The H1299 human NSCLC cell line was obtained from American Type Culture Collection (Manassas, VA, USA; CRL-5803™) and cultured in RPMI medium (Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% v/v fetal bovine serum and 1% v/v antibiotic solution (growth medium). The culture medium used for the other cell lines has been described elsewhere (9). The cells were incubated at 37°C with 5% CO₂. The other cell lines used for sequencing are listed in Table I.

Total RNA was extracted from the lung cancer cells using the miRNeasy Mini kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's protocol. An mRNA library was prepared using the TruSeq RNA Sample Prep kit v2 (Illumina, Inc., San Diego, CA, USA). A small RNA library was prepared using TruSeq Small RNA Sample Prep kit (Illumina, Inc.). mRNA and small RNA sequence data were acquired using the HiSeq2500 system (Illumina, Inc.). All raw sequence data were deposited in the DNA Data Bank of Japan under the accession numbers DRA001846 (RNA-Seq) (9) and DRA003587 (small RNA-Seq).

An miRNA duplex corresponding to miR26a, pcDNA3-miR26a2, was purchased from Addgene, Inc. (Cambridge, MA, USA). The pcDNA3.1 control vector used for transfection was purchased from Invitrogen; Thermo Fisher Scientific, Inc. HMGA1 small interfering RNA (siRNA) was synthesized by Sigma-Aldrich; Merck Millipore (Darmstadt, Germany). AllStars Negative Control siRNA was purchased from Qiagen GmbH and used as a control.

The cells (2.5×105 cells per well) were plated in 2 ml of growth medium in 6-well plates 1 day prior to transfection. Oligonucleotide transfection was performed using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). Briefly, 7.5 µl of Lipofectamine 3000 was diluted in 125 µl of Opti-MEM1 (Gibco; Thermo Fisher Scientific, Inc.) and mixed with 125 µl of DNA diluted in P3000 reagent with Opti-MEM1 medium. The mixed solutions were maintained at room temperature for 5 min and subsequently added to each well. The 6-well plate was then incubated at 37°C with 5% CO₂ prior to use in further experiments.

At 3 days post-transfection, the cells were harvested and lysed in 500 µl RIPA buffer (Wako Pure Chemical Industries, Ltd., Osaka, Japan) supplemented with 10% Complete Mini solution (Roche Diagnostics, Basel, Switzerland). The protein concentration was determined using a Pierce BCA Protein Assay kit (Thermo Fisher Scientific, Inc.). Lysate aliquots containing equal quantities of protein (1 µg per lane) were separated by 5–20% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then electrophoretically transferred onto Immun-Blot PVDF membranes (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The membranes were blocked in a solution containing 5% milk protein and 0.1% Tween 20 in Tris-buffered saline (TBS), and then incubated overnight with the appropriate primary antibody at 4°C. The blots were then washed and incubated with an appropriate horseradish peroxidase-linked species-specific whole secondary antibody diluted 1:100,000 in TBS with 0.1% Tween 20 and 5% milk protein for 1 h at room temperature. Secondary antibody binding was visualized using the Las-3000 mini system (FujiFilm, Tokyo, Japan) and an ECL Prime Western Blotting detection system (GE Healthcare Life Sciences, Chalfont, UK). The following primary antibodies were purchased from Abcam (Cambridge, UK) and used for western blot analysis: Anti-tubulin (cat. no. ab4074; 1:1,000 dilution) and anti-HMGA1 (cat. no. ab129153; 1:10,000). The anti-rabbit horseradish peroxidase-linked secondary antibody (cat. no. NA934-100UL) was purchased from GE Healthcare Life Sciences.

