The HepG2 cell line was obtained from the American Type Culture Collection (cat. no. HB-8065; Manassas, VA, USA) and maintained at 37°C, 95% humidity and 5% CO2 in Dulbecco's modified Eagle's medium (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) with 10% fetal bovine serum (FBS) (Invitrogen; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin. The HepG2 cell line is a hepatoblastoma cell line previously misidentified as hepatocellular carcinoma (23). The HepG2 cells were seeded in a 24-well plate at a concentration of 1.5×105 cells/well (diluted with antibiotic-free complete medium) 1 day prior to transfection with miRNAs.

TargetScan (www.targetscan.org), DIANA micro T (www.diana.imis.athena-innovation.gr), microRNA.org (www.microrna.org) and RepTar (reptar.ekmd.huji.ac.il) were used to predict the miRNAs that potentially bind to the 3′-UTR of CKα mRNA, using the default settings of the respective programs. The minimum free energy of the predicted miRNAs was calculated using RNAfold (rna.tbi.univie.ac.at), mFold (unafold.rna.albany.edu) and KineFold (kinefold.curie.fr). The site features of the mRNA-miRNA pairings in terms of G:C, A:U and G:U matches, mismatches and bulges in miRNAs and mRNAs were analyzed manually to search for high complementarity at the seed region and the minimum number of bulges. The potential miRNAs were additionally analyzed based on the classification of miRNA target sites, according to Brennecke et al (24), types of matching to seed region, according to Lewis et al (25)and site contexts, based on Grimson et al (26).

Non-targeting (miRIDIAN microRNA Mimic Negative Control #1; GE Healthcare Dharmacon, Inc., Lafayette, CO, USA), GAPDH-targeting (Mimic Housekeeping Positive Control #2; GE Healthcare Dharmacon, Inc.), miR-876-5p (MISSION^® MicroRNA mimic HMI0929, 5′-UGGAUUUCUUUGUGAAUCACCA-3′; Sigma Aldrich; Merck KGaA, Darmstadt, Germany) and miR-646 (MISSION^® MicroRNA mimic HMI0877, 5′-AAGCAGCUGCCUCUGAGGC-3′; Sigma Aldrich; Merck KGaA) miRNAs were re-suspended in 1X miRNA buffer (GE Healthcare Dharmacon, Inc.) to prepare a stock solution of 20 µM. The non-targeting (negative control) miRNA sequence was based on the Caenorhabditis elegans cel-miR-67 that was confirmed to exhibit minimal sequence identity with miRNAs from humans, mice and rats. The GAPDH-targeting miRNA (referred to as miRNA-GAPDH) targets the 3′UTR of GAPDH mRNA. The miRNA stock solution was further diluted with RNase-free water to obtain a 2 µM working solution. Prior to the transfection, the miRNAs were diluted with serum free Opti-Minimum Essential Medium (MEM)^® I (Thermo Fisher Scientific, Inc.), such that the final concentrations in the treatment plate were 25, 50 and 100 nM. In a separate tube, DharmaFECT 2 transfection reagent (GE Healthcare Life Sciences, Little Chalfont, UK) was additionally diluted with the serum free Opti-MEM^®, according to the manufacturer's protocol. The diluted miRNAs and transfection reagent were mixed for 20 min at room temperature to form transfection complexes. The culture medium from the 24-well plate was removed and 400 µl complete medium plus 100 µl transfection complexes were added into each well and cultured at 37°C with 5% CO₂, for a duration of 24 to 48 h. For screening of miRNA mimics for the downregulation of CKα mRNA levels in the HepG2 cancer cell line, the cells were transfected for 48 h.

