All 43 human glioma tissues (age range, 26–71 years; mean age, 49 years), as well as their paired adjacent non-cancerous tissues, were obtained from patients who underwent surgical resection at The First Affiliated Hospital of Nanjing Medical University (Nanjing, China) between January 2013 and December 2016. The distance between the tumor and the matched normal adjacent tissue was ~2 cm. The inclusion criteria were as follows: i) Diagnosed with glioma; and ii) had not received pre-operative radiotherapy, chemotherapy or other adjuvant treatments before surgery. The exclusion criteria were as follows: i) Diagnosed with other diseases; and ii) failed to cooperate with researchers. Histological grade was classified by pathologists using the World Health Organization criteria (28). All tissue samples were collected during surgery, frozen immediately in liquid nitrogen, and stored at −80°C for total RNA or protein extraction. The clinicopathological characteristics of the patients are summarized in Table I. The present study was approved by the Institutional Review Board and Ethics Committee of Nanjing Medical University (Nanjing, China) and Fourth Military Medical University (Xi'an, China), and written informed consent was obtained from all patients.

A total of six human glioblastoma (GBM) cell lines, including U87MG [GBM of unknown origin; American Type Culture Collection HTB-14; short-tandem repeat (STR) profiling was performed], LN229, H4, U251, U118 (derived from the U138MG astrocytoma cell line; American Type Culture Collection HTB-15; STR profiling was performed) and A172, were obtained from The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences. The human 293T cell line was obtained from American Type Culture Collection. All cells were sustained in DMEM (HyClone; Cytiva) supplemented with 10% FBS (HyClone; Cytiva), 100 U/ml penicillin and 100 ng/ml streptomycin. Normal human astrocytes (NHAs) were purchased from Lonza Group, Ltd. and cultured in the provided astrocyte growth media (Lonza Group, Ltd.) supplemented with 0.1% recombinant human epidermal growth factor, 0.25% insulin, 0.1% ascorbic acid, 0.1% GA-1000, 1% L-glutamine (all Lonza Group, Ltd.) and 5% FBS. All cells were cultured at 37°C in a humidified atmosphere with 5% CO₂.

Total RNA was extracted from glioma tissues or cultured cell lines using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. Single-stranded cDNA was synthesized using the Primerscript RT Master mix (Takara Bio, Inc.) according to the manufacturer's protocol. Subsequently, qPCR analysis was performed with SYBR Premix Ex Taq (Takara Bio, Inc.) according to the manufacturer's protocols. RT-qPCR was performed using an Applied Biosystems 7900 Sequence Detection system (Thermo Fisher Scientific, Inc.). The thermocycling conditions were as follows: 95°C for 3 min, followed by 40 cycles at 95°C for 1 min and 57°C for 45 sec, and 72°C for 2 min. All primers used for RT and RT-qPCR were purchased from Guangzhou RiboBio Co., Ltd. and the sequences were as follows: NEAT1 forward, 5′-TGGCTAGCTCAGGGCTTCAG-3′ and reverse, 5′-TCTCCTTGCCAAGCTTCCTTC-3′; miR-324-5p forward, 5′-CGCGGATCCGGGTGGATGTAAGGGATGAG-3′ and reverse, 5′-CCGGAATTCTTGGGCTGATCCAGGAGAAG-3′; KCTD20 forward, 5′-CGGGATCCATGAATGTTCACCGTGGCAG-3′ and reverse, 5′-CGAATTCCTAATCCTGAAAGTCGTTAGAAGC-3′; GAPDH forward, 5′-GACTCATGACCACAGTCCATGC-3′ and reverse 5′-AGAGGCAGGGATGATGTTCTG-3′; and U6 forward, 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′. The relative expression levels of NEAT1 and KCTD20 were normalized to those of GAPDH, while U6 was used as an internal control for miRNA. The expression levels of NEAT1, miR-324-5p and KCTD20 were calculated using 2^−ΔΔCq analysis (29).

