Fresh liver cancer tissue samples and neighboring non-cancerous liver tissue samples were obtained from 52 patients who underwent primary surgical resection of liver tumors at the Shanghai Eastern Hepatobiliary Surgery Hospital affiliated to the Second Military Medical University (Shanghai, China) between March and September 2015. Non-cancerous tissues were obtained 5–10 cm away from the primary tumor. Follwoing resection, all the samples were snap-frozen in liquid nitrogen and stored at −80°C prior to RNA extraction. Patients who received preoperative treatment were excluded from the study. All included patients provided written informed consent prior to enrollment. The protocols involving human samples were conducted in conformity with the ethical principles of research and approved by the Human Resources Ethics Committee of Shanghai East Hospital affiliated to Tongji University (Shanghai, China). The histopathological diagnosis of all samples was respectively verified by two pathologists. The clinical staging was based on the 7th edition of the AJCC Cancer Staging Manual. The main demographic and clinicopathological characteristics of the patients are presented in Table I.

The human HCC cell lines, Bel-7402, Sk-hep-1, Hep3B, Huh7 and HepG2, were obtained from the Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, China). The liver cell line, L02, was purchased from Li Yandong Research Group Shanghai East Hospital affiliated to Tongji University School of Medicine.

Each cell line was cultured in minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS; HyClone, Logan, UT, USA) as well as 100 U/ml penicillin and 100 µg/ml streptomycin. All the above-mentioned cells were maintained in a humidified incubator at 37°C with 5% CO₂.

Protein was extracted from the cells and tissues using RIPA lysis buffer [1% NP40, 0.1% sodium dodecyl sulfate (SDS), 100μg/ml phenylmethylsulfonyl fluoride, 0.5% sodium deoxycholate, in PBS] on ice. The super-natants were collected following centrifugation at 12,000 × g at 4°C for 20 min. The protein concentration was determined using a BCA protein assay kit (Bio-Rad, Shanghai, China), and whole lysates were mixed with 4X SDS loading buffer [125 mmol/l Tris-HCl, 4% SDS, 20% glycerol, 100 mmol/l dithiothreitol (DTT) and 0.2% bromophenol blue] at a ratio of 1:3. The samples were heated at 100°C for 5 min and were separated on SDS-polyacrylamide gels. The separated proteins were then transferred onto PVDF membranes (Millipore, Bedford, MA, USA). The membranes were first probed with a primary antibody. After blocking with 5% skim milk for 2 h, the membranes incubated with primary antibodies [rabbit anti-KLF4 (1:500, Santa Cruz Biotechnology, Santa Cruz, CA, USA). The secondary antibody was mouse anti-GAPDH (1:3000, Santa Cruz Biotechnology).

Total RNA was isolated from the tissue samples and cultured cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's insstructions. The concentration of extracted RNA was measured using a NanoDrop ND-1000 Spectrophotometer (Agilent, Santa Clara, CA, USA). RNA was reverse transcribed into complementary DNA (cDNA) using the PrimeScript RT reagent kit with gDNA Eraser (Takara, Dalian, China). Quantitative PCR (qPCR) was performed using the SYBR PrimeScript RT-PCR kit (Takara, Shiga, Japan) and the ABI 7500 System (Applied Biosystems, Foster City, CA, USA) according to manufacturer's instructions. The relative expression was calculated via the comparative cycle threshold (CT) method and was normalized to the expression of U6 small RNA. The primers used are listed in Table II. The differential expression level was calculated using the 2^−ΔΔCt formula. All the experiments were conducted at least 3 times.

The miR-31 mimic, inhibitor, Klf4 siRNA and negative controls (mimic NC or inhibitor NC) were synthesized by GenePharma, Shanghaim, China. The previously described plasmid Klf4 and the empty control vector plasmid were obtained from Xie Keping Research Group Shanghai East Hospital affiliated to Tongji University School of Medicine. The transfection of the plasmids and siRNAs, respectively, into the HCC cells was performed using Lipofectamine 2000 (Invitrogen) transfection reagent. The relative levels of Klf4 and miR-31 in the transfected cells were examined by RT-qPCR. The cells were transfected with the plasmids or siRNAs at various concentrations as indicated 48 h before the performance of functional assays.

A total of 2×10⁶ Hep3B cells grown in the exponential phase were seeded in 6-well plates and cultured for 2 days. The cells were then harvested for total RNA using the miRNeasy kit (Qiagen). A total of 100 ng of total RNA was assayed using the Human nCounter miRNA Assay 2.0 kit following the manufacturer's instructions (NanoString Technologies, Seattle, WA, USA). Differences in miRNA expression were analyzed using the NanoSTRIDE software program with default settings. Clustering of the differentially expressed genes and heatmap generation was performed using the GenePattern Server (genepattern.broadinstitute.org). The volcano plot displaying the significance of the miRNA differences was produced using R version 3.0.2.

