ABCA8 is regulated by miR-374b-5p and inhibits proliferation and metastasis of hepatocellular carcinoma through the ERK/ZEB1 pathway

We collected matched HCC and adjacent non-tumor tissue from patients who underwent hepatectomy in the First Affiliated Hospital of Harbin Medical University between August 2010 and September 2014. We invited senior pathologists with senior professional titles to perform pathological diagnosis on paraffin sections, and only patients with pathological results of HCC were included in this study. All patients who participated in this study provided informed consent. This study was approved by the Research Ethics Committee. Detailed clinicopathological features of 105 HCC specimens involved in this study are shown in Table 1.

Human HCC cell lines, Huh7, HepG2, HCCLM3, and SK-Hep-1were obtained from the Chinese Academy of Science (Shanghai, China). Normal liver cell line WRL-68 was obtained from AcceGen (Fairfield, USA). All cell lines were cultured in Dulbecco’s Modified Eagle Medium (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco, USA), 100 U/mL penicillin and 100 μg/mL streptomycin. All cells were incubated in incubators containing 5% CO₂ at 37 °C.

Lentiviral vectors for ABCA8 and miR-374b-5p gene up-regulation (Lv-ABCA8, Lv-miR-374b-5p), down-regulation (Lv-shABCA8, Lv-anti-miR-374b-5p), empty vectors (Lv-NC) and encoding human firefly luciferase were manufactured and obtained from GeneChem (Shanghai, China). Details of the short hairpin RNA sequence against ABCA8 are listed in Additional file 1: Table S1.

Briefly, the protein samples extracted from cells or tissues were loaded, separated and then transferred onto nitrocellulose membranes (Invitrogen, Carlsbad, USA). Subsequently, 5% bovine serum albumin was used to block the nitrocellulose membrane for 1 h. Finally, the primary antibody and conjugated secondary antibody were added. Protein blots were detected using enhanced chemiluminescence (Beyotime, Shanghai, China). Details of the antibodies are listed in Additional file 1: Table S2.

Total RNA was isolated from fresh frozen tissue and logarithmically growing cells using an RNA Miniprep Kit (Axygen, Jiangsu, China), and cDNA was synthesized using either a High Capacity RT Kit or TaqMan® MicroRNA RT kit (Applied Biosystems, Carlsbad, USA). qPCR was performed using SYBR Green (Roche, Indianapolis, USA) or TaqMan® qPCR Master Mix (Applied Biosystems, Carlsbad, USA) on a 7500 Fast PCR System. Then the expression levels of mRNA and miRNA were normalized to GAPDH and U6, respectively. Details of the primers and probes for qPCR are listed in Additional file 1: Table S3.

After a series of processes including dewaxing, rehydration, and antigen repair, the carcinoma and adjacent non-tumor sections were blocked with secondary antibody source serum. After blocking, the sections were incubated with primary antibodies overnight. The following day, the sections were incubated with secondary antibodies and stained with diaminobenzidine. The protein staining intensity score was calculated according to previously described methods [10]. ABCA8 staining intensity was scored as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong). The staining extent was scored based on the percentage of positive cells using the following scale: 0 (negative), 1 (0.01–25%), 2 (25.01–50%), 3 (50.01–75%), and 4 (75.01–100%). The histologic score (H score) for each section was calculated with the following formula: histologic score = proportion score × intensity score. Thus, the total score could be 0, 1, 2, 3, 4, 6, 8, 9, or 12, and the staining could be classified as negative/low (0, 1, 2, 3, 4) or positive/high (6, 8, 9, 12).

Stably transfected cells were inoculated in 96-well plates and cultured overnight for attachment. The following day, the culture solution was replaced with a solution which contained the CCK-8 reagent and cultured for 2 h in the dark. Then, the absorbance at 450 nm was measured for each well.

Stably transfected cells in the logarithmic growth phase were seeded in 6-well plates for two weeks. The colonies were then fixed and stained for easy observation, and photographs were taken.

Stably transfected cells were inoculated in 6-well plates and cultured until cell fusion occurred. A straight cut was made at the bottom of the plate with a 10-μL pipette tip. The floating cells were washed away and the wound closure was photographed at 0 and 24 h.

