In silico analyses was performed to identify the putative miRNA targets using TargetScan Human V5.2 (http://www.targetscan.org/) (Whitehead Institute for Biomedical Research, Cambridge, MA, USA), a database of conserved 3′UTR targets. TargetScan provides accurate ranking of the predicted targets of miRNA based on total context+ score, which is the sum of the contribution of six targeting factors including site type, site number, site location, local AU content, 3′-supplementary pairing, target site abundance and seed-pairing stability. The total context+score predicts the relative repression of mRNA with 3′UTR, with low context scores being more favorable (23). The web tool Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.7 (http://david.abcc.ncifcrf.gov/summary.jsp) (SAIC-Frederick, Inc., Frederick, MD, USA), which is made up of an integrated biological knowledgebase and analytic tools (24), was then employed, using default parameters, to perform gene-annotation enrichment analyses on TargetScan's predicted miRNA targets that has a total context+ score of <0. Data from TargetScan and DAVID were combined to generate a hypothetical pathway of the relationship between the miRNAs and their gene targets.

Human lung adenocarcinoma cell line A549 [Cancer Research Initiative Foundation (CARIF), Subang Jaya Medical Centre, Subang Jaya, Malaysia] was cultured in RPMI-1640 (SH30027.01; HyClone Laboratories-GE Healthcare Life Sciences, Pittsburgh, PA, USA) whereas SK-LU-1 cells (LA-HL-045; AseaCyte, Pvt. Ltd., Kuala Lumpur, Malaysia) were cultured in MEM-α (32561037; Gibco, Waltham, MA, USA). All cells were supplemented with 10% fetal bovine serum (FBS) (SV30160.03; HyClone Laboratories-GE Healthcare Life Sciences) and maintained at 37°C in a humidified incubator containing 5% CO₂.

Cells were seeded 24 h prior to transfection with miR-608 mimics (C-300933-01-0010; GE Healthcare Dharmacon, Lafayette, CO, USA), non-specific mimic controls (mimic NC) (CN-001000-01-20; GE Healthcare Dharmacon), miR-608 inhibitors (IH-300933-03-0010; GE Healthcare Dharmacon) or non-specific antimiR controls (inhibitor NC) (IN-001005-01-20; GE Healthcare Dharmacon) at a final concentration of 80.0 nM using DharmaFECT reagent (T-2001-03; GE Healthcare Dharmacon), as per the manufacturer's protocol.

Wild-type 3′UTR of AKT2 containing predicted miR-608 binding sites and/or its corresponding mutant sequences were cloned into the pmirGLO Dual-Luciferase miRNA expression vector (E1330; Promega, Madison, WI, USA). A549 cells were plated 24 h prior to co-transfection with 40.0 ng of pmirGLO constructs and 80.0 nM of miR-608 mimic/inhibitor or mimic NC/inhibitor NC using DharmaFECT reagent. Luciferase activity was analyzed 48 h post-transfection using the Dualluciferase reporter assay system (E2920; Promega), as per the manufacturer's protocol and detected on the GloMax Multi Luminescence Multimode Reader (Promega). Relative luciferase activity was normalized against Renilla luciferase activity.

Protein was extracted using the NE-PER^® Nuclear and Cytoplasmic Extraction kit (78833; Thermo Fisher Scientific, Waltham, MA, USA) 48 h post-transfection, as per the manufacturer's protocol. Protein lysates were separated by electrophoresis in 12% SDS-PAGE and then electrophoretically transferred to nitrocellulose membranes. Membranes were blocked in 1X Tris-buffered saline (TBS) with 0.05% Tween-20 and 5% non-fat skim milk powder (115363; Merck, Kenilworth, NJ, USA) for 1 h at room temperature and then immunostained overnight at 4°C with primary monoclonal rabbit antibodies: AKT (4691, 1:1,000 dilution; Cell Signaling Technology, Danvers, MA, USA) or GAPDH (2118, 1:10,000 dilution; Cell Signaling Technology). The following day membranes were washed and incubated with secondary goat anti-rabbit IgG HRP-linked antibody (7074, 1:1,000 dilution; Cell Signaling Technology) and anti-biotin HRP-linked antibody (7075, 1:1,000 dilution; Cell Signaling Technology). Bands were visualized using WesternBright Quantum (K-12042-D10; Advansta, Inc., Menlo Park, CA, USA) on the Fusion FX7 system (Vilber Lourmat GmbH, Eberhardzell, Germany) and quantified using the ImageJ Analyst software (National Institutes of Health, Bethesda, MD, USA), with band intensities normalized to GAPDH.