Following the transfection, the medium was removed and the cells were trypsinized, harvested and washed with PBS prior to RNA extraction using the RNeasy Mini kit (Qiagen GmbH, Hilden, Germany), according to the manufacturer's protocol. Any residual DNA was eliminated by adding RNase-Free DNase I (Qiagen GmbH) onto the spin column and incubating for 20 min at room temperature. The concentration of the purified total RNA was measured at 260 nm and the OD_260/280 was determined using a BioPhotometer Plus (Eppendorf, Hamburg, Germany). The RNA integrity and size distribution were assessed by running the purified RNA samples on a 1% agarose gel.

The first strand cDNA was synthesized from the extracted RNA using the RevertAid™ H Minus First Strand cDNA Synthesis kit (Fermentas; Thermo Fisher Scientific, Inc.) with a MyCycler Thermal Cycler (Bio-Rad Laboratories, Inc., Hercules, CA, USA). A total of 1 µg total RNA was mixed with 0.5 µl each of oligo(dT)18 (50 µM) and random hexamers (50 µM), with the final volume made up to 12 µl with distilled water. The mixture was incubated at 65°C for 5 min. Subsequently, 4 µl 5X reaction buffer, 1 µl RiboLock™ RNase Inhibitor (20 U/µl), 2 µl dNTP mix (10 mM) and 1 µl RevertAid™ H Minus M-MuLV Reverse Transcriptase (200 U/µl) was added into the mixture, and incubated at 25°C for 5 min and 42°C for 1 h; the reaction was terminated by a final incubation at 70°C for 5 min. The synthesized cDNA was stored at −20°C until use.

The qPCR reactions were performed using an ABI 7500 Fast Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The reactions were performed in 96-well plates (Axygen; Corning Incorporated, Corning, NY, USA). A relative quantification method with ubiquitin C (UBC) and tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein ζ (YWHAZ) as reference genes was used. A negative control without template DNA was run together with the other samples to detect possible contamination or non-specific amplification in the reaction. The PCR reactions with a final volume of 25 µl consisted of 12.5 µl Power SYBR-Green I Master Mix (Applied Biosystems; Thermo Fisher Scientific, Inc.), 1.5 µl gene specific primers, 1 µl 1:2 diluted cDNA and water. The cycling parameters were 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 10 sec at 95°C and 1 min at 60°C. Following the amplification, PCR specificity was verified by melting curve analysis with temperatures ranging between 60 and 95°C with 0.1°C increments. All reactions were run in triplicate. The reference genes (UBC and YWHAZ) were selected based on a previous study (27). The forward primer sequence for UBC was 5′-CTGATCAGCAGAGGTTGATCTT-3′ and its reverse primer sequence was 5′-GTCTTGCCAGTGAGTGTCTT-3′. For YWHAZ, the forward primer sequence was 5′-TTCTTGATCCCCAATGCTTC-3′ and the reverse primer sequence was 5′-AGTTAAGGGCCAGACCCAGT-3′. The forward and reverse primer sequences for GAPDH were 5′-CAAGGTCATCCATGACAACTTTG-3′ and 5′-GTCCACCACCCTGTTGCTGTAG-3′, respectively. The forward primer sequence for total CKα was 5′-TCAGAGCAAACATCCGGAAGT-3′ and the reverse primer sequence for total CKα was 5′-GGCGTAGTCCATGTACCCAAAT-3′. The forward and reverse primer sequences for CKα2 specific amplification were 5′-GGCCTTAGCAACATGCTGTTC-3′ and 5′-AGCTTGTTCAGAGCCCTCTTT-3′, respectively. Relative gene expression levels normalized to the geometric mean of UBC and YWHAZ C_q values were determined by the 2^−ΔΔCq method (28).

The MTT cell viability assay was performed by seeding 1.5×10⁴ HepG2 cells in 100 µl antibiotic-free complete medium into each well of a 96-well plate and incubating at 37°C with 5% CO₂ 1 day prior to the transfection with miR-876-5p mimic. The cells were transfected with 25 nM miR-876-5p for 48 h with a final volume of 100 µl in each well. The MTT assay was performed by replacing the medium following transfection with 100 µl fresh culture medium, followed by the addition of 10 µl 12 mM MTT stock solution and incubation at 37°C for 4 h. Following that, 85 µl medium was removed and 50 µl dimethyl sulfoxide (DMSO) as the solubilizing agent to dissolve the formazan dye was added to each well and thoroughly mixed by pipetting, prior to further incubation at 37°C for 10 min. Finally, the absorbance at 540 nm was read by using microplate reader (Bio-Rad Laboratories, Inc.) and the percentage cell viability was calculated based on the average absorbance values for the samples and control.