Matrigel-coated (BD Biosciences, Franklin Lakes, NJ) or non-Matrigel-coated Transwells were used to examine the invasion and migration ability of cells. Stably transfected cells were inoculated in the upper chamber of the transwells and serum-free media was added. Normal media was injected into the plate wells. After a 24 h incubation period, the cells in the upper layer filter were removed and the cells in the bottom layer were fixed, stained and counted.

Cells were inoculated in 24-well plates, and the wild-type or mutated 3′-UTR sequence of ABCA8 were cotransfected with pRL-TK Renilla. After incubation for 48 h, luciferase activities were measured by the dual luciferase reporter assay kit.

Stably transfected cells were inoculated on glass sheets and incubated overnight. After attachment, the cells were fixed, permeabilized, and blocked with normal goat serum and incubated with primary antibodies overnight at 4 °C. The following day, cells were incubated with a fluorescent secondary antibody for 1 h. Finally, nuclei were counterstained with DAPI, and the images were photographed under a fluorescence microscope.

Male BALB/c nude mice (4–6 weeks old) were purchased from the Experimental Animal Center of Shanghai Institute. Subcutaneous xenograft tumors were established as follows: 5 × 10⁶ cells were dissolved in 0.15 mL phosphate buffered saline, then subcutaneously injected into the flanks of the mice. The size of the hypodermic ectopic neoplasms was observed weekly. After 6 weeks the nude mice were sacrificed and the xenograft tumor was excised. The volume of the tumor was measured by the following calculation: V=W × L × H/2.

The excised subcutaneous xenograft tumor was divided into 1 mm³ cubes and transplanted into the livers (left lobes) of mice from the same line to establish the orthotopic xenograft nude mice model. The mice were sacrificed after 6 weeks, after which the tumors were resected.

The pulmonary metastases nude mice model was established as follows: 4 × 10⁶ cells dissolved in 0.15 mL phosphate buffered saline were injected into the tail veins of each mouse. After 6 weeks, the mice were sacrificed and their lungs were collected. Animal experiments were approved by the Animal Ethics Committee of Harbin Medical University, and each step was carried out in accordance with animal care and use standards.

To verify the link between ABCA8 expression and HCC, qPCR and western blots were performed. The results of these assays (Fig. 1a and b) were mostly consistent with those in the TCGA database (https://​www.​cancer.​gov/​; Fig. 1c). Additionally, immunohistochemical staining of 105 pairs of hepatocellular carcinoma and adjacent non-tumor tissue was performed (Fig. 1d) and analyzed in combination with clinicopathological characteristics. Decreased ABCA8 expression in HCC was positively correlated to tumor diameter (p = 0.0121), metastasis (p = 0.0103) and TNM stage (p = 0.0106) (Table 1). We also found that patients with low ABCA8 expression had poorer overall survival than patients with high ABCA8 expression (Fig. 1e), which was supported by data from UALCAN (http://​ualcan.​path.​uab.​edu/​index.​html; Additional file 2: Figure S1a). To confirm the reliability of our results, we consulted a Kaplan-Meier plotter (http://​kmplot.​com/​analysis/​; Additional file 2: Figure S1b), which also demonstrated that low ABCA8 expression in patients with HCC had poor prognoses. In addition, the outcomes of qPCR and western blots demonstrated that the expression of ABCA8 was lower in HCC cell lines than in normal liver cells and that the expression of ABCA8 decreased as the malignancy of the cells increased (Fig. 1f, g). These results demonstrated that ABCA8 is frequently reduced in HCC, and low expression of ABCA8 linked to poor prognosis.

To understand the effect of ABCA8 in HCC, we transfected Huh7 and HepG2 cell lines with short hairpin RNAs (shRNA) by lentivirus vectors to silence the expression of ABCA8. In addition, we transfected HCCLM3 and Sk-Hep-1 cell lines with lentivirus to overexpress ABCA8. The results of transfection showed that shABCA8–2 had the most significant silencing effect (Fig. 2a and Additional file 3: Figure S2), and therefore we utilized shABCA8–2 for subsequent experiments. Meanwhile, the untreated cells were uesd as blank groups (Bg).

CCK8 experimental results revealed that overexpression of ABCA8 can inhibit the activity of HCCLM3 cells and reduce their ability to proliferate. This result was also observed in Sk-Hep-1 cells. Silencing the expression of ABCA8 can promote the vitality of Huh7 and HepG2 cells and enhance their proliferation ability (Fig. 2b). Colony formation also demonstrated that up-regulating expression of ABCA8 can reduce the number of clones formed by HCCLM3 and Sk-Hep-1 cell lines, while down-regulating expression of ABCA8 can increase the number of clones of Huh7 and HepG2 cell lines (Fig. 2c).

