In the present study, rat C6 and human U87MG glioma cells (of unknown origin) (HTB-14™; The American Type Culture Collection, Manassas, VA, USA) were used. The cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Life Technologies; Thermo Fisher Scientific, Inc., Waltham, A, USA) containing 10% fetal bovine serum (FBS; HyClone Laboratories; GE Healthcare, Chicago, IL, USA), 0.1% β-mercaptoethanol (Life Technologies; Thermo Fisher Scientific, Inc.), 1% penicillin/streptomycin (P/S) (Life Technologies; Thermo Fisher Scientific, Inc.), 1% non-essential amino acids (NEAA; Life Technologies; Thermo Fisher Scientific, Inc.) and 1% sodium pyruvate (Life Technologies; Thermo Fisher Scientific, Inc.), and incubated at 37°C in a humidified atmosphere containing 5% CO₂.

Glioma cells were seeded on 1% basement membrane matrix-coated plates (BD Biosciences, San Jose, CA, USA) at a density of 1×10³ cells/cm². After incubating the cells at 37°C for 24 h, we replaced the medium with glial differentiation medium, which consisted of neurobasal medium:advanced DMEM/F12 (1X) (1:1), 1% N2 supplement (100X), 2% B27 supplement (50X), 0.05% bovine serum albumin (BSA), 1% P/S, 1% glutamax-I (100X), 0.1% β-mercaptoethanol (1000X) (all of which were from Life Technologies; Thermo Fisher Scientific, Inc.), BDNF (10 ng/ml), GDNF (10 ng/ml) and NT-3 (10 ng/ml) (all from PeproTech, Inc., Rocky Hill, NJ, USA). We then added a combination of small molecules, composed of either forskolin (Tocris Bioscience, Bristol, UK) and rapamycin (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) or forskolin, rapamycin and T3 (Sigma-Aldrich; Merck KGaA). After 7 days, we replaced the medium with only I-BET151 (Tocris Bioscience) for maturation of the differentiated cells. The cells were then incubated at 37°C for 7 additional days. We replaced the media every 2–3 days. In some experiments, we used temozolomide (TMZ; 50 µM) to compare the growth inhibition effect of the small-molecule combination.

The small molecules were dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich; Merck KGaA) and diluted in glial differentiation medium to the following final concentrations: forskolin, 100 µM; rapamycin, 100 nM; T3, 100 nM; I-BET151, 1 µM; and TMZ, 50 µM. We tested the glial cell differentiation into glioma in three independent experiments.

For measuring cell proliferation, we used the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Sigma-Aldrich; Merck KGaA). Cells from each group, containing the same volume of DMSO as that in the differentiation media, were differentiated for 7 or 14 days in 24-well plates. Then, we added 5 mg/ml MTT solution to the media and incubated them at 37°C for 3 h. The media were removed, and formazan crystals were dissolved using DMSO. The samples were then incubated at 37°C for 10 min and transferred to 96-well plates, and the absorbance was measured at 490 nm using VersaMax ELISA microplate reader (Molecular Devices, Sunnyvale, CA, USA). Results were acquired from three independent experiments.

Media were removed and the cells were washed with phosphate-buffered saline (PBS; HyClone Laboratories; GE Healthcare). The cells were then fixed with 4% paraformaldehyde (Millipore, Temecula, CA, USA), pH 7.2, for 10 min at room temperature. Subsequently, they were washed thrice with 0.3% Tween-20 (Life Technologies; Thermo Fisher Scientific, Inc.) in PBS for 3 min. The blocking procedure was performed in PBS with 10% normal donkey serum and 0.3% Triton-X 100 for 30 min at room temperature. Primary antibodies, diluted in the blocking buffer, namely, mouse monoclonal anti-Nestin (dilution 1:1,000; cat. no. ab6142; Abcam, Cambridge, MA, USA), rabbit polyclonal anti-NG2 (dilution 1:250; cat. no. ab5320; Millipore), rabbit polyclonal anti-Olig2 (dilution 1:100; cat. no. ab42453; Abcam), rabbit polyclonal anti-PDGFRa (dilution 1:500; cat. no. sc-338; Santa Cruz Biotechnology, Santa Cruz, CA, USA;) mouse monoclonal anti-MBP (dilution 1:1,000; cat. no. ab2404; Abcam), mouse monoclonal anti-CNP (dilution 1:1,000; cat. no. NE1020; Millipore), mouse monoclonal anti-oligodendrocyte (RIP; dilution 1:50,000; cat. no. MAB1580; Millipore), rabbit polyclonal anti-GFAP (dilution 1:1,000; cat. no. ab7260; Abcam), mouse monoclonal anti-GFAP (dilution 1:1,000; cat. no. G3893; Sigma-Aldrich; Merck KGaA), mouse monoclonal anti-O4 (dilution 1:100; cat. no. MAB345; Millipore) and rabbit polyclonal anti-Ki-67 (dilution 1:250; cat. no. ab15580; Abcam), were allowed to react with the samples for 60 min at room temperature. After washing with PBS with 0.3% Tween-20, secondary antibodies diluted in the blocking buffer were added and allowed to react for 30 min at room temperature: FITC donkey anti-rabbit (cat. no. 711-095-152) or anti-mouse (cat. no. 715-095-151) and Cy3 donkey anti-rabbit (cat. no. 711-165-152) or anti-mouse (cat. no. 715-165-151) (dilution 1:500; all of them are from Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) were used as secondary antibodies. The samples were then washed thrice and stained using 4′,6′-diamino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA). The samples were then covered with glass coverslips and examined using a confocal laser scanning microscope (LSM 700; Carl Zeiss, Oberkochen, Germany).

