VDAC3 3′UTR was mutated (Fig. 3a, right panel). Subsequently, RT-qPCR and Western blot analyses showed that miR-3928v decreased VDAC3 mRNA and protein levels in HepG2, Huh7 and HepG2.2.15 cells (Fig. 3b). Furthermore, VDAC3 mRNA levels were down-regulated in these cell lines (Fig. 3c) and HCC tissues (Fig. 3d). We also examined VDAC3 expression in tissues from liver cancer patients and nude mice by immunohistochemistry (IHC) and found that VDAC3 expression was lower in liver cancer tissues than in adjacent normal tissues (Fig. 3e, left panel) and lower in miR-3928v-overexpressing tumor tissues than in control tissues (Fig. 3e, right panel). Similarly, RT-qPCR analysis showed that VDAC3 mRNA levels in tumor tissue from nude mice were almost 45% lower in the pri-miR-3928v group than in the control group (Fig. 3f). Taken together, these results indicate that miR-3928v directly targets and down-regulates VDAC3.

To study the role of VDAC3 in HCC malignancy mediated by miR-3928v, we performed gain- and loss-of-function analyses in HCC cell lines. Cell viability, cell growth, migration/invasion and VM, as demonstrated by MTT (Fig. 4a), colony formation (Fig. 4b), migration/invasion (Fig. 4c) and tube formation assays (Fig. 4d), were significantly promoted by shRNA-mediated knockdown of VDAC3 in HepG2 and Huh7 cells and were suppressed by forced VDAC3 expression in HepG2.2.15 cells. Furthermore, FACS analysis revealed that VDAC3 increased apoptosis (Additional file 1: Figure S2) and suppressed the G1/S transition (Fig. 4e) in HCC cells. EMT assays showed that VDAC3 reduced vimentin expression and enhanced E-cadherin protein expression, indicating that VDAC3 inhibits EMT (Fig. 4f). Moreover, VDAC3 reduced MMP2 and MMP9 expression, suggesting that VDAC3 suppresses VM (Fig. 4g). In conclusion, VDAC3 suppressed cell viability, growth and migration/invasion by inducing apoptosis and inhibiting cell cycle progression and EMT. In addition, VDAC3 reduced MMP2 and MMP9 expression to suppress VM. These data indicate that VDAC3 may function as a tumor suppressor in HCC cells.

To verify that miR-3928v promotes the malignancy of HCC cells by directly down-regulating VDAC3, we performed rescue experiments with a VDAC3 expression vector (pVDAC3) containing the VDAC3 open reading frame without the 3′UTR. When HepG2 cells were co-transfected with pri-miR-3928v or a control vector and pVDAC3, the miR-3928v-induced inhibition of VDAC3 mRNA and protein expression (Fig. 5a) and increases in cell viability, cell cycle, migration/ invasion and VM were abrogated (Fig. 5b-f). These results indicate that miR-3928v functions as an oncogene through VDAC3 and that VDAC3 is a functional target of miR-3928v in HCC cells.

It has been reported that HBx contributes to HCC progression [4]. To determine whether HBx-induced miR-3928v mediates the oncogenic activities of HCC cells caused by HBV, we first expressed HBx to observe its effects in HepG2 cells. As shown in Additional file 1: Figure S3A-D (left panel), HBV/HBx promoted HepG2 cell viability, cell growth, migration/invasion and VM. Next, rescue experiments were performed in HCC cells by co-transfection with the HBx expression plasmid and ASO-miR-3928v. As expected, the ability of HBx to promoted HepG2 cell viability, cell growth, migration/invasion and VM was abolished by ASO-miR-3928v (Additional file 1: Figure S3A-D, right panel). These data suggest that miR-3928v may contribute to the malignancy of HCC cells caused by HBV infection.

To determine the mechanism underlying the up-regulation of miR-3928v after HBV infection, we analyzed its location in the human genome. As shown in Fig. 6a, miR-3928v is located on Homo sapiens chromosome 22 (NC_000022, chr22: 31,160,061–31,160,083). The Promoter 2.0 Prediction Server (http://​www.​cbs.​dtu.​dk/​services/​Promoter/​) was used to predict the promoter region upstream of miR-3928v. Then, the miR-3928v promoter was cloned into pGL3-basic to generate P1307, with the first base of the miR-3928v precursor being set as + 1 (Fig. 6a, left). Dual-luciferase reporter assays showed that P1307 increased the relative luciferase activity in Huh7 cells (Fig. 6a, right). Furthermore, to map the core promoter of miR-3928v, we constructed four additional fragments of P1307 with 5′ terminal deletion mutations, P972, P637, P335 and P302 (Fig. 6a, left). As shown in the right panel of Fig. 6a, P972, P637 and P302 displayed high luciferase activity, with P302 having the highest activity, whereas P335 showed no luciferase activity, perhaps suggesting the presence of a silencer. Therefore, P1307 and P302 were used in further studies. To investigate whether the candidate promoter fragments are associated with HBV, P1307 and P302 were transiently transfected into HepG2.2.15, HepG2 and L02 cells. HBV (+) cells had the highest promoter activity (Fig. 6b, left), indicating that HBV may positively regulate the miR-3928v promoter. The effect of HBV on the miR-3928v promoter was further confirmed by transient transfection of HBV1.3copy/HBx in Huh7 cells, which resulted in increased luciferase activity (Fig. 6b, right), indicating that HBV/HBx positively regulates miR-3928v promoter activity. Transcription factors are known to be critical promoter-activating elements. PromoterScan was used to predict transcription factors that may bind to P302. EGR1 was found to have a putative binding site located at − 194~ − 186 bp in the miR-3928v promoter region (Fig. 6c, right). To study the effects of EGR1 on the miR-3928v promoter, Huh7 cells were co-transfected with P1307 or P302 and an EGR1 expression plasmid (pEGR1); EGR1 increased miR-3928v promoter activity (Fig. 6c, left). When the EGR1 binding site in the miR-3928v promoter was mutated, the luciferase activity decreased (Fig. 6c, right). ChIP and EMSA analyses demonstrated that EGR1 directly binds to the miR-3928v promoter (Fig. 6d). In addition, RT-qPCR showed that EGR1 enhanced miR-3928v expression (Fig. 6e). Furthermore, shRNA-mediated knockdown of EGR1 abolished HBx-induced miR-3928v up-regulation (Fig. 6f), demonstrating that EGR1 is a critical factor that mediates the HBx-induced promotion of miR-3928v expression.

It has been reported that HBx can induce NF-κB signaling in HCC [28, 29] and that tumor necrosis factor (TNF)-α, an NF-κB signaling activator, enhances EGR1 promotor activity [30, 31]; therefore, we hypothesized that HBx may increase EGR1 expression. As shown in Fig. 6g, EGR1 expression was higher in liver cancer tissues than in adjacent non-tumor tissues. Western blot analyses revealed that HBx increased EGR1 expression and promoted its nuclear import (Fig. 6h). An immunofluorescence assay showed similar results (Fig. 6i). Taken together, our results indicate that HBx may facilitate the activation of miR-3928v expression via EGR1 in an NF-κB signaling-dependent manner.