To further prove that ABCA8 inhibits the growth of HCC in vivo, we constructed subcutaneous and orthotopic xenograft models. Larger HCC tumors were observed in the Huh7-shABCA8 group and smaller HCC tumors were observed in the HCCLM3-ABCA8 group when compared to the corresponding control groups and black groups (Fig. 2d, e). Subcutaneous tumors were sectioned and immunohistochemical analysis showed that higher levels of Ki-67 were observed in the Huh7-shABCA8 group and lower levels of Ki-67 were observed in the HCCLM3-ABCA8 group when compared to the corresponding control groups and black groups (Fig. 2f). In short, our results showed that ABCA8 has a tumor inhibition effect both in vitro and in vivo.

Metastasis is the major cause of poor prognoses for HCC patients. Therefore, we explored the effect of ABCA8 on the invasive and migratory ability of HCC cells. A wound-healing experiment showed that silencing ABCA8 accelerated the area of scratch growth in HCC cells, whereas, overexpression of ABCA8 slowed the growth of scratches (Fig. 3a and Additional file 4: Figure S3a). A transwell assay (Matrigel-coated or non-Matrigel-coated) showed that the invasive and migratory abilities of HCC cells were increased by ABCA8 silencing, whereas ABCA8 overexpression weakened the ability of cells to migrate and invade (Fig. 3b and Additional file 4: Figure S3b).

We constructed a pulmonary metastasis model (as described in our methods) to demonstrate this characteristic in vivo. We found that the Huh7-shABCA8 group had more numerous and larger pulmonary metastatic nodules than the control group. In contrast, the size and number of pulmonary metastatic nodules were suppressed in the HCCLM3-ABCA8 group compared to the control group (Fig. 3c).

Epithelial to mesenchymal transformation (EMT) is the first stage in the migration of individual and collective cells. It was thought to involve cells from the static outer membrane losing their apical polarity and cellular adhesion characteristics, eventually becoming single cells and migrating between tissues with mesenchymal phenotypes [11]. To clarify the relationship between ABCA8 and EMT, markers of EMT were evaluated by qPCR and western blots. When ABCA8 was silenced in Huh7 cells, the expression of E-cadherin (an epithelial marker) was decreased and the expression of N-cadherin and vimentin (mesenchymal markers) were increased (Fig. 4a, b and Additional file 5: Figure S4). Inversely, ABCA8 overexpression enhanced the expression of E-cadherin and decreased that of N-cadherin and vimentin in HCCLM3 cells compared to control cells. The results of immunofluorescence experiments corroborated that the reduction of ABCA8 can induce the EMT process (Fig. 4c). As we know, Snail, Slug, ZEB1/2, and Twist are important upstream transcription regulators of EMT [12‐14]. However, only ZEB1 expression decreased or increased with overexpression or silencing of ABCA8, respectively (Fig. 4d). This result was confirmed by qPCR (Additional file 6: Figure S5). To clarify the molecular mechanism by which ABCA8 regulates EMT, we sought to detect key molecules in several pathways. We found that only the expression of p-ERK was up- and down-regulated by ABCA8 silencing and overexpression, respectively. Changes in p-AKT and TGF-β1 were not observed (Fig. 4e). In order to detect the expression of ERK, p-ERK, ZEB1 and EMT-related markers in vivo models, proteins were extracted from subcutaneous tumors for Western blots. The results showed that the expression of ERK and other proteins in subcutaneous tumors was basically consistent with that in cells (Additional file 7: Figure S6).

SCH722984, the novel specific inhibitor of ERK1/2 [15, 16], was used to further investigate the mechanisms of ABCA8-induced HCC progression. Western blots demonstrated that the expression of p-ERK was partially reduced by treating Huh7-shABCA8 cells with SCH722984 (Fig. 5a). An analogous effect was detected in HCCLM3-ABCA8 cells (Fig. 5a). The CCK8 assay showed that cell viability was dramatically decreased when treated with SCH722984 (Fig. 5b). The same result was observed in colony formation experiments (Fig. 5c). Similarly, when treated with SCH722984, the invasion and migration capacity of Huh7-shABCA8 and HCCLM3-ABCA8 cells was partially blocked (Fig. 5d, e and Additional file 8: Figure S7a, b). These outcomes suggest that the ERK signaling pathway is vital for ABCA8-induced HCC progression and that ERK is a key molecule in ABCA8-induced EMT.