The corresponding negative controls (NCs) and NEAT1 [small interfering RNA (si/siRNA)NEAT1], KCTD20 (siKCTD20), miR-324-5p inhibitor and miR-324-5p mimics were synthesized by Shanghai GenePharma Co., Ltd. The sequences were as follows: si control (Ctrl), 5′-CCCACCAGUUUGAGACUCCACAAAU-3′; siNEAT1, 5′-GGTCTGTGTGGAAGGAGGAAGGCAG-3′; siKCTD20, 5′-GGGAGGAAUAUUCCCAAAUTT-3′; miR-324-5p mimic, 5′-CGCAUCCCCUAGGGCAUUGGUGU-3′; miR-324-5p inhibitor, 5′-ACACCAAUGCCCUAGGGGAUGCG-3′; mimic NC (miR-NC), 5′-UUCUCCGAACGUGUCACGU-3′; and inhibitor NC (Anti-Ctrl), 5′-CAGUACUUUUGUGUAGUACAA-3′. When cells reached a confluence of 80%, cells were transfected with siRNAs (100 nM), mimics (100 nM) or miRNA inhibitors (100 nM) using Lipofectamine^® 2000 transfection reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The cells were then cultured using DMEM containing 10% FBS, 100 U/l penicillin and 100 mg/l streptomycin at 5% CO₂ and 37°C for 48 h. The transfected cells were harvested 48 h post transfection and used for subsequent experiments. To overexpress lncRNA NEAT1 in GBM cells, 3 µg pcDNA3.1-NEAT1 (pcDNA-NEAT1) and 3 µg negative control (empty Vector) (both Shanghai GenePharma Co., Ltd.) were transfected into cells using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The cells were incubated in complete medium supplemented with 10% FBS, 100 U/l penicillin and 100 mg/l streptomycin for 48 h at 37°C with 5% CO₂, and subsequent experiments were performed after 48 h of transfection. For stable transfection, the shRNA lentiviral vectors targeting NEAT1 (shNEAT1, 5′-CATGGACCGTGGTTTGTTACT-3′) and KCTD20 (shKCTD20, 5′-GCTTCCAAAGTGGGAATAAAC-3′) were synthesized by Shanghai GenePharma Co., Ltd. Lentiviruses were produced using a second-generation lentiviral system in 293T cells. 293T cells in the logarithmic growth phase were harvested and re-suspended in antibiotic-free and serum-free medium. Subsequently, cells (5×10⁵ cells/ml) were cultured in DMEM (HyClone; Cytiva) supplemented with 10% FBS (HyClone; Cytiva), 100 U/ml penicillin and 100 ng/ml streptomycin in a 6-well plates at 37°C with 5% CO₂ until the cells reached 80% confluence. Briefly, shNEAT1 or shKCTD20 was inserted into the pLKO.1 vector (BioSettia, Inc.) and transfected into 293T cells (American Type Culture Collection) together with 0.75 µg psPAX2 (BioSettia, Inc.) and 0.25 µg pMD2.G envelope plasmids (BioSettia, Inc.; lentiviral plasmid: Packaging vector: Envelope=4:3:1) using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) to produce an shRNA-containing lentivirus. The media containing lentiviral particles were harvested at 48 h after transfection. Next, cells in the media were removed using a 0.45-µm filter (EMD Millipore) and viral particles were collected. Subsequently, U251 cells were transduced with shNC, shNEAT1 or shKCTD20 lentiviruses (multiplicity of infection, 10) for 24 h at 37°C with 5% CO₂ in the presence of 5 µg/ml polybrene (Shanghai GenePharma Co., Ltd.). After 48 h of lentivirus infection, cells were selected using 5 mg/ml puromycin (Sigma-Aldrich; Merck KGaA) and maintained in puromycin-containing medium (5 mg/ml) for 14 days to establish stably transduced cell lines.

At 48 h after transfection, cells were seeded at 2,000 cells per well in 96-well plates and cultured. First, the cell proliferation rate was detected at the indicated time points (0, 24, 48, 72 and 96 h) using a CCK-8 assay (Dojindo Molecular Technologies, Inc.) according to the manufacturer's protocols. After 1 h of incubation with CCK-8 at 37°C, the absorbance (optical density value) at a wavelength of 450 nm was detected and used to calculate cell proliferation.

The experimental procedures were performed according to the method described in our previous study (30). Briefly, cells were harvested 48 h after transfection and then seeded into a 6-well plate (200 cells/well). Following culture for ~2 weeks until colony formation was observed, visible colonies were fixed with 100% methanol for 20 min at room temperature and stained with 0.1% crystal violet (Sigma-Aldrich; Merck KGaA) for 15 min at room temperature. The colonies were counted using an inverted light microscope (Olympus X71; Olympus Corporation). The numbers of colonies were then counted and measured using ImageJ software (version 1.51; National Institutes of Health). Colony formation efficiency was calculated as the number of colonies/plated cells ×100%.

For the EdU proliferation assay, the Cell-Light EdU labeling detection kit was purchased from Thermo Fisher Scientific, Inc. As aforementioned, U251 and LN229 cells in 96-well plates (2×10³ cells/well) were transfected with plasmid DNA or siRNA for 48 h. Then, 10 µM EdU was added to the 96-well plates, and they were incubated for 24 h at 37°C with 5% CO₂. The medium was then discarded and cells were washed twice with PBS at room temperature for 3 min each time. The cells were fixed with 4% paraformaldehyde in PBS for 25 min at room temperature and 0.5% Triton X-100 in PBS for 20 min at room temperature. Then, cells were stained with the Alexa-Fluor 594 reaction cocktail for EdU. After removal of the reaction cocktail, cells were washed with 3% BSA (Thermo Fisher Scientific, Inc.) in PBS at room temperature. Mounting medium with DAPI (2 µg/ml; cat. no. ab104139; Abcam) was used to label the cell nuclei for 15 min at room temperature, the cells were then washed with PBS at room temperature three times (30 sec each time). A total of five visual fields were selected randomly using a fluorescence microscope (magnification, ×200; Leica Microsystems GmbH). The EdU-positive cells and DAPI stained cells (total cells) were counted. Cell proliferation rate=number of proliferative cells/number of total cells ×100%.