FITC Annexin V apoptosis detection kit (556547; BD Biosciences, San Jose, CA, USA) was used to detect cell death 72 h post-transfection, as per the manufacturer's protocol. Signals were detected from 1.0×10⁴ cell population using the BD FACSCanto™ II flow cytometer (BD Biosciences) and examined on the BD FACSDiva™ software (BD Biosciences).

Caspase-Glo 3/7 assay kit (G8090; Promega) was utilized to analyze caspase-3 and -7 activity, 48 h post-transfection as per the manufacturer's protocol. Samples were incubated at 25°C for 1 h in the dark and luminescence was then detected using the GloMax Multi Luminescence Multimode Reader.

Flow cytometry was used to analyze cell cycle using the BD Cycletest™ Plus DNA kit assay (340242; BD Biosciences) 48 h post-transfection, as per the manufacturer's protocol. Signals were detected from 1.0×10⁴ cell population using the BD FACSCanto™ II flow cytometer and examined on the BD FACSDiva™ software. Results were then analyzed using the ModFit LT v3.2.1 (Verity Software House, Inc., Topsham, ME, USA) and the percentage of the cells in G0/G1, S and G2/M phase were counted and compared.

Experiments involving zebrafish were approved by the University of Malaya, Faculty of Medicine, Institutional Care of Use Committee (FOM IACUC) (Ethics reference number: 2015-181006/IBS/R/NO) and complied with all relevant animal welfare laws, guidelines and policies. Wild-type Danio rerio zebrafish embryos were cared for and maintained using standard husbandry practices.

Zebrafish embryos were injected with A549 cells transfected with 80.0 nm miR-608 mimics, inhibitors, or their corresponding negative controls at the superficial location of the yolk near to the perivitelline space of the embryos using a FemtoJet Microinjector (Eppendorf, Hamburg, Germany) and InjectMan NI 2 Micromanipulator (Eppendorf) with constant injection pressure and injection time. The injection volume and cell suspension was calibrated to be ~100–200 cells/injection in each embryo. After transplantation, embryos were immediately placed at 37°C overnight.

Embryos were fixed in 4% paraformaldehyde at 4°C overnight followed by 2-h dehydration in methanol at −20°C. Following rehydration, embryos were washed with 1% dimethyl sulfoxide (DMSO), 0.1% Triton in PBS (1X PDT) and blocked with 10% FBS, 2% BSA in PBST (blocking buffer) for 1 h at room temperature. Embryos were then stained with purified rabbit anti-active caspase-3 antibody (559565, 1:500 dilution; BD Biosciences) for 2 h at room temperature followed by washes in PDT. Again embryos were incubated with blocking buffer, and then stained with anti-rabbit IgG Fab2 Alexa Fluor 647 Conjugate (4414, 1:500 dilution; Cell Signaling Technology) overnight at 4°C. The following day, embryos were washed with PDT before visualization and imaging using the Leica confocal laser-scanning microscope SPII and Leica Application Suite (LAS) software v5.0 (Leica Microsystems, Wetzlar, Germany). Fluorescence was quantified using ImageJ Analyst software. Threshold was set to eliminate background fluorescence and embryos were analyzed to generate arbitrary fluorescence units.

Silencing of the AKT2 gene was performed using a set of three unique 27 mer siRNA duplexes at a final concentration of 10.0 nM (siRNA A: GCAUCAUA AAUUGGUAGUUUCCUGC, siRNA B: AGCGUGUGAAUA CAUCAAGACCTG, siRNA C: ACAGCAAAGCAGGAG UAUAAGAAAG) (SR300144; Origene Technologies, Inc., Rockville, MD, USA). A universal scrambled negative control siRNA (siRNA NC) was used as a control. At 48 h post-transfection, silencing efficiency was assessed by western blot analysis. Amongst the three siRNAs utilized, the siRNA with the greatest silencing efficiency was selected for further downstream work and referred to as siAKT2. The effects of miR-608 mediated apoptosis via AKT2 was validated through transfection of 80.0 nM miR-608 inhibitors, followed by transfection with 10.0 nM siAKT2 directed against the human AKT2 gene (siAKT2) 6 h later. AKT2 protein levels were determined via western blot analysis 48 h post-transfection while apoptosis was detected using the FITC Annexin V apoptosis detection kit and Caspase-Glo 3/7 assay kit.