The morphological alterations in HepG2 cells following treatment with either DMSO or the desired miRNAs for 48 h were observed under a scanning electron microscope (Quanta 200; FEI; Thermo Fisher Scientific, Inc.). Following transfection, cells were grown on a plastic cover slip (Thermo Fisher Scientific, Inc.) and were fixed with McDowell-Trump fixative solution [1% (v/v) glutaraldehyde and 4% (v/v) formaldehyde in 0.1 M phosphate buffer; pH 7.2] (29) at 4°C for 24 h. The following day, the samples were washed with 0.1 M PBS and incubated in 1% osmium tetroxide at room temperature for 1–2 h for secondary fixation. The samples were dehydrated via sequential incubation of 15 min each in 50, 75, 95 and 100% ethanol. The dehydrated samples were immersed in hexamethyldisilazane (HMDS): acetone (1:1 ratio) for 15 min and in 100% HMDS three times, for 15 min each, prior to being left overnight in a desiccator. The dried samples were mounted onto the sample stub prior to being coated with gold and viewed under the scanning electron microscope.

All data were analyzed using Student's t-test and one-way analysis of variance with the Bonferroni post hoc test. P<0.05 was considered to indicate a statistically significant difference, and all analyses were performed using SPSS software version 22.0 (IBM Corp., Armonk, NY, USA). All data are presented as the mean ± standard error of mean from three independent experiments.

A total of 54 non-repeating miRNAs were predicted to target CKα mRNA by the online programs used in the present study. Of these, 22 miRNAs were shortlisted based on their sequence conservation among different species. Subsequently, the miRNAs were selected based on favorable minimum free energy (≤-1.0 kcal/mol), miRNA-mRNA binding site features at the seed region, complementarity (either the stronger 5′dominant or the weaker 3′compensatory), types of matching to the seed region (with site efficacy following this order: 8mer>7mer-m8>7mer-A1>6mer>Offset 6 mer) and other site contexts that improve efficacy (extra Watson-Crick base pairing at nucleotides 12–17 of the miRNA, and target sites that are located ≥15 nucleotides downstream of the stop codon, near either end of the 3′-UTR).

Based on the selection criteria described above, two miRNA candidates (miR-876-5p and miR-646) were selected for experimental validation. The principal features of these two miRNAs and the locations of their target sites at the 3′-UTR of CKα mRNA are presented in Fig. 1. In addition to exhibiting other favorable characteristics, miR-876-5p was selected based on its 8mer site type and two Watson-Crick base pairings at nucleotides 12–17. Although miR-646 only has a 7mer-m8 site type, it was selected for the four Watson-Crick base pairings at nucleotides 12–17 and the high GC base pairing throughout the whole miRNA sequence with its target site. In summary, the two miRNAs possess favorable seed types, pairings at nucleotides 12–17, distances from the stop codon, minimum free energy, and the absence of GU wobble pairs and mismatches at the seed region. The two miRNAs additionally possess the 5′dominant canonical target site type, which is more effective than the 5′dominant seed and 3′compensatory types.