To determine the number of positive cells stained by each antibody (GFAP, CNP, RIP and Ki-67), we performed flow cytometric analyses. Briefly, the cells were harvested using 0.25% trypsin/EDTA (Life Technologies; Thermo Fisher Scientific, Inc.). For fixation, 4% paraformaldehyde was added to the cell pellet and incubated for 10 min. After centrifugation at 400 × g for 5 min, the supernatant was discarded and the cell pellet was washed twice with 0.3% Tween-20 in PBS for 3 min. Next, the primary antibodies (GFAP, 1:1,000; CNP, 1:1,000; RIP, 1:50,000; and Ki-67, 1:250) were added to the pellet and allowed to react for 60 min at room temperature. Following a second wash using the same procedure, secondary antibodies (FITC donkey anti-rabbit or anti-mouse and Cy3 donkey anti-rabbit or anti-mouse; dilution 1:500) were added to the cell pellet and allowed to react for 30 min at room temperature. Finally, the pellet was washed and re-suspended in PBS. Samples were analyzed using a FACSCalibur (Becton-Dickinson; BD Biosciences, San Jose, CA, USA) and CellQuest software (Becton-Dickinson; BD Biosciences). Results were acquired from three independent experiments.

Microarray analysis was performed according to the manufacturer's instructions. After total RNA isolation, cDNA was synthesized using the GeneChip WT (Whole Transcript) Amplification kit (Thermo Fisher Scientific, Inc.). Labeled DNA target was hybridized to the Affymetrix GeneChip Array (Affymetrix; Thermo Fisher Scientific, Inc.). Hybridized arrays were washed and stained on a GeneChip Fluidics Station 450 and scanned on a GCS3000 Scanner (Thermo Fisher Scientific, Inc.). Analysis was performed using Affymetrix^® GeneChip Command Console^® Software (AGCC; Affymetrix; Thermo Fisher Scientific, Inc.). Microarray data were deposited in a public database (https://www.ncbi.nlm.nih.gov/geo/), GSE101337 (For undifferentiated C6 and SMiG), and 101338 (For single clone no. 4 and single clone no. 8).

To determine the proliferation rate of untreated glioma cells, i.e., those without the small molecules, we seeded them at a density of 1×10³ cells/cm² on a 24-well plate. When the cells reached confluence, they were harvested using 0.25% trypsin/EDTA. After being centrifuged at 268 × g for 3 min and re-suspended in 1 ml culture medium, the cells were diluted to half concentration in trypan blue (Life Technologies; Thermo Fisher Scientific, Inc.) and counted using the manual cell counting method. All the cells were then seeded on 12-well plates (Corning Inc., Corning, NY, USA). The cells were counted when they reached confluence, and were re-plated on wider plates, using the same method. Results were acquired from three independent experiments.

We performed a serial dilution of glioma cells for isolation of a single colony. Briefly, we added 100 µl culture media into all the wells of a 96-well plate. Approximately 100 cells were mixed with 100 µl culture medium and added into the first well of a 96-well plate. Then, 100 µl was taken from the first well, and then serially diluted to the next well, and so on. After confirming the single cell isolation, cells were incubated until a colony was formed. After 14 days, each colony was treated with trypsin for 5 min, and maintained in the culture media.

We tested the glial differentiation efficiency with a few isolated glioma. We used the clone no. 4 and 8 glioma for the present study.

A two-tailed Student's t-test was performed to assess differences between two groups. Two-way analysis of variance (ANOVA) followed by Tukey's post hoc test was performed to assess differences among three groups or more. Data are presented as mean ± standard error of the mean (SEM) and P-values <0.05 were considered to indicate a statistically significant result.

A schematic diagram was provided to facilitate understanding of the present study (Fig. 1A). First, we confirmed that undifferentiated glioma cells were positive for specific markers of glial progenitor cells, namely, Nestin, NG2, Olig2 and PDGFRa (Fig. 1B). However, the glioma cells were negative for other markers of glial cells, namely, GFAP, CNP, RIP and O4 (Fig. 1B).

