The existing GEO chip data from the GEO database was first analyzed using the ‘GBM miRNA expression’ as a search condition. The miRNA expression profile dataset GSE103229 and GSE13140 (17) were downloaded from the Gene Expression Omnibus database (GEO; http://www.ncbi.nlm.nih.gov/geo). GSE103229 was used for personal purpose while GSE13140 has reported the miRNA expressions in the mice with or without IL-4 stimulation. Expression profiling was analyzed using the Exiqon human V3 microRNA PCR panel I+II platform (Qiagen, Inc.) according to the manufacturer's protocol. GEO2R analysis was performed on the website provided directly by the GEO project (https://www.ncbi.nlm.nih.gov/geo/geo2r/).

To predict miR-138-5p target mRNA, TargetScan version 7.2 (http://www.targetscan.org/vert_72) was used with ‘miR-138-5p’ as the key word to identify the candidate genes.

A total of 20 GBM tissues and adjacent normal tissues were collected from patients who underwent surgical resection between March 2018 and March 2019 at The First Affiliated Hospital of Zhejiang Chinese Medical University (Hangzhou, China). The mean age was 61±13 years old (range, 27–83 years; 16 males, 4 females) The key inclusion criteria included patients with newly diagnosed GMB (histologically confirmed by biopsy or resection) within 4 weeks of diagnosis and >18 years old. The present study excluded patients unwilling to abide by the protocol as well as being legally incapacitated. Patients who had either received radiotherapy or chemotherapy prior to the surgical resection procedure were also excluded. The present study was approved by the Ethics Committee of The First Affiliated Hospital of Zhejiang Chinese Medical University and all human tissue samples were obtained following written informed consent. Fresh tissues were immediately frozen in liquid nitrogen and stored at −80°C before RNA extraction.

Three human GBM cell lines, including U87, U251 and U373 were purchased from the American Type Culture Collection (ATCC), and the normal brain glial cell line HEB was purchased from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences. The U373 and U87 (ATCC^® HTB-14™) cells lines were authenticated by STR profiling and identified as GBM cells of unknown origin. All cell lines were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc.), supplemented with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.), 100 µg/ml streptomycin and 100 U/ml penicillin. The cells were incubated at 37°C in a humid incubator containing 5% CO₂.

In total, 1×10⁵ of U87 and U251 cells were seeded into 6-well plates and incubated at 37°C in a humid incubator overnight until 50–60% confluence was reached. Then, the cells were transiently transfected with 20 nM small interfering RNA (siRNA/si) (Shanghai GenePharma Co., Ltd.) or 50 nM miRNA mimics using Lipofectamine^® 3000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. The cells were collected for subsequent experimentation following 24 h of transfection at 37°C. The following siRNA sequences were used: Si-CCND3 sense, 5′-GGAUCUUUGUGGCCAAGGATT-3′ and antisense, 5′-UCCUUGGCCACAAAGAUCCTT-3′; and si-negative control (NC) sense 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense, 5′-ACGUGACACGUUCGGAGAATT-3′. Synthetic miRNA mimics targeting the sequence of miR-138-5p (5′-AGCUGGUGUUGUGAAUCAGGCCG-3′) and miR-NC (5′-ACUCUAUCUGCACGCUGACUU-3′) were produced by Invitrogen; Thermo Fisher Scientific, Inc.

Total RNA from GBM cells and tissues was extracted using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Total RNA was reverse transcribed into cDNA using High-Capacity cDNA Reverse Transcription kit with RNase Inhibitor (Fermentas; Thermo Fisher Scientific, Inc.) and Maxima SYBR green/ROX qPCR Master Mix was used (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturers' protocols. qPCR was subsequently performed using a QuantStudio 6 Flex Real-Time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The following thermocycling conditions were used for the qPCR: Initial denaturation at 95°C for 10 min; followed by 40 cycles at 92°C for 15 sec and 60°C for 1 min. The following primer pairs were used for the qPCR: GAPDH forward, 5′-ATTCCATGGCACCGTCAAGGCTGA-3′ and reverse, 5′-TTCTCCATGGTGGTGAAGACGCCA-3′; miR-138-5p forward, 5′-TGCAATGGGTTTGGCGTAGAAC-3′ and reverse, 5′-CCAGTGCCGCAGGGTAGGT-3′; CCND3 forward, 5′-TACCCGCCATCCATGATCG and reverse, 5′-AGGCAGTCCACTTCAGTGC; U6 forward, 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′. U6 was used as the loading control for miRNA expression levels, while GAPDH was used as the loading control for mRNA expression levels. Relative quantification of gene expressions were calculated with the 2^−ΔΔCq method (18).