Total RNA was extracted from the cells with the miRNeasy Mini kit according to the manufacturer's protocol. Primers were synthesized by Exigen (Tokyo, Japan). The miRNA expression levels were detected via SYBR Green-based real-time PCR with the mir-X miRNA qRT-PCR SYBR kit (Clontech Laboratories, Inc., Palo Alto, CA, USA) according to the manufacturer's protocol using 200–400 ng cDNA and 0.5 µl 100 nM primers. The mRNA expression levels were detected with SYBR SELECT Master mix (Applied Biosystems, Inc.; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol using 400–800 ng cDNA and 2 µl 100 nM primers. The following specific primers were used for RT-qPCR: miR-26a, forward 5′-TTCAAGTAATCCAGGATAGGCT-3′ and reverse mRQ 3′primer (a component of the mir-X miRNA qRT-PCR SYBR kit); HMGA1 pair A, forward 5′-GGCCCAAATCGACCATAAAGG−3′ and reverse 5′-GGACAAATCATGGCTACCCCT−3′; HMGA1 pair B, forward 5′-CAGCGAAGTGCCAACACCTAA−3′ and reverse 5′-GTTTTTGCTTCCCTTTGGTCG−3′; and U6 forward and reverse primers (provided in the mir-X miRNA qRT-PCR SYBR kit). The qPCR analysis was performed under the following conditions: 2 min at 50°C, 10 min at 95°C, and 40 cycles at 95°C for 15 sec and 60°C for 1 min on a 7900HT Fast Real Time PCR system (Applied Biosystems, Inc.; Thermo Fisher Scientific, Inc.). The relative expression levels of miR-26a and HMGA1 were calculated and quantified using the ΔΔCq method (12) following normalization to the expression levels of U6 small nuclear RNA.

The CytoSelect 96-well cell migration assay (8 µm; fluorometric format; Cell Biolabs, Inc., San Diego, CA, USA) and CytoSelect 96-well cell invasion assay basement membrane (fluorometric format; Cell Biolabs, Inc.) were used to analyze cell migration and invasion activities, respectively. Equal numbers of cells (1.0×10⁵ cells per chamber), which had been transfected with RNA duplex or plasmid for 48 h, were seeded into the upper chambers containing serum-free medium. Medium with 10% serum was added to the lower chambers as a chemoattractant. After 12 h incubation at 37°C, 4X lysis buffer/CyQuant GR dye solution (Cell Biolabs, Inc.) was added to the lower chambers. The numbers of migrated and invaded cells were counted via fluorescence detection at 480/520 nm, using a FLUOstar OPTIMA microplate reader (BMG Labtech, GmbH, Ortenberg, Germany).

The cells were counted using a Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan). In brief, 10 µl of CCK-8 solution was added to the cells in each well of a 96-well plate, followed by incubation for 2 h at 37°C in 5% CO₂. The solution absorbance was measured at 450 nm in an iMark microplate absorbance reader (Bio-Rad Laboratories, Inc.).

To construct the luciferase reporter vectors, a 3′untranslated region (UTR) fragment of wild-type (WT) human HMGA1, which contained miR-26a binding sites, was amplified from genomic DNA using the following primers: Forward 5′-TCGAAAGCTTCCCAAATCGACCATAAAGGGTG-3′ and reverse 5′-TCGAACTAGTTCCAGAAAAGGATATTTTTTTTATTCAAG−3′. The amplification was performed with Phusion High-Fidelity DNA Polymerase (New England BioLabs, Inc., Ipswich, MA, USA) under the following conditions: 30 sec at 98°C; and 30 cycles at 98°C for 10 sec, 65°C for 30 sec and 72°C for 90 sec, followed by 72°C for 10 min; on a GeneAmp PCR System 9700 (Applied Biosystems; Thermo Fisher Scientific, Inc.). The amplified fragment was inserted into the pMIR-REPORT miRNA expression reporter vector (Applied Biosystems, Inc., Thermo Fisher Scientific, Inc.) at the HindIII/SpeI sites. Mutations were generated in the miR-26a binding sites using the KOD-Plus-Mutagenesis kit (Toyobo Co., Ltd., Osaka, Japan) and the following primers: Forward 5′-ATGAACTATAAAAAAAATATCCTTTTCTGGAA-3′ and reverse 5′-CTGCAAATAGGAAACCAGAGAGGG−3′. The sequences of the resulting plasmids containing the WT or mutated (Mut) 3′UTR of HMGA1 were verified by DNA sequencing. The Renilla luciferase-encoding plasmid, pRL-TK (Promega Corporation, Madison, WI, USA) was used as an internal control. The H1299 cells (1.2×10³ cells per well) were plated in 100 µl of growth medium in 96-well plates one day prior to transfection. At room temperature, the H1299 cells were co-transfected with either the miR-26a expression vector or the control vector along with the pRL-TK vector and pMIR-REPORT miRNA, the expression reporter vector containing the WT or Mut 3′UTR of HMGA1. After 48 h, the cells were harvested and subjected to luciferase analysis using the Dual-Luciferase reporter assay system (Promega Corporation) in a Centro LB 960 (Berthold Technologies, Bad Wildbad, Germany).

