The human glioblastoma cell line A172 was purchased from American Type Culture Collection (ATCC; Manassas, VA, USA) and maintained in Dulbecco’s modified Eagle’s medium (DMEM; HyClone, Logan, UT, USA) containing 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin at 37°C in a 5% CO₂ atmosphere. A172 cells were maintained in the above medium (Nm) and grown as monolayers on polystyrene-treated tissue culture plates (NP). The cells (1×10⁵ cells) were cultured under the following conditions for 72 h: i) Nm or serum-free glioblastoma sphere medium (GBM) containing EGF (10 ng/ml) and bFGF (10 ng/ml) (both from R&D Systems, Minneapolis, MN, USA) as previously described (10,23), and ii) NP, ultra-low attachment plates (UP; Corning, Tewksbury, MA, USA), 20 mg/ml poly-HEMA-coated plates (HP) and 1% agar-coated plates (AP). The morphology of cells was examined using an inverted phase contrast microscope (x200, magnification). To observe the effects of ROS scavengers on the induction of SOX-2, cells were treated for 72 h with 10 mM N-acetyl-l-cysteine (NAC; Sigma-Aldrich, St. Louis, MO, USA).

Non-coated 6-well plates were coated with 1 ml of poly-HEMA (20 mg/ml; Sigma-Aldrich), dried overnight at room temperature, and washed twice with phosphate-buffered saline (PBS) before use. For the agar plates, 1% agar (Sigma-Aldrich) solution in PBS was added to 6-well plates and allowed to solidify for 1 h. The ultra-low attachment plates were purchased from Corning (Corning, NY, USA) and the polystyrene tissue culture plates were from SPL Life Sciences (Pocheon, Gyeonggi-do, korea).

Cell proliferation was assessed as a function of metabolic activity using an EZ-Cytox Cell Viability Assay kit (ItsBio, Seoul, korea). The assay is based on reduction of tetrazolium chloride to the water-soluble formazan by succinate-tetrazolium reductase, which forms part of the mitochondrial respiratory chain. After treatment with 20 µl/well, cells were incubated for 2 h at 37°C in a 5% CO₂ atmosphere. Absorbance was measured on a microplate reader (Bio-Rad 680; Bio-Rad, Hercules, CA, USA) at 450 nm. After subtraction of the background, the viability was determined as the ratio relative to the control and reported as the mean ± standard error (SE).

Total RNA was isolated using AccuZol (Bioneer, Daejeon, korea) and first-strand cDNA was synthesized by reverse transcription with 1 µg of total RNA using a ReverTra Ace qPCR kit (Toyobo, Osaka, Japan). Quantitative real-time PCR was performed to validate the expression level using SYBR^® Premix Ex Taq™ (Takara Bio, Otsu, Shiga, Japan) with specific primers (Table I) on an Applied Biosystems 7300 machine (Applied Biosystems, Carlsbad, CA, USA). The relative values for specific mRNA were calculated after normalization to the Ct value of β-actin in the same sample using the 2^−ΔΔCt method.

Spheres were trypsinized for 5 min at 37°C in a 5% CO₂ atmosphere, neutralized with complete medium containing FBS, and centrifuged at 3,000 × g for 5 min. Single cells were then fixed with 70% ethanol at −20°C overnight and incubated with a solution containing 5 µg/ml RNase A and 10 µg/ml propidium iodide (PI) (both from Sigma-Aldrich) for 30 min at room temperature. After washing twice with cold PBS, the cells were analyzed using a FACSCalibur (Becton-Dickinson, Bedford, MA, USA). Mean fluorescence intensities were obtained using CellQuest software (Becton-Dickinson).

Intracellular ROS levels in cells were measured using 5-(and-6)-car-boxy-2′,7′-dichlorodihydrofluorescein diacetate (H₂DCF-DA; Invitrogen, Carlsbad, CA, USA). Briefly, the cells were washed in PBS, trypsinized, and then neutralized with PBS-containing FBS. After two washes with PBS and further centrifugation, the cells were resuspended in PBS with 20µM DCF-DA and incubated for 20 min in the dark at 37°C. After the cells were resuspended in PBS, the fluorescence was measured using a FACSCalibur (excitation at 488 nm and emission at 515–545 nm). The mean fluorescence intensity for 10,000 cells/sample was determined using CellQuest software.

Cells were washed twice with PBS and lysed with RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 50 mM Tris-HCl and pH 8.0) with protease inhibitor (Roche Diagnostics, Penzberg, Germany) on ice for 30 min. Protein concentration was determined using a BCA Protein Assay kit (Pierce Biotechnology, Rockford, IL, USA). Equal amounts of protein were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto nitrocellulose membranes (Millipore, Billerica, MA, USA). The membranes were incubated for 1 h with 5% dried skim milk in TBST (20 mM Tris, 137 mM NaCl and 0.1% Tween 20) buffer and then incubated with primary antibodies to SOX-2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), SOD1 (Enzo Life Sciences, Farmingdale, NY, USA) and SOD2 (Abcam, Cambridge Science Park, Cambridge, UK). Anti-β-actin antibody (Sigma-Aldrich) was used as an internal control. After incubation with horseradish peroxidase-conjugated secondary IgG (Santa Cruz Biotechnology), the immunoreactive bands were visualized on an enhanced chemiluminescence substrate (Thermo Fisher Scientific, Waltham, MA, USA). For quantification of the relative protein levels, densitometric analysis was carried out using Image J software (National Institutes of Health, Bethesda, MD, USA). The expression ratio relative to the control was calculated after normalization with the intensity of β-actin from each group.