For the binding of Klf4 to the miR-31 promoter, the 1.5 kb region directly upstream of miR-31 transcription binding site was amplified by PCR and inserted into the pGL3 vector (Promega, San Luis Obispo, CA, USA). The Hep3B and Sk-hep-1 cells were co-transfected with either pcDNA3.1 (Promega) or the pcDNA3.1-Klf4 and miR-31 promoter. At 48 h after transfection, luciferase activity was analyzed according to the Dual-Luciferase reporter assay system (Promega), using a GloMax fluorescence reader (Promega).

The Sk-hep-1 and Bel-7402 cells were transfected with the indicated miR-31 mimic or inhibitor (50 nM), the Klf4 overexpression plasmid or and Klf4 siRNA (siKlf4; 120 ng), plated in 96-well plates at 3,000 cells/well and maintained in culture medium. They were assessed at days 1–6 of culture with 0.5 mg/ml cell counting kit 8 (CCK8; Dojindo Laboratories, Kumamoto, Japan) at 37°C for 3 h and then shaken for 20 min. The optical density was determined at 450 nm using an enzyme-linked immunosorbent assay reader (Dasit, Milan, Italy). Each experiment was conducted for at least 3 times and the average of the results was analyzed.

The transfected HCC cells (5×10⁴) in serum-free medium were placed into the upper part of a Transwell chamber in a 24-well format with 8 mm diameters (Corning, Corning, NY, USA). In the bottom chamber, 800 µl of normal MEM medium containing 10% FBS were added as a chemoattractant and the chambers were incubated for 24–48 h at 37°C and 5% CO₂. The cells on the upper part of the Transwell chamber were removed using a cotton swap, and the cells that had migrated through the membrane were stained with 0.05% crystal violet for 2 h. Finally, the migrated cells were counted in 5 random fields under a microscope (Olympus, Shanghai, China) and the average number of 5 fields was calculated. All assays were performed in triplicate and repeated 3 times.

For the wound healing assay, 3.5×10⁵ of examined HCC cells were seeded into 6-well plates to achieve 90% confluence. Wounds were produced in confluent monolayer cells using a plastic tip, the cell debris were removed using phosphate-buffered saline (PBS), and 0.5% FBS-containing MEM was added. The scratched cells were then incubated at 37°C with 5% CO₂. The initial scratched gap breadth (0 h) and the residual scratched gap breadth (48 h) were measured using a light microscope (Nikon, Tokyo, Japan).

Quantitative data are presented as the means ± SEM. Parameters of two-tailed, 95% CI were used for statistical analysis. Only P-values <0.05 were considered to indicate statistically significant differences.

To identify candidate miRNAs that may play a role in HCC, RT-qPCR and western blot analysis of Klf4 expression in the Hep3B cells transfected with the Klf4 plasmid were initially performed (Fig. 1A). We found that the mRNA and protein expression of Klf4 was significantly increased following the overexpression of Klf4. We then compared miRNA expression levels between the empty vector- and Klf4 overexpression vector-transfected Hep3B cells using the nCounter NanoString platform. Using RNA derived from the Hep3B cells transfected with the empty vector or Klf4 overexpression vector under standard culture conditions, we found that a number of miRNAs were differentially expressed. Of the 2,565 miRNAs assayed by this method, a significant difference in expression was observed in 436 miRNAs (17.0%; P<0.05, 1.5-fold), including 216 that were downregulated and 220 that were upregulated in the Hep3B cells transfected with the empty vector or Klf4 overexpression vector. The heatmap identified the 18 most differentially regulated miRNAs from the miRNA array of Hep3B cells transfected with the empty vector or Klf4 overexpression vector, including 8 upregulated and 10 downregulated miRNAs (Fig. 1B). We found that the miR-31 cluster was significantly upregulated in the Hep3B cells (fold change, 1.92; P-value, 0.009). At last, Volcano plot revealed the profile of the differentially expressed miRNAs in the Hep3B cells. This plot outlines the fold change (x-axis) and significance level of expression as the log₁₀ P-value (y-axis). The red dots on the left side represent the downregulated miRNAs in the Hep3B cells transfected with the Klf4 overexpression vector compared to the cells transfected with the empty vector; the red dots on the right side represent the upregulated miRNAs in the Hep3B cells transfected with the Klf4 overexpression plasmid compared with the cells transfected with the empty vector. The black dots indicate miRNAs which did not exhibit a significant change in expression. Significance was determined with a P-value cut off of 0.05 and a 1.5-fold change (Fig. 1C). We found that miR-31 (the red dots on the right) was significantly upregulated in the Hep3B cells.