Therefore, we aimed to ascertain whether T3, which is typically used to differentiate glial progenitor cells into oligodendrocytes, could convert gliomas into glial cells (i.e., oligodendrocytes). However, the T3-treated glioma cells were not transformed into a glial-specific morphology (Fig. 1C). Next, we added T3 and forskolin to the glioma cells, but could not confirm glial-specific morphology until 7 days of incubation (Fig. 1C). Nevertheless, we persisted in our efforts to convert C6 glioma cells into glial cells by adding rapamycin to T3 and forskolin, and incubating the cells for 7 days. With this, we were finally able to observe glial-specific morphology (Fig. 1C), as well as glial-specific markers such as GFAP, CNP, and RIP (Fig. 1D).

After several experiments, we confirmed that this glial-specific morphology could be maintained using medium containing rapamycin, T3 and I-BET151 (Fig. 1E). To further confirm whether the converted cells were fully differentiated into glial cells, they were cultured in the absence of small molecules for 3 days. Consequently, we observed that the glial-specific morphology was retained; the GFAP-, CNP- and RIP-positive cells did not react with Ki-67 (Fig. 1F).

Next, we investigated the growth inhibitory effect of the small molecules. The undifferentiated glioma cells grew rapidly, despite a very low cell density (Fig. 1G). In contrast, treatment with a combination of forskolin, T3 and rapamycin conspicuously decreased the cell proliferation (Fig. 1H). Further experimentation confirmed that I-BET151 played a major role in the growth inhibition after glial induction (Fig. 1I).

To determine the major compound among forskolin, T3, rapamycin and I-BET151, responsible for glial induction, we performed further experiments. We confirmed that glioma cells were converted into GFAP-, CNP-, RIP- and O4-positive cells with glial-specific morphology, even in the absence of T3 (Fig. 2A and B). These converted cells did not react with Ki-67 (marker for dividing cells) (Fig. 2B). Thus, T3 did not affect the glial conversion of glioma cells.

We further confirmed that glial-specific morphology was maintained in the presence of I-BET151 after induction of glial differentiation using forskolin and rapamycin, for 7 days (Fig. 2C). In contrast, it was difficult to maintain glial-specific morphology in the absence of I-BET151 (Fig. 2D). As a quantitative result, the number of CNP-, RIP- and GFAP-positive cells clearly increased in the SMiGs converted using forskolin, rapamycin and I-BET151 (Fig. 2E). We also observed that glial-specific morphology was retained even when small molecules were withdrawn for 3 days (Fig. 2F).

We performed microarray analysis to compare the gene expression pattern between glioma cells and SMiGs. In a microarray analysis, gene expression profile related to oligodendrocyte differentiation and myelination was significantly induced in the SMiGs. In contrast, gene expression profiles relating to cell division and mitosis, or extracellular matrix (ECM) and vessel development were markedly upregulated in glioma cells (Fig. 2G). Gene Ontology related to glial differentiation, including oligodendrocyte differentiation and myelination, was significantly different between undifferentiated glioma cells and SMiGs (Fig. 2H). These results revealed that specific characteristics of malignant gliomas can be converted into those of glial cells by treatment with a combination of forskolin, rapamycin and I-BET151.

We investigated whether the strong proliferation potency of malignant gliomas can be inhibited by glial conversion. On day 7 after treatment with a combination of forskolin and rapamycin, we confirmed that the proliferation of glioma cells was significantly reduced compared to that of the single-molecule treatment group, and the rate of Ki-67-positive cells was significantly decreased compared to that in the DMSO group (Fig. 3A). We also confirmed that the proliferation of U87MG glioma cells was significantly reduced compared to that of the single-molecule treatment group (Fig. 3B). Further treatment with I-BET151 was more effective in inhibiting cell proliferation (Fig. 3C and D). This pattern was similar even in the absence of the small molecules (Fig. 3E). These results indicated that the strong proliferation capacity of malignant glioma cell can be controlled by glial conversion.

Some cells were not positive for GFAP, CNP, or RIP after glial induction by small molecules (Fig. 4A). We, therefore, selected each single colony considering the glioma's heterogeneity (Fig. 4B). Of the many colonies we isolated, colony no. 4 did not react with GFAP, CNP or RIP after glial induction with forskolin, rapamycin and I-BET151; it was only positive for Ki-67 (Fig. 4C). However, most of colony no. 8 was positive for GFAP, CNP and RIP after glial induction with forskolin, rapamycin and I-BET151 (Fig. 4C). GFAP-, CNP- and RIP-positive cells did not react with Ki-67 (Fig. 4C). We, then, analyzed the gene expression patterns between colony no. 4 and 8, using a microarray analysis (Fig. 4D). The top 10 most-expressed genes in colony no. 8 were prkcb, postn, mpz, pros1, cd55, ct55, plxdc2, antxr1, fmr1nb and igfbp5 (Fig. 4E and F). We then compared the proliferation-inhibition effect of the small-molecule combination with that of a standard medication such as temozolomide (TMZ). For colony no. 4, cell proliferation was reduced by TMZ, but not by the small-molecule combination (Fig. 4G), although, our small molecules inhibited proliferation more effectively than TMZ in colony no. 8 (Fig. 4G).