A Cell Counting Kit-8 (CCK-8) assay (Beyotime Institute of Biotechnology) was used to investigate the cell viability according to the manufacturer's protocol. Briefly, following transfection, 5×10³ cells/well were seeded in 96-wells plates and cultured at 37°C for 0, 24, 48, 72 and 96 h. Subsequently, 10 µl CCK-8 reagent was added/well and further incubated at 37°C for 2 h in the dark. The optical density value was measured using a microplate reader at a wavelength of 450 nm to determine the cell viability.

Following transfection, 1×10³ U87 and U251 cells/well were plated into 6-well plates and incubated at 37°C for 14 days. A colony was defined as a clump of cells that could be clearly distinguished from another clone after staining. Then, the formed cell colonies were fixed with 10% formaldehyde for 20 min at room temperature and stained at 37°C with 0.1% Coomassie Brilliant Blue R250 for 2 min. Cell colonies were visualized using a camera and counted.

The CCND3 3′-UTR wild-type (wt) sequence was amplified using PCR and cloned into the restrictive site between XhoI and Bg1II in the firefly luciferase reporter vector repGL3 (Promega Corporation). DNA sequencing was performed to verify the cDNA sequence. The following primers were used for the cDNA amplification: CCND3-3′-UTR-wt-upstream, 5′-CCCTGGAGAGGCCCTCTGGA-3′ and CCND3-3′UTR-wt-downstream, 5′-TTCCAAGAAGCCAAAGCCA-3′. Subsequently, the QuickMutation™ Plus Site Mutation kit (Beyotime Institute of Biotechnology) was used to mutate the miR-138-5p putative binding site in the 3′-UTR-containing vector.

For the dual luciferase reporter gene assay, 5×10³ cells/well were seeded into 96-wells plates and transfected with wt or mutated (mut) reporter vector or empty plasmid (p-con) for 48 h at 37°C using Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.). The concentration of plasmids used above was 1 µg/µl. The firefly luciferase activity was measured using a dual luciferase reporter assay system (Beyotime Institute of Biotechnology). The mean of the results from cells co-transfected with p-con and miR-138-5p NC was set as 100 and the frefly luciferase activity was calculated as mean ± standard deviation following normalization to Renilla luciferase activity.

Flow cytometric analysis was performed using a Cell Cycle Analysis kit (Beyotime Institute of Biotechnology). Briefly, 10×10⁵ cells were washed with PBS twice, collected by trypsinization and fixed with 70% ethanol overnight at 4°C. Subsequently, 100 µl propidium iodide staining buffer was added to the cells following the resuspension in ice-cold PBS containing 50 µg/ml of RNase. After incubation for 30 min at 37°C in the dark, the cell cycle distribution was analyzed using a BD FACSLyric™ flow cytometer (BD Biosciences) with a laser beam at 488 nm. The date were analyzed by FlowJo version 7.6 Software (BD Biosciences).