In this study, we explored the relative expression levels of Klf4 and miR-31 in 52 pairs of HCC tissues and neighboring noncancerous liver tissues by RT-qPCR. The expression of Klf4 and miR-31 was significantly downregulated in the cancerous compared to the non-cancerous tissues (Fig. 2A and B). The downregulation of Klf4 was found to inversely correlate with the clinical stage, T classification and hepatitis B infection, whereas the downregulation of miR-31 was directly associated with the clinical stage in patients with HCC (Table I). Of note, the decreased expression of miR-31 in HCC was observed in 65.4% (34/52) of the cases (Fig. 2C). Additionally, the endogenous expression of Klf4 and miR-31 was analyzed by qRT-PCR in six chosen human liver cell lines, including an immortalized normal hepatic cell line (Fig. 2D). The liver cancer cell lines (Bel-7402, Sk-hep-1, Hep3B, Huh7 and HepG2) exhibited relatively low Klf4 and miR-31 expression levels compared to the non-cancer cell line, L02. These findings again suggest that the expression of Klf4 and miR-31 is suppressed in HCC and that a low expression of Klf4 and miR-31 is associated with the biological process of tumorigenesis in HCC.

The expression of Klf4 in HCC was examined by RT-qPCR and western blot analysis. The Klf4 mRNA and protein levels were significantly increased when the Sk-hep-1 and Bel-7402 cells were transfected with the Klf4 overexpression plasmid (Fig. 3A). We then examined the expression of miR-31 by RT-qPCR, and found that it was significantly increased and decreased when the Hep3B, Bel-7402 and Sk-hep-1 cells were transfected with the Klf4 overexpression plasmid and or siKlf4, respectively (Fig. 3B–D). Moreover we examined the pri-miR-31 levels by RT-qPCR, and the results revealed that the levels were significantly increased and decreased when the Hep3B, Bel-7402 and Sk-hep-1 cells were transfected with the Klf4 overexpression plasmid and siKlf4, respectively (Fig. 3E–G).

To explore this regulation of miR-31 expression, we analyzed the 1.5 kb region of miR-31 upstream and found 3 binding sites of Klf4 (mut1-3). We then examined the effects of the transcription factor Klf4 on miR-31 promoter-driven luciferase activity in Hep3B cells (Fig. 3H). The mutation of the Klf4 binding site 3 (mut3) abolished the effects of Klf4 on the relative luciferase activity in the Hep3B and Sk-hep-1 cells, while mutations of other binding sites had no significant effect (Fig. 3I and J). Thus, Klf4 regulates the expression of miR-31 by binding to site 3, indicating that Klf4 is involved in the transcriptional regulation of miR-31 by directly binding to the miR-31 promoter. A positive correlation of Klf4 and miR-31 was also revealed in the HCC tissues (Fig. 3K). Taken together, these findings suggest that the transcription factor Klf4 is a functional regulator of miR-31 in HCC.

To ascertain the biological functions of Klf4 and miR-31 in HCC, following the ectopic expression of Klf4 and miR-31, we examined the growth rate of two different liver cancer cell lines, Sk-hep-1 and Bel-7402, by CCK8 assay. Transfection with siKlf4 promoted Sk-hep-1 cell proliferation, and transfection with miR-31 mimic attenuated the growth-promoting effects induced by transfection with siKlf4 (Fig. 4A). Klf4 overexpression suppressed Bel-7402 cell proliferation, while transfection with miR-31 inhibitor attenuated the growth inhibitory effects of Klf4 (Fig. 4B). The Sk-hep-1 and Bel-7402 cells were seeded onto 96-well plates and respectively transfected with Klf4 or siKlf4, miR-31 mimic or inhibitor for 1–6 days. Cell viabilities were determined by CCK8 assays. Cell growth was measured every 24 h. Mimic NC and inhibitor NC represents negative control miRNA.

As shown in Fig. 4C, transfection with miR-31 mimic decreased the number of Sk-hep-1 migrating cells, which had been increased by transfection with siKlf4. Similarly, transfection with miR-31 inhibitor increased the number of Bel-7402 migrating cells, which had been decreased by transfection with Klf4 overexpression vector. Consistent with the Transwell assay results, following transfection with the miR-31 mimic, the Sk-hep-1 cells migrated more slowly than the controls and the siKlf4-transfected cells, as shown by wound healing assays (Fig. 4D). By contrast, the Bel-7402 cells transfected with miR-31 inhibitor exhibited a more rapid migration rate compared with the controls and the cells transfected with the Klf4 overexpression plasmid, as shown by wound healing assays (Fig. 4D). These results demonstrate that miR-31 contributes to the regulation of liver cancer cell motility and progression.