The human glioma cell lines U251 and A172 were obtained from Cell Resource Center of Shanghai and cultured in Dulbecco's Modified Eagle Medium (DMSO; Gibco, Grand Island, NY, USA) supplemented with 10% fetal calf serum (FBS; Gibco), 100 U/ml penicillin and 100 mg/ml streptomycin (Life Technologies, Grand Island, NY, USA) at 37℃ in 5% CO₂.

Glioma gene expression arrays with survival data were obtained from the National Cancer Institute Repository for Molecular Brain Neoplasia Data (NCI REMBRANDT), the Cancer Genome Atlas (TCGA), and gene expression omnibus datasets. TCGA data were extracted directly from the web site (https://portal.gdc.cancer.gov/). For assessing the overall survival (OS) of glioma patients included in REMBRANDT, we used project Betastasis (http://www.betastasis.com). For assessing the OS of glioma patients included in GSE16011, we used R2: Genomics Analysis and Visualization Platform (http://r2.amc.nl). OS was defined as the period from the date of the pathological diagnosis to death.

Small interfering RNA (siRNA) targeting FoxD2-AS1 and negative control siRNA were synthesized by Biomics Biotechnologies Co., Ltd. (Nantong, China). Sense siFoxD2-AS1: 5′-GCGAAGAGUACGUUGCUAUTT-3′, antisense siFoxD2-AS1 : 5′-AUAGCAACGUACUCUCGCTT-3′; Sense siNC: 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense siNC: 5′-ACGUGACACGUUCGGAGAATT-3′. Cells were grown on six-well plates to confluency and transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Briefly, each well was supplemented with 5 µl Lipofectamine 2000 and 100 pmol siRNA. About 48 h later, cells were acquired for the following experiments.

siFoxD2-AS1 or siNC transfected cells were treated with TMZ for 48 h, and then were washed once with phosphate buffer saline (PBS) and dissolved in Protein Extraction Reagent (Boster Bioengineering, Wuhan, China) containing 1 mM phenylmethanesulfonyl fluoride (PMSF; Roche Molecular Biochemicals, Indianapolis, IN, USA). Protein samples were separated by 10%–12% SDS-PAGE, transferred onto the surface of polyvinylidene fluoride membrane and immunoblotted with the indicated primary antibodies. Using the enhanced chemiluminescence reagent (Amersham Biosciences, Piscataway, NJ, USA), visualization of the protein bands was conducted in the Omega Lum G system (Aplegen, Pleasanton, CA, USA). The primary antibodiy to β-actin was purchased from Hua Bio (Hangzhou HuaAn Biotechnology Co., Ltd., Hangzhou, China), and the antibodys to Caspase-3, Bax, MGMT were purchased from (Proteintech Group, Chicago, IL, USA).

The DNA was extracted using Takara MiniBEST Universal Genomic DNA Kit (Takara, Shiga, Japan) and then 1 µg of extracted DNA underwent bisulfite modification using a DNA Methylation Kit (CoWin Biosciences, Beijing, China) according to the manufacturer's instructions. The MS-PCR was performed under the following conditions: 95℃ for 10 min; 35 cycles of 95℃ for 30 sec, the annealing temperature 60℃ for 30 sec, and 72℃ for 30 sec; and a final extension of 10 min at 72℃. Primers for the methylated MGMT were 5′-TTTCGACGTTCGTAGGTTTTCGC-3′ (forward) and 5′-GCACTCTTCCGAAAACGAAACG-3′ (reverse); for the un-methylated MGMT were 5′-TTTGTGTTTTGATGTTTGTAGGTTTTT GT-3′ (forward) and 5′-AACTCCACACTCTTCCAAAAACAAAACA-3′ (reverse). The PCR product was directly loaded onto 2.5% agarose gels, stained with ethidium bromide, and visualized with the aid of ultraviolet light. The density of each band was quantified using imaging analysis and the relative band density values were calculated as the ratio of methylated MGMT to that of methylated plus un-methylated MGMT.

Cell viability was evaluated by using MTT assay. Glioma cells were seeded into 96-well plates at the concentration of 2 × 10³ cells/well. Cells were treated with different concentrations of TMZ (MedChem Express, Monmouth Junction, NJ, USA) for 48 h. Then, 10 µl MTT (5 mg/ml) was added to each well and incubated in the dark at 37℃ for another 4 h. Absorbance was determined at a wavelength of 570 nm using a SpectraMax M3 microplatereader (Molecular Devices, Sunnyvale, CA, USA).

Cells were seeded (3–5 × 10⁵ cells/well) and then transfected with siFoxD2-AS1 or siNC alone. After 24 h post-transfection, cells were treated with TMZ for another 24 h, washed, serially diluted, plated in triplicate into 6 well plates, allowed 6–8 days of cell growth for colony formation, stained with 5% crystal violet. Finally, colonies were photographed, and the number of colonies was counted manually.

In the migration assay, cells were suspended in DMEM and the density was adjusted to approximately 1 × 10⁵ cells/ml. Cell suspension was then seed to the upper chamber (Corning, Corning, NY, USA) of the inserts (0.2 ml/well), and DMEM containing negative control siRNA were synthesized by Biomics Biotechnologies Co., Ltd. (Nantong, China). Sense siFoxD2-AS1: 5′-GCGAAGAGUACGUUGCUAUTT-3′, antisense siFoxD2-AS1 : 5′-AUAGCAACGUACUCUCGCTT-3′; Sense siNC: 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense siNC: 5′-ACGUGACACGUUCGGAGAATT-3′. Cells were grown on six-well plates to confluency and transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Briefly, each well was supplemented with 5 µl Lipofectamine 2000 and 100 pmol siRNA. About 48 h later, cells were acquired for the following experiments.