Student’s t-test was used for comparison of differences between two groups. Each experiment was repeated at least three times. In all analyses, P<0.05 was taken to indicate statistical significance.

A172 cells were cultured for 72 h in various culture environments differing in plate and media conditions, as described in the Materials and methods. The morphologies of cells from each group were examined under an inverted microscope. Fig. 1A shows that the A172 cells in Nm on NP (designated as controls) grew as a monolayer, while those in serum-free medium supplemented with growth factors (GBM, 10 ng/ml bFGF and EGF) on NP appeared loosely attached on the plates with some cells showing stellate projections. In contrast, cells cultured on the other plates (UP, HP and AP) in the Nm condition formed floating aggregates, while those cultured on UP and HP formed sphere-like colonies in GBM. To examine whether they could re-adhere to NP, cells or colonies from each group were dissociated using trypsin and seeded onto NP under the same media conditions. After 4 h, most cells cultured in Nm on NP, UP, HP and AP showed stable adherence to NP, whereas those cultured in GBM showed delayed adherence to NP irrespective of the previous plate conditions (Fig. 1B). After 24 h, most cells in Nm or GBM on the four surface types adhered to NP, but those cultured in GBM showed increased populations with elongated and fibroblast-like morphologies (Fig. 1C).

To examine cellular viability under the various conditions, equal numbers of cells from each condition were seeded on plates (1×10⁴/well) and incubated with tetrazolium chloride to monitor mitochondrial reductase activity. Fig. 2A shows that the relative viability of cells in GBM on NP (61±6.9%) was ~40% lower than that of the control. Moreover, cells in Nm and GBM on three types of plates (UP, HP and AP) had significantly lower activities than those in GBM on NP. Specifically, cellular activity on AP was the lowest among the conditions examined (9.3±5.32% in Nm and 9.39±3.88% in GBM). We also performed cell cycle analysis by flow cytometry after staining the cells with PI (Fig. 2B). Compared with control cells (2.05±0.03%), there was a ~10% increase in the subG1 proportion of cells under GMB culture conditions (GBM/NP, 11.4±3.2%) together with decreased G2/M phase. However, on the other plates, incubation in GBM did not increase the subG1 proportion (Nm/UP, 19.88±3.96%; GBM/UP, 14.64±8.12%; Nm/HP, 9.16±2.22%; GBM/HP, 8.82±3.33%; Nm/AP, 20.13±9.83%; and GBM/AP, 17.76±6.58%).

To determine whether combined culture conditions affected the self-renewal and EMT abilities of these cells, the expression levels of four stem cell markers and seven EMT-related markers were examined by real-time PCR using specific primers (Table I). Compared with the control, SOX-2 mRNA levels showed the greatest increase in GBM on both UP and HP (fold-change: 321 and 369, respectively). CD133 expression also increased in GBM on UP and HP but to a lesser extent, which was not statistically significant. Among the EMT markers, SNAIL expression decreased on three plate types (UP, HP and AP). The mRNA levels of the other EMT markers appeared to be dependent on the plate type rather than the medium type (Fig. 3).

The effects of different culture conditions on ROS levels were investigated because ROS may be an important factor in regulation of stem cells and the chemoresistance of cancer stem cells (32–34). Performance was assessed by the conversion of non-fluorescent H₂DCF to the fluorescent form (35). Fig. 4 shows that the level of ROS generation in the cells cultured on UP or HP in normal medium was about double that in those cultured on normal plates. Serum deprivation GBM on UP or HP further increased ROS generation by 8-fold. Interestingly, ROS levels from cells cultured in both Nm and GBM on AP were 6- and 8-fold higher, respectively, than the control. These data indicate that the effect of serum deprivation on ROS accumulation is greater than that of the non-adherent condition.

As ROS generation was increased by serum deprivation, which was also effective for SOX-2 induction, we examined whether alterations in SOD1 and SOD2 expression were responsible for ROS accumulation and SOX-2 induction under serum deprivation or non-adherent conditions. Western blotting indicated that the level of SOX-2 expression was higher in the GBM condition than in normal medium on all four plate types, which was similar to the mRNA profile (Fig. 5A). While SOD1 expression showed no prominent pattern with respect to serum deprivation or plate condition, the SOD2 expression profile exhibited an inverse relationship with that of SOX-2 (Fig. 5B and C). However, with the exception of values from UP in Nm or GBM, the correlations were not statistically significant. Dot plot analysis was performed using the mean values of the SOX-2 and SOD2 protein levels. Fig. 5D shows that the two protein values from each group could be classified into a significantly larger SOX-2-expressing group (gray-dotted circle; GBM on three plate types) and a SOD2-expressing group (black ellipse; Nm on UP, HP and AP). SOX-2 expression was shown to be dependent on the GBM condition, while SOD2 expression was dependent on the non-adherent conditions. Finally, cells were treated with NAC, a general ROS scavenger, to determine whether NAC could reverse the SOX-2 induction observed under our culture conditions. Western blotting was performed using antibodies to SOX-2, SOD1, and SOD2 after treatment with 10 mM NAC under the same conditions for 72 h. Fig. 6 shows that NAC treatment did not block the induction of SOX-2 or SOD2, suggesting that it was independent of ROS. Data from qRT-PCR analyses also revealed no significant recovery of SOX-2 with NAC treatment (data not shown).