Cell cycle analysis was performed as previously described (31). For cell cycle analysis, U251 and LN229 cells were transfected with the plasmids or siRNA for 48 h and then collected by centrifugation for 5 min at 1,500 × g at room temperature. Subsequently, cells were washed twice with PBS and fixed with 75% ice-cold ethanol at 4°C for 24 h. The collected cells were re-suspended in PBS containing 25 mg/ml PI, 0.1% Triton and 10 mg/ml RNase [Hangzhou Multi Sciences (Lianke) Biotech Co., Ltd.] and incubated for 30 min in the dark, before being analyzed by flow cytometry. The apoptotic cells were stained with Annexin V-FITC/PI, the cells were analyzed with a Gallios flow cytometer (Beckman Coulter, Inc.) equipped with Kaluza 2.1.1 software (Beckman Coulter, Inc.). Cells were divided into dead cells (B1), late apoptotic cells (B2), negative control normal cells (B3) and early apoptotic cells (B4) according to the different state of the cell. In each experiment, the total percentage of early and late apoptotic cells was compared with the controls.

Briefly, cultured cells were lysed on ice for 30 min in RIPA buffer (Nanjing KeyGen Biotech Co., Ltd.). Subsequently, total protein extraction of different transfected U251 and LN229 cells was evaluated using a BCA Protein Assay kit (Pierce; Thermo Fisher Scientific). Protein (30 µg per lane) was loaded, subjected to 10% SDS-PAGE and then transferred onto PVDF membranes (EMD Millipore) for western blotting. After blocking the membranes with 5% non-fat for 2 h at room temperature, the membranes were probed with antibodies against CDK4 (dilution, 1:1,000; cat. no. ab108357; Abcam), cyclin D1 (dilution, 1:1,000; cat. no. ab40754; Abcam), Bcl-2 (dilution, 1:1,000; cat. no. ab59348; Abcam), Bax (dilution, 1:1,000; cat. no. ab32503; Abcam), KCTD20 (dilution, 1:500; cat. no. ab122094; Abcam), NIPBL (dilution, 1:500; cat. no. ab225908; Abcam), ELAVL1 (dilution, 1:1,000; cat. no. ab200342; Abcam), KLF3 (dilution, 1:500; cat. no. ab154531; Abcam), MMP19 (dilution, 1:500; cat. no. ab53146; Abcam), ZFX (dilution, 1:500; cat. no. 5419; Cell Signaling Technology, Inc.) and GAPDH (dilution, 1:1,000; cat. no. AF5009; Beyotime Institute of Biotechnology) at 4°C overnight. Membranes were then incubated with corresponding secondary antibodies goat anti-mouse (dilution, 1:1,000; cat. no. A0216; Beyotime Institute of Biotechnology) and goat anti-rabbit (dilution, 1:1,000; cat. no. A0208; Beyotime Institute of Biotechnology) for 1 h at room temperature, and processed using ECL Western Blot Detection reagents (EMD Millipore). An ECL detection system (Thermo Fisher Scientific, Inc.) was used to visualize the protein. GAPDH was used as an internal control. Signals were examined by densitometric scans using ImageJ software (version 1.51; National Institutes of Health).

For the RIP assay, pSL-MS2-12X (Addgene, Inc.) was double digested using BamHI and XhoI, and the MS2-12X fragment was inserted to the pcDNA3.1-NEAT1 vector to form pSL-MS2-NEAT1. Next, U251 and LN229 cells were co-transfected with pMS2-GFP and pSL-MS2-NETA1 or pSL-MS2-12× vectors (Addgene, Inc.). After 48 h, the cells were used for the RIP assay with a green fluorescent protein antibody (Roche Diagnostics). The RIP assay was performed using the Magna RIP RNA-Binding Protein Immunoprecipitation kit (EMD Millipore) according to the manufacturer's protocols. Briefly, U251 and LN229 cells were collected by centrifugation at 800 × g for 5 min at 4°C and then lysed using RIP lysis buffer (EMD Millipore). One RIP reaction required 100 µl of cell lysate from ~2.0×10⁷ cells. Subsequently, cell lysates were incubated with 50 µl A/G magnetic beads conjugated with human anti-argonaute RISC catalytic component 2 (Ago2) antibody (dilution, 1:150; cat. no. ab32381; Abcam) for 30 min at room temperature, while normal rabbit IgG (dilution, 1:150; cat. no. 12-370; EMD Millipore) served as a negative control. Next, the compounds of antibody and magnetic beads in 900 µl RIP Immunoprecipitation Buffer (EMD Millipore) were co-incubated overnight at 4°C with cell lysates supernatants (100 µl) after centrifugation at 13,000 × g for 10 min at 4°C. Next, the beads (50 µl) were collected with a magnetic separator, washed three times with cold RIP Wash Buffer (EMD Millipore). Then, the complexes were incubated with 0.1% SDS/0.5 mg/ml proteinase K (EMD Millipore) for 30 min at 55°C to remove proteins and immunoprecipitated RNA was isolated. NEAT1 and miR-324-5p expression was assessed by RT-qPCR, as aforementioned.

The complementary DNA fragment containing the wild-type (WT) or mutant (MUT) sequences of NEAT1 or KCTD20 were subcloned downstream of the luciferase gene within the pGL3-luciferase reporter plasmid vectors (Promega Corporation). Plasmids containing the sequences of miR-324-5p mimic (5′-CGCAUCCCCUAGGGCAUUGGUGU-3′) and miR-NC (5′-UUCUCCGAACGUGUCACGU-3′) were purchased from Shanghai GenePharma Co., Ltd. Briefly, cells were seeded into 24-well plates and cultured overnight. The cells were then co-transfected with WT- or MUT-NEAT1 and the 3′ untranslated region (UTR) of the KCTD20 fragment, as well as equal amounts of miR-NC and miR-324-5p using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocols. At 48 h after transfection, the relative luciferase activity in the cells was measured using a Dual Luciferase Reporter Assay System (Promega Corporation). The relative luciferase activity was normalized to the Renilla luciferase activity. All experiments were performed in triplicate.