Many studies have shown that miRNAs can bind to the 3’UTR of mRNA, affecting its stability and function. Moreover, miRNAs are associated with many biological processes, including tumor progression [17‐19]. Therefore, in order to further understand the regulatory mechanism of ABCA8 as it relates to HCC, we identified miRNAs that could be upstream regulators of ABCA8 through the publicly accessible databases, TargetScan (http://​www.​targetscan.​org/​vert_​72/​), miRDB (http://​mirdb.​org/​) and miTarBase (http://​mirtarbase.​mbc.​nctu.​edu.​tw/​php/​index.​php). We screened six candidate miRNAs with highly conserved binding sequences to ABCA8 (Additional file 9: Figure S8). The miRNA inhibitors were transfected into HCC cell lines and the mRNA expression of ABCA8 was detected by qPCR (Additional file 10: Figure S9). Among the six miRNAs tested, ABCA8 levels only increased when miR-374b-5p was inhibited. Previous studies have reported that miR-374b-5p is highly expressed in HCC and that high levels of miR-374b-5p expression can promote tumor proliferation and metastasis [20]. This result in combination with our own data led us to focus our subsequent experiments on miR-374b-5p. Through a luciferase reporter assay, we tested for a link between miR-374b-5p and ABCA8. As expected, luciferase activity was suppressed by miR-374b-5p in the wild-type (WT) ABCA8 3′-UTR group, but no change was detected for the mutant group (Fig. 6a). qPCR illustrated that miR-374b-5p was upregulated in HCC tissues (Fig. 6b) and HCC cell lines (Additional file 11: Figure S10). Our results are consistent with previous reports [20]. Furthermore, clinicopathological analysis showed that high levels of miR-374b-5p expression were positively correlated with tumor diameter (p = 0.0166), metastasis (p = 0.0021) and TNM stage (p = 0.0090, Table 2). Patients with high levels of miR-374b-5p and low levels of ABCA8 had lower survival rates compared with patients with low levels of miR-374b-5p and high levels of ABCA8 (Fig. 6c). We transfected Huh7 and HepG2 cell lines with lentivirus-miR-374b-5p and HCCLM3 and SK-Hep-1 cell lines with lentivirus-anti-miR-374b-5p, followed by western blotting to verify changes in ABCA8 expression levels (Fig. 6d). These results demonstrated that miR-374b-5p directly targets ABCA8 in HCC. The high expression of miR-374b-5p combined with low expression of ABCA8 was correlated to a poor prognosis for HCC patients.

The outcome of western blots demonstrated that the expression of ABCA8 and E-cadherin were up-regulated after miR-374b-5p silencing, while the expression of p-ERK, N-cadherin, Vimentin and ZEB1 were down-regulated. ABCA8 silencing can partially restore this effect, while the expression of total ERK was not affected by either miR-374b-5p silencing or ABCA8 silencing in HCCLM3 cell line (Fig. 7a). Overexpression of miR-374b-5p in the Huh7 cell line had the opposite effect and these effects could be reversed by overexpression of ABCA8.

In order to detect the role of miR-374b-5p on cell proliferation, colony formation experiments were performed. Overexpression of miR-374b-5p promoted HCC cell proliferation, which could be partially blocked by overexpression of ABCA8. Silencing miR-374b-5p had the opposite effect and that effect is partially reversed by silencing ABCA8 (Fig. 7b). Transwell assays (migration or invasion) indicated that silencing miR-374b-5p inhibited the migratory and invasive capability of HCCLM3 cells, and the overexpression of miR374b-5p facilitated Huh7 cell migration and invasion. The migratory and invasive capabilities, that are strengthened or weakened by overexpression or silencing of miR-374b-5p, can be partially reversed by overexpression or silencing of ABCA8, respectively (Fig. 7c and Additional file 12: Figure S11). These outcomes illustrate that miR-374b-5p promotes the development of HCC through the ABCA8/ERK/Zeb1 axis (Fig. 7d).