The levels of CKα mRNA were determined by reverse transcription-qPCR following transfection of HepG2 cells with 25 nM different miRNA mimics for 48 h. Fig. 2 illustrates that miR-876-5p mimic significantly (P<0.05) downregulated the expression of total CKα and CKα2 by ~0.8 and 0.7-fold, respectively. miR-646 did not significantly affect the expression levels of total CKα and CKα2 mRNA. The successful transfection of HepG2 cells with the miRNAs was confirmed by the GAPDH-targeting miRNA (positive control miRNA), which resulted in strong downregulation of GAPDH to ~10% compared with the negative control (90% downregulation). The results additionally demonstrated that CKα2 was the predominant CKα isoform, since the levels of downregulation by miR-876-5p for total CKα and CKα2 were very similar. Due to the significant downregulation of CKα mRNA by miR-876-5p, the effects of transfection concentration and duration were subsequently determined in the present study.

A total of three concentrations of miR-876-5p were tested at the 48 h transfection duration. According to the results presented in Fig. 3, the total CKα and CKα2 mRNA expression levels were significantly downregulated to ~0.7 fold (P<0.05) compared with the negative control when the cells were transfected with 25 nM miR-876-5p. The total CKα and CKα2 mRNA expression levels were significantly downregulated (P<0.05) to 86% and 88%, respectively, compared with the negative control following the transfection with 50 nM miR-876-5p mimic. Transfection with 100 nM miR-876-5p mimic did not significantly downregulate the total CKα and CKα2 mRNA expression levels compared with the negative control (P>0.05). Comparison between the three concentrations demonstrated that there were significant differences between CKα2 levels in cells treated with 25 and 50 nM, and 25 and 100 nM (P<0.05). Notably, the downregulation caused by miR-876-5p may be underestimated for all the concentrations tested, as the level of GAPDH mRNA was not effectively downregulated by the positive control miRNA as in previous experiments. However, the overall results demonstrated that the concentration of 25 nM miR-876-5p mimic was able to produce the most marked downregulation of CKα gene expression.

HepG2 cells were transfected with 25 nM miR-876-5p mimic at three different transfection durations. Fig. 4 illustrates that the expression levels of total CKα mRNA were significantly downregulated to 26, 31 and 38% compared with the negative control following transfection for 24, 36 and 48 h, respectively. The expression levels of CKα2 mRNA were additionally downregulated to 38, 34 and 25% compared with the negative control at the 24, 36 and 48 h transfection durations, respectively. All the downregulations were significant when compared with the negative control (P<0.05). Comparison between the three transfection durations demonstrated that there were significant differences in CKα2 expression levels between all the durations tested, while total CKα expression levels were only significantly different between cells transfected for 24 and 48 h (P<0.05). The results were more reliable compared with the experiments for the optimization of miR-876-5p concentration, as the GAPDH mRNA levels were markedly downregulated by the positive control miRNA at all transfection durations tested. Since human CKα2 is more active than the CKα1 isoform (11) and the strongest downregulation of CKα2 mRNA expression levels in the HepG2 cell line was obtained with 48 h transfection of 25 nM miR-876-5p mimic, these parameters were selected for subsequent experiments. Taken together, the results affirm the ability of miR-876-5p to downregulate CKα gene expression.

The viability of HepG2 cells following transfection with miR-876-5p was determined by MTT cell proliferation assay. As presented in Fig. 5, the viability of HepG2 cells transfected with 25 nM miR-876-5p mimic for 48 h was ~25% lower compared with cells transfected with the non-targeting miRNA. There was a significant difference (P<0.05) between the downregulation levels of miR-876-5p and the negative control miRNA.

Scanning electron microscopy was used to investigate the effect of miR-876-5p transfection on HepG2 cellular morphology, particularly the effect on the cell membrane structure. Fig. 6 presents representative images of non-treated and miR-876-5p-treated HepG2 cells. The majority of the non-treated cells (upper panels) appeared healthy, exhibiting a round shape with filopodia structures at the edges of their surfaces. By contrast, the cells transfected with miR-876-5p exhibited signs of apoptosis, including cell shrinkage, membrane blebbing and loss of filopodia on their membranes. Since the non-treated cells were not exposed to the liposomal transfection reagent, it must be noted that the apoptosis-indicating morphology observed in miR-876-5p-transfected cells may be due to the transfection reagent.