A total of 24 male immunodeficient nude mice (n=6 for each group; age, 6 weeks; weight, ~20 g) were purchased from Shanghai Laboratory Animal Research Center. The mice were maintained under specific pathogen-free conditions in a temperature-controlled room (~20°C; humidity, 20%), with a 12-h light/dark cycle, and with ad libitum access to commercially available mouse food and sterilized water. To examine tumor growth in the orthotopic xenograft model, a total of 100 µl of PBS containing U251 cells (2.5×10⁵) were stably transfected with shNEAT1 or shKCTD20 and the corresponding negative control (shNC) was injected intracranially into the striatum of NOD/SCID mice by a stereotactic device (coordinates, 2 mm behind the anterior fontanel, 2 mm lateral to the sagittal suture, at a 3-mm depth from the dura). A bioluminescence imaging system (IVIS Spectrum; PerkinElmer, Inc.) was used to confirm tumor formation and measure tumor growth weekly. At 7, 14, 21 and 28 days after cell implantation, the mice were injected intraperitoneally with D-luciferin at 50 mg/ml and then subjected to in vivo bioluminescence imaging to visualize tumor growth for 10–120 sec. The whole experiment lasted ~2 months. During this process, according to development of neurological signs (hunching, weight loss, rough coat), the moribund mice were anesthetized by an intraperitoneal injection of 10% chloral hydrate (400 mg/kg body weight). The mice showed no signs of peritonitis, pain or other discomfort, and were then euthanized by means of cervical dislocation. Their brains were harvested, cut into sections, paraffin-embedded and stained with H&E to confirm the presence of tumors, and subjected to immunohistochemistry (IHC). For H&E staining, brain tissues were fixed with 10% formaldehyde at room temperature for 24 h. Subsequently, paraffin-embedded specimens were cut into 5-µm sections and dewaxed with xylene and cleared with a series of changing alcohol concentrations. The samples were washed three times in PBS (5 min each time) at room temperature, and then stained with hematoxylin (Sigma-Aldrich; Merck KGaA) for 5 min at room temperature. Sections were stained with eosin (Sigma-Aldrich; Merck KGaA) for 2 min at room temperature to observe the clarity of the nucleus and cytoplasm under a light microscope (Leica Microsystems GmbH). The images were assessed using Image-Pro Plus 6.0 (Media Cybernetics, Inc.). For IHC, tumor tissues were fixed with 10% formaldehyde for 24 h at room temperature and sectioned into 5-µm-thick slides, followed by incubation with citrate buffer (Thermo Fisher Scientific, Inc.) under high pressure to repair the antigen for 3 min. The slides were incubated with 0.3% H₂O₂ for 20 min at room temperature to quench the endogenous peroxidase. Next, blocking reagent (3% BSA; Thermo Fisher Scientific, Inc.) was used to reduce the non-specific binding of antibodies at room temperature for 1 h. Primary antibodies against Ki-67 (dilution, 1:200; cat. no. ab16667; Abcam) and KCTD20 (dilution, 1:100; cat. no. ab122094; Abcam) were used for incubation at 4°C overnight. After washing with PBS, the slides were incubated with biotinylated secondary antibody (dilution 1:500; cat. no. ab207995; Abcam) for 1 h at room temperature. The slides were incubated with HRP-conjugated streptavidin for 40 min at room temperature and then the DAB chromogen (Promega Corporation) was used for visualization. Slides were imaged under a light microscope (Leica Microsystems GmbH). The maximum diameter of the tumor was ~1.1 cm and the tumor volume 0.56 cm³. Tumor volume was calculated using the following formula: Tumor volume=length × width²/2. In addition, the overall survival of the mice was monitored during the experimental period. All animal experimental procedures were approved by the Fourth Military Medical University Institutional Committee for Animal Research (Xi'an, China), and were in accordance with the Animal Management Rule of the Chinese Ministry of Health (document 55, 2001) (32).