Total protein was extracted from cells using RIPA lysis buffer (Beyotime Institute of Biotechnology). Total protein was quantified using a bicinchoninic acid assay kit (Beyotime Institute of Biotechnology) and 20 µg protein/lane was separated via 10% SDS-PAGE. The separated proteins were subsequently transferred onto polyvinylidene fluoride membranes (MilliporeSigma) and blocked with 5% non-fat milk in TBS-Tween-20 (TBST; 0.1% Tween-20) at room temperature for 1 h. The membranes were then incubated at 4°C overnight with the following primary antibodies: Anti-CCND3 (1:1,000; cat. no. ab28283; Abcam) and anti-GAPDH (1:1,000; cat. no. ab181602; Abcam). Following the primary antibody incubation, the membranes were rinsed five times with TBST and incubated with the secondary antibodies: Horseradish, peroxidase-conjugated goat anti-rabbit secondary antibody (1:10,000; cat. no. ab205718; Abcam) and horseradish peroxidase-conjugated goat anti-mouse secondary antibody (1:10,000; cat. no. ab205719; Abcam) for 1 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (Thermo Fisher Scientific, Inc.) reagent and densitometric analysis was performed using ImageJ version 1.50d software (National Institutes of Health).

Transwell insert chambers with an 8-µm pore size membrane (Costar; Corning, Inc.) as well as chambers coated with Matrigel (BD Biosciences) were used for Transwell assay. A total of 3×10⁴ cells were seeded in the upper chamber, maintaining in the medium without serum. Medium containing 20% FBS (Gibco; Thermo Fisher Scientific, Inc.) was added to the lower chamber as chemoattractant. Subsequently, the cells were incubated at 37°C for 24 h. Non-invading cells were removed using cotton swabs while the cells migrated to the bottom of the membrane were fixed with cold methanol and stained with 0.1% crystal violet at 37°C for 20 min. The stained cells were captured under a light microscope in five random fields with 100× magnification, and the average number of migratory cells was calculated. Data analysis was performed using GraphPad Prism 7.0 software (GraphPad Software, Inc.).

All experiments were repeated in triplicate. Statistical analysis was performed using SPSS 19.0 software (IBM Corp.) and data are presented as the mean ± SD. The normality assumption of data distribution was assessed using a Kolmogorov-Smirnov test. For normally distributed data, a paired or unpaired Student's t-test was used for two groups depending on whether the data were paired. A one-way ANOVA followed by a Tukey's post hoc test was used for multiple group comparisons. The Kruskal-Wallis test followed by a Dunn's post hoc test was used for non-parametric statistical analysis. As the data analyzed was non-parametric, correlation analysis was performed using Spearman's correlation coefficient. P<0.05 was considered to indicate a statistically significant difference.

The existing GEO chip data from the GEO database was first analyzed using the ‘GBM miRNA expression’ as a search condition. Two databases were identified: GSE103229 and GSE13140. After analyzing the data in the two data sets, the expression level of miRNA in glioma and the expression in normal tissues adjacent to the tumor were compared, and a data set with lower miRNA expression level was obtained. Finally, we took the intersection of the results obtained in the two databases, and obtained four miRNAs, namely miR-124-3p, miR-129-5p, miR-138-5p and miR-338-3p. As the role of miR-124-3p, miR-129-5p and miR-338-3p in GBM has been previously reported, the role of miR-138-5p in GBM was further investigated in the present study. miR-138-5p expression levels in GBM tissues were discovered to be notably downregulated compared with the normal tissues (Fig. 1A). To validate this result, RT-qPCR was performed to analyze the miR-138-5p expression levels in 20 paired GBM tissues and adjacent normal tissues collected from The First Affiliated Hospital of Zhejiang Chinese Medical University. Similar to the results obtained from the GEO datasets, miR-138-5p expression levels were significantly downregulated in the GBM tissues compared with the normal tissues (Fig. 1B). In addition, miR-138-5p expression levels were investigated in different GBM cell lines, including U87, U251 and U373. The results illustrated that miR-138-5p expression levels were also significantly downregulated in the GBM cell lines compared with the normal brain glial cell line, HEB (Fig. 1C). These results suggested that miR-138-5p expression levels may be significantly downregulated in GBM tissues and cells.

To investigate the biological function of miR-138-5p in GBM, miR-138-5p mimics and a miR-NC were transfected into U87 and U251 cells. The transfection efficiency was evaluated using RT-qPCR; the miR-138-5p mimic-transfected cells had significantly upregulated expression levels of miR-138-5p compared with the miR-NC-transfected cells (Fig. 2A).