All in vitro experiments were performed in triplicate independent experiments. In vivo experiments were performed with sample size of 15 zebrafish embryos per treatment group. All data were presented as mean ± standard deviation (25). Paired Student's t-test was used to determine the statistical significance of results, whereby a P-value of ≤0.05 was considered significant.

A previous study conducted by our laboratory determined that the expression of miR-608 was significantly downregulated following the silencing of BCL-XL in lung adenocarcinoma cell line A549. Results also indicated that miR-608 played a tumour suppressor role in regulating the apoptotic properties of A549 and a secondary lung adenocarcinoma cell line SK-LU-1 (22). To determine the molecular mechanism by which miR-608 regulates the apoptotic properties in NSCLC cell lines, we performed an in silico bioinformatics analysis to identify the putative miR-608 gene targets through the use of the TargetScan Human v5.2 algorithm, followed by functional annotation using the web tool DAVID v6.7, which lists the predicted targets of miR-608 according to their apoptosis-related pathways. miR-608 was found to be associated with various signaling pathways involved in cancer, including the phosphatidylinositol 3-kinase/protein kinase B (PI3K/AKT), wingless-type MMTV integration site family (WNT), transforming growth factor (TGF-β), mitogen activated protein kinase (MAPK) and the intrinsic and extrinsic pathway (Fig. 1).

The 3′UTR of V-Akt Murine Thymoma Viral Oncogene Homolog 2 (AKT2) contains two miR-608 binding sites, and is involved with apoptosis and proliferation and was thus chosen for further validation. To verify whether AKT2 3′UTR was a direct target of miR-608, the wild-type and mutated AKT2 3′UTR were cloned into the pmirGLO Dual-luciferase miRNA target expression vector (Fig. 2A). Luciferase reporter assay confirmed that miR-608 mimics had a significant inhibitory effect on wild-type 3′UTR but not on the mutant 3′UTR of AKT2 luciferase activity, while mimic NC had no effect on either the wild-type or mutant luciferase activity (Fig. 2B). This result suggests that miR-608 directly binds to the binding sequence of AKT2 3′UTR, and this was further verified by a decrease in AKT2 protein levels in response to miR-608 mimic transfection, as analyzed by western blot analysis. Conversely, the expression of AKT2 was significantly increased when miR-608 was inhibited (Fig. 2C).

We have previously demonstrated that miR-608 plays an important role in the regulation of apoptosis, and presently identified miR-608 as a direct regulator of AKT2. It was thus hypothesized that low expression of miR-608 in NSCLC may result in suppression of its inhibitory effects towards AKT2 causing AKT2 expression to be upregulated, which in turn blocks apoptosis. To investigate this hypothesis, co-transfection of miR-608 inhibitors and siRNA inhibiting AKT2 was performed in A549 and SK-LU-1 cells. siRNAs were provided as a set of three siRNA duplexes; therefore to evaluate the silencing efficiency of the siRNAs, densitometry analysis of western blot bands was performed to evaluate the AKT2 protein expression in siRNA-transfected in comparison to siRNA NC transfected cells. Amongst the three siRNAs utilized, siRNA C was able to significantly decrease AKT2 protein levels in A549 and SK-LU-1 cells (Fig. 3A). As siRNA C (referred to as siAKT2 henceforth) had the greatest silencing efficiency amongst the three siRNAs, it was selected for further downstream work. Results indicated that silencing of AKT2 was able to partially rescue the inhibition of apoptosis and caspase-3/7 activation that was induced by miR-608 inhibitors (Fig. 3B and C). Collectively, these results demonstrate the tumour suppressor role of miR-608 in NSCLC is at least partially through its inhibition of AKT2.

The in vivo effect of miR-608 on apoptosis was determined through utilization of zebrafish embryos as an animal model. miR-608 mimic, inhibitor or negative control transfected A549 cells were microinjected into zebrafish. Embryos were then visualized using a Leica confocal microscope following immunostaining with anti-active caspase-3 monoclonal antibodies (Fig. 4A). Results of fluorescent image analysis using the ImageJ software indicated that detection of caspase-3 was significantly increased in zebrafish injected with miR-608 in comparison to negative control injected zebrafish (Fig. 4B). This suggests that miR-608 is able to induce apoptosis in vivo through caspase activation.