lncRNA NEAT1 expression in glioma tissues and normal samples was examined by FISH. Briefly, tissues were fixed with 10% formaldehyde at room temperature for 2 h and embedded in paraffin, and then cut into 5-µm tissue slices. The NEAT1 sequence (5′-CGAGAAACGCACAAGAAGGCAGGCAAACAG-3′) was synthesized by Wuhan Servicebio Technology Co., Ltd. (probe type, oligonucleotide; template source, human), and marked with 5′ digoxigenin (DIG; cat. no. GDP1070; Wuhan Servicebio Technology Co., Ltd.) at 37°C for 16 h. Treated sections were digested with Proteinase K (20 µg/ml; Sigma-Aldrich; Merck KGaA) for 3 min at 37°C, washed three times for 5 min each with deionized water and PBS at room temperature. Next, the sections were post-fixed in 10% paraformaldehyde fixative in PBS for 20 min at room temperature and washed three times (5 min each time) with PBS. And then tissue samples were incubated with pre-hybridization solution (Wuhan Servicebio Technology Co., Ltd.) in an incubator at 37°C for 1 h. Next, after removing the pre-hybridization solution, the hybridization solution (containing 6 ng/µl NEAT probe; Wuhan Servicebio Technology Co., Ltd.) was added onto the slides, followed by incubation at 37°C overnight. Tissue sections were washed with preheated 2X sodium citrate (SSC) for 10 min, 1X SSC (twice for 5 min each) and 0.5X SSC in sequence for 10 min at 37°C to remove non-specific and repetitive RNA hybridization. Subsequently, sections were blocked with 3% BSA (Thermo Fisher Scientific, Inc.) at room temperature for 30 min, followed by incubation with anti-DIG-488 antibody (dilution, 1:300; Jackson ImmunoResearch Laboratories, Inc.) for 50 min and washed three times (3 min each) with PBS at 37°C. Afterwards, the FITC-TSA reagent (dilution, 1:3,000; Wuhan Servicebio Technology Co., Ltd.) was used for incubation for 5 min at room temperature in the dark, followed by washing with PBS three times for 5 min each, and incubation with aqueous fluoroshield mounting medium with DAPI (2 µg/ml; cat. no. ab104139; Abcam) for nuclei staining in the dark for 15 min at room temperature. Finally, the slides were removed from the plate and fixed with 50% glycerol in PBS at room temperature for 30 min. Sections were then examined with a Zeiss LSM 700 confocal microscope (Zeiss AG). The images were analyzed using ImageJ software (version 1.51; National Institutes of Health).

All statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software, Inc.). Gene expression profiles for patients with glioma [low grade glioma (LGG) and high-grade glioma (HGG; GBM)] were obtained from The Cancer Genome Atlas (TCGA; http://tcga-data.nci.nih.gov/tcga/) (33) and clinical data for overall survival were also download from TCGA (34). Gliomas were categorized as LGG [World Health Organization (WHO) Grade I–II] and HGG (WHO Grade III–IV) (35). China Glioma Genome Atlas (CGGA; http://www.cgga.org.cn/) (36) and Gene Expression Omnibus (dataset accession no. GSE16011; http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE16011) (37) were used to select differentially expressed lncRNAs, miRNAs and target genes. Furthermore, StarbaseV2.0 (http://starbase.sysu.edu.cn) and lncRNASNP2 (http://bioinfo.life.hust.edu.cn/lncRNASNP/#!/) (38) were used to predict the potentially targeting relationship between lncRNAs and miRNAs. Based on the dual luciferase reporter assay results, lncRNASNP2 was used to predict the binding site between NEAT1 and miR-324-5p. Three online prediction tools, including Targetscan 7.1 (http://www.targetscan.org/), miRDB (http://mirdb.org) (39) and miRWalk 2.0 (http://mirwalk.umm.uni-heidelberg.de), were used to search for specific miRNA-mRNA relationships. The different results of online data prediction were analyzed and used to draw Venn diagrams using the Venn online analysis software (http://bioinformatics.psb.ugent.be/webtools/Venn/) (40). In addition, putative binding sites between miR-324-5p and KCTD20 from Targetscan were used for miRNA target validation analysis.

All experiments were performed in triplicate and data are presented as the mean ± SD. All statistical analysis was carried out using GraphPad Prism 5.0 (GraphPad Software, Inc.). Comparisons between tumor and adjacent normal tissues were performed using a paired Student's t-test and the experimental and control groups were compared using an unpaired Student's t-test. One-way ANOVA was used to compare multiple different groups followed by Bonferroni's test. Survival curves were analyzed using Kaplan-Meier analysis, and significance was determined using the log-rank test. For TCGA Kaplan-Meier analysis, the median expression level was used as the cut-off value (n=314 >median; n=314 ≤median). The χ² test was used to analyze the association between NEAT1 and the clinicopathologic characteristics of patients. The correlation between NEAT1, miR-324-5p and KCTD20 expression was determined by Spearman's rank correlation analysis. P<0.05 was considered to indicate a statistically significant difference.

To evaluate aberrantly expressed lncRNAs in glioma and to identify those that may be involved in tumorigenesis, the glioma datasets LGG and HGG (GBM) from TCGA was analyzed. Among the dysregulated lncRNAs detected, NEAT1 expression was significantly upregulated in high-grade gliomas (HGG; n=172) compared with in low-grade gliomas (LGG; n=530; Fig. 1A). In addition, RT-qPCR analysis of NEAT1 levels in 43 matched pairs of clinical specimens revealed significantly elevated NEAT1 expression in glioma tissues compared with adjacent normal brain tissues (n=43; Fig. 1B). Additionally, NEAT1 expression was examined in NHAs and a panel of six human glioma cell lines (U87MG, LN229, H4, U251, U118 and A172), and significantly higher NEAT1 expression was detected in all six tumor cell lines, particularly U251 and LN229 cells, compared with in NHAs (Fig. 1C). Furthermore, FISH and IHC demonstrated that the expression levels of NEAT1 and Ki-67, a proliferation marker, were upregulated in glioma sections compared with in normal brain sections (Fig. 1D), which was consistent with the RT-qPCR results (Fig. 1B).