An apoptosis analysis kit (NanJing KeyGen Biotech Co., Ltd., Nanjing, China) was utilized to detect the cell apoptotic rates. Briefly, the Annexin V and PI were firstly added into the binding buffer, and the treated cells were resuspended in the binding buffer. After incubating in the dark for 15 min, the cells were washed and analyzed using a Guava easyCyte 5 Flow Cytometer (EMD Millipore, Bellerica, MA, USA).

All statistical data were analyzed using the SPSS 21.0 software (IBM Co., Armonk, NY, USA) and GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA). p-values < 0.05 were considered statistically significant. Data are expressed as mean ± standard deviation. Statistical significance was determined using Student's t-test or one-way analysis of variance. The Kaplan–Meier method with the log-rank test was applied for OS.

Three independent glioma gene expression datasets (TCGA, GSE16011, and REMBRANDT) were employed to examine the association between FoxD2-AS1 expression levels and clinical outcome of patients with glioma. First, we compared the FoxD2-AS1 expression in recurrent glioma tissues with non-recurrent tissues obtained from TCGA and found that FoxD2-AS1 expression is significantly higher in recurrent than in non-recurrent tissues (p < 0.01, Fig. 1A). Further Kaplan–Meier and log-rank test analysis showed that FoxD2-AS1 was associated with OS of glioma patients included in TCGA, GSE16011, and REMBRANDT, patients with FoxD2-AS1 higher expression had a significant shorter overall survival time than those with FoxD2-AS1 lower expression (Fig. 1B–D). Moreover, high expression of FoxD2-AS1 is significantly correlated with poor patient outcome in all tested cases, namely, IDH wildtype and mutant tumors, astrocytoma, and oligodendroglioma (Fig. 1E–H). Collectively, these data suggest that high FoxD2-AS1 expression is associated with poor outcome in patients with glioma and potentially contributing to the aggressive tumor biology of glioma.

Evidence of a correlation between FoxD2-AS1 expression and the clinicopathological status of patients with glioma was searched in TCGA. The results of a chi-square test indicated that methylation of MGMT is significantly less frequent in high FoxD2-AS1 expression patients. As shown in Table 1, methylation of MGMT was found in 81.6% (266/326) of patients with low FoxD2-AS1 expression, but in 66.7% (205/307) of patients with high FoxD2-AS1 expression (p < 0.001). Significant correlations were also detected between FoxD2-AS1 expression and certain clinicopathological features, including IDH mutation status and neoplasm histologic grade (Table 1).

To asked whether FoxD2-AS1 could modulate the sensitivity of glioma cells to TMZ, the siNC or siFoxD2-AS1 transduced A172 and U251 cells were treated with different doses of TMZ, respectively. The results of the MTT assay shown that down-regulation of FoxD2-AS1 reduced the viability of glioma cells after treatment with TMZ (Fig. 2A). Moreover, we found that the IC₅₀ values for TMZ in siFoxD2-AS1 transduced glioma cells were much lower than that in their control cells (56.66 µg/ml in A172 cells, 52.27 µg/ml in A172/siFox cells, 49.15 µg/ml in U251 cells, and 45.31 µg/ml in U251/siFox cells). Next, the colony formation assay, flow cytometry, and western blot analysis were performed to further determine the effect of FoxD2-AS1 on TMZ resistance of glioma cells. As shown in Fig. 2B, the colony formation for siFoxD2-AS1 transduced cells was lower than that of the siNC group by treatment with 20 and 40 µg/ml TMZ. The results of flow cytometry analysis revealed that FoxD2-AS1 knockdown significantly increased glioma cell apoptosis caused by TMZ treatment (Fig. 2C). Furthermore, Western blot analysis showed that TMZ-induced cellular apoptosis was greatly enhanced by FoxD2-AS1 knockdown compared to NC (Fig. 2D). Above all, it was implied that down-regulation of FoxD2-AS1 increased the sensitivity of glioma cells to TMZ.

Previous studies reported that the arisen of chemoresistance may associated with the high metastatic ability of cancer cells [20], so we used transwell system to investigate the role of FoxD2-AS1 on the migration and invasion ability of glioma cells. As shown in Fig. 3, the cell counts of U251 and A172 were significantly decreased in both migration and invasion systems after transfection with siFoxD2-AS1, which suggested that FoxD2-AS1 knockdown could inhibit the metastatic ability of glioma cells.

MGMT promoter methylation is the key mechanism of MGMT gene silencing and predicts a favorable TMZ chemotherapy outcome in glioblastoma (GBM) patients. To further clarify the relationship between FoxD2-AS1 expression and methylation status of the MGMT promoter in glioma cells, DNA was isolated from siNC or siFoxD2-AS1 transduced A172 and U251 cells and MGMT methylation was determined by MS-PCR. As shown in Fig. 4A, the ratio of methylated promoter DNA was much higher in siFoxD2-AS1 transduced than siNC transduced glioma cells. We further detected the MGMT protein expression in siNC or siFoxD2-AS1 transduced glioma cells by western blot, the results shown that the MGMT protein level was much lower in siFoxD2-AS1 transduced cells than that in siNC group (Fig. 4B). These results demonstrated that FoxD2-AS1 mediated TMZ resistance in glioma cells by regulating the methylation status of the MGMT promoter region.