The sequence data were mapped to the human reference genome, hg19, in the UCSC genome browser (https://genome.ucsc.edu) using the Efficient Large-Scale Alignment of Nucleotide Databases tool (Illumina, Inc.), as described in a previous study (9). Using perl script, the parts per million (ppm) and reads per kilobase per million were calculated as miRNA and mRNA expression levels, respectively. Correlation coefficients were calculated using Pearson product-moment correlation. A heat map was generated using heatmap.2 in the gplots of R.

The present study assessed small RNAs in 26 human lung adenocarcinoma cell lines (Table I) using next-generation sequencing. A total of 1,341 miRNAs were selected from these data, and their expression levels were calculated in ppm. To determine the associations between miRNA and mRNA expression throughout the genome, the correlation coefficient was calculated between each miRNA and individual mRNAs using mRNA data from a previous study (9). A heat map of these correlation coefficients is shown in Fig. 1A. As shown in the histogram in Fig. 1A, a higher number of mRNAs were negatively correlated with miRNAs, compared with the number of mRNAs positively correlated with miRNAs. This suggested the predominantly repressive effects of miRNAs on mRNAs in these cell lines. Among the miRNAs, hsa-miR-26a-1, the major precursor RNA of miR-26a, exhibited the most marked negative correlation with overall mRNA expression (the 10 miRNAs most negatively correlated with overall mRNA expression are shown in Table II), suggesting that miR-26a may have the most marked effect within these adenocarcinoma cell lines. Subsequently, the present study aimed to identify the mRNA correlated most significantly with hsa-miR-26a-1. It was found that the HMGA1 mRNA exhibited the most marked negative correlation with hsa-miR-26a-1, suggesting that it was affected most by miR-26a. HMGA1 is a master chromatin structural regulator, which primarily mediates its gene regulatory activity by interacting directly with A/T-rich DNA sequences located in promoter and enhancer regions (13). HMGA1 also exhibited a negative correlation with another miR-26a precursor, hsa-miR-26a-2 (Fig. 1B and Table III).

To determine the functionality of the negative correlation between miR-26a and HMGA1, miR-26a was overexpressed in the H1299 NSCLC cell line, which has low and high endogenous levels of miR-26a and HMGA1, respectively. The expression level of miR-26a was increased >500-fold following transfection with the miR-26a expression vector (Fig. 2A). This led to reduced protein levels of HMGA1, as shown in Fig. 2B. The mRNA level of HMGA1 also decreased in response to the overexpression of miR-26a (Fig. 2C). Thus, miR-26a appeared to suppress the protein expression of HMGA1 by reducing the mRNA level of HMGA1.

To examine whether HMGA1 is a direct target of miR-26a, the present study performed a luciferase reporter assay using a plasmid construct containing the 3′UTR of HMGA1 mRNA. The 3′UTR contained putative miR-26a binding sites, and a reporter plasmid containing mutated miR-26a binding sites was used as a control (Fig. 3A). These reporter constructs, containing either miR-26a or a control expression vector, were co-transfected into H1299 cells. The luciferase activity of the reporter harboring the HMGA1 WT 3′UTR was reduced by 50% following co-transfection with the miR-26a expression vector. By contrast, the activity of the Mut construct was not altered by co-transfection with miR-26a. These results suggested that HMGA1 mRNA is a direct target of miR-26a (Fig. 3B).

To investigate the role of miR-26a in the capacities of cells to metastasize and invade, miR-26a was overexpressed in H1299 cells and the resulting cellular phenotype was evaluated. In the cells overexpressing miR-26a, 60% reductions were observed in the migration and invasion capacities, compared with the control (Fig. 4A). These data suggested that miR-26a suppressed migration and invasiveness in H1299 cells. Furthermore, reduced cell growth was observed in response to the overexpression of miR-26a, compared with the control, which suggested that miR-26a repressed the proliferation of the H1299 cells (Fig. 4B).

To identify cellular phenotypes, which varied with altered expression of HMGA1, cell invasion and migration activity assays were performed. HMGA1 was knocked down using siRNA and the protein expression of HMGA1 was successfully suppressed by 80% (Fig. 5A). Upon HMGA1 knockdown, a 60% decrease in migration capacity and 50% decrease in invasiveness were observed, compared with the control (Fig. 5B). These data indicated that a high expression level of HMGA1 enhanced cell migration and invasiveness in H1299 cells. Furthermore, a reduction in the rate of cell growth was observed following HMGA1 knockdown, suggesting a positive role of HMGA1 in the promotion of H1299 cell growth (Fig. 5C).