To assess the effect of NEAT1 upregulation on the prognosis of patients with glioma, Kaplan-Meier survival analysis was conducted using the median NEAT1 expression level as the cut-off value for dichotomization of patients from the dataset obtained from TCGA. It was revealed that high NEAT1 expression was significantly associated with a shorter overall survival (Fig. 1E). Furthermore, 43 patients were divided into two groups with high or low NEAT1 expression, using the median expression level as the cut-off value (n=22 >median; n=22 ≤median). The results demonstrated that high NEAT1 expression was significantly associated with a larger tumor size (P=0.047) and advanced World Health Organization glioma stage (P=0.044; Table I). In combination, these data suggested that NEAT1 expression was significantly upregulated in glioma and could potentially serve as a prognostic marker for patients with glioma.

To explore the potential mechanisms via which elevated NEAT1 expression may influence glioma cell biology, gain- and loss-of-function experiments with U251 and LN229 cell lines, which had the highest NEAT1 expression levels (Fig. 1C), were performed by transfecting them with a control sequence (siCtrl) or NEAT1-specific siRNA (siNEAT1), or with a NEAT1 overexpression vector (pcDNA 3.1-NEAT1) or empty vector. RT-qPCR was performed 48 h after transfection and demonstrated that NEAT1 siRNA significantly reduced NEAT1 expression in U251 and LN229 cells, while transfection of pcDNA-NEAT1 increased NEAT1 expression (Fig. 2A).

To analyze the effects of NEAT1 regulation on glioma cell function, the proliferation of transfected cells was assessed using CCK-8, EdU staining and colony formation assays. The results revealed significantly reduced proliferation in cells expressing siNEAT1 compared in cells transfected with siCtrl (Fig. 2B, D and F), as reflected in all three assays. Conversely, the overexpression of NEAT1 promoted cell proliferation in both U251 and LN229 cells in all three assays (Fig. 2C, E and G). Collectively, these results suggested that NEAT1 may act as an oncogene in glioma.

To determine the potential mechanism through which NEAT1 may promote glioma cell proliferation, the effects of knockdown or overexpression of NEAT1 on the cell cycle distribution and apoptotic rate of U251 and LN229 cells were analyzed using flow cytometry. Downregulation of NEAT1 expression resulted in a marked accumulation of U251 and LN229 cells at the G₀/G₁ phase of the cell cycle, with a concomitant decrease in the proportion of cells at the S phase (Fig. 3A). Conversely, an increase in cell cycle progression of G₁-S transition was observed in NEAT1-overexpressing U251 (S phase increased from 21.1 to 34.6%) and LN229 cells (S phase increased from 23.2 to 41.4%; Fig. 3B). Consistent with these results, the proportion of cells undergoing apoptosis was increased significantly following knockdown of NEAT1 and decreased following overexpression of NEAT1 (Fig. 3C and D). These results were supported by the analysis of the expression levels of the cell cycle-related proteins CDK-4 and cyclin D1, the pro-apoptotic protein Bax and the anti-apoptotic protein Bcl-2. Western blotting demonstrated that the protein expression levels of CDK-4, cyclin D1, Bcl-2 and Bax were upregulated or downregulated in a manner consistent with the functional effects of knockdown and overexpression of NEAT1 on cell cycle progression and apoptosis. As shown in Fig. 3E, knockdown of NEAT1 was associated with decreased expression levels of cycle-related proteins (CDK-4 and cyclin D1) and apoptosis-related protein Bcl-2 but the expression levels of apoptosis-related protein Bax were increased. By contrast, the overexpression of NEAT1 inhibited the expression of Bax and promoted the expression of CDK-4, cyclin D1 and Bcl-2 (Fig. 3E).

Subsequently, the present study examined whether the role of NEAT1 in the promotion of glioma cell proliferation in vitro could also be observed in vivo. U251 cell lines stably expressing shNC or shNEAT (n=6 for each group) were generated, and the cells were injected intracranially into nude mice. As shown in Fig. 3F and G, silencing of NEAT1 markedly inhibited the growth of intracranial tumors at all examined time points and also significantly increased the survival of mice. Furthermore, RT-qPCR revealed that the shNEAT1 transfection group had lower NEAT1 levels compared with the control group (Fig. 3H). H&E and IHC staining of tumors excised from experimental mice demonstrated the effect of NEAT1 knockdown on tumor growth. Tumors derived from shNEAT1-expressing U251 cells were markedly smaller and expressed much lower levels of Ki-67 compared with control tumors (Fig. 3I). These data indicated that silencing of NEAT1 inhibited glioma cell proliferation by inducing cell cycle arrest and promoting apoptosis, and demonstrated the carcinogenic activity of NEAT1 in glioma in vivo.

To determine whether the effects of NEAT1 in glioma might be mediated by sponging of ≥1 miRNAs, the present study searched for potential candidate miRNAs with sequences complementary to NEAT1 using the online target prediction tools StarbaseV2.0 and lncRNASNP2. Based on this analysis, 13 miRNAs that were identified by both tools were selected (Fig. 4A), and their expression levels were analyzed in normal and glioma tissues using the TCGA and CGGA datasets. The results identified four miRNAs (miR-125a, miR-324-5p, miR-495-3p and miR-504) that were expressed at significantly lower levels in GBM tissues compared with normal brain tissues (Figs. 4B and S1A), and in HGG compared with LGG (Fig. 4C).