Following the successful transfection, a CCK-8 assay was used to evaluate the effect of miR-138-5p upregulation on cell viability. The results identified that the cell viability of U87 and U251 cells transfected with the miR-138-5p mimic was significantly decreased compared with the miR-NC-transfected cells following 48–96 h of incubation (Fig. 2B). The colony formation assay revealed that the number of colonies formed in cells transfected with the miR-138-5p mimics was significantly decreased compared with the miR-NC group in both cell lines (Fig. 2C), suggesting a suppressive effect of miR-138-5p on GBM colony formation. The results from the Transwell assay also revealed that the invasive ability of the miR-138-5p mimic group was significantly decreased compared with the miR-NC group in both cell lines (Fig. S1). These data indicated that miR-138-5p may serve as a tumor suppressor and inhibit the viability, colony formation and invasion of GBM cells.

It is well known that CCND3 is one of the highly conserved cyclin family members that serves a critical role in the cell cycle (19,20). As a result, flow cytometric analysis was performed to determine whether miR-138-5p impacted the GBM cell cycle. The results revealed that compared with the miR-NC, a significant increase was observed in the number of cells arrested at the G₀/G₁ phase in the cells transfected with the miR-138-5p mimics, while significantly fewer cells were located in the S phase (Fig. 2D).

It is well established that miRNAs bind to the 3′-UTR of target mRNAs to suppress their expression in order to participate in various physiological activities (21–23). To identify miR-138-5p target mRNA, the TargetScan database was used to predict the potential target genes, which revealed CCND3 as a potential miR-138-5p target (Fig. 3A). For further validation, RT-qPCR was performed to investigate CCND3 expression levels in GBM tissues. The results discovered that CCND3 expression levels were significantly upregulated in GBM tissues compared with the adjacent normal tissues (Fig. 3B), suggesting a negative correlation between CCND3 and miR-138-5p expression levels, which was subsequently confirmed by correlation analysis (Fig. 3C). As expected, RT-qPCR and western blotting analysis demonstrated that CCND3 expression levels were also markedly upregulated in various GBM cell lines compared with the normal brain glial HEB cells (Fig. 3D). To further confirm that CCND3 was a target gene of miR-138-5p in GBM, miR-138-5p mimics were used to overexpress miR-138-5p and then CCND3 mRNA and protein expression levels were analyzed in U87 and U251 cells. The results revealed that CCND3 mRNA and protein expression levels were both markedly downregulated in the cells transfected with the miR-138-5p mimics compared with the miR-NC-transfected cells (Fig. 3E).

A dual luciferase reporter gene assay was performed to validate the direct interaction between miR-138-5p and CCND3. The supposed 3′-UTR binding site of CCND3 was cloned into the luciferase reporter vector as a wt version [plasmid (p)-CCND3-wt], while a mut type fragment of the 3′-UTR was also constructed as a mutant version (p-CCND3-mut). The cells were co-transfected with miR-138-5p mimics and luciferase reporter plasmids containing the wt or mut type of CCND3 3′-UTR. The results identified a significantly reduced luciferase activity in the U87 and U251 cells co-transfected with the p-CCND3-wt and miR-138-5p mimic compared with the p-CCND3-mut (Fig. 4A). Taken together, these results suggested that miR-138-5p may target CCND3 in GBM cells.

To determine whether CCND3 was responsible for the tumor suppressive role of miR-138-5p in GBM cells, CCND3 was knocked down in U87 and U251 cells using siRNA. The successful transfection efficiency was confirmed using RT-qPCR and western blotting (Fig. 4B and C). Subsequently, colony formation and CCK-8 assays were performed, which identified that the colony forming ability and cell viability were significantly decreased when U87 and U251 cells were transfected with si-CCND3 compared with si-NC (Fig. 4D and E). These data indicated that CCND3-knockdown inhibits the colony formation ability and viability of GBM cells.

Taken together, the present study revealed that CCND3 expression levels were negatively regulated by miR-138-5p in GBM, which indicated that decreased miR-138-5p expression levels may promote CCND3 and the consequent progression from the G₀/G₁ phase to the S phase, resulting in increased cell proliferation and colony formation in GBM (Fig. 5).