The ability of miR-125a, miR-324-5p, miR-495-3p and miR-504 to interact directly with NEAT1 was analyzed using luciferase reporter assays. Co-transfection of cells with a NEAT1-driven luciferase expression vector and miR-125a or miR-324-5p mimics significantly inhibited luciferase activity compared with that of cells co-transfected with controls (Fig. 4D), indicating that these two miRNAs bound to the NEAT1 sequence. Of the two inhibitory miRNAs, miR-324-5p had the greatest inhibitory activity and was therefore selected for further analysis. To verify binding between NEAT1 and miR-324-5p, the control or miR-324-5p mimics were co-transfected along with luciferase vectors driven by the WT-NEAT1 sequence or a mutated sequence (MUT-NEAT1) carrying mismatched residues in the predicted miR-324-5p binding site (Fig. 4E). As shown in Fig. 4F, the luciferase activity of U251 and LN229 cells co-transfected with WT-NEAT1 and miR-324-5p was significantly reduced compared with that of cells transfected with miR-NC, but there was no significant change in the MUT-NEAT1 group, which demonstrated a direct association between NEAT1 and miR-324-5p. These results were substantiated by the results of the RIP experiments in U251 and LN229 cells. The RIP analysis revealed that NEAT1 was highly enriched in Ago2 immmunoprecipitates from cells transfected with the miR-324-5p mimics compared with the anti-IgG group (Fig. 4G).

The assessment of the functional association between NEAT1 and miR-324-5p by RT-qPCR, and it was revealed that knockdown of NEAT1 was associated with the upregulation of miR-324-5p in U251 and LN229 cells. Furthermore, overexpression of NEAT1 was associated with the opposite results (Fig. 4H). The downregulation or upregulation of miR-324-5p had no significant effect on NEAT1 expression (Fig. S1B). These results demonstrated that NEAT1 may directly bind to and modulate the expression levels of miR-324-5p in glioma cells. The expression levels of miR-324-5p were detected in NHA, U251 and LN229 cells by RT-qPCR, and the results revealed that, compared those in with NHA cells, the expression levels of miR-324-5p in U251 and LN229 were significantly reduced (Fig. 4I). Furthermore, miR-324-5p expression was significantly lower in human glioma tissues compared with in normal tissues (Fig. 4J), and Spearman's correlation analysis demonstrated a significant negative correlation between the expression levels of NEAT1 and miR-324-5p in glioma tissues (Fig. 4K). In combination, these data demonstrated that NEAT1 functions as a ceRNA to modulate miR-324-5p levels in glioma.

Having established an association between NEAT1 and miR-324-5p, it was next examined whether miR-324-5p mediates the NEAT1-induced changes in glioma cell biology. For these experiments, U251 and LN229 cells were transfected with a miR-324-5p inhibitor (Fig. 5A) with or without concomitant NEAT1 knockdown. Compared with those of control glioma cells, transfection of the miR-324-5p inhibitor alone markedly increased proliferation and colony formation (Fig. 5B and C). In addition, the experiments demonstrated that overexpression of miR-324-5p significantly inhibited the progression of glioma in vitro. miR-324-5p mimics were transfected into U251 and LN229 cells (Fig. S1C). Colony formation and EdU assays revealed that overexpression of miR-324-5p significantly inhibited the cell proliferation rate (Fig. S1D and E). Additionally, the cell cycle was blocked at the G₀/G₁ phase (Fig. S1F). Additionally, upregulation of miR-324-5p resulted in an increase in the number of apoptotic cells (Fig. S1G). Western blotting was used to analyze the expression levels of cycle-related proteins and apoptosis-related proteins after cells were transfected with miR-324-5p mimics (Fig. S1H). As shown in Fig. 5E-J, glioma cells were co-transfected with siNEAT1 and the miR-324-5p inhibitor or siCtrl. These experiments demonstrated that miR-324-5p inhibition partially reversed the effects of transfection with siNEAT1 on U251 and LN229 cell proliferation and colony formation (Fig. 5D and E), as well as cell cycle arrest and apoptosis (Fig. 5F-I), and had a corresponding rescue effect on the expression levels of cell cycle and apoptosis-related proteins (Fig. 5J). Collectively, these data suggested that a potential interaction between NEAT1 and miR-324-5p may be involved in the development of glioma.

The present study next sought to identify the target mRNAs of miR-324-5p through which NEAT1 and miR-324-5p regulate glioma cell proliferation. Using the online informatics tools TargetScan, miRDB and miRWalk, which predict miR-324-5p target mRNAs based on sequence complementarity, 52 shared target genes with potential binding sites for miR-324-5p were identified from the three online informatics tools (Fig. 6A). Among the 52 genes, six were expressed at significantly higher levels in glioma tissues compared with normal tissues in the dataset from TCGA (Fig. S2A and B). The mRNA and protein expression levels of the six genes were analyzed in U251 cells transfected with a control sequence or the miR-324-5p inhibitor. This analysis revealed that miR-324-5p suppression significantly increased the expression levels of all six genes. However, the most striking effect was that on KCTD20 expression (Fig. 6B). Notably, KCTD20 expression was consistently upregulated in glioma tissues compared with normal brain tissues in the clinical specimens, as well as the TCGA, CGGA and GSE16011 datasets (Figs. 6C and S2B). Similarly, IHC staining showed that the abundance of KCTD20 protein in human glioma samples was increased compared with that in normal brain tissue (Fig. S2C). Therefore, KCTD20 was selected for further analysis of the NEAT1-miR-324-5p-mRNA regulatory relationship in glioma cells.

To verify that KCTD20 mRNA was directly regulated by miR-324-5p in glioma cells, a luciferase reporter assay was performed in cells expressing plasmids with the WT 3′-UTR of KCTD20 or MUT 3′-UTR carrying mutations in the putative binding site for miR-324-5p (Fig. 6D). Co-transfection with miR-324-5p mimics significantly repressed luciferase activity in U251 and LN229 cells expressing the WT, but not the MUT, KCTD20 3′-UTR (Fig. 6E). Furthermore, KCTD20 protein and mRNA expression was increased following the transfection of glioma cells with the miR-324-5p inhibitor and suppressed by their transfection with miR-324-5p mimics (Fig. 6F). These data demonstrated that miR-324-5p functionally regulated KCTD20 expression in glioma cells and suggested that KCTD20 expression may therefore be regulated by changes in NEAT1 levels. To test this, KCTD20 expression was examined by western blotting and RT-qPCR in cells transfected with siNEAT1 or a control sequence. Notably, KCTD20 expression was markedly reduced following NEAT1 knockdown (Fig. 6G). However, co-silencing of miR-324-5p abolished the inhibitory effects of siNEAT1 on KCTD20 mRNA and protein expression (Fig. 6H). IHC staining revealed reduced KCTD20 expression in tumor sections from mice injected with shNEAT1-expressing glioma cells compared with in mice injected with control cells (Fig. S2D). Finally, a significant positive correlation was identified between the mRNA expression levels of NEAT1 and KCTD20 in the glioma specimens (Fig. 6I). Since NEAT1 could sponge miR-324-5p, the present study next determined whether NEAT1 could regulate KCTD20 expression by binding to the same site in miR-204-5p. Luciferase reporter assays demonstrated that miR-324-5p could bind to NEAT1 and the 3′ UTR of KCTD20. To determine whether miR-324-5p served a role in the relationship between NEAT1 and KCTD20, cells were co-transfected with siNEAT1 and the miR-324-5p inhibitor. Knockdown of NEAT1 also significantly reduced KCTD20 mRNA and protein expression in U251 and LN229 cells (Fig. 6G). Additionally, the downregulation of KCTD20 protein expression induced by siNEAT1 was effectively reversed by the miR-324-5p inhibitor. The experimental results also showed that changes in the expression of NEAT1 could affect the expression levels of miR-324-5p; however, changes in the expression levels of miR-324-5p had no effect on NEAT1 (Fig. S1B). The expression levels of NEAT1 and KCTD20 in clinical tissues were positively associated (Fig. 6I), consistent with the existence of a NEAT1-miR-324-5p-KCTD20 regulatory axis. In combination, these data suggested that NEAT1 regulated the expression levels of KCTD20 in glioma cells by post-transcriptional modulation of miR-324-5p.

Subsequently, it was determined whether the functional effects of NEAT1 and miR-324-5p on glioma cell biology were mediated through their control of KCTD20 expression. Transfection of U251 and LN229 cells with siKCTD20 effectively decreased KCTD20 mRNA and protein expression compared with that in control cells (Fig. 7A). Notably, the silencing of KCTD20 significantly inhibited U251 and LN229 proliferation, based on the results of CCK-8, colony formation and EdU incorporation assays (Fig. 7B-D), Furthermore, the results demonstrated that more cells were in the G₀/G₁ phase and fewer cells were in S phase of the cell cycle after cells were transfected with siKCTD20 (Fig. S3A), Depletion of KCTD20 significantly increased the efficiency of apoptosis (Fig. S3B). Similarly, the protein expression levels of cyclin D1, CDK4 and Bcl-2 were downregulated, while the expression levels of Bax protein were upregulated (Fig. S3C). These results were consistent with the observed effects of knockdown of NEAT1 or overexpression of miR-324-5p on glioma cell functions.

The findings were verified by examining the effects of KCTD20 knockdown or concomitant KCTD20 and miR-324-5p knockdown on glioma cell biology. First, U251 cells stably expressing luciferase and either shKCTD20 or control shNC (n=6 for each group) were intracerebrally injected into nude mice, and the mice were followed up by imaging examinations for up to 28 days. Consistent with the results of the in vitro assays, silencing of KCTD20 markedly inhibited glioma growth and significantly improved the survival time of tumor-bearing mice (Fig. S3D and E). As presented in Fig. S3F, immunoblot analysis indicated that KCTD20 expression was downregulated in shKCTD20-inoculated tumor tissues. In addition, transfection of miR-324-5p inhibitor in addition to knockdown of KCTD20 in glioma cell lines partially reversed the effects of KCTD20 knockdown (Figs. 7E, F and H and S3G and H), demonstrating the functional relationship between miR-324-5p and KCTD20. Furthermore, the concomitant knockdown of miR-324-5p and KCTD20 in U251 and LN229 cells abrogated the effects of knockdown of KCTD20 on cell cycle arrest (Fig. 7G), apoptosis (Fig. 7I), and CDK4, cyclin D1, Bcl-2 and Bax expression (Fig. 7J). Finally, a statistically significant inverse correlation was detected between the expression levels of miR-324-5p and KCTD20 in the clinical glioma tissues (r=−0.4582; P=0.002; Fig. S3I). In combination, the data presented in the current study provided substantial evidence that NEAT1 is oncogenic in glioma cells and exerts its effects through the modulation of miR-324-5p-regulated